Advanced Structural Chemistry: Tailoring Properties of Inorganic Materials and their Applications [1 ed.] 3527349006, 9783527349005

Citation preview

Advanced Structural Chemistry

Advanced Structural Chemistry Tailoring Properties of Inorganic Materials and their Applications

Edited by Rong Cao

Volume 1

Advanced Structural Chemistry Tailoring Properties of Inorganic Materials and their Applications

Edited by Rong Cao

Volume 2

Advanced Structural Chemistry Tailoring Properties of Inorganic Materials and their Applications

Edited by Rong Cao

Volume 3

Editor

Prof. Rong Cao Fujian Institute of Research on the Structure of Matter Chinese Academy of Sciences Yangqiao West Road 155# Gulou District 350002 Fuzhou China Cover Image: © koto_feja/Getty Images

All books published by Wiley-VCH are carefully produced. Nevertheless, authors, editors, and publisher do not warrant the information contained in these books, including this book, to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate. Library of Congress Card No.:

applied for British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library. Bibliographic information published by the Deutsche Nationalbibliothek

The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at . © 2021 WILEY-VCH GmbH, Boschstr. 12, 69469 Weinheim, Germany All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form – by photoprinting, microfilm, or any other means – nor transmitted or translated into a machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law. Print ISBN: 978-3-527-34900-5 ePDF ISBN: 978-3-527-83173-9 ePub ISBN: 978-3-527-83174-6 oBook ISBN: 978-3-527-83175-3 Typesetting SPi Global, Chennai, India Printing and Binding

Printed on acid-free paper 10 9 8 7 6 5 4 3 2 1

This book is dedicated to the 60th anniversary of Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, and its founder director, Professor Jiaxi Lu, on the memory of his 105th birthday. The authors would like to extend their appreciation to Professors Xin-Tao Wu and Mao-Chun Hong for their valuable advice.

vii

Contents Volume 1 1

Introduction 1 Jian Zhang, Guo-Cong Guo, Rong Cao, and Xin-Tao Wu

2

Coordination-Assembled Metal–Organic Macrocycles and Cages 9 Xiao-Zhen Li, Yu-Ling Liang, Li-Peng Zhou, Li-Xuan Cai, Qiang-Yu Zhu, Zhuo Wang, Xiao-Qing Guo, Dan-Ni Yan, Shao-Jun Hu, Shao-Chuan Li, Shi-Yu Wu, Shi-Long Han, Ran Chen, Pei-Ming Cheng, Kai Cheng, Xiao-Shan Feng, Tian-Pu Sheng, Can He, Feng-Rong Dai, and Qing-Fu Sun Introduction 9 Metallacycles 9 Dinuclear Metallamacrocycles 10 Triangles 10 Rectangle 12 Hexagons 14 Irregular Metallacycles 20 Multilayered Metallacycles 22 Polygon-Based Polymers 28 Responsive Dynamic Metallacycles 34 Metallacages 35 Helicates 35 Tetrahedron 38 Truncated Tetrahedron 46 Triangular Prism 47

2.1 2.2 2.2.1 2.2.2 2.2.3 2.2.4 2.2.5 2.2.6 2.2.7 2.2.8 2.3 2.3.1 2.3.2 2.3.3 2.3.4

viii

Contents

2.3.5 Cubes 50 2.3.6 Octahedron 54 2.3.7 Dodecahedron 61 2.3.8 Cuboctahedrons 65 2.3.9 Hexadecahedrons 65 2.3.10 Barrel-Shaped Cages 67 2.3.10.1 Calixarene Constructed Barrel-Shaped Cages 2.3.10.2 Dimetallic Clips-Constructed Barrel-Shaped Cages 68 2.3.11 Multiple Structural Cages 71 2.3.12 Other Cages 74 2.4 Conclusion 77 Acknowledgments 77 References 77 3

3.1 3.1.1 3.1.2 3.1.2.1 3.1.2.2 3.1.3 3.1.3.1 3.1.3.2 3.1.3.3 3.1.4 3.1.4.1 3.1.4.2 3.1.5 3.1.6 3.1.7 3.1.8

67

Structural Chemistry of Metal-Oxo Clusters 81 Xiaofeng Yi, Weihui Fang, Jinying Liu, Cheng Chen, Mingyan Wu, and Lei Zhang Oxo Clusters of Transition Metal 81 Introduction 81 General Synthetic Approaches and Experimental Methods 83 General Synthetic Approaches 83 Experimental Methods 83 Polyoxotitanates (POTis) 84 Diverse Structures of POTis 84 Tuneable Properties of POTis: Bandgap Engineering and Photo-Related Activities 89 Potential Application of POTis 92 Polyoxovanadates (POVs) 95 Diverse Structure of POVs 95 Tunable Properties and Potential Applications of POVs 101 Polyoxoniobates (PONbs) 102 Polyoxomolybdates (POMos) 104 Polyoxopalladates (POPs) 105 Polyoxotungstates (POTs) 107

Contents

3.1.8.1 3.1.8.2 3.1.9 3.2 3.2.1 3.2.2 3.2.2.1 3.2.2.2 3.2.2.3 3.2.2.4 3.2.3 3.2.3.1 3.2.3.2 3.2.3.3 3.2.4 3.2.4.1 3.2.4.2 3.3 3.3.1 3.3.2 3.3.2.1 3.3.2.2 3.3.2.3 3.3.2.4 3.3.2.5 3.3.3 3.3.4

Transition-Metals-Substituted-POTs (TMSPs) 107 Inorganic–Organic Hybrid TMSPs 107 Polyoxotantalates (POTas) 109 Oxo Clusters of Main Group Metal 111 Introduction 111 Synthesis of Borates 111 Inorganic Templated Borates 112 Organic-Templated Borates 113 TMC-Templated Borates 113 Templated Synthesis of Aluminoborates 115 Synthesis of Germinates and Borogermanates 116 Templated Synthesis of Germinates 116 Templated Synthesis of Germinates 117 Self-Polymerization and Induced Congregation of Lanthanide Germanate Lusters 118 Aluminum Oxo Clusters Hydrolysis and Condensation 119 Aluminum Oxo Clusters Isolated from Organic Solutions 119 Aluminum Oxo Clusters Via Aqueous Synthetic Routes 120 Oxo Clusters of Lanthanides 122 Introduction 122 High-Nuclearity Clusters of Lanthanides 123 High-Nuclearity Lanthanide Clusters Supported by O-Donor Ligands 123 High-Nuclearity Lanthanide Clusters Supported by N-Donor Ligands 127 High-Nuclearity Lanthanide Clusters Supported by Multiple N,O-Donor Ligands 129 High-Nuclearity Lanthanide Clusters Supported by Calix[n]arenes Ligands 136 High-Nuclearity Lanthanide Clusters Supported by Other Donor Ligands 139 Monometallic Lanthanide-Based Single-Molecule Magnets 140 Heterometallic 3d–4f Clusters 142

ix

x

Contents

3.4

4

4.1 4.2

4.3 4.3.1

4.3.2

4.3.3

4.4 4.5

5 5.1 5.2 5.3

Conclusion 149 Acknowledgments References 149

149

Synthetic Approaches, Structural Diversities, and Biological Aspects of Molybdenum (Tungsten)-Based Heterometallic Sulfide Clusters and Coordination Polymers 163 Du Shaowu and Wu Xintao Introduction 163 Synthesis of Mo–Fe–S Cuboidal Clusters for the Structural Modeling of the Iron–Molybdenum Cofactor (FeMoco) 167 Rationally Designed Synthesis of Mo(W)–Cu(Ag)–S Clusters 173 Unit Construction Method for the Synthesis of Simple Mo(W)–Cu(Ag)–S Clusters Starting from Thiomolybdates or Thiotungstates Building Units 173 Unit Construction Method for the Synthesis of Single Cubane- and Cage-like Mo(W)–Cu(Ag)–S Clusters Starting from Tri- and Dinuclear Thiomolybdates and Thiotungstates 175 Unit Construction Method for the Synthesis of Mo(W)–Cu(Ag)–S Clusters Having Multiple Cubane-like Structures 178 Rationally Designed Synthesis of W(Mo)–Ag–S Coordination Polymers 182 Conclusion 190 Acknowledgments 191 References 191 Group 11–15 Metal Chalcogenides 195 Jian-Rong Li, Mei-Ling Feng, Bing Hu, and Xiao-Ying Huang Introduction 195 Inorganic–Organic Hybrid Chalcogenides Based on II–VI Semiconductors 198 Discrete Chalcogenide Clusters Synthesized in Ionic Liquids 209

Contents

5.3.1 5.3.2 5.4 5.4.1 5.4.2 5.5 5.5.1 5.5.2 5.5.3 5.5.3.1 5.5.3.2 5.5.4 5.5.4.1 5.5.4.2 5.6 5.6.1 5.6.1.1 5.6.1.2 5.6.2 5.6.2.1 5.6.2.2 5.6.3 5.7

Discrete Chalcogenido Tn Clusters 209 Other Discrete Chalcogenide Clusters 216 Chalcogenidostannates 218 Chalcogenidostannates 219 Heterometallic Chalcogenidostannates Containing Ag+ 228 Chalcogenidoantimonates 234 Thioantimonates 234 Chalcogenidometalates Containing Group 12(II) Ions and Antimony(III) 236 Chalcogenidometalates Containing Group 13(III) Ions and Antimony(III) 245 Ga–Sb–S Compounds 246 In–Sb–Q (Q = S and Se) Compounds 248 Chalcogenidometalates Containing Group 14(IV) Ions and Antimony(III) 254 Ge–Sb–S Compounds 254 Sn–Sb–S Compounds 259 Selected Properties 260 Properties of Selected Inorganic–Organic Hybrid Metal Chalcogenides 260 Optical and Electric Properties 261 Thermal Expansion Behavior 263 Photocatalytic Property 264 Photocatalytic Hydrogen Production 265 Photodegradation of Organic Dye Molecules 267 Ion Exchange Property 269 Conclusion 275 Acknowledgments 276 References 277 Volume 2

6

The Structures of Metal–Organic Frameworks Yuehong Wen, Xintao Wu, and Qi-Long Zhu

7

Structural Design and Rational Synthesis Qipu Lin

391

283

xi

xii

Contents

8

Structural Topologies and Interpenetration in the Coordination Polymers 425 Fei Wang, Hai-Xia Zhang, and Jian Zhang

9

Inorganic Chalcogenides: From Zero-Dimensional Clusters to Three-Dimensional Frameworks 465 Hua Lin, Xin-Tao Wu, and Qi-Long Zhu

10

Relationship Between Structure and Electroluminescent, Photochromic, or Second-Order Nonlinear Optical Property 531 Ming-Sheng Wang, San-Gen Zhao, Qian Wang, and Zhong-Ning Chen

11

Relationship Between Structure and Ferroelectric Properties 601 Wuqian Guo, Zhihua Sun, Zhiyun Xu, Shiguo Han, and Junhua Luo Volume 3

12

The Relationship Between Structure and Electric Property 669 Guan-E Wang, Yonggang Zhen, Guo-Dong Wu, Huanli Dong, and Gang Xu

13

Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets 777 Zhu Zhuo, Guo-Ling Li, and You-Gui Huang

14

Relationship Between MOF Structures and Gas Absorption Properties 833 Qi Yin and Tian-Fu Liu

15

Relationship Between Structure and Separation Property 881 Zhanfeng Ju, El-Sayed M. El-Sayed, and Daqiang Yuan

Contents

16

Catalysis in Metal–Organic Frameworks: Relationship Between Activities and Structures 953 Yuan-Biao Huang, Teng Zhang, and Rong Cao Index 995

xiii

vii

Contents Volume 1 1

Introduction 1 Jian Zhang, Guo-Cong Guo, Rong Cao, and Xin-Tao Wu

2

Coordination-Assembled Metal–Organic Macrocycles and Cages 9 Xiao-Zhen Li, Yu-Ling Liang, Li-Peng Zhou, Li-Xuan Cai, Qiang-Yu Zhu, Zhuo Wang, Xiao-Qing Guo, Dan-Ni Yan, Shao-Jun Hu, Shao-Chuan Li, Shi-Yu Wu, Shi-Long Han, Ran Chen, Pei-Ming Cheng, Kai Cheng, Xiao-Shan Feng, Tian-Pu Sheng, Can He, Feng-Rong Dai, and Qing-Fu Sun

3

Structural Chemistry of Metal-Oxo Clusters 81 Xiaofeng Yi, Weihui Fang, Jinying Liu, Cheng Chen, Mingyan Wu, and Lei Zhang

4

Synthetic Approaches, Structural Diversities, and Biological Aspects of Molybdenum (Tungsten)-Based Heterometallic Sulfide Clusters and Coordination Polymers 163 Du Shaowu and Wu Xintao

5

Group 11–15 Metal Chalcogenides 195 Jian-Rong Li, Mei-Ling Feng, Bing Hu, and Xiao-Ying Huang Volume 2

6 6.1 6.2 6.2.1 6.2.2 6.2.3

The Structures of Metal–Organic Frameworks Yuehong Wen, Xintao Wu, and Qi-Long Zhu Introduction 283 One-dimensional MOFs 284 Linear Chains 284 Zigzag Chains 287 Helical Chains 291

283

viii

Contents

6.2.4 6.2.5 6.3 6.3.1 6.3.2 6.3.3 6.3.4 6.3.5 6.4 6.4.1 6.4.1.1 6.4.1.2 6.4.1.3 6.4.1.4 6.4.2 6.4.2.1 6.4.2.2 6.4.2.3 6.4.3 6.4.3.1 6.4.3.2 6.4.3.3 6.4.3.4 6.4.4 6.4.4.1 6.4.4.2 6.4.4.3 6.4.4.4 6.4.5 6.4.5.1 6.4.5.2 6.4.6 6.5

Double Chains 300 Ladder-like Chains 304 Two-dimensional MOFs 309 Triangular-Grid Networks 309 Square-Grid Networks 311 Honeycomb, Brick-Wall, and Herringbone Networks 319 Bilayer Networks 325 Other 2D MOFs 333 Three-dimensional MOFs 342 Carboxylate Linkers 343 Ditopic Carboxylate Ligands 343 Tritopic Carboxylate Linkers 347 Tetratopic Carboxylate Linkers 350 Hexatopic or Octatopic Carboxylate Linkers 351 N-heterocyclic Linkers 353 Imidazolates-based MOFs 353 Pyrazolate/Triazolate/Tetrazolate-based MOFs 355 Pyridine and Other N-heterocyclic Based MOFs 357 Metallo-linkers 358 Metallo-Linkers Based on Porphyrins 359 Metallo-Linkers Based on Salen 360 Metallo-Linkers Based on BINOL 362 Other Metallo-linkers 363 Mixed N-/O-donors 365 Ligands with Mixed N- and O-Donor Atoms 365 Mixed Carboxylate Linkers 367 Mixed Metallo-ligand and Organic Ligand 367 Mixed N-donated Linkers and O-donated Linkers 369 Phosphonate or Sulfonate Linkers 373 Phosphonate Linkers 373 Sulfonate Linkers 375 POM-Based MOFs 377 Conclusion 380 Acknowledgments 380 References 380

7

Structural Design and Rational Synthesis 391 Qipu Lin Introduction 391 Polyoxometalate Clusters 392 Inclusion of Small Anions/Cations 392 Inclusion of Inorganic Anions/Cations 392 Inclusion of Organic Anions/Cations 393 Multivariate Metal Mixing 393 Nitrogen Alkylation 395 Coordination with Organic Ligands 395 POM-Based Cages 395

7.1 7.2 7.2.1 7.2.1.1 7.2.1.2 7.2.2 7.2.3 7.2.4 7.2.4.1

Contents

7.2.4.2 7.2.5 7.2.5.1 7.2.5.2 7.2.5.3 7.2.5.4 7.2.5.5 7.2.5.6 7.3 7.3.1 7.3.2 7.3.3 7.3.4 7.3.5 7.3.6 7.3.7 7.3.8 7.3.9 7.3.10 7.3.10.1 7.3.10.2 7.3.10.3 7.3.10.4 7.3.10.5 7.4 7.4.1 7.4.1.1 7.4.1.2 7.4.1.3 7.4.1.4 7.4.1.5 7.4.2 7.4.2.1 7.4.2.2 7.4.2.3 7.4.3 7.4.3.1 7.4.3.2 7.4.3.3 7.4.4 7.4.4.1 7.4.4.2 7.4.4.3 7.4.4.4 7.4.4.5 7.5 7.5.1

POM-Based Nets 396 Others 396 Using Precursors 396 Templating by Clusters 397 Charge-Balancing by Complexes 397 Single-Crystal to Single-Crystal Transformations 397 Reaction pH Control 397 Networked Reactor System 397 Chalcogenidometalate Superlattices 398 Mixing Metals for Valance–Balance 398 Using Surface-Capping Ligands 399 Termination with Superbases 399 Using Structure-Directing Agents (SDAs) 400 Using Metal Complexes as Templates 400 Mimicking Zeolites 401 Amine-Based Solvothermal Reaction 401 Ionothermal Synthesis 403 Using Surfactants 403 Others 403 Insertion of O2− Anions 403 Linkage by Pyridines/Imidazoles 404 Reaction Parameters 404 Postsynthetic Insertion of Heterometals 404 Phase Transformation 404 Polygonal/Polyhedral Complexes 405 Design Based on Platonic Polyhedra 405 Tetrahedra 406 Cubic Systems 406 Octahedra 406 Dodecahedra 407 Icosahedra 407 Design Based on Archimedean Polyhedra 407 Cuboctahedra 407 Truncated Tetrahedra 408 Rhombicuboctahedra 408 Design Based on Other Shapes 408 Prism-Like Systems 408 Goldberg Polyhedra 409 Stellated Polyhedra 410 Linkage Modes 410 Linkage with Pd/Pt–Nitrogen 410 Linkage with Ga/Fe–Catechol 410 Linkage with Ln–Tridentate Ligand 410 Linkage with Metal–Carboxylate 410 Nodes with Transition Metal–Calixarene 411 Metal–Organic Frameworks 411 Reticulating Chemistry (Scale Chemistry) 412

ix

x

Contents

7.5.2 7.5.2.1 7.5.2.2 7.5.2.3 7.5.2.4 7.5.2.5 7.5.2.6 7.5.2.7 7.5.2.8 7.5.3 7.5.3.1 7.5.3.2 7.5.3.3 7.5.3.4 7.5.3.5 7.5.3.6 7.5.4 7.5.4.1 7.5.4.2 7.5.4.3 7.5.5 7.5.5.1 7.5.5.2 7.5.5.3 7.5.5.4 7.5.5.5 7.5.6 7.5.6.1 7.5.6.2 7.5.6.3 7.6

Inorganic Nodes (SBUs) 414 Mononuclear Units 414 Dinuclear Units 414 Trinuclear Units 414 Tetranuclear Units 414 Hexanuclear Units 415 Octanuclear Units 415 Mixing SBUs 415 Rod-Shaped Chains 416 Organic Linkers with Carboxylates 416 Dicarboxylates 416 Tricarboxylates 417 Tetracarboxylates 417 Hexacarboxylates 417 Octacarboxylates 417 Mixing Ligands 417 Organic Linkers with Other Functional Groups 418 Other Oxygen-Containing Ligands 418 Nitrogen–Heterocyclic Linkers 418 Mixing Hetero-Linkers 419 Synthetic Strategies 419 Solvothermal Reactions 419 Template-Directed Synthesis 419 Seed-Mediated Approach 419 Microwave-Assisted Synthesis 420 High-Throughput Methods 420 Others 420 Tailor-Made Approach 420 Design Based on MOPs 420 Merged Nets Approach 421 Conclusion 421 Acknowledgments 421 References 421

8

Structural Topologies and Interpenetration in the Coordination Polymers 425 Fei Wang, Hai-Xia Zhang, and Jian Zhang Introduction 425 Supramolecular Assembly and Reticular Synthesis 425 Geometric Basis of Crystallization Chemistry 426 Dual Nets 430 Topological Network of Metal–Organic Framework 430 Two-Dimensional Network 431 3-Connected Topology 431 4-Connected Topology 432 (3,4)-Connected Topology 434 6-Connected Topology 435

8.1 8.2 8.3 8.4 8.5 8.5.1 8.5.1.1 8.5.1.2 8.5.1.3 8.5.1.4

Contents

8.5.1.5 8.5.2 8.5.2.1 8.5.2.2 8.5.2.3 8.5.2.4 8.5.3 8.5.3.1 8.5.3.2 8.5.3.3 8.5.3.4 8.5.3.5 8.5.3.6 8.5.3.7 8.5.3.8 8.5.4 8.5.5 8.5.6 8.5.7 8.5.8 8.5.9 8.5.10 8.6 8.7

(3,6)-Connected Topology 436 3-Connected Three-Dimensional Topological Network 436 srs(SrSi2 ) Topological Networks 437 ths(ThSi2 ) Topological Networks 438 utp((10,3)-d) Topological Networks 438 Other 3-Connected Topological Networks 438 4-Connected Three-Dimensional Topological Network 439 Dia (= Diamond) Topological Network 440 cds (= CdSO4 ) Topological Network 440 NbO Topological Network 441 Quartz Topological Network 441 Moganite and PtS Topological Network 442 86 (= (8,4)) Topological Network 443 Zeolite-Type Topology Networks 443 Other 4-Connected Topological Networks 447 (3,4)-Connected Three-Dimensional Topological Network 448 5-Connected Three-Dimensional Networks 449 6-Connected Three-Dimensional Networks 449 8-Connected Three-Dimensional Topological Networks 451 (3,6)-Connected Three-Dimensional Topological Networks 453 (4,8)-Connected Three-Dimensional Topological Networks 455 12-Connected Three-Dimensional Topological Networks 456 Rod-Packing Topology Networks 457 Conclusion 459 References 460

9

Inorganic Chalcogenides: From Zero-Dimensional Clusters to Three-Dimensional Frameworks 465 Hua Lin, Xin-Tao Wu, and Qi-Long Zhu Introduction 465 Inorganic Chalcogenides 466 Zero-Dimensional (0D) Cluster Chalcogenides 466 Ba12 In4 S19 466 Ba23 Ga8 Sb2 S38 469 Ba3 (BS3 )1.5 (MS3 )0.5 (M = Sb, Bi) and Ba3 (BQ3 )(SbQ3 ) (Q = S, Se) 470 (Cs6 Cl)2 Cs5 [Ga15 Ge9 Se48 ] 470 BaHgSe2 471 Ba12 Sn4 S23 and Ba7 Sn3 S13 472 Ba8 Sn4 S15 473 Ba4 M3 Q9 Cl2 (M = Si, Ge; Q = S, Se) 474 One-Dimensional (1D) Chain Chalcogenides 475 Ba4 In2 S8 and Ba4 Ga2 S8 475 Ln4 GaSbS9 (Ln = Pr, Nd, Sm, Gd, Tb, Dy, Ho) 475 La4 InSbS9 477 Ba3 La4 Ga2 Sb2 S15 and BaLa3 GaSb2 S10 478 Ba8 Zn4 Ga2 S15 479 A4 Ge4 Se12 (A = Rb, Cs) 479

9.1 9.2 9.2.1 9.2.1.1 9.2.1.2 9.2.1.3 9.2.1.4 9.2.1.5 9.2.1.6 9.2.1.7 9.2.1.8 9.2.2 9.2.2.1 9.2.2.2 9.2.2.3 9.2.2.4 9.2.2.5 9.2.2.6

xi

xii

Contents

9.2.2.7 9.2.2.8 9.2.2.9 9.2.3 9.2.3.1 9.2.3.2 9.2.3.3 9.2.3.4 9.2.3.5 9.2.3.6 9.2.3.7 9.2.3.8 9.2.3.9 9.2.3.10 9.2.3.11 9.2.4 9.2.4.1 9.2.4.2 9.2.4.3 9.2.4.4 9.2.4.5 9.2.4.6 9.2.4.7 9.2.4.8 9.2.4.9 9.2.4.10 9.2.4.11 9.2.4.12 9.2.4.13 9.2.4.14 9.2.4.15 9.2.4.16 9.2.4.17 9.2.4.18 9.2.4.19 9.2.4.20 9.2.4.21 9.2.4.22 9.2.4.23 9.2.4.24 9.2.4.25 9.2.4.26 9.2.4.27 9.2.4.28 9.2.4.29 9.2.4.30

BaGeOSe2 480 Ba5 In4 Te4 S7 482 Ba8 Ga2 Sn7 Se18 and Ba10 Ga2 Sn9 Se22 483 Two-Dimensional (2D) Layer Chalcogenides 484 La4 FeSb2 Q10 (Q = S, Se) 484 Ln2 Mn3 Sb4 S12 (Ln = Pr, Nd, Sm, Gd) 485 CsLnCdTe3 (Ln = La, Pr, Nd, Sm, Gd–Tm, and Lu) 486 (Cs6 Cl)6 Cs3 [Ga53 Se96 ] 486 Cs2 [Mn2 Ga3 S7 Cl] 487 Ba2 Cr4 GeSe10 489 A2 Ge4 Se10 (A = Rb, Cs) 490 Na6 Zn3 III2 Q9 (III = Ga, In; Q = S, Se) 490 RECuTe2 (RE = Tb–Tm) 491 CsMnInTe3 491 Ba4 F4 XGa2 S6 (X = Cr, Mn, Fe) and Ba4 F4 MnIn2 S6 493 Three-Dimensional (3D) Framework Chalcogenides 494 PbMnIn2 S5 494 Ax RE2 Cu6 − x Te6 (A = K–Cs; RE = La–Nd) 495 CsRE2 Ag3 Te5 (RE = Pr, Nd, Sm, Gd–Er) and RbRE2 Ag3 Te5 (RE = Sm, Gd–Dy) 496 Cs[RE9 Mn4 Se18 ] (RE = Ho–Lu) 497 Cs[RE9 Cd4 Se18 ] (RE = Tb–Tm) 498 Cs3 Cu20 Te13 499 A–II4 –III5 –Q12 Type 500 A–III–Sn2 –Se6 Type 501 Ba3 AGa5 Se10 Cl2 and Its Derivatives 502 Ba6 Li2 CdSn4 S16 504 La2 CuSbS5 504 Ba6 Zn7 Ga2 S16 506 Na2 In4 SSe6 , NaGaIn2 Se5 , and NaIn3 Se5 507 PbGa2 MSe6 (M = Si, Ge) 508 SnGa4 Q7 (Q = S, Se) 509 Na2 ZnGe2 S6 510 Ba–Na2 –IV–Q4 Type (IV = Ge, Sn; Q = S, Se) 511 Ba–Li2 –IV–Q4 Type (IV = Ge, Sn; Q = S, Se) 512 CsAg5 Te3 513 Ba5 Cu8 In2 S12 513 Cs[Lu7 Q11 ] and (ClCs6 )[RE21 Q34 ] (RE = Dy, Ho; Q = S, Se, Te) 515 Yb6 Ga4 S15 and Lu5 GaS9 516 BaAg2 GeS4 and BaAg2 SnS4 517 Sr5 ZnGa6 S15 519 Cd4 GeS6 520 Cs2 [RE8 InS14 ] (RE = Ho–Lu) 521 Na7 IISb5 S12 (II = Zn, Cd, Hg) 522 Cs2 Ge3 M6 Te14 (M = Ga, In) 523 Na2 Ga2 MQ6 (M = Ge, Sn; Q = S, Se) 524 Sn2 Ga2 S5 525

Contents

9.2.5 9.2.5.1 9.2.5.2 9.3

10

10.1 10.2 10.2.1 10.2.1.1 10.2.1.2 10.2.2 10.2.2.1 10.2.2.2 10.2.3 10.3 10.3.1 10.3.1.1 10.3.1.2 10.3.2 10.3.2.1 10.3.2.2 10.3.2.3 10.3.2.4 10.3.2.5 10.3.3 10.4 10.4.1 10.4.2 10.4.3 10.4.4 10.4.5 10.4.6 10.4.7 10.4.8 10.4.9

Mixed-Dimensional (MD) framework Chalcogenides 526 Ln3 M0.5 (Ge0.5 /M0.5 )S7 (Ln = La, Sm; M = Ga, In) and Ln3 In0.33 GeS7 (Ln = La, Sm, Gd) 526 CsCu5 S3 527 Conclusion 528 Acknowledgments 528 References 528 Relationship Between Structure and Electroluminescent, Photochromic, or Second-Order Nonlinear Optical Property 531 Ming-Sheng Wang, San-Gen Zhao, Qian Wang, and Zhong-Ning Chen Introduction 531 Organic Electroluminescent Materials–d8 /d10 Heteronuclear Metal Complexes 532 The Photoluminescence of d8 /d10 Heteronuclear Metal Complexes 533 d8 –d10 Heteronuclear Alkynyl Complexes 534 d10 –d10 Heteronuclear Alkynyl Complexes 537 The Electroluminescence of d8 /d10 Heteronuclear Metal Complexes 539 The Electroluminescence of d8 –d10 Heteronuclear Complexes 539 The Electroluminescence of d10 –d10 Heteronuclear Complexes 548 Prospective 549 Photochromic Materials–Viologen Compounds and their Analogs 551 Viologen Compounds and their Analogs 551 Metalloviologen Compounds 553 N-Substituted Monocyclic Aromatic Ion-Templated Compounds 558 Optical Applications of Viologens and their Analogs 560 Opto-Optical Switching 560 Radiation Detection 565 Photocatalysis 568 Electrical Applications of Viologens and their Analogs 569 Molecular Recognition Applications of Viologens and their Analogs 572 Prospective 574 NLO Materials 575 KBBF Family 575 KBBF Derivatives 577 SBBO Family 583 AB4 O6 F Family 584 Apatite-Like Borates 586 ABCO3 F Family 588 A3 VO(O2 )2 CO3 Family 590 AXII 4 XIII 5 Se12 Family 593 Conclusions and Prospects 593 References 595

xiii

xiv

Contents

11

11.1 11.1.1 11.1.1.1 11.1.1.2 11.1.1.3 11.1.1.4 11.1.2 11.1.3 11.1.4 11.1.4.1 11.1.4.2 11.1.4.3 11.1.4.4 11.1.4.5 11.2 11.2.1 11.2.2 11.2.3 11.2.4 11.3

Relationship Between Structure and Ferroelectric Properties 601 Wuqian Guo, Zhihua Sun, Zhiyun Xu, Shiguo Han, and Junhua Luo Concepts and Fundamentals 601 Structural Phase Transitions 601 Introduction 601 Thermodynamic Theory 602 Microscopic Theory 609 Research Methods 611 Symmetry Breaking of Ferroelectrics 614 Classification of Ferroelectrics 614 Characterization of Ferroelectrics 616 Dielectric and Dielectric Switch 616 NLO and NLO Switch 617 Pyroelectric Properties 619 Polarization Switching – Ferroelectricity 621 Domain Motions 624 Recent Advance of Molecular Ferroelectrics 626 Organic Ferroelectrics 626 Binary Molecular Ferroelectrics 636 Organic–Inorganic Hybrid Ferroelectrics 638 Metal–Organic Framework Ferroelectrics 647 Conclusion and Perspective 662 References 663 Volume 3

12

The Relationship Between Structure and Electric Property 669 Guan-E Wang, Yonggang Zhen, Guo-Dong Wu, Huanli Dong, and Gang Xu

13

Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets 777 Zhu Zhuo, Guo-Ling Li, and You-Gui Huang

14

Relationship Between MOF Structures and Gas Absorption Properties 833 Qi Yin and Tian-Fu Liu

15

Relationship Between Structure and Separation Property 881 Zhanfeng Ju, El-Sayed M. El-Sayed, and Daqiang Yuan

16

Catalysis in Metal–Organic Frameworks: Relationship Between Activities and Structures 953 Yuan-Biao Huang, Teng Zhang, and Rong Cao Index 995

vii

Contents Volume 1 1

Introduction 1 Jian Zhang, Guo-Cong Guo, Rong Cao, and Xin-Tao Wu

2

Coordination-Assembled Metal–Organic Macrocycles and Cages 9 Xiao-Zhen Li, Yu-Ling Liang, Li-Peng Zhou, Li-Xuan Cai, Qiang-Yu Zhu, Zhuo Wang, Xiao-Qing Guo, Dan-Ni Yan, Shao-Jun Hu, Shao-Chuan Li, Shi-Yu Wu, Shi-Long Han, Ran Chen, Pei-Ming Cheng, Kai Cheng, Xiao-Shan Feng, Tian-Pu Sheng, Can He, Feng-Rong Dai, and Qing-Fu Sun

3

Structural Chemistry of Metal-Oxo Clusters 81 Xiaofeng Yi, Weihui Fang, Jinying Liu, Cheng Chen, Mingyan Wu, and Lei Zhang

4

Synthetic Approaches, Structural Diversities, and Biological Aspects of Molybdenum (Tungsten)-Based Heterometallic Sulfide Clusters and Coordination Polymers 163 Du Shaowu and Wu Xintao

5

Group 11–15 Metal Chalcogenides 195 Jian-Rong Li, Mei-Ling Feng, Bing Hu, and Xiao-Ying Huang Volume 2

6

The Structures of Metal–Organic Frameworks Yuehong Wen, Xintao Wu, and Qi-Long Zhu

7

Structural Design and Rational Synthesis 391 Qipu Lin

283

viii

Contents

8

Structural Topologies and Interpenetration in the Coordination Polymers 425 Fei Wang, Hai-Xia Zhang, and Jian Zhang

9

Inorganic Chalcogenides: From Zero-Dimensional Clusters to Three-Dimensional Frameworks 465 Hua Lin, Xin-Tao Wu, and Qi-Long Zhu

10

Relationship Between Structure and Electroluminescent, Photochromic, or Second-Order Nonlinear Optical Property 531 Ming-Sheng Wang, San-Gen Zhao, Qian Wang, and Zhong-Ning Chen

11

Relationship Between Structure and Ferroelectric Properties 601 Wuqian Guo, Zhihua Sun, Zhiyun Xu, Shiguo Han, and Junhua Luo Volume 3

12

12.1 12.2 12.2.1 12.2.1.1 12.2.1.2 12.2.2 12.2.2.1 12.2.2.2 12.2.2.3 12.2.2.4 12.2.3 12.2.3.1 12.2.3.2 12.2.4 12.2.4.1 12.2.4.2 12.2.4.3 12.2.4.4 12.2.5 12.2.5.1 12.2.5.2 12.3

The Relationship Between Structure and Electric Property 669 Guan-E Wang, Yonggang Zhen, Guo-Dong Wu, Huanli Dong, and Gang Xu Introduction 669 Structure and Electrical Properties of Inorganic Conductive Materials 670 Elemental Conducting Materials 670 General Introduction of Elemental Conducting Materials 670 Structure and the Property of Elemental Conducting Materials 672 Transparent Conducting Oxides (TCOs) 673 Materials Designed for n-Type TCOs 675 Typical n-Type TCOs and Their Electronic Structure 676 Materials Designed for p-Type TCOs 678 Typical p-Type TCOs and Their Electronic Structure 679 Nitride Conducting Materials 683 General Introduction of Nitride Conducting Materials 683 Structure and the Property in Typical Nitride Conducting Materials 686 Carbide Conducting Materials 691 The General Structure and the Property of SiC 691 The Doping of Silicon Carbides and Its Electronic Property 692 Materials Designed for New SiC 694 Materials Designed for Transition-Metal Carbides 695 Sulfide Conductive Materials 697 General Introduction of Sulfide Conductive Materials 697 The Structure and the Property of Typical Sulfide Conductive Materials 699 Structure and Electrical Properties of Organic Conductive Materials 704

Contents

12.3.1 12.3.1.1 12.3.1.2 12.3.2 12.3.2.1 12.3.2.2 12.3.3 12.3.3.1 12.3.3.2 12.3.3.3 12.3.4 12.3.4.1 12.3.4.2 12.4 12.4.1 12.4.1.1 12.4.1.2 12.4.1.3 12.4.2 12.4.2.1 12.4.2.2 12.4.2.3

13

13.1 13.2 13.2.1 13.2.1.1 13.2.1.2 13.2.2 13.2.2.1 13.2.2.2 13.2.2.3 13.3 13.3.1 13.3.2 13.3.3

Growth Methods and Packing Arrangements of Organic Single Crystals 705 Organic Small Molecule Crystals 705 Packing Arrangements in Organic Crystals 707 Structure–Property Relationship in Organic Conductor Crystals 707 Doped Systems 708 Single-Component Systems 710 Structure–Property Relationship in Organic Semiconductor Crystals 710 Linear-Shaped Molecules 711 Cylinder-Like or Disk-Like Molecules 713 Bowl-Shaped Molecules 715 Structure–Property Relationship Beyond Single Component 716 Organic Cocrystals 716 Solid Solution 722 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials 725 Structure and Electrical Properties of MOFs 725 Three-Dimensional (3D) Metal–Organic Frameworks Through Different Linkers 726 Two-Dimensional (2D) Metal–Organic Frameworks Through Different Linkers 737 One-Dimensional (1D) Metal–Organic Frameworks Through Different Linkers 747 Structure and Electrical Properties of Organic–Inorganic Hybrids 752 Three-Dimensional (3D) Organic–Inorganic Hybrids 754 Two-Dimensional (2D) Organic–Inorganic Hybrids 759 One-Dimensional (1D) Organic–Inorganic Hybrids 768 References 770 Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets 777 Zhu Zhuo, Guo-Ling Li, and You-Gui Huang Introduction 777 Magneto-Structural Correlations in Spin-Crossover Compounds 779 Influence of Structure on the Occurrence of Spin Crossover 781 Molecular Structure 781 Crystal Packing and Intermolecular Steric Contacts 786 Influence of Structure on the Cooperativity of Spin Crossover 788 Influence of Molecular Structure on the Cooperative Behavior 789 Influence of Crystal Packing on the Cooperative Behavior 790 Cooperative Spin-Crossover Coordination Polymers 790 Magneto-Structural Correlations of Low-Dimensional Magnets 791 Polynuclear Single-Molecule Magnets 792 Lanthanide Single-Ion Magnets 794 3d-Block Single-Ion Magnets 801

ix

x

Contents

13.3.4 13.4 13.4.1 13.4.1.1 13.4.1.2 13.4.1.3 13.4.1.4 13.4.2 13.4.3 13.5 13.5.1 13.5.2 13.5.3 13.5.4 13.5.4.1 13.5.4.2 13.5.4.3 13.5.4.4

14

14.1 14.2 14.2.1 14.2.2 14.2.3 14.2.4 14.3 14.3.1 14.3.1.1 14.3.1.2 14.3.2 14.3.2.1 14.3.2.2 14.3.3 14.3.4 14.4 14.4.1 14.4.2 14.4.3

Single-Chain Magnets 803 Influence of Ligand Bridging Modes on the Mediated Magnetic Coupling 804 Azido-Mediated Systems 805 Ferromagnets 806 Ferrimagnets 809 Antiferromagnets 810 Other Magnetic Behaviors 810 Monocarboxylate-Mediated Systems 811 Oxalate-Mediated Systems 812 Geometrically Frustrated Molecular Magnetic Materials 814 Geometric Magnetic Frustration 814 Evaluation of Geometric Frustration 815 Ground States of Frustrated Magnets 815 Molecular Materials Showing Spin Frustration 816 Isolated Molecules 816 Δ-Chain Lattice 819 2D Lattice 821 3D Lattice 822 Acknowledgments 823 References 823 Relationship Between MOF Structures and Gas Absorption Properties 833 Qi Yin and Tian-Fu Liu Introduction 833 Basic Conceptions About Gas Absorption 833 Physical Adsorption and Chemical Adsorption 833 Adsorption Curves 834 Langmuir Monolayer Adsorption Isotherms 835 BET Multilayer Adsorption Isotherms 836 The Correlation Between Physical Adsorption Isotherms and MOFs Structures 837 Effect of Pore Size on N2 Adsorption Isotherms 838 The Microporous MOFs Sorption Isotherms 838 The Mesoporous MOFs Sorption Isotherms 838 Effect of Pore Shape on Sorption Isotherms 840 The Gas Sorption Isotherms for MOFs with Channels 840 The Gas Sorption Isotherms of MOFs with Cage Structures 841 Effect of Activation on Sorption Isotherms 841 Activating Solvent Effects on Adsorption Isotherm 842 Effect of Material Defects on Gas Adsorption 842 Defect Definition 843 Types of Defect 843 Examples of the Effect of Defects on Adsorption 844

Contents

14.5 14.5.1 14.5.2 14.5.2.1 14.5.2.2 14.5.2.3 14.5.2.4 14.6 14.6.1 14.6.2 14.6.2.1 14.6.2.2 14.6.2.3 14.6.3 14.6.3.1 14.6.3.2 14.6.4 14.6.4.1 14.6.4.2 14.6.5

15 15.1 15.2 15.2.1 15.2.2 15.2.3 15.2.4 15.2.5 15.3 15.3.1 15.3.2 15.3.3 15.3.4 15.4 15.4.1 15.4.2 15.4.3 15.4.4 15.4.5 15.4.6

Representative MOFs with Ultrahigh Surface Area and the Factors Effecting on Surface Area 852 Representative MOFs with Ultrahigh Surface Area 853 The Factors Effecting on MOFs Surface Areas 855 Influence of Ligand Length 855 Influence of Interpenetration 859 Influence of Metal Moieties 861 Influence of Functionality 862 Adsorption Enthalpy of MOF Materials 863 Adsorption Enthalpy 863 Determination of Adsorption Enthalpy 863 Method 1 863 Method 2 864 Method 3 864 Definition of Adsorption Amount and Usable Adsorption Amount 864 The Definition of Usable Adsorption 865 Effect of Adsorption Enthalpy on Usable Adsorption 866 Factors Affecting MOFs Gas Adsorption 867 MOFs Containing Open Metal Sites 868 MOFs with Functional Organic Ligand 868 Effect of the Guest Molecules into the MOFs on Gas Adsorption 871 Acknowledgments 871 References 872 Relationship Between Structure and Separation Property 881 Zhanfeng Ju, El-Sayed M. El-Sayed, and Daqiang Yuan Introduction 881 CO2 Capture and Separation 882 Utilizing Open Metal Sites 883 Introducing Polar Functional Groups 885 Lewis Base Incorporation in Metal Sites 889 Functionalization of Ligands 893 Pore Size and Function Control 895 Separation of Hydrocarbons 900 Separation of C2 Hydrocarbons 901 Separation of C3 Hydrocarbons 913 Separation of Long-Chain Hydrocarbon 917 Separation of Alkane with Different Carbon Atoms 920 Separation of Noble Gases 922 HKUST-1 923 The MOF-74 (M-DOBDC) 923 SBMOF-1 924 CROFOUR-1-Ni and CROFOUR-2-Ni 925 [Co3 (C4 O4 )2 (OH)2 ]⋅3H2 O 927 SCU-11 929

xi

xii

Contents

15.4.7 15.5 15.5.1 15.5.2 15.5.3 15.6 15.6.1 15.6.2 15.6.3

FMOF-Cu 929 Separation of Hydrogen Isotopes 930 Kinetic Quantum Sieving 931 Chemical Affinity Quantum Sieving 934 Special Quantum Sieving 937 Enantioselective Separation 939 Chiral MOF as Absorbent in SPE and Solutions 940 Chiral MOFs as Stationary Phase in HPLC 943 Chiral MOFs as Stationary Phase in GC 945 References 949

16

Catalysis in Metal–Organic Frameworks: Relationship Between Activities and Structures 953 Yuan-Biao Huang, Teng Zhang, and Rong Cao Introduction 953 Catalysis by Metal Nodes 954 Lewis Acid Catalysis 954 Oxidation Reaction 956 Suzuki–Miyaura Coupling Reaction 958 Other Reactions 959 Catalysis by Functionalized Linkers 959 Catalysis by Species in Pores 963 MNPs Supported in MOFs for Catalysis 963 Polyoxometalates Encapsulated in MOFs for Catalysis 967 Metal Complexes Trapped in MOFs for Catalysis 968 Enzyme Trapped in MOFs for Catalysis 969 Asymmetric Catalysis in MOFs 970 Photocatalysis in MOFs 974 Photo-Degradation of Pollutants 974 Organic Photocatalysis 977 Photocatalytic Water Splitting and Artificial Photosynthesis 980 Multi-Component Catalysis Using MOFs 983 CUMs and Functional Organic Linkers in the MOF Backbone 984 Mixed Metal Centers as Active Sites for Synergistic Catalysis 984 Mixed Linkers as Bifunctional Active Sites for Tandem Reactions 984 Metal Nodes and Linkers as Bifunctional Active Sites for Synergistic Catalysis 986 Active Guest sites and Active Sites in the MOF Backbone 988 MNP@MOF Composites for Synergistic Catalysis and Tandem Reactions 988 POM@MOF Composites for Synergistic Catalysis and Tandem Reactions 989 References 991

16.1 16.2 16.2.1 16.2.2 16.2.3 16.2.4 16.3 16.4 16.4.1 16.4.2 16.4.3 16.4.4 16.5 16.6 16.6.1 16.6.2 16.6.3 16.7 16.7.1 16.7.1.1 16.7.1.2 16.7.1.3 16.7.2 16.7.2.1 16.7.2.2

Index 995

1

1 Introduction Jian Zhang, Guo-Cong Guo, Rong Cao, and Xin-Tao Wu Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, State Key Laboratory of Structural Chemistry, 155 Yangqiao Road West, Fujian, Fuzhou 350002, P.R. China

The syntheses of new substances with specific functions are fundamental to the modern civilization, providing an inherent impetus for social, economic, scientific, and technological advancements. Worldwide governments and scientists are devoted to the discovery of new substances with specific functions and structures. The creation of new functional substances, which is the strategic highland for an innovative country, represents the core competency of this country. The creation of new substances belongs to major basic research. So far, there are still many important scientific issues unsolved at this stage, and we are far away from the key goal, that is, to develop the substances with desired properties. Structural chemistry fabricates this aim and leads the creation of new substances. Structural chemistry is a discipline focusing on the bonding of atoms, molecules, and crystal packing as well as the correlations between the structure and the macroscopic chemical and physical properties. Therefore, structural chemistry serves as a source of the innovation of materials sciences. Although more than 23 million of substances have been synthesized or separated in the past century, only a small portion has found real-world applications. Currently, it remains a grand challenge for the scientific community to produce new substances with desired functions on a rational basis. One of the core tasks in the field of structural chemistry is to reveal the relationship between composition structure–function for creating new substances with expected properties through better molecular or structural design. In fact, the quantitative relationship between the structure and the property of a substance represents one of the challenges in chemistry in the twenty-first century. On the other hand, the synthetic processes must be economical, safe, resource-efficient, energy-saving, and environmentally benign. Chemists should strive for developing “perfect reaction chemistry,” i.e. synthesizing target compounds with a 100% yield and 100% selectivity without producing any waste. To create the new substances with desired functions, it is highly desirable to carry out function-directed structural design and structure-directed precise synthesis. That is the new developing direction of structural chemistry. Advanced Structural Chemistry: Tailoring Properties of Inorganic Materials and their Applications, First Edition. Edited by Rong Cao. © 2021 WILEY-VCH GmbH. Published 2021 by WILEY-VCH GmbH.

2

1 Introduction

Infinite 3D framework compounds, (small cation or small anion)

H Cluster compounds or cage compounds (small cation or small anion)

H

H

Layer structures (2D compounds) (small cation or small anion)

H

H

H

Chain structures (1D compounds) (small cation or small anion)

The classification of the ionic-covalent compounds

According to the bonding characteristics and structural dimensions, chemical substances may be divided into four categories. The first category includes the molecules, cluster compounds, or cage compounds with discrete structures, and such compounds can be assembled into chain structures (one-dimension = 1D, the second category), layer structures (2D, the third category), and framework structures (3D, the fourth category) through ionic or covalent bonding modes. One of the main bottlenecks lies in the inadequacy of the original designing, methods, and theories for the creation of new functional substances with different structures. Therefore, by strengthening the research on the structure- and function-guided creation of new substances, this situation could be drastically changed. However, two key issues should be considered in the area of structural chemistry: (i) how to accomplish the oriented synthesis for the specific structure of matter? (ii) How to design the matter’s structure with specific function? In this book, we intend to reveal the relationship between the structure and function of matter and develop efficient and precise synthetic methodologies and theoretical tools for new functional substances. This book is written by the experts engaged in structural chemistry research of Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences. It aims to reflect the recent research progress of structural chemistry in China and abroad. The book mainly includes two parts: synthesis and structure, and structure and property. It focuses on the structural design and properties of clusters, porous materials for gas sorption, separation and catalysis, and photoelectric or magnetic materials. This book clarifies the scientific connotation and subject development direction of structural chemistry.

Introduction

In the first chapter, the metal–organic macrocycles containing regular/irregular polygons, including multilayered cyclic structures are discussed. In addition, helicates, tetrahedrons, triangular prisms, cubes, octahedrons, hendecahedrons, dodecahedrons, cuboctahedrons, hexadecahedrons, and barrel- and ball-shaped cages are reviewed. Discrete metal-organic metallasupramolecular structures have drawn increasing attention due to their special esthetic 2D or 3D structures and potential applications in molecular recognition, sensing, separation, and biomimetic catalysis. In general, metallasupramolecular structures can be defined as metallacycles and metallacages, while many borderline cases exist concerning the complexity of geometry of the ligands and shape of the internal cavity. To facilitate discussion, herein the metal centers are taken as vertices and the organic ligands as edges or faces in the polygons and polyhedra to define these metallasupramolecular structures as macrocycles and cages, respectively. In Chapter 2, we summarize the development of synthetic strategies, analytical techniques, fundamental structure characteristics, and the typical applications of the metal-oxo clusters, which have been divided into three families of the transition metal-, main group metal-, and lanthanide-oxo clusters. The metal-oxo clusters with atomical precise structures are of importance in several disciplines relevant to synthetic chemistry, analytical chemistry, catalysis, biology, medicine, magnetism, and materials science. In the oxo clusters of transition metal part, structural fundamentals and novelties of the crystalline transition metal-oxo clusters, including Ti, V, Nb, Mo, Pd, W, and Ta, will be introduced. Besides, their tunable properties and potential applications related to the enormous structural diversities will also be briefly covered. In the oxo clusters of main group metal part, an overview on the main group borates, geminates, and aluminum elements will be presented. Both inorganic and organic ligand-supported oxo cluster structures will be introduced. Finally, in the oxo clusters of lanthanides part, we will present the recent advances in the high-nuclearity lanthanide clusters, which are classified by the protecting ligands including O-donor ligands, N-donor ligands, multiple N,O-donor ligands, calix[n]arenes, and other donor ligands such as Se- or C-donor ligands. Based upon these sufficient discussions, the dependence of the synthetic strategy of lanthanide-oxo clusters on the ligands, anions, and templating effects will also be summarized. We believe that these structurally well-defined metal-oxo clusters can provide models for decoding the mechanistic insight of the metal-based materials at the molecular level. In turn, the performance of these clusters could further feedback to the guidance of designing materials in a more controllable way. Chapter 3 focuses on the synthetic and structural studies on the transition metal sulfide clusters, in particular, the heterometallic Mo–Fe–S and Mo(W)–Cu(Ag)–S clusters. Benefitting from the proposed structural models of FeMoco and inspired by the versatility of the unit construction concept, we synthesized a range of metal sulfide clusters exhibiting various structural types. Special attention has been paid to the designed synthesis of the molybdenum and tungsten heterometallic thioclusters with one or more cuboidal cluster units, as well as some novel mixed-metal sulfide coordination polymers. These results demonstrate how flexible the unit construction

3

4

1 Introduction

method can be when it comes to making cluster compounds with desired structures based on the diverse coordination modes of the sulfur atom and the knowledge of structural chemistry. The ideas behind these syntheses are of significant value not only to the structural modeling of the biological catalytic active sites but also to the creation of novel molecular functional materials. Chapter 4 summarizes the syntheses, crystal structures, and selected properties of a number of group 11–15 metal chalcogenides that are mainly obtained by Huang et al. Metal chalcogenides are a class of compounds characteristic of the covalent bonding between metal cations and chalcogen Q (Q = S, Se, Te) anions. Metal chalcogenides are important materials showing excellent properties in ion exchange, semiconductor optoelectronics/thermoelectric, nonlinear optics, photocatalysis, etc. Neutral amine molecules have been successfully incorporated into the group 12 metal chalcogenides, say II–VI semiconductors, forming a novel class of organic–inorganic hybrid nanostructures with a general formula of [(MQ)n (L)x ] (MQ = ZnS, ZnSe, ZnTe, CdS, CdSe, etc.; L = mono- or di-amines or hydrazine; n = 1, 2; x = 0.5, 1, 2). The hybrid materials exhibit strong quantum confinement effects and outstanding properties such as largely modified optic absorptions, white light emission, and nearly zero/negative thermal expansion. Some of the obtained chalcogenidometalates demonstrate superior ion-exchange properties of selectively capturing hazardous metal ions such as radioactive Cs+ , Sr2+ , UO2 2+ , and Ln3+ ions from complex solutions. In Chapter 5, various metal–organic frameworks (MOFs) are categorically highlighted on the basis of their dimensions, with a focus on the construction and structural analysis, which provides an introduction of recent achievements in MOFs not only for novices but also for experienced researchers. It is expected to guide the design and preparation of more novel and advanced functional MOF materials. MOFs possess the unique features over traditional inorganic or organic materials, including structural tunability, ultrahigh porosity, large surface areas, etc., which make them being widely applied in various areas such as storage and separation, energy transfer, catalysis, enzyme inhibitor, sensing, drug delivery, and so on. MOFs have received great attention in the last two decades, and in a short time, they have matured and grown to a huge system. According to the extended manner, the structures of MOFs can be described as one-, two-,or three-dimensional infinite networks. In Chapter 6, we give a brief survey of recent advances in the rational assembly of polyoxometalate (POM)/chalcogenidometalate-based clusters and superlattices, polyhedral coordination supermolecules, and MOFs to show the knowledge on their synthesis mechanism and procedures. First, we present the synthetic and design approaches to the many POM types encompassing iso-POMs, hetero-POMs, organically derived POMs by alkylation or coordination, and POM–MOF hybrids. Next, we summarize the design and synthesis of crystallographically defined chalcogenidometalate clusters, including supertetrahedral series of Tn, Pn, Cn, Tm,n, and oxychalcogenide types, most of which are allowed to be fabricated by the mixed-metal and cationic-template strategies to deal with the local and the global charge issues. Then, we outline the recent developments of coordination-driven

Introduction

self-assembly, with a focus on discrete architectures, viz metallacages of polyhedral shapes through well-defined directional-bonding assembly strategies (e.g. edge-directed, face-directed, and symmetry-adapted approaches) including diverse linking modes. Finally, we provide an overview of secondary building units (SBUs) mainly of metal carboxylates (classified by their geometry and the number of metal atoms), and the linkers bearing other binding groups, such as pyridyls, azoles, phenols, and their representative network examples, also highlight some synthetic strategies, including modulated synthesis, isoreticular expansion, topology-guided design, and multivariate chemistry. In Chapter 7, we summarize the typical topological networks of MOFs with some prominent examples. Over the past decade, great progress of coordination polymers including MOFs has been made and tens of thousands of new compounds have been synthesized. Topological analysis supplies a convenient tool to understand and simplify a large number of complicated compounds at the early stage. With the development of new concept “reticular synthesis,” the design and synthesis MOFs have gradually tended to be practical and theoretical. The so-called “reticular synthesis” refers to the process of assembling well-designed rigid molecular building units into predetermined regular structures (or networks) through strong bonding. “Reticular synthesis” provides a reasonable way to synthesize solid materials with high stability, extended structure, predesigned molecular building units, and special properties. Two important aspects of “reticular synthesis” are the rational design of SBUs and framework topologies. Numerous MOFs constructed by various SBUs and organic ligands but with same underlying topological networks can be designed and synthesized. Therefore, topological analysis has become a powerful tool to design and predict new MOFs. In Chapter 8, about 60 classical examples selected from the literature on a wide range of inorganic chalcogenide classes are presented, including centrosymmetric (CS) and non-centrosymmetric (NCS) crystal structures. These chalcogenides display a rich structure diversity based on the BBUs and can be divided into five classes according to the dimensional features: (i) the zero-dimensional (0D) discrete clusters; (ii) one-dimensional (1D) chains; (iii) two-dimensional (2D) layers; (iv) three-dimensional (3D) frameworks; and (v) mixed-dimensional (MD) structures. In this chapter, we focus on the unit cell, space group, and dimensional change as well as the structural assembly of selected chalcogenides. This work gives some exploring strategies for novel chalcogenides with diverse dimensions. In Chapter 9, we wish to present advances in the past 10 years for three kinds of optical materials: electroluminescent materials, photochromic materials, and second-order nonlinear optical (NLO) materials. Optical materials refer to materials that may modify parameters of input light (such as phase, intensity, and frequency) or switch reciprocally optical and nonoptical signals (such as electric, heat, and sound). Electroluminescent materials are materials that can be excited to excited states and then return to ground state with the release of light. They have been widely applied to LEDs or OLEDs, which have brought revolution of both lighting and display fields. Photochromic materials are bistable materials that may be switched between two ground states with at least one direction being excited

5

6

1 Introduction

by light. They have been used for glassware (such as sun glasses, automobile rearview mirror, building window, etc.) and chromic ink (for anti-fake, cloth decorating, etc.) in the market. Other applications, such as switching, memory, bioimaging, and radiation monitoring, have also been demonstrated in laboratory. Second-order NLO materials are usually crystals without inversion symmetry that are capable of generating a second harmonic. They play a significant role in the field of laser-related science and technology, such as semiconductor manufacturing, photolithography, optical storage, and high-capacity communication networks. This chapter emphasizes the description of relations between structures and electroluminescent properties for d8 /d10 heteronuclear metal complexes, photochromic properties for viologen compounds and their analogs, and second-order NLO properties for the KBBF, SBBO, AB4 O6 , apatite-like borates, ABCO3 , A3 VO(O2 )2 CO3 , and AXII 4 XIII 5 Se12 families’ derivatives. Chapter 10 will discuss the relationship between the symmetry-breaking crystal structures and ferroelectric properties, and the recent advances of molecular ferroelectrics are also systematically summarized. As an important family of electroactive materials, ferroelectrics with the polar structures are characterized by spontaneous polarization, of which the direction can be reversibly switched under external electric field. Ferroelectrics demonstrate diverse physical attributes including piezoelectricity, pyroelectricity, NLO effect, mechanical functions, and dielectric properties, occupying an indispensable position in the field of condensed matter physics. All these fascinating properties and wide applications of ferroelectric materials are closely related to their unique structural characteristics. Essentially, spontaneous polarization of ferroelectrics can only exist below the Curie temperature point (T c ), coupling with the remarkable symmetry breaking during the paraelectric-to-ferroelectric transitions. Various physical properties for ferroelectric materials exhibit giant anomalies in the vicinity of T c , stemming from their structural changes. Compared with traditional inorganic ferroelectrics, the counterpart of molecular ferroelectrics displays many distinct characteristics, such as light weight, easy processing, mechanical flexibility, tunable structure, biocompatibility, etc. Therefore, molecular ferroelectrics are currently becoming one of the research hotspots. In Chapter 11, three series of inorganic, organic, and inorganic–organic hybrid conductive materials were selected from the literature. An overview of some representative fantastic advancement in terms of crystal growth methods, inorganic/organic crystalline conductors, and semiconductors is presented along with current challenges and future research directions provided finally. Crystalline conductive materials have been receiving increasing attention in recent years, not only for their unique physical properties and their potential applications in future electronics. They can be divided into inorganic conductive materials, organic conductive materials, and inorganic–organic hybrid conductive materials. Inorganic conductive materials are the earliest emerging conductive materials and now becoming one of the most crucial functional materials. Organic conductive material is an interdisciplinary research field concerning the design, synthesis,

Introduction

characterization, and application of organic small molecules or polymers that show desirable electronic properties such as conductivity, semiconductivity, and even superconductivity. Organic–inorganic hybrid conductive material is a crystalline hybrid material composed of organic linker bridged metal ions or inorganic clusters. The deepened study of structure–property relationships is necessary to help guide the design and synthesis of novel functional conductive materials. In Chapter 12, we will give a comprehensive overview of magnetostructural correlations that have been illustrated in the field of molecular magnets, in particular those reported by our institute. We will examine the environments of spin carriers, low-dimensional magnets, the magnetic exchanges between nearest-neighbor spin carriers, and the diversity of topologies of magnetic frameworks. The correlations between these structural features and the seemingly complex magnetic behaviors including spin crossover, magnetic slow relaxation, magnetic ordering, and spin frustration will be discussed. In Chapter 13, we present a comprehensive and detailed introduction of gas uptake properties of MOFs. First, we provide the basic concepts of gas adsorption, such as physical adsorption, chemical adsorption, Langmuir monolayer adsorption, and BET multilayer adsorption, and expound the correlation between the type of adsorption isotherms and MOF structures. Second, we classify the pore sizes of MOFs into three kinds based on IUPAC definition and discuss the relationship between pore size and sorption isotherms. Third, we describe the defects in MOFs and their measuring technologies and further expatiate on the connection between defects and adsorption capacities. Fourth, we summarize some representative MOFs with high surface area and discuss four universal factors and their influence on the surface area of MOFs. Finally, we introduce the concept of adsorption enthalpy and elaborate three methods to calculate adsorption enthalpy based on the experiments. Besides, we distinguish the difference between adsorption amount and usable adsorption amount and discuss the factors influencing adsorption enthalpy of materials and the influence of adsorption enthalpy on adsorption capacities and work capacities in practical application. We expect this chapter can deliver significant guidance for researchers to design/discover novel MOFs with excellent properties of gas storage and separation and therefore expand their practical application. In Chapter 14, we introduce the CO2 capture and separation, separation of hydrocarbons, noble gases, hydrogen isotopes, and the enantioselective separation based on MOFs. Meanwhile, the relationship between structure and separation during these processes will be discussed in details. In Chapter 15, we review and summarize the typical strategies and examples in MOF catalysis to give a brief but broad scope overview of the MOF catalysis area. We will first discuss catalysis with different active sites: (i) open metal nodes (SBUs) and modified SBUs for Lewis acid catalysis, oxidation reaction, Suzuki–Miyaura coupling reaction, and so on; (ii) linkers functionalized by Brønsted acids, unsaturated metal complexes, or organic bases; and (iii) pore-encapsulated guest species such as metal nanoparticles, POMs, metal complexes, and enzymes. Three major approaches are discussed next to show the unique advantage of MOFs compared

7

8

1 Introduction

to traditional heterogeneous and homogeneous catalytic systems: heterogeneous asymmetric catalysis based on privileged ligands or chiral pore structure; cooperative or sequential catalysis that involves more than one active sites; and synergistic photocatalysis for CO2 reduction, water splitting, and organic reactions. This book is dedicated to the 60th anniversary of Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, and its founder director, Professor Jiaxi Lu, on the memory of his 105th birthday.

9

2 Coordination-Assembled Metal–Organic Macrocycles and Cages Xiao-Zhen Li, Yu-Ling Liang, Li-Peng Zhou, Li-Xuan Cai, Qiang-Yu Zhu, Zhuo Wang, Xiao-Qing Guo, Dan-Ni Yan, Shao-Jun Hu, Shao-Chuan Li, Shi-Yu Wu, Shi-Long Han, Ran Chen, Pei-Ming Cheng, Kai Cheng, Xiao-Shan Feng, Tian-Pu Sheng, Can He, Feng-Rong Dai, and Qing-Fu Sun Chinese Academy of Sciences, State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, 155 Yangqiao Road West, Fuzhou, Fujian 350002, P.R. China

2.1 Introduction Metal–organic supramolecular macrocycles and cages have attracted considerable attention due to the ease of preorganization self-assembly and unique physicochemical properties, such as host–guest chemistry, molecular recognition, etc. Similarly with the construction of highly complex supramolecular assemblies with biological importance using simple and identical subunit in nature, coordination-driven self-assembly provides a facile way for the construction of highly organized supramolecular architectures, assembling different components into predesigned assemblies with well-defined sizes and shapes. Since the early exploration of two-component complexes involving Pt– or Pd–pyridine interactions, transition metals, lanthanides, and massive donor units are subsequently adopted, generating multicomponent macrocycles and cages with diverse structures and functionalities. Herein, this chapter reviews the recent developments in metal–organic coordination macrocycles and cages, with a focus on the construction and functionalities. It is organized based on the configuration of the structures and key design principles, and their functionalizations will be discussed particularly.

2.2 Metallacycles Metallacycles stand for one unique class of discrete two-dimensional (2D) structures, which can be constructed by using metal centers (such as Pt, Zn, Cu, Au, Ag, Pd, Ru, Rh, Ir, etc.) assembled with suitable ligands. Ever since the first report on the tetranuclear Bpy4 (enPd)4 “molecular square” by the Fujita group in 1990, overwhelming examples of beautiful metallacycles have been documented in the Advanced Structural Chemistry: Tailoring Properties of Inorganic Materials and their Applications, First Edition. Edited by Rong Cao. © 2021 WILEY-VCH GmbH. Published 2021 by WILEY-VCH GmbH.

10

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

literature. Representative reports from Chinese researchers in the past two decades are summarized below.

2.2.1

Dinuclear Metallamacrocycles

In 2012, Yu and coworkers reported a bistable self-assembled M2 L2 metal–organic macrocycle synthesized from Pd(II) with di(1H-naphtho[2,3-d]imidazol1-yl)methane (L) (Figure 2.1) [1]. The macrocycle undergoes intramolecular conformational motion switched reversibly with anions by multiple hydrogen-bonding interactions. When employing nitrate precursor, two ligands adopt the cis-conformation and form a bowl-shaped structure which is locked by the nitrate anion through quadruple hydrogen bonds between the oxygen atom of the nitrate and the hydrogen atoms from the naphthanoimidazolium moieties. When the counter-anion is changed to tetraphenylborate, all ligands switch to the trans-conformation and a chair complex is obtained. Remarkably, two acetonitrile molecules are located along the central axis of the complex with C–H· · ·π interactions. Interestingly, the platinum analog of this complex retains a fixed bowl-shaped conformation, and no conformation change is observed regardless of the anions owing to the inert Pt—N bonds.

2.2.2

Triangles

In 2018, Huang’s group constructed supramolecular triangles with excellent luminescent properties by using bodipy-based bridging ligands assembled with Pt(II) (Figure 2.2) [2]. These triangles exhibit outstanding anticancer activity due to the presence of the vertical Pt(II) ions. Simultaneously, its transportation into cancer cells can be visualized because of the fluorescence property of bodipy cores. This study demonstrates that the formation of triangles can improve their anticancer efficacy compared to their metal precursors.

Unlock BPh4– NO3– Lock

Figure 2.1 Fine-tuning conformational motion of a self-assembled metal–organic macrocycle by multiple C–H⋅⋅⋅anion hydrogen bonds. Source: Reproduced with permission Xie et al. [1]. © 2012, Wiley-VCH.

2.2 Metallacycles

H3g N OTf Pt N OTf

H4c

H4a

H3a H3b N

H3d

4

N

N

H4b H4d

H3f H3e

H3c

N B F F

Et3P OTf Pt Et3P OTf

3

5

[3+3] 60 °C, 24 h, CH2Cl2/DMA

3+4

1

3+5

2

Figure 2.2 Supramolecular triangles for cancer cells transportation. Source: Zhou et al. [2]. © 2018, American Chemical Society.

In 2015, Chan’s group reported a series of triangular metallacycles using three unsymmetrical bisterpyridine aligands with different lateral lengths assembled with Zn2+ [3]. The self-assembly processes are strongly geometry dependent. A ligand with the largest geometrical difference between two coordinating moieties has higher degree of cross-ligand self-selection compared with the complexation reaction of other ligands, giving the single head-to-tail (HT) triangular species. The isomeric mixtures are characterized and readily differentiated by gradient MS/MS coupled with TWIM-MS in the gas phase. HT isomers are found to be more stable than the head-to-head (HH) isomers, which are confirmed by DTD shift toward a longer drift time with increasing trap voltage and their corresponding CCSs agree reasonably with the computational values. Li and coworkers prepared a triangular Cu3 and its extended framework [Cu3 ]n by using pyrazole-based ligands [4]. Two luminescent coordination compounds, [Cu(Pz)]3 and [Cu2 (Bpz)]n , were isolated

11

12

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

from solvothermal reactions of Cu(NO3 )2 with 3,5-dimethylpyrazole (HPz) and 3,3′ ,5,5′ -tetramethyl-4,4′ -bipyrazole (H2 Bpz), respectively, in the presence of NH3 .

2.2.3

Rectangle

In 2019, Han’s group demonstrated that new supramolecular metallacycles were capable of undergoing photochemical reactions in solution utilizing metal-carbene as templates [5]. These new rectangular metallacycles are generated by employing dinuclear metal-carbene organometallic clips and olefin-functionalized bridging ligands through coordination-driven self-assembly. Photolysis of these molecular metallacycles in situ leads to structural interconversion and release of the cyclobutane products quantitatively (Figure 2.3). This supramolecular-assisted synthetic methodology enables the preparation of new types of functionalized cyclobutanes. Jin’s group had also a long-standing interest in the self-assembly of two preorganized binuclear half-sandwich metal (Rh or Ir) molecular clips and pyridyl ligands. They reported two molecular organometallic box (3a, 3b) compounds by applying 6,11-dihydroxy-5,12-naphthacenedione (H2 dhnq) and pyrazine spacing ligands assembled with half-sandwich iridium/rhodium (Figure 2.4) [6]. Tetranuclear complexes [Cp* 4 M4 (m-pyrazine)2 (m-L)2 ]-(OTf)4 (M = Ir (3a), Rh (3b); L = dhnq2) were obtained in high yields from 1a or 1b by reacting with pyrazine in the presence of AgOTf (Tf = O2 SCF3 ), respectively. The highly symmetrical structure of this rectangle, composed of four Ir/Rh clusters as the four vertices and two pyridines and two dhnq2 as edges, was confirmed by NMR spectroscopy and (a)

nBu

nBu

nBu

N

N

N

Au-Cl

Au

Au

N

AgOTf

+ N

N N

Au-Cl

N N

N

in situ –2 AuOTf

3a–g

N

(b)

N

nBu

N

2+ – N nBu 2OTf

N

N

N

N

Au

+ nBu

N

Au n N Bu

nBu

nBu

2a–g

hv 365 nm

N

N

nBu 1a

N Au

Au

N

4+ – 4OTf

5a–g N

N

N

N N

4 N N

N

N

N

N

N

N

N

5a

N

N

N

N

N

N

N 5e

N

N

N N

5b

N

N

N N

N

N

N

N

N 5f

N

N

N

N

N

5c

N

N

O

N N

N N

HO

N N

N N

O

N

5g

N

N

5d

O OH

OH OH

O 5h

Figure 2.3 (a) Coordination-driven self-assembly of organometallic metallacycles 3a–g and their photolytic transformation to afford cyclobutane derivatives 5a–g. (b) The structures of cyclobutane derivatives 5a–h. Source: Ma et al. [5]. © 2019, Wiley-VCH.

2.2 Metallacycles 4+ 2

Cl

Cl

O

O

M

M O

O

O

O

O CH3OH – AgCl CH3OH

O

AgOTf

M

1a: M = Ir 1b: M = Rh

CH3OH M

O O

M N

O O M N

N

N

CH3OH

N O O M

N O O

M

2a: M = Ir 2b: M = Rh

3a: M = Ir 3b: M = Rh

Figure 2.4

Stepwise formation of 3a and 3b. Source: Han et al. [6]. © 2009, Wiley-VCH.

Figure 2.5 Size-dependent formation of Borromean link structures. Source: Huang et al. [7]. © 2013, American Chemical Society.

L1 = N

N

L2 = N

L3 = N

N N 2–

L

Cu

=O O

L1,L2,L3

N N Cu O O Rh

1, 2, 3

L= L4 = N L5 = N

O N H

O O

= L4,L5

= H N

N

O

4, 5

N

single-crystal structure analyses. More interestingly, organometallic molecules display face-to-face π–π interactions between the planes of adjacent dhnq2 ligands, generating one-dimensional (1D) coordination frameworks, and the Cl2 C=CCl2 guests are shown to drive the transformation of the 1D assemblies to a 2D framework through CH· · ·Cl interactions. The 1D rows are perfectly aligned to form the 2D framework, in which the face-to-face π–π interactions between the planes of the monomeric complexes are remained. Because of the openness of coordinated metal centers, aromatic D–A–π interactions, guest-induced interactions, etc., the metallacycles can form Borromean rings (BRs) through adjusting their arm length (Figure 2.5) [7]. Arising from the combination of open Cu centers and favorable cavity space, [(Cp* Rh)4 (bpe)2 [Cu(opba)⋅2MeOH]2 ]4(OTf)⋅6MeOH shows extraordinary catalytic abilities with high efficiency and wide substrate selectivity in the acyl-transfer reaction. When longer linear pyridine ligand (L2) was used to react with dirhodium precursor 1a, a tetranuclear metallarectangle 5a was achieved, as confirmed by NMR, ESI-MS, and single-crystal X-ray crystallographic analyses (Figure 2.6) [8]. However, when diiridium precursor 2a and L2 used to build the metallarectangles, quantitative self-assembly of molecular rectangle 6a was noticed. Similarly, interlocked BRs (6a-BRs) complexes were produced when halogenated diiridium precursors 2b, 3, 4 were used. Halogen atoms are proposed to play an important role in forming BRs structure. It is worth to note that formation of BRs is concentration-dependent, and reversible conversion of the two structures has been achieved by adjusting the solvent conditions. By using an elongated Cp* Rh-based dynamic conformational binuclear building block, a class of tetranuclear rectangles with multiple aromatic stacks was achieved by a guest-induced constitutional rearrangement [9].

13

14

2 Coordination-Assembled Metal–Organic Macrocycles and Cages 2+

HO O

M

L1

M Cl

Cl

O

O OH 2 AgOTf 2 NaOH

L2 N

N

O

5a: M = Rh 5b: M = Ir

O (OTf)2

O H2O

H2O

M 6a: M = Rh, X = Cl 6b: M = Ir, X = Cl 7: M = Rh, X = F 8: M = Rh, X = Br

1a: M = Rh 1b: M = Ir

Cl

M Cl

2+

HO O X

M O

X

H2O O

X O OH X X= F, Cl, or Br O O (OTf)2 2 AgOTf M H2O 2 NaOH 2a: M = Rh X = Cl 2b: M = Ir X = Cl 3: M = Rh, X = F 4: M = Rh, X = Br

L2 Dilution with MeOH

Increase H2O to MeOH (6a and 8)

6a–BRs: M = Rh, XR = Cl 6b–BRs: M = Ir, X = Cl 7–BRs: M = Rh, X = F 8–BRs: M = Rh, X = Br

Figure 2.6 Synthesis of interconvertible metallarectangles and Borromean rings based on different dinuclear precursors. Source: Lu et al. [8]. © 2017, Cell Press.

Owing to the often-similar physical and chemical properties of structural isomers of organic molecules, facile separation of regioisomeric compounds remains difficult. An organometallic capsule was synthesized based on binuclear half-sandwich metal (Rh or Ir) molecular clips, in which two silver centers are rigidly separated from each other by two tetranuclear [Rh4 ] pyramidal frustums. Due to the shape and size matching, the heterometallic capsule could wrap a series of aromatic compounds [10]. This research provides an important solution for the separation of aromatic isomer molecules in industry. Air- and moisture-stable coordinatively unsaturated organometallic complexes were synthesized based on Cp–M–S fragments [11]. A series of 16-electron M2 L2 -type metallacycles were more inclined to bind with small ligands such as MeCN, Cl− , CO, pyridine, and so on (Figure 2.7). The coordination geometry around the metal center displays a two-legged piano stool, and the S—M—S bond angle is approximately 90∘ . More interestingly, [(Cp* M)2 (μ-L1)2 ](OTf)4 can reversibly convert into [(Cp* M)2 (μ-L1)3 ](OTf)4 , along with [(Cp* M)2 H3 ](OTf)4 and [(Cp*M)2 (μ-L1)3 ](OTf)4 becoming 18-electron M2 L3 -typed cylinders. The authors propose a mechanism that once hydride iridium species are formed, the released free ligands can coordinate to coordinatively unsaturated metal centers, rapidly leading to 18-electron M2 L3 -typed complexes.

2.2.4

Hexagons

Chiral NH functionalities play a key role in controlling the homochirality in nature’s chemical armory, such as the fabrication of DNA, RNA, and enzymes. In 2017, Cui and coworkers reported the design and synthesis of a series of chiral NH-functionalized cyclic structures through introducing partially reduced analogs of zinc–salen complexes as building blocks (Figure 2.8) [12]. In the asymmetric unit,

2.2 Metallacycles

R N Cl

Cl M

(i) 4 AgX

M Cl

Cl

Y

M

MeOH

M

(X)4 S

M = Ir, Rh

S

Y

N

R N

4+

S

S

(ii) 2 L

N R

N

N

N

N R

L1–L6: R = Me L7: R = Et

Y=

L1

L3

L2

L4

L5

L6

L7

Figure 2.7 Synthesis of coordinatively unsaturated 16-electron organometallic metallacycles. Source: Han et al. [11]. © 2014, American Chemical Society.

N N

O

Zn

N

H

O

N N

N

O

Zn

N O

N

Zn(salen)

Zn(salalen)

1:1

(a)

(b)

(c)

Figure 2.8 Self-assembly of chiral NH-controlled supramolecular metallacycles (R)-1 (a) from Zn(salalen) and (R)-3 (c) from Zn(salen). (R)-2 (b) was achieved by Zn(salalen) and Zn(salen) with 1 : 1 ratio. Source: Dong et al. [12]. © 2017, American Chemical Society.

the Zn ions adopt a distorted square pyramidal geometry coordination environment and adjacent Zn ions are bridged by peripheral pyridyl groups, resulting in a hexameric metallacyclic complex. Moreover, the metallacycles are packed through multiple supramolecular interactions and form an interlocking structure via strong hydrophobic interactions between cyclohexyl groups. Some functional metallacycles were reported by Huang’s group. As shown in Figure 2.9, model 2a solution exhibits clear-to-opaque transformation above 80 ∘ C and model 2 shows no lower critical solution temperature (LCST) behavior. A thermosensitive amphiphilic hexagonal metallacycle (1) with good LCST behavior

15

(a)

(b) O

O

O

O

O

O O

O

O O

O

O

O

O

O

O

O O

O

O H3 O

O O

Ha O

O

T>

H4 O

R

O

R

H2

T
K LC ST 2 T< LC ST 2

LCST1 > LCST2

Figure 2.9 (a) Self-assembly of 2 and 3 to give an amphiphilic discrete organoplatinum(II) metallacycle. (b) Cartoon illustration of its thermosensitivity and potassium cation responsiveness. Source: Wei et al. [13]. © 2014, American Chemical Society.

2.2 Metallacycles

is fabricated by introducing hydrophilic and thermal-responsive tri(ethylene glycol) (Tg) around the hydrophobic core and then self-assembling in D2 O/acetone-d6 mixed solution. Tg chains of 1 that form H-bonding interactions with solvent molecules below clouding point (T cloud ) become nonpolar and intermolecular H-bonding interactions above T cloud , which is an extremely important reason why 1 solution exhibits LCST behavior. Moreover, the LCST of 1 solution degrades with the increase of the concentration and the addition of K+ . The K+ could bind to the Tg chains, which disrupts the H-bonding interactions between Tg chains and water molecule and thus decreases T cloud . Opaque 1 solution is clear upon cooling below T cloud , which shows the excellent thermal reversibility of metallacycle solution [13]. Tetraphenylethylene (TPE) has excellent aggregation-induced emission (AIE) property and is a good building block for the construction of functional metallacycles. Multiple TPE units were used to construct rhomboid metallacycle 7 or different sized hexagons 8–11 based on coordination-driven self-assembly between 60∘ (2), 120∘ (3, 4), and 180∘ (5, 6) organoplatinum(II) acceptors with 120∘ TPE-based dipyridyl ligand (Figure 2.10a) [14]. Figure 2.10b,c shows that TPE-based 2D hexagonal metallacycles and three-dimensional (3D) drumlike metallacages with three different counter-anions are prepared by coordination-driven self-assembly, respectively [15]. Compared with TPE molecule, these multi-TPE metallacycles and metallacages display novel AIE property and higher quantum yields, due to the rotatable external pendant phenyl rings on the metallacycles and the locked TPE-based ligands, respectively, which are beneficial to eliminate the nonradiative decay pathways. Moreover, these metallacycles can be used as sensors due to the optical changes when interacting with electron-deficient substrates, such as picric acid. In 2012, a series of light-responsive multi-bisthienylethene (BTE) hexagons were reported by Yang’s group, with precise control over the structural shapes, sizes,

(b)

(a)

(c)

Figure 2.10 (a) Self-assembly of 1 with 2–6 to give rhomboid 7 and hexagons 8–11. (b) Self-assembly of 3 with 4 to give 5. (c) Self-assembly of 3 with 6 to give 7. Source: Yan et al. [14, 15]. © 2015, 2016, American Chemical Society.

17

18

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

o s o a N

b

e

g f

S

BTO

d

S

c N

h

Et3P Pt TfO PEt 3

120°

O

1

PEt3 i PEt3 Pt OTf PEt3 TfO Pt Pt PEt3 PEt3 OTf Et3P 2

UV

1 Vis +

3

BTO

PSS of 3

180° UV

2

Vis 4

PSS of 4

Figure 2.11 Graphical representation of self-assembled multi-bisthienylethene hexagons and their structural transformations. Source: Chen et al. [16]. © 2012, American Chemical Society.

and the number of photochromic units [16]. The 120∘ BTE-based donor and 120∘ or 180∘ diplatinum (II) acceptors are used as building blocks, generating [3+3] hexagon 3 and [6+6] hexagon 4, containing different numbers of photochromic units (Figure 2.11). These multi-BTE hexagons are highly sensitive and responsive to photo-stimuli as a result of the photoinduced switchable property and good fatigue resistance of the photochromic BTE system. Quantitative reversible conversions from the ring-open forms (3, 4) to the ring-closed forms (PSS of 3, 4) have been realized by alternating UV (365 nm) and visible-light (>510 nm) irradiation in degassed CH2 Cl2 , with higher conversion yields (99%) than that of the free photochromic BTE units (88%) due to the coordination-induced preference for the antiparallel configuration of the BTE moieties in metallacycle structures. The 2D or 3D metal–organic assemblies provide excellent scaffold for the construction of functional materials attributing to the enhanced intermolecular interactions. Reaction of alkynylplatinum(II)-containing donor ligand with diplatinum(II) acceptor yielded a vapochromic metallacycle, which exhibits highly selective color change toward CH2 Cl2 vapor [17]. Upon exposure to CH2 Cl2 vapor, the color of the metallacycle changes from yellow to red and retains in air for several months at room temperature, which is quite different from the behaviors of many known vapochromic materials. The chair conformation of the metallacycle in the solid state is crucial to its vapochromic behavior. It favors the close molecular stacking through intermolecular Pt…Pt and π–π stacking interactions triggered by CH2 Cl2 vapor molecules. The stacking mode affects the Pt…Pt distance and the HOMO-LUMO energy gap of the supramolecule and further results in the vapochromic phenomena. Moreover, mechanically grinding provides sufficient energy to disrupt the close molecular stacking, enabling a reversible color change.

2.2 Metallacycles

In 2014, they further designed three different kinds of peripherally functionalized 120∘ dendritic donor ligands to construct a series of metallodendrimers featuring a well-defined hexagonal metallacycle at the cores through coordination-driven self-assembly with 120∘ diplatinum (II) acceptor [18]. The multiple intermolecular interactions (e.g. π–π stacking, CH–π interactions, and hydrogen bonds) imposed by the peripherally DMIP-functionalized poly(benzyl ether) dendrons in hexagonal metallodendrimers facilitate the further formation of monodisperse vesicle-like nanostructures. Based on the dynamic nature of metal–ligand bonds, disassembly and reassembly of the hexagonal core are observed by the addition and removal of bromide anions. Therefore, well-controlled encapsulation and release of fluorescent dye have been successfully realized, which may be used as smart nanocapsules for guest capsulation and controlled release. Based on the high sensitivity and efficiency of fluorescence-resonance energy transfer (FRET) technique, 7-(diethylamino)-coumarin (donor)-decorated dipyridyl ligand 1 and rhodamine (acceptor)-functionalized diplatinum (II) unit 2 were employed to construct supramolecular metallacycles for real-time monitoring of the process and dynamics of coordination-driven self-assembly (Figure 2.12) [19]. (a)

N

O

N

N

O O O

O O

O

Et3P N

N 1

Pt O2NO PEt3

Et2P

2

PEt3 Pt ONO2

hvʺ (b)

hv

Coordination-driven self-assembly

+ Donor ligand

ET

FR

Acceptor ligand

Metallacycle

Figure 2.12 (a) Structure of building blocks 1 and 2. (b) Cartoon presentation of coordination-driven self-assembly of the donor ligand and the acceptor ligand. Source: Huang et al. [19]. © 2017, American Chemical Society.

19

20

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

Shorter distance between the coumarin and rhodamine moieties facilitates the FRET process, which further results in higher Forster energy transfer efficiency (FET ) in the system. The FRET phenomenon allows for further investigation of solvent effects in the self-assembly process, the dynamic ligand exchange between metallasupramolecular architectures, the anion-induced disassembly and reassembly of metallacycles, and the stability of metallasupramolecular structures in different solvents. FRET is capable of providing much more information about the process and dynamics of the supramolecular metallacycles and is quite helpful for the understanding of the coordination-driven self-assembly process.

2.2.5

Irregular Metallacycles

Duan’s group combined biomimetic hydrogenation with in situ regeneration of the active site in a one-pot transformation using light as a clean energy source for enzyme-like catalysis [20]. By modulating the active site of a nicotinamide adenine dinucleotide (NADH) model, this redox-active molecular flask facilitates the encapsulation of benzoxazinones for biomimetic hydrogenation within the inner space of the flask. The redox-active metal centers provide an active hydrogen source by light-driven proton reduction outside the pocket and further allow the in situ regeneration of the NADH models under irradiation, exhibiting high efficiency in the light-driven hydrogenation of benzoxazinones and quinoxalinones (Figure 2.13). In 2005, Yu et al. reported a luminescent supramolecular chiral Au16 ring that self-assembled from a tetrameric array of achiral {[(dppm)Au2 ]2 (pipzdtc)} (1; dppm = bis(diphenylphosphino)-methane, pipzdtc = piperazine-1,4dicarbodithiolate) subunits based on Au(I)…Au(I) interactions [21]. Each monomer 1 is linked by a bridging pipzdtc2− ligand with an average intramolecular Au(I)…Au(I) separation of 2.90 Å. Alternatively, the tetramer 14 can be viewed as made up of two halves, namely, a dimer-to-dimer coupling. The two monomers, [(dppm)Au(15)Au(16)-pipzdtc-Au(3)Au(4)(dppm)] and [(dppm)Au(1)Au(2)pipzdtc-Au(5)Au(6)(dppm)], spanned each other in space and are linked together by three intermolecular Au(I)…Au(I) interactions (Au(16)…Au(1), Au(2)…Au(3), and Au(4)…Au(5)), affording a chiral dimer, although the monomer itself is achiral. Similarly, the other two monomers formed another dimer. The two dimers with the same chirality were further coupled by two intermolecular Au(I)…Au(I) interactions (Au(6)…Au(7) and Au(14)…Au(15)) to form a closed chiral macrocycle with a framework of 16 Au(I) centers, directly linked via short Au…Au contacts, with a perimeter of up to 4.822 nm. The average intermolecular Au(I)…Au(I) separation in the ring is 3.12 Å, about 0.22 Å longer than the intramolecular ones. Both are significantly shorter than the sum of van der Waals radii for gold (3.32 Å), showing that Au(I)…Au(I) interactions exist in the molecule. On the basis of UV-vis and crystal structure studies, Yu speculated that upon an increase in the concentration, two monomers 1 which exist mainly in dilute solutions pair up to adopt a “cross-shaped” structure and result in chirality under driving of the intermolecular

2.2 Metallacycles

O

O O

O N

NH Encapsulation Departure Enzymatic hydrogenation

O

O

N O

NH

O O

N H

N O

O N H

In situ regeneration

N H

Ph

NH

O

Ph O H N

O

Ph

O N H

N Ph

Production active H-atoms

NADH model Ru(I) HA–

NAD+ model Ru(II) hυ

*Ru(II)

Figure 2.13 Schematic of the redox-active macrocycle with NADH mimics. Source: Zhao et al. [20]. © 2017, John Wiley & Sons.

aurophilic interactions. Then two dimers with the same handiness aggregate further with better packing to afford the chiral tetramer, probably induced and facilitated by the presence of Au(I)…Au(I) interactions. This unusual aggregate also provided a nice example of homochiral and aurophilicity-directed self-assembly. A crown-like Au36 ring was further developed through a spontaneous hierarchical hetero-chiral self-assembly as directed by strong AuI · · ·AuI bonding interactions from achiral components (Figure 2.14) [22]. In a first step, three bidentate dithiocarbamate units coordinate to six AuI centers in a HT cyclic manner to form a three-bladed propeller-shaped monomer 1 with D3 symmetry in dilute solution. 1 is equimolarly formed in the form of Δ or Λ. These six AuI centers are coplanar and form two homocentric parallel equilateral gold triangles with outer sides averaging 7.84 Å and inner sides averaging 3.05 Å. Upon increasing the concentration, three alternating Δ-Au6 and Λ-Au6 units then crystallize into a racemic cyclic hexamer to afford a giant crown of Au36 with a diameter of 2.25 nm and a perimeter of 6.88 nm, analogous to Pedersen’s [18]crown-6, with three monomers above and three below the mean plane of the gold atoms in the driving of strong intermonomer AuI · · ·AuI bonding interactions (average: 2.890 Å).

21

22

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

Figure 2.14 Au36 crown: a macrocyclization directed by metal–metal bonding. Source: Yu et al. [22]. © 2008, John Wiley & Sons.

2.2.6

Multilayered Metallacycles

A series of metallasupramolecular multilayered rings derived from multivalent 2,2′ :6′ ,2′′ -terpyridine (tpy) ligands were reported by Chan’s group and allowed the fundamental understanding of self-assembly behavior of sophisticated and functional supramolecular architectures (Figure 2.15) [23]. The flexible linker was found to be the key factor for controlling the architecture and stability of resulted assemblies. As a result of the geometrical constraint caused by the short C4 linker, complexation of L1 with CdII ions gave the complex [Cd6 L1 2 ] with a dimer inner ring. No expected [Cd9 L1 3 ] was formed (Figure 2.15a). With the C6 and C8 linkers, ligands L2 and L3 afforded the desired ring-in-ring structures [Cd9 L2 3 ] and [Cd9 L3 3 ] (Figure 2.15b), respectively. However, the self-assembly composition derived from L2 with CdII was found to be concentration- and temperature-dependent, and the proportion of [Cd12 L2 4 ] was observed to increase with either increasing concentration or decreasing temperature. When extending the scaffold of L3 horizontally along the linker axis, the multivalent ligand L5 was synthesized and achieved a spiderweb structure unambiguously when treated with CdII ions (Figure 2.15c). Self-selective ligation provides an efficient strategy for the preparation of more elaborate and diverse self-assembled structures. A pair of complementary tpy-based ligands selectively complexate with CdII ions into enlarged triangle and bilayered ditrigon structures (Figure 2.16) [24]. The auxiliary ion–dipole interactions between

(a)

(c)

rt

(b)

Figure 2.15 (a) Self-assembly of ring-in-ring structures from L1 , L2 , and L3 . (b) The molecular structure of [Cd9 L3 3 ]. (c) Self-assembly of spiderweb [Cd15 L5 3 ]. Source: Fu et al. [23]. © 2015, John Wiley & Sons.

(a)

(c)

(b)

(d)

Figure 2.16 (a) Self-selective coordination of L1 and L2 with CdII ions and (b) the crystal structure of [CdL1 L2 ]. (c) Self-assembly of Ttriangle [Cd6 L3 3 L4 3 ], ditrigon [Cd15 L3 6 L5 3 ], and (d) crystal structure of [Cd6 L3 3 L4 3 ]. Source: Wang et al. [24]. © 2016, American Chemical Society.

2.2 Metallacycles

the methoxy groups and the CdII center make L1 serve as a pseudo-pentadentate ligand to facilitate the heteroleptic complexation. The X-ray crystallographic analysis displays that two dimethoxyphenyl substituents and the central pyridine unit of L2 are parallel to each other with an average π–π stacking distance of 3.6 Å, and this donor–acceptor–donor stacking provides the additional thermodynamic stabilization for the heteroleptic structure (Figure 2.16b). The reaction of ditopic ligands L3 and L4 with two molar equivalents of CdII yields a multicomponent trigonal structure (Figure 2.16c). The crystal structure shows that each pseudo-octahedral CdII center is coordinated by two respective tpy units from L3 and L4 , resulting in an alternating ligand arrangement in the metallomacrocycle (Figure 2.16d). A more sophisticated ditrigon architecture is obtained through the self-assembly of CdII ions with ligands L3 and L5 . In 2014, Li’s group employed tritopic and tetratopic terpyridine ligands reacting with Zn(II) ions to produce two supramolecular hexagon wreaths [Zn9 LA6 ] and [Zn12 LB6 ], respectively [25]. Compared with conventional macrocycles that use more flexible ditopic building blocks and thus leading to a mixture of multiple structures, multitopic ligands with high density of coordination sites (DOCSs) provide more geometric constraints to form single thermodynamically stable architectures. The shapes, sizes, and structures of these stable assemblies were confirmed by NMR, ESI-MS, TWIM-MS, and TEM. Actually, such hexagon wreaths are constructed by small subunit of the dimer hexagons in a recursion process around a central hexagon, exhibiting the self-similarity of the fractal geometry. By applying this approach, more sophisticated 2D supramolecular architectures can be precisely designed. Based on AIE property of TPE, luminescent materials with similarly sophisticated structures were attained employing the third-generation AIE-active ligands with full conjugation of TPE with tpy. TPE–tpy ligands were assembled with Cd(II) to construct rosette-like metallosupramolecules ranging from G1 to G3 (Figure 2.17). A mixture of macrocycles was obtained by using ditopic ligand, while discrete double-layered hexameric and triple-layered heptameric rosettes were assembled from tetratopic and hexatopic ligands because such multitopic building blocks provided more geometric constraints in self-assembly processes to reach the most thermodynamically favorable structures. Additionally, because the intramolecular rotation around the TPE groups was further restricted after coordination, the assemblies exhibited high emission efficiency in both aggregation and solution states and displayed tunable emissive properties, especially, pure white-light emission for G2. This study provides an efficient strategy to construct various complexes and functional supramolecular architectures [26]. To increase the overall DOCS of macrocycles, a series of tetratopic ligands based on 2,2′ :6′ ,2′′ -terpyridine (tpy) were used to design and assemble a concentric hexagon system, namely, hexagon in hexagon. Hybrid concentric hexagons could also be obtained by combining different tetratopic ligands. Furthermore, these supramolecules were utilized as building blocks to hierarchically self-assemble supramolecular metal–organic nanoribbons (SMON) at highly oriented pyrolytic graphite (HOPG) surface through the stronger π–π interaction [27].

25

26

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

(a)

NN N

N N N

L1

Cd

G1

(b)

C6H13O

OC6H13

N N N

N N N N N N

L2

Cd G2

NN N

(c) C6H13O

OC6H13 OCH3 H3CO

N N N

N N N

OC6H13 C6H13O

Cd

N N N

G3

N N N N N N

L3

NN N

Figure 2.17 Self-assembly of supramolecular rosettes of (a) G1, (b) G2 and (c) G3. Source: Yin et al. [26]. Licensed under CC BY 4.0.

Inspired by Kandinsky circles, three generations of giant nested supramolecules were designed and assembled by Li’s group [28]. They overcome the challenge of the synthesis of multitopic ligands by using modular pyrylium salts followed by consecutive condensation reaction with primary amines. Furthermore, the nested supramolecules display high antimicrobial activity against Gram-positive bacteria MRSA and negligible toxicity to eukaryotic cells. This work paves a new avenue into the development and application of antimicrobial agents. To create more complex artificial nanostructures, two giant rigid 2D rhombus star-shaped supramolecules were designed and assembled by stepwise multicomponent self-assembly. The super snowflakes complexes encompass three types of tpy ligands and two metal ions, Ru(II) and Zn(II), in each individual structure. Additionally, dynamic ligand exchange of two preassembled architectures S2 and S1 is observed to form a series of hybrid snowflakes (Figure 2.18) [29]. The Star of David, a historical Hebrew symbol known as the Shield of David or Magen David, is a fascinating structure. But the construction of this topology always is a challenge in supramolecular chemistry fields. Due to the predictability

2.2 Metallacycles

6x

L5 12 x

S1

L3 12 x

6x L6

Zn Ru S2

Figure 2.18 Self-assembly of snowflakes S1 and S2. Source: Zhang et al. [29]. © 2017, American Chemical Society.

of coordination geometry among the coordination-driven self-assembly, not only effective synthetic strategies but also appropriate ligand designs play critical roles in the formation of highly ordered architectures. In 2017, Li and Wang’s groups constructed a supramolecular pentagram and a supramolecular hexagram by carefully designing metalloorganic ligands through a stepwise synthesis approach. By introducing robust Ru-polyterpyridyl moieties to ligand LA (VRu2+ X, V = bisterpyridine, X = tetraterpyridine, Ru = ruthenium) or LB [V(Ru2+ X)2 ], the multi-nuclear pentagonal and hexagonal architectures were achieved, respectively [30]. In the same year, they successfully synthesized 2D and 3D Star of David structures by one step self-assembly of a tetratopic pyridyl ligand with a 180∘ diplatinum-(II) motif and PdII ions. The 3D structure shows remarkable stability owing to its high DOCS analyzing by gradient tandem mass spectrometry (gMS2 ) [31]. A nut-like hexagonal bismetallo-architecture with a central hollowed Star of David was further synthesized by stepwise self-assembly. The expected hexagonal metallostructure T12 K6 Fe30 could not be obtained by one-pot self-assembly of T and K with Fe2+ directly. However, a higher ordered superstructure [T2 Ru]6 K6 Fe24 could be prepared by the reaction of Ru-dimer metalloorganic ligand (T2 Ru) with tetrakisterpyridine (K) and Fe2+ [32].

27

28

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

´ Sierpinski triangle, a fractal geometry, has attracted increasing attention in many fields including science, technology, engineering, math, and art. In 2017, Li and Wang’s groups reported the construction of different generations (G1 to G3) ´ of Sierpinski triangles using the based metalloorganic ligands as the building blocks to assemble with Zn(II) or Cd(II) in a stepwise strategy (Figure 2.19). Because of the structural and conformational complementarity of building blocks, or molecular puzzling, as well as highly reversible coordination, the stepwise self-assembly approaches were employed to overcome the formation of undesirable metallosupramolecules and coordination polymers [30]. ´ Compared to the low generation (≤3) of Sierpinski metallasupramolecular fractals, it still remains a formidable challenge to assemble high generation of fractal geometries due to the challenges of design synthesis and separation. In 2018, Li’s group used well-documented 2,2′ :6′ ,2′′ -terpyridine (tpy)-based coordination-driven self-assembly to construct discrete supramolecular fractals ranging from generation 1 (G1) to 5 (G5) in high yields (Figure 2.20). Except for G1 and G4, all other fractals were constructed using a series of Ru(II)-organic building blocks (ROBBs) to assemble with weak coordination metal, i.e. Zn(II). Those structures contain the same repeating shape unit, G1, but display at different scales, leading to different molecular geometry and symmetry. Moreover, all the geometry of G1 to G4 can be found as a proportion in G5, so they demonstrate the concept of “supramolecular fractal” [33]. Subsequently, they designed different generations of discrete polycyclic supramolecules using the same strategy, which were obtained in high yields by precise coordination of organic or ROBBs with Cd(II), Fe(II), or Zn(II) ions. Such 2D supramolecules with precisely controlled shapes and size have attracted considerable attention in materials science based on their hierarchical self-assembly behaviors both in solution and on the liquid–solid surface [34].

2.2.7

Polygon-Based Polymers

Coordination-driven hierarchical self-assembly provides an efficient way for the construction of dendronized polymers (DPs), which exhibit promising applications as electronic materials, liquid crystals, and siRNA delivery. In 2013, Yang’s group reported the dendronized rhomboidal organoplatinum(II) metallacyclic polymers constructed through hierarchical self-assembly of 60∘ organo-Pt(II) acceptors decorated with [G1]–[G3] Frechet-type dendrons (1a–c) with 2-ureido-4-pyrimidinone (UPy)-functionalized 120∘ dipyridyl ligand 2 (Figure 2.21) [35]. PM6 semi-empirical molecular orbital analysis indicates that the rhomboidal metallacycle has planar rhomboidal backbones with exohedral functionalization by the pendant UPy motifs and dendronized subunits. The highly directional and well-defined quadruple H-bonding of UPy facilitates the polymerization of the rhomboids into dendronized organoplatinum(II) metallacyclic polymers, which possess the structural features of conventional DPs as well as the dynamic reversibility of supramolecular polymers. The size of the polymers is highly dependent on the generation number of the attached dendrons, as confirmed by DOSY and DLS experiments. TEM morphological studies suggest that [G3]-metallodendrimer 3c aggregates to form single

R 1 R1

R1 R1

N N N

N N

Ru N

N

N N N N

R1 R1

N

N

Cd N N

N

R2

N

N

N

N

Zn N N

N

N

N

R1 R1

N

N

N

N

N

Ru N N

N

N

N

N

R1R1 R1 R

N 1

R2

N

N N

N

N

N

Cd N N

N

N

N

N N

R2 R2

N

N N

Zn N

N

N

N

N N

R1 R1

R1

N

N

Ru N

N

N

Zn N N

N

N N

N N N Ru N N N

R2 R2

N N

Zn N

N Zn N N N

N

N

Zn N N

N

N

Zn N

N

N

N N

N

Ru N N

N

N N

R2

N

R2 R2

N

Zn N

N N N

N

R2

N

Cd N N

R1

Figure 2.19

N

Zn N N N

Cd N N N

N

R2

N

N

Cd N N

N

Zn N N

N

N

Zn N N

N N

N

N

Ru N N N

R2 R2

N

Cd N N

N

N

Ru N N

R2

6PF6–

N

N

36PF6–

18PF6–

N

N

N

Zn N N

N

N

Zn N N N

N

R2 R2

N

N

N N N Zn N N N

N N Zn N N N

N N N Ru N N N

N N N

R2 R2

´ Chemical illustration of three Sierpinski triangles G1, G2, and G3. Source: Jiang et al. [30]. © 2017, John Wiley & Sons.

Ru N N

N

R1 R1

L

LA

LD 2×

LA

×1

LB

×1

3×

×2

6×

LC

12 × ×6

LD

LE

R = –OC6H13

Figure 2.20

×6

= Ru(II)

12 ×

G1 [Zn2LA2]

G2 [Zn3LB]

G3 [Zn6LC2]

G4 [Zn12LD6]

G5 [Zn12LE6]

= Zn(II)

Different approaches to prepare the ligands and fractal structures. Source: Wang et al. [33]. © 2018, American Chemical Society.

2.2 Metallacycles

OO O

O

O

O

O

O

O

O

O

O O

O

O

R

Pt R O2NO

Pt

R

R

R

ONO2

Pt

O2NO

R

1a

O

O

R

O O

O

R

Pt

ONO2

1b

R

1c

Pt

O2NO

Ha

R = PEt3 O

He

N

H3 H2

H1 R

R

Pt

R

ONO2

O

[2+2]

Hc Hb Hβ

Hd N

N H

O

Hf

O HN N H

O

+

DMSO-d6

N

2

Hα

O HN

O

O

N H

N H

N

O

O

O

O O

O O

O O

O O O

R O

N Pt R

O

R

N

Pt R

3c

Pt N O

O

N

H N

R N Pt

H N

O

O

O

O

O

O

R

R

O

O

R

O O O O

O

H-bonding

O

NH O

O

R H N

N

H N

O

[G3]-Dendronized organoplatinum metallacyclic polymer

HN

N H

N

N H R

O

NH

O

Figure 2.21 Formation of [G3]-DOMPs by hierarchical self-assembly of 60∘ [G3]-dendronized organo-Pt(II) acceptor 1c and 120∘ UPy-functionalized ligand 2. Source: Yan et al. [35]. © 2013, American Chemical Society.

31

32

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

polymeric chain nanofibers, and these fibers subsequently form laterally associated bundles containing several like strands, giving rise to widths of 6.0–50 nm. A new class of multiple stimuli-responsive polymer gels based on hexagonal metallacycles were also prepared through hierarchical self-assembly and host–guest interactions. The reaction of a new 120∘ pillar [5] arene containing dipyridyl donor ligand with complementary 180∘ di-Pt(II) acceptors with different lengths resulted in the formation of hexakis-pillar [5] arene metallacycles of different sizes [36]. PM6 semi-empirical molecular orbital calculations revealed a structure of roughly planar hexagonal ring at the core surrounded by six rigid pillar [5] arenes with the internal diameters of 5.0 (H2) and 5.8 nm (H3). Host–guest complexations of H2 and H3 with neutral ditopic guests were observed which led to the formation of analogous concentration-dependent cross-linked supramolecular polymers H2 ⊃ (G5)3 and H3 ⊃ (G5)3 . When the concentrations were increased to 22–23 mM, the cross-linked supramolecular polymers transformed into stable supramolecular gels with different colors. By taking advantages of the dynamic nature of the host–guest interactions and Pt–N coordination bonds, reversible multiple stimuli-responsive gel–sol phase transition of the polymer gels was realized through temperature control, the addition and removal of competitive guests and halide. In 2015, the Yang’s group further presented an example of counter polyanioninduced hierarchical self-assembly of discrete metallacycle [37]. Heparin, a sulfated glycosaminoglycan polymer with multiple negative charges that has been widely used as an anticoagulant drug, was selected to induce hierarchical self-assembly. The introduction of AIE active moiety, TPE, endowed the resultant metallacycle with interesting sensing functionality. The aggregation between the metallacycle and heparin, induced by multiple electrostatic interactions, resulted in dramatic fluorescence enhancement. Entangled pearl-necklace networks were obtained by multiple electrostatic interactions. Obvious emission intensity changes accompanied by the hierarchical self-assembly process with a linear relationship demonstrated that the functionalized metallacycle could work as a turn-on sensing probe for heparin. The metallacycle displayed high selectivity and sensitivity to heparin in comparison with structurally similar analogs such as chondroitin-4-sulfate (ChS) and hyaluronic acid (HA). Hexagonal metallacycles decorated with three CTAs at alternative vertexes were also synthesized through the coordination-driven self-assembly of 120∘ dipyridyl building block 1 and the corresponding complementary 120∘ di-Pt(II) acceptors 2 or 3 (Figure 2.22) [38]. The tris-CTAs metallacycles were further used as RAFT agent for the preparation of stimuli-responsive and self-healing supramolecular star polymers through the controlled radical polymerization of N-isopropylacrylamide (NIPAAM). The resultant star polymers exhibited LCST behavior in water because of the introduction of PNIPAAM moieties. Moreover, the supramolecular polymer hydrogels cross-linked by discrete metallacycles were obtained from the star polymers at room temperature without heating–cooling process, and the existence of metallacyclic scaffold endowed the obtained hydrogels with bromide anion stimuli-responsive and self-healing properties by taking advantage of the dynamic nature of the Pt—N coordination bonds. This study opens a

(a)

O O

Hα

Hβ

S

S S H2 H 3

H4 H 6 H5 H 7

Et3P

H1

N

TfO

N

1

O

H10 H9

Pt Et3P

PEt3

Pt

PEt3 OTf

or

Et3P Pt TfO PEt3

2

+ 1

Br

H8

PEt3 Pt OTf

3

Acetone

Nipam, AIBN

rt, 8 h

Acetone 60 °C, 4 h

2 or 3

Et3P

S C12H25 S

Sn H11 HN

H13 H14 O

O O

H12

4 or 5

6 or 7

(b) Cut Selfhealing

rt 24 h

Cut

Cut

Selfhealing

rt 24 h

Figure 2.22 (a) Graphical representation of synthesis of organoplatinum(II) metallacycles and star supramolecular polymers. (b) Rapid macroscopic self-healing experiments of the supramolecular polymeric hydrogel. Source: Zheng et al. [38]. © 2016, American Chemical Society.

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

new avenue to the construction of novel functional materials based on well-defined metallasupramolecular assemblies, in particular “smart” soft matters. Supramolecular hydrogels containing discrete organoplatinum(II) metallacycles are also promising candidates as biomaterials [39]. N,N-dimethylaminoethyl methacrylate (DMAEMA) and NIPAAM were selected as block segments due to the typical LSCT behavior of PNIPAAM and the CO2 -responsiveness and low cytotoxity of PDMAEMA. The star-shaped supramolecular copolymers with well-defined hexagonal metallacycles as cores were prepared through coordination-driven self-assembly and stepwise polymerizing DMAEMA and NIPAAM. The resultant copolymers featured CO2 stimuli-responsive properties including CO2 -triggered morphology transition and CO2 -induced thermoresponsive behavior. Reversible gel-to-sol transition of the supramolecular hydrogels with metal–organic macrocycles as junctions near physiological temperature can be realized through the removal and addition of CO2 . Control experiments show that both block copolymers and well-defined metallacycles are key factors in the realization of the CO2 -promoted hydrogel formation. The resultant polymeric hydrogel also displays injectable and cytocompatible properties.

2.2.8

Responsive Dynamic Metallacycles

In 2017, Sun’s group reported an artificial assembly system that features adaptive self-assembly and induced-fit transformation properties in the presence of anionic guests (Figure 2.23) [40]. Five metal–organic macrocycles with empirical formula

One simple semirigid ligand Guest-adaptive self-assembly

Pd7L14

Induced-fit transformation Pd6L12

(5)

Pd5L10 (4) 7 O 6– 24

PF6 – or OTf –

–

Mo

Pd4L8

(6)

BF 4

Dimension

34

(3)

HSO4– (2)

– NO 3

Pd3L6

N N

(1)

N N

PdII Number of components

Figure 2.23 Self-assembly of anion-binding metal–organic macrocycles. Source: Zhang et al. [40]. © 2017, Nature Publishing Group.

2.3 Metallacages

of Mn L2n (n = 3, 4, 5, 6, 7) were selectively obtained starting from one simple benzimidazole-based ligand and square-planar palladium(II) ions, either by direct anion-adaptive self-assembly or induced-fit transformations. Hydrogen bonding between the inner surface of the macrocycles and the bound guests, different anions in this case, dictated the shape and size of the final product. Moreover, a comprehensive map showing all the transformations across the macrocyclic species was drawn, with a representative reconstitution process from Pd7 L14 to Pd3 L6 traced in detail by titration experiments, revealing a gradual ring-shrinking mechanism. Moreover, this new class of donut-shaped assemblies could provide unique tunable hydrogen-binding pockets, where molecular sensing/catalysis may be possible to take place.

2.3 Metallacages Metallacages that are one class of discrete 3D structures have achieved magnificent attention in recent 20 years, not only for their esthetic attributes but also because of their wide applications in host–guest chemistry, supramolecular polymers, catalysis, biomedicine, and so on. Many domestic researchers have already developed a series of rationally designed metallacages with various shapes via coordination-driven self-assembly.

2.3.1

Helicates

In 2015, Sun’s group succeeded in designing a cationic M2 L4 cage by quantitative self-assembly of four anthracene-bridged benzimidazole ligands and two PdII ions [41]. The cage showed a D4 symmetry but exhibited excellent capability in the selective encapsulation of nitrate. Solid structural confirmation of Pd2 L4 was provided by X-ray crystallographic analysis. Crystallographic data show that four ligands in Pd2 L4 arrange in a quadruple helicate conformation due to the steric repulsion between the anthracene panels, resulting in a D4 symmetry of the host framework with inherent P or M helicity (Figure 2.24). More interestingly, the four anthracene walls of the ligand wrap up a very concise hydrophobic cavity where all the benzimidazole protons are pointing inward, forming a perfect bind pocket that is occupied by a nitrate anion. The binding constant (K+ anion) for the inclusion of NO3− is at least 2 orders of magnitude higher than all the other anions screened. Such a big difference can be attributed to the presence of maximal hydrogen-bonding interactions between the nitrate and the host cage in spite of the mismatching on symmetry. On the contrary, the lack of hydrogen bonding weakens this binding between anions and host even if their symmetries are better matching when single-atom halide anions are placed in the D4 symmetrical host. Similarly, little exchange for NO2− was observed. The constitutional control of structural interconversions of supramolecular architectures via molecular recognition or templating effects is of fundamental and high applicative interest. In 2016, Su’s group reported fully structural interconversions between monomeric Pd2 L4 and interlocked dimeric Pd4 L8 cages [42]. They discussed the origin of the thermodynamic stability and the driving force for

35

2.3 Metallacages

of Mn L2n (n = 3, 4, 5, 6, 7) were selectively obtained starting from one simple benzimidazole-based ligand and square-planar palladium(II) ions, either by direct anion-adaptive self-assembly or induced-fit transformations. Hydrogen bonding between the inner surface of the macrocycles and the bound guests, different anions in this case, dictated the shape and size of the final product. Moreover, a comprehensive map showing all the transformations across the macrocyclic species was drawn, with a representative reconstitution process from Pd7 L14 to Pd3 L6 traced in detail by titration experiments, revealing a gradual ring-shrinking mechanism. Moreover, this new class of donut-shaped assemblies could provide unique tunable hydrogen-binding pockets, where molecular sensing/catalysis may be possible to take place.

2.3 Metallacages Metallacages that are one class of discrete 3D structures have achieved magnificent attention in recent 20 years, not only for their esthetic attributes but also because of their wide applications in host–guest chemistry, supramolecular polymers, catalysis, biomedicine, and so on. Many domestic researchers have already developed a series of rationally designed metallacages with various shapes via coordination-driven self-assembly.

2.3.1

Helicates

In 2015, Sun’s group succeeded in designing a cationic M2 L4 cage by quantitative self-assembly of four anthracene-bridged benzimidazole ligands and two PdII ions [41]. The cage showed a D4 symmetry but exhibited excellent capability in the selective encapsulation of nitrate. Solid structural confirmation of Pd2 L4 was provided by X-ray crystallographic analysis. Crystallographic data show that four ligands in Pd2 L4 arrange in a quadruple helicate conformation due to the steric repulsion between the anthracene panels, resulting in a D4 symmetry of the host framework with inherent P or M helicity (Figure 2.24). More interestingly, the four anthracene walls of the ligand wrap up a very concise hydrophobic cavity where all the benzimidazole protons are pointing inward, forming a perfect bind pocket that is occupied by a nitrate anion. The binding constant (K+ anion) for the inclusion of NO3− is at least 2 orders of magnitude higher than all the other anions screened. Such a big difference can be attributed to the presence of maximal hydrogen-bonding interactions between the nitrate and the host cage in spite of the mismatching on symmetry. On the contrary, the lack of hydrogen bonding weakens this binding between anions and host even if their symmetries are better matching when single-atom halide anions are placed in the D4 symmetrical host. Similarly, little exchange for NO2− was observed. The constitutional control of structural interconversions of supramolecular architectures via molecular recognition or templating effects is of fundamental and high applicative interest. In 2016, Su’s group reported fully structural interconversions between monomeric Pd2 L4 and interlocked dimeric Pd4 L8 cages [42]. They discussed the origin of the thermodynamic stability and the driving force for

35

36

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

(a)

(b)

Figure 2.24 (a) The crystal structure of Pd2 L4 with one encapsulated NO3− anion. (b) Hydrogen-bonding interactions between the NO3− and the benzimidazole ligands. Source: Zhou et al. [41]. © 2015, The Royal Society of Chemistry.

structural transformations via thorough investigations on anion exchange behaviors, anion-bonding affinities, kinetics, and thermodynamics in monomerization and dimerization processes, which revealed the unstable nature of the interlocked structure itself but subject to modulation through variable host–guest interactions dictated by the anion-templating effect. Among the chiral lanthanide assemblies, triple-stranded dimetallic helicates Ln2 L3 possessing two opposite helical senses (P or M helicity) have been widely investigated. Stereocontrolled self-assembly of dinuclear triple-stranded europium helicate (Eu2 L3 ) based on DTE-functionalized ligands has been achieved via the chiral-induction strategy (Figure 2.25) [43]. The point chirality of the ligands is transferred to give either Δ or Λ metal-centers and hence leads to an overall P or M helical senses. Moreover, the helicates in solution feature a reversible photocyclization and cycloreversion, offering an opportunity to develop as chiroptical switches. Reversible photochromism has been revealed for both ligands and helicates, with distinct change observed on the circular dichroism (CD) signals for the latter. A pair of enantiopure pyrene-functionalized C2 symmetrical bis(tridentate) ligands [(R,R)-/(S,S)-L] were also synthesized, which led to the diastereoselective formation of P- and M-type Ln2 L3 helicates [44]. The self-assembly process was followed by NMR and ESI-Q-TOF-MS. Moreover, commercially available Δ-TRISPHAT proved to be an effective NMR chiral resolving agent to differentiate between the two enantiomers of the helicate. Wu’s group also reported a self-assembly triple-anion helicate featuring a cavity resembling that of the choline-binding protein ChoX, as revealed by crystal and DFT-optimized structures, which bind choline in a unique dual-site-binding mode (Figure 2.26) [45]. Of particular significance is this host system displays a unique high selectivity toward choline over the closely related acetylcholine

2.3 Metallacages F2 F2

F2 m S

N

3+

pcamEu

O

×2

×3

O

pc am

HN *

S n

HN

o-LRR/SS

P-Eu2(o-LSS)3

M-Eu2(o-LRR)3

Figure 2.25 Stereocontrolled self-assembly of Eu2 L3 triple helicates from ligands o-LRR/SS . Source: Cai et al. [43]. © 2018, The Royal Society of Chemistry.

(a)

(b)

P 6.263 Å N

(c)

6.422 Å II

P

Bending angle I

Figure 2.26 (a) Crystal structure of A2 L3 . (b) Hydrogen bonds formed between a PO4 3− ion and six urea units. (c) The aromatic box (site-I) trapping a TMA+ through cation–π interactions (purple-dashed lines) and a potential hydrogen-bonding site(II), which together resembles the structure of Ch+⊂ChoX. Source: Jia et al. [45]. © 2017, Nature Publishing Group.

37

38

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

(K choline /K acetylcholine > 15). The triple helicate is able to act as a fluorescence displacement sensor for discriminating choline, acetylcholine, L-carnitine, and glycine betaine effectively. Based on the molecular helicate constructed by hydrothermal reaction of 3,5-bis(2-pyridyl)pyrazole (HL) with NiSO4 ⋅6H2 O (or ZnSO4 ⋅7H2 O) and NaSCN, further solvothermal reaction with CuSCN in the presence of PPh3 led to the formation of helical coordination polymer complex [46]. The coordination polymer could also be obtained solvothermally in acetonitrile by the one-pot reaction of CuSCN, Ni2+ (or Zn2+ ) salt, HL, and PPh3 . The intermolecular noncovalent interactions between the helicates and the copper–pseudohalide components respond to the stereochemical induction of the double-helical coordination polymer, resembling the effect of the intramolecular coordination bonds in the assembly of discrete helicates.

2.3.2

Tetrahedron

In 2015, Sun’s group reported the first example of stereoselective self-assembly of chiral luminescent europium coordination tetrahedral cages and their novel self-sorting behavior (Figure 2.27) [47]. Two pairs of R and S ligands are designed on the basis of the pyridine-2,6-dicarboxamide coordination unit, bis(tridentate) L1 and tris(tridentate) L2 . Corresponding chiral Eu4 (L1 )6 and Eu4 (L2 )4 topological tetrahedral cages are independently assembled via edge- and face-capping design strategies, respectively. Narcissistic self-sorting was observed in the self-assembly process when two differently shaped ligands L1 and L2 were mixed. More impressively, distinct self-sorting behavior between Eu4 L6 and Eu4 L4 coordination cages has been observed for the first time when racemic mixtures of ligands are used. And the results reveal that ligands with different shapes and geometry are easily discriminated.

Figure 2.27 Schematic diagram of coordination self-assembly and self-sorting of lanthanide metal chiral cages. Source: Yan et al. [47]. © 2015, American Chemical Society.

Case 1: Narcissistic self-sorting +

+

Case 2a: Dynamic mixture

+

+

+

+

Case 2b: Narcissistic self-sorting +

+

2.3 Metallacages

Moreover, L2 shows unprecedented self-assembly capacity with metal ions spanning across the periodic table, including alkaline earth (CaII ), transition (CdII ), and all the lanthanide (LnIII ) metal ions (M), ascribed to the appropriate rigidity of the C3 symmetrical scaffolding, high assembly adaptability, and adequate chelating affinity with lanthanide ions of the neutral coordination motif [48]. More importantly, this versatile ligand displays rare and rather high discrimination between metal ions with identical coordination geometries as well as extremely small ionic difference, arising from supramolecular multivalent cooperativity, resulting in absolute or highly efficient metal ion self-recognition during mixed-metal self-assembly process. A supramolecular lanthanide extraction separation method was proposed based on the strong synergistic metal ion self-identification characteristics of M4 L4 cage, providing a new design principle for the next generation of efficient lanthanide separation. Concentration-triggered transformation from helicate to tetrahedron cage was obtained using oxazoline-based bis(tridentate) ligand (Figure 2.28) [49]. The improved stability and luminescent property of this new generation of Eu4 L6 tetrahedral cage facilitated their efficient sensing toward highly explosive nitroaromatic compounds. Moreover, luminescent sensing of nitroaromatic explosives was demonstrated, featuring selective and efficient detection of PA at ppb level. The bright luminescent self-assembled tetrahedral compound Eu4 L4 can behave as dual-responsive and highly selective luminescent probe toward both anions and cations, taking advantage of the intraligand charge transfer (ILCT) sensitization mechanism [44]. To improve the luminescent properties of the Ln-based tetrahedral cages, a group of bright luminescent lanthanide organic tetrahedrons with record-high emission quantum yields has been constructed by using two fully conjugated ligands featuring the triazole-pyridine-amido (L2) chelating moiety easily accessible from the Click reaction [50]. In addition, using the synergistic energy transfer (ET) between the lanthanide vertices, the ratio luminescent thermometer covering the physiological temperature range in the Eu/Tb mixed tetrahedron is

(a) N

O N

6 O

i

h

g

N H

f H N

d e

c N

O

L1

N

3O

N

N H

Eu4(L1)6

(b)

O

L2

Eu2(L1)3

O 4 Eu3+ = b N CH3CN a

H N

O

2x

N

O N

2 Eu3+ = CH3CN Eu2(L2)3

Figure 2.28 Self-assembly of either a Eu2 (L1)3 helicate or a Eu4 (L1)6 tetrahedron from ligand L1, and self-assembly of a Eu2 (L2)3 helicate structure from ligand L2. Source: Liu et al. [49]. © 2017, The Royal Society of Chemistry.

39

40

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

Previous reports

This work

N

O N

N N

Ln N

L1

N

L3

O

N

HN N

N

N N

HN

N

L2 N

Ln N N

N

N Ln

O

N N

Ln N

L4

O HN

Figure 2.29 Strongly luminescent lanthanide–organic tetrahedral Ln4 L4 cages based on tris(tridentate) ligands with different chelating groups. Source: Wu et al. [51]. © 2019, American Chemical Society.

proved to be a potential application prospect as a single-molecule cell thermometer. It is well documented that increasing the rigidity of ligands can effectively restrain the consumption of ET process on lanthanide–organic complexes. Thus, they replaced the inner amide bond with triazole on C3 symmetric ligands to increase both the rigidity and conjugation of ligands. In 2019, they further reported fine-tuned luminescence on a series of Ln4 L4 -type cages constructed by two new ligands (L3 and L4) with the triphenyltriazine core (Figure 2.29) [51]. They found that the fully conjugated L3, with 2,4,6-tris(4-pyridyl)-1,3,5-triazine (TPT) chelating groups, manifests excellent sensitization ability toward all the emitting lanthanide ions (Ln = Pr, Nd, Sm, Eu, Tb, Dy, Ho, Er, Tm, Yb) covering both the visible and the NIR regions. A record luminescent quantum yield for Tb4 (L3)4 is reached (Φ = 82%). As a control, L4 with an amido-pyridine-triazole (APT) chelate that differs from the known L2 only in the relative position of the amido and triazole groups exhibits weaker sensitization ability on the lanthanide-organic polydrons (LOPs). Finally, white-light emission systems have been obtained by combining the desired luminescent compounds. The study could offer important design principles toward strongly NIR-emitting LOPs. Moreover, these strongly luminescent LOPs provide new candidates for photochemical supramolecular devices. Anion coordination chemistry has developed rapidly in recent years because it plays an important role in biological, environmental, and chemical processes. Some pioneer studies in anion coordination chemistry show that anion displays similar properties with transition metals in coordination chemistry, such as coordination geometry and coordination number. These analogies provide promising ideas for the self-assembly of novel supramolecular systems based on anion coordination. Wu’s

2.3 Metallacages

(a)

(b)

12− Figure 2.30 (a) The tetrahedral cage [(PO3− in the crystal structure, with the 4 )4 L4 ] dark-blue facial ligand truncated so that the interior and opposite PO3− 4 corner can be seen. (b) Hydrogen bonds around the PO3− 4 ion. Source: Wu et al. [52]. © 2013, Wiley-VCH.

group from Northwest University in China has made a series of characteristic works in anion coordination chemistry from 2013. They developed a series of ortho-phenylene-bridged oligourea ligands, which displayed excellent affinity and complementarity to the tetrahedral phosphate anions. Moreover, the fully deprotonated phosphate anion could form 12 hydrogen bonds with ligand that achieves a theoretical and experimental coordination saturation. In the beginning, they designed the tris(bisurea) ligand by attaching three bisurea moieties to a central C3 -symmetric triphenylamine platform and reported the first tetrahedral anion cage ([A4 L4 ]-type (A = anion)) from ligand and PO4 3− ions. The new A4 L4 crystallizes from diethyl ether/acetonitrile in the centrosymmetric cubic space group P43n (Figure 2.30) [52]. The structure had the ideal T symmetry with one-twelfth of the tetrahedral molecular cage (one-third of a phosphate ion, one third of the ligand, one TMA as the counter-cation) appearing in the asymmetric unit. The phosphate ions occupied the vertices and the ligands lay on the faces. However, the tri(bisurea) group limited the cage’s inner space which was estimated to be 121 Å3 by DFT and could not encapsulate any guest. To solve the problem of cage’s small inner space, Wu’s group designed ampliative ligand to build a larger anion coordination cage. They used 1,3,5-triphenylbenzene as linker to replace triphenylamine to amplify its inner space, which was estimated to be about 229 Å3 [53]. Then this new anion coordination cage was used to encapsulate some chlorinated hydrocarbons at room temperature to explore anion cage’s host–guest chemistry and attempted to broaden its application. To make further research about A4 L4 cage’s host–guest chemistry, Wu’s group chose white phosphorus (P4 ) and yellow arsenic (As4 ) as guests, which were very unstable and difficult to storage and transport [54]. In this work, they demonstrated a new design strategy for constructing finite cages based on anion coordination chemistry. The phosphate coordination-based tetrahedral cages could readily accommodate the tetrahedral guests P4 and As4 , which is facilitated by the shape and size

41

42

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

Figure 2.31 Crystal structure of cage ⊃ TEA+ . Source: Zhang et al. [55]. © 2018, American Chemical Society.

complementarity as well as favorable σ–π and lone-pair–π interactions. Moreover, the latter case represented the first example of As4 inclusion in a well-defined tetrahedral cage. In 2018, they reported a C3 symmetric trisbis(urea) ligand based on the 2,4,6-triphenyl-1,3,5-triazine spacer, which assembled with phosphate anions to form a new A4 L4 -type tetrahedral cage, with unusually high packing coefficients (up to 99.5% for the best substrate) [55]. This cage was able to adjust its size and shape (from 136 to 216 Å3 ) by bending of the triphenyltriazine plane (Figure 2.31). This allowed it to accommodate relatively large guests. In the case of DABCO, inclusion within the cage allowed the degree of methylation to be controlled and the monomethylated product to be isolated cleanly under conditions where mixtures of the mono- and dimethylated adduct were obtained in the absence of cage. Except for the common tetrahedra based on anion coordination chemistry, Wu’s group also designed other characteristic anion complexes with different shapes. In 2014, they designed a tetrakis(bisurea)-decorated tetraphenylethene ligand and constructed a special cage with fluorescence “turn-on” over a wide concentration range, from dilute to concentrated solution and to the solid state [56]. This “anion-coordination-induced emission” (ACIE) is another approach for fluorescence turn-on in addition to AIE. As the simplest platonic polyhedron, tetrahedral cages have emerged a wealth of research prospect in supramolecular chemistry; thus, exploring new novel strategy for the synthesis of tetrahedral assembly is also meaningful. Yuan’s group first chose trinuclear zirconocene cluster as vertex to construct two types of tetrahedra in a one step [57]. The trinuclear cluster demonstrated a C3 symmetry with three C2 axes along the carboxylate ligands forming an angle of about 60∘ , which provided a possibility to act as vertex in constructing coordination polyhedra. Subsequently, according to edge- and face-directed strategies, both V4 E6 (V = vertex and E = edge) and V4 F4 (F = face) tetrahedra could be synthesized in in situ method. As shown in Figure 2.32, the V4 E6 tetrahedral geometry, four Cp3 Zr3 O(OH)3 second building units (SBUs) sit on the vertices and four benzene-1,4-dicarboxylic acid ligands are disposed along the edges, while four benzene-1,3,5-tricarboxylic acid ligands span the faces in V4 F4 tetrahedron. As an expansion of structure based on reticular chemistry, extended version of tetrahedra can be easily obtained using the same method. Further gas adsorption study demonstrates that these tetrahedral cages could maintain their architectural rigidity and permanent porosity after activation.

+

(a) O Cl Zr Cl

HO

Hydrolysis O

OH O

O Zr O

(b)

O Zr

=

Zr O H

O

O

O

O

O

+

V4E6 O

O

+ O

O O

O V4F4

Figure 2.32 (a) Schematic representations of the formation of trinuclear zirconocene nodes. (b) Configurations of [4+6] V4 E6 and [4+4] V4 F4 tetrahedra. Source: Liu et al. [57]. © 2013, American Chemical Society.

44

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

Rational manipulation of catalytically active ligands in constructing polyhedral cages not only can regulate the catalytic performance but also promote the design of more sophisticated and smart assemblies for enantioselective reaction. In 2018, Cui’s group reported the design and synthesis of five chiral single- and mixed-ligand tetrahedral supramolecular cages constructed by metallosalen ligands (a privileged chiral catalyst) as edges and four C3 symmetry Cp3 Zr3 clusters as vertices (Figure 2.33) [58]. These tetrahedral cages demonstrate inherent nanoscale hydrophobic cavities and various catalytically active sites, which endow them with prominent catalytic activity and enantioselectivity over sequential asymmetric alkene epoxidation/epoxide ring-opening reactions with up to 99.9% ee. In 2011, Duan’s group reported a novel cerium-based neutral molecular tetrahedron by incorporating amide-containing tridentate chelating sites into the three arms of a triphenylamine fragment which acts as a bright blue emitter (Figure 2.34) [59]. Single-crystal X-ray structural analysis verified the formation of Ce4 (H2 TTS)4 tetrahedron in a crystallographic C3 symmetry. The tetrahedron involves four vertical metal centers and four deprotonated H2 TTS ligands. Each cerium ion is chelated by three tridentate chelating groups from three different ligands which individually position on the four faces of the tetrahedron. Interestingly, the tetrahedron encapsulates molecules of nitric oxide (NO) and nitronyl nitroxide (PTIO) within the cavity and can be developed as a luminescent detector of PTIO, a specific spin-labeling NO trapper for biological imaging in living cells. The tetrahedron can also be applied in quantificational detection of free tryptophan in serum since the synergistic effects of hydrogen bonding, aromatic stacking, and size/shape matching [60]. The metal–organic tetrahedron can act as a redox vehicle to encapsulate organic dyes for photocatalytic proton reduction [61]. The assembled neutral Co-TFT tetrahedron consists of four ligands and four cobalt ions, in which each cobalt ion is coordinated in a fac-configuration by three bidentate thiosemicarbazone NS chelators from three different ligands. The three anionic sulfur donors are positioned on one side with great potential for attaching a proton via hydrogen-bonding interactions. Once organic dyes are encapsulated in the pocket as photosensitizer, these essential components are forced closer together and facilitate the photoinduced electron transfer from the excited state of the photosensitizer to the cobalt-based catalytic sites via a powerful pseudo-intramolecular pathway, resulting in excellent TON for the generation of hydrogen from water. When the ligand is coordinated with redox Ni, redox active Ni-TFT octahedron pocket is formed [62]. The encapsulation of organic dye within the pocket of the redox active vessel modifies photocatalytic proton reduction in the inner space of the pocket and gives molecular hydrogen and oxidized dye. The oxidized dye leaves the pocket and causes sulfide oxidation outside the cavity to give element sulfur, which combines the photocatalytic hydrogen evolution and sulfide oxidation. In 2017, Zhang’s group fabricated a water-soluble and ultra-stable Ti4 L6 tetrahedron based on the titanium-oxo clusters assembling with embonic acid ligands. This tetrahedron has excellent selective activities for dye photodecomposition through N–H⋅⋅⋅O hydrogen bonding and also shows stepwise assembly function with other metal ions (such as Co2+ , Ln3+ ) to form novel Ti4 L6 -Co3 cage and Ti4 L6 -Ln cage (Figure 2.35) [63].

(a)

O HO C O

N N Mn O O

O C OH

HO C O

N O

Cr

N O

O C OH

HO C O

N

N Fe O O

O C OH

HO

O

Zr

O Zr O

OH

Zr

O

O O H O

(b)

1Mn

1Cr

1MnCr

1MnFe

1CrFe

(c)

Figure 2.33 (a) Structures of the ligands and trimetallic cluster. (b) Schematic representations of cages 1Mn , 1Cr , 1MnCr , 1MnFe , and 1CrFe . (c) Single-crystal X-ray structure of 1Mn and its space-filling model (light blue, Mn; green, Zr; blue, N; red, O; gray, C; the cavities are highlighted by yellow spheres). Source: Jiao et al. [58]. © 2018, American Chemical Society.

46

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

OH

O

N

N

HN

N H

OH

O N

NH

O

N HO

Figure 2.34 Structure of H6 TTS, constructional fragments of the Ce4 (H2 TTS)4 tetrahedron. Source: Wang et al. [59]. © 2011, American Chemical Society.

COO–

+

O– O–

M4+

M = Ti, Zr, Hf COO– = Co2+, Ln3+ O O

O O

O

O

O

M O O

Figure 2.35 Self-assembly of water-soluble and ultra-stable Ti4 L6 tetrahedron for potential metal ion trapping. Source: He et al. [63]. © 2017, American Chemical Society.

2.3.3

Truncated Tetrahedron

In 2016, Wang’s group reported the first example of alkoxopolyoxovanadate-based SBU and its application in assembly to build metal–organic tetrahedrons [64]. They developed a new family of V4 E6 - and V4 F4 -type (V = vertex, E = edge, and F = face) metal–organic tetrahedrons using unprecedented Anderson-like

2.3 Metallacages

alkoxo-polyoxovanadate [V6 O6 (OCH3 )9 -(μ6 -SO4 )(COO)3 ]2− polyanions as 3-connected SBUs. A rational synthetic strategy was further developed to construct two supramolecular isomers based on polyoxovanadate (POV) organic polyhedra with tetrahedral symmetries (Figure 2.36) [65]. VMOP-α, a product of low temperature, has an extremely large cell volume (470 842 Å3 ), which is one of the top three for well-defined metal–organic polyhedra (MOPs). The “corner-to-corner” packing of tetrahedra leads to a pretty low density of 0.174 g/cm3 with 1D channels about 5.4 nm. While for the high-temperature outcome VMOP-β, the cell volume is only 15 513 Å3 . The packing mode of tetrahedra is “corner-to-face,” giving rise to a high-density architecture (1.324 g/cm3 and the channel is 0.8 nm). Most strikingly, supramolecular structural transformation between VMOP-𝛂 and VMOP-𝛃 could be reversibly achieved by temperature-induced solvent-mediated transformation.

2.3.4

Triangular Prism

Metal–organic hosts provide an ideal platform to mimic the highly evolved and finely tuned natural photosynthetic systems [66]. When a quinhydrone (QHQ) cofactor was encapsulated in the inner pocket of cobalt triangular prisms, extensive electron delocalization and decrease of the over potential of the metal sites for proton reduction were induced by host–guest interactions and allowed the tandem reductions being combined to efficiently reduce nitrobenzene using active H-sources from the light activation of water (Figure 2.37). Thiacalix[4]arenes, representing a versatile class of macrocyclic containers in supramolecular and molecular recognition chemistry, are structural analogs of calixarenes by replacing the methylene groups with sulfur linkages [67]. The thiacalixarene scaffold is a unique host with vast possibilities for functionalization not only at the upper and lower rim but also at the bridging sulfide groups. The corresponding sulfinyl and sulfonyl derivatives are accessible by direct oxidation of thiacalix[4]arene using the hydrogen peroxide or sodium perborate as the oxidant under mild condition. Thiacalix[4]arenes and their oxidized derivatives are excellent multidentate ligands for constructing polynuclear coordination complexes with various metal ions through their O- and S-donor atoms. For example, the shuttlecock-like tetranuclear clusters [M4 (calix)(μ4 -Y)] (Y = Cl, OH, H2 O, or CH3 OH) (Figure 2.38a), consisting of four metal ions, four acetates, a calixarene ligand, and a μ4 -Y (Y = Cl, OH, H2 O, or CH3 OH) were well stabilized with each metal center octahedrally coordinated by two phenolic O atoms, one sulfur atom or sulfonyl oxygen atom, two carboxylate O atoms, and one μ4 -Y. This type of tetranuclear cluster subunit is a useful building block, serving as a unique vertex, to construct coordination cages through substitution of acetate with suitable bridging ligands. In 2016, Liao’s group obtained a trigonal prismatic coordination cage from solvothermal reaction of H4 L5, NiCl2 ⋅6H2 O, and 2,5-thiophenedicarboxylic acid (H2 TDC) under a basic condition to confine them (Figure 2.38b) [68]. In the structure of Ni24 (L5)6 (TDC)12 (H2 O)6 , three tetranuclear units are bridged by three

47

48

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

(a)

Bridging-ligand substitution

(b)

O O

O edb

O

VMOP

(c)

Synthon A

Synthon B

Synthon C

VMOP-β

VMOP-γ

(d)

VMOP-α

Figure 2.36 (a) Ball-and-stick view of the hexavanadate cluster that corresponds to a trigonal vertex. (b) A VMOP built from four [V6 (SO4 )] clusters and six edb ligands with the dimension of 26.34 Å, which can be simplified as a truncated tetrahedron. (c) Three types of supramolecular synthon. (d) The diverse superstructures formed by the packing of truncated tetrahedra. Source: Gong et al. [65]. © 2019, Wiley-VCH.

2.3 Metallacages

hν Ru(bpy)32+ O OH

Co2+

Co2+

2+

*Ru(bpy)3 H2A

e–

e–

Co+

Ru(bpy)3+

Co+ O OH

H2A+

Figure 2.37 The construction of the host–guest charge-transfer systems for the combined light-driven proton reduction and biomimetic hydrogenation of nitrobenzene. Source: Zhao et al. [66]. © 2017, Wiley-VCH. (a)

R

(b)

X

+ O

Y

M HO

S

O OH

Figure 2.38 (a) Calixarene-based tetranuclear cluster building block possessing four points of extension, (b) assembly of trigonal prismatic coordination cage from 2,5-thiophenedicarboxylic acid. Source: Kumagai et al. [67, 68]. © 1997, Elsevier, © 2016, American Chemical Society.

TDC ligands into a coordination triangular framework and two such triangles are pillared by three pairs of TDC ligands to form a trigonal prism. The nanocage has an inner cavity of about 12.1 × 12.1 × 9.6 Å3 measured between opposite μ4 -oxygen atoms of the coordinated water molecules, featuring five apertures on the side facets and two bases. 3D polyimidazolium cages may act as receptors for advanced applications, and it has been suggested that they can be considered a new generation of cryptands for the encapsulation of anions. However, the method of preparing large 3D polyimidazolium derived from conventional strategy, based on the cyclization reaction of imidazole precursors with di(bromomethyl)benzene derivatives or related halogenated alkanes, is found to be unsuitable. In 2018, Han’s group employed a template-controlled strategy to accomplish large 3D polyimidazolium cages [69]. Given the simplicity and high efficiency of the described procedure, this strategy could be used in the future particularly for large-scale preparation of biomedicals and functional materials.

49

50

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

They also reported the narcissistic and social self-sorting during the formation of homo- and heteroligand poly-NHC metal assemblies [70]. The formation of donor–acceptor π· · ·π stacking interactions between the electron-rich and electron-poor ligand backbone groups was assumed to be responsible for the formation of the heteroligand assembly. Upon formation of the heteroligand assembly [Ag3 (A1)(D1)](BF4 )3 , a donor–acceptor interaction between the two central building blocks could be formed via aromatic stacking. This led to an additional stabilization for [Ag3 (A1)(D1)](BF4 )3 . An equimolar mixture of the imidazolium salts H3-A1(BF4 )3 , H3-B1(BF4 )3 , H3-C1(BF4 )3 , and H3-D1(BF4 )3 reacted with a slight excess of Ag2 O in acetonitrile at 70 ∘ C for two days in the dark. The reaction mixture containing three trinuclear silver(I) assemblies was obtained (Figure 2.39).

2.3.5

Cubes

Coordination-driven self-assembly acts as a powerful chemical approach for the construction of a series of 3D structures with precise geometries and sizes to mimic nature’s activities. Up to date, many previous studies of tpy–M(II)–tpy only concentrated on linear and 2D structures. But few 3D supramolecular cages and prisms were reported using tpy-based building blocks. Giant 3D highly symmetric cubes [M12 L8 ] were self-assembled using three-armed tpy-based ligands constructed on adamantine, with the ligands acting as directing unit in the vertices and metal ions as edge gluing elements (Figure 2.40) [71]. In 2018, Cui’s group selected two TPE-derived tetraamines as ligands for subcomponent self-assembly of octanuclear Zn8 L6 cages with tunable cavity sizes (Figure 2.41) [72]. The high π-electron density of assembled walls on the coordination cages endow them with prominent catalytic activity to promote a cascade condensation and cyclization to produce nonplanar 2,3-dihydroquinazolinones. Notably, the reaction is highly efficient with high rate enhancements (up to kcat/kuncat = 38 000) and multiple turnovers compared to the bulk reaction mixture. Sun’s group reported a sensitive structural switching phenomenon during the stereocontrolled self-assembly of a group of Ln2n L3n (Ln for lanthanides, L for organic ligands, and n = 1, 2, 4) compounds (Figure 2.42) [73]. This is the first time the boundary between the helical complex and the tetrahedral switches clearly revealed by using the eutectic structure. Three pairs of enantiopure ligands were synthesized by stepwise amide formation reactions according to an established method starting from dimethylpyridine-2,6-dicarboxylate, where the peripheral chiral amide groups were introduced first, followed by coupling of the central diamine spacers. Furthermore, M8 L12 -type supramolecular cubes were obtained for the first time by taking advantage of a concentration-dependent self-assembly process. Researchers have also focused on how to assemble a desired cage with a certain geometry and size and to explore its inner cavities (e.g. host–guest chemistry). However, approaches for organizing cages into more complex or architecturally controlled frameworks through supramolecular interactions to achieve new

+

+

H3-A1(BF4)3

H3-B1(BF4)3

+

H3-C1(BF4)3

H3-D1(BF4)3

CH3CN Ag2O 70 °C, 2 d

(BF4)3

(BF4)3

[Ag3(A1)(D1)](BF4)3

Figure 2.39

(BF4)3

[Ag3(B1)2](BF4)3 versus

(BF4)3

+

+

+

Heteroligand assembly Social self-sorting

(BF4)3

[Ag3(C1)2](BF4)3

Homoligand assemblies Narcissistic self-sorting

[Ag3(A1)2]3+

[Ag3(D1)2]3+ Not observed

Self-sorting of organometallic assemblies from tris-NHC ligands with different backbones. Source: Wang et al. [70]. © 2018, Wiley-VCH.

52

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

Figure 2.40 Self-assembly process of [M12 L8 ]. Source: Wang et al. [71]. © 2014, The Royal Society of Chemistry. (a)

H2N

NH2

6

O

+ 24 H2 N

L1

N

+ 8 Zn(OTf)2

NH2

16+

16OTf–

= Zn2+ (b) H2N

NH2 O

+ 24

6

H 2N

L2

N

+ 8 Zn(OTf)2

NH2

16+

16OTf–

Figure 2.41 Subcomponent self-assembly. (a) TPE-1 and (b) TPE-2. Source: Jiao et al. [72]. © 2018, Nature Publishing Group.

2.3 Metallacages

This Work

Previous Reports R

(i)

(ii)

HN

and

O N Lnlll

O

NH

HN

Lnlll O N O

(iii)

NH

(iv)

or

R

Figure 2.42 Self-assembly of L with Ln(OTf)3 . Source: Li et al. [73]. © 2017, American Chemical Society. O 12

NH

+ 6 H2N

NH2

N 8 Co2+

6 H2O

BF4–/CIO4–/PF6– (= )

(=

)

Figure 2.43 Illustration of the hierarchical self-assembly of the supramolecular framework. Source: Luo et al. [74]. © 2015, Wiley-VCH.

advanced functions are still unusual. Li’s group have developed the solvothermal subcomponent self-assembly of metal–imidazolate coordination cubes, which further assembled hierarchically with anions to yield supramolecular frameworks (Figure 2.43) [74]. Each hierarchical architecture was assembled from small components (metal ions, amines, aldehydes, anions, etc.) through multiple interactions including covalent bonds, dative bonds, and weak C–H⋅⋅⋅X (X = O, F, and P) hydrogen bonds. The supramolecular frameworks with giant cavities of mesoporous size are capable of taking up large molecules (an organic dye, metal coordination cages, or the biomolecule vitamin B12), and thus, a procedure has been developed in which aggregations give rise to a desired function. A mesoporous supramolecule constructed from large aggregates of ordered structures will open up a general avenue for using such similar strategies and provide a potential platform for obtaining advanced crystalline materials.

53

54

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

As a continuing study of metal–imidazolate cages and gyroidal metal–organic frameworks (MOFs) constructed using the subcomponent self-assembly technology under solvothermal reaction conditions, Li’s group successfully obtained a series of mesoporous supramolecular frameworks by the self-assembly of Co–imidazolate cages and inorganic anions (BF4 − , ClO4 − , and PF6 − ) through unconventional C–H⋅⋅⋅X (X = O, F, and π) hydrogen bonds. The combination of nickel(II) salts, 5-methyl-4-formylimidazole, m-xylylenediamine, and judiciously chosen inorganic anions resulted in a series of cages or cage-based products featuring diverse structures ranging from discrete cages to 2D sheets [75]. They anticipated that plenty of coordination polymers or MOFs with interesting structures and advanced functions can be constructed by this technology and approach. In 2018, Zhang’s group reported a Ti8 L12 cube, which could be further applied to encapsulate an octahedral [Ti(H2 O)6 ] species based on host–guest interaction. The obtained [Ti(H2 O)6 ]@Ti8 L12 complex was demonstrated to be stable both in solid and in solution states by X-ray diffraction and ESI-MS analysis (Figure 2.44) [76]. Compared with Ti8 L12 cube, [Ti(H2 O)6 ]@Ti8 L12 complex exhibits slightly longer lived excited state due to the influence of Ti(H2 O)6 species. This work provides an interesting supramolecular model for the host–guest photophysical study.

2.3.6

Octahedron

MOPs constructed through the coordination of metal ions and organic linkers often display unique molecular recognition properties and glamorous enzyme-like reactivity due to their high symmetry, stability, confined microenvironment, and rich chemical/physical properties and have attracted considerable attentions. In 2008, Duan’s group developed a new strategy for the preparation of octahedral nanocages (Figure 2.45) [77]. Metal-variable isostructural octahedral nanocages with considerable stabilities were constructed using a C3 symmetric facial ligand with three tridentate coordinating sites and various metal ions. The disk-shaped ligand is composed of three quinoline groups by meta-substitution of a central benzene ring. The ligands were alternatively arranged onto four of the eight triangle faces of the S4 symmetric octahedron defined by the six metal ions which were coordinated with two planar tridentate N2 O chelators in a mer configuration. Each quinoline group acted as both chromophore and fluorophore, thus serving as an N2 O tridentate coordinating site for the amide groups, which were introduced as trigger sites to achieve efficient guest interactions and a consequently good signal response. Once guest was encapsulated and formed donor-type hydrogen bonds with the amide groups, the electronic distribution on the conjugated backbone of the ligand was perturbed. As a result, host–guest interactions affected the charge transfer associated with the quinoline units and led to significant changes in the optical properties. Appropriately modulating the quinoline groups to pyridine groups, the several lanthanide ions based octahedral nanocages were gained. Each metal was coordinated by two N2 O tridentate chelating units and two bidentate coordinating nitrate anions.

2.3 Metallacages

(a)

(b)

(c)

(d)

Figure 2.44 Structure of the free (a) Ti8 L12 cube and (b,c) [Ti(H2 O)6 ]@Ti8 L12 . (d) Selected strong host–guest hydrogen bonds inside the cube of [Ti(H2 O)6 ]@Ti8 L12 . Source: Zhu et al. [76]. © 2018, Wiley-VCH.

In particular, when using paramagnetic Gd3+ as metal node, each Gd3+ ion was coordinated to two tridentate chelators from two different ligands and one disordered nitrate anion and one disordered water molecule, which possibly led to the increase in the proton relaxivity [78]. The addition of glucosamine caused a significant decrease in the longitudinal relaxivity due to the decreased number of bound water molecules replaced by the coordination of glucosamine. Thus, such a robust gadolinium-based metal–organic octahedron achieved high selective and sensitive MRI responses toward glucosamine. Except for the above octahedrons with a single metal as center or single ligands as edges, there are other octahedrons with special structures, such as POV-based octahedrons, heterometallic coordination octahedrons, and so on. V-based MOPs exhibit beautiful structures. Since 2016, Wang’s group has reported a series of

55

56

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

N N NH

N

O N

O

N H Metal-tunable functional site and signaling unit HN

O

N Guest interaction site and communicator N

Window

C3-symmetric facial ligand Metal ion CoII, NiII, ZnII, EuIII, TbIII

Figure 2.45 Structure of the ligand H3 L1 and constitutive/constructional fragments of the functional octahedral cages. Source: He et al. [77]. © 2008, Wiley-VCH.

excellent structures built by POVs. They successfully synthesized three isostructural POV-based MOP under solvothermal conditions [79]. These compounds were characterized by single-crystal X-ray diffraction (SCXRD), PXRD, IR, TGA, and N2 adsorption. By the same logic, VMOP-4, VMOP-5, and VMOP-6 have been synthesized by employing different carboxylate ligands to assemble with vanadium–oxygen clusters (Figure 2.46) [80]. Though all of the three hybrids feature the same [VV VIV 4 ] units, their structures exhibit differences changing from truncated triangular prism to truncated quadrangular prism to the octahedron, mainly depending on the nature of carboxylate ligands. By introducing a tritopic tricarboxylate ligand of 1,3,5-benzenetricarboxylic acid (BTC) or 4,4′ ,4′′ -benzene-1,3,5-triyl-tribenzoic acid (BTB), discrete calixarene-based coordination cages (Figure 2.47) were successfully synthesized through the self-assembly of metal ions, thiacalixarenes or sulfonylcalixarenes, and tricaboxylate

{V5O9CI}

2,5-H2 TDA (a)

2,6-H2 NDC

m-H2BDC (b)

(c)

Figure 2.46 Schematic representation of the preparation of VMOP-4 (a), VMOP-5 (b), and VMOP-6 (c), highlighting the ligand-directed structure diversities. Source: Zhang et al. [80]. © 2016, American Chemical Society.

58

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

ligands [81]. They adopted a octahedral geometry where six [M4 (calix)(μ4 -Y)] tetranuclear units were located on the vertices and eight tricarboxylate ligands resided on the triangular facets. Each [M4 (calix)(μ4 -Y)] unit was bonded by four tricarboxylate ligands through four quadrangle metal borders, while each tricarboxylate ligand was bonded to three [M4 (calix)(μ4 -Y)] units with its three carboxyl groups. The internal cavities of cages could be significantly increased from 0.55 to 2.75 nm3 when the tricarboxylate ligand was changed from BTC to BTB elongating with a phenyl ring, while the peripheral diameters of the cages were tunable from 3.3 nm for the BTC-bridged cages to 4.7 nm in BTB-bridged cages (Figure 2.47). In addition to the abovementioned face-directed octahedral cages formed by the [6+8] condensation of [M4 (calix)(μ4 -Y)] units and tricarboxylate ligands, cages mimicking the shape of an edge-directed octahedron were also feasible when linear dicarboxylate bridging ligands were employed. Hong and coworkers [82] and Wang and coworkers [83] introduced 1,4-benzenedicarboxylic acid (1,4-BDC) ligands into this system and obtained a new edge-directed octahedral coordination cage from the self-assembly of cationic tetranuclear-metal building blocks and linear anionic ligands (1,4-BDC) through a [6+12] condensation process. Six [M4 (calix)(μ4 -Y)] units as the vertices were connected by 12 1,4-BDC ligands locating on the octahedron edges. The outer diameter of this type of cages was comparable to that observed in the BTC-bridged face-directed octahedral cages. However, they featured a larger internal cavity volume that is twice as big as that of corresponding BTC-bridged cages, possessing even wider apertures than those in BTC-bridged cages. Recently, Cui’s group further utilized a cage extension strategy to construct chiral permanent porous hydrogen-bonded frameworks (HOFs) (Figure 2.48) [84]. Multiple noncovalent interactions including hydrogen bonds and hydrophobic interactions between tert-butyl groups direct the hierarchical assembly of the cages into a permanent porous material. Moreover, the HOFs possess channels decorated with chiral phosphoric acid groups which make them as potential catalysts to promote [3+2] coupling of indoles with quinone monoimine and F–C alkylations of indole and aryl aldimines with high enantioselectivities. Porous coordination cages (PCCs) are nanoscopic structures assembled by metal clusters and organic linkers, which bear intrinsic porosity in both solid and solution states. In 2018, Zhou’s group reported a series of PCCs synthesized by the solvothermal method [85]. PCC-1 and PCC-2 hold similar carboxylic acid ligand and transition metal cluster. The difference between them comes from the charge. For PCC-2, anionic sulfate groups (SO3 − ) replaced tert-butyl groups (tBu) of PCC-1, making the whole structure negatively charged. And PCC-3 cage was synthesized from pyridine ligand and noble metal knot (Figure 2.49). This cage shows distinct positive net charge and hydrophilic surface. These PCCs show different subcellular distributions as a result of their varied charge and affinity features. In addition, the PCC-2 cage was applied to study catalytic activity [86].

3.3 nm

CIAC-101 COO– 3.7 nm –OOC

COO– CIAC-102

COO–

4.4 nm

–OOC

COO–

CIAC-103

4.7 nm

CIAC-104

Figure 2.47 Assembly of octahedral calixarene-based coordination cages from [M4 (calix)(μ4 -Y)] and tricarboxylate ligands. Source: Dai et al. [81]. © 2012, American Chemical Society.

(a)

(b)

H4L

M4-TBSC (M = Ni/Co)

(c)

Figure 2.48 (a) Self-assembly of cages 1-Ni and 1-Co (only half of the HL3− ligand is shown in the cage for clarity). (b) The single-crystal structure of the octahedral cage in 1-Ni. (c) The space-filling model with an elliptical shape viewed along the short axis (sky-blue, Ni; green, P; yellow, S; gray, C; red, O). The cavities are highlighted by colored spheres. Source: Gong et al. [84]. © 2019, Nature Publishing Group.

2.3 Metallacages

(a)

V1

V2

V3

L1

L2

L3

PCC-1: PCC-2: PCC-3: (b)

(c)

(d)

Figure 2.49 Schematic structure and X-ray structure of PCCs. (a) Cartoon of octahedron cage PCC and the cage components. Single-crystal X-ray crystal structure of (b) PCC-1, (c) PCC-2, and (d) PCC-3. Source: Fang et al. [85]. © 2018, Wiley-VCH.

2.3.7

Dodecahedron

Natural systems are capable of fabricating supramolecular ensembles with complexity and diversity via a spontaneous self-assembly protocol. The polyhedral coordination cage 1 with O symmetry and formulated as [Ni14 L24 ]⋅4NO3 ⋅xguest [HL = N-methyl-1-(4-imidazolyl)methanimine] was obtained by the solvothermal reaction of 24 L and 14 Ni2+ or the one-pot assembly of 62 subcomponents, including 24 methylamine, 24 4-formylimidazole, and 14 Ni2+ (Figure 2.50) [87]. The 62-component assembly involved synchronized formation of both dynamic covalent bonds and coordination bonds, demonstrating the vigorousness of the subcomponent self-assembly technique in complicated supramolecular self-assembly involving a large number of units. There are two types of Ni2+ centers in the nonsymmetric unit (Figure 2.50b): Ni1 adopts an octahedral geometry chelated by three L (located on the threefold axis), and Ni2 adopts a square geometry bound by four L (located on the fourfold axis). It is notable that the square Ni2 is unsaturated and probably acts as an active site for further attack by other molecules, providing the possibility of postmodification of the polyhedron. Li’s group also described the control of assembly for a series of neutral cubic nickel(II)–imidazolate cages formulated as Ni8 L12 X4 (1) by the variation of substituents and anions [88].

61

62

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

(a)

(b)

Method 1

14

Ni1

+ 24

Ni2 Method 2

14

+ 24

+ 24

= Ni2+

–H2O

HN

– NH2 =

=

N

N HN

O

=

C H N Ni

N

Figure 2.50 (a) Self-assembly of 1 by two methods. (b) Single-crystal structure of 1. Source: Zhou et al. [87]. © 2012, American Chemical Society.

The structures of this multicomponent system are sensitive to both substituents and anions. A change of a substituent of imidazole ligand or anion switches the final structure observed across the ensemble of building blocks between cubic Ni8 L12 X4 and dodecahedral Ni14 L24 (4) cages. Cubic cage 1 can transform to a rhombic dodecahedral cage 4 in the presence of methylamine molecules. The transformation is irreversible and accompanies an obvious color change. The cage-to-cage transformation involving the stereoelectronic preference of metal ions and color is discovered, which might provide a unique example to understand the supramolecular dynamic transformation process. Heterometallic coordination cages incorporating multiple metal centers have attracted increased attention on account of their novel properties and advanced function. In 2014, Su’s group constructed nanosized heterometallic metal–organic cages [Pd6 (RuL3 )8 ]28+ (MOC-16) from photoredox-active and stereogenic Ru(II)-metalloligand and naked Pd(II) ion using a stepwise assembly approach (Figure 2.51) [89]. The metalloligand RuL3 possesses spatially triangular C3 symmetry, which was prepared by treating RuCl3 with 2-(pyridin-3-yl)-1H-imidazo[4,5-f ][1,10]phenanthroline (L) in a microwave method, while the naked Pd(II) has coplanar-squared D4 symmetry. Crystal structures of the cages indicate that the cages show the shape of a rhombic dodecahedron defined by Pd6 Ru8 centers and have 12 rhombic boxlike windows, with 6 Pd occupying the vertices of a truncated octahedron and RuL3 lying on each face. The overall cage size was estimated to be 3.1 × 3.4 × 3.4 nm3 , possessing six Pd vertices with separations of 27.9, 29.5, and 29.5 Å and a large cavity of 5350 Å3 . In this work, formation of racemic Δ/Λ-Pd6 (RuL3 )8 (rac-Δ-/Λ-MOCs-16) cages from mixed precursors has been accomplished. Large hydrophobic cavity and relatively narrow windows of the MOC-16 cage were capable of a maximum of 18 phen guests inclusion, as well as 12 pyrene or anthracene molecules encapsulation. Later in 2016, they developed a general approach to assemble enantiopure Δand Λ-Pd6 (RuL3 )8 cages (Δ-/Λ-MOCs-16) from pre-resolved Δ-3 and Λ-3 metalloligands [90]. This is the first example of constructing enantiopure MOCs from predetermined chiral metalloligands. X-ray crystallographic analysis verified the formation of absolute ΔΔΔΔΔΔΔΔ and ΛΛΛΛΛΛΛΛ homoconfigurations in Δ-MOC-16 and Λ-MOC-16, respectively. Stereoselective separation of two types of racemic organic guests has been examined by 1 H NMR enantiodifferentiation

2.3 Metallacages

N N

N

N

NH

N

N

RuCl3-3H2O N Micorwave 400 W, 190 °C

H N N

N NH

2+

N N Ru N N N N

HN

L

2+

RuL3

+ N

Pd

N

N N N N Pd N

N

28+

HN N N

N Ru N N N

N HN

Guest

N N Pd N N

N

N

HN N

Pd

NN

Pd N

N

N

Pd

N N

N

N N Pd N N

Guest ⊂ cage

Pd6(RuL3)8 MOC

Figure 2.51 Stepwise assembly of Pd6 (RuL3 )8 metal–organic cages via metalloligand strategy. Source: Li et al. [89]. © 2014, American Chemical Society.

experiments of host–guest diastereomers, which demonstrated better stereoselectivity toward chiral compound with C2 symmetric chirality than that containing C* stereocenters. In the same year, photocatalytic H2 production has been explored using this metal–organic cage MOCs-16, containing eight Ru2+ photo-centers and six catalytically active Pd2+ -centers [91]. The efficient hydrogen production might derive from the directional electron transfers through multiple channels owing to proper organization of the photo- and catalytic multi-units within the octahedral cage. Since enantiopure Δ/Λ-MOCs-16 cages can provide achiral coordination space for chiral guest recognition and separation, such active MOCs could serve as dual-functional photoredox- and stereochemically active nanoreactors for stereoselective photoreactions. In 2017, they reported an unprecedented photoinduced regio- and enantioselective coupling of naphthols and derivatives using this RuII metalloligand-based cage (Figure 2.52) [92]. Naphthol guests encapsulated in the confined chiral coordination space undergo a regiospecific 1,4-coupling, rather than the normal 1,1-coupling, to form 4-(2-hydroxy-1-naphthyl)-1,2-napthoquinones. This unusual dimerization constitutes a very rare example of asymmetric induction in biaryl coupling by making use of coordination cages with dual functionality-photoredox reactivity and stereoselectivity. Recently, Su’s group expanded this stepwise strategy to the stereolabile D3 symmetry tris-chelate-Fe-type metalloligand, demonstrating the successful assembly of configurationally stable [Pd6 (FeL3 )8 ]28+ (Δ/Λ-MOCs-42) homochiral octahedral cages via strong face-directed stereochemical coupling and facile chiral-induced resolution processes based on stereodifferentiating host–guest dynamics [93]. Excellent enantiopurity of Δ/Λ-MOCs-42 is achieved via an in situ assembly process for

63

64

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

N N N

Blue LED N

N

N

N H

Ru N

N Pd

OH

N Pd

N NH

Br

3 Photo-hydrogen-evolving (PHE) unit

N N N Ru N N N

HN N

Photo-active Redox-active Chiral-confined

N Pd

N N H

Pd

Regioselective enantioselective

N Pd

O O Br OH

Self-active coordination space Pd

[Pd6(RuL3)8]28+ Metal–organic cage (Δ-MOC-16)

Br

S-4

Figure 2.52 The photoinduced biaryl coupling of 3-bromo-2-naphthol to give its 3-bromo derivative proceeds with unique selectivity using homochiral MOCs-16. Source: Guo et al. [92]. © 2017, Wiley-VCH.

enantioresolution of rac-MOCs-42, of which the chiral-induced precipitation leads to kinetic resolution, while cocrystallization leads to thermodynamic resolution, both are superior to the normal post-assembly process. Recyclable enantioseparation of racemic atropisomers has been successfully accomplished, giving good ee value. Cui’s group utilized privileged ligand (Mn(salen)-derived dicarboxylic acids) as linear linkers and six Zn4 -p-tert-butylsulfonylcalix[4]arene clusters as tetravalent four-connected vertices to synthesize a chiral octahedral cage 1, which contained a large hydrophobic cavity of about 3944 Å3 decorated by catalytically active metallosalen moieties (Figure 2.53) [94]. In the single crystal, six shuttle-like [Zn4 (μ4 -O)(TBSC)] sit on the vertexes of the octahedron, and each Mn(salen) ligand acts as the edge bridging two vertices. Notably, this cage could efficiently catalyze the oxidative kinetic resolution (OKR) of racemic secondary alcohols with ee value up to 99.3%. The strategy of incorporating metallosalen into rigid polyhedral cage not only provides a hydrophobic cavity for concentrating reactants but also avoids the deactivation of catalytically active Mn(salen), resulting in obviously enhanced catalytic activities and enantioselectivities. Similarly, a series of asymmetric metallosalen catalysts were incorporated into a highly stable Zr-based UiO-68 MOF via ligand exchange strategy. They demonstrated that MOFs with potentially acid-labile chiral catalysts could be synthesized via postsynthetic exchange [95]. This could not be synthesized through direct solvothermal synthesis. The single-M(salen)-linked MOFs display active and efficient catalytic activity for asymmetric cyanosilylation of aldehydes, ring opening of epoxides, OKR of secondary alcohols and aminolysis of stilbene oxide, and the mixed-M(salen) linker variants are active for sequential asymmetric alkene

2.3 Metallacages

(b) (a)

Figure 2.53 (a) Structures of the Mn(H2 L) linker and calix[4]arene-based metal node. (b) Single-crystal X-ray structure of the octahedral cage in 1 (brown, Zn; green, Mn; blue, N; red, O; purple, S; gray, C). For clarity, the coexisting Zn4 clusters and the coordinated DMF and H2 O molecules in 1 are not shown. The cavity was highlighted by a yellow polyhedron. Source: Tan et al. [94]. © 2018, Wiley-VCH.

epoxidation/epoxide ring-opening reactions. Furthermore, the chiral MOF catalysts are highly enantioselective and completely heterogeneous and recyclable, which make them attractive catalysts for eco-friendly synthesis of fine chemicals.

2.3.8

Cuboctahedrons

In 2010, Zhou’s group reported a porous coordination cuboctahedron nanocage constructed from the Cu paddlewheel clusters and bridged by the isophthalate moieties of 5-((triisopropylsilyl)ethynyl) isophthalic acid (TEI) (Figure 2.54) [97]. The CuTEI cage has eight triangular and six square windows. The bulky triisopropylsilyl (TIPS) groups left outside are highly likely to function as gates by undergoing thermal vibration within CuTEI, giving this material the thermosensitive gate opening property. The discrete CuTEI molecule tends to move around due to the lack of a strong holding force. Such movement leads to an amorphous structure. As a result, channels and openings toward the inner void in the perfect crystal model are partially reduced and blocked by the close packing of adjacent nanocages and extruded TIPS groups in the activated sample, yielding abundant “kinetically closed pores.” By grafting with azide-terminated polyethylene glycol (PEG) through “click chemistry” or adhering to each other, they made the surface of the CuTEI cage functional [96].

2.3.9

Hexadecahedrons

Comparing to tricarboxlate ligands, 5-(pyridin-4-yl)isophthalate (PIP) ligand is also a tritopic ligand that is still able to connect metal ions in three directions but has a lower symmetry (C2v ). Liao and coworkers reported a novel J 17 Johnson hexadecahedronal coordination cage, in which 10 calixarene-based clusters are located on vertices and 16 PIP ligands are served as the triangular tiles (Figure 2.55) [98]. It is constructed by pillaring two square pyramids with a square antiprism. The 10 Ni4 (L1)(μ4 -Cl) units adopt three types of coordination modes: (1) two

65

66

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

(a)

(b)

Si

HOOC

COOH

Figure 2.54 (a) Ligand TEI. (b) The porous coordination nanocage CuTEI. Source: Liu et al. [96]. © 2011, Wiley-VCH.

N

HOOC

COOH

Ni S CI C N O

Figure 2.55 Assembly of J17 Johnson hexadecahedronal coordination cage from 5-(pyridin-4-yl)isophthalate ligand. Source: Hang et al. [98]. © 2016, American Chemical Society.

Ni4 (L1)(𝜇 4 -Cl) units have the same coordination modes through chelated coordination to four –OCO– groups from four PIP ligands, forming the similar tetranuclear units established in the octahedral cages. Therefore, each nickel atom is surrounded by two oxygen atoms and one sulfur atom in L1, one chlorine ion, and two oxygen atoms from two adjacent PIP ligands. (2) For another four Ni4 (L1)(μ4 -Cl) units, each chelated to three –OCO– groups from three PIP ligands with two pyridines in two

2.3 Metallacages

PIP ligands coordinating to two nickel atoms, fulfilling the octahedral coordination geometry of each nickel atom. (3) In the remaining four tetranuclear units, there are only five PIP ligands coordinated to four nickel atoms, in which two –OCO– groups bridge three nickel atoms, two pyridines coordinate to two nickel atoms separately, and one carboxylate group of the PIP ligand chelates a nickel atom.

2.3.10 Barrel-Shaped Cages 2.3.10.1 Calixarene Constructed Barrel-Shaped Cages

A novel calixarene-based wheel-like coordination cage, [Ni18 Cl6 (L1)6 (MNA)6 ] (H2 MNA = 2-mercaptonicotinic acid), was successfully synthesized by Liao and coworkers in 2018 (Figure 2.56) [99]. A unique shuttlecock-like Ni3 (L1)(μ3 -Cl) building block (Figure 2.56a) is formed by the coordination of an L1 to three nickel atoms through four phenolic oxygen atoms and three sulfur atoms and the three nickel atoms are further capped by a μ3 -Cl anion. The wheel-like entity consists of six Ni3 (L1)(μ3 -Cl) units bridging by six MNA ligands, forming a hexagonal prismatic inner cavity. The nickel ions are six coordinated by two phenoxy oxygen (a)

(b)

Ni S CI O N C (c)

(d)

Figure 2.56 A shuttlecock-like Ni3 (L1) building block capped by a Cl− anion (a) and crystal structure of calixarene-based coordination wheel (b–d). Source: Wang et al. [99]. © 2018, American Chemical Society.

67

68

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

atoms, one μ3 -Cl, one sulfur atom from L1, and two oxygen atoms or one sulfur and one nitrogen or one sulfur and one oxygen of a same MNA ligand. The dimensions of the coordination cage are approximately 25.7 × 27.5 × 15.1 Å3 , and the size of the hexagonal prismatic pore is about 9.0 × 9.0 × 13.3 Å3 . Besides the linear dicarboxylate ligand, the angular dicarboxylate ligand is also applied to construct calixarene-based coordination cages. When the linear 1,4-BDC ligand was replaced by 1,3-benzenedicarboxylic acid (1,3-BDC), barrel-shaped coordination cages were isolated [100]. They can be thought to form by cutting two opposite vertices of the BTC-bridged coordination octahedron, consisting four tetranuclear units that are bridged by eight 1,3-BDC linkers. Therefore, the internal cavity features two wide opposite square windows. Calix[4]resorcinarene is another important subset of calix[4]arenes, featuring a bowl-shaped aromatic cavity that can be easily modified with various functional groups. Yang and coworkers demonstrated a design strategy for the assembly of metal-coordinated calix[4]resorcinarene cavitands and cages [101]. First, two new calix[4]resorcinarene cavitands with four chelating 2-(4H-pyrazol-3-yl)pyridine moieties at the upper rim were designed. The assembly of the calix[4]resorcinarene cavitands (TPC4R) with zinc ions and isophthalic acid analogs (Figure 2.57a) provided a series of intriguing isostructural metal coordinated cavitands [Zn4 (TPC4R)(1,3-BDCs)4 ] with highly deep cavities ranging from 10.7 to 12.5 Å. Each zinc atom is chelated by a 2-(4H-pyrazol-3-yl)pyridine group and two carboxylate groups from two 1,3-BDCs liners. By introducing tetracarboxylic acid analogs with appropriate spacers between two isophthalic acids (Figure 2.57a), a family of coordination cages was successfully achieved. The [Zn8 (TPC4R)2 (Ln)4 ] nanocages were constructed by connecting two [Zn4 (TPC4R)] units with four semiflexible tetracarboxylic acids (Figure 2.57b). The internal cavities of these coordination cages are tunable in shape and size by simple modification of the tetracarboxylic acids. With the increasing length of the tetracarboxylic acids, the internal cavities were changed from the spindle shape to ellipsoid shape with the longitudinal dimensions increasing from approximately 2.3 to 3.0 nm. 2.3.10.2 Dimetallic Clips-Constructed Barrel-Shaped Cages

In 2005, Yu et al. reported a series of nanosized cavity-containing metallomacrocycles by the self-assembly of the bipyrazolate ligands and dimetal (Pd/Pt) centers in aqueous solution [102]. During the assembly, the pyrazole spontaneously deprotonated and coordinated to the dimetal (Pd/Pt) centers and the pH value of the solution sharply decreased. However, these complexes exhibit high kinetic stability in acidic aqueous solution due to the strong bonding between the dimetal centers and the anionic ligand. Next, many fascinating 2D and 3D supramolecular architectures have been built based on these dimetallic clips [103]. In view of these highly stabile dimetallic clips, they constructed homo- or hetero-metallomacrocycles and organo-heterometallic cages through a programmable, stepwise self-assembly process [104].

N

(a)

N N

O

N O N

N O

R

R

R

O N

O N ON O

OH

HO

N

HO

OH O

TPC4R

O

OH O

O

Isophthalic Acid Analogs O

O

HO

O

O OH

O

HO

O

NH2

N

O

OH

HO

R = Me, Pen

O R

O N

(b) N

HO

OH

O

O

OH O O

O

HO

OH O

O

O

HO

OH O

OH

HO

O

O

O

Tetracarboxylic acid analogs

Figure 2.57 (a). Structures of TPC4R, isophthalic acid analogs, and tetracarboxylic acid analogs. (b). Structures of metal-coordinated cavitands [Zn4 (TPC4R)(1,3-BDCs)4 ] (top) and [Zn8 (TPC4R)2 (Ln)4 ] cages (bottom). Source: Pei et al. [101]. © 2017, American Chemical Society.

70

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

+2

2

1 (i) (a)

(b)

3 (ii)

PdII

(iii)

4a

(iii)

4

5

Figure 2.58 Representation for the hierarchical self-assembly of segmental (4a) and continuous (5) MMONTs starting from dimetallic clips 1 and ligand 2. Source: Zhang et al. [105]. © 2015, Wiley-VCH.

Based on the similar dimetallic Pd clip, Sun’s group reported the rational bottom-up self-assembly of discrete hexagonal nanobarrels 4 formulated as [Pd30 L1 24 L2 24 ] which consists of as many as 78 components (Figure 2.58) [105]. They reported the crystallization-driven conversion into segmental or continuous mesoporous MONTs (MMONTs) with a large out-diameter of up to 4.5 nm and a channel size of 2.4 nm. An unprecedented crystallization-driven cross-linking between discrete nanobarrel building units by spontaneous loss of the capping ligands to form infinite nanotubes was observed. Formation of either segmental or continuous MMONTs can be easily controlled by choosing appropriate starting materials.

2.3 Metallacages

2.3.11 Multiple Structural Cages Although coordination patterns or chelating moieties are the same, there may be different structural cages achieved because of ligands with different length, angle, and so on. Some ligands assembling with metal centers can also form multiple structural cages in different conditions. It may be possible to successfully realize their transformations among different structures in replying external conditions. When researchers need to synthesis highly symmetric coordination polyhedra, the dinuclear paddlewheel unit is a popular building block [106]. The combination of a ditopic bridging ligand with 120∘ bend angle and such a dinuclear paddlewheel unit leads to a coordination cuboctahedron. In the same way, a 90∘ bend angle bridging ligand gave rise to a coordination octahedron. With the utilization of different bend angle ditopic bridging ligands and dinuclear metal nodes, coordination polyhedra with an odd number of faces and an odd number of vertices can be assembled. Cage structures constructed by the low-coordination-number Cu+ ion have been rarely explored, probably due to the redox instability of Cu+ at ambient atmosphere. In 2009, Su’s group reported the self-assembly of copper(I) cuboctahedral coordination cages with distinct counter-anions of varied shapes and sizes, which were constructed from the L ligand possessing a rigid triangular N3 plane and the trigonal Cu+ ion providing another kind of CuN3 triangular face (Figure 2.59) [107]. CuI 4 L4 cage was considered to be a truncated tetrahedron with trigonal Cu+ centers at the apices and triangular L ligands at the faces. These Cu+ cages show host–guest-dependent redox activity. Complexes 1–2a are stable and redox inert because of good adaptability of spherical ClO4 − and I− to the cuboctahedral cavity, thereby stabilizing the cage and protecting the four Cu+ ions against O2 attack. In contrast, complexes 3–4a with linear CF3 SO3 − or planar MeC6 H4 SO3 − are redox active and Cu+ could be slowly oxidized to Cu2+ complexes with concomitant hydroxylation of the ligand, resulting in the structural conversion to the dinuclear complex 3b and the tetranuclear complex 4b. As we can imagine, limited by geometrical constraints, a series of discrete M3n L2n with different sizes of internal cavities may be obtained by changing the length, (a)

(b)

(c)

(d)

Figure 2.59 (a) Representation of the [CuI 4 L4 ]4+ cages in 1–4a showing a rectified cube. Benzyl groups and hydrogen atoms are omitted for clarity. The internal cavity is indicated by a yellow ball, and rectification is demonstrated by trigonal Cu+ coordination plane and central benzene plane. (b) Crystal structure of 1a showing tetrahedral arrangement of Cu+ ions and ClO4 − guest anion as a space-filling model. (c) Crystal structure of 3b. (d) Crystal structure of 4b. Source: He et al. [107]. © 2009, Wiley-VCH.

71

72

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

(a)

M3L2

M6L4

M12L8

M30L20

(b) 4 S

N

N

N

N

1 N

S

N N

N N

N NN

8x 12 x Zn2+ Complex A |Zn12LA6|

N

N

N

S

N

4x 6 x Zn2+ Complex B |Zn6LB4|

N

N

3

S

N N N

LA

8x

N

2

N N

N N N

N N

N

N N N N

LB

N N

LC

2x 3 x Zn2+ Complex C |Zn3LC2|

Figure 2.60 (a) Self-assembly of M3n L2n family with metals (M) as edges and tritopic ligands (L) as vertices, respectively. (b) Synthetic route of ligands and complexes Zn3n L2n . Source: Wang et al. [108]. © 2014, American Chemical Society.

angle, and rigidity of adamantane-based terpyridine ligand (Figure 2.60a). Thus, Li’s group successfully achieved the construction of the trigonal bipyramidal-like dimer (M3 L2 ), tetrahedron (M6 L4 ), and cubes (M12 L8 ) with the ligands as corner directing units and Zn2+ metal ions as edges by changing the angles of the linkers between adamantane and tpy (Figure 2.60b). Meanwhile, they also explored the behavior of self-sorting among the binary mixtures of LA and LC or LB and LC [108]. Chiral coordination cages feature both chirality and defined inner space, providing advanced molecular materials. However, such resulting cages tend to form racemic crystals or racemic conglomerates of homochiral crystals in the absence of chiral additive. Thus, Li’s group reported the subcomponent self-assembly of a series of emerging polyhedral metal–imidazolate cages (Figure 2.61), resulting from the self-assembly of 72 components featuring an unusual tetartoid (tetragonal

2.3 Metallacages

NH2

NH2 Co2+ R2

O

HN

N R1 = H

R1 = CH3 R2 = H or CH3 Compound: 1, 2, 3, 4, and 5

R1

R2 = H or CH3 Compound: 6

Figure 2.61 Subcomponent self-assembly of metal–imidazolate tetartoids and cubes. Source: Luo et al. [109]. © 2018, American Chemical Society.

pentagonal dodecahedron) geometry [109]. Spontaneous resolution of racemic tetartoidal cage into a conglomerate of homochiral crystals is confirmed by both SCXRD analyses and solid-state CD spectra. Additionally, enantiomerically pure ΔΔΔΔ-1 and ΛΛΛΛ-1 can be obtained by homochiral crystallization through chiral induction of (D)- and (L)-enantiomers of menthol, providing a way to prepare homochiral cages by chiral induction. The 2-methyl group on the imidazolyl plays a structure-direct role for forming the tetartoidal cages, while cubic cages are obtained in the absence of steric effect of the 2-methyl group. Bridging-ligand-substitution reaction and “polyhedral projection labeling” provide excellent synthetic and theoretical analysis strategies for the preparation and isolation of novel MOPs based on a square four-connected Cu2 (O2 CR)4 unit and various carboxylate ligands acting as building blocks [110]. In the nine kinds of MOPs as shown in Figure 2.62, we can divide them into two types. With metal nodes being viewed as vertices and ligands as edges, the cuboctahedron was found in MOPs 1, 2, 3, 5, 6 and 9, and octahedron was associated with 4 and 8 (Figure 2.62). The ligands can be arranged and linked properly to form new polygonal rings, with each corner considered as a ligand. Various stimuli has been used to control supramolecular structural transformations, including light, electricity, temperature, concentration, ligand and metal substitution, host–guest interaction, and so on. The Yuan’s group reported a special example of solvent-responsive structural transformation process between two V24 metal–organic nanocapsules (MONCs) (Figure 2.63) [111]. The geometries of these two nanocapsules are completely different, one displays an expanded ball-shaped molecular cage (V24 -ball) with the cavity of about 1400 Å3 , and another one presents an octahedral cage (V24 -oct) with the cavity of about 1000 Å3 . Interestingly, a solvent-controlled interconversion can be reversibly realized following the addition of different solvents. As the structural transformation, the quasi-isomers

73

74

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

Figure 2.62 Schematic of the syntheses and crystal structures of compounds 1–10. Source: Li et al. [110]. © 2010, Nature Publishing Group.

demonstrate obviously different magnetic properties. This research work provides a new strategy for the design and synthesis of the new isomers and quasi-isomers and informs the relationship between transformation process and structural property. In 2018, the first four examples of nanoscale Goldberg MOPs constructed by isolation of a pentagonal [WV5 O11 (SO4 )6 ]8− cluster reported by Wang’s group [112]. The largest Goldberg MOP-4 exhibits a diameter of 4.3 nm and can trap fullerene C60 molecules in its interstitial cavities.

2.3.12 Other Cages In 2014, Yu’s group reported a new class of gold(I)-containing metallasupramolecular cages [Au8 L2 ] (L = tetrakis-dithiocarbamato-calix[4]arene) featuring a quadruple-stranded helicate structure (Figure 2.64) [113]. The flexible polydentate bridging dithiocarbamate ligands functionalized on the upper rim of the bowl-shaped calix[4]arene have been employed to coordinate to four sets of dinuclear gold(I) units to form a 3D cage, [Au8 L2 ]. In the metal–organic gold(I)-centered

2.3 Metallacages

+

VOSO4

Interconversion

Figure 2.63 Controlled self-assembly and interconversions of V24 capsules. Chemical structure of hexameric pyrogallol[4]arene V24 octahedron and V24 ball from 6 PgC3 ligands and 24 vanadium ions. Vanadium is green, oxygen red, and carbon blue. Source: Su et al. [111]. © 2018, Nature Publishing Group.

(a)

(b)

Au N O S C H

Figure 2.64 AuI ⋅⋅⋅AuI bonding interaction driven the self-assembly from [AuI ⋅⋅⋅AuI ]· · ·coupled cages (a) to the cage-built 2-D [AuI ⋅⋅⋅AuI ] arrays (b). Source: Jiang et al. [113]. © 2014, American Chemical Society.

75

76

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

cage-like structure, the freely tunable calix[4]arene-based cavitands mainly adopt a favorable 1,3-expanded and 2,4-contracted conformation to meet the steric demand, leading to the formation of a quadruple-stranded helicate structure. Besides four AuI · · ·AuI interactions inside the dinuclear AuI units, two weaker interactions are also observed between the neighboring two dinuclear AuI units to hold the chiral helicate conformation. Interestingly, these functional supramolecular metalloorganic cages aggregate into 2D sheet-like molecular arrays in the solid state, consisting of 1D molecular wires with extended intra- and intermolecular AuI · · ·AuI bonding interactions. Moreover, small variations on the periphery lead to a dramatic difference in their 3D crystal packing mode. When phenylmethyl substituent is used, a zigzag linear packing mode has been observed in the crystal structure which is driven by the third type of AuI · · ·AuI interactions existing between the neighboring cages, leading to the formation of single-layer 2D molecular arrays composing of two kinds of Au(I) chains (Au8 and [Au8 ]∞ ) with continuous AuI · · ·AuI interaction along the a-axis. For a slightly larger phenylethyl substituent, no polymeric 2D molecular array has been observed in the crystal structure due to steric effect. However, a continuous AuI · · ·AuI interaction-directed 2D molecular array composing of two infinite chains of [Au8 ]∞ without any breakage has been observed when a 2-thienylmethyl substituent is used in the ligand. These layers are loosely packed along the a-axis. The remarkable packing differences in the complexes offer a useful empirical rule for the design of supramolecular gold(I)-containing solid-state materials. These novel gold(I) supramolecular cages exhibit green phosphorescence and have been shown to serve as highly selective proof-of-concept luminescent sensors toward AgI cation among various competitive transition-metal ions. In 2018, Sun’s group reported the self-assembly, photochromic, redox, and host–guest functions of a Pd4 L2 -type nanocapsule made of four cis-blocked palladium corners and two pyridinium-functionalized bis-bidentated ligands, which were synthesized from two TPT with a p-xylene linker (Figure 2.65) [114]. As a near relative of the previous Pd6 L4 cage reported by Fujita group, the Pd4 L2 cage kept highly +12 charged thus is highly water soluble. Moreover, introduction of

N +

N N

N

Redox-active

Pd N

N

N

Large cavity

N +

N

N N N

N N

Pd N

N N N

N

Pd

POMs encapsulation

N

N Pd N + N

Photochromism

Selective catalysis

+ N

N

Pd

N

Pd

Figure 2.65 Pd6 L4 cage and its properties. Source: Cai et al. [114]. © 2014, American Chemical Society.

References

pyridinium moieties not only enhances the electron-deficient nature of the TPT panels but also imparts photochromic and redox activities into the cage.

2.4 Conclusion In summary, we have discussed the recent advances in the field of supramolecular metallacycles and metallacages with particular architectures and functions. Rational design and selection of the building blocks and the orthogonality of the metal–ligand coordination are key factors in determining the final structures. The incorporation of stimuli-responsive scaffolds or catalytic group endows the systems with various functionalities. The inherent cavities of the macrocycles and cages have also been exploited for host–guest chemistry, separation, or catalysis. Ongoing exploration of new design strategies and optimization of the performances provide a strong outlook for the future development of metal–organic coordination macrocycles and cages and their applications in sensing, catalysis, biological systems, and yet unforeseen uses.

Acknowledgments This work was supported by the National Natural Science Foundation of China (Grant No. 21825107, 21801241, 21901245, 21673239), the Strategic Priority Research Program of the Chinese Academy of Sciences (Grant No. XDB20000000), the National Postdoctoral Program for Innovative Talents (BX20180305), and the China Postdoctoral Science Foundation (2018M642579).

References 1 2 3 4 5 6 7 8 9 10 11

Xie, T.Z., Guo, C., Yu, S.Y., and Pan, Y.J. (2012). Angew. Chem. Int. Ed. 51: 1177. Zhou, J., Zhang, Y., Yu, G. et al. (2018). J. Am. Chem. Soc. 140: 7730. Liang, Y.P., He, Y.J., Lee, Y.H., and Chan, Y.T. (2015). Dalton Trans. 44: 5139. He, J., Yin, Y.G., Wu, T. et al. (2006). Chem. Commun.: 2845. Ma, L.L., An, Y.Y., Sun, L.Y. et al. (2019). Angew. Chem. Int. Ed. 58: 3986. Han, Y.F., Jia, W.G., Lin, Y.J., and Jin, G.X. (2009). Angew. Chem. Int. Ed. 48: 6234. Huang, S.L., Lin, Y.J., Hor, T.S., and Jin, G.X. (2013). J. Am. Chem. Soc. 135: 8125. Lu, Y., Deng, Y.-X., Lin, Y.-J. et al. (2017). Chem 3: 110. Zhang, L., Lin, L., Liu, D. et al. (2017). J. Am. Chem. Soc. 139: 1653. Zhang, W.-Y., Lin, Y.-J., Han, Y.-F., and Jin, G.-X. (2016). J. Am. Chem. Soc. 138: 10700. Han, Y.F., Zhang, L., Weng, L.H., and Jin, G.X. (2014). J. Am. Chem. Soc. 136: 14608.

77

78

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53

Dong, J., Tan, C., Zhang, K. et al. (2017). J. Am. Chem. Soc. 139: 1554. Wei, P., Cook, T.R., Yan, X. et al. (2014). J. Am. Chem. Soc. 136: 15497. Yan, X., Wang, H., Hauke, C.E. et al. (2015). J. Am. Chem. Soc. 137: 15276. Yan, X., Wang, M., Cook, T.R. et al. (2016). J. Am. Chem. Soc. 138: 4580. Chen, S., Chen, L.-J., Yang, H.-B. et al. (2012). J. Am. Chem. Soc. 134: 13596. Jiang, B., Zhang, J., Ma, J.-Q. et al. (2016). J. Am. Chem. Soc. 138: 738. Chen, L.-J., Zhao, G.-Z., Jiang, B. et al. (2014). J. Am. Chem. Soc. 136: 5993. Huang, C.-B., Xu, L., Zhu, J.-L. et al. (2017). J. Am. Chem. Soc. 139: 9459. Zhao, L., Wei, J., Lu, J. et al. (2017). Angew. Chem. Int. Ed. 56: 8692. Yu, S.Y., Zhang, Z.X., Cheng, E.C.C. et al. (2005). J. Am. Chem. Soc. 127: 17994. Yu, S.-Y., Sun, Q.-F., Lee, T.K.-M. et al. (2008). Angew. Chem. Int. Ed. 47: 4551. Fu, J.H., Lee, Y.H., He, Y.J., and Chan, Y.T. (2015). Angew. Chem. Int. Ed. 127: 6329. Wang, S.Y., Fu, J.H., Liang, Y.P. et al. (2016). J. Am. Chem. Soc. 138: 3651. Wang, M., Wang, C., Hao, X.Q. et al. (2014). J. Am. Chem. Soc. 136: 6664. Yin, G.Q., Wang, H., Wang, X.Q. et al. (2018). Nat. Commun. 9: 567. Wang, M., Wang, K., Wang, C. et al. (2016). J. Am. Chem. Soc. 138: 9258. Wang, H., Qian, X., Wang, K. et al. (2018). Nat. Commun. 9: 1815. Zhang, Z., Wang, H., Wang, X. et al. (2017). J. Am. Chem. Soc. 139: 8174. Jiang, Z., Li, Y., Wang, M. et al. (2017). Angew. Chem. Int. Ed. 56: 11450. Song, B., Zhang, Z., Wang, K. et al. (2017). Angew. Chem. Int. Ed. 56: 5258. Li, Y., Jiang, Z., Wang, M. et al. (2016). J. Am. Chem. Soc. 138: 10041. Wang, L., Liu, R., Gu, J. et al. (2018). J. Am. Chem. Soc. 140: 14087. Song, B., Kandapal, S., Gu, J. et al. (2018). Nat. Commun. 9: 4575. Yan, X., Jiang, B., Cook, T.R. et al. (2013). J. Am. Chem. Soc. 135: 16813. Li, Z.-Y., Zhang, Y., Zhang, C.-W. et al. (2014). J. Am. Chem. Soc. 136: 8577. Chen, L.-J., Ren, Y.-Y., Wu, N.-W. et al. (2015). J. Am. Chem. Soc. 137: 11725. Zheng, W., Chen, L.-J., Sun, B. et al. (2016). J. Am. Chem. Soc. 138: 4927. Zheng, W., Yang, G., Shao, N. et al. (2017). J. Am. Chem. Soc. 139: 13811. Zhang, T., Zhou, L.-P., Guo, X.-Q. et al. (2017). Nat. Commun. 8: 15898. Zhou, L.-P. and Sun, Q.-F. (2015). Chem. Commun. 51: 16767. Li, Y.H., Jiang, J.J., Fan, Y.Z. et al. (2016). Chem. Commun. 52: 8745. Cai, L.-X., Yan, L.-L., Li, S.-C. et al. (2018). Dalton Trans. 47: 14204. Zhang, T., Zhang, G.-L., Zhou, L.-P. et al. (2017). Tetrahedron: Asymmetry 28: 550. Jia, C., Zuo, W., Yang, D. et al. (2017). Nat. Commun. 8: 938. Hou, J.Z., Li, M., Li, Z. et al. (2008). Angew. Chem. Int. Ed. 47: 1711. Yan, L.-L., Tan, C.-H., Zhang, G.-L. et al. (2015). J. Am. Chem. Soc. 137: 8550. Yan, Q.-Q., Zhou, L.-P., Zhou, H.-Y. et al. (2019). Dalton Trans. 48: 7080. Liu, C.L., Zhou, L.P., Tripathy, D., and Sun, Q.F. (2017). Chem. Commun. 53: 2459. Guo, X.-Q., Zhou, L.-P., Cai, L.-X., and Sun, Q.-F. (2018). Chem. Eur. J. 24: 6936. Wu, S.-Y., Guo, X.-Q., Zhou, L.-P., and Sun, Q.-F. (2019). Inorg. Chem. 58: 7091. Wu, B., Cui, F., Lei, Y. et al. (2013). Angew. Chem. Int. Ed. 52: 5096. Yang, D., Zhao, J., Zhao, Y. et al. (2015). Angew. Chem. Int. Ed. 54: 8658.

References

54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86

87 88 89 90 91 92 93 94

Yang, D., Zhao, J., Yu, L. et al. (2017). J. Am. Chem. Soc. 139: 5946. Zhang, W., Yang, D., Zhao, J. et al. (2018). J. Am. Chem. Soc. 140: 5248. Zhao, J., Yang, D., Zhao, Y. et al. (2014). Angew. Chem. Int. Ed. 53: 6632. Liu, G., Ju, Z., Yuan, D., and Hong, M. (2013). Inorg. Chem. 52: 13815. Jiao, J., Tan, C., Li, Z. et al. (2018). J. Am. Chem. Soc. 140: 2251. Wang, J., He, C., Wu, P. et al. (2011). J. Am. Chem. Soc. 133: 12402. He, C., Wang, J., Wu, P. et al. (2012). Chem. Commun. 48: 11880. Jing, X., He, C., Yang, Y., and Duan, C. (2015). J. Am. Chem. Soc. 137: 3967. Jing, X., Yang, Y., He, C. et al. (2017). Angew. Chem. Int. Ed. 56: 11759. He, Y.P., Yuan, L.B., Chen, G.H. et al. (2017). J. Am. Chem. Soc. 139: 16845. Zhang, Y.-T., Wang, X.-L., Li, S.-B. et al. (2016). Chem. Commun. 52: 9632. Gong, Y., Zhang, Y., Qin, C. et al. (2019). Angew. Chem. Int. Ed. 58: 780. Zhao, L., Wei, J., Zhang, J. et al. (2017). Angew. Chem. Int. Ed. 56: 15284. Kumagai, H., Hasegawa, M., Miyanari, S. et al. (1997). Tetrahedron Lett. 38: 3971. Wang, S., Gao, X., Hang, X. et al. (2016). J. Am. Chem. Soc. 138: 16236. Sun, L.Y., Sinha, N., Yan, T. et al. (2018). Angew. Chem. Int. Ed. 57: 5161. Wang, Y.S., Feng, T., Wang, Y.Y. et al. (2018). Angew. Chem. Int. Ed. 57: 15767. Wang, C., Hao, X.-Q., Wang, M. et al. (2014). Chem. Sci. 5: 1221. Jiao, J., Li, Z., Qiao, Z. et al. (2018). Nat. Commun. 9: 4423. Li, X.-Z., Zhou, L.-P., Yan, L.-L. et al. (2017). J. Am. Chem. Soc. 139: 8237. Luo, D., Zhou, X.P., and Li, D. (2015). Angew. Chem. Int. Ed. 54: 6190. Luo, D., Zhou, X.P., and Li, D. (2015). Inorg. Chem. 54: 10822. Zhu, B.C., Fang, W.H., Wang, J. et al. (2018). Chem. Eur. J. 24: 14358. He, C., Lin, Z., He, Z. et al. (2008). Angew. Chem. Int. Ed. 47: 877. He, C., Wu, X., Kong, J. et al. (2012). Chem. Commun. 48: 9290. Zhang, Y.-T., Wang, X.-L., Zhou, E.-L. et al. (2016). Dalton Trans. 45: 3698. Zhang, Y., Wang, X., Li, S. et al. (2016). Inorg. Chem. 55: 8770. Dai, F.-R. and Wang, Z. (2012). J. Am. Chem. Soc. 134: 8002. Xiong, K., Jiang, F., Gai, Y. et al. (2012). Chem. Sci. 3: 2321. Dai, F.-R., Sambasivam, U., Hammerstrom, A.J., and Wang, Z. (2014). J. Am. Chem. Soc. 136: 7480. Gong, W., Chu, D., Jiang, H. et al. (2019). Nat. Commun. 10: 600. Fang, Y., Lian, X., Huang, Y. et al. (2018). Small 14: 1802709. (a) Fang, Y., Li, J., Togo, T. et al. (2018). Chem 4: 555. (b) Fang, Y., Xiao, Z., Li, J. et al. (2018). Angew. Chem. Int. Ed. 57: 5283. (c) Fang, Y., Xiao, Z., Kirchon, A. et al. (2019). Chem. Sci. 10: 3529. Zhou, X.P., Liu, J., Zhan, S.Z. et al. (2012). J. Am. Chem. Soc. 134: 8042. Zhou, X.P., Wu, Y., and Li, D. (2013). J. Am. Chem. Soc. 135: 16062. Li, K., Zhang, L.Y., Yan, C. et al. (2014). J. Am. Chem. Soc. 136: 4456. Wu, K., Li, K., Hou, Y.J. et al. (2016). Nat. Commun. 7: 10487. Chen, S., Li, K., Zhao, F. et al. (2016). Nat. Commun. 7: 13169. Guo, J., Xu, Y.W., Li, K. et al. (2017). Angew. Chem. Int. Ed. 56: 3852. Hou, Y.-J., Wu, K., Wei, Z.-W. et al. (2018). J. Am. Chem. Soc. 140: 18183. Tan, C., Jiao, J., Li, Z. et al. (2018). Angew. Chem. Int. Ed. 57: 2085.

79

80

2 Coordination-Assembled Metal–Organic Macrocycles and Cages

95 Tan, C., Han, X., Li, Z. et al. (2018). J. Am. Chem. Soc. 140: 16229. 96 (a) Zhao, D., Tan, S., Yuan, D. et al. (2011). Adv. Mater. 23: 90. (b) Liu, T.-F., Chen, Y.-P., Yakovenko, A.A., and Zhou, H.-C. (2012). J. Am. Chem. Soc. 134: 17358. 97 Zhao, D., Yuan, D., Krishna, R. et al. (2010). Chem. Commun. 46: 7352. 98 Hang, X., Liu, B., Zhu, X. et al. (2016). J. Am. Chem. Soc. 138: 2969. 99 Wang, S., Gao, X., Hang, X. et al. (2018). J. Am. Chem. Soc. 140: 6271. 100 (a) Liu, M. and Liao, W. (2012). CrystEngComm 14: 5727. (b) Dai, F.-R., Becht, D.C., and Wang, Z. (2014). Chem. Commun. 50: 5385. 101 Pei, W.Y., Xu, G.H., Yang, J. et al. (2017). J. Am. Chem. Soc. 139: 7648. 102 Yu, S.-Y., Huang, H.-P., Li, S.-H. et al. (2005). Inorg. Chem. 44: 9471. 103 (a) Yu, S.-Y., Jiao, Q., Li, S.-H. et al. (2007). Org. Lett. 9: 1379. (b) Sun, Q.-F., Wong, K.M.-C., Liu, L.-X. et al. (2008). Inorg. Chem. 47: 2142. (c) Ning, G.-H., Yao, L.-Y., Liu, L.-X. et al. (2010). Inorg. Chem. 49: 7783. (d) Qin, L., Yao, L.-Y., and Yu, S.-Y. (2012). Inorg. Chem. 51: 2443. (e) Jiang, X.-F., Hu, J.-H., Tong, J., and Yu, S.-Y. (2013). Inorg. Chem. Commun. 36: 232. (f) Yao, L.-Y., Yu, Z.-S., Qin, L. et al. (2013). Dalton Trans. 42: 3447. (g) Jiang, X.-F., Deng, W., Jin, R. et al. (2014). Dalton Trans. 43: 16015. (h) Jiang, X.-F., Deng, W., Jin, R., and Yu, S.-Y. (2014). Inorg. Chem. Commun. 43: 98. 104 (a) Tong, J., Yu, S.-Y., and Li, H. (2012). Chem. Commun. 48: 5343. (b) Chen, Z.-M., Cui, Y., Jiang, X.-F. et al. (2017). Chem. Commun. 53: 4238. (c) Sun, W.-Q., Tong, J., Lu, H.-L. et al. (2018). Chem. Asian J. 13: 1108. 105 Zhang, G.-L., Zhou, L.-P., Yuan, D.-Q., and Sun, Q.-F. (2015). Angew. Chem. Int. Ed. 127: 9982. 106 Li, J.-R. and Zhou, H.-C. (2009). Angew. Chem. Int. Ed. 48: 8465. 107 He, Q.T., Li, X.P., Liu, Y. et al. (2009). Angew. Chem. Int. Ed. 48: 6156. 108 Wang, M., Wang, C., Hao, X.Q. et al. (2014). J. Am. Chem. Soc. 136: 10499. 109 Luo, D., Wang, X.Z., Yang, C. et al. (2018). J. Am. Chem. Soc. 140: 118. 110 Li, J.-R. and Zhou, H.-C. (2010). Nat. Chem. 2: 893. 111 Su, K., Wu, M., Yuan, D., and Hong, M. (2018). Nat. Commun. 9: 4941. 112 Zhang, Y., Gan, H., Qin, C. et al. (2018). J. Am. Chem. Soc. 140: 17365. 113 Jiang, X.-F., Hau, F.K.-W., Sun, Q.-F. et al. (2014). J. Am. Chem. Soc. 136: 10921. 114 Cai, L.-X., Li, S.-C., Yan, D.-N. et al. (2018). J. Am. Chem. Soc. 140: 4869.

81

3 Structural Chemistry of Metal-Oxo Clusters Xiaofeng Yi # , Weihui Fang# , Jinying Liu# , Cheng Chen# , Mingyan Wu, and Lei Zhang Fujian Institute of Research on the Structure of Matter, State Key Laboratory of Structural Chemistry, 155 Yangqiao Road West, Fuzhou 350002, Fujian, P.R. China

3.1 Oxo Clusters of Transition Metal 3.1.1

Introduction

Transition metal-oxo clusters with well-defined structures are unique in terms of molecular and electrical structural versatility and are of importance in several disciplines relevant to analytical chemistry, catalysis, biology, medicine, and materials science [1–5]. Thereinto the traditional polyoxometalates (POMs), i.e. discrete anionic molecular metal-oxo clusters, concern in particular the early transition metal ions in their higher oxidation states, such as MoV/VI , WVI , VIV/V , NbV , and TaV , represent a remarkable and classical class of transition metal-oxo clusters [6]. Such species typically form in condensation or self-assembly reactions of mononuclear metal oxo ions MOx n− in aqueous or (less common) organic medium by controlling a number of reaction parameters, e.g. reagents concentration, pH, solvent, and temperature, which result in a wide variety of accessible topologies, sizes, and compositions. The metal ions in such POMs typically possess predominantly octahedral coordination environment and the metal-oxo polyhedra in POM structures could be linked with each other in a corner-, edge-, or in rare cases also in a face-sharing manner (Figure 3.1). The gate of classical POMs chemistry was first opened in 1826 by a groundbreaking discovery of (NH4 )3 [PMo12 O40 ] in a reaction of ammonium molybdate with orthophosphoric acid by Berzelius [7]. However, the principles of the metal addenda ions connection in polyanions were first well understood only in 1933 after Keggin has reported results on first POM crystallographical characterization performed for H3 [PW12 O40 ]⋅5H2 O polyoxotungstate [8]. The POM families exhibit several archetypal structures, e.g. Lindqvist [9], Keggin, Wells-Dawson, Anderson-Evans, Dexter-Silverton, and Waugh structural types [6, 10] (Figure 3.2). # These authors contributed equally to this chapter and should be considered co-first authors. Advanced Structural Chemistry: Tailoring Properties of Inorganic Materials and their Applications, First Edition. Edited by Rong Cao. © 2021 WILEY-VCH GmbH. Published 2021 by WILEY-VCH GmbH.

82

3 Structural Chemistry of Metal-Oxo Clusters

Corner-sharing

Figure 3.1

Edge-sharing

Face-sharing

The possible manner between two MO6 octahedral units.

Lindqvist

Anderson-Evans

Keggin

Dexter-Silverton

Wells-Dawson

Waugh

Figure 3.2 Polyhedral representation of some examples of polyanions: an isopolyanion of Lindqvist ([M6 O19 ]n– ) structural type and heteropolyanions of Keggin ([XM12 O40 ]n– ), Wells-Dawson ([X2 M18 O62 ]n– ), Anderson-Evans ([XM6 O24 ]n– ), Dexter-Silverton ([XM12 O42 ]n– ), and Waugh ([XM9 O32 ]n– ) structural types. Color code: MO6 rose octahedra, XOn lime polyhedra.

Due to the development of synthetic strategies and approaches as well as analytical techniques, the last several decades have witnessed a remarkable development and progress related to fundamental and applied domains in traditional POMs chemistry, in the meantime, great efforts have also been made to extend this research realm to other transition metal-oxo clusters such as the most prosperous in recent years representative of polyoxotitanate (POTi), which is expected to be a promising photo-related sustainable materials with the purpose to ease the environmental and energy challenges. The nuclearity of POTis till now is rapidly keep growing to Ti52 [11], and numerous appropriate POTis nano-building blocks have been reliably utilized in the construction of extended metal–organic frameworks (MOFs) and functional materials, e.g. MIL-22, MIL-125, and MOF-901 [12]. The breakthrough discovery of late-transition-metal palladium-oxo polyanions opens a novel area in transition metal-oxo clusters. The frontier area of structurally well-defined polyoxopalladates (POPs), serving as the ideal models to decode the mechanistic insight of the Pd-containing catalytic materials in the molecular level, was initiated by Kortz and coworkers in 2008 with the pioneering illustration of the first POP [H6 Pd13 O8 (AsO4 )8 ]8− [13]. Their consecutive investigations on POPs diagnose the hypothetical fundamental building units, which have been employed for the construction of structural archetypes from cube, star, bowl, dumbbell, wheel to open-shell matrix.

3.1 Oxo Clusters of Transition Metal

According to the above mentioned, this chapter is directed to concentrate more on the structural fundamentals and novelties in the field of crystalline transition metal-oxo clusters, including Ti, V, Nb, Mo, Pd, W, and Ta. Besides, their tunable properties and potential applications related to the enormous structural diversities will be also briefly covered.

3.1.2

General Synthetic Approaches and Experimental Methods

A dramatic increase in the number of structure-determined transition metal-oxo clusters has been achieved due to the developments in novel synthetic approaches, methods, and instrumentations. 3.1.2.1 General Synthetic Approaches

Two main synthetic approaches for preparation of transition metal-oxo clusters have been developed and exploited. 3.1.2.1.1 One-Pot Self-Assembly Approach

The one-pot self-assembly pathway is based on the simultaneous reactions of all reagents and the corresponding resulting species are usually unpredictable as they are subject to various reaction parameters such as the reagents ratios and their concentrations, pH, ionic strength, temperature, as well as the presence or absence of reducing or templating agents. 3.1.2.1.2 Step-by-Step Approach

This path implies a step-by-step functionalization, where at the first step preformed precursors are isolated and then they are further used in the following reactions. The observed results suggest that this strategy could provide more control over topology and composition of the polyanions. 3.1.2.2 Experimental Methods 3.1.2.2.1 Strict Inert Condition Synthesis

Special transition metal-oxo clusters such as POTi needs to be synthesized under strict conditions in terms of the easy hydrolysis of reagents. Thus, it is essential to carry out the reaction under the argon atmosphere by means of standard Schlenk and glove box techniques. 3.1.2.2.2 Conventional Solution Synthesis

The conventional synthesis, the most popular method for synthesis of transition metal-oxo clusters, is characteristic of its intrinsic advantages as following: (i) relatively easy operation owing to the simple equipment and mild reaction conditions; (ii) the progress of reaction is conveniently observed along with reaction process; and (iii) the products are usually of high yield and may dissolve in water or organic solvents, which benefit the characterization and further investigation of the obtained compounds. The syntheses utilizing this method are performed by reacting initial components in various aqueous media at room temperature. A choice of the reaction

83

84

3 Structural Chemistry of Metal-Oxo Clusters

medium (present simple ions and their concentrations, ionic strength as well as pH) is of particular importance. The crystallization is achieved by evaporation of the filtered mother liquids with and without the addition of counteractions and takes from several days up to several weeks, depending on solubility of the obtained products. 3.1.2.2.3

Hydro(solvo)thermal Method

The hydro(solvo)thermal method, which can largely improve the solubility and reactivity of the starting materials on account of high pressure and temperature, has been demonstrated to be useful and efficient for producing crystalline clusters. Thus, more components such as aliphatic or aromatic ligands could be employed into the reaction system contributing greatly to the compounds with various structural diversities and abundant functions. 3.1.2.2.4

Inothermal and Eutectic Solvent Synthesis

The ionic liquids (ILs) not only act as solvent but also provide the templates or structure-directing agents (SDAs) in the formation of products [14]. In addition, the synthesis happens at ambient pressure in a light of little even no volatility of ILs eliminating safety concerns compared with the hydro(solvo)thermal reactions. Furthermore, the higher solubility of inorganic precursors in ILs can significantly enhance their reactivity during the reaction in contrast to hydro(solvo)thermal synthesis [15]. An organic salt capable of melting at a particular temperature mixed with other compound, known as eutectic solvents, can decrease the melting point to produce liquids consisting of predominantly ionic species. In recent years, it was demonstrated that the preparation of crystalline transition metal-oxo clusters has been realized via ILs or eutectic solvents.

3.1.3

Polyoxotitanates (POTis)

As a promising photo-related sustainable material, titanium dioxide (TiO2 ) was expected to help ease the energy crisis and environmental issues as well as pollution problems. The exploration of crystalline POTis stems from the molecular model of TiO2 materials but suffers from daunting and challenging synthesis. The classical and latest synthetic strategies, discussed in Section 3.2.2, were employed to promote the control of preparation of crystalline POTis. Thus, an increasing number of POTis with structural diversities were realized owing to the development and improvement of the synthetic techniques and further encourage us to investigate their tunable properties as well as potential application in many areas. 3.1.3.1 Diverse Structures of POTis

The achievement of POTis characteristic of structural diversities benefits from the development and improvement of synthetic procedures. In addition to the earlier reviews by Rosez and coworkers as well as Zhang and coworkers [1, 16, 17], here the state-of-the-art progress of POTis will be summarized according to the bridging ligands based on the recent extensive developments in particular from the group of Zhang in the field of synthesis, characterization, and utilization of POTis for the construction of advanced hybrid materials [1].

3.1 Oxo Clusters of Transition Metal

A nucleophilic reaction of titanium alkoxides with water, which could be introduced to the reaction system by traces of water or generated by an esterification reaction of acetic acid and alcohol, results frequently in the POTis where Ti centers are typically fused together via oxo bridges from alkoxides and carboxylates. The recent structures discussed here start from stabilization by carboxylates followed by the presence of other O-donor organic ligands, such as the phosphinates and phosphonates, in the condensation process of Ti(OR)4 (R = alkyl). Furthermore, an overview of the additional employment of N-donor organic ligands will be also displayed in conjunction with their ability in the self-assembly of POTis. Moreover, the POTis associated with inorganic species instead of organic ligands will be described later. 3.1.3.1.1 Carboxylate-Supported POTis

The earlier critical review by Rozes and Sanchze concentrates on titanium clusters containing oxo bridges fall into three structural types marked as titanium oxoalkoxides clusters, titanium oxo-carboxo-alkoxides clusters, and titanium oxocarboxo clusters [16]. In particular, the titanium oxo-carboxo-alkoxides are the most prosperous class among them undergoing not only large isostructural exploration essential for bandgap engineering but also rapid high nuclearity evolution. The carboxylate-containing ligands supply for water source via esterification and act as the oxo-bridges leading to the performed clusters much less sensitive to moisture makes their investigation of photo-related properties possible. Recently, Dai and coworkers introduced a series of carboxylate-containing ligands to the POTis exemplified as benzenedicarboxylates (labeled as BDC), salicylate (labeled as SAL), benzoic acid (labeled as BA) assistance with alizarin (labeled as Az), and ferrocene-1-carboxylate (labeled as FcCO2 ) accompanied with catechol (labeled as Cat) [18–21]. The representative example could be the non-sphere Ti13 cluster with formula Ti13 O10 (BDC)4 (SAL)4 (Oi Pr)16 possessing paddle arrangement with an S4 symmetry [18] (Figure 3.3a). Meanwhile, Zhang and coworkers contributed largely to this system by virtue of formic acid, propionic acid (PA), SAL, 5-fluorosalicylic acid, 1-hydroxy-2-naphthoic acid, terephthalic acid and its derivatives, as well as a series of dicarboxylate ligands with {Ti3 (μ3 -O)} building units [11, 22–25]. Nevertheless, sometimes it is unexpected that the disappearance of carboxylates in the final compounds, they are still requisite in the condensation process of cluster such as fullerene-like [Ti42 (μ3 -O)60 (Oi Pr)42 (OH)12 ]6− [22] (Figure 3.3b). Utilization of PA instead of formic acid achieves the isolation of cluster Ti52 (μ-OH)2 (μ-O)14 (μ3 O)50 (μ4 -O)8 (PA)34 (Oi Pr)28 , which presents the largest size record in the family of POTis by far [11] (Figure 3.3c). Rather than the self-assemble strategy namely one pot reaction, post functionalization is another efficient method for the discovery of novel carboxylates-attached POTis as exemplified by the functionalization of {I@Ti22 } host–guest cluster with catecholate and carboxylate ligands leading to cluster [I@Ti22 (μ2 -O)11 (μ3 -O)20 (μ1 -Oi Pr)16 (μ2 -Oi Pr)2 (t BuCOO)9 (OH2 )H3 ] [26]. 3.1.3.1.2 Phosphinate and Phosphonate-Stabilized POTis

At the early stage, phophonates serve as the protective ligands for the stabilization of titania nanoparticles and gradually involved in the aggregation of POTis on account

85

86

3 Structural Chemistry of Metal-Oxo Clusters

(a)

(b)

(c)

Figure 3.3 Structures of Ti13 O10 (BDC)4 (SAL)4 (Oi Pr)16 (a) Source: Modified from Hou et al. [18], [Ti42 (μ3 -O)60 (Oi Pr)42 (OH)12 ]6− (b) Source: Modified from Gao et al. [22], and Ti52 (μ-OH)2 (μ-O)14 (μ3 O)50 (μ4 -O)8 (PA)34 (Oi Pr)28 (c) Source: Modified from Fang et al. [11]. Color code: C, gray; O, red spheres; TiO5 and TiO6 , green; TiO7 , tan polyhedra. All H atoms and alkoxides ligands are omitted for clarity.

of the stable Ti—O—P bonds [27, 28]. In addition, the versatile coordination mode of tridentate phosphonates could largely expand topologies of POTis. In 2013, Schbert and coworkers elucidated the influence of organic phosphonate ligands on the structures of POTis. The strong coordination ability of phosphonates to Ti centers allows further emergence in not only [Ti7 (μ3 -O)2 (μ2 -Oi Pr)6 (Oi Pr)6 (O3 PCH2 CH2 CH2 Cl)6 ] but also Ti10 (μ2 -Oi Pr)2 (Oi Pr)14 (OMc)4 (O3 PCH2 CH=CH2 )10 (McOH = methacrylic acid) and Ti10 O2 (EtO)32 (AEP)2 (AEP = 2-aminoethylphoshonate) [27, 29, 30]. At the same time, Coppens and coworkers focused on the investigations that the binding mode of the phosphonate anchor is strongly dependent on the structure of the underlying substrate with successful isolation of larger POTis [Ti25 O26 (OEt)36 (C6 H5 PO3 )6 ] and [Ti26 O26 (OEt)39 (C6 H5 PO3 )6 ]Br [31] (Figure 3.4a,b). Furthermore, the combination of phosphate groups and carboxylate ligands enables the production of [Ti5 O(Oi Pr)11 (CH3 COO)(O3 PCH2 CH2 CH2 Br)3 ] and H2 [Ti18 (μ3 -O)14 (μ2 -O)6 (C6 H5 PO3 )2 (PA)16 (i OPr)14 ] [32, 33] (Figure 3.4c). Recently, Zhang and coworkers implied on the utilization of phosphite-stabilized Ti3 (μ3 -O) units as building blocks for the construction of coordination cages with topologies as [Ti6 (HPO3 )2 (i OPr)10 (μ3 -O)2 (μ2 -O)2 (NA)2 ] (PTC-75, HNA = nicotinic acid) cluster and [Ti18 (HPO3 )6 (i OPr)42 (μ3 -O)6 (BPDS)3 ] (PTC-78 BPDS = biphenyldisulfonic acid) capsule further revealed that the coordination behavior and spatial orientation of organic ligands can significantly influence the assembly of polymetallic Ti—O units [34] (Figure 3.4d).

3.1 Oxo Clusters of Transition Metal

(a)

(c)

(b)

(d)

Figure 3.4 Structures of [Ti25 O26 (OEt)36 (C6 H5 PO3 )6 ] (a) and [Ti26 O26 (OEt)39 (C6 H5 PO3 )6 ]Br (b). Source: Modified from Chen et al. [31], H2 [Ti18 (μ3 -O)14 (μ2 -O)6 (C6 H5 PO3 )2 (PA)16 (i OPr)14 ] (c). Source: Refs. [32, 33], and [Ti18 (HPO3 )6 (i OPr)42 (μ3 -O)6 (BPDS)3 ] (d). Source: Modified from Fan et al. [34]. Color code: C, gray; O, red; P, rose; S, gold spheres; TiO6 , green polyhedra. All H atoms and alkoxides as well as carboxylate ligands are omitted for clarity.

3.1.3.1.3 N-Donor Ligands Participated POTis

Nitrogen doping to POTis is a promising strategy for improving their photo-related activity. The pioneering work, related to the direct Ti—N bonds in POTis, by Schubert and coworkers suggest that mono- and dioximates are versatile ligands for the modification of titanium alkoxides [35]. The existence of oxime ligands could largely enhance the visible light harvest which is further evidenced by nuclearity structures in the range Ti4 , Ti5 , Ti6 , Ti7 , Ti12 , and Ti18 under the efforts of Zhang and coworkers [36]. Moreover, Zhang and coworkers continue to explore the property of Ti—N functionalization in POTis via combination of phophonates and azole-containing ligands and reveal an unprecedented approach for connecting robust titanium-oxo units into multiple cluster series and the ligands dependent rapid molybdenum blue (MB) degradation activities under normal sunlight might be attributed to the direct Ti–N functionalization [37, 38]. Phase isomerism is an important pheromone in inorganic materials, it is crucial to mimic the molecular isomerism of their nanocluster models. Zhang and coworkers reported the first pair of isomeric titanium-oxo clusters with assistance of the N-containing ligands 8-OQ by obtaining Ti20 -oxo clusters with vertical ([Ti20 (μ2 -O)8 (μ3 -O)20 (PA)14 (8-OQ)10 ] labeled as PTC-49V ; PA, 8-OQ = 8-hydroxyquinoline) and horizontal ([Ti20 (μ2 -O)10 (μ3 -O)16 (μ4 -O)2 (PA)14 (8-OQ)10 ] labeled as PTC-49H ) core configurations through combining pentagonal {Ti(Ti5 )} building units in corner-sharing or edge-sharing forms, where the vertical-type cluster exhibits better photocatalytic H2 evolution activity, higher photocurrent response, and faster charge transfer to external acceptors than its horizontal isomer [39] (Figure 3.5a,b).

87

88

3 Structural Chemistry of Metal-Oxo Clusters

(a)

(c)

(b)

(d)

Figure 3.5 Structures of [Ti20 (μ2 -O)8 (μ3 -O)20 (PA)14 (8-OQ)10 ] of PTC-49V (a) Source: Modified from Fan et al. [39], [Ti20 (μ2 -O)10 (μ3 -O)16 (μ4 -O)2 (PA)14 (8-OQ)10 ] of PTC-49H (b) Source: Modified from Fan et al. [39], [Ti28 (μ4 -O)4 (μ3 -O)20 (μ2 -O)24 (PhenO)14 (1,10-phn)14 ]2+ of PTC-64 (c) Source: Modified from Narayanam et al. [40], and [Ti10 (μ2 -O)4 (μ3 -O)8 (PZ)12 (Oi Pr)8 ] of PTC-197 (d) Source: Modified from Fan et al. [41]. Color code: C, gray; O, red; N, blue spheres; TiO5 and TiO6 , green; TiO7 , tan polyhedra. All H atoms and alkoxides as well as carboxylate ligands are omitted for clarity.

The blossoming of synthetic approaches keep dedicating efforts to the essential structural exploration. More recently, Zhang adopts ionothermal synthesis using deep eutectic solvents to produce the POTis with high nuclearity [Ti28 (μ4 -O)4 (μ3 -O)20 (μ2 -O)24 (PhenO)14 (1,10-phn)14 ]Cl2 ⋅6PhenOH (PTC-64) (PhenOH = phenol; 1,10-phn = 1,10-phenanthroline) and realizes the complete functionalization of Ti—O cores by conjugated ligands, which endow the obtained clusters with high water-phase stability making investigation to their photocatalytic H2 production activities possible [40] (Figure 3.5c). They also propose another innovative but universal pyrazole-thermal approach to synthesize POTis doped with abundant nitrogen with successful isolation of [Ti10 (μ2 -O)4 (μ3 -O)8 (PZ)12 (Oi Pr)8 ] (PTC-197) (PZ = pyrazole) exhibiting reversible photochromism due to the reduction of Ti4+ to Ti3+ by the photogenerated electrons [41] (Figure 3.5d). 3.1.3.1.4

Inorganic Ligands Bridged POTis

Most known POTis have been assembled from nonaqueous solvothermal synthesis by using organic ligands in the periphery, such compounds are usually passivated with alkoxide ligands or chelate ligands such as carboxylates or phosphonates. However, a small dosage of Ti(OR)4 in water can generate immediately a large amount of white precipitate, which prevents its practical application from synthesis in the water phase. One possibly efficient strategy is to provide the acidic solution to protect TiIV from hydrolysis, thus a battery of POTis performed in water phase is obtained. On the basis of this synthetic approach, Fu et al. carried out the reaction of Ti(OC4 H9 )4 with sulfuric acid in the presence of organo-amines resulted in

3.1 Oxo Clusters of Transition Metal

(a)

(b)

(c)

Figure 3.6 Structures of [TiO(SO4 )2 ]8 16− rings (a) Source: Modified from Fu et al. [42], and [Ti18 O27( OH2 )30 (SO4 )6 ]6+ in perpendicular views (b, c) Source: Modified from G. Zhang et al. [43]. Color code: C, gray; O, red; S, yellow spheres; TiO6 , green; TiO7 , tan polyhedra. All H atoms are omitted for clarity.

anionic [TiO(SO4 )2 ]8 16− rings [42] (Figure 3.6a). Wang and coworkers further used the solubility control to crystallize the {Ti18 } cores, where the 18 TiIV ions are uniquely connected with μ-oxo ligands into a tripledecked pentagonal prism [43] (Figure 3.6b,c). 3.1.3.2 Tuneable Properties of POTis: Bandgap Engineering and Photo-Related Activities

The photo-related material of TiO2 has been widely used to help ease the energy crisis and environmental pollution. However, its application is restrained to UV light range on account of the large bandgap at around 3.20 eV. The POTis were considered as the structure and reactivity model compounds of bulk nanoscale TiO2 materials. Therefore, investigation on the bandgap engineering of POTis will do benefit to their photocatalytic behavior and also help to elucidate the mechanisms of TiO2 modification. The ligands functionalization and metal/non-metal dopants have been employed to not only modify the electronic band structure but also enhance the photoreactivity and others [17, 44]. 3.1.3.2.1 Ligand Modification

POTis functionalized with organic ligands have been considered as ideal model compounds for spectroscopic investigation on light absorption of a photosensitizer on semiconductor surfaces [45]. Recently, a series of dye molecule attached POTis for the research of photocurrent responses were prepared. Some typical examples therefore will be summarized and demonstrated here to show a better understanding of how it works for the reader. In 2012, Coppens and coworkers functionalized the polyoxotitanate Ti17 O24 (i OPr)20 (labeled as Ti17 ) with four p-nitrophenyl acetylacetone ligands (labeled as NPA) possessing precise structural information, which enables theoretical simulations to predict the interfacial electron transfer (IET) in the compounds and allows to find the evidence of photoinduced IET revealed by EPR spectroscopy. The EPR spectra demonstrated that the excitation wavelength for the onset of charge separation lies between 345 and 295 nm for Ti17 ; meanwhile, the charge-separated state for Ti17 -NPA4 clusters is observed with irradiation above 400 nm due to the

89

90

3 Structural Chemistry of Metal-Oxo Clusters

addition of the NPA chromophore [45]. In addition, they continued to sensitize the Ti17 with covalent ligation of p-nitrophenyl-acetylacetonate or coumarin 343 (labeled as C343) adsorbates and find that hole injection into surface states can be induced in Ti17 NPA4 and Ti17 C3432 by photoexcitation with wavelengths at the threshold of the photoabsorption spectrum but the minimum energy triggers electron injection [46]. Recently, Zhang et al. provided a systematic investigation on the bandgap engineering of POTis. They proposed to utilize the robust phosphonate-stabilized Ti6 P2 cluster, [Ti6 O4 (i OPr)10 (O3 PR)2 ]2+ (labeled as Ti6 P2 ) has been reported by Schubert and coworkers [33] as a platform to investigate their bandgap modulations. Correspondingly, 14 O-donor ligands, including carboxylates, phosphonates, and sulfonates, have been successfully attached to the active coordination sites of the Ti6 P2 cluster core. The increasing electron-withdrawing effect of the organic ligands allows the gradual reduction of the bandgaps of functionalized Ti6 P2 species. In addition, transition-metal ions are introduced to organize the functionalized Ti6 P2 clusters into polymeric structures and indicate that the coordination environments of the applied metal ions significantly influence their property of visible light adsorption [47] (Figure 3.7). Bandgap engineering through ligand modification can also be found in extended frameworks based on polyoxotitanate building blocks. MIL-125, constructed from Ti—O clusters and 1,4-benzenedicarboxylate (labeled as BDC) linkers, exhibits an optical bandgap in the UV region (c. 3.6 eV) [48]. Walsh and coworkers modified the

(a) PTC-1

PTC-2

O

PTC-3

O

O

O HO

HO N

Co2+ N

N

N HO

N

HO

P HO

PTC-16

O

O

OH

HO

O

O

HO

HO

Br

{Ti6P2}

O

Br

HO

F

F

O

F HO

F

PTC-13

PTC-7

HO

O

PTC-15

PTC-6

O

OH

Co2+ N

O

HO

O

NO2

O

N

OH

N N

Cu2+ N

PTC-14

PTC-5

N

HO

PTC-17

PTC-4

F

PTC-12

F F F

O HO

PTC-11

s O

PTC-8 PTC-9

O HO

PTC-10

Figure 3.7 Illustration of the ligand modification of phosphonate-stabilized [Ti6 O4 (i OPr)10 (O3 PR)2 ]2+ . Source: Based on Czakler et al.[33]. Copyright 2014, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

3.1 Oxo Clusters of Transition Metal

bandgap of MIL-125 with aminated linkers, in specular the diaminated BDC-(NH2 )2 as the linking units was predicted to lower the bandgap of MIL-125 to 1.28 eV [49]. This work confirms the efficient strategy to lower the bandgap is to increase the electron density in BDC linking units. Inspired by this opinion, Uribe-Romo and coworkers prepared a series of MOFs isoreticular to MIL-125-NH2 , displaying a gradual decrease in the optical bandgap from 2.56 eV in MIL-125-NH2 to 2.29 eV in MIL-125-NHCy [50]. 3.1.3.2.2 Heterometallic Doping

Apart from ligands modification, attention has also turned into the doping of POTis with low levels of metal and non-metal (e.g. N) atoms, which give rise to significant reduction of the bandgap and facilitate therefore photoactivity into visible light region. In addition to the non-metal-doped POTis were already discussed in Section 3.1.3. Relevance to the metal-doped strategy presented hereinafter, a broad family of recent studies has shown that doping of POTis with metal ions, such as alkaline/alkaline earth metals [51], transition metals [52–57], lanthanides [58–61], and noble metals ions [62], leads to crucial improvement of photocatalytic- and optical-limiting activities. On the basis of connectivity between the core of POTis and doped metal centers, the metal-doped POTis could be divided into two kind of doping models. On the one hand, the doped metal centers are incorporated to the Ti—O cluster via μ-O bridge (marked as type I). On the other hand, the other manner is the functionalization of POTis with heterometal-oxo clusters to form intercluster compounds (marked as type II). The intrinsic nature of the doped metal ions has a significant effect on the bandgap, which is indicated in the series of hollow cage-like mono-substituted Ti12 clusters. Coppens and coworkers carried out the first systematic investigation on homodisperse-doped POTis with formula Ti11 (MX)O14 (i OPr)17 (M = Mn, Fe, or Co; X = Cl, Br, or I), and all compounds exhibit redshifted bands compared with parent Ti12 cluster, in particular the Fe-containing complex possessing pronounced absorption up to about 800 nm [52]. In most circumstances, the metal-doped POTis could be assigned to the type I, and the introduction of polymetallic components to POTis leading to type II remains rare and challenging. Copper ions serving as dopants have proven to be an effective methodology for decreasing bandgap in the field of POTis materials [57, 63]. Recently, Zhang and coworkers realized the successful isolation of a series of crystalline supersalts including both POTis and copper halide, [Ti12 (μ3 -O)14 (Oi Pr)18 ]⋅2[CuI Cl2 ]⋅2HOi Pr, [Ti12 (μ3 -O)14 (Oi Pr)18 ]⋅[CuII 2 Cl4 (μ-Cl)2 ], [Ti12 (μ3 -O)14 (Oi Pr)18 ]⋅[CuI 4 Br6 ], and [Ti12 (μ3 -O)14 (Oi Pr)18 ]⋅[CuI 5 I7 ]. Interestingly, they consist of the same cationic [Ti12 (μ3 -O)14 (Oi Pr)18 ]2+ cluster but scale-up copper halides anions with the increment of halide radius and exhibit significant visible light absorption and lower bandgap compared with the previously reported undoped Ti12 clusters [64] (Figure 3.8).

91

92

3 Structural Chemistry of Metal-Oxo Clusters

Salts

Supersalts

Figure 3.8 Copper halide clusters doped Ti12 series. Source: Fang et al.[64]. Copyright 2017, Royal Society of Chemistry.

The distance between two different types of clusters is largely dependent of the stacking mode of the structures in the above polymetallic component introduction. Therefore, the judicious selection of organic ligands as linkers is necessary in order to better control polymetallic doping with tunable distance. Zhang and coworkers continued to imply on the employment of the robust Ti6 P2 with labile coordination sites as the platform to investigate their assembly with polymetallic components [47, 65]. And the bifunctional N/O-donor ligands are introduced as organic linkers in light of the soft and hard acid–base theory in coordination chemistry that Cu and Ti ions have affinity to nitrogen and oxygens, respectively. Thus, Ti6 P2 clusters have been linked to a variety of polymetallic copper halides ranging from mononuclear to tetranuclear leading to 1D chain, 2D layer, and 3D diamond framework materials. The achievement of polymetallic and multidimensional doping has a crucial effect on the POTis concerning the light absorption and bandgap structures. Furthermore, the designable construction of core–shell materials with atomically precise structures is realized on the basis of this molecular assembly method. 3.1.3.3 Potential Application of POTis

The crystalline POTis materials, characteristic of precisely atomic structure, structural diversity, tunability, solubility, and photoactivity, have found numerous applications in the last decade ranging from carbon dioxide adsorption [66, 67], anticancer agents [68], catalyst, and other photo-related activities [19, 62, 69–72]. Attributed to synergistic effect, the obtained ternary composites, the preformed CdS/MIL-101 composites were loaded with a series of polyoxotitanate clusters, are all highly efficient H2 evolution photocalysts. Moreover, their activities can be significantly improved by increasing the conjugated effect or aromatic decoration of the organic ligands in POTis [69] (Figure 3.9). Furthermore, Zhang and coworkers applied a modified liquid phase epitaxy (LPE) approach to load homochiral Ti4 (OH)4 (R/S-BINOL)6 clusters (BINOL = 1,1′ -bi-2-naphthol) inside the pores of an achiral HKUST-1 MOF thin film, which can be used to recognize and separate (+)-MeLt and (−)-MeLt (MeLt = methyl lactate) [70] (Figure 3.10). More recently, the photo-related optical limiting activity could also be realized through combination of noble metal cluster with POTis. For this purpose, Zhang and

3.1 Oxo Clusters of Transition Metal

H2

Light Energy band matching

Ligand sensitization

H2O CdS

OHHO

O

O HO

O HO

H2O

O

N O

Br

N

Polyoxo-Ti clusters

H2

CdS/MIL-101

Polyoxo-Ti clusters

Figure 3.9 Strategy for constructing ternary PTC/CdS/MIL-101 photocatalysts with efficient H2 -evolution activities. Source: Jiang et al. [69]. Copyright 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

(200)

Ti-MOC

Pristine HKUST-1 (400) Ti-MOC@HKUST-1 Ti C O

I

Experimental

HKUST-1 Ti-MOC@HKUST-1

(a)

Simulation

1.66 nm (c)

6

8

10

12

2θ (°)

14

16

Ti-MOC@HKUST-1 1459 1323 1385 1240

18

20

Absorbance

1653

1800

(b)

Ti-MOC@HKUST-1

(d)

982

HKUST-1

Ti-MOC

1600

1400

1200

1000

Wavenumbers (cm–1)

Figure 3.10 (a) Structure of the R-Ti-MOC cluster. (b) Schematic presentation of in situ layer-by-layer (lbl) growth of enantiopure Ti-MOC-loaded HKUST-1 thin film using LPE approach. Characterization of chiral Ti-MOC-loaded HKUST-1 thin film via an in situ lbl LPE approach: (c) XRD and (d) infrared reflection absorption spectroscopy (IRRAS). Source: Gu et al. [70]. Copyright 2016, American Chemical Society.

93

3 Structural Chemistry of Metal-Oxo Clusters

coworkers prepared two unprecedented Ag6 @Ti16 nanoclusters with precise atomic structures. The octahedral Ag6 core can adopt diverse geometric configurations inside the Ti16 shell and differ by about 45∘ rotation in compounds PTC-47 and PTC-48 resulting from different acidic/redox synthetic conditions. Correspondingly, structural differences have a crucial effect on their optical limiting effects that PTC-47 displays better optical limiting activity toward 532 nm laser light than PTC-48 [71] (Figure 3.11). In addition, single atom supported catalysts 1.05

Normalized transmittance

1.00 0.95 0.90 PTC-47 PTC-48

0.85 0.80 0.75 0.70 –100

(a)

–50

0 Z (mm)

50

100

1.10 1.05 1.00 0.95 0.90 0.85 1 × 10–4 M 2 × 10–4 M 3 × 10–4 M 4 × 10–4 M

0.80 0.75 0.70

Transmittance reduction (%)

Normalized transmittance

94

0.65 0.60

45 40 35 30 25 20 15

1

0.55 (b)

–100

–50

0 Z (mm)

50

2 3 C(10–4 M)

4

100

Figure 3.11 The open aperture Z-scan (points) and theoretical fit (solid lines) curves of PTC-47 and PTC-48 at 532 nm (upper) and different concentrations of PTC-47 (lower). Inset: transmittance reduction versus concentration plot. Source: Chen et al. [71]. Copyright 2018, Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim.

3.1 Oxo Clusters of Transition Metal

Active metal site

Single cluster

Oxide cluster support

Single atom

synergism

Atomic resolution Metal stabilization Charge distribution Unique performance

Figure 3.12 Illustration of the proposed molecular system consisting of a Ti–O cluster supporting single active metal sites with synergistic performance. Source: Chen et al. [62]. Copyright 2019, Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim.

have been emerging as a promising material in a variety of catalytic applications. Recently, Zhang and coworkers successfully obtained a new kind of single Ag sites doped POTi Ag10 Ti28 (μ2 -O)12 (μ3 -O)30 (BC)38 (HBC)2 (CH3 CN)6 (H2 O)2 (PTC-80 HBC = benzoic acid), which holds four exposed and six embedded Ag sites. DFT calculations indicate that all doped Ag atoms contribute to the reactivity of PTC-80, where the embedded Ag sites distributing charges through Ti–O–Ag moieties and exposed ones acting as active single sites. At the same time, PTC-80 exhibits a moderate CO binding capacity similar to the metallic copper-based catalysts. Correspondingly, this work reveals that POTi-supported single noble metal sites can act as a promising class of single atom catalysts [62] (Figure 3.12).

3.1.4

Polyoxovanadates (POVs)

The chemistry of polyoxovanadates (POVs), characterizing homo- or hetero-valent V centers exhibiting such as square-pyramidal [VV O5 ]5− and [VIV O5 ]6− , octahedral [VV O6 ]7− , and [VIV O6 ]8− and tetrahedral [VV O4 ]3− coordination geometries, is a prosperous field of research and generally divided into fully oxidized (VV ), mixed-valent (VV /VIV and VIV /VIII ), fully reduced (VIV ), and highly reduced (VIII ) POVs families possessing a number of intriguing properties, which enable their wide application and relevance in various branches of chemical, physical, and biological sciences [2, 73–75]. Herein, the key developments in the structural chemistry as well as the reactivity of POVs functionalized heterogroups will be reviewed. 3.1.4.1 Diverse Structure of POVs

The POVs are capable to self-assemble in both aqueous and organic phases [76], where small {VOx } fragments aggregate to form a vast number of vanadium-oxo clusters with various compositions and geometries. They possess rich electronic properties with an exceptional capacity to form mixed-valence species and are generally divided into fully oxidized (VV ), mixed-valent (VV /VIV and VIV /VIII ), fully reduced (VIV ), and highly reduced (VIII ) POVs families [2]. The fully oxidized classes are popular with the X-ray single-crystal diffraction determined structures of [V2 O7 ]4− , [V3 O9 ]3− , [V4 O12 ]4− , [V5 O14 ]3− , [V10 O28 ]6− ,

95

96

3 Structural Chemistry of Metal-Oxo Clusters

[V4O12]4–

[V5O14]3–

[V10O28]6–

[V12O32]4–

[V13O34]3–

[V16O42]4–

Figure 3.13 Structures of fully oxidized POVs with different topologies. Color code: O, red spheres; VV Ox , yellow-tan polyhedra.

[V12 O32 ]4− , [V13 O34 ]3− , [V15 O42 ]9− , and [V16 O42 ]4− [76–84] (Figure 3.13). Likewise, the crystallographically characterized [VIV VV 12 O40 ]16− , [VIV 8 VV 7 O36 ]5− , [VIV 5 VV 12 O42 ]4− , [VIV 8 VV 10 O44 ]6− , [VIV 8 VV 14 O54 ]6− , and [VIV 16 VV 18 O82 ]10− constitute the classes of mixed-valent (VV /VIV and VIV /VIII ) species [85–94] (Figure 3.14). The archetypal [VIV 18 O42 ]12− polyanions represent the family of fully reduced POVs species [95]. The VIII -containing POVs (mixed-valence [VIV /VIII ] POVs and highly reduced [VIII ] POVs families) are usually further stabilized with ligands [96, 97]. 3.1.4.1.1

Template-Effect-Determined POVs

In recent years, there has been an increasing interest in POVs due to its capacity of form vanadium-oxo clusters encapsulating neutral or charged molecules functioning as templates, and the central templates certainly determine the shape of the POVs shell [98–100]. Achim Müller isolated the polyanions featuring an approximately spherical [V18 O42 ]3− shell as a unique host system for spherical guest such as halide ions [88]. While the shape-determining templates turned to linear-type N3 − ions resulting in the corresponding oval-type POVs shell due to the structural complementarity between the cluster shell and the guest ions. Furthermore, the bigger ClO4 − ion-induced gourd-shaped POVs shell with higher nuclearity, which is formulated as [ClO4 @V22 O54 ]6− [86] (Figure 3.15). This family of inclusion complex proves that the POVs cluster shell formation is controlled via the molecular recognition of a template. During the host–guest complexation in POVs system, the different types of linking of an inorganic square-pyramidal VO5 unit is controlled by a central anionic template, which can influence the shape and size of cluster shells [85, 89, 92, 99–103].

3.1 Oxo Clusters of Transition Metal

[VIVVV12O40]16–

[VIV8VV7O36]5–

[VIV5VV12O42]4–

[VIV8VV10O44]6–

[VIV8VV14O54]6–

[VIV16VV18O82]10–

Figure 3.14 Structures of mixed-valence (VV /VIV ) POVs-shell with different topologies. Color code: O, red spheres; VV Ox and VIV Oy with same color, aqua polyhedra, because the valence of some V centers is not unambiguously determined or intermediate between the VV and the VIV ions.

[I@VIV10VV8O42]4–

[N3@VIV8VV10O44]7–

[CIO4@VIV8VV14O54]7–

Figure 3.15 Structures of host–guest POVs with different topologies. Color code: O, red; N, blue; Cl, light-blue; I, purple; VV/IV , aqua spheres. The shape-determining templates inside the POVs shell are shown enlarged.

3.1.4.1.2 Metal-Doped POVs

The metal ion can be employed as an effective seminal seed in the oligomerization process of {VOx } units to facilitate the formation of the larger POVs species [91, 104–107]. For instance, Hayashi and coworkers isolated the hexavanadate [PdV6 O18 ]4− , octavanadate [Cu2 V8 O24 ]4− , and decavanadate [Ni4 V10 O30 (OH)2 (H2 O)6 ]4− polyanions, which define a new class of cyclic POVs species featuring novel macrocyclic coordination geometries between POVs and various metal ions. These inorganic crown-ether-type inclusion complexes can potentially extract transition metal ions selectively within their ring cavities [106] (Figure 3.16a–c). Recently, they manipulated the formation of sandwich-, ring-, and

97

98

3 Structural Chemistry of Metal-Oxo Clusters

[PdV6O18]4–

[Cu2V8O24]4–

[Ni4V10O38]4–

[Y4V20O62]12–

Figure 3.16 Structures of metal-doped POVs. Color code: O, red; N, blue; Pd, violet; Cu, light-blue; Ni, sky-blue; Y, light-green spheres; VV O4 , yellow-tan tetrahedra.

cage-type Y-doped POVs via controlling the stoichiometry of Y3+ and [VO3 ]− [107] (Figure 3.16d). 3.1.4.1.3

Ligands-Participated POVs

The introduction of oxygen-, nitrogen-, phosphorus-, arsenic-, antimony-, silicon-, and germanium-containing donor ligands to POVs allows charge compensation via substitution of bridging or terminal oxo ligands to stabilize otherwise unstable POVs architectures and further provide novel building units for the assembly of POVs systems with higher nuclearity. Oxygen- and Nitrogen-Supported POVs The utilization of multidentate and bulky

ligands largely not only contributes to the aggregation process of metal-oxo but also favor the resulting clusters aside from each other by locating at their periphery [2, 75, 108–111]. The oxygen- and nitrogen-containing ligands seem particularly attractive. For instance, the phenol oxygen in bulky calixarenes macrocycles could act as an available platform for coordination to vanadium centers. Luneau and coworkers reported the polyoxo(alkoxo)hexavanadate anion [V6 O6 (OCH3 )8 (calix) (CH3 OH)]− (calix = p-tert-butylcalix[4]arene) exhibiting a Lindqvist-type mixedvalent {VIII VIV 5 O19 } core incorporated to the calix macrocycle with the assistance of conjugated acid of the base. The study of the magnetic behavior shows that the VIII ⋅⋅⋅VIV interactions are found to be ferromagnetic while the VIV ⋅⋅⋅VIV are antiferromagnetic. This novel family of polyanions opens a new direction for the investigations to the properties of POVs species [109]. The POVs can act as the promising second building units (SBUs) in reticular Chemistry. Wang and coworkers employed the highly symmetric polyatomic SBUs combined with carboxylate ligands to construct fascinating metal–organic polyhedral (MOP) [112–114] (Figure 3.17). Furthermore, Mclnnes and coworkers have developed the use of N-containing ligands, such as 5,6-dimethylbenzotriazole (labeled as Me2 btaH), that serve as rigid-bridging ligands in an effort to isolate low-valent POVs cages with useful magnetic properties. Correspondingly, mixed-valent paramagnetic cage-like polyanion [(VIV O)8 VIII 2 (Me2 bta)8 (OH)4 (OMe)10 ] was unexpectedly generated [110] (Figure 3.18).

3.1 Oxo Clusters of Transition Metal

Figure 3.17 (a) Ball-and-stick view of the hexavanadatecluster that corresponds to a trigonal vertex. (b) A V-MOP built from four {V6 (SO4 )}clusters and six edb ligands with the dimension of 26.34 Å, which can be simplified as a truncated tetrahedron. (c) Three types of supramolecular synthon. (d) The diverse superstructures formed by the packing of truncated tetrahedra. Source: Gong et al. [112]. Copyright 2019, Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim.

Bridging-ligand substitution

(a)

VMOP

(b)

(c) Synthon A

(d) VMOP-α

(a)

edb

Synthon B

Synthon C

VMOP-β

VMOP-γ

(b)

Figure 3.18 Structures of oxygen- and nitrogen-containing ligands for stabilization of POVs: [V6 O6 (OCH3 )8 (calix)(CH3 OH)]− (a) [109], [(VIV O)8 VIII 2 (Me2 bta)8 (OH)4 (OMe)10 ] (b) [110]. Color code: C, gray; O, red; N, blue spheres; VIV O6 , aqua; VIII O6 , lavender octohedra. All H atoms are omitted for clarity. Source: Based on Aronica et al. [109].

99

100

3 Structural Chemistry of Metal-Oxo Clusters

Phosphorus-, Arsenic-, and Antimony-Containing Ligands Incorporated POVs The VIII -containing POVs have been considered as promising magnetic materials on account of large zero-field splitting (ZFS) of the VIII ion. Xu and coworkers took advantage of the synergistic effect of alkylamine and phosphoric acid for the attempt to design and synthesize novel VIII -containing POVs skeleton with formula [VIII 3 VIV 18 P6 O60 (DACH)3 ]9− (DACH = 1,2-diaminocyclohexane) (Figure 3.19a) [115]. To explore a set of VIV -based magnetic POVs clusters, Bensch and coworkers in 2008 isolated the first discrete antimony POVs: [V16 Sb4 O42 ]8− , [V14 Sb8 O42 ]4− , and [V15 Sb6 O42 ]6− [118]. Recently, they were able to realize the configurational isomerization in Sb-POVs [VIV 14 SbIII 8 O42 (H2 O)]4− [117] (Figure 3.19d). The assembly of metal–ligand-mediated molecular capsules to date constitutes one of the most vibrant areas of chemistry. A key issue for the formation of such species is the recognition of preorganized and stable building entities, which provide ligand-available coordination sites for directing the blocks into desired capsule molecules. Schmitt and coworker employed organic arsonates and phosphonates with coordination to POVs building blocks leading to cage-like compounds [V10 O18 (O3 PC12 H8 PO3 )4 ]10− [116] (Figure 3.19b) [(VIV O)16 (OH)8 (O4 AsC6 H5 )2 (O3 AsC6 H5 )8 ] [98] (Figure 3.19c).

(a)

(c)

(b)

(d)

Figure 3.19 Structures of phosphorus-, arsenic-, and antimony-containing ligands for stabilization of POVs: (a) [VIII 3 VIV 18 P6 O60 (DACH)3 ]9− . Source: Modified from Wang et al. [115], (b) [V10 O18 (O3 PC12 H8 PO3 )4 ]10− . Source: Based on Breen and Schmitt [116], (c) [(VIV O)16 (OH)8 (O4 AsC6 H5 )2 (O3 AsC6 H5 )8 ]. Source: Modified from Zhang and Schmitt [98], (d) [VIV 14 SbIII 8 O42 (H2 O)]4− . Source: Modified from Mahnke et al. [117]. Color code: C, gray; O, red; N, blue; P, rose; As, orange; Sb, pink spheres; VV O5 , yellow-tan; VIV O5 , aqua square pyramid; VIII O6 , lavender octohedra. All H atoms are omitted for clarity.

3.1 Oxo Clusters of Transition Metal

(a)

(b)

(c)

(d)

(e)

(f)

Figure 3.20 Structures of silicon- and germanium-containing ligands for stabilization of POVs: (a) [VIV 18 O42 ]12− . Source: Based on Johnson and Schlemper [95], (b) [VIV 15 SiIV 6 O48 ]12− . Source: Modified from Gao et al. [119], (c) [VIV 16 SiIV 4 O46 ]12− . Source: Based on Wang et al. [120], (d) [Ge8 V12 O48 ]16− , (e) [Ge8 V14 O50 ]12− , and (f) [Ge4 V16 O46 ]12− . Source: (d, e, f) Modified from Whitfield et al. [121]. Color code: C, gray; O, red; Si, dark-rose; Ge, light-orange spheres.; VIV O5 , aqua polyhedra.

Silicon- and Germanium-Containing Ligands Stabilized POVs In comparison to

P-POVs, As-POVs, and Sb-POVs, the chemistry of Si-POVs and Ge-POVs is still more underdeveloped and can be typically considered as the being derived from the fully reduced [VIV 18 O42 ]12− archetype [95] (Figure 3.20a). The family of Si-POVs is characterized by the assemblies with nuclearity V15 , V16 , and V18 [119, 120, 122], such as [VIV 15 SiIV 6 O42 (OH)6 (Cl)]7− [119] and [VIV 16 SiIV 4 O46 ]12− [120] (Figure 3.20b,c), while the Ge-POVs are classified as V6 , V9 , V12 , V14 , V15 , and V16 [119, 121–124] (Figure 3.20d–f). 3.1.4.2 Tunable Properties and Potential Applications of POVs

Incorporation of organic groups into POV building blocks would endow the POV-based organic–inorganic hybrids with additional functionality and promising applications [75, 125, 126]. In particular, the POVs-based MOFs with nanostructure represent a class of promising heterogeneous catalysts. In 2005, Hong and coworkers explored the first mixed-valent tetradeca-vanadate synthesized from organic solution, which possesses half-open basket-shaped structure and displays an intense blue luminescent emission in solution might be active in photo-oxidation reactions and serve as a potential photocatalyst [127]. Liu and coworkers adopted the nanasheet [Ni(4,4′ -bpy)2 ]2 [VIV 7 VV 9 O38 Cl]⋅(4,4′ -bpy)⋅6H2 O (NENU-MV-1a) as the catalyst in olefin epoxidation in air and exhibited excellent catalytic activity (95% conversion) [125] (Figure 3.21). In addition, Niu and coworkers elucidated that the mixed-valent POVs-based carboxylate derivative K6 H[VV 17 VIV 12 (OH)4 O60 (OOC(CH2 )4 COO)8 ]⋅nH2 O possesses catalytic properties for the oxidation of sulfides under mild conditions [75] (Figure 3.22).

101

102

3 Structural Chemistry of Metal-Oxo Clusters

Figure 3.21 Representation of formation, structure, and proposed mechanism for the epoxidation of cyclohexene on NENU-MV-1a. Source: Reproduced with permission Wang et al. [125]. Copyright 2019, American Chemical Society. Figure 3.22 Representation of the structure and catalytic property for the oxidation of sulfides of [VV 17 VIV 12 O64 (OOC(CH2 )4 COO)8 ]11− . Conv. = conversion; Selec. = selectivity. Source: Wang et al. [75]. Copyright 2017, American Chemical Society.

TBHP 60 °C Conv. 81% Selec. 100%

S S

3.1.5

O

O

Polyoxoniobates (PONbs)

In comparison to group VI (Mo, W) yields fascinating high nuclearity metal-oxo clusters, possessing rich diversity of size and topologies accessible by means of pH-control and redox chemistry to promote rearrangement, polyoxoniobates (PONbs) chemistry predominantly confined to alkaline pH and relative redox inert to date. The Lindqvist-type polyanion [Nb6 O19 ]8– is the predominant species in aqueous solution above pH 12. Thus, exploitation of the [Nb6 O19 ]8− as a precursor is a key approach for the assemblies of higher-nuclearity PONbs via controlling various reaction parameters, such as reducing the pH value, addition of hetero-metal ions, and increasment of temperature. Pioneering work of Cronin group unveiled

3.1 Oxo Clusters of Transition Metal

that [Nb6 O19 ]8− subjected to hydrothermal conditions in the presence of sodium dibenzyldithiocarbamate giving rise to [HNb27 O76 ]16− and [H10 Nb31 O93 (CO3 )]23− polyanions constructed by pentagonal building blocks in their respective architectures [128]. Wang and coworkers contribute significantly to the high nuclearity PONbs via investigation on the self-assembly and photocatalytic properties of POVs: {Nb24 O72 }, {Nb32 O96 }, and {Nb96 O288 } clusters, featuring the construction of [Nb7 O22 ]9− fundamental building blocks [129]. Great efforts have been made by Zheng and coworkers for the isolation of high-nuclearity PONb clusters: {Nb52 O150 }, {Nb81 O225 }, and {Nb114 O316 } [130]. Recently, they continued to elucidate a protein-sized windmill-like cluster {Nb288 O768 (OH)48 (CO3 )12 } (labeled as Nb288 ), which is by far the highest nuclearity PONb and can be viewed as the aggregation of six {Nb47 O128 (OH)6 (CO3 )2 } (labeled as Nb47 ) units further joined together by six additional Nb centers. The Nb47 segments in particular could be generated in situ and further linked by Cu centers to form 3-D PONb framework [131] (Figure 3.23). Furthermore, they have successfully isolated the rare and to date largest Cr-substituted PONb [Cr2.5 Nb27.5 O66 (OH)20 (H2 O)2 ]7− with high ionic conductivity [132]. In addition to [Nb6 O19 ]8− , Nyman and coworkers expand to [Nb10 O28 ]6− as precursor, which transforms to oligomers of (Nb24 O72 )24− upon addition of solely alkali chloride salts [133]. They also illustrate a reaction pathway for control over speciation of that is driven by countercations instead of pH by formation of [Nb14 O40 (O2 )2 H3 ]14− , [((UO2 )(H2 O))3 Nb46 (UO2 )2 O136 H8 (H2 O)4 ]24− , and [(Nb7 O22 H2 )4 (UO2 )7 (H2 O)6 ]22− [134] (Figure 3.24). Nb10

Nb3

(b) Nb47 Nb17 =

(a)

+ Nb O C

(c)

Cu N C

Figure 3.23 Structure of Nb47 (a), Nb288 (b) and Nb47 joined together into 3D-frameworks via Cu centers (c). Source: Wu et al. [131]. Copyright 2018, Wiley-VCH Verlag GmbH& Co. KGaA, Weinheim.

103

104

3 Structural Chemistry of Metal-Oxo Clusters

{Nb24}

{Nb7}

{Nb10}

(a)

7

9

{Nb6} 12

(c)

(b) pH

(d)

Figure 3.24 Representation of (a) the most common PONbs as a function of pH, (b) [Nb14 O40 (O2 )2 H3 ]14− , (c) [(Nb7 O22 H2 )4 (UO2 )7 (H2 O)6 ]22− , (d) [((UO2 )(H2 O))3 Nb46 (UO2 )2 O136 H8 (H2 O)4 ]24− . Source: Martin et al. [134]. Copyright 2019, Wiley-VCH Verlag GmbH& Co. KGaA, Weinheim.

3.1.6

Polyoxomolybdates (POMos)

During the past three decades, numerous attempts have been made to understand, rationalize, and control the formation of gigantic polyoxomolybdates (POMos). MB compounds represent a family of giant mixed-valent POMos clusters, which are almost instantaneously obtained by the reduction of MoVI species in acid solutions (pH < 3) [135–137]. The MB species possessing a huge variety of topologies and compositions have been isolated by the group of Müller in one-pot reactions, such as the archetype cage-like {Mo132 } and wheel-type {Mo154 } and {Mo176 } as well as lemon-shaped {Mo368 } featuring inside {(MoVI )MoVI 5 } moieties as virtual building blocks constructed from a central pentagonal bipyramid {MoVI O7 } surrounded by five {MoVI O6 } octahedra [138–140] (Figure 3.25). The {(MoVI )MoVI 5 } units are abundant in several other well-known MB-derivatives, such as {MoVI 72 M30 } Keplerates (M = MoV , FeIII , VIV ) [141], and wheel-like {Mo248 }, {Mo128 Eu4 } as well as half-sealed {Mo120 Ce6 } elliptical wheel [142–144]. In majority of cases, the wheel-like MB polyanions are constructed from basic {Mo1 }, {Mo2 }, and {Mo8 } building blocks, thereinto {Mo2 } units are the reactive sites and possess strong affinity to amino acid ligands or available to be replaced by electrophiles. Thus, this was convinced by Cronin and coworkers by grafting cysteine and tyrosine onto {Mo154 } and exploiting synergy between coordinating lanthanides ions as symmetry breakers to produce MB with chiral frameworks decorated by amino acids ligands [145] (Figure 3.26).

3.1 Oxo Clusters of Transition Metal

[MoVI72MoV60O372 (CH3COO)30(H2O)72]42–

[Mo154(NO)14O448(H2O)70H28]14–

[Mo176O528(H2O)80H32]

Figure 3.25 Structures of cage-like and wheel-type MB species. Color code: O, red sphere; MoOx , rose, lime, gray, orange, and dark blue polyhedra. {MO154} Self-assembly

D-Amino acid

D7d

{MO1}

L-Amino acid

Centrosymmetric {MO11} C2 {MO8} Racemic Compound {MO2}

LnIII

Building blocks

Symmetry breaker

C2 {MO124Ce4}

Λ

D-Amino acid

Stereoselective Sythesis L-Amino acid

Chirality induction

Δ Chiral Mo Blue

Figure 3.26 Schematic of the stereoselective synthesis of chiral MB by using lanthanides as “Symmetry Breaker” and amino acids as chiral ligands. Source: Xuan et al. [145]. Licensed under CC BY 4.0

Besides, Xu and coworkers dedicated great efforts to the development of hybrid mixed-valent MoIV -containing POMos for their unusual molecular and electronic structures and potential application as redox bifunctional catalysis by the synergistic effect of both MoIV -reducing and MoVI -oxidizing active sites in the single POMos skeleton [146–148]. For instance, they adopted the bottom-up strategy by employing [MoIV 3 O4 (H2 O)9 ]4+ as precursor for generating the MoIV -based POMo [147].

3.1.7

Polyoxopalladates (POPs)

POPs, a blossoming class of precisely structurally determined Pd-oxo clusters act as ideal models to elucidate the working mechanism of Pd-based catalyst, have been provoking extensive attention [149, 150]. In 2008, Kortz and coworkers first reported the parent trideca-palladate with formula [Pd13 As8 O34 (OH)6 ]8− (labeled as Pd13 As8 ), which is composed of cubic

105

106

3 Structural Chemistry of Metal-Oxo Clusters

cage {Pd12 }, a central Pd center as guest and eight {AsO4 } capping units [13]. Their consecutive investigations on the substitution of arsonate capping units in Pd13 As8 with phenylarsonates and lone electron pair-containing selenites have resulted in the trideca-palladate derivatives [Pd13 O8 (PhAsO3 )8 ]6− and [Pd13 O8 (SeO3 )8 ]6− [151], while the employment of phosphates as capping ligands leads to the formation of star-shaped pentadeca-palladate [Pd15 O10 (PO4 )10 ]20− [152, 153]. Meanwhile, it has been identified that the central cavity inside the Pd12 and Pd15 archetypal shells can accommodate various metal ions [152–156].. The POPs polyanions are sensitive to pH in the process of self-assembled reactions in the aqueous solution, which is convinced by the isolation of cage-like [CuPd12 P8 O36 (OH)4 ]10− and dumbbell-like [H4 Cu2 Pd22 O16 (PO4 )12 (OH)4 ]20− polyanions by only pH adjustment [156, 157]. In addition, the slower hydrolysis–condensation process at room temperature could stimulate the formation of giant wheel-shaped clusters [Pd84 O42 (CH3 CO2 )28 (PO4 )42 ]70− , which consist of seven {Pd6 (μ4 -O)2 (μ2 -O)(PO4 )3 (OAc)2 }2 building blocks and possesses an overall D7d symmetry, with assistance of auxiliary ligands instead of POPs with low-nuclearity (e.g. 80 ∘ C) [158, 159]. Besides, Kortz and coworkers introduced Ag and Au into the POPs system evidenced by isolation of {Ag5 Pd15 } and [NaAu4 Pd8 O8 (AsO4 )8 ]11− , respectively [160] (Figure 3.27). More recently, they reported the first case of POP-based MOF, which comprises the trideca-palladate [Pd13 O8 (AsO4 )8 H6 ]8− grafted by Ba2+ centers and further connected with each other by means of rigid linear ligands such as p-carboxyphenylarsonate into 3D POP-MOF [161].

r St

+

ity ers div l a tur uc

{PdO4}

{L = RXO3}

Hos t–g ue st as Gu se est m me bly tal i on

Cube {MPd12L8} Cube {Pd13L8}

Dumbbell {Pd22L12}

Star {Pd15L10}

Bowl {Pd7L6}

Wheel {Pd84L42L′28}

Open-Shell {MPd12L6L′3}

Star {MPd15L10}

Figure 3.27 Schematic of controlled self-assembly processes of square-planar {PdO4 } units in the presence of external groups (RXO3 ) leading to POPs with a large structural variety. Source: Yang and Kortz [150]. Copyright 2018, American Chemical Society.

3.1 Oxo Clusters of Transition Metal

3.1.8

Polyoxotungstates (POTs)

Contrast to the behavior of other classical POMs systems, the polyoxotungstates (POTs) especially the lacunary POTs represent perfect candidates for the elaboration of functionalized POTs via the step-by-step procedure demonstrated by isolation of a library of lacunary, stable, and functionalizable POTs. 3.1.8.1 Transition-Metals-Substituted-POTs (TMSPs)

The oxygen-enriched surfaces of POTs fragments allow them with perfect nucleophilic ability to chelate almost all metal ions for constructing novel materials with different compositions and topologies and potential application ranging from catalysis, magnetism, medicine, photochemistry to materials science [4, 162–164]. 3.1.8.1.1 TMSPs via Step-by-Step Strategy

The step-by-step strategy is believed to be an efficient approach for establishing transition-metals-substituted-POTs (TMSPs) evidenced by reporting one of the novel Zr24 -based-POTs [Zr24 O22 (OH)10 (H2 O)2 (W2 O10 H)2 (GeW9 O34 )4 (GeW8 O31 )2 ]32− from the group of Yang by utilization of the preformed [GeW9 O34 ]10− as precursor reacting with ZrOCl2 leading to the largest [Zr24 O22 (OH)10 (H2 O)2 ] cluster by far in all the Zr-substituted POTs. 3.1.8.1.2 TMSPs via Self-Assembly Approach

Self-assembly of gigantic and discrete POTs by condensation of simple tungstate salts with assistance of heterometal ions is a topic of continuing interest. Cronin group and coworkers contributed greatly to this topic exemplified by the isolation of {W200 Co8 O660 } cluster with saddle-shaped structure [165–169]. 3.1.8.2 Inorganic–Organic Hybrid TMSPs

It is feasible to direct the formation of inorganic–organic hybrid TMSPs via the utilization of lacunary POTs as SDAs to assemble with TMs in the presence of organic ligands. 3.1.8.2.1 Hybrid TMSPs via Same Lacunary POTs Units

Remarkable progress is seen in the field of inorganic–organic hybrid structure chemistry of TMSPs. Yang group made great efforts into developing feasible and systematic hydrothermal strategy for creating hybrid TMSPs (Figure 3.28). In 2007, Yang and coworkers generated a series of hybrid Ni6 -substituted POTs {Ni6 (OH)3 (H2 O)6 (L)3 (B-𝛼-XW9 O34 )} (L = ethylenediamine (labeled as en), 1,2-diaminopropane (labeled as dap); X = P, As, Ge) comprising a {B-𝛼-XW9 O34 } segment attached by a planar core {Ni6 (OH)3 (L)3 O7 (H2 O)6 } [170, 171] (Figure 3.28c). Furthermore, the Ni6 -core can be covalently decorated with tripodal alcohol ligands [172] (Figure 3.28a). In addition to Ni2+ ions, Cu6 -core can also be introduced into this system which can further be served as the building blocks for the construction of infinite chains even 3-D frameworks [173, 174] (Figure 3.28b,e,g).

107

108

3 Structural Chemistry of Metal-Oxo Clusters

(a)

Decorated by various organic ligands

(b)

(c)

From Ni6 to Cu6

(d)

From Ni6 to Ni7

From clusters to extended structures

(f) (e)

(g)

Figure 3.28 View of structures based on Ni6 /Cu6 -substituted POTs. Source: Zheng and Yang [4]. Copyright 2012, The Royal Society of Chemistry.

3.1 Oxo Clusters of Transition Metal

3.1.8.2.2 Hybrid TMSPs via Synergistic Effect of Different Lacunary POTs Units

Contrasts to the structure directing effect of single vacant POT, synergistic directing effect of different lacunary POT segments can be employed to induce the formation of inorganic–organic hybrid TMSPs in dimer, oligomer, or polymer topologies. Yang and coworkers uncovered a Ni20 -based-POT comprising three kinds of lacunary POT units {B-𝛼-PW9 O34 }, {PW6 O26 } and {W4 O16 } joined together via different nickel cores {Ni6 O15 (OH)2 (H2 O)(dap)2 } and {Ni4 O10 (H2 O)2 (dap)2 } [175]. 3.1.8.2.3 TMSPs-Organic Frameworks

The design and synthesis of TMSPs-organic frameworks might merge the merits of POTs and MOFs into the resulting species with special properties. Yang and coworkers discovered a TMSPs-organic molecular cage with formula [Ni(en)2 (H2 O)2 ]6 {Ni6 (Tris)(en)3 (BTC)1.5 (B-𝛼-PW9 O34 )}8 (en = ethylenediamine, BTC = 1,3, 5-benzenetricarboxylate) composed of eight tris-grafted-{Ni6 PW9 } building blocks and 12 BTC linkers [176] (Figure 3.29). 3.1.8.2.4 Supramolecular POTs-Templated MOFs

Understanding the inclusion mechanism certainly enables to deliberate novel generations of supramolecular POTs-templated MOFs. Lu and coworkers carried out an investigation on the ternary Ag+/trz/Keggin-PW12 system and discovered the first example of 3D polycatenated framework with Keggin-PW12 cluster as template inside the nanocages [177] (Figure 3.30).

3.1.9

Polyoxotantalates (POTas)

The development of polyoxotantalates (POTas) is extremely torpid owing to their unique features of high surface-charge densities and alkaline dependence Figure 3.29 Structure of molecular cage in [Ni(en)2 (H2 O)2 ]6 {Ni6 (Tris)(en)3 (BTC)1.5 (B-α-PW9 O34 )}8 (en = ethylenediamine, BTC) [176]. Color code: O, red; N, blue; C, gray spheres; NiO6 , aqua; PO4 , rose polyhedra. All H atoms are omitted for clarity.

109

110

3 Structural Chemistry of Metal-Oxo Clusters

Figure 3.30 Arrangement of polyoxometalates and {Ag24 (trz)18 }6+ nanocages. Source: Kuang et al. [177]. Copyright 2010, Macmillan Publishers Limited.

(a)

(b)

(c)

that Ta-precursors tend to convert easily to gel-like species or precipitate and hetero-metal ions poorly exist under strong alkaline media [5, 178]. POTas are mainly restricted to Lindqvist-type hexatantalate [Ta6 O19 ]8− [9] and its derivatives [Ta10 O28 ]6− [179] as well as heterometal-doped [H5 Co8 Ta24 O80 ]15− [180], [Ti2 Ta8 O28 ]8− and [Ti12 Ta6 O44 ]10− [181]. The induction of POTs is a feasible and controllable way to access POTs-incorporated POTas [182–188]. Liu and coworkers first time reported the tris-(peroxotantalum)-substituted Dawsonand Keggin-type tetrameric-oligomer derivatives comprising an respective {Ta12 } and {Ta16 } core [188]. Recently, Su and coworkers produced the largest {Ta18 }- and {Ta18 Yb2 }-based polyanions formulated as [Ta18 P12 W90 (OH)6 (H2 O)2 O360 ]36− and [Yb2 Ta18 P12 W90 (OH)6 (H2 O)16 O360 ]30− , respectively. Both of them display significant photocatalytic water splitting activities [185] (Figure 3.31).

3.6 nm

3.2 Oxo Clusters of Main Group Metal

(a)

(b)

(c)

Figure 3.31 Polyhedral view of [Ta18 P12 W90 (OH)6 (H2 O)2 O360 ]36− . Source: Modified from Huang et al. [185]. Copyright 2016, The Royal Society of Chemistry.

3.2 Oxo Clusters of Main Group Metal 3.2.1

Introduction

Main group elements are the most abundant elements on the earth. The main group includes the elements (except hydrogen, which is sometimes not included) in groups 1 and 2 (s-block) and groups 13–18 (p-block). In Sections 3.2.2–3.2.4, an overview on main group borates, geminates, and aluminum elements will be presented. Both non-metal and metal-oxo clusters and inorganic and ligand-supported structures will be introduced. In addition to homometallic clusters comprising only one metal type, a large amount of heterometallic structures will be described.

3.2.2

Synthesis of Borates

A non-centrosymmetric structure is a prerequisite for crystals to exhibit efficient second-order nonlinear optical (NLO) effects that enable the manufacture of second-harmonic generating (SHG), electro-optical, and photorefractive devices [189]. As known, borates are excellent NLO materials because planar ionic groups with π-conjugated systems such as BO3 trigonal planes are responsible for the large SHG coefficients of these materials [190]. The most well-known example is the discovery of β-BaB2 O4 (BBO) [191], LiB3 O5 (LBO) [192], CsB3 O5 (CBO) [193], and Sr2 Be2 B2 O7 (SBBO) [194]. However, these NLO borate materials are made by solid-state reactions, no systematic investigation on NLO borates has been carried out under hydro(solvo)thermal conditions. Since 2003, Yang’s group have made considerable efforts to study hydro(solvo)thermal synthesis of borates. They have made some advances in the systems of inorganic, organic, and transitional metal complexes (TMCs) template borates. In addition, heterometallic structures of the same group metals are also included.

111

3.6 nm

3.2 Oxo Clusters of Main Group Metal

(a)

(b)

(c)

Figure 3.31 Polyhedral view of [Ta18 P12 W90 (OH)6 (H2 O)2 O360 ]36− . Source: Modified from Huang et al. [185]. Copyright 2016, The Royal Society of Chemistry.

3.2 Oxo Clusters of Main Group Metal 3.2.1

Introduction

Main group elements are the most abundant elements on the earth. The main group includes the elements (except hydrogen, which is sometimes not included) in groups 1 and 2 (s-block) and groups 13–18 (p-block). In Sections 3.2.2–3.2.4, an overview on main group borates, geminates, and aluminum elements will be presented. Both non-metal and metal-oxo clusters and inorganic and ligand-supported structures will be introduced. In addition to homometallic clusters comprising only one metal type, a large amount of heterometallic structures will be described.

3.2.2

Synthesis of Borates

A non-centrosymmetric structure is a prerequisite for crystals to exhibit efficient second-order nonlinear optical (NLO) effects that enable the manufacture of second-harmonic generating (SHG), electro-optical, and photorefractive devices [189]. As known, borates are excellent NLO materials because planar ionic groups with π-conjugated systems such as BO3 trigonal planes are responsible for the large SHG coefficients of these materials [190]. The most well-known example is the discovery of β-BaB2 O4 (BBO) [191], LiB3 O5 (LBO) [192], CsB3 O5 (CBO) [193], and Sr2 Be2 B2 O7 (SBBO) [194]. However, these NLO borate materials are made by solid-state reactions, no systematic investigation on NLO borates has been carried out under hydro(solvo)thermal conditions. Since 2003, Yang’s group have made considerable efforts to study hydro(solvo)thermal synthesis of borates. They have made some advances in the systems of inorganic, organic, and transitional metal complexes (TMCs) template borates. In addition, heterometallic structures of the same group metals are also included.

111

112

3 Structural Chemistry of Metal-Oxo Clusters

3.2.2.1 Inorganic Templated Borates

In 2004, a new potassium templated acentric manganese borate namely K7 [(BO3 ) Mn-{B12 O18 (OH)6 }]⋅H2 O (KMnBO) with NLO properties has been made [195]. For the first time the B and Mn atoms can link together in a remarkable way via both vertex-sharing and edge-sharing (Figure 3.32a). Recent years, they reported a series of alkaline borate, alkaline-earth-metal borates, as well as mixed-alkaliand alkaline-earth-metal borates from 1D chain to 2D layer to 3D frameworks [196–199, 201–205]. Nine to twenty-one membered-ring windows exist in these compounds. LiSr2 [B10 O16 (OH)3 ] is a 2D layer with nine-membered-ring windows

(b)

(c)

(a)

(d)

(g)

(e)

(h)

(f)

(i)

Figure 3.32 Inorganic templated borates: (a) view of the coordination environments for B and Mn in the structure unit of 1, showing the atom-labeling scheme and 50% thermal ellipsoids. Atom labels having “A” refer to symmetry-generated atoms. Source: Zhang et al. [195]. Copyright © 2004 Elsevier Inc. All rights reserved. (b, c) View of the 2D B–O layer constructed from B10 O16 (OH)3 clusters and the 3D B–O framework made of the B3 O7 SBUs along the c axis. Source: Wu et al. [196]. Copyright © The American Chemical Society 2013. (d) View 9R and 13R channels along the b axis. Source: Wang et al. [197]. Copyright © The American Chemical Society 2017. (e, f) 14R channels along the c axis and SHG intensity of Li2 CsB7 O10 (OH)4 and KDP versus particle size. Source: Huang et al. [198]. Copyright © The American Chemical Society 2019. (g) Framework structure viewed along the [001] direction showing 21R channels. Source: Wei et al. [199]. Copyright © 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim. (h, i) The projection of 3D supramolecular cages assembled from {Cu4 @B20 } clusters. Source: Wang et al. [200]. Copyright © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

3.2 Oxo Clusters of Main Group Metal

built up of B10 O19 (OH)3 9− cluster units (Figure 3.32b), while LiBa[B9 O15 ] exhibits a 3D framework with 12-membered-ring channels composed of B3 O7 cluster units (Figure 3.32c) [196]. Na2 B9 O15 (H2 O)(H3 O) possesses 13-ring channels constructed from different cluster units of B3 O7 and B6 O13 (Figure 3.32d) [197]. Li2 CsB7 O10 (OH)4 consists of unique bird-shaped [B7 O12 (OH)4 ]7− clusters with 14-ring pores (Figure 3.32e). Its SHG signal is 2.5KDP (KH2 PO4 ), with its short ultraviolet (UV) cutoff edge indicating that this crystal is a potential deep-UV transparent NLO material (Figure 3.32f) [198]. LiBa3 (OH)[B9 O16 ][B(OH)4 ] displays an acs-type net with large 21-ring channels (Figure 3.32g). The SHG measurement shows that it is a type I phase-matchable material with a strong SHG signal intensity of ∼3.1 times that of KDP [199]. In 2017, two new copper borates with mesoscale cubic supramolecular cages assembled from {Cu4 @B20 } clusters were reported. Owing to extra H3 BO3 molecules participated in building the supramolecular framework, H6 [(μ4 -O)-Cu4 @B20 O32 (OH)8 ]⋅34H2 O⋅8H3 BO3 has a larger cubic cage size and higher non-framework volume, leading to the cage size extended to mesoporous size set as a version of “H6 [(μ4 -O)Cu4 @B20 O32 (OH)8 ]⋅25H2 O plus” (Figure 3.32h) [200]. 3.2.2.2 Organic-Templated Borates

Organic solvents play important roles as templates as well as linkers in borates. In the presence of trans-1,4-diaminocyclohexane acting as an SDA, the first example of layered borates of [H3 N(C6 H10 )NH3 ][B5 O8 (OH)] was isolated [206]. The structure consists of layers of 3,9-membered boron rings constructed from pentaborate anion groups [B5 O8 (OH)]2− (Figure 3.33a). In 2007, the first 1D borate [NH3 CH2 CHCH3 NH3 ][B8 O11 (OH)4 ]⋅H2 O template by H2 enMe has been made [207]. Its structure consists of infinite open-branched borate chains built by [B3 O6 (OH)] units, onto which the [B5 O7 (OH)3 ] clusters are grafted (Figure 3.33b). By employing 1,4-diaminobutane as SDA, two new borates [H2 dab][B7 O9 (OH)5 ]⋅2H2 O and [H2 dab][B7 O10 (OH)3 ] were obtained [208]. The former contains the first example of organic templated heptaborate, [B7 O9 (OH)5 ] unit (Figure 3.33c), while the latter consists of [B14 O20 (OH)6 ] unit (Figure 3.33d) built by the dehydration of the [B7 O9 (OH)5 ] unit. In 2007, a novel hybrid borate B3 O4 (OH)⋅0.5C4 H10 N2 was first made under milder hydro(solvo)thermal conditions (Figure 3.33e) [209]. The co-templating approach has been proven to be an effective way to obtain inorganic microporous structures [212]. In 2015, a three-dimensional open-framework Na2 B10 O17 ⋅H2 en constructed from layered borate co-templated by inorganic cations and organic amines was reported (Figure 3.33f) [210]. Its framework displays a unique 5-connected net constructed by B5 O11 clusters and emits blue luminescence. Two years later, they reported a novel supramolecular magnesoborate framework with snowflake-like channels built by unprecedented huge B69 cluster cages (Figure 3.33g) [211]. Up to now, the discovery of B69 cluster sets a new record in cluster sizes among borates. 3.2.2.3

TMC-Templated Borates

Since 2004, the TMCs have been introduced into the structure of the borates [213]. Even though [Zn(dien)2 ][B5 O6 (OH)4 ]2 and [B5 O7 (OH)3 Zn(tren)] (dien = diethylenetriamine; tren = tris(2-aminoethyl)amine) are two pentaborates,

113

114

3 Structural Chemistry of Metal-Oxo Clusters

(c)

(b)

(d)

(a)

(e)

(f)

(g)

Figure 3.33 Organic templated borates: (a) view along the b-axis of 2 showing layers with 3, nine-membered boron rings. Source: Wang et al. [206]. Copyright © 2004 Elsevier Inc. All rights reserved. (b) View of the open-branched borate chains in Ref. [207]. Source: Pan et al. [207]. Copyright © 2007 Elsevier Inc. All rights reserved. (c) Polyhedral representation of the two FBBs in Ref. [208]. Source: Pan et al. [208]. Copyright © 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim. (d) Polyhedral view of a hybrid sheet formed by two strictly alternating borate helices of opposite chirality cohered by piperazine rings. L/R: left/right-handed helix, respectively. Source: Pan et al. [208]. Copyright © 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim. (e) Hybrid pillar-layered compound of B3 O4 (OH)⋅0.5C4 H10 N2 . Source: Liu et al. [209]. Copyright © 2007 Elsevier Inc. All rights reserved. (f) View of the packing structure of Na2 B10 O17 ⋅H2 en along the b-axis. Source: Wang et al. [210]. Copyright © The Royal Society of Chemistry 2015. (g) View of 3D supramolecular framework of [Mg7 @B69 O108 (OH)18 ]⋅6NH2 (CH3 )2 ⋅7NH3 CH3 ⋅5H2 O along the c-axis, showing the snowflake-like channels. Source: Wang et al. [211]. Copyright © The Royal Society of Chemistry 2017.

the former consists of isolated [B5 O6 (OH)4 ]− anion and [Zn(dien)2]2+ cation (Figure 3.34a), the latter is built by two distinct motifs of a [B5 O7 (OH)3 ]2− cluster and a supporting [Zn(tren)]2+ complex (Figure 3.34b) [214]. In 2006, three cobalt complex template borates, [Co(dien)2 ][B5 O6 (OH)4 ]2 , [B5 O7 (OH)3 Co(tren)], and [Co2 (teta)3 ][B5 O6 (OH)4 ]4 (teta = triethylenetetramine), have been reported [218]. Besides, a series of inorganic–organic hybrid cadmium borates with novel Cd-centred [Cd@B14 O20 (OH)6 ]2− clusters were obtained (Figure 3.34c) [215]. As a continued work to explore the structure-directing and extending effects of 1,6-diaminehexane (dah), a new zincoborate Zn4 (dah)(BO3 )2 (B3 O7 H2 )⋅H3 O⋅1.5H2 O was isolated [216]. It is a novel framework structure with zincoborate layers pillared by dah molecules (Figure 3.34d). They represent the first hybrid cadmium borates that exhibit 3D supramolecular open frameworks with different topologies. M(1,4-dab)[B5 O7 (OH)3 ] [M = Zn/Cd, 1,4-dab = 1,4-diaminobutane] and Co(1,3-dap)[B4 O7 ] (1,3-dap = 1,3-diaminopropane) are a series of 3d inorganic– organic hybrid borates constructed by 1D transition–metal complexes and different inorganic boron oxides (Figure 3.34e) [217].

3.2 Oxo Clusters of Main Group Metal

(a)

(b)

Cd O H B N C

(c)

b

a

a

c b

c

c a

(e)

(f)

Zn B O N C

Co B O N C

(g)

Figure 3.34 TMCs-templated borates: (a, b) ORTEP view of the asymmetric unit of [Zn(dien)2 ][B5 O6 (OH)4 ]2 and [B5 O7 (OH)3 Zn(tren)] showing the 30% probability displacement ellipsoids; atom labels with “A” refer to symmetry-generated atoms. Source: Wang et al. [214]. Copyright © 2004 Elsevier Inc. All rights reserved. (c) The anionic cluster of {(py)2 Cd@[B14 O20 (OH)6 ]}2− and {(AImH)2 Cd@[B14 O20 (OH)6 ]}. Source: Wei et al. [215]. Copyright © The Royal Society of Chemistry 2016. (d) View of the zincoborate layers pillared by DAH molecules. Color code: zinc, cyan; boron, green; oxygen, red; nitrogen, blue; carbon, gray. Source: Zhao et al. [216]. Copyright © The Royal Society of Chemistry 2014. (e, f) View of 3D framework of M(1,4-dab)[B5 O7 (OH)3 ] along the b-axis and Co(1,3-dap)[B4 O7 ] along the c-axis. Source: Zhi et al. [217]. Copyright © The American Chemical Society 2018.

3.2.2.4 Templated Synthesis of Aluminoborates

Being in the same group as B, Al exist tetrahedral, square pyramid, and octahedral coordination geometry. Al was introduced into borate system in 1973 giving porous aluminoborates (ABOs) analogous to zeolites [219]. Despite several ABO phases were revealed, their structures remain unknown limited by the synthetic difficulties of crystals suitable for structure determination [220–222]. Lin and coworkers made a series of PKU open-framework ABOs through the boric acid flux method [223–226]. Yang and coworkers chose aluminum isopropoxide as an Al source based on the following considerations: (i) the chiral Al center may form from three-coordinated Al(i PrO)3 transforming to four-coordinated AlO4 group via the hydrolysis of Al(i PrO)3 in the crystallization process; (ii) the characteristic of acentric B—O clusters formed in situ via self-polymerization of H3 BO3 can be transferred into the inorganic frameworks; and (iii) the synergistic combination between chiral AlO4 groups and acentric B—O clusters will not only greatly increase the likelihood of producing new acentric ABOs but also offer us the opportunity for choosing efficient NLO materials from these products. The field of ABOs experienced rapid development. Accordingly, the first examples of organic template ABOs with open frameworks were successfully made (Figure 3.35a) [227]. Notably, a cancrinite-type aluminoborate (BIT-1) with

115

116

3 Structural Chemistry of Metal-Oxo Clusters

(a)

Al B O

b c

(b)

a

Al B O

(d)

b c a

Al/GaO4 + B5O10 N,N’-Bis(3-aminopropyl)ethylenediamine

Diethylene- N-Ethyldiaminotriamine propane

(e)

(c)

Figure 3.35 Templated synthesis of aluminoborates: (a) views of the linkage of B5 O10 and AlO4 groups. Source: Rong et al. [227]. Copyright © 2009 American Chemical Society. (b) Framework structure of BIT-1 viewed along the [001] direction showing 24R, 9R, and 8R channels, respectively. Source: Cao et al. [228]. Copyright © The Royal Society of Chemistry 2016. (c) A series of alumino/galloborates templated by organic compounds. Source: Cheng and Yang [229]. Copyright © 2018 American Chemical Society. (d) Framework structure of BIT-1 viewed along the [001] direction showing 24R, 9R, and 8R channels, respectively. Source: Cheng et al. [230]. Copyright © 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim. (e) Top view of a 3-D bilayer showing the interlaced 9-MR windows along the c-axis. Color code: M/MO5 , purple; BO4 /BO3 , cyan. Source: Wei et al. [231]. Copyright © 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

gigantic 24-ring channels has been made under solvothermal conditions using Al(i PrO)3 as the Al source and amines as the SDAs (Figure 3.35b) [228]. Recently, a series of alumino/galloborates, including (H3 apea)6 [AlB5 O10 ]9 ⋅12H2 O (apea = N,N ′ -bis(3-aminopropyl)ethylenediamine), (Hdeta)2 (H2 deta)2 [AlB5 O10 ]3 (deta = diethylenetriamine), and (H2 edap)[GaB5 O10 ]⋅H2 O (edap = N-ethyldiaminopropane), were solvothermally synthesized (Figure 3.35c) [229]. In the presence of the Ba2+ cations, two series of novel porous layered metal borates, Ba3 M2 [B3 O6 (OH)]2 [B4 O7 (OH)2 ] (M = Al/Ga) (Figure 3.35d) [230] and Ba[MB4 O8 (OH)]⋅H2 O (M = Al/Ga) have been produced [232]. Furthermore, extended inorganic and organic templates were also introduced in ABO systems, resulting in a series of TMC-templated ABOs [233–236]. Especially, a 3D ABO open framework templated by 2D zinc-amine coordination polymer networks (Figure 3.35e), [Zn(dap)2 ][AlB5 O10 ] [231] and a novel ABO [In(dien)2 ][Al2 B7 O16 H2 ] [237] with large chiral cavities, has been first obtained.

3.2.3

Synthesis of Germinates and Borogermanates

3.2.3.1 Templated Synthesis of Germinates

Since the discovery of the first germanate with open framework in 1991 [238], a number of germanates with 2D, 3D, or isolated structures have been made by

3.2 Oxo Clusters of Main Group Metal

[Ge7O13(OH)2F3]3– (a)

[H2en][Ge(B4O9)] (h) H3BO3/en

Ni2+/dien/F– [Ge7O14F3]3– (b)

In3+/dien/F–

Ni2+/en or dap [Ni@Ge14O24(OH)3

]4–

[Ge3B2O9(OH)2]2+ (g)

GeO2

H3BO3/dach H3BO3/deta [(BO2)2(GeO2)4]2– (f)

K2B4O7 (c)

KBGe2O6 (d)

K2B4O5(OH)4 K2[Ge(B4O9)] (e)

Figure 3.36 View of a series of borogermanates constructed from different oxoboron cluster units. Source: Refs [239–243].

inorganic or organic templates. However, TMCs had not been applied to make germanates. [Ge7 O13 (OH)2 F3 ]3 ⋅Cl− ⋅[Ni(dien)2 ]2+ (Figure 3.36a) represents the first chainlike germinate templated by a TMC. Interestingly, the chirality of the Ni(dien)2 2+ ions transfer into adjacent germanate chains. This compound not only fills the gap left by the absence of 1D structures in germanate family but also is the first chiral germinate [244]. Ge7 O14 F3 ⋅0.5In(dien)2 ⋅0.5H3 dien⋅2H2 O (Figure 3.36b) is a 2D germanate template by indium complexes, in which the unique germanate layers were built by Ge7 O14 F3 clusters [245]. By using a racemic mixture of a NiL3 Cl2 (L = en/enMe) complex as a template, two novel germanates Ni@Ge14 O24 (OH)3 ⋅2Ni(L)3 (FJ-1a/1b) have successfully been made (Figure 3.36c). FJ-1 represents the first example of porous materials having Ge–Ni–Ge linkages. In addition, FJ-1 exhibits stereospecificity and chiral molecular recognition between chiral TMC template and inorganic structural motif. 3.2.3.2 Templated Synthesis of Germinates

KBGe2 O6 (FJ-9) is the first chiral zeotype borogermanate with 7-ring channels (Figure 3.36d) [239]. It is built from Ge2 O7 clusters and B2 Ge2 O12 ring clusters. FJ-16 (K2 [Ge(B4 O9 )]⋅2H2 O) is a 3D borogermanate with two pairs of interweaving double helical channels built by B4 O9 clusters and GeO4 units (Figure 3.36e) [240]. Since it is an acentric structure, it exhibits distinct NLO properties. Instead of templated by K+ ions in FJ-9 and FJ-16, (C4 N3 H15 )[(BO2 )2 (GeO2 )4 ] (FJ-17) is the first organically templated 3D borogermanate (Figure 3.36f) [241]. It is a novel zeolite-type framework with 1D 12 rings and large channels constructed from the cyclic Ge8 O24 clusters with eight rings and B2 O7 dimers. Meanwhile, another two organically templated 2D borogermanates, [H2 dach][Ge3 B2 O9 (OH)2 ] (dach = trans-1,4-diaminocyclohexane) and [H2 bappz][Ge3 B2 O9 (OH)2 ]⋅1.5H2 O (bappz = trans-1,4-bis(3-aminopropyl)piperazine) have been obtained (Figure 3.36g) [242]. Though its SBUs are the same as FJ-16, the structure of GeB4 O9 ⋅H2 en (FJ-18) is distinct from it because of different templates (Figure 3.36h) [243]. Here, the concept of host–guest symmetry and charge matching [246] was further been extended from tetrahedral and tetrahedral–octahedral frameworks to tetrahedral– triangular frameworks. Under solvothermal/surfactant-thermal, centrosymmetric open-framework (Hdima)2 [Ge5 B3 O15 (OH)] (dima = dimethylamine) containing an unusual basket-shaped Ge5 B3 O18 (OH) clusters was obtained [247]. CsBx Ge6−x O12

117

118

3 Structural Chemistry of Metal-Oxo Clusters

(x = 1) is a zeolite sodalite-type borogermanate with a high Ge/B ratio by partial boron substitution [248]. 3.2.3.3 Self-Polymerization and Induced Congregation of Lanthanide Germanate Lusters

Till now, no systematic investigation on Ln germanates has been carried out under mild hydrothermal method although several Ln germinates have been obtained by both the flux-growth method and high-temperature, high-pressure hydrothermal method in supercritical water. This is because of the fact that the conditions for solubility of traditional germanium and Ln source are incompatible in solution: acidic and basic solutions are favorable reaction systems for Ln2 O3 and GeO2 , respectively. After persistent exploration and efforts, Yang’s group chosen H2 E2 Ge2 O3 (E = –CH2 CH2 COO− ) with hydrophilic tails as the germanium source, successfully isolated frameworks containing Ge–O and Ln–O clusters under mild hydrothermal conditions: [Nd8 Ge12 (μ3 -O)24 E12 (H2 O)7 ]⋅13H2 O (FJ-19) and [Ln11 Ge12 (μ3 -O)24 E12 (pa)6 (H2 O)10 ]⋅(Cl,2OH)⋅nH2 O (Hpa = 2-picolinic acid, FJ-20, Ln = Pr/Nd/Eu/Gd, FJ-20, n = 19 for Pr/Nd, n = 15/14 for Eu/Gd) (Figure 3.37) [249]. These two novel 3D Ln–Ge–O cluster-organic frameworks contain high-nuclearity cage-shaped building blocks, Nd8 Ge12 for FJ-19 and Ln11 Ge12 for FJ-20. FJ-19 is a twofold interpenetrating enantiomorphic pair nets. To avoid interpenetration, a rigid chelating ligand with considerable steric hindrance, Hpa, was introduced into the reaction system. As a result, open framework of FJ-20 has been made. It is reasonable to believe that the present work will be important in expanding the study of germanates with open frameworks. In 2016, they report a series of lanthanide-germanate cluster organic frameworks based on {Ln8 Ge12 } clusters from one-dimensional chains to twodimensional layers and three dimensional [250, 251]. The successful preparation of these frameworks suggests the importance of the second ligands. Interestingly, the gradual replacement of active water sites located at equatorial and polar positions on the hypothetical {Ln8 Ge12 } core with oxygen or nitrogen atoms from organic ligands can be observed.

{Ln8Ge12}

HOOC

N

+

{Ln11Ge12}

Figure 3.37 View of the transformation from the twofold interpenetrating nets of FJ-19 built by Nd8 Ge12 cluster units to the noninterpenetrating networks of FJ-20 constructed from Ln11 Ge12 cluster units through introducing the second ligand. Source: He et al. [249]. Copyright 2009 American Chemical Society.

3.2 Oxo Clusters of Main Group Metal

3.2.4

Aluminum Oxo Clusters Hydrolysis and Condensation

Aluminum is the most widely studied one among the group 13 elements for its high natural abundance on the one hand and the technological importance together with environmental issues on the other hand. Beyond these direct applications of aluminum complexes as a sequestration agent, clay pillaring agent, other research work has focused on the synthetic of aluminum complexes as mineral mimics. The structure and reactivity of minerals are poorly understood for the surface of bulk samples is generally hard to determine [252]. However, the use of synthetic small molecule analogues for these structures enables the prediction of some of the natural processes taking place at the surface of natural minerals and the controlled nature of a laboratory and the smaller sizes of the cluster compounds can provide a clearer understanding of the more complicated processes in nature [253]. Although aluminum alkoxides have found widespread applications in sol–gel chemistry, the number of isolated and fully characterized examples of distinct metal oxido alkoxides is rather small [252, 253]. 3.2.4.1 Aluminum Oxo Clusters Isolated from Organic Solutions

Douglas has reviewed oligomeric aluminum oxo clusters in detail [253]. Hence, we herein restrict our discussion to clusters with nuclearities of 10 titanium atoms and larger. Schmidbaur and coworkers presented a study on the ammonolysis of aluminum triisopropoxide [254]. Unexpectedly, novel cluster [Al10 (μ4 -O)2 (μ3 -O)4 (μ-Oi Pr)2 (Oi Pr)16 (NH3 )8 ] as a result of partial hydrolysis rather than ammonolysis was observed. The cluster might be described as being composed of the hexanuclear core [Al6 (μ4 -O)2 (μ3 -O)4 (μ-Oi Pr)2 (Oi Pr)4 (NH3 )8 ] further coordinated by four molecules Al(Oi Pr)3 via the bridging oxygen atoms (Figure 3.38a). The Al10 O6 i Bu16 (μ-H)2 is a high aluminum content cluster produced from neat octakisisobutyltetraluminoxane (Al4 O2 i Bu8 ) at 80 ∘ C in six to eight hours followed by slow crystallization [257]. Murugavel and Kuppuswamy successfully use a phosphoric acid mono-aryl ester to synthesize cage-like molecules rather than cyclic or 1D polymeric phosphates. [Al10 {μ3 -O3 P(OR)}12 (μ3 -O)2 (Oi Pr)2 (thf)4 ]⋅6C7 H8 is made up of two [Al5 (O3 P(OR))6 ] units connected together across the crystallographic inversion center (Figure 3.38b) [255]. Through “controlled hydrolysis,” Kessler and coworkers reported an aluminum oxo cluster [Al11 (μ4 -O)2 (μ3 -O)2 (μ-O)2 (μ-On Pr)10 (μ-Oi Pr)2 (μ-ROH)2 (Oi Pr)8 (OR)] (R = i Pr, n Pr), which is made up of two connected pentanuclear cores connected together via a central [Al(μ3 -O)2 (μ-O)2 (OR)] group (Figure 3.38c) [256]. Kemnitz and coworkers studied the sol–gel chemistry of aluminum alkoxides in anhydrous hydrogen fluoride [258, 259]. Stabilized by the incorporation of additional external donor molecules like pyridine (py), they isolated [Al10 F16 (μ4 -O)2 (μ-Oi Pr)10 (py)4 ] [71]. Their further work included the successful isolation of [Al16 F20 (μ4 -O)4 (μ-Oi Pr)20 ] without external stabilizing donor molecules [259]. The former might be described as being built up by two interconnected [Al4 F5 (μ4 -O)(μ-Oi Pr)5 ] via fluorido ligands and two AlF3 (py)2 fragments building units and quite similarly the latter is composed of four times the same building block.

119

120

3 Structural Chemistry of Metal-Oxo Clusters

(a)

(b)

(c)

(d)

Figure 3.38 Currently, large aluminum oxo clusters isolated in organic solvents. Ball-stick-view of the molecular structure of [Al10 (μ4 -O)2 (μ3 -O)4 (μ-Oi Pr)2 (Oi Pr)16 (NH3 )8 ] (a), [Al10 {μ3 -O3 P(OR)}12 (μ3 -O)2 (Oi Pr)2 (thf)4 ] (b), [Al11 (μ4 -O)2 (μ3 -O)2 (μ-O)2 (μ-On Pr)10 (μ-Oi Pr)2 (μ-ROH)2 (Oi Pr)8 (OR)]. Source: Based on Murugavel and Kuppuswamy [255]. (c) and [Al16 F20 (μ4 -O)4 (μ-Oi Pr)20 ] (d). Source: Modified from Starikov et al. [256].

3.2.4.2 Aluminum Oxo Clusters Via Aqueous Synthetic Routes

A variety of spectacular structures of aluminum oxo clusters is accessible via aqueous synthetic routes. These aluminum hydroxide clusters fall into two broad structural classes. The first one is planar clusters. [Al13 (μ3 -OH)6 (μ-OH)12 (heidi)6 (H2 O)6 ] (NO3 )3 [260] (Figure 3.39a) possess the same metal ion arrangement of inner Al7 disk core as the pure inorganic Al13 compound [262]. A planar core of seven octahedral Al atoms is observed, all bridged by μ3 -OH ligands. The outer shell consists of six additional Al atoms, alternating above and below the plane of the inner core. The most familiar ones are derivatives of the Baker–Figgis–Keggin isomers that have tetrahedral central metals and the other class is characteristic cores of edge-shared Al(O)6 octahedra similar to the mineral brucite (Mg(OH)2 ) [252]. There are five Baker–Figgis isomers (α-, β-, γ-, δ-, and ε-forms) of the Keggin ions [261, 263–265]. The ε-isomer is the most stable one, which is accessible in combination with different counterions. In 1960, Johansson described the first ε-type Keggin structure in (NH4 )7 [Al13 O4 (OH)24 (H2 O)12 ](SO4 )7 in aqueous solution (Figure 3.39b) [261]. The basic structural motif is the central tetrahedral metal cation surrounded by 12 edge-sharing octahedral metal oxido units. In other words, such clusters might be described as being composed of four trinuclear metal oxido clusters bound to a central tetrahedrally coordinated metal atom. Polyaluminum clusters with metal atoms more than 20 are all built up by a combination of two Keggin-type {Al13 } clusters. The most simple combination is found in Al2 (μ4 -O8 )(Al24 (μ2 -OH)50 (H2 O)20 )(2,6-NDS)6 (H2 O)12.4 (Al26 ; 2,6-NDS = 2,6naphtalene disulfonate) [266]. The polycation observed in the Al26 compound

3.2 Oxo Clusters of Main Group Metal

(a)

(c)

(b)

(d)

(e)

Figure 3.39 Currently, large aluminum oxo clusters via aqueous synthetic routes. Ball-stick-view of the molecular structure of [Al13 (μ3 -OH)6 (μ-OH)12 (heidi)6 (H2 O)6 ]3+ (a), polyhedral view of polycations [Al13 O4 (OH)24 (H2 O)12 ]7+ (b), (Al2 (μ4 -O8 )(Al24 (μ2 -OH)50 (H2 O)20 )12+ . Source: Based on Johansson [261]., (c), [Al30 O8 (OH)56 (H2 O)26 ]18+ (d), and [Al32 (μ4 -O)8 (μ-OH)60 (H2 O)28 (SO4 )2 ]8+ (e).

undergoes clusters condense via peripheral coordinated water molecules to result in two bridging hydroxyl groups (Figure 3.39c). Four additional aluminum hydroxide moieties attached to the {Al13 } clusters give rise to [Al30 O8 (OH)56 (H2 O)26 ]18+ (Al30 ) (Figure 3.39d) aluminum polycations, which can be balanced by sulfate salt, disulfonate/chloride salts, and cucurbit[6]uril. Almost simultaneously, Taulelle and Nazar reported the first examples of Al30 compounds, compensated by nine sulfate ions [264, 267]. In 2006, Fedin and coworkers report Al30 supramolecular compounds with the organic macrocyclic cavitand cucurbit[6]uril [268]. In 2012, Forbes and coworkers focus on a directed supramolecular approach by using 2,6-NDS to crystallize the Keggin-type polyaluminum cations and successfully isolated another Al30 compound formulated as (Al2 (μ4 -O8 )Al28 (μ2 -OH)56 (H2 O)26 )(2,6-NDS)8 Cl2 (H2 O)34 [266]. Interestingly, it is possible to substitute two aluminum atoms of Al30 by tungsten atoms to give the [W2 Al28 (μ4 -O)8 (μ3 -O)4 (μ-O)4 O2 (μ-OH)48 (H2 O)24 ]12+ heterobimetallic cluster [269]. [Al32 (μ4 -O)8 (μ-OH)60 (H2 O)28 (SO4 )2 ]Cl2 (SO4 )7 and [{Al(IDA)(H2 O)}2 Al30 (μ4 -O)8 (μ-OH)60 (H2 O)22 ]Cl4 (2,6-NDS)4 (SO4 )2 (H2 IDA = iminodiacetic acid) {Al32 } represents the largest homometallic aluminum oxo clusters till now [270, 271].

121

122

3 Structural Chemistry of Metal-Oxo Clusters

The structure of is similar to that of Al30 , but the units are linked by two [Al(OH)2 (H2 O)3 (SO4 )]− groups [270] and [{Al(IDA)(H2 O)}2 ]+ [271] fragments (Figure 3.39e). It is noteworthy that additional ligand-stabilized metals might attach to the cluster core as demonstrated by the isolation of [{(Zn(NTA)H2 O}2 {Al(NTA)(μ-OH)2 }2 Al30 (μ4 -O)8 (μ3 -OH)54 (μ-OH)6 (H2 O)20 ](2,6-NDS)5 (H3 NTA = nitrilotriacetic acid) {Zn2 Al32 } [271].

3.3 Oxo Clusters of Lanthanides 3.3.1

Introduction

The exploration of high-nuclearity lanthanide clusters is still one of the most fascinating research frontiers in both inorganic chemistry and material science, owing to their esthetically pleasing structures as well as potential applications such as molecular magnets and coolers [272], luminescence [273], sensor [274], and catalysis [275]. However, the synthesis of high-nuclearity pure lanthanide clusters continues to be a challenge. The most important reason may be that lanthanide ions have variable and high coordination numbers as well as small energy differences in the various coordination geometries. Another reason is that lanthanide ions are pH sensitive and tend to form insoluble hydroxides or oxides. Until now, only a few examples of lanthanide clusters with the number of lanthanide ions no less than 10, such as Ln10 [276], Ln11 [277], Ln12 [278], Ln13 [279], Ln14 [280], Ln15 [281], Ln17 [282], Ln19 [278f, 283], Ln20 [284], Ln22 276b], Ln24 [285], Ln26 [286], Ln27 [287], Ln28 [288], Ln36 [289], Ln38 [290], Ln48 [290, 291], Ln60 [292], and Ln104 [293], have been documented. Therefore, it is urgent to summarize the recent progresses in the high-nuclearity pure lanthanide clusters, which may be helpful for the development of lanthanide chemistry. In recent years, lanthanide-based single-molecule magnets (Ln-SMMs) have become new star molecules and high-performance single-molecule magnetisms (SMMs), because of their potential applications in ultra-high density data storage, quantum computing, and spintronics [294]. The new developing lanthanide organometallic chemistry and lanthanide magnetochemistry have promote a new direction in SMM by combining the nonclassical organometallic synthetic approach with the traditionally distinct field of molecular magnetism. An advantage of organometallic chemistry is that the bridged ligand can vary over a much larger range of donor atoms from p-block, which is superior to the classical coordination chemistry. It endows crystal field changes and exchange interactions to be explored periodically. Until now, the magnetic properties of these cyclopentadienyl compounds have been studied in very few cases. In this feature article, we summarize and enclose our topic mainly in the pure lanthanide clusters with the number of lanthanide ions no less than 10 that have been reported recently, and the magnetic studies pertinent to the pursuit of materials with a large magnetocaloric effect (MCE). In addition, some of classical single-molecule magnets based on lanthanides are also discussed in this chapter to emphasize

3.3 Oxo Clusters of Lanthanides

the importance of studying the slow magnetic relaxation. And, some typical high-nuclearity lanthanide–transition–metal compounds are summarized, especially the hybrid core–shell nanoparticle fabricated by individual high-nuclearity lanthanide–transition–metal compounds. At the same time, the magnetic and luminescent properties of the above complexes will be also briefly discussed. In principle, one effective and general strategy for the syntheses of high-nuclearity lanthanide clusters is to control the excessive hydrolysis of lanthanide ions in the presence of supporting ligands. Therefore, the suitable organic ligands are extremely important since they act as not only the counter anions but also the organic shells to stabilize the inner cluster cores. We thus elaborately present the recent advances in the high-nuclearity lanthanide clusters synthesized by ligands. The ligands are further classified into O-donor ligands, N-donor ligands, multiple N,O-donor ligands, calix[n]arenes and other donor ligands such as Se- or C-donor ligands by considering the different coordination atoms. Based upon these sufficient discussions, the dependence of the synthetic strategy on the ligands, anions, and templating effects is also summarized.

3.3.2

High-Nuclearity Clusters of Lanthanides

3.3.2.1 High-Nuclearity Lanthanide Clusters Supported by O-Donor Ligands

According to the hard and soft acids and bases principle (HSAB), the hard lanthanide cations prefer to the coordination of the hard O-donor ligands. As we know, the carboxylate ligands and β-diketone ligands are extensively used in the syntheses of the multinuclear lanthanide clusters. The carboxylate group, [RCO2 ]− , is one of the versatile ligands and presents rich coordination modes with the metal ions. Accordingly, the [RCO2 ]− group is usually used in the syntheses of the transition metal clusters, which can act as not only the terminal ligands but also the bridge ligands. In 2014, Long and coworkers synthesized three Ln104 cluster complexes through the hydrolysis of rare earth perchlorate salts in the presence of acetate ligands [293]. Even though there is a little difference in the number and coordination modes of the outer acetate ligands, the inner spherical Ln104 cores of the above three clusters are very similar (Figure 3.40). These cluster cores express a rarely seen four-shell arrangement, abbreviated as Ln8 @Ln48 @Ln24 @Ln24 . From inside to outside, there are one platonic cube and three Archimedean solids, i.e. a truncated cuboctahedron, a truncated octahedron, and a rhombicuboctahedron. More importantly, the Gd104 cluster complex exhibits a large magnetic entropy change of 46.9 J/(kg K) at 2 K with ΔH = 7 T, which sets the record among the previously known lanthanide-exclusive cluster complexes. In continuation of the previous work, Long and coworkers discovered two isostructural Gd27 and Dy27 clusters by replacing the acetate ligands with longer propionate ligands [287]. These two Ln27 clusters are the largest odd-numbered lanthanide clusters reported up to now (Figure 3.41) [287]. Notably, these cage-like Ln27 structures are templated by eight CO3 2− and one ClO4− groups. Magnetic studies show that Gd27 complex has a large MCE, whereas Dy27 complex exhibits a slow relaxation of its magnetization.

123

124

3 Structural Chemistry of Metal-Oxo Clusters

Figure 3.40 The cluster core of Ln104 complex. Ln, olive; oxygen, red. For clarity, the terminal ligands are omitted. Source: Modified from Peng et al. [293]. 2014, © American Chemical Society.

Figure 3.41 The cationic Ln27 cluster templated by the ClO4 − anion. Ln, olive; oxygen, red; carbon, gray; chloride, yellow. The templated ClO4 − anion is shown in the space-filling mode. Source: Modified from Zheng et al. [287]. 2016, © The Royal Society of Chemistry.

In 2017, Long and coworkers reported a lanthanide-exclusive cluster {Gd140 }, with a diameter of 6.0 nm. As far as we know, this is the largest lanthanide-exclusive cluster complex to date. The gigantic 140-metal core showed a beautiful wheel-like structure and displayed 10-fold symmetry. Including all the bridging oxygen atoms, the metal-oxo core of the {Gd140 } cluster can also be considered as consisting of 20 well-known cubane-like {Gd4 O4 } building blocks and 50 defected cubane-like {Gd3 O4 } units sharing adjacent vertices (Figure 3.42) [295]. The {Gd140 } cluster represents a new member of the molecular wheel family and provides a possibility to explore the synthesis of large wheel-like lanthanide clusters on building blocks with available symmetries. Excellent work has also been done by Tong and coworkers. Using the chloroacetic acid as organic supporters, they synthesized two huge lanthanide clusters {Gd38 (ClO4 )6 } and {Gd48 (Cl)2 (NO3 )} in the presence of different anion templates

3.3 Oxo Clusters of Lanthanides

≈ 3.4 nm

≈ 6.0 nm

Figure 3.42 Crystal structure of the cation cluster of [Gd140 (CO3)20 (μ3 -OH)100 (CH3 COO)80 (LH3 )40− (H2 O)200 ]80+ . Color code: purple, Gd; red, O; gray, C; white, H. Source: Zheng et al. [295]. 2017, © American Chemical Society.

≈ 1.5 nm

(a)

(b)

(Figure 3.43) [290]. The cage-like {Gd38 (ClO4 )6 } cluster is composed of 12 vertexsharing Gd4 tetrahedrons and one Gd· · ·Gd pillar encapsulates, which encapsulates six ClO4 − anions in the cavity. Additionally, it is the first time to observe the rare linear M–O–M′ fashion and the unique 𝜇 8 -ClO4 mode in pure lanthanide complexes. By comparison, the barrel-like {Gd48 (Cl)2 (NO3 )} cluster consists of 12 vertex-sharing Gd4 tetrahedrons and six Gd5 pyramids. Interestingly, the {Gd38 (ClO4 )6 } cluster can convert into the {Gd48 (Cl)2 (NO3 )} cluster in the presence of Cl− and NO3 − anions. Thanks to the dominant weak antiferromagnetic interactions, the high Gd density, and the relatively compact crystal lattice, both clusters display large MCEs and may be used as molecular magnetic coolers.

Cl + NO3

Figure 3.43 The {Gd38 (ClO4 )6 } and {Gd48 (Cl)2 (NO3 )} complexes. Ln, olive; oxygen, red; carbon, gray; chloride, yellow. Source: Modified from Guo et al. [290]. 2013, © Wiley-VCH.

125

126

3 Structural Chemistry of Metal-Oxo Clusters

The controlled condensation of small lanthanide clusters, as another effect approach, has already been employed to obtain the high-nuclearity lanthanide clusters. By this method, Kögerler and coworkers synthesized two charge-neutral mixedvalence cerium(III/IV) clusters from a simple dinuclear cerium(III) precursor under aerobic condition [276b]. From their work, one can note that the different organic chelating ligands such as N-methyldiethanolamine and 1,1,1-tris(hydroxymethyl)ethane can result in different Ce10 and Ce22 clusters, respectively. Additionally, the structure-directing effects of the ligands are extremely important, even though the same precursor is used. Recently, Winpenny and coworkers employed this method to construct two novel cyclic phosphonate–carboxylate Ln10 cages (Ln = Dy and Gd), in which the cluster core was built by a nine-metal ring and a tenth inner lanthanide site [276h]. During the synthesis process of the high-nuclearity lanthanide clusters, the carboxylate ligands are found not only being derived from the reactants but also being generated in situ. For example, in 2013, Zhao and coworkers synthesized two discrete Ln24 clusters (Ln = Gd and Dy) by solvothermal reaction, in which the dimethylcarbamic acid anions were generated from in situ decomposition of DMF that played an important role in stabilizing the Ln24 clusters [285]. As shown in Figure 3.44, the backbone of the Ln24 cluster is composed of two kinds of Ln3 triangles and three kinds of Ln6 hexagons, possessing 18 tetragons and eight trigons. In addition, each tetragon anchors a 𝜇 4 -CO3 2− anion generated in situ. Importantly, MCE studies reveal that Gd24 is a promising candidate as the molecular magnetic cryogenic material.

Figure 3.44 (a) The discrete Ln24 cluster complex. Ln, olive; oxygen, red; nitrogen, blue; carbon, gray. (b) The skeleton of the Ln24 cluster core. Two kinds of Ln3 triangles are in yellow and cyan, respectively. Three kinds of Ln6 hexagons are in olive, red, and purple, respectively. Source: Modified from Chang et al. [285]. 2013 © The Royal Society of Chemistry. (a)

(b)

3.3 Oxo Clusters of Lanthanides

Similar to the carboxylate ligands, β-diketone is also a well-known O-donor ligand that has been extensively explored as the effective sensitizer to the lanthanide luminescence. In 2002, Wang and coworkers constructed the first Ln14 clusters by hydrolyzing Ln(acac)3 ⋅4H2 O (Ln = Tb, Eu; acac = acetylacetonato) in mixed CH2 Cl2 –hexane solvent with 2,2′ -dipyridylsulfide as the organic base [280a]. The Ln14 cluster cores contain an octahedral Ln6 unit sharing two opposing apexes with two square pyramidal Ln5 units. It should be noted that they are the first examples which have a “hollow” octahedral Ln6 unit without the support of a 𝜇 6 -oxo ligand. Similar Ln14 clusters were also synthesized by Roesky and coworker by means of partial hydrolysis of anhydrous lanthanum trichloride with potassium o-nitrophenolate [280d]. These Ln14 cluster complexes show a common structural feature that they are crystallized in the centrosymmetric space group. While in 2011, Li et al. found a new Ln14 cluster complex that is crystallized in the monoclinic chiral space group C2 [280b]. In general, the dimensions of the Ln-𝛽-diketone cages depend heavily on the ligands. Usually, small ligands, such as acetylacetonate and benzoylacetone, tend to form high-nuclearity lanthanide cages. However, Junk and coworkers reported an La12 cluster through the reaction of the bulky dibenzoylmethane with LaCl3 in the presence of an excess of trimethylamine [278a]. Interestingly, Ph2 acac was oxidized into phenylglyoxylate with hydroxide clusters as catalysts. Additionally, this La12 cage appeared to be templated by two CO3 2− anions by fixation of atmospheric CO2 . In addition to the β-diketone ligand, the introduction of the second ligand can form novel lanthanide cluster. In 2011, Schepers and coworkers introduced the cell-penetrating peptide monomer, 2-[{3-(((tert-butoxycarbonyl)amino)methyl)benzyl}amino]acetic acid hydrochloride (Pep-CO2 H⋅HCl), into the Ln–dibenzoylmethane system and obtained two novel Eu15 and Tb15 cluster complexes that were featuring intense red and green luminescence (Figure 3.45a) [272a]. In these Ln15 cluster cores, five [Ln4 (𝜇 3 -OH)4 ]8+ heterocubane subunits are fused together by an edge-sharing arrangement to form a pentagonal ring, which is stabilized by a centered 𝜇 5 -Cl− anion (Figure 3.45b). Due to the distinct biological compatibility of the Pep-CO2 ligand, the Eu15 and Tb15 clusters possess a pronounced capability to cell penetration. Polyalcohols are another important class of O-donor ligands for the construction of high-nuclearity lanthanide clusters. However, there are few reports on the use of these ligands. Recently, Long and coworkers got some positive results by utilizing polyalcohols as supporting ligands for lanthanide hydrolysis under hydro/ solvothermal conditions. With 1,2,3-cyclohexanetriol (H3 L) as organic ligand, the double-cage-like structures [(CO3 )2 @Ln37 (H3 L)8 (CH3 COO)21 (CO3 )12 (μ3 -OH)41 (μ2 -H2 O)5 (H2 O)40 ]⋅(ClO4 )21 (Ln = Gd, Dy) were obtained [296]. Zheng and coworkers also reported a nanoscopic cluster [Dy72 (mda)24 -(mdaH)8 (OH)120 (O)8 (NO3 )16 ]⋅(NO3 )8 by using N-methyldiethanolamine (mdaH2 ), which is a bisalcohol derivative of amine that can efficiently control the hydrolysis of Dy(III) [297]. 3.3.2.2 High-Nuclearity Lanthanide Clusters Supported by N-Donor Ligands

Compared with the hard O-donor, N atom is softer and more difficult to coordinate to the hard lanthanide cations. Usually, the multidentate N-donor ligands, such

127

128

3 Structural Chemistry of Metal-Oxo Clusters

(a)

(b)

Figure 3.45 (a) The discrete cationic Ln15 cluster complex. Ln, olive; oxygen, red; nitrogen, blue; carbon, gray; chloride, yellow. (b) Polyhedral representation of the skeleton of the Ln15 cluster templated by a μ5 -Cl− anion. Source: Based on Ren and Zheng [272a].

as 2,2′ -dipyridyl and 1,10-phenanthroline, tend to ligate to the lanthanide cations in the chelating coordination mode. For instance, Batten and coworkers reported three tridecanuclear lanthanoid clusters (Ln = La, Ce and Pr) supported by 1,10-phenanthroline (Figure 3.46a) [279]. As anticipated, 18 1,10-phenanthroline molecules as well as six auxiliary carbamoylcyanonitrosomethanide anions were coordinated to the lanthanide cations at the surface of the Ln13 cluster. These Ln13 clusters express unique core structures, in which the pseudo-spherical Ln13 core is divided into two parts, i.e. an inner lanthanoid metal ion and an outer distorted icosahedron. And interestingly, a novel [Ln(CO3 )6 ] moiety is also found in this Ln13 core. Based on this excellent work, Batten and coworkers further synthesized two larger tetradecanuclear polycarbonatolanthanoid clusters (Ln14 ) by using smaller Gd(II) and Dy(III) cations (Figure 3.46b) [280e]. The Ln14 cluster core can be viewed as a [Ln10 (CO3 )6 (OH)] moiety with another four lanthanoid cations on the periphery. In 2013, Tong research group synthesized a novel Dy10 cluster by utilitilizing 3,5-bis(pyridine-2-yl)-1,2,4-trizole as multidentate chelating ligands. This Dy10 cluster is constructed by four edge-sharing [Dy4 (𝜇 4 -O)] tetrahedral units [276c]. To maintain the continuity with the triazole–lanthanide system, Tong and coworkers adopted the pyridyl-substituted 3,5-bis[3-(pyrid-2-yl)-1,2,4-triazolyl]-pyridine (3,5-H2 bptp) as the organic ligand to construct a series of isostructural icosanuclear clusters by the solvothermal reaction [284]. There is an unexpected but interesting result that the 3,5-H2 bptp is methylated and oxidized in situ to 1-methyl-3,5-bis[3-(pyrid-2-yl)-1,2,4-triazolyl]pyridone (H2 MebptpO), and the Ln20 cluster shows a T 3 supertetrahedral geometry (Figure 3.47). Additionally, note that the benzoic acid here plays an important role in the formation of supertetrahedral clusters. These clusters, to our knowledge, are the first examples of supertetrahedral lanthanide oxide clusters to date. It is no doubt that this work will open a window to the design and synthesis of larger supertetrahedral clusters.

3.3 Oxo Clusters of Lanthanides

Figure 3.46 (a) The structure of cationic Ln13 cluster complex. The central Ln cation is in the polyhedral mode. Outer Ln, olive; central Ln, purple; oxygen, red; nitrogen, blue; carbon, gray; chloride, yellow. Source: Based on Chesman et al. [279]. (b) The structure of cationic Ln14 cluster complex. Ln on the periphery, olive; inner Ln, purple; oxygen, red; nitrogen, blue; carbon, gray; chloride, yellow. Source: Modified from Chesman et al. [280e]. 2012, © The Royal Society of Chemistry.

(a)

(b)

3.3.2.3 High-Nuclearity Lanthanide Clusters Supported by Multiple N,O-Donor Ligands

The ligands possessing the O-donor and N-donor simultaneously have been demonstrated to be good ligands for the synthesis of the 3d and 3d/4f clusters. Likewise, the use of multiple N,O-donor ligands as building blocks is also a very efficient strategy for the synthesis of high-nuclearity lanthanide clusters. Usually, these mixed-donor ligands can be classified into the amino acids and Schiff bases. It is worth mentioning that the amino acids, especially the natural amino acids, exhibit variable coordination modes under different conditions and thus are extensively employed to construct clusters. At low pH values, they will coordinate to the metal centers through only the oxygen atoms of the carboxylate moiety. The amino group remains protonated and uncoordinated, which can affect the final structure of the complex by electrostatic and/or steric effects. At higher pH values, however, chelation of a metal ion may be achieved by both the amino and the carboxylate groups.

129

130

3 Structural Chemistry of Metal-Oxo Clusters

Figure 3.47 (a) The structure of cationic Ln20 cluster complex. Ln, olive; oxygen, red; nitrogen, blue; carbon, gray; chloride, yellow. (b) The supertetrahedral skeleton of the Ln20 cluster core. Source: Based on Lin et al. [284]. 2013, © Wiley-VCH.

(a)

(b)

In 1999, Zheng and coworkers discovered a pentadecanuclear Eu(III) complex by the reaction of Eu(ClO4 )3 and tyrosine at the high pH value (c. 6) [281b]. The Eu15 core is a “wheel” of five corner-sharing cubanes, which is similar to that reported by Schepers, Bräse, and Roesky [281a]. A particularly salient feature of the Eu15 cluster is that a 𝜇 5 -Cl− ion is trapped at the center of the ring-shaped cluster core and serves as a template. Detailed researches demonstrate that both Cl− and Br− anions can direct the formation of Ln15 clusters. However, for the bigger I− anion, only the dodecanuclear lanthanide complexes can be isolated instead [281c], in which the Ln12 core consists of four vertex-sharing cubane-like [Ln4 (𝜇 3 -OH)4 ]8+ units and appears to be a square-shaped cyclic structure. Besides, one I− anion is located on each side of the square plane through the hydrogen bonding with four 𝜇 3 -OH− groups. However, a smaller face-capped octahedral cluster can be obtained by replacing the halogen anion with the NO3 − anion under the similar conditions. Therefore, it is concluded that not only the size but also the shape of the anions play an important role in the formation of lanthanide clusters. With continuous researches on Ln-amino acid clusters, in 2009, Zheng and coworkers synthesized bigger chiral Er60 complex by replacing L-tyrosine with

3.3 Oxo Clusters of Lanthanides

Figure 3.48 (a) The structure of cationic Er60 cluster complex. Ln, olive; oxygen, red; nitrogen, blue; carbon, gray. (b) The Er60 cluster core templated by μ6 -CO3 2− anions. Source: Based on Kong et al. [292]. 2009 © American Chemical Society.

(a)

(b)

L-threonine (Figure 3.48) [292]. Note that Er60 is the largest Er(III) cluster to date. The cluster core is templated by a 𝜇 6 -CO3 2− anion and can be viewed as a hexagonal cubane wheel built from 24 [Er4 (𝜇 3 -OH)4 ] units by sharing corners with three identical neighbors. If considering the [Er4 (𝜇 3 -OH)4 ] units as a 3-connected node, the Er60 cluster can be simplified as a truncated octahedral sodalite cage. Alternatively, this Er60 cluster can be viewed as a fascinating double-shell polyhedral structure, i.e. an outer truncated octahedral shell of 24 Er cations and an inner doubly truncated octahedral shell of 36 Er cations. While using D-tyrosine with L-threonine, a chiral Er60 enantiomer can also be obtained. It should be noted that they are the largest Er(III) clusters to date. Besides the natural amino acids, the pyridine carboxylate ligands such as nicotinic acid (HNIC) and isonicotinic acid (HIN) are also proved to be good ligands for the construction of lanthanide clusters. For the syntheses of high-nuclearity clusters, the “surface modification” strategy proposed by Xue and coworkers shows two advantages. The incorporation the surface modifiers such as NO3 − anions into the cluster core will lead to the decrease of the

131

132

3 Structural Chemistry of Metal-Oxo Clusters

Figure 3.49 The skeleton of the Ln26 cluster templated by I− anion. For clarity, the terminal ligands are omitted. The Ln cation of trigonalbipyramid, olive; the Ln cation sitting at the triangular facets, orange; oxygen, red; nitrogen, blue; iodine, purple. Source: Based on Gu and Xue [286a]. 2007 © American Chemical Society.

positive charge of the lanthanide cluster, and thereby to stabilize the whole complex. Moreover, the expansion of the cluster surface will permit more lanthanide cations to be incorporated into the cluster core. Following the “surface modification” strategy, they reported a series of Ln26 complexes by the reaction of Ln2 O3 , AgI, HIN, and HNO3 under hydrothermal condition (Ln = Dy and Er) [286a, b]. Among these clusters, two kinds of Dy26 complexes are of great interest (Figure 3.49). In these two complexes, the Dy26 cores are almost the same. Five Dy4 cubane-like clusters are assembled in the edge-sharing mode into a Dy20 trigonal bipyramid. Furthermore, other six Dy cations sitting on six triangular facets of the trigonal bipyramid, respectively, are connected to the Dy20 fragment by sharing 𝜇 3 -O atoms and the oxygen atoms from the NO3 − ligands. Thus, a Dy26 cluster core is formed. However, there is obviously difference in these two Dy26 complexes upon expansion. In one complex, two Dy26 cluster units are linked by two Dy4 cubane-like units through the coordination of IN ligands. Instead, in the other complex four Dy26 cluster units are connected together by the IN linkers to form an unusual rhombic tetramer. Interestingly, though the Ag+ ion is absent from the final structure, a free I− anion is encapsulated into the cavity of the cluster as template. Similar results were also acquired by Xu and coworkers. Replacing AgI and HNO3 with Mn(OAc)2 ⋅4H2 O and HCOOH, they achieved two discrete Ln26 complexes [286c]. The Ln26 cluster cores are similar to those discussed above, while the template anion is the CO3 2− anion rather than I− anion, and the surface modifier is CO3 2− anion rather than NO3 − anion. For the whole structure, a distinct change is that the number of the IN ligands decreases due to the introduction of the CH3 COO− and CO3 2− anions to the cluster surface. It means that the steric hindrance between the Ln26 cluster units also decreases and the cluster units may

3.3 Oxo Clusters of Lanthanides

be connected together into polymeric one-dimension (1D) chains, two-dimension (2D) nets, and three-dimension (3D) frameworks. In place of Mn(OAc)2 ⋅4H2 O with Zn(OAc)2 , they also prepared two 3D coordination frameworks based on the nanosized Ln26 clusters. As anticipated, the Ln26 units are connected together through the coordination of the N atoms of the pyridine rings to the Zn(II) cations [286d]. Further researches show that besides the Zn(II) cations, the Cu(I) and Ag(I) ions could also link these huge Ln26 cluster units into 3D frameworks [286e], even if the NIC anions substitute for the IN ligands. Certainly, aggregation of the Ln26 units could also come true without the introduction of the second 3d or 4d metal ions. For example, in 2011 Koner and coworkers reported a 3D framework based on the Gd26 units, which were linked by NIC ligands [286f]. This Gd26 complex has good capability of catalyzing the epoxidation of various olefinic substrates including α,β-unsaturated ketones in heterogeneous media. The above results demonstrate that the “surface modification” strategy is an effective approach to synthesize high-nuclearity clusters. However, this strategy exposes a shortcoming, namely, the products may be mixed with small tiny particles of rare earth oxide due to the fact that the source of the lanthanide cations is from the rare earth oxide, which will have some influence on the further performance studies. As for the traditional ligand-controlled hydrolysis, it is usually performed at the low temperature (no more than 100 ∘ C) with the strong bases such as NaOH. Under this condition, the huge lanthanide clusters tend to precipitate immediately due to the low solubility. Taking the above factors into consideration, we suggest a plan to control the hydrolysis of lanthanide cations at the higher temperature with weaker bases. Higher reaction temperature will increase the solubility of lanthanide cluster and avoid the immediate precipitation. Weaker bases will decrease the speed of the hydrolysis of lanthanide cations, which will enable the crystallization of lanthanide clusters. NaN3 is a good candidate as the inorganic base since it has good solubility in water and can be hydrolyzed into NaOH with release of N2 at high temperature. Following the above idea, we made an attempt on the synthesis of the highnuclearity lanthanide clusters. Under hydrothermal condition, we gained two polymeric 36-metal pure lanthanide nanosize clusters Gd36 and Dy36 (Figure 3.50a), which were on behalf of as the highest nuclearity Gd(III) and Dy(III) clusters reported at that time [289]. The spheric Ln36 core can be viewed as the aggregation of one Ln24 wheel and two identical tripod-like Ln6 units (Figure 3.50b). Different from the reported metallacrowns or wheels, in this Ln24 unit six tetrahedral Ln4 clusters adopt the up and down arrangement and form a cyclohexane chair-like structure. It should be noted that these two Ln36 complexes are not isolative. They extend the structures through the coordination of the IN− ligands with the Ln(III) cations to form a square layer. These polymeric frameworks are rarely seen because the high-nuclearity lanthanide clusters may be either prevented from the further aggregation by big hydrophobic organic ligands or surrounded by supporting ligands without further coordination sites. Magnetic studies show that the Gd36 complex possesses a large MCE of 39.66 J/(kg K), and the Dy36 complex exhibits slow relaxation of the magnetization.

133

134

3 Structural Chemistry of Metal-Oxo Clusters

(a)

(b)

+

+

(c)

Figure 3.50 (a) The 36-nuclearity Gd(III) cluster. (b) For clarity, only the bridge atoms and Gd(III) cations are kept. Green, Gd; red, oxygen; blue, nitrogen. (c) Illustration of the structure of the Gd36 cluster. The Gd(III) cations of the wheel-like Gd24 units are in red. The Gd(III) cations of the two tripod-like Gd6 units are in blue. Source: Wu et al. [289]. 2013 © The Royal Society of Chemistry.

Using the Er cations instead of Gd or Dy cations, we achieved a polymeric tube-like Er48 complex (Figure 3.51) [291a], whose cluster core is alike to the Gd48 cluster reported by Tong and coworkers [290]. The cationic Er48 tube is not a regular cylinder and can be divided into three parts, i.e. one ring-like Er12 unit and two identical wheel-like Er18 units. Different from the cyclohexane chair-like Ln24 wheel discussed above, six tetrahedral Er4 clusters are joined together through the corner-sharing mode and adopt the parallel arrangement to construct an Er18 wheel. At the “belly” of the tube, six V-shape Er3 subunits link to each other in the corner-sharing mode to generate an Er12 ring. Further, two wheel-like Er18 units sandwich one ring-like Er12 unit to form the ultimate Er48 tube. A particularly salient feature of this Er48 cluster is that two kinds of anions are trapped within the nanosize tube. At each side of the tube, one spherical Cl− anion is found at the center of the Er18 wheel, which is hydrogen bonding with six 𝜇 3 -OH− groups from six different Er4 O4 tetrahedra, respectively. In the middle of the tube, one triangular NO3 − anion is also found hydrogen bonding with six 𝜇 4 -OH− groups, where each oxygen atom of the NO3 − anion is anchored by two 𝜇 4 -OH− groups. We think the double- or multi-anion templating effects will throw a light on the research of syntheses of high-nuclearity lanthanide clusters. Through the coordination of the IN− and N3 − ligands, the Er48 clusters form a rarely seen layer structure, which is the coordination polymer based on highest nuclearity lanthanide clusters to date.

3.3 Oxo Clusters of Lanthanides

Cl–

+ NO3–

+ Cl–

Figure 3.51 The sketch map of the Cl− and NO3 − anions templating the formation of the nano-size Er48 tube through hydrogen bonding. Light green, Er(III) cations; green, chloride; red, oxygen; blue, nitrogen; black, hydrogen; purple, Er(III) cations in the ring-like Er12 unit. Source: Wu et al. [19]. 2014 © The Royal Society of Chemistry.

A similar 2-D coordination polymer constructed from drum-like pure Ho48 clusters was also reported by Xu and coworkers [291b]. However, an obvious feature is that the Ho48 cluster unit is an anion rather than a cation, which is also rarely seen in the lanthanide clusters. For synthesis of lanthanide clusters, another kind of extensively used N,O-donor ligands is Schiff bases, which contain a imine group formed via the condensation of a primary amine and an active carbonyl such as aldehydes and ketones. Salicylaldehyde and its derivatives are usually chosen to condense with various amines not only because they are cheap and easily available but also the phenolic hydroxyl group can chelate the 3d or 4f metal cation along with the N-atom of the imine group. Zhang and coworkers have reported a few lanthanide clusters based on the Schiff bases obtained from the o-vanillin moiety. For example, two kinds of Ln10 clusters were prepared by the reaction of 2-(((2-hydroxy-3methoxyphenyl)methylene)amino)-2-(hydroxymethyl)-1,3-propa-nediol and the lanthanide acetate salts [276d]. In the Dy10 cluster, there are two sets of vertexsharing Dy3 triangles. However, if the lanthanide cation is changed into Pr(III) or Nd(III) cation, an unprecedented assembly of Ln10 aggregates containing two Ln5 pentagons can be isolated [276e]. The formation of Ln5 subunit is templated by the 𝜇 5 -CO3 2− via spontaneous fixation of atmospheric carbon dioxide. If using the Schiff base condensated from the o-vanillin and diethylenetriamine, a dodecanuclear dysprosium wheel was obtained [278g]. For clarity, the large Dy12 cluster core can be viewed as the aggregation of six vertex-sharing Dy3 triangle motifs. Six vertexsharing Dy3 subunits are linked by six CO3 2− anions to form a Dy12 wheel with an approximately equilateral hexagonal six-membered ring. The adjacent capped

135

136

3 Structural Chemistry of Metal-Oxo Clusters

Dy(III) ions of the Dy3 triangles are alternately arranged in the up and down fashion to form a staggered Dy12 double crown, which is much different from the Dy12 cluster reported by Tong. In comparison, the Dy12 cluster reported by Tong and coworkers is spherical, which can be regarded as the fusion of four vertex-sharing cubane-like Dy4 units [278c]. Using the multichelating 1,10-phenanthroline-2,9-dicarbal-dehyde dioxime (H2 phendox) as the ligand, Tong and coworkers discovered a Dy11 cluster complex featuring two Dy4 cubanes and two face-sharing defective Dy4 cubanes [277]. In this synthesis, the H2 phendox was oxidized into the 1,10-phenanthroline-2,9dicarboxylic acid (H2 phenda) in situ. This Dy11 complex features a 3D supramolecular architecture stabilized by offset π⋅⋅⋅π interactions. Big polygon channels can be found along the c axis with the dimension of 10 × 10 Å2 . Additionally, this guest-free supramolecular framework has good capability of absorbing CO2 at 195 K and 1 atm, with the uptake of 86.4 cm3 /g, which shows a promising application for the separation of CO2 . 3.3.2.4 High-Nuclearity Lanthanide Clusters Supported by Calix[n]arenes Ligands

Calixarenes and their derivatives, representing a class of cyclic oligomers, provide versatile multidentate coordination sites for constructing multinuclear lanthanide complexes. The calix[n]arenes and its derivatives stand out among various candidates by affording multidentate and nucleating coordination modes (Scheme 3.1). Calixarenes have been proved to show great advantages in constructing polynuclear 4f complexes [298]. Intriguingly, the sulfurated calixarenes analogous, named the thiacalix[n]arenes, also demonstrate considerable potentials in behaving as a bifunctional multidentate ligand to coordinate to the metal ions, because of not only the phenol oxygen atoms on the lower rim position but also the sulfoxide or sulfone groups on the cyclic framework [298a]. For example, p-tert-butylthiacalix[4]arene (H4 BTC4A) could provide four phenoxyl groups and four bridge sulfur atoms to bridge a large number of M(II) metal ions, and thus construct polynuclear compounds like Ni32 [299] and Co32 [300] species. Simultaneously, each H4 BTC4A molecule binding to four divalent M(II) R

X

X

Calix[n]arene

OH OH HO

R

R

Thiacalix[n]arene

OH X

X

R

Scheme 3.1

n

X

n

CH2

≥4

S

R

4,6,8

tBu

Sulfmylcalix[n]arene

SO

4,

tBu

Suifonylcalix[n]arene

SO2

4,

tBu

Calix[n]arenes and its derivatives.

3.3 Oxo Clusters of Lanthanides

cations was found to form a MII 4 -BTC4A entity as an secondary building unit (SBU). Moreover, the extension can be continued by various bridging ligands and further fascinate high-nuclearity compounds in the end. The in situ formed ditetrazolate ligands are able to generate versatile polynuclear compounds, ranging widely from sandwich-like Co8 [301], saddle-like M12 (M = Co, Ni) [302], elongated octahedral M24 (M = Co, Fe) coordination cages [303], large Co24 metallamacrocycles [304], discrete giant spherical Co32 clusters [300] to tetragonal-prismatic Co32 nanoscale cages [305]. In 2006, Kajiwara et al. reported a series of wheel-like octalanthanide clusters constructed by the p-tert-butylsulfonylcalix[4]arene (H4 BSC4A) ligand with the help of acetate groups [306]. Under similar conditions, while replacing the mono-carboxyl acetate ligands with the di-carboxyl malonate ligands, the larger wheel-like Ho12 cluster compound has been synthesized [307]. Fraser, Massi, and Ogden reported the Dy12 and Dy19 clusters supported by a tetrazole-functionalized calix[4]arene. A rod-like Dy12 cluster complex was obtained in the presence of benzoate anions as co-ligands. The cluster core was found to be built by one Dy4 cubane and two Dy6 trigonal bipyramid in the vertex-sharing mode. While the benzoate anions were replaced with acetate anions, the elongated Dy19 clusters can be generated under similar condition. Three central Dy6 trigonal bipyramids and two side Dy4 cubanes were linked together in the vertex mode to form a Dy19 rod. M.C. Hong has also taken continuous interests in constructing various metal–organic supramolecular complexes or clusters by the calixarenes ligands and its derivatives [308]. In this part, we highlight the journey from 2010 until date in pursuit of lanthanide clusters constructed by calixarene and its derivatives, and in turn, summarize the synthetic strategy, structural characteristics, and their potential properties. Herein, M.C. Hong group has taken long-term interests in constructing highnuclearity clusters by the phosphate and phosphonate ligands. Compared with the carboxylate ligands, they afford abundant oxygen atoms available, diverse anionic forms and richer coordination modes, which have been confirmed to be advantageous candidate. While used as co-ligands in constructing polymetallic compounds, a small library of high-nuclearity transition metal compounds, such as chair-like Ni8 , tri-capped trigonal prismatic Co9 , sphere-shaped M12 (M = Ni or Co), drum-like Mn14 , diamond-like Mn16 clusters, helmet-shaped Co20 , and truncated octahedral Co24 cages, have been generated as well as two large alkali-metaltemplated Na2 Co24 and KCo24 clusters [h, 308c–e]. Therefore, they have been considered as proper ligands in the formation of high-nuclearity lanthanide clusters. As anticipated, three attractive high-nuclearity neodymium compounds; Nd10 , Nd11 , and Nd19 clusters have been solvothermally synthesized based on sulfonylcalix[4]arene and methylphosphonic acid (MePO3 H2 )/phenylphosphonic acid (PhPO3 H2 ) ligands (Figure 3.52) [308j]. The Nd10 complex features an oval-shaped Nd10 core, where four Nd3 -BSC4A subunits are linked together by sharing the Nd(III) ions to constitute a –[Nd–O]– repeating unit with a large 16-membered wheel-like framework. Different from the BSC4A4− ligands in Nd10 complex, three cone-like BSC4A4− ligands in Nd11 complex coordinate to four Nd(III) ions to generate Nd4 -BSC4A subunits. The nearby Nd4 -BSC4A subunits are connected by

137

138

3 Structural Chemistry of Metal-Oxo Clusters

Nd10

Nd11

Nd19

Figure 3.52 Molecular structure of complexes Nd10 , Nd11 , and Nd19 ; hydrogen atoms are omitted for clarity. Source: Su et al. [308j]. 2016, © The Royal Society of Chemistry.

sharing one Nd(III) ion and two PhPO3 2− ligands to form the body of the rugby, whose vertexes are further capped by two NdIII ions, respectively. In comparison, the Nd19 complex has a large irregular Nd19 core paneled by in situ generated carbonato and formate anions. It should be noted that the Nd19 complex represents the highest-nuclearity lanthanide cluster based on H4 BSC4A ligand to date. Meanwhile, the Nd10 , Nd11 , and Nd19 complexes also exhibit luminescent and magnetic behaviors. In addition, while replacing the sulfonylcalix[4]arene with p-tert-butylcalix[8] arene (H8 TBC8A), two dumbbell-like Ln10 clusters Pr10 and Nd10 can also be obtained in the presence of the PhPO3 2− ligands [308f] (Figure 3.53). One fully deprotonated TBC8 A8− anion in a double-cone conformation chelates four rare earth cations by eight oxygen atoms derived from lower-rim phenolic groups. Such two tetranuclear Ln-calixarene entities are linked together in a head-to-head style by one HCOO− anion, one HCO3 − anion, two Ln(III) ions, and four PhPO3 2− ligands. To the best of our knowledge, these two dumbbell-like Ln10 clusters represent the largest pure lanthanide aggregates supported with H8 TBC8A ligands.

Figure 3.53 Molecular structure of complexes Ln10 ; hydrogen atoms are omitted for clarity. Source: Xiong et al. [308p]. 2012, © The Royal Society of Chemistry.

Ln Ni Cl S Na O N C

3.3 Oxo Clusters of Lanthanides

3.3.2.5 High-Nuclearity Lanthanide Clusters Supported by Other Donor Ligands

Compared with O and N elements, the chalcogen elements such as S, Se, and Te show more electronegativities. Therefore, the lanthanide complexes constructed by the chalcogen–donor ligands will present their novel and unique physical and chemical properties, which have been arousing great interest in exploring this kind of high-nuclearity lanthanide clusters. In 2008, Brennan and coworkers synthesized the isostructural Ln28 cluster complexes (Ln = Pr and Nd) by the reaction of the Ln(SePh)3 with NH4 F in the presence of pyridine (Figure 3.54a) [288]. The Ln28 cluster has a central Ln4 core, where each Ln cation is 12-coordinated by the F− anions. The Ln4 core is further encapsulated by the Ln24 layer through the bridging F− anions to form a spherical Ln28 cluster core. The remaining coordination sites of the external Ln cations are saturated by the pyridine and SePh− ligands. In the presence of Na ions, Brennan Figure 3.54 (a) The skeleton of the Ln28 cluster. For clarity, the terminal ligands are omitted. The Ln cation of inner tetrahedron, orange; the outer Ln cation, olive; selenium, purple; fluoride, cyan. Source: Modified from Romanelli et al. [288]. (b) The skeleton of spheric Ln17 cluster. For clarity, only bridging atoms are retained. The Ln central cation, orange; the outer Ln cation, olive; selenium, purple. Source: Based on Moore et al. [282].

(a)

(b)

139

140

3 Structural Chemistry of Metal-Oxo Clusters

and coworkers also selectively precipitated a series of Ln17 complexes (Ln = Ce, Pr and Nd) (Figure 3.54b) [282]. By contrast, without the Na ions, either Ln8 clusters or polychalcogen species can be found under similar condition. An 8-coordinated Ln cation sits at the center of Ln17 cluster core, which is surrounded by 16 outer Ln cations to form the cluster surface. Besides being coordinated to the Se2− anions, the Ln cations also ligate to the SePh− and pyridine ligands. As anticipated, the above Nd28 and Nd17 complexes both exhibit intensive NIR emissions with the quantum efficiency up to 41% and 35%, respectively. In addition to the ligands containing chalcogen elements, the monocyclopentadienyl (CP) ligand can also facilitate the formation of the high-nuclearity lanthanide clusters. For example, treatment of the THF adducts [CpSmCl2 (thf)3 ]with hot toluene and simultaneous removal of the evolved tetrahydrofuran by distillation led to the Sm12 complex [278d]. Different from cyclic Ln12 clusters [278c, g, 281c], herein 12 Sm cations build up an icosahedral cluster skeleton, which is rarely seen in lanthanide clusters complexes. The Sm cations are bridged by the 24 𝜇3 -Cl anions. Notably, the 60 C atoms of the 12 Cp ligands form a truncated icosahedron, which is little alike to the C60 fullerene.

3.3.3

Monometallic Lanthanide-Based Single-Molecule Magnets

Further exploring the magnetic properties of lanthanide organometallic chemistry remains as an enormous challenge. Recently, one of the most exciting discoveries in lanthanide magnetochemistry is the SMMs, a phenomenon in which coordination compounds show slow relaxation of the magnetization in a manner that does not rely on cooperative interactions across magnetic domains. The slow magnetic relaxation properties of single-molecule magnets endowed them with capability of storing magnetic information, quantum computing, and spintronics at the level of single molecules and even single atoms. However, the low operating temperature relates to the blocking temperature (T B ), and the effective energy barrier (U eff ) restricts their use in application. The design of a high spin ground state and a large magnetic anisotropy are helpful to enhance the U eff . Nevertheless, in most of reported polynuclear systems [309], the compensation caused by different spin centers has dramatically reduced the magnetic anisotropy, resulting in a negligible improvement of the SMM properties. Based on these concerns, LnIII ions are introduced for their large magnetic anisotropy from the strong spin–orbit coupling [310]. The lanthanide elements, especially dysprosium, play a crucial role in the development of potential nanoscale applications of SMMs. The cyclopentadienyl ligands play a vital role in the understanding of lanthanide organometallic chemistry because the bonding in lanthanide compounds is predominantly electrostatic. And, as an interesting subgroup, the monometallic SMMs or single-ion magnets (SIMs), i.e. the dysprosium metallocene attracted more and more attention [311]. Layfield and coauthors reported that two cyclopentadienyl ligands can provide Dy3+ with a strongly axial coordination environment [312]. The equatorial ligands can limit U eff but do harm the hysteresis. So, it is rational to propose [Cp2 Dy]+

3.3 Oxo Clusters of Lanthanides

itself as an SMM material, which is in agreement with Gao and coworkers’ report. That the ab initio calculations on the hypothetical cation [Cp* 2 Dy]+ predicted exceptional magnetic axiality and an energy barrier of >1000 cm−1 [313]. In 2017, Layfield and coauthors reported the synthesis of [(Cpttt )2 Dy]+ [B(C6 F5 )4 ]− by reacting [(Cpttt )2 DyCl] with the superelectrophile [(Et3 Si)2 H][B(C6 F5 )4 ]. The bulky ligand 1,2,4-tri(tert-butyl)-cyclopentadienyl (Cpttt ) can prevent the formation of a contact ion pair. The structural changes of removing chloride ion are significant, particularly the reduction in the Dy–Cpcent distances from 2.413(2) Å in [(Cpttt )2 DyCl] to 2.324(1) Å and 2.309(1) Å in [(Cpttt )2 Dy]+ [B(C6 F5 )4 ]− and the increase in the Cp–Dy–Cp angle from 147.59(7)∘ to 152.845(2)∘ (Figure 3.55) [314]. These structural changes lead to the wider angle subtended at dysprosium in [(Cpttt )2 Dy]+ [B(C6 F5 )4 ]− and further resulted in the axial crystal field strengthening, complete eliminated of the equatorial field, and stronger axiality.

⊕

Me3C

Me3C CMe3

Me3C

Dy

Me3C

CI

+

[B(C6F5)4]⊝

CMe3

Me3C

CMe3

Me3C

[(Et3Si)2(μ–H)]⊕

Dy –Et3SiH –Et3SiCI

CMe3

Me3C Me3C

(a)

[B(C6F5)4]⊝ tBu

t t

Bu

t

Bu

Dy

tBu

CI tBu

[H(SiEt3)2][B(C6F5)4], Benzene

t

t

Bu t

–HSiEt3, –CISiEt3

Bu

Dy

Bu

t

(b)

t

t

Bu 1400

Energy (cm–1)

1200 1000 800

[B(C6F5)4]

Bu

Bu (1)

Bu

~ | ± 1/2⟩ ~ | ± 3/2⟩ ~ | ± 5/2⟩ ~ | ± 7/2⟩ ~ | ± 9/2⟩ ~ | ± 11/2⟩

600 400

~ | ± 13/2⟩

200 0

(c)

(d)

~ | ± 15/2⟩ –10

–5 0 Magetic moment (β)

5

10

Figure 3.55 (a, b) Synthesis of [(Cpttt )2 Dy]+ [B(C6 F5 )4 ]− . (c) Main magnetic axes in the ground Kramers doublet. (d) Calculated magnetic relaxation barrier. Source: Guo et al. [314]. 2017, © Wiley-VCH.

141

142

3 Structural Chemistry of Metal-Oxo Clusters

The [(Cpttt )2 DyCl] showed no clear maxima in the 𝜒 ′′ (𝜈) plots and only very narrow hysteresis loops even at 1.8 K. Astonishingly, the 𝜒 ′′ (𝜈) data for [(Cpttt )2 Dy]+ [B(C6 F5 )4 ]− showed explicit maxima up to a significantly high temperature 111 K, and the energy barrier is as high as 1277 cm−1 . The T B of [(Cpttt )2 Dy]+ [B(C6 F5 )4 ]− was determined to be 60 K from three separate experiments, i.e. the 100 seconds blocking temperature, the maximum temperature of open M(H) hysteresis loops (scan rate of 3.9 mT/s), and the temperature of FCZFC magnetic susceptibilities diverge (cooling rate of 2 K/min). The coercive field recorded as 0.06 T at 60 K was also a significant result. These observations are consistent with the parameters reported for the same system [(Cpttt )2 Dy]+ [B(C6 F5 )4 ]− (Cpttt = {C5 H2 t Bu3 -1,2,4} and t Bu = C(CH3 )3 ) in a separate and independent study by Chilton and coworkers [315]. Additionally, the computational studies also prove that the Dy(III) ion placed between two bis(cyclopentadienyl) ligands with large negative charges in axial contribution leading to exceptional magnetic axiality [29]. In light of these advances, Dy(III) ion is the most promising paramagnetic source to construct high-performance SMMs. In 2018, Tong and coworkers report a chemical strategy to access the dysprosium metallocene cation [(CpiPr5 )Dy(Cp*)]+ (CpiPr5 , penta-iso-propylcyclopentadienyl; Cp*, pentamethylcyclopentadienyl), which displays magnetic hysteresis above liquid-nitrogen temperatures. Magnetic measurements reveal that [(CpiPr5 )Dy(Cp*)][B(C6 F5 )4 ] is an SMM with a record anisotropy barrier up to U eff = 1541 cm−1 and a record magnetic blocking temperature of T B = 80 K, for this [(CpiPr5 )Dy(Cp*)] cation overcomes an essential barrier toward the development of nanomagnet devices that function at practical temperatures (Figure 3.56). Because of the strong coupling between Cp* out-of-plane vibrations and the methyl vibrations, it is conceivable that their energies can be fine-tuned by substituent changes. Such an approach should lead to further improvements in SMM performance beyond this [(CpiPr5 )Dy(Cp*)]+ cation and therefore enhance the possibility of its practical application in magnetic information storage materials [316].

3.3.4

Heterometallic 3d–4f Clusters

Due to the different coordination behaviors of Lanthanide and transition metal ions, the classical method to provide heterometallic 3d–4f cluster complexes is to control the subsequent hydrolysis of lanthanide ions by using preformed transition metal complexes [317]. The key of this approach is first to form a stable transition metal complex with chelating ligands and then control lanthanide hydrolysis by the complex as a metalloligand. In 2006, Wu and coworkers used amino acid ligands L-proline (Pro), glycine (Gly), and 2-methylalanine (mAla) to obtain four high-nuclear 3d–4f heterometallic clusters: [Gd6 Cu24 (μ3 -OH)30 (mAla)16 (ClO4 )(H2 O)22 ]⋅(ClO4 )17 ⋅(OH)2 ⋅(H2 O)20 , Na4 [Tb6 Cu26 (μ3 -OH)30 (Gly)18 (ClO4 )(H2 O)22 ]⋅(ClO4 )25 ⋅(H2 O)42 , and {[Ln6 Cu24 (μ3 OH)30 (Pro)12 (Ac)6 (ClO4 )(H2 O)13 ]2 Cu(Pro)2 }⋅(ClO4 )18 ⋅(OH)16 ⋅(H2 O)55 (Ln = Sm, Gd). All of them are based on the {Ln6 Cu24 } octahedral unit [318]. In 2009, they reported two Ln3+ –Cu2+ –glycine (Hgly) clusters with a 44 -net 2D open framework, composed

3.3 Oxo Clusters of Lanthanides i

i

Pr

iPr

iPr

i

Pr

[HB(μ-H)3]

i

Dy

[HB(μ-H)3]

1 (a)

Pr

O

Pr

i

iPr

KCp* –K(BH4) –THF

Pr

i

iPr

Dy

Pr H H

2

[(Et3Si)2(μ-H)] [B(C6F5)4] BH2

+

iPr

i

Pr

i

Pr i

Pr

–2 Et3SiH –0.5 B2H6

3

i

Pr

Dy

[B(C6F5)4]

–

(b)

Figure 3.56 (a) Synthesis of [(Cpttt )2 Dy]+ [B(C6 F5 )4 ]− . (b) The principal magnetic axis of the ground Kramers’ doublet. Source: Guo et al. [316]. 2018, © Science.

of novel [Ln6 Cu22 ] core nodes connected by trans-Cu(gly)2 linkers, and a 3D open framework with a novel 36 ⋅418 ⋅53 ⋅6 topology, made up of [Er6 Cu24 ] cluster nodes and trans-Cu(gly)2 linkers [319]. Long and coworkers have successfully applied α-amino acids and H2 IDA for creating lanthanide hydroxide clusters. By using the typical transition metal complex [M(IDA)] as a metalloligand, they obtained an interesting heterometallic cluster [Gd54 Ni54 (IDA)48 (OH)144 (CO3 )6 (H2 O)25 ]⋅(NO3 )18 under hydrothermal conditions (Figure 3.57a,b). Such framework features a nesting Russian-doll-like four-shell structure, which can be described as Gd2 Ni6 @Gd20 @Gd32 @Ni48 that six bridging CO3 2− ions contributed total of 12 O atoms to the coordination spheres of the edge Gd(III) ions [321]. Interestingly, when they carried out the same reaction under ambient pressure, a new lanthanide–transition–metal compound [Gd52 Ni56 (IDA)48 (OH)154 (H2 O)38 ]⋅(NO3 )18 was obtained (Figure 3.57c) [320]. The structure is very similar to [Gd54 Ni54 (IDA)48 (OH)144 (CO3 )6 (H2 O)25 ]⋅(NO3 )18 , while two Gd(III) and six CO3 2− ions in [Gd54 Ni54 (IDA)48 (OH)144 (CO3 )6 (H2 O)25 ]⋅(NO3 )18 are replaced by two Ni(II) ions and 10 OH− groups in the new cluster. With the same metalloligand, one giant four-shell, 136-metal 3d–4f heterometallic cluster [La60 Ni76 (IDA)68 (μ3 -OH)158 (NO3 )4 (H2 O)44 ]⋅(NO3 )34 was obtained (Figure 3.58a). The 136 metal ions are organized into four distinct shells, that is, 8 Ni(II) ions in Shell 1 are encaged by Shell 2 of La20 Ni4 , followed by Shell 3 of La40 and Shell 4 of Ni64 , forming a structure approximating a rectangular parallelepiped with a dimension of 23 × 23 × 31 Å3 (Figure 3.58b). The magnetic studies revealed weak ferromagnetic interactions between the Ni(II) ions [322]. The successful use of the metalloligand strategy has also been demonstrated by Zheng and coworkers.

143

144

3 Structural Chemistry of Metal-Oxo Clusters

Gd Ni O C N

(a)

(b)

(c)

Figure 3.57 (a) A ball-and-stick view of the cationic cluster of [Gd54 Ni54 (IDA)48 (OH)144 (CO3 )6 (H2 O)25 ]⋅(NO3 )18 . (b) A four-shell presentation of the cluster showing only the metal frameworks. (c) The structure of the cationic cluster of [Gd52 Ni56 (IDA)48 (OH)154 (H2 O)38 ]18+ . Source: Modified from Liu et al. [320]. 2016, © Wiley-VCH.

(a)

La Ni O N C

(b)

Figure 3.58 (a) Ball-and-stick plot of the cationic cluster complex [La60 Ni76 (IDA)68 (μ3 -OH)158 (NO3 )4 (H2 O)44 ]34+ along the bc plane. (b) Ball-and-stick plot of the La60 Ni76 framework with its four shells distinctly shown. Source: Kong et al. [322]. 2009, © The Royal Society of Chemistry.

In 2016, they used a mixture of Ln(NO3 )3 , Ni(OAc)2 , H2 IDA, and DMPA to construct the largest 3d–4f clusters, [Ni64 Ln96 (μ3 -OH)156 (IDA)66 (DMPA)12 (CH3 COO)48− (NO3 )24 (H2 O)64 ]⋅Cl24 (Ln = Gd, Dy, Y; DMPA = 2,2-dimethylol propionic acid) [322]. These giant clusters possess a unique porous cube-like structure and represent the highest nuclearity heterometallic cluster complexes. These porous solids also exhibit high selectivity for CO2 over CH4 or N2 at room temperature and large MCEs (42.8 J/(kg K) at 3 K with ΔH = 7 T) (Figure 3.59a). Tong and coworkers reported the wheel-shaped clusters [Co16 Ln24 (OH)50 (pyacac)16 (NO3 )18 (H2 O)12 ][Ln(H2 O)8 ]2 (NO3 )16 (OH)10 by using the 1,3-bis(2-pyridyl)-1,3-propanedione (pyacacH) as organic ligand to form a metalloligand. The cluster shows a nanoscale wheel-like structure and the pyacac− ligands act as organic shells wrapping up the inorganic core (Figure 3.59b) [323]. In 2018, Long and coworkers reported a giant lanthanide-titanium oxo cluster of [Eu24 Ti8 (L)31 (HL)42 (CH3 CN)11− (H2 O)8 ]⋅nH2 O (H2 L = salicylic acid) under the guidance of the HRESI-MS. Single-crystal structural analysis reveals that the cluster has a wheel-like structure with diameter of c. 4.1 nm and is the highest nuclearity lanthanide–titanium oxo cluster reported so far

3.3 Oxo Clusters of Lanthanides

(a)

(b)

≈4.1 nm

≈0.9nm

(c)

3.2 nm

Figure 3.59 (a) Ball-and-stick view of the cationic {Ni60 Gd96 } core with H atoms removed for clarity. Gd, purple; Ni, cyan; N, green; O, orange; C, gray. (b) Ball-and-stick plot of {Co16 Ln24 }, Dy, violet; Co, turquoise; O, red; N, blue; C, gray; excluding hydrogen atoms. (c) Ball-and-stick representation of the crystal structure of {Eu24 Ti8 }. Purple Eu, green Ti, red O, gray C, white H, blue N. Source: Based on Zhang et al. [323]. 2013, © The Royal Society of Chemistry.

(Figure 3.59c). The time- and temperature-dependent HRESI-MS ancillary investigation indicates the molecular self-assembly can be realized within short time and provides a fast and efficient method for guiding the molecular self-assembly [324]. In recent years, to explore the magnetic interactions between individual clusters within the bulk material for practical applications have captured much attention. To evaluate such interactions, Long and coworkers prepared a core–shell Gd52 Ni56 @SiO2 hybrid material by the lanthanide–transition metal compound [Gd52 Ni56 (IDA)48 (OH)154− (H2 O)38 ]⋅(NO3 )18 . The monodisperse core–shell nanoparticles were successfully made by microemulsion method, with each consisting of one core of cluster encapsulated by one nanoshell of silica [325]. Its maximum magnetic entropy change of 44.6 J/(kg K) reflects a 10% enhancement over the value of the corresponding parent cluster (40.5 J/(kg K)). Theoretical studies revealed that shelling the molecular cluster effectively reduces magnetic interactions between the cluster units. Then, the research team has prepared four isomorphic lanthanide–transition metal clusters Ln52 Ni56 (Ln = Eu, Pr, Nd, and Gd) and loaded them atomically precise onto CdS photoabsorber surface to achieve enhanced photocatalytic H2 production [326]. Investigations of the photocatalytic hydrogen evolution reaction show that the Eu52 Ni56−x Cdx /CdS-3 composite exhibits the highest photocatalytic performance of 25 353 μmol/(h g), which can be attributed to the multipath charge transfer process generated by Eu2+ with synergistic effects (Figure 3.60). In 2017, they have reported the synthesis and crystal structure of the largest Ln–Fe cluster of Gd12 Fe14 (Figure 3.61) [327]. Single-crystal analysis reveals that the cluster mainly consists of a cationic metal cluster of [Gd12 Fe14 (μ3 -OH)12 (μ4 -OH)6 (μ4 -O)12 -(TEOA)6 (CH3 COO)16 (H2 O)8 ]2+ . The cationic metal cluster can be viewed as a hexagram-like metal core of [Gd12 Fe14 (μ3 -OH)12 (μ4 -OH)6 (μ4 -O)12 ]36+ stabilized by six triply deprotonated triethanolamine, 16 acetate ligands, and eight coordination water molecules. Significantly, such cluster is very stable in aqueous solution that can be processed into monodisperse core–shell structure of Gd12 Fe14 @SiO2 through the facile one-pot microemulsion method. The

145

After 30 min visible-light irradiation Before irradiation

Ln52Ni56–xCdx Cluster 180

°C/1

CH

Ln52Ni56 cluster

(a)

g = 2.003

–1

g = 1.922

3200 (b)

– – – –

2H+

0.67 + + + +

1

HOMO 1.51 + + + +

2

+Cd 2+

Ln52Ni56–xCdx /CdS

e– Eu Ni 52 56–xCdx H2 –0.29 e– LUMO

Eu3+/Eu2+ e–

0

8h

4 N2 S

CdS –0.81 CB

– – – – – –0.35 e

V vs NHE

180 °C/18 h

Intensity (a.u.)

+Cd2+

3400

3600 3800 H [G]

VB

3

4000 (c)

Figure 3.60 (a) One-step synthesis of Ln52 Ni56−x Cdx and Ln52 Ni56−x Cdx /CdS-3 composites. Green Ni, yellow Cd, red O, purple Ln, pink S. (b) EPR spectra of Eu52 Ni56−x Cdx /CdS-3 before and after 30 minutes of visible-light irradiation. (c) Charge separation and transfer in Eu52 Ni56−x Cdx /CdS-3 composite under visible-light irradiation. VB = valence band, CB = conduction band. Source: Chen et al. [326]. 2018, © Wiley-VCH.

3.3 Oxo Clusters of Lanthanides

150 χmT (cm3 K/mol–1)

140 130 120 110 100

(a)

(b)

25 nm

0

50

100

150 T(K)

200

250

300

Figure 3.61 (a) Ball-and-stick view of the metal cluster core of [Gd12 Fe14 (μ3 -OH)12 (μ4 -OH)6 (μ4 -O)12 -(TEOA)6 (CH3 COO)16 (H2 O)8 ]2+ . Green Fe, purple Gd, blue N, red O, gray C. (b) Experimental plots of 𝜒 M T versus T of [Gd12 Fe14 ] and [Gd12 Fe14 ]@SiO2 and fitting curves (solid line) for 𝜒 M T versus T for [Gd12 Fe14 ] (red) and [Gd12 Fe14 ]@SiO2 (blue). Source: Zheng et al. [327]. 2017, © Wiley-VCH.

experimental and theoretical studies on the Gd12 Fe14 and Gd12 Fe14 @SiO2 systems demonstrated that the shielding effect of the SiO2 not only effectively decreases the intermolecular magnetic interactions but also significantly increases the ZFS effect of the outer layer Fe(III) ions. The core–shell nanostructure can improve the process ability and stabilize the cluster materials for practical applications and also provide a platform for investigating magnetic exchange coupling within a particular cluster unit. Two high-nuclearity 3d–4f heterometallic cluster-based compounds, such as [Na2 Ni12 Ln2 (BTC4A)3 (μ7 CO3 )3 (μ3 OH)4 (μ3 Cl)2 (OAc)6 (dma)4 ]⋅2OAc⋅0.5dma⋅3CH3 CN⋅8DMA (Ln = Dy for Dy and Tb for H4 BTC4A = p-tert-butylthiacalix[4]arene, dma = dimethylamine, and DMA = N,N ′ -dimethylacetamide), were synthesized [308p]. Every four Ni2+ ions binds to one fully deprotonated BTC4 A4 ligand, in which one carbonato anion acts as the cork base to form a shuttlecock-like NiII 4 -BTC4 A building block. Three subunits are linked together in an up-to-up coordination mode, in which the linkage between four cations (two sodium ions and two Ln3+ ions) and other anions (two chloride anions, four hydroxide anions, and three acetate anions) leads to a pseudo-trigonal planar entity of heterometallic Na2 Ni12 Ln2 cluster. Three BTC4 A4 ligands are located on the trigonal plane of the tricubane core. It is worth to note that such vertex-fused tricubane unit possessing more than one type of metal element has not been reported before. In view of the crystal packing, the complex exhibits a bilayer structure with the pseudo-trigonal planar entities sitting in an up–down mode (Figure 3.62). Magnetic studies reveal that only the DyIII complex shows the slow relaxation of the magnetization with expected SMM behavior. By utilizing calixarene as molecular building blocks (MBBs), a series of cationic trigonal prismatic heterometallic organic nanocages (HMONCs): Na4 Ni24 Dy4 with tunable sizes can be obtained through a stepwise method [308m]. More specially, in

147

148

3 Structural Chemistry of Metal-Oxo Clusters

(a)

(b)

(c)

Figure 3.62 (a) Molecular structure of Na2 Ni12 Ln2 cluster (left); vertexfused tricubane core (middle); the green triangle stands for the MBB (right). (b) Linear dicarboxylic acid linkers. (c) Schematic representation of the formation of the HMONCs. Source: Su et al. [308m]. 2015, © The Royal Society of Chemistry.

every HMONCs, three linear acetate ligands substitute the peripheral coordination of two Na2 Ni12 Ln2 clusters to form an unprecedented Na4 Ni24 Ln4 HMONC through a M2 L3 condensation. Moreover, the Na2 Ni12 Dy2 core exhibits a slow magnetic relaxation behavior and gas sorption behavior to a certain extent. A hydrothermal reaction of H4 BTC4 A⋅CHCl3 , Zn(OAc)2 ⋅2H2 O, and Ln(OAc)3 ⋅6H2 O in mixed DMF/CH3 OH solvent at 120 ∘ C yielded four isomorphous colorless prismatic crystals, named ZnII LnIII 3 (Ln = Gd, Tb, Dy, Ho) [308i]. The single-crystal X-ray diffraction reveals that compound ZnII GdIII 3 contains four unique metal ions, in which one ZnII ion adopts six-coordinated mode while three GdIII ions show eight coordinations. These four metal ions are bridged by oxygen species, forming Figure 3.63 The kite-like heterometallic tetranuclear ZnII LnIII cluster. Source: Su et al. [308i]. © 2013 American Chemical Society.

ZnII Ln3III

References

a kite-like tetranuclear core, as shown in Figure 3.63. The photoluminescent analyses reveal that the H4 BTC4A is an efficient sensitizer for Tb3+ ions in compound ZnII TbIII 3 . The magnetic properties of complex ZnII DyIII 3 exhibit slow magnetization relaxation typical for single-molecule magnets.

3.4 Conclusion The high-nuclearity cluster-type metal complexes have not only interesting structures but also potential applications and will continue to be the hot topic of the cluster science. In this chapter, the facile one-pot self-assembly approach and step-by-step strategy are accessible to accomplish enormous generation of well-defined oxo clusters of transition metal, main group metal and lanthanides, with structural and compositional novelty via controlling the process of hydrolysis and condensation. The excellent stability of given metal-oxo clusters in the solid state and solution as well as gas phase facilitates their applications in deverse fields ranging from catalysis, magnetism, energy conversion, and storage to material science owing to their intrinsic features of structural tunability and robustness accompanied with thermal and redox stability, etc. Thereinto the structurally well-defined metal-oxo clusters providing models for decoding mechanistic insight of metal-based materials in the molecular level, in turn, the performance of materials feedback to the guidance of designing materials in a more controllable way.

Acknowledgments The authors acknowledge the financial support from National Natural Science Foundation of China (21673238, 21901241, 21922111, 21771181, 21871266, and 21731006), the Strategic Priority Research Program of the Chinese Academy of Sciences (XDB20000000), Key Research Program of Frontier Science CAS (QYZDYSSW-SLH025), and Youth Innovation Promotion Association CAS (2017345).

References 1 Fang, W.H., Zhang, L., and Zhang, J. (2017). Chem. Soc. Rev. 47: 404–421. 2 Monakhov, K.Y., Bensch, W., and Kogerler, P. (2015). Chem. Soc. Rev. 44: 8443–8483. 3 Pope, M.T. and Müller, A. (2002). Polyoxometalate Chemistry from Topology Via Self-Assembly to Applications. New York, Boston, MA, Dordrecht, London, Moscow: Kluwer Academic Publishers. 4 Zheng, S.T. and Yang, G.Y. (2012). Chem. Soc. Rev. 41: 7623–7646. 5 Abramov, P.A. and Sokolov, M.N. (2017). Russ. J. Coord. Chem. 43: 421–432. 6 Pope, M.T. and Müller, A. (1991). Angew. Chem. Int. Ed. Engl. 30: 34–48. 7 Berzelius, J.J. (1826). Ann. Phys. Chem. 6: 369–392.

149

150

3 Structural Chemistry of Metal-Oxo Clusters

8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38

Keggin, J.F. (1933). Proc. R. Soc. London, Ser. A 144: 75–100. Lindqvist, I. (1953). Ark. Kemi 5: 247–250. Long, D.L., Burkholder, E., and Cronin, L. (2007). Chem. Soc. Rev. 36: 105–121. Fang, W.H., Zhang, L., and Zhang, J. (2016). J. Am. Chem. Soc. 138: 7480–7483. Assi, H., Mouchaham, G., Steunou, N. et al. (2017). Chem. Soc. Rev. 46: 3431–3452. Chubarova, E.V., Dickman, M.H., Keita, B. et al. (2008). Angew. Chem. Int. Ed. 47: 9542–95466. Cooper, E.R., Andrews, C.D., Wheatley, P.S. et al. (2004). Nature 430: 1012–1016. Ahmed, E. and Ruck, M. (2012). Angew. Chem. Int. Ed. 51: 308–309. Rozes, L. and Sanchez, C. (2011). Chem. Soc. Rev. 40: 1006–1030. Coppens, P., Chen, Y., and Trzop, E. (2014). Chem. Rev. 114: 9645–9661. Hou, J.L., Luo, W., Wu, Y.Y. et al. (2015). Dalton Trans. 44: 19829–19835. Wu, Y.Y., Luo, W., Wang, Y.H. et al. (2014). Inorg. Chem. 51: 8982–8988. Guo, Y., Hou, J.-L., Luo, W. et al. (2017). J. Mater. Chem. A 5: 18270–18275. Hou, J.L., Huo, P., Tang, Z.Z. et al. (2018). Inorg. Chem. 57: 7420–7427. Gao, M.Y., Wang, F., Gu, Z.G. et al. (2016). J. Am. Chem. Soc. 138: 2556–2559. Ding, Q.R., Liu, J.X., Narayanam, N. et al. (2017). Dalton Trans. 46: 16000–16003. Zhu, B.-C., Fang, W.-H., Emayavaramban, P. et al. (2018). CrystEngComm 20: 5964–5968. Gao, M.Y., Fan, X., Zhang, L., and Zhang, J. (2018). Inorg. Chem. 57: 5642–5647. Hou, J., Hu, J., Sun, Q. et al. (2016). Inorg. Chem. 55: 7075–7078. Czakler, M., Artner, C., and Schubert, U. (2013). Eur. J. Inorg. Chem. 2013: 5790–5796. Guerrero, G., Mehring, M., Mutin, P.H. et al. (1999). J. Chem. Soc., Dalton Trans.: 1537–1538. Seisenbaeva, G.A., Melnyk, I.V., Hedin, N. et al. (2015). RSC Adv. 5: 24575–24585. Czakler, M., Artner, C., and Schubert, U. (2015). Monatsh. Chem. 146: 1249–1256. Chen, Y., Trzop, E., Sokolow, J.D., and Coppens, P. (2013). Chem. Eur. J. 19: 16651–16655. Fang, W.H., Zhang, L., and Zhang, J. (2017). Dalton Trans. 46: 803–807. Czakler, M., Artner, C., and Schubert, U. (2014). Eur. J. Inorg. Chem. 2014: 2038–2045. Fan, X., Narayanam, N., Gao, M. et al. (2018). Dalton Trans. 47: 663–665. Baumann, S.O., Bendova, M., Fric, H. et al. (2009). Eur. J. Inorg. Chem. 2009: 3333–3340. Chen, S., Fang, W.H., Zhang, L., and Zhang, J. (2018). Inorg. Chem. 57: 8850–8856. Narayanam, N., Chintakrinda, K., Fang, W.H. et al. (2016). Inorg. Chem. 55: 10294–10301. Gao, M.-Y., Fang, W.-H., Wen, T. et al. (2017). Cryst. Growth Des. 17: 3592–3595.

References

39 Fan, X., Wang, J., Wu, K. et al. (2019). Angew. Chem. Int. Ed. 58: 1320–1323. 40 Narayanam, N., Fang, W.H., Chintakrinda, K. et al. (2017). Chem. Commun. 53: 8078–8080. 41 Fan, X., Fu, H., Zhang, L., and Zhang, J. (2019). Dalton Trans. 48: 8049–8052. 42 Fu, Y., Xu, Z., and Zhang, F. (2008). J. Mol. Struct. 873: 168–172. 43 Zhang, G., Liu, C., Long, D.L. et al. (2016). J. Am. Chem. Soc. 138: 11097–11100. 44 Kment, S., Riboni, F., Pausova, S. et al. (2017). Chem. Soc. Rev. 46: 3716–3769. 45 Snoeberger, R.C. III, Young, K.J., Tang, J. et al. (2012). J. Am. Chem. Soc. 134: 8911–8917. 46 Negre, C.F., Young, K.J., Oviedo, M.B. et al. (2014). J. Am. Chem. Soc. 136: 16420–16429. 47 Liu, J.X., Gao, M.Y., Fang, W.H. et al. (2016). Angew. Chem. Int. Ed. 55: 5160–5165. 48 Dan-Hardi, M., Serre, C., Frot, T. et al. (2009). J. Am. Chem. Soc. 131: 10857–10859. 49 Hendon, C.H., Tiana, D., Fontecave, M. et al. (2013). J. Am. Chem. Soc. 135: 10942–10945. 50 Logan, M.W., Ayad, S., Adamson, J.D. et al. (2017). J. Mater. Chem A 5: 11854–11863. 51 Chen, Y., Trzop, E., Makal, A. et al. (2013). Inorg. Chem. 52: 4750–4752. 52 Chen, Y., Jarzembska, K.N., Trzop, E. et al. (2015). Chem. Eur. J. 21: 11538–11544. 53 Hu, J., Zhan, L., Zhang, G. et al. (2016). Inorg. Chem. 55: 8493–8501. 54 Chen, Y., Sokolow, J., Trzop, E. et al. (2013). J. Chin. Chem. Soc. 60: 887–890. 55 Chen, Y., Trzop, E., Makal, A. et al. (2014). Dalton Trans. 43: 3839–3841. 56 Lv, Y., Cheng, J., Matthews, P.D. et al. (2014). Dalton Trans. 43: 8679–8689. 57 Artner, C., Koyun, A., and Schubert, U. (2015). Monatsh. Chem. 146: 1777–1780. 58 Lv, Y., Willkomm, J., Leskes, M. et al. (2012). Chem. Eur. J. 18: 11867–11870. 59 Lu, D.F., Kong, X.J., Lu, T.B. et al. (2017). Inorg. Chem. 56: 1057–1060. 60 Wang, S., Su, H.C., Yu, L. et al. (2015). Dalton Trans. 44: 1882–1888. 61 Yue, L., Wang, S., Zhou, D. et al. (2016). Nat. Commun. 7: 10742. 62 Chen, S., Chen, Z.N., Fang, W.H. et al. (2019). Angew. Chem. Int. Ed. 58: 10932–10935. 63 Wang, C., Liu, C., Li, L.J., and Sun, Z.M. (2019). Inorg. Chem. 58: 6312–6319. 64 Fang, W.H., Zhang, L., and Zhang, J. (2017). Chem. Commun. 53: 3949–3951. 65 Fang, W.-H., Wang, J.-F., Zhang, L., and Zhang, J. (2017). Chem. Mater. 29: 2681–2684. 66 Hong, K., Bak, W., and Chun, H. (2014). Inorg. Chem. 53: 7288–7293. 67 Hong, K. and Chun, H. (2013). Inorg. Chem. 52: 9705–9707. 68 Cini, M., Bradshaw, T.D., and Woodward, S. (2017). Chem. Soc. Rev. 46: 1040–1051. 69 Jiang, Z., Liu, J., Gao, M. et al. (2017). Adv. Mater.: 29. 70 Gu, Z.G., Fu, H., Neumann, T. et al. (2016). ACS Nano 10: 977–983. 71 Chen, S., Fang, W.H., Zhang, L., and Zhang, J. (2018). Angew. Chem. Int. Ed. 57: 11252–11256.

151

152

3 Structural Chemistry of Metal-Oxo Clusters

72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100

Bocchini, S., Fornasieri, G., Rozes, L. et al. (2005). Chem. Commun.: 2600–2602. Livage, J. (1998). Coord. Chem. Rev. 178–180: 999–1018. Hayashi, Y. (2011). Coord. Chem. Rev. 255: 2270–2280. Wang, K., Niu, Y., Zhao, D. et al. (2017). Inorg. Chem. 56: 14053–14059. Day, V.W., Klemperer, W.G., and Yaghi, O.M. (1989). J. Am. Chem. Soc. 111: 4518–4519. Marrot, J., Barthelet, K., Simonnet, C., and Riou, D. (2005). C.R. Chim. 8: 971–976. Hamilton, E.E., Fanwick, P.E., and Wilker, J.J. (2002). J. Am. Chem. Soc. 124: 78–82. Hou, D., Hagen, K.S., and Hill, C.L. (1993). J. Chem. Soc., Chem. Commun.: 426–428. Hou, D., Hagen, K.S., and Hill, C.L. (1992). J. Am. Chem. Soc. 114: 5864–5866. Day, V.W., Klemperer, W.G., and Yaghi, O.M. (1989). J. Am. Chem. Soc. 111: 5959–5961. Bensch, W., Hug, P., Reller, A., and Oswald, H.R. (1989). Mater. Res. Bull. 24: 403–415. Day, V.W., Klemperer, W.G., and Maltbie, D.J. (1987). J. Am. Chem. Soc. 109: 2991–3002. Fuchs, J., Mahjour, S., and Pickardt, J. (1976). Angew. Chem. Int. Ed. Engl. 15: 374–375. Müller, A., Krickemeyer, E., Penk, M. et al. (1987). Angew. Chem. Int. Ed. Engl. 26: 1045–1046. Müller, A., Krickemeyer, E., Penk, M. et al. (1991). Angew. Chem. Int. Ed. Engl. 30: 1674–1677. Müller, A., Rohlfing, R., Döring, J., and Penk, M. (1991). Angew. Chem. Int. Ed. Engl. 30: 588–590. Müller, A., Sessoli, R., Krickemeyer, E. et al. (1997). Inorg. Chem. 36: 5239–5250. Suber, L., Bonamico, M., and Fares, V. (1997). Inorg. Chem. 36: 2030–2033. Hayashi, Y., Fukuyama, K., Takatera, T., and Uehara, A. (2000). Chem. Lett. 29: 770–771. Hayashi, Y., Miyakoshi, N., Shinguchi, T., and Uehara, A. (2001). Chem. Lett. 30: 170–171. Chen, Y., Gu, X., Peng, J. et al. (2004). Inorg. Chem. Commun. 7: 705–707. Hao, N., Qin, C., Xu, Y. et al. (2005). Inorg. Chem. Commun. 8: 592–595. Ishaque Khan, M., Ayesh, S., Doedens, R.J. et al. (2005). Chem. Commun.: 4658–4660. Johnson, G.K. and Schlemper, E.O. (1978). J. Am. Chem. Soc. 100: 3645–3646. Batchelor, L.J., Shaw, R., Markey, S.J. et al. (2010). Chem. Eur. J. 16: 5554–5557. Khanra, S., Kloth, M., Mansaray, H. et al. (2007). Angew. Chem. Int. Ed. 46: 5568–5571. Zhang, L. and Schmitt, W. (2011). J. Am. Chem. Soc. 133: 11240–11248. Fang, X., Kogerler, P., Isaacs, L. et al. (2009). J. Am. Chem. Soc. 131: 432–433. MÜller, A. (1991). Nature 352: 115–115.

References

101 Müller, A., Reuter, H., and Dillinger, S. (1995). Angew. Chem. Int. Ed. Engl. 34: 2328–2361. 102 Pope, M.T. (1992). Nature 355: 27–27. 103 Sole-Daura, A., Notario-Estevez, A., Carbo, J.J. et al. (2019). Inorg. Chem. 58: 3881–3894. 104 Liu, S., Li, D., Xie, L. et al. (2006). Inorg. Chem. 45: 8036–8040. 105 Ichida, H., Nagai, K., Sasaki, Y., and Pope, M.T. (1989). J. Am. Chem. Soc. 111: 586–591. 106 Kurata, T., Uehara, A., Hayashi, Y., and Isobe, K. (2005). Inorg. Chem. 44: 2524–2530. 107 Maruyama, T., Kawabata, H., Kikukawa, Y., and Hayashi, Y. (2019). Eur. J. Inorg. Chem. 2019: 529–533. 108 Schurr, B.E., Nachtigall, O., VanGelder, L.E. et al. (2019). J. Coord. Chem. 72: 1267–1286. 109 Aronica, C., Chastanet, G., Zueva, E. et al. (2008). J. Am. Chem. Soc. 130: 2365–2371. 110 Laye, R.H., Wei, Q., Mason, P.V. et al. (2006). J. Am. Chem. Soc. 128: 9020–9021. 111 Zhang, J., Huang, Y., Li, G., and Wei, Y. (2019). Coord. Chem. Rev. 378: 395–414. 112 Gong, Y., Zhang, Y., Qin, C. et al. (2019). Angew. Chem. Int. Ed. 58: 780–784. 113 Xu, N., Gan, H., Qin, C. et al. (2019). Angew. Chem. Int. Ed. 58: 4649–4653. 114 Zhang, Y., Gan, H., Qin, C. et al. (2018). J. Am. Chem. Soc. 140: 17365–17368. 115 Wang, C.L., Zhang, Z.B., Fu, J. et al. (2012). Chem. Eur. J. 18: 11909–11912. 116 Breen, J.M. and Schmitt, W. (2008). Angew. Chem. Int. Ed. 47: 6904–6908. 117 Mahnke, L.K., Kondinski, A., Warzok, U. et al. (2018). Angew. Chem. Int. Ed. 57: 2972–2975. 118 Wutkowski, A., Nather, C., Kogerler, P., and Bensch, W. (2008). Inorg. Chem. 47: 1916–1918. 119 Gao, Y., Xu, Y., Li, S. et al. (2010). J. Coord. Chem. 63: 3373–3383. 120 Wang, X., Liu, L., Zhang, G., and Jacobson, A.J. (2001). Chem. Commun.: 2472–2473. 121 Whitfield, T., Wang, X., and Jacobson, A.J. (2003). Inorg. Chem. 42: 3728–3733. 122 Tripathi, A., Hughbanks, T., and Clearfield, A. (2003). J. Am. Chem. Soc. 125: 10528–10529. 123 You, L.-S., Zhu, Q.-Y., Zhang, X. et al. (2013). CrystEngComm 15: 2411–2415. 124 Chen, Y.-M., Wang, E.-B., Lin, B.-Z., and Wang, S.-T. (2003). Dalton Trans.: 519–520. 125 Wang, S., Liu, Y., Zhang, Z. et al. (2019). ACS Appl. Mater. Interfaces 11: 12786–12796. 126 Santoni, M.P., La Ganga, G., Mollica Nardo, V. et al. (2014). J. Am. Chem. Soc. 136: 8189–8192. 127 Chen, L., Jiang, F., Lin, Z. et al. (2005). ChemInform 127: 8588–8589. 128 Tsunashima, R., Long, D.L., Miras, H.N. et al. (2010). Angew. Chem. Int. Ed. 49: 113–116. 129 Huang, P., Qin, C., Su, Z.M. et al. (2012). J. Am. Chem. Soc. 134: 14004–14010.

153

154

3 Structural Chemistry of Metal-Oxo Clusters

130 Jin, L., Zhu, Z.K., Wu, Y.L. et al. (2017). Angew. Chem. Int. Ed. 56: 16288–16292. 131 Wu, Y.L., Li, X.X., Qi, Y.J. et al. (2018). Angew. Chem. Int. Ed. 57: 8572–8576. 132 Guo, Z.W., Chen, Y., Zhao, D. et al. (2019). Inorg. Chem. 58: 4055–4058. 133 Sures, D., Segado, M., Bo, C., and Nyman, M. (2018). J. Am. Chem. Soc. 140: 10803–10813. 134 Martin, N.P., Petrus, E., Segado, M. et al. (2019). Chem. Eur. J. 25: 10580–10584. 135 Müller, A. and Serain, C. (2000). Acc. Chem. Res. 33: 2–10. 136 Muller, A. and Gouzerh, P. (2012). Chem. Soc. Rev. 41: 7431–7463. 137 Miras, H.N., Long, D.L., and Cronin, L. (2017). Exploring self-assembly and the self-organization of nanoscale inorganic polyoxometalate clusters. In: Advances in Inorganic Chemistry, Polyoxometalate Chemistry, vol. 69 (eds. R.V. Eldik and L. Cronin), 1–28. Zoe Kruze. 138 Müller, A., Krickemeyer, E., Meyer, J. et al. (1995). Angew. Chem. Int. Ed. Engl. 34: 2122–2124. 139 Müller, A., Krickemeyer, E., Bögge, H. et al. (1998). Angew. Chem. Int. Ed. Engl. 37: 1220–1223. 140 Müller, A., Krickemeyer, E., Bögge, H. et al. (1998). Angew. Chem. Int. Ed. Engl. 37: 3359–3363. 141 Botar, B., Ellern, A., Sougrati, M.T., and Kögerler, P. (2009). Eur. J. Inorg. Chem. 2009: 5071–5074. 142 Müller, A., Shah, S.Q.N., Bögge, H., and Schmidtmann, M. (1999). Nature 397: 48–50. 143 Cronin, L., Beugholt, C., Krickemeyer, E. et al. (2002). Angew. Chem. Int. Ed. 41: 2805–2808. 144 Xuan, W., Pow, R., Long, D.L., and Cronin, L. (2017). Angew. Chem. Int. Ed. 56: 9727–9731. 145 Xuan, W., Pow, R., Watfa, N. et al. (2019). J. Am. Chem. Soc. 141: 1242–1250. 146 Zhao, J. and Xu, L. (2011). Eur. J. Inorg. Chem. 2011: 4096–4102. 147 Chen, W.P., Sang, R.L., Wang, Y., and Xu, L. (2013). Chem. Commun. 49: 5883–5885. 148 Xu, X., Luo, B.L., Wang, L.L., and Xu, L. (2018). Dalton Trans. 47: 3218–3222. 149 Izarova, N.V., Pope, M.T., and Kortz, U. (2012). Angew. Chem. Int. Ed. 51: 9492–9510. 150 Yang, P. and Kortz, U. (2018). Acc. Chem. Res. 51: 1599–1608. 151 Izarova, N.V., Dickman, M.H., Biboum, R.N. et al. (2009). Inorg. Chem. 48: 7504–7506. 152 Xu, F., Scullion, R.A., Yan, J. et al. (2011). J. Am. Chem. Soc. 133: 4684–4686. 153 Yang, P., Xiang, Y., Lin, Z. et al. (2014). Angew. Chem. Int. Ed. 53: 11974–11978. 154 Barsukova, M., Izarova, N.V., Biboum, R.N. et al. (2010). Chem. Eur. J. 16: 9076–9085. 155 Izarova, N.V., Biboum, R.N., Keita, B. et al. (2009). Dalton Trans.: 9385–9387. 156 Barsukova-Stuckart, M., Izarova, N.V., Barrett, R. et al. (2012). Chem. Eur. J. 18: 6167–6171.

References

157 Barsukova-Stuckart, M., Izarova, N.V., Jameson, G.B. et al. (2011). Angew. Chem. Int. Ed. 50: 2639–2642. 158 Christie, L.G., Surman, A.J., Scullion, R.A. et al. (2016). Angew. Chem. Int. Ed. 55: 12741–12745. 159 Christie, L.G., Asche, S., Mathieson, J.S. et al. (2018). J. Am. Chem. Soc. 140: 9379–9382. 160 Yang, P., Xiang, Y., Lin, Z. et al. (2016). Angew. Chem. Int. Ed. 55: 15766–15770. 161 Bhattacharya, S., Ayass, W.W., Taffa, D.H. et al. (2019). J. Am. Chem. Soc. 141: 3385–3389. 162 Oms, O., Dolbecq, A., and Mialane, P. (2012). Chem. Soc. Rev. 41: 7497–7536. 163 Ma, X., Li, H., Chen, L., and Zhao, J. (2016). Dalton Trans. 45: 4935–4960. 164 Monakhov, K.Y., Moors, M., and Kögerler, P. (2017). Perspectives for polyoxometalates in single-molecule electronics and spintronics. In: Advances in Inorganic Chemistry, Polyoxometalate Chemistry, vol. 69 (eds. R.V. Eldik and L. Cronin), 251–286. Zoe Kruze. 165 Yan, J., Gao, J., Long, D.L. et al. (2010). J. Am. Chem. Soc. 132: 11410–11411. 166 Yan, J., Long, D.L., and Cronin, L. (2010). Angew. Chem. Int. Ed. 49: 4117–4120. 167 Surman, A.J., Robbins, P.J., Ujma, J. et al. (2016). J. Am. Chem. Soc. 138: 3824–3830. 168 Liu, J.X., Izarova, N.V., and Kogerler, P. (2019). Chem. Commun. 55: 10744–10747. 169 de la Oliva, A.R., Sans, V., Miras, H.N. et al. (2012). Angew. Chem. Int. Ed. 51: 12759–12762. 170 Zhao, J.-W., Zhang, J., Song, Y. et al. (2008). Eur. J. Inorg. Chem. 2008: 3809–3819. 171 Zheng, S.T., Yuan, D.Q., Jia, H.P. et al. (2007). Chem. Commun.: 1858–1860. 172 Li, X.-X., Zheng, S.-T., Fang, W.-H., and Yang, G.-Y. (2011). Inorg. Chem. Commun. 14: 1541–1545. 173 Zhao, J.W., Jia, H.P., Zhang, J. et al. (2007). Chem. Eur. J. 13: 10030–10045. 174 Li, B., Zhao, J.W., Zheng, S.T., and Yang, G.Y. (2009). Inorg. Chem. 48: 8294–8303. 175 Zheng, S.T., Zhang, J., Clemente-Juan, J.M. et al. (2009). Angew. Chem. Int. Ed. 48: 7176–7179. 176 Zheng, S.T., Zhang, J., Li, X.X. et al. (2010). J. Am. Chem. Soc. 132: 15102–15103. 177 Kuang, X., Wu, X., Yu, R. et al. (2010). Nat. Chem. 2: 461–465. 178 Yu, H. and Zheng, S.-T. (2018). Chin. Sci. Bull. 63: 3277–3285. 179 Matsumoto, M., Ozawa, Y., Yagasaki, A., and Zhe, Y. (2013). Inorg. Chem. 52: 7825–7827. 180 Liang, Z., Zhao, S., Ma, P. et al. (2018). Inorg. Chem. 57: 12471–12474. 181 Son, J.H. and Casey, W.H. (2016). Chem. Eur. J. 22: 14155–14157. 182 Peng, Q., Li, S., Wang, R. et al. (2017). Dalton Trans. 46: 4157–4160. 183 Huang, P., Wu, H.Y., Huang, M. et al. (2017). Dalton Trans. 46: 10177–10180. 184 Huang, P., Wang, X.-J., Qi, J.-J. et al. (2017). J. Mater. Chem A 5: 22970–22974. 185 Huang, P., Qin, C., Zhou, Y. et al. (2016). Chem. Commun. 52: 13787–13790.

155

156

3 Structural Chemistry of Metal-Oxo Clusters

186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218

Huang, P., Han, X.-G., Li, X.-L. et al. (2016). CrystEngComm 18: 8722–8725. Molina, P.I., Sures, D.J., Miro, P. et al. (2015). Dalton Trans. 44: 15813–15822. Li, S., Liu, S., Liu, S. et al. (2012). J. Am. Chem. Soc. 134: 19716–19721. Williams, D.J. (1984). Angew. Chem. Int. Ed. 23: 690–703. Chen, C.T. and Liu, G.Z. (1986). Annu. Rev. Mater. Sci. 16: 203–243. Chen, C.T., Wu, B.C., Jiang, A.D., and You, G.M. (1985). Sci. Sin., Ser. B 28: 235–243. Chen, C.T., Wu, Y.C., Jiang, A.D. et al. (1989). J. Opt. Soc. Am. B: Opt. Phys. 6: 616–621. Wu, Y.C., Sasaki, T., Nakai, S. et al. (1993). Appl. Phys. Lett. 62: 2614–2615. Chen, C.T., Ye, B.W., Wu, B.C. et al. (1995). Nature 373: 322–324. Zhang, H.-X., Zhang, J., Zheng, S.-T., and Yang, G.-Y. (2004). Inorg. Chem. Commun. 7: 781–783. Wu, H.-Q., Ju, P., He, H. et al. (2013). Inorg. Chem. 52: 10566–10570. Wang, E.-R., Huang, J.-H., Yu, S.-J. et al. (2017). Inorg. Chem. 56: 6780–6783. Huang, J.-H., Jin, C.-C., Xu, P.-L. et al. (2019). Inorg. Chem. 58: 1755–1758. Wei, Q., Wang, J.-J., He, C. et al. (2016). Chem. Eur. J. 22: 10759–10762. Wang, J.-J., Wei, Q., Yang, B.-F., and Yang, G.-Y. (2017). Chem. Eur. J. 23: 2774–2777. Wang, J.-H., Cheng, J.-W., Wei, Q. et al. (2014). Eur. J. Inorg. Chem.: 4079–4083. Wei, Q., Cheng, J.-W., He, C., and Yang, G.-Y. (2014). Inorg. Chem. 53: 11757–11763. Wang, J.-J., Wei, Q., and Yang, G.-Y. (2017). ChemistrySelect 2: 5311–5315. Wei, Q., Sun, L., Zhang, J., and Yang, G.-Y. (2017). Dalton Trans. 46: 7911–7916. Wei, Q., Sun, S.-J., Zhang, J., and Yang, G.-Y. (2017). Chem. Eur. J. 23: 7614–7620. Wang, G.-M., Sun, Y.-Q., and Yang, G.-Y. (2004). J. Solid State Chem. 177: 4648–4654. Pan, C.-Y., Wang, G.-M., Zheng, S.-T., and Yang, G.-Y. (2007). J. Solid State Chem. 180: 1553–1558. Pan, C.-Y., Wang, G.-M., Zheng, S.-T., and Yang, G.-Y. (2007). Z. Anorg. Allg. Chem. 633: 336–340. Liu, G.Z., Zheng, S.T., and Yang, G.Y. (2007). Inorg. Chem. Commun. 10: 84–87. Wang, J.-H., Wei, Q., Cheng, J.-W. et al. (2015). Chem. Commun. 51: 5066–5068. Wang, J.-J. and Yang, G.-Y. (2017). Chem. Commun. 53: 10398–10401. Xing, H.Z., Li, J.Y., Yan, W.F. et al. (2008). Chem. Mater. 20: 4179–4181. Zhang, H.X., Zheng, S.T., and Yang, G.Y. (2004). Acta Crystallogr., Sect. C 60: M241–M243. Wang, G.-M., Sun, Y.-Q., and Yang, G.-Y. (2005). J. Solid State Chem. 178: 729–735. Wei, Q., Zhang, Y.-J., Song, Y. et al. (2016). Dalton Trans. 45: 13937–13943. Zhao, P., Lin, Z.-E., Wei, Q. et al. (2014). Chem. Commun. 50: 3592–3594. Zhi, S.-C., Wang, Y.-L., Sun, L. et al. (2018). Inorg. Chem. 57: 1350–1355. Wang, G.-M., Sun, Y.-Q., and Yang, G.-Y. (2006). J. Solid State Chem. 179: 1545–1553.

References

219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254

Lehmann, H.A. and Teske, K. (1973). Z. Anorg. Allg. Chem. 400: 169–175. Yu, J.H., Chen, J.S., Xu, R.R. et al. (1996). Polyhedron 15: 4127–4132. Yu, J.H., Xu, R.R., Kan, Q.B. et al. (1993). J. Mater. Chem. 3: 77–82. Yu, J.H., Xu, R.R., Chen, J.S., and Yue, Y. (1996). J. Mater. Chem. 6: 465–468. Ju, J., Lin, J.H., Li, G.B. et al. (2003). Angew. Chem. Int. Ed. 42: 5607–5610. Ju, J., Yang, T., Li, G.B. et al. (2004). Chem. Eur. J. 10: 3901–3906. Su, J., Wang, Y., Wang, Z., and Lin, J. (2009). J. Am. Chem. Soc. 131: 6080–6081. Su, J., Wang, Y., Wang, Z. et al. (2010). Inorg. Chem. 49: 9765–9769. Rong, C., Yu, Z., Wang, Q. et al. (2009). Inorg. Chem. 48: 3650–3659. Cao, G.-J., Wei, Q., Cheng, J.-W. et al. (2016). Chem. Commun. 52: 1729–1732. Cheng, L. and Yang, G.-Y. (2018). Inorg. Chem. 57: 13505–13512. Cheng, L., Wei, Q., Wu, H.Q. et al. (2013). Chem. Eur. J. 19: 17662–17667. Wei, L., Wei, Q., Lin, Z.E. et al. (2014). Angew. Chem. Int. Ed. 53: 7188–7191. Wei, Q., Li, L., Cheng, L. et al. (2014). Dalton Trans. 43: 9427–9430. Zhou, J., Zheng, S.T., Zhang, M.Y. et al. (2009). CrystEngComm 11: 2597–2600. Zhou, J., Fang, W.H., Rong, C., and Yang, G.Y. (2010). Chem. Eur. J. 16: 4852–4863. Liu, Y., Pan, R., Cheng, J.W. et al. (2015). Chem. Eur. J. 21: 15732–15739. Wei, Q., Wang, J.-J., Zhang, J., and Yang, G.-Y. (2017). Inorg. Chem. 56: 6630–6637. Cheng, L. and Yang, G.Y. (2014). Chem. Commun. 50: 344–346. Cheng, J. and Xu, R.R. (1991). J. Chem. Soc., Chem. Commun.: 483–485. Lin, Z.E., Zhang, J., and Yang, G.Y. (2003). Inorg. Chem. 42: 1797–1799. Zhang, H.X., Zhang, J., Zheng, S.T. et al. (2004). Inorg. Chem. 43: 6148–6150. Zhang, H.X., Zhang, J., Zheng, S.T., and Yang, G.Y. (2005). Inorg. Chem. 44: 1166–1168. Wang, G.M., Sun, Y.Q., and Yang, G.Y. (2005). Cryst. Growth Des. 5: 313–317. Pan, C.Y., Liu, G.Z., Zheng, S.T., and Yang, G.Y. (2008). Chem. Eur. J. 14: 5057–5063. Zhang, H.X., Zhang, J., Zheng, S.T., and Yang, G.Y. (2003). Inorg. Chem. 42: 6595–6597. Liu, G.Z., Zhang, H.X., Lin, Z.E. et al. (2007). Chem. Asian J. 2: 1230–1239. Bu, X.H., Feng, P.Y., and Stucky, G.D. (1997). Science 278: 2080–2085. Yu, S.-J., Gu, X.-Y., Deng, T.-T. et al. (2017). Inorg. Chem. 56: 12695–12698. Pan, R., Cheng, J.-W., Yang, B.-F., and Yang, G.-Y. (2017). Inorg. Chem. 56: 2371–2374. He, H., Cao, G.J., Zheng, S.T., and Yang, G.Y. (2009). J. Am. Chem. Soc. 131: 15588–15589. Li, L.-L., Cao, G.-J., Zhao, J.-W. et al. (2016). Inorg. Chem. 55: 5671–5683. Li, L.-L., Pan, R., Zhao, J.-W. et al. (2016). Dalton Trans. 45: 11958–11967. Casey, W.H. (2006). Chem. Rev. 106: 1–16. Mensinger, Z.L., Wang, W., Keszler, D.A., and Johnson, D.W. (2012). Chem. Soc. Rev. 41: 1019–1030. Tripathi, U.M., Schier, A., and Schmidbaur, H. (1998). Z. Naturforsch., B 53: 434–437.

157

158

3 Structural Chemistry of Metal-Oxo Clusters

255 Murugavel, R. and Kuppuswamy, S. (2006). Angew. Chem. Int. Ed. 45: 7022–7026. 256 Starikova, Z.A., Kessler, V.G., Turova, N.Y. et al. (2004). Polyhedron 23: 109–114. 257 Wu, F.J., Simeral, L.S., Mrse, A.A. et al. (2007). Inorg. Chem. 46: 44–47. 258 Dimitrov, A., Koch, J., Troyanov, S.I., and Kemnitz, E. (2009). Eur. J. Inorg. Chem. 2009: 5299–5301. 259 Konig, R., Scholz, G., Veiczi, M. et al. (2011). Dalton Trans. 40: 8701–8710. 260 Jordan, P.A., Clayden, N.J., Heath, S.L. et al. (1996). Coord. Chem. Rev. 149: 281–309. 261 Johansson, G. (1960). Acta Chem. Scand. 14: 771–773. 262 Wang, W., Wentz, K.M., Hayes, S.E. et al. (2011). Inorg. Chem. 50: 4683–4685. 263 Baur, W.H. and Ohta, T. (1982). Acta Crystallogr., Sect. B 38: 390–401. 264 Rowsell, J. and Nazar, L.F. (2000). J. Am. Chem. Soc. 122: 3777–3778. 265 Smart, S.E., Vaughn, J., Pappas, I., and Pan, L. (2013). Chem. Commun. 49: 11352–11354. 266 Abeysinghe, S., Unruh, D.K., and Forbes, T.Z. (2012). Cryst. Growth Des. 12: 2044–2051. 267 Allouche, L., Gerardin, C., Loiseau, T. et al. (2000). Angew. Chem. Int. Ed. 39: 511–514. 268 Mainicheva, E.A., Gerasko, O.A., Sheludyakova, L.A. et al. (2006). Russ. Chem. Bull. 55: 267–275. 269 Son, J.H., Kwon, Y.U., and Han, O.H. (2003). Inorg. Chem. 42: 4153–4159. 270 Sun, Z., Wang, H., Tong, H., and Sun, S. (2011). Inorg. Chem. 50: 559–564. 271 Abeysinghe, S., Unruh, D.K., and Forbes, T.Z. (2013). Inorg. Chem. 52: 5991–5999. 272 (a) Ren, M. and Zheng, L.-M. (2015). Acta Chim. Sin. 73: 1091. (b) Sorace, L., Benelli, C., and Gatteschi, D. (2011). Chem. Soc. Rev. 40: 3092. (c) Ungur, L., Lin, S.Y., Tang, J.K., and Chibotaru, L.F. (2014). Chem. Soc. Rev. 43: 6894. (d) Zheng, Y.Z., Zhou, G.J., Zheng, Z.P., and Winpenny, R.E.P. (2014). Chem. Soc. Rev. 43: 1462. (e) Feltham, H.L.C. and Brooker, S. (2014). Coord. Chem. Rev. 276: 1. (f) Huang, Y.G., Jiang, F.L., and Hong, M.C. (2009). Coord. Chem. Rev. 253: 2814. (g) Liu, J.L., Chen, Y.C., Guo, F.S., and Tong, M.L. (2014). Coord. Chem. Rev. 281: 26. (h) Thompson, L.K. and Dawe, L.N. (2015). Coord. Chem. Rev. 289–290: 13. (i) Wang, H.L., Wang, B.W., Bian, Y.Z. et al. (2016). Coord. Chem. Rev. 306: 195. (j) Liu, J.L., Chen, Y.C., and Tong, M.L. (2016). Chem. Rec. 16: 825. 273 (a) Binnemans, K. (2009). Chem. Rev. 109: 4283. (b) Feng, J. and Zhang, H.J. (2013). Chem. Soc. Rev. 42: 387. (c) Liu, Y.S., Tu, D.T., Zhu, H.M., and Chen, X.Y. (2013). Chem. Soc. Rev. 42: 6924. (d) Rocha, J., Carlos, L.D., Almeida Paz, F.A., and Ananias, D. (2011). Chem. Soc. Rev. 40: 926. (e) Buenzli, J.-C.G. (2015). Coord. Chem. Rev. 293: 19. (f) Kumar, G.A., Riman, R.E., and Brennan, J.G. (2014). Coord. Chem. Rev. 273–274: 111. 274 (a) Tsukube, H. and Shinoda, S. (2002). Chem. Rev. 102: 2389. (b) Butler, S.J. and Parker, D. (2013). Chem. Soc. Rev. 42: 1652. (c) Cui, Y.J., Chen, B.L., and Qian, G.D. (2014). Coord. Chem. Rev. 273: 76.

References

275 (a) Inanaga, J., Furuno, H., and Hayano, T. (2002). Chem. Rev. 102: 2211. (b) Shibasaki, M. and Yoshikawa, N. (2002). Chem. Rev. 102: 2187. (c) Edelmann, F.T. (2012). Chem. Soc. Rev. 41: 7657. (d) Zhao, J.W., Li, Y.Z., Chen, L.J., and Yang, G.Y. (2016). Chem. Commun. (Camb.) 52: 4418. 276 (a) Westin, L.G., Kritikos, M., and Caneschi, A. (2003). Chem. Commun. 39: 1012. (b) Malaestean, I.L., Ellern, A., Baca, S., and Kogerler, P. (2012). Chem. Commun. 48: 1499. (c) Guo, P.H., Liao, X.F., Leng, J.D., and Tong, M.L. (2013). Acta Chim. Sin. 71: 173. (d) Ke, H.S., Xu, G.F., Zhao, L. et al. (2009). Chem. Eur. J. 15: 10335. (e) Ke, H., Zhao, L., Xu, G.F. et al. (2009). Dalton Trans. 38: 10609. (f) Das, S., Dey, A., Kundu, S. et al. (2015). Chem. Eur. J. 21: 16955. (g) Langley, S.K., Moubaraki, B., and Murray, K.S. (2013). Polyhedron 64: 255. (h) Zangana, K.H., Pineda, E.M., McInnes, E.J.L. et al. (2014). Chem. Commun. 50: 1438. (i) Kornienko, A., Emge, T.J., Kumar, G.A. et al. (2005). J. Am. Chem. Soc. 127: 3501. 277 Miao, Y.L., Liu, J.L., Li, J.Y. et al. (2011). Dalton Trans. 40: 10229. 278 (a) Andrews, P.C., Beck, T., Forsyth, C.M. et al. (2007). Dalton Trans. 36: 5651. (b) Ren, Y.X., Zheng, X.J., Li, L.C. et al. (2014). Inorg. Chem. 53: 12234. (c) Miao, Y.L., Liu, J.L., Leng, J.D. et al. (2011). CrystEngComm 13: 3345. (d) Kretschmer, W.P., Teuben, J.H., and Troyanov, S.I. (1998). Angew. Chem. Int. Ed. Engl. 37: 88. (e) Kajiwara, T., Katagiri, K., Takaishi, S. et al. (2006). Chem. Asian J. 1: 349. (f) D’Alessio, D., Sobolev, A.N., Skelton, B.W. et al. (2014). J. Am. Chem. Soc. 136: 15122. (g) Zhao, L., Xue, S.F., and Tang, J.K. (2012). Inorg. Chem. 51: 5994. 279 Chesman, A.S.R., Turner, D.R., Moubaraki, B. et al. (2009). Chem. Eur. J. 15: 5203. 280 (a) Wang, R.Y., Song, D.T., and Wang, S.M. (2002). Chem. Commun. 38: 368. (b) Li, X.L., He, L.F., Feng, X.L. et al. (2011). CrystEngComm 13: 3643. (c) Burgstein, M.R., Gamer, M.T., and Roesky, P.W. (2004). J. Am. Chem. Soc. 126: 5213. (d) Burgstein, M.R. and Roesky, P.W. (2000). Angew. Chem. Int. Ed. 39: 549. (e) Chesman, A.S.R., Turner, D.R., Moubaraki, B. et al. (2012). Dalton Trans. 41: 10903. (f) Tian, H., Bao, S.S., and Zheng, L.M. (2016). Chem. Commun. 52: 2314. 281 (a) Thielemann, D.T., Wagner, A.T., Rosch, E. et al. (2013). J. Am. Chem. Soc. 135: 7454. (b) Wang, R.Y., Zheng, Z.P., Jin, T.Z., and Staples, R.J. (1999). Angew. Chem. Int. Ed. 38: 1813. (c) Wang, R.Y., Selby, H.D., Liu, H. et al. (2002). Inorg. Chem. 41: 278. 282 Moore, B.F., Kumar, G.A., Tan, M.C. et al. (2011). J. Am. Chem. Soc. 133: 373. 283 Su, K.Z., Jiang, F.L., Wu, M.Y. et al. (2016). CrystEngComm 18: 4928. 284 Lin, W.Q., Liao, X.F., Jia, J.H. et al. (2013). Chem. Eur. J. 19: 12254. 285 Chang, L.X., Xiong, G., Wang, L. et al. (2013). Chem. Commun. 49: 1055. 286 (a) Gu, X.J. and Xue, D.F. (2007). Inorg. Chem. 46: 3212. (b) Gu, X.J., Clerac, R., Houri, A., and Xue, D.F. (2008). Inorg. Chim. Acta 361: 3873. (c) Chen, L., Huang, L., Wang, C.L. et al. (2012). J. Coord. Chem. 65: 958. (d) Huang, L., Han, L.J., Feng, W.J. et al. (2010). Cryst. Growth Des. 10: 2548. (e) Zhang, Y.,

159

160

3 Structural Chemistry of Metal-Oxo Clusters

287 288 289 290 291 292 293 294

295 296 297 298

299 300 301 302 303 304 305 306 307 308

Huang, L., Miao, H. et al. (2015). Chem. Eur. J. 21: 3234. (f) Sen, R., Hazra, D.K., Mukherjee, M., and Koner, S. (2011). Eur. J. Inorg. Chem. 2011: 2826. Zheng, X.Y., Peng, J.B., Kong, X.J. et al. (2016). Inorg. Chem. Front. 3: 320. Romanelli, M., Kumar, G.A., Emge, T.J. et al. (2008). Angew. Chem. Int. Ed. 47: 6049. Wu, M.Y., Jiang, F.L., Kong, X.J. et al. (2013). Chem. Sci. 4: 3104. Guo, F.S., Chen, Y.C., Mao, L.L. et al. (2013). Chem. Eur. J. 19: 14876. (a) Wu, M.Y., Jiang, F.L., Yuan, D.Q. et al. (2014). Chem. Commun. 50: 1113. (b) Chen, L., Guo, J.Y., Xu, X. et al. (2013). Chem. Commun. 49: 9728. Kong, X.J., Wu, Y., Long, L.S. et al. (2009). J. Am. Chem. Soc. 131: 6918. Peng, J.B., Kong, X.J., Zhang, Q.C. et al. (2014). J. Am. Chem. Soc. 136: 17938. (a) Sessoli, R., Gatteschi, D., Caneschi, A., and Novak, M. (1993). Nature 365: 141. (b) Sessoli, R., Tsai, H.L., Schake, A.R. et al. (1993). J. Am. Chem. Soc. 115: 1804. (c) Wernsdorfer, W. and Sessoli, R. (1999). Science 284: 133. (d) Leuenberger, M.N. and Loss, D. (2001). Nature 410: 789. (e) Bogani, L. and Wernsdorfer, W. (2010). Nanoscience and Technology: A Collection of Reviews from Nature Journals, 194. World Scientific. (f) Mannini, M., Pineider, F., Sainctavit, P. et al. (2009). Nat. Mater. 8: 194. (g) Thiele, S., Balestro, F., Ballou, R. et al. (2014). Science 344: 1135. Zheng, X.-Y., Jiang, Y.-H., Zhuang, G.-L. et al. (2017). J. Am. Chem. Soc. 139: 18178. Zhou, Y., Zheng, X.-Y., Cai, J. et al. (2017). Inorg. Chem. 56: 2037. Qin, L., Yu, Y.Z., Liao, P.Q. et al. (2016). Adv. Mater. 28: 10772. (a) Kajiwara, T., Iki, N., and Yamashita, M. (2007). Coord. Chem. Rev. 251: 1734. (b) Ovsyannikov, A., Solovieva, S., Antipin, I., and Ferlay, S. (2017). Coord. Chem. Rev. 352: 151. (c) Bi, Y., Du, S., and Liao, W. (2014). Coord. Chem. Rev. 276: 61. Bilyk, A., Dunlop, J.W., Fuller, R.O. et al. (2010). Eur. J. Inorg. Chem. 2010: 2106. Bi, Y., Wang, X.-T., Liao, W. et al. (2009). J. Am. Chem. Soc. 131: 11650. Bi, Y., Liao, W., Xu, G. et al. (2010). Inorg. Chem. 49: 7735. Xiong, K., Jiang, F., Gai, Y. et al. (2012). Cryst. Growth Des. 12: 3335. Tan, H., Du, S., Bi, Y., and Liao, W. (2014). Inorg. Chem. 53: 7083. Bi, Y., Xu, G., Liao, W. et al. (2010). Chem. Commun. 46: 6362. Bi, Y., Wang, S., Liu, M. et al. (2013). Chem. Commun. (Camb.) 49: 6785. Kajiwara, T., Wu, H., Ito, T. et al. (2004). Angew. Chem. Int. Ed. 43: 1832. Kajiwara, T., Katagiri, K., Takaishi, S. et al. (2006). Chem. Eur. J. 1: 349. (a) Su, K., Wu, M., Tan, Y. et al. (2017). Chem. Commun. 53: 9598. (b) Su, K., Jiang, F., Qian, J. et al. (2014). Inorg. Lett. 1: 1. (c) Su, K., Jiang, F., Qian, J. et al. (2014). Cryst. Growth Des. 14: 3116. (d) Su, K., Jiang, F., Qian, J. et al. (2015). RSC Adv. 5: 33579. (e) Su, K., Jiang, F., Qian, J. et al. (2014). Cryst. Growth Des. 14: 5865. (f) Su, K., Jiang, F., Qian, J. et al. (2015). Inorg. Chem. Commun. 54: 34. (g) Su, K., Jiang, F., Qian, J. et al. (2015). CrystEngComm 17: 1750. (h) Su, K., Jiang, F., Qian, J. et al. (2014). Inorg. Chem. 53: 18. (i) Su, K., Jiang, F., Qian, J. et al. (2013). Inorg. Chem. 52: 3780. (j) Su, K., Jiang, F.,

References

309

310 311

312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327

Wu, M. et al. (2016). CrystEngComm 18: 4921. (k) Su, K., Wu, M., Wang, W. et al. (2018). Sci. China Chem. 61: 664. (l) Su, K., Wu, M., Yuan, D., and Hong, M. (2018). Nat. Commun. 9: 4941. (m) Su, K.Z., Jiang, F.L., Qian, J.J. et al. (2015). Inorg. Chem. 54: 3183. (n) Xiong, K., Jiang, F., Gai, Y. et al. (2012). Aust. J. Basic Appl. Sci. 3: 2321. (o) Xiong, K., Jiang, F., Gai, Y. et al. (2012). Inorg. Chem. 51: 3283. (p) Xiong, K., Wang, X., Jiang, F. et al. (2012). Chem. Commun. 48: 7456. (q) Xiong, K.-C., Jiang, F.-L., Gai, Y.-L. et al. (2012). Chem. Eur. J. 18: 5536. (r) Yuan, D.Q., Wu, M.Y., Wu, B.L. et al. (2006). Cryst. Growth Des. 6: 514. (a) Tasiopoulos, A.J., Vinslava, A., Wernsdorfer, W. et al. (2004). Angew. Chem. Int. Ed. 43: 2117. (b) Ochsenbein, S.T., Murrie, M., Rusanov, E. et al. (2002). Inorg. Chem. 41: 5133. (c) Brechin, E.K., Boskovic, C., Wernsdorfer, W. et al. (2002). J. Am. Chem. Soc. 124: 9710. Liu, J.-L., Chen, Y.-C., and Tong, M.-L. (2018). Chem. Soc. Rev. 47: 2431. (a) Liu, J., Chen, Y.-C., Liu, J.-L. et al. (2016). J. Am. Chem. Soc. 138: 5441. (b) Chen, Y.-C., Liu, J.-L., Ungur, L. et al. (2016). J. Am. Chem. Soc. 138: 2829. (c) Feng, X., Hwang, S.J., Liu, J.-L. et al. (2017). J. Am. Chem. Soc. 139: 16474. Pugh, T., Chilton, N.F., and Layfield, R. (2016). Angew. Chem. Int. Ed. 55: 11082. Meng, Y.S., Zhang, Y.Q., Wang, Z.M. et al. (2016). Chem. Eur. J. 22: 12724. Guo, F.S., Day, B.M., Chen, Y.C. et al. (2017). Angew. Chem. Int. Ed. 56: 11445. Goodwin, C.A., Ortu, F., Reta, D. et al. (2017). Nature 548: 439. Guo, F.-S., Day, B.M., Chen, Y.-C. et al. (2018). Science 362: 1400. Zheng, Z. and Burns, P.C. (2017). Recent Development in Clusters of Rare Earths and Actinides: Chemistry and Materials. Springer-Verlag. Zhang, J.-J., Hu, S.-M., Xiang, S.-C. et al. (2006). Inorg. Chem. 45: 7173. Xiang, S., Hu, S., Sheng, T. et al. (2009). Chem. Eur. J. 15: 12496. Liu, D.-P., Lin, X.-P., Zhang, H. et al. (2016). Angew. Chem. Int. Ed. 55: 4532. Kong, X.J., Ren, Y.P., Chen, W.X. et al. (2008). Angew. Chem. Int. Ed. 47: 2398. Kong, X.-J., Long, L.-S., Huang, R.-B. et al. (2009). Chem. Commun.: 4354. Zhang, Z.-M., Pan, L.-Y., Lin, W.-Q. et al. (2013). Chem. Commun. 49: 8081. Zheng, H., Du, M.-H., Lin, S.-C. et al. (2018). Angew. Chem. Int. Ed. 57: 10976. Chen, W.-P., Liao, P.-Q., Yu, Y. et al. (2016). Angew. Chem. Int. Ed. 55: 9375. Chen, R., Yan, Z.-H., Kong, X.-J. et al. (2018). Angew. Chem. Int. Ed. 57: 16796. Zheng, X.-Y., Zhang, H., Wang, Z. et al. (2017). Angew. Chem. Int. Ed. 56: 11475.

161

163

4 Synthetic Approaches, Structural Diversities, and Biological Aspects of Molybdenum (Tungsten)-Based Heterometallic Sulfide Clusters and Coordination Polymers Du Shaowu 1 and Wu Xintao 2 1 Minjiang University, Fujian Key Laboratory of Functional Marine Sensing Materials, College of Physics & Electronic Information Engineering, 200 Xiyuangong Road, Fuzhou, Fujian, 350108, P.R. China 2 Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, State Key Laboratory of Structural Chemistry, 155 Yangqiao Road West, Fuzhou, Fujian, 350002, P.R. China

4.1 Introduction The importance of transition metal sulfide clusters has long been recognized for their structural diversity and their unusual properties associated with facile electron and proton transfer [1]. Among them, the family of heterometallic sulfide clusters with biological metals like Mo and W are of particular interest because they can be potentially used as structural models for various biological and industrial catalytic processes [2]. Molybdenum heterometallic thioclusters, for example, have been found in the active sites of some metalloenzymes, many of which play essential roles in living organisms. The most widely studied of these clusters is the FeMo cofactor of nitrogenase, a metal–sulfur cluster with a {MoFe7 S9 } core structure, which promotes the conversion of dinitrogen into ammonia [3]. Apart from the nitrogen biogeochemical cycle, molybdenum heteronuclear clusters also participate in the global carbon cycle, maintaining safe levels of carbon monoxide in our atmosphere. A typical example is the carbon monoxide dehydrogenase, Mo–Cu CODH, which consists of a unique binuclear Mo–Cu–S cofactor capable of catalyzing CO oxidation to CO2 [4]. In addition to biological catalysis, the development of heterometallic sulfide clusters has also been driven by their potential applications in industrial catalysis. For instance, as nickel and cobalt are excellent catalyst promoters for MoS2 , some cuboidal clusters with Mo2 M2 ′ S4 and Mo3 M′ S4 cores (M′ = Co, Ni) can serve as molecular models to mimic the Mo–Ni(Co)–S sites of heterogeneous catalysts for industrial hydrodesulfurization process based on the so-called “cluster-surface analogy” [5]. Furthermore, it has been shown recently that several Mo(W)–Cu(Ag)–S clusters possess strong nonlinear optical behaviors such as optical limiting effects and third-order NLO susceptibilities [6]. The most convenient and straightforward approach for the synthesis of metal sulfide clusters involves the spontaneous self-assembly of metal cations and sulfide anions by one-pot reactions, usually with stabilizing ancillary ligands. Such one-pot Advanced Structural Chemistry: Tailoring Properties of Inorganic Materials and their Applications, First Edition. Edited by Rong Cao. © 2021 WILEY-VCH GmbH. Published 2021 by WILEY-VCH GmbH.

164

4 Synthetic Approaches, Structural Diversities, and Biological Aspects of Molybdenum (Tungsten)

reactions often lead to a mixture of products, and in most cases, the outcome of such reactions is unpredictable. This is even more so for the synthesis of heterometallic sulfide clusters which usually results in a mixture of homometallic sulfide clusters. The “Unit construction” strategy, on the other hand, has evolved to rationally synthesize heterometallic sulfide clusters [7]. This approach treats reactive metal sulfide units as building blocks, which allows for the sequential installation of heterometals with atomic precision into the heterometallic assemblies. Accordingly, a metal sulfide unit which contains lone-pair electrons on the sulfur atom(s) may readily combine with another metal complex carrying unsaturated coordination sites or easily removable ligands. These possibilities arise mainly due to the versatile coordination modes of the sulfide S2− ligand, which is able to coordinate in a terminal, bridging, and encapsulated fashion, enabling multiple metal atoms to be bridged through a single S2− ligand. As the coordination mode moves from terminal to μ2 -S and to μ3 -S or μ4 -S(a) bridges, the reactivity of sulfido ligands declines, whereas for μ4 -S(b), μ5 -S, and μ6 -S bridges, the reactivity may completely disappear, a result consisting with the gradual reduction of lone-pair electrons (Table 4.1) [21]. The constructive reactivity of sulfido ligands with various coordination modes is summarized as follows: (1) terminal sulfido ligand (term-S). One typical example is the tetrathiomolybdate dianion MoS4 2− in which sulfur atoms are each doubly bonded to the Mo center. Such a terminal S atom is highly reactive toward metal ions because it only binds to one metal ion and possesses two lone-pairs of electrons. (2) μ2 -Bridging sulfido ligand (μ2 -S). There are three subtypes of this coordination configuration (a, b, and c). In type a, the sulfido ligand bridges two metal atoms with a metal–metal bond, forming a so-called M2 S island, a triangular fragment that is the fundamental structural unit of a metal sulfide cluster. For type b, two sulfido ligands span two metal atoms to produce a M2 S2 rhomb that can be viewed as two M2 S triangles fused by sharing the M–M edge, as observed in Roussins red salt. Due to the existence of d–p bonding between metal and sulfur, the number of lone-pair electrons on the sulfur of both type a and type b is equal to or less than four. The type c is a disulfido ligand with two or three lone-pair electrons, in which each sulfur is similar to a μ3 -S atom. The S—S bond is easily disrupted and its reactivity is similar to the type b sulfido ligand. (3) μ3 -Bridging sulfido ligand (μ3 -S). This type of sulfido ligand donates four electrons in total to form three M–μ3 –S covalent bonds, leading to reduced reactivity. The remaining one pair of lone-pair electrons occupies one corner of a tetrahedral geometry of the sulfur atom. The μ3 -S ligand exists widely in nature because of its high stability. This type of sulfido atom can be categorized into four different subtypes (a, b, c, and d). Type a is exemplified by two well-known clusters M3 S4 4+ (M = Mo and W), both of which contain a M3 (μ2 -S)3 triangle capped by a μ3 -S atom. Type b can be found in the triangular metal clusters with strong metal–metal interactions such as Co3 S2 , where two μ3 -S atoms capped above and below the Co3 plane, respectively. This type of cluster is usually stable and easily available. Type c is commonly observed in the cubane-like clusters, in which four metal atoms and four sulfur atoms occupy alternate corners in a cube structure. Each sulfur atom at one vertex caps a trimetallic plane and each metal at one vertex bonds to the three sulfur atoms. Sulfido atoms of the final type d can be

4.1 Introduction

Table 4.1 The relationship between coordination configuration of the sulfide ligands and constructive reactivity.

Type of Coordination No. coordination configuration

No. of lone-pair C. N. of Assembly electrons S atoms activity Example

1

Term-S

S

4

1

MoS4 2−

[8]

2

μ2 -S

S

≤4

2

Pt2 S(PPh3 )4

[9]

≤4

2

[Mo2 O2 S2 (S2 )(S4 )]2−

[10]

2–3

3

Fe2 S2 (CO)6

[11]

2

3

[Mo3 S4 ]4+

[12]

[13]

M

(a)

References

M

M

(b)

S M

M S

(c)

S M

M S

3

μ3 -S

(a)

S M

M M

Increase (b)

S M

2

3

Co3 S2 (C5 H5 )3

2

3

[Mo3 CuS4 ]5+ [14]

2

3

Mo6 S8 (PEt3 )6 [15]

2

4

Co4 S2 (CO)10 [16]

M

M S

(c)

S

M

S M

S

(d)

M S M

S

S M S S M M M M S S M S S

(a)

S M

M

M

M

(Continued)

165

166

4 Synthetic Approaches, Structural Diversities, and Biological Aspects of Molybdenum (Tungsten)

Table 4.1

(Continued)

Type of Coordination No. coordination configuration

No. of lone-pair C. N. of electrons S atoms

4

0

4

0

μ4 -S

(b)

M

5

[17]

5

Os5 (CO)15 (μ5 -S) [W(CO)4 PPh3 ]

[18]

M S M

M

6

μ6 -S

0

6

[(μ6 -S)Cu6 S6 [19] (S2 )6 Mo6 O6 ]2−

0

6

[{Mo2 Cu2 O2 S2 (edt)2 }3 (μ6 -S2 )]2−

M

M

μ5 -S

(a)

(b)

M

M

M

M S

M M

M M

M

M

S S

M M

No activity

References

Fe4 (CO)12 (SR)2

S M

Assembly activity Example

M

M

[20]

Source: Wu et al. [21].

seen in the stable hexanuclear cluster [Mo6 S8 (PEt)6 ], where the eight sulfido ligands capped the triangular plane of the octahedral core formed by six Mo atoms. This arrangement permits each sulfur to bond to three Mo atoms simultaneously. (4) μ4 -Bridging sulfido ligand (μ4 -S). This type of sulfido ligand acts as a bridge across four metal atoms. They can be a four-electron donor capping the tetrametallic plane (type a) or tetrahedrally coordinated by four metal atoms (type b). The former has some reactivity due to the presence of one lone-pair electrons, whereas the latter has no reactivity because all six of its valence electrons have been used up for making four metal–sulfur bonds. (5) μ5 -Bridging sulfido ligand (μ5 -S). This type of sulfido atom appears in clusters containing M5 S structural units, as displayed in metal carbonyl clusters Ru4 (CO)7 (μ-CO)2 (PMe2 Ph)(μ4 -S)(μ5 -S)[W(CO)4 PMe2 Ph] and Os5 (CO)15 (μ5 -S)[W(CO)4 PPh3 ], in which the μ5 -S ligand serves as a six-electron donor and has a tetragonal-pyramidal (TP) coordination geometry. (6) μ6 -Bridging sulfide ligand (μ6 -S). This type of ligand contains two subtypes. The first type (type a) functions as a six-electron donor and is present in the clusters with an M6 S structural unit. It is unusual but not without precedent. For instance, it was observed in the dodecanuclear cage-like Mo–Cu(Ag)–S clusters such as [Et4 N]2 [{Mo2 Cu2 O2 S2 (S2 )2 }3 (μ6 -S)] and [Et4 N]2 [{Mo2 Ag2 O2 S2 (edt)2 }3 (μ6 -S)] (edt = 1,2-ethanedithiolate)⋅CH3 CN. As in μ4 -S(b) and μ5 -S, all six electrons of a μ6 -S atom participate in forming six M—S bonds. The second type (type b) is a disulfido

4.2 Synthesis of Mo–Fe–S Cuboidal Clusters for the Structural Modeling

M

{3 + 1}

M′

+ S

M′

S

M

S

M

M

S

M

S

S

M

S

S S

M′ S

M

M

S

+

M′ S

S

{2 + 2}

M′

S M

S

S

M

M′

S

M S

S M

+ M′ + M′

{2 + 2[1]}

S

Figure 4.1

Unit construction synthetic approach for heterometallic cubane clusters.

ligand in which a bond between two sulfur atoms exists. Each sulfur atom is similar to a μ4 -S(b). The whole μ6 -S2 ligand contributes 10 electrons to form six M—S bonds, as demonstrated by the cage cluster [Et4 N]2 [{Mo2 Cu2 O2 S2 (edt)2 }3 (μ6 -S2 )]⋅CH2 Cl2 , in which the interstitial disulfido ligand binds six Ag atoms. The above discussion indicates that by this unit construction method, one may combine different building blocks with chemical and structural complementarities in a much more controllable way, producing a series of heterometallic sulfide clusters with diverse nuclearities and structural arrangements, such as butterflies, cubanes, cages, squares, and even coordination polymers. For instance, as illustrated in Figure 4.1, the cubane cluster formed by adding a monomeric complex to a triangular metal cluster, or by a combination of two dinuclear rhombuses, is designated as {3 + 1} or {2 + 2}. They can also be prepared from a dinuclear cluster and two monomeric complexes through a {2 + 2[1]} route. Repeating unit construction steps will then result in the formation of polymeric compounds.

4.2 Synthesis of Mo–Fe–S Cuboidal Clusters for the Structural Modeling of the Iron–Molybdenum Cofactor (FeMoco) The structural and mechanistic study of the FeMoco active center in nitrogenases, the only type of enzymes known in nature capable of reduction atmospheric dinitrogen to ammonia, has remained a central topic in chemistry and biology owing to their importance for global nitrogen cycle, agriculture, and fertilizer manufactures. The FeMoco active center is a remarkable heterometallic sulfide cluster composed of seven irons and one molybdenum organized by nine sulfide atoms to form a linked incomplete double cubane cluster {MoFe7 S9 }. In the structure of {MoFe7 S9 }, the two cuboidal subunits {Fe4 (μ3 -S)3 } and {MoFe3 (μ3 -S)3 } are bridged by three μ2 -S atoms

167

168

4 Synthetic Approaches, Structural Diversities, and Biological Aspects of Molybdenum (Tungsten)

S

S S

Fe

Fe

Fe

S

Fe

Mo

Fe S

S

Fe

c

Fe

Fe

S

Mo

S

S

N

Fe

Fe

S

S

S

Fe

S

S

S

S

(c) S

Fe S

Fe

N

Fe

Fe

S

Fe N

S

(b) S

Mo

Fe

Fe

S

Fe

Fe

N

Fe

N

Mo

S

S S

S

S

(d)

S

Fe

Fe

S S

Fe

Fe

Y

Fe

S

S

N

(a)

S

Fe

Fe

Fe

Mo

Fe

S

(e)

Fe

Fe

S

S

Figure 4.2 (a) The structure of FeMoco cluster core. (b) Activation of dinitrogen in Fuzhou model I. (c) View of Fuzhou model II. (d) Structural model for the coordination activation of N2 in the core of the FeMoco cavity as proposed by Rees et al. (Y = O, N). (e) New Fuzhou model proposed for the coordination activation of N2 . Source: Based on Refs. [22–32].

and a central interstitial light atom that was only until recently identified as a carbide atom by combined spectroscopic and structural analysis (Figure 4.2a) [22–24]. The terminal Mo atom is also chelated by a homocitrate molecule and the whole cluster is anchored to the protein only through one cysteine and one histidine residues at each end of the double cubane. Although the atomic structure of FeMoco center is now finally clarified and the electron transfer pathway from the [4Fe–4S] cluster to the [8Fe–7S] P-cluster and then to the FeMoco during the dinitrogen reduction has become widely accepted, the detailed mechanism for this multielectron catalytic reaction that occurred at the FeMoco is still unknown. Therefore, the development of synthetic model studies on the FeMoco allows us to understand at a molecular level how such a complex biosystem can be assembled and how it operates to activate the strong bond of dinitrogen under ambient conditions. However, its incredible complexity in terms of the metal–sulfur connectivity which, together with the presence of mysterious yet important carbide atom, makes the FeMoco the greatest challenge for modeling chemistry. During the 1970s and 1980s, before the first crystallographic structure of the FeMoco was revealed at 2.7 Å resolution [28, 29], a range of proposed structural models for the FeMoco active center have been put forward as deduced from the spectroscopic and biochemical evidences, as well as from the structural criteria for dinitrogen activation based on the coordination chemistry of transition metal–dinitrogen complexes. Among these are structural models called Fuzhou models suggested by Professor Lu Jiaxi, which represent the earliest and most successful cluster models (Fuzhou model I, Figure 4.2b and Fuzhou model II, Figure 4.2c) [25]. Enlightened by the structure of monoanionic black Roussinate and taking into consideration that multiple active metal sites are essential for

4.2 Synthesis of Mo–Fe–S Cuboidal Clusters for the Structural Modeling

simultaneous promotion of N2 binding and reduction, both models were proposed to contain a {MoFe3 S3 } “string-bag” unit (a MoFe3 S4 cube with one sulfur atom missing from a corner of the cube). Fuzhou mode II was proposed according to the elemental composition of the FeMoco Mo/Fe/S ≈ 1 : 8 : 6 [30] and was a Siamese-twin of two units of Fuzhou model I. Theoretical investigations suggested that the N2 molecule tended to vertically insert into these “string-bag” units, and then activated by the synergic action of the metal centers through one end-on and three side-on metal–dinitrogen interactions, as is shown in Figure 4.2b. The combined action of end-on and side-on binding may enhance the activation of the inert N≡N triple bond. Furthermore, multiple side-on binding is absolutely necessary to prevent any of them from possible isomerization into the end-on type. In the early 1990s, Kim, Rees and coworker published the crystal structure of nitrogenase from Azotobacter vinelandii with a resolution of 2.7 Å, and subsequently, the structure at 2.2 Å resolution was also obtained [26]. Meanwhile, Bolin, coworkers reported the crystal structure of nitrogenase from Clostridium pasteurianum with the resolution refined to 2.2 Å [31]. Based on the crystallographic results, they were able to propose structural models for the active center of the FeMoco. Both their proposed models are made up of two “string-bag” clusters {MoFe3 S3 } and {Fe4 S3 } coupled in a “neck-to-neck” fashion. A quasi-mirror S–Y–S plane is located between the two clusters, resulting in an approximate C3h point group symmetry (Figure 4.2d). The only difference between Ree’s model and Bolin’s model is the triad of bridging atoms, namely, two S atoms and one “Y” atom (either O or N) for the former and three S atoms for the latter. Soon after they reported the structure of nitrogenase, Chan et al. suggested that the N2 molecule was likely to be accommodated in the central cavity of the active center of FeMoco and activated only by interacting with six belt Fe ions (Figure 4.2d) [26]. It is worth pointing out that the structure of {MoFe3 S3 } “string-bag” cluster in Rees and Bolin’s models as well as the mechanism thereupon proposed for the activation of N2 were quite similar to the Fuzhou model I suggested by Professor Lu Jiaxi nearly two decades ago. In view of the structural features of Fuzhou models I and II, in combination with those of Chan–Kim–Rees (CKR) and Bolin’s models, a new N2 coordination activation model, namely, New Fuzhou model, was thus proposed (Figure 4.2e) [32]. The active center of the FeMoco was taken from the CKR model and the Bolin’s model by coupling two “string-bag” clusters {MoFe3 S3 } and {Fe4 S3 } with three μ2 -sulfur bridges to form a double “string-bag” cage. Such an arrangement created three puckered octagonal {Fe4 S4 } windows through which an N2 molecule was able to penetrate into the cavity of the cage. Once inside the cage, one N atom approached close to the Mo atom, creating an end-on Mo—N bond. At the same time, the other N atom would orient itself to interact with three nearby Fe atoms of the {MoFe3 S3 } “string-bag” by side-on coordination, pretty much the same as that in Fuzhou model I. The other “string-bag” cluster {Fe4 S3 }, meanwhile, only served for the purpose of electron storage and transfer in the reduction of the coordinated N2 molecule. Although at present there is an increasing evidence that the front face rather than the central cavity of FeMoco is the catalytic reaction zone [33], these earlier proposed structural models have spurred the development of heterometallic

169

170

4 Synthetic Approaches, Structural Diversities, and Biological Aspects of Molybdenum (Tungsten)

Mo–Fe–S clusters with particular emphasis on the designed synthesis of those having a cubane-like structure. The most common method to synthesize heterometallic Mo–Fe–S cubane-like clusters is via the spontaneous self-assembly reactions of simple inorganic salts, most often tetrathiomolybdate and iron chloride, with thiolate ligands. With monodentate thiolate ligands, the reactions usually led to double cubane clusters with low cluster core oxidation states, as exemplified by [Mo2 Fe6 S8 (SR)9 ]3− , the first Mo–Fe–S clusters containing a cubane structure, obtained from the reactions of (NH4 )2 MoS4 with FeCl2 and RS− (Figure 4.3a) [34]. These clusters are composed of two identical [MoS4 Fe3 (SR)3 ] subunits triply bridged by three μ2 -SR ligands across the Mo atoms. There was, however, one exception of this type of reaction that yielded a single cubane [MoFe4 S4 (SC6 H11 )7 ]2− when bulky cyclohexanethiolate was used (Figure 4.3b) [35]. This cluster has a single cubane MoFe3 S4 core with an additional Fe atom hanging outside the cubane. After being treated with citric acid, it converted to the normal double cubane cluster [Mo2 Fe6 S8 (SC6 H11 )9 ]3− (Figure 4.3c). When using bidentate thiolates such as dithiocarbamate R2 dtc− , the reactions of (NH4 )2 MoS4 with FeCl2 afforded exclusively Mo–Fe–S single cubanes. They all have a similar {MoFe3 S4 } cubane-like structure, but different cluster core oxidation states depending on the thiolate substituents as well as the solvents and metal to thiolate ligand ratios used. For example, the reaction with a large (threefold) excess of Et2 dtc− gave the monoanionic cubane [MoFe3 S4 (Et2 dtc)5 ]− as a major product, whereas with a twofold excess, a neutral cubane [MoFe3 S4 (Et2 dtc)5 ] was isolated (Figure 4.3d) [36]. The two clusters have similar overall cubane geometries, but displaying +4 and +5 cluster core oxidation states, respectively. Another single cubane [MoFe3 S4 (Me2 dtc)6 ] with +6 core oxidation state could be NR2 C S Cl

[Fe(DMF) ] 6

Cl

Fe

Mo

Cl

S S

S

R2dtc‒

Fe

R2NC

(g)

Cl S [Fe(DMF)6][MoS4(FeCl2)2]

(h)

SR S Fe S

S Fe

Mo

R S R S

S Mo

S R

S

SR

S

S

(a)

R = Et

(e) SR

SR‒ NR2

(b)

S

SR Fe

Fe S

S Fe

S

Mo

R S R S

R2NC S Fe

SR

S R

[Fe(DMF)6]

Mo S

S S

Cl Fe Cl

SR [Et4N2[MoFe4S4(SR)7] R = C6H11

S

n = 2: [MoFe3S4(R2dtc)5]

(c) Citric acid SR S

S S Fe CNR2 S

n = 3: [MoFe3S4(R2dtc)5]‒

nR2dtc‒

(NH4)2MoS4 + FeCl2

Fe

S

[Mo2Fe6S8(SR)9]3‒

SR

SR‒

CNR2

S

Fe

R2NC

(d)

Fe

Fe

S Fe

S

S

R2dtc‒

SR Fe

S

S S

Mo

S

(f) SR‒ SR

S

[Fe(DMF)6][MoS4(FeCl2)]

S S

S S

S Fe S NR S 2 S C S Fe S S S Fe S S R2NC S CNR2 Mo

CNR2

[MoFe3S4(R2dtc)6] R = Me

Figure 4.3 Synthetic routes for the cuboidal Mo–Fe–S clusters. Source: Modified from Refs. [34–37].

4.2 Synthesis of Mo–Fe–S Cuboidal Clusters for the Structural Modeling

prepared in an analogous manner by using Me2 dtc− instead of Et2 dtc− (Figure 4.3e) [36]. Compared with the monodentate thiolates, bidentate R2 dtc− are weaker reducing agents that explained their preferable formation of cubane cores with higher oxidation states. Moreover, the peripheral metal binding sites of clusters might be fully blocked by the intracubane bridging and chelating R2 dtc− ligands, which results in the prevented coupling of the single cubanes. In addition to the above one-pot self-assembly reactions which sometimes required fractional crystallization processes to separate mixtures of cluster compounds. These cubane-like Mo–Fe–S clusters could also be prepared more efficiently. It was discovered that the di- or trinuclear linear Mo–Fe–S clusters could be readily converted into cubane clusters when reacted with thiolate ligands. Whether the product was a double cubane or a single cubane again depended on the nature of thiolates. Thus, upon treatment with RS− , the linear trinuclear cluster [Fe(DMF)6 ][MoS4 (FeCl2 )2 ] was readily transferred to the double cubane-like cluster [Et4 N]3 [Mo2 Fe6 S8 (SR)9 ] (Figure 4.3f), while with R2 dtc− , the single cubanes [MoFe3 S4 (R2 dtc)5 ] formed (Figure 4.3g) [37]. Experimental evidence has suggested that this linear cluster could be the intermediate for the formation of cubane clusters. In fact, linear dinuclear cluster like [Fe(DMF)6 ][MoS4 (FeCl2 )] also works well for this cluster conversion reaction (Figure 4.3h) [37]. Such linear clusters all contain {MoS2 Fe} rhombs, a basic face for the {MoFe3 S4 } cubane, which in principle should be able to combine with {FeS2 Fe} rhombs to constitute a cubane-type structure. In order to test this hypothesis, linear clusters [Mg(DMF)6 ][Cl2 FeS2 FeCl2 ] and [Mg(DMF)6 ][(FeCl2 )MoS4 ] were mixed in acetonitrile followed by the addition of R2 dtcNa, which, after a few hours reaction, indeed produced the single cubanes [MoFe3 S4 (R2 dtc)5 ] (Figure 4.4a) [27]. Here Mg(DMF)6 2+ was used to replace [Fe(DMF)6 ]2+ to avoid the interference of free Fe2+ ions. Following the same S Fe

Fe S Fe

S

S Fe

M = M′ = Fe

(c) S S

M

Fe

Mo

M = Mo; M′ = Fe

S Fe

S

S Fe

(a)

S

Fe S

M = M′ = Mo

S

(b)

Fe S

Fe

Mo

S

Fe

S

S Mo

M′

Figure 4.4 A {2 + 2} unit construction route for cubane-like Mo–Fe–S clusters. Source: Based on Liu et al. [27].

171

172

4 Synthetic Approaches, Structural Diversities, and Biological Aspects of Molybdenum (Tungsten)

idea, cubane clusters, such as [Fe4 S4 (R2 dtc)4 ] and [Mo2 Fe2 S4 (R2 dtc)5–6 ], could be facilely synthesized by simply coupling two {FeS2 Fe} and two {MoS2 Fe} rhombic units, respectively (Figure 4.4b,c) [27]. All these results eventually suggest that cubane clusters could be rationally synthesized by combining two dinuclear cluster building units through a {2 + 2} route, a strategy designated as a unit construction method [38]. For a long time until recently, most structural models of the FeMoco have been limited to the single cubanes. Even though the above MoFe3 S4 single cubanes mimic quite well the molybdenum half of FeMoco, modeling the full FeMoco cluster has been proven to be challenging task not only because of the asymmetric arrangement of metals within the double cubane with Mo and Fe on the opposite ends but also due to the central carbide that has no precedent in metal sulfide clusters. In fact, attempt to couple two cuboidal halves of FeMoco, i.e. MoFe3 S4 and Fe4 S3 , has never been achieved. Notwithstanding these obstacles, efforts have been made during the last three decades to develop synthetic chemistry strategies for this cofactor. However, up to now, there have been only a few synthetic model clusters with a similar topology of the FeMoco. Two notable examples are the double cubane-like clusters [(Tp)2 Mo2 Fe6 S9 (SH)2 ]2 and [(DmpS)Fe4 S3 ]2 (μ-SDmp)2 (μ-STip)(μ6 -S) reported by Holm et al. in 2003 and Tatsumi and coworkers in 2007 (Tp = hydro(trispyrazolyl)borate(1–), Dmp = 2,6-di(mesityl)phenyl, Tip = 2,4,6-tri(isopropyl)phenyl) (Figure 4.5a,b). They excellently reproduced the topological structures of all-ferrous PN cluster and FeMoco but were not chemically perfect for mimicking the latter due to the presence of two Mo or Fe ends and the central atom being a μ6 -S rather than a carbon atom [44, 45]. Nearly 10 years later, Tatsumi and coworkers were successful in isolating a double cubane cluster [Fe8 S6 O] encapsulated an O atom inside the central cavity (Figure 4.5c), making a step forward in mimicking the light heteroatom in FeMoco [46]. Very recently, an elegant way to symmetrically introduce a N central atom into [WFe3 S3 N] cluster has been reported to prepare a [(Tp*)WFe3 S3 (NSiMe3 )] cluster SH

HS Fe

S S

TpMo

S

Fe

S

(a)

S

Dmp S

S

Fe S (b)

S

S

Fe S Fe

Fe Cl

Tp*W

MoTp 2003

S

Fe

S

Cl

Fe

S

Fe

S

S

Cl CPh3

S

O Fe

S Fe

Fe S

S Fe Dmp S

Fe

Dmp

S

Dmp Fe

SiMe3

Fe

(d)

Tip

Fe

N

2018

S

Dmp S

S

S

Fe

2007 S

S

2012 (c)

Fe Fe

Fe O

S Fe Dmp S

S

Fe

Dmp Fe

Fe S

S

S

Dmp

Figure 4.5 Selected FeMoco model cubanes in the period of 1978–2019. Source: Based on Refs. [39, 44–46] .

4.3 Rationally Designed Synthesis of Mo(W)–Cu(Ag)–S Clusters

(Tp* = tris(3,5-dimethyl-1-pyrazolyl)hydroborate(1–)) with a nitrogen atom closing the single cubane (Figure 4.5d) [39]. This cluster is highly similar topologically to the [MoFe3 S3 C] fragment of FeMoco as confirmed by the result of least-squares structural superposition. The corner N atom plays the role of the FeMoco carbide and also causes the distortion of the cubane as observed in FeMoco. Although O and N atoms have been successfully inserted into the central cavity of model clusters, embedding an interstitial μ6 -C atom into a cluster core seems to be a tough challenge. Future modeling work will have to focus more on the synthesis of model clusters that contain a central carbide ion.

4.3 Rationally Designed Synthesis of Mo(W)–Cu(Ag)–S Clusters Mo(W)–Cu(Ag)–S clusters whose chemistry can be traced back to the earlier studies on the Mo–Cu antagonism constitute a novel branch of the heterometallic sulfide clusters [47]. Besides their importance as a fundamental class of new complexes with unprecedented structures and potential applications in biological systems and optical materials, the effective synthetic approaches developed for the synthesis of these materials are also of significant value to the development of cluster chemistry and the creation of novel molecular functional materials.

4.3.1 Unit Construction Method for the Synthesis of Simple Mo(W)–Cu(Ag)–S Clusters Starting from Thiomolybdates or Thiotungstates Building Units The idea of unit construction, first introduced by Professor Lu Jiaxi in the synthetic chemistry of Mo–Fe–S clusters, was meanwhile successfully extended to include the Mo(W)–Cu(Ag)–S system by Professor Wu Xintao [48]. The most useful building units for the synthesis of Mo(W)–Cu(Ag)–S clusters are the thiomolybdates and their tungsten analogs MOn S4−n 2− (M = Mo, W; n = 0, 1). Tetrathiomolybdate MoS4 2− , for instance, contains four terminal sulfides arranged in a tetrahedral geometry, in which each sulfur atom is coordinated to the central Mo atom, leaving two free electron pairs capable of binding other metal ions. These species are a kind of soft bases and have a strong tendency to react with low-charge-density transition metals, such as Cu(I) and Ag(I), which are considered as soft acids. Typically, MoS4 2− can combine up to six Cu(I) cations, accompanied by changes in the coordination mode of each sulfide from terminal to μ2 , μ3 , and then to μ4 . As the coordination number (CN) of sulfur atom increases from 1 to 4, it contributes more and more electrons to the metal–sulfur bonding, thereby drastically reducing its reactivity as an electron donor. Once the CN is greater than 4, the sulfur atom would lose the complex reactivity, resulting from the exhaustion of free electron pairs (Figure 4.6) [40]. Mo(W)–Cu(Ag)–S clusters can also be prepared in a more controllable way by a stepwise unit construction method, again starting from MOn S4−n 2− anionic units (M = Mo, W). The final products obtained this way may include not

173

174

4 Synthetic Approaches, Structural Diversities, and Biological Aspects of Molybdenum (Tungsten)

S Mo S

Terminal-S(CN = 1)

S 6Cu(I)

4Cu(I)

2Cu(I) Cu

S Cu

2–

S

S

Cu

Cu

Mo

Cu

Mo

S

S

S

Cu

S S

Cu

Mo

S

S

Cu

S Cu

S

S Cu Cu µ4–S(CN = 4)

Cu µ3–S(CN = 3)

µ2–S(CN = 2)

Decreasing reactivity at S-site

Figure 4.6 Stepwise additions of Cu(I) ions to the tetrathiomolybdate anion. Source: Based on Wu et al. [40].

only diheterometallic but also triheterometallic complexes. The {1 + 1} combination of MS4 2− and CuCN, the simplest assembly for the Mo(W)–Cu(Ag)–S clusters, afforded heterometallic dinuclear rhombs [S2 MS2 Cu(CN)]2− in which MS4 2− acted as monochelating agents (Figure 4.7a) [41]. These complexes still have two remaining terminal sulfides and can apparently serve as heterometalloligands toward the third metal atom. Reactions of [S2 MS2 Cu(CN)]2− with [Ag(PPh3 )2 NO3 ] through a {2 + 1} route produced exclusively linear trimetallic complexes [Et4 N][(PPh3 )2 AgS2 MS2 Cu(CN)] (Figure 4.7b), where no statistical disorder, a major problem for heterometallic compounds, was found between the Cu and Ag atoms. The oxotrithiomolybdate MoOS3 2− behaved in the same manner as MS4 2− in these triheterometallic assemblies, generating a triheterometallic S CuCN

NC

{1 + 1}

Cu

S M S

S E=S

M S

{2 + 1}

NC

Cu

S

CuCN E = O

S

Cu

(b) O Mo S

S

M = Mo, W

{2 + 1}

{1 + 1}

Fe

S

Cu

Mo

Cl S

(e)

+

S

{2 + 1}

PPh3 Ag

S

PPh3

(d) S

Cl

Cu(PPh3)3

S

Mo

NC

S

Cl FeCl2

S

Ag(PPh3)2+

(c) E=S

PPh3

S

O NC

Ag

M

S {1 + 1}

PPh3

S

(a) E

S

S Ag(PPh3)2+

Fe

PPh3

S Cu

Mo S

Cl

PPh3

S

(f)

Figure 4.7 Unit construction methods for the synthesis of triheterometallic clusters. Source: Based on Refs. [41–43].

4.3 Rationally Designed Synthesis of Mo(W)–Cu(Ag)–S Clusters

butterfly cluster [Et4 N][(PPh3 )2 AgS2 MoOSCu(CN)] (Figure 4.7c,d) [42]. Similarly, the dinuclear precursor [Et4 N]2 [(Cl)2 FeS2 MoS2 ] (Figure 4.7e) was also reacted with Cu(PPh3 )3 Cl to form a triheterometallic complex [Et4 N][(PPh3 )2 CuS2 MoS2 Fe(Cl)2 ] (Figure 4.7f) [43].

4.3.2 Unit Construction Method for the Synthesis of Single Cubane- and Cage-like Mo(W)–Cu(Ag)–S Clusters Starting from Tri- and Dinuclear Thiomolybdates and Thiotungstates Cubane-like heterometallic sulfide clusters have attracted particular interest due to their possible relevance as models for cluster sites in proteins or on surfaces of industrial catalysts. The most common Mo(W)–Cu(Ag)–S heterocubanes have been conveniently synthesized by simply reacting monooxo complexes MOS3 2− instead of the MS4 2− with triphenylphosphine complexes of Cu(I) and Ag(I), giving rise to cubane-like clusters of the general formula [MOS3 M3 ′ (PPh3 )3 (X)] (M = Mo, W; M′ = Cu, Ag; X = halide ions). (Figure 4.8a). An alternative way to achieve Mo(W)–Cu–S cubanes is to start from a series of trinuclear Mo(W)–S clusters with unique cuboidal cores {M3 (μ2 -S)3 (μ3 -S)4 }4+ and loosely coordinated peripheral ligands (Figure 4.8b). These puckered quasi-aromatic cores with conjugated three-centered two-electron d(M)–p(S)–d(M) π-bonds exhibit some benzene-like structural features and chemical properties such as ligand substitution, addition, and oxidation reactions [49]. Addition of monomeric Cu(I) to the active μ2 -S atoms of these triangular complexes formed, by a facile {3 + 1} construction process, a series of isostructural cubane-like clusters, with general formulas [M3 CuS4 ][S2 P(OEt)2 ]3 ⋅I⋅(μ2 -OOCR)⋅L (R = CH3 , C6 H5 , CCl3 ; L = H2 O, DMF, DMSO, Py, CH3 CN) (Figure 4.8c) [14, 50]. The insertion of Cu(I) increases the electron count of the cluster cores from 48 to 60, in consistent with the presence (EtO)2P

Ph3P

S PPh3

M′ S

X

M′

S (EtO)2P

M′ S

PPh3

M S O

[MOS3M3′(PPh3)3(X)]

S

L M

S C

S M S

CuI P(OEt)2

S

S S

M'

P(OEt)2

O

S

S S

S

C

S S

R [M3CuS4][S2P(OEt)2]3.I.(µ2-OOCR).L

Ph3P

S

M

M

S O

M

S

M = Mo, W M′= Cu, Ag

S

[Et4N][M2M′S4(edt)2(PPh3)]

M S

S

S

S

(EtO)2P

(c) {3 + 1}

O

{2 + 1}

S

I S Cu

[M′(PPh3)n]+

[Et4N]2

L

S

R (b)

(a)

S M

S

S

S

S M O

(EtO)2P

S

S M

M S

S

(d)

Ph3P

S

PPh3

Cu

S

S Cu

+

2[Cu(PPh3)3] [Et4N]2[M2S4(edt)2]

{2 + 2[1]}

S

S

S

M

M S

S

S

M = Mo, W [M2Cu2S4(edt)2(PPh3)2]

Figure 4.8 Unit construction methods for cuboidal Mo(W)–Cu(Ag) clusters. Source: Based on Refs. [14, 50–53].

175

176

4 Synthetic Approaches, Structural Diversities, and Biological Aspects of Molybdenum (Tungsten)

of three M—M and three M—Cu bonds based on the extended Wade rule for the skeletal bond–electron pairs (Wade index I w = 6). Furthermore, the outer ligands do not affect the structural parameters of the metal sulfide cores, indicating the high stability and structural intactness of these heterometallic thiocubanes. Apart from the trinuclear {M3 (μ2 -S)3 (μ3 -S)4 }4+ cluster units, dinuclear thiomolybdates and thiotungstates can also serve as starting building blocks for this purpose by using their lone-pair electrons of the sulfur atoms. The perfect example of such dimeric building blocks for this chemistry is those containing thiolate ancillary ligands, e.g. the syn isomers of [Et4 N]2 [M2 S4 (edt)2 ] (M = Mo, W). These complexes each contain a thiometallic core with one terminal sulfur, one bridging sulfur, and one bidentate edt ligand at each M atom. Such an arrangement of active S atoms offers an ideal coordination pole, chemically and sterically suitable for forming cuboidal structures through {2 + 1} or {2 + 2[1]} building block strategies when combined with one or two heterometals (Figure 4.8d). Thus, when [Et4 N]2 [Mo2 S4 (edt)2 ] reacts with 1 equiv of [Cu(PPh3 )2 (dtp)], where dtp = O,O′ -diethyl thiophosphate, at room temperature under nitrogen atmosphere, the incomplete cubane-like cluster with one missing corner atom, [Et4 N][Mo2 CuS4 (edt)2 (PPh3 )], forms [51]. This reaction follows a {2 + 1} pathway associated with changing the coordination mode of two sulfides from terminal to μ2 and one sulfide from μ2 to μ3 . The resulting two μ2 -S atoms and one μ3 -S atom constitute three vertices of the tetrahedron centered at Cu(I), leaving the rest vertex occupied by an undissociated triphenylphosphine ligand. The metal sulfide core of this cluster is similar to its homometallic trimolybdenum analog [Mo3 (μ2 -S)3 (μ3 -S)]4+ , but with a much more distorted C3v symmetry due to the introduction of a heterometal, as reflected by the obvious difference in the distance between two Cu–μ2 –S bonds (2.313 versus 2.409 Å) and two Mo—Cu bonds (2.760 versus 2.802 Å). It seems that the configuration of the remaining three μ2 -S atoms in [Et4 N][Mo2 CuS4 (edt)2 (PPh3 )] allows further addition of the second Cu(I) in a way similar to the addition of Cu(I) atom to [Mo3 (μ2 -S)3 (μ3 -S)]4+ (Figure 4.8c), but to our surprise, such a {3 + 1} construction has been unsuccessful. Nevertheless, with the addition of 2 equiv of [Cu(PPh3 )3 Cl] to the dinuclear thiomolybdate, the desired cubane-like cluster [Mo2 Cu2 S4 (edt)2 (PPh3 )2 ] was obtained through a {2 + 2[1]} route [52]. In both clusters, the coordination sphere around the Mo atom consists of three bridging sulfides and two S atoms from a edt ligand, arranged in a TP geometry. It is worth noting that the configuration of the dimeric [Mo2 S2 (μ2 -S)2 ] unit in the starting dinuclear thiomolybdate is preserved after the addition of heterometals, allowing one to rationally design and predict the structure of the reaction product. Indeed, a similar reaction using [Ag(PPh3 )3 NO3 ] instead of [Cu(PPh3 )2 (dtp)] gave the corresponding incomplete cubane-like cluster [Et4 N][Mo2 AgS4 (edt)2 (PPh3 )] [53] whose structure was analogous to [Et4 N][Mo2 CuS4 (edt)2 (PPh3 )], except that the Mo–Ag distances (3.010 and 3.017 Å) are too long to indicate any significant bonding. Besides, these {2 + 1} and {2 + 2[1]} building block strategies also worked well with dinuclear thiotungstate [Et4 N]2 [W2 S4 (edt)2 ], which led to the formation of incomplete cubane-like clusters [Et4 N][W2 MS4 (edt)2 (PPh3 )] (M = Cu and Ag)

4.3 Rationally Designed Synthesis of Mo(W)–Cu(Ag)–S Clusters

and cubane-type cluster [W2 Cu2 S4 (edt)2 (PPh3 )2 ], all of them were isostructural with their molybdenum analogs. Cyclic voltammetry studies showed that the isomorphic clusters [Mo2 Cu2 S4 (edt)2 (PPh3 )2 ] and [W2 Cu2 S4 (edt)2 (PPh3 )] displayed irreversible reduction waves at −0.76 and −1.2 V, respectively, aligning with the fact that the tungsten complex is more electron-rich and hence more difficult to reduce. The structure of incomplete cubane-like cluster [Mo2 CuS4 (edt)2 (PPh3 )] deserves to be further discussed as an exemplary instance for Mo(W)–Cu(Ag)–S clusters. The distances of two newly formed Mo–μ2 –S bonds (2.169 and 2.161 Å) are close to the Mo=S terminal bonds of 2.10 Å in the parent dinuclear complex, as reflected by the small shift of Mo=S vibration in the IR spectrum from 515 to 495 cm−1 . Meanwhile, the bond length of Cu–μ3 –S (2.265 Å) is shorter than the normal Cu—S single bond, suggesting its partially double-bond character. These three quasi-double bonds are approximately parallel to each other, creating a so-called d–p–π stereo-conjugation system, a bit different from that in homotrimetallic [Mo3 S4 ]4+ core due to the presence of d10 Cu(I) ion. It is this large π conjugation system on the Mo–S–Cu bridge that renders Mo–Cu–S clusters particularly stable. A similar phenomenon is also observed in other cuboidal Mo(W)–Cu–S clusters. The oxothiomolybdate dimer [Et4 N]2 [Mo2 O2 S2 (edt)2 ], whose structure is similar to that of [Et4 N]2 [Mo2 S4 (edt)2 ], except that the terminal apical Mo = S is replaced by Mo = O, can also react with Cu(I) and Ag(I) ions. In addition to the μ2 -S, the sulfur atom of the bidentate edt ligand can also bind to heterometals, generating somewhat unexpected cage-like clusters. For example, when [Et4 N]2 [Mo2 O2 S2 (edt)2 ] reacted with 2 equiv of Cu(I) (generated in situ by reduction of Cu(II) salt with NaBH4 ) followed by the addition of (NH4 )2 Sx , a dodecanuclear cage-like cluster [Et4 N]2 [{Mo2 Cu2 O2 S2 (edt)2 }3 (μ6 -S2 )]⋅CH2 Cl2 was isolated (Figure 4.9a) [20]. A crystallographic study established that its unexpected structure contained O

2 O S

S

S 2M′+

S

M

M S

S

M O

O

S

M′

M = Mo, W

S S

O

S S

M

S

X3 M′

S

S

S

S

Mo

Mo S

S

2–

O

S

S

Trimerization M′ = Ag

M′ = Cu

Cu(I) S2–

S22 O S S

M M

O

S

M O

M S

Cu

S

M

M S

S

M O

O

O

S S [Et4N]2[{Mo2Cu2O2S2(edt)2}3(µ6-S2)].CH2Cl2

(a)

Ag

S

O

M S

S Ag

S

S

Ag

S

M

M S

S

O

O

O

Mo O

S

S

2– S S

S

Cu Cu

S S

S Cu S

S

S

S

Mo

[Et4N]2[{Mo2Ag2O2S2(edt)2}3(µ6-S)].CH3CN

(b)

Mo

Cu

S

S

S

S

Mo

S Cu

Ag

Ag

S

S

S Cu

S S

S

S

Cu S

SS

S

O

O S

M M

Ag

Cu S

Cu S

S S

S

S

2–

O

SS

Cu S

O

2–

O

Cu S

S

Mo

Mo S

O

[Et4N]2[{Mo2Cu2O2(S)2}3(µ6-S)]

(c)

Figure 4.9 Synthesis of the Mo(W)–Cu(Ag)–S cages. Source: Modified from Refs. [19, 20, 54].

O

177

178

4 Synthetic Approaches, Structural Diversities, and Biological Aspects of Molybdenum (Tungsten)

three identical mixed-metal Mo2 Cu2 cores, which were assembled together by a μ6 -S2 2− anion located in the center of the cavity thus formed. Although the detailed mechanism is unknown, the formation of this structure may reasonably be viewed as resulting from a {2 + 2[1]} assembly of dimeric molybdenum building blocks with two Cu(I) ions followed by a trimerization of the resultant [Mo2 Cu2 O2 S2 (edt)2 ] fragments. The key principle involved in this synthetic route is the introduction of (NH4 )2 Sx , which acts as a source for “naked” disulfide anion to bind six Cu(I) atoms. Monosulfide anions, meanwhile, can also function in the same way as disulfide for this type of cluster core aggregation. This was demonstrated by the reaction of AgNO3 with [Et4 N]2 [Mo2 O2 S2 (edt)2 ], which, after treatment with a small amount of aqueous (NH4 )2 S, gave a similar cage-like cluster [Et4 N]2 [{Mo2 Ag2 O2 S2 (edt)2 }3 (μ6 -S)]⋅CH3 CN (Figure 4.9b) [54]. Both Cu(I) and Ag(I) ions in these clusters are in a distorted tetrahedral geometry bonded to four sulfur atoms, one from a μ2 -S bridge, two from the edt ligands of different oxomolybdenum dimers, and one from the central S2 2− or S2− anion. Tungsten analogs of these cage-like clusters have also been synthesized in the similar manner except that the inaccessible oxothiotungstate dimer was generated through pre-oxidation of all-sulfur [Et4 N]2 [W2 S4 (edt)2 ] with H2 O2 . Alternatively, the above “naked” sulfide anions can also be generated in situ by sulfur deprivation from ancillary polysulfide ligands of the oxothiomolybdate dimer. Treatment of [Et4 N]2 [Mo2 O2 S2 (S2 )(S4 )] with two equimolar amounts of Cu(I) afforded dodecanuclear cage [Et4 N]2 [{Mo2 Cu2 O2 S2 (S2 )2 }3 (μ6 -S)] (Figure 4.9c) [19]. Its overall structure again features a central μ6 -S atom, but the Cu(I) ion is bonded to disulfides on Mo rather than edt ligands as above cages. The μ6 -S is not frequently observed for metal sulfide clusters. The empty d orbitals of μ6 sulfur in these cases are assumed to involve in the M—S bonding, adopting an sp3 d2 hybridization instead of conventional sp3 .

4.3.3 Unit Construction Method for the Synthesis of Mo(W)–Cu(Ag)–S Clusters Having Multiple Cubane-like Structures As we can see, the unit construction method shows great flexibility in the synthesis of heterometallic sulfide clusters. The starting building blocks can be either homometallic or heterometallic units provided that they still contain reactive sulfur atoms. One example of heterometallic building unit is the trinuclear butterfly species such as [MOS3 Cu2 (PPh3 )3 ] (M = Mo, W) (Figure 4.10a), which were found to combine with copper(I) thiolates in two different ways depending on the dentation of the thiolate ligands [55]. Despite their poor solubility, both CuSBut (monodentate thiolate) and CuS2 COEt (bidentate thiolate) can react smoothly with [MOS3 Cu2 (PPh3 )3 ], yielding double cubane-like clusters [MOS3 Cu3 (PPh3 )2 (SBut )]2 (Figure 4.10b) and single cubane-like clusters [MOS3 Cu3 (PPh3 )3 (S2 COEt)] (Figure 4.10c). The reactions probably involved cuboidal intermediates [MOS3 Cu(PPh3 )3 (SR)] (Figure 4.10d) formed by incorporation of one Cu(I) to the butterfly [MOS3 Cu2 (PPh3 )3 ] through a {3 + 1} route. In the case of monodentate ligand, instead of forming single cubane, coupling

4.3 Rationally Designed Synthesis of Mo(W)–Cu(Ag)–S Clusters PPh3 S

S

Cu

M

O

PPh3

CuSR

S Cu

S

O

M S

PPh3

SR

S Cu

{3 + 1}

PPh3 S

Cu

Cu

O

Cu S

M

PPh3

PPh3 [MOS3Cu2(PPh3)3] (a)

PPh3

Cu

S

PPh3

SR

Cu

PPh3

[MOS3Cu(PPh3)3(SR)] (d) R = SBut

R = S2COEt

2{3 +1} PPh3

Ph3 P S O

M

t Cu Bu

S Cu S

Cu Ph3P

PPh3

S But

Cu

S

S

Cu

S

S Cu Cu

M

S S

PPh3

[MOS3Cu3(PPh3)2(SBut)]2 (b)

O

O

M

S S

Cu Cu

S

C

OEt

PPh3

PPh3 [MOS3Cu3(PPh3)3(S2COEt)] (c)

Figure 4.10 Unit construction methods for Mo(W)–Cu–S cubane-type clusters. Source: Based on Du et al. [55].

of two intermediates took place in a “mouth-to-mouth” fashion through a pair of triply bridging thiolate ligands that spanned two cubane halves. This could be explained by the tension in the structure of cuboidal intermediate that may not permit the μ3 -SBut to occupy one corner of the cubane as the halide does in cubane clusters [MoOS3 Cu3 (PPh3 )3 X], where the halide only loosely mounted on the corner, acting mainly as a counter-anion (Figure 4.8a). For bidentate ligand, however, the two thiolate sulfurs tended to form intramolecular Cu—S bonds, generating a close cubane cluster after the migration of a PPh3 to the incoming Cu atom. In the structure of Mo–Cu–S double cubane-like cluster, the Cu atoms with PPh3 adopt a distorted tetrahedral geometry, while those without PPh3 are in an approximately planar triangular environment (the sum of three S–Cu–S angles is close to 360∘ ). The Mo—Cu(triangular) bonds (2.668 Å) are much shorter than Mo—Cu(tetrahedral) bonds (2.750 and 2.763 Å), in line with the stronger d(Mo)–p(S)–p(Cu) π conjugation on the Mo–S–Cu bridge for the trigonal Cu. The formation of these double cubane-like clusters can be regarded as a 2{3 + 1} construction, that is, a {3 + 1} process followed by dimerization. Similar double cubanes without PPh3 at copper can also be obtained by a stepwise unit construction approach, which gives a clearer view of how these clusters might be realized [56]. Reactions of MOS3 2− with 2 equiv of CuSBut in CH2 Cl2 formed, initially, double butterfly clusters [M2 O2 S6 Cu4 (SBut )2 ]2− (Figure 4.11a) isolated as tetraethylammonium salts (M = Mo, W). They can be viewed as consisting of two {MOS3 Cu2 } butterfly units linked via a pair of μ2 -SBut bridges. Further addition of two more Cu atoms to these butterfly dimers led to the production of double cubane-like clusters

179

180

4 Synthetic Approaches, Structural Diversities, and Biological Aspects of Molybdenum (Tungsten) 3CuSBut / [Cu(NCMe)4]PF6 / Et4NBr

But O

S

2

M S S M = Mo, W

S

2CuSBut Et4NBr

But

S

O M

S

Cu

Cu S

S M O

S

2{1 + 2[1]}

3CuSBut

Cd(NO3)2.4H2O

Cu

Cu

S

S

CuSBut [Cu(NCMe)4]PF6 {6 + 2[1]}

O M

S

3Cu(dtp) But4NI

S O But Cu M Cu S S S Cu But Cu Cu S S S Cu Cu I S S M M O Cu Cu O S S S

S

t

P

But [Bu4N]2[{M3O3S9Cu9(SBut)4][I] (c)

Cu

Cu

M O

S

S

t

[Et4N][M2O2S6Cu6(SBu )3] (b)

S

S

S

S

O S P Mo Cu S S SS Cu Cu P S S Cu S Mo Mo S Cu P S S Cu S Cu Mo Cu S S S Cu Cu P S S S S O P S

O

Cu

S

Cu Cu S

But

[Et4N]2[M2O2S6Cu4(SBu )2] (a)

S

S

But

S Cu

But

But4NI

But

S Cu

Cu

[Bu4N]2[{MoOS3}4Cu12(dtp)6] (d)

S O

S S Fe

Mo S

Fe

Fe

Fe

C S

S

Fe S Fe

S

Fe S

FeMoco

Figure 4.11 Unit construction methods for double and triple incomplete cubane-type clusters. For cluster (d), the ethoxyl groups on P are not shown. Source: Modified from Li et al. [57].

[Et4 N][M2 O2 S6 Cu6 (SBut)3 ] (Figure 4.11b) through a {6 + 2[1]} route. In these clusters, the two cuboidal {MOS3 Cu3 } units are connected by three μ2 -SBut , and all the Cu atoms have an approximately trigonal planar coordination. Again, due to the more effective π-conjugation, the Mo–Cu distances (av. 2.671 Å) are shorter than those found for tetrahedral Cu atoms in other Mo–Cu–S clusters. Unlike the double cubane-like clusters with PPh3 in which two Cu3 triangle faces are staggered with respect to each other, the openings of two {MOS3 Cu3 } halves here are aligned directly opposite to each other, leading to an approximate trigonal prismatic arrangement of the six central Cu atoms, very much similar to that found for the six belt iron atoms in FeMoco. Furthermore, the coordination geometry of Cu ions is also similar to that adopted by the belt iron atoms of FeMoco, where each iron atom is distorted away from the usual tetrahedral and toward nearly trigonal planar with three sulfide ions, and weakly bonded to the central carbide atom with mainly ionic interaction (see Figure 4.11, the structure of FeMoco). Alternatively, such double cubanes can be directly obtained by the reactions of MOS3 2− with stoichiometric amount of CuSBut and 1 equiv of [Cu(MeCN)4 ]PF6 that serves to eliminate excess thiolates. If Cd(NO3 )2 ⋅4H2 O was used in place of [Cu(MeCN)4 ]PF6 for the same purpose, incomplete cubane-like cluster trimers [Bu4 N]2 [{MOS3 }3 Cu9 (SBut )4 ][I] (Figure 4.11c) were isolated, in which an iodide ion was located in the center of the cavity defined by three thiolate-bridged cuboidal {MOS3 Cu3 } units. As described above, the introduction of monothiolate ligands into the Mo(W)–Cu–S system resulted in the formation of multiple incomplete cubanes through thiolate bridges. It is expected that dithiolate ligands could also function in the same way as monothiolate ligands. Indeed, the reaction of Cu(dtp) as a copper source with MoOS3 2− gave rise to a new cluster [Bu4 N]2 [{MoOS3 }4 Cu12 (dtp)6 ] (Figure 4.11d) [57]. The structure contains four cuboidal units {MoOS3 Cu3 }, in which two of them are fused together into a sphere-like core {O2 Mo2 S6 Cu6 } sitting

4.3 Rationally Designed Synthesis of Mo(W)–Cu(Ag)–S Clusters

between the other two. The middle core can be viewed as a ball-type polyhedron consisting of six {MoS2 Cu} and four {CuS2 Cu} rhombic faces. Each of the six μ4 -S atoms is coordinated to one Mo and three Cu atoms. Despite not having the iron composition, the two types of double incomplete cubane-like clusters [MOS3 Cu3 (PPh3 )2 (SBut )]2 and [Et4 N][M2 O2 S6 Cu6 (SBut )3 ] share some similarities with FeMoco catalytically active center from a structural point of view. They share a common structural characteristic, i.e. a closed capsule-type linked incomplete double cubane structure with FeMoco active center, and both have been regarded as structural models for the active center of FeMoco, in particular the latter cluster, which contains three doubly bridging μ2 -SBut groups around the central belt [61]. However, in the active center of FeMoco, the two cuboidal units are connected by inorganic sulfides rather than thiolate ligands. Therefore, it is interesting to make Mo(W)–Cu–S cuboidal subunits linked only with inorganic sulfide bridges. Although direct addition of sulfide anions to the reaction system containing metal salts may, in general, cause rapid precipitation of metal sulfide, the treatment of MoOS3 2− with CuBr and Bun 4 NBr produced, with an addition of a suitable amount of sulfide anion at the appropriate stage of the reaction, a tetradecanuclear Mo–Cu–S cluster [Bun 4 N]4 [Mo4 O3 S16 Cu10 ]⋅H2 O (Figure 4.12a) [58, 59]. This cluster is composed of a cuboidal {MoOS3 Cu3 } unit, a trigonal-prismatic {MoS4 Cu4 } unit, and two butterfly {MoOS3 Cu2 } units. The four cluster subunits are linked by two μ3 -S atoms each from one butterfly and a μ4 -sulfide anion at the center of the cavity. With WOS3 2− , the tungsten analog with the terminal ligand on one of the W atoms being half oxygen and half sulfur was isolated (Figure 4.12b). Further inspection of this reaction system revealed that the S2− ion played an important role in the formation of the tetradecanuclear cluster by substituting a bromide ligand at Cu(I), generating a solvent ligated Cu(I) complex. In fact, such a highly reactive Cu(I) complex could also be produced in situ by the reduction of Cu(NO3 )2 ⋅3H2 O with KBH4 in acetonitrile, which subsequently reacted with a mixture of MS4 2− and MO2 S2 2− to afford butterfly cluster tetramers MOS32– + CuBr + Li2S

MS42– + MO2S22

[M [Mo2O2S9]

2–

WOS32–

Cu(NO3)2.3H2O2 + TMEN

E M S

O

M

Cu

S

S Cu

Cu S

Cu S H

S

M

M S

S H

O M

Cu S

S

S

S

S

S Cu

Cu

Cu S

S

S

S

O

Cu Cu

Cu

S

S

S

O

S

S

S

S

S Cu

M O

S

M S

S Cu S

O

S

O M Cu S Cu

Cu

Cu

Cu M

S O Cu M S S S

M O Cu

Cu S

S Cu M S

S

O

[Bun4N]4[M4O2S16Cu10E].H2O

(a) M = Mo, E = O; (b) E = W, E = 1/2 O + 1/2S]

(c) [Et4N]4[M4O4S12Cu4]

(d) [M4O4S12Cu8(TMEN)4]

Figure 4.12 Synthesis of multiple cuboidal Mo(W)–Cu–S clusters. Source: Modified from Refs. [58–60].

181

182

4 Synthetic Approaches, Structural Diversities, and Biological Aspects of Molybdenum (Tungsten)

[Et4 N]4 [M4 O4 S12 Cu4 ] (M = Mo, W) (Figure 4.12c) [60]. The four {MOS3 Cu2 } butterfly subunits in these tetramers share wingtip Cu atoms with each other, forming an approximately square ring rarely observed for the Mo(W)–Cu–S clusters. As cuboidal clusters are readily formed by adding one more metal atom to their corresponding butterfly precursors, these butterfly tetramers should be able to further react with four Cu atoms to form cuboidal cluster tetramers. Unfortunately, they are sparingly soluble in common solvents and hence cannot be used as starting precursors. Nevertheless, the desirable cuboidal cluster tetramers [M4 O4 S12 Cu8 (TMEN)4 ] (TMEN = N,N,N ′ ,N ′ -tetramethylethylenediamine) (Figure 4.12d) could be directly synthesized through a similar reaction starting from [Me4 N]2 [Mo2 O2 S8 ] or WOS3 2− without the addition of KBH4 . In these cases, however, copper nitrate was apparently reduced by the starting thiometallates [60].

4.4 Rationally Designed Synthesis of W(Mo)–Ag–S Coordination Polymers The early Cu(I) and Ag(I) precursors used for the synthesis of W(Mo)–Cu(Ag)–S clusters usually contain semi-protected ligands, most often triphenylphosphine or halide anions. They are labile enough to be partially replaced by sulfur donors, leaving one or two still attached to the metal ions in the final products to prevent further reactions between the two active units. In the absence of such protecting agents, sequential assembly of active units may result in the formation of coordination polymers, especially when tetrathiomolybdate or tetrathiotungstate is used. For example, unlike MOS3 2− whose one coordinate direction is blocked by an oxo ligand, the reaction of MS4 2− with unprotected Cu(I) ions led almost invariably to unidentified amorphous substances and some of the resulting 1D and 2D coordination polymers have only been structurally characterized by analysis of powder X-ray diffraction data [62, 63]. Surprisingly, in the case of Ag+ , the situation was quite different. The reactions of MS4 2− with free Ag+ ions produced 1D coordination polymers with various structures depending on the types of cations present or the solvents used. It has been found that the cations not only function to neutralize the negative charges but also act as templates to induce the crystallization of polymeric W(Mo)–Ag–S anions. The cations can be solvated divalent and trivalent thiophobic metal ions or protonated small organic molecules containing N and O atoms. The latter may form hydrogen-bonded supramolecular networks that contribute strongly to the stabilization of polymeric structures. The proper choice of cations and solvents has led to the successful isolation of a range of W(Mo)–Ag–S heterometallic coordination polymers with diverse heterometallic cluster monomers and chain structures. The variety of monomers in these polymeric systems indicates that the valence state and the size of the cation strongly affect the configuration of the anionic W(Mo)–Ag–S chains. A summary of the synthetic routes is given in Scheme 4.1, and the structural motifs of the anionic chains are listed in Table 4.2. The basic W–Ag–S linear polymeric chain [WAgS4 ⋅NH4 ]n was prepared by the reaction of (NH4 )2 WS4 , AgNO3 , and D,L-valine (1 : 1 : 1) in a mixture of

4.4 Rationally Designed Synthesis of W(Mo)–Ag–S Coordination Polymers NH4+ DMF/H2O TMEN DMF/H2O NH3C(CH2OH)3+ DMF HNEt3+ DMF

Et3(PhCH2)N+ DMF/H2O Ca2+ MS42–+Ag+

DMF

n-

[WAgS4] n

(linear chain) 2n-

[W2 Ag2S8] n

(linear chain) EtOH

n-

[WAgS4] n

H2O

(linear chain)

Ca2+

n[WAgS4] n

DMSO

(linear chain)

n-

[MoAgS4] n

n-

[WAgS4] n

(double chain) 4n-

[W 4Ag4S16] n

(zigzag chain)

(Wave-like chain)

2n-

[Mo 2Ag2S8] n

(Zigzag chain)

[W2Ag2S8]n2n–

(Zigzag chain)

(M = Mo,W)

Zn(4,4′-bipy)2 DMF/DMSO La3+ DMAC La3+ DMSO Nd3+ DMSO Nd3+ DMSO

Scheme 4.1

[W3Ag3S12]n3n– (Helical chain) [W6Ag6S24]n6n– (8-likehelical chain) [W3Ag3S12]n3n– (Loose-helical chain) [Mo3Ag3S12]n3n– (Alternative helical chain)

Ln3+ DMF

[W4Ag5S16]n3n– (Pane chain: polymer with a single octanuclear square repeat unit)

Ln3+ DMF

[W8Ag10S32]n6n–

Ca2+ DEAC

[W4Ag6S16]n2n– (Hangling ladder structure)

(Double pane chain: polymer with a double octanuclear square repeat unit)

Synthetic routes to W(Mo)–Ag–S coordination polymers.

DMF and water. Similar reactions with N,N,N′ ,N′ -tetramethylethylenediamine (TMEN), tris(hydroxymethyl)aminomethane (Tris), and triethylamine (Et3 N), respectively, instead of D,L-valine also yielded linear chains with analogous anionic polymeric W–Ag–S structures (Figure 4.13a) (Table 4.1, entries 1–4) [64–67]. The linear chain is composed of a repeating building block –S2 WS2 Ag– in which the tetrahedral W and Ag atoms are alternately bridged by a pair of sulfide atoms with the neighboring WS2 Ag rhombuses posing perpendicularly to each other. Such an infinite linear chain was first identified in (PPh4 )Ag(MoS4 ) through resonance Raman spectroscopy [72, 73] and later confirmed by crystal structures

183

Table 4.2

Structure types of W(Mo)–Ag–S polymeric chains.

Compound number

Formula

Cation used

Repeat unit

Structure type

References

1

[WAgS4 ⋅NH4 ]n

NH4 +

[WAgS4 ]−

Linear chain

[64]

2

[W2 Ag2 S8 (tmenH2 ) (TMEN)⋅(H2 O)]n

tmenH2 2+

[WAgS4 ]2 2−

Linear chain

[65]

3

[WAgS4 ⋅HNEt3 ⋅DMF]n

HEt3 N+

[WAgS4 ]−

Linear chain

[66]

4

[WAgS4 ⋅Htris⋅2DMF]n

HTris+

[WAgS4 ]−

Linear chain

[67]

5

[WAgS4 ⋅Htris⋅H2 O]n

HTris+

[WAgS4 ]−

Double chain

[67]

6

[W4 Ag4 S16 ⋅2Ca(DMSO)6 ]n

[Ca(DMSO)6 ]4+

[W4 Ag4 S16 ]4−

Zigzag chain

[68]

7

[W4 Ag5 S16 ]n [M(DMF)8 ]n (M = Nd, La)

[M(DMF)6 ] +

[W4 Ag5 S16 ]3−

Pane chain

[64]

8

[W4 Ag6 S16 ]n [Ca(DEAC)6 ]n

[Ca(DEAC)6 ]2+

[W4 Ag4 S16 ]2−

Hanging ladder

[75]

9

[W8 Ag10 S32 ]n [{M(DEF)8 }2 ]n (M = La, Nd)

[M(DEF)8 ]3+

[W8 Ag10 S32 ]6−

Double pane chain

[69]

10

[W3 Ag3 S12 ]n [Nd(DMSO)8 ]n

[Nd(DMSO)8 ]3+

[W3 Ag3 S12 ]3−

Helical chain

[70]

11

[W3 Ag3 S12 ]n [La(DMAC)5 (H2 O)3 ⋅(DMAC)4 ]n

[La(DMAC)5 (H2 O)3 ]3+

[W3 Ag3 S12 ]3−

Helical chain

[69]

Zigzag chain

[75]

12

′

{[W2 Ag2 S8 ]⋅[Zn(4,4 -bipy)2 (DMF)2 (DMSO)2 ]}n

3

′

[Zn(4,4 -bipy)2 (DMF)2 (DMSO)2 ]2+

2−

[W2 Ag2 S8 ]

13

{[W2 Ag2 S8 ]⋅[Zn(4,4′ -bipy)2 (DMSO)4 ](DMSO)}n

[Zn(4,4′ -bipy)2 (DMSO)4 ]2+

[W2 Ag2 S8 ]2−

Linear chain

[75]

14

[N(C6 H5 CH2 )(C2 H5 )3 ]n [MoAgS4 ]n

[N(C6 H5 CH2 )(C2 H5 )3 ]+

[MoAgS4 ]−

Wave-like chain

[71]

15

[Mo2 Ag2 S8 ]n [Ca(DMSO)6 ]n

[Ca(DMSO)6 ]2+

[Mo2 Ag2 S8 ]2−

Zigzag chain

[71]

16

[Mo3 Ag3 S12 ]n [Nd(DMSO)8 ]n

[Nd(DMSO)8 ]3+

[Mo3 Ag3 S12 ]3−

Helical chain

[71]

186

4 Synthetic Approaches, Structural Diversities, and Biological Aspects of Molybdenum (Tungsten)

(a) Ca(NO)3.4H2O DMSO EtOH H2O

W

Ag

S

(c)

(b)

Figure 4.13 (a) The anionic polymeric structure of linear chains [WAgS4 ⋅NH4 ]n , [W2 Ag2 S8 (tmenH2 )(TMEN)⋅(H2 O)]n , [WS4 Ag⋅HNEt3 ⋅DMF]n , and [WS4 Ag⋅Htris⋅H2 O]n . (b) The structure of the double chain [WAgS4 ⋅Htris⋅H2 O]n . (c) The structure of zigzag chain [W4 Ag4 S16 ⋅2Ca(DMSO)6 ]n . Source: Modified from Refs. [67, 68].

of complexes [AgMoS4 ⋅α-PyH]n , [AgMoS4 ⋅γ-MePyH]n , and [AgWS4 ⋅γ-MePyH]n [74]. Interestingly, when the linear chain [WAgS4 ⋅HNEt3 ⋅DMF]n (Table 4.1, entries 3) was recrystallized from EtOH with a few drops of H2 O, this single chain was converted to a double chain [WAgS4 ⋅Htris⋅H2 O]n (Figure 4.13b) (Table 4.2, entry 5). The double chain can be viewed as two [SWSAg]n zigzag chains double-bridged by μ2 -S atoms at intervals. Such a chain breakage-reunion process was presumably driven by water, which formed 2D hydrogen-bonded networks with HTris cations between which parallel inorganic W–Ag–S chains were arrayed [67]. Upon addition of calcium nitrate to a DMSO solution of [WAgS4 ⋅HNEt3 ⋅DMF]n , the linear chain transferred to a zigzag chain [W4 Ag4 S16 ⋅2Ca(DMSO)6 ]n with concomitant replacement of HNEt3 cation by Ca(DMSO)6 2+ (Figure 4.13c) (Table 4.2, entry 6) [68]. The transformation again underwent chain fragmentation in solution and repolymerization induced by Ca(DMSO)6 2+ through a charge–charge or other nonbonding interactions. The basic construction unit of the zigzag chain is an L-shaped linear cluster [W4 Ag4 S16 ]4− , with the short edge being formed by –S2 WS2 AgS2 WS2 – and the long edge by –AgS2 WS2 AgS2 WS2 Ag–. The divalent Ca(DMSO)6 2+ cation presents here certain template effect in terms of spatial fit and charge balance required by the shape of W–Ag–S anionic chain. It would be expected that higher valent metal cations may induce the formation of entirely new W–Ag–S polymers. Trivalent cations Nd(III) and La(III) indeed could induce the formation of one-dimensional anionic polymer chains [W4 Ag5 S16 ]n [M(DMF)8 ]n (M = Nd and La) with a trianionic repeating unit –W4 Ag5 S16 – (Figure 4.14a) (Table 4.2, entry 7). The overall structure of these W–Ag–S chains can be regarded as square-like octanuclear clusters [W4 Ag4 S16 ]4− linked through Ag+ ions at the diagonal corners to form pane chains. In each square octanuclear cluster, the W atoms occupy the corners of the square, while the Ag atoms lie at the center of edges. The metal atoms are coordinated by three types of S atoms, namely, terminal sulfur atoms, μ2 -S atoms, and μ3 -S atoms. The Nd(III) or La(III) ions are ligated by eight DMF molecules and well separated from each other along both sides of each cluster anion [64].

4.4 Rationally Designed Synthesis of W(Mo)–Ag–S Coordination Polymers

(a)

(b)

(c)

W

Ag

S

Figure 4.14 Cation-induced structural diversity of W–Ag–S polymers. (a) The pane chain structure with a single octanuclear square repeat unit. (b) The hanging ladder-like structure. (c) The pane chain structure with a double octanuclear square repeat unit. Source: Modified from Refs. [69, 75].

It has been discovered that like cations, solvents also play the role of structure-directing agent, exerting a profound impact on the structure of W–Ag–S polymers. For example, using N,N′ -diethylacetamide (DEAC) instead of DMSO in the synthesis of zigzag chain afforded a hanging ladder-like polymer [W4 Ag6 S16 ]n [Ca(DEAC)6 ]n (Figure 4.14b) (Table 4.2, entry 8) [75]. While it also contains a basic W–Ag–S square cluster monomer resembling that found in the pane chain [W4 Ag5 S16 ]n [M(DMF)8 ]n , the monomers are linked by edge sharing rather than corner sharing. It can also be regarded as two identical W–Ag–S helical chains alternately cross-linked by Ag atoms. Similarly, when trivalent cations were employed but with N,N′ -diethylformamide (DEF) as a solvent, two such W–Ag–S square clusters could be joined together to form a double square repeating unit by sharing an Ag atom at the corner zone, giving rise to a zigzag chain (also called double pane chain) [W8 Ag10 S32 ]n [{M(DEF)8 }2 ]n (M = La and Nd) (Figure 4.14c) (Table 4.2, entry 9) [69]. As the solvent changed from DEF to DMSO and N,N′ -dimethylacetamide (DMAC), reactions with the same trivalent cations gave two helical chains [W3 S12 Ag3 ]n [Nd(DMSO)8 ]n (Figure 4.15a) (Table 4.2, entry 10) [70] and [W3 S12 Ag3 ]n [La(DMAC)5 (H2 O)3 ⋅(DMAC)4 ]n (Figure 4.15b) (Table 4.2, entry 11) [69]. Although these two helical chains have the same repeating unit –W3 Ag3 S12 –, the connectivity pattern of their helical chains is quite different. In the former polymer, the metal atoms are connected by μ2 -S and μ3 -S bridges, while in the latter, all the metal atoms are coordinated by μ2 -S atoms as in the linear W–Ag–S chains. Subsequently, the helical chain of the former is slightly more compact

187

188

4 Synthetic Approaches, Structural Diversities, and Biological Aspects of Molybdenum (Tungsten)

W

Ag

s

(a)

(b)

(c)

Figure 4.15 Cation-induced structural diversity of W–Ag–S polymers. (a) The helical chain structures of [W3 S12 Ag3 ]n [Nd(DMSO)8 ]n . (b) The helical structure of [W3 S12 Ag3 ]n [La(DMAC)5 (H2 O)3 ⋅(DMAC)4 ]n . (c) Simplified diagram of the anionic helical chain of [W3 S12 Ag3 ]n [La(DMAC)5 (H2 O)3 ⋅(DMAC)4 ]n . Source: Modified from Refs. [69, 70].

than that of the latter, with the screw pitch being 1.7 and 1.8 nm, respectively. Furthermore, unlike the well-separated trivalent cations [Nd(DMSO)8 ]3+ in the former helical chain, the cations in the latter also form a helical cationic chain through strong hydrogen bonds between oxygen atoms of coordination water molecules and solvent DMAC molecules (O· · ·O = 2.8 Å). Such structural variations in W–Ag–S chains may be related to the size and the connectivity mode of cationic complexes which, when having the same cationic metal ions, depends on what solvent molecules are bound to the cationic metal center. In some cases, anionic W–Ag–S chains and other cationic chains can coexist in the same crystal structure to produce a composite polymer. For example, a linear cationic chain [Zn(4,4′ -bipy)2 (DMF)2 (DMSO)2 ]n 2n+ induces a zigzag W–Ag–S anionic chain, whereas a zigzag cationic chain [Zn(4,4′ -bipy)2 (DMSO)4 ]n 2n+ elicits a linear W–Ag–S anionic chain, yielding respectively composite polymers {[W2 Ag2 S8 ]⋅[Zn(4,4′ -bipy)2 (DMF)2 (DMSO)2 ]}n and {[W2 Ag2 S8 ]⋅[Zn(4,4′ -bipy)2 (DMSO)4 ](DMSO)}n (Table 4.2, entries 12 and 13) [75]. In both polymers, the anionic and cationic polymeric chains simultaneously exist in the unit cell. The above results provide nice examples evidencing how the cations and solvents can modify the chain structure of W–Ag–S polymers. Despite the great success that has been accomplished on the designed synthesis of W–Ag–S polymers, very little has been achieved so far in isolating their corresponding molybdenum analogs. This might be because Ag+ has a higher affinity for MoS4 2− than for WS4 2− , and as a result, the reaction of MoS4 2− with Ag+ is much more difficult to control. One impressive assembly of such complexes involves the use of Ag2 S, a highly insoluble solid which, at first glance, might not be able to react with

4.4 Rationally Designed Synthesis of W(Mo)–Ag–S Coordination Polymers

(a)

W

Ag

s

(b)

(c)

Figure 4.16 (a) The structure of the wave-like Mo–Ag–S chain. (b) The structure of the zigzag Mo–Ag–S chain. (c) The structure of the helical structure of Mo–Ag–S chain showing only Mo and Ag atoms. Source: Modified from Ref. [71].

MoS4 2− in organic solvents. However, it turned out that Ag2 S could slowly release free Ag+ ions to react with MoS4 2− in DMF, affording three Mo–Ag–S polymers, namely, the wave-like chain [N(C6 H5 CH2 )(C2 H5 )3 ]n [MoS4 Ag]n , the zigzag chain [Mo2 S8 Ag2 ]n [Ca(DMSO)6 ]n , and the helical chain [Mo3 S12 Ag3 ]n [Nd(DMSO)8 ]n in the presence of N(C6 H5 CH2 )(C2 H5 )3 Cl, calcium nitrate, and niobium nitrate, respectively (Table 4.2, entries 14–16) [71]. In the structure of wave-like chain, the Mo and Ag atoms are arranged in an alternating sequence and are connected with each other in a wave-like way by all μ2 -S bridges (Figure 4.16a). Compared to the nearly linear anion chains of [AgMoS4 ⋅γ-MePyH]n and [AgWS4 ⋅γ-MePyH]n with similar monovalent organic cations in which the Ag–Mo–Ag and Mo–Ag–Mo angles range from 179.3(2)∘ to 180.0(1)∘ [74], the wave-like chain exhibits much shorter Ag–Mo–Ag and Mo–Ag–Mo angles (157.14(5)∘ and 156.47(5)∘ , respectively). In the zigzag chain, the basic building unit is a butterfly cluster {SMoAgS3 } which shares wigtip Ag atoms with two adjacent butterflies. The anionic chain can be regarded as a string of “butterflies” that adopt the alternating “up-down” orientation, and all the Mo and Ag atoms of one chain form an equilateral zigzag arrangement (Figure 4.16b). The overall topology of helical Mo–Ag–S chain is the same as that of the zigzag chain, but with a more compact and twisted configuration, resulting in a helical chain structure (Figure 4.16c). These results indicate that like W–Ag–S chains, the Mo–Ag–S chain is also highly flexible which can be stretched, compressed, or screwed to a certain degree, to best suit the charge, the size, or the shape of the cationic counterpart present. Metal chalcogenide coordination polymers have received great attention in recent years owing to their intriguing structures and potential applications in catalysis

189

190

4 Synthetic Approaches, Structural Diversities, and Biological Aspects of Molybdenum (Tungsten)

Table 4.3

Conductivities and energy gaps of some W–Ag–S polymers.

Compounds

Structural type

Conductivity Energy gap, (S/cm) E g (eV) References

{[W2 Ag2 S8 ]⋅ [Zn(4,4′ -bipy)2 (DMF)2 (DMSO)2 ]}n

Zigzag chain 5.43 × 10−7

3.87

[75]

[W3 Ag3 S12 ]n [La(DMAC)5 (H2 O)3 ⋅ (DMAC)4 ]n

Helical chain 1.16 × 10−4

3.96

[69]

{[W6 Ag6 S24 ]⋅ [La(DMSO)8 ]2 }n

Helical chain 6.96 × 10−7

3.81

L. Chen (Beijing Normal University, Beijing, P. R. China) and X. Wu (Fujian Institute of Research on the Structure of Matter, Fuzhou, P. R. China) (unpublished results)

{[W4 Ag5 S16 ]2 ⋅ [La(DEF)2 (DMF)6 ]⋅ [La(DEF)4 (DMF)4 ]}n

Pane chain

3.14 × 10−7

3.73

[76]

[W8 Ag10 S32 ]n [{La(DEF)8 }2 ]n

Double pane 2.58 × 10−4 chain

3.77

[69]

[W8 Ag10 S32 ]n [{Nd(DEF)8 }2 ]n

Double pane 2.58 × 10−4 chain

3.77

[69]

[W4 Ag6 S16 ]n [Ca(DEAC)6 ]n

Hanging ladder

2.92 × 10−5

3.70

[75]

and conductivity. Some of the abovementioned W–Ag–S polymers also exhibit considerable semiconducting properties. The electronic conductivity measured was performed using samples in the form of pressed pellets (two probes), and the electronic structures were calculated by the NNEW3 program package applying extended Hückel theory. As can be seen from Table 4.3, the conductivities of these W–Ag–S polymers are relatively low (1.16 × 10−4 to 3.14 × 10−7 S/cm) but can be still considered as semiconductors base on the Kittel’s proposal that the conductivity of semiconductors is between 10−9 and 10−2 S/cm. The small conductivities are consistent with their large energy gaps (3.70–3.96 eV). It should be noted, however, as the properties of W–Ag–S polymers are anisotropic in view of their 1D chain structures, the powder conductivity data could be used for references only.

4.5 Conclusion This chapter summarized selected achievements of previous studies on the chemistry of metal sulfide clusters at Fujian Institute of Research on the Structure of Matter. Initially inspired by the proposed models of FeMoco, several [MoFe3 S4 ] single cubane clusters with high core oxidation states have been synthesized using

References

bidentate thiolates. Owing to the chelated nature of bidentate thiolate ligands, these single cubanes could be readily isolated from one-pot self-assembly reactions without otherwise having to undergo the cleavage of double cubanes. Based on the structures and experimental results, a unit construction strategy was developed which allowed for the rational synthesis of Mo–Fe–S cubane clusters, together with a variety of Mo(W)–Cu(Ag)–S clusters. Flexible application of the unit construction method opens up tremendous possibilities for designing specifically targeted clusters through smart choice of appropriate building blocks and controllable multistep construction processes. The introduction of additional thiolate or sulfide bridging ligands to the reaction system, or by using metal ions themselves, the initially formed cluster subunits could be aggregated or condensed into higher nuclear clusters in a more predictable way, giving rise to the formation of Mo(W)–Cu(Ag)–S clusters with intriguing structures such as butterfly, cubane, multiple incomplete cubane, square, and cage clusters. Finally, the use of unprotected silver(I) ions with tetrathiotungstate and tetrathiomolybdate generated a range of one-dimensional coordination polymers whose structures were modulated by the types of cations and the solvents used. The underlying principles as well as the simplicity and efficiency of the unit construction approach may pave a way to new applications in other synthetic systems.

Acknowledgments The authors gratefully acknowledge financial support from the Natural National Science Foundation of China (21972060), and the State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences.

References 1 Dance, I. and Fisher, K. (1994). Metal chalcogenide cluster chemistry. In: Progress in Inorganic Chemistry, vol. 41 (ed. K.D. Karlin), 637. New York: Wiley. 2 Braunstein, P. and Rosé, J. (1999). Heterometallic clusters in catalysis. In: Metal Clusters in Chemistry, Chapter 2.2, vol. 2 (eds. P. Braunstein, L.A. Oro and P.R. Raithby), 616. Verlag GmbH, Weinheim, Germany: Wiley-VCH. 3 Hille, R., Schulzke, C., and Kirk, M.L. (eds.) (2017). Molybdenum and Tungsten Enzymes: Bioinorganic Chemistry, RSC Metallobiology Series, vol. 6. The Royal Society of Chemistry. 4 Dobbek, H., Gremer, L., Kiefersauer, R. et al. (2002). Proc. Natl. Acad. Sci. USA 99: 15971. 5 Muetterties, E.L., Rhodin, T.N., Band, E. et al. (1979). Chem. Rev. 79: 91. 6 Zhang, W.-H., Ren, Z.-G., and Lang, J.-P. (2016). Chem. Soc. Rev. 45: 4995. 7 Wu, X.-T. and Lu, J.-X. (1989). J. Struct. Chem. 8: 399.

191

192

4 Synthetic Approaches, Structural Diversities, and Biological Aspects of Molybdenum (Tungsten)

8 Müller, A., Jørgensen, C.K., and Diemann, E. (1972). Z. Anorg. Allg. Chem. 391: 38. 9 Chatt, J. and Mingos, D.M.P. (1970). J. Chem. Soc. A: 1243. 10 Huang, L.-R., Wang, B., and Wu, X.-T. (1984). Huaxue Tongbao 7: 15. 11 Wei, C.H. and Dahl, L.F. (1965). Inorg. Chem. 4: 1. 12 Lin, X.-T., Lin, Y.-H., Huang, J.-L., and Huang, J.-Q. (1986). Kexue Tongbao 31: 509. 13 Frisch, P.D. and Dahl, L.F. (1972). J. Am. Chem. Soc. 94: 5082. 14 Lu, S., Zhu, N., Wu, X. et al. (1989). J. Mol. Struct. 197: 15. 15 Saito, T., Yamamoto, N., Yamagata, T., and Imoto, H. (1988). J. Am. Chem. Soc. 110: 1646. 16 Markó, L., Bor, G., and Almásy, G. (1961). Chem. Ber. 94: 847. 17 Coleman, J.M., Wojcicki, A., Pollick, P.J., and Dahl, L.F. (1967). Inorg. Chem. 6: 1236. 18 Adams, R.D., Babin, J.E., Natarejan, K. et al. (1987). Inorg. Chem. 26: 3708. 19 Wu, X., Wang, B., Zheng, Y., and Lu, J. (1988). Chin. J. Struct. Chem. 7: 47. 20 Lin, P., Wu, X., Huang, Q. et al. (1998). Inorg. Chem. 37: 5672. 21 Wu, X., Huang, Q., Wang, Q. et al. (1996). Transition Metal Sulfur Chemistry: Biological and Industrial Significance, Chapter 17, ACS Symposium Series, vol. 653 (eds. E.I. Stiefel and K. Matsumoto). Washington, D.C.: American Chemistry Society. 22 Einsle, O., Tezcan, F.A., Andrade, S.L.A. et al. (2002). Science 297: 1696. 23 Spatzal, T., Aksoyoglu, M., Zhang, L. et al. (2011). Science 334: 940. 24 Lancaster, K.M., Roemelt, M., Ettenhuber, P. et al. (2011). Science 334: 974. 25 Lu, J.-X. (1992). Research on the chemical modelling of biological nitrogen fixation in the New China-an overview of research carried out at the Fujian Institute during the 1970s and the early 1980s. In: The Nitrogen Fixation and its Research in China, Section I: Chemistry of Nitrogen Fixation, Chapter 1 (ed. G.-F. Hong). Berlin and Heidelberg: Springer-Verlag. 26 Chan, M.K., Kim, J., and Rees, D.C. (1993). Science 260: 792. 27 Liu, Q., Huang, L., and Lu, J. (1990). Sci. China, Ser. B 20: 897. 28 Kim, J. and Rees, D.C. (1992). Science 257: 1677. 29 Kim, J. and Rees, D.C. (1992). Nature 360: 553. 30 Lu, J.-X. (1980). Composite “String bag” cluster model for the active center of nitrogenase. In: Nitrogen Fixation: Free-Living System and Chemical Models, vol. 1 (eds. W.E. Newton and W.H. Orme-Johnson), 343. Baltimore: University Park Press. 31 Bolin, J.T., Campobasso, N., Muchmore, S.W. et al. (1993). Structure and environment of metal clusters in the nitrogenase molybdenum-iron protein from Clostridium pasteurianum. In: Molybdenum Enzymes, Cofactors and Model Systems (eds. E.I. Stiefel, D. Coucouvanis and W.E. Newton), 186. Washington, D.C.: American Chemistry Society. 32 Wu, X.-T. and Lu, J.-X. (1996). Chin. Sci. Bull. 41: 825. 33 Dance, I. (2019). Dalton Trans. 48: 1251. 34 Wolff, T.E., Berg, J.M., Warrick, C. et al. (1978). J. Am. Chem. Soc. 100: 4630.

References

35 36 37 38 39 40 41 42 43 44 45 46 47 48

49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72

Chen, C., Wen, T., Li, W. et al. (1999). Inorg. Chem. 38: 2375. Liu, Q., Huang, L., Liu, H. et al. (1990). Inorg. Chem. 29: 4131. Liu, Q., Kang, B., Chen, C. et al. (1988). Sci. China, Ser. B Chem. 18: 920. Lu, J.-X. and Zhuang, B.-T. (1989). J. Struct. Chem. 8: 233. Xu, G., Wang, Z., Ling, R. et al. (2018). Proc. Natl. Acad. Sci. USA 115: 5089. Wu, X., Chen, P., Du, S. et al. (1994). J. Cluster Sci. 5: 265. Müller, A., Diemann, E., Josters, R., and Bögge, H. (1981). Angew. Chem. Int. Ed. Engl. 20: 934. Du, S., Zhu, N., Chen, P. et al. (1992). J. Chem. Soc., Dalton Trans.: 339. Shen, T., Du, S., and Wu, X. (1993). Polyhedron 12: 111. Zhang, Y. and Holm, R.H. (2003). J. Am. Chem. Soc. 125: 3910. Ohki, Y., Ikagawa, Y., and Tatsumi, K. (2007). J. Am. Chem. Soc. 129: 10457. Ohta, S., Ohki, Y., Hashimoto, T. et al. (2012). Inorg. Chem. 51: 11217. Müller, A., Bögge, H., Tölle, H.G. et al. (1980). Angew. Chem. Int. Ed. Engl. 19: 654. Chen, L. and Wu, X.-T. (2005). Molybdenum(tungsten)-copper(silver) thiolates: rationally designed syntheses from “reactive” building blocks. In: Inorganic Chemistry highlights in Focus II (eds. G. Mayer, D. Naumann and L. Wesemann), 207. Weinheim: Wiley-VCH. Lu, J. and Chen, Z. (1994). Int. Rev. Phys. Chem. 13: 85. Zhan, H., Zheng, Y., Wu, X., and Lu, J. (1989). Inorg. Chim. Acta 156: 277. Zhu, N., Zheng, Y., and Wu, X. (1990). Inorg. Chem. 29: 2705. Zhu, N., Zheng, Y., and Wu, X. (1990). J. Chem. Soc., Chem. Commun. 780. Zhu, N., Wu, X., and Lu, J. (1991). J. Chem. Soc., Chem. Commun. 235. Lin, P., Wu, X., Zhang, W. et al. (1997). Chem. Commun.: 1349. Du, S., Zhu, N., Chen, P., and Wu, X. (1992). Angew. Chem. Int. Ed. Engl. 31: 1085. Li, Z., Du, S., and Wu, X. (2004). Dalton Trans.: 2438. Li, Z., Du, S., and Wu, X. (2004). Inorg. Chem. 43: 4776. Guo, J., Wu, X., Zhang, W. et al. (1997). Angew. Chem. Int. Ed. Engl. 36: 2464. Guo, J., Sheng, T., Zhang, W. et al. (1998). Inorg. Chem. 37: 3689. Huang, Q., Wu, X., Wang, Q. et al. (1996). Inorg. Chem. 35: 893. Kure, B., Ogo, S., Inoki, D. et al. (2005). J. Am. Soc. Chem. 127: 14366. Binnie, W.P., Redman, M.J., and Mallio, W.J. (1970). Inorg. Chem. 9: 1449. Pruss, E.A., Snyder, B.S., and Stacy, A.M. (1993). Angew. Chem. Int. Ed. Engl. 32: 256–257. Huang, Q., Wu, X., Wang, Q. et al. (1996). Angew. Chem. Int. Ed. Engl. 35: 868. Huang, Q., Wu, X., Sheng, T., and Wang, Q. (1996). Acta Cryst. C52: 29. Huang, Q., Wu, X., Wang, Q., and Sheng, T. (1996). Acta Crystallogr C52: 795. Huang, Q., Wu, X., Sheng, T., and Wang, Q. (1995). Inorg. Chem. 34: 4931. Huang, Q., Wu, X., and Lu, J. (1996). Inorg. Chem. 35: 7445. Chen, L., Wu, X., Guo, X. et al. (1999). J. Chem. Soc., Dalton Trans. 4303. Huang, Q., Wu, X., and Lu, J. (1997). Chem. Commun.: 703. Yu, H., Zhang, W., Wu, X. et al. (1998). Angew. Chem. Int. Ed. Engl. 37: 2520. Müller, A., Jaegermann, W., and Hellmann, W. (1983). J. Mol. Struct. 100: 559.

193

194

4 Synthetic Approaches, Structural Diversities, and Biological Aspects of Molybdenum (Tungsten)

73 74 75 76

Müller, A. and Hellmann, W. (1985). Spectrochim. Acta 41A: 359. Lang, J., Li, J., Bao, S., and Xin, X. (1993). Polyhedron 12: 801. Chen, L., Yu, H., Wu, L. et al. (2000). J. Solid State Chem. 151: 286. Chen, L., Wu, X., and Lin, P. (1999). J. Chem. Crystallogr. 29: 629.

195

5 Group 11–15 Metal Chalcogenides Jian-Rong Li, Mei-Ling Feng, Bing Hu, and Xiao-Ying Huang Chinese Academy of Sciences, State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, Fuzhou, Fujian 350002, P.R. China

5.1 Introduction Metal chalcogenides are a class of compounds characteristic of the bonding between metal cations and chalcogen Q (Q = S, Se, and Te) anions. Metal chalcogenides have received enormous attention due to their excellent performances in ion exchange [1], semiconductor optoelectronics/thermoelectric [2, 3], nonlinear optics [4], photocatalysis [5], to name a few. Among them, the V/Mo/W/-S cluster-based compounds have been extensively studied since 1970s initially due to their relevance to biological systems [6], whereas two-dimensional (2D) transition metal chalcogenide compounds have continuously been a research hotspot owing to their importance as functional nanomaterials [7]. By contrast, group 11–15 metal ions can form a variety of compounds with chalcogen ions with varied functionalities [8]. For example, II–VI compounds of ZnS, ZnSe, ZnTe, CdS, CdSe, and CdTe are important compound semiconductors [9]; AgGaS2 and AgGaSe2 with a polar structure are promising materials for nonlinear optical applications operating in the infrared region of the electromagnetic spectrum [10], while SnSe crystal is a good thermoelectric material exhibiting ultralow thermal conductivity and high thermoelectric figure of merit [11]. Crystalline group 11–15 metal chalcogenides exhibit rich structural chemistry. To conveniently describe their structures, let us first introduce the concepts of building unit (BU), primary building unit (PBU), secondary building unit (SBU), and tertiary building unit (TBU). The concept of BU has been widely used to analyze the single crystal structures of compounds [8]. Metal chalcogenide coordination polyhedra as the minimum assemblies in the structures of metal chalcogenides are usually viewed as the PBUs, while the oligomerizations of such “brick”-like PBUs by sharing vertices and/or edges lead to SBUs. Occasionally, further interconnections of SBUs or mixed PBUs/SBUs lead to more complicated assemblies that can be termed TBUs. All the types of BUs mentioned above, varying from metal chalcogenide coordination polyhedra, clusters to chain-like or even layered motifs, can undergo condensation with other groups (identical or different) by a variety of modes to generate the Advanced Structural Chemistry: Tailoring Properties of Inorganic Materials and their Applications, First Edition. Edited by Rong Cao. © 2021 WILEY-VCH GmbH. Published 2021 by WILEY-VCH GmbH.

196

5 Group 11–15 Metal Chalcogenides

diverse structures of crystalline group 11–15 metal chalcogenides, briefly described as follows. Group 11(I) metal ions (e.g. Cu+ and Ag+ ) can adopt different coordination modes with chalcogen anions ranging from linear, trigonal, to tetrahedral geometries. The tetrahedral geometry is preferred when group 12(II) metal ions Zn2+ and Cd2+ are coordinated by chalcogen anions. As a result, binary group 12(II) metal chalcogenides (known as II–VI compounds) crystallize in either zinc blende or würtzite structural types. Occasionally, other geometries with higher coordination numbers of 5 or 6 are possible for Zn2+ and Cd2+ due to the steric requirements of organic ligands. Group 12(II) Hg2+ ion has flexible coordination geometries, ranging from the low-coordinate linear pattern to the high-coordinate octahedron, possibly due to its larger size and greater polarizability. The group 13(III) and 14(IV) metal chalcogenides are dominated by the heavier metal ions Ga3+ , In3+ , Ge4+ , and Sn4+ due to their greater affinity for Q (in most known cases, Q = S and Se) atoms than that of lighter elements (B3+ , Al3+ , C4+ , and Si4+ ). These four metal ions, especially Ga3+ and Ge4+ , are inclined to adopt tetrahedral coordination to form either discrete ortho-anion or oligomeric or polymeric entities. On the other hand, the larger size of Q atoms compared to that of oxygen can easily extend its coordination number up to 4. This makes it possible to connect such {MQ4 } (M = Zn, Cd, Ga, In, Ge, and Sn; Q = S, Se, and Te) tetrahedra in a similar arrangement of a cubic ZnS lattice, resulting in supertetrahedral clusters (Tn, Pn, and Cn; n is the number of metal sites on each edge of the cluster) with diverse sizes and compositions. Significant development on the further assembly of such supertetrahedral clusters as SBUs to fabricate microporous chalcogenides has been made in the last two decades [12]. In addition, higher coordination numbers of 5 and 6 have also become possible for In3+ and Sn4+ due to their larger ionic radius. For example, {SnQ5 } (Q = S, Se) trigonal bipyramid and {SnQ6 } octahedron have proven to be critical BUs for the construction of 2D and three-dimensional (3D) Sn–Q anionic structures. Overall, the group 11(I) (Cu+ and Ag+ ), 12(II) (Zn2+ , Cd2+ , and Hg2+ ), 13(III) (Ga3+ and In3+ ), and 14(IV) (Ge4+ and Sn4+ ) ions exhibit a wide range of polyhedral geometries (see Scheme 5.1 for some examples) when forming metal chalcogenides, depending on the size and polarizability of metal center. By contrast, the coordination of group 15(III) metal ion Sb3+ by Q (Q = S and Se in most known cases; and Q = Te in relatively few cases) atoms tends to exhibit unique pseudo-tetrahedral 𝜓-{SbQ3 } (trigonal pyramid) and pseudo-trigonal bipyramidal 𝜓-{SbQ4 } (see-saw shape) (Scheme 5.1) due to the stereo-activity of the lone electron pair (LEP). Similar to the formation of Sn–Q SBUs as exemplified in Scheme 5.2a, self-condensation of these pseudo coordination polyhedra by corner- and/or edge-sharing leads to further complex oligomeric groups, such as {Sb2 Q4 }, {Sb2 Q5 }, and {Sb3 Q6 } (Scheme 5.2b), which contributes to the great structural diversity of chalcogenidoantimonates(III) that has been well demonstrated before [13, 14]. On the other hand, the development of metal chalcogenides structural chemistry largely relies on the advance of synthetic methods or routes for metal chalcogenides. Early high-temperature solid-phase routes and medium-temperature reactive molten salt method are mainly used to synthesize inorganic metal

5.1 Introduction

Group 11 (I) metal

Group 12 (II) metal

Linear

Hg

Group 13 (III) metal

Trigonal Tetrahedron M = Cu, Ag

Hg

Linear

M

M

Ag

M

Zn

Trigonal plane Tetrahedron Trigonal bipyramid M = Zn, Cd, Hg

In

M

In

Tetrahedron Trigonal bipyramid Octahedron M = Ga, In

Group 14 (IV) metal

M

Sn

Sn

Tetrahedron Trigonal bipyramid Octahedron M = Ge, Sn

Antimony (III)

Sb Trigonal pyramid

Sb See-saw

Scheme 5.1 Common coordination geometries of the group 11(I), 12(II), 13(III), and 14(IV) metal ions and antimony(III) in the solid-state structures of metal chalcogenides.

chalcogenides in which dense solid-state structures are dominant [15, 16]. Since the pioneer works by Sheldrick and Wachhold [17], Kanatazidis and coworker [18], Li et al. [19], and Parise [20] in 1990s, softer chemical synthetic methods such as hydrothermal and solvothermal methods with relatively mild conditions have resulted in a lot of novel chalcogenide compounds with structures ranging from 0D to 3D, in which polarizing solvents such as water, alcohol, and organic amines are applied at relatively low temperatures (T ≈ 100–200 ∘ C). In such synthesis, the organic compounds, especially organic amines, play important roles acting as solvents, structure directing agents (SDAs), and even ligands. Room−/low-temperature solution methods are often powerful for the synthesis of chalcogenide cluster compounds [6, 21]. Classification of amines used in the synthesis of metal chalcogenides in this chapter is presented in Scheme 5.3. They may enter the final chalcogenide structures in two ways in terms of the interactions

197

198

5 Group 11–15 Metal Chalcogenides

{Sn2Q6}

{Sn2Q7} {Sn3Q10}

(a)

cis–{Sb2Q4}

trans–{Sb2Q5} {Sb3Q6}

(b)

Scheme 5.2 Representative Sn(IV)–Q and Sb(III)–Q SBUs (Q = S, Se, and Te). Source: (b) Sheldrick and Wachhold [13]; Seidlhofer et al. [14].

between the organic and inorganic components. One is to enter the structure through either ionic bonding and/or relatively weak H-bonding and van der Waals interactions serving as SDAs, templates, space-filling agents, and charge-balancing agents (CBAs), and thus forming organically templated chalcogenidometalates [19]. In this case, sometimes the organic cations in chalcogenidometalates can be removable and thus replaced by other cations, showing the ion exchange property [1, 8]. The other is to form an organic–inorganic hybrid architectures via coordinating to metal ions of the inorganic moiety as neutral ligands, which can change or modify the structure and physical properties of metal chalcogenides such as optical absorption edges [22]. Nevertheless, there was still an urgent need for developing new synthetic methods and new templates that may offer unique capabilities for producing unprecedented compounds with new properties. Indeed, since 2009, new synthetic methods or strategies such as ionothermal synthesis [23], surfactant thermal synthesis [24], and deep eutectic solvothermal synthesis [25] have been applied one after another. This chapter is not intended to describe the structures of all metal chalcogenides with great diversity. Instead, it will be focused on summarizing the structures of organic hybrid II–VI metal chalcogenides, discrete Tn clusters, chalcogenidostannates and chalcogenidoantimonates, which were mostly reported by the authors.

5.2 Inorganic–Organic Hybrid Chalcogenides Based on II–VI Semiconductors II–VI compound semiconductors are of great importance in the fields such us optoelectronics and spintronics [26]. In 1998, Rao’s group reported a series of II–VI-based mesostructures using long-chain amines as amphiphiles [27]. However,

5.2 Inorganic–Organic Hybrid Chalcogenides Based on II–VI Semiconductors H N

H2 N

H N

H2N

n-butylamine (ba)

n-Proptlamine (pa)

Dimethylamine [(Me)2NH]

Methylamine (ma)

H 2N

H2N

H2N

Benzylamine (bza)

Dipropylamine [(CH3CH2CH2)2NH]

H 2N

H2N

1,2-Ethanediamine (en)

Tetrahydrofurfurylamine (thafa)

NH2

H2N

NH2

1,3-Propanediamine (pda)

Hydrazine (hdz)

NH2

H2N

NH2

O

N H2N

1,2-Diaminopropane (1,2-pda)

N-Dimethylethylenediamine (N,N-dmen) H2N

NH2 H2N

H 2N

1,4-Butanediamine (bda)

NH2

H 2N

1,5-Pentanediamine (ptda)

H N

H 2N

NH2

N H

Triethylenediamine (Dabco)

p-Xylylenediamine (PDXA)

NH2

N H

H2N

Triethylenetetramine (teta)

Diethylenetriamine (deta)

H 2N

NH

N

H N NH2

trans-1,2Diaminocyclohexane (chxn)

N

NH2

m-Xylylenediamine (mxda)

H2N

1,6-Hexanediamine (hda)

H2N H 2N

NH2

O

Propyleneurea (pu)

H2N N

NH2

N-(2-Aminoethyl)piperazine (aep)

N NH2

H 2N

tris(2-Aminoethyl)amine (tren)

H 2N

N

H N

N H 2N

N-(3-Aminopropyl)-imidazole (api)

H N N H

NH2

Tetraethylenepentamine (tepa)

Scheme 5.3 Classification of typical amines used in the synthesis of metal chalcogenides in this chapter.

neither their single crystal structures nor the optical properties were studied. Since last two decades, an entirely new class of II–VI-based inorganic–organic hybrid crystalline nanostructures with different combinations of II–VI and primary amines have been developed, which can be denoted as [(MQ)n (L)x ] (MQ = ZnS, ZnSe, ZnTe, CdS, CdSe, CdTe, and MnSe; L = mono- or di-amines or hydrazine [hdz]; n = 1 and 2; x = 0.5, 1, and 2) [28–46]. These nanostructures are composed of a fragment (e.g. mono- or double-atomic slabs or chains) of II–VI semiconductor and coordinatively bonded primary amine molecules. The di- and mono-amine molecules with low

199

200

5 Group 11–15 Metal Chalcogenides

melting points are selected as the organic component including alkyl-diamines with number of carbon atoms of 2–8, alkyl-monoamines with number of carbon atoms ranging from 1 to 16, and other amines such as cyclohexylamine (cha) and m-xylylenediamine (mxda); later on, hydrazine has also been introduced into the structure. The –NH2 group in these primary amines has strong ability to bind the transition metals such as Mn, Zn, and Cd. As a result, hybrid nanostructures based on most of the II–VI binaries have been obtained. Table 5.1 lists the crystallographic parameters of selected II–VI hybrids. Typical crystal structures of II–VI-based inorganic–organic hybrid compounds are plotted in Figure 5.1. As illustrated in Figure 5.1, all the structures feature neutral inorganic II–VI fragments (slabs or chains) sandwiched or terminated by coordinatively bonded amine spacers. However, significant overall crystal structure changes were realized by varying the topology and dimensionality of inorganic structural motifs of MQ accompanied by the incorporation of different types of amines. The structural dimensionality of the hybrids varies from one-dimensional (1D) chain, 2D slab, to 3D network as a result of the different connection modes of the amine molecules and dimensionality of inorganic structural motifs. Figure 5.1a shows the structure of 3D-α-[ZnTe(en)0.5 ] (en, 1,2-ethanediamine), a representative of the most common α-type hybrid structures [28, 33, 43, 46] that features inorganic single layers of α-[MQ]n interconnected by diamine or hydrazine molecules. In the α-[MQ]n , alternating three connected M 2+ and Q2− ions form a puckered 63 (honeycomb) net, which can be regarded as a distorted “slice” taken from the (110) cleavage of the würtzite-type II–VI structure. Figure 5.1b shows the structure of 3D-β-[ZnTe(en)0.5 ] [28], in which single layers of β-[ZnTe]n resemble the (110) crystal cleavage of the zinc-blende-type ZnTe structure and are interconnected by ethanediamine molecules. Figure 5.1c illustrates the crystal structure of 2D-α′ -[ZnTe(hdz)], in which the α′ -[ZnTe]n single layers also resemble the (110) crystal face of the würtzite-type ZnTe structure, however, in a less distorted form compared to the α-[ZnTe]n . Thus, according to the configuration, the honeycomb nets of single layer [MQ]n can be classified into three types, that is, α-, α′ -, and β-[MQ]n . The α-type is highly distorted (110) cleavage of the würtzite-type structure and found in most of the 3D-[(MQ)(diamine)0.5 ] and 2D-[(MQ)(monoamine)] hybrid structures, while the α′ -type is also similar to the (110) crystal face of the würtzite structure but is less distorted one compared to the α-type, which is only present in 2D-[(ZnTe)(hdz)] [33]; 3D-β-[ZnTe(en)0.5 ] [28] is the only structure where β-[ZnTe]n presents. Note that the relative orientation and connectivity between the inorganic slabs and organic pillars in 3D-α and β-[ZnTe(en)0.5 ] are different as a result of the difference in the topologies of the 63 nets, resulting in the different space groups they adopt (Pbca and Pbcm). Figure 5.1d shows the typical structure of 2D-[(M 2 Q2 )(monoamine)], where the double-layer inorganic slabs of [M 2 Q2 ]n resemble “two slices” taken from the (110) face of the würtzite structure. Figure 5.1e shows the structure of 1D-[ZnTe(pda)] (pda = 1,3-propanediamine) [33], in which Zn and Te ions are two connected to each other forming a chain-like structure and the two additional coordination bonds of Zn are occupied by a chelated pda

Table 5.1

Crystallographic data for selected II–VI-based hybrid compounds.

Formula

D

a (Å)

b (Å)

c (Å)

𝜶 (∘ )

𝜷 (∘ )

𝜸 (∘ )

V (Å3 )

Space group

References

Note

α-[ZnTe(en)0.5 ]

3

7.061

6.927

17.524

90

90

90

857.1

Pbca

[28]

SXRD

β-[ZnTe(en)0.5 ]

3

5.660

17.156

4.336

90

90

90

421.0

Pnnm

[28]

SXRD SXRD

α-[ZnTe(pda)0.5 ]

3

20.169

7.038

6.882

90

90

90

976.9

Cmc21

[28]

α-[ZnTe(bda)0.5 ]

3

7.0723

6.9258

22.3497

90

90

90

1094.72

Pbca

[43]

SXRD

γ-[ZnTe(bda)0.5 ]

3

7.0996

6.9439

20.4912

90

90

90

1010.19

Pbca

[43]

SXRD SXRD

α-[ZnTe(ptda)0.5 ]

3

25.2760

7.0590

6.9005

90

90

90

1231.21

Cmc21

[43]

α-[ZnTe(hda)0.5 ]

3

7.0714

6.9201

27.051

90

90

90

1323.7

Pbca

[43]

SXRD

δ-[ZnTe(hda)0.5 ]

3

7.1091

34.612

25.091

90

90

90

6174

P21 21 21

[43]

SXRD

γ-[ZnTe(hda)0.5 ]

3

7.1353

6.9616

23.7318

90

90

90

1178.83

Pbca

[43]

SXRD

α-[ZnTe(hdz)0.5 ]

3

6.9167

6.8033

13.2853

90

90

90

625.16

Pbca

[46]

SXRD

α-[ZnTe(ma)]

2

7.179

6.946

18.913

90

90

90

943.1

Pbca

[33]

SXRD

α′ -[ZnTe(hdz)]

2

4.2222

6.9057

7.3031

90

98.92

90

210.36

P21

[33]

SXRD

α-[ZnTe(hdz)2 ]

1

7.2157

11.5439

7.3909

90

101.30

90

603.72

P21

[38]

SXRD

β-[ZnTe(hdz)2 ]

1

8.1301

6.9580

10.7380

90

91.703

90

607.17

Pn

[38]

SXRD

[ZnTe(pda)]

1

9.997

6.997

10.332

90

90

90

722.7

Pbcm

[33]

SXRD

α-[ZnSe(en)0.5 ]

3

6.6326

6.4630

17.3540

90

90

90

743.9

Pbca

[29]

PXRD

α-[ZnSe(pda)0.5 ]

3

19.9731

6.6268

6.4394

90

90

90

852.3

Cmc21

[29]

PXRD

α-[ZnSe(bda)0.5 ]

3

6.646

6.473

22.31

90

90

90

961.2

Pbca

[31]

PXRD

α-[ZnSe](deta)0.5

3

24.7200

6.6239

6.4426

90

90

90

1055.36

Cmc21

[39]

PXRD

α-[ZnSe(mxda)0.5 ]

3

6.7346

24.9804

6.4428

90

90

90

1083.9

Ccm21

[41]

PXRD

α-[ZnSe(hda)0.5 ]

3

6.6252

6.4505

27.138

90

90

90

1159.8

Pbca

[31]

PXRD

[(Zn2 Se2 )(ba)]

2

6.8035

6.5194

41.894

90

90

90

1858.2

Pbca

[42]

PXRD

α-[(ZnSe)(ba)]

2

6.6746

6.4642

34.217

90

90

90

1476.3

Pbca

[42]

PXRD

α-[ZnS(en)0.5 ]

3

17.263

6.393

6.205

90

90

90

684.78

Pbca

[35]

PXRD (continued)

Table 5.1

(Continued)

c (Å)

V (Å3 )

Space group

References

Note

90

1018.3

Ccm21

[45]

PXRD

90

1019.3

Pcab

[45]

PXRD

90

1597.29

Pbca

[44]

PXRD

90

90

1957.85

Pbca

[44]

PXRD

90

90

90

1610.81

Pbca

[44]

PXRD

90

90

90

1591.54

Pbca

[44]

PXRD

𝜶 (∘ )

𝜷 (∘ )

Formula

D

a (Å)

b (Å)

[ZnS(mxda)0.5 ]

2

6.6087

24.77

6.2193

90

90

[ZnS(PXDA)0.5 ]

2

6.6014

24.840

6.2162

90

90

[Zn2 S2 (bza)]

2

6.0948

6.4108

40.880

90

90

[Zn2 S2 (mbza)]

2

6.1409

6.218

51.274

90

[Zn2 S2 (fbza)]

2

6.0944

6.2940

41.994

[Zn2 S2 (pca)]

2

6.297

6.2598

40.376

𝜸 (∘ )

[Zn2 S2 (thfa)]

2

5.6095

6.1814

39.678

90

90

90

1375.82

Pbca

[44]

PXRD

α-[CdTe(en)0.5 ]

3

7.484

7.204

16.821

90

90

90

906.90

Pbca

[32]

PXRD PXRD

α-[CdSe(en)0.5 ]

3

7.0949

6.795

16.7212

90

90

90

806.17

Pbca

[33]

α-[CdSe(en)0.5 ]

3

7.0848

6.7856

16.6940

90

90

90

802.56

Pbca

[32]

SXRD

α-[CdSe(pda)0.5 ]

3

20.6660

6.8900

6.7513

90

90

90

961.31

Cmc21

[33]

PXRD

α-[CdSe(hda)0.5 ]

3

6.8852

6.7894

27.4113

90

90

90

1281.38

Pbca

[37]

PXRD

α-[CdS(en)0.5 ]

3

6.841

6.548

16.659

90

90

90

746.32

Pbca

[32]

PXRD

α-[MnSe(en)0.5 ]

3

6.711

6.614

17.720

90

90

90

786.5

Pbca

[29]

SXRD

α-[MnSe(pda)0.5 ]

3

20.384

6.719

6.565

90

90

90

899.1

Cmc21

[29]

SXRD

D, dimensionality; PXRD, powder X-ray diffraction; SXRD, single crystal X-ray diffraction; en, 1,2-ethanediamine; pda, 1,3-propanediamine; bda, 1,4-butanediamine; ptda, 1,5-pentanediamine; hda, 1,6-hexanediamine; deta, diethylenetriamine; hdz, hydrazine; ma, methylamine; ba, n-butylamine; mxda, m-xylylenediamine; PXDA, p-xylylenediamine; bza, benzylamine; mbza, 4-methoxybenzylamine; fbza, 4-flurobenzylamine; pca, 3-picolylamine; thfa, tetrahydrofurfurylamine.

5.2 Inorganic–Organic Hybrid Chalcogenides Based on II–VI Semiconductors

c

(a)

Te N C Zn a 0

b

(b) b Zn Te N C

(c)

c

N Te Zn

b

0

(d) Se N C

Zn

0

(e)

b

C Te

N Zn

Figure 5.1 Typical crystal structures of II–VI-based inorganic–organic hybrid compounds and corresponding organic and inorganic components. Source: (b) Based on Huang et al. [28]; (e) Based on Huang et al. [33].

203

204

5 Group 11–15 Metal Chalcogenides

molecule. It is worth noting that in all hybrid structures containing monolayer [MQ]n slabs, the metal ions adopt distorted tetrahedral configuration bonded to three Q and one N atoms, while the chalcogen anions are three connected to metal ions. However, there are four-coordinated metal and chalcogen atoms in the double-layer hybrid compounds. In all cases, the M—Q bonds change little compared to that of the corresponding parent II–VI semiconductors. Another notable feature is that, for the 3D-[(MQ)(diamine)0.5 ], the structures containing diamines with an even number of carbons n (n = 0, 2, 4, and 6) crystallize in centrosymmetric space group Pbca (No. 61), while those with an odd number of diamines carbons n (n = 3 and 5) crystallize in the noncentrosymmetric space group Cmc21 (No. 36). In fact, most of the hybrid compounds crystallize in the centrosymmetric space group, with exceptions of 3D-α-[ZnQ(mxda)0.5 ] [41, 45], 3D-α-[ZnQ(deta)0.5 ] (Q = S and Se; deta = diethylenetriamine) [39], 3D-[MQ(diamine)0.5 ] structures with odd number of carbon atoms between the two NH2 groups (space group Cmc21 ), 2D-α′ -ZnTe(hdz) (space group P21 ) [33], 1D-α-[ZnTe(hdz)2 ] (space group P21 ) [38], and 1D-β-[ZnTe(hdz)2 ] (space group Pn) [38] and δ-[ZnTe(hda)0.5 ] (space group P21 21 21 ) [43]. Figure 5.2 shows the crystal structures of a series of 3D α-ZnTe(diamine)0.5 compounds in which the inorganic monoatomic α-ZnTe slabs are interconnected via diamine molecule pillars from en to 1,6-hexanediamine (hda) in different lengths. The diamines adopt an all-trans conformation in these structures. Obviously, the distance between inorganic layers increases with the increasing of the length of diamine. It was found that not only the coordination mode of the amines affects the final structures but also their conformations may lead to different crystal structures. Moon et al. have theoretically investigated the structural stability and total energy differences of en-based hybrids. It is suggested that there must be two polymorphs of 3D hybrids due to the conformation difference of the en molecules [40]. The calculations showed that αI phase is the most stable isomorph, while the β and αII phases are less stable. For instance, the conformations of en molecules in 3D-αI-[ZnTe(en)0.5 ] are TTT (or a GTG′ very close to TTT) (T = Trans, G = Gauche

Figure 5.2 Crystal structural diagrams of structures 3D α-ZnTe(diamine)0.5 (from left to right, diamine = en, pda, bda, ptda, and hda).

5.2 Inorganic–Organic Hybrid Chalcogenides Based on II–VI Semiconductors

for various conformations of the free en molecule), while the conformations of en molecules in 3D-αII-[CdSe(en)0.5 ] are GTG′ (T = Trans, G = Gauche for various conformations of the free en molecule), leading to a much shorter c-axis and smaller inter-layer distance than that in 3D-αI-[ZnTe(en)0.5 ] [40]. The study implies the important role of kinetic effect in the formation of different phases in the actual synthesis. The effect of conformation of diamine on the structure of hybrids can be further evidenced by polymorphism and phase transition of hybrids containing 1,4-butanediamine (bda) and hda [43], as shown in Figure 5.3. However, the phenomena have only been observed in 3D [ZnTe(L)0.5 ] structures in which the ligand L possesses an even number of carbon atoms (centric and polar, with I site symmetry). No conformational change induced polymorphism, and phase transitions have been found for those containing L with an odd number of carbon atoms (acentric and polar, with m site symmetry). The all-trans conformation is dominant for the diamines in room-temperature 3D single-layer structures, exemplified by the 3D-α-[ZnTe(bda)0.5 ] (Figure 5.3a) and 3D-α-[ZnTe(hda)0.5 ] (Figure 5.3c). However, the conformations and site symmetries of the diamine molecules can be different in hybrids upon different temperatures, leading to new phases of 3D-[ZnTe(diamine)0.5 ] phases, e.g. γ-[ZnTe(bda)0.5 ], δ-[ZnTe(hda)0.5 ], and γ-[ZnTe(hda)0.5 ]. The local environment (i.e. the bonding between zinc and tellurium atoms) of the inorganic ZnTe slabs and the connectivity between the ZnTe slabs and hda molecules of the low-temperature γ-phases are similar to those of the corresponding α-phases. However, instead of an all-trans (T) form in α-[ZnTe(bda)0.5 ] and α-[ZnTe(hda)0.5 ] (namely TTT and TTTTT, respectively), the diamine molecules in γ-[ZnTe(bda)0.5 ] and γ-[ZnTe(hda)0.5 ] adopt the GTG and GTTTG configurations, respectively, resulting in shorter diamine lengths and thus shorter interlayer distances between the adjacent inorganic slabs compared to the corresponding α phases. Moreover, in δ-[ZnTe(hda)0.5 ], two of the five crystallographically independent I molecules have the all-trans (TTTTT) conformation and the other three take alternate conformations of GTTTG and GTGTT (mixture of trans (T) and gauche (G)). Note that the linear extents of the I molecules vary from about 10.3 Å (TTTTT), c. 9.5 Å (GTGTT), to about 9.2 Å (GTTTG). Consequently, the shortest interlayer distances between the adjacent inorganic ZnTe slabs are 6.25 Å in δ-[ZnTe(hda)0.5 ] (using the covalent radius of tellurium), compared with that of 8.12 Å in α-[ZnTe(hda)0.5 ]. Thus, the fivefold commensurate supercell (here 5b) in δ-[ZnTe(hda)0.5 ] has a c-axis that is shorter by about 2 Å than that of α-[ZnTe(hda)0.5 ]. More interestingly, α-[ZnTe(bda)0.5 ], α-[ZnTe(hda)0.5 ], and δ-[ZnTe(hda)0.5 ] could undergo temperature-induced phase transitions. The phase transition temperatures are about 130 K for α-[ZnTe(bda)0.5 ] → γ-[ZnTe(bda)0.5 ] and 225 K for δ-[ZnTe(hda)0.5 ] → γ-[ZnTe(hda)0.5 ]. The structural transformations are reversible, except that both α-[ZnTe(hda)0.5 ] and δ-[ZnTe(hda)0.5 ] transform to γ-[ZnTe(hda)0.5 ] upon cooling but then γ-[ZnTe(hda)0.5 ] changes back to δ-[ZnTe(hda)0.5 ] only upon warming. More interestingly, although the crystals of α-[ZnTe(hda)0.5 ] rarely survive multiple fractures during the initial cooling from

205

(a)

(c)

Figure 5.3

(b)

(d)

(e)

Structural diagrams for two polymorphs of 3D-[ZnTe(bda)0.5 ] (α (a) and γ (b)) and three polymorphs of 3D-[ZnTe(hda)0.5 ] (α (c), γ (d), and δ (e)).

5.2 Inorganic–Organic Hybrid Chalcogenides Based on II–VI Semiconductors

the synthesis at 473 K, the single crystals with good quality can be reliably obtained by slow heating the single crystals of δ-[ZnTe(hda)0.5 ] to above room temperature. Polymorphism and phase transition also exist in the hybrid structures containing hdz [33, 38, 46] or pda [28, 33]. As shown in Figure 5.4, there are four polymorphic phases for hdz-based hybrids. Similar to 3D hybrids composed of single atomic slabs of ZnTe and diamine of different lengths mentioned above, the structure of 3D-α-[ZnTe(hdz)0.5 ] contains single slabs interconnected by the shortest diamine (n = 0) and also crystallizes in space group of Pbca [46], while the 2D-α′ -[ZnTe(hdz)] is composed of [ZnTe(hdz)]n layers in which the less-distorted inorganic α′ -[ZnTe] slabs are sandwiched by terminally coordinated hdz molecules [33]. Both 1D-αand β-[ZnTe(hdz)2 ] structures contain [ZnTe] chains and two terminal hydrazine molecules [38]; however, the conformation and packing of chains are somewhat different. All the low-dimensional structures are further stabilized by interlayer or inter-chain VDM’s interactions, whereas both 1D-α- and β-[ZnTe(hdz)2 ] have additional hydrogen-bonding interactions among the pendent hydrazine molecules of the adjacent chains. The crystallinity, colour, and habit of the crystals and the morphology of the hybrid compounds are dependent on the II–VI compositions, metal and chalcogen sources, reaction temperature, and the use of mixed solvents. In most of the cases, a variety of M 2+ salts and elemental chalcogens were chosen as II-group metal ion and VI-group anion sources, respectively. Other reactant sources were also used, but then the optimal reaction temperature, crystalline size and morphology, and the (a)

(b)

(c)

a c b

(d)

a

(f) a

c

a

c

b

(e)

b

b b

(g)

c

a

a c

b

Figure 5.4 The 1∼3D ZnTe-based hybrid structures synthesized in hydrazine. (a) Framework of 3D-α-[ZnTe(hdz)0.5 ]; (b) the packing mode of layers in 2D-α′ -[ZnTe(hdz)]; (c) inorganic honeycomb 63 net of [ZnTe] in 2D-α′ -[ZnTe(hdz)]; (d) and (e) the chain structures in α-[ZnTe(hdz)2 ]; (f) and (g) the chain structures in β-[ZnTe(hdz)2 ].

207

208

5 Group 11–15 Metal Chalcogenides

phase purity of the final products would be different. For example, using metal powder as M 2+ source usually produces hybrids with smaller crystal size, whereas A2 Q (A = Li, Na, K, Rb, or Cs; Q = S, Se, and Te) as a chalcogen source would lower down the solvothermal reaction temperature, resulting in products of less crystallinity. The reactant source normally would not affect the phases produced. An exception was the isolation of 3D-α-[ZnTe(en)0.5 ] and 3D-β-[ZnTe(en)0.5 ] via using ZnCl2 and Zn(NO3 )2 ⋅6H2 O as the Zn2+ source, respectively [28, 44]. The hydrazine or mono- and di-amine molecules with low melting points are used as reactive solvents in the synthesis of II–VI hybrids, for they serve as both organic reactants and solvents. Sometimes, mixed solvents were used in the syntheses. For instance, pure 3D-α-[ZnTe(hdz)0.5 ] crystals were only obtained in a mixture of hydrazine and methylamine as the solvent [46], whereas 1D-ZnTe(pda) was synthesized in a mixture of pda and hydrazine [33]. The departure of half of hdz molecules from 1D-α-[ZnTe(hdz)2 ] at ∼100 ∘ C during thermogravimetric (TG) experiment led to the formation of 2D-α′ -[ZnTe(hdz)], demonstrating an interesting solid-to-solid phase transition [33]. Some hybrids could also be prepared as uniform nanobelts in ternary solutions [39]. Most of the hybrid nanostructures were synthesized under solvothermal conditions, whereas a few were only obtained at ambient conditions or in heated solutions in the presence of hydrazine, such as 1D-α- and 1D-β-[ZnTe(hdz)2 ] [38]. Alternatively, the hybrids can be prepared by exchange of organic ligands, for example, pure polycrystalline sample of 2D-[ZnSe(ba)] (ba = n-butylamine) can be prepared by solvothermal reaction of 2D-[ZnSe(pa)] with n-butylamine [42]. Reaction temperature has proved to be one of the critical parameters affecting the formation of final products under solvothermal conditions. The optimal range of temperature varied from 50 to 210 ∘ C, depending on the II–VI combinations and the nature of amines used. Normally, the desirable single-phased hybrid compounds could only be obtained at a relatively small range of temperatures. Higher temperatures would result in II–VI binaries as the major phase, whereas lower temperatures would lead to incomplete reactions. More interestingly, by slightly adjusting the reaction temperature, pure 2D-[MQ(monomine)] and 2D-[M 2 Q2 (monomine)] with monolayer and double layer of MQ, respectively, have been isolated [42]. Another noticeable trend is, as the elemental chalcogen sources change from S, Se, to Te, higher temperatures were needed for the formation of hybrid phases. Most of the S- and Se-based hybrids can only be obtained in polycrystalline form, and the cadmium-based hybrids exhibit lower crystallinity than their zinc-based analogues. The ZnTe- and MnSe-based hybrids usually form crystals of fairly large sizes. By carefully adjusting the reaction conditions, the single crystals of CdSe(en)0.5 have also been isolated [32, 33]. Therefore, their crystal structures could be determined by single crystal X-ray diffraction (SCXRD). For the hybrids that could only be prepared in powder form, Rietveld refinements on their powder X-ray diffraction (PXRD) patterns have been used to determine the crystal structures [47]. This method has been successfully applied in the crystal structure determination of a number of hybrids including 3D-[CdSe(L)0.5 ] (L = en, pda, and hda) [32, 33], 3D-α-[ZnS(en)0.5 ] [35], 3D-α-[ZnSe(L)0.5 ] (L = en, pda, bda, hda, and deta)

5.3 Discrete Chalcogenide Clusters Synthesized in Ionic Liquids

[29, 31, 39], 2D-α-[ZnSe(ba)], 2D-[(Zn2 Se2 )(ba)] [42], and 3D-α-[ZnSe(mxda)0.5 ] [41, 48]. The crystal structures of the remaining phases were proposed by investigating the similarity of their PXRD patterns with that of the hybrids with known structures.

5.3 Discrete Chalcogenide Clusters Synthesized in Ionic Liquids Ionic liquids (ILs) are commonly defined as salts consisting of predominantly ionic species that are fluid around or below 100 ∘ C, which possess generally negligible vapor pressure, wide liquid range, good solubility, good chemical, and thermal stability [49]. Ionothermal synthesis, the application of ILs as both solvent and template [50], is a new approach developed in recent decades for the synthesis of crystalline metal chalcogenides [51]. Compared with traditional molecular solvents, the ILs provide unique ionic reaction environment, which may contribute to the formation of novel metal chalcogenide compounds that are inaccessible by using molecular solvents [52]. Although the application of ionothermal techniques in chalcogenide chemistry is still in an early stage, some significant successes have already been achieved [51]. In 2009, Kanatzidis and coworkers reported on preparing the novel cationic chalcogenide clusters of [Sb7 S8 Br2 ](AlCl4 )3 in the lewis acidic IL (Emim)Br-AlCl3 (Emim = 1-ethyl-3-methylimidazolium) [23]; Dehnen and coworkers synthesized the largest-known discrete polyanion “zeotype” clusters of [Bmmim]24 [Sn36 Ge24 Se132 ] (Bmmim = 1-butyl-2,3-dimethylimidazolium) and [Bmim]24 [Sn32.5 Ge27.5 Se132 ] (Bmim = 1-butyl-3-methylimidazolium) that consist of only main-group elements [53], while Huang’s group has prepared a number of discrete chalcogenido Tn clusters [54–59] and novel crystalline selenidostannates [60–65] with diverse composition and structures in ILs. Selected discrete metal chalcogenides synthesized in ILs are listed in Table 5.2. Scheme 5.4 summarizes typical ILs used in the synthesis of crystalline chalcogenides.

5.3.1

Discrete Chalcogenido Tn Clusters

Significant progress has been achieved on the construction of chalcogenidometalate with open framework structures based on supertetrahedral clusters [12], while the research on the synthesis of discrete supertetrahedral clusters has been relatively rarely reported [67]. The electrical and optical properties of discrete supertetrahedral clusters can be adjusted by changing the composition and size of clusters, and they can also be used as precursors for further assembly to obtain porous or functional materials with open framework structures. In particular, compared with supertetrahedral cluster-based compounds with extended structures, discrete supertetrahedral Tn clusters can be regarded as fragments of cubic ZnS-type lattices, which are analogous to the smallest semiconductor quantum dots (QDs). Thus, their highly ordered lattice arrangements can provide an accurate model for studying the structure–property relationship [68].

209

Table 5.2

Crystallographic data for selected chalcogenide clusters synthesized in ILs.

V (Å3 )

Space group

References

90

18 218

C2/c

[55]

90

19 573

C2/c

[55]

107.76

C2/c

[55]

Formula

Tn

a (Å)

b (Å)

c (Å)

𝜶 (∘ )

𝜷 (∘ )

𝜸 (∘ )

[Bmmim]5 [In10 S16 Cl3 (Bim)]

T3

39.879

16.750

29.064

90

110.22

[Bmmim]5 [In10 S7.12 Se8.88 Cl3 (Bim)]

T3

41.360

17.025

29.322

90

108.563

[Bmmim]5 [In10 Se16 Cl3 (Bim)]

T3

41.314

16.952

28.921

90

107.76

90

[Bmmim]5 [In10 Se13.80 Te2.20 Cl3 (Bim)]

T3

42.578

17.262

29.765

90

108.609

90

20 733

C2/c

[55]

[Bmmim]6 [In10 Se16 Cl4 ](MIm)2

T3

19.726

19.726

31.169

90

90

120

10 503

P63 /m

[59]

[Bmmim]5 [(Me)2 NH2 ]4 [NH4 ]

T4

20.851

20.851

36.270

90

90

120

13 657

P63 /m

[56]

T4

20.872

20.872

35.981

90

90

120

13 575

P63 /m

[56]

T4

20.945

20.945

36.312

90

90

120

13 796

P63 /m

[56] [57]

[Mn4 In16 S31 (SH)4 ]⋅6H2 O [Bmmim]5 [(Me)2 NH2 ]4 [NH4 ] [Zn4 In16 S31 (SH)4 ]⋅6H2 O [Bmmim]5 [(Me)2 NH2 ]4 [NH4 ] [Cd4 In16 S31 (SH)4 ]⋅6H2 O [Bmmim]9 [Cd3 In17 S31 Cl4 ]

T4

21.603

21.603

38.273

90

90

120

15 469

P63 /m

[Bmmim]9 [Cd3 In17 S13 Se18 Cl4 ]

T4

20.095

20.477

23.679

66.161

71.126

64.248

7896

P-1

[57]

[Bmmim]9 [Cd3 In17 Se31 Cl4 ]⋅

T4

21.923

21.923

43.837

90

90

120

18 247

P63 /m

[57]

T5

23.617

40.808

46.252

90

94.748

90

44 423

C2/c

[54]

T5

23.018

23.018

42.321

90

90

120

19 418

P63 /m

[54]

T5

22.414

23.022

23.782

95.818

116.72

114.54

9322

P-1

[54]

[4,4′ -bpy] [Bmmim]12 [NH4 ] [Cu5 In30 S52 (SH)2 Cl2 ] [Bmmim]10 [NH4 ]3 [Cu5 Ga30 S52 (SH)4 ] [Bmmim]8 [NH4 ]3 [Cu5 Ga30 S52 (SH)2 (Bim)2 ]

[Bmmim]9.5 [NH4 ]2 [Cu5 Ga30 S52 (SH)1.5 Cl(Bim)1.5 ]

T5

22.168

22.846

24.663

115.17

111.39

96.407

9984

P-1

[54]

[Bmmim]12 [Cu5 In30 Se52 Cl3 (Im)]

T5

24.171

25.988

24.286

63.618

62.129

87.155

11 838

P-1

[58]

[Bmmim]12 [Cu5 In30 Se48.5 S3.5 Cl3 (Im)]

T5

24.132

25.953

24.400

62.875

61.868

86.119

11 781

P-1

[58]

[Bmmim]11 [Cd6 In28 Se52 Cl3 (MIm)]

T5

23.870

24.072

25.627

64.595

62.937

65.038

11 364

P-1

[58]

[Bmmim]11 [Cd6 In28 Se28.5 S23.5 Cl3 (MIm)] T5

23.753

23.813

23.813

63.601

61.814

79.282

10 759

P-1

[58]

[Bmmim]11 [Cd6 In28 Se16 S36 Cl3 (MIm)]

T5

23.573

23.574

23.884

63.361

61.697

79.190

10 440

P-1

[58]

[Bmmim]11 [Cd6 In28 Se8 S44 Cl1 (MIm)3 ]

T5

23.384

23.522

23.948

62.914

61.762

78.673

10 329

P-1

[58] [23]

[Sb7 S8 Br2 ](AlCl4 )3

—

11.989

16.896

17.378

90

90

90

3520.2

P21 21 21

[Emim]2 [Sn2 As2 S4 (S2 )2 Br2.43 Cl1.56 ]

—

15.189

12.716

16.435

90

90

90

3174.3

Pbca

[66]

[Bmmim]24 [Sn36 Ge24 Se132 ]

—

37.130

36.288

37.287

90

92.074

90

50 206

P21 /c

[53]

[Bmim]24 [Sn32.5 Ge27.5 Se132 ]

—

24.320

24.401

26.744

115.84

97.969

108.90

12 761

P-1

[53]

Bmmim, 1-butyl-2,3-dimethylimidazolium; Bmim, 1-butyl-3-methylimidazolium; Bzmim, 1-benzyl-3-methylimidazolium; Emim, 1-ethyl-3-methylimidazolium; en, ethanediamine; 4,4′ -bpy, 4,4′ -dipyridine; [(Me)2 NH2 ], dimethylamine; Bim, 1-butyl-2-methylimidazole; MIm, 1-methylimidazole; Im, imidazole.

212

5 Group 11–15 Metal Chalcogenides

N

+

N

N

+

N

N

1-Propyl-2, 3-dimethyl-imidazolium + [Prmmim]

N

N

1-Propyl-3-methyl-imidazolium [Prmim]+

1-Ethyl-3-methyl-imidazolium [Emim]+

N

+

+

N

F –

CI

Br

–

F

N

1-Butyl-3-methyl-imidazolium [Bmim]+

N

1-Butyl-2,3-dimethyl-imidazolium + [Bmmim]

+

+

N

1-Pentyl1-2,3-dimethyl-imidazolium + [Pmmim]

–

B F F

Tetrafluoroborate

Scheme 5.4 Typical cations and anions of ionic liquids used in the synthesis of crystalline chalcogenides.

The group 13 metal ion indium(III) commonly adopts tetrahedral coordination geometry with chalcogen atoms. It has been widely utilized as a component in the synthesis of supertetrahedral chalcogenide clusters. Among them, the [In10 Q20 ]10− (Q = S and Se) T3 clusters have been found to be one of the most popular SBUs for the construction of multidimensional chalcogenides [69–71]; the discrete T3 cluster, in contrast, is scarce [72]. In 2018, four isostructural chalcogenidometalate compounds were reported, namely [Bmmim]5 [In10 Q16 Cl3 (Bim)] (Q = S, S7.12 Se8.88 , Se, Se13.80 Te2.20 ; Bmmim; Bim = 1-butyl-2methylimidazole) [55]. As shown in Figure 5.5, these compounds consist of T3 anionic units of isolated [In10 Q16 Cl3 (Bim)]5− surrounded by [Bmmim]+ cations. [Bmmim]5 [In10 Se16 Cl3 (Bim)] represents the first example of discrete T3 InSe clusters, and the InTe-based discrete T3 cluster ([Bmmim]5 [In10 Se13.80 Te2.20 Cl3 (Bim)]) was characterized for the first time. These compounds were obtained by the reactions of In and chalcogens in the IL of [Bmmim]Cl, in which the IL acted as both solvent and reactants providing its cation [Bmmim]+ . More importantly, the decomposition of [Bmmim]+ cation released the [Bim] as a neutral terminal ligand for the corner In atom of T3 cluster. Noticeably, these discrete anionic T3 clusters are mostly stabilized by organic capping ligands, while completely inorganic T3 clusters are rare. Recently, an indium selenide [Bmmim]6 [In10 Se16 Cl4 ]⋅(MIm)2 (MIm = 1-methylimidazole) has been synthesized and characterized [59]. It features an organic-ligand free discrete T3 cluster with all four corners occupied by chloride ions. Packing diagram of [In10 Se16 Cl4 ]6− is deposited in Figure 5.5d, showing that the adjacent T3 clusters exhibit an up−down inverted arrangement along both the a- and b-axes. Inorganic chalcogenide nanoclusters are particularly desired in photocatalytic hydrogen evolution, as they may show better charge delocalization and transfer property compared with those capped with organic ligands.

5.3 Discrete Chalcogenide Clusters Synthesized in Ionic Liquids

(a)

(b) In Se CI N C

c a CI

(c)

(d) In Se

b a

c

Figure 5.5 (a) Anionic T3 cluster in [Bmmim]5 [In10 Se16 Cl3 (Bim)]; (b) view of [Bmmim]5 [In10 Se16 Cl3 (Bim)] along the b-axis showing the arrangement of zipper-like structures and the [Bmmim]+ cations; hydrogen atoms and the [Bmmim]+ cations within the zipper-like structures are omitted for clarity; (c) thermal ellipsoid ORTEP drawing of the [In10 Se16 Cl4 ]6− anionic cluster in [Bmmim]6 [In10 Se16 Cl4 ]⋅(MIm)2 ; displacement ellipsoid was drawn at 50% probability; (d) packing diagram of compound [Bmmim]6 [In10 Se16 Cl4 ]⋅(MIm)2 viewed along the c-axis showing the arrangement way of anions.

According to the global charge balance rule, however, introducing low-valence metal ions (e.g. Zn2+ , Cd2+ , Mn2+ , and Cu+ ) could help augment the cluster sizes from small T3 to larger Tn (n > 3). By means of such a mixed-metal synthetic strategy, three discrete T4 clusters [Bmmim]5 [(Me)2 NH2 ]4 [NH4 ][M 4 In16 S31 (SH)4 ]⋅6H2 O (M = Mn, Zn, and Cd) were synthesized in the IL [Bmmim]Cl in the presence of dimethylamine as an auxiliary solvent [56]. Interestingly, both the IL cations [Bmmim]+ and protonated dimethylamines [(Me)2 NH2 ]+ act as charge balance agents in the structures and stabilize the discrete T4 clusters via anion–π interactions and N–H· · ·S interactions in addition to ionic interactions, Figure 5.6. In the continuous efforts in preparing the molecular Tn clusters with different sizes and compositions, recently, three discrete Cd–In–Q (Q = S, S/Se, and Se) T4 clusters were synthesized by using an ionic-liquid-assisted precursor method, namely [Bmmim]9 [Cd3 In17 S31 Cl4 ], [Bmmim]9 [Cd3 In17 S13 Se18 Cl4 ], and [Bmmim]9 [Cd3 In17 Se31 Cl4 ]⋅(4,4′ -bpy) (4,4′ -bpy = 4,4′ -dipyridine) [57]. Their structures feature discrete T4 clusters of [Cd3 In17 Q31 Cl4 ]9− (Q = S, S/Se, and Se) with the imidazolium [Bmmim]+ cations as the CBAs. The presence of Cl− anions as the terminal ligands successfully prevents the T4 cluster from forming an extended framework and decreases the excessive charge of T4 clusters. The syntheses of large discrete Tn clusters are desirable because they can be useful for studying the quantum size effect. The large size of supertetrahedral

213

214

5 Group 11–15 Metal Chalcogenides

(a)

(c) Zn In S

(b)

c b

Figure 5.6 (a) The distribution of divalent Zn2+ ions and trivalent In3+ ions in T4 cluster of [Bmmim]5 [(Me)2 NH2 ]4 [NH4 ][Zn4 In16 S31 (SH)4 ]⋅6H2 O. (b) Discrete T4 cluster is surrounded by [Bmmim]+ cations, dimethylamines, and ammonium ions in [Bmmim]5 [(Me)2 NH2 ]4 [NH4 ] [Zn4 In16 S31 (SH)4 ]⋅6H2 O. (c) Discrete T4 clusters take an up–down inverted arrangement along the ab-plane and are sandwiched between the dimethylamine and ammonium layers. Source: Yang et al. [56]. Copyright 2018, the Royal Society of Chemistry.

clusters usually leads to a higher negative charge originating from the increasing number of unsaturated bonded chalcogen sites on the surface as the cluster gets larger. It is difficult to stabilize the large Tn clusters, and more geometrically matched cations are required to balance the charge. The largest discrete Tn clusters have been T4 clusters for a long time [73]. In 2012, a series of discrete T5 supertetrahedral chalcogenides based on group 13 Ga/In synthesized by ionothermal or solvothermal approaches were reported by Huang’s group, which are known as the largest discrete clusters of the supertetrahedral Tn family thus far, namely [Bmmim]10 [NH4 ]3 [Cu5 Ga30 S52 (SH)3 Cl], [Bmmim]12 [NH4 ][Cu5 In30 S52 (SH)3 Cl], [Bmmim]8 [NH4 ]3 [Cu5 Ga30 S52 (SH)2 (Bim)2 ], [Bmmim]9.5 [NH4 ]2 [Cu5 Ga30 S52 (SH)1.5 Cl(Bim)1.5 ], and [Bmmim]10 [NH4 ]3 [Cu5 Ga30-x Inx S52 (SH)4 ] (x = 0, 6.6, 14.5, 23.8, and 30) [54]. One notable feature of these compounds is the presence of monodentate groups (SH− or Cl− ) or neutral ligands (Bim) rather than bridging S2− ions at four corners of the T5 cluster, which might be very important for the formation of discrete clusters. Interestingly, the discrete T5 clusters in compounds [Bmmim]12 [NH4 ][Cu5 In30 S52 (SH)3 Cl] and [Bmmim]10 [NH4 ]3 [Cu5 Ga30 S52 (SH)3 Cl] (Figure 5.7a) are arranged in distorted hexagonal diamond type superlattices (Figure 5.7c), while those in [Bmmim]9.5 [NH4 ]2 [Cu5 Ga30 S52 (SH)1.5 Cl(Bim)1.5 ] and [Bmmim]10 [NH4 ]3 [Cu5 Ga30-x Inx S52 (SH)4 ] (x = 0, 6.6, 14.5, 23.8 and 30) (Figure 5.7b) are distorted cubic diamond-type superlattices (Figure 5.7d). The use of precursor is a key for the syntheses of such discrete cluster-based compounds. However, the poor solubility of regular polymeric precursor will result in the impure phase of target compounds with a low yield. Thus, the step by step precursor method is worthy of exploration in synthesizing new crystalline chalcogenides in ILs. In 2020, six new M-In-Q (M = Cu or Cd; Q = Se or Se/S) chalcogenide compounds with discrete T5 anionic cluster were reported, namely [Bmmim]12 [Cu5 In30 Q52 Cl3 (Im)] (Q = Se and

5.3 Discrete Chalcogenide Clusters Synthesized in Ionic Liquids

(a)

(b) In

Cu/In

(c)

(d)

b b c

c

a

a

Figure 5.7 (a) Discrete [Cu5 In30 S52 (SH)2 Cl2 ]13− T5 cluster surrounded by [Bmmim]+ cations in [Bmmim]12 [NH4 ][Cu5 In30 S52 (SH)3 Cl]; (b) discrete [Cu5 Ga30 S52 (SH)2 (Bim)2 ]11− T5 cluster in [Bmmim]8 [NH4 ]3 [Cu5 Ga30 S52 (SH)2 (Bim)2 ]. Illustration of a hexagonal diamond-type and cubic diamond-type arrangement fashion of clusters in [Bmmim]12 [Cu5 In30 Se52 Cl3 (Im)] (Im = imidazole) (c) and [Bmmim]11 [Cd6 In28 Se52 Cl3 (MIm)] (d) via connecting the adjacent centers of clusters, respectively. Source: Xiong et al. [54]. Copyright 2012, the Royal Society of Chemistry.

Se48.5 S3.5 ), [Bmmim]11 [Cd6 In28 Q52 Cl3 (MIm)] (Q = Se, Se28.5 S23.5 , and Se16 S36 ), and [Bmmim]9 [Cd6 In28 Se8 S44 Cl(MIm)3 ]. [Bmmim]12 [Cu5 In30 Se52 Cl3 (Im)] and [Bmmim]11 [Cd6 In28 Se52 Cl3 (MIm)] represent the largest molecular supertetrahedral Tn selenide clusters to date [58]. The optimal crystallization conditions are provided in Scheme 5.5. The first step was to form the Cu-In-Q-IL or In-Q-IL precursors via microwave-assisted ionothermal processes. The use of precursor is crucial to the crystallization of these discrete supertetrahedral Tn selenide clusters. Without using the precursors, when other reaction conditions were kept the same, the one-step reactions resulted in the formation of mixed phases and/or targeted phases with poor crystallinity. By contrast, the microwave-assisted ionothermal processes could produce uniform gelatinous precursors, which might prevent the formation of thermodynamically stable binary phases (e.g. In2 Q3 , Cu2 Q, and CdQ; Q = S/Se) and promote the crystallization of discrete supertetrahedral Tn selenide clusters.

215

216

5 Group 11–15 Metal Chalcogenides

4.5 In + 2.2 Se + 0.13 CuSCN

N+ N – CI

MV, 200 °C, 1h

11 CH3CN + 11 Mlm + 3.0 DBU 150 °C, 4 d 11 CH3CN + 11 Mlm + 3.0 DBU + 1.3 thiourea 150 °C, 4 d

In + 2.2 Se

(Bmmim)12[Cu5In30Se52CI3(Im)] (Bmmim)12[Cu5In30Se48.5S3.5CI3(Im)]

(Bmmim)11[Cd6In28Se52CI3(MIm)]

In + 1.1 Se + 1.1 S In + 0.60 Se + 1.6 S

3.3

N+N – CI

MV, 200 °C, 1 h

12 Mlm + 0.30 CdCl2 + 1.3 thiourea 150 °C, 3 d

In + 0.30 Se + 1.9 S

(Bmmim)11[Cd6In28Se28.5S23.5CI3(MIm)] (Bmmim)11[Cd6In28Se16S36CI3(MIm)] (Bmmim)9[Cd6In28Se8S44CI(MIm)3]

Scheme 5.5 Typical crystallization processes of supertetrahedral T5 selenide clusters in IL; the amount of reactant is in the unit of mmol. Source: Wang et al. [58]. Copyright 2020, Wiley-VCH.

As shown in Figure 5.8a, the anions in [Bmmim]12 [Cu5 In30 Se52 Cl3 (Im)] are discrete units of [Cu5 In10 Se52 Cl3 (Im)]12− . The central metal site surrounded by four tetrahedral Se2− ions was assigned to a Cu+ ion. Four other Cu+ ions are distributed over twelve possible metal sites on the four faces of the super tetrahedron, exhibiting a 1 : 3 Cu/In co-occupancy. Remaining metal sites were designated as In3+ ions. Its another notable feature is that the four vertexes of T5 are occupied by Im or Cl, which is very important for the formation of discrete cluster. The formation of terminal In—Cl and In—N bonds in vertexes not only brings less negative charge than M—Se bonds but also prevents them from polymerizing, thereby promoting the formation of discrete clusters. [Bmmim]11 [Cd6 In28 Se52 Cl3 (MIm)] is characteristic of a coreless T5 anionic cluster with one 1-methylimidazole molecule and three Cl− as terminal ligands (Figure 5.8b). There are twelve metal sites inside the cluster co-occupied by Cd2+ and In3+ in a ratio of 0.5 : 0.5. The remaining metal sites are designated as In3+ ions. The ratio of Cd:In (6 : 28) in it is in good agreement with that in the typical discrete T5 Cd-In-S clusters [67]. By connecting the nearest adjacent points (the distance of M–M is about 15.5 Å if treating each T5 cluster as a pseudo-tetrahedral atom), all the arrangement patterns of these T5 clusters are a cubic diamond-type superlattice (Figure 5.8c and d). In summary, a series of discrete Tn (T3 to T5) chalcogenides have been synthesized in ILs, in which the IL acts as a reactive solvent or provides [Bmmim]+ as the charge balancing cation and Cl− as the terminal ligand and even as reactant to release a neutral terminal ligand (e.g. Bim and Im). The obtained discrete T5 chalcogenide clusters are the largest discrete supertetrahedral Tn clusters to date. This work represents an important step in the synthetic development of discrete clusters.

5.3.2

Other Discrete Chalcogenide Clusters

Using the IL (Emim)Br–AlCl3 as the solvent, Kanatzidis and coworkers synthesized a chalcogenide crystal of [Sb7 S8 Br2 ](AlCl4 )3 , which represents the first cationic chalcogenide cluster synthesized in ILs [23]. The compound consists of cationic [Sb7 S8 Br2 ]3+ clusters and charge-balanced [AlCl4 ]− anions in a ratio of 1 : 3 (Figure 5.9). Each cationic cluster adopts a double cubane structure in which two distorted cubic clusters connect by sharing one corner (the Sb1 site). The other

5.3 Discrete Chalcogenide Clusters Synthesized in Ionic Liquids

(a)

(b)

InCu In Cu Se Cl N C

(c)

InCd In Se Cl N C

(d)

a c

a b

b

c

Figure 5.8 Anionic T5 clusters in [Bmmim]12 [Cu5 In30 Se52 Cl3 (Im)] (a) and [Bmmim]11 [Cd6 In28 Se52 Cl3 (MIm)] (b); illustration of the cubic diamond-type arrangement fashion of clusters in [Bmmim]12 [Cu5 In30 Se52 Cl3 (Im)] (c) and [Bmmim]11 [Cd6 In28 Se52 Cl3 (MIm)] (d) via connecting the adjacent centers of clusters. Source: Wang et al. [58]. Copyright 2020, Wiley-VCH. Figure 5.9 The structure of compound [Sb7 S8 Br2 ] (AlCl4 )3 .

Sb2 Br1

Sb4

Sb Br

Sb3

S

Sb1

Al Cl

Sb7 Sb5 Br2

Sb6

corners are alternately occupied by Sb and S atoms. Two Sb sites (Sb2 and Sb6) have terminal Sb—Br bonds projecting out of the cluster structure. The compound exhibits nonlinear optical properties with the second harmonic generation intensity about one-third that of potassium dihydrogen phosphate (KDP) at 700 nm and comparable to that of KDP above 900 nm. Metal sulfides with double cubane-like cluster structures have been widely studied due to their geometrical relevance to the active site structures of enzymes and potential catalytic applications [74]. In 2013, by controlling the mixed solvents, [Cr7 S8 Cl2 (NH3 )14.5 (H2 O)1.5 ]Cl3 ⋅H2 O and [Emim]2 [Sn2 As2 S4 (S2 )2 Br2.43 Cl1.56 ]

217

218

5 Group 11–15 Metal Chalcogenides

(a) S1

Cr4 N1B/C11

(b) Br1/C11

Cr1

Cr3 S2 S3

C12

As1

S4

Br2/C12

S1

S3 Sn1 S2

C13/O2 S4 Cr2 C11B/O1

Figure 5.10 The structures of compounds [Cr7 S8 Cl2 (NH3 )14.5 (H2 O)1.5 ]Cl3 ⋅H2 O (a) and [Emim]2 [Sn2 As2 S4 (S2 )2 Br2.43 Cl1.56 ] (b).

[66] were obtained. [Cr7 S8 Cl2 (NH3 )14.5 (H2 O)1.5 ]Cl3 ⋅H2 O features a cationic double cubane-like cluster of [Cr7 S8 Cl2 (NH3 )14.5 (H2 O)1.5 ]3+ (Figure 5.10a). The edge-sharing Cr3+ octahedra are arranged in two [Cr4 S4 ] cubane-like clusters, which are further jointed by a Cr(1)3+ ion located in an inversion center to form a centrosymmetrically double cubane-like cluster of [Cr7 S8 ] surrounded by the terminal NH3 , Cl− , and H2 O ligands. Though a number of corner-shared double cubane-like chalcogenide clusters have been reported, [Cr7 S8 Cl2 (NH3 )14.5 (H2 O)1.5 ]Cl3 ⋅H2 O provides the first example with NH3 -rich terminal ligands. [Emim]2 [Sn2 As2 S4 (S2 )2 Br2.43 Cl1.56 ] is the first tin arsenic chalcohalide compound incorporating organic cation. In its anionic cluster, one 𝜓-AsS3 and two distorted [SnS4 BrCl] octahedra form an incomplete cubane-like cluster. And then two such centrosymmetrically related clusters share the Sn(1)–S(2)–Sn(1)–S(2) face to form an incomplete double cubane-like cluster (Figure 5.10b). Dehnen and coworkers have extended the bottom-up strategy for the synthesis of microporous chalcogenides by using precursors containing [Ge4 Se10 ]4− units in ILs. Two discrete polyanions “Zeoball” [Sn36 Ge24 Se132 ]24− and [Sn32.5 Ge27.5 Se132 ]24− have been isolated (Figure 5.11) [53]. The polyanion comprises 192 atoms with an outer diameter of 2.83 nm (including van der Waals radii of the surface atoms), representing the largest known discrete polyanion consisting only of main-group elements; charge neutrality is achieved by the inclusion of 24 countercations, [Bmmim]+ in [Sn36 Ge24 Se132 ]24− and [Bmim]+ in [Sn32.5 Ge27.5 Se132 ]24− . Notably, the ILs used in the synthesis of chalcogenides thus far are ILs based on imidazolium cations. Other discrete clusters from ILs have also been reported, however, will not be described in detail here [51].

5.4 Chalcogenidostannates Crystalline open framework chalcogenidometalates are desirable for applications such as ion exchange [1], photocatalysts [5], and ion conductivity [75]. In particular, the research on thio- and selenidostannates in the past few decades has resulted

218

5 Group 11–15 Metal Chalcogenides

(a) S1

Cr4 N1B/C11

(b) Br1/C11

Cr1

Cr3 S2 S3

C12

As1

S4

Br2/C12

S1

S3 Sn1 S2

C13/O2 S4 Cr2 C11B/O1

Figure 5.10 The structures of compounds [Cr7 S8 Cl2 (NH3 )14.5 (H2 O)1.5 ]Cl3 ⋅H2 O (a) and [Emim]2 [Sn2 As2 S4 (S2 )2 Br2.43 Cl1.56 ] (b).

[66] were obtained. [Cr7 S8 Cl2 (NH3 )14.5 (H2 O)1.5 ]Cl3 ⋅H2 O features a cationic double cubane-like cluster of [Cr7 S8 Cl2 (NH3 )14.5 (H2 O)1.5 ]3+ (Figure 5.10a). The edge-sharing Cr3+ octahedra are arranged in two [Cr4 S4 ] cubane-like clusters, which are further jointed by a Cr(1)3+ ion located in an inversion center to form a centrosymmetrically double cubane-like cluster of [Cr7 S8 ] surrounded by the terminal NH3 , Cl− , and H2 O ligands. Though a number of corner-shared double cubane-like chalcogenide clusters have been reported, [Cr7 S8 Cl2 (NH3 )14.5 (H2 O)1.5 ]Cl3 ⋅H2 O provides the first example with NH3 -rich terminal ligands. [Emim]2 [Sn2 As2 S4 (S2 )2 Br2.43 Cl1.56 ] is the first tin arsenic chalcohalide compound incorporating organic cation. In its anionic cluster, one 𝜓-AsS3 and two distorted [SnS4 BrCl] octahedra form an incomplete cubane-like cluster. And then two such centrosymmetrically related clusters share the Sn(1)–S(2)–Sn(1)–S(2) face to form an incomplete double cubane-like cluster (Figure 5.10b). Dehnen and coworkers have extended the bottom-up strategy for the synthesis of microporous chalcogenides by using precursors containing [Ge4 Se10 ]4− units in ILs. Two discrete polyanions “Zeoball” [Sn36 Ge24 Se132 ]24− and [Sn32.5 Ge27.5 Se132 ]24− have been isolated (Figure 5.11) [53]. The polyanion comprises 192 atoms with an outer diameter of 2.83 nm (including van der Waals radii of the surface atoms), representing the largest known discrete polyanion consisting only of main-group elements; charge neutrality is achieved by the inclusion of 24 countercations, [Bmmim]+ in [Sn36 Ge24 Se132 ]24− and [Bmim]+ in [Sn32.5 Ge27.5 Se132 ]24− . Notably, the ILs used in the synthesis of chalcogenides thus far are ILs based on imidazolium cations. Other discrete clusters from ILs have also been reported, however, will not be described in detail here [51].

5.4 Chalcogenidostannates Crystalline open framework chalcogenidometalates are desirable for applications such as ion exchange [1], photocatalysts [5], and ion conductivity [75]. In particular, the research on thio- and selenidostannates in the past few decades has resulted

5.4 Chalcogenidostannates

Ge Sn

Sn

Se Ge

Se

Figure 5.11 Two different orientations of the [Sn36 Ge24 Se132 ]24− anionic structure of [Bmmim]24 [Sn36 Ge24 Se132 ], disregarding partial Sn/Ge disorder. Source: Lin et al. [53]. Copyright 2012, the American Chemical Society.

in the isolation of a lot of crystalline compounds, especially the open framework compounds with abundant structural types [13]. Huang’s group has carried out a detailed study on the effects of structure-directing cations, auxiliary solvents, and temperature on the construction of novel thio- and selenidostannates; selected compounds are listed in Table 5.3.

5.4.1

Chalcogenidostannates

A layer compound, namely [(Me)2 NH2 ]4/3 [(Me)3 NH]2/3 [Sn3 S7 ]⋅1.25H2 O (FJSMSnS), was obtained in large scale by a facile, one-pot, and economically solvothermal method [76]. In the structure, all the Sn atoms are five coordinated with S to form [SnS5 ] trigonal bipyramids. Three [SnS5 ] trigonal bipyramids are fused into a [Sn3 S10 ] unit with a [Sn3 S4 ] semi-cubane-like core. Then, each [Sn3 S10 ] unit connects to other three such units by edge-sharing, resulting in a 2D [Sn3 S7 ]n 2n− anionic layer parallel to the ab plane (Figure 5.12a), in which there are windows formed by 24-membered [Sn12 S12 ] rings from six [Sn3 S4 ] cores. The layers are stacked in AA sequence along the c-axis (Figure 5.12b). The interlayer distance is estimated to be 7.258 Å. The [(Me)2 NH2 ]+ and [(Me)3 NH]+ cations are located at the interlayer spaces. ILs have been used to prepare porous materials, such as zeolites and metal–organic frameworks [52], while their use for preparing porous chalcogenides is scarce. Choosing selenidostannates as a model system, Huang and coworkers carried out systematic studies on reactions of Sn and Se in a 1 : 2.5 ratio in imidazolium-based ILs [60]. By using various ILs and adjusting the weight fraction of ILs and N2 H4 ⋅H2 O, three open framework selenidostannates with high crystallinity have been obtained, namely 3D-[Bmim]4 [Sn9 Se20 ], 3D-[Bmmim]4 [Sn9 Se19 (Se2 )0.9 Se0.1 ], and 2D-[Pmmim]4 [Sn17 Se38 ], Figure 5.13. Remarkably, compounds [Bmim]4 [Sn9 Se20 ] and [Bmmim]4 [Sn9 Se19 (Se2 )0.9 Se0.1 ]

219

Table 5.3

Crystallographic data for selected chalcogenidostannates.

Formula

D

a (Å)

b (Å)

c (Å)

𝜶 (∘ )

𝜷 (∘ )

𝜸 (∘ )

V (Å3 )

S. G.

References

[(Me2 )NH2 ]4/3 [(Me)3 NH]2/3

2

22.565

13.067

14.806

90

101.23

90

4282.3

C2/c

[76]

[Bmim]4 [Sn9 Se20 ]

3

19.974

26.854

14.575

90

105.07

90

7549

Cc

[60]

[Bmmim]4 [Sn9 Se19 (Se2 )0.9 Se0.1 ]

3

20.594

11.083

36.208

90

104.94

90

7985

P21 /c

[60]

[Sn3 S7 ]⋅1.25H2 O

[Pmmim]4 [Sn9 Se19 (Se2 )0.93 Se0.07 ]

3

20.358

11.615

36.055

90

105.05

90

8232.5

P21 /c

[60]

[Pr mmim]2 [Sn3 Se7 ]

2

13.947

13.947

27.855

90

90

90

4692.5

P32 21

[61]

[Bmmim]2 [Sn3 Se7 ]

2

13.938

13.938

27.774

90

90

90

4672.7

P32 21

[61]

[Pmmim]8 [Sn17 Se38 ]

2

20.351

21.901

22.832

113.68

110.83

93.22

8472

P-1

[60]

[Pr mmim]4 [Sn9 Se20 ]

2

18.652

20.353

20.196

90

105.24

90

7397

P21 /c

[61]

[Bmmim]2 [Sn3 Se7 ]

1

24.538

13.961

19.434

90

94.539

90

6637

P21 /c

[61]

[Pr mmim]2 [Sn3 Se7 ]

1

14.270

15.549

14.377

90

92.749

90

3186

P21 /c

[61]

[Bmmim]3 [Mn(en)3 ]2 [Sn9 Se21 ]Cl

2

19.449

19.511

26.615

90

109.59

90

9514.9

P21 /c

[62]

[Bmmim]6 [Mn(deta)2 ]2 [Sn15 Se35 ]

2

13.426

29.904

19.585

90

106.64

90

7533.4

P21 /m

[62]

[Bmmim]2 [Ni(teta)(en)][Sn3 Se7 ]2

2

24.375

13.378

17.789

90

93.949

90

5786.9

P21 /c

[65]

[Bmmim]1.5 [detaH]0.5

3

31.088

10.539

19.971

90

99.966

90

6444.7

P21 /c

[65]

[Ni(deta)2 ] [Sn4 Se9 ]2 [Bmmim]3 [Ni(en)3 ]2 [Sn9 Se21 ]Cl

2

19.488

19.626

26.793

90

109.93

90

9634

P21 /c

[64]

[Bmmim]8 [Ni2 (teta)2

2

18.976

25.108

20.573

90

100.87

90

9626.4

P21 /n

[64]

(μ-teta)] Sn18 Se42 [Bmmim]2 [Ni(1,2-pda)3 ]Sn8 Se18 ]

3

27.791

27.791

47.428

90

90

120

31 724

Pn

[64]

[Bmmim]4 [Ni(tepa)Cl]2 [Ni(tepa)Sn12 Se28 ]

2

21.223

13.499

23.872

90

112.46

90

6320.3

R3c

[64]

[CH3 NH3 ]6 Ag12 Sn6 S21

3

18.865

19.912

14.313

90

100.12

90

5292.5

P21 /c

[77]

[CH3 NH3 ]2 Ag4 SnIV 2 SnII S8

3

19.378

7.390

13.683

90

90

90

1959.5

Pnma

[78]

[NH4 ]3 AgSn3 Se8

1

8.276

8.276

13.461

90

90

90

922.0

P4/nbm

[79]

[(Me)2 NH2 ]0.75 [Ag1.25 SnSe3 ]

3

13.805

13.805

9.671

90

90

90

1843.0

P4 21 m

[80]

[(Me)2 NH2 ]0.05 Cs0.70 [Ag1.25 SnSe3 ]⋅0.25H2 O

3

13.989

13.989

8.853

90

90

90

1732.4

P4 21 m

[80]

[(Me)2 NH2 ]0.10 Rb0.65

3

13.998

13.998

8.685

90

90

90

1701.9

P4 21 m

[80]

3

13.909

13.909

8.862

90

90

90

1714.3

P4 21 m

[80]

3

13.635

13.635

9.149

90

90

90

1700.9

P4 21 m

[77] [77]

[Ag1.25 SnSe3 ]⋅0.75H2 O [(Me)2 NH2 ]0.25 [NH4 ]0.5 [Ag1.25 SnSe3 ]⋅0.50H2 O [CH3 NH3 ]2 [H3 O][Ag5 Sn4 Se12 ]⋅ EtOH [NH4 ]4 [Ag12 Sn7 Se22 ]

3

54.634

6.818

13.518

90

99.955

90

4959.1

C2/c

[Bmmim]7 [AgSn12 Se28 ]

2

10.898

13.555

23.091

88.904

86.70

68.23

3162.6

P-1

[61]

(Bmmim)8 In8 Sn8 Se30 (Se4 )2

2

19.593

19.5931

37.836

90

90

90

14 525

I41 /a

[63]

(Bmmim)8 CsIn9 Sn7 Se30 (Se4 )2

2

19.644

19.644

37.705

90

90

90

14 549

I41 /a

[63]

(Bmmim)8 RbIn9 Sn7 Se30 (Se4 )2

2

19.622

19.622

37.690

90

90

90

14 511

I41 /a

[63]

Bmmim, 1-butyl-2,3-dimethylimidazolium, Bmim, 1-butyl-3-methylimidazolium, Bzmim, 1-benzyl-3-methylimidazolium, Pr mmim, 1-propyl-2,3-dimethyl-imidazolium, Pmmim, 1-pentyl-2,3-dimethylimidazolium, Emim, 1-ethyl-3-methylimidazolium, en, 1,2-ethanediamine, deta, diethylenetriamine, teta, triethylenetetramine, tepa, tetraethylenepentamine, 1,2-pda, 1,2-diaminopropane, CH3 NH3 , methylamine, [(Me)2 NH], dimethylamine, [(Me)3 N], trimethylamine.

5 Group 11–15 Metal Chalcogenides

a

b c

Sn1 Sn3 S1 Sn2 b

a Sn S N C

c

(a)

(b)

Figure 5.12 (a) A 2D [Sn3 S7 ]n 2n− anionic layer parallel to the ab plane in the compound [(Me)2 NH2 ]4/3 [(Me)3 NH]2/3 [Sn3 S7 ]⋅1.25H2 O (FJSM-SnS); (b) packing of the layers in FJSM-SnS in a perspective view along the b-axis. The H2 O molecules and H atoms of organic amines are omitted for clarity. Source: Qi et al. [76]. Copyright 2015, the Royal Society of Chemistry.

N

Sn + Se + N2H4∙H2O N

N N N

N

222

3D-[Bmmim]4[Sn9Se19(Se2)0.9Se0.1] 3D-[Bmim]4[Sn9Se20]

2D-[Pmmim]4[Sn17Se38]

Figure 5.13 Three novel crystalline selenidostannates synthesized in imidazolium-based ionic liquids with a small amount of hydrazine monohydrate as additive. Source: Li et al. [60]. Copyright 2011, Wiley-VCH.

represent the first examples of IL-directed 3D open framework based on binary selenidostannates, and compound [Pmmim]4 [Sn17 Se38 ] features 2D microporous structure composed of inorganic selenidostannate nanotubes. The preparation of these open framework selenidostannates will help understand crystallization and phase selectivity of chalcogenides in ILs. Introducing amine to the IL-containing reaction systems adds more reaction variables, offering opportunities for obtaining new crystalline chalcogenides that are inaccessibly obtained by traditional methods. Indeed, the structures of selenidostannates can further be tuned by simply changing the types of ILs and amines, as well as reaction temperature and time. As a result, four crystalline selenidostannates have been obtained, namely 1D-[Pr mmim]2 [Sn3 Se7 ], 1D-[Bmmim]2 [Sn3 Se7 ], 2D-[Bmmim]2 [Sn3 Se7 ], and 2D-[Pr mmim]2 [Sn3 Se7 ] [61]. Compounds 1D-[Pr mmim]2 [Sn3 Se7 ], 1D-[Bmmim]2 [Sn3 Se7 ], and 2D-[Bmmim]2 [Sn3 Se7 ] possess the same anionic formula of [Sn3 Se7 ]n 2n− ; however, their structures are different. The structure of 1D-[Pr mmim]2 [Sn3 Se7 ] features a 1D chain of anionic

5.4 Chalcogenidostannates

[Sn3 Se7 ]n 2n− (Figure 5.14a), while in 1D-[Bmmim]2 [Sn3 Se7 ], the [Sn3 Se7 ] units are interconnected to each other by edge-sharing two Se atoms resulting in a infinite double chain of [Sn3 Se7 ]n 2n− (Figure 5.14b). Both the 2D-[Bmmim]2 [Sn3 Se7 ] and 2D-[Pr mmim]2 [Sn3 Se7 ] belong to the trigonal chiral space group P32 21, whose structures feature a well-known 63 net [Sn3 Se7 ]n 2n− (Figure 5.14c) lamellar structure with [Pr mmim]+ and [Bmmim]+ cations located in the inter-layer spaces, respectively. The [Sn3 Se7 ]n 2n− double chain is possibly a “intermediate product” of

(a)

(b)

Sn Se a b

c

(c)

Figure 5.14 a) View of one [Sn3 Se7 ]n 2n− chain in 1D-[Pr mmim]2 [Sn3 Se7 ]; b) view of one [Sn3 Se7 ]n 2n− double chain in 1D-[Bmmim]2 [Sn3 Se7 ]; c) view of one [Sn3 Se7 ]n 2n− layer in 2D-[Bmmim]2 [Sn3 Se7 ] and 2D-[Pr mmim]2 [Sn3 Se7 ] along the c-axis.

223

224

5 Group 11–15 Metal Chalcogenides

the reaction pathway from single chain to honeycomb layer, which has never been isolated before when using molecular solvent as reaction media. The effectiveness of the strategy of using the mixture of IL and auxiliary amine as a designable solvent system to synthesize novel crystalline selenidostannates has been further proved by the successful isolation of a novel lamellar selenidostannate, namely [Pr mmim]4 [Sn9 Se20 ] [61]. Its structure features a [Sn9 Se20 ]n 4n− layer with the inter-layer spaces filled by the [Bmmim]+ cations, Figure 5.15. In the structure, there are two kinds of SBUs, i.e. the double [Sn6 Se10 ] semicube constructed by two [Sn3 Se4 ] semicubes edge-sharing two Se atoms and the [Sn3 Se10 ] unit consisting of three corner-sharing [SnSe4 ] tetrahedra. The two kinds of SBUs are interconnected to each other to form a (44 ⋅44 ) 2D anionic network of [Sn9 Se20 ]n 4n− along the ac plane. Things become more interesting when introducing metal-coordinate complexes (typically metal–amine complexes, denoted as MACs) as SDAs in the synthesis of selenidostannates in ILs. Previously, a number of compounds containing anionic [Sn3 Q7 ]n 2n− (Q = S and Se) layers have been synthesized with different cations as SDAs (e.g. alkali metals, protonated organic amines, cation of ILs, and metal-coordinated complexes), in which the regular hexagonal six-membered ring is commonly found in the lamellar [Sn3 Q7 ]n 2n− . While in the coexistence of IL cation and MAC, it is of particular interest that not only hexagonal six-membered ring, but also an elliptic six-membered ring and even larger eight-membered ring can be established, as exemplified in compounds [Bmmim]6 [Mn(deta)2 ]2 [Sn15 Se35 ] [62] and [Bmmim]8 [Ni2 (teta)2 (μ-teta)][Sn18 Se42 ] (teta = triethylenetetramine) [64] (Figure 5.16). This might be due to the formation of larger aggregated cationic complex as a supramolecular inclusion complex or liquid clathrate through hydrogen-bonding networks between MACs and imidazolium cations. In addition, from the synthesis point of view, all of the selenidostannates that contain both the

(a)

(b)

[Sn6Se10] (c)

a b

c

[Sn3Se10]

Figure 5.15 The structure of [Pr mmim]4 [Sn9 Se20 ]. (a) View of one [Sn9 Se20 ]n 4n− layer along the b-axis; (b) a [Sn6 Se10 ] double-semicube unit; (c) a [Sn3 Se10 ] unit. Source: Li et al. [61]. Copyright 2013, the Royal Society of Chemistry.

5.4 Chalcogenidostannates

+

x2

Regular hexagonal six-membered ring

x2

Distorted hexagonal six-membered ring

x2

Compressed six-membered ring

x2

Elliptic six-membered ring

x2

Heart-shaped eight-membered ring

Figure 5.16 Structural evolution of the lamellar [Sn3 Se7 ]n 2n− based on six- or eight-membered rings formed by direct fusion of single [Sn3 Se7 ]n 2n− zigzag chains or insertion of [Sn3 Se9 ] units between chains. Source: Du et al. [62]. Copyright 2015, the Royal Society of Chemistry.

MACs and imidazolium cations were sensitive to the mole ratio of ILs and MACs. This might also be a proof for the existence of larger aggregated cationic complex. Huang et al. concluded that there was a synergy of MACs and IL cations as mixed SDAs during the formation of selenidostannates. Huang’s group further explored the synergy of ILs and the MACs as mixed SDAs. They found that various selenidostannates with structures varying from 2D layers to 3D frameworks could be obtained, such as 2D-[Bmmim]3 [Mn(en)3 ]2 [Sn9 Se21 ]Cl [62], 2D-[Bmmim]6 [Mn(deta)2 ]2 [Sn15 Se35 ] [62], 2D-[Bmmim]2 [Ni(teta)(en)] [Sn3 Se7 ]2 [65], 2D-[Bmmim]1.5 [detaH]0.5 [Ni(deta)2 ][Sn4 Se9 ]2 [65], 2D-[Bmmim]3 [Ni(en)3 ]2 [Sn9 Se21 ]Cl [64], 2D-[Bmmim]8 [Ni2 (teta)2 (μ-teta)][Sn18 Se42 ] [64], 2D[(Bmmim)]4 [Ni(tepa)Cl]2 [Ni(tepa)Sn12 Se28 ] (tepa = tetraethylenepentamine) [64], 3D-[Bmmim]2 [Ni(1,2-pda)3 ][Sn8 Se18 ] [64], and 3D-[Bmmim]1.5 [detaH]0.5 [Ni(deta)2 ] [Sn4 Se9 ]2 [65] (Figure 5.17), which might be ascribed to different configurations of MACs. Along with the growth of polyamine chains from en, 1,2-pda, deta, teta to tepa, the MACs exhibit different coordination configurations resulting in the variation of the hydrogen bonding of MACs. Supramolecular interactions (including hydrogen bonding and anion–π interaction) are found between IL cations and MACs, clearly revealing a synergistic structure-directing effect on directing the formation of selenidostannate structures. It’s worth mentioning that even when there is no direct bonding between MACs and (Bmmim)Cl, the synergistic structure-directing effect could still be observed. Figure 5.18 compares the crystal structures of 2D-[Bmmim]3 [Ni(en)3 ]2 [Sn9 Se21 ]Cl

225

226

5 Group 11–15 Metal Chalcogenides c

a

Synergistic structure-directing effect

b

[Ni(pda)3]2+

[Ni(teta)(en)]2+ a

a

c

b a

[Ni(deta)2]2+

c

[Ni(en)3]2+ –

+

Ionic Liquid

2D layer

3D framework

Figure 5.17 2D-[Bmmim]2 [Ni(teta)(en)][Sn3 Se7 ]2 , [Bmmim]3 [Ni(en)3 ]2 [Sn9 Se21 ]Cl and 3D-[Bmmim]2 [Ni(1,2-pda)3 ][Sn8 Se18 ], [Bmmim]1.5 [detaH]0.5 [Ni(deta)2 ][Sn4 Se9 ]2 structures directed by the synergistic IL cation and corresponding MAC. Se(15)

Sn(6)

Se(11)

Sn Se Ni N C

Sn(4)

Ni(2)

Sn Se Ni N C

Sn(5)

Ni(1)

Se(8)

Ni(2)

3.896 Å N(15)

Cl(1)

Cl(1)

N(15)

Se(15)

Ni(1) Ni(2)

Se(15)

a

c Ni(1) 43 3.7

Å

N(18)

Cl(1) Cl(2) b a

(a)

(b)

N(17)

b c

a

08

Å

Ni(3)

3.5

(c)

Figure 5.18 Crystal structures of 2D-[Bmmim]3 [Ni(en)3 ]2 [Sn9 Se21 ]Cl (a) and 2D[(Bmmim)]4 [Ni(tepa)Cl]2 [Ni(tepa)Sn12 Se28 ] (b). Comparison of the supramolecular interactions between [Bmmim]Cl and MACs of [Ni(en)3 ]2+ and [Ni(tepa)Cl]+ is shown in (c). Source: Du et al. [64]. Copyright 2016, John Wiley & Sons.

[64] and 2D-[(Bmmim)]4 [Ni(tepa)Cl]2 [Ni(tepa)Sn12 Se28 ] [64], with corresponding aggregated cationic complexes shown in Figure 5.18c. The latter represents a rare example of 2D lamellar selenidostannates with organic-decorating groups. In brief, a synergistic structure-directing effect and reaction mechanism of ILs and MACs were demonstrated in depth by Huang’s group. Although it is difficult to predict the resulting products prepared by this mixed-cation method, the research develops an effective route toward the synthesis of novel metal chalcogenides by utilizing the synergistic effect between ILs and MACs. Huang and coworkers successfully expanded ionothermal method to synthesize heterometalllic selenidostannates, such as 2D-[Bmmim]7 [AgSn12 Se28 ] [61] (the structure will be described in Section 5.4.2) and [Bmmim]8 [In8 Sn8 Se30 (Se4 )2 ] [63]. [Bmmim]8 [In8 Sn8 Se30 (Se4 )2 ] features an infinite 2D-anionic [In8 Sn8 Se30 (Se4 )2 ]n 2n− layer composed of ternary core-less In8 Sn8 Se34 -T2,2 super-supertetrahedral clusters interconnected by polyselenium Se4 chains along the ab plane. The asymmetric

5.4 Chalcogenidostannates

unit is composed of four [MSe4 ] tetrahedra (M = 0.5 In + 0.5 Sn), two [Bmmim]+ cations, and half of the polyselenium Se4 chain. Since the In3+ and Sn4+ atoms are isoelectronic, they cannot be distinguished directly by X-ray diffraction. The In/Sn ratios in the clusters were established to be 1 : 1 by energy-dispersive X-ray (EDX) spectroscopy and ICP-AES. In a In8 Sn8 Se34 -T2,2 cluster, the [MSe4 ] tetrahedra are interconnected each other by sharing vertical Se atoms to form a hollow cage structure (Figure 5.19a). Based on the results of structural refinements and electroneutrality principle, the atomic site at the center of the supertetrahedron is vacant, as also confirmed by EDX and elemental analysis. Noteworthy, it is also the first case that the polyselenium Se4 chain was combined with supertetrahedral cluster. The layers are further packed into supramolecular 3D structure along the c-axis, in which the [Bmmim]+ cations are filled in the inter-laminar spaces as the charge compensator accompanied by the hydrogen-bonding interactions and electrostatic interactions between the clusters and [Bmmim]+ cations. Interestingly, the [Bmmim]+ cations completely enclose every core-less In8 Sn8 Se34 -T2,2 cluster and form a double-shell ivory ball-like T2,2@Bmmim cage (Figure 5.19b). The distance of the adjacent In8 Sn8 Se34 -T2,2 clusters is about 1.74 nm, which are connected by Se4 bridge to form 2D-layer (Figure 5.19c). Moreover, by means of a

InSe4

(a)

SnSe4

T2

T2,2 cluster cage

1.6 nm

1.738 nm

Sn/In Se C N Cavity (b)

(c)

Figure 5.19 (a) The In8 Sn8 Se34 -T2,2 cluster built up of InSe4 and SnSe4 tetrahedra; (b) the T2,2 cluster with [Bmmim]+ cations shell; (c) the connection of In8 Sn8 Se34 -T2,2 cluster cages in (Bmmim)8 In8 Sn8 Se30 (Se4 )2 via Se4 bridges. Source: Du et al. [63]. Copyright 2015, the American Chemical Society.

227

228

5 Group 11–15 Metal Chalcogenides

one-pot ionothermal method, an alkali metal ion (Cs+ or Rb+ ) could be precisely doped into the central cavity of the cluster, forming an alkali@T2,2@Bmmim quaternary cluster.

5.4.2

Heterometallic Chalcogenidostannates Containing Ag+

The open frameworks constructed from SBUs based on tetrahedra have been well documented [12], whereas those based on heterometallic SBUs integrating tetrahedra and other geometry configurations are less explored. Ag+ can adopt a variety of coordination modes with chalcogen anions ranging from linear, trigonal to tetrahedral geometries. Particularly, the trigonal [AgQ3 ] units (Q = S and Se) sometimes form clusters themselves or generate novel SBUs by integrating with other types of metal polyhedra, especially in the Ag+ -rich chalcogenide phases 3D-[CH3 NH3 ]6 Ag12 Sn6 S21 [77] (Figure 5.20) and 3D-[CH3 NH3 ]2 Ag4 SnIV 2 SnII S8 [78] (Figure 5.21) have been solvothermally synthesized by utilizing [CH3 NH3 ]+ as SDAs and CBAs. [CH3 NH3 ]6 Ag12 Sn6 S21 crystallizes in the monoclinic space group of P21 /c and features an unprecedented 3D open framework anionic structure of [Ag12 Sn6 S21 ]n 6n− . Its asymmetric unit contains one formula unit. In the framework of [Ag12 Sn6 S21 ]n 6n− , all the Sn atoms are tetrahedrally coordinated by sulfur atoms, whereas the Ag+ ions adopt three different coordination modes including trigonal planar coordination geometry, tetrahedrally coordination and linear coordination geometry. One [Ag(9)S2 ] dumbbell, one [Ag(11)S3 ] triangle, and two [AgS4 ] tetrahedra (i.e. [Ag(1)S4 ] and [Ag(4)S4 ]) are condensed by corner- or edge-sharing to form a unique [Ag4 S8 ] cluster, denoted as C1 (Figure 5.20a). Whereas three trigonal planar [AgS3 ] units (i.e. [Ag(3)S3 ], [Ag(6)S3 ], and [Ag(7)S3 ]) and one [Ag(2)S4 ] tetrahedron are jointed together via corner- or edge-sharing to produce another unique [Ag4 S8 ] cluster, denoted as C2 (Figure 5.20b). Such two [Ag4 S8 ] clusters are interconnected by sharing S(7) and S(3) atoms to form a [Ag8 S14 ] moiety. As shown in Figure 5.20c, there is a [Ag3 S8 ] moiety with a six-membered [Ag3 S3 ] ring defined by one trigonal planar [AgS3 ] unit and two [AgS4 ] tetrahedra via vertex-sharing. The [Ag8 S14 ] and [Ag3 S8 ] moieties are interconnected via corner-sharing S(11) and S(16) atoms to constitute one [Ag11 S20 ] moiety, which further connects two adjacent such [Ag11 S20 ] moieties through sharing S(19) atoms to form a [Ag11 S19 ]n 27n− ribbon. Such two centrosymmetric [Ag11 S19 ]n 27n− ribbons are interconnected by the tetrahedral Ag(10) atoms to form a [Ag12 S19 ]n 26n− double ribbon (Figure 5.20d,e), which further links four neighboring identical ones via sharing S(21) atoms to form a 3D [Ag12 S17 ]n 22n− anionic framework. Then, the [Sn2 S7 ] dimers (Figure 5.20f) are decorated in the [Ag12 S17 ]n 22n− framework by corner-sharing S atoms to result in a more complex 3D open framework of [Ag12 Sn6 S21 ]n 6n− with 1D channels (Figure 5.20g). The channels parallel to the a-axis are distorted cross-shaped with a cross section of 6.40 × 10.13 Å2 , which are composed of a 20-membered ring defined by two [AgS3 ], two [AgS4 ], and six [SnS4 ] units via corner-sharing. Interestingly, compound [CH3 NH3 ]6 Ag12 Sn6 S21 is a rare example containing three types of basic building blocks of [AgS2 ] dumbbells, [AgS3 ] triangles, and [AgS4 ] tetrahedra in one compound.

5.4 Chalcogenidostannates

Ag(1)

Ag(11) (a)

Ag(9)

Ag(4) S(7)

S(19)

S(3)

(d)

Ag(6)

Ag(2) (b)

(c)

Ag(3)

S(11)

S(16) Ag(10)

Ag(5)

Ag(7)

Ag(12) Ag(8) S(21)

S(19)

Ag(10)

b

a (f)

S(21)

S(21)

S(21)

(e)

S(21)

S(21)

(g)

c b

Sn Ag S

Figure 5.20 Crystal structure of [CH3 NH3 ]6 Ag12 Sn6 S21 . (a) The [Ag4 S8 ] cluster C1; (b) another unique [Ag4 S8 ] cluster C2; (c) the [Ag3 S8 ] unit containing a six-membered [Ag3 S3 ] ring; (d) the [Ag12 S19 ]n 26n− double ribbon extending along the a-axis; (e) the [Ag12 S19 ]n 26n− double ribbon viewed along the a-axis; (f) the [Sn2 S7 ] unit; and (g) perspective view of the anionic framework along the a-axis. C, N, and H atoms are omitted for clarity. Source: Zhang et al. [77]. Copyright 2017, the American Chemical Society.

Compound [CH3 NH3 ]2 Ag4 SnIV 2 SnII S8 represents the first Ag–Sn–S compound characteristic of mixed-valence of Sn(IV) and Sn(II). It belongs to the orthorhombic space group of Pnma with half a formula unit in the crystallographically asymmetric unit. Its structure features a 3D anionic open framework of [Ag4 SnIV 2 SnII S8 ]n 2n− with 1D channels along the b-axis where the [CH3 NH3 ]+ cations as SDAs or templates reside. The asymmetric unit of [CH3 NH3 ]2 Ag4 SnIV 2 SnII S8 contains two unique Ag(I) ions (Ag(1) and Ag(2)), two halves Sn(IV) ions ((Sn(1) and

229

230

5 Group 11–15 Metal Chalcogenides

S(1) S(1) S(1) S(3) Ag(2) S(2) Ag(1) S(6) S(1) S(5) Ag(1) Ag(2) S(2)

Sn(1) S(2) Sn(2) S(2)

(a)

(c)

(d)

(b)

S(6) Sn(3) S(4)

S(5) S(6)

(e)

Sn(3) Sn(3)

a

Ag Sn S

c

(f)

Figure 5.21 (a) [Ag4 S7 ]n 10n− ribbon where the [Ag4 S4 ] rings are filled in green; (b) [SnIV SnII S5 ]n 4n− chain; (c) two centrosymmetric [Ag4 SnIV SnII S8 ]n 6n− ribbons; (d) [Ag4 SnIV SnII S8 ]n 6n− ribbon; (e) [Sn(3)S4 ] tetrahedron; and (f) perspective view of the open framework structure of [CH3 NH3 ]2 Ag4 SnIV 2 SnII S8 along the b-axis with one of the [Ag4 SnIV SnII S8 ]n 6n− ribbons circled. C, N, and H atoms are omitted for clarity. Source: Zhang et al. [78]. Copyright 2016, the American Chemical Society.

(Sn(3)), half a Sn(II) ion (Sn(2)), two and four halves S2− ions, and two halves [CH3 NH3 ]+ cations. Except for Ag(1), Ag(2), S(2), and S(6), all the non-hydrogen atoms are located at 4c sites with mirror symmetry. The Ag(1) atom is in a trigonal environment bonding to three S atoms, while Ag(2) is in a distorted tetrahedral coordination geometry with four S atoms. The Sn(1) and Sn(3) atoms are tetrahedrally coordinated by four S atoms, while the Sn(2) atoms adopt an infrequent trigonal pyramidal coordination geometry.

5.4 Chalcogenidostannates

Vertex-sharing of two [Ag(1)S3 ] and two [Ag(2)S4 ] units produces an eight-membered [Ag4 S4 ] ring; such rings are interconnected via corner-sharing S(1) and S(3) atoms to constitute a 1D [Ag4 S7 ]n 10n− ribbon extending along the b-axis (Figure 5.21a), while a [SnIV SnII S5 ]n 4n− chain is formed by alternating arrangement of [Sn(2)S3 ] pyramids and [Sn(1)S4 ] tetrahedra via vertex-sharing S(2) atoms along the b-axis (Figure 5.21b). Then, the [Ag4 S7 ]n 10n− ribbon and the [SnIV SnII S5 ]n 4n− chain are condensed by sharing the S(1) and S(3) atoms to give rise to a [Ag4 SnIV SnII S8 ]n 6n− ribbon (Figure 5.21c), two of which are centrosymmetric and further fused via sharing the S(1) atoms to form a complex [Ag4 SnIV SnII S8 ]n 6n− ribbon extending along the b-axis (Figure 5.21d). Each [Ag4 SnIVSnII S8 ]n6n− ribbon further connects four adjacent [Ag4 SnIV SnII S8 ]n6n− ribbons via tetrahedral Sn(3)4+ ions (Figure 5.21e) to result in the final 3D anionic open framework structure of [Ag4 SnIV 2 SnII S8 ]n 2n− with 1D channels (Figure 5.21f). An intriguing structural feature of [CH3 NH3 ]2 Ag4 SnIV 2 SnII S8 is that there exists mixed-valent Sn atoms in the ratio of Sn(II)/Sn(IV) = 1/2. In 2011, Huang’s group reported on a novel silver–tin–selenide compound, namely [(Me)2 NH2 ]0.75 [Ag1.25 SnSe3 ] (Figure 5.22) [80], which was obtained by solvothermal reaction of AgCl, Se, and Sn in a mixed solvent of N,N ′ -dimethylformamide (DMF) and hydrazine hydrate at 160 ∘ C for six days. In the preparation, the in situ generation of dimethylammoniun ions is essential. An isomorphic compound of [CH3 NH3 ]2 [H3 O]Ag5 Sn4 Se12 ⋅EtOH [77] was isolated by a similar reaction for [(Me)2 NH2 ]0.75 [Ag1.25 SnSe3 ] except replacing the mixed solvent of DMF and hydrazine hydrate by methylamine (33–40% alcohol solution). As shown in Figure 5.22, the open framework 3D-[(Me)2 NH2 ]0.75 [Ag1.25 SnSe3 ] features 3D channels occupied by [(Me)2 NH2 ] cations. The Sn4+ ion is tetrahedrally coordinated to Se atoms. The Ag(1) ion has a trigonal planar coordination geometry, while the Ag(2) atom is tetrahedrally coordinated. The Se2− ions adopt two different coordination modes; Se(1) and Se(2) are μ3 –Se atoms connecting one Sn1 and two Ag ions, while Se(3) and Se(4) are μ2 –Se atoms linking two Sn ions. In the structure, the four trigonal planar [Ag(1)Se3 ] units are jointed together via corner-sharing to form a unique [Ag4 Se8 ] cluster. The Ag· · ·Ag distances range from 3.1825(16) to 3.2414(16) Å. The clusters are interconnected by tetrahedral [Ag(2)Se4 ] units via edge-sharing to form an infinite [Ag5 Se8 ] chain along the c-axis. Each [Ag5 Se8 ] chain further connects to four adjacent [Ag5 Se8 ] chains via [Sn2 Se2 ] units resulting in a 3D network. [(Me)2 NH2 ]0.75 [Ag1.25 SnSe3 ] exhibits rapid and high efficient ability for the capture of Cs+ over a wide pH range, while [CH3 NH3 ]2 [H3 O]Ag5 Sn4 Se12 ⋅C2 H5 OH exhibits photocatalytic degradation of crystal violet (CV) under visible light irradiation. Single crystals of the [NH4 ]+ salts of [AgSn3 Se8 ]3− [77] and [Ag12 Sn7 Se22 ]4− [79] were solvothermally synthesized with the reactants of Sn, Se, and AgCl in mixed solvents containing hydrazine hydrate at 160 ∘ C for five to seven days. The N2 H4 ⋅H2 O might act as a solvent and also decompose into NH4 + in the reactions. [NH4 ]3 [AgSn3 Se8 ] crystallizes in the tetragonal space group of P4/nbm featuring an infinite linear anionic chain of [AgSn3 Se8 ]n 3n− , Figure 5.23. Edge-sharing of three tetrahedral [SnSe4 ] units produces linear [Sn3 Se8 ]4− building blocks, which

231

232

5 Group 11–15 Metal Chalcogenides

Ag2 Ag1 (b) Se

Sn1 (c)

Sn Ag Se

b

N C

a (a)

Figure 5.22 (a) View of the structure of [(Me)2 NH2 ]0.75 [Ag1.25 SnSe3 ] along the c-axis; (b) view of the [Ag5 Se8 ] chain extended along the c-axis; and (c) the [Sn2 Se6 ] unit. Source: Li et al. [80]. Copyright 2011, Royal Society of Chemistry. Figure 5.23 (a) View of the [AgSn3 Se8 ]n 3n− anionic chain extended along the c-axis. (b) Perspective view of the structure of [CH3 NH3 ]6 Ag12 Sn6 S21 along the c-axis. H atoms are omitted for clarity. Source: Zhang et al. [77]. Copyright 2017, the American Chemical Society. Sn(1) Sn(2) Sn(1)

b a

(a)

(b)

Ag Sn Se N

5.4 Chalcogenidostannates

are further interconnected by Ag+ ions to form the infinite straight [AgSn3 Se8 ]n 3n− chain. The [AgSn3 Se8 ]n 3n− anionic chains propagate along the c-axis are well separated by charge balancing NH4 + cations. Compound (NH4 )4 Ag12 Sn7 Se22 belongs to the space group of C2/c. Its anionic [Ag12 Sn7 Se22 ]n 4n− framework is constructed from arrays of [Sn3 Se9 ]n 6n− chains interconnecting [SnAg6 Se10 ]n 10n− layers by sharing Se(3), Se(4), and Se(6), and [Ag3 Se4 ]n 5n− layers by sharing Se(7), Se(9), Se(10), and Se(11) (Figure 5.24a). As illustrated in Figure 5.24b, the [SnAg6 Se10 ]n 10n− exhibits a tri-layer structure extended along the bc plane. The Ag(1)Se4 and Ag(3)Se4 tetrahedra are interconnected to form a (63 ) [Ag2 Se5 ]n 8n− layer by core-sharing the Se(1), Se(2), and Se(3) atoms, and then two such [Ag2 Se5 ]n 8n− layers are alternately interconnected by a pair of [Ag(2)Se3 ] by corner-sharing Se(1), Se(2), and Se(3) and [Sn(1)Se4 ] by corner-sharing Se(1) and Se(2) to form the sandwich-like [SnAg6 Se10 ]n 10n− layer, in which the twofold rotation axes passing through the Sn(1) atoms are parallel to the b-axis. Figure 5.24c shows a [Sn3 Se9 ]n 6n− chain extended along the b-axis formed by [Sn(3)Se4 ] connecting [Sn(2)Se4 ] and [Sn(4)Se4 ] via corner-sharing the Se(5), Se(6), and Se(8) atoms. As shown in Figure 5.24d, the planar trigonal [Ag(5)Se3 ] and [Ag(6)Se3 ] corner-share the Se(9), Se(10), and Se(11) to form a [Ag2 Se3 ]n 4n− ribbon; the ribbons are interconnected into a [Ag3 Se4 ]n 5n− layer by [Ag(4)Se3 ] through

b c

a Se11 Se9 Se10

Se6 Se7 Se4

Se3

(a)

Sn2 Ag3 Ag2 Se5 Se3 Se1 Sn3 Sn4 Se2 Sn1 Ag1 Se6 Se8

Se10 Se11

Ag5 Ag6 Ag4 Se11 Se10 Se9

b c

a

(b)

b c

b c

Se11

a

a

(c)

(d)

Figure 5.24 (a) 3D-(NH4 )4 Ag12 Sn7 Se22 structure viewed along the b-axis. H atoms are omitted for clarity. (b) [SnAg6 Se10 ]n 10n− layer. The arrows represent the twofold axes. (c) [Sn3 Se9 ]n 6n− chain. (d) [Ag3 Se4 ]n 5n− layer. Source: Du et al. [79]. Copyright 2016, the American Chemical Society.

233

234

5 Group 11–15 Metal Chalcogenides

Se Sn Ag

a

b c

Ag Sn Se

Figure 5.25 View of one [AgSn12 Se28 ]n 7n− layer in [Bmmim]7 [AgSn12 Se28 ] running along the [81] direction. Source: Li et al. [61] Copyright 2013, the Royal Society of Chemistry.

corner-sharing Se(10) and Se(11) with Ag(6) and Ag(5), respectively. Compound (NH4 )4 Ag12 Sn7 Se22 is toxic metal free and exhibits an obvious photosensitivity in the near-infrared range. Remarkably, a lamellar silver-selenidostannate of [Bmmim]7 [AgSn12 Se28 ] was obtained by ionothermal method [61]. The layer of [AgSn12 Se28 ]n 7n− contains the [Sn12 Se28 ]n 8n− double chain similar to that found in 1D-[bmmim]2 [Sn3 Se7 ] [61]. Interestingly, the Ag+ ion adopts the two-coordinate linear [AgSe2 ] coordination geometry and interconnects the double chains into a lamellar structure (Figure 5.25). It is noteworthy that the Ag+ ion in the known inorganic silver chalcogenides normally has trigonal planar or tetrahedral coordination geometries. The compound represents a rare example of heterometallic chalcogenides containing linear-coordinate Ag+ ion. The synthesis of open framework chalcogenidostannates often requires the addition of organic amines as SDAs, referring to the complementary shapes of the SDAs and the frameworks they direct. It seems that the smaller organic amine may be suitable for the synthesis of open framework chalcogenidostannates. Scheme 5.6 compares the synthesis of some Ag–Sn–Q (Q = S, Se) compounds.

5.5 Chalcogenidoantimonates 5.5.1

Thioantimonates

The research on crystalline chalcogenidoantimonate(III) compounds was very fruitful in the past three decades [13, 14]. Because of the stereochemical effect of the lone pair of electrons and the wide range of coordination number of Sb(III) from 3 to 6

5.5 Chalcogenidoantimonates

[(CH3)2NH2]+ DMF/hdz (160 °C) NH4+ DAMSI/hdz (160 °C) [Bmmim]+ hdz (160 °C) Sn + Se/S + AgCl

NH4+ Formamide/hdz (160 °C) [CH3NH3]+ Methylamine (140 °C) [CH3NH3]+ Methylamine (160 °C)

3D-[(CH3)2NH2]0.75[Ag1.25SnSe3] 3D-(NH4)4Ag12Sn7Se22

2D-(Bmmim)7AgSn12Se28 1D-(NH4)3AgSn3Se8

3D-[CH3NH3]6Ag Sn6S21 12

3D-[CH3NH3]2Ag SnIV2SnIIS8 4

DMF = N, Nʹ-dimethylformamide; Bmmim = 1-buty1-2,3-dimethyl-imidazolium hdz = NH2NH2.H20; DAMSI = 4-(4-(dimethylamino)styry1)-1-methylpyridinium iodide

Scheme 5.6 Typical crystallization processes of Ag–Sn–Q (Q = S and Se) compounds directed by small molecular amines or/and cation of ILs.

and the application of various SDAs or templates (typically organic amines or their transition metal complexes), chalcogenidoantimonate(III) compounds exhibit large structural and compositional diversity. Many thioantimonates(III) have been reported [14]. More of them possess low-dimensional structures. By contrast, relatively little progress has been made on the preparation of 3D-thioantimonates(III) if taking no account of the weaker secondary Sb–S interactions. Limited examples of 3D-thioantimonates include K2 Sb4 S7 [82], [Ni(aepa)2 ]Sb4 S7 [83], [M(en)3 ][Sb12 S19 ] (M = Co and Ni) [84], [Ni(cyclam)][Sb4 S7 ] [85], and [Co(cyclam)]x [cyclamH2 ]1-x [Sb4 S7 ] (0.08 ≦ x ≦ 0.74) [85]. In 2012, Huang’s group reported a polar 3D thioantimonate [Ni(phen)3 ]2 Sb18 S29 (phen = 1,10-phenanthroline) [86]. Significantly, the first utilization of [M(phen)3 ]2+ as the template and the Sb:S ratio (1 : 1.611) of [Ni(phen)3 ]2 Sb18 S29 are unique amongst those in the reported thioantimonates. Its structure features a 3D framework with the largest channels in thioantimonates filled with two distinct arrays of [Ni(phen)3 ]2+ complexes (Figure 5.26). Except for Sb(1) and Sb(2), the other sixteen antimony atoms are coordinated to three sulfur atoms to form the SbS3 trigonal pyramidal geometries with the Sb–S distances ranging from 2.336(3) to 2.539(2) Å, whereas both the Sb(1) and Sb(2) are four-coordinated, with two short (2.423(2)–2.478(2) Å) and two long (2.616(2)–2.868(2) Å) Sb—S bonds. The Sb(9–16)S3 trigonal pyramids are interconnected by sharing the corners to form a Sb8 S8 ring (R1), whereas another similar Sb8 S8 ring (R2) consisting of the Sb(9), Sb(11–13), Sb(15–18), and eight sulfur atoms is also observed which shares three SbS3 units with R1. The alternating arrangements of the two types of {Sb8 S8 } rings

235

236

5 Group 11–15 Metal Chalcogenides S24

Sb18

Sb14

Sb4

Sb5

Sb15

Sb13

S1 S19 Sb16 R2 R1 Sb12 Sb11 S8 Sb9 Sb17 S12 Sb10

(d)

Sb6

Sb8

S24

S29

Sb3

Sb7

S14

b S14

+(b)

(a)

Sb1

S6

Sb2

(a) (b)

S10 R3

S3 Sb1

a c

S4

18 Å

Sb2 S7

(d)

8.5 Å

Layer

S2 Ribbon Sb S

(c)

(e)

Figure 5.26 The structure of [Ni(phen)3 ]2 Sb18 S29 . (a) 1D ribbon of [Sb10 S18 ]n 6n− ; (b) trans-[Sb2 S6 ] dimer; (c) [Sb12 S20 ]n 4n− layer; (d) [Sb6 S11 ] unit; and (e) 3D framework viewed along the a-axis.

result in a 1D ribbon ([Sb10 S18 ]n 6n− ) along the a-axis (Figure 5.26a). The Sb(1)S4 and Sb(2)S4 unit share one edge (S(4) and S(6)) forming a trans-[Sb2 S6 ] dimer (Figure 5.26b). Interconnections of 1D ribbons of [Sb10 S18 ]n 6n− by the tetradentate bridging [Sb2 S6 ] dimers along the c-axis lead to a [Sb12 S20 ]n 4n− layer along the ac plane by sharing four sulfur atoms (S(2), S(3), S(7), and S(10)) (Figure 5.26c). The new eight-membered Sb8 S8 rings (R3) are observed in the layer due to the Sb2 S6 dimers bridging 1D ribbons. In addition, two [Sb3 S6 ]3− semicubes constructed by three SbS3 trigonal pyramids (Sb(3–5)S3 and Sb(6–8)S3 , respectively) are connected into a [Sb6 S11 ] unit by sharing the S(29) atom (Figure 5.26d). Then, the [Sb6 S11 ] units acting as linkers further crosslink the [Sb12 S20 ]n 4n− layers into a 3D [Sb18 S29 ]n 4n− network (Figure 5.26e). An inorganic–organic hybrid compound Mn2 (api)Sb2 S5 (api = N-(3-aminopropyl)imidazole) was reported by Huang et al. [87]. Its structure consists of {Mn2 Sb2 S5 }n layers interlinked by the neutral api ligands via Mn—N coordination bonds (Figure 5.27). Two Sb1 and two Mn1 are connected by six μ3 -S2− ions to form a di-semi-cubane-like {Mn2 Sb2 S6 } core, while one Sb2 and one Mn2 couple into a {MnSbS2 } core with a common-shared S3–S4 edge. Alternating interlinkages of {Mn2 Sb2 S6 } cores and {MnSbS2 } cores in a 1 : 2 ratio give rise to a honeycomb-like {Mn2 Sb2 S5 }n layer, with concomitant formation of an eight-memberred ring (8-MR) of {Mn4 Sb4 S8 }. The layers stack in an ABAB sequence along the c-axis. The api linkers weave the inorganic layers into a 3D framework by connecting Mn2 ions from adjacent layers.

5.5.2 Chalcogenidometalates Containing Group 12(II) Ions and Antimony(III) Versatile structures of chalcogenidometalates based on group 12(II) ions (Zn2+ , Cd2+ , and Hg2+ ) and antimony(III) varying from 0D to 3D have been reported,

5.5 Chalcogenidoantimonates

Mn1 Mn2 Sb1

Sb2

8-MR

b c b a

(a)

Figure 5.27 layer (b).

Sb Mn S N C

(b)

The 3D-hybrid framework of Mn2 (api)Sb2 S5 (a) and its inorganic {Mn2 Sb2 S5 }n

some of which are listed in Table 5.4. In these structures, the group 12(II) metal ions are coordinated exclusively by chalcogen atoms and usually adopt tetrahedral coordination geometry. The combination of such tetrahedral {MQ4 } (M = Zn, Cd, and Hg; Q = S and Se) with the 𝜓-{SbQx } (Q = S, Se; x = 3, 4) or the self-assemblies of the {MQ4 } tetrahedra can result in various types of SBUs and TBUs (Figure 5.28). Zn2+ ion can also present the trigonal bipyramidal geometry due to steric constraints from the organic ligands, whereas Hg2+ ion can variously adopt linear, trigonal planar, or tetrahedral geometries due to its larger polarizability. On the other hand, a longer mean M—Q bond length is expectable along with a higher coordination number as well as a larger covalent radius of the coordinating Q atom [106]. In the linear {HgS2 } unit, the Q–Hg–Q angle may deviate from 180∘ because of the distortion caused by asymmetric coordination environment of the metal center. Thus far, however, all the {HgS2 } units in the Hg–Sb–S inorganic moieties adopt an ideal linear geometry due to the location of a crystallographic inversion center in Hg2+ ion. The Hg2+ in {HgS2 } normally functions as bridging atom connecting the adjacent 0D or 1D BUs. For example, [(Me)2 NH2 ]2 [HgSb8 S14 ] [88] and [detaH2 ][HgSb8 S14 ] [107] feature the identical 1D-[HgSb8 S14 ]n 2n− ribbon (Figure 5.29a) constructed by interconnecting two parallel, centrosymmetrically related {Sb4 S7 }n ribbons through di-coordinated Hg2+ ions. The adjacent Hg⋅⋅⋅Hg distance in this ladder-shaped ribbon is 10.02 Å. By contrast, Hg2+ ions in [Mn(phen)]2 HgSb2 S6 link {[Mn(phen)]2 Sb2 S6 } clusters into an infinite ribbon with adjacent Hg· · ·Hg distance of 7.75 Å (Figure 5.29b) [89]; the {[Mn(phen)]2 Sb2 S6 } cluster contains two identical 𝜓-{SbS3 } units, different from the organically coordinated clusters of double-semicube in 1D-{[Mn(L)]3 (AsV S4 )2 }n (L = 2,2′ -bipyridine [bipy] or phen) [108, 109] and 1D-{[Mn(phen)]3 (AsV S4 )(AsIII S3 )}n [110]. Linear {HgS2 } and tetrahedral {HgS4 } coexist in [TM(deta)2 ][Hg3 Sb4 S10 ] (TM = Mn, Co, and Ni) [111]. Such coexistence is scarce, as was only found previously in Hg–As–Q phases such as [Ph4 P]2 [Hg2 As4 Se11 ] (Ph = phenyl) [81]. Further, one {Hg3 S8 } group containing one {HgS2 } and two {HgS4 } polyhedra is

237

Table 5.4

Crystallographic data for selected chalcogenidoantimonates.

Formula

D

a (Å)

b (Å)

c (Å)

𝜶 (∘ )

𝜷 (∘ )

[Ni(phen)3 ]2 Sb18 S29

3

11.18

41.30

12.3700

90

106.56

Mn2 (api)Sb2 S5 –100 K

3

12.4377

12.2724

21.0852

90

90

Mn2 (api)Sb2 S5 –150 K

3

12.4568

12.2988

21.0020

90

90

Mn2 (api)Sb2 S5 –200 K

3

12.4920

12.3011

20.9270

90

Mn2 (api)Sb2 S5 –273 K

3

12.5276

12.3171

20.8638

Mn2 (api)Sb2 S5 –323 K

3

12.5694

12.3552

Mn2 (api)Sb2 S5 –373 K

3

12.5883

Mn2 (api)Sb2 S5 –473 K

3

[(Me)2 NH2 ]2 HgSb8 S14

𝜸 (∘ )

V (Å3 )

S. G.

References

90

5471.8

P21

[86]

90

3236.5

Pbca

[87]

90

3217.6

Pbca

[87]

90

90

3215.8

Pbca

[87]

90

90

90

3219.4

Pbca

[87]

20.8407

90

90

90

3236.5

Pbca

[87]

12.3520

20.8164

90

90

90

3236.8

Pbca

[87]

12.6068

12.3619

20.8422

90

90

90

3248.1

Pbca

[87]

1

7.140

27.938

8.102

90

98.04

90

1600.3

P21 /n

[88]

[enH2 ]0.5 HgSbS3

2

8.880

8.976

18.883

90

90

90

1505.2

Cmca

[88]

[tetaH2 ]0.5 HgSbS3

2

9.013

8.856

25.757

90

90

90

2056

Cmca

[88]

[1,2-pdaH]HgSbS3

2

9.476

11.355

19.349

90

90

90

2082.0

Cmca

[88]

[Ni(en)3 ]0.5 HgSbS3

2

19.875

7.768

14.736

90

106.64

90

2179.8

C2/c

[88]

[Mn(phen)]2 HgSb2 S6

1

7.7501

8.6677

11.4926

106.7

92.72

102.94

715.30

P-1

[89]

[tetaH2 ]0.25 Rb0.5 HgSbSe3

3

19.4511

8.6547

10.0138

90

90

90

1685.8

Pnma

[88]

[Mn(tren)]HgSb2 Se5

1

7.5404

10.8268

12.5959

69.40

78.061

86.215

941.73

P-1

[89]

[Fe(tren)]HgSb2 Se5

1

7.5282

10.7335

12.5982

69.49

78.66

86.608

934.80

P-1

[89]

[Co(tren)]HgSb2 Se5

1

7.5076

10.6782

12.5866

69.63

78.97

86.992

928.37

P-1

[89]

[Ni(1,2-pda)3 ]HgSb2 Se5

1

11.2626

11.5370

12.1698

105.0

97.17

118.83

1267.7

P-1

[88]

[Mn(deta)2 ]HgSb2 Se5

1

11.1774

20.3003

20.8304

90

90

90

4726.5

Pbca

[90]

[Ni(en)3 ]Hg2 Sb2 Se6

2

20.4656

7.9377

15.2026

90

106.52

90

2367.8

C2/c

[90]

[Ni(en)(teta)]Hg2 Sb2 Se6

2

15.1754

7.9178

20.6589

90

90

90

2482.3

Pna21

[(Me)2 NH2 ][Hg3 Sb3 Se8 ]

3

11.3895

15.5501

11.2968

90

90

90

2000.8

Abm2

[90] [90]

[Ni(en)3 ][Ga2 Sb2 S7 ]

2

13.0650

9.2588

18.918

90

95.72

90

2277.1

P21 /c

[91]

[(Me)2 NH2 ]2 [Ga2 Sb2 S7 ]

2

9.988

12.597

15.547

90

91.17

90

1955.7

P21 /c

[91]

[(Me)2 NH2 ]2 [Ga2 Sb2 S7 ]⋅H2 O

2

9.9843

12.6207

16.1903

90

90

90

2040.1

P21 21 21

[92]

[(Me)2 NH2 ]2 [Ga2 Sb2 S7 ]⋅H2 O

2

9.1374

10.0229

11.704

90

106.31

90

1028.8

P21

[1]

[enH2 ][Ga2 Sb2 S7 ]en

2

9.524

12.698

16.078

90

91.57

90

1943.8

P21 /c

[93]

[puH]2 [Ga2 Sb2 S7 ]

2

10.057

9.963

12.872

90

112.71

90

1189.8

P21

[93]

[Haep]2 [Ga2 Sb2 S7 ]

2

9.9274

12.9201

21.2254

90

90

90

2722.4

P21 21 21

[94]

[(Et)2 NH2 ]2 [Ga2 Sb2 S7 ]⋅H2 O

2

11.7253

11.7253

34.469

90

90

90

4739

I41 /a

[92]

[maH]4 Ga4 SbS9 S0.28 O0.72 H

3

13.6350

13.6350

13.6350

90

90

90

2534.9

P21 3

[95]

[(Me)2 NH2 ]2 [In2 Sb2 S7 ]

2

19.812

6.9607

16.818

90

123.61

90

1931.6

C2/c

[96]

[(Me)2 NH2 ]2 [In2 Sb2 S4.8 Se2.2 ]

2

20.0561

7.0325

16.8343

90

123.49

90

1980.3

C2/c

[96]

[(Me)2 NH2 ]2 [In2 Sb2 S2.8 Se4.2 ]

2

20.2182

7.1175

16.974

90

123.62

90

2034.1

C2/c

[96]

[(Me)2 NH2 ]2 [In2 Sb2 Se7 ]

2

20.3648

7.1872

17.1677

90

123.55

90

2094.1

C2/c

[96]

[Ni(en)3 ][InSbS4 ]

1

29.3601

29.3601

11.3830

90

90

120

8497.7

R3c

[97]

[maH]4 [In4 SbS9 SH]

3

13.9748

13.9748

13.9748

90

90

90

2729.2

P21 3

[98]

[paH]3 [In3 Sb2 S9 ]

2

12.4629

7.1421

34.2641

90

96.458

90

3030.5

P21 /n

[99]

Fe(en)3 ][In2 Sb2 S7 ]

2

11.2266

17.1624

13.2954

90

112.80

90

2361.6

P21 /n

[99]

[Mg(deta)2 ][In2 Sb2 S7 ]⋅0.5H2 O

2

17.4725

17.4610

18.0777

90

90

90

5515

Aba2

[99]

In2 (N,N-dmen)2 Sb2 S6

0

9.1428

9.5960

13.115

90

102.29

90

1124.3

P21 /c

[99] (continued)

Table 5.4

(Continued)

Formula

D

a (Å)

b (Å)

c (Å)

𝜶 (∘ )

𝜷 (∘ )

𝜸 (∘ )

V (Å3 )

S. G.

References

{[In(chxn)2 ]2 Sb4 S8 }Cl2

0

29.0259

6.7896

24.2023

90

99.524

90

4703.9

C2/c

[100]

[(Me)2 NH2 ]2 [GeSb2 S6 ]

3

10.988

10.988

13.808

90

90

90

1667.0

P41 21 2

[101]

[(Me)2 NH2 ]6 [(Ge2 Sb2 S7 )(Ge4 S10 )]

0

9.334

16.941

17.70

116.3

100.64

98.53

2382

P-1

[102]

[(Me)2 NH2 ][DabcoH]2 [Ge2 Sb3 S10 ]

1

27.827

6.370

9.507

90

99.276

90

1663

C2

[102] [102]

[Ni(en)3 ][GeSb2 S6 ]

2

13.283

17.465

18.481

90

90

90

4287.3

Pbca

[Co(en)3 ][GeSb2 S6 ]

2

13.3233

17.5003

18.5366

90

90

90

4322.0

Pbca

[102]

[(Me)2 NH2 ]2 [GeSb2 S6 ]

3

10.988

10.988

13.808

90

90

90

1667.0

P41 21 2

[102]

[aepH2 ][GeSb2 S6 ]⋅CH3 OH

1

6.7183

18.3065

31.5007

90

90

90

3874.2

Pbca

[103]

[maH]20 Ge10 Sb28 S72 ⋅7H2 O

2

29.2964

29.3261

41.6006

90

100.08

90

35 189

C2/c

[104]

[(CH3 CH2 CH2 )2 NH2 ]3 Ge3 Sb5 S15 ⋅

2

9.7628

15.7590

17.0313

79.868

75.010

81.094

2475.2

P-1

[104]

0

10.801

16.336

14.311

90

99.107

90

2493.3

P21 /n

[105]

0.5(EtOH) [La(en)4 SbSnS5 ]2 ⋅0.5H2 O

phen, 1,10-phenanthroline, api, N-(3-aminopropyl)-imidazole, teta, triethylenetetramine, 1,2-pda, 1,2-diaminopropane, tren, tris(2-aminoethyl)amine, deta, diethylenetriamine, maH, methylamine, [(Me)2 NH2 ], dimethylamine, [(Et)2 NH2 ], diethylamine, pu, propyleneurea, aep, N-(2-aminoethyl)piperazine, pa, n-propylamine, [(CH3 CH2 CH2 )2 NH], dipropylamine, N,N-dmen, N,N-dimethylethylenediamine, Dabco, triethylenediamine, deta, diethylenetriamine, pda, 1,3-propane diamine.

5.5 Chalcogenidoantimonates

SBUs

TBUs

Dinuclear {HgSbS5} {HgSbS4}n–1 {MSb2Q7} (M = Zn, Cd, Hg) {MSbS4}n–2 (M = Cd, Hg)

Trinuclear {M2SbS8}–1 (M = Cd, Hg)

{HgSbQ4}n–3 {Hg2Sb2Se8}–2

Tetranuclear

{Hg2Sb2Q10}

{Hg4Se12}

{Hg2SbSe6}n

{Hg3Se8}n

Figure 5.28 Representations of some ternary M-Sb-Q (M = Zn, Cd, and Hg; Q = S and Se) or binary Hg–Se SBUs and TBUs. Color code: M (turquoise), Sb (pink), and Q (yellow). Note: the labels “1,” “2,” and “3” after the formulae represent different types of architectures with the same M:Sb:Q ratio. Source: Wang et al. [8]. Copyright 2016, Elsevier B.V.

combined together with two {Sb2 S5 } groups to give a {Hg3 Sb4 S14 } cluster as SBU; the SBUs were further connected into an {Hg3 Sb4 S12 }n ribbon-like TBU. Therefore, the 2D-[Hg3 Sb4 S10 ]n 2n− layer is constructed by interlinking {Hg3 Sb4 S12 }n ribbons in an ABAB mode (Figure 5.29c). To date, the {HgQ3 } trigonal planar geometry was relatively rarely observed in chalcogenidometalates, except for a few examples such as [Ni(1,2-pda)3 ]2 [HgSb3 S7 ]Cl [111], KHgSbS3 [112], (Ph4 P)2 [Hg2 As4 S9 ] [113], [(Me)4 N][HgAsSe3 ] [81], [(Et)4 N][HgAsSe3 ] (Et = ethyl) [81], and [Ph4 P]2 [Hg2 As4 Se11 ] [81]. In [Ni(1,2-pda)3 ]2 [HgSb3 S7 ]Cl, {HgS3 } and {Sb3 S7 } are interconnected into a 1D-[HgSb3 S7 ]n 3n− zig-zag chain with a pitch of 20.514 Å (Figure 5.29d). Hg—S bond lengths vary in the range of 2.345(3)–2.638(3) Å, in agreement with the distortion of the {HgS3 } from a regular trigonal planar environment. The isomeric [MSb2 Q5 ]n 2n− moieties (Figure 5.30) includes three types: (a) 1D-[MSb2 Q5 ]n 2n− (M = Zn and Cd; Q = Se; M = Hg, Q = S and Se) (HgSb2 Q5 -I)

241

242

5 Group 11–15 Metal Chalcogenides

{HgS2}

{HgS2}

{HgS4}

1D-[HgSb8S14]n2n– (a)

(c) 2D-[Hg3Sb4S10]n2n– {HgS2} 20.514 Å

(b) 1D-{[Mn(phen)]2HgSb2S6}n

(d)

{HgS3}

1D-[HgSb3S7]n3n–

Figure 5.29 View of (a) 1D-[HgSb8 S14 ]n 2n− ; (b) 1D-{[Mn(phen)]2 HgSb2 S6 }n ; (c) 2D-[Hg3 Sb4 S10 ]n 2n− ; and (d) 1D-[HgSb3 S7 ]n 3n− . Color code: Hg (turquoise), Sb (pink), S (yellow), N (blue), and C (gray). Color-highlighted moieties: for (a), {Sb4 S7 }n ribbon (dark red), for (b), {[Mn(phen)]2 Sb2 S6 } cluster (red), for (c), {Hg3 Sb4 S12 }n ribbon (red), and for (d), {Sb3 S7 } cluster (dark red). Hydrogen atoms are omitted for clarity. Source: Wang et al. [8]. Copyright 2016, Elsevier B.V.

in compounds [TM(deta)2 ][MSb2 Se5 ] (TM = Mn, M = Zn, Cd, Hg; TM = Co, M = Cd) [90], [TM(deta)2 ][HgSb2 S5 ] (TM = Mn, Co, and Ni) [111], and [Ni(1,2-dap)3 ][HgSb2 Se5 ] [90]; (b) 1D-[HgSb2 Q5 ]n 2n− (Q = S and Se) (HgSb2 Q5 -II) in quaternary inorganic–organic hybrids [TM(tren)]HgSb2 Q5 (TM = Mn, Co, and Ni; Q = S; TM = Mn, Co, and Ni, Q = Se, tren = tris(2-aminoethyl)amine) [89, 111]; (c) 2D-[HgSb2 S5 ]n 2n− (HgSb2 Q5 -III) in compound [Ni(1,2-dap)3 ][HgSb2 S5 ] [111]. Differences between these isomers can be clarified from the following three aspects. (i) SBUs or TBUs are different. HgSb2 Q5 -I is made of trinuclear {MSb2 Q7 } (M = Zn, Cd, and Hg) SBUs, where each pair of adjacent units is interconnected in a centrosymmetric fashion through two Q atoms. HgSb2 Q5 -II holds the same dimensionality and charge as HgSb2 Q5 -I but features special “terminal” Q atoms (Q1) that are further bonded to unsaturated [TM(tren)]2+ complexes. A centrosymmetric “S”-shaped {[TM(tren)]2 Hg2 Sb4 Q12 } SBU was identified for these hybrid compounds. Whereas the HgSb2 Q5 -III shows a 2D anionic network that can be viewed as the interconnection of {HgSbS4 }n -1 chain-like TBUs (composed of {HgSbS5 } SBUs) through trans-{Sb2 S4 } dimeric bridges. (ii) Connection modes are different. In HgSb2 Q5 -I, the {MQ4 } and ψ-{SbQ3 } polyhedra are solely connected via corner-sharing mode, while both corner-sharing and edge-sharing modes are present in HgSb2 Q5 -II and HgSb2 Q5 -III. The presence of edge-sharing connections in HgSb2 Q5 -III enhances the multiplicity of the interconnection between two neighboring metal centers and simultaneously decreases the overall

5.5 Chalcogenidoantimonates

(c)

1D-[MSb2Q5]n2n– (M = Zn, Cd, Q = Se; M = Hg, Q = S, Se) (HgSb2Q5-I

(a) Q1

Q2

2D-[HgSb2S5]n2n– (HgSb2Q5-III) 1D-{[TM(tren)]HgSb2Q5]n (TM = Mn, Co, Ni; Q = S, Se) (with 1D-[HgSb2Q5]n2n– (HgSb2Q5-II) backbone)

(b)

Figure 5.30 View of 1D-[MSb2 Q5 ]n 2n− (M = Zn, Cd, Q = Se; M = Hg, Q = S, Se) (HgSb2 Q5 -I) (a), 1D-{[TM(tren)]HgSb2 Q5 }n (TM = Mn, Co, Ni; Q = S, Se) (with 1D-[HgSb2 Q5 ]n 2n− (HgSb2 Q5 -II) backbone) (b) and 2D-[HgSb2 S5 ]n 2n− (HgSb2 Q5 -III) (c). Color code: M (turquoise), Sb (pink), TM (green), Q (yellow), N (blue), and C (gray). Color-highlighted moieties: for (a), {MSb2 Q7 } cluster (red); for (b), {[TM(tren)]2 Hg2 Sb4 Q12 } cluster (red); and for (c), {HgSbS4 }n -1 chain (dark red) and trans-{Sb2 S4 } dimer (blue). Hydrogen atoms are omitted for clarity. Source: Wang et al. [8]. Copyright 2016, Elsevier B.V.

number of connected metal pairs, resulting in a greater openness of the network to some degree. Similarly, the anionic network of [Ni(deta)2 ]1.5 [In3 Sb2 S9 ]⋅H2 O [114] possesses larger windows than that in [paH]3 [In3 Sb2 S9 ] [99]. (3) μ3 -Q atoms are present. As for HgSb2 Q5 -II, it requires the occurrence of equivalent μ3 -Q atoms (Q2) to the “terminal” Q (Q1) atoms as complementary species to remain the overall balanced composition (Hg:Sb:Q = 1 : 2 : 5) similar to that of HgSb2 Q5 -I (Figure 5.30b). [MSbQ3 ]n n− moieties also present the isomerism. They show as many as four isomers (Figures 5.31 and 5.32a) including: (i) 2D-[MSbS3 ]n n− (M = Cd, Hg) (HgSbQ3 -I) in compounds [1,4-dabH2 ]0.5 [CdSbS3 ] [115], [enH2 ]0.5 [HgSbS3 ] [88], and [tetaH2 ]0.5 [HgSbS3 ] [88]; (ii) 2D-[HgSbS3 ]n n− (HgSbQ3 -II) in compound [1,2-dapH][HgSbS3 ] [88]; (iii) 2D-[HgSbQ3 ]n n− (HgSbQ3 -III) in compounds [TM(en)3 ]0.5 [HgSbQ3 ] (TM = Co, Ni, Q = S; TM = Ni, Q = Se) [88, 90, 111] and [Ni(en)(teta)]0.5 [HgSbSe3 ] [90]; and (iv) 3D-[HgSbSe3 ]n n− (HgSbQ3 -IV) in compound [tetaH2 ]0.25 Rb0.5 [HgSbSe3 ] [88]. The three isomers feature layered structures based on different ribbon-like TBUs. HgSbQ3 -I shows a condensed grid network constructed by interlinking parallel {MSbS4 }n -2 (M = Cd and Hg) ribbons, whereas HgSbQ3 -II and HgSbQ3 -III display the interconnections of {HgSbQ4 }n 3− (Q = S and Se) ribbons in parallel and antiparallel fashions, respectively. In the three types of layers, the ribbon-like TBUs are linked by the bridging Q atoms, accompanied by the formation of closely packed 8-MRs of {M 2 Sb2 Q4 } in the interspace. The overall numerical ratio of μ3 -Q:μ2 -Q of all the three types of layers is 1 : 2. Thus, the structural discrepancies can be further revealed by their distinct TBUs. Obviously, {M 2 SbS8 }-1 SBUs interlinked by sharing

243

244

5 Group 11–15 Metal Chalcogenides

{HgSbQ4}n-3 ribbons

2D-[HgSbS3]nn– (HgSbQ3-II)

2D-[MSbS3]nn– (M = Cd, Hg)

2D-[HgSbS3]nn– (Q = S, Se)

(HgSbQ3-I)

(a)

(HgSbQ3-III)

(b)

(c)

Figure 5.31 View of (a) 2D-[MSbS3 ]n n− (M = Cd, Hg) (HgSbQ3 -I), (b) 2D-[HgSbS3 ]n n− (HgSbQ3 -II) and (c) 2D-[HgSbQ3 ]n n− (Q = S, Se) (HgSbQ3 -III). Color code: M (turquoise), Sb (pink), Q (yellow), N (blue), and C (gray). Color-highlighted moieties: for (a), {MSbS4 }n -2 ribbon (red); for (b, c), {HgSbQ4 }n -3 ribbon (red and blue). Source: Wang et al. [8]. Copyright 2016, Elsevier B.V.

3D-[HgSbSe3]nn– (HgSbQ3-IV)

3D-[Hg3Sb3Se8]nn–

+ {Hg4Sb2Se12}n (a)

+ {SbSe3}

{Hg6Se16}n

+

{Sb2Se5} {SbSe3}

(b)

Figure 5.32 View of 3D-[HgSbSe3 ]n n− (HgSbQ3 -IV) (a), 3D-[Hg3 Sb3 Se8 ]n n− (b), and the combinations of their BUs. Color code: Hg (turquoise), Sb (pink), and Se (gold). Color-highlighted moieties: for (a), {Hg2 SbSe6 }n ribbon (red), ψ-{SbSe3 } (blue); for (b), {Hg3 Se8 }n ribbon (red), {Sb2 Se5 } dimer (blue), and ψ-{SbSe3 } (green). Source: Wang et al. [8]. Copyright 2016, Elsevier B.V.

the M vertices form a {MSbS4 }n -2 ribbon in HgSbQ3 -I. The infinite arrangement of 6-MRs of {M 2 SbS3 } is formed by interconnecting {MS4 } and ψ-{SbS3 } polyhedra via corner-sharing mode. On the contrary, the interconnection of {Hg2 Sb2 Q10 } SBUs by sharing the edge of two adjacent {HgQ4 } polyhedra forms the {HgSbQ4 }n -3 ribbon in HgSbQ3 -II and HgSbQ3 -III, resulting in the formation of 4-MRs of {Hg2 Q4 } equivalent to the larger 8-MRs of {Hg2 Sb2 Q4 }.

5.5 Chalcogenidoantimonates

In fact, the numerical relationship of Q atoms and n-MRs in one M-Sb-Q layer (each M/Sb—Q bond is shared exclusively by two n-MRs [n ≥ 4]) can be revealed in Eq. (5.1): ∑ ∑ (p × q) = (n × m)∕2 (5.1) in which q denotes the number of bridging 𝜇p -Q (p ≥ 2) ligands that connect M/Sb metal ions, while m represents the total number of n-MR in the network. For example, HgSbQ3 -I contains equivalent 6-MR of {M 2 SbS3 } and 8-MR of {M 2 Sb2 S4 }, whereas HgSbQ3 -II includes the distributed 4-MR of {Hg2 S2 } and 8-MR of {Hg2 Sb2 S4 } with a ratio of 1 : 3. Based on Eq. (5.1), the numerical ratio of 6-MR (HgSbQ3 -I):8-MR (HgSbQ3 -I):4-MR (HgSbQ3 -II):8-MR (HgSbQ3 -II) can be easily determined to be 2 : 2 : 1 : 3 within a specific elemental composition. The chalcogenide layered structures containing alkali metal cations with the Hg:Sb:Q ratio of 1 : 1 : 3 composition have also been reported, such as in KHgSbS3 [112], KHgSbSe3 [116], AHgSbSe3 (A = Rb and Cs) [116, 117], and RbHgSbTe3 [118]. Some of these layers show distinct structures from those mentioned above, and a systematic comparison has been made before [88]. It is worth nothing that the chain−/ribbon-like TBUs and discrete {Sbx Qy } PBUs/SBUs can be combined in a “line + node” mode, often resulting in frameworks with large windows or channels. For instance, the aforementioned 2D-[HgSb2 S5 ]n 2n− (HgSb2 Q5 -III) (Figure 5.30c) [111] features a large 28-MR of {Hg6 Sb8 S14 } with a cross section of approximate 13.29 × 9.85 Å2 , which is formed by interconnecting {HgSbS4 }n -1 chains with trans-{Sb2 S4 } dimers. Figure 5.32 shows two more 3D framework structures constructed by such a “line + node” combination. One is 3D-[HgSbSe3 ]n n− (HgSbQ3 -IV) (Figure 5.32a) framework in compound [tetaH2 ]0.25 Rb0.5 [HgSbSe3 ] [88] constructed by interconnecting parallel {Hg2 SbSe6 }n ribbons through 𝜓-{SbSe3 } groups. Mixed Rb+ /[tetaH2 ]2+ cations are located in the channels with a cross section of approximate 15.59 × 8.65 Å2 . The other is 3D-[Hg3 Sb3 Se8 ]n n− framework (Figure 5.32b) in [(Me)2 NH2 ][Hg3 Sb3 Se8 ] [90] constructed by interlinking {Hg3 Se8 }n ribbons by {Sb2 Se5 } and 𝜓-{SbSe3 } groups. In this framework, ordered channels are observed along the direction parallel to the {Hg3 Se8 }n ribbons, in which [(Me)2 NH2 ]+ are present as counter cations. Although 1D-[HgSb8 S14 ]n 2n− [88] and 3D-[Hg3 Sb3 Se8 ]n n− are both directed by the [(Me)2 NH2 ]+ ions, they are very different in composition and structure. This is probably due to the different synthetic conditions coupled with some potential distinctions between S and Se in their chemical nature, such as atomic radius and electronegativity.

5.5.3 Chalcogenidometalates Containing Group 13(III) Ions and Antimony(III) Heterometallic chalcogenidometalates including group 13(III) metal ions (Ga3+ and In3+ ) and Sb(III) have been also reported. Similar to the group 12(II) metal ions,

245

246

5 Group 11–15 Metal Chalcogenides

Ga3+ and In3+ ions also prefer adopting tetrahedral coordination with chalcogen atoms forming {MQ4 } (M = Ga, Q = S; M = In, Q = S, Se) units. Then, {MQ4 } can be further combined with 𝜓-{SbQx } (Q = S, Se; x = 3, 4) into various SBUs and TBUs (Figure 5.33). Moreover, the larger ionic radius of In3+ comparing to that of Ga3+ facilitates the chelation by polyamines; thus, the coordination number for In3+ can increase to six. 5.5.3.1 Ga–Sb–S Compounds

The first two Ga–Sb–S compounds were reported in 2009, namely [(Me)2 NH2 ]2 [Ga2 Sb2 S7 ] and [Ni(en)3 ][Ga2 Sb2 S7 ] [91], whose inorganic moieties feature an identical 2D-[Ga2 Sb2 S7 ]n 2n− network (Figure 5.34). Ga3+ ion exhibits a tetrahedral coordination with Ga–S distances ranging from c. 2.25 to 2.31 Å. Two {GaS4 } and two 𝜓-{SbS3 } units are interconnected by corner-sharing mode to form tetranuclear clusters of {Ga2 Sb2 S9 } as SBUs for the anionic inorganic layer. This {Ga2 Sb2 S9 } SBUs

Trinuclear

{In2SbS8}–1

TBUs

{In4Sb4S18}

{In2SbS8}–2 {In2SbS6}n

Tetranuclear

{In2SbQ10} {InSbQ6}n {M2Sb2S9} (M = Ga, In) {InSb4S6}n

Pentanuclear

{In3Sb2S12}

{In3Sb4S14}n

Figure 5.33 View of some typical M–Sb–Q (M = Ga, In; Q = S, Se) SBUs and TBUs. Color code: M (bright green), Sb (pink), and Q (yellow). Note: the labels “1” and “2” after the formulae represent different types of architectures with the same M:Sb:Q ratio. Source: Wang et al. [8]. Copyright 2016, Elsevier B.V.

5.5 Chalcogenidoantimonates

Figure 5.34 View of the 2D-[Ga2 Sb2 S7 ]n 2n− in [(Me)2 NH2 ]2 [Ga2 Sb2 S7 ] and [Ni(en)3 ][Ga2 Sb2 S7 ]. Color code: Ga (teal), Sb (pink), and S (yellow). Color-highlighted moieties: {Ga2 Sb2 S9 } cluster (red). Source: Feng et al. [91]. Copyright 2009, the American Chemical Society.

2D-[Ga2Sb2S7]n2n–

cluster resembles the {Ga4 Q9 (en)2 } cluster in the [Ga4 Q7 (en)2 ]2− (Q = S and Se) layers [119, 120], merely by replacing the two {Ga(en)} groups with two Sb3+ ions. Subsequently, such a 2D-[Ga2 Sb2 S7 ]n 2n− layer has also been isolated in [enH2 ][Ga2 Sb2 S7 ]⋅en [93], [puH]2 [Ga2 Sb2 S7 ] [121], [aepH]2 [Ga2 Sb2 S7 ] [94], and [(Me)2 NH2 ]2 [Ga2 Sb2 S7 ]⋅H2 O [1]. Among them, [(Me)2 NH2 ]2 [Ga2 Sb2 S7 ]⋅H2 O (space group: P21 ) exhibits an interesting flytrap-like response of the windows on the 2D-[Ga2 Sb2 S7 ]n 2n− layer to the uptake of Cs+ ions. In 2018, Huang’s group reported two new gallium thioantimonates, [(Me)2 NH2 ]2 [Ga2 Sb2 S7 ]⋅H2 O (FJSM-GAS-1, space group: P21 21 21 ) and [(Et)2 NH2 ]2 [Ga2 Sb2 S7 ]⋅H2 O (FJSM-GAS-2, space group: I41 /a), Figure 5.35 [92]. All the above gallium thioantimonates present similar inorganic moieties of 2D-[Ga2 Sb2 S7 ]n 2n− . However, they crystallize in different space groups, and the conformations of layers differ from each other as a result of the differences in the interactions between layers and organic cations; especially, FJSM-GAS-2 has a square grid-like layer other than an oval grid-like layer present in other gallium thioantimonates [1, 91, 93, 94]. Numerous open frameworks with diversified topologies are constructed by interconnecting supertetrahedral Tn clusters [12]. But it is still a challenge for the incorporation of such Tn clusters with 𝜓-{SbQx } (x = 3 and 4). [CH3 NH3 ]4 Ga4 SbS9 S0.28 O0.72 H [95] represents a limited example of structures of this type, which is isostructural to the ammonium-directed compounds reported by Kanatzidis in 2009 [122]. It features a 3D anionic open framework of [Ga4 SbS9 QH]n 4n− (Q = S/O) with intersected channels parallel to the a-, b-, and c-axes (Figure 5.36). Vertex-sharing of three [Ga(1)S4 ] and one [Ga(2)S3 (QH)] units produces an adamantane-like [Ga4 S9 (QH)]7− T2 cluster (Figure 5.36a). Each [Ga4 S9 (QH)]7− T2 cluster is covalently bound to three Sb atoms via S(1) atoms (Figure 5.36b), and the Sb atoms bridge three separate T2 clusters via corner-sharing S(1) atoms, creating a threefold rotation axis around Sb. The interconnection of [Ga4 S9 (QH)]7− T2 clusters and {SbS3 } units results in a 3D open framework of [Ga4 SbS9 QH]n 4n– , in which the charge balancing [CH3 NH3 ]+ cations

247

248

5 Group 11–15 Metal Chalcogenides

50 μm

50 μm

(a)

(b) a

a

b

c b

c Sb Ga S

Sb Ga S

(c)

(d) c

c a

a b

Sb Ga S N C O

Sb Ga S N C O

(e)

b

(f)

Figure 5.35 SEM images of a FJSM-GAS-1 (a) and a FJSM-GAS-2 crystal (b). 2D grid-like layers of [Ga2 Sb2 S7 ]n 2n− along the ab plane in FJSM-GAS-1 (c) and FJSM-GAS-2 (d). Packing of layers in FJSM-GAS-1 (e) and FJSM-GAS-2 (f) in a perspective view along the b-axis; hydrogen atoms are omitted for clarity. Source: Reproduced with permission Feng et al. [92]. Copyright 2018, the American Chemical Society.

reside (Figure 5.36c). [CH3 NH3 ]4 Ga4 SbS9 S0.28 O0.72 H exhibited interesting Ni2+ ion exchange properties. 5.5.3.2 In–Sb–Q (Q = S and Se) Compounds

To date, many In–Sb–Q (Q = S and Se) compounds have been isolated, those of which reported by Huang’s group are listed in Table 5.4. Let’s start with the comparison of the structures of two 1D anionic In–Sb–S moieties constructed from trinuclear {In2 SbS8 } as SBUs, i.e. 1D-[In2 Sb4 S11 ]n 4n− ribbon (Figure 5.37a) in [TM(deta)2 ]2 [In2 Sb4 S11 ] (TM = Co and Ni) [114, 123] and 1D-[InSbS4 ]n 2n− (InSbS4 -I) ribbon (Figure 5.37b) in [Ni(en)3 ][InSbS4 ] [97]. The ribbons exhibit stereoisomerism of {In2 SbS8 } SBUs, in which the conformational difference lies in the “inwards” (noted as {In2 SbS8 }-1) or “outwards” (noted as {In2 SbS8 }-2) orientation of the terminal S atom in the 𝜓-{SbS3 }. Equivalent {In2 SbS8 }-1 and {Sb3 S7 } units are alternating connected into a {In2 Sb4 S13 }n chain. Such chains are further dimerized in an antiparallel fashion to give a 1D-[In2 Sb4 S11 ]n 4n− ribbon along with the alternating arrangements of equivalent 8-MR of {In2 Sb2 S4 } and 10-MR of {In2 Sb3 S5 }.

5.5 Chalcogenidoantimonates

(a)

(b)

a

c

Sb Ga S O/S T2 cluster (c)

(d)

Figure 5.36 (a) The {Ga4 S9 (QH)} T2 cluster built up of three [Ga(1)S4 ] and one [Ga(2)S3 (QH)] tetrahedra; (b) the interconnection modes of {Ga4 S9 (QH)} T2 clusters and {SbS3 } units in [CH3 NH3 ]4 Ga4 SbS9 S0.28 O0.72 H; (c) perspective view of the open framework structure of [CH3 NH3 ]4 Ga4 SbS9 S0.28 O0.72 H along the b-axis; (d) the tiling of [CH3 NH3 ]4 Ga4 SbS9 S0.28 O0.72 H. Source: Zhang et al. [95]. Copyright 2018, Wiley-VCH.

By comparison, 1D-[InSbS4 ]n 2n− ribbon presents a novel left-handed helix made of {In2 SbS8 }-2 clusters interlinked by sharing In3+ ions; charge balancing [Ni(en)3 ]2+ complexes with Λ configuration are found between these helical 1D-[InSbS4 ]n 2n− ribbons. The crystallization of [Ni(en)3 ][InSbS4 ] in the non-centrosymmetric space group R3c probably renders this material a ferroelectric property. Compared with 1D ribbon-like structures, the In–Sb–Q (Q = S and Se) compounds with a 2D layered structure are more popular. A noticeable feature is the common presence of “node + node” or “line + line” combination in terms of the BU dimensionality in these structures. The typical structural types of 2D In–Sb–Q compounds are exemplified by two categories based on their composition, that is, 2D-[In2 Sb2 Q7 ]n 2n− (Q = S, Se or mixed S/Se) and 2D-[In3 Sb2 S9 ]n 3n , where interesting isomerism is present. As shown in Figure 5.38, 2D-[In2 Sb2 Q7 ]n 2n− (Q = S, Se or mixed S/Se) layers are constructed from tetranuclear {In2 Sb2 Q10 }/{In2 Sb2 S9 } SBUs. Oligomeric/polymeric derivatives of the {In2 Sb2 Q10 } clusters as TBUs can be further found. Huang’s group reported two isomorphic indium chalcogenidoantimonates(III) and their quaternary solid solution analogues [(Me)2 NH2 ]2 [In2 Sb2 S7-x Sex ] (x = 0, 2.20, 4.20,

249

250

5 Group 11–15 Metal Chalcogenides

“Inward” S atom

(a)

1D-[In2Sb4S11]n4n–

“Outward” S atom 1D-[InSbS4]n2n– (InSbS4-I) (b)

Figure 5.37 View of (a) 1D-[In2 Sb4 S11 ]n 4n− and (b) 1D-[InSbS4 ]n 2n− (InSbS4 -I). Color code: In (bright green), Sb (pink), and S (yellow). Color-highlighted moieties: for (a), {InSb2 S8 }-1 cluster (red); for (b), {InSb2 S8 }-2 cluster (red). Source: Wang et al. [8]. Copyright 2016, Elsevier B.V.

2D-[In2Sb2S7–xSex]n2n– (In2Sb2Q7-I)

(a)

2D-[In2Sb2S7]n2n– (In2Sb2Q7-II)

(b)

2D-[In2Sb2S7]n2n– (In2Sb2Q7-III)

(c)

Figure 5.38 View of (a) 2D-[In2 Sb2 S7-x Sex ]n 2n− (In2 Sb2 Q7 -I), (b) 2D-[In2 Sb2 S7 ]n 2n− (In2 Sb2 Q7 -II), and (c) 2D-[In2 Sb2 S7 ]n 2n− (In2 Sb2 Q7 -III). Color code: In (bright green), Sb (pink), and Q (yellow). Color-highlighted moieties: for (a), {InSbQ4 }n ribbon (red); for (b), {In2 Sb2 S9 } cluster (red); and for (c), {In4 Sb4 S18 } cluster (red). Source: Wang et al. [8]. Copyright 2016, Elsevier B.V.

and 7) [96], which feature 2D-[In2 Sb2 S7−x Sex ]n 2n− (In2 Sb2 Q7 -I) layers (Figure 5.38a) based on an interconnection of {InSbQ4 }n ribbon-like TBUs in antiparallel fashion by sharing μ2 -Q atoms. Each {InSbQ4 }n ribbon contains tetranuclear {In2 Sb2 Q10 } clusters formed by alternatingly connecting two {InQ4 } and two 𝜓-{SbQ3 } units in a centrosymmetric fashion. Such a {In2 Sb2 Q10 } cluster is structurally similar to other tetranuclear units in [Mn(tren)]4 Mn2 Sb4 S12 [94], [Fe(tren)]FeSbS4 [124], [La(en)4 ]2 Sn2 Sb2 S10 ⋅0.5H2 O [105], and [Mn(tren)]InAsS4 [125]. Although the

5.5 Chalcogenidoantimonates

{In2 Sb2 Q10 } cluster and the {InSbQ4 }n ribbon have never been isolated in bulk material, they are general building fragments for the In-Sb-Q structures [99, 126, 127]. [Mg(deta)2 ][In2 Sb2 S7 ]⋅0.5H2 O features an anionic 2D-[In2 Sb2 S7 ]n 2n− (In2 Sb2 Q7 -II) layer (Figure 5.38b) with alkaline earth MACs [Mg(deta)2 ]2+ as the counter cation [99], which is analogous to the aforementioned 2D-[Ga2 Sb2 S7 ]n 2n− [1, 91, 93, 94]. The {In2 Sb2 S9 } SBU in In2 Sb2 Q7 -II is isostructural with the {Ga2 Sb2 S9 } cluster in [(Et)2 NH2 ]2 [Ga2 Sb2 S7 ]⋅H2 O. In 2D-[In2 Sb2 S7 ]n 2n− (In2 Sb2 Q7 -II) layer, there are rectangular 16-MRs of {In4 Sb4 S8 } with a dimension of 6.89 × 7.02 Å2 . Compounds [TM(en)3 ][In2 Sb2 S7 ] (TM = Fe and Ni) [99, 126] present the third type of 2D-[In2 Sb2 S7 ]n 2n− (In2 Sb2 Q7 -III) layer (Figure 5.38c) with a novel {In4 Sb4 S18 } TBU based on a trimeric assembly of {In2 Sb2 S10 } SBUs by sharing two In–S–Sb edges. Each {In4 Sb4 S18 } shares eight S vertices with four adjacent ones in nearly perpendicular orientations, finally generating the 2D-[In2 Sb2 S7 ]n 2n− (In2 Sb2 Q7 -III). Further investigation of the structures of In2 Sb2 Q7 -I, In2 Sb2 Q7 -II, and In2 Sb2 Q7 -III reveals that the isomerism might be due to the varied interlinkage modes of tri-coordinated Sb3+ ions and dimeric {In2 Q7 } bridging groups units with various geometries (Figure 5.39). (i) In In2 Sb2 Q7 -I and In2 Sb2 Q7 -II, the {In2 Q7 } groups in a regular fashion with C2 symmetry are observed and a twofold rotation axis goes through the bridging Q atom, whereas the {In2 S7 } group with a slight distortion in In2 Sb2 Q7 -III reduces the symmetry to C1 . (ii) two {In2 Q7 } groups C2 In2Sb2Q7-I

(a)

(Mode 1)×2

(Mode 2)×2 Principal axis C2

In2Sb2Q7-II (Mode 1)

(Mode 3)

(b) In2Sb2Q7-III

(Mode 1)×2 (Mode 1) + (Mode 2)

(Mode 2)×2

(c)

Figure 5.39 View of different linkage modes between two adjacent Sb3+ ions in (a) In2 Sb2 Q7 -I, (b) In2 Sb2 Q7 -II, and (c) In2 Sb2 Q7 -III. Color code: In (bright green), Sb (pink), and Q (yellow). Source: Wang et al. [8]. Copyright 2016, Elsevier B.V.

251

252

5 Group 11–15 Metal Chalcogenides

are connected by a pair of adjacent Sb3+ ions in In2 Sb2 Q7 -I and In2 Sb2 Q7 -III, where the two bridging Q atoms in each group are linked to either the same In3+ ion (mode 1) or different In3+ ions (mode 2). Different combinations of mode 1 and mode 2 lead to the distinction between these two layered structures. By contrary, the case for In2 Sb2 Q7 -II is different. For one, in mode 1, a single {In2 S7 } group only bridges a pair of adjacent Sb3+ ions along the larger 16-MR of {In4 Sb4 S8 }. For another, the adjacent Sb3+ ions in one {In2 Sb2 S9 } unit are bridged by one {In2 S7 } group via four S atoms (mode 3). Compounds [paH]3 [In3 Sb2 S9 ] [99] and [Ni(deta)2 ]1.5 [In3 Sb2 S9 ]⋅H2 O [114] show two isomeric 2D-[In3 Sb2 S9 ]n 3n− layers (abbreviated as In3 Sb2 S9 -I and In3 Sb2 S9 -II). In3 Sb2 S9 -I (Figure 5.40a) displays the alternating interconnection of {InSbS4 }n and {In2 SbS6 }n ribbon-like TBUs, whereas the puckered In3 Sb2 S9 -II layer contains pentanuclear {In3 Sb2 S12 } SBUs. In3 Sb2 S9 -I can be viewed as a reorganization of the known 2D-[In2 Sb2 S7 ]n 2n− (In2 Sb2 Q7 -I) layer by inserting a novel {In2 SbS6 }n ribbon based on the trinuclear {In2 SbS8 }-2 clusters into each pair of adjacent {InSbS4 }n ribbons. The original C2 symmetric characteristic of the 2D-[In2 Sb2 S7 ]n 2n− (In2 Sb2 Q7 -I) has been not found in In3 Sb2 S9 -I due to the lower symmetry of the inserted {In2 SbS6 }n ribbon. In the case of In3 Sb2 S9 -II (Figure 5.40b), a novel {In6 S18 } moiety is evidenced, which is constructed from two {In2 S7 } units and one {In2 S6 } unit built up by corner- and edge-sharing two {InS4 } tetrahedra, respectively. Noticeably, it is very rare that {InS4 } tetrahedra form oligomers via a mixed corner−/edge-sharing mode [114], as the corner- and edge-sharing linkages of {InS4 } tetrahedra alone tend to form supertetrahedral Tn clusters [12] and infinite [InS2 ]n n− chains [128], respectively. In addition to the “node + node” or “line + line” combination, the “line + node” combination of BUs also contributes to the structural variety of the In–Sb–Q system as demonstrated by the following three examples. Compounds [M(1,2-dap)3 ] [InSb3 S7 ] (M = Co and Ni) [129] show an anionic 3D-[InSb3 S7 ]n 2n− framework

2D-[In3Sb2S9]n3n– (In3Sb2S9-I) (a)

2D-[In3Sb2S9]n3n– (In3Sb2S9-II) (b)

Figure 5.40 View of (a) 2D-[In3 Sb2 S9 ]n 3n− (In3 Sb2 S9 -I) and (b) 2D-[In3 Sb2 S9 ]n 3n− (In3 Sb2 S9 -II). Color code: In (bright green), Sb (pink), and S (yellow). Color-highlighted moieties: for (a), {InSbS4 }n ribbon (red), {In2 SbS6 }n ribbon (blue); for (b) {In3 Sb2 S12 } cluster (red) and {In6 S18 } moiety (dark red). Source: Wang et al. [8]. Copyright 2016, Elsevier B.V.

5.5 Chalcogenidoantimonates

constructed by interconnecting {InSb2 S6 }n ribbon-like TBUs through mononuclear 𝜓-{SbS4 } “nodes” (Figure 5.41a). The {InSb2 S6 }n TBU is made of {In2 Sb2 S10 } SBUs linked by sharing In3+ vertices. 8-, 12-, and 16-MR intersected channels are present in the framework. Compound [dpaH]5 [In5 Sb6 S19 ]⋅1.45H2 O features an anionic 2D-[In5 Sb6 S19 ]n 5n− layer constructed by {In2 Sb2 S9 } clusters as “nodes” interconnecting {In3 Sb4 S14 }n ribbons that are made of {In3 Sb2 S12 } and {In2 Sb2 S10 } clusters with shared In3+ vertices (Figure 5.41b) [130]. In this layer, a large elliptical 24-MR of {In6 Sb6 S12 } with dimensions of 16.93 × 7.37 Å2 is observed. In addition, the {In2 Sb2 S9 } “nodes” connect the {InSbS4 }n ribbons into a 3D-[In6 Sb6 S21 ]n 6n− (In2 Sb2 Q7 -IV) framework with 20-, 28-, and 40-MR intersected channels in [Co(en)3 ]3 (en)[In6 Sb6 S21 ]⋅H2 O [127] and [Ni(en)3 ]3 (en)[In6 Sb6 S21 ] [123] (Figure 5.41c). [maH]4 [In4 SbS9 SH] [98] is isostructural to its Ga-Sb-S analogues [95]. Its 3D framework of [In4 SbS9 SH]n 4n− is constructed by the alternating connection of equivalent 𝜓-{SbS3 } pyramids and T2-{In4 S9 SH} clusters. Considering both 𝜓-{SbS3 } and T2-{In4 S9 SH} as 3-connected node, an overall 103 topology of the 3D framework is evidenced. Organic amine normally enter the In-Sb-Q structures as protonated cations, while occasionally they can function as neutral coordinating ligands bonding to In3+ ions of the In-Sb-Q moiety, as exemplified in {[In(chxn)2 ]2 Sb4 S8 }Cl2 [100] (chxn = trans-1,2-diaminocyclohexane) and [In(N,N-dmen)]2 Sb2 S6 (N,N-dmen = N,N-dimethylethylenediamine) [99], Figure 5.42. In both compounds, In3+ ion adopts a distorted octahedral geometry hexa-coordinated mixed S/N atoms. In—S bond length (∼2.50–2.70 Å) is obviously longer than that in {InS4 } tetrahedron (∼2.40–2.50 Å) due to the higher coordination number of the In3+ ion. In {[In(chxn)2 ]2 Sb4 S8 }Cl2 , each In3+ ion is coordinated by two S atoms from the {Sb4 S8 }

3D-[InSb3S7]n2n–

2D-[In5Sb6S19]n5n–

+ {In2Sb4S12}n

(a)

3D-[In6Sb6S21]n6n– (In2Sb2Q7-IV)

+

ψ-{SbS4}

{In3Sb4S14}n

(b)

+

{In2Sb2S9}

{In2Sb2S8}n

{In2Sb2S9}

(c)

Figure 5.41 View of (a) 3D-[InSb3 S7 ]n 2n− , (b) 2D-[In5 Sb6 S19 ]n 5n− , (c) 3D-[In6 Sb6 S21 ]n 6n− (In2 Sb2 Q7 -IV), and the combinations of their BUs. Color code: In (bright green), Sb (pink), and S (yellow). Color-highlighted moieties: for (a), {InSb2 S6 }n ribbon (red), ψ-{SbS4 } (blue); for (b), {In3 Sb4 S14 }n ribbon (red) and {In2 Sb2 S9 } cluster (blue); and for (c) {InSbS4 }n ribbon (red), the {In2 Sb2 S9 } cluster (blue). Source: Reproduced with permission [8]. Copyright 2016, Elsevier B.V.

253

5 Group 11–15 Metal Chalcogenides

N3

N2 N4

2.5

4Å

In1

2.5

3Å

N1

S1

S2

0D-{[In(chxn)2]2Sb4S8}2+

(a)

N1

N2

2.55 Å

2.55 Å S1 9Å

S2ʹ

In1

2.6

7Å

S3 2.6

254

S2

(b)

0D-{[In(N,N-dmen)]2Sb2S6}

Figure 5.42 View of (a) 0D-{[In(chxn)2 ]2 Sb4 S8 }2+ and (b) 0D-{[In(N,N-dmen)]2 Sb2 S6 }. Color code: In (bright green), Sb (pink), S (yellow), N (blue), and C (gray). Hydrogen atoms are omitted for clarity. Source: Wang et al. [8]. Copyright 2016, Elsevier B.V.

units and four N donor atoms from two chxn ligands. In [In(N,N-dmen)]2 Sb2 S6 , each In3+ ion is bonded to only one N,N-dmen ligand forming an unsaturated [In(N,N-dmen)]2+ complex; then, a pair of such complexes dimerizes via two 𝜓-{SbS3 } bridges into a double-semicube 0D-{[In(N,N-dmen)]2 Sb2 S6 } molecular species. Interestingly, the higher oxidation state of In3+ ion than that of Mn2+ render the electrically neutral 0D-{[In(N,N-dmen)]2 Sb2 S6 } cluster rather than anionic {[Mn(phen)]2 Sb2 S6 } [89] or {[Mn(phen)]2 As2 S6 } [131] clusters, which are further connected by cationic Hg2+ or [Mn(phen)]2+ to give the infinite neutral chains. Similarly, 0D-{[In(chxn)2 ]2 Sb4 S8 }2+ in {[In(chxn)2 ]2 Sb4 S8 }Cl2 is cationic and is charge-balanced by two Cl− anions.

5.5.4 Chalcogenidometalates Containing Group 14(IV) Ions and Antimony(III) 5.5.4.1 Ge–Sb–S Compounds

Most of organic amines directed germanium sulfantimonates were reported thus far by Huang’s group, including 0D-[(Me)2 NH2 ]6 [(Ge2 Sb2 S7 )(Ge4 S10 )] [102], 1D-[(Me)2 NH2 ][DabcoH]2 [Ge2 Sb3 S10 ] [102], 1D-[aepH2 ][GeSb2 S6 ]⋅CH3 OH [103], 2D-[Ni(en)3 ][GeSb2 S6 ] [102], 2D-[Co(en)3 ][GeSb2 S6 ] [102], 2D-[CH3 NH3 ]20 Ge10 Sb28 S72 ⋅7H2 O [104], 2D-[(CH3 CH2 CH2 )2 NH2 ]3 Ge3 Sb5 S15 ⋅0.5(EtOH) [104], and 3D-[(Me)2 NH2 ]2 [GeSb2 S6 ] [101]. Normally, the Ge4+ ion adopts a tetrahedral coordination geometry in chalcogenidometalates. Various combinations of {GeS4 }

5.5 Chalcogenidoantimonates

and 𝜓-{SbSx } (x = 3, 4) give a great structural variety of Ge–Sb–S compounds, some of which are listed in Table 5.4. Figure 5.43 shows typical heterometallic SBUs and TBUs based on {GeS4 } and 𝜓-{SbSx }, respectively. Among them, the trinuclear {GeSb2 S7 } cluster contains one {GeS4 } tetrahedron and one {Sb2 S5 } group, in which the lone pair electrons of two Sb3+ are oriented on the same side of the plane created by the three metal ions, structurally different from the aforementioned {MSb2 Q7 } (M = Zn, Cd, and Hg; Q = S and Se) cluster [90, 107, 111, 132], where the lone pair electrons of two Sb3+ ions set on the opposite orientations. Two types of 2D anionic networks can be formed by such a {GeSb2 S7 } cluster combined with 4-connected {Sb4 S10 } or 3-connected 𝜓-{SbS3 } groups (Figure 5.44), such as [Co(deta)2 ]2 [GeSb4 S10 ] [97] and [Ni(deta)2 ]3 [Ge3 Sb8 S21 ]⋅0.5H2 O [133]. The two layer structures can be viewed as the nets with 44 ,⋅62 , and 63 topologies, respectively, when considering {GeSb2 S7 } clusters as a 2-connected bridging spacer. The 2D-[GeSb2 S6 ]n 2n− (GeSb2 S6 -I) layers in compounds [M(en)3 ][GeSb2 S6 ] (M = Mn, Co, Ni, and Ge) [102, 134, 135] are formed by interconnection of {GeSb2 S8 } clusters (Figure 5.45a), in which each {GeSb2 S8 } cluster shares four terminal S atoms with three adjacent ones. GeSb2 S6 -I layer is an overall 63 network from a topological viewpoint. There are two structural differences between {GeSb2 S8 } and {GeSb2 S7 } cluster [133] mentioned above. First, the lone pair electrons of two Sb3+ ions point toward two opposite sides of the {GeSbSb} plane in {GeSb2 S8 }. Second, SBUs Tetranuclear

TBUs

SBUs Dinuclear

{GeSbS7} {GeSb3S11}

{GeSb3S9}-1

{GeSbS6}n

{GeSb3S9}n

Trinuclear

{GeSb2S7}

{GeSb2S8}

Pentanuclear

{GeSb3S9}-2

{GeSb3S8}n {GeSb4S14}

{Ge2Sb2S10}

{Ge2Sb2S7}

{Ge3Sb4S14}n

{Ge2Sb3S12}

Figure 5.43 View of some typical Ge–Sb–S SBUs and TBUs. Color code: Ge (blue gray), Sb (pink), and S (yellow). Source: Wang et al. [8]. Copyright 2016, Elsevier B.V.

255

256

5 Group 11–15 Metal Chalcogenides

{Sb4S10}

ψ-{SbS3}

2D-[GeSb4S10]n4n–

(a)

2D-[Ge3Sb8S21]n6n–

(b)

Figure 5.44 Representations of (a) 2D-[GeSb4 S10 ]n 4n− and (b) 2D-[Ge3 Sb8 S21 ]n 6n− . Color code: Ge (blue gray), Sb (pink), and S (yellow). Color-highlighted moieties: for (a) {GeSb2 S7 } cluster (red), {Sb4 S10 } group (blue); for (b) {GeSb2 S7 } cluster (red), ψ-{SbS3 } (blue). Source: Wang et al. [8]. Copyright 2016, Elsevier B.V.

one of the two Sb3+ ions adopts a tetra-coordinated geometry by four S atoms in {GeSb2 S8 } with a see-saw coordination geometry. The rich isomerisms of [GeSb2 S6 ]n 2n− are further demonstrated by two more isomers exhibiting distinct dimensionalities, BUs, and linkage modes. One is the unique 1D-[GeSb2 S6 ]n 2n− (GeSb2 S6 -II) double ribbon in [aepH2 ][GeSb2 S6 ]⋅CH3 OH [103] constructed from two adjacent {GeSb3 S9 }n ribbon-like TBUs composed of tetranuclear {GeSb3 S11 } SBUs sharing infinite {SbS}n edges, (Figure 5.45b). This GeSb2 S6 -II double ribbon is structurally related to that of earlier reported 1D-[Sb9 S15 ]n 3n− [136]. Note that the tetranuclear {GeSb3 S11 } cluster is quite unique among tetranuclear Ge-Sb-Q SBUs. It contains one {GeS4 }, one 𝜓-{SbS3 }, and two 𝜓-{SbS4 }, completely different from {M 2 Sb2 S9 } (M = Ga and In) [1, 91, 93, 94, 99, 126, 127, 130], and {M 2 Sb2 Q10 } (M = In and Sn; Q = S and Se) [96, 99, 105, 126, 127, 129], that are both based on two {MQ4 } and two 𝜓-{SbQ3 }. The first chiral microporous germanium–antimony sulfide [(Me)2 NH2 ]2 [GeSb2 S6 ] (space group P41 21 2) features a 3D-[GeSb2 S6 ]n 2n− (GeSb2 S6 -III) framework with chiral channels (Figure 5.45c and 5.46) [101]. In the structure, 𝜓-SbS4 trigonal bipyramids are interconnected to each other to form a left-handed helical Sb–S–Sb chain running along the 41 -axis (Figure 5.46b). Another type of left-handed helical Ge–S–Sb chain along the 21 -axis is formed by the alternation array of GeS4 tetrahedra and 𝜓-SbS4 trigonal bipyramids via corner-sharing (Figure 5.46c). Each Sb–S–Sb left-handed helix connects to four Ge–S–Sb left-handed helices by sharing one-fourth of the 𝜓-SbS4 polyhedra to generate the 3D-[GeSb2 S6 ]n 2n− framework (Figure 5.46a). The anionic 3D-[GeSb2 S6 ]n 2n− framework can also be viewed as a “line + node” combination of parallel {GeSbS6 }n chain-like TBUs, which is composed of dinuclear {GeSbS7 } SBUs, and discrete 𝜓-{SbS4 } interlinkers (Figure 5.45c). 3D-[GeSb2 S6 ]n 2n− also presents intersected channels parallel to the a- or b-axis (Figure 5.46d). Three SbS4 groups give one Sb3 S10 unit by corner-sharing; the Sb3 S10

5.5 Chalcogenidoantimonates

1D-[GeSb2S6]n2n– (GeSb2S6-II) 2D-[GeSb2S6]n2n– (GeSb2S6-I) (a)

(b)

3D-[GeSb2S6]n2n– (GeSb2S6-III)

2D-[Ge3Sb5S15]n3n–

+ {GeSbS6}n (c)

+

ψ-{SbS4}

{Ge3Sb4S14}n

trans-{Sb2S6}

(d)

Figure 5.45 View of (a) 2D-[GeSb2 S6 ]n 2n− (GeSb2 S6 -I), (b) 1D-[GeSb2 S6 ]n 2n− (GeSb2 S6 -II), (c) 3D-[GeSb2 S6 ]n 2n− (GeSb2 S6 -III), (d) 2D-[Ge3 Sb5 S15 ]n 3n− , and the combinations of their BUs. Color code: Ge (blue gray), Sb (pink), and S (yellow). Color-highlighted moieties: for (a) {GeSb2 S8 } cluster (red); for (b) {GeSb3 S9 }n ribbon (red); for (c) {GeSbS6 }n ribbon (red) and ψ-{SbS4 } (blue); for (d) {Ge3 Sb4 S14 }n ribbon (red) and trans-{Sb2 S6 } dimer (blue). Source: Wang et al. [8]. Copyright 2016, Elsevier B.V.

units are inter-linked by GeS4 tetrahedra to form a left-handed helix (Figure 5.46e), while GeS4 tetrahedra interconnect the middle SbS4 groups of Sb3 S10 units by corner-sharing resulting in a right-handed helix (Figure 5.46f). One full turn of the left- and right-handed helixes include eight (two GeS4 and six SbS4 ) and four polyhedra (two GeS4 and two SbS4 ), respectively. The dimethylammonium cations are located in the channels and form N—H· · ·S hydrogen bonds with the S atoms of the anionic framework. The solvent accessible volume excluding dimethylammonium cations in 1 is 45.2%.

257

258

5 Group 11–15 Metal Chalcogenides

b a Sb Ge S N C

(a)

(b)

(c) L

a c

(e)

Sb

Figure 5.46 (a) View of the structure of [(Me)2 NH2 ]2 [GeSb2 S6 ] along the c-axis. (b) A left-handed helical chain running along the 41 -axis parallel to the c-axis formed by corner-sharing ψ-SbS4 trigonal bipyramids. (c) A left-handed helical chain running along the 21 -axis parallel to the c-axis by the alternate array of GeS4 tetrahedra and ψ-SbS4 trigonal bipyramids via corner-sharing. (d) View of the structure of [(Me)2 NH2 ]2 [GeSb2 S6 ] along the b-axis, the dimethylammonium cations are located in the chiral channels. The left- (e) and right-handed (f) helical chains are parallel to the b-axis.

Ge S

R

N C

(d)

(f)

Figure 5.47 View of T2-[Ge4 S10 ]4− (left) and 0D-[Ge2 Sb2 S7 ]2− (right) in compound 0D-[Me2 NH2 ]6 [(Ge2 Sb2 S7 )(Ge4 S10 )]. Color code: Ge (blue gray), Sb (pink), and S (yellow). Source: Wang et al. [8]. Copyright 2016, Elsevier B.V. T2-[Ge4S10]4–

0D-[Ge2Sb2S7]2–

The only discrete Ge–Sb–S species is present in [(Me)2 NH2 ]6 [(Ge2 Sb2 S7 )(Ge4 S10 )] (Figure 5.47) [102], where the tetranuclear [Ge2 Sb2 S7 ]2− and T2-[Ge4 S10 ]4− cluster coexist. A rare Sb—Sb bond is found in [Ge2 Sb2 S7 ]2− , whose length of 2.82 Å is comparable with those reported in other thioantimonates, e.g. [Ph4 P]2 [Sb4 S6 ] (dSb–Sb = 2.86 Å) [137] and [bapenH4 ]0.5 [Sb7 S11 ] (dSb–Sb = 2.92 Å) [138], as well as to the shortest Sb–Sb distances (2.91 Å) in elemental α-Sb [139]. Two types of pentanuclear Ge–Sb–S clusters as SBUs, namely, {Ge2 Sb3 S12 } and {GeSb4 S14 }, are present in [(Me)2 NH2 ][DabcoH]2 [Ge2 Sb3 S10 ] [102] and [maH]20 [Ge10 Sb28 S72 ]⋅7H2 O [104], Figure 5.48. Both pentanuclear Ge–Sb–S clusters are formed by integrating two trinuclear subunits through a shared metal vertex; however, they differ in composition, architecture, and contribution to the formation of solid-state structures. Sb3+ ion adopts a see-saw coordination polyhedron in {Ge2 Sb3 S12 }, in which the equatorial and apical S ligands are shared with two {GeS4 } and two 𝜓-{SbS3 }. A 1D-[Ge2 Sb3 S10 ]n 3n− ribbon is formed by the interlinkage of {Ge2 Sb3 S12 } clusters (Figure 5.48a), whereas the central {GeS4 } is

5.5 Chalcogenidoantimonates

{GeSb3S8} ribbon n

(a)

1D-[Ge2Sb3S10]

3n– n

{GeS4}

(b)

2D-[Ge10Sb28S72]

20n– n

{GeSb4S11} layer n

Figure 5.48 View of (a) 1D-[Ge2 Sb3 S10 ]n 3n− and (b) 2D-[Ge10 Sb28 S72 ]n 20n− . Color code: Ge (blue gray), Sb (pink), and S (yellow). Color-highlighted moieties: for (a), {Ge2 Sb3 S12 } cluster (red); for (b), top right, {GeSb3 S9 }-1 cluster (red), bottom right, {GeSb4 S14 } cluster (red). Source: Wang et al. [8]. Copyright 2016, Elsevier B.V.

connected with two dimeric {Sb2 S7 } groups in the {GeSb4 S14 } cluster, where all Sb3+ ions adopt the see-saw coordination geometries. An infinite single layer of {GeSb4 S11 }n is formed by the interconnection of each {GeSb4 S14 } cluster with the adjacent four clusters. Then, two single layers of {GeSb4 S11 }n are linked in parallel through additional bridging {GeS4 } tetrahedra to form a double layer. Finally, the overall 2D-[Ge10 Sb28 S72 ]n 20n− anionic layer in [maH]20 [Ge10 Sb28 S72 ]⋅7H2 O is generated with both faces of the double layer decorated by additional {GeSb3 S8 }n ribbons arranged in parallel way (Figure 5.48b). 5.5.4.2 Sn–Sb–S Compounds

Sn–Sb–S compounds are limited compared with Ge–Sb–S compounds. In [La(en)4 ]2 Sn2 Sb2 S10 ⋅0.5H2 O [105], a tetranuclear {Sn2 Sb2 S10 } cluster is formed by the alternating connection of two {SnS4 } and two 𝜓-{SbS3 }, which is identical to {In2 Sb2 Q10 } (Q = S and Se) in configuration [96, 99, 126, 127]. The {Sn2 Sb2 S10 } cluster is further decorated by two unsaturated [La(en)4 ]3+ complexes through the La–S–Sn bridges, resulting in a quaternary neutral 0D-{[La(en)4 ]2 Sn2 Sb2 S10 } (Figure 5.49a). However, Sn4+ ions adopt the octahedral coordination geometries with six S ligands in [TM(en)3 ][SnSb4 S9 ] (TM = Co and Ni) [140]. The anionic 3D-[SnSb4 S9 ]n 2n− framework is formed by the slightly distorted {SnS6 } units connecting discrete {Sb4 S9 } units (Figure 5.49b), where there are two types of chiral channels with cross sections of approximate 8.54 × 8.54 Å2 and 6.14 × 6.14 Å2 , respectively.

259

260

5 Group 11–15 Metal Chalcogenides

0D-{[La(en)4]2Sn2Sb2S10} (a)

{SnS6}

(b)

3D-[SnSb4S9]n2n–

{Sb4S9}

Figure 5.49 View of (a) 0D-{[La(en)4 ]2 Sn2 Sb2 S10 } and (b) 3D-[SnSb4 S9 ]n 2n− . Color code: Sn (dark blue), Sb (pink), S (yellow), N (blue), and C (gray). Hydrogen atoms are omitted for clarity. Source: Wang et al. [8]. Copyright 2016, Elsevier B.V.

5.6 Selected Properties 5.6.1 Properties of Selected Inorganic–Organic Hybrid Metal Chalcogenides Inorganic–organic hybrid metal chalcogenides have shown rich physical and chemical properties, such as tunable optic properties, nearly zero thermal (NZT) expansion, reduced thermal conductivity, white light emission and dilute magnetic property. The most striking feature of II–VI-based hybrid is their giant tunability of optical bandgap (0.1–2 eV) and large enhancement of absorption coefficients with respect to the parent II–VI semiconductors [29, 43]. Moreover, some of the ZnTe-based compounds exhibit NZT expansion [141], reversible or irreversible structural transformation, and interesting thermal conductivity [43], whereas double-layered CdS-amine compounds can emit white light [142]. Magnetic ions such as Mn2+ , Fe2+ , and Co2+ have been successfully incorporated into the host lattice of 3D-[ZnSe(L)0.5 ], 3D-[CdSe(L)0.5 ], 2D-[ZnSe(ba)], and 2D-[(Zn2 Se2 )(ba)] systems [29, 31–33, 37, 39, 41, 42]. Another function of the II–VI-based hybrid compounds is their utility as precursors for the preparation of II–VI nanoparticles and nanostructures with different sizes and morphologies [39, 143]. 3D-[ZnSe(en)0.5 ]

260

5 Group 11–15 Metal Chalcogenides

0D-{[La(en)4]2Sn2Sb2S10} (a)

{SnS6}

(b)

3D-[SnSb4S9]n2n–

{Sb4S9}

Figure 5.49 View of (a) 0D-{[La(en)4 ]2 Sn2 Sb2 S10 } and (b) 3D-[SnSb4 S9 ]n 2n− . Color code: Sn (dark blue), Sb (pink), S (yellow), N (blue), and C (gray). Hydrogen atoms are omitted for clarity. Source: Wang et al. [8]. Copyright 2016, Elsevier B.V.

5.6 Selected Properties 5.6.1 Properties of Selected Inorganic–Organic Hybrid Metal Chalcogenides Inorganic–organic hybrid metal chalcogenides have shown rich physical and chemical properties, such as tunable optic properties, nearly zero thermal (NZT) expansion, reduced thermal conductivity, white light emission and dilute magnetic property. The most striking feature of II–VI-based hybrid is their giant tunability of optical bandgap (0.1–2 eV) and large enhancement of absorption coefficients with respect to the parent II–VI semiconductors [29, 43]. Moreover, some of the ZnTe-based compounds exhibit NZT expansion [141], reversible or irreversible structural transformation, and interesting thermal conductivity [43], whereas double-layered CdS-amine compounds can emit white light [142]. Magnetic ions such as Mn2+ , Fe2+ , and Co2+ have been successfully incorporated into the host lattice of 3D-[ZnSe(L)0.5 ], 3D-[CdSe(L)0.5 ], 2D-[ZnSe(ba)], and 2D-[(Zn2 Se2 )(ba)] systems [29, 31–33, 37, 39, 41, 42]. Another function of the II–VI-based hybrid compounds is their utility as precursors for the preparation of II–VI nanoparticles and nanostructures with different sizes and morphologies [39, 143]. 3D-[ZnSe(en)0.5 ]

5.6 Selected Properties

under hydrostatic pressure has been studied using a first-principles pseudopotential method with mixed-basis set [29], suggesting the highly compressed chemical bonds of the inorganic [ZnSe] layer rather than those of the en. In addition, the study indicates that the band gap as well as other properties of the hybrid compounds may show unusual pressure dependence. 5.6.1.1 Optical and Electric Properties

The absorption spectra of selected hybrids along with that of the parent II–VI binaries are plotted in Figure 5.50. Some noticeable trends can be observed by comparison of the absorption spectra of the selected hybrids. First, all hybrids except for 3D-[MnSe(L)0.5 ] exhibit a large blue shift in their optical absorption edge (0.7–2.0 eV) and enhanced band edge absorption with respect to their corresponding parent II–VI bulk materials; second, the length of the organic molecules and the overall structural dimensionality (for 2D and 3D structures) show negligible effect on the band gap and blue shift of the hybrids; third, a larger blue shift in the optical absorption edge is observed for the 1D-[ZnQ(pda)] compounds compared to that of corresponding 2D and 3D hybrids; finally, a smaller blue shift is found for the optical absorption edge of double-layered 2D-[(M 2 Q2 )(L)] hybrids compared to that of the respective single-layer structures. The greatly enhanced band-edge absorption of II–VI hybrid crystalline nanostructures with respect to their bulk binaries is mostly due to the quantum confinement effect (QCE) [28]. The insulating organic spacers (amines) break down the bulk ZnSe 3D-[ZnSe(en)1/2] 3D-[ZnSe(bda)1/2] 3D-[ZnSe(hda)1/2]

10

ZnTe 12

Kubelka-Munk function

Kubelka-Munk function

12

8 6 4 2

1D-[ZnTe(pda)] 2D-α-[ZnTe(ma)]

10

0

3D-α-[ZnTe(en)1/2]

8 6 4 2 0

0

(a)

1

2

3

4

0

Energy (eV)

2

3

4

Energy (eV)

ZnSe

11 10

Kubelka-Munk function

1

(b) 2D-[(Zn2Se2)(ba)]

9

2D-[ZnSe(ba)]

8 7 6 5 4 3 2 1 0 0

(c)

Figure 5.50 binaries.

1

2

3

4

5

Energy (eV)

The absorption spectra of selected hybrids along with that of the parent II–VI

261

262

5 Group 11–15 Metal Chalcogenides

II–VI semiconductor host lattices into well-ordered low-dimensional neutral segments such as double-layered [M 2 Q2 ], single-layered [MQ] slabs or single-atomic [MQ] chains, very much resembling the semiconductor superlattices. As the layer thickness of the II–VI slabs or the cross section of the II–VI chains in the hybrids fall into the sub-nanometer region (typically 4–5 Å in the single-layered hybrids) and is notably smaller than the sizes of the QDs, QCE is induced and the extent of QCE is significantly greater than that of the smallest nanoparticles (blue shift: ∼1 eV) [29, 33, 42]. In the 1D hybrids, the II–VI chains are confined in two dimensions; therefore, a larger blue shift is expected in the bandgap, compared to that of single-layer hybrid structures in which the II–VI slab are confined in one dimension. Since the thickness of the [M 2 Q2 ] slab in a double-layer 2D-[M 2 Q2 (L)] hybrid is about twice of the [MQ] slab in a corresponding single-layer 2D-[MQ(L)] hybrid, relative less confinement is expected, resulting in a smaller blue shift for the former. The very small bandgap blue shift (0.1–0.2 eV) for the 3D-α-[MnSe(L)0.5 ] hybrids can be ascribed to the high localization nature of the Mn 3d bands which are insensitive to the confinement effect. The origin of QCE has been confirmed and explained by density functional calculations (DFTs) with local density approximation on a number of hybrid structures [144, 145]. DFT calculations confirm that the organic barriers induce a strong confinement of both electrons and holes to the inorganic region and thus the contributions to the valence band and conduction band region come predominantly from the atomic states of the inorganic (II and VI) components. Therefore, changes in the length of the organic molecules would not affect much on the band gap of the resultant hybrids. The successful synthesis of hybrids with different thicknesses of inorganic II–VI slabs implies a new strategy to tune the bandgaps of semiconductors. Obviously, one can systematically tune the bandgaps of hybrids by simply varying the thickness (n) of the II–VI slabs, analogous to the size change in colloidal quantum dots. As outlined in Scheme 5.7, the structure of a 2D-[MQ(L)] is sketched at the right side where n = 1 (single slab of II–VI) with strongest confinement; at the left side is the II–VI bulk (zinc-blende or würtzite) where n = ∞ with no confinement. The structures of 2D-[(MQ)2 (L)] (double layer) and 2D-[(MQ)4 (L)] (quadruple layer) are also sketched in the scheme. Construction of hybrid structures with intermediate n values between n=∞

n=4 n

MQ

n=2

n=1

+ L MQ + L

MQ(L)1/8

MQ(L)1/4

MQ(L)1/2

Tunable band gap (Eg)

Scheme 5.7 II–VI slabs.

The bandgap of II–VI hybrids can be tuned with thickness of inorganic

5.6 Selected Properties

Emission (a.u.)

Absorption (a.u.)

Absorption Emission

300

(a)

400

500

600

Wavelength (nm)

700

800

(b)

Figure 5.51 Absorption and emission spectra of the double-layer 2D-[Cd2 S2 (ba)] based structures (a), which can be used as a single-material white-light-emitting source in LEDs (b).

the two extreme cases will enable controllable and systematic tuning of the band gap, and thus, the electronic and optical properties. More importantly, in the hybrids, the organic spacers are blended with their II–VI into a completely ordered structure, say crystal of macroscopic dimensions, leading to a structure-induced, rather than size-induced, QCE, whereas in the organic-stabilized colloidal QDs, the QDs normally arrange randomly thus lacking long-range order. The band edge excitonic emission is observed in a number of these hybrid nanostructures. Most interestingly, 2D-[Cd2 S2 (L)] (L = ba, pa (n-propylamine, pta (n-pentylamine), and ha (n-hexylamine)) show a broad emission that covers the entire visible spectrum upon excitation of 𝜆ex = 360 nm [142]. It is believed that the very large number of surface sites within each crystal due to the nature of its layered structure are responsible for the significant reduction of the band edge emission and the very broad emission, similar to that observed in very small CdSe NCs. The highest photoluminescence (PL) intensity was achieved for a Mn-doped [Cd2 S2 (ba)] sample at a dopant level of 0.5 mol% of Mn (Figure 5.51a). The fluorescence quantum yield is on the order of 4–5%, comparable to that for CdSe and ZnSe nanocrystals. The perfectly ordered and extended structures, the systematically tunable light emitting properties, possible high carrier conductivity and mobility, all make these hybrids very promising for high-efficiency light-emitting diodes (LEDs) (Figure 5.51b). Luminescence of 3D-Cd1-x Mnx Se(L)0.5 (L = en or hda) was also investigated [37]. The growth of single crystals of the ZnTe-based hybrids in fairly large crystal size (1–2 mm) allows for study of polarization dependence of PL [141]. Polarized PL was measured at different temperatures on selected single crystals of 1D-ZnTe(pda) and 3D-β-[ZnTe(en)0.5 ]. The free exciton–polariton emission at the fundamental band edge was observed at both LT and RT for the 3D-β-[ZnTe(en)0.5 ] structure for the first time. 5.6.1.2 Thermal Expansion Behavior

Materials that may not respond to a temperature change (neither expanding nor contracting) (ZTE) [146–148] are extremely rare but are both fundamentally and practically important because they are in great demand for applications in which a constant length and/or volume is required [149, 150]. Commonly, by mixing a

263

264

5 Group 11–15 Metal Chalcogenides

positive thermal expansion (PTE) and a NTE, composites with ZTE may be realized [151]. However, grain boundary associated cracks often hamper their performance. The II–VI hybrid materials have unique modular features in the crystal structures, since the bulk II–VI with PTE and amines with ZTE [46, 141, 152] are combined into a single crystal lattice, which may give rise to an overall ZTE (or near ZTE) by compensation of PTE and NTE effect of the two individual modules. As an illustrative example, the thermal expansion properties of five 3D-α-ZnTe(L)0.5 single crystals (L = hdz (i), en (ii), pda (iii), bda (iv), and 1,5-pentanediamine [ptda] (v)) over a temperature range of 200 ∘ C have been investigated [46]. All five structures are composed of ZnTe single layers interconnected by dative bonded diamines. The two short axes parallel to the inorganic layers, L1 and L2, expand as the temperature increases for all five compounds; the PTE effect generally decrease as the length of amines increases. The axis perpendicular to the ZnTe layers (long axis, L3) exhibits a significantly smaller thermal expansion, a very small but PTE with shortest amine hydrazine and NTE with increasing magnitude as diamine gets longer. The difference in the thermal behaviour of the organic and inorganic structure components leads to the sign and magnitude change of thermal expansion coefficient. The thermal expansion behaviour of 3D-β-[ZnTe(en)0.5 ] was studied in both single crystal and powder form, revealing ZTE in a very broad temperature range of 4–400 K. The possible mechanisms were suggested by calculation of the photon spectrum employing a linear response theory. While the vast majority of known NTE or ZTE materials are insulators [146], the hybrid II–VI compounds are a remarkable group of semiconductor materials that combine ZTE and systematically tuneable electronic, optical, properties by manipulating two distinctly different structural components in a single crystal lattice. Uniaxial NTE was also observed in another hybrid Mn2 (api)Sb2 S5 [87], which takes an average 𝛼 c (100–373 K) value of −46.7 × 10−6 K−1 (Figure 5.52). The NTE mechanism was investigated at the atomic level using variable-temperature single crystal XRD techniques. The inorganic {Mn2 Sb2 S5 }n layer shows an typical PTE on the planar directions, whereas the organic api ligand is characteristic of NTE due to the partial conformation transfer. The distinct thermal expansions of two components contribute synergistically to the overall NTE along the layer packing direction, by virtue of the specific zigzag linkage of the interlamellar api molecules. Considering the synergistic effect in this hybrid system, a novel concept of the “elevatorplatform” movement was put forward to develop the typical lattice fencing-like mechanism from the geometrical flexibility viewpoint.

5.6.2

Photocatalytic Property

Photocatalytic water splitting and elimination of organic pollutants represent attractive strategies to address the global energy crisis and environmental pollution issues, respectively. Metal chalcogenides are particularly promising as they combine visible-light absorption with good charge transfer properties and appropriate band alignments with the potentials of relevant redox reactions. Over the past years, a large number of nanostructure chalcogenides, such as quantum dots,

5.6 Selected Properties

{Mn2Sb2S5}n layer

api Sb Mn S N C

𝛥d 𝛿

d

𝛿

(a) 002

(b) Low temperature 020

5

(c)

10

15

High temperature

006

20

25

30

35

40

473 Ks 473 K 423 K 373 K 323 K 293 K 200 K 150 K 100 K 100 K-s

2θ (°)

Figure 5.52 (a) Side view of the 3D framework of the Mn2 (api)Sb2 S5 . (b) Representation of the “elevatorplatform” mechanism for the uniaxial NTE behavior of Mn2 (api)Sb2 S5 . (c) Temperature-dependent powder XRD patterns. The top and bottom are simulated patterns from single-crystal XRD data collected at 473 and 100 K. Hydrogen atoms are omitted for clarity. Source: Wang et al. [87]. Copyright 2014, the Royal Society of Chemistry.

have been employed as highly efficient photocatalytic materials [153]. Crystalline chalcogenide photocatalytic materials, however, are far less explored. 5.6.2.1 Photocatalytic Hydrogen Production

In 2020, Li and coworkers report on the metal chalcogenide solid solutions based on larger T5 selenide clusters, that is, [Bmmim]11 [Cd6 In28 Q52 Cl3 (MIm)] (Q = Se (T5–1), Se28.5 S23.5 (T5–2), Se16 S36 (T5–3)) and [Bmmim]9 [Cd6 In28 Se8 S44 Cl(MIm)3 ] (T5–4) [58]. The band gaps estimated from the absorption edges are 2.67 eV for T5–3, 2.79 eV for T5–4, 2.85 eV for T5–5, and 3.01 eV for T5–6. With the increase of the content of doped S element in the T5 clusters, the colors of the obtained crystals become lighter gradually. The photocatalytic H2 evolution activity of these Cd–In–Q compounds under visible light irradiation was tested by using crystalline samples. As shown in Figure 5.53, the amount of H2 evolution for T5 clusters increased almost linearly with the irradiation time. The H2 production rates of T5–3, T5–4, T5–5, and T5–6 were estimated 8.1, 17.3, 44.5, and 59.3 μmol/h g, respectively. Obviously, the hydrogen evolution activity of the T5 clusters increased with increasing the content of doping S element in the compound.

265

5 Group 11–15 Metal Chalcogenides

60

600

T5-3 T5-4 T5-5 T5-6

50

400

40

300

30

200

20

100

10

0

H2 evolution rate (μmol/g h)

500 H2 evolution (μmol/g)

266

0 0

2

4 6 Time (h)

8

10

T5-3 T5-4

T5-5 T5-6

Figure 5.53 Comparison of the photocatalytic H2 productions of T5-3, T5-4, T5-5, and T5-6 systems upon Xe-lamp irradiation with a cut-off filter (𝜆 ≥ 420 nm). Source: Wang et al. [58]. Copyright 2020, Wiley-VCH.

The discrete supertetrahedral chalcogenido Tn clusters can be regarded as a type of QDs with precise structure and uniform size. They were commonly studied in the solid state due to their poor solubility or highly negative charge that leads to instability in common solvents. These drawbacks limit their potential applications as efficient photocatalysts. Recently, Li and coworkers reported on three discrete cluster-based compounds, [Bmmim]9 [Cd3 In17 S31 Cl4 ] (T4–1), [Bmmim]9 [Cd3 In17 S13 Se18 Cl4 ] (T4–2) [Bmmim]9 [Cd3 In17 Se31 Cl4 ](4,4′ -bpy) (T4–3) [57]. T4–1 could not be dispersed in common solvents, while T4–2 and T4–3 could be dispersed uniformly in DMSO after ultrasonic treatment. The high dispersion of T4 clusters was further confirmed by transmission electron microscopy, and the stability of solvents were determined by electrospray ionization mass spectrometry. Their hydrogen evolution activities were tested under visible light irradiation. The stable H2 evolution rate was estimated to be 15.9 μmol/h g for T4–3 in the solid state, while those for T4–1 and T4–2 were 3.5 and 5.6 μmol/h g, respectively. As the T4–1 could not be well scattered in common solvents, stable H2 evolution rate of processed T4–1 was about 6.1 μmol⋅h−1 ⋅g−1 after being further ultrasonicated in DMSO. Comparatively, the stable H2 evolution rates were 27.5 μmol/h g for soluble T4–2 and 91.5 μmol/h g for soluble T4–3, respectively (Figure 5.54, right). Because more efficient channels of electron transfer can be opened after Tn clusters are highly dispersed in solution, the H2 evolution activities for T4 in the dissolved state were about fivefold of that in the solid powder state [57]. The discrete T3 clusters of [Bmmim]6 [In10 Se16 Cl4 ]⋅(MIm)2 could also be stably dispersed in DMSO, exhibiting photocatalytic hydrogen evolution activity (91.5 μmol/h g), at least six times that (15.9 μmol/h g) of the pristine in the solid state due to exposure of more active sites.

5.6 Selected Properties

100

H2 HO

‥

O

e– Pt

NH

+ OH OH

HO

‥

OH

h+

N

HO

Cd-In-Q T4 clusters (Q = S, S/Se, Se)

H2 evolution rate (μmol/g h)

Primary

hv

Dispersed

80 60 40 20 0 T4-1

T4-2

T4-3

Figure 5.54 (a) Discrete supertetrahedral chalcogenido T4 clusters and comparison of the photocatalytic H2 productions of the title T4 systems upon Xe-lamp irradiation with a cut-off filter (𝜆 ≥ 420 nm). Source: Hao et al. [57]. Copyright 2019, the American Chemical Society.

5.6.2.2 Photodegradation of Organic Dye Molecules

Since 2010, the heterogeneous photodegradation of organic dye molecules (e.g. methyl orange (MO) and CV) by using these chalcogenidometalates as catalysts has been investigated. A series of T3 cluster-based compounds of [Bmmim]5 [In10 Q16 Cl3 (Bim)] (Q = S, S/Se, Se, and Se/Te) having tunable optical absorption edges exhibited photodegradation of MO under the irridiation of both UV and visible lights [55]. The degradation ratio of MO over [Bmmim]5 [In10 S16 Cl3 (Bim)] reached nearly 95.4% after 80 minutes under UV light irradiation. By introducing other metal for constructing heterometallic chalcogenides, Huang and coworkers found that the compound (Bmmim)2 In2 Sn2 Se7.5 (Se4 )0.5 with 2D microporous structure possessed the ability of rapid photodegradation of MO under acidic condition [63]. The pH-dependent photocatalytic activity of [Bmmim]2 [In2 Sn2 Se7.5 (Se4 )0.5 ] was carried out in MO solutions at different pH values (pH = 1.5, 3, 4.5, 6.8, 8, 10, and 11.8). As shown in Figure 5.55, the degradation ratios of MO were 0.9% (pH = 6.8), 33.6% (pH = 4.5), 95.7% (pH = 3), and 79.5% (pH = 1.5), after 10 minutes under visible light illumination. Compared to that in neutral condition (removal rate: 96% after three hours), the photodegradation rate was accelerated (99% in 30 minutes for pH < 4.5) and was even six-times faster when pH was below 3 (Figure 5.55a). The pH-dependent photodecomposition can mainly be attributed to the variations of surface charge properties of the photocatalyst. The adsorption of MO is favored in the acidic solution due to its anionic configuration. As for the photodegradation process, the close contact of MO with the catalyst could facilitate its oxidative degradation by positive holes or hydroxyl radicals. The pH-dependent photodecomposition of MO was observed for layered indium chalcogenidoantimonates(III) solid solutions [(Me)2 NH2 ]2 [In2 Sb2 S7-x Sex ] (x = 0, 2.20, 4.20, and 7) [96]. As the proportions of Se increase, the compounds showed a red-shift of their optical absorption edges and exhibited tunable photocatalytic

267

5 Group 11–15 Metal Chalcogenides

1.0

1.0 pH = 1.5 pH = 3

0.8

pH = 4.5 pH = 6.3 pH = 8.5

0.6

pH = 10 pH = 11.8 pH = 1.5 without catalyst

0.4

C / C0

C / C0

0.8

pH = 6.8 without catalyst pH = 11.8 without catalyst

0.2

0.6 0.4 0.2

0.0

0.0 0

10

20 Time (min)

(a)

30

40

0

20

40

(b)

60

80 100 120 140 160 180 Time (min)

Figure 5.55 Photodegradation of MO in different pH values by [Bmmim]2 [In2 Sn2 Se7.5 (Se4 )0.5 ] monitored as the normalized change in concentration as a function of irradiation time under visible light (Xe lamp, 420–780 nm). (a) The partial enlarged detail for the first 45 minutes of photodegradation process and (b) the full graph. Source: Du et al. [63]. Copyright 2015, the American Chemical Society.

activity for degradation of MO with a shift of optical response from UV to the visible light region. [Bmmim]12 [Cu5 In30 Q52 Cl3 (Im)] (Q = Se, Se48.5 S3.5 ) exhibit the ability of photocatalytic degradation of CV under visible light irradiation; the photocatalytic activity of sulfur-doped is clearly significantly higher than that of pure selenium [58]. Chalcogenidometalates containing silver and tin lies in the energy range suitable for visible light photocatalytic applications. For example, [CH3 NH3 ]2 Ag4 SnIV 2 SnII S8 [78] and [CH3 NH3 ]2 [H3 O]Ag5 Sn4 Se12 ⋅C2 H5 OH [77] exhibit the optical band gaps of 2.10 eV and 2.10, respectively. Their photocatalytic activity was evaluated by degradation of CV as the test pollutant under visible light irradiation. As shown in Figure 5.56a, the photolysis of CV without the photocatalyst could be neglected. The degradation ratio of CV reached 92% in 200 minutes for [CH3 NH3 ]2 Ag4 SnIV 2 SnII S8 . For [CH3 NH3 ]2 [H3 O]Ag5 Sn4 Se12 ⋅EtOH, the degradation ratio of CV reached 96% when exposed to visible light irradiation for 10 minutes and achieved nearly 100% after 60 minutes, resulting in complete decolorization (Figure 5.56b). Without photocatalyst Photocatalyst

1.0

1.0

0.8

0.8

0.6

0.6

C / C0

C / C0

268

0.4

0.4 0.2

0.2 0.0 0

(a)

Without photocatalyst Photocatalyst

30

60

90 120 150 180 210 240 Time (min)

0.0 0

(b)

10

20

30 40 Time (min) Time (min)

50

60

Figure 5.56 Photodegradation of CV by [CH3 NH3 ]2 Ag4 SnIV 2 SnII S8 (a) and [CH3 NH3 ]2 [H3 O]Ag5 Sn4 Se12 ⋅EtOH (b) monitored as the normalized change in concentration as a function of irradiation time. Source: Zhang et al. [77]. Copyright 2016 and 2017, the American Chemical Society.

5.6 Selected Properties

5.6.3

Ion Exchange Property

Today, both environmental protection and energy saving are of vital significance for the development of human society. Uranium is an essential element with radioactivity and high chemical toxicity in the nuclear fuel cycle. Uranyl cations UO2 2+ can dissolve in water and cause environmental and human health problems [154]. Uranium harvesting is also of interest in nuclear energy generation. About 4.5 billion tons of uranium (about 3.3 ppb) presents in the ocean, and it is very appealing to find ways to capture it [155, 156]. 137 Cs (t1/2 = 30.17 years) and 90 Sr (t1/2 = 28.80 years) are produced by nuclear fission of 235 U and/or 239 Pu after having absorbed neutrons in a nuclear reactor. They are one of the main hazardous non-actinide isotopes present in nuclear wastes owing to emitting high-energy β and γ radiations [157]. And Cs+ and Sr2+ ions having high solubility and strong biological toxicity could easily release into aqueous systems and thus posing long-term threat to environmental safety and human health. Radioactive lanthanides can cause long-term environmental pollution due to long half-life and non-biodegradable nature. For instance, europium as the representative of lanthanides has relatively long-lived radioisotopes (152 Eu: t1/2 ∼ 13.3 years; 154 Eu: t1/2 ∼ 8.6 years) that may penetrate into human body and significantly alter normal physiological processes by inhibiting enzymes, regulating synaptic transmitters and blocking membrane receptors [158]. Importantly, Eu3+ is typically employed as the simulant for long-lived radiotoxic actinides (An3+ ) such as Am3+ for their similar adsorption properties However, the effective sequestration of radioactive ions from complex radioactive wastes remains a serious challenge. Ion exchange treatment is an attractive method for its high selectivity, minimum solidified waste, and reductive radioactive discharge. Chalcogenidometalates are a new class of promising ion exchange materials due to their many advantages compared to oxides. For instance, the S2− or Se2− anions in their structures acting as soft base have the strong affinity for soft Lewis acidic metal ions [1, 76, 80, 101, 159–165]. What is more, their frameworks are more flexible and dynamic because of the wider chemical and bonding flexibility and can even display “breathing action” to capture the certain ions [1, 160]. In this field, the early research was focused on the inorganic crystalline chalcogenido ion exchangers, such as KBi3 S5 [166, 167], K14 Cd15 Sn12 Se46 [161], K6 Sn[Zn4 Sn4 S17 ] [160, 162], (NH4 )4 In12 Se20 [163], K2x Mnx Sn3−x S6 (KMS−1 ) [164, 165], and K2x Mgx Sn3−x S6 (KMS-2) [168]. In recent years, the important progress has been made in the search for chalcogenido ion exchange materials with organic amine cations as counterions and SDAs [1, 76, 80, 98, 101, 104, 130, 169], as the organic amines have advantages of large size tunability and conformational flexibility. In 2008, Huang’s group reported the first chiral microporous Ge–Sb–S compound, namely, 3D-[(Me)2 NH2 ]2 GeSb2 S6 (Figure 5.46) [101] with high ion exchange capacity and high selectivity for Cs+ ion. Nearly hexagon-shaped channels with a cross section of 7.68 × 5.65 Å2 parallel to the a- or b-axis in the structure are observed. [(Me)2 NH2 ]+ cations are located in the channels and form N—H· · ·S hydrogen bonds with the S atoms of the anionic network. The experiments proved that [(Me)2 NH2 ]+ cations in 3D-[(Me)2 NH2 ]2 GeSb2 S6 were able to be partially

269

270

5 Group 11–15 Metal Chalcogenides

exchanged by alkali metal cations such as Cs+ , Rb+ , K+ , and Na+ ions in aqua solutions. The removal rates of Cs+ , Rb+ , K+ , and Na+ were 93%, 85%, 75%, and 34%, respectively. Interestingly, the competitive ion exchange experiments indicated that the 3D-[(Me)2 NH2 ]2 GeSb2 S6 had a strong preference for Cs+ ions. For instance, the competitive exchange experiment with a 20 : 1 M ratio of Na+ : Cs+ gave the product with only Cs+ exchange. The excellent ion exchange property can be attributed to its open framework with 3D chiral channels which enable the guest cations to diffuse in and out. The ion exchange capacity falls along with the decreasing radii of Cs+ , Rb+ , K+ , and Na+ ions, i.e. the softer Lewis acidic cations are preferred over harder ones related to the soft basic nature of the framework. Then, Huang’s group also observed the interesting structural flexible response upon ion exchange. For instance, 3D-[(Me)2 NH2 ]0.75 [Ag1.25 SnSe3 ] features a novel microporous framework containing in situ-generated dimethylammonium in the intersecting 3D channels (Figure 5.57a) [80]. The channels parallel to the c-axis are petal-shaped with a cross section of 7.6 × 7.1 Å2 . Single-ion exchange experiments indicated that its [(Me)2 NH2 ]+ could be readily replaced by Cs+ , Rb+ , and NH4 +

b

a c

Sn Ag Se N C

(a) 118°

133°

5.8 Å

6.0 Å 6.1 Å

7.1 Å

3.65 Å

4.00 Å

86°

98°

(b)

(c)

Figure 5.57 (a) View of the structure of [(Me)2 NH2 ]0.75 [Ag1.25 SnSe3 ] along the [110] direction; (b) and (c) show the cross sections of channels along the [110] in the pristine compound and the Cs+ -exchanged product, respectively. Source: Li and Huang [80]. Copyright 2011, the Royal Society of Chemistry.

5.6 Selected Properties

in exchange yields of 93%, 87%, and 67%, respectively. Interestingly, the c-axis of the unit cell of the Cs+ -exchanged crystals shrinked ∼0.8 Å and the cell volume constricted ∼6% (∼111 Å3 ) compared to that of the pristine. The cross section of the channels running along the [110] or [1–10] directions was observed to be 6.1 × 6.0 Å2 , indicating a significant reduction compared to that in the pristine (7.1 × 5.8 Å2 ) (Figure 5.57b,c). On the contrary, the cross section of the channels running along the c-axis was 7.9 × 7.6 Å2 , indicating an expansion compared to 7.6 × 7.1 Å2 in the pristine. The ion exchange-induced structural variation might stem from the strong interaction between Cs+ ions and Se2− ions in the soft basic framework, which seemed to also play an important role in the high preference for Cs+ ion. Such framework flexibility has also been observed for the Rb+ - and NH4 + -exchanged products of 3D-[(Me)2 NH2 ]0.75 [Ag1.25 SnSe3 ]. In addition, the phenomenon of framework flexibility upon ion exchange has also been observed in 3D-[CH3 NH3 ]4 [In4 SbS9 SH] [98]. Detailed ion exchange reactions demonstrated its high selectivity for Rb+ ion adsorption as well as significant role of the pore size on the ion exchange selectivity. Under the time-dependent Rb+ exchange treatments, the compound exhibited framework flexibility, in which the increasing exchange yields accompanied with gradual contraction of its unit cell. It is well known that the practical application of any materials greatly depends on the cost and efficiency. Therefore, it is of significance to prepare the ion exchangers with remarkable function by facile methods. Accordingly, a 2D microporous thiostannate, namely, 2D-[(Me)2 NH2 ]4/3 [(Me)3 NH]2/3 [Sn3 S7 ]⋅1.25H2 O (FJSM-SnS) was prepared in large-scale by a facile, one-pot and economically solvothermal method. The structure of FJSM-SnS features a 2D [Sn3 S7 ]n 2n− anionic layer with large windows formed by 24-membered [Sn12 S12 ] rings from six [Sn3 S4 ] cores (Figure 5.58a). The [(Me)2 NH2 ]+ , [(Me)3 NH]+ cations and lattice water molecules are located at the interlayer spaces. FJSM-SnS presents excellent Cs+ and Sr2+ ion exchange properties. The single-crystal X-ray crystallography data of Cs+ -exchanged FJSM-SnS have been collected (Cs2 Sn3 S7 ⋅4.5H2 O, denoted FJSM-SnS-Cs) to further understand the ion exchange mechanism. For a clearer comparison, a hexagonal window was brought out with [Sn3 S4 ] as a node and (μ–S)2 as a ligand, respectively. All the lengths of sides shrinked 0.1∼0.3 Å after Cs+ exchange, while the diagonal lengths of the hexagonal window mentioned above changed from 15.04 × 15.08 × 15.08 Å3 to 14.86 × 15.09 × 15.09 Å3 (Figure 5.58b). The interlayer distance condensed from 7.258 to 6.709 Å (Figure 5.58c), which was in accordance with the shift toward the higher 2𝜃 angles in the PXRD patterns after ion exchange. The structural variation described above might spring from the strong affinity of Cs+ to the soft basic framework (Figure 5.58d). FJSM-SnS has many advantages as an ion exchange material. First, its flexible framework has an intermediate character between the layered chalcogenidometalates and porous molecular sieves, which plays an important role in the extraordinary ion exchange performance. Second, the ion exchange equilibrium for Cs+ and Sr2+ can be reached within five minutes, which is comparable to those of the other ion exchangers including the commercial UOP IONSIV IE-911 whose equilibrium time is more than four hours [170]. Third, the maximum Cs+ and Sr2+ ion exchange capacities (qm ) are

271

5 Group 11–15 Metal Chalcogenides

B2 7.5 12 Å

0Å

(a)

7.530 Å

A3

Å 12 7.5

A4

b

14.860 Å

Å

a

2 7.51

a c

B5

A2

7.5 40 Å

b

b

Å 40 7.5

Sn2

15.041 Å

B1 7.513 Å Å 092 15.

1Å

Sn1

0Å

.0 8 15

A5 Sn3 S1

B6

4 7.5

7.530 Å

7.5 12 Å

A1

A6 7.5 4

272

a

B4 7.513 Å

B3

(b)

Sn S Cs

Cs+ or Sr2+

c a

(c)

(d)

Figure 5.58 (a) View of a 2D [Sn3 S7 ]n 2n− anionic layer parallel to the ab plane; the simplified hexagon with Sn3 S4 as a node and (μ-S)2 as a ligand in FJSM–SnS and FJSM–SnS–Cs (b); packing of FJSM–SnS–Cs (c); the capture of Cs+ or Sr2+ ion by FJSM–SnS (d).

408.91 and 65.19 mg/g, respectively. It is worth noting that the qm of FJSM-SnS for Cs+ ion is 1.8 times of that of KMS-1 (226 mg/g) [165] and far more than those of commercial AMP-PAN (81 mg/g) and TAM-5 (191.8 mg/g) which are currently marketed by UOP as IONSIV IE-910 and IE-911 [170–173]. Fourth, the FJSM-SnS particularly shows wide pH resistance (0.7∼12.7) which makes it outstanding amongst the ion exchangers. Finally, the ion exchange chromatographic column is first studied for chalcogenido ion exchange materials, that is, a column filled with 3.0 g of FJSM-SnS could remove 96–99% of Cs+ ion and near 100% of Sr2+ ion at low ionic concentrations in 900 bed volumes solutions, respectively. All these advantages make it a promising material for radionuclide Cs+ and Sr2+ remediation. FJSM-SnS also presents the first example of chalcogenides in which organic amine cations can be used for selective UO2 2+ ion exchange (Figure 5.59) [174]. FJSM-SnS has high exchange capacity, affinity, and selectivity for UO2 2+ ions with the excellent super acid and alkali resistance (pH = 2.1–11), which are attributed to the strong affinity of soft Lewis basic S2− ions for relatively softer Lewis acidic UO2 2+ ions as well as its 2D microporous and flexible open framework. Its maximum uranium exchange capacity is 338.43 mg/g, which is comparable to those of the best reported uranium adsorbents [175–177]. It is worth noting that the qm of FJSM-SnS for U is much higher than commercial resin products such as commercial phosphinic acid resin, Tulsion CH-96 (70 mg/g) [178], strong base AMBERSEP 920 U Cl Resin

5.6 Selected Properties

Ion exchange

Ln3+

Me2NH2+/Me3NH+ Out

Cs+/Sr2+ In

Sn S

[Me2NH2]+ [Me3NH2]+

UO22+

2D-[(Me2NH2)4/3(Me3NH)2/3Sn3Se7•1.25H2O] (FJSM-SnS) for Cs+, Sr2+, UO22+, Ln3+-exchange Cs+

[Me2NH2]+

(Me)2NH2+

In

Radioactive metal ion

Cs+

Ion-exchange Out

R+

3D-[(Me)2NH2]0.75[Ag1.25SnSe3]

3D-[(Me)2NH2]2GeSb2S6 Cs+

Rb+

[CH3NH3]+

3D-[CH3NH3]4[In4SbS9SH]

[CH3NH3]+

2D-[CH3NH3]20Ge10Sb28S72∙7H2O

2D-[Me2NH2]2[Ga2Sb2S7]∙H2O (FJSM-GAS-1) 2D-[Et2NH2]2[Ga2Sb2S7]∙H2O (FJSM-GAS-2)

Figure 5.59 Selected ion exchange chalcogenidometalates containing organic amine cations reported by Huang’s group. Source: Feng et al. [174]. Copyright 2016, the American Chemical Society.

(50 mg/g) [179], and ARSEN-Xnp Purolite Resin (47 mg/g) [180]. In addition, it could efficiently capture UO2 2+ ions in the presence of high concentrations of Na+ , Ca2+ , or HCO3 − (the highest K d value reached 4.28 × 104 ml/g). It is very effective for the removal of trace levels of U even against Na+ (the relative amounts of U removed are close to 100%). What is more, uranyl in corresponding exchanged products can be easily eluted with a cost-affordable and environmentally friendly method. To date, crystalline metal sulfides have been investigated as excellent ion exchangers for radioactive elements, heavy metals, and actinides removal [1, 162, 163, 175, 176, 181, 182]. Huang’s group studied the efficient Ln3+ recovery performance of FJSM-SnS, which is the first chalcogenide example as a superior Ln3+ ion exchanger from very complex aqueous solutions (Figure 5.59) [183]. The results indicate that FJSM-SnS exhibits rapid kinetics (five minutes), high capacity (139 mg/g for Eu, 147 mg/g for Tb, 126 mg/g for Nd), wide pH resistance (1.9∼8.5), the largest distribution coefficient (K d ) value of 6.5 × 106 ml/g, and good selectivity against Al3+ , Fe3+ , and Na+ ions. In addition, the high recovery rate (> 99%) in ion exchange column experiments and the efficient regeneration by elution with KCl solution indicate that FJSM-SnS is an excellent rare earth element (REE) trapping material. 133 Ba2+ , 63 Ni2+ , and 60 Co2+ with long half-life and high environmental mobility are hazardous radioisotopes. Ba2+ is also employed as the simulant for highly radiotoxic 226 Ra2+ . It is important to study the chemically selective scavengers for 133 Ba2+ , 63 Ni2+ , and 60 Co2+ from complex waste water for radionuclide remediation and human health. FJSM-SnS also exhibits excellent capture properties for Ba2+ ,

273

274

5 Group 11–15 Metal Chalcogenides

Ni2+ , and Co2+ with high capacities (qm Ba = 289.0 mg/g; qm Ni = 83.27 mg/g; qm Co = 51.98 mg/g), fast kinetics (within five minutes), wide pH durability, and outstanding Ba2+ selectivity. It exhibits high removal efficiencies (>99%) for these ions in ion exchange column experiments. The material possesses radiation resistance with good structural and crystal stability, and it could survive in highly acidic conditions [184]. It was also found by Huang’s group that the organically directed layered Ge–Sb–S compound, namely, 2D-[CH3 NH3 ]20 Ge10 Sb28 S72 ⋅7H2 O, can exhibit excellent Cs+ ion exchange property despite the presence of excess competitive cations, such as Na+ , K+ , Mg2+ , and Ca2+ ions [104]. Its maximum exchange capacity for Cs+ ion was found to be 230.91 mg/g, and its structural stability is impressive over a wide range of pH (2.8 ∼ 11.0). In addition, the ion exchange equilibrium can be reached within two minutes. Such fast removal efficiency is comparable with that of FJSM-SnS [76] and KMS-1 [165] and significantly higher than that of the commercial zeolite A [185] and AM-2 [186]. This should be attributed to the strong interactions between acidic Cs+ cations and S2− ions of the basic framework, and high mobility of the small organic amines. Besides, the organically directed layered Ge–Bi–S compound, namely, 2D-[(Me)2 NH2 ][BiGeS4 ], shows ion exchange ability for Rb+ ion [187]. The chalcogenidometalate as Ni2+ ion exchanger has also been studied. An open framework chalcogenidoantimonate, namely, [CH3 NH3 ]4 Ga4 SbS9 S0.28 O0.72 H, showed Ni2+ ion exchange properties and wide pH resistance, with a maximum exchange capacity of 76.9 mg/g and high K d of 1.65 × 105 ml/g [95]. This is the first example of amine-directed 3D heterometallic chalcogenidometalates for the highly selective Ni2+ ion capture. In addition, [TAEAH][TAEAH2]0.6 Ga2.2 Sn1.8 S8 ⋅H2 O (GaSnS-1; TAEA = tris(2-aminoethyl)amine) presents mercury uptake performances [188]. Its structure features a 3D zeolite-typed framework of [Ga2.2 Sn1.8 S8 ]n 2.2n− that is constructed by corner-sharing of supertetrahedral [Ga2.2 Sn1.8 S10 ]6.2− T2 clusters. The equilibrium model study indicates that the maximum Hg2+ saturation capacity of GaSnS-1 was 213.9 mg/g. GaSnS-1 possessed extremely rapid adsorption kinetics following the pseudo-second-order model with k2 of 5.65 × 102 g/(mg min). Particularly, GaSnS-1 exhibited an excellent selectivity for Hg2+ ion with the high distribution coefficient K d value of 1.62 × 107 ml/g and high removal efficiency of close to 100%. The superior Hg2+ ion adsorption performance was also impressive despite the presence of excessive competing cations and the acidic/basic conditions. Furthermore, a simple chromatographic column loaded with GaSnS-1 microcrystals is capable of rapidly and effectively capturing Hg2+ ion far below the upper limit (2 ppb, USA-EPA) of drinking water. These advantages make GaSnS-1 a promising candidate for the fast and efficient remediation of Hg2+ -contaminated water sources. Recently, Huang’s group prepared water-stable K+ -activated porous sulfide K@GaSnS-1 by K+ ion exchange of GaSnS-1. K@GaSnS-1 shows excellent UO2 2+ ion exchange properties with high exchange capacity (qm U = 147.6 mg/g) and wide pH resistance (pH = 2.75–10.87). In particular, it can efficiently capture UO2 2+ ion even under the presence of excessive Na+ , K+ , Mg2+ , and Ca2+ ions. The highest distribution coefficient

5.7 Conclusion

K d value, signifying the affinity and selectivity for UO2 2+ ion, reaches as high as 1.24 × 104 ml/g. More importantly, the uranium in corresponding exchanged samples can be facilely and effectively eluted by a low-cost and eco-friendly method. These merits make K@GaSnS-1 promising for the effective and selective removal of uranium from complex contaminated water. Two new gallium thioantimonates, namely, [(Me)2 NH2 ]2 [Ga2 Sb2 S7 ]⋅H2 O (FJSM-GAS-1) and [(Et)2 NH2 ]2 [Ga2 Sb2 S7 ]⋅H2 O (FJSM-GAS-2), featuring the inorganic anionic 2D networks with organic amine cations located at the interlayer spaces (Figures 5.34), present excellent ion exchange properties for UO2 2+ , Cs+ , and Sr2+ ions [92]. They exhibit high ion exchange capacities for UO2 2+ , Cs+ , and Sr2+ ions (qm U = 196.10 mg/g, qm Cs = 164.23 mg/g and qm Sr = 79.79 mg/g for FJSM-GAS-1, qm U = 143.74 mg/g for FJSM-GAS-2) and short equilibrium time for UO2 2+ ion exchange (5 minutes for FJSM-GAS-1 and 15 minutes for FJSM-GAS-2, respectively). Both compounds display active ion exchange with UO2 2+ in the pH range of 2.9–10.5. Moreover, they could maintain high K d U even in the presence of excess Na+ , Ca2+ , and HCO3 − . The K d U of 6.06 × 106 ml/g exhibited by FJSM-GAS-1 is the highest among the reported U adsorbents. The UO2 2+ -laden products can be recycled by conveniently eluting the uranium with a low-cost method. These advantages coupled with facile synthesis, as well as β and γ radiation resistance, make them new promising materials for nuclear remediation. Selected chalcogenidometalates containing organic amine cations with ion exchange properties reported by Huang’s group are shown in Figure 5.59.

5.7 Conclusion This chapter highlights the recent development of group 11–15 metal chalcogenides. A unique class of hybrid semiconductor materials with a general formula of [(MQ)n (L)x ] (MQ = ZnS, ZnSe, ZnTe, CdS, CdSe, CdTe, and MnSe; L = mono- or di-amines or hydrazine; n = 1, 2; x = 0.5, 1, and 2) are described. These compounds incorporate organic and inorganic components into a single crystal lattice via covalent (coordinative) bonding to form extended 1D, 2D, and 3D network structures composed of subnanometer sized II–VI semiconductor segments (inorganic component) and amine molecules (organic component) arranged into perfected ordered arrays. They exhibit very large blue shift in their optical absorption edge, as a result of strong structure-, rather than size-induced QCE. Such confinement can be systematically tuned by modifying the crystal structure, dimensionality, and thickness of the inorganic motifs. In addition, the blending of the organic and inorganic components has led to a number of enhanced properties, as well as new phenomena and new functionality into the hybrid structures that are not possible with the individual components alone. These include an exceptionally broad bandgap tunability and very strong band edge absorption, unique conformational, structural, and thermal properties, and so on. Known examples of structure-induced (particle size independent) quantum confinement are rare, largely because crystal growth of periodic, covalent bonded modular hybrid structures remains to be

275

276

5 Group 11–15 Metal Chalcogenides

highly challenging. Thus, we hope that our discovery and understanding of this unprecedented type of II–VI-based inorganic–organic hybrid semiconductors will provide useful and valuable information toward the future development of hybrid nanostructured materials of similar nature. Ionothermal synthesis has shown promising for the preparation of crystalline chalcogenides. The chapter has presented the preparation and structural variation of selenidostannates and discrete chalcogenides Tn clusters from ILs. The structures of selenidostannates and Tn clusters can be tuned by simply changing the variety of ILs and amines, as well as reaction temperature and time. The use of mixture of IL and auxiliary amine as a designable solvent system may open up new opportunities in the preparation of novel crystalline metal chalcogenides (e.g. 3D extra-large microporous selenidostannates and largest discrete chalcogenides Tn clusters) that are inaccessible in traditional molecular solvents. Chalcogenidometalates with organically directing chalcogenidometalates represent a fascinating subclass of materials within inorganic compound community. Variations of structural directing organic amine/IL cations as well as change in reaction conditions have result in Ag–Sn–Q (Q = S and Se) with unique crystal structural types and benign properties. By virtue of the BU concept, the coordination chemistry of the heterometallic chalcogenidoantimonates can be regarded as the combination of the regular-polyhedral {M(Q/N)x } (M = 12(II) of Zn, Cd, and Hg; 13(III) of Ga and In; 14(IV) of Ge and Sn; Q = S and Se; N = N donor atom of the ligand; x = 2, 3, 4, 5, and 6) and the pseudo-polyhedral 𝜓-{SbQx } (x = 3 and 4; Q = S and Se). This offers a clear structural analysis from the coordination characteristics of PBUs, to the condensation of PBUs to form SBUs or even TBUs, and to their further assemblies into the final solid-state structures. The protonated amine cations present powerful features for structure-directing and charge-balancing roles. The results in the research on optical, electric, and ion exchange properties rendering this family of materials promising applications in energy utilization, environmental remediation, health, and materials science that are related to the sustainable development of human society. At last, although diverse structures have been already prepared, the research in this area still faces challenging items: (i) the way to better understand the structure-directing roles of the organic species and put forward guidelines for the synthesis of desirable chalcogenide materials; (ii) the relationship between the intrinsic structures, compositions, and the corresponding properties; and (iii) the experimental/theoretical pathways can be taken to understand better the real mechanism for these perspective properties.

Acknowledgments The authors thank the current and former research group members Professor Wei-Wei Xiong, Professor Ke-Zhao Du, Dr. Kai-Yao Wang, Dr. Cheng-Feng Du, Dr. Bo Zhang, Xin-Hui Qi, Yu-Jie Gao, De-Nian Kong, Professor Zai-Lai Xie, Dr. Xiao-Wu He, Yan-Qi Wang, Min-Ting Hao, Dr. Qian-Qian Hu, Dr. Nan-Nan Shen,

References

Dong Ye, Dr. Yu-Long Wang, and Dr. Guo-Dong Zou for their contributions to this research. The authors also thank Ying-Chen Peng and Jun-Hao Tang for proofreading the whole chapter. XH thanks Professor Jing Li for supervising the research of II–VI hybrid chalcogenides. MF thanks Professor Mercouri G. Kanatzidis for supervising some research of ion exchange properties. Grants from the NNSF of China (Nos. 21771183, 21521061, 21371001, 21373223, 21221001, 21001104, 21171164, 20873149, 20803081, 21905279, 91127011, and 20771102), the NSF of Fujian Province (Nos. 2018J01027, 2016J01083, and 2010J01056), the 973 programs (Nos. 2014CB845603, 2012CB821702, and 2006CB932904), Chunmiao project of Haixi institute of Chinese Academy of Sciences (CMZX-2014-001), the FJIRSM&IUE Joint Research Fund (No. RHZX-2018-005), and the Youth Innovation Promotion Association of Chinese Academy of Sciences (2011220) are acknowledged.

References 1 Ding, N. and Kanatzidis, M.G. (2010). Nat. Chem. 2: 187. 2 Panthani, M.G., Akhavan, V., Goodfellow, B. et al. (2008). J. Am. Chem. Soc. 130: 16770. 3 Heremans, J.P., Jovovic, V., Toberer, E.S. et al. (2008). Science 321: 554. 4 Guo, S.P., Chi, Y., and Guo, G.C. (2017). Coord. Chem. Rev. 335: 44. 5 Zheng, N.F., Bu, X.H., Vu, H., and Feng, P.Y. (2005). Angew. Chem. Int. Ed. 44: 5299. 6 Lu, J.X. (2000). Some New Aspects of Transition-Metal Cluster Chemistry. Beijing/New York: Science Press. 7 Lu, Q.P., Yu, Y.F., Ma, Q.L. et al. (2016). Adv. Mater. 28: 1917. 8 Wang, K.Y., Feng, M.L., Huang, X.Y., and Li, J. (2016). Coord. Chem. Rev. 322: 41. 9 Kumar, S. and Nann, T. (2006). Small 2: 316. 10 Tell, B. and Kasper, H.M. (1971). Phys. Rev. B 4: 4455. 11 Chang, C., Wu, M.H., He, D.S. et al. (2018). Science 360: 778. 12 Feng, P.Y., Bu, X.H., and Zheng, N.F. (2005). Acc. Chem. Res. 38: 293. 13 Sheldrick, W.S. and Wachhold, M. (1998). Coord. Chem. Rev. 176: 211. 14 Seidlhofer, B., Pienack, N., and Bensch, W. (2010). Z. Naturforsch., B 65: 937. 15 Sunshine, S.A., Kang, D., and Ibers, J.A. (1987). J. Am. Chem. Soc. 109: 6202. 16 Kanatzidis, M.G. (1990). Chem. Mater. 2: 353. 17 Sheldrick, W.S. and Wachhold, M. (1997). Angew. Chem. Int. Ed. Engl. 36: 206. 18 Liao, J.H. and Kanatzidis, M.G. (1990). J. Am. Chem. Soc. 112: 7400. 19 Li, J., Chen, Z., Wang, R.J., and Proserpio, D.M. (1999). Coord. Chem. Rev. 190: 707. 20 Parise, J.B. (1991). Science 251: 293. 21 Fuhr, O., Dehnen, S., and Fenske, D. (2013). Chem. Soc. Rev. 42: 1871. 22 Li, J., Huang, X.Y., and Yuen, T. (2010). Towards dilute magnetic semiconductors: doped inorganic–organic hybrid nanostructured materials based on

277

278

5 Group 11–15 Metal Chalcogenides

23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59

II-VI semiconductors. In: Doped Nanomaterials and Nanodevices, vol. 3 (ed. W. Chen), 149–171. American Scientific Publishers. Zhang, Q.C., Chung, I., Jang, J.I. et al. (2009). J. Am. Chem. Soc. 131: 9896. Xiong, W.W., Athresh, E.U., Ng, Y.T. et al. (2013). J. Am. Chem. Soc. 135: 1256. Wang, K.Y., Ding, D., Zhang, S. et al. (2018). Chem. Commun. 54: 4806. Achermann, M., Petruska, M.A., Kos, S. et al. (2004). Nature 429: 642. Neeraj, C.N.R. and Rao, J. (1998). Mater. Chem. 8: 279. Huang, X.Y., Li, J., and Fu, H.X. (2000). J. Am. Chem. Soc. 122: 8789. Huang, X.Y., Heulings, H.R., Le, V., and Li, J. (2001). Chem. Mater. 13: 3754. Deng, Z.X., Wang, C., Sun, X.M., and Li, Y.D. (2002). Inorg. Chem. 41: 869. Huang, X.Y. and Li, J. (2002). Mat. Res. Soc. Symp. Proc. 728: S1.7.1. Deng, Z.X., Li, L.B., and Li, Y.D. (2003). Inorg. Chem. 42: 2331. Huang, X.Y., Li, J., Zhang, Y., and Mascarenhas, A. (2003). J. Am. Chem. Soc. 125: 7049. Lu, J., Wei, S., Peng, Y.Y. et al. (2003). J. Phys. Chem. B 107: 3427. Ouyang, X., Tsai, T.Y., Chen, D.H. et al. (2003). Chem. Commun.: 886. Huang, X.Y., Heulings, H.R., Li, J. et al. (2005). J. Nanosci. Nanotechnol. 5: 1487. Lu, J., Wei, S., Yu, W.C. et al. (2005). Chem. Mater. 17: 1698. Mitzi, D.B. (2005). Inorg. Chem. 44: 7078. Yao, W.T., Yu, S.H., Huang, X.Y. et al. (2005). Adv. Mater. 17: 2799. Moon, C.Y., Dalpian, G.M., Zhang, Y., and Wei, S.H. (2006). Chem. Mater. 18: 2805. Gawel, B., Lasocha, W., and Zieba, M. (2007). J. Alloys Compd. 442: 77. Huang, X.Y. and Li, J. (2007). J. Am. Chem. Soc. 129: 3157. Huang, X.Y., Roushan, M., Emge, T.J. et al. (2009). Angew. Chem. Int. Ed. 48: 7871. Wang, S.J. and Li, J. (2015). J. Solid State Chem. 224: 40. Luberda-Durnas, K., Guillen, A.G., and Lasocha, W. (2016). J. Solid State Chem. 238: 162. Li, J., Bi, W.H., Ki, W. et al. (2007). J. Am. Chem. Soc. 129: 14140. Larson, A.C. and Von Dreele, R.B. (1998). GSAS, Generalized Structure Analysis System. Los Alamos, NM: Los Alamos National Laboratory. Lin, Y., Massa, W., and Dehnen, S. (2012). Chem. Eur. J. 18: 13427. Rogers, R.D. and Seddon, K.R. (2003). Science 302: 792. Cooper, E.R., Andrews, C.D., Wheatley, P.S. et al. (2004). Nature 430: 1012. Santner, S., Heine, J., and Dehnen, S. (2016). Angew. Chem. Int. Ed. 55: 876. Parnham, E.R. and Morris, R.E. (2007). Acc. Chem. Res. 40: 1005. Lin, Y.M., Massa, W., and Dehnen, S. (2012). J. Am. Chem. Soc. 134: 4497. Xiong, W.W., Li, J.R., Hu, B. et al. (2012). Chem. Sci. 3: 1200. Shen, N.N., Hu, B., Cheng, C.C. et al. (2018). Cryst. Growth Des. 18: 962. Yang, D.D., Li, W., Xiong, W.W. et al. (2018). Dalton Trans. 47: 5977. Hao, M.T., Hu, Q.Q., Zhang, Y.F. et al. (2019). Inorg. Chem. 58: 5126. Wang, Y.Q., Zhu, Z.P., Sun, Z.F. et al. (2020). Chem. Eur. J. 26: 1624. Wang, Y.Q., Hu, Q.Q., Jin, J.C. et al. (2020). Dalton Trans. 49: 5020.

References

60 61 62 63 64 65 66 67 68

69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95

Li, J.R., Xie, Z.L., He, X.W. et al. (2011). Angew. Chem. Int. Ed. 50: 11395. Li, J.R., Xiong, W.W., Xie, Z.L. et al. (2013). Chem. Commun. 49: 181. Du, C.F., Li, J.R., Feng, M.L. et al. (2015). Dalton Trans. 44: 7364. Du, C.F., Li, J.R., Zhang, B. et al. (2015). Inorg. Chem. 54: 5874. Du, C.F., Shen, N.N., Li, J.R. et al. (2016). Chem. Asian J. 11: 1555. Du, C.F., Shen, N.N., Li, J.R. et al. (2016). Dalton Trans. 45: 9523. Du, K.Z., Feng, M.L., Li, J.R., and Huang, X.Y. (2013). CrystEngComm 15: 5594. Wu, T., Bu, X.H., Liao, P.H. et al. (2012). J. Am. Chem. Soc. 134: 3619. Levchenko, T.I., Huang, Y.N., and Corrigan, J.F. (2017). Large metal chalcogenide clusters and their ordered superstructures via solvothermal and ionothermal syntheses. In: Clusters - Contemporary Insight in Structure and Bonding, vol. 174 (ed. S. Dehnen), 269. Switzerland: Springer International Publishing. Cahill, C.L., Ko, Y., and Parise, J.B. (1998). Chem. Mater. 10: 19. Li, H.L., Eddaoudi, M., Laine, A. et al. (1999). J. Am. Chem. Soc. 121: 6096. Zhang, X., Luo, W., Zhang, Y.P. et al. (2011). Inorg. Chem. 50: 6972. Zhang, Y.P., Zhang, X., Mu, W.-Q. et al. (2011). Dalton Trans. 40: 9746. Wu, T., Wang, L., Bu, X. et al. (2010). J. Am. Chem. Soc. 132: 10823. Holm, R.H. (1981). Chem. Soc. Rev. 10: 455. Zheng, N.F., Bu, X.H., and Feng, P.Y. (2003). Nature 426: 428. Qi, X.H., Du, K.Z., Feng, M.L. et al. (2015). J. Mater. Chem. A 3: 5665. Zhang, B., Feng, M.L., Li, J. et al. (2017). Cryst. Growth Des. 17: 1235. Zhang, B., Li, J., Du, C.F. et al. (2016). Inorg. Chem. 55: 10855. Du, K.Z., Qi, X.H., Feng, M.L. et al. (2016). Inorg. Chem. 55: 5110. Li, J.R. and Huang, X.Y. (2011). Dalton Trans. 40: 4387. Chou, J.H. and Kanatzidis, M.G. (1996). J. Solid State Chem. 123: 115. Graf, H.A. and Schafer, H. (1972). Z. Naturforsch., B 27: 735. Seidlhofer, B., Djamil, J., Näther, C., and Bensch, W. (2011). Cryst. Growth Des. 11: 5554. Lees, R.J.E., Powell, A.V., and Chippindale, A.M. (2007). J. Phys. Chem. Solids 68: 1215. Powell, A.V., Lees, R.J.E., and Chippindale, A.M. (2006). Inorg. Chem. 45: 4261. Du, K.Z., Feng, M.L., Li, L.H. et al. (2012). Inorg. Chem. 51: 3926. Wang, K.Y., Feng, M.L., Zhou, L.J. et al. (2014). Chem. Commun. 50: 14960. Kong, D.N., Xie, Z.L., Feng, M.L. et al. (2010). Cryst. Growth Des. 10: 1364. Wang, K.Y., Zhou, L.J., Feng, M.L., and Huang, X.Y. (2012). Dalton Trans. 41: 6689. Wang, K.Y., Ye, D., Zhou, L.J. et al. (2013). Dalton Trans. 42: 5454. Feng, M.L., Xie, Z.L., and Huang, X.Y. (2009). Inorg. Chem. 48: 3904. Feng, M.L., Sarma, D., Gao, Y.J. et al. (2018). J. Am. Chem. Soc. 140: 11133. Kong, D.N., Feng, M.L., Ye, D., and Huang, X.Y. (2010). Chin. J. Struct. Chem. 29: 905. Lin, Z.E., Bu, X.H., and Feng, P.Y. (2010). Microporous Mesoporous Mater. 132: 328. Zhang, B., Li, W.A., Liao, Y.Y. et al. (2018). Chem. Asian J. 13: 672.

279

280

5 Group 11–15 Metal Chalcogenides

96 Wang, K.Y., Feng, M.L., Kong, D.N. et al. (2012). CrystEngComm 14: 90. 97 Feng, M.L., Li, P.X., Du, K.Z., and Huang, X.Y. (2011). Eur. J. Inorg. Chem. 2011: 3881. 98 Wang, K.Y., Feng, M.L., Li, J.R., and Huang, X.Y. (2013). J. Mater. Chem. A 1: 1709. 99 Wang, K.Y., Zhang, B., Qi, X.H. et al. (2015). Cryst. Growth Des. 15: 29. 100 Quiroga-Gonzalez, E., Nather, C., and Bensch, W. (2010). Solid State Sci. 12: 1235. 101 Feng, M.L., Kong, D.N., Xie, Z.L., and Huang, X.Y. (2008). Angew. Chem. Int. Ed. 47: 8623. 102 Feng, M.L., Xiong, W.W., Ye, D. et al. (2010). Chem. Asian J. 5: 1817. 103 Feng, M.L., Hu, C.L., Wang, K.Y. et al. (2013). CrystEngComm 15: 5007. 104 Zhang, B., Feng, M.L., Cui, H.H. et al. (2015). Inorg. Chem. 54: 8474. 105 Feng, M.L., Ye, D., and Huang, X.Y. (2009). Inorg. Chem. 48: 8060. 106 Holloway, C.E. and Melnik, M. (1994). Main Group Met. Chem. 17: 799. 107 Tang, W.W., Tang, C.Y., Wang, F. et al. (2013). J. Solid State Chem. 199: 287. 108 Fu, M.L., Guo, G.C., Cai, L.Z. et al. (2005). Inorg. Chem. 44: 184. 109 Wang, X., Sheng, T.L., Hu, S.M. et al. (2009). J. Solid State Chem. 182: 913. 110 Liu, G.N., Jiang, X.M., Wu, M.F. et al. (2011). Inorg. Chem. 50: 5740. 111 Yue, C.Y., Lei, X.W., Liu, R.Q. et al. (2014). Cryst. Growth Des. 14: 2411. 112 Imafuku, M., Nakai, I., and Nagashima, K. (1986). Mater. Res. Bull. 21: 493. 113 Chou, J.H. and Kanatzidis, M.G. (1995). Chem. Mater. 7: 5. 114 Zhou, J.A., Yin, X.H., and Zhang, F. (2010). Inorg. Chem. 49: 9671. 115 Gang, G., Baiyin, M.H., Wang, X.J. et al. (2013). Chin. J. Inorg. Chem. 29: 979. 116 Chen, Z. and Wang, R.J. (2000). Acta Chim. Sini. 58: 326. 117 Chen, Z., Wang, R.J., and Li, J. (1998). MRS Proc. 547: 419. 118 Li, J., Chen, Z., Wang, X., and Proserpio, D.M. (1997). J. Alloys Compd. 262–263: 28. 119 Zhou, J., An, L., Hu, F. et al. (2012). CrystEngComm 14: 5544. 120 Zhou, J., Hu, F.L., An, L.T. et al. (2012). Dalton Trans. 41: 11760. 121 Zhou, J., An, L.T., Liu, X. et al. (2012). Chem. Commun. 48: 2537. 122 Mertz, J.L., Ding, N., and Kanatzidis, M.G. (2009). Inorg. Chem. 48: 10898. 123 Liu, X. (2011). Inorg. Chem. Commun. 14: 437. 124 Seidlhofer, B., Antonova, E., Wang, J. et al. (2012). Z. Anorg. Allg. Chem. 638: 2555. 125 Wang, Z., Zhang, H., and Wang, C. (2009). Inorg. Chem. 48: 8180. 126 Zhou, J., Liu, X., An, L. et al. (2013). Dalton Trans. 42: 1735. 127 Zhou, J. and An, L.T. (2011). CrystEngComm 13: 5924. 128 Zhou, J., Bian, G.Q., Zhang, Y. et al. (2007). Inorg. Chem. 46: 6347. 129 Zhou, J., An, L.T., and Zhang, F. (2011). Inorg. Chem. 50: 415. 130 Ding, N. and Kanatzidis, M.G. (2007). Chem. Mater. 19: 3867. 131 Liu, G.N., Guo, G.C., Chen, F. et al. (2011). Inorg. Chem. 51: 472. 132 Jiang, H., Wang, X., Sheng, T.L. et al. (2011). Sci. China Chem. 41: 726. 133 Zhou, J., Liu, X., Liang, G.M. et al. (2013). Inorg. Chem. Commun. 27: 92. 134 Zhou, J., An, L.T., Liu, X. et al. (2011). Dalton Trans. 40: 11419.

References

135 Powell, A.V. and Mackay, R. (2011). J. Solid State Chem. 184: 3144. 136 Spetzler, V., Kiebach, R., Nather, C., and Bensch, W. (2004). Z. Anorg. Allg. Chem. 630: 2398. 137 Martin, T.M., Schimek, G.L., Pennington, W.T., and Kolis, J.W. (1995). Dalton Trans.: 501. 138 Powell, A.V., Boissiere, S., and Chippindale, A.M. (2000). Chem. Mater. 12: 182. 139 Barrett, C.S., Cucka, P., and Haefner, K. (1963). Acta Cryst. 16: 451. 140 Yue, C.Y., Lei, X.W., Ma, Y.X. et al. (2014). Cryst. Growth Des. 14: 101. 141 Zhang, Y., Islam, Z., Ren, Y. et al. (2007). Phys. Rev. Lett. 99: 215901. 142 Ki, W. and Li, J. (2008). J. Am. Chem. Soc. 130: 8114. 143 Gao, M.R., Xu, Y.F., Jiang, J., and Yu, S.H. (2013). Chem. Soc. Rev. 42: 2986. 144 Reimann, S.M. and Manninen, M. (2002). Rev. Mod. Phys. 74: 1283. 145 Fluegel, B., Zhang, Y., Mascarenhas, A. et al. (2004). Phys. Rev. B 70: 205308. 146 Salvador, J.R., Gu, F., Hogan, T., and Kanatzidis, M.G. (2003). Nature 425: 702. 147 Margadonna, S., Prassides, K., and Fitch, A.N. (2004). J. Am. Chem. Soc. 126: 15390. 148 Mary, T.A., Evans, J.S.O., Vogt, T., and Sleight, A.W. (1996). Science 272: 90. 149 Evans, J.S.O., Hu, Z., Jorgensen, J.D. et al. (1997). Science 275: 61. 150 Sleight, A.W. (1998). Annu. Rev. Mater. Sci. 28: 29. 151 Barrera, G.D., Bruno, J.A.O., Barron, T.H.K., and Alan, N.L. (2005). J. Phys. Condens. Matter 17: R217. 152 Wang, L., Dove, M.T., Shi, J. et al. (2020). Dalton Trans. 49: 719. 153 Li, X.B., Tung, C.H., and Wu, L.Z. (2018). Nat. Rev. Chem. 2: 160. 154 Craft, E.S., Abu-Qare, A.W., Flaherty, M.M. et al. (2004). Sep. Sci. Technol. 7: 297. 155 Kim, J., Tsouris, C., Mayes, R.T. et al. (2013). Sep. Sci. Technol. 48: 367. 156 Davies, R.V., Kennedy, J., Mcilroy, R.W. et al. (1964). Nature 203: 1110. 157 Romanovskiy, V.N., Smirnov, I.V., Babain, V.A. et al. (2001). Solvent Extr. Ion Exch. 19: 1. 158 Palasz, A. and Czekaj, P. (2000). Acta Biochim. Pol. 47: 1107. 159 Li, H.L., Laine, A., O’Keeffe, M., and Yaghi, O.M. (1999). Science 283: 1145. 160 Manos, M.J., Iyer, R.G., Quarez, E. et al. (2005). Angew. Chem. Int. Ed. 44: 3552. 161 Ding, N. and Kanatzidis, M.G. (2006). Angew. Chem. Int. Ed. 45: 1397. 162 Manos, M.J., Chrissafis, K., and Kanatzidis, M.G. (2006). J. Am. Chem. Soc. 128: 8875. 163 Manos, M.J., Malliakas, C.D., and Kanatzidis, M.G. (2007). Chem. Eur. J. 13: 51. 164 Manos, M.J., Ding, N., and Kanatzidis, M.G. (2008). Proc. Natl. Acad. Sci. U.S.A. 105: 3696. 165 Manos, M.J. and Kanatzidis, M.G. (2009). J. Am. Chem. Soc. 131: 6599. 166 Mccarthy, T.J., Tanzer, T.A., and Kanatzidis, M.G. (1995). J. Am. Chem. Soc. 117: 1294. 167 Chondroudis, K. and Kanatzidis, M.G. (1998). J. Solid State Chem. 136: 328. 168 Mertz, J.L., Fard, Z.H., Malliakas, C.D. et al. (2013). Chem. Mater. 25: 2116. 169 Lan, Y., Su, Z., Li, X.L. et al. (2007). J. Radioanal. Nucl. Chem. 273: 99.

281

282

5 Group 11–15 Metal Chalcogenides

170 Huckman, M.E., Latheef, I.M., and Anthony, R.G. (1999). Sep. Sci. Technol. 34: 1145. 171 Tranter, T.J., Herbst, R.S., Todd, T.A. et al. (2002). Adv. Environ. Res. 6: 107. 172 Zheng, Z., Philip, C.V., Anthony, R.G. et al. (1996). Ind. Eng. Chem. Res. 35: 4246. 173 Park, Y., Lee, Y.C., Shin, W.S., and Choi, S.J. (2010). Chem. Eng. J. 162: 685. 174 Feng, M.L., Sarma, D., Qi, X.H. et al. (2016). J. Am. Chem. Soc. 138: 12578. 175 Manos, M.J. and Kanatzidis, M.G. (2012). J. Am. Chem. Soc. 134: 16441. 176 Sarma, D., Malliakas, C.D., Subrahmanyam, K.S. et al. (2016). Chem. Sci. 7: 1121. 177 Ma, S.L., Huang, L., Ma, L.J. et al. (2015). J. Am. Chem. Soc. 137: 3670. 178 Wnkatesan, K.A., Shyamala, K.V., Antony, M.P. et al. (2008). J. Radioanal. Nucl. Chem. 275: 563. 179 Cheira, M.F., El-Didamony, A.M., Mahmoud, K.F., and Atia, B.M. (2014). IOSR J. Appl. Chem. 7: 32. 180 Panturu, R.I., Jinescu, G., Panturu, E. et al. (2011). Rev. Chim. 62: 814. 181 Manos, M.J. and Kanatzidis, M.G. (2009). Chem. Eur. J. 15: 4779. 182 Manos, M.J. and Kanatzidis, M.G. (2016). Chem. Sci. 7: 4804. 183 Qi, X.H., Du, K.Z., Feng, M.L. et al. (2017). J. Am. Chem. Soc. 139: 4314. 184 Gao, Y.J., Sun, H.Y., Li, J.L. et al. (2020). Chem. Mater. 32: 1957. 185 El-Kamash, A.M. (2008). J. Hazard. Mater. 151: 432. 186 Dobelin, N. and Armbruster, T. (2007). Microporous Mesoporous Mater. 99: 279. 187 Feng, M.L., Qi, X.H., Zhang, B., and Huang, X.Y. (2014). Dalton Trans. 43: 8184. 188 Zhang, B., Li, J., Wang, D.N. et al. (2019). Inorg. Chem. 58: 4103.

283

6 The Structures of Metal–Organic Frameworks Yuehong Wen, Xintao Wu, and Qi-Long Zhu State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, 155 Yangqiao Road West, Fuzhou 350002, Fujian, China

6.1 Introduction Metal–organic frameworks (MOFs), also known as porous coordination polymers (PCPs), a class of emerging crystalline materials, are constructed from periodically alternate metal ions/clusters and organic linkers. The term MOFs is technically suitable for any extended array comprised of metal nodes (either mono- or polynuclear) and organic units; however, it is commonly referred to as coordination polymers with porosity capable of including guest molecules [1]. The history of MOFs can be dated back to the eighteenth century. Nonetheless, the research in this field has rapidly developed in only the past two decades, attributing to the crystal engineering and routinely accessible X-ray crystallographic techniques. Nowadays, MOFs have grown to an unprecedented level of activity, because of not only their diverse and tunable structures, but also their great potential in many applications such as material storage and separation, delivery, sensing, biomedicine, catalysis, nonlinear optics, and so on [2–15]. MOFs are built from metal containing nodes and organic building units via coordination bonds, so the choice of metal sources and organic ligands plays a key role for MOFs construction. Most metals have been employed for MOF synthesis, in contrast to inorganic porous materials, which are based on only a few metals. The organic ligands are carboxylates, phosphonate, sulfonate, heterocyclic compounds, metallo-ligands, etc. or their mixtures, which can be rigid or flexible and ditopic, tritopic, tetratopic, hexatopic, or octatopic according to the different coordination modes to metal nodes. Usually, the metal building units and coordination bonds are formed in situ, while the linkers are predesigned and keep their integrity during the assemble process. Therefore, the structure of the resulting MOFs can be dictated by the geometry and connectivity of linker. The structures of MOFs are primarily dependent upon the metal moieties and organic ligands; however, other factors such as solvents, templates, counterions, temperature, pH values, and concentration have an unpredictable impact on their Advanced Structural Chemistry: Tailoring Properties of Inorganic Materials and their Applications, First Edition. Edited by Rong Cao. © 2021 WILEY-VCH GmbH. Published 2021 by WILEY-VCH GmbH.

284

6 The Structures of Metal–Organic Frameworks

final structures. Given the huge choice of possible building units, along with the varied experimental conditions and synthetic methods, the diversity of MOF materials can be easily imagined. Therefore, a great number of MOFs have been reported, and the numbers is continuing to rise. Generally, the MOFs can be categorized into three types based on their dimensionalities: one-, two-, or three-dimensional infinite networks. In this chapter, various MOFs are categorically highlighted on the basis of their dimensions, with a focus on the construction and structural analysis, which provides an introduction of recent achievements in MOFs.

6.2 One-dimensional MOFs In one-dimensional (1D) MOFs, ligands and node centers generally alternate in one direction. 1D chains are considered to be the simplest and the least interesting structures in the coordination polymeric materials; however, noncovalent interactions between such 1D infinite chains can result in the formation of interesting architectures. The most common 1D chains are shown in Figure 6.1, which are linear chains, zigzag chains, helical chains, double chains, ladder and railroad-like chains, etc. [16].

6.2.1

Linear Chains

Linear chains are obviously the simplest structural motif in 1D MOFs, which can easily be synthesized by design from linear spacer ligands coordinated to the metal ions in a linear (trans) fashion. There are numerous examples of linear MOFs. For instance, complex [Co(1,4-bis(4-pyridyl)butadiyne)(NO3 )2 (H2 O)2 ⋅2H2 O]n shows a straight-chain polymeric architecture via two ligands bonding the metal center in a trans fashion. Such linear chains are associated by hydrogen bonding into pairs that crisscross each other, leading to a layer architecture (Figure 6.2) [17].

Linear chain

Helix

Zigzag chain

Double chain

Ladder Railroad

Figure 6.1

1D coordination polymer motifs. Source: Based on Leong and Vittal [16].

6.2 One-dimensional MOFs

Figure 6.2 Packing diagram of [Co(1,4-bis(4-pyridyl)butadiyne)(NO3 )2 (H2 O)2 ⋅2H2 O]n . Disordered water molecules located in the cavities are shown as isolated circles. Source: Zaman et al. [17]. © 2001 American Chemical Society.

Complex [Ni4 (oba)4 (4,4′ -bpy)4 (H2 O)4 ]⋅(4,4′ -bpy)2 (H2 oba = 4,4′ -oxybis(benzoic acid); 4,4′ -bpy = 4,4′ -bipyridine) displays a very unusual and interesting 1D train-like box motif, in which four linear 1D chains [Ni(oba)] are linked together by 4,4′ -bpy ligands to form [Ni4 (4,4′ -bpy)4 ] box. Each Ni(II) is linked to two oba ligands in trans manner, two 4,4′ -bpy ligands in cis fashion, and terminated by two water ligands. These 1D train-like boxes are entangled through the hydrogen bonds, generating a 3D porous supramolecular structure with free 4,4′ -bpy ligands locating in the channels [18]. An inclusion complex {[(C4AS)2 Ag3 (μ-2,2′ -bpy)2 (2,2′ -bpy)2 ] [Ag5 (l-2,2′ -bpy)4 (2,2′ -bpy)4 ]⋅20H2 O}n (2,2′ -bpy = 2,2′ -bipyridine, C4AS = p-sulfonatocalix[4]arene) consists of a 1D infinite silver(I) chain[Ag5 (l-2,2′ -bpy)4 (2,2′ -bpy)4 ] and a C4AS-trisilver block [(C4AS)2 Ag3 (μ-2,2′ -bpy)2 (2,2′ -bpy)2 ]. C4AS-trisilver block is generated by two C4AS molecules and a trinuclear silver(I) unit via coordination interactions, which further forms hydrophobic layers. The 1D infinite silver(I) chains serve as counterion and occupy the cavities. The dissociative water molecules compose 1D infinite water belts, which function as fences of clathrate-like units to entrap these silver(I) chains (Figure 6.3). None of the π· · ·π stacking interactions exists between calixarenes; instead, the hydrogen bonding interactions play an important role [19]. A polyoxometalates (POMs)-based 1D MOF, [Cu2 (H2 tda)2 (H2 O)4 ]2 ⋅[Cu2 (tda)2 (H2 O)4 ]⋅[HPW12 O40 ]⋅18H2 O has been prepared from 1,2,3-triazole-4,5-dicarboxylic acid (H3 tda), CuCl2 , counterions TMAH (tetramethylammonium hydroxide),

285

286

6 The Structures of Metal–Organic Frameworks c b

Ag S O N C

Figure 6.3 Clathrate-like structure, where water belts acting as fences isolate every two C4AS cavities into clathrate-like units, and then the 1D infinite silver(I) chains reside in there. Source: Xiong et al. [19]. © 2009 Royal Society of Chemistry.

and H3 PW12 O40 at 100 ∘ C. In the structure, the PW12 cluster is linked by Cu3 (II) dinuclear secondary building units (SBUs) to form a 1D linear chain, while the Cu1 (II)–Cu2 (II) dinuclear SBUs are distributed around the 1D chains (Figure 6.4a). By replacing the H3 PW12 O40 POM with H4 SiW12 O40 , another MOF [Cu2 (H2 tda)2 (H2 O)4 ]⋅[SiW12 O40 ]⋅(TMA)2 ⋅3H2 O has been obtained, in which the SiW12 cluster is linked by Cu(II) dinuclear SBUs to form a 1D chain with TMA counterions distributed around the 1D chains (Figure 6.4b). Other POM-MOF complexes with various dimensionalities and topologies have also been built by Cu3 unit

Cu1–Cu2 Discrete SBUs TMA

(a)

(b)

Figure 6.4 (a) 1D chains surrounded by discrete SBUs. (b) 1D chains surrounded by the TMA counter ion. Source: Sun et al. [20]. © 2016 Royal Society of Chemistry.

6.2 One-dimensional MOFs

(a)

(b)

Figure 6.5 (a) Coordination environment of Ag(I) in complex [Ag(pyppt)(NO2 )](CHCl3 ). (b) 1D chain array. Source: Li et al. [21]. © 2017 Royal Society of Chemistry.

changing the POM species, temperature, and counterions [20]. The charges of POM polyoxoanions control the coordination number of POMs; the high temperature tends to form high dimensionalities; the counterions efficiently hinder building blocks from assembling, which benefits the formation of lower dimensionalities. A 1D chain [Ag(pyppt)(NO2 )](CHCl3 ) has been constructed by the reaction of 3-(2-pyridyl)-4-(4-pyridyl)-5-(3-pyridyl)-4H-1,2,4-triazole (pyppt) and AgNO2 in CHCl3 /CH3 CN. In the structure, each pyppt ligand bonds three Ag(I) centers via the 2-pyridyl, 3-pyridyl, and triazole groups as a μ3 -bridge, and the Ag(I) centers are locked by nitrite anions, forming 1D polymeric array (Figure 6.5), which are stacked in a parallel fashion with the separation of 9.355 Å. The void is sufficient for the free rotation of 1D SBUs. Thus, the spontaneous SC-SC transformation of such 1D chain to a 3D cationic MOF is achieved by soaking it in the aqua solution of CF3 COONa. Such conversion is followed by a tandem pathway [21].

6.2.2

Zigzag Chains

Zigzag-type chain is also ubiquitous in 1D MOFs, which can be constructed from flexible exoditopic ligands and linear or cis-coordinated octahedral metal centers or tetrahedral metal ions. A 1D zigzag chain structure of

287

288

6 The Structures of Metal–Organic Frameworks b

c

O1

O1W

N2

O18

O4 N6

Figure 6.6 2D supramolecular layer formed by linking 1D zigzag chains through the direct (O4⋅ ⋅ ⋅N6) and indirect (O1W: N2⋅ ⋅ ⋅O1W; O1⋅ ⋅ ⋅O1W; O18⋅ ⋅ ⋅O1W) hydrogen interactions. Source: Deng et al. [22]. Copyright 2012, American Chemical Society.

[Cd2 (H2 pidc)4 (H2 O)2 ]n ⋅2nH2 O is built by 2-propylimidazole-4,5-dicarboxylate (H3 pidc) ligand and Cd(II). The H2 pidc− anions exhibit three coordination modes, connecting Cd(II) centers to generate the zigzag chains, which are linked together by direct hydrogen bond interactions (O4· · ·N6) and indirect hydrogen interactions (O1W: N2· · ·O1W; O1· · ·O1W; O18· · ·O1W) to generate a two-dimensional (2D) supramolecular layer (Figure 6.6). These adjacent layers are further stacked together by the direct hydrogen bond interactions to form a 3D supramolecular structure [22]. A copper(I)/copper(II)-salen coordination polymer {[Cu(SalImCy)](CuI)2 ⋅DMF}n (SalImCy = N,N ′ -bis-[(imidazol-4-yl)methylene]cyclohexane-1,2-diamine) has been prepared by solvothermal reactions of the copper(II)-salen ligand [Cu (SalHImCy)](NO3 )2 and CuI in a molar ratio of 1 : 2. The prominent structural feature of this complex contains the square-planar 4-coordinate Cu(II) in the metallo-salen ligand and the 3-coordinate Cu(I) in the rhomboid Cu2 I2 clusters. The copper(II)-salen with an angle of 112.5∘ between two deprotonated imidazol N atoms is linked by two Cu2 I2 nodes to form zigzag chains, which stack together to form tubular channels to trap DMF molecules [23]. Moreover, the 1D chain MOF could be used to catalyze the asymmetric three-component strecker reaction. The structures of two solvated compounds, {[Zn(ebic)2 ]⋅EtOH}n and {[Co(ebic)2 ]⋅H2 O}n (Hebic = 2-ethyl-1H-benzo[d]imidazole-5-carboxylic acid), display 3D supramolecular networks composed by stacking 1D [M(ebic)2 ]n chains through N—H· · ·O hydrogen bonds and π· · ·π stacking interactions (Figure 6.7). These 1D chains can be regarded as the aggregates of [2+2] {M2 (ebic)2 } metallo-cycles or two 1D [M(ebic)]n zigzag chains by sharing the M(II) centers. The 1D channels of the 3D networks are filled with solvent molecules, which can be

6.2 One-dimensional MOFs

c b

a

d = 3.7 Å

d = 3.58 Å (a)

(b)

Figure 6.7 (a) The stacking of 1D [M(ebic)]n chains via hydrogen bonding and π· · ·π interactions. (b) 3D supramolecular network showing the 1D channels. Source: Yu et al. [24]. © 2015 American Chemical Society.

removed reversibly without loss of crystallinity. Their desolvated materials display efficient iodine uptake capacity, and the iodine-loaded crystals showed 80 times higher electrical conductivity [24]. A 1D zigzag Cu(I)–MOF is prepared from CuBr2 and the fluorene-bridged imidazole capped ligand under solvothermal reactions. In the structure, the Cu(I) nodes adopt linear coordination modes, which are connected by the bidentate ligands to generate a zigzag polymer chain. The [CuBr2 ]− moieties are alternately adhered at both sides of the chain backbone by weak Cu· · ·Cu interaction. Therefore, this polymer could be considered as coordination polymer supported CuBr2 species, acting as a highly effective multifunctional heterogeneous catalyst for a series of organic transformations [25]. Three MOFs with sinusoidal or zigzag chain structural architectures have been constructed from a rigid ligand 1,2-bis(3-pyridyl)ethyne (3,3′ -DPA) (Figure 6.8). In [Cd(3,3′ -DPA)(NO3 )2 (H2 O)2 ]n , the 3,3′ -DPA ligands adopt a trans fashion, and the pyridyl rings rotate almost 90∘ relative to one another. Such coordination geometry, in combination with the orientation of nitrogen atoms at the 3-position, leads to a sinusoidal chain, which is intertwined via intralayer hydrogen bonds between the coordinated water molecules and the nitrate ions to form a 2D layer. MOF [Cu(3,3′ -DPA)(CH3 OH)(NO3 )2 ]n shows a sinusoidal chain-type architecture as well; however, the five coordinate {CuO3 N2 } square–pyramidal arrangement of the copper compound is quite different from the octahedral coordination environment for the cadmium atom in [Cd(3,3′ -DPA)(NO3 )2 (H2 O)2 ]n . Similarly, the sinusoidal polymeric chains are stacked pair-wise by hydrogen bonds between the coordinated methanol molecule and the nitrate ions. In [Cu(hfac)2 (3,3′ -DPA)1.5 (NO3 )2 ]n (hfac = hexafluoroacetylacetonate), each copper center is connected by 3,3′ -DPA ligands into a one-dimensional zigzag chain. Neighboring polymeric chains stack closely in an AB fashion, and the fluorine atoms of the hfac ligand and hydrogen on the 3,3′ -DPA ligand of the neighboring chains are involved in weak interactions [26].

289

290

6 The Structures of Metal–Organic Frameworks

(a)

(b)

(c)

Figure 6.8 (a) View of the one-dimensional sinusoidal chain architecture of [Cu(3,3′ -DPA)(CH3 OH)(NO3 )2 ]n . (b) One sinusoidal chain in [Cd(3,3′ -DPA)(NO3 )2 (H2 O)2 ]n . (c) A one-dimensional zigzag chain in [Cu(hfac)2 (3,3′ -DPA)1.5 (NO3 )2 ]n . Source: Zaman et al. [26]. © 2002 American Chemical Society.

In complex [HgI2 (bpmbpt)] (bpmbpt = 2,6-bis(4-pyridinylmethyl)-benzo[1,2-c:4, 5-c′ ]dipyrrole-1,3,5,7(2H,6H)-tetrone), each semirigid ditopic “Z” type ligand coordinates to two Hg(II) ions and each Hg(II) ion is coordinated by two bpmbpt to construct a 1D zigzag chain. These zigzag chains are running in perpendicular directions to interweave in a “two-over/two-under” fashion to give the cloth-like sheet structure (Figure 6.9). At each Hg(II) ion, the two coordinated arms of the ligands pass either both above or both below two other perpendicular 1D chains [27]. Another example of entanglement of zigzag chains was found in complex [Zn(bac)2 (bpp)]⋅1.5H2 O (bac = 1-benzoylacetone; bpp = 1,3-bis(4-pyridyl)propane). The flexible zigzag chain is obtained by two bac ligands chelating one Zn(II) to form a linear unsaturated coordinated building block, which is further linking by bpp ligands. These chains propagate in two nearly perpendicular directions, interweaving to 2D entangled network, which is stabled by hydrogen bonding interactions between guest water molecules and host framework [28]. Complex [Zn(phen)(sdc)] (phen = 1,10-phenathroline; H2 sdc = trans-stilbene4,4′ -dicarboxylic acid) is an unprecedented example of 3D entanglement of zigzag 1D MOFs containing permanent microporosity. In the structure, the quasi-linear sdc ligands connect Zn(II) ions to generate infinite 1D zigzag chains, which arrange

6.2 One-dimensional MOFs

Figure 6.9 View of “two-over/two-under” interwoven 2D network in [HgI2 (bpmbpt)]. The independent chains are distinguished by different colors. Source: Li et al. [27]. © 2003 Royal Society of Chemistry.

around the crystallographic fourfold axes and propagate along four noncoplanar directions, making the chains to interweave in a complicated fashion to yield a 3D framework with open channels (Figure 6.10). The zigzag shape plays a key role for the 2D interwoven motif. All the aromatic rings are involved in π· · ·π interactions, which are the main forces to organize the chains in position and stabilize the 3D entanglement [29].

6.2.3

Helical Chains

Helicity and chirality are essential elements of life. Many biopolymers such as DNA and peptides have helical structures. For better realization of the origin of the helicity of biopolymers, tremendous efforts have been devoted to the design and construction of helical MOFs. Therefore, many helical metal–organic architectures including the single-, double-, triple-, and multiple-stranded helix have been constructed [30, 31]. Helical chain MOFs can be assembled from chiral or achiral building blocks. When achiral building blocks are used, usually both right- and left-handed helices have been obtained in equal amounts as racemates. In other cases, spontaneous resolution into enantiomeric chiral crystals has been achieved. A pair of homochiral 1D-helical MOFs has been obtained by spontaneous resolution from achiral substrates. Reaction of cucurbit[5]uril with nitrate salts of dysprosium results in a 1D linear chain. In the presence of hydroquinone, enantiomorphs [Dy(H2 O)4 Q[5]](C6 H6 O2 )(NO3 )3 ⋅7H2 O are produced. For these two compounds, cucurbit[5]uril serving as a bidentate ligand coordinates to one Dy3+ ion center, while each Dy3+ ion is embraced by two bidentate ligands, which

291

292

6 The Structures of Metal–Organic Frameworks

C4

b

c a

(a)

a

(c)

a

(b)

c

b

b

c a (d)

Figure 6.10 (a) Views showing four 1D zigzag chains arranged around a fourfold axis. (b) The entanglement of the chains. (c) The 2D warp-and-woof weave sheet. (d) The 3D entanglement of the 2D sheets. Source: Cheng et al. [29]. © 2007, Royal Society of Chemistry.

is similar to the blades in a twofold propeller, leading to helical D/L chirality. In each isomer, the metal centers have the same D- or L-configuration, leading to the homochiral right- or left-handed helical chain (Figure 6.11). Each turn of the helix contains six units of cucurbit[5]uril molecules and Dy3+ ions. The high affinity of the hydroquinone molecule to cucurbit[5]uril and its steric effect induce an appropriate coordination mode of Dy3+ ion, which is essential for the formation of the chiral helix [32]. By changing the heavy lanthanides with calcium cations and hydroquinone inducer with p-hydroxybenzoic acid, another homochiral 1D-helical coordination polymer has also been assembled [33]. Similarly, homochiral crystallization of single-stranded helical MOFs has been achieved by the structure of auxiliary ligands or spontaneous symmetry breaking. Three MOFs with homochiral single-stranded helical chains were assembled by using V-shaped bridging ligand dicyanamide and chelate side-chain auxiliary ligands. When homochiral auxiliary ligands are employed, the origin of chirality of the polymers stems from these auxiliary ligands; whereas, the homochirality feature stems from the spontaneous symmetry breaking in crystallization from a stirred solution, when the achiral ligand was used [34].

6.2 One-dimensional MOFs

010A 02W 04W

09A

01W Dy1 03W 01 02

05

04

03

O10 O9

Dy1

O1 O2

06

07

09

08

01WA

010

∆ chirality

Dy1A 04WA

02A 01A 03WA

Figure 6.11 Illustration of the chiral environment around Dy and the homochiral right-handed helical structure. Source: Chen et al. [32]. © 2012 Royal Society of Chemistry.

A 1D single-strand helical chain has been constructed from μ2 -N,N ′ -bga and 𝜇 1 -(η2 -N,O)-pzc ligands (bga = benzoguanamine, Hpzc = pyrazine-2-carboxylic acid). The formation of a helical chain is attributed to a favorable combination of the skewed conformer of bga and the suitable geometry around the N—Ag—N bond, while the pzc auxiliary ligands bound to the Ag(I) ions point away from the helical chain and do not contribute to the extension of the 1D helical chain to any higher dimensionalities. The alternate left- and right-handed helical chains link together through complementary N—H· · ·N hydrogen bonds to the form a racemic 2D supramolecular sheet. The adjacent sheets interdigitated with each other to generate a 3D supramolecular framework through π· · ·π stacking between pyrazinyl rings. When H2 pzdc (H2 pzdc = pyrazine-2,3-dicarboxylic acid) was used to replace pzc, another 1D U-shaped chain has been assembled instead. In each chain, the [Ag2 (bga)2 (pzdc)(H2 O)] subunits are staggered to produce large interstices between adjacent bga ligands, which are occupied by adjacent identical chains through mutual interdigitation. As a result, a 1D → 2D interdigitated sheet is formed. The 2D sheet further interdigitates with its opposite one to form a 2D double sheet. Hydrogen bonds play a significant direction role in the formation of such interesting interdigitated sheet. The structural differential between the two 1D chain structures is caused by the different auxiliary ligands and different coordination modes of bga ligand (Figure 6.12) [35]. Besides these simple single helical strands, there are a few examples of meso-helix MOFs. For example, the 1D coordination polymer {[Cu(𝜇-pydz)2 ][PF6 ]}∞ prepared from achiral dinuclear [Cu2 (H3 CCN)2 (𝜇-pydz)3 ][PF6 ]2 (pydz = pyridazine) exhibits a rare meso-helix [36]. Four flexible bis(imidazole) ligands-based MOFs afford similar meso-helical structures, in spite of their different crystal space groups and ligand sets. In all these compounds, neighboring metal centers are bridged together via trans bis(imidazole) ligands to construct the

293

294

6 The Structures of Metal–Organic Frameworks

Ag2O 1D helical chain

1D

2D interdigitation

Figure 6.12 Preparation of {[Ag(bga)(pzc)]⋅0.5H2 O}n and [Ag2 (bga)2 (pzdc)(H2 O)]n , and their simplified structural motifs. Source: Sun et al. [35]. © 2012 Royal Society of Chemistry.

meso-helical chains, which contain both left- and right-handed helical loops in each chain and display a “∞” shape [37]. The reaction of racemic macrocyclic compounds [Ni(α-rac-hmta)](ClO4 )2 (containing equal amounts of SS and RR enantiomers) (hmta = 5,5,7,12,12,14-hexamethyl-1,4,8,11-tetraazacyclotetradecane) with K[Ag(CN)2 ] yields a 1D meso-helical chain, which is constructed by [Ni(f RR-hmta)][Ag(CN)2 ]2 enantiomers alternately connecting with [Ni(f -SS-hmta)] [Ag(CN)2 ]2 enantiomers through intermolecular argentophilic interactions (Figure 6.13). The 1D helical chains are further linked through the interchain hydrogen bonds to form a 2D network. When chiral [Ni(α-SS-hmta)](ClO4 )2 and [Ni(α-RR-hmta)](ClO4 )2 units were used, two supramolecular stereoisomers have been obtained, which display homochiral right-handed and left-handed helical chain motif, respectively, and the 1D homochiral helical chains are connected by the interchain hydrogen bonds to form a 3D structure [38]. The divergent “Σ-shaped” organic spacers together with linear or trigonal metal nodes tend to give rise to the helical skeleton. A double-stranded helix has been assembled from oxadiazole-bridging ligands (containing two parallel symmetric 3-pyridinecarboxylate arms that are perpendicular to the basal 2,5-bisphenyl-1,3,4-oxadiazole to form a “Σ-shaped” spacer) (Figure 6.14a) and Ag(I). The Ag(I) ions are bridged by the “Σ-shaped” ligands to form a 1D helical chain. Two right-handed single strands are offset by one-half of a pitch and intertwine together to generate a right-handed double-stranded helix through two sets of interchain π· · ·π interactions between the oxadiazole–phenyl and phenyl–phenyl pairs (Figure 6.14b). The counterions are located between the double-stranded helical chains and linked together through the weak hydrogen bonding. Changing the BF4 − and ClO4 − counterions with larger and polar SO3 CF3 − , another corresponding helical structure with both M- and P-helical chains is presented. The left- and right-handed chains are arranged alternatively and linked through CF3 SO3 − anions into a 1D ladder-like chain (Figure 6.14c) [39].

6.2 One-dimensional MOFs

Ag1 Ag2

Figure 6.13 The 1D meso-helical structure constructed via argentophilic interactions. Source: Zheng et al. [38]. © 2008 American Chemical Society.

N

N C O O

C O N N

O

O

(a)

(b)

(c)

Figure 6.14 (a) The structure of “Σ-shaped” ligand. (b) The double-stranded helix formed by the intertwining of two single helical chains. (c) The CF3 SO3 − anions are located between left- and right-handed helical chains. Source: Dong et al. [39]. © 2006 American Chemical Society.

For the double helix, if the two strands are of the same chirality and all parallel double helices are also of the same chirality, the polymer would be noncentrosymmetric and chiral even if it contains no chiral molecular unit in the helices. For example, a homochiral infinite double-stranded helical polymer {[Zn2 I4 (tmdp)2 ]n ⋅[Zn2 I4 (tmdp)2 ]n } has been prepared by hydrothermal reaction of

295

296

6 The Structures of Metal–Organic Frameworks

Figure 6.15 (a) The intertwined infinite double-helical structure. (b) Parallel packing of adjacent double helices with the same chirality. Source: Han et al. [40]. © 2007 American Chemical Society.

(a) c 0

a

b

(b)

4,4-trimethylenedipyridine (tmdp) with ZnI2 , which consists of two single-stranded infinite helices of the same M handedness. The achiral flexible tmdp ligand shows two different conformations in each [Zn2 I4 (tmdp)2 ]n unit, which plays an important role in the formation of the double-helical structure. The two strands of this double helix are arranged together through intermolecular forces, van der Waals interactions, as there are no obvious hydrogen-bonding and π· · ·π-stacking interactions. The adjacent chiral double helices are of the same chirality and packed in a parallel fashion, leading to a homochiral solid (Figure 6.15) [40]. Two unique DNA-mimic double-helical chains isomers, L-{Cu4 [N(CN)2 ]2 (hmp)4 (CH3 COO)2 ⋅CH3 CN}∞ and D-{Cu4 [N(CN)2 ]2 (hmp)4 (CH3 COO)2 ⋅CH3 CN}∞ have been assembled from dicyanamido (dca, N(CN)2 − ), 2-(hydroxymethyl)pyridine (Hhmp), and Cu(CH3 COO)2 in solvent CH3 CN/CH3 OH, where the Cu—O bonds between the two strands replace the hydrogen-bonding interactions in DNA. In their structures, four 𝜇 3 -O atoms from four hmp− ligands connect four Cu cations to form a [Cu4 O4 ] cubane, which are bonded alternately by the twisted dca ligands in 𝜇 2 -1,5 mode to generate an infinite right-handed helical chain. These helices are paired and linked via Cu–O coordination to a unique double-helical structure. They exhibit homochirality in the entire crystal. By changing the solvent and supplementary ligand, the structures can be tuned. When water is used as solvent and Cu(ClO4 )2 is used instead of Cu(CH3 COO)2 , the Cu(II) ions are connected by Hhmp and dca to form dinuclear copper clusters, which are further linked by hydrogen bonds between the protonated Hhmp ligands and the additional dca ligands to

6.2 One-dimensional MOFs

OH N N(CN)2– Cu2+

Left/right

Left/right

Right

Left

Figure 6.16 Self-assembly of DNA-mimic homochiral 1D helical Cu(II) chain from achiral flexible ligand. Source: Zhang et al. [41]. © 2016 American Chemical Society.

generate a 1D meso quasi-double-helical chain Cu[N(CN)2 ]2 (Hhmp) (Figure 6.16). Changing solvent from H2 O to CH3 CN/H2 O, the Cu(II) ions are only connected by one dca ligand alternately, leading to 1D single helices of {Cu[N(CN)2 ]2 (Hhmp)}∞ ; however, the left- and right-handed helical chains are coexistent and no optical activity was observed [41]. A 1D meso-helical structure, which consists of two adjacent hydrogen-bonded three-stranded single-helical (P and M) chains, has been constructed from flexible multidentate hydrogenated Schiff base ligand tris{2-[(4′ -pyridylmethyl)amino] ethyl}amine (hsb-1). Intermolecular CH (or NH)/Cl hydrogen bonds play important roles in the formation of the triple helices and meso-helical structure. In the structure, one type of the two Cd(II) centers are chelated and held by the internal nitrogen of ligand to form a subunit. Each subunit with three monodentate pyridyl groups is linked by another Cd(II) ions to build an infinite helical chain, with an uncoordinated pyridyl group extending outward from the helical microchannel. Three left-handed helical chains and three right-handed helical chains are wrapped together respectively by hydrogen bonding, resulting in a meso-helical structure (Figure 6.17), which are further arranged to a 2D framework [42]. By changing the metal ions, 2D layer networks composed of alternate right-handed and left-handed helical chains by sharing metal ions have also been obtained. The first cases of actinide triple helices, UO2 (dbsf)(phen) and [UO2 (dbsf) (phen)]⋅H2 O, have been synthesized with the combination of a semirigid V-shaped 4,4′ -dicarboxybiphenyl sulfone (H2 dbsf) and a bulky rigid aromatic base phen (Figure 6.18). They are identified to be π· · ·π stacking directed conformational supramolecular isomers, which are dependent on the pH conditions. Neutral condition tends to give UO2 (dbsf)(phen) with intra-chain π· · ·π stacking, while acidic solution tends to generate [UO2 (dbsf)(phen)]⋅H2 O with interchain π· · ·π stacking. In UO2 (dbsf)(phen), bidentate dbsf ligands with a large torsion angle interconnect the uranyl motifs to generate a chiral helix, three of which intertwine together to form unique uranyl triple-stranded helices. The phen ligand also plays

297

298

6 The Structures of Metal–Organic Frameworks

b

b ac

a

(a)

c

(b)

Figure 6.17 (a) Three left-handed helical chains are wrapped together by hydrogen bonding to form a triple helix. (b) One-dimensional meso-helical structure built by the arrangement of adjacent three-stranded single-helical (P and M) chains through NH/Cl hydrogen bonding. Source: Wen et al. [42]. © 2012 Royal Society of Chemistry.

cb a

c

b

b

a

Crystailize induced by fast cooling

Dissolve in solution at 180 °C

left-handed [UO2(dbsf)(phen)]-H2O

c a

Mother solution

Right-handed

UO2(dbsf)(phen)

Figure 6.18 A possible transformation pathway through the conformational turnover of phen motifs in actinide triple helices. Source: An et al. [43]. © 2015 Royal Society of Chemistry.

6.2 One-dimensional MOFs

a critical role in the construction of triple-stranded helices, which coordinates to uranyl centers at an out-of-the-equatorial-plane angle and is located in parallel at both of sides of the triple-stranded helices, pointing up to the helical axis. Such orientation affords the face-to-face detachment of 3.15 Å between the inter-strand phen ligands in the triple-stranded helices, indicating strong aromatic π· · ·π stacking interactions between adjacent chains of the triple-stranded helices. One significant difference between the two complexes is the orientation of aromatic phen motifs. A subtle change on the orientation of phen in [UO2 (dbsf)(phen)]⋅H2 O expands the face-to-face detachment of phen motifs to a larger distance (7.03 Å), enabling the adjacent triple-strand helices to be inserted into the space via phen located at the middle of the aromatic space. As a result, adjacent triple-stranded helices arrange in mirror-image forms and lead to a mesomeric effect. Interestingly, crystal [UO2 (dbsf)(phen)]⋅H2 O can transfer to UO2 (dbsf)(phen) by incubating it in the mother solution at 180 ∘ C for 24 hours, followed by fast cooling to room temperature. The pathway of kinetics transformation concerning a conformational rearrangement of organic base is assigned to dissolution–crystallization process [43]. A 1D silver MOF with the unprecedented 1D twofold interpenetrating motifs constructed by 1D triple helical chains has been prepared by the reaction of AgPF6 and the asymmetric ligand 1,6-dihydro-2-methyl-6-oxo-(3,4′ -bipyridine)-5-carbonitrile. In the structure, each ligand connects two Ag(I) ions to form infinite right-handed helical chains, three of which are linked together by the Ag(I) centers to form triple helix with lengthening nanosized cages. Two parallel triple helical chains entangle together to compose a unique twofold interpenetration framework. The nanocage is equally divided into two ball-like cages with a diameter of about 10 Å. Through interchain hydrogen-bonding interaction, a 3D supramolecular motif is obtained, which has triangle-like 1D channels. The PF6 − anions can be divided into two types: the unencapsulated ones act as counteranions; the other ones encapsulated in cages play a templating role for directing the formation of the twofold interpenetration framework [44]. Interesting interlocked quintuple helices are built from C2 -symmetric bipyridyl ligands and linear metal-connecting points. In [Ni(acac)2 (envp)]⋅3CH3 CN⋅6H2 O (envp = (S)-2,2′ -diethoxy-1,1′ -binaphthyl-6,6′ -bis(4-vinyl-pyridine)), the Ni-(acac)2 units are bridged by binaphthyl backbones of envp to form an infinite helical chain (Figure 6.19a). Five infinite helical chains associate in parallel to form quintuple helices with a tetragonal nanotube (with an opening of ∼2 × 2 nm2 ) (Figure 6.19b). Each quintuple helix further intertwines with four other helices to give an interlocked architecture, leading to a 3D chiral framework with the eclipsing of nanotube corners (Figure 6.19c). Partially eclipsed nanotubes have open channels of 1.7 × 1.7 nm2 in dimensions, which are filled with CH3 CN and water guest molecules (Figure 6.19d). The framework is stabilized by two types of strong π· · ·π stacking interactions among the intertwined vinylnaphthyl groups. Modifying the ligand with chiral crown ethers also provides a similar tubular architecture, with the tubular channels decorated by chiral crown ethers [45].

299

300

6 The Structures of Metal–Organic Frameworks

b a

(a)

(b)

(c)

(d)

Figure 6.19 (a) Lefthanded 41 helical chain. (b) Parallel association of five helices into a chiral nanotube. (c) A schematic illustrating the interlocking of adjacent helical chains. (d) A spacefilling model showing the open channels within the 3D chiral framework of [Ni(acac)2 (envp)]⋅3CH3 CN⋅ 6H2 O. Source: Cui et al. [45]. © 2003 American Chemical Society.

A 3D chiral framework, [Cd(bpea)(phen)2 ] (bpea = biphenylethene-4,4′ dicarboxylic acid) built from achiral components, exhibits a remarkable assembly of ninefold interlocked homochiral helices. The Cd center is coordinated by two chelating phen ligands to form [Cd(phen)2 ]2+ molecular corners, which are linked by linear spacers to construct an infinite helical chain. Each helical chain is chemically independent but physically interwoven with the other chains in all directions. Therefore, each helical chain is interlocked by 8 equiv polymeric units to give an interlocked 3D chiral network, which can be described as an infinite interlocked array originating from ninefold interwoven homochiral helices (Figure 6.20) [46]. The framework is stabilized by strong π· · ·π stacking interactions between the interwoven aryl groups.

6.2.4

Double Chains

Double chains are generated from flexible ligand backbone in bent/gauche conformations, which are interconnected through metal nodes to form macrocycle. The double-chain motifs contain cavities/pores as in higher-dimensional MOFs, which can accommodate guest molecules, so they are of special interest in the coordination networks. Complex [{Cu(dpds)2 (H2 O)}⋅2NO3 ⋅3H2 O]n (dpds = 4,4′ -dipyridyl disulfide) displays a double-stranded chain architecture. In the double chains, each copper center is bridged by four dpds ligands, and two dpds ligands cooperate with two copper atoms to build distorted square metallacycles with the dimensions of the cavity c. 5 × 5 Å2 . The two pyridyl rings of the dpds ligand are almost perpendicular. Two

6.2 One-dimensional MOFs

Figure 6.20 A schematic illustration of the ninefold interlocked homochiral helices of [Cd(bpea)(phen)2 ]. Source: Wang et al. [46]. © 2004 John Wiley & Sons.

such double-stranded chains lie reversibly and are connected into a ladder structure through hydrogen bonds, which are further linked into a bilayer through hydrogen bonds between free NO3 − anions and the lattice water molecules [47]. Two exotic 1D molecular tapes with serial and parallel cyclic azido-bridged eight membered copper rings have been assembled from 2-benzoylpyridyl (bzp) and azido–copper in solution with methanol, either with or without the addition of acetonitrile. [Cu4 (N3 )8 (CH3 CN)3 (bzp)2 ]n consists of neutral 1D azido–copper chains with serial octanuclear repeating units, while [Cu5 (N3 )10 (bzp)2 ]n is another interesting 1D molecular tape with parallel eight-membered copper rings [48]. A ribbon polymer [Cd(envp)2 (ClO4 )2 ]⋅11EtOH⋅6H2 O, displaying high permanent porosity and undergoing SCSC transformation induced by solvent exchange, has been obtained from envp and Cd(ClO4 )2 ⋅6H2 O. Adjacent Cd(II) centers are bridged by two envp ligands to form 46-membered rhombic macrocycles with an opening of approximately 19.3 × 19.3 Å2 (Figure 6.21a). Neighboring 1D double chains interdigitate to each other to form a 2D layer via strong π· · ·π stacking interactions (Figure 6.21b). The 62 operation generates an ⋅ ⋅ ⋅ABCABC⋅ ⋅ ⋅ stacking pattern for all the 1D chains. The 1D chains in the adjacent layers along the c axis are rotated by 120∘ with respect to each other, generating chiral, 1D pseudo-hexagonal channels with dimension of 16.77 Å (Figure 6.21c). The effective volume for the inclusion is 4900.6 Å3 per unit cell, which is 53.6% of the crystal volume. Interestingly, the polymer undergoes reversible SCSC transformation in the presence of benzene vapor. The most significant structural difference lies in the relative

301

302

6 The Structures of Metal–Organic Frameworks

b a

b a (a)

(b)

(c)

(d)

Figure 6.21 (a) A view of the packing of the 1D double chains. (b) Space-filling view of 1D hexagonal channels. (c) Schematic and space-filling views of the 62 -helical arrangement of the 1D chains. (d) Schematic and space-filling views of the 61 -helical arrangement of the 1D chains after benzene exchange. Source: Wu et al. [49]. © 2005 John Wiley & Sons.

orientation of the ethoxy groups of the envp ligands, the coordinating perchlorate groups, and the 1D polymeric chains. As a result, the 1D double chains stack in an ⋅ ⋅ ⋅ ABCDEFABCDEF ⋅ ⋅ ⋅ pattern with the macrocycle planes slightly tilted away (by 0.2∘ ) from the ab plane (Figure 6.21d) [49]. Two homochiral porous cadmium(II) MOFs based on 1D double chains have been synthesized from (S)-2,2′ -dimethoxy-1,1′ -binaphthyl-3,3′ -bis(4-vinylpyridine) (mnvp, similar to envp) under slightly different conditions. Reaction of Cd(ClO4 )2 ⋅ 6H2 O and mnvp in DMF/o-C6 H4 Cl2 /EtOH affords [Cd(mnvp)2 (ClO4 )2 ]⋅3EtOH⋅ H2 O, which displays a 1D double chains structure. Two mnvp ligands bond two Cd(II) centers to form a 38-membered macrocycle, which are propagated by the Cd(II) nodes into 1D double chains. These chains are further stacked together via π· · ·π interactions, generating large chiral open channels of dimensions of 9.9 × 12.2 Å2 . [Cd(mnvp)2 (ClO4 )(H2 O)](ClO4 )⋅1.5(o-C6 H4 Cl2 )⋅3EtOH⋅6H2 O has been prepared when a small amount of water was added into the reaction mixture of [Cd(mnvp)2 (ClO4 )2 ]⋅3EtOH⋅H2 O, which also shows a 1D double chain structure built up from 38-membered macrocycles. However, the Cd(II) centers coordinate to an aqua ligand instead of a perchlorate anion in one of the axial positions. Strong π· · ·π interactions link the 1D macrocycles into a 3D open framework with large channels along the 𝛼 axis (8.0 × 16.0 Å2 ) and the 110 direction (6.6 × 16.7 Å2 ), which is occupied by perchlorate anion and dichlorobenzene, ethanol, and water molecules [50]. Another two 1D double chains have been constructed by a flexible and different N-donor Ligands. In [Cd(ppbe)(phen)] (ppbe = dicarboxylate (3-carboxyl-phenyl)-

6.2 One-dimensional MOFs

(a) c b

(c)

(b)

(d)

Figure 6.22 (a) View of the 1D double chain of [Cd(ppbe)(phen)]. (b) The 2D supramolecular layer of [Cd(ppbe)(phen)]. (c) The 3D supramolecular architecture via π⋅⋅⋅π interactions in [Cd(ppbe)(phen)]⋅2H2 O. (d) The 2D supramolecular layer constructed in [Cd2 (ppbe)2 (biim-2)]⋅H2 O. Source: Niu et al. [51]. © 2012 American Chemical Society.

(4-(2′ -carboxyl-phenyl)-benzyl) ether), each ppbe showing 𝜇2 -𝜂 1 :𝜂 1 coordination modes and a “U-shaped” conformation connects four Cd(II) ions to form the double chain, attaching with the bidentate phen ligands (Figure 6.22a). The chains are further extended into the 2D supramolecular layer through π⋅⋅⋅π interactions between pyridyl rings and phenyl rings of the phen ligands from adjacent chains (Figure 6.22b). [Cd(ppbe)(phen)]⋅2H2 O shows a similar 1D double chain; however, the coordination mode of carboxylate groups is different. One carboxylate group of the ppbe exhibits a bidentate chelating mode (𝜇 2 -𝜂 1 :𝜂 1 ), while the other one shows a tridentate coordination mode (𝜇 3 -𝜂 1 :𝜂 2 ). There are three kinds of π· · ·π stacking interactions associated with phenyl rings of ppbe anions and pyridyl rings of phen ligands in the adjacent layers, and the intermolecular π· · ·π interactions transform these 1D double chains into a 3D supramolecular architecture (Figure 6.22c). When phen was replaced by biim-2 (biim-2 = 1,2-bis(imidazol-1′ -yl)ethane) ligand, polymer [Cd2 (ppbe)2 (biim-2)]⋅H2 O has been obtained, which also consists of 1D double chains. Two Cd(II) atoms are bridged by two ppbe anions with two carboxylate groups in 𝜇 2 -𝜂 1 :𝜂 1 modes and two ppbe anions with one of two carboxylate groups in a 𝜇 3 -𝜂 1 :𝜂 2 coordination mode and the other one in a 𝜇 4 -𝜂 2 :𝜂 2 coordination mode to form a binuclear [Cd2 (ppbe)4 ] unit, which is further linked by biim-2 ligands to construct the 1D double chains. Adjacent chains are linked by π· · ·π interactions to a 2D supramolecular layer (Figure 6.22d) [51]. Two double-chain MOFs composed of two different kinds of M2 L2 macrocyclic rings or two kinds of 1D chains have been achieved from the mixed ligands of 1,3-di(1H-imidazol-4-yl)benzene (imb) and 1,4-phenylenediacetic acid (H2 pda) (Figure 6.23). In the structure of [Ni(imb)(pda)(H2 O)], two Ni(II) atoms connect two imb or two pda2− ligands to generate two different kinds of M2 L2 macrocyclic rings, which are further joined together by the coordination of Ni(II) to

303

304

6 The Structures of Metal–Organic Frameworks

(a)

(b)

Figure 6.23 (a) 1D double chain of [Ni(imb)(pda)(H2 O)] is composed of two different kinds of M2 L2 macrocyclic rings. (b) 1D double chain of [Cd(imb)(pda)] is formed by cross-linking together two kinds of 1D chains. Source: Liu et al. [52]. © 2019 Elsevier.

construct an infinite 1D double chain. These double chains are further extended into a 3D supramolecular architecture through hydrogen-bonding interactions. In [Cd(imb)(pda)], each imb bridges two Cd(II) to afford an infinite 1D chain, and each pda2− bonds two Cd(II) using its two carboxylate groups each with (𝜇 1 -𝜂 1 :𝜂 1 )-( 𝜇 1 -𝜂 0 :𝜂 0 )-pba2− mode to generate another 1D chain. Such two kinds of 1D chains cross-link together via the coordination of Cd(II) to form a 1D double chain, which is further extended into a 3D supramolecular architecture through hydrogen-bonding interactions. The different structures of the two double chains are ascribed to the distinct metal centers [52]. A 1D chain containing nanotubes has been observed in complex [Ag7 (tpst)4 (ClO4 )2 (NO3 )5 (DMF)2 ] (tpst = 2,4,6-tris[(4-pyridyl)methyl-sulfanyl]-1,3,5-triazine). In the structure, each tpst ligand coordinates to three silver(1) centers through the N atoms of the three pyridyl groups, and each silver(1) ion is in turn surrounded by the pyridyl group of another tpst ligand to form an [Ag3 (tpst)2 ] nanosized ring. Two adjacent rings are connected by Ag—N and Ag—S bonds from one N and one S atom of the trithiocyanuric spacer to form the basic nanosized tube unit with the dimensions of 1.34 × 0.96 × 0.89 nm3 , which accommodates two DMF molecules and two perchlorate ions. The tube units are linked to an infinite chain via share Ag(1) by Ag—N and Ag—S bonds. All tpst ligands act as tetradentate ligands but exhibit two kinds of coordination mode: one type coordinates to four silver atoms through three nitrogen atoms of different pyridyl groups and a sulfur atom of the thioether moiety, and the other through three pyridyl nitrogen atoms and one nitrogen atom of the triazine ring. The silver centers exhibit two kinds of coordination environment: one is the normal linear coordination mode of AgN2 and the other is tetracoordinate with a slightly distorted square-planar AgS2 N2 unit (Figure 6.24) [53].

6.2.5

Ladder-like Chains

Ladder-like chains can be obtained by metal ions as nodes and spacer ligands as rails and rungs to form “T-shaped” building blocks. The cavities of ladders are influenced

6.2 One-dimensional MOFs Ag(1) S N Cl O

Ag(2A) Ag(4)

Ag(3A)

Ag(3)

Ag(4A) Ag(2)

M

M

Ag(1A)

(a)

(b)

Figure 6.24 (a) View of the nanosized tube unit. (b) View of the 1D chain composed of nanotubes. Source: Hong et al. [53]. © 2000 John Wiley & Sons.

I

VIII

II

III

IV

IX

V

VI

VII

X

Figure 6.25 Schematic representation of different types of common molecular ladder motifs. Different kinds of spacers are shown in different colors. Metallophilic bonds are represented as dashed lines. Source: Based on Leong and Vittal [16].

by the length, shape, and orientation of spacer ligands. Hence, a desired ladder structure can be built by judicious choice of these components. A great number of molecular ladders have been reported [16, 54–56] and the common ladder motifs are listed in Figure 6.25. A 1D infinite non-interpenetrated molecular ladder with large cavities has been constructed from 4,4′ -bis(imidazol-1-methyl)-biphenyl (bimb) and Pb(II). In complex {[Pb(bimb)1.5 (NO3 )2 ](DMF)}n , all bimb ligands adopt a trans conformation and link adjacent Pb(II) centers to form a 3 : 2 ligand/metal infinite non-interpenetrated molecular ladder. Two such ladders are juxtaposed to generate sheets, which stack in a close-packed stair-like manner [57]. Another two novel infinite1D molecular ladder complexes are based on a conjugated bispyridyl-based Schiff base ligand. In [M2 (bpdb)2 (OAc)⋅2(MeOH)]n (bpdb = 1,4-bis(4-pyridyl)-2,3-diaza-1,3-butadiene, M = Zn(II) or Mn(II)), each zinc or manganese ion is coordinated to two pyridyl nitrogen atoms from two bpdb ligands, extending to a linear polymeric chain.

305

306

6 The Structures of Metal–Organic Frameworks

Figure 6.26 (a) View of the stacking of 1D molecule ladder chains in [M2 (bpdb)2 (OAc)⋅2(MeOH)]n . Source: Zhang et al. [58]. © 2006 American Chemical Society. (b) View of the 1D ladder-like chain in a triazole–Ag MOF. Source: Wang et al. [59]. © 2007 American Chemical Society. (a)

(b)

One of the acetate anions strongly chelates to the metal center with two Zn—O bonds, while two other acetate anions bind two distinct metal centers and serve as bridges to connect two infinite bpdb-M chains to the 1D molecular ladder. These molecule ladders stack in a parallel fashion to generate a layer. Such layers arrange in a nearly perpendicular direction, resulting in a 3D array (Figure 6.26a) [58]. The self-assembly of silver(I) salts with bitopic triazole 4-(2-pyridinyl)-1,2,4-triazole produced another 1D molecular-ladder MOFs (Figure 6.26b). Different silver(I) centers are bridged by triazole ligands, water molecules, and perchlorate anions, resulting in the ladder motif, which are linked to a 3D supramolecular network via hydrogen bonds [59]. POM-based inorganic–organic hybrid compounds H2 [Cu(en)2 H2 O]8 [Cu(en)2 ]3 [{(α-SiW11 O39 )Ce(H2 O)(η2 ,𝜇-1,1)-CH3 COO}4 ]⋅22H2 O (en = 1,2-ethylenediamine) represent the first example in POM chemistry of a 1D ladder-like polyanion chain consisting of monovacant Keggin-type polyanions, lanthanide complexes, and transition metal complexes. In such structure, [Cu(en)2 H2 O]2+ is grafted on [(α-SiW11 O39 )Ce(H2 O)(η2 ,𝜇-1,1)-CH3 COO]6− polyanions and further connected by [Cu(en)2 ]2+ groups, constructing the ladder-like chain [60]. Another 1D ladder-like POM-based MOF [Cu(bipy)]2 [HPMo12 O40 ] has been constructed via Keggin-type [PMo12 O40 ]3− anion bridging two single Cu-bipy chains (Figure 6.27). Such ladder chains are further connected via weak interactions to obtain a 2D structure [61]. Complex [Cd2 {(R)-Hdmpa}2 {(S)-Hdmpa}2 (dpe)3 ]n (H2 dmpa = 6,6′ -dimethyl-1,1′ biphenyl-2,2′ -dicarboxylic acid, and dpe = 1,2-di(4-pyridyl)ethylene) is an example of ligand self-discrimination in an assembly process, in which dpe connects the [Cd{(R)-Hdmpa}{(S)-Hdmpa}] building units to generate a 1D ladder structure [62]. [Ni(bpp)1.5 (H2 O)(OH-bdc)]n displays an interdigitated structure from the interpenetrated long arms of 1D molecular ladders. The bpp components adopt two types

6.2 One-dimensional MOFs

cb a

(a)

(b)

Figure 6.27 (a) The ladder-like chain with [PMo12 O40 ]3− anion as “middle rails.” (b) Schematic view of the bitrack Cu(bipy) chain-modified POMs. Source: Sha et al. [61]. © 2009 American Chemical Society.

of conformation with different coordination modes: one possesses a TG conformation bridging two Ni(II) centers to form the 1D neutral molecular ladder; the other shows a TT conformation and serves as a terminal ligand and as a lateral arm of the ladder. The long arms of TT bpp are threaded into [Ni4 (bpp)2 (OH-bdc)2 ] rectangles of an adjacent molecular ladder. Each rectangle is oppositely penetrated by two lateral arms from different molecular ladders, resulting in a polythreaded polymeric structure [63]. A 1D ladder and a zigzag linear chain are existing simultaneously in complex [Cd(bppd)(NO3 )2 (CH3 OH)2 ⋅Cd2 (bppd)3 (NO3 )4 ]⋅{4(HCCl3 )⋅2H2 O}n (bppd = N,N ′ -bis(4-pyridylmethyl)-pyromellitic diimide) (Figure 6.28). The ladder-like 1D

Figure 6.28 The zigzag linear chain packing between the 1D ladder-like chains. Source: Li et al. [64]. © 2012 Royal Society of Chemistry.

307

308

6 The Structures of Metal–Organic Frameworks

chain [Cd2 (bppd)3 (NO3 )4 ] is constructed by the Z-mode conformation of bppd acting as a bridge that links two Cd2+ ions, which in turn connect three different ligands; while the zigzag linear chain [Cd(bppd) (NO3 )2 (CH3 OH)2 ] is formed by the bppd ligand bridging two Cd2+ ions, which in turn connect two different ligands. A large simple 1D and a rare polycatenated 3D molecular ladders have been assembled from T-shaped building blocks of N,N ′ -bis-(4-pyridinylmethylene)1,5-naphthalenediamine (nbpy4). In both molecular ladders [Co2 (nbpy4)3 (NO3 )4 ]⋅ solvents and [Cd2 (nbpy4)3 (NO3 )4 ], metal centers (Co or Cd) are hepta-coordinated by three nbpy4 ligands to form a “T-joint” at the metal center, whose remaining coordination sites are occupied by two bidentate nitrates. The 1D ladders are constructed from the T-shaped building blocks, with nbpy4 as rungs and side rails. In [Co2 (nbpy4)3 (NO3 )4 ]⋅solvents, ladders run parallelly and stack in an offset way, with disorder CH2 Cl2 and MeOH molecules accommodating inside the cavities (Figure 6.29a). In [Cd2 (nbpy4)3 (NO3 )4 ], every ring of the ladder is interlocked by two parallel ladders above and by two parallel ladders below. Such staggered interpenetration by the inclined ladders leads to the 1D to 3D structure (Figure 6.29b) [65]. A hierarchically ordered homochiral MOF based on a long chiral bipyridine bridging ligand exhibits simultaneous interlocking and interpenetration of 1D ladders. In [Cu3 (ddbbe)4 (DMF)6 (H2 O)3 (ClO4 )][ClO4 ]5 ⋅10DMF⋅10EtOH⋅ 7H2 O (ddbbe = (R)-6,6′ -dichloro-2,2′ -diethoxy-1,1′ -binaphthyl-4,4′ -bis(p-ethynylpyridine)), one Cu(II) center links two adjacent ddbbe ligands to form the long edge of a rectangle, the other two Cu(II) centers form the corners of an exceptionally large rectangle of 24.8 × 48.6 Å2 , which further extends into a 1D ladder. The adjacent ladders interlock each other with strong π ⋅ ⋅ ⋅ π interactions to generate a

(a)

(b)

Figure 6.29 (a) Molecular structure of [Co2 (nbpy4)3 (NO3 )4 ]⋅solvents showing cavities hosting guest molecules. (b) Structural view of [Cd2 (nbpy4)3 (NO3 )4 ] showing fourfold catenation (left) and the simplified staggered polycatenation (right). Source: Su et al. [65]. © 2004 Royal Society of Chemistry.

6.3 Two-dimensional MOFs

Figure 6.30 (a) Interlocking of the 1D ladders to form a 2D network. (b) Interlocking of the 1D ladders that are 46.6∘ from each other. Source: Wu et al. [66]. © 2008 American Chemical Society.

(a)

(b)

2D lamellar framework (Figure 6.30a). The interlocking 1D ladders further interpenetrate with equivalent 1D ladders that are 46.6∘ from each other (Figure 6.30b). The 1D ladder motifs are thus simultaneously interlocking and interpenetrating with their neighbors to form a 3D structure with large voids that are filled with perchloride ions and solvent molecules [66].

6.3 Two-dimensional MOFs 2D MOFs, another type of low-dimensional materials, have attracted great interest in recent years, due to their unique attributes in numerous fields. 2D MOFs are constructed by four (or three) ligands coordinating to the metal ion and extending in two directions. In order to describe and understand the structures more easily and intuitively, the crystal structures of 2D MOFs can be reduced to networks (or nets). Up to now, a large number of 2D networks have been reported. The simplest 2D networks, known as regular tilings or tessellations, are composed by just one kind of regular polygon based upon triangles, squares, and hexagons (Figure 6.31).

6.3.1

Triangular-Grid Networks

In regular triangle-grid networks, six triangles meet at a node with angles of 60∘ with the corresponding Schläfli topology symbols 36 . To form such networks, the nodes must have sixfold rotational symmetry (Figure 6.31a). Few metal ions are capable of generating such a six-connected planar metal-centered node. Therefore, the regular triangular-grid networks are rare in 2D MOFs.

309

6.3 Two-dimensional MOFs

Figure 6.30 (a) Interlocking of the 1D ladders to form a 2D network. (b) Interlocking of the 1D ladders that are 46.6∘ from each other. Source: Wu et al. [66]. © 2008 American Chemical Society.

(a)

(b)

2D lamellar framework (Figure 6.30a). The interlocking 1D ladders further interpenetrate with equivalent 1D ladders that are 46.6∘ from each other (Figure 6.30b). The 1D ladder motifs are thus simultaneously interlocking and interpenetrating with their neighbors to form a 3D structure with large voids that are filled with perchloride ions and solvent molecules [66].

6.3 Two-dimensional MOFs 2D MOFs, another type of low-dimensional materials, have attracted great interest in recent years, due to their unique attributes in numerous fields. 2D MOFs are constructed by four (or three) ligands coordinating to the metal ion and extending in two directions. In order to describe and understand the structures more easily and intuitively, the crystal structures of 2D MOFs can be reduced to networks (or nets). Up to now, a large number of 2D networks have been reported. The simplest 2D networks, known as regular tilings or tessellations, are composed by just one kind of regular polygon based upon triangles, squares, and hexagons (Figure 6.31).

6.3.1

Triangular-Grid Networks

In regular triangle-grid networks, six triangles meet at a node with angles of 60∘ with the corresponding Schläfli topology symbols 36 . To form such networks, the nodes must have sixfold rotational symmetry (Figure 6.31a). Few metal ions are capable of generating such a six-connected planar metal-centered node. Therefore, the regular triangular-grid networks are rare in 2D MOFs.

309

310

6 The Structures of Metal–Organic Frameworks

(a)

Figure 6.31

(b)

2D networks based on triangles, squares, and hexagons.

Zn1

(a)

(c)

Zn2

Zn3

(b)

Figure 6.32 The structure of the trinuclear building block (a) and the architecture of the 36 tessellated 2D sheet in {[Zn3 (bdc)3 (DEF)2 ]⋅DEF}∞ (b). Source: Williams et al. [67].

Polynuclear metal centers as potential nodes of high connectivity have been employed to assemble triangle-grid networks. Reaction of Zn(NO3 )2 and 1,4-benzenedicarboxylic acid (H2 bdc) at 100 ∘ C in diethylformamide (DEF) yielded a 2D 36 tessellated framework polymer, {[Zn3 (bdc)3 (DEF)2 ]⋅DEF}∞ , which based on trinuclear zinc(II) building blocks bridged by bdc (Figure 6.32) [67]. The trinuclear zinc(II) building block consists of two five-coordinated zinc(II) centers and a six-coordinated central zinc(II) center. Three carboxylates of three bdc ligands connect each pair of zinc centers, adopting monodentate coordination to the central zinc(II) center and either monodentate or asymmetric chelating bidentate coordination to the terminal zinc(II) centers, to generate a triangular-grid network. Different nets stack along the a axis with coordinated DEF molecules on the terminal zinc(II) centers interdigitating with the triangular cavities of the adjacent sheets. The remaining space in the structure is occupied by disordered solvent molecules. 2D MOFs [Zn2 M(bpdc)3 (DMF)2 ]⋅4DMF (M = Co(II), Ni(II) or Cd(II), DMF = N,N ′ -dimethylformamide) are built from the mixtures of transition metal ions and 4,4′ -biphenyldicarboxylic acid (H2 bpdc) via solvothermal syntheses. The asymmetric unit contains a hetero-trinuclear cluster of two Zn centers in a linear array and one Co(II) center, which are bridged by six bpdc ligands through the carboxylate oxygen atoms (Figure 6.33). Two DMF molecules act as the terminal ligands to complete the tetrahedral coordination geometry of Zn centers. When the

6.3 Two-dimensional MOFs

Zn1

Co1

Zn1A

(a)

(b)

Figure 6.33 The coordination environments of the metal centers (a) and the structure of the triangular meshes in [Zn2 M(bpdc)3 (DMF)2 ]⋅4DMF (b).

trinuclear cluster is represented by solid rods and the bpdc ligand by flat ribbons, the 2D layered structures are simplified as triangular meshes. Such adjacent 2D layers are stacked in an ⋅ ⋅ ⋅ ABC ABC ⋅ ⋅ ⋅ staggering fashion, with the distance of 14.369(1) Å [68].

6.3.2

Square-Grid Networks

Square-grid (or rhombohedral) networks are the most commonly structural motifs in 2D architectures, with the corresponding Schläfli topology 44 . The proportion of metal to ligand usually is 1 : 2 with the metal centers coordinated to four different ligand molecules; therefore, the nodes necessarily have fourfold rotational symmetry (Figure 6.31b). The square-grid polymer based on bpy ligands was firstly reported by Robson and coworkers in 1990, which is interpenetrated and nonporous [69]. In 1994, Fujita et al. prepared an non-interpenetrated square-grid network {[Cd(4,4′ -bpy)2 ](NO3 )2 }n , consists of an edge-sharing, perfectly planar square with a Cd(II) ion and 4,4′ -bpy at each corner and side. The polymer contains catalytic active metal centers in the network, and it acts as a heterogeneous catalyst for cyanosilylation of aldehydes and imines (Figure 6.34a) [70, 72]. A great number of similar square-grid polymers were subsequently constructed from such bipyridine or related ligands. In polymer {[Cu(4,4′ -bpy)2 (H2 O)][Cu(2-pySO3 )3 ](NO3 )}⋅H2 O (2-pySO3 = 2-pyridine-sulfonate), Cu(II) centers are bridged by 4,4′ -bpy ligands to form a square-grid cationic layer. Adjacent layers are parallelly stacked with an ⋅⋅⋅ABAB⋅⋅⋅ motif, generating 6.78 × 6.78 Å2 rectangular channels, which are threaded by a chain of [CuII (2-pySO3 )3 ]− anions [73]. The square grid can be enlarged by using longer bipyridine ligands. For example, extraordinarily big grid dimensions of 20 × 20 Å2 were obtained from 4,4′ -bis(4-pyridyl)biphenyl ligand and Ni(NO3 )2 (1 : 2 ratio). Each grid is occupied by six o-xylene guest molecules. The packing of the grids created big rectangular channels (c. 10 × 20 Å2 ) (Figure 6.34b) [71]. When the extremely long N,N ′ -type ligand,

311

312

6 The Structures of Metal–Organic Frameworks

(a)

(b)

Figure 6.34 (a) Non-interpenetrated square-grid network in {[Cd(4,4′ -bpy)2 ](NO3 )2 }n . (b) The packing of grids showing the formation of rectangular channels in a non-interpenetrated square-grid network. Source: Ohmori and Fujita [70]. © 2004 Royal Society of Chemistry; Biradha et al. [71]. © 2000 John Wiley & Sons.

2,5-bis(4′ -(imidazol-1-yl)benzyl)-3,4-diaza-2,4-hexadiene (ImBNN), was used to react with different metal salts, a series of large, non-interpenetrated square-grid coordination polymers were synthesized. In complex [Mn(ImBNN)2 (NO3 )2 ]∞ , the right- and left-handed helical chains intersect via Mn(II) atoms, forming 2D layers with (4,4) topology. The size of the M4 L4 square is about 2.1 × 2.1 nm2 . Although the layers are non-interpenetrated, the ABCDABCD stacking reduced the void space in the structure. The dimensions of the square grids in the two layers of {[Cd(ImBNN)2( NO3 )2 ]⋅solvate}∞ are both c. 2.2 × 2.2 nm2 , and the ABCABC stacking of them creates infinite channels occupied by many disordered solvent molecules. In MOF {[Zn(ImBNN)2 ](BF4 )2 ⋅(C6 H6 )2.564 ⋅(DMF)1.576 ⋅(MeOH, H2 O)3.454 }∞ , half of the 𝜇2-ImBNN ligands connect adjacent zinc centers into rightand left-handed helical chains ([Zn(ImBNN)]n ) and the remaining ligands connect adjacent helical chains to form a (4,4) topological sheet with the dimensions of each grid about 2.1 × 2.1 nm2 . Differently, the layers are severely distorted and nonplanar. Viewed from the side, the layer looks like a 1D ladder. The stacking of the layers significantly reduces the accessible channel dimensions as well [74]. Two 2D 44 -grid networks [Fe(bpb)2 (NCS)2 ]⋅X (X = 2.5 H2 O and 0, respectively) were assembled from 1,4-bis(pyrid-4-yl)benzene (bpb) and Fe(NCS)2 . [Fe(bpb)2 (NCS)2 ]⋅2.5 H2 O is square-grid planar network with the square dimension of 15.9 × 16.0 Å2 . Neighboring layers are stacked alternately in an ABAB mode, existing 1D channels parallel to a axis filling with solvent molecules. The structure of [Fe(bpb)2 (NCS)2 ] is a 2D 44 rhombic-grid network with the grid dimension of 15.3 × 15.7 Å2 as well. But the adjacent layers are arranged in the offset rather than face-to-face fashion, resulting in a nonporous close-packing structure [75]. Long bridging ligands are expected to create large cavity sizes. But in many cases, they favor the formation of interpenetrating structures. The long rigid ligand ImBNN reacts with M(CF3 SO3 )2 (M = Cd and Mn) salts to assemble large meshes; however, the guest-free rhombic-grid coordination polymers are triply interpenetrated. When the aromatic guest molecules are included, such triply interpenetrated networks changed to doubly interpenetrated ones accompanied by

6.3 Two-dimensional MOFs

Figure 6.35 Schematic views of the formation of threefold interpenetrated grid networks and twofold interpenetrated ones. Source: Wang et al. [76]. © 2009 American Chemical Society.

N N N N

N

N

Without guests

+ M2+

With guests

grids of contracting/expanding change. The resulting twofold entangled networks display a strong preference and remarkable tolerance toward various aromatic guests, such as benzene, o-/p-xylenes, naphthalene, phenanthrene, and pyrene (Figure 6.35) [76]. A doubly interpenetrated square-grid coordination polymer {[Cd-(ImBNN)2 (CF3 SO3 )2 ]⊃guest}n (guest = C7 H8 ) was also self-assembled from ImBNN and Cd(CF3 SO3 )2 . This framework is dynamic, expanding or shrinking with the variation of temperature and exhibiting temperature-dependent ligand exchange driven by guest release/uptake [77]. In order to inhibit the interpenetration of the networks, the bulky elongated organic ligands have been employed to form huge square grids with larger cavity size. Long N,N ′ -bipyridine-based ligands containing side chains, 9,9-diethyl-2,7bis(4-pyridylethynyl)fluorene (dbpyf) and chiral 9,9-bis[(S)-2-methylbutyl]-2,7bis(4-pyridylethynyl)fluorene (bmbpyf), reacted with copper nitrate, leading to exceptionally large, non-interpenetrating, square-grid polymers with grid dimensions of 25 × 25 Å2 . In MOF [Cu(dbpyf)2 (NO3 )2 ], the short side chains of the ligand lead to an open-grid space. Although the ABAB stacking of the layers caused reduced void, very large infinite channels of dimensions 16 × 16 Å2 still remained. The solvent-accessible volume calculated with the PLATON program is 54.1%. The channels are occupied by disordered solvent molecules and nitrate anions. [Cu(bmbpyf)2 (NO3 )2 ] is the first chiral non-interpenetrating square-grid coordination polymer. Different layers are stacked in the ABCABC fashion, which significantly reduces the accessible channel dimension to about 8 × 8 Å2 . Different to the virtually planar grids in [Cu(dbpyf)2 (NO3 )2 ], the grid network in [Cu(bmbpyf)2 (NO3 )2 ] is undulate [78]. A bipyridyl linker that features photochromic 2,2-diphenylbenzopyran (DP) as the backbone was prepared and used to coordinate with Cd(II), leading to a 2D neutral square-grid framework [Cd2 Cl2 (DP)3 ]⋅(CHCl3 )4 (DMF)4 EtOH with large void spaces. The layers stack each other with some offset. The phenyl rings at the C2 position of the benzopyran moiety lie perpendicular to the layers such that they interject into the voids of the neighboring ones in a ridge-in-a-furrow fashion (Figure 6.36) [79]. Despite the rod-like N,N′ -type ligands, other N-containing ligands have also been widely applied to compose square-grid structures. Chiral ligand 5-[(2S)-pyrrolidine-2-yl]-1H-tetrazole ((S)-HPTZ) from commercially available L-proline assembled with Cd2+ ion and Zn2+ to construct two new 2D homochiral

313

314

6 The Structures of Metal–Organic Frameworks

= M2+ Self-assembly

Figure 6.36 Cartoon representation the developing 2D layered MOFs by using linkers with a bulged core, which may hinder self-assembly in the orthogonal direction. Source: Mukhopadhyay et al. [79]. © 2018 Elsevier.

isostructural MOFs, [Cd((S)-PTZ)2 ]n and [Zn((S)-PTZ)2 ]n , respectively. Each (S)-PTZ ligand connects two Cd2+ (or Zn2+ ) and acts as a κ 3 -linker. The whole frameworks can be topologically represented as a four-connected net with point (Schläfli) symbol of (84 ⋅ 122 ) (8)2 , regarding the (S)-PIA ligands and the metal ions as the 2- and four-connected nodes, respectively [80]. The first porphyrin-based polymer [Cu2 (AcO)4 (CuTPyP)1/2 ]⋅CHCl3 , formed by the reaction of 5,10,15,20-tetra-4-pyridyl-21H,23H-porphine (H2 TPyP) and dicopper(II) tetraacetate [Cu2 (AcO)4 ], exhibits a 2D infinite 22.2 × 22.2 Å2 square-grid coordination network, with dinuclear Cu2 (AcO)4 moiety as a linear linker motif and H2 TPyP as a four-connected vertex. The 2D layers stack in an ABAB sequence along the c axis, so the structure does not show large channels (Figure 6.37) [81]. Flexible multidentate hydrogenated Schiff base hsb-1, derived from tris(2aminoethyl)amine (tren) and 4-pyridinecarboxaldehyde, has been used to prepare

N

N

N

N Cu N N

N

N

N

N

N Cu N N

N

N Cavity

22.2 Å

N

N

N

N Cu N N

N

N

=

N

N

N

22.2 Å

N

N Cu N N

Me O OO Me O Cu Cu O Me O O O Me

N

N

Figure 6.37 A 2D MOF with 22.2 × 22.2 Å2 square-grid coordination network. Source: Ohmura et al. [81]. © 2006 American Chemical Society.

6.3 Two-dimensional MOFs

2D layer networks, {[Cd1.5 (hsb-1)(NO3 )2 (H2 O)]NO3 ⋅2H2 O}n and {[Mn1.5 (hsb-1)Cl2 (H2 O)]Cl⋅6H2 O}n . Both display 2D layer structures possessing two kinds of infinite helical chains, in which the right-handed and left-handed helical chains are in an alternate arrangement by sharing the Cd (or Mn) ions. Two pyridyl groups of the ligand are coordinated to different Cd (or Mn) atoms respectively and the remaining pyridyl group points outward (or inside) from the helical chain. The 2D net can be described as a 44 -sql topology [42]. Another type of hydrogenated Schiff base, 1,2-bis(4′ -pyridylmethylamino)ethane (hsb-2), can exhibit three different types of conformations, linear and V-shape according to the direction of two pyridyl groups. When it coordinates to different metal salts, many 2D MOFs were obtained. [Ag(hsb-2)]ClO4 ⋅CH3 CN, [Cu(hsb-2)]NO3 are infinite 2D cationic networks [82, 83]. In [Zn(hsb-2)(H2 O)2 ⋅2ClO4 ⋅4H2 O]n , the ligand hsb-2 adopts a V-shape and functions as a bridging linker, resulting in a 2D net with a square grids. The disordered ClO4 2− anions are filled between neighboring layers, while in 2D network of [Zn(hsb-2)Ac(H2 O)⋅Ac⋅2H2 O]n , the coordinated Ac− of two adjacent layers is interdigitated inward, and the other uncoordinated Ac2− is distributed randomly outside (Figure 6.38) [84]. When hsb-2 reacted with Zn(NO3 )2 in H2 O/EtOH at room temperature, another 2D layered structure [Zn(hsb-2)(NO3 )2 ⋅2H2 O]n with square windows was afforded. Interestingly, the adjacent layers have a “mirror symmetry,” which can be “separated” through the “pillaring” strategy to form 3D chiral MOFs [85]. Counterion plays an important role in the formation of those layered complexes. Oxygen-containing ligands such as phthalic acid and their analogues, phosphonates, etc. are another type of building blocks for 2D square-grid MOFs. Solvothermal reactions of Cu(II) nitrate with isophthalic acid (H2 IPT) in different solvents create various 2D non-interpenetrated square-grid layers. In these networks, paddle-wheel Cu2 clusters act as nodes and IPT as linkers, and guest molecules are trapped via either the coordination to open Lewis acid copper sites or inclusion in open channels. The solvents affect the arrangement and separation of the adjacent layers. In [Cu(IPT)(H2 O)](H2 O)2 and Cu(IPT)(C2 H5 OH), the 2D layers are eclipsed each other to form a microporous framework, and the layer-to-layer separations (d1 and d2 ) are 6.764 and 6.743 Å; adjacent parallel layers of Cu2 (IPT)2 (DEF)(H2 O) are almost

(a)

(b)

Figure 6.38 Two 2D networks based on hydrogenated Schiff base ligand hsb-2. Source: Wen et al. [84]. © 2013 Royal Society of Chemistry.

315

316

6 The Structures of Metal–Organic Frameworks

d3

d1,d2

(a)

(b) d5

d4

(c)

(d)

Figure 6.39 Views of the space arrangements of the adjacent 2D layers in [Cu(IPT)(H2 O)](H2 O)2 and Cu(IPT)(C2 H5 OH) (a), Cu2 (IPT)2 (DEF)(H2 O) (b), [Cu4 (IPT)4 (DMF)2 (C2 H5 OH)(H2 O)](DMF)2 (c), and [Cu2 (IPT)2 (CH3 CN)(H2 O)](H2 O)2.25 (d). Source: Zhong et al. [86]. © 2011 Royal Society of Chemistry.

eclipsed each other with the layer-to-layer separation of d3 (6.918 Å), so it shows poor porosity; in MOFs [Cu4 (IPT)4 (DMF)2 (C2 H5 OH)(H2 O)](DMF)2 and [Cu2 (IPT)2 (CH3 CN)(H2 O)](H2 O)2.25 , the adjacent 2D layers interleaved each other, which results in poor porosities. The layer-to-layer separations are 7.742 and 7.657 Å, and the valid space between the adjacent layers is occupied by guest molecules (Figure 6.39) [86]. Six-layered metal phosphonates were prepared by the hydrothermal reactions of ZnII or MnII ion with three different phosphonic acids. They feature 2D sandwich-like frameworks with Zn–O–P or Mn–O–P layers. Each [ZnO4 ] tetrahedron (or [MnO6 ] octahedron) shares three corners (or four corners) with three neighboring [PCO3 ] tetrahedra to generate a Zn–O–P layer (or Mn–O–P layer) containing eight- and four-member rings (or eight-member rings). The organic groups hang on two sides of a Zn–O–P or Mn–O–P layer to form 2D sandwich-like frameworks [87]. PPF-1, i.e. Zn2 (ZnTCPP)⋅3H2 O⋅2DEF, assembled from the Zn-tetra-(4-carboxyphenyl)porphyrin (ZnTCPP) and Zn paddle-wheel clusters [Zn2 (COO)4 ], exhibits a 2D square-grid network. These grids are stacked in an AB packing pattern, generating square channels with dimensions of approximately 11.8 × 11.8 Å2 . The porphyrin zinc atoms are arranged in line with the paddle-wheel zinc atoms, which is disordered as slightly above or below the porphyrin plane [88]. A bulky elongated cross-shaped dicarboxylate ligand was used to prepare a highly porous square-grid MOF with the unprecedentedly large grid dimensions of 25.5 × 25.5 Å2 . The grids are virtually planar and non-interpenetrating; however, the adjacent layers are stacked in an offset ABCABC manner, which reduces the free space, giving rise to two minor channels with dimensions of only 5 × 5 Å2 and 2 × 2 Å2 viewed from the grid plane. Interestingly, a rare hydrogen bonded 64 ⋅82 -nbo network is formed by the stacking of the 44 -sql layers because of the NH· · ·N interactions in layers [89]. Flexible triazine-based carboxylic ligands have been used to construct a series of 2D square-grid MOFs. MOF {[Zn2 (CCTA)(DMF)2 ]⋅(DMF)⋅(H2 O)2 }n

6.3 Two-dimensional MOFs

(H4 CCTA = 2,4-bis(4-carboxyphenylamino)-6-bis(carboxymethyl)amino-1,3,5triazine) exhibits a 2D network with sql topology, which consists of tetranuclear zinc units. The adjacent layers adapt an ABAB packing model from the side view. In the structure of MOF {[Zn2.5 (CCTA)(OH)(H2 O)]⋅DMF⋅H2 O}n , the carboxylates and μ3 -OH− are coordinated to the metal ions forming a pentanuclear Zn(II) cluster SBUs, which further propagate to a 1D infinite cluster chain via bridging carboxylates. The inorganic chains are then gripped by CCTA4− to compose a novel 2D architecture with unprecedented point (Schläfli) symbol of (418 ⋅610 )(45 ⋅6)2 . Different layers are arranged in parallel rows [90]. Another two triazine-based flexible polycarboxylate ligands, 1,3,5-triazine-2-iminodiacetic acid-4,6-bis(L-alanine) and 1,3,5-triazine-2-iminodiacetic acid-4,6-biglycine, have been prepared and utilized for the construction of many 2D lanthanide coordination polymers. Adjacent layers are held together via intermolecular π· · ·π interactions [91]. Many organic ligands containing both N- and O-(or S-) donors as building blocks have been employed to construct various 2D square-grid networks. Ligand 4-(imidazol-1-yl)-benzoic acid possessing both imidazole and carboxylate moiety was treated with cobalt nitrate in the presence of sodium azide under hydrothermal condition, yielding a perfect square-grid network with the dimensions of 11.42 × 11.42 Å2 . Different layers are stacked in an ABCD pattern and the interlayer separation is c. 7.0 Å, which are further stabilized by π· · ·π stacking interaction between adjacent imidazole rings and between adjacent phenyl rings [92]. A wave-like 2D Ag-MOF was assembled from 4-cyanobenzoate and AgNO3 . The 2D rhombohedral-grid layers are stacked in parallel without interpenetration to generate 10.832 × 6.650 Å2 channels. There are interlayer Ag · · · Ag interactions and π ⋅ ⋅ ⋅ π stacking interactions among the layers. It is noteworthy that this MOF emits tunable yellow-to-white photoluminescence by variation of excitation light [93]. Chiral, acentric, nonlinear optical (NLO)-active MOFs based on 2D square (or rhombohedral) coordination networks have been synthesized by using unsymmetrical linking groups. MOF bis(nicotinato)zinc shows an infinite square grid composed of six-coordinate Zn centers and bridging nicotinate, which crystallizes in the chiral space group P43 21 2. Bis-{3-[2-(4-pyridyl)ethenyl]benzoate}cadmium crystallizes in the acentric space group Fdd2, displays an infinite 2D rhombohedral grid. They both exhibit powder SHG efficiency [94]. Complex Co(4-pyt)2 (4-pyt = pyridine-4-thiolate) has been prepared by in situ generation of a 4-pyt ligand from a 4,4′ -dithiodipyridine precursor under solvothermal conditions, which shows a 2D square-grid structure. Each 4-pyt ligand links two Co ions through N and S atom; thus, each Co ion connects four ligands to form a (4,4)-net with dimensions of 7.5 × 7.5 Å2 . Adjacent layers are stacked to give a non-interpenetrating topology [95]. Employing the in situ ligand formation strategy, 2D homochiral square-grid framework was obtained from the in-situ formed chiral amino acid–tetrazole ligands [96]; 44 -sql layer structure was constructed from the in situ oxidative dehydrogenation of 4-(3,5-(dicyano-2,6-dipyridyl)dihydropyridyl)benzoic acid [97]. The 2D square grid networks become more abundant by using the mixed-ligand strategy. Rigid linear pyridine ligands with alkoxy functional groups incorporating rigid carboxyl-containing auxiliary ligands have been applied to

317

318

6 The Structures of Metal–Organic Frameworks

build a series of 2D square-grid networks. In complex {[Zn(dmopdy)(IPT)]}n (dmopdy = 4,4′ -(2,5-dimethoxy-1,4-phenylene) dipyridine), the IPT ligands link the Zn(II) ions to form a 1D chain, which are further extended through dmopdy ligands to furnish (4,4)-connected sql nets with point symbol {44 ⋅62 }. In complex {[Co(dmopdy)(TPA)-(H2 O)2 ]⋅2DMF}n (H2 TPA = terephthalic acid), mixed ligands are coordinated to the Co(II) ion to form planar layers with rectangular grid with the dimensions about 15.684 × 11.500 Å2 . These layers are held together by hydrogen bonding and π· · ·π stacking interactions in AAA mode, leading to infinite 1D channels, which are occupied by disordered DMF molecules [98]. A series of 44 -sql 2D layer MOFs have been prepared from flexible carboxylate and bis(pyridine) Ligands. For example, homochiral polymers [Zn2 (hsb-2)(NCG)2 (H2 O)2 ⋅4H2 O]n and [Zn(bpe)(NCG)⋅3H2 O]n based on flexible N-carbamyl-Lglutamate (NCG) and bis(pyridine) ligand hsb-2 and 1,2-bis(4-pyridyl)ethane, respectively, are wave-like networks. The adjacent layers are stacked in an AA type arrangement. [Zn(bpe)(NCG)(NaNO3 )0.5 H2 O]n displays an interesting 2D double-layered framework, adjacent layers are connected via hydrogen bonding and aromatic π· · ·π stacking interactions. Complex [Zn(bpp)(NCG)⋅3.5H2 O]n (bpp = 1,3-bis(4-pyridyl)propane) contains two uncommon helical chains of bpp and NCG, with two flexures in each single strand. Different layers are stacked in an offset ABAB fashion [99]. The (4,4) topological 2D MOFs based on mixed rigid and flexible ligands (abstracted as lines and arcs, respectively) can be diversified to five types, which are named a, b, c, and d that have a single-arched bridge and type e that has double-arched bridges (Figure 6.40). Five 2D coordination polymers with four types of edge-transitive square lattices (sql)4 were prepared from rigid ligand 4,4′ -azodibenzoic acid (adc) or 3,3′ -dichloro-4,4′ -azodibenzoic acid (ClH2 adc) and flexible ligand bpp. Complexes [Co(adc)(bpp)(H2 O)]n , [Ni(Cladc)(bpp)(H2 O)]n , and [Co(Cladc)(bpp)(H2 O)]n feature 2D undulated layers (belonging to type a). All arched bpp ligands display a TT conformation bent in one direction, exhibiting a wave mode. Complex {[Ni(adc)(bpp)2 (H2 O)]2 ⋅bpp}n features two types of 2D undulated layers, which are heterointerpenetrated. Complex [Zn(adc)(bpp)⋅DMF]n shows a step-shaped 2D (4,4) network. The layers belongs to type d, in which the orientations of tetrahedral Zn(II) centers alternately change along adc2− ligands, following which bpp ligands protrude from both sides of the layer. The structure represents a typical 2D → 3D parallel interpenetration, in which each layer is entangled with two adjacent identical layers [100]. By the hydrothermal reactions of copper nitrate, rigid 4,4′ -bpy, and flexible bgxH4 (bgxH4 = α,α′ -bis(N-glutamyl)-p-xylene), a 2D MOF [Cu4 (bpgxH2 )(4,4′ -bpy)4 (NO3 )2 ]⋅4H2 O (bpgxH2 = α,α′ -bis(N-pyroglutamyl)p-xylene) composed of rhombohedral networks has been assembled. Ligand bpgxH2 was formed by the in situ intramolecular cyclization of bgxH4 precursor. This structure composed of rhombohedral networks can be described as adjacent rigid 4,4′ -bpy chains connected by flexible bpgxH2 linkers (belonging to type a), encapsulating free nitrate anions. The anionic 2D MOF displays

6.3 Two-dimensional MOFs

360˚

(4,4) Topology

Type a

Type c

Type b

Type d

Type e

Figure 6.40 Five types of reported 2D layers constructed of mixed flexible and rigid ligands with (4,4) topology. Source: Guo et al. [100]. © 2013 American Chemical Society.

extensive visual colorimetric changing ranges toward a great number of anionic species in aqueous solutions [101]. In the structure of complexes [Cd(BCbpy)(bpdc)0.5 X]⋅7H2 O (BCbpy = 1-(4-carboxybenzyl)-4,4′ -bipyridinium, H2 bpdc = 4,4′ -biphenyldicarboxylic acid; X = Cl or Br), the [Cd2 (μ2 -X)2 ] cores are in turn linked by four μ2 -BCbpy ligands and simultaneously in the vertical direction bridged by two μ2 -bpdc ligands to form a 2D sheet (belonging to type e). Two such nets are further interpenetrated in a 2D → 2D parallel fashion, leading to a polyrotaxane-like interpenetrating network [102].

6.3.3

Honeycomb, Brick-Wall, and Herringbone Networks

In honeycomb structural motifs, three hexagons meet at a node in a 2D network with angles of 120∘ , and the corresponding Schläfli topology symbol is 63 (Figure 6.31c). The nodes in honeycomb networks need to have threefold rotational symmetry. Loss of this symmetry results in alternative 2D networks: three-connected nodes with angles of 180∘ , 90∘ , and 90∘ in brick-wall or herringbone architectures. Both the brick-wall and herringbone networks also belong to the 63 topologies (Figure 6.41). Two isomeric honeycomb frameworks, Zn2 (TMTA)(H2 O)2 ⋅NO3 ⋅6H2 O⋅DEF (α-1) and Zn2 (TMTA)(H2 O)2 ⋅NO3 ⋅2H2 O⋅0.5DMA (𝛽-1), have been prepared by the solvothermal reactions of Zn(NO3 )2 ⋅6H2 O and 4,4′ ,4′′ -(2,4,6-trimethylbenzene1,3,5-triyl)tribenzoic acid (H3 TMTA) in DEF or DMA [103]. Complex α-1 crystallizes in the orthorhombic space group Cmcm, while complex β-1 crystallizes in

319

320

6 The Structures of Metal–Organic Frameworks

(a)

(b)

Figure 6.41

(a) Brick-wall networks. (b) Herringbone networks. HO O

+ OH

HO O

H3TMTA

O DEF

DMA

Chiral ABCDEF stacking

Achiral ABAB stacking

2D honeycomb layer

α-1

β-1

Figure 6.42 Two isomeric supramolecular architectures from TMTA and Zn2 (COO)3 SBUs consist of 2D honeycomb layers. The layers in α-1 and β-1 show different stacking arrangements (ABAB versus ABCDEF). Source: Zhao et al. [103]. © 2010 American Chemical Society.

chiral space group P6122 . All of the zinc ions in two complexes are four-coordinated by three oxygen atoms from different TMTA ligands and one water molecule in a tetrahedral geometry. Each Zn2 (COO)3 SBU links to three TMTA ligands, and every TMTA connects three Zn2 (COO)3 SBUs, generating cationic honeycomb networks (Figure 6.42). Complexes α-1 and β-1 display similar 2D frameworks, but their stacking arrangements of the layers are quite different (Figure 6.42). In α-1, the cationic layers stack in an ABAB fashion with a distance between layers of 10.3 Å, resulting in a 3D achiral supramolecular architecture. No significant interactions among the layers have been observed. While the 2D cationic layers in β-1 adopt ABCDEF stacking with a layer-to-layer distance of 4.63 Å (significantly shorter than that in α-1), giving a 3D chiral supramolecular architecture. The CH· · ·π interactions (3.529 and 3.597 Å)

6.3 Two-dimensional MOFs

between the side benzene rings in one layer and the central benzene rings in its adjacent layer have been found, which connect the TMTA ligands in different layers to form a 1D 61 helical chain. All TMTA ligands are arranged around the helical axes. The rigidity of the layer and the directionality of CH· · ·π interactions make the same chirality of the helical chains preserved; therefore, the supramolecular structure is chiral. According to the different stacking modes (ABAB versus ABCDEF), polymer α-1 shows more void volume than that of β-1. Two periodic layers with honeycomb topologies have been hydrothermally prepared from cucurbit[6]uril-based pseudorotaxanes, rigid carboxylate ligands and transition metal ions [104]. The combination of [PR64]2+ ⋅2[NO3 ]− , H2 BPDC, and Cd2+ leads to 2D wavy layers of complex [Cd(BPDC)(PR64)0.5 Cl]⋅2H2 O (Figure 6.43a). The hexa-nuclear cadmium-organic loop is formed by six Cd2+ ions, four BPDC2− anions, two [PR64]2+ , and six Cl atoms. All Cd2+ ions are coordinated by four oxygen atoms from two different BPDC ligands in a bidentate mode, one N atom from the [PR64]2+ , and one coordinated chloride, in a distorted octahedral mode. The two terminal pyridine groups in trans-[PR64]2+ are in parallel. The framework is so wavy that it takes three other layers to offset the gap with a height of 21.84 Å and a dihedral angle of 62.5∘ . The adjacent layers are staggered in a tiling array manner via π· · ·π stacking interactions of CB[6]s and BPDC2− . All CB[6]s in different layers are well separated and arrayed in a line. The combination of [PR64]2+ ⋅2[NO3 ]− , m-H2 bdc and ZnCl2 results in a complex [Zn(PR64)0.5 (m-bdc)Cl]⋅3H2 O with 2D flat layers (Figure 6.43b). Zn centers are coordinated by two oxygen atoms of two m-bdc2− ligands, one N atom of the trans-[PR64]2+ , and one coordinated chloride atom in tetrahedral mode. The

OH O N

O H2N

6 H2 N C

N N 3

O Cd

HO

N H2 N C

N NH2 O

(a)

[PR64]2+ HO

O

Zn

O OH

(b)

Figure 6.43 Synthesis and schematic views of (a) the 2D wavy layer in [Cd(BPDC)(PR64)0.5 Cl]⋅2H2 O and (b) the 2D flat (6,3) net in [Zn(PR64)0.5 (m-bdc)Cl]⋅3H2 O. Source: Liang et al. [104]. © 2016 Royal Society of Chemistry.

321

322

6 The Structures of Metal–Organic Frameworks

HN N

N HN

NH N

N NH

Figure 6.44 Schematic view of {[Cd(H2 MDP)(1,4-H2 bdc)](H2 O)3/2 }n with 2D 63 -honeycomb topology composed of M2 L2 -type SBUs and 1D zigzag chains. Source: Based on Guo et al. [105].

structure also displays honeycomb topology composed of oblate hexagon rings, in which adjacent layers are closely stacked in an inlaid manner. Using a flexible ligand of methylenebis(3,5-dimethylpyrazole) (H2 MDP) and a rigid ligand of 1,4-H2 bdc, another network with 63 -honeycomb topology has been obtained (Figure 6.44) [105]. In the structure of {[Cd(H2 MDP)(1,4-H2 bdc)] (H2 O)3/2 }n , the asymmetric unit contains one H2 MDP ligand, one Cd(II) ion, and two independent half 1,4-H2 bdc ions. Each Cd(II) ion adapts an octahedra coordination geometry completed by two nitrogen atoms from two H2 MDP ligands and four oxygen atoms from two 1,4-H2 bdc ligands in bidentate mode. Two H2 MDP ligands connecting two Cd(II) ions form the M2 L2 -type closed loops as SBUs, which connect 1D zigzag chains of 1,4-H2 bdc ligands and Cd(II) to form a 2D framework. Two M2 L2 -type motifs and four 1,4-H2 bdc ligands build a hexagon, which further compose the honeycomb topology layer by simplifying Cd(II) ions as three-connected nodes, and H2 MDP double bridges and 1,4-H2 bdc ligands as linkers. The adjacent layers are non-interpenetrated, and they are piled up in an offset fashion by the π· · ·π stacking interactions associated with benzene rings and pyrazole rings between the neighboring layers. A brick-wall layer was reported by Fujita et al. in 1995, which is composed of heptacoordination as well as T-shaped connection of Cd(II) and a fluorinated bipyridine-based ligand. Three independent molecular bricks interpenetrate each other to form a thick infinite molecular sheet [106]. Classical 2D 63 topological brick-wall sheets, that is, [Cu2 (pyz)(CN)2 ], have been assembled from CuCl2 ⋅2H2 O, K4 Fe(CN)6 ⋅3H2 O, and pyrazine (pyz) [107]. In the layered structure, all Cu(I) centers are coordinated by one N from pyrazine and one C and one N from two cyanides, displaying trigonal geometries. In each layer, the brick-wall nets arrange in ABAB rows. Different brick-wall sheets are stacked in an offset mode, resulting in small 1D channels c. 4.1 × 4.1 Å2 , which are occupied by linear [Cu(CN)]∞

6.3 Two-dimensional MOFs

) 6] (CN Fe °C [ K 4 70 1

(6,3) brick wall net

CuCl2 + pyrazine K3 [F e(C 160 N)6 ] °C

(5,34) Shubnikov net

Figure 6.45 Synthesis and schematic views of two 2D layers with brick-wall net and Shubnikov net.

chains to form a 3D pseudocatenate-like supermolecular structure. Interestingly, unprecedented (3,4)-connected Shubnikov-type (5,3 4 ) sheets are obtained by the reaction of CuCl2 ⋅2H2 O, K3 [Fe(CN)6 ], and pyz, which contain three-connected and four-connected nodes of Cu(I) and metallo-ligand CN–Cu(I)–CN linkers. The smallest units in the (5,3 4 ) sheet are [Cu6 (CN)4 (pyz)2 ] bricks with size of 9.60 × 8.57 Å2 , which stand side by side. Adjacent brick rows are arranged in ABBA to form the final (3,4)-connected Shubnikov-type networks. Similarly, the layers are stacked parallelly in AB fashion possessing channels of c. 4.0 × 4.0 Å2 , which are penetrated by [Cu(CN)]∞ chains to form a 3D pseudopolyrotaxane isomer (Figure 6.45). Two enantiomeric MOFs with brick-wall layered structures, Zn-BCIP1 and Zn-BCIP2, have been synthesized via the solvothermal reaction of L-Ntertbutoxycarbonyl-2-(imidazole)-1-pyrrolidine (L-BCIP) (or D-BCIP), 4,4′ ,4′′ tricarboxyltriphenylamine (H3 TCA), and Zn(NO3 )3 ⋅6H2 O (Figure 6.46). The Figure 6.46 Synthesis and schematic views of 2D brick-wall network of Zn-BCIP1. Source: Wu et al. [108]. © 2012 American Chemical Society.

O OH N N HO O H3TCA

+ O OH

Zn

+

N Boc L-BCIP

N

323

324

6 The Structures of Metal–Organic Frameworks

independent layers are built from three-connected binuclear zinc nodes and 4,4′ ,4′′ -nitrilotribenzoate linkers. In the binuclear SBUs, one of the Zn(II) adopts a distorted trigonal–bipyramidal geometry by coordination with one bridged hydroxyl anion, one water molecule, one bidentate, and one bimonodentate carboxylate group of two different TCA3− anions, while the other zinc atom is coordinated by two oxygen atoms from two different TCA3− anions, one bridged hydroxyl anion, and one nitrogen atom from the L-BCIP in a distorted-tetrahedral geometry. Two adjacent layers are stacked together face to face, which are further packed to generate 1D 12 × 16 Å2 channels, with L-BCIP molecules located within the channels. The tert-butoxycarbonyl (Boc) protects the pyrrolidine moiety and precludes framework interpenetration, which can be removed by just heating Zn-BCIP1 in a dry DMF employing microwave irradiation. The new MOF Zn-PYI1, containing chiral organocatalytic D-/L-PYI moiety and photosensitizer triphenylamine moiety, acts as efficient heterogeneous asymmetric catalysts for light-driven asymmetric α-alkylation of aldehydes [108]. A brick-wall layer and a herringbone layer network, [Bi2 (mdhbqdc)(ox)2 (DMF)4 ] and [Bi2 (mdhbqdc)2 (ox)(DMF)4 ] (H2 mdhbqdc = dimethyl 3,6-dihydroxy-2,5benzoquinone-1,4-dicarboxylate, ox = oxalate), have been prepared via the solvothermal in situ ligand synthesis methodology [109]. In complex [Bi2 (mdhbqdc) (ox)2 (DMF)4 ], each Bi(III) center is coordinated by eight oxygen atoms of one mdhbqdc, two oxalates, and two DMF molecules in triangular dodecahedral geometry, which is linked by ox and mdhbqdc to give a 2D 63 brick-wall layer with ratio of mdhbqdc to ox 1 : 2. Adjacent layers are packed in an overlapped AA mode. The asymmetric unit of [Bi2 (mdhbqdc)2 (ox)(DMF)4 ] contains one Bi(III), one mdhbqdc, half oxalate, and two DMF molecules. The Bi(III) centers also display triangular dodecahedral geometry, but the eight coordinated oxygen atoms are from two mdhbqdc, one oxalate, and two DMF molecules. Each Bi(III) connected three other ones by ox and mdhbqdc bridges with molar ratio 2 : 1, yielding 2D 63 -topological herringbone networks (Figure 6.47). Hydrothermal reaction of cadmium salt, 3,3′ ,4,4′ -diphenylsulfonetetracarboxylic acid (H4 dpstc) and 1,10-phenanthroline (phen) under pH = 10 affords another 2D herringbone-like network [Cd2 (dpstc)(phen)2 ]n ⋅nH2 O [110]. The asymmetric unit contains two cadmium atoms, one dpstc4− ligand, two phen molecules, and one water guest molecule. One of the Cd(II) centers adopts a distorted octahedra geometry, while the other one displays a triangular prism geometry, both of which are six-coordinated by two nitrogen atoms of phen and four atoms of dpstc4− ligand. Binuclear cadmium clusters are formed by sharing one oxygen atom O6, four of which are further linked by four dpstc4− ligands to give a dumbbell-shaped ring. The two phthalic moieties in dpstc4− ligand exhibit different connection types: binding to one binuclear unit and bridging two binuclear clusters. As a result, 2D herringbone-like net is constructed (Figure 6.48).

6.3 Two-dimensional MOFs

(a)

(b)

(c)

(d)

Figure 6.47 View of the coordination environments of Bi(III) ions (a) and 2D brick-wall layer (b) in [Bi2 (mdhbqdc)(ox)2 (DMF)4 ]. View of the coordination environments of Bi(III) ions (c) and 2D herringbone layer (d) in [Bi2 (mdhbqdc)2 (ox)(DMF)4 ].

6.3.4

Bilayer Networks

More and more bilayer network MOFs have been assembled and reported, which make the 2D MOF structures more diverse. Usually, the bilayer networks are fabricated by T-shaped building blocks (Figure 6.49). The bilayer architecture can also be constructed by non-T-shape building block, seen in many networks of 2D pillared bilayers, interpenetrated or interdigitated layers. A 2D double-layered network, {[Co(bpt)(m-bdc)]⋅3H2 O}n , was built from bent dipyridyl ligand 4-amino-3,5-bis(4-pyridyl)-1,2,4-triazole (bpt) and isophthalate (m-bdc). In the structure, each Co(II) center adopts a distorted octahedral geometry, which is complete by four equatorial oxygen atoms from three m-bdc and two axial nitrogen atoms from two bpt ligands. The Co(II) centers are linked by the m-bdc ligand to generate 1D ribbons, which are further connected by exo-bidentate bpt ligands to form a 2D double-layer. Adjacent double layers are stacked in the interdigitating arrangement (Figure 6.50). The structure of {[Cd2 (bpt)(m-bdc)2 (H2 O)4 ]⋅6H2 O}n exhibits a rare interpenetrated and interdigitated framework based on the T-shaped bilayer networks. Two independent Cd(II) centers display distorted pentagonal–dipyramidal and octahedral geometry,

325

326

6 The Structures of Metal–Organic Frameworks

Figure 6.48 (a) View of the coordination environments of Cd(II) ions. (b) 2D herring-bone layer.

O6A

(a)

(b)

Figure 6.49 networks.

2D bilayer

respectively. Each Cd(II) center function as a three-connected T-shaped nodes to construct an uncommon 2D bilayer with 82 ⋅10 topology. Two neighboring bilayers in the same orientation parallel interpenetrate and the opposite “double bilayers” interdigitate in a “tongue-and-groove” stacking fashion [111]. The structure of MOF {[Zn2 (m-bdc)2 (4,4′ -bpt)(H2 O)3 ]⋅H2 O}n (4,4′ -bpt = 1H3,5-bis(4-pyridyl)-1,2,4-triazole) displays the alternately interdigitation of bilayers. The asymmetric unit has two independent Zn(II) centers, two m-bdc, one 4,4′ -bpt, and three coordinated water molecules. One of the Zn(II) centers is coordinated by two O atoms of two m-bdc and one N atom of 4,4′ -bpt in the equatorial plane and two water molecules at the axial sites, showing a distorted trigonal–dipyramidal geometry, while the other Zn(II) is tetrahedral geometry formed by three O atoms from two m-bdc and one water molecule and one

6.3 Two-dimensional MOFs

(a)

(c)

(b)

(d)

Figure 6.50 The metal coordination (a) and perspective view of the 2D double-layered framework (b) of {[Co(bpt)(m-bdc)]⋅3H2 O}n . The three-connected T-shaped nodes (c) and the alternately interpenetrated and interdigitated architecture of the bilayer motifs in {[Cd2 (bpt)(m-bdc)2 (H2 O)4 ]⋅6H2 O}n . Source: Du et al. [111]. © 2007 American Chemical Society.

N atom from 4,4′ -bpt ligand. They are connected by the m-bdc ligands. Therefore, each Zn(II) displays a T-shaped coordination, with two m-bdc in the horizontal positions and one 4,4′ -bpt in the vertical orientation, which extend to form a 82 ⋅10 topological 2D bilayer network (Figure 6.51). Two such adjacent bilayers are alternately interdigitated in a “tongue-and-groove” stacking pattern. Similarly, another 2D double-layer network has been prepared by replacing the 4,4′ -bpt ligand with 3,4′ -bpt (3,4′ -bpt = 1H-3-(3-pyridyl)-5-(4-pyridyl)-1,2,4-triazole). In the structure of MOF [Zn(m-bdc)(3,4′ -bpt)]n , the Zn(II) centers are coordinated by two N atoms of two 3,4′ -bpt ligands and four O atoms of three m-bdc, exhibiting distorted octahedral environments, which are bridged by m-bdc ligands to form 1D double chains. The 3,4′ -bpt ligands connect the double chains to constructed a 2D double-layered network [112]. A 2D + 2D → 2D interpenetrated bilayer based on a 2D (3,4) motif has been prepared from mixed-ligands of 1,3,5-tris(1-imidazolyl)benzene (tib) and bdc under solvothermal conditions [113]. In MOF [Cd2 (tib)2 (bdc)(NO3 )2 ⋅2H2 O⋅4DMF]n , the Cd centers are seven-coordinated in pentagonal–bipyramidal geometries. Adjacent Cd(II) centers are connected by tib ligands to form ladder chains, which are linked by bdc to form a (3,4)-connected 2D layer with large window of 11.238 × 13.739 Å2 . Due to the large rectangular windows, two adjacent individual layers are interpenetrated with each other to generate 2D + 2D → 2D interpenetrated bilayer networks (Figure 6.52). An interdigitated and interpenetrated bipillared-bilayer MOF, {[Cu6 (pybz)8 (OH)2 ]⋅I5 − ⋅I7 − }n , was prepared from 4-pyridyl benzoate (pybz) and Cu(II) salt by

327

328

6 The Structures of Metal–Organic Frameworks

(a)

(b)

(c)

Figure 6.51 (a) The coordination environments of the Zn(II) atoms in {[Zn2 (m-bdc)2 (4,4′ -bpt)(H2 O)3 ]⋅H2 O}n . (b) The 2D coordination bilayer motif. (c) Schematic representation of the bilayer motif with 82 ⋅10 topology. Figure 6.52 Simplified 2D + 2D → 2D interpenetrated bilayer network in MOF [Cd2 (tib)2 (bdc)(NO3 )2 ⋅2H2 O⋅4DMF]n . Source: Xue et al. [113]. © 2015 Royal Society of Chemistry.

using iodine as a template. The asymmetric unit of such complex contains six independent Cu2+ ions, eight pybz ligands, and 12 iodine atoms, in which two μ3 -OH-bridged tri-copper units are composed of two five-coordinated Cu(II) and one four-coordinated Cu(II) [114]. These trimers are linked by four pybz ligands to a rhombohedral-grid sheet with window size of c. 15.6 × 15.6 Å2 . Two such sheets are further connected by two pybz ligands to form a bipillared bilayer. The rest of pybz ligands hang on the sheet as dangling arms, by the carboxylate groups bridging pairs of Cu(II) ions and the nitrogen atom keeping uncoordinated. The bipillared-bilayer

6.3 Two-dimensional MOFs

Dangling arm N

O N

OH

Cu O

Å 15.6

Cu O O Cu OO

O

14.5 Å

O

N

15.6 Å

(a)

(b)

Polyiodide driven Interdigitation

Layer

Two fold interpenetration

Assembly

(c)

Figure 6.53 (a) Perspective view of the μ3 -OH-bridged tri-copper unit. (b) Perspective view of the bipillared bilayer with cuboidal boxes and dangling pybz arms. (c) Proposed structural evolution from twofold interpenetration and interdigitation into 3D framework (the tri-copper units are simplified to polyhedrons). Source: Yin et al. [114]. © 2012 American Chemical Society.

motif is composed of face-sharing cuboidal boxes (c. 15.6 × 15.6 × 14.5 Å3 ) from another point of view. These bilayers interpenetrate each other in a parallel/parallel inclined mode leading to a 2D + 2D → 2D architecture. The adjacent interpenetrated pillar bilayers are interdigitated to generate a stable 3D supramolecular structure via strong hydrogen bonding and weak offset π· · ·π interaction (Figure 6.53). Although the framework has mutual interpenetration and interdigitation, there still remain two kinds of 1D channels, which are filled with I5 − and I7 − ions, respectively. It’s worth mentioning that this compound has displayed multifunctionality. A double-(6,3)-layer cationic MOF, [Zn2 (Tipa)2 (OH)]⋅3NO3 ⋅12H2 O (FIR-53, FIR denotes Fujian Institute of Research, Tipa = tris(4-(1H-imidazol-1-yl)phenyl)amine), was assembled from neutral Tipa ligand and Zn(II) [115]. In the structure of FIR-53, each Zn(II) center is coordinated by three nitrogen atoms of three Tipa ligands and one oxygen atom of the μ2 -OH group, displaying a tetrahedral geometry. The Zn(II) centers are connected by the μ3 -Tipa ligands to form a (6,3)-layer. Adjacent layers are linked to a double-layer network by the μ2 -OH bridges (Figure 6.54). Such double-(6,3)-layers are stacked together via π· · ·π interactions, possessing 1D ellipse

329

330

6 The Structures of Metal–Organic Frameworks

O1

(a)

(b)

(c) Figure 6.54 (a) The coordination environment in FIR-53. (b) The double layer in FIR-53. (c) Crystal-packing diagram of FIR-53 with large channels.

channels with the window size of 18 × 13 Å2 . The accessible volume of FIR-53 is 41.6%. FIR-53 was applied to trap Cr2 O7 2− with high capacity via the SC-SC ion exchange. Two bilayer-network MOFs have been directionally constructed from Co(II) and a flexible tricarboxylate ligand by virtue of solvent templates. [Co3 (BTTB)2 (H2 O)2 ]⋅(DMF)4.5 ⋅(H2 O)4 (470-MOF, H3 BTTB = 4,4′ ,4′′ -(benzene-1,3,5-triyl-tris (oxy))tribenzoic acid) was synthesized by heating H3 BTTB and CoCl2 ⋅6H2 O at 120 ∘ C in DMF [116]. Similar to that of 470-MOF, [Co3 (BTTB)2 (H2 O)2 ]⋅(DMF)3 ⋅ (H2 O)4 ⋅(dioxane)2.5 (471-MOF) was prepared by adding 1,4-diethylene dioxide, which functions as the templates. The asymmetric units for the two MOFs consist of two independent Co(II) ions, 1/3 BTTB ligand, and one coordinated water. One

6.3 Two-dimensional MOFs

Co(II) center adopts a octahedral coordination geometry, which is completed by six oxygen atoms from six BTTB ligands. The other Co(II) center is tetrahedral sphere and coordinated by three oxygen atoms of three BTTB ligands and one terminal water. Three Co(II) centers are linked by two triple carboxylate bridges to from a trinuclear [Co3 (O2 CR)6 (H2 O)2 ] unit, which connects with six other one through six three-connected BTTB to form a 2D (3,6)-connected bilayer network. Seen parallel to the ab plane, such bilayers in 470-MOF and 471-MOF are different, caused by the diverse conformation of BTTB ligands and Co3 SBUs in the two MOFs: hexagonal pores with sizes of 13.0 × 13.0 Å2 are observed in 470-MOF, while it shows distorted quadrangular meshes with dimensions of 14.0 × 2.0 Å2 in 471-MOF. Serving Co3 SBUs as six-connected nodes and BTTB ligands as three-connected nodes, the 2D bilayer network can be simplified as a (3,6)-connected MoS2 -H net with Schläfli symbol of (43 ⋅612 )(43 )2 and kgd network with (43 )2 (46 ⋅66 ⋅83 ) for 470-MOF and 471-MOF, respectively. Adjacent bilayers arrange in –(AB)n – and –(ABCDEF)n –mode for 470-MOF and 471-MOF, respectively, which reduce the sizes of channels (Figure 6.55). Homochiral bilayer MOFs with nanoscale channel have been obtained from the mixed ligands of proline derivatives and bipyridines. MOF 1-L and 1-D, formula as Co2 [(PBP) (bpy)3 (OH)2 (H2 O)2 ]⋅bpy (PBP = phenyl-4,4′ -bis[carbonyl-N-(proline)]), are obtained by the reaction of L-PBP or D-PBP with bpy, respectively [117]. The asymmetric unit of 1-L or 1-D contains two independent octahedral Co(II) centers, one PBP, four bpys, two hydroxyl ions, and two coordinated water molecules. 1-L and

(a)

(b)

(c)

(d)

(e)

(f)

Figure 6.55 (a, d) View of the 2D bilayer network along the c-axis of 470-MOF and 471-MOF. (b, e) Schematic view of the (3,6)-connected 2D network. (c, f) Crystal-packing diagram. Source: Chen et al. [116]. © 2015 Royal Society of Chemistry.

331

332

6 The Structures of Metal–Organic Frameworks O

OH N

O

O

N

HO O PBP

(a)

N

OH

O N

O N

O

N

N BPY

HO

N BPEP

O BPBP

c b a

(b)

a

(c)

Figure 6.56 (a) Structure of linkers and their extension modes. (b) Crystal-packing diagram of 1-L and simplified topology of the 1-L network. (c) Crystal-packing diagram of 2-L and simplified topology of the 2-L network. Source: Zhuo et al. [117]. © 2017 American Chemical Society.

1-D exhibit bilayer networks. When viewed from the side, the bilayers look like ladders, having two different channels along the b axis. The arms composed of Co(II) and bpy are parallel in the same plane, which interlaced with those in the adjacent planes (Figure 6.56). The bilayer networks can be simplified as topological type 4L11 (4^3.6^3). The interspace between bilayer is occupied by free bpys via π· · ·π interactions. The pore size of above bilayers is broadened by extending the lengths of both linkers. Two novel 2D bilayer MOFs with very large channels, Cd2 [(BPBP)2 (BPEB) (H2 O)] (2-L and 2-D), are constructed from corresponding long ligands of biphenyl-4,4′ -bis[carbonyl-N-(proline)] (BPBP) and 1,4-bis[2-(4-pyridyl)ethenyl] benzene (BPEB). Different from 1-L and 1-D, Cd(II) ions in 2-L and 2-D adopt two different coordination modes: distorted-square–pyramidal geometries and distorted-octahedral geometries. The structures of 2-L and 2-D display similar 2D ladder-like bilayer networks; however, the arms of the ladders are composed of L-BPBP and BPEB alternatively, and the bars are composed of only L-BPBP. The point symbol for the 2-L and 2-D is {6^3}{6^5.8}. Adjacent bilayers stack closely in an AB mode. Those homochiral MOFs have been applied to enantioseparate diols of different sizes in good enantiomeric excess (ee%). A trilayered polythreading 2D MOF, hierarchically assembled from 2D square-grid networks, has been synthesized by a hydrothermal reaction of Cd(OH)2 and mixed ligands squaric acid and 1,2-bis(4-pyridyl)ethane (bpe) in water. In the assembly process, the protonated bpe cations function as controllable structural directing agent [118]. The structure of MOF [Cd(C4 O4 )2 (H2 O)2 ][Cd(C4 O4 )0.5 (Hbpe)(H2 O)3 ] contains two independent polymeric motifs with opposite charge. The anionic one is fabricated by Cd(II) and μ2 -1,3-squarate, displaying a 8.26 × 8.34 Å2 2D square-grid sheet. The other cationic

6.3 Two-dimensional MOFs

(a)

(c)

(b)

Figure 6.57 (a) 2D anionic square grid layer. (b) 2D cationic motif showing side arms. (c) 2D trilayered polythreading network involving a 2D cationic layer with dangling arms, simultaneously interlocked with two 2D anionic square grid layers.

motif is built from Cd atoms, squarate, and monoprotonated bpe, which is a (4,4)-topological 2D sheet with dangling arms (Hbpe+ ). The Hbpe+ arms on both sides of the 2D cationic layer intercalate the grids of two adjacent anionic sheet (above and below), resulting in a 2-nm-thick trilayered network (Figure 6.57). Such structure represents an unprecedented polythreading 2D(+) + 2 × 2D(−) → 2D architecture.

6.3.5

Other 2D MOFs

In contrast to triangles, squares, and hexagons, the regular pentagons cannot tile the plane, thus the incorporation of such five-sided polygonal into 2D MOFs is undeveloped. Nonetheless, the convex pentagons can tile the plane; there are 14 different tilings containing such convex pentagons: five (5,3) nets, eight (5,3 4 ) nets, and one (5,3 6 ) net [119–121]. Three of the 14 tilings consist of congruent semiregular pentagons (Figure 6.58). The first Shubnikov-type (5,3 4 ) network (Figure 6.58b) was observed in MOF [Cu2 (pyz) (CN)2 ]⋅[CuCN]∞ , which has been described earlier in Section 3.3.3 [107]. The structure of MOF [(HMTA)3 (Cu2 (μ-O2 CCH2 CH3 )4 )5 ]n (HMTA =

(a)

Figure 6.58

(b)

(c)

Three of the 14 tilings that contain congruent pentagons.

333

334

6 The Structures of Metal–Organic Frameworks

Figure 6.59 View of the (5,3 4 )-network in [(HMTA)3 (Cu2 (𝜇-O2 CCH2 CH3 )4 )5 ]n . Source: Based on B. Moulton et al. [122].

hexamethylenetetramine) represents the first example of the (5,3 4 )-network (Figure 6.58a). Blue crystals were obtained from the tetrahedral HMTA and a dimetal tetracarboxylate cluster, i.e. Cu2 (μ-O2 CR)4, via diffusion methodology. In the structure, HMTA nodes display a tetrahedral geometry and coordinate to either three or four dinuclear copper carboxylate spacers in a 2 : 1 ratio, forming a 2D curved sheet with pentagon nets (Figure 6.59) [122]. Similar network was also observed in MOF {[Cu(tpt)(MeCN)](ClO4 )}∞ (tpt = 2,4,6-tris(4-pyridyl)-1,3,5-triazine) (Figure 6.60) [123]. This MOF was prepared from [Cu(MeCN)4 ](ClO4 ) and excess tpt via solvothermal synthesis method. In the structure, the fragments of a [Cu(tpt)4/3 ] fragment and a [Cu(tpt)2/3 (MeCN)2 ]

Figure 6.60 et al. [123].

View of the (5,3 4 )-network in {[Cu(tpt)(MeCN)](ClO4 )}∞ . Source: Based on Liu

6.3 Two-dimensional MOFs

are connected by a tpt ligand to form the [{Cu(tpt)4/3 }(Cu(MeCN)2 (tpt)2/3 )]2+ cation. The Cu(I) center in the [Cu(tpt)4/3 ] unit is coordinated by four N atoms of four tpt, while the Cu(I) in the [Cu(tpt)2/3 (MeCN)2 ] shows a distorted tetrahedral geometry completed by two N atoms of two tpt ligands and two MeCN. Each tpt ligand functions as a three-connecting node linking three Cu(I) atoms. Each [Cu(tpt)4/3 ] unit connects four [Cu(tpt)2/3 (MeCN)2 ] units by four tpt ligands, and each [Cu(tpt)2/3 (MeCN)2 ] unit links four [Cu(tpt)4/3 ] units by four tpt ligands, forming the [{Cu(tpt)4/3 }(Cu(MeCN)2 (tpt)2/3 )]2 4+ unit, which interconnects each other by sharing tpt ligands to construct the rare (5,3 4 )-network with the topology of Schläfli symbol (53 )2 (54 82 ). The perchlorate anions occupy the ring centers. Different layers are stacked along the c-axis via the strong π· · ·π interactions (3.24 Å) between the pyridyl groups in adjacent layers. Besides these networks with uniform grids, there are many other 2D layers having nonuniform nets, such as the Archimedean networks in which all nodes adopt the same connection type but have different polygons and the Catalan networks in which all nodes are connected by uniform polygons but have two or more types of nodes (the (5,3 4 )-sheets discussed earlier belong to the Catalan networks as well) [119]. The CdI2 -type network consists of uniform quadrangles like the (4,4)-square net; however, it has two types of nodes in 2 : 1 ratio: three-connecting as in a (6,3) net and six-connecting as in a (3,6) net. Hence, such topology can be represented by (4,3 6 ), which is another one of the Catalan nets. The formation of (4,3 6 ) CdI2 net can be regarded as the offset overlap of two (6,3) nets, which is an intermediate net between the (6,3) and (3,6) net. As the interpenetration usually takes place in (6,3) net and the (3,6) net is often too dense, the intermediate (4,3 6 ) CdI2 net may be fitted for achieving good porosity. Two cationic CdI2 -type topological MOFs, {[Cd(TIPT)2 ](OTs)2 ⋅H2 O}n and {[Cd(TIPT)2 ](OTf)2 ⋅3.5DMF⋅2.5H2 O}n (TIPT = 2,4,6-tris[4-(1H-imidazole-1-yl) phenyl]-1,3,5-triazine, OTs = p-CH3 C6 H5 SO3 − , OTf = CF3 SO3 − ), have been built up from the rigid triangular TIPT ligand and Cd2+ [124]. In both structures, each Cd2+ ion is octahedral coordination geometry coordinated by six nitrogen donors from six TIPT ligands, and each TIPT ligand links three Cd2+ nodes, leading to 2D cationic networks. From topological analysis, they reveal the CdI2 -type net with the topology of Schläfli symbol (46 66 83 )(43 )2 if the TIPT ligands are simplified as a three-connecting nodes and the Cd2+ ions as six-connecting nodes (Figure 6.61). In the CdI2 nets, the TIPT ligands are noncoplanar, so the 2D layers are concavo-convex, which renders porosity with a potential solvent area volume of 1860 Å3 (14%) and 3017.7 Å3 (33.6%), respectively. An anionic CdI2 -type topological net, [Cu2 (tci)2 ]2− (tci = tris-(2-carboxyethyl) isocyanurate), has been self-assembled by the flexible C3 -symmetric H3 tci ligand and copper acetate [125]. In the 2D layered network, each independent CuII center shows a distorted tetragonal–pyramid geometry completed by five oxygen atoms from five tci ligands. The tci ligand serves as a pentadentate ligand, two carboxylate groups of which adopt the same syn–syn coordination fashion, while the other one links one CuII ion in a monodentate mode. Four coordination

335

336

6 The Structures of Metal–Organic Frameworks

(a)

(b)

(4, 36)net

Figure 6.61 Perspective view of a 2D layer in MOFs {[Cd(TIPT)2 ](OTs)2 ⋅H2 O}n and {[Cd(TIPT)2 ](OTf)2 ⋅3.5DMF⋅2.5H2 O}n (a) and schematic view of CdI2 -type topology (b). Figure 6.62 View of a 2D layer in [Cu5 (tci)2 (OH)2 (H2 O)8 ][Cu2 (tci)2 ]⋅11H2 O.

carboxylates (syn–syn coordination mode) connect two adjacent copper ions to yield paddle-wheel clusters (Figure 6.62). As a result, a non-interpenetrating anionic CdI2 -type network is constructed by simplifying the paddle-wheel Cu(II) clusters as six-connecting nodes and the tci as three-connecting nodes, which can be represented by Wells notation (4,3 6 ) belonging to the Catalan nets. Interestingly, such 2D anionic layers, and the pentanuclear copper cluster cation [Cu5 (tci)2 (OH)2 (H2 O)8 ]2+ located in the cavities between adjacent layers, are connected by the cyclic (H2 O)18 clusters via hydrogen bonds to form a supramolecular framework [Cu5 (tci)2 (OH)2 (H2 O)8 ][Cu2 (tci)2 ]⋅11H2 O possessing 1D channels with a channel diameter of 6.7 Å. A binodal (3,6)-connected net with a kgd topology has also been built from tci ligand [126]. The 2D MOF, formulated as [(Me2 NH2 )Zn(tci)⋅0.5DMF]n , was prepared by the reaction of tci ligand and Zn(II) in mixed solvents of DMF–EtOH–H2 O. In the structure, Zn(II) ion adopts a ZnO5 square–pyramidal geometry, which is

6.3 Two-dimensional MOFs

Figure 6.63

View of a binodal (3,6)-connected net with a kgd topology.

coordinated by five oxygen atoms of five carboxyl groups from five tci ligands. Adjacent Zn(II) centers are bridged to a paddle-wheel SBU by four carboxyl groups (syn–syn coordination mode). The tci ligand displays a cis–cis–trans conformation, which connects the paddle-wheel SBUs to construct a non-interpenetrating 2D layer. If the tci ligands are simplified to three-connecting nodes and paddle-wheel SBUs to six-connecting node, the layer is regarded as a rare kgd network (Figure 6.63). The three-connecting nodes (vertex symbol of 43 ) are twice as many as the six-connecting nodes (vertex symbol of 46 .66 .83 ), so the network can be represented as the Schläfli symbol of {43 }2 {46 .66 .83 }. Another example of kgd network can be seen in the 2D layered MOF of [Cd(PPA)2 ]n (HPPA = 3-pyridinepropionic acid) [127]. In the structure, the Cd(II) ion displays a distorted octahedal coordination geometry, which is completed by four oxygen atoms from monodentate carboxylate of four independent PPA ligands and two nitrogen atoms from other two PPA ligands. Adjacent Cd(II) centers are linked by two PPA to generate a loop, which is connected by sharing bidentate carboxylate and Cd(II) centers, leading to a 2D layer with the channels of 9.54 × 3.99 Å2 . The network contains one three-connected (PPA ligands) and one six-connected (CdII centers) topologically nodes, exhibiting a (3,6)-connected kgd topology with the Schläfli symbol of {43 }2 {46 .66 .83 } as well. A 2D Archimedean cem layer has been built from a multidentate N-donor and a biphenyl-based polycarboxylate [128]. MOF {[Cd(Hpptpz)(bpba)]⋅2H2 O}n (Hpptpz = 2-(3-(4-(pyridin-4-yl)phenyl)-1H-1,2,4-triazol-5-yl)pyrazine, H2 bpba = 3,4′ -biphenylbicarboxylic acid) was prepared from Hpptpz, H2 bpba, and Cd(II) under hydrothermal conditions. In the structure, the Cd(II) center is coordinated by three oxygen atoms from three bpba ligands and three nitrogen atoms from two Hpptpz ligands, to furnish a CdN3 O3 octahedral coordination geometry. Two inversion-related Cd(II) ions are linked to a centrosymmetric [Cd2 (Hpptpz)2 ] subunit by a pair of head-to-tail arranged Hpptpz ligands, which is further extended to the final framework by the bpba ligands. If each Cd(II) is simplified into five-connected node, such structure can be simplified to a five-connected cem Archimedean layered net with (33 ⋅44 ⋅53 ) topology, which is composed of triangles and rectangles.

337

338

6 The Structures of Metal–Organic Frameworks

2D layered MOFs, in which a single layer is a planar macromolecule constructed by strong coordination bonds and their overall framework is formed by the stacking of these layers via weak van der Waals forces, hydrogen bonding, π· · ·π stacking, etc. have received an extensive attention in recent years due to their unique features. A great number of novel 2D MOFs are coming out constantly. The structure of 2D MOF is mostly dependent on the network of the layer; however, the minor adjustment of their stacking mode, interlayer space, interlayer distance can also affect the overall frameworks. A series of novel layered MOFs with intercalated structures have been assembled by inserting guest species into the interlayer region of 2D MOF (W1) based on hydrogenated Schiff base hsb-2 and Zn(ClO4 )2 . In MOF W1, disordered ClO4 − anions are filled in the spaces between neighboring layers, half of which are replaced by chiral molecules (L/D-pyroglutamate) through ion exchange resulting in two chiral isomers, [Zn(hsb-2)(H2 O)2 ⋅S⋅ClO4 ⋅3H2 O] (1-S; S = (S)-5-oxopyrrolidine-2-carboxylate) and [Zn(hsb-2)(H2 O)2 ⋅R⋅ClO4 ⋅3H2 O] (1-R; R = (R)-5-oxopyrrolidine-2-carboxylate) [129]. Both of 1-S and 1-R adapt a similar asymmetric unit, having one Zn(II), one hsb-2, two coordinated water molecules, one ClO4 − , one chiral guest S or R, and three free water molecules. Ligand hsb-2 shows a V-shape, bridging Zn(II) to generate a 2D net with a square window. Interestingly, the 2D layers of 1-S and 1-R have a “mirror symmetry,” indicating that the addition of the chiral guest induces chirality in the layered hosts. Different layers are stacked orderly in an ABCD mode to compose a 3D framework of 1-S and 1-R. In 1-S, layer B overlaps exactly with layer A by rotating 90∘ , with layer C by rotating 180∘ , layer D by rotating 270∘ , and layer A of the next cycle by rotating 360∘ anticlockwise, respectively, and so forth. This phenomenon is the same in 1-R, except for the counterclockwise rotation changing to clockwise rotation. Between the same two adjacent layers in 1-S or 1-R, the arrangement of each chiral guest is identical. The chiral guests between layer A and B in 1-S overlap with those between layer B and layer C by rotating 90∘ anticlockwise, and so on, similar to the layered hosts. Likewise, they are arranged in the same way in 1-R except for the counterclockwise rotation changing to clockwise rotation (Figure 6.64). The overall 3D structures of 1-S and 1-R have a “mirror symmetry” too, which illustrates that they are supramolecular enantiomers. The interlamellar space of the 2D layered MOF W1 is flexible and available to accommodate guest molecules with different sizes. A great number of sulfonates guests have been successfully introduced to prepare several intercalated compounds with multifunctional properties, which can be tuned by altering the guest molecules or their arrangements [130]. Compounds [Zn(hsb-2)(H2 O)2 ⋅G1⋅DMF⋅2H2 O]n , [Zn(hsb-2)(H2 O)2 ⋅G2⋅DMF⋅2H2 O]n , and [Zn(hsb-2)(H2 O)2 ⋅G3⋅DMF⋅2H2 O]n , [Zn(hsb-2)(H2 O)2 ⋅2G4]n (G1 = naphthalene-2,7-disulfonate; G2 = 1-naphthol-3,6disulfonate; G3 = naphthalene-1,6-disulfonate; G4 = p-aminobenzenesulfonate) were synthesized by addition of sulfonate G1–G4 to the layered MOF in situ prepared from the hsb-2 and Zn(ClO4 )2 . In those structures, the hsb-2 ligands and Zn(II) centers construct the similar layered networks with square windows. The interlayer distances are 9.2, 9.2, 9.3, and 8.9 Å, respectively, longer than that of

6.3 Two-dimensional MOFs

(a)

1-S

Figure 6.64

b

(b)

1-R

Crystal packing diagrams of 1-S and 1-R.

Zn S C N O

c a

b

9.2 Å

c

Zn S C N O

a

(a)

9.2 Å

(b)

b a

c Zn S C N O

c

9.3 Å

b

(c)

a

Zn S C N O

8.9 Å

(d)

a a

b

b c

c

(e)

17.1Å

12.4 Å

(f)

Figure 6.65 Crystal structures of the intercalated layered MOFs with different interlayer distances (a–f). Source: Wen et al. [130]. © 2016 John Wiley & Sons.

the original 2D MOF (7.5 Å) (Figure 6.65a–d). Different sulfonates guests occupy the space between adjacent layers. Other sulfonate guest molecules such as bis(2-sulfonatostyryl) biphenyl (Figure 6.65e), acid orange 10, 2-anthraquinone sulfonate (Figure 6.65f), (1R or 1S)-(−)-10-camphorsulfonate can also be intercalated into layered MOF W1. The

339

340

6 The Structures of Metal–Organic Frameworks

interlayer distances vary from 10.4 to 17.1 Å according to the different sizes of these accommodated guest molecules. Importantly, the properties of the guest molecules were tuned and some new performances were obtained after confining them into the interlayer region. Such a hybrid approach provides an efficient strategy to design and prepare multifunctional materials. Very recently, the transformation of 2D MOF materials to nanosheets has attracted considerable attention. 2D MOF nanosheets are the 2D MOFs with a single layer or several layers in the thickness direction less than 10 nm, which have larger surface area and more accessible active sites on the surface, so they display many unique physical and chemical properties. In general, there are two strategies to prepare the 2D MOF nanosheets: top-down methods, i.e. exfoliation strategies, and bottom-up methods, including wet-chemical synthesis and chemical vapor deposition [131]. Top-down methods represent the simple and efficient strategy to prepare 2D MOF nanosheets. In 2D layered MOF {Zn2 (bdc)4 (H2 O)2 ⋅2DMF}n (MOF-2), each layer is constructed by the paddle-wheel Zn2 clusters and bdc, which are stacked together by hydrogen-bonding interactions. Its nanosheet is prepared by the ultrasonic treatment of the bulk crystals of MOF-2 in acetone at room temperature. The exfoliation yield of MOF-2 is about 10%. The formation of MOF-2 nanosheets via top-down delamination has been determined by the Tyndall effect, the SEM and AFM measurements. The thickness of MOF-2 nanosheets is 1.5–6.0 nm and the lateral size is from 100 nm to several micrometers. MOF-2 nanosheets restack after volatilizing the solvent, and the PXRD of its two reflection peaks can be ascribed to the (001) and (002) of MOF-2, indicating the crystal structure of the nanosheets [132]. An intercalation/chemical exfoliation method has been developed to synthesize ultrathin MOF nanosheets with high yield [133]. A chemically labile dipyridyl ligand, 4,4′ -dipyridyl disulfide (dpds), is intercalated into a known layered porphyrinic MOF, i.e. Zn2 (PdTCPP) (PPF-1), via one pyridinic N of dpds connecting to the Zn2 (COO)4 SBU. As a result, a new intercalated MOF Zn2 (PdTCPP)(dpds) is formed with the interlayer distance varied from 9.8 to 22.6 Å, which can be exfoliated into single-layer freestanding MOF nanosheets in 57% yield by selectively cleaving the disulfide bond of dpds using trimethylphosphine (Figure 6.66). The yield is much higher than the Zn2 (PdTCPP) nanosheets prepared by the sonication exfoliation.

Disulfide ligand Breakage

Intercalation

Exfoliation

4,4ʹ-dipyridyl disulfide

Layered MOF crystals

Trimethylphosphine

Intercalated MOF crystals

2D MOF nanosheets

Figure 6.66 Schematic illustration of the process for the synthesis of 2D MOF nanosheets via an intercalation and chemical exfoliation approach. Source: Ding et al. [133]. © 2017 American Chemical Society.

6.3 Two-dimensional MOFs

The high resolution TEM image clearly shows the lattice fringes of the exfoliated multilayer MOF nanosheets with an interplanar distance of 1.65 nm, belonging to the (100) plane of the intercalated crystals, which indicates that the crystalline structure is maintained. A bottom-up method, i.e. a surfactant-assisted synthetic method, has been firstly developed to prepare uniform ultrathin 2D MOF nanosheets with sub-10 nm thickness [134]. In the structure of the bulk Zn-TCPP MOF, each layer is built up of Zn2 (COO)4 paddle-wheel SBU and TCPP ligands, and different layers are stacked in an AB pattern along the vertical direction to form the overall framework. When the polyvinylpyrrolidone (PVP) is used as surfactant during the preparation of Zn-TCPP, it selectively attaches on the Zn2 (COO)4 metal nodes of the MOF nucleus, weakening the interlayer interactions and preventing the stacking of the layers along the vertical direction, leading to the anisotropic growth of MOFs and then formation of ultrathin MOF nanosheets (Figure 6.67). STEM and AFM image reveals the Zn-TCPP nanosheets with lateral size of 1.2 ± 0.4 μm and thickness of 7.6 ± 2.6 nm. The structure of Zn-TCPP nanosheets coincides with that of bulk Zn-TCPP MOFs, which is confirmed by XRD and SAED. A high-quality conductive MOF membrane has been prepared by an air−liquid interfacial growth method. The semiconductive MOF, i.e. Ni3 (HITP)2 , consists of nickel ions and HITP ligands (HITP = 2,3,6,7,10,11-hexaiminotriphenylenesemiquinonate), shows a 2D graphene-like honeycomb network. Adjacent layers stack in an AB mode with an interval of 3.5 Å, possessing 1D channels with an open window size of ∼1.4 nm [135]. The Ni3 (HITP)2 membrane was prepared by heating of HATP⋅6HCl (HATP = 2,3,6,7,10,11-hexaaminotriphenylene), NiCl2 ⋅6H2 O, and trimethylamine to 60 ∘ C. The in situ generated HITP self-assembles with Zn(II) to form Ni3 (HITP)2 nanoparticles at the air–liquid interface, which acts as a “smooth substrate” and closely packs to form a nanometer-thick uniform layer due to the interface-confining effect. The MOF membrane generates after the air–liquid interface is fully covered with this Ni3 (HITP)2 layer, the thickness Traditional synsthesis Isotropic Nucleation

crystal growth

Zn2(COO)4 Bulk crystal

+

TCPP

Surfactant

Anisotropic

Nucleation

crystal growth Ultrathin nanosheets Surfactant-assisted synthesis

Figure 6.67 Scheme of the traditional synthesis of Zn-TCPP bulk crystal and surfactant-assisted synthesis of Zn-TCPP nanosheets. Source: Zhao et al. [134]. © 2015 John Wiley & Sons.

341

342

6 The Structures of Metal–Organic Frameworks

of which is dependent on the reaction time. A ∼100-nm-thick membrane with a compact and smooth surface was obtained after reaction for three minutes. Transferring the MOF membrane to SiO2 /Si substrates, the field effect transistors (FETs) with a crystalline microporous MOF channel layer were fabricated [136].

6.4 Three-dimensional MOFs 3D MOFs, which generally exhibit more complicated geometries, structural flexibility, and intriguing topologies as compared to the low-dimensional ones, possess highly-order crystalline porous frameworks with ultrahigh porosity, large surface areas, unique host–guest dynamics, thermal stability and mechanical flexibility, etc. The crystal structures of some representative 3D MOFs with high porosity and surface areas are shown in Figure 6.68 [137]. Owing to their structural and

MOF-5

HKUST-1

[CuSiF6(4,4ʹ-bpy)]

MOF-14

MOP-1

ELM-11

MIL-47

MIL-53

MIL-88

MOF-177

Cr-MIL-100

Cr-MIL-101

Ni-CPO-27

UiO-66

ZIF-8

PCN-14

DO-MOF

[Be12(OH)12(BTB)4]

UMCM-2

NOTT-116

MOF-200

UTSA-20

IRMOF-74-XI

NU-125

Figure 6.68 Crystal structures of representative 3D MOFs with high porosity and surface areas. Source: Silva et al. [137]. © 2015 Royal Society of Chemistry.

342

6 The Structures of Metal–Organic Frameworks

of which is dependent on the reaction time. A ∼100-nm-thick membrane with a compact and smooth surface was obtained after reaction for three minutes. Transferring the MOF membrane to SiO2 /Si substrates, the field effect transistors (FETs) with a crystalline microporous MOF channel layer were fabricated [136].

6.4 Three-dimensional MOFs 3D MOFs, which generally exhibit more complicated geometries, structural flexibility, and intriguing topologies as compared to the low-dimensional ones, possess highly-order crystalline porous frameworks with ultrahigh porosity, large surface areas, unique host–guest dynamics, thermal stability and mechanical flexibility, etc. The crystal structures of some representative 3D MOFs with high porosity and surface areas are shown in Figure 6.68 [137]. Owing to their structural and

MOF-5

HKUST-1

[CuSiF6(4,4ʹ-bpy)]

MOF-14

MOP-1

ELM-11

MIL-47

MIL-53

MIL-88

MOF-177

Cr-MIL-100

Cr-MIL-101

Ni-CPO-27

UiO-66

ZIF-8

PCN-14

DO-MOF

[Be12(OH)12(BTB)4]

UMCM-2

NOTT-116

MOF-200

UTSA-20

IRMOF-74-XI

NU-125

Figure 6.68 Crystal structures of representative 3D MOFs with high porosity and surface areas. Source: Silva et al. [137]. © 2015 Royal Society of Chemistry.

6.4 Three-dimensional MOFs

functional properties, 3D MOFs have been widely used for many applications such as gas storage, chemicals separation, water purification, optoelectronics, proton conduction, dielectrics, drug delivery, chemical sensors, and biomedical applications [138–146]. Over the past two decades, the most efforts have been devoted to the designed construction of carboxylate- and pyridyl-based 3D MOFs. Meanwhile, zeolite-like MOFs, particularly zeolitic imidazolate frameworks (ZIFs), which involve the use of imidazolate ligands, represent a very important family of 3D MOFs. However, the use of phosphonate and sulfonate ligands for the formation of the framework structures is less studied. The hallmark characteristic of the 3D MOFs is their permanent porosity, which is mainly determined by the geometry and connectivity of the chosen organic ligands. As a result, the adjustments of the geometry, length, functional group of a ligand are important for tuning the pore size and shape and internal surface property of the resulting MOF. In this section, we highlight the advances in 3D MOF synthesis focusing on the selection/design of ligands, which are categorized as carboxylate (ditopic, tritopic, and polytopic), N-heterocyclic, metallo-linker, mixed N-/O-donors, phosphonate, and sulfonate linkers, and others. The construction of some 3D MOFs with unique properties, through ligand design, and the relationship between organic ligands and MOF structures will be described.

6.4.1

Carboxylate Linkers

6.4.1.1 Ditopic Carboxylate Ligands

Since ditopic carboxylate ligands are readily accessible, particularly those commercially available, the ditopic carboxylate ligands have been well investigated. It is known that the formation of M–O–C metal clusters or SBUs is highly dependent on the reaction conditions. As a result, the same precursors (metal salts and organic ligands) can generate different crystal structures through tuning the reaction conditions, which further enrich the structural diversity of the MOFs. Among the diverse organic linkers, carboxylate ligands are of particular interest owing to their preference to stabilize the MOFs through in situ formed SBUs, such as paddle-wheel M2 (CO2 )4 and octahedral Zn4 O(CO2 )6 clusters. Terephthalate (bdc) is the most commonly used ditopic carboxylate ligand and the reactions between bdc and different metal salts gave many 3D MOFs with various crystal structures. One of the most famous examples is Zn-based MOF-5, [Zn4 O(bdc)3 ], reported by Yaghi et al., in which the octahedral Zn4 O(CO2 )6 clusters, known as basic zinc acetate for decades, act as the SBUs to construct the 3D network with elegant cubic structure (Figure 6.69) [148]. This star MOF has unprecedented high apparent surface area (SLangmuir = 2900 m2 /g) and porosity and can be stable up to 300 ∘ C. By replacing bdc ligand with other ditopic carboxylate ligands with different functionalities or lengths as shown in Figure 6.69, a variety of isoreticular metal-organic frameworks (IRMOFs) with the same network topology can be obtained, which can not only introduce diverse functionalities, but also precisely regulate the pore sizes [147]. Besides MOF-5, the use of bdc as ligand has generated many other well-known 3D MOFs, such as MIL series (metal ions can be Cr3+ , V3+ , Fe3+ , Ga3+ , and Al3+ )

343

344

6 The Structures of Metal–Organic Frameworks ⊖

⊖

COO

COO

⊖

⊖

COO

COO

⊖

COO

⊖

COO

⊖

COO

⊖

COO

⊖

COO

S

R ⊖

S

COO

⊖

COO

R-bdc

ttdc

⊖

COO

ndc

O O O O O ⊖

COO

bpdc

⊖

⊖

COO

COO

hpdc

pdc

O O O O O ⊖

COO

tpdc

⊖

COO

⊖

COO

tpdc

bdc

IRMOF-1 (MOF-5)

Zn4O(CH3COO)6 IRMOF-16

Figure 6.69 Some representative ditopic carboxylate ligands and illustration of formation of extended IRMOF networks by replacing acetate with rigid dicarboxylates. Color scheme: Zn (turquoise polyhedra); O (red); C (black). Source: Lu et al. [147]. © 2014 Royal Society of Chemistry.

[149–152], Zr-based UiO series [153], etc. As a particular example, MIL-101(Cr), [Cr3 F(H2 O)2 O(bdc)3 ]⋅nH2 O, which is made from the linkage of bdc ligands and Cr-based trimeric SBUs, displays incredibly large cage diameters of 29 and 34 Å and hexagonal windows of 12 and 16 Å opening [152]. It possesses the ultrahigh BET and Langmuir surface areas larger than 4100 and 5900 m2 /g, respectively (Figure 6.70a). MIL-101(Cr) can remain stable when treated with water, various organic solvents, and even highly acidic solution (such as sulfuric acid) at ambient temperature or in solvothermal conditions, which could be due to the fact that Cr3+ , as hard Lewis acid, can bond strongly to carboxylate. The structures of MIL-53, [M(OH)(bdc)] (M = Cr3+ , Al3+ , Fe3+ , V3+ ), made of chains of corner-sharing MO4 (OH)2 octahedra interconnected through bdc linkers, feature 3D frameworks containing 1D large pore channels with nanometer dimensions [149]. Intriguingly, a very large breathing effect can be observed upon water or CO2 adsorption, resulting from the transition between two major forms, namely the large pore and narrow pore forms (Figure 6.70b). As Zr4+ has a high affinity for hard oxygen donor ligands, Zr-based MOFs usually show high coordination clusters to act as inorganic nodes. UiO-66 exhibits a cubic structure closely packed with octahedral cages (free diameter: 11 Å), which is built up by connecting Zr6 O4 (OH)4 inorganic nodes with bdc linkers [153]. It also shows remarkable thermal and chemical stability.

6.4 Three-dimensional MOFs

19.69 Å ~16 Å

~14.7 Å

~12 Å 7.85 Å

Pentagonal windows

Hexagonal windows +H2O

–H2O

16.83 Å

~29 Å

~34 Å 13.04 Å

Mesoporous cages

(a)

(b)

Figure 6.70 (a) Ball-and-stick view and free dimensions of the pentagonal and hexagonal windows and the two cages in MIL-101(Cr). Source: Serre et al. [149]. © 2002 American Chemical. (b) Transition between the large pore and narrow pore forms of MIL-53. Source: Based on Férey et al. [152]. Copyright 2005, American Association for the Advancement of Science.

Subsequent studies have demonstrated that this unique regular octahedral cage can also be easily expanded with increasing length of the linkers. 2,5-Dioxido-1,4-benzene-dicarboxylate (dobdc) is the organic strut for MOF-74, in which both the aryloxide and carboxylate moieties bonded to the metal sites. Although dobdc is tetra-anionic, it is considered as a ditopic linker because the aryloxide and adjacent carboxylate coordinate to the same SBU. In MOF-74, helical Zn–O–C rods of composition [O2 Zn2 ](CO2 )2 are constructed from six-coordinated Zn2+ centers. The infinite inorganic rod-type SBUs are linked by the benzene units of the dobdc to produce bnn parallel rod packing and 1D channels of dimensions 10.3 × 5.5 Å2 [154]. The systematic expansion of MOF-74, from its original link of one phenylene ring (I), 2, 3, 4, 5, 6, 7, 8, and 11 (II to XI, respectively), afforded an isoreticular series of MOF-74 structures (termed IRMOF-74-I to XI) with pore apertures ranging from 14 to 98 Å (Figure 6.71) [155]. The rigidity, geometry, and shapes of the bridging ligands have strong influence on the structures of the resulting MOFs [156]. The careful choice of the ligands, which can adopt different coordination modes and conformations during the self-assembly process, could result in a wide range of structural diversities. In this regard, the use of the ditopic carboxylate ligands with different bend angles opened a new path toward rich coordination architectures [157]. Chun and Jung have reported the synthesis of a Zn4 O(CO2 )6 -based 3D MOF, [Zn4 O(mip)3 ], with the simple, nonlinear dicarboxylate ligand, 5-methylisophthalate (mip), which is featured by complicated mesh-like pores with narrow passages [158]. Compared to rigid ditopic carboxylate ligands, semirigid ones can adopt diverse coordination modes and exhibit a variety of configurations according to geometric requirements of metal ions/clusters due to their certain flexibility. Xu and coworkers reported the solvothermal reaction of

345

346

6 The Structures of Metal–Organic Frameworks

Figure 6.71 Crystal structure of MOF-74 and chemical structure of organic links used in the synthesis of a series of nine IRMOFs (MOF-74I to XI). Source: Reprinted with permission from Ref. [155]. Copyright 2012, American Association for the Advancement of Science.

Zn(II) nitrate and a V-shaped ligand, 4,4′ -(hexafluoroisopropylidene)bis(benzoic acid) (H2 hfipbb), to generate two interesting enantiomeric 3D MOFs, which exhibit similar (4,4)-connected 3D network with two types of helically twofold and eightfold nanochannels along the c-axis and ultralong helical pitches in the eightfold helical channels [159]. However, the presence of the bent ditopic carboxylate ligands usually lowers the tendency of the MOF structures to form high-dimensional carboxylate-bridged networks [160]. Therefore, the auxiliary ligands, such as 4,4′ -bipyridine, have been widely utilized to update the dimension of the MOFs constructed with bent ditopic carboxylate ligands [161, 162]. For example, Cao and coworkers reported two 3D MOFs based on trinuclear and pentanuclear cobalt cluster units, respectively, which were synthesized from V-shaped H2 hfipbb and linear N-donor 1,2-di(4-pyridyl)ethane under solvothermal conditions at different temperatures [163]. In their structures, 1,2-di(4-pyridyl)ethane as auxiliary bridged ligand played an important role in the formation of high-dimensional networks. Camphoric acid (H2 cam) is an enantiopure ditopic organic linker with rich and colorful coordination modes, which has been widely applied to construct 3D chiral MOFs [164]. For example, reaction of Mg(NO3 )2 ⋅6H2 O with H2 cam results in the immediate formation of dimetallic [Mg2 (Hcam)3 ]+ threefold paddle wheels, which associate together to build the body-centered cubic (bcu) net in [Mg2 (Hcam)3 ⋅3H2 O]⋅NO3 ⋅MeCN [165]. The framework consists of a series of fused Mg12 cages that have 12 water molecules at their centers, creating isolated 0D cavities within the structure. Two series of metal–camphorate frameworks [Ln(D-cam)(CH3 COO)(H2 O)] and [Ln(D-cam)(HCOO)] were synthesized from

6.4 Three-dimensional MOFs

D-cam with acetate (early lanthanides La–Dy) or formate (late lanthanides Tb–Lu and Y) as the auxiliary ligand, respectively [166]. 6.4.1.2 Tritopic Carboxylate Linkers

1,3,5-Benzenetricarboxylate (btc) represents the most common tritopic carboxylate linker in the construction of 3D MOFs. In HKUST-1, each btc linker connects to three dicopper paddle-wheel SBUs to form a T d octahedron. Four linkers occupy alternating triangular faces and six SBUs locate at vertices of the T d -octahedron. Further connection with other units through corner sharing of the octahedron forms a cubic framework with tbo topology [167]. MIL-100(Cr) is built from Cr3 O(CO2 )6 cluster and btc. The oxido-centered chromium trimers are interconnected by the btc linkers to form the “supertetrahedra (ST)” with four chromium trimmers as the vertices and four organic linkers as the triangular faces. The ST is further connected together to obtain the augmented zeolite Mobil Thirty-Nine (MTN) type of framework, containing two types of mesoporous cages [168]. Lanthanide MOF Y(btc)(H2 O)⋅4.3H2 O is a tetragonal porous framework with channels about 5.8 × 5.8 Å2 , built by btc bridging Y atoms. Y-btc can be stable up to 490 ∘ C and can retain the framework after thermal activation [169]. Cage-within-cage porous In-carboxylate framework is also constructed from btc (Figure 6.72). In [(CH3 )2 NH2 ][In3 O(btc)2 (H2 O)3 ]2 [In3 (btc)4 ]⋅7DMF⋅23H2 O (denoted as CPM-5, CPMs = crystalline porous materials), the large Archimedean cage (truncated octahedral cage, denoted as the In24 cage) is formed by 24 mononuclear In3+ sites, each of which is eight-coordinated to oxygen atoms, but serves as a four-connected node. Any two adjacent In3+ sites on the In24 cage are bridged by a btc3− anion using two of its three carboxylic groups. The third carboxylic group of each btc

+

BTC

_

~8.7 Å

~26 Å

BTC

(a)

In12@in24

(b)

Figure 6.72 (a, b) Structures of monomeric In3+ ion, In24 cage, trimeric [(In3 O)(H2 O)3 ] unit, In12 cage, and In12 @In24 cage in CPM-5, respectively. Source: Zheng et al. [170]. © 2010 American Chemical Society.

347

348

6 The Structures of Metal–Organic Frameworks

interconnects the outer cage with the inner In12 tetrahedron cage, which is built via four trimeric [(In3 O)(H2 O)3 ] clusters joining together by four btcs [170]. When btc was elongated to 4,4′ ,4′′ -benzene-1,3,5-triyl-tribenzoicacid (H3 btb), some other 3D MOFs have also been prepared. Cu3 (btb)2 (H2 O)3 ⋅(DMF)9 (H2 O)2 (MOF-14) reveals a pair of interwoven frameworks that are mutually reinforced. It contains remarkably large pores of 16.4 Å in diameter [171]. Zn4 O(btb)2 ⋅(DEF)15 (H2 O)3 (MOF-177) is a (6,3)-connected qom net with the center of the octahedral Zn4 O(CO2 )6 cluster as the site of six-connection and the center of the btb linker as the site of three-connection. MOF-177 has extra-large pores capable of binding polycyclic organic guest such as fullerene and a number of dyes [172]. Further extensions of the btb linkers, two IRMOFs to MOF-177 have been constructed. MOF-180 with 4,4′ ,4′′ (benzene-1,3,5-triyl-tris(ethyne-2,1-diyl))tribenzoate (bte) as linker and MOF-200 with 4,4′ ,4′′ -(benzene-1,3,5-triyl-tris(benzene-4,1-diyl))tribenzoate (bbc) as linker. The cell volumes of MOF-180 and MOF-200 are 1.8 and 2.7 times that of MOF-177, respectively. Both MOF-180 and-200 have extremely high porosity (89% and 90%, respectively) and ultra-high BET surface areas (Figure 6.73) [173]. Other tritopic carboxylate linkers by replacing the central benzene ring with triazine, heptaazaphenalene, a nitrogen atom, a carbon atom, etc. have also been used to prepare 3D MOFs. For example, mesoMOF-1 is based on 4,4′ ,4′′ -s-triazine-1,3,5-triyltri-p-aminobenzoate (tatab) [174]; PCN-htb (PCN = porous coordination polymer) is built from 4,4′ ,4′′ -(1,3,4,6,7,9,9-heptaazaphenalene2,5,8-triyl)tribenzoate (htb); PCN-6′ is assembled from 4,4′ ,4′′ -s-triazine-2,4,6triyltribenzoate (tatb) [175]. They are IRMOFs to HKUST-1. Remarkably, the cell volume of PCN-6′ is 5.5 times that of HKUST-1. Two enantiomorphic frameworks, {Na3 [Cd2 (L-TTA)2 (μ2 -Cl)](H2 O)6 }n and {Na3 [Cd2 (D-TTA)2 (μ2 -Cl)](H2 O)6 }n , have been self-assembled from L- and D-H3 TTA (L-H3 TTA = N,N′ ,N′′ -1,3,5-triazine-2,4,

(a)

(b)

Figure 6.73 A qom net shown in augmented form and crystal structures of representative MOFs in its net. Source: Reprinted with permission from Ref. [173]. Copyright 2010, American Association for the Advancement of Science.

6.4 Three-dimensional MOFs

6-triyltris(L-alanine)) (L-H3 TTA) and transition metal ions under hydrothermal conditions. They display particular (3,4)-connected (63 )(63 .103 ) networks with fascinating architectures built from the distorted trigonal–bipyramid cages. Furthermore, the 3D chiral double helices along three axes lead to the homochiral frameworks [176]. [Zn2 (OH)(tcpa)]⋅2DMF⋅2H2 O (FIR-1; FIR denotes Fujian Institute of Research) is obtained from a long trigonal bridging ligand tris[(4-carboxyl)phenylduryl]amine (tcpa), which shows extraordinary assembly from a 3D ths net to a 3D self-penetrating 3,5-connected net. Such a net-to-net assembly leads to a stable microporous framework with notable CO2 /N2 separation capacity [177]. A redox-active Ca-MOF, {[Ca5 (tcpa)3 (H2 O)6 ]+ }n ⋅nNO3 − (FIR-29) was also synthesized from tcpa. In the structure of FIR-29, each rod-like Ca–COO chain is connected to other six chains by long tcpa to form a 3D rod-packing architecture with a honeycomb lattice of hexagonal channels. The consecutive slipped π· · ·π stacking interactions of photoactive tcpa groups on the surface of hexagonal channels make it an effective photocatalyst [178]. A 3D thorium MOF, [Th(tcpmb)(H2 O)2 ]⋅5.5DMF⋅4(H2 O) (SCU-11, tcpmb = 4-[tris(4-carboxyphenyl)methyl]benzoate) was synthesized from tritopic carboxylate tcpmb. SCU-11 contains a series of cages with an effective size of c. 21 × 24 Å2 (Figure 6.74). The Th4+ in SCU-11 is 10-coordinate with a new bicapped square prism coordination geometry. The bicapped position is occupied by two coordinated water molecules that can be removed to afford a very unique open Th4+ site. The degassed phase (SCU-11-A) exhibits a Brunauer–Emmett–Teller surface area of 1272 m2 /g, enabling it to sufficiently retain water vapor, Kr, and Xe with uptake capacities of 234 cm3 /g, 0.77 mmol/g, 3.17 mmol/g, respectively

(a)

(b)

Figure 6.74 The molecular cage (a) and 3D framework of SCU-11 (b). Source: Wang et al. [179]. © 2018 John Wiley & Sons.

349

350

6 The Structures of Metal–Organic Frameworks

and a Xe/Kr selectivity of 5.7 [179]. Wang et al. have also reported a rare 3D uranyl organic framework, [(CH3 )2 NH2 ][UO2 (dcpb)]⋅0.5DMF⋅15H2 O based on 3,5-di(4′ -carboxylphenyl) benzoic acid (H3 dcpb). In such a structure, three sets of crystallographically equivalent (6,3) wavy nets are entangled together resulting in an extremely rare case of a 2D + 2D → 3D polycatenated framework. However, when btb is used as the linker, a 2D layered structure with graphene-like (6,3) net is afforded [180].

6.4.1.3 Tetratopic Carboxylate Linkers

Tetratopic carboxylate linkers are another type of intriguing building units in MOF constructions and have gained great attention in recent years. MOFs with large porosities could be constructed with tetrahedral carboxylate and symmetrically compatible SBUs. Four-connected square planar clusters and tetrahedral carboxylate ligands in Zn2 (mtba)(H2 O)2 ⋅(DMF)6 (H2 O)5 (MOF-36, mtba = methanetetra(4-benzoate)) are linked into the PtS network [181]. A flu-topology MOF, PCN-521, was formed by the combination of tetrahedral linkers 4′ ,4′′ ,4′′′ ,4′′′′ -methanetetrayltetrabiphenyl-4-carboxylate (mtbc) and the eight-connected [Zr6 (𝜇 3 -OH)8 (OH)8 ](CO2 )8 SBUs in a 2 : 1 ratio. The size of the octahedral cavity of the fluorite structure is 20.5 × 20.5 × 37.4 Å3 . The calculated solvent accessible is 78.50%, and its BET surface area is 3411 m2 /g [182]. Many rigid/flexible nonregular tetrahedral linkers have been used to construct various 3D MOFs. A magnesium MOF, Mg5 (OH)2 (btec)2 (H2 O)4 ⋅11H2 O (H4 btec = 1,2,4,5-benzenetetracarboxylic acid), was prepared from a simple tetratopic carboxylate ligand H4 btec. Such 3D MOF with 1D channels is built by btec ligands and rare pentanuclear magnesium clusters as SBUs. The Mg-MOF demonstrates high fluorescence sensing for carbon disulfide and nitroaromatic compounds [183]. By employing biphenyl-3,3′ ,5,5′ -tetracarboxylate (bptc), an anionic large porous [Et2 NH2 ][In(bptc)]⋅5H2 O (InOF-2) has been obtained, featuring a 3D uninodal four-connected unc-type topological network with a Schläfli symbol of {66 }, which is constructed from tetrahedral mononuclear [In(COO)4 ] nodes bridged by the topologically equal four-connected bptc4− ligands [184]. Further extension of the H4 btec to 1,1′ :4′ ,1′′ :4′′ ,1′′′ -quaterphenyl-3,3′′′ ,5,5′′′ -tetracarboxylic acid (H4 qptca) occurs, which reacted with Eu(NO3 )3 via a solvothermal method affording 3D framework of [Eu2 (qptca)(NO3 )2 (DMF)4 ]⋅(CH3 CH2 OH)3 (ZJU-88). ZJU-88 adopts a four-connected topology with an RCSR symbol of lvt when the binuclear Eu2 (COO)4 SBU and tetracarboxylate ligand qptca are both regarded as four-connected nodes. ZJU-88 possesses 1D channels about 8 × 12 Å2 , which are large enough to encapsulate perylene molecules [185]. A series of 3D soft MOFs (FJI-H11-R, R = Me, Et, iPr) have been assembled from tetra-carboxylate tpcb ligands bearing dangling side groups (4,4′ -di(substituent) oxybiphenyl-3,3′ ,5,5′ -tetra-(phenyl-4-carboxylic acid)) (Figure 6.75a) and Cu(II) paddle-wheel SBUs. FJI-H11-R are isoreticular frameworks, which display the (4,4)-c NbO-type network with the topological point symbol of {64 .82 }. Such 3D hinged framework structure consists of 2D rhombic fences (Figure 6.75b,c). These

6.4 Three-dimensional MOFs O HO

R

O O

OH 71.2°

HO

O

O (a)

R O R= Me, Et, iPr

OH

(b)

Contraction Expansion

(c)

(d)

Figure 6.75 (a) The chemical structure of tetra-carboxylate tpcb ligands bearing dangling side groups. (b) The distorted NbO-type topology with the angle of 71.2∘ . (c) Perspective view of the framework. (d) The illustration shows the hinged-framework motif responding to external stimuli. Source: Pang et al. [186]. © 2016 John Wiley & Sons.

soft crystal materials show interesting dynamic features in response to thermal treatment and solvent molecules as a result of the rotation and bending of the organic linkers. Importantly, the successive intermediate states of both thermal expansion and desolvation can be visualized by in situ X-ray snapshot analyses (Figure 6.75d) [186]. A series of Zr-MOFs (PCN-605, PCN-606, PCN-608) are obtained from tpcb with different substitutes. The tetratopic ligands with various substituents adopt three different symmetries and further bridge the eight-connected Zr6 clusters to form PCN-605, PCN-606, or PCN-608 series with flu, scu, csq topologies [187, 188]. 6.4.1.4 Hexatopic or Octatopic Carboxylate Linkers

Hexatopic linkers usually contain 1,3-benzenedicarboxylate units, 4,4′ -azanediyldibenzoate units, etc. [Cu3 (bhb)] (UTSA-20; UTSA = University of Texas at San Antonio) with novel trinodal (3,3,4) net of zyg topology is formed by the self-assembly from a hexacarboxylate linker H6 bhb (3,3′ ,3′′ ,5,5′ ,5′′ -benzene-1,3,5triyl-hexabenzoic acid) and the paddle-wheel Cu2 (COO)4 SBU. The three isophthalate groups of each linker are noncoplanar because the four phenyl rings within bhb are tilted with respect to each other owing to the steric hindrance between the central and peripheral phenyl rings, which lead to the final zyg topology structure (Figure 6.76) [189]. A series of isoreticular (3,24)-connected mesoporous MOFs (PCN-61, PCN-66, PCN-69, and PCN-610) were assembled from dicopper paddle-wheel units and the corresponding hexatopic linkers. The isophthalate-sustained cuboctahedra in the (3,24)-connected network prohibits framework interpenetration, leading to MOFs with high surface areas [190–192].

351

352

6 The Structures of Metal–Organic Frameworks

+

Figure 6.76 The SBUs and the assembly of these shapes in UTSA-20. Source: Guo et al. [189]. © 2011 John Wiley & Sons.

The reaction of H6 dpot (5,5′ ,5′′ -((1,3,5-triazine-2,4,6-triyl)tris(oxy))tri-isophthalic acid) and Zn(NO3 )2 ⋅6H2 O gives 3D chiral MOF FJI-H2 O, formulated as {[Zn4 O(dpot)](H2 O)2 (DEF)4 }n . In the structure, Zn4 O(CO2 )6 SBUs and dpot6− linkers can both be considered as 6-c nodes, and the framework of FJI-H2 O can be simplified as the pcu topology. The framework of FJI-H2 O should also be described as the (3,6)-c network with point symbol of {62 .8}3 {63 }{69 .83 .103 } if the dpot6− nodes are split into four three-connected nodes (mcs topology, a pcu-derived net) [193]. A 3D porous network composed of single right-handed helix and unusual triple molecular necklace-like helix has been built from hexatopic carboxylate {6,6′ ,6′′ ,6′′′ -((5′ -(4-(bis(6-carboxynaphthalen-2-yl)amino)phenyl)-[1,1′ :3′ ,1′′ terphenyl]-4,4′′ -diyl)bis(azanetriyl))tetrakis(2-naphthoic acid)}. The 3D framework shows the 1D channel with the average diameter of ∼9.4 Å, which is a (3,4)-connect with the (63 )6 (6⋅82 )2 (62 ⋅8⋅102 ⋅12)3 topological symbol if each central benzene ring and each N atom of ligand are considered as three-connected nodes, and each [Cu2 (CO2 )4 ] cluster acts as a four-connected node. Whereas if each ligand is considered as a six-connected node and each binuclear [Cu2 (CO2 )4 ] cluster as a four-connected node, it displays a (4,6)-connected framework with the topological symbol of (43 ⋅63 )3 (49 ⋅66 )2 [194]. Zn-based MOF FJI-C8 based on π-conjugated aromatic ligand H6 tdpat (2,4,6-tris(3,5-dicarboxylphenylamino)-1,3,5-triazine) is a 3D cubic porous network, which can be regarded as 3-nodal 4,6,6-c topology network with the point symbol of {42 ⋅84 }3 {46 ⋅89 }2 {49 ⋅66 }4 by considering tdpat as 6-c nodes and Zn3 O clusters and mono-Zn cores as 6-c and 4-c nodes [195]. MOFs with octatopic carboxylate linkers are rare possibly because of the lack of linkers themselves. Hong and coworkers reported a stable and porous 3D Cu-MOF FJI-H8 based on Cu2 (COO)4 SBUs and octacarboxylates ligands (H8 tddb, 3,3′ ,5,5′ -tetra(3,5-dicarboxyphenyl)-4,4′ -dimethoxy-biphenyl) containing m-benzenedicarboxylate units. FJI-H8 adopts the rare (4,4,8)-connected URJ network when two types of Cu2 SBUs are simplified as two kinds of four-connected nodes and the tddb ligands as eight-connected nodes. There are three types of polyhedral nanocages in the framework: one regular cuboctahedron (Cage-A) with the pore diameter around 15 Å, one distorted octahedron (Cage-B) with a dimension of ∼8 Å, and one distorted cuboctahedron (Cage-C) with a dimension

6.4 Three-dimensional MOFs

COOH

COOH

O HOOC

COOH

HOOC

COOH O COOH

COOH

H8tddb (a)

(b)

Figure 6.77 The structure of ligand H8 tddb (a) and FJI-H8 (b). Source: Based on Pang et al. [196]. Licensed under CC BY 4.0.

of ∼12 Å, respectively. On the whole, Cage-A is connected by six Cage-C through six rhombic faces and eight Cage-B through eight triangular faces. Similarly, Cage-C is linked by six Cage-A through six rhombic faces and eight Cage-B by sharing eight m-benzenedicarboxylate moieties. However, Cage-B is linked by four Cage-A through four triangular faces and four Cage-C by sharing four m-benzenedicarboxylate moieties (Figure 6.77). Importantly, FJI-H8 exhibits a record-high acetylene uptake under ambient conditions at that time [196]. They later prepared another octatopic carboxylate–based 3D MOF, FJI-H19, which is prepared from 5′ ,5′′ -bis(3,5-dicarboxyphenyl)-2′ ,2′′ ,4′ ,6′′ -tetramethoxyacid 4′′ ,6′ -dimethyl-[1,1′ :3′ ,1′′ :3′′ ,1′′′ -quaterphenyl]-3,3′′′ ,5,5′′′ -tetracarboxylic (H8 btdta) and Cu(NO3 )2 ⋅2H2 O. FJI-H19 contains two kinds of different cages, i.e. the distorted cuboctahedron and the distorted octahedron. Much different from FJI-H8, it adopts a distinct (4,8)-connected scu topological structure with the point symbol of {416 ⋅612 }{44 ⋅62 }2 . FJI-H19 also exhibits good thermal and water stability and possesses permanent porosity with the BET surface area of 1928 m2 /g [197].

6.4.2

N-heterocyclic Linkers

Various organic linkers containing nitrogen donor such as azole and pyridine derivatives have been widely used to construct 3D MOFs via nitrogen–metal coordination. Azoles are a class of five-membered aromatic N heterocycles, much more diverse than pyridine ring with only one nitrogen atom, which includes imidazole (Him), pyrazole (Hpz), 1,2,4-triazole (Htz), 1,2,3-triazole (Hta), and tetrazole (Httz). 6.4.2.1 Imidazolates-based MOFs

Metal–organic zeolites are an important branch of MOFs that combine the advantages of zeolites and MOFs, such as high surface area and porosity as well as the exceptional stability of zeolites [198, 199]. Chen and coworkers reported a

353

354

6 The Structures of Metal–Organic Frameworks

great number of imidazolates-based MOFs. For example, a porous zinc(II) benzimidazolate (MAF-3) has a distorted zeolitic sodalite topology with small pores (c. 6.3 × 4.1 Å2 ) [200]. By changing the side group to a smaller one and using the mixed-ligand strategy, more porous metal–organic zeolites, SOD-[Zn-(mim)2 ] (MAF-4), ANA-[Zn(eim)2 ] (MAF-5), and RHO-[Zn-(eim/mim)2 ] (MAF-6), with regular zeolitic topologies are obtained [201]. MAF-6 possesses large surface area, pore volume, pore size, and aperture size and exhibits exceptional hydrophobicity on both the internal pore and external crystal surfaces, which are originated from its ethyl-lined pore surface and nanoscaled corrugation (Figure 6.78) [202]. Yaghi and coworkers developed the ZIFs employing the imidazole ligand. ZIF-1 to ZIF-12 have been synthesized from either Zn(II) (ZIF-1 to ZIF-4, ZIF-6 to ZIF-8, and ZIF-10 to ZIF-11) or Co(II) (ZIF-9 and ZIF-12) with imidazolate-type links. The structures are based on the nets of seven distinct aluminosilicate zeolites: tetrahedral Si(Al) and the bridging O are replaced with transition metal ion and imidazolate link, respectively. ZIF-5 is formed via the mixed-coordination imidazolate of Zn(II) and In(III) based on the garnet net. Among these ZIFs, the sod topology ZIF-8 (or called MAF-4, Zn(mIM)2 ) possesses large pores (11.6 Å in diameter) but small apertures with diameters of 3.2 Å [203]. ZIF-8 represents one of the most studied MOFs due to its exceptional stability, high porosity, facile/diversified preparation methods, and inexpensive/nontoxic components. Using bulky amides as the structure-directing agents, two ZIFs CAN-[Zn(Im)2 ] and AFI-[Zn(Im)2 ], mimicking the CAN and

Figure 6.78 Perspective view of the framework and pore surface structures of MAF-6. Source: He et al. [202]. © 2015 American Chemical Society.

6.4 Three-dimensional MOFs

AlPO-5 (AFI) zeotypes with 12-membered ring pore openings are obtained, respectively. AFI-[Zn(Im)2 ] has the largest pore apertures for ZIF materials at that time [204]. Zhang and coworkers reported the hybrid zeolitic imidazolate frameworks (HZIFs) with integrated structural features and functions of zeolites and ZIFs. The assembly of tetrahedral TO4 (MoO4 or WO4 ) and [M(im)3 O] units fabricates four-connected zeolite-type topologies with the general framework composition M4 (im)6 TO4 . These HZIFs have distinct framework topology, unusual high thermal stability, and catalytic properties [205]. 6.4.2.2 Pyrazolate/Triazolate/Tetrazolate-based MOFs

In contrast to imidazolates, the simple pyrazolate is difficult to form 3D frameworks; therefore, many bipyrazole or polypyrazolate ligands have been used to construct 3D MOFs. Xiang and coworkers reported the FJU-66 based on the triangular [Cu3 (μ-Pz)3 ] cluster (Pz, pyrazolate) and linear bipyrazole ligand 2,7-bis(3,5-dimethyl)dipyrazol-1,4,5,8-naphthalene-tetracarboxydiimide, showing a 3D pcu-type network with the topological point symbol of 412 .63 . Two such networks interlock each other, forming a twofold interpenetrating array with 1D channels of c. 5.6 Å in diameter. FJU-66 can maintain its structure up to 803 K and in aqueous solutions with pH values in the range of 3–14 [206]. In addition to the linear bipyrazole ligands, tetrapyrazole ligands can also be employed to compose 3D Pz-based MOFs. A Pz-based porphyrinic MOF PCN-601 is built up of [Ni8 (OH)4 (H2 O)2 Pz12 ] nodes and 5,10,15,20-tetra(1H-pyrazol-4-yl)porphyrin (H4 tpp) ligands. PCN-601 can retain its structure in saturated sodium hydroxide solution (∼20 mol/l) at room temperature and 100 ∘ C [207]. Another Pz-based porphyrinic MOF PCN-602 is based on 12-connected [Ni8 (OH)4 (H2 O)2 Pz12 ] cluster and pyrazolate-based porphyrin ligand, 5,10,15,20-tetrakis(4-(pyrazolate-4-yl)phenyl)porphyrin. PCN-602 can remain stable in aqueous solutions of HCl and NaOH with pH between 4 and 14 and shows good stability in 1 M KF, 1 M Na2 CO3 , and 1 M K3 PO4 aqueous solutions at room temperature [208]. PCN-601 and 602 adopt a ftw-α topology featuring large cubic cages within the frameworks. The diameters of the cages in PCN-601 and 602 are 1.5 and 2.1 nm, respectively. PCN-602 with elongated tetrapyrazole ligands has a larger window size (6.3 × 14.2 Å2 ) (Figure 6.79). The BET surface area and total pore volume of PCN-602 (2219 m2 /g and 1.36 cm3 /g, respectively) are larger than those of PCN-601 (1309 m2 /g and 0.78 cm3 /g). Lu and coworkers prepared infinite 3D MOFs {[Ag2 (trz)2 ][Ag24 (trz)18 ]}[PW12 O40 ]2 from1,2,4-triazole (trz). Its structure is composed of catenated polyhedral cages {Ag24 (trz)18 }6− built by 24 Ag+ cations, 12 μ3 1,2,4-triazole ligands, and 6 μ2 1,2,4-triazole ligands. Each adamantane-like cage is catenated by six other ones through its vertices to form a NaCl-type α-Po topological framework, by regarding each nanocage as a six-connected node and the catenation as linkage between nodes. Two independent NaCl-type α-Po polycatenated frameworks interpenetrate one other to form a highly ordered supramolecular aggregate. The window-to-window arrangement of the {Ag24 (trz)18 }6− nanocages between the interpenetrating frameworks generates nanosized pores loading with [PW12 O40 ]3− counteranions. The

355

356

6 The Structures of Metal–Organic Frameworks

8.0 2.1

TPPP4–

PCN-601

14.0

TPP4–

6.0

PCN-602

Figure 6.79 Structure of PCN-601 and PCN-602. Source: Lv et al. [208]. © 2017 American Chemical Society.

Keggin polyanions function not only as counteranions but also as templates to help direct the formation of the polycatenated framework [209]. A 3D cationic MOF [Ag2 (btr)2 ]⋅2ClO4 ⋅3H2 O (ABT⋅2ClO4 , btr = 4,4′ -bis(1,2,4triazole)) consisting of nanoscale cages has been prepared from a triazole linker. The octahedral cage is built up of 12 Ag(I) ions and 12 btr ligands, with the overall edge length about 2.0 nm and the inner diameter and calculated pore volume about 4.18 Å and 305.9 Å3 , respectively. Each octahedral cage has eight trigonal windows with the aperture of 4.0 Å. The tetrahedral cage is composed of eight Ag(I) ions and eight btr ligands and its edge length is about 1.0 nm. Each octahedral cage is linked to six tetrahedral cages through the sharing of two Ag(I) ions and two btr ligands, while every tetrahedral cage shares Ag(I) ions and btr ligands with four octahedral cages, resulting in a 3D cationic framework with 1D channels [210]. A flexible MOF with pairs of uncoordinated triazolate N-donors, [Zn2 (btm)2 ]⋅4H2 O (MAF-23), has been constructed by using a bis-triazolate ligand bis(5-methyl-1H-1,2,4-triazol-3-yl)methane (H2 btm) and tetrahedral Zn(II) ion. In MAF-23, each Zn(II) ion is coordinated by four N atoms from three btm2− ligands, and each btm2− ligand coordinates to three Zn2+ ions in a bisimidazolate mode, resulting in the 3D framework with 1D narrow channels (void volume 23.4%). Each triazolate uses only two N atoms for coordination, leaving the third N atom as a guest binding site, exhibiting strong CO2 adsorption and high CO2 /N2 selectivity [211].

6.4 Three-dimensional MOFs

The M3 [(M4 Cl)3 (BTT)8 ]2 (M-BTT; BTT3− = 1,3,5-benzenetristetrazolate) series are constructed from tetrazolate linker. M-BTT displays an expanded sodalite-type structure, in which truncated octahedral cages share square faces to compose a cubic (3,8)-net. The square faces correspond to an [M4 Cl]7+ unit, in which the axially bound solvent molecule at the metal centers can be removed to yield an exposed metal cation site. Cu-BTT can be fully desolvated, while 70% and 30% of metal centers retain a MeOH molecule within Mn- and Fe-BTT, respectively. Mn-BTT has the highest BET surface area of 2100 m2 /g compared with Fe-BTT (2010 m2 /g) and Cu-BTT (1710 m2 /g) [212]. Tetrazolate BTT linker has also been applied to construct porous 3D framework Zn(Hbtt) (CPF-6). In comparison with Mn-BTT (or Cu-BTT), three tetrazole groups in CPF-6 show two types of binding modes, one bidentate and two unidentate fashions, leading to a 3D network of CPF-6 with a (3,6)-rtl (rutile) topology, which has single-walled channels with a diagonal separation of about 1.5 nm [213]. 6.4.2.3 Pyridine and Other N-heterocyclic Based MOFs

A 3D Cu(I) MOF, {[Cu(pytpy)]⋅NO3 ⋅CH3 OH}∞ , has been synthesized from a pyridine-based linker 2,4,6-tris(4-pyridyl)pyridine (pytpy) and Cu(NO3 )2 ⋅3H2 O, the structure of which can be simplified as an unusual uninodal (10,3)-b (also denoted ThSi2 ) net by regarding the Cu(I) ion and pytpy as three-connected nodes. The pytpy ligands connect adjacent Cu(I) ions to generate Cu–pytpy helix chains. Such alternately arranged right- and left-handed hexagonal helices stack in the (10,3)-b frameworks by sharing a common edge. Four individual ThSi2 nets interpenetrated together because of the large void in the single net and the self-dual nature of this topology to form a rare fourfold interpenetrated 3D architecture with helical channels in dimensions of ∼5 × 6 Å2 (Figure 6.80). The guest methanol molecules in the channels can be replaced by water molecules via an SC-SC process [214]. Another 3D MOF, {[Cd(4-btapa)2 (NO3 )2 ]⋅6H2 O⋅2DMF}n , is assembled from a three-connector pyridine-based ligand with amide groups. In the structure, the Cd(II) center is octahedrally coordinated to six pyridyl nitrogen atoms from six different 1,3,5-benzene tricarboxylic acid tris[N-(4-pyridyl)amide] (4-btapa), which is very rare. Three octahedral Cd(II) moieties are linked together by three 4-btapa units to form a large six-membered ring, which expand in three directions to form a 3D framework. The two frameworks mutually interpenetrate to produce 3D running channels (so-called 3D pores) with dimensions of 4.7 × 7.3 Å2 . The channel is not straight but zigzags and the amide groups inside the pores act as important functional organic sites in the interaction between the host and guest molecules [215]. A 3D four-connected CdSO4 (cds) framework [Zn(hsb-2)(H2 O)2 ⋅G⋅2H2 O]n (G = naphthalene-2,7-disulfonate) has been prepared from a hydrogenated Schiff base hsb-2. All the Zn(II) ions in the MOF are six-coordinated by four nitrogen atoms of three different hsb-2 ligands and two oxygen atoms from two waters to afford the distorted octahedral geometry ZnN4 O2 . Ligand hsb-2 exhibits a “V” shape and functions as a twisted bridging linker. As a result, a 3D framework is generated, which is composed of alternated right-handed and left-handed helical chains. Topologically, the 3D network can be described as cds net with a point

357

358

6 The Structures of Metal–Organic Frameworks

c a

Figure 6.80 A space-filling diagram of the fourfold interpenetration in {[Cu(pytpy)]⋅NO3 ⋅CH3 OH}∞ and views of the single fourfold helical channel. Source: Chen et al. [214]. © 2013 Royal Society of Chemistry.

symbol {65 .8}. Naphthalene-2,7-disulfonate molecules are located in the channels through intermolecular hydrogen bonds between sulfonates and water molecules. Tuning the amount of DEF solvent in the preparation of the Zn-hsb MOF, another 2D layered network with disulfonate anions intercalated to the interlayer region or another 1D infinite linear chain can also be obtained, which highlights a subtle solvent-induced self-assembly in crystal engineering [216]. Cluster-organic framework [Cu4 I4 (dabco)2 ]n (COZ-1; dabco = 1,4-diazabicyclo[2.2.2]octane) from N-heterocyclic dabco and tetrahedral Cu4 I4 clusters features a zeolite MTN-type framework with two types of giant 64 512 and 512 cages. The larger 64 512 cage has an inner diameter of 2.6 nm and contained 28 Cu4 I4 clusters as nodes and 42 dabco ligands as linkers, with pore volume of 9.2 nm3 , while the smaller 512 cage contains 20 Cu4 I4 clusters and 30 dabco ligands and has an inner diameter of 2.0 nm. The two giant cages are adjacent to each other, sharing the pentagonal windows [217].

6.4.3

Metallo-linkers

3D MOFs with metallo-linkers have gained more and more attention due to their potential applications in heterogeneous catalysis as the secondary metal centers can function as catalytic active sites to promote a wide range of organic reactions. There are so many metallo-linkers such as metallo-porphyrins, metallo-BINOL, metallo-bipyridine, metallo-salen, and so on. Several strategies have been developed to construct 3D metallo-linkers-based MOFs: (i) introduce metal centers via postsynthetic modifications; (ii) self-assemble MOFs from preformed metallo-ligands; (iii) postsynthetic linker exchange.

6.4 Three-dimensional MOFs

6.4.3.1 Metallo-Linkers Based on Porphyrins

The porphyrin core allows pre-metalation, in situ metalation, or post-metalation by various metal ions forming a great number of metallo-porphyrin-based linkers, which have been used to construct diverse 3D MOFs [218]. Lin reported the first 3D metallo-porphyrin-MOF possessing large cavities stable after the removal of solvent molecules, which is built from tetra(4-pyridyl)porphyrin and Co(II) or Mn(II) metal ions [219]. Tcpp represents one of the most used metallo-porphyrin linker used in the MOFs construction. Metallo-porphyrin MOF [Co(TCPP)Co1.5 ] (PIZA-1) was prepared from cobalt chloride and free meso-tetra(4-carboxyphenyl)porphyrin, in which the porphyrins were in situ metallated. PIZA-1 is a stable and robust framework with a large void volume composed of hydrophilic channels accessible to guest molecules. The porphyrin metal centers functionalized the walls of the channels, thus PIZA-1 displays repeatable size-, shape-, and functional group-selective sorption property [220]. Zhou and coworkers reported a series of zirconium metallo-porphyrin. PCN-222 (Fe) from Fe-TCPP and Zr6 clusters nodes, which has a large 1D open channels (with a diameter up to 3.7 nm). PCN-222 (Fe) is extraordinarily stable and can remain the framework upon treatment with boiling water and even concentrated aqueous HCl solution for 24 hours [221]. Unlike 12-connected Zr6 clusters in the UiO series, four of the surrounding carboxylate groups connected to the Zr6 cluster are replaced by hydroxyl groups in PCN-222. Thus, the Oh symmetry of the Zr6 cluster is reduced to D4h , which generates more space for catalysis. By carefully tuning the reaction conditions, a series of novel porphyrinic MOFs (PCN-224) containing D3d symmetric Zr6 cluster were prepared, which have 3D channels and high BET surface area (up to 19 Å and 2600 m2 /g, respectively). PCN-224 can remain intact over a wide range of pH in aqueous solution [222]. They further engineered PCN-224 (Fe) by the insertion of ethyl, bromo, chloro, and fluoro substitutions into the β-position of TCPP ligand to tune the environment around catalytic cores [223]. Su and coworkers utilized iridium–porphyrin ligand Ir(TCPP)Cl to react with HfCl4 to produce a stable Ir(III)–porphyrin MOF, [(Hf6 (μ3 -O)8 (OH)2 (H2 O)10 )2 (Ir(TCPP)Cl)3 ]⋅solvents (Ir-PMOF-1(Hf)). The framework features a (4,6)-connected she net with the short vertex symbol of 4.4.4.4.8(26).8(26), possessing two types of open cavities (1.9 × 1.9 × 1.9 nm3 and 3.0 × 3.0 × 3.0 nm3 ) cross-linked through orthogonal channels (1.9 × 1.9 nm2 ) in three directions (Figure 6.81) [224]. Bu and coworkers have synthesized four cubic zirconium–porphyrin frameworks, CPM-99 (H2 , Zn, Co, Fe) by solvothermal reactions of tcbpp-X (tcbpp = tetrakis(4-carboxybiphenyl)porphyrin, X = H2 , Zn, Co, Fe), ZrOCl2 ⋅8H2 O, and benzoic acid in DEF, which feature a binodal cubic net composed of 12-connected Zr6 O4 (OH)4 cuboctahedra linked by tcbpp. The 3D frameworks contain large cubic cages with an edge length as large as 2.5 nm. Each Zr6 cluster resides on one vertex and each face of the cube is capped by one tcbpp. Such cavities are packed in a primitive cubic lattice. Apart from the cubic cage, the ftw-type structure has another kind of cage, slightly distorted octahedron with a cavity diameter of ∼1.1 nm, comprising two Zr6 clusters in the axial sites and four tcbpp linkers in the equatorial plane [225].

359

360

6 The Structures of Metal–Organic Frameworks

COOH

HOOC

HOOC

N Cl N Ir N

HfCl4

N

120 °C

Ir(TCPP)CI

COOH

c-pore

Ir-PMOF-1(Hf)

a-pore

b-pore

Figure 6.81 3D Ir-PMOF-1(Hf) based on six-connected Hf6 (Hf6 (μ3 -O)8 (OH)2 (H2 O)10 ) clusters and four-connected Ir(TCPP)Cl ligands, showing two types of cavities. Source: Wang et al. [224]. © 2017 Royal Society of Chemistry.

Ma and coworkers reported a 3D MOF MMPF-3 with fcu topology net based on metallo-porphyrin linker of Co-dcdbp (dcdbp = 5,15-bis(3,5-dicarboxyphenyl)10,20-bis(2,6-dibromophenyl)porphyrin). MMPF-3 contains three types of polyhedral cage: a cubohemioctahedron (window size of c. 5.9 Å and inner dimensions of about 7.3 Å); a truncated tetrahedron; a truncated octahedron (window size of about 9.2 Å and a 4000 Å3 cavity). The three types of polyhedral cage are interconnected, and MMPF-3 possesses a solvent-accessible volume of 60% [226]. A series of porous metallo-porphyrinic isostructural MOFs ZJU-18 to ZJU-20 have been self-assembled from metallo-porphyrin octacarboxylates 5,10,15,20-tetrakis(3,5-biscarboxylphenyl)porphyrin (M-H8 ocpp, M = MnIII Cl and NiII ) by Wu and coworkers. They are three-periodic, binodal, edge-transitive nets with Reticular Chemistry Structure Resource symbol tbo, having intercrossed pore windows of about 11.5 Å and pore cages about 21.3 Å in diameter [227].

6.4.3.2 Metallo-Linkers Based on Salen

Compared with metallo-porphyrins, it is more convenient to obtain chiral metallo-salens, just by choosing the enantiopure diamines such as 1,2cyclohexanediamine as the start materials. Chiral metallo-salens are well-known privilege catalysts. Immobilization of chiral metallo-salen on MOFs is a promising way to combine the catalytic properties of the salen complexes with porous MOF materials [228].

6.4 Three-dimensional MOFs

A family of isoreticular chiral MOFs with the primitive cubic network topology has been assembled from [Zn4 (μ4 -O)(O2 CR)6 ] SBUs and systematically elongated chiral Mn-Salen catalytic subunit. The open channels and pore sizes can be tuned by changing the spacers of dicarboxylate struts and controlling the framework catenation [229]. Cui and coworkers constructed two chiral porous MOFs by using dicarboxyl-functionalized chiral Ni(salen) and Co(salen) ligands. The Cd-Ni(salen) MOF is a 3D chiral nanoporous framework and crystallizes in the chiral space group C2 . All Ni(salen) units display an exo-pentadentate coordination mode including one bridging bidentate and one chelating-bridging tridentate carboxylate groups. Each Ni ion is coordinated in a distorted square-planar geometry with two nitrogen atoms and two oxygen atoms from the salen ligand. Each tetranuclear Cd4 cluster is linked by eight Ni(salen) ligands, and each Ni(salen) ligand is linked to three Cd(II) ions to give a chiral porous 3D framework with channel cross sections of ∼1.2 × 0.8 nm2 . MOF Cd–Co(salen) is isostructural to Cd–Ni(salen), the 3D network of which is built from Cd4 clusters and Co(salen) linkers with open channels of ∼1.2 × 0.8 nm2 [230]. They recently prepared an isostructural series of twofold interpenetrated multivariate (MTV) MOFs by controlling incorporation of one, two, or three different enantiopure metallo-salen-based linkers (Figure 6.82). Interpenetration of the frameworks brings catalytically active M(salen) units into a dense arrangement, allowing cooperative activation, leading to improved efficiency and enantioselectivity [231]. A rare eightfold interpenetrated 3D framework is based on a four-connecting Cu4 I4 cluster and a two-coordinating Ni-(salen) ligand ((R, R)-N,N ′ -bis(3-tert-butyl5-(4-pyridyl)salicylidene)-1,2-diphenylethylenediamine nickel(II)). In spite of interpenetration, the framework still has two types of 1D chiral hydrophobic

O

OH

O

OH

N O V O N O

N O Cu N O

O

OH

N O Cr CI N O

O

OH

O

OH

O

OH

O

OH

O

OH

O

OH

+

N O MnCI N O

O

(a)

OH

N O FeOAc N O

O

OH

N O Co OAc N O

O

8.2 Å

OH

(b)

Figure 6.82 (a) The structures of metal nodes and different metallo-salen linkers. (b) The structure of twofold interpenetrated multivariate MOFs. Source: Xia et al. [231]. © 2017 American Chemical Society.

361

362

6 The Structures of Metal–Organic Frameworks

channels with pore window sizes of 6.77 × 8.64 Å2 and 6.09 × 10.96 Å2 . All Ni(salen) moieties lie inside of these channels and the empty coordination sites of Ni2+ orient to the cavities. This Cu–Ni(salen) MOF is extremely stable under high temperature (>400 ∘ C), in water vapor (90% relative humidity), in acid/base aqueous solution (pH 0–14), and in saturated NaOH solution at 100 ∘ C, as well as in 30 wt% H2 O2 and 70 wt% tert-butyl hydroperoxide solution [232]. By incorporating such Ni(salen) and metallo-porphyrin, a porphyrin-salen based chiral MOF has been constructed. In the structure, each Cd(TCPP) (or H2 TCPP) acts as a four-connecting bridge to link four separate SBUs to propagate a 2D lamellar network, which are pillared by Ni(salen) at the axial positions, generating a 3D framework. In spite of twofold interpenetration, there still exist 1D channels with a cross section of 6.7 × 10.5 Å2 . This porphyrin–salen-based MOF is used for the asymmetric cyanosilylation of aldehydes. Metallo-porphyrin units are responsible for Lewis acid activation, and metallo-salen units are in charge of chiral induction [233]. 6.4.3.3 Metallo-Linkers Based on BINOL

Axially chiral binaphthol (BINOL) or biphenol is a good chelating ligand, which can bind different metals to form catalytic moieties. The incorporation of such complexes (metallo-linkers) into MOFs can produce chiral heterogeneous MOFs catalysts. Lin and Wu constructed a chiral MOF from a BINOL-derived ligand (R)-6,6′ -dichloro-2,2′ -dihydroxy-1,1′ -binaphthyl-4,4′ -bipyridine. The BINOL ligands link one of Cd(II) centers into a 2D square grid of dimensions 20.3 × 20.3 Å2 and link the other type of Cd(II) centers to form 1D zigzag chains. The 2D grids and 1D zigzag polymeric chains are joined to each other by the bridging nitrate groups to form a 3D framework, which are twofold interpenetrated. Despite the interpenetration, it still possesses large interconnected channels with dimensions of 4.9 × 13.1 Å2 and 13.5 × 13.5 Å2 . Treatment of the BINOL-MOF with excess Ti(OiPr)4 affords a catalytic active MOF containing the Ti-BINOL linkers [234]. They also constructed many isoreticular chiral MOFs with similar non-interpenetrating structures but different sizes of open channels, from different BINOL-derived tetracarboxylate linkers and copper paddle-wheel SBUs. After PSM, these MOFs graft the Ti(IV) complex exhibiting highly catalytic activities (Figure 6.83) [235]. BINOL-derived tetracarboxylate ((R)-2,2′ -dihydroxy-1,1′ -binaphthyl-4,4′ ,6,6′ -tetrabenzoate) has been used to prepare an interpenetrating chiral 3D Zn-MOF. In the structure, Zn(II) atoms are triply bridged by carboxylate groups to form {Zn2 (μ2 -CO2 )3 (μ1 -CO2 )} SBUs. The fourth carboxylate group of BINOL-ligand coordinates to the square–pyramidal Zn atom in a monodentate fashion. Consequently, each BINOL ligand, as well as each Zn2 unit, acts as a four-connecting linker that leads to a 3D unc network, which is twofold interpenetrated. After interpenetration, it still possesses large channels (c. 1.5 × 2.0 nm2 ). The dihydroxy groups in the framework can be transformed into an active Lewis acidic catalyst by PSM with Ti(OiPr)4 [236]. Cui and coworkers presented a homochiral BINOL-based MOF decorated with the lithium salts. The BINOL-derived tetracarboxylate ligand exhibits an exooctadentate coordination fashion, binding to four Zn4 O clusters via four bidentate carboxylate groups. Three pairs of 12 Zn4 O clusters are bridged by carboxylate groups merging at

6.4 Three-dimensional MOFs CO2H CO2H HO2C

CO2H CO2H

HO2C

HO2C

HO2C OR OR

OR OR

OR OR

OR OR

HO2C CO2H

HO2C

HO2C CO2H

HO2C CO2H

R = Et or H

CO2H

(a)

OiPr O Ti O OiPr

(b)

Figure 6.83 (a) The structures of the ligands used for isoreticular chiral MOFs. The carboxylic acid groups are used to form chiral MOFs, whereas the dihydroxy groups react with Ti(OiPr)4 via PSM to form asymmetric catalysts. (b) The structure of the chiral MOF/Ti-BINOL. Source: Ma et al. [235]. © 2010 Springer Nature.

two more Zn4 O cores to generate a D3 -symmetric cage. The cage can also be viewed as a triply stranded structure with two Zn4 O cores located at opposite positions surrounded by three metallo-macrocycles, each of which is built of 10 Zn4 O clusters bridged by isophthalate linkers. The cage has an irregular open cavity with a maximum inner width of ∼1.8 nm. The hexahedral aperture on each face has a diagonal distance of ∼1.6 × 1.6 nm2 . Sharing the hexahedral windows with neighboring cages leads to multidirectional zigzag channels in the framework. The cage cavities are periodically decorated with the dihydroxy groups of biphenyl backbones, which can be partially exchanged by Li ions, forming a highly efficient and recyclable heterogeneous catalyst [237]. 6.4.3.4 Other Metallo-linkers

Other metallo-linkers based on bipyridyl, BINAP, N-heterocyclic carbenes, catechol, etc. are also widely observed in many 3D MOFs. Usually, MOFs bearing these open chelate sites are primarily constructed, then the metals are introduced

363

364

6 The Structures of Metal–Organic Frameworks

)] 2

[Ir

N

F TH

(CH

N

SO

3 CN

)4 ][

(M = Ir(COD)(OMe)) M

DM

[Pd

BPY-UiO

Bpy-UiO-Ir

Me

)(O

D (CO

Bpy-UiO-Pd (M = Pd(DMSO)2[BF4]2)

BF

4 ]2

Figure 6.84 Postsynthetic metalation of bpy-UiO to form bpy-UiO-Ir and bpy-UiO-Pd. Source: Manna et al. [238]. © 2014 American Chemical Society. ′

via postsynthetic metalation. For example, 2,2 -bipyridyl-derived bpy-UiO reacted with [Ir(COD)(OMe)]2 (COD = 1,5-cyclooctadiene) and [Pd(CH3 CN)4 ][BF4 ]2 affording bpy-UiO-Ir and bpy-UiO-Pd respectively (Figure 6.84). Both the metalated MOFs showed reduced BET surface areas (365.0 and 457.5 m2 /g versus 2277 m2 /g) and pore sizes (5.6 and 6.7 Å versus 7.2 Å), due to the import of Ir/Pd sites in the pores [238]. A series of other robust and porous UiO topology MOFs with open bipyridine or phenanthroline sites, BPV-MOF, mBPV-MOF, and mPT-MOF, from bipyridyl-functionalized dicarboxyate linker, mixed bipyridylor phenanthryl-functionalized and unfunctionalized dicarboxylate linkers were metalated with [Ir(COD)(OMe)]2 to form BPV-MOF-Ir, mBPV-MOF-Ir and mPT-MOF-Ir, respectively, which functioned as single-site catalysts [239]. By incorporating BINAP into MOFs, novel single-site solid catalysts can be formed. Lin and coworkers employed a BINAP-MOF with UiO-66 topology based on the Zr6 O4 (OH)4 (O2 CR)12 cluster SBUs and a BINAP-derived dicarboxylate linker as a precursor for postsynthetic metalation. A series of Ru- or Rh-functionalized BINAP MOFs have been obtained [240, 241]. Wu and coworkers constructed two MOFs from a double azolium derivative. After the deprotonation of the imidazolium moieties to form NHCs via PSM, active palladium sites were anchored [242]. They then reported a metal–organic nanotube (MONT) based on the similar bent double azolium ligands. The MONT was composed of a large exterior wall (4.91 nm in diameter) and an interior channel (3.32 nm in diameter). A 3D chiral framework with 1D channels (2.0 nm in diameter) was formed by interlocking of the nanotubes. The imidazolium moieties in the pore walls can be modified to form NHCs and immobilized palladium atoms [243]. Cohen and coworkers decorated UiO-66 with catecholate or thiocatecholate by employing postsynthetic ligand exchange (PSE). Metalation of the catechol or thiocatecholate functionalized UiO-66, yielding Fe-monocatecholato, Cr-monocatecholato, and Pd-mono(thiocatecholato) species. [FeFe]-(dcbdt)(CO)6 (dcbdt = 1,4-dicarboxylbenzene-2,3-dithiolate) has also been incorporated into UiO-66 by the PSE method, while the direct solvothermal synthesis does not work due to the thermally unstable nature of [FeFe](bdt)(CO)6 moiety [244–246]. Ma and coworkers chose MIL-101-Cr–SO3 H as the platform to react with AlCl3 . The

6.4 Three-dimensional MOFs

resultant MIL-101-Cr–SO3 H⋅Al(III) (Lewis acid–Brønsted acid MOF) displayed a decrease in BET surface area from 1571 to 1449 m2 /g and pore volume from 0.686 to 0.605 cm3 /g, respectively [247].

6.4.4

Mixed N-/O-donors

Diverse N- or O-donor ligands have been designed and synthesized to compose several 3D MOFs; however, employing the ligand containing only one kind of N- or O-donor cannot realize the specific functionality of frameworks sometimes. Thus, it is necessary to design a ligand containing N- and O-donors simultaneously or mixed ligands with the same or different donors. 6.4.4.1 Ligands with Mixed N- and O-Donor Atoms

Hong and coworkers have designed and synthesized a Cu(II) MOF FJI-H14 from 2,5-di(1H-1,2,4-triazol-1-yl)terephthalic acid (H2 btta) ligand, which contains both N- and O-donor atoms. In FJI-H14, each Cu(II) ion adopts a square–pyramidal coordination geometry, surrounding by two imine N atoms from two different 1,2,4-triazole groups and two O atoms from two different carboxylate groups in the equatorial plane, together with one O atom of the water molecule in the vertex. Topologically, the Cu(II) ion is a planar four-connected node, further linked by four tetradentate btta2− ligands into a 3D network with Kagome-like USF topology with hexagonal 1D channels (Figure 6.85). The pore limiting diameter and the maximum pore diameter predicted by the program Poreblazer for the fully evacuated FJIH14 are 5.95 and 7.62 Å, respectively and the evacuated FJI-H14 has a theoretical porosity of 44.4% according to PLATON calculations and has a total concentration of active sites as high as 9.22 mol/l. Furthermore, FJI-H14 exhibits unusual acid and base stability and high volumetric uptake (171 cm3 /cm3 ) of CO2 under ambient conditions (298 K, 1 atm) [248]. Two Co-MOFs, [Co3 (2,4-pdc)2 (μ3 -OH)2 ]⋅9H2 O (CUK-1) and [Co2 (6-mna)2 ]⋅3H2 O (CUK-2), were prepared by the hydrothermal reaction of Co(II) cations with 2,4-pyridinedicarboxylate (2,4-pdc) and 6-mercapto-3-pyridinecarboxylate (6-mna) anions. In CUK-1, the cobalt–hydroxide chains act as undulating pillars around which the ligands are arranged. In CUK-2, Co(II) is linked by single-atom thiolate bridges to form cobalt–thiolate helical chains. Adjacent 6-mna ligands provide carboxylate bridges between different Co(II), forming an extended structure similar to that in CUK-1 [249]. A unique fourfold interpenetrated (10,3)-b 3D Cu-MOF, [Cu(inaip)]⋅2H2 O, has been constructed from 5-(isonicotinamido) isophthalate (inaip). In the framework, each Cu(II) adopts a distorted square-planar coordination geometry, while each inaip ligand employs its one pyridyl and two carboxylate groups in turn to connect three metal atoms. 1D Zigzag chain is firstly formed by the connections between the carboxylate groups and Cu(II) by neglecting pyridyl group, which are joined together by 5-isonicotinamido groups to form 3D framework. In order to stabilize the framework, one such 3D net combines with three other ones to generate a fourfold parallel interpenetrated structure. The 3D Cu-MOF displays reversible

365

366

6 The Structures of Metal–Organic Frameworks

HO

O

N N

N

N

N N

O

OH

(a)

(b)

(c)

(d)

Figure 6.85 (a) The structure of ligand H2 btta. (b) The coordination environment of the Cu(II) ions as four-connected nodes and BTTA also as a four-connected node. (c) Structural illustration of FJI-H14, which contains 1D nanoporous channels. (d) The framework of USF topology. Source: Liang et al. [248]. Licensed under CC BY 4.0.

dehydration and rehydration in an SC-SC process, and the dehydrated material can encapsulate CH3 OH molecules in an SC-SC fashion as well [250]. Solvothermal reactions of MnCl2 with pyridylbenzoic acids gave three 3D MOFs, MCF-34, MCF-43, and MCF-44. [Mn(34-pba)2 ] (MCF-34, 34-Hpba = 3-(pyridin-4-yl) benzoic acid) consists of 1D zigzag Mn-carboxylate chains. In the structure, the ligand 34-Hpba serves as a tripodal node to connect the octahedral Mn(II) to form a (3,6)-connected ant (anatase) topology, furnishing 1D ultramicroporous channels (void = 16.4%, Fmin/max = 3.9/4.4 Å) parallel to the Mn-carboxylate chains. MCF-34 displays exceptional guest-responsive thermal-expansion properties [251]. By using ligands 4-(pyridin-3-yl) benzoic acid (43-Hpba) and 4-(pyridin-4-yl)benzoic acid (44-Hpba), [Mn(43-pba)2 ] (MCF-43) and [Mn(44-pba)2 ]⋅2.5DMF (MCF-44) are obtained. MCF-43 exhibits a nonporous framework that can also be simplified as a (3,6)-connected ant topology, if the 43-pba ligands and Mn2+ ions are regarded as three- and six-connected nodes. MCF-44 can be also simplified as a (3,6)-connected; however, it possesses a rtl (rutile) topology by detailed analysis. These metal carboxylate frameworks possess 1D Mn-carboxylate chains, which can be interpreted as flexible rods of edge-sharing octahedra [252].

6.4 Three-dimensional MOFs

6.4.4.2 Mixed Carboxylate Linkers

Numerous 3D MOFs are constructed by mixed carboxylate linkers. Reaction of bdc with Zn(II) affords MOF-5, while btb affords MOF-177 under same synthetic conditions. Combining these two linkers leads to a new MOF, Zn4 O(bdc)(btb)4/3 (UMCM-1), which consists of Zn4 O clusters linked together by two bdc and four btb linkers arranged in an octahedral geometry. Two bdc linkers are adjacent, leaving the other four positions occupied by btb linkers, and these octahedra assemble into a structure containing both micropores and mesopores [253]. Two major factors are considered to be key roles to build such MOFs: the geometric factor (length ratio between two linkers) and the statistical factor (mole fraction of two linkers). Minor changes of these factors may lead to significantly different MOF structures (UMCM-1, 2, 3, and 4) or failure of coassembly [254]. MTV MOFs are constructed from links with different functional groups whose orientation, number, relative position, and ratio along the backbone (metal-oxide and phenylene units) can be controlled by virtue of the unchanged length of the link and its unaltered connectivity. Eighteen MTV MOF-5-type structures containing up to eight distinct functionalities in one phase have been synthesized from bdc and its derivatives –NH2 , –Br, –(Cl)2 , –NO2 , –(CH3 )2 , –C4 H4 , –(OC3 H5 )2 , and –(OC7 H7 )2 . The backbone (zinc oxide and phenylene units) of these structures is ordered, but the distribution of functional groups is disordered [255]. Three-ring-based zeolite-type MOFs, CPM-2 and CPM-3, have been prepared from bpdc and m-bdc (or its derivatives). The bent dicarboxylates induce the assembly of In3 three-rings, while the linear dicarboxylates provide the cross-links between three-rings to generate an overall 3D architecture (Figure 6.86). The synthesis of CPM-2 and CPM-3 provides a new pathway for the construction of zeolite-type MOFs using only carboxylate ligands, instead of the currently popular practices that rely on heterocyclic compounds [256]. 6.4.4.3 Mixed Metallo-ligand and Organic Ligand

Mixed-metal–organic framework (M′ MOF) Zn3 (bdc)3 [Cu(SalPyen)]⋅(G)x ′ (M MOF-1; G = guest molecules) has been built from mixed ligands of bdc and Cu-(SalPyen) (PyenH2 = 5-methyl-4-oxo-1,4-dihydro-pyridine-3-carbaldehyde). M′ MOF-1 is a 3D framework composed of the trinuclear Zn3 (COO)6 SBUs. These SBUs are bridged by bdc to form the 36 tessellated Zn3 (bdc)3 2D sheets that are further pillared by the Cu(SalPyen) to construct a 3D framework [257]. By using chiral metallo-ligand Cu(SalPyCy), enantiopure M′ MOF Zn3 (bdc)3 [Cu(SalPycy)]⋅(G)x (M′ MOF-2) has been obtained, which is isostructural to the nonchiral M′ MOF-1. Replacing bdc with cdc (cdc = 1,4-cyclohexanedicarboxylate), Zn3 (cdc)3 [Cu(SalPycy)]⋅(G)x (M′ MOF-3) has been formed. M′ MOF-2 and M′ MOF-3 are isostructural 3D frameworks, in which Zn3 (COO)6 SBUs are linked by bdc or cdc to form the 36 2D tessellated Zn3 (bdc)3 or Zn3 (cdc)3 sheets, which are further pillared by the Cu(SalPyCy) to hexagonal primitive networks (Schäfli symbol 36 418 53 6) (Figure 6.87). The use of chiral metallo-ligand leads to enantiopure M′ MOF-2 and M′ MOF-3 with two chiral pore cavities of about 6.4 Å in diameter. M′ MOF-2 and M′ MOF-3 have the pore accessible volume of 51.7%

367

368

6 The Structures of Metal–Organic Frameworks

3-Ring

4.6 Å

(a)

(b)

12-Ring

b a (c)

Figure 6.86 (a) Structure of the {In3 (aip)3 } three-ring in CPM-2-NH2 . (b) Coordination of the {In3 (aip)3 } three-ring to adjacent units. (c) 3D view of the NPO-zeolite framework of CPM-2-NH2 . Source: Zheng et al. [256]. © 2011 John Wiley & Sons.

and 48.1%, respectively [258]. Further changing the metallo-salen ligands, four similar isostructural M′ MOFs Cd3 (bdc)3 [Cu(SalPyMeCam)]⋅(G)x (M′ MOF-4), Zn3 (cdc)3 [Cu(SalPyMeCam)]⋅(G)x (M′ MOF-5), Cd3 (bdc)3 [Cu(SalPytBuCy)]⋅(G)x (M′ MOF-6), and Zn3 (cdc)3 [Cu(SalPytBuCy)]⋅(G)x (M′ MOF-7) are also constructed [259]. A chiral porous zeolite-like MOF was constructed by using mixed ligands of dipyridyl-functionalized chiral Ti(salan) (salan = reduced salen) and bpdc. The framework containing salan-bound Ti4 O6 clusters consists of both hydrophobic and amphiphilic mesocages. The hydrophobic or amphiphilic cage shares its quadrilateral faces with three neighboring amphiphilic/hydrophobic cages, generating a 3D porous structure with 1D square channels of 0.5 nm × 1.5 nm [260].

6.4 Three-dimensional MOFs

RR N

N Cu N

O

O

N

+ Zn(NO3)2.6H2O

O OH

OH O H

H

HO

O O

HO

(a) MʹMOF-2

MʹMOF-3

(b)

MʹMOF-3

Figure 6.87 (a) Illustration of the syntheses of M′ MOF-2 and M′ MOF-3. (b) The structure of M′ MOF-3. Source: Xiang et al. [258]. © 2011 Springer Nature.

6.4.4.4 Mixed N-donated Linkers and O-donated Linkers

Wu and coworkers prepared three chiral “pillar-layered” MOFs from hsb-2 ligand and the angular XO4 2− (X = S, Mo or Cr). In the structure, hsb-2 ligand adopts a V-shape and functions as a bridging linker, resulting in a 2D network with a rhombic window. Different layers are connected by XO4 2− coordinating to Zn(II) to form 3D pillar-layered frameworks. Spontaneous resolution occurred during the “pillar–layer” process, which can be induced to asymmetric crystallization in the presence of enantiopure camphorsulfonic acid. The combination of the “pillaring” strategy and chiral induction provides a novel strategy to prepare homochiral 3D MOFs from achiral precursors [85]. By using bdc as the pillar, another 3D “pillar-layered” MOF (HSB-W1) has been constructed. Topologically, the 3D network of HSB-W1 can be described as a six-connected cubic net with a point symbol {412 .63 }, possessing two types of rhombic channels with diagonal distances of about 16.8 × 12.5 Å2 and 15.5 × 14.1 Å2 and a 1D channel about 14.3 × 10.5 Å2 , respectively. The void spaces are occupied by abundant disordered solvent molecules and the accessible void is estimated to be 55.4%, without consideration of the solvent molecules. HSB-W1 services as a good host matrix to incorporate red-green-blue fluorescent dyes for tunable white light emission [261]. Utilizing the mixed-ligands strategy, they also prepared a novel fourfold interpenetrated 3D homochiral MOF with rare pair quadruple-stranded helices from bpee and NCG. Changing the carbamyl substituent of NCG with benzoyl group (NBzG), a non-interpenetrated 3D homochiral MOF composed of alternate right-handed and left-handed single helix has been obtained (Figure 6.88). When p-tolylsulfonyl substituent was used instead, a homochiral linear structure was formed from mixed-ligand bpee and N-p-tolylsulfonyl-L-glutamate (NTsG), with all individual NTsG being lined up orderly. The steric hindrance of N-substituent of L-glu has a tremendous impact on the construction of these diverse frameworks [262]. Similarly, a new stable 3D MOF {Cu3 (bbbca)2 (4,4′ -bpy)[NH(CH3 )2 ]⋅7DMA⋅12H2 O}n (YCZ-1) was fabricated by mixed ligands of 4,4′ -bpy and 3,6-bis(4-benzoicacid)-N-(4-benzoic acid) carbazole (H3 bbbca). YCZ-1 displays a doubly interpenetrated 3,5,6-connected

369

370

6 The Structures of Metal–Organic Frameworks

b

a c

b c

a

(a) a b

c b c

a

R R L (b)

Figure 6.88 (a) Fourfold-interpenetrated 3D framework of [Zn(bpee)(NCG)⋅1.5H2 O]n with pair quadruple-stranded helices. (b) Non-interpenetrated 3D structure of [Zn(bpee)(NBzG)]n with left-handed and right-handed helical chains arranged alternately. Source: Wen et al. [262]. © 2015 American Chemical Society.

3D framework with a new Schläfli symbol of {512 ⋅83 }⋅{52 ⋅8}4 ⋅{56 ⋅84 }2 . The crystalline material YCZ-1 has been utilized as stationary phases for high-performance liquid chromatography (HPLC) separations of small organic molecules [263]. Wu and coworkers reported another mixed-ligands MOF, ZnATZ-btb, assembled from O-donated linker btb and N-donated linker Hatz (5-amino-1-H-tetrazolate). ZnATZ-btb possesses a twofold interpenetrated 3D network. By considering Zn(II) centers as tetrahedral four-connected nodes, and the btb and atz ligands as triangular three-connecting nodes and a two-connecting spacer, respectively, ZnATZ-btb displays a (3,4)-connected dinodal net with the Schläfli symbol of (62 ⋅84 )3 (63 )2 [264]. The reaction of Hatz with Zn(OH)2 affords a 2D 66 pillar-layer array of [Zn(atz)2 ]n . In the network, each atz ligand is bonded to two Zn(II) ions in the μ2 -1,4 bridging mode to give a 2D planar sheet. Adjacent sheets are pillared by the atz ligand to form a wavy 2D bilayer. After undergoing exchange with different dicarboxylate

6.4 Three-dimensional MOFs

pillars, four pillar-layer non-interpenetrated and twofold interpenetrated 3D frameworks were obtained from [Zn(atz)2 ]n as a layer precursor under solvothermal conditions [265]. Hong and coworkers have combined btb with different N-donated ligand to prepare 3D MOFs FJI-1 and FJI-2. [Zn6 (btb)4 (4,4′ -bpy)3 ]solv (FJI-1) based on mixed ligands of btb and 4,4′ -bpy is the non-interpenetrated analogue of MOF-14. The Zn(II) paddle-wheel SBUs in FJI-1 can be regarded as four-connected nodes and btb ligands as three-connected nodes, so the whole framework has an augmented Pt3 O4 topology, which contains two different types of pores (Figure 6.89). The solvent accessible volume is 81.9%, much higher than 67% in MOF-14 [266]. [Zn9 (btb)4 (odabco)3 (μ3 -O)3 (μ2 -H2 O)6 ]⋅16DEF (FJI-2; odabco = N-oxide-1,4-diazabicyclo[2.2.2]-octane) is built from btb and odabco. FJI-2 consists of 2.4 nm double-walled octahedral cages and topologically features a binodal network. The calculated free volume of FJI-2 is ∼50.0% by PLATON [267]. Zhang and coworkers prepared two 3D homochiral MOFs, (H3 O)2 [Cd8 (S-btatp)6 (bpy)3 (H2 O)4 ] and (H3 O)2 [Cd8 (R-btatp)6 (bpy)3 (H2 O)4 ], from enantiopure tricarboxylate ligands (2S,2′ S,2′′ S)-2,2′ ,2′′ -(benzenetricarbonyltris (azanediyl))tripropanoic acid (S-H3 btatp) or R-H3 btatp and 4,4′ -bpy. In the framework, double helical chains based on enantiopure ligands and single helical structures braided by the achiral bpy and cadmium clusters are integrated. The chirality transfer and reproduction process results from the chirality of the molecular building blocks in the form of interweaving double helical chains and the single helices. The

30 Å

Figure 6.89 Chemistry.

14 Å

Crystal structure of FJI-1. Source: Han et al. [266]. © 2011 Royal Society of

371

372

6 The Structures of Metal–Organic Frameworks

structure is topologically represented as a (3,12)-connected net with a vertex symbol of (3⋅4⋅5)3 (36 ⋅46 ⋅518 ⋅630 ⋅76 ) by regarding tetranuclear Cd units and btatp ligands as 12- and three-connected nodes, respectively, while the bpy ligands as linkers [268]. Lu and coworkers reported a 3D Ni-MOF, [Ni2 (μ2 -OH)(bpdc)(tpt)2 ][NO3 ]⋅3DMA⋅ 4CH3 OH⋅6H2 O, from a dicarboxylate ligand of bpdc and a tri-pyridinyl-based ligand of 2,4,6-tri(4-pyridinyl)-1,3,5-triazine (tpt). In the structure, adjacent Ni(II) ions is held together by a bridging μ2 -OH and two carboxylate groups from independent bpdc2− ligands to form a binuclear [Ni2 (μ2 -OH)(COO)2 ] SBU. Six such SBUs are linked together by two bpdc2− ligands and eight tpt ligands to form an octahedral cage, which packed closely to construct the overall structure. By regarding the binuclear-Ni SBUs and tpt ligands as eight- and three-connecting nodes, respectively, the framework can be simplified into a rare (3,8)-connected 3D network with the Schläfli symbol of (3⋅52 )2 (34 ⋅42 ⋅56 ⋅611 ⋅74 ⋅8) [269]. There are some example of 3D MOFs built from pyridine ligands and fluorides. Reaction of 4,4′ -dipyridylacetylene (dpa) with CuSiF6 affords MOF [Cu(dpa)2 (SiF6 )]n (SIFSIX-2-Cu), which is a primitive-cubic net with square channels of pore dimensions 13.05 Å. SIFSIX-2-Cu-i is composed of doubly interpenetrated nets that are isostructural to SIFSIX-2-Cu. The independent nets are staggered with respect to one another, resulting in 5.15 Å pores [270]. When a shorter organic linker of 4,4′ -azopyridine (azpy, 9.0 Å) was used instead of dpa (9.6 Å), an isoreticular [Cu(azpy)2 (SiF6 )]n (SIFSIX-14-Cu-i/UTSA-200) with a smaller pore size of 3.4 Å has been constructed (Figure 6.90) [271]. Two another

N

9.0 Å N N

N

Azpy UTSA-200a

Figure 6.90 Structure description of UTSA-200a (desolvated UTSA-200). Source: Li et al. [271]. © 2017 John Wiley & Sons.

6.4 Three-dimensional MOFs

SIFSIX-type MOFs, CPM-131 and CPM-132, have been prepared from ZnSiF6 ⋅xH2 O and tetra(4-pyridyl)metalloporphyrin (TPyP-M, M = Fe and Zn, respectively). CPM-131 ([(TPyP-Fe)Zn(SiF6 )]n ) can be simplified into a 4,6-connected fsc network, built of heterometalloporphyrin (4,4)-grid sheets and SiF6 2− pillar, which possesses 3D interconnecting channels with a solvent accessible volume of 44.3%. CPM-132 ([(TPyP-Zn)2 F2 Zn2 (SiF6 )]n ) is simplified into a 5,6-connected fsx net, which is also view as F-intercalated TPyP-Zn bilayers further pillared by SiF6 2− giving rise to a 3D framework [272].

6.4.5

Phosphonate or Sulfonate Linkers

6.4.5.1 Phosphonate Linkers

In comparison with carboxylates and pyridyl-based ligands, MOFs based on phosphonates are less studied. Some reasons have hindered their development: they often precipitate rapidly to insoluble phases during the growth of single crystals; the coordination chemistry of phosphonates is much more complicated and less predictable owing to their more possible ligating modes and three possible states of protonation [273, 274]. Shimizu and coworkers use a tetrahedral phosphonate ligand, 1,3,5,7tetrakis(4-phosphonatophenyl)adamantane, to construct a doubly interpenetrated diamondoid MOF (Figure 6.91). Despite the interpenetration, a considerable open void space still remained, with pore entrance widths of 4.1 Å and ovalshaped pores with a minimum width of 8.8 Å and a maximum width of 16.5 Å. The ligand directs the formation of trimetallic copper clusters as tetrahedral nodes, which, along with the adamantyl ligand, enable the diamondoid structure [275]. The same group also prepared another 3D MOF, PCMOF-5, from phosphonate ligand of 1,2,4,5-tetrakisphosphonomethylbenzene. PCMOF-5 adopts a modified

PO3H2

H2O3P

PO3H2 b

H2O3P c (a)

(b)

Figure 6.91 (a) The structure of 1,3,5,7-tetrakis(4-phosphonatophenyl)adamantane. (b) A space-filling representation of the doubly interpenetrated diamondoid network. Source: Taylor et al. [275]. © 2007 John Wiley & Sons.

373

374

6 The Structures of Metal–Organic Frameworks

pillared-layered motif. In the network, the 1D La(III) phosphonate chains are connected into a 3D framework through three of the four phosphonate groups on the tetrakisphosphate ligand, leaving the fourth group, a diprotic phosphonic acid, uncoordinated and protruding into the hydrated channel. The hydrated channel is completely lined with phosphonic acid groups; three-coordinated water molecules do not line the channel and appear to be involved in structural hydrogen bonds with the 1D La(III) phosphonate columns. Free water molecules fill the channel, alternating between and bridging the phosphonic acid groups through hydrogen bonds. PCMOF-5 conducts protons above 10−3 S/cm at 60 ∘ C and 98% relative humidity [276]. Uranium phosphonates, an important class of actinide–organic coordination polymers, exhibit an exceptionally diverse structures [277]. Albrecht-Schmitt and coworkers reported a 3D MOF, UO2 Np(H2 O)2 [CH2 (PO3 )(PO3 H)]2 (UNpC1P2–1) built from UO6 tetragonal bipyramids, NpO8 distorted dodecahedra, and monoprotonated methylenediphosphonate anions [278]. They also prepared two 3D uranyl phenyldiphosphonate MOFs from 1,4-benzenebisphosphonic acid. [Ba(H2 O)3 ]{(UO2 )3 [C6 H4 (PO3 )2 ]2 (O)}⋅5(H2 O) (BaUbbp) contains UO7 units that are bridged by the phosphonate group to form a pillared layered 3D framework. In [Sr(H2 O)3 ]{(UO2 )2 [C6 H4 (PO3 ) 2 ](OH)2 (H2 O)}⋅3(H2 O) (SrUbbp), the uranyl cations are bridged through the phosphonate moiety into a layered 1D uranyl chain, which are subsequently cross-linked by the rigid phenyl spacers into pillared layered networks. The voids in BaUbbp and SrUbbp are filled with water molecule [279]. Sun and coworkers reported a 3D uranium phosphonates, Zn(H2 tib)(UO2 )2 (EDP)(HEDP)(H2 EDP)0.5 ⋅3H2 O (EDP-ZnU1, H4 EDP = ethane-1,2-diyldiphosphonic acid). EDP-ZnU1 comprises dimeric U2 O12 unit condensed by two UO7 pentagonal bipyramids, which are further connected by Zn-centered polyhedral and EDP ligands to generate a 3D framework [280]. By hydrothermal reaction of EDP and zinc uranyl acetate in the presence of organic template 1,10-phenanthroline, another 3D uranium phosphonates EDP-U4 has been obtained. In EDP-U4, the diphosphonate ligands adopt two types of coordination mode: one is in a cis-mode and chelates three uranyl cations; the other one is in a trans-mode and directly bridges two uranyl cations. The connection of UO7 polyhedra by cis- and trans-EDP ligands results into a 3D open-framework with nanosized channels [281]. Wu and coworkers have prepared 10 isomorphous 3D zinc diphosphonates from 1-hydroxyl-2-(3-pyridyl)ethylidene-1,1-diphosphonic acid. In their structures, [ZnO3 N], [ZnO4 ] and [PO3 C] tetrahedra are connected through corner sharing to form a 2D layer, which consists of 4 and 6 MRs, as well as ellipsoid-like 12 MRs. These layers are further pillared by phosphate ligand into 3D frameworks with 12 MR ellipsoid-like channels, filling with protonated amine guests [282]. Hydrothermal reactions of Cd(II) or Mn(II) with 4-pyridyl-CH2 N(CH2 COOH)(CH2 PO3 H2 ), another two isomorphous 3D transition metal phosphonates are afforded. In the frameworks, the phosphonate exhibits a hexadentate mode to combine five Cd(II) or Mn(II) ions through three phosphonate and one carboxylate oxygen atoms along with two nitrogen donors. Each heptanuclear Cd(II) or

6.4 Three-dimensional MOFs

(a)

(b)

(c)

Figure 6.92 (a) Ball-and-stick view of the Eu–O–P–C chain. (b) Polyhedral view of the layer. (c) Polyhedral view of the 3D framework. Source: Fu et al. [285]. © 2017 Royal Society of Chemistry.

Mn(II) cluster contacts with surrounding six clusters through 12 phosphonate anions to form the 3D open framework [283]. Seven lanthanide phosphonates have been prepared by the hydrothermal reactions of lanthanide oxide or lanthanide chloride with 3-ammonium-1-hydroxypropylidene-1,1-diphosphonic acid (+ H3 NCH2 CH2 C(OH)(PO3 H2 )(PO3 H− )). They display a similar 3D open framework with eight-membered ring channels, which are occupied by organic pendants of CH2 CH2 NH3 [284]. Recently, they presented another five 3D lanthanide phosphonates based on a α-hydroxyphosphonic acid with formula of 1,4-C6 H4 (CH(OH)(PO3 H2 ))2 . In these structures, the ligands exhibit a bidentate and a hexadentate mode to link two and four lanthanide atoms, respectively. Each lanthanide atom is connected to neighboring two by four O–P–O and four O–C–P–O groups, forming a chain, which are cross-linked by the hexadentate phosphonate anions into a layer. The layers are further interconnected by the bidentate phosphonate anions to form a 3D framework with two type channels (Figure 6.92). The 3D framework is composed of a large solvent accessible void of 542.5 Å3 that constitutes volume 37.8% per unit cell [285]. 6.4.5.2 Sulfonate Linkers

MOFs with sulfonate linkers are much less, as the coordination ability of sulfonate anions is relatively weak. Shimizu and coworkers reported a crystalline sulfonate MOF, Na3 (2,4,6-trihydroxy-1,3,5-benzenetrisulfonate) (β-PCMOF2). β-PCMOF2 crystallizes in a honeycomb structure with 1D pores lined with sulfonate groups. In β-PCMOF2, the hexagonal sheets of sulfonate molecules cross-link in the third dimension by Na ions to form channels with a diameter of 5.65–5.91 Å. β-PCMOF2 can be loaded with amphoteric heterocycles to exhibit proton conduction [286]. They prepared a series of isostructural mixed-metal networks from

375

376

6 The Structures of Metal–Organic Frameworks

4,4′ -disulfo-2,2′ -bipyridine-N,N ′ -dioxide (dsbpdo). Firstly, trivalent lanthanides (Sm, Eu, Gd, Tb, Dy) have been complexed to the dianionic ligand in a 3 : 1 ratio to form trianionic complex building blocks [Eu(dsbpdo)3 (H2 O)2 ]3− , which are then cross-linked into a 3D network solid by addition of BaCl2 . They demonstrate permanent microporosity but also photophysical properties [287]. Later, they employed the same ligand to prepare five lanthanide frameworks. The structures are constructed via the one-pot self-assembly of the anionic cubic metallo-ligand, [Ln(dsbpdo)4 ]5− , and subsequent cross-linking with sodium cations and chloride to form 3D architectures, which contain 2D pores (9.93 Å × 5.05 Å). Both coordinated and uncoordinated water molecules, up to ∼80%, can be reversibly removed, leaving a rigid and porous framework with a Dubinin–Radushkevich surface area of 426 m2 /g [288]. By combining a rigid sulfonic ligand 1,5-naphthalenedisulfonate (1,5-nds) with a rare earth (RE) ion, a series of isomorphic 3D frameworks were synthesized by Wu and coworkers. In the frameworks of RE3 (OH)7 (1,5-nds) (RE = Y, Gd, Er, Yb), RE ion is octa-coordinated to seven μ3 -(OH) groups and an oxygen atom of a 1,5-nds ligand in an YbO8 triangulated dodecahedron. Three of YbO8 polyhedra share the (OH)–(OH) edge and the (OH)–(OH)–(OH) triangle planes with each other, generating a trimetric SBU, which are joined through S atoms and also share the (OH)–(OH) edges with each other, resulting in a 1D chain. Adjacent chains arrange alternately in a zigzag-type pattern to construct a 2D plane, further connecting by the whole 1,5-nds ligands to the final 3D frameworks. These MOFs have high thermal stability even at 351 ∘ C [289]. Sulfonate–carboxylate ligands have been widely explored to construct 3D architectures. For example, a tetranuclear copper cluster-based MOF has been synthesized by Zhang and coworkers from 5-sulfoisophthalic acid ligand. In such structure, every CuII 4 core is surrounded by six ligands, generating a [Cu4 (OH)2 (CO2 )4 (SO3 )2 ] cluster. Regarding the cluster as a six-connected node and ligand as a three-connected node, the whole framework can be simplified as a (3,6)-connected 3D non-interpenetrating network possessing 1D irregular channels of diameter 7.0 Å [290]. Zang et al. use sulfonate–carboxylate linker disodium-2,2′ -disulfonate-4,4′ -oxydibenzoic acid (Na2 H2 dsoa) to react with Tb3+ ions, affording a 3D porous terbium–organic framework {[Tb4 (OH)4 (dsoa)2 (H2 O)8 ]⋅(H2 O)8 }n (Tb-dsoa), which features porous robust framework and 1D open hydrophilic channels decorated by uncoordinated sulfonate oxygen atoms and aqua ligands [291]. The dsoa4− possessing two aromatic carboxylates at the terminal and two sulfonates in the middle and having flexibility is apt to construct porous and hydrophilic networks. Another 3D MOF Cu-dsoa has also been assembled by Zang and coworkers. The framework of Cu-dsoa has tetrameric copper clusters with a [Cu4 (𝜇 3 -OH)2 ]6+ core, surrounding by six dsoa4− ligands with each dsoa4− bridging three [Cu4 (μ3 -OH)2 ]6+ units. Therefore, dsoa4− ligands and [Cu4 (μ3 -OH)2 ]6+ can be viewed as three- and six-connected nodes, respectively, and Cu-dsoa displays a binodal network with Schäfli {4^2.6^10.8^3} (Figure 6.93) [292].

6.4 Three-dimensional MOFs

Cu S O C

(a)

b a (b)

Figure 6.93 (a) A view of a tetrameric copper cluster. (b) The 3D framework of Cu-dsoa. Source: Dong et al. [292]. © 2013 Royal Society of Chemistry.

6.4.6

POM-Based MOFs

The introduction of POM clusters as SBUs to construct extended polyoxometalate– organic frameworks (POMOFs) has attracted great interest, because such materials not only integrate the advantages of both POMs and MOFs, but also exhibit good prospects in photo-/electrocatalysis, molecular recognition, and so on [293, 294]. Dolbecq and coworkers carried out a study on the design and simulation of a family of zeolitic MOFs (named Z-POMOFs) based on ε-type Keggin POM SBUs and bdc ligand. Among these POMOFs, the cristobalite-like structure was predicted to be the most stable structure. Their prediction was validated by the synthesis of the first experimental {ε-Zn4 PMo12 }-based POMOF (NBu4 )3 [PMo12 O36 (OH)4 Zn4 (bdc)2 ]⋅2H2 O (Z-POMOF1). Z-POMOF1 shows a threefold interpenetrated dia-type topology and its crystallinity is maintained up to 180 ∘ C. Tetrabutylammonium cations play the role of counterions and space-filling agents in this 3D interpenetrated framework. By the incorporation of the POM within a framework, Z-POMOF1 exhibits remarkable electroactivity in the reduction of bromate [295]. Lan’s group prepared two porous POMOFs NENU-500 and NENU-501, employing two tritopic carboxylate ligands, btb and 1,1′ -biphenyl-3,4′ ,5-tricarboxylate, to link the {ε-Zn4 PMo12 } SBUs. In NENU-500, each Zn-ε-Keggin fragment is coordinated by four btb3− linkers, and each btb3− connects three Zn-ε-Keggin units, resulting a 3D porous framework with a ctn topology. Such single network penetrates with another identical one, to form a twofold interpenetrated ctn array. In NENU-501, the aggregation of two monomeric Zn-ε-Keggin units generates the dimeric Zn-ε-Keggin fragment. Each 1,1′ -biphenyl-3,4′ ,5-tricarboxylate ligand bridges three dimeric fragments to yield a 3D (3,6)-connected network with flu-3,6-C2/c topology [296]. Later, they prepared another two isostructural POMOFs with diamond topology, NENU-506 and NENU-507, comprising ε-Keggin polymolybdate units capped by four Zn ions linked through the linear rigid bifunctional ligands of isonicotinic acid and 4-(pyridin-4-yl) benzoic acid, respectively.

377

378

6 The Structures of Metal–Organic Frameworks

In NENU-506, the single dia network is penetrated by another identical one to generate a twofold interpenetrated diamond array, having two infinite helical chains from two individual diamondoid nets entangled each other. NENU-507 displays a fourfold interpenetrated diamond networks, by using the longer 4-(pyridin-4-yl) benzoic acid ligand instead of isonicotinic acid ligand. There are four helical chains form four individual dia nets, which can be considered as two pairs of helices of opposite helicity. NENU-506 and NENU-507 show excellent thermal and chemical stability, and NENU-507 used as an anode material in lithium-ion battery exhibits high reversible capacity of 640 mA/h g after 100 cycles [297]. Xu and coworkers reported two 3D chiral POMOFs based on Zn-ε-Keggin unit and achiral ligands bpp. In their frameworks, the alternate connection of Zn-ε-Keggin cluster and bpp ligands generates helical infinite chains, while each single spiral chain is further interlinked to adjacent neighboring units to construct a 3D chiral architecture with a qtz topology. The two POMOFs are stable in both acid and base aqueous solutions [298]. Recently, Lan and coworkers synthesized a series of stable PMOFs, [PMoV 8 MoVI 4 O35 (OH)5 Zn4 ]2 [M-TCPP][2H2 O][1.5NBu4 OH] (M = Fe, Co, Ni, and Zn), by applying Zn-ε-Keggin cluster and metallo-porphyrin as building block and linker, respectively. In their structures, Zn-ε-Keggin and M-TCPP display tetrahedral and quadrilateral connection modes, respectively. One Zn-ε-Keggin cluster with four Zn(II) centers connects with adjacent Zn-ε-Keggin clusters with two Zn—O bonds to generate cluster chain, which are further linked by M-TCPP ligands to produce 3D networks with twofold interpenetrated mog topology (Figure 6.94) [299].

Co-TCPP

Fe-TCPP

Zn-ɛ-keggin

Ni-TCPP Chain

Zn-TCPP

Figure 6.94 Schematic illustration of the structures of M-PMOFs (M = Co, Fe, Ni, Zn). Source: Wang et al. [299]. Licensed under CC BY 4.0.

6.4 Three-dimensional MOFs

379

Yang and Wang reported two Anderson-type POM-based POMOFs via a step-by-step synthetic strategy. Firstly, Anderson-type POM (NBu)3 [MnMo6 O18 (tri)2 ] was constructed by the reaction of a rigid bifunctional organic ligand H3 tri (2-(hydroxymethyl)-2-(pyridin-4-yl)-1,3-propanediol) and (NBu)4 [Mo8 O26 ], which can be viewed as two deprotonated tri3− ligands capped on the both sides of a well-known B-type Anderson-type cluster [MnMo6 O18 (OH)6 ]3− . Then, [MnMo6 O18 (tri)2 ]3− assembled with different cuprous halide clusters produced two novel Anderson-based POMOFs, (NBu)6 [Cu4 I4 ][MnMo6 O18 (tri)2 ]2 ⋅2DMA and (NBu)6 [Cu2 I2 ][MnMo6 O18 (tri)2 ]2 ⋅4DEF. They have solvent-accessible volumes of 69.4% and 71.4% and extra-large intriguing channels with size of about 2.78 × 2.92 nm2 and 2.45 × 2.30 nm2 , respectively (Figure 6.95) [300]. Zheng and coworkers extended the reaction system by combining the Anderson-type POM SBUs with different transition metal–oxygen clusters. They prepared two 3D Anderson-based POMOFs, (CH3 NH3 )3 [Cu4 (μ3 -O)2 (peb)4 ][MnMo6 O18 (tri)2 ]⋅4NMF⋅ 5H2 O and [(CH3 )2 NH2 ]3 [Co5 (μ3 -O)2 (peb)6 ][MnMo6 O18 (tri)2 ]⋅5DMF (peb = 4-(2-(4-pyridyl)ethenyl)benzoic acid; NMF = N-methylformamide), via the reactions of [MnMo6 O18 (tri)2 ]3− with different transition-metal salts in the presence of ligand peb [301]. N

OH OH OH

2.78 nm

(a)

2.30 nm

2.92 nm

2.45 nm

c

b c (b)

b (c)

Figure 6.95 (a) The synthetic route of Anderson-type POM cluster [MnMo6 O18 (tri)2 ]3− . (b, c) POMOFs based on Anderson-type POM clusters and metal–halide clusters. Source: Li et al. [300]. © 2016 John Wiley & Sons.

380

6 The Structures of Metal–Organic Frameworks

Recently, Yang and Wang reported another two 3D POMOFs based on Lindqvist-type POM SBUs. They have been prepared via step-by-step synthetic strategy from Lindqvist-type polyoxovanadate clusters. POMOF (NBu)3 Cu[V6 O13 (tri)2 ]2 ⋅ 4DEF displays a 3D sixfold interpenetrated framework with diamond topology, while (NBu)Ag[V6 O13 (tri)2 ] solvent exhibits a non-interpenetrated 3D framework structure with nanoscale 1D hexagonal channels [302].

6.5 Conclusion MOFs are well-known organic–inorganic hybrids, which can be of different dimensionalities: 1D, 2D, and 3D. According to the coordination mode of the metal ions/clusters and geometry of the organic linkers, the structures of MOFs extend to different dimensions and adopt diverse topologies. There are so many available metal ions/clusters and organic ligands including carboxylate, N-heterocyclic, metallo-linkers, phosphonate or sulfonate, or their combination, etc. so countless MOFs have been reported and more and more new structures are being produced. Despite the different choice of the metal ions/clusters and ligands, the other factors such as solvents, templates, counterions, temperature, pH, concentration, synthetic method, etc. also have great influence on the structural prediction, which increase the number of MOFs. Though significant progress has been achieved in the construction of MOF architectures, there are still many challenges, for example, how to control the conformation of ligands, adopting a relative orientation suitable for the coordinated to sphere of metal centers; the mechanisms of solvent effect, template effect, etc. are poorly understood. Therefore, there is a long way to establish a successful approach to predict and control the MOF structures and fully understand the MOFs construction. In the future, research on MOFs will still be actively pursued; with the help of advanced techniques and artificial intelligence, further developments are to be expected.

Acknowledgments The authors are grateful for the financial support of the One Thousand Young Talents Program under the Recruitment Program of Global Experts, the National Natural Science Foundation of China (NSFC) (21771179 and 21233009), the Strategic Priority Research Program of CAS (XDB20010200), and the Fujian Province (2019J01126).

References 1 Cook, T.R., Zheng, Y.-R., and Stang, P.J. (2013). Chem. Rev. 113: 734–777. 2 Furukawa, H., Cordova, K.E., O’Keeffe, M., and Yaghi, O.M. (2013). Science 341: 1230444.

References

3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39

Li, J.R., Sculley, J., and Zhou, H.-C. (2012). Chem. Rev. 112: 869–932. Wu, H., Gong, Q., Olson, D.H., and Li, J. (2012). Chem. Rev. 112: 836–868. Sumida, K., Rogow, D.L., Mason, J.A. et al. (2012). Chem. Rev. 112: 724–781. Zhao, X., Wang, Y., Li, D.-S. et al. (2018). Adv. Mater. 30: 1705189. Cui, Y., Yue, Y., Qian, G., and Chen, B. (2012). Chem. Rev. 112: 1126–1162. Wang, C., Zhang, T., and Lin, W. (2012). Chem. Rev. 112: 1084–1104. Liu, J., Chen, L., Cui, H. et al. (2014). Chem. Soc. Rev. 43: 6011–6061. Chang, Z., Yang, D.-H., Xu, J. et al. (2015). Adv. Mater. 27: 5432–5441. Huang, Y.-B., Liang, J., Wang, X.-S., and Cao, R. (2017). Chem. Soc. Rev. 46: 126–157. Zhao, M., Ou, S., and Wu, C.-D. (2014). Acc. Chem. Res. 47: 1199–1207. Zhu, Q.-L. and Xu, Q. (2014). Chem. Soc. Rev. 43: 5468–5512. Liang, J., Huang, Y.-B., and Cao, R. (2019). Coord. Chem. Rev. 378: 32–65. Wen, Y.H., Zhang, J., Xu, Q. et al. (2018). Coord. Chem. Rev. 376: 248–276. Leong, W.L. and Vittal, J.J. (2011). Chem. Rev. 111: 688–764. Zaman, Md.B., Smith, M.D., and zur Loye, H.-C. (2001). Chem. Mater. 13: 3534–3541. Sun, C.-Y., Zheng, X.-J., Gao, S. et al. (2005). Eur. J. Inorg. Chem. 2005: 4150–4159. Xiong, K., Wu, M., Zhang, Q. et al. (2009). Chem. Commun.: 1840–1842. Sun, J.-W., Yan, P.-F., An, G.-H. et al. (2016). Dalton Trans. 45: 1657–1667. Li, C.-P., Zhou, H., Wang, S. et al. (2017). Chem. Commun. 53: 9206–9209. Deng, J.-H., Zhong, D.-C., Luo, X.-Z. et al. (2012). Cryst. Growth Des. 12: 4861–4869. Hou, Y.-L., Sun, R.W.-Y., Zhou, X.-P. et al. (2014). Chem. Commun. 50: 2295–2297. Yu, F., Li, D.-D., Cheng, L. et al. (2015). Inorg. Chem. 54: 1655–1660. Dong, Y., Zhu, N., Zhao, C. et al. (2016). RSC Adv. 6: 108645–108653. Zaman, M.B., Smith, M.D., Ciurtin, D.M., and zur Loye, H.-C. (2002). Inorg. Chem. 41: 4895–4903. Li, Y.-H., Su, C.-Y., Goforth, A.M. et al. (2003). Chem. Commun.: 1630–1631. Han, L. and Zhou, Y. (2008). Inorg. Chem. Commun. 11: 385–387. Cheng, A.-L., Liu, N., Yue, Y.-F. et al. (2007). Chem. Commun.: 407–409. Han, L. and Hong, M.-C. (2005). Inorg. Chem. Commun. 8: 406–419. Zheng, X.-D. and Lu, T.-B. (2010). CrystEngComm 12: 324–336. Chen, K., Hu, Y.-F., Xiao, X. et al. (2012). RSC Adv. 2: 3217–3220. Chen, K., Liang, L.-L., Liu, H.-J. et al. (2012). CrystEngComm 14: 8049–8056. Ding, S., Gao, Y., Ji, Y. et al. (2013). CrystEngComm 15: 5598–5601. Sun, D., Hao, H.-J., Liu, F.-J. et al. (2012). CrystEngComm 14: 480–487. Plasseraud, L., Maid, H., Hample, F., and Saalfrank, R.W. (2001). Chem. Eur. J. 7: 4007–4011. Qi, Y., Luo, F., Batten, S.R. et al. (2008). Cryst. Growth Des. 8: 2806–2813. Zheng, X.-D., Jiang, L., Feng, X.-L., and Lu, T.-B. (2008). Inorg. Chem. 47: 10858–10865. Dong, Y.-B., Sun, T., Ma, J.-P. et al. (2006). Inorg. Chem. 45: 10613–10628.

381

382

6 The Structures of Metal–Organic Frameworks

40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73

Han, L., Valle, H., and Bu, X.H. (2007). Inorg. Chem. 46: 1511–1513. Zhang, X.Y., Li, B., and Zhang, J.P. (2016). Inorg. Chem. 55: 3378–3383. Wen, Y.H., Sheng, T.L., Zhu, Q.L. et al. (2012). CrystEngComm 14: 2879–2885. An, S.-W., Mei, L., Wang, C.-Z. et al. (2015). Chem. Commun. 51: 8978–8981. Niu, C.-Y., Wu, B.-L., Zheng, X.-F. et al. (2007). Dalton Trans. (48): 5710–5713. Cui, Y., Lee, S.J., and Lin, W.B. (2003). J. Am. Chem. Soc. 125: 6014–6015. Wang, X.-L., Qin, C., Wang, E.-B. et al. (2004). Angew. Chem. Int. Ed. 43: 5036–5040. Luo, J., Hong, M., Wang, R. et al. (2003). Eur. J. Inorg. Chem. (19): 3623–3632. Zhang, Y.-Z., Wei, H.-Y., Pan, F. et al. (2005). Angew. Chem. Int. Ed. 44: 5841–5846. Wu, C.D. and Lin, W.B. (2005). Angew. Chem. Int. Ed. 44: 1958–1961. Wu, C.-D. and Lin, W.B. (2005). Inorg. Chem. 44: 1178–1180. Niu, D., Yang, J., Guo, J. et al. (2012). Cryst. Growth Des. 12: 2397–2410. Liu, Z.-Q., Zhao, Y., Liu, X.-H. et al. (2019). Polyhedron 167: 33–38. Hong, M., Zhao, Y., Su, W. et al. (2000). Angew. Chem. Int. Ed. 39: 2468–2469. Sokolov, A.N. and MacGillivray, L.R. (2006). Cryst. Growth Des. 6: 2615–2624. Barnett, S.A. and Champness, N.R. (2003). Coord. Chem. Rev. 246: 145–168. Khlobystov, A.N., Blake, A.J., Champness, N.R. et al. (2001). Coord. Chem. Rev. 222: 155–192. Sui, B., Zhao, W., Ma, G.H. et al. (2004). J. Mater. Chem. 14: 1631–1639. Zhang, G.Q., Yang, G.Q., and Ma, J.S. (2006). Cryst. Growth Des. 6: 1897–1902. Wang, Y., Ding, B., Cheng, P. et al. (2007). Inorg. Chem. 46: 2002–2010. Du, D.Y., Qin, J.S., Li, S.L. et al. (2010). Aust. J. Chem. 63: 1389–1395. Sha, J.Q., Peng, J., Zhang, Y. et al. (2009). Cryst. Growth Des. 9: 1708–1715. Wang, R.H., Hong, M.C., Yuan, D.Q. et al. (2004). Eur. J. Inorg. Chem. 2004: 37–43. Li, X.-J., Cao, R., Bi, W.-H. et al. (2005). Polyhedron 24: 2955–2962. Li, G.-B., He, J.-R., Pan, M. et al. (2012). Dalton Trans. 41: 4626–4633. Su, C.-Y., Goforth, A.M., Smith, M.D., and zur Loye, H.-C. (2004). Chem. Commun.: 2158–2159. Wu, C.-D., Ma, L.Q., and Lin, W.B. (2008). Inorg. Chem. 47: 11446–11448. Williams, C.A., Blake, A.J., Hubberstey, P., and Schröder, M. (2005). Chem. Commun. (43): 5435–5437. Wang, Y.H., Bredenkötter, B., Rieger, B., and Volkmer, D. (2007). Dalton Trans.: 689–696. Gable, R.W., Hoskins, B.F., and Robson, R. (1990). J. Chem. Soc., Chem. Commun. (23): 1677–1678. Ohmori, O. and Fujita, M. (2004). Chem. Commun. (14): 1586–1587. Biradha, K., Hongo, Y., and Fujita, M. (2000). Angew. Chem. Int. Ed. 39: 3843–3845. Fujita, M., Kwon, Y.J., Washizu, S., and Ogura, K. (1994). J. Am. Chem. Soc. 116: 1151–1152. Jiang, Y.-M., Yin, Z., He, K.-H. et al. (2011). Inorg. Chem. 50: 2329–2333.

References

74 Chen, C.-L., Goforth, A.M., Smith, M.D. et al. (2005). Inorg. Chem. 44: 8762–8769. 75 Wu, X.-R., Yang, X., Wei, R.-J. et al. (2014). Cryst. Growth Des. 14: 4891–4894. 76 Wang, Q., Zhang, J., Zhuang, C.-F. et al. (2009). Inorg. Chem. 48: 287–295. 77 Zhuang, C.-F., Zhang, J., Wang, Q. et al. (2009). Chem. Eur. J. 15: 7578–7585. 78 Pschirer, N.G., Ciurtin, D.M., Smith, M.D. et al. (2002). Angew. Chem. Int. Ed. 41: 583–585. 79 Mukhopadhyay, A., Maka, V.K., Savitha, G., and Moorthy, J.N. (2018). Chem 4: 1059–1079. 80 Xu, Z.-X., Liu, Y., and Zhang, J. (2016). Inorg. Chem. Commun. 67: 44–46. 81 Ohmura, T., Usuki, A., Fukumori, K. et al. (2006). Inorg. Chem. 45: 7988–7990. 82 Fei, B.L., Sun, W.Y., Yu, K.B., and Tang, W.X. (2000). J. Chem. Soc., Dalton Trans.: 805–811. 83 Sun, W.-Y., Fei, B.-L., Okamura, T.-a. et al. (2001). Eur. J. Inorg. Chem. 2001: 1855–1861. 84 Wen, Y.H., Sheng, T.L., Hu, S.M. et al. (2013). CrystEngComm 15: 2714–2721. 85 Wen, Y.H., Sheng, T.L., Sun, Z.H. et al. (2014). Chem. Commun. 50: 8320–8323. 86 Zhong, R.-Q., Zou, R.-Q., and Xu, Q. (2011). CrystEngComm 13: 577–584. 87 Fu, R.B., Hu, S.M., and Wu, X.T. (2015). Cryst. Growth Des. 15: 3004–3014. 88 Choi, E.-Y., Wray, C.A., Hu, C.H., and Choe, W.Y. (2009). CrystEngComm 11: 553–555. 89 Tsai, C.-C., Luo, T.-T., Yin, J.-F. et al. (2009). Inorg. Chem. 48: 8650–8652. 90 Li, J., Sheng, T.L., Bai, S.Y. et al. (2014). CrystEngComm 16: 2188–2195. 91 Huang, Y.H., Zhu, Q.L., Sheng, T.L. et al. (2013). CrystEngComm 15: 3560–3567. 92 Zhang, J.-Z., Cao, W.-R., Pan, J.-X., and Chen, Q.-W. (2007). Inorg. Chem. Commun. 10: 1360–1364. 93 Wang, M.-S., Guo, S.-P., Li, Y. et al. (2009). J. Am. Chem. Soc. 131: 13572–13573. 94 Lin, W.B., Evans, O.R., Xiong, R.-G., and Wang, Z.Y. (1998). J. Am. Chem. Soc. 120: 13272–13273. 95 Han, L., Bu, X.H., Zhang, Q.C., and Feng, P.Y. (2006). Inorg. Chem. 45: 5736–5738. 96 Wang, S.-H., Zheng, F.-K., Zhang, M.-J. et al. (2013). Inorg. Chem. 52: 10096–10104. 97 Huang, T., Wang, Y.-L., Yin, Q. et al. (2016). CrystEngComm 18: 2742–2747. 98 Zhang, C.L., Zhang, M.D., Qin, L., and Zheng, H.G. (2014). Cryst. Growth Des. 14: 491–499. 99 Wen, Y.H., Sheng, T.L., Xue, Z.Z. et al. (2014). Cryst. Growth Des. 14: 6230–6238. 100 Guo, J.-S., Xu, G., Guo, G.-C., and Huang, J.-S. (2013). Cryst. Growth Des. 13: 2642–2649. 101 Bai, S.Y., Sheng, T.L., Tan, C.H. et al. (2013). J. Mater. Chem. A 1: 2970–2973. 102 Jin, X.-H., Sun, J.-K., Cai, L.-X., and Zhang, J. (2011). Chem. Commun. 47: 2667–2669. 103 Zhao, X.L., He, H.Y., Dai, F.N. et al. (2010). Inorg. Chem. 49: 8650–8652. 104 Liang, J., Wu, X.-S., Wang, X.-L. et al. (2016). CrystEngComm 18: 2327–2336.

383

384

6 The Structures of Metal–Organic Frameworks

105 Guo, X.-G., Yang, W.-B., Wu, X.-Y. et al. (2013). Dalton Trans. 42: 15106–15112. 106 Fujita, M., Kwon, Y.J., Yamaguchi, O.K., and Ogura, K. (1995). J. Am. Chem. Soc. 117: 7287–7288. 107 Qin, Y.-L., Liu, J., Hou, J.-J. et al. (2012). Cryst. Growth Des. 12: 6068–6073. 108 Wu, P.Y., He, C., Wang, J. et al. (2012). J. Am. Chem. Soc. 134: 14991–14999. 109 Gao, H. and Zhang, X.-M. (2012). Dalton Trans. 41: 1562–1567. 110 Zhang, S.-Q., Jiang, F.-L., Wu, M.-Y. et al. (2013). CrystEngComm 15: 3992–4002. 111 Du, M., Jiang, X.-J., and Zhao, X.-J. (2007). Inorg. Chem. 46: 3984–3995. 112 Huang, F.-P., Tian, J.-L., Chen, G.-J. et al. (2010). CrystEngComm 12: 1269–1279. 113 Xue, Z.Z., Sheng, T.L., Wang, Y.L. et al. (2015). CrystEngComm 17: 2004–2012. 114 Yin, Z., Wang, Q.-X., and Zeng, M.-H. (2012). J. Am. Chem. Soc. 134: 4857–4863. 115 Fu, H.-R., Xu, Z.-X., and Zhang, J. (2015). Chem. Mater. 27: 205–210. 116 Chen, M., Zhao, H., Liu, C.-S. et al. (2015). Chem. Commun. 51: 6014–6017. 117 Zhuo, C., Wen, Y.H., Hu, S.M. et al. (2017). Inorg. Chem. 56: 6275–6280. 118 Zheng, B. and Bai, J.F. (2009). CrystEngComm 11: 271–273. 119 Wells, A.F. (1977). Three-Dimensional Nets and Polyhedra. New York: Wiley. 120 Wells, D. (1991). The Penguin Dictionary of Curious and Interesting Geometry. London: Penguin. 121 Grünbaum, B. and Shephard, G.C. (1987). Tilings and Patterns. New York: W. H. Freeman. 122 Moulton, B., Lu, J., and Zaworotko, M.J. (2001). J. Am. Chem. Soc. 123: 9224–9225. 123 Liu, G.-F., Zhang, W.-H., Chen, Y. et al. (2007). Inorg. Chem. Commun. 10: 1049–1053. 124 Zheng, S.-R., Yang, Q.-Y., Liu, Y.-R. et al. (2008). Chem. Commun.: 356–358. 125 Han, Z.-B., Zhang, G.-X., Zeng, M.-H. et al. (2010). Inorg. Chem. 49: 769–771. 126 Cui, P.P., Wu, J.L., Zhao, X.L. et al. (2011). Cryst. Growth Des. 11: 5182–5187. 127 Zhou, Y.F., Xu, W.T., Wu, M.Y. et al. (2012). Inorg. Chem. Commun. 15: 140–145. 128 Sun, D., Han, L.-L., Yuan, S. et al. (2013). Cryst. Growth Des. 13: 377–385. 129 Wen, Y.H., Sheng, T.L., Hu, S.M. et al. (2013). Chem. Commun. 49: 10644–10646. 130 Wen, Y.H., Sheng, T.L., Zhu, X.Q. et al. (2016). Chem. Eur. J. 22: 5327–5334. 131 Zhao, M.T., Huang, Y., Peng, Y.W. et al. (2018). Chem. Soc. Rev. 47: 6267–6295. 132 Li, P.-Z., Maeda, Y., and Xu, Q. (2011). Chem. Commun. 47: 8436–8438. 133 Ding, Y., Chen, Y.-P., Zhang, X. et al. (2017). J. Am. Chem. Soc. 139: 9136–9139. 134 Zhao, M.T., Wang, Y.X., Ma, Q.L. et al. (2015). Adv. Mater. 27: 7372–7378. 135 Sheberla, D., Sun, L., Blood-Forsythe, M.A. et al. (2014). J. Am. Chem. Soc. 136: 8859–8862. 136 Wu, G., Huang, J., Zang, Y. et al. (2017). J. Am. Chem. Soc. 139: 1360–1363. 137 Silva, P., Vilela, S.M.F., Tomé, J.P.C., and Paz, F.A.A. (2015). Chem. Soc. Rev. 44: 6774–6803. 138 Yang, Q., Liu, D., Zhong, C., and Li, J.-R. (2013). Chem. Rev. 113: 8261–8323.

References

139 Barea, E., Montoro, C., and Navarro, J.A.R. (2014). Chem. Soc. Rev. 43: 5419–5430. 140 de Voorde, B.V., Bueken, B., Denayer, J., and Vos, D.D. (2014). Chem. Soc. Rev. 43: 5766–5788. 141 Yang, S., Ramirez-Cuesta, A.J., Newby, R. et al. (2015). Nat. Chem. 7: 121–129. 142 Kreno, L.E., Leong, K., Farha, O.K. et al. (2012). Chem. Rev. 112: 1105–1125. 143 Horcajada, P., Gref, R., Baati, T. et al. (2012). Chem. Rev. 112: 1232–1268. 144 Zhu, Q.-L. and Xu, Q. (2016). Chem 1: 220–245. 145 Cui, Y.J., Zhang, J., He, H.J., and Qian, G.D. (2018). Chem. Soc. Rev. 47: 5740–5785. 146 Yang, Q., Xu, Q., and Jiang, H.-L. (2017). Chem. Soc. Rev. 46: 4774–4808. 147 Lu, W., Wei, Z., Gu, Z.-Y. et al. (2014). Chem. Soc. Rev. 43: 5561–5593. 148 Li, H., Eddaoudi, M., O’Keeffe, M., and Yaghi, O.M. (1999). Nature 402: 276. 149 Serre, C., Millange, F., Thouvenot, C. et al. (2002). J. Am. Chem. Soc. 124: 13519–13526. 150 Surblé, S., Serre, C., Mellot-Draznieks, C. et al. (2006). Chem. Commun.: 284–286. 151 Latroche, M., Surblé, S., Serre, C. et al. (2006). Angew. Chem. Int. Ed. 45: 8227–8231. 152 Férey, G., Mellot-Draznieks, C., Serre, C. et al. (2005). Science 309: 2040–2042. 153 Cavka, J.H., Jakobsen, S., Olsbye, U. et al. (2008). J. Am. Chem. Soc. 130: 13850–13851. 154 Rosi, N.L., Kim, J., Eddaoudi, M. et al. (2005). J. Am. Chem. Soc. 127: 1504–1518. 155 Deng, H., Grunder, S., Cordova, K.E. et al. (2012). Science 336: 1018–1023. 156 Liu, T.-F., Lü, J., and Cao, R. (2010). CrystEngComm 12: 660–670. 157 Li, J.-R. and Zhou, H.-C. (2009). Angew. Chem. Int. Ed. 48: 8465–8468. 158 Chun, H. and Jung, H. (2009). Inorg. Chem. 48: 417–419. 159 Jiang, H.-L., Liu, B., and Xu, Q. (2010). Cryst. Growth Des. 10: 806–811. 160 Zhang, Y.-X., Lin, H., Wen, Y., and Zhu, Q.-L. (2019). Cryst. Growth Des. 19: 1057–1063. 161 Ye, B.-H., Tong, M.-L., and Chen, X.-M. (2005). Coord. Chem. Rev. 249: 545–565. 162 Yang, Y.-Y., Lin, Z.-J., Liu, T.-T. et al. (2015). CrystEngComm 17: 1381–1388. 163 Han, L.-W., Lü, J., Lin, Z.-J., and Cao, R. (2014). CrystEngComm 16: 1749–1754. 164 Gu, Z.-G., Zhan, C.H., Zhang, J., and Bu, X.H. (2016). Chem. Soc. Rev. 45: 3122–3144. 165 Rood, J.A., Boggess, W.C., Noll, B.C., and Henderson, K.W. (2007). J. Am. Chem. Soc. 129: 13675–13682. 166 Zhao, X., Wong, M., Mao, C.Y. et al. (2014). J. Am. Chem. Soc. 136: 12572–12575. 167 Chui, S.S.-Y., Lo, S.M.-F., Charmant, J.P.H. et al. (1999). Science 283: 1148–1150. 168 Férey, G., Serre, C., Mellot-Draznieks, C. et al. (2004). Angew. Chem. Int. Ed. 43: 6296–6301. 169 Luo, J.H., Xu, H.W., Liu, Y. et al. (2008). J. Am. Chem. Soc. 130: 9626–9627. 170 Zheng, S.-T., Bu, J.T., Li, Y.F. et al. (2010). J. Am. Chem. Soc. 132: 17062–17064.

385

386

6 The Structures of Metal–Organic Frameworks

171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204

Chen, B., Eddaoudi, M., Hyde, S.T. et al. (2001). Science 291: 1021–1023. Chae, H.K., Siberio-Perez, D.Y., Kim, J. et al. (2004). Nature 427: 523–527. Furukawa, H., Ko, N., Go, Y.B. et al. (2010). Science 329: 424–428. Wang, X.-S., Ma, S., Sun, D. et al. (2006). J. Am. Chem. Soc. 128: 16474–16475. Ma, S., Sun, D., Ambrogio, M. et al. (2007). J. Am. Chem. Soc. 129: 1858–1859. Zhu, Q.L., Sheng, T.L., Fu, R.B. et al. (2010). Chem. Commun. 46: 9001–9003. He, Y.-P., Tan, Y.-X., Wang, F., and Zhang, J. (2012). Inorg. Chem. 51: 1995–1997. Tan, Y.-X., He, Y.-P., Yuan, D.Q., and Zhang, J. (2018). Appl. Catal., B 221: 664–669. Wang, Y.L., Liu, W., Bai, Z.L. et al. (2018). Angew. Chem. Int. Ed. 57: 5783–5787. Wang, Y.L., Liu, Z.Y., Li, Y.X. et al. (2015). J. Am. Chem. Soc. 137: 6144–6147. Kim, J., Chen, B., Reineke, T.M. et al. (2001). J. Am. Chem. Soc. 123: 8239–8247. Schaate, A., Roy, P., Godt, A. et al. (2011). Chem. Eur. J. 17: 6643–6651. Wu, Z.-F., Tan, B., Feng, M.-L. et al. (2014). J. Mater. Chem. A 2: 6426–6431. Qian, J.J., Jiang, F.L., Yuan, D.Q. et al. (2013). J. Mater. Chem. A 1: 9075–9082. Cui, Y.J., Song, R.J., Yu, J.C. et al. (2015). Adv. Mater. 27: 1420–1425. Pang, J.D., Liu, C.P., Huang, Y.G. et al. (2016). Angew. Chem. Int. Ed. 55: 7478–7482. Pang, J.D., Yuan, S., Qin, J.S. et al. (2017). J. Am. Chem. Soc. 139: 16939–16945. Pang, J.D., Yuan, S., Du, D.Y. et al. (2017). Angew. Chem. Int. Ed. 56: 14622–14626. Guo, Z., Wu, H., Srinivas, G. et al. (2011). Angew. Chem. Int. Ed. 50: 3178–3181. Zhao, D., Yuan, D., Sun, D., and Zhou, H.-C. (2009). J. Am. Chem. Soc. 131: 9186–9188. Yuan, D., Zhao, D., Sun, D., and Zhou, H.-C. (2010). Angew. Chem. Int. Ed. 49: 5357–5361. Yuan, D., Zhao, D., and Zhou, H.-C. (2011). Inorg. Chem. 50: 10528–10530. Xue, H., Zhou, K., Liu, L. et al. (2019). Chem. Commun. 55: 4611–4614. Li, S.-L., Lan, Y.-Q., Sakurai, H., and Xu, Q. (2012). Chem. Eur. J. 18: 16302–16309. Wang, X.-S., Li, L., Yuan, D.-Q. et al. (2018). J. Hazar. Mat. 344: 283–290. Pang, J.D., Jiang, F.L., Wu, M.Y. et al. (2015). Nat. Commun. 6: 7575. Hu, F., Huang, P., Di, Z. et al. (2019). Chem. Commun. 55: 10257–10260. Tan, Y.-X., Wang, F., and Zhang, J. (2018). Chem. Soc. Rev. 47: 2130–2144. Zhang, J.-P., Zhang, Y.-B., Lin, J.-B., and Chen, X.-M. (2012). Chem. Rev. 112: 1001–1033. Huang, X.C., Zhang, J.P., and Chen, X.M. (2003). Chin. Sci. Bull. 48: 1531–1534. Huang, X.C., Lin, Y.Y., Zhang, J.P., and Chen, X.M. (2006). Angew. Chem. Int. Ed. 45: 1557–1559. He, C.-T., Jiang, L., Ye, Z.-M. et al. (2015). J. Am. Chem. Soc. 137: 7217–7223. Park, K.S., Ni, Z., Côté, A.P. et al. (2006). Proc. Natl. Acad. Sci. U.S.A. 103: 10186–10191. Shi, Q., Xu, W.-J., Huang, R.-K. et al. (2016). J. Am. Chem. Soc. 138: 16232–16235.

References

205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235

Wang, F., Liu, Z.-S., Yang, H. et al. (2011). Angew. Chem. Int. Ed. 50: 450–453. Li, Z., Zhang, Z., Ye, Y. et al. (2017). J. Mater. Chem. A 5: 7816–7824. Wang, K., Lv, X.L., Feng, D. et al. (2016). J. Am. Chem. Soc. 138: 914–919. Lv, X.-L., Wang, K., Wang, B. et al. (2017). J. Am. Chem. Soc. 139: 211–217. Kuang, X.F., Wu, X.Y., Yu, R.M. et al. (2010). Nat. Chem. 2: 461. Li, X.X., Xu, H.Y., Kong, F.Z., and Wang, R.H. (2013). Angew. Chem. Int. Ed. 52: 13769–13773. Liao, P.-Q., Zhou, D.-D., Zhu, A.-X. et al. (2012). J. Am. Chem. Soc. 134: 17380–17383. Sumida, K., Stück, D., Mino, L. et al. (2013). J. Am. Chem. Soc. 135: 1083–1091. Lin, Q.P., Wu, T., Zheng, S.-T. et al. (2012). J. Am. Chem. Soc. 134: 784–787. Chen, Y.-Q., Li, G.-R., Chang, Z. et al. (2013). Chem. Sci. 4: 3678–3682. Hasegawa, S., Horike, S., Matsuda, R. et al. (2007). J. Am. Chem. Soc. 129: 2607–2614. Wen, Y.H., Sheng, T.L., Zhuo, C. et al. (2016). Inorg. Chem. 55: 4199–4205. Kang, Y., Wang, F., Zhang, J., and Bu, X.H. (2012). J. Am. Chem. Soc. 134: 17881–17884. Gao, W.-Y., Chrzanowski, M., and Ma, S.Q. (2014). Chem. Soc. Rev. 43: 5841–5866. Lin, K.-J. (1999). Angew. Chem. Int. Ed. 38: 2730–2732. Kosal, M.E., Chou, J.-H., Wilson, S.R., and Suslick, K.S. (2002). Nat. Mater. 1: 118–121. Feng, D., Gu, Z.-Y., Li, J.-R. et al. (2012). Angew. Chem. Int. Ed. 51: 10307–10310. Feng, D., Chung, W.-C., Wei, Z. et al. (2013). J. Am. Chem. Soc. 135: 17105–17110. Huang, N., Yuan, S., Drake, H. et al. (2017). J. Am. Chem. Soc. 139: 18590–18597. Wang, Y.X., Cui, H., Wei, Z.-W. et al. (2017). Chem. Sci. 8: 775–780. Lin, Q.P., Bu, X.H., Kong, A.G. et al. (2015). J. Am. Chem. Soc. 137: 2235–2238. Meng, L., Cheng, Q.G., Kim, C. et al. (2012). Angew. Chem. Int. Ed. 51: 10082–10085. Yang, X.-L., Xie, M.-H., Zou, C. et al. (2012). J. Am. Chem. Soc. 134: 10638–10645. Yuan, G., Jiang, H., Zhang, L. et al. (2019). Coord. Chem. Rev. 378: 483–499. Song, F., Wang, C., Falkowski, J.M. et al. (2010). J. Am. Chem. Soc. 132: 15390–15398. Zhu, C., Yuan, G., Chen, X. et al. (2012). J. Am. Chem. Soc. 134: 8058–8061. Xia, Q.C., Li, Z.J., Tan, C.X. et al. (2017). J. Am. Chem. Soc. 139: 8259–8266. Li, J., Ren, Y., Qi, C., and Jiang, H. (2017). Dalton Trans. 46: 7821–7832. Li, J., Ren, Y., Qi, C., and Jiang, H. (2017). Chem. Commun. 53: 8223–8226. Wu, C.-D. and Lin, W.B. (2007). Angew. Chem. Int. Ed. 46: 1075–1078. Ma, L.Q., Falkowski, J.M., Abney, C., and Lin, W.B. (2010). Nat. Chem. 2: 838–846.

387

388

6 The Structures of Metal–Organic Frameworks

236 Ma, L., Wu, C.-D., Wanderley, M.M., and Lin, W. (2010). Angew. Chem. Int. Ed. 49: 8244–8248. 237 Mo, K., Yang, Y.H., and Cui, Y. (2014). J. Am. Chem. Soc. 136: 1746–1749. 238 Manna, K., Zhang, T., and Lin, W. (2014). J. Am. Chem. Soc. 136: 6566–6569. 239 Manna, K., Zhang, T., Greene, F.X., and Lin, W. (2015). J. Am. Chem. Soc. 137: 2665–2673. 240 Falkowski, J.M., Sawano, T., Zhang, T. et al. (2014). J. Am. Chem. Soc. 136: 5213–5216. 241 Sawano, T., Thacker, N.C., Lin, Z. et al. (2015). J. Am. Chem. Soc. 137: 12241–12248. 242 Kong, G.-Q., Xu, X., Zou, C., and Wu, C.-D. (2011). Chem. Commun. 47: 11005–11007. 243 Kong, G.-Q., Ou, S., Zou, C., and Wu, C.-D. (2012). J. Am. Chem. Soc. 134: 19851–19857. 244 Fei, H., Shin, J., Meng, Y.S. et al. (2014). J. Am. Chem. Soc. 136: 4965–4973. 245 Fei, H. and Cohen, S.M. (2015). J. Am. Chem. Soc. 137: 2191–2194. 246 Pullen, S., Fei, H., Orthaber, A. et al. (2013). J. Am. Chem. Soc. 135: 16997–17003. 247 Li, B., Leng, K., Zhang, Y. et al. (2015). J. Am. Chem. Soc. 137: 4243–4248. 248 Liang, L.F., Liu, C.P., Jiang, F.L. et al. (2017). Nat. Commun. 28: 1233. 249 Humphrey, S.M., Chang, J.-S., Jhung, S.H. et al. (2007). Angew. Chem. Int. Ed. 46: 272–275. 250 Chen, M.-S., Chen, M., Takamizawa, S. et al. (2011). Chem. Commun. 47: 3787–3789. 251 Zhou, H.-L., Lin, R.-B., He, C.-T. et al. (2013). Nat. Commun. 4: 2534. 252 Zhou, H.-L., Li, M., Li, D. et al. (2014). Sci. China Chem. 57: 365–370. 253 Koh, K., Wong-Foy, A.G., and Matzger, A.J. (2008). Angew. Chem. Int. Ed. 47: 677–680. 254 Koh, K., Wong-Foy, A.G., and Matzger, A.J. (2010). J. Am. Chem. Soc. 132: 15005–15010. 255 Deng, H., Doonan, C.J., Furukawa, H. et al. (2010). Science 327: 846–850. 256 Zheng, S.-T., Zuo, F., Wu, T. et al. (2011). Angew. Chem. Int. Ed. 50: 1849–1852. 257 Chen, B.L., Zhao, X.B., Putkham, A. et al. (2008). J. Am. Chem. Soc. 130: 6411–6423. 258 Xiang, S.-C., Zhang, Z.J., Zhao, C.-G. et al. (2011). Nat. Commun. 2: 204. 259 Das, M.C., Guo, Q.S., He, Y.B. et al. (2012). J. Am. Chem. Soc. 134: 8703–8710. 260 Xuan, W.M., Ye, C.C., Zhang, M.N. et al. (2013). Chem. Sci. 4: 3154–3159. 261 Wen, Y.H., Sheng, T.L., Zhu, X.Q. et al. (2017). Adv. Mater. 29: 1700778. 262 Wen, Y.H., Sheng, T.L., Xue, Z.Z. et al. (2015). Inorg. Chem. 54: 3951–3957. 263 Chen, D., Zhuo, C., Wen, Y. et al. (2018). Mater. Chem. Front. 2: 1508–1514. 264 Zhang, H., Lin, C.S., Sheng, T.L. et al. (2016). Chem. Eur. J. 22: 4460–4468. 265 Zhang, H., Sheng, T.L., Hu, S.M. et al. (2016). Cryst. Growth Des. 16: 3154–3162. 266 Han, D., Jiang, F.L., Wu, M.Y. et al. (2011). Chem. Commun. 47: 9861–9863. 267 Qian, J.J., Jiang, F.L., Su, K.Z. et al. (2013). J. Mater. Chem. A 1: 10631–10634.

References

268 Wu, X., Zhang, H.-B., Xu, Z.-X., and Zhang, J. (2015). Chem. Commun. 51: 16331–16333. 269 Zhang, L., Qian, J.J., Yang, W.B. et al. (2015). J. Mater. Chem. A 3: 15399–15402. 270 Nugent, P., Belmabkhout, Y., Burd, S.D. et al. (2013). Nature 495: 80. 271 Li, B., Cui, X.L., O’Nolan, D. et al. (2017). Adv. Mater. 29: 1704210. 272 Lin, Q.P., Mao, C.Y., Kong, A.G. et al. (2017). J. Mater. Chem. A 5: 21189–21195. 273 Gagnon, K.J., Perry, H.P., and Clearfield, A. (2012). Chem. Rev. 112: 1034–1054. 274 Goura, J. and Chandrasekhar, V. (2015). Chem. Rev. 115: 6854–6965. 275 Taylor, J.M., Mahmoudkhani, A.H., and Shimizu, G.K.H. (2007). Angew. Chem. Int. Ed. 46: 795–798. 276 Taylor, J.M., Dawson, K.W., and Shimizu, G.K.H. (2013). J. Am. Chem. Soc. 135: 1193–1196. 277 Yang, W.T., Parker, T.G., and Sun, Z.-M. (2015). Coord. Chem. Rev. 303: 86–109. 278 Nelson, A.-G.D., Bray, T.H., and Albrecht-Schmitt, T.E. (2008). Angew. Chem. Int. Ed. 47: 6252–6254. 279 Adelani, P.O. and Albrecht-Schmitt, T.E. (2012). Cryst. Growth Des. 12: 5800–5805. 280 Yang, W.T., Yi, F.-Y., Tian, T. et al. (2014). Cryst. Growth Des. 14: 1366–1374. 281 Tian, T., Yang, W.T., Wang, H. et al. (2013). Inorg. Chem. 52: 7100–7106. 282 Fu, Ru.B., Hu, S.M., and Wu, X.T. (2011). CrystEngComm 13: 6334–6336. 283 Fu, Ru.B., Hu, S.M., and Wu, X.T. (2013). CrystEngComm 15: 802–807. 284 Fu, Ru.B., Hu, S.M., and Wu, X.T. (2014). Cryst. Growth Des. 14: 6197–6204. 285 Fu, R., Hu, S., and Wu, X. (2017). J. Mater. Chem. A 5: 1952–1956. 286 Hurd, J.A., Vaidhyanathan, R., Thangadurai, V. et al. (2009). Nat. Chem. 1: 705. 287 Chandler, B.D., Cramb, D.T., and Shimizu, G.K.H. (2006). J. Am. Chem. Soc. 128: 10403–10412. 288 Chandler, B.D., Yu, J.O., Cramb, D.T., and Shimizu, G.K.H. (2007). Chem. Mater. 19: 4467–4473. 289 Li, H.R., Sheng, T.L., Xue, Z.Z. et al. (2017). CrystEngComm 19: 2106–2112. 290 Meng, X., Song, S., Song, X. et al. (2015). Chem. Commun. 51: 8150–8152. 291 Zang, S., Dong, X., Wang, R. et al. (2015). J. Mater. Chem. A 3: 641–647. 292 Dong, X.-Y., Wang, R., Li, J.-B. et al. (2013). Chem. Commun. 49: 10590–10592. 293 Li, X.-X., Zhao, D., and Zheng, S.-T. (2019). Coord. Chem. Rev. 397: 220–240. 294 Fang, W.-H. and Yang, G.-Y. (2018). Acc. Chem. Res. 51: 2888–2896. 295 Marleny Rodriguez-Albelo, L., Ruiz-Salvador, A.R., Sampieri, A. et al. (2009). J. Am. Chem. Soc. 131: 16078–16087. 296 Qin, J.-S., Du, D.-Y., Guan, W. et al. (2015). J. Am. Chem. Soc. 137: 7169–7177. 297 Wang, Y.-Y., Zhang, M., Li, S.-L. et al. (2017). Chem. Commun. 53: 5204–5207. 298 Cheng, W., Shen, F.-C., Xue, Y.-S. et al. (2018). ACS Appl. Energy Mater. 1: 4931–4938. 299 Wang, Y.-R., Huang, Q., He, C.-T. et al. (2018). Nat. Commun. 9: 4466. 300 Li, X.-X., Wang, Y.-X., Wang, R.-H. et al. (2016). Angew. Chem. Int. Ed. 55: 6462–6466. 301 Li, X.-X., Deng, C.-C., Zhao, D. et al. (2018). Dalton Trans. 47: 16408–16412. 302 Li, X.-X., Zhang, L.-J., Cui, C.-Y. et al. (2018). Inorg. Chem. 57: 10323–10330.

389

391

7 Structural Design and Rational Synthesis Qipu Lin State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, 155 Yangqiao Road West, Fuzhou 350002, Fujian, P.R. China

7.1

Introduction

During the past decades, great efforts have been devoted to the design and synthesis of polyoxometalate (POM)/chalcogenidometalate clusters/superlattices and metal–organic cages/networks, whose composition and properties can be adapted to specific applications. In this chapter, we give a brief survey of recent advances in the rational assembly of POMs, cluster-based chalcogenidometalates, polyhedral coordination supermolecules, and metal–organic frameworks (MOFs) to show the knowledge on their synthesis mechanism and procedures. First, we present the synthetic design approaches to the many POM types encompassing iso-POMs, hetero-POMs doped by main-group metals, transition metals, rare-earth metals, organically derived POMs by nitrogen alkylation, and coordination with ligands such as carboxylates, POM–MOF hybrids, etc. Next, we summarize the design and synthesis of crystallographically defined chalcogenidometalate clusters, among which supertetrahedral series that bear the closest resemblance to structurally precise fragments of the cubic ZnS-type lattice, termed as Tn (n is the number of metal layers, n = 1–6), are the most fundamental form, which can evolve other supertetrahedral series including penta-supertetrahedral (Pn, n = 1–2) and capped-supertetrahedral (Cn, n = 1–3), super-supertetrahedral (Tm,n, i.e. T2,2, T3,2, T4,2), and oxychalcogenide-supertetrahedral (containing O2− anion inside cluster, i.e. OTn and TOn) types, most of which are allowed to be fabricated by the mixed-metal and cationic-template strategies to deal with the local and the global charge issues. Then, we outline the recent developments of coordination-driven self-assembly, with a focus on discrete architectures, viz metallacages of polyhedral shapes through well-defined directional-bonding assembly strategies (e.g. edge-directed, face-directed, and symmetry-adapted approaches) including diverse linking modes, like square-planar Pd2+ /Pt2+ with pyridyl, octahedral Ga3+ /Fe3+ /V3+ with catechol/pyrogallol, nona-coordinated Ln3+ with tridentate ligand, and square dimeric Cu2+ or calixarene-based tetrameric Advanced Structural Chemistry: Tailoring Properties of Inorganic Materials and their Applications, First Edition. Edited by Rong Cao. © 2021 WILEY-VCH GmbH. Published 2021 by WILEY-VCH GmbH.

392

7 Structural Design and Rational Synthesis

moieties/η5 -C5 H5 -terminated Zr4+ with carboxylates. Finally, we provide an overview of inorganic secondary building units (SBUs, mainly of discrete metal carboxylates, classified by their geometry and the number of metal atoms) and the linkers bearing other binding groups, such as phenols, and their representative network examples also highlight some synthetic strategies, including modulated synthesis, isoreticular expansion, topology-guided design, multivariate (MTV) chemistry, and reaction conditions (e.g. solvo/hydro/ionothermal methods, slow diffusion, sono-chemical, and high-throughput protocol).

7.2

Polyoxometalate Clusters

POMs are a unique class of metal–oxygen clusters, whose structural types and physical properties could be designed by the building block principle. In this section, the synthetic and design approaches to the many POM types are presented encompassing iso-POMs, hetero-POMs doped by main-group metals, transition metals, and lanthanide metals, POMs with N-alkylation, POMs coordinated by ligands (e.g. POM-based cages), POM–MOF hybrids (including POM@MOF), etc. Much of the interest in POMs has arisen because such clusters represent a paradigm in the discovery of systems that could grow up to the nanoscale. Their versatile nature in terms of structure, size, redox/photochemistry, and charge distribution has made the developments of POM chemistry continuing at a rapid pace. The large number of structural types in POMs could be broadly split into three subsets: (i) inorganic iso-POMs, which are metal–oxide clusters that include anions such as SO4 2− , PO4 3− , MoO4 2− /WO4 2− , like the archetypal systems of the Keggin {XM12 O40 } and the Wells–Dawson {X2 M18 O62 } (M = Mo, W, and X is a template); (ii) inorganic hetero-POMs, composed of an MOF, but with the heteroatoms including main-group, transition, and rare-earth metals, which are often much more stable than their iso-POM counterparts; and (iii) organically derived POMs, which are organic–inorganic hybrid isolated systems and POM–MOF/POM@MOF hybrids. The synthetic variables of greatest importance in synthesizing POMs are (i) concentration/type of oxometallate reactant, (ii) pH and type of acid/electrolyte, (iii) introduction of reducing agent or additional ligand, and (iv) other basic parameters such as solvent, reaction temperature, and time.

7.2.1 7.2.1.1

Inclusion of Small Anions/Cations Inclusion of Inorganic Anions/Cations

Growing processes in POM-type systems are mostly based on nucleophiles that start “growing” by directing the formation of electrophiles. Compared with iso-POMs, hetero-versions have been observed with higher nuclearities, e.g. {W48 O152 (PO4 )8 } or {Ce16 (H2 O)36 W148 O488 (AsO3 )12 }, which are all exclusively templated by heteroatom anions such as PO4 3− and AsO3 3− , acting as linking groups and also affecting the nucleophilicity of the polyoxotungstate clusters [1]. Müller et al. reported the formation of a rather giant molybdenum oxide-based nanocluster, the size of hemoglobin

7.2 Polyoxometalate Clusters

(diameter approximately 6 nm), by introduction of SO4 2− to increase the negative charge for preventing uncontrolled linking and initiating further protonation as a prerequisite. The negative charge in the case is supported by the abundance of SO4 2− coordinated to the intermediates and the final cluster species (such situation does not occur in the presence of weaker coordinating anions, such as ClO4 − , or stronger coordinating anions, such as PO4 3− ) [2]. By introducing AsO3 3− and SbO3 3− , Dickman and coworkers also developed a tungstoarsenate, {As6 W65 O217 (H2 O)7 }26− , and an antimony analog, {Sb6 W65 O217 (H2 O)7 }26− , isolated in an acidic medium (pH = 2) with good yield [3]. In POMs’ chemistry, Na+ and K+ ions are also used widely as counterions for making lots of POMs. The assembly of {(H2 O)4 K ⊂ [H12 W36 O120 ]}11− requires trace amounts of K+ ions, which is similar to the inclusion of potassium in [(Mo6 O6 )K] motifs that are observed in the spherical {K20 ⊂ Mo80 V22 } cluster [1]. Zheng and coworkers also constructed a high-symmetry POM, {K42 Ge8 W72 O272 }, via incorporating high-nuclearity K42 cluster [4]. 7.2.1.2

Inclusion of Organic Anions/Cations

Cronin and coworkers used squarate and molybdate to obtain a family of huge-size polyoxothiometalates (e.g. {Mo45 }, {Mo47 }, {Mo55 }, {Mo68 }, {Mo96 }, all contain a unique isomer of the Lindqvist-type [Mo5 O18 ]6− ). The assembly of the system is also dependent on the nature of the alkali-metal cations used (in the presence of K+ for {Mo55 }, but Cs+ for {Mo68 }). The introduction of the squarate template is crucial for the self-condensation of [Mo2 O2 S2 (H2 O)6 ]2+ to form these macrocycle structures [5]. Through a remarkable symmetry-breaking process, Müller et al. reported the formation of a rather giant molybdenum oxide-based nanocluster, the size of hemoglobin (diameter approximately 6 nm), and containing wheel {Mo176 } and ball {Mo102 } clusters [2].

7.2.2

Multivariate Metal Mixing

The introduction of other metals (e.g. 3d-transition, Tm, and 4f–lanthanide metals, Ln) into iso-POMs is an effective route for the construction of POMs with diverse chemical compositions, structural architecture, and versatile functionalities. For instance, this strategy has achieved great success in creating Tm/Ln-substituted POMs based on tungstate/molybdate, phosphate/arsenate, silicate/germanate, or their combination. Especially, this strategy can allow access to a greater level of self-assembly to form giant heterometallic clusters or cycles or cages, such as the recently reported examples {Zr24 W4 Ge6 W52 }, {Ni25 P6 Si6 W54 }, and also {Dy30 Co8 Ge12 W108 }. By the introduction of Cu2+ ions, Zheng and coworkers synthesized a class of high-nuclearity cupric-niobite clusters, including {K ∩ Cu3 Nb78 } and {K ∩ Cu4 Nb78 } [6]. Through adding cobalt nitrates, Wang and coworkers fabricated four stable clusters of Co2+ -doped POMs, {Co4 P4+4 W36 }, {Co4 P4 Si4 W36 }, {Co4 P4 Ge4 W36 }, and {Co4 P4 As4 W36 } [7]. Some hetero-POMs, e.g. {P8 W48 }, have proven to be a fruitful building unit in the construction of POM-based infinite porous lattices, one of the most striking examples of which is the Mn-linked cubic framework, {Mn8 P8 W48 }, as shown in Figure 7.1. Individual {P8 W48 } rings are

393

394

7 Structural Design and Rational Synthesis

Hetero-POMs; XaMn

Iso-POMs; Mn Lindqvist ions; M6

XM18

W11

X5M30

W36 Wheel

X8M48

Sphere

Network

TM, LN, ... Iso-POM

Organic ligand Hetero-POM POM-Organic hybrid

Figure 7.1 The different classes of polyoxometalates, including iso-POMs, hetero-POMs, POM-based cages/networks, and POM–MOF hybrids. Source: Mitchell et al. [8]; Miras et al. [9].

connected through Mn2+ centers to form a face-directed cube, with internal cavity volume of approximately 7.2 nm3 [8]. A sodium salt of [ErW10 O36 ]9− was prepared by Coronado and coworkers [10]. Boskovic and coworkers also reported two discrete terbium-containing POMs, {Tb2 As2 W16 W3 } and {Tb8 As4 W32 W6 } [11]. By Gd3+ acting as efficient linkers, Patzke and coworkers reported a gadolinium-containing polyoxotungstate, {Gd8 As12 W124 O432 }60− [12]. Among known heterometallic Nb-POMs, the major structure types are derived from isopolyniobates [Nb6 ], [Nb10 ], or heteropolyniobates [XNb12 ] (X = Si, Ge, P, V). Zheng and coworkers reported a series of the multi-metal composite {Ln12 W12 O36 (Nb6 O19 )12 } (Ln = Y, La, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Yb), presenting multilayered structures and also containing

7.2 Polyoxometalate Clusters

sodalite-type cages. Notably, {Ln12 W12 Nb72 } can act as a giant synthon for anionic three-dimensional inorganic–organic hybrid framework, cluster-in-cluster-like [Cu(en)2 ]4 ⋅{Ln12 W12 Nb72 O264 }, of which the sodalite cages bridged together by the complex of [Cu(en)2 ]2+ (en = ethylenediamine) and further charge-balanced by Na+ and K+ [13]. Powell and coworkers built a giant super-tetrahedral heterometallic POM, {Dy30 Co8 Ge12 W108 O480 }56− , which shows single-molecule magnet behavior and is also an example to incorporate 3d–Tm and 4f–Ln assemblied within a same POM. This work indicates that the key to isolating such a hybrid 3d–4f POM nanocluster lies in finding the reaction conditions that favor the formation of the component units and the self-assembly reaction to give the product [14].

7.2.3

Nitrogen Alkylation

Peng and coworkers reported the synthesis of a family of N-alkylated POMs through Pd-catalyzed coupling reactions and demonstrate, for the first time, that iodo-functionalized hexamolybdates can undergo coupling reactions with ethynylarenes. Such reaction (under the protection of N2 , monitored by thin-layer chromatography) has led to a variety of hybrid materials containing covalently bonded POM clusters and organic conjugated moieties, which can be prepared in a controllable way. The products exhibit excellent solubility in common organic solvents such as dichloromethane, chloroform, acetone, acetonitrile, THF, and N,N ′ -dimethylformamide (DMF) [15]. Wei and coworkers developed a hierarchical assembly strategy involving both coordination and noncovalent interactions successively, which results in the formation of nanoscaled complexes, tetrapolyoxometalate square {Mo6 }4 {Cu2 }, by using a soluble organoimido-derivatized hexamolybdate containing a carboxyl terminus, {O18 Mo6 ≡ NC6 H4 –COOH}, as subunits and Cu2+ as the coordinating metal ions [16]. Wei and coworkers also built chiral metallamacrocycles based on POM (viz {Mo6 O17 (≡NC6 H4 OCn H2n OC6 H4 N≡)} (n = 4, 6, 8), some of which undergo spontaneous resolution upon crystallization) from the achiral Lindqvist hexamolybdate and bisarylamines [17].

7.2.4 7.2.4.1

Coordination with Organic Ligands POM-Based Cages

Under hydrothermal conditions, Lan and coworkers made a crystalline POM-based solid, [TPT]5 {Zn4 PMo8 Mo4 O40 } (NNU-11, TPT = tris-(4-pyridyl)triazine). In NNU-11, each Zn-ε-Keggin connects to four neighbors by π–π interaction of the pyridyl rings from different TPT [18]. Yang and coworkers reported an unprecedented POM-based cubic molecular cage, [Ni(en)2 ]6 {Ni6 (Tris)(en)3 (BTC)1.5 (PW9 O34 )}8 (en = ethylenediamine, BTC = 1,3,5-benzenetricarboxylate), by covalently anchoring Tris onto the surface of Anderson-type {Ni6 PW9 }. Notably, the incorporation of Tris causes the {Ni6 PW9 } to not easily agglomerate because of the steric hindrance of the Tris. Metal–organic polyhedra (MOPs) with tetrahedral and octahedral symmetry can be designed by the use of suitable triangular or square

395

396

7 Structural Design and Rational Synthesis

molecular subunits [19]. As developed by Zaworotko and coworkers, reaction of triangular and tetragonal subunits with carboxylate spacers afforded two groups of Goldberg MOPs (tetrahedral and octahedral MOPs) under solvothermal conditions. Wang and coworkers also fabricated another family of Goldberg MOPs by using pentagonal {WV5 } assembled with linear or triangular ligands of different lengths to generate four examples of two distinct types (Gv(1,1) are composed of 12 {WV5 } subunits and 30 linear ligands, and Gv(2,0) are composed of 12 {WV5 } subunits and 20 triangular ligands) [20]. Wang and coworkers constructed two supramolecular isomers based on polyoxovanadate organic polyhedra, V-MOP-α and V-MOP-β (in a truncated tetrahedral geometry, having unit-cell volume of 470 842 and 15 513 Å3 , respectively), by a bottom-up approach. Moreover, V-MOP-α and V-MOP-β can undergo solvent-mediated isomerization induced by temperature to achieve reversible structural transformation [21]. Wang and coworkers also listed a series of theoretically predicted polyoxovanadate MOPs obeying the minimal transitivity principle [22]. 7.2.4.2

POM-Based Nets

Su and coworkers developed two 3D chiral POM-based frameworks, D/L-{KCu3 BW12 } (built from homochiral intertwined double helices), by utilizing enantiopure proline and Keggin-type [BW12 O40 ]5− anions as precursors [23]. Yang and coworkers constructed a series of POMOFs (viz POM-based MOFs) by the use of a {Ni6 PW9 } cluster (formed in situ and containing six terminal water molecules offering the possibility for the replacement with carboxylates) as subunits and multi-carboxylates as linkers under hydrothermal conditions [24]. Lan and coworkers also developed another two 3D POM-based MOFs, [TBA]3 {Zn4 PMoV8 Mo4 ][BTB]4/3 } (NENU-500, BTB = benzene tribenzoate, TBA = tetrabutylammonium) and [TBA]3 {Zn4 PMo8 Mo4 ][BPT] (NENU-501, BPT = [1,1′ -biphenyl]-3,4′ ,5-tricarboxylate), by the hydrothermal technique [25]. Su and coworkers fabricated a stable (4,12)-connected 3D coordination framework, (NH4 )[Cu24 I10 ][PMo12 O40 ]3 L12 (L = 4-[3-(1H-1,2,4-triazol-1-yl)propyl]-4H-1,2,4triazole), which is built of two traditional but distinct nanoclusters by hydrothermal reaction of {Cu24 I10 } and Keggin anions of [PMo12 O40 ] (c. 10.5 Å in diameter) [26].

7.2.5 7.2.5.1

Others Using Precursors

Synthesis by using precursors has opened up a range of perspectives for the development of much larger, more spherical cluster architectures. Cronin and coworkers reported two polyoxoniobate cluster anions of {Nb27 O76 }16− and intrinsically chiral {Nb31 O93 (CO3 )}23− , both of which are synthesized by subjecting {Nb6 } of precursor to hydrothermal conditions in the presence of sodium dibenzyldithiocarbamate [27]. This strategy is also successfully extended to the polyoxotungstate chemistry by using {W36 } cluster as a building unit for larger architectures with other electrophiles [1]. By the same way, three polyoxoniobates, {Nb24 }-A, {Nb24 }-B, and {Nb32 } (all use heptaniobate [Nb7 O22 ]9− as precursor), were isolated by Wang

7.2 Polyoxometalate Clusters

and coworkers under the conventional aqueous conditions [28]. Using lacunary {GeW9 O34 } as a precursor, a high-symmetry POM, {K42 Ge8 W72 O272 }, was made by Zheng and coworkers [4]. By utilizing enantiopure proline and Keggin-type [BW12 O40 ]5− anions as precursors, Su and coworkers fabricated two 3D chiral POM-based frameworks, D/L-{KCu3 BW12 }, built from homochiral intertwined double helices [23]. 7.2.5.2

Templating by Clusters

Cronin and coworkers employed a flow system to study the steps underlying assembly of an ellipsoidal molybdenum oxide cluster@wheel, {Mo36 } ⊂ {Mo150 }, of 3.6 nm in diameter. They observed crystallization of an intermediate structure in which a central {Mo36 } cluster appears to template the assembly of the surrounding {Mo150 } wheel. The template’s role in the self-assembly mechanism is further confirmed by the deliberate addition of the template to the reaction mixture, which greatly accelerates the assembly time of the {Mo150 } wheel and increases the yield [29]. 7.2.5.3

Charge-Balancing by Complexes

A nanocluster of {Ni(enMe)2 }3 {Ni20 P4 W34 }, with idealized Ci symmetry, was made by Yang and coworkers under hydrothermal reactions in water and by lacunary POT as structure-directing agents (SDAs). This sandwich-type structure contains {Ni(enMe)2 }2+ complexes serving as counter-cations [30]. 7.2.5.4

Single-Crystal to Single-Crystal Transformations

Zheng and coworkers constructed two intriguing matryoshka doll-like four-shell structures of {Ln ∩ W6 ∩ Ln26 ∩ W100 } (Ln = La and Ce), which contain unique elliptic [Ln26 ] clusters incorporated into 10 lacunary [GeW10 O38 ]12− . Intriguingly, {Ln27 Ge10 W106 } can undergo single-crystal to single-crystal transformations to extended inorganic–organic hybrid frameworks using [Ln28]-containing POMs as building nodes [31]. 7.2.5.5

Reaction pH Control

Wang and coworkers synthesized three polyoxoniobates ({Nb24 }-A, {Nb24 }-B, and {Nb32 }, all built from heptaniobate [Nb7 O22 ]9− units and crystallized with only alkali-metal counterions). The optimal pH ranges of isolation of {Nb24 }-A, {Nb24 }-B, and {Nb32 } are 10.5–11.2, 10–11.5, and 10.6–11.5, respectively [28]. 7.2.5.6

Networked Reactor System

Cronin and coworkers applied a networked reactor system (allowing one-pot reactions to be probed or expanded over a number of reaction vessels, rather than relying on one single vessel) for the discovery of POMs (chain-linked {W11 Co}n and {W200 Co8 } with over 4 nm in diameter). The self-assembly processes are controlled by modulation of the reaction conditions, thus allowing different libraries of building blocks from each of the reactors to be condensed together within this system [32].

397

398

7 Structural Design and Rational Synthesis

7.3

Chalcogenidometalate Superlattices

Compared to insulating oxides, like zeolites, crystallographically defined chalcogenidometalate clusters and open framework solids have attracted considerable interest because of their intrinsic optoelectronic properties resulting from their narrow bandgaps, often in semiconducting range. Among them, supertetrahedral series that bear the closest resemblance to tetrahedrally shaped fragments of the cubic ZnS-type lattice, termed as Tn (n is the number of metal layers), are the most fundamental form, which could evolve other supertetrahedral series including penta-(Pn, which are derived from assemblage of four Tn clusters tetrahedrally distributed onto four faces of one anti-Tn cluster) and capped-(Cn, which are composed of a Tn cluster at the core covered with a shell of atoms, whose stoichiometry is also related to the Tn) types and others (e.g. super-supertetrahedral Tm,n and oxygen-stuffed variants of OTn and TOn). In this section, the synthetic and design approaches to these supertetrahedral chalcogenidometalate moieties and their superlattices are presented. For an underlying Tn (n > 3), its low-coordinated anions especially on the edges are prone to bond to higher-valent metals, whereas its μ3/4 -coordinated chalcogens at the face/core favor the cations in lower valence, which is related to the Pauling’s electrostatic valence rule (viz the charge of an anion is balanced locally by its adjacent cations). Up to now, most studies are focused on the mixed-metal and cationic-template strategies for the charge-density matching. According to the classification of guest cationic species, the chalcogenidometalates can also generally be divided into two categories: (i) ones with organic quaternary ammonium or phosphonium cations, such as R4 N+ and R4 P+ (P = Me, Et, Pr, etc.), or with various protonated organic amines, which act as templating or SDAs, and (ii) ones with alkali metals (like Li+ , Na+ ) or transition/lanthanide metal complexes. Chalcogenidometalates are typically synthesized from a basic solution under hydrothermal/solvothermal conditions with alkali-metal cations, molecular metal complexes, or protonated organic amines as the templates or SDAs. Alternatively, chalcogenidometalates can also be prepared by using the ionothermal approach or surfactants as promising reaction media.

7.3.1

Mixing Metals for Valance–Balance

It is recognized that the local and/or whole charge-density matching plays a key role in the crystallization of supertetrahedral metal–chalcogenide clusters. For a basic supertetrahedral cluster of II–VI Tn with the theoretical formula of [Mn(n+1)(n+2)/6 S(n+1)(n+2)(n+3)/6 ](n+1)(n+2)− , its global charge increases with the increase of cluster size (while T4-ZnS has the formula of [Zn20 S35 ]30− , T5-ZnS is of [Zn35 S56 ]42− ). Such highly negative charge origins from low-coordinated μ2 -S on edges and μ3 -S sites on faces of the cluster. The method that avoids the use of surface organic groups but can deal with global charge issue is to replace divalent metal (M2+ ) with higher-valent metal (M3+ or M4+ ). However, this method creates another issue: local charge balance (the distribution of total cationic charge around each anion), especially for a huge Tn, because μ4 -S sites do not favor high-valent metal

7.3 Chalcogenidometalate Superlattices

ions. According to Pauling’s electrostatic valence rule, coordination configuration such as (M3+/4+ )4 (μ4 -S) is unstable due to the excessive cationic charge around S2− . Such instability becomes exacerbated for larger Tn clusters because there would be more μ4 -S sites. To reconcile the needs for both whole and local charge balance, a mixed-metal strategy was adopted, i.e. replacing some divalent metal sites with high-valent metal ions while keeping metal sites surrounding μ4 -S low-valent ions (e.g. (M2+ )4 (μ4 -S) or (M+ )2 (M3+ )2 (μ4 -S)). Numerous efforts on heterometal combination have led to T1-to-T5 of supertetrahedral chalcogenidometalate clusters in various compositions and intercluster connectivity [33]. In addition, the ratios of mixed-metal cations also need to be varied to be in sync with the increased number of core μ4 -S sites and also with the charge density of SDAs. Lately, by controlling the high ratio of Zn2+ to In3+ , Wu and coworkers realized the so-far largest supertetrahedral cluster (T6, [Zn25 In31 S84 ]25− ) [34].

7.3.2

Using Surface-Capping Ligands

To address the charge issue, the use of surface-capping ligands has also proved effective to reduce global charge and sometimes results in the formation of other cluster types such as capped-Cn. Each Cn cluster is composed of a core that is a regular fragment of the cubic zinc blende-type phase, viz Tn, and four corner barrelanoid cages possessing the features of the hexagonal wurtzite-type phase. Nearly all Cn clusters are synthesized in the form of molecular crystals or covalent superlattices through corner-sharing thiolates. They are mainly synthesized under solvothermal conditions at temperatures between 60 and 120 ∘ C, much lower than that for other cluster-based metal–chalcogenide structures. In a typical synthetic process as MOL-1 of C3,4-{Cd54 S32 (SPh)48 }, thiourea is used as the S2− source, and Cd2+ and SPh− sources are usually in the form of precursor Cd(SPh)2 [35]. By replacing thiourea with selenourea, another core–shell-type C3,4 cluster, [Cd54 Se32 (SPh)48 (H2 O)4 ]4− , can be synthesized as well. It is also possible to incorporate transition metal cations (e.g. Co2+ and Fe2+ ) into nanoclusters, further increasing the diversity and potential applications of materials that may be realized through this synthetic approach. Similar molecular crystals have also been grown from clusters of other sizes by using this approach, indicating the broad applicability of the surface-capping method in synthesizing chalcogenidometalates of quantum dots [36].

7.3.3

Termination with Superbases

Apart from the methods aforementioned (introducing higher-valent metal ions, e.g. In3+ and Sn4+ , or using surface-capping ligands, e.g. SPh− and SePh− , onto the surface), there are other ways to reduce the negative charge and to help stabilize the clusters: (i) balancing the charge through metal complex cations or (ii) terminating the cluster corners with neutral ligands. To stabilize the rapidly increasing negative charge on the cluster as the size of the cluster gets bigger, organic superbases were also used to terminate the vertexes of Tn clusters. The choice of the organic

399

400

7 Structural Design and Rational Synthesis

superbases is crucial because the charge reduction, charge stabilization, and charge balance should be concerned. The affinity of the base with the metal ion in the formation of M—N bond can reduce the cluster charge by terminating its corners, and the basicity of the superbase in the protonated form can stabilize the negative cluster. Some single bases (e.g. lutidine, DBU (= 1,8-diazabicyclo[5.4.0]-7-undecene), DBN (= 1,5-diazabicyclo[4.3.0]5-nonene), and imidazolium) can meet both roles mentioned earlier. A family of discrete supertetrahedral metal–chalcogenide clusters with different sizes (termed as ISC-n, built of Tn, respectively, with four corners terminated by DBU/DBN, where n = 3, 4, 5) have been made by Feng and coworkers [37]. Huang and coworkers have also made molecular T5 clusters terminated by the nitroalkylated imidazole (IM) generated in situ from the decomposition of the ionic liquid (IL), the used IL serves as both solvent and stabilizer [38].

7.3.4

Using Structure-Directing Agents (SDAs)

Similar to a synthetic procedure of artificial metal–oxide zeolites, alkali (earth)-metal and organic amine cations can be used as SDAs [39]. ICF-26, built from a metal–chalcogenide cluster of P2-[Li4 In22 S44 ]18− , was prepared by using Li+ cations as SDAs. For most other case, organic amine species are needed as surface-stabilizing ligands and/or extraframework SDAs. Compared with the O⋅⋅⋅H—N bonding in oxide frameworks, the hydrogen bonding between chalcogenide frameworks and protonated amines (e.g. S⋅⋅⋅H–N) is much weaker. The co-assembly of chalcogenidometalate clusters with guest molecules depends to a large extent on the host–guest Coulombian interaction. This helps to explain that metal–chalcogenide open frameworks generally have a rather negative framework and neutral or nearly neutral cluster-based metal–chalcogenide structures are few. To match the highly negative host networks, guest molecules in open framework based on chalcogenidometalate clusters usually have a high charge density. An example of a typical preparation of UCR-20GaGeS-TAEA is as follows: gallium metal, germanium oxide, sulfur powder, and tris(2-aminoethyl)amine were mixed in a Teflon-lined stainless steel autoclave for about 20 minutes. The vessel was then sealed and heated at 190 ∘ C for six days. The success with the supertetrahedral chalcogenidometalate cluster-based system are mainly attributed to the use of the amine-directed synthesis method [40].

7.3.5

Using Metal Complexes as Templates

Metal complexes are effective SDAs to control the organization of metal– chalcogenide clusters in the 3D space. While [Fe(1,10-phenanthroline)3 ]2+ controls the formation of MOL-6, larger [Fe(3,4,7,8-tetramethyl-1,10-phenanthroline)3 ]2+ controls the packing mode in MOL-7 (both MOL-6 and MOL-7 consist of C2,1-[Cd32 S14 (SC6 H5 )38 ]2− ). Compared to commonly used protonated amines or inorganic cations, the metal-chelate dyes have larger size and lower charge density, which makes them ideal to template the formation of superlattices that

7.3 Chalcogenidometalate Superlattices

are also built of low-charge density clusters, such as Cn. The use of metal-chelate molecules as templates is essential for the synthesis of the covalently linked frameworks based on large Cn clusters, as in the COV-10/11 (twofold interpenetrating diamond-type network with C2-[Cd32 S14 (SC6 H5 )38 ]2− at the tetrahedral node). The large extraframework space is occupied by the cationic metal complexes. In addition, the hydrophobic surface of the metal-chelate dyes also matches well with the hydrophobic surface of the nanoclusters. Just due to the low charge density of Cn clusters, the frameworks of COV-10/11 have considerably lower charge density than those from Tn and other clusters. The structural and compositional diversity of metal complexes offers many possibilities in the controlled assembly of different low-charged nanoclusters into open architectures that are not accessible by other templating methods [41].

7.3.6

Mimicking Zeolites

Tetrahedral clusters are usually linked together via single bridge (e.g. S2− or Se2− ). If considering each cluster as a node, an open metal–chalcogenide framework could be simplified as a four-connection topological net, just like a zeotype. In recent two decades, a great progress has been made in the synthesis of open frameworks built from supertetrahedral clusters. Among them, the single and double diamond-type nets are the most common. For small clusters (e.g. T2 and T3), both single and double diamond-type structures have been synthesized; while for structures with T3 or larger clusters, only double diamond frameworks are known. In this case, the net interpenetration and guest molecules (or cations) combine to fill the large void space. Currently, about 248 framework topological types are approved by the Structure Commission of International Zeolite Association (IZA-SC). However, up to now only 5 out of 248 zeotype frameworks are composed by metal–chalcogenides (viz SOD, BCT, ABW, RWY, and NAB, some of which are plotted in Figure 7.2). Feng and coworkers synthesized a family of metal–chalcogenide zeolite analogs by simultaneous triple substitutions of O2− with S2− or Se2− , Si4+ with Ge4+ or Sn4+ , and Al3+ with Ga3+ or In3+ [43]. The metal–sulfide/selenide open frameworks can be enhanced by mimicking high-silica zeolites, and their crystallization is facilitated by introducing a small amount of divalent metal ions during the synthetic process [44].

7.3.7

Amine-Based Solvothermal Reaction

Up to now, several synthetic methods have been widely used in the preparation of crystalline chalcogenidometalates, including high-temperature solid-state synthesis, molten-flux techniques, low-temperature solution processing, and hydro(solvo)thermal synthesis. Except for solid-state synthesis, reaction solutions (such as molten salts, water, organic solvents, or organic amines) are usually used in these synthetic strategies, in which alkali (alkaline earth)-metal cations, metal complexes, and organic amines act as counterions, SDAs, and templates in the construction of metal–chalcogenide frameworks. One of the reasons for

401

402

7 Structural Design and Rational Synthesis

T1 T2 Metal Chalcogen

T3 T4

Oxygen

T5 T6

OT5

P2 C2 Types of supertetrahedral cluster

TO2

Types of framework

RWY

NAB

CRB

Figure 7.2 The different types of supertetrahedral metal–chalcogenide clusters (Tn, Cn, Pn, OTn, TOn) and their superlattices by interconnection. Source: Based on Feng et al. [42].

the difficulty in synthesizing supertetrahedral metal–chalcogenide clusters is their usually negative charge that increases dramatically as the size of the cluster gets bigger. As the cluster gets larger, the increase of μ2/3 -chalcogen sites on the surface of the cluster contributes to a dramatic increase in the magnitude of the negative charge per cluster. Accordingly, solvothermal techniques using amines have proven to be one of the most effective routes to lower the negative charge of cluster-based chalcogenidometalates. A variety of organic amines have been introduced into the structures of chalcogenidometalates, in which the organic species may play two different roles in terms of the interactions with the inorganic components. One is to serve as SDAs, templates, space-filling agents, and charge-balancing agents through either ionic bonding and/or relatively weak H-bonding and van der Waals interactions. The other is to form an organic–inorganic hybrid architecture via coordination to metal ions of the inorganic cluster as ligands [45].

7.3 Chalcogenidometalate Superlattices

7.3.8

Ionothermal Synthesis

The increasing attraction in using ILs is due to their excellent solvating nature, which has the high thermal stability, a wide liquidus range, and the ability to dissolve a large variety of chemicals. Dehnen and coworkers summarize all the cations and anions that have been used in the formation of the reported ILs for chalcogenidometalate synthesis [46]. Four discrete T5 clusters with Cu–(Ga/In)–S composition have been synthesized in [Bmmim]Cl (Bmmim = 1-butyl-2,3-dimethyl-imidazolium). The ionothermal environment and specific structural effect provided by ILs could be advantageous for the formation of larger isolated Tn clusters (termination by bonding with imidazolium not only adds less negative charge to the clusters compared to that of regular clusters with chalcogen located at the apexes but also prevents them from polymerization, thus promoting the opportunity to form large discrete clusters) [38].

7.3.9

Using Surfactants

Compared with ILs, surfactants are much cheaper and also have diversified characters such as acidic, basic, neutral, cationic, and anionic. It is well known that surfactants can not only tailor the sizes, shapes, and surface properties of nanocrystals but also control the pore sizes and phases of mesoporous frameworks. These advantages make surfactants potential to be as reaction media for the synthesis of crystalline metal–chalcogenides and offer more choices for controlling the crystal growth [47]. Under surfactant-thermal conditions, single crystals of the poly(ethylene glycol) (PEG)–selenidostannate composite have been produced [48]. The PEG chains occupy the channels surrounded by the honeycomb [Sn3 Se7 ]n 2n− layers and form multiple hydrogen bonding with amine cations and [Sn3 Se7 ]n 2n− anionic layers. The H-bonded network formed between PEG chains and inorganic layers and amine cations at lower temperature play an important role in the crystallization of the PEG–selenidostannate organization. Surfactants can act as counterions and templates (owing to their cationic, anionic, and neutral properties) in the formation of chalcogenide frameworks [49]. Surfactants can be used as reactants in reactions with inorganic precursors to produce crystalline chalcogenides under solution conditions at room temperature. The surfactant environment is quite different from molecular-solvent conditions in the preparation of crystalline chalcogenides. Besides their use as templates or counterions in the synthesis of crystalline chalcogenides, surfactants can also be used as promising reaction media to control the crystal growth of chalcogenides.

7.3.10 Others 7.3.10.1 Insertion of O2− Anions

According to Pauling’s electrostatic valence rule or Brown’s equal valence rule, regular T3 and even larger supertetrahedral chalcogenidometalate clusters of single trivalent or tetravalent metals are unstable, as there would be an overabundance of

403

404

7 Structural Design and Rational Synthesis

the total bond valence for face and core (if any) chalcogen sites. Filling the cluster with oxygen helps to dilute some of the bond valence away from the chalcogen anions, as in the filled T3 clusters, [Sn10 O4 S20 ]8− , in which each adamantane-type cage envelopes one O2− anion. In 2016, a three-dimensional open framework based on the pseudo-T4 [In4 Sn16 O10 S34 ]12− cluster is developed by Zhang and coworkers [50]. By using suitable structure-directing and charge-balancing templates, Lin and coworkers realized a network built from the filled T4 [Sn20 O10 S34 ]8− (denoted as 3D-T4-SnOS) [51]. The introduction of O2− anions is to eliminate excessively high charge on the μ3 -S2− anions and to stabilize the Sn4+ ions at the face sites. Two T5-InSO clusters in IOS-1 and IOS-2 are formed with the indium core surrounded by eight O2− ions [52]. Moreover, doping with nonmetal elements (e.g. O) into metal–chalcogenide clusters is also recognized as an effective tool to tailor band gap and enhance the stability. 7.3.10.2

Linkage by Pyridines/Imidazoles

Neutral organic ligands (e.g. multi-pyridine) have been shown to function as cross-linkers between supertetrahedral clusters (of which the size is limited to T3 with 10 metal sites), and the superlattices made are mainly low-dimensional, such as 1D/2D. Metal–chalcogenide clusters can also be interlinked by various imidazolate ligands, as realized by Feng and coworkers. Such co-assembly brings about the interface chemistry between metal–chalcogenide clusters and open zeolitic imidazolate frameworks (ZIFs). Integration of organic components into metal–chalcogenide open frameworks offers opportunities not only for new structure types but also for synergistic properties resulting from inorganic and organic components through their uniform integration at the molecular level [53]. 7.3.10.3

Reaction Parameters

Reaction temperature, solvent, surface ligand type, and chalcogenide source are key synthetic parameters, affecting the size and network topology of resulting metal–chalcogenide materials. As reported by Feng and coworkers, co-crystallization of helices of the same handedness or opposite handedness can be tuned by the synthetic conditions [54]. 7.3.10.4

Postsynthetic Insertion of Heterometals

A two-step synthesis strategy is applied by Wu and coworkers to realize an atomically precise doping of Mn2+ ion into the core site of the nanoclusters and to achieve uniform distribution of Mn2+ dopants in the crystal lattice. The method is to utilize a well-defined isolated supertetrahedral nanocluster, [Cd6 In28 S52 (SH)4 ], whose core metal site is unoccupied in as-synthesized pristine form [55]. 7.3.10.5

Phase Transformation

The 3D open framework covalently linked with T4-[Mx Ga18−x Sn2 Q35 ]12− (x = 2 or 4; M = Mn, Cu, Zn; Q = S, Se) could be transformed into the corresponding discrete molecular clusters (done by Feng and coworkers). The driving force for such a phase

7.4 Polygonal/Polyhedral Complexes

transformation is the perfect match in both charge density and geometry between chalcogenide clusters and protonated amines, leading to the higher stability of the discrete clusters.

7.4

Polygonal/Polyhedral Complexes

Crystal-engineering self-assembly has witnessed a variety of discrete metallacages with the shapes of polygon, Platonic polyhedron (e.g. tetrahedron, cube, octahedron, dodecahedron, icosahedron), Archimedean polyhedron (e.g. cuboctahedron, truncated tetrahedron, rhombicuboctahedron), and other versions (e.g. prism, Goldberg type, Stellated type). The inherent cavities of metallacages have been exploited for host–guest chemistry, leading to applications in sensing, catalysis, and other studies. According to the interactions used in the assembly process, supramolecular chemistry can be classified into many branches, but in this section, we are focused on those that just use strong bondings between the multibranched chelating ligands and the metals with satisfied coordination geometries. Ligands used during assembly process could be bidentate, tridentate, and tetradentate which combine with matched metal node to give the desired polyhedral complexes. The shape and the size of the final self-assembled structures are largely dependent on the number and the bent angle of the coordination orientations of the ligands, for instance, bidentate ligands with bent angle 120∘ , 135∘ , and 149∘ could produce spherical molecules with the geometry of cuboctahedron, rhombicuboctahedron, and icosidodecahedron, respectively. We illustrate the designed synthesis of supramolecular architectures of different shapes (mainly polyhedral cages) based on coordination interactions (e.g. square-planar Pd2+ /Pt2+ with pyridyl, octahedral Ga3+ /Fe3+ /V3+ with catechol/pyrogallol, nona-coordinated Ln3+ with tridentate ligand, square dimeric {Cu2+ 2 } or calixaren–transition metal clusters/η5 -C5 H5 -terminated trimer of Zr4+ with carboxylic groups), of which directional bonding, symmetry matching, molecular paneling/block building approaches are the most extensively used and adopted. In addition, the orthogonality of metal–ligand bonding formation with other interaction, like hydrogen bonding, electrostatic chemistry, templating effect, etc., has also motivated the fabrication of supramolecular materials.

7.4.1

Design Based on Platonic Polyhedra

The field of metal–organic coordination polyhedra has witnessed rapid growth in recent two decades, as the understanding has been gained as to the architectural principles. The concepts allow outcomes to be controlled by the stoichiometry, symmetry, and geometry of the components utilized. A variety of high-symmetry metal–organic capsules based on Platonic (faces consisting of one regular polygon) and Archimedean (faces consisting of two or more regular polygons) solids have been successfully prepared.

405

406

7 Structural Design and Rational Synthesis

7.4.1.1

Tetrahedra

According to the principles of geometric analysis, combination of the C3 -axisymmetric node with the V-shaped connecter would result in the formation of a tetrahedrally shaped supermolecule. Embonic acid (H4 L) contains two pairs of carboxylate and phenol groups to allow it with strong coordination affinity toward metal cations and has an sp3 carbon atom which is able to rotate freely to adopt desired bend angle for the potential coordination cage construction. PTC-101/102, developed by Zhang and coworkers, are built of Ti4 L6 . Four Ti4+ cations situate at the vertices and the six L4− anionic linkers define the edges of a tetrahedral geometry. There are two types of absolute configuration (viz Δ and Λ) of Ti4+ chelated by three carboxyl–phenol groups from three different ligands. Triggered by the C2 -axisymmetric embonate anion (L4− ) in a twisted form, the four coordination vertices have the same coordination stereochemistry and bring about the chirality of cages (has two isomers, ΔΔΔΔ-[Ti4 L6 ] or ΛΛΛΛ-[Ti4 L6 ]) [56]. C3 -axisymmetric oxazoline-based ligands can also be utilized to form a tetrahedral cage by coordination with lanthanide cations [57]. 7.4.1.2

Cubic Systems

Combination of the C4 -axisymmetric tetrakis-bidentate ligand (L) with the C3 -axisymmetric iron2+ tris(pyridylimine) moiety would lead to the fabrication of an O-symmetric cubic cage (formula, Fe8 L6 ) of which the corners are defined by the Fe2+ centers and the faces by the ligands. By this design strategy, a closed-face metallosupramolecular cubic cage (with six porphyrin macrocycles located on the faces) was made by Nitschke and coworkers [58]. Tetraphenylethylene (TPE) displays either clockwise or anticlockwise rotational patterns in its P or M rotational configurations, respectively, due to the steric hindrance between the phenyl rings. Through dynamic covalent chemistry, chiral organic cages can be constructed by restricting the P or M configuration of TPE faces [59]. 7.4.1.3

Octahedra

Macrocyclic molecules of Cn-alkylpyrogallol[4]arenes (which are composed of 1,2,3-trihydroxybenzene, also abbreviated as PgCn, n is for the length of the alkyl tail) have been determined to be versatile ligands for the construction of supramolecular complexes. Hong and Yuan and coworkers produced two coordination-based capsule-like quasi-isomers with distinct geometries. These two V24 capsules (a contracted octahedral capsule, octahedral {V24}, with the inner cavity of 1000 Å3 , and an expanded ball-shaped capsule, ball-like {V24}, with the inner cavity of 1400 Å3 ), are composed of the same number of subcomponents, 24 vanadium centers and 6 PgC3 units [60]. Another type of calixarenes, sulfonylcalix[4]arenes, also can react with metal ions but to form a shuttlecock-like tetrametallic cluster, which prove to be a suitable four-connected node in making huge cages by coordinating with secondary organic linkers. Cui and coworkers constructed a chiral octahedral coordination cage by using six Zn4 -p-tert-butylsulfonylcalix[4]arene clusters as tetravalent four-connected vertices and 12 enantiopure metallosalen-derived dicarboxylic acids as linear linkers, featuring a large hydrophobic cavity [61].

7.4 Polygonal/Polyhedral Complexes

7.4.1.4

Dodecahedra

Being the most intricate one of the five Platonic polyhedra, the dodecahedron contains 12 fused five-membered rings that possess the highest symmetry group I h . Its 12 pentagonal faces are built of 20 vertexes and 30 edges, so it can be prepared via edge-directed assembly from 20 tridentate angular subunits with c. 108∘ directing angles combined with 30 bidentate linear subunits [62]. Tri(4′ -pyridyl)methanol, which can be prepared from di(4-pyridyl)-ketone in high yield, has the tetrahedral directing angles, close to 108∘ . By using tri(4′ -pyridyl)methanol and linear bidentate units (e.g. bis[1,4-(trans-Pt(PPh3 )2 OTf)]ethynylbenzene, bis[4,4′ -(transPt(PR1 3 )2 OTf)]R2 , where R1 = Et, Ph; R2 = benzene, biphenyl), Stang et al. developed two dodecahedra with outer dimensions of ∼5 and ∼8 Å, respectively. As the size of these self-assembled cage structures augments, the rigidity of the individual building units becomes essential for the transmission of directing effects over larger distances. Accordingly, flexible linear linkers cause the formation of oligomers rather than discrete supermolecules with defined shape. 7.4.1.5

Icosahedra

Metal–organic capsules are mainly constructed using metal centers and organic ligands to define the symmetry axes of Platonic polyhedra, including tetrahedral, cubic, octahedral, and dodecahedral organization. Through subcomponent self-assembly, apart from a tetrahedral face-capped [Fe4 L4 ], an icosahedral capsule (with formula of [Fe12 L12 ]) was made, in which the coordinative (N → Fe) and dynamic-covalent (C—N) bonds are formed during the reaction of the 1,3,5-tri(4-aminophenyl)benzene, 2-formylpyridine, and iron triflimide in the presence of a template of cyclohexane. This method has proven a fruitful technique for the synthesis of metal–organic capsules. Unlike other Platonic cases, the symmetry axes of capsule [Fe12 L12 ] are coincident neither with its ligands nor with its metal centers. This arrangement is similar to the structures of icosahedral viral capsids, wherein the protein subunits are asymmetric, and symmetry axes are defined by linkages between capsid-forming proteins [63].

7.4.2 7.4.2.1

Design Based on Archimedean Polyhedra Cuboctahedra

The shape of a cuboctahedron could be achieved by combining tridentate planar faces with 108∘ turning angles via bidentate angular components. As 109.5∘ is close to 108∘ , the bidentate angular subunit could have a tetrahedral atom connected to a linear donor or acceptor site. Conversely, the tridentate building unit could have three donor or acceptor sites located in one plane at 120∘ relative to each other. The ratio of tridentate to bidentate ligands must be 2 : 3 with a total of 20 suitable subunits in order to obtain the closed structure. Driven by the coordination to transition metal ions of Pt2+ , two cuboctahedral capsules (characterized by nuclear magnetic resonance (NMR) and electrospray mass spectrometry) were fabricated from 8 tridentate and 12 bidentate subunits in a single step [64]. The tridentate unit, 1,3,5-tris(4′ -iodophenyl)ethynylbenzene, was synthesized from

407

408

7 Structural Design and Rational Synthesis

1,3,5-triethynylbenzene via cross-coupling; the bidentate angular donor unit with the required tetrahedral carbon, 4,4′ -bispyridylacetal, was prepared from 4,4′ -bispyridylketone. 7.4.2.2

Truncated Tetrahedra

Via a rational coordination directed self-assembly paradigm, Stang and coworkers developed various truncated tetrahedra, derived from the tritopic tris(pyridylethynyl)benzene or tris(p-cyanophenylethynyl)benzene linkers with cis-platinum and cis-palladium bistriflate salts [65]. Combination of tritopic trisplatinum and cis-bis(p-pyridyl)porphyrin would also result in the formation of a complementarily self-assembled system with the same T d symmetry. A variety of bis(diphenylphosphino)ferrocene- and bis(triethylphosphino)-substituted macrocyclic three-dimensional cage compounds were prepared. The crystals are not suitable for X-ray diffraction analysis, due to the large cage size of these supermolecules as well as the high solvent content within the crystals (as confirmed by the structural modeling). 7.4.2.3

Rhombicuboctahedra

Similar geometrical constraints have frequently been employed in artificial multicomponent self-assembly. The formation of pseudo-spherical polyhedra with a general formula of Mn L2n is predicted when metal ions with square-planar coordination sphere (M) and rigid angled ligands (L) are defined at the vertices and on the edges, respectively, of the polyhedra. For entropically favored regular or semiregular polyhedra, n is limited by geometrical constraints to be 6, 12, 24, 30, or 60 (Figure 7.3) [66]. From square-planar Pd2+ ions and bent dipyridylthiophene bridges, a giant M24 L48 spherical capsule is produced of a 72-component system which is greatly sensitive to the ligand geometry. Emergent behavior in an artificial system is on similar scale to that observed in the assembly of biological structures. In addition, the scale feasibility of using self-assembly has been extended as a powerful bottom-up technology for the construction of discrete nanoscale systems. Mapping the 48 ligands to edges and 24 metal ions to vertices, a capsule M24 L48 forms a rhombicuboctahedron, which is an Archimedean solid with 8 triangular and 18 rectangular faces, with 1 triangle and 3 rectangles meeting at each vertex.

7.4.3 7.4.3.1

Design Based on Other Shapes Prism-Like Systems

Fujita and coworkers discuss the self-assembly of large prism-like cage in which end-capped Pt2+ cations link two panel-like ligands with three pyrazine pillars. With the interplane separation (about 3.5 Å), this cage binds aromatic guests, which in turn facilitate to selectively acquire the desired cage from multicomponents. For example, triphenylene derivative efficiently templates the selective multicomponent assembly of cage of a trigonal prism. This cage is stable even when the template is removed [67].

7.4 Polygonal/Polyhedral Complexes

Rʹ

R

48 N

S

6

12

30

4

6

12

4

8

20

24 pd2+

24 N

N

12 pd2+ O

N

a: R = Rʹ = H; b: R = H, Rʹ = Br; c: R,Rʹ = –OCH2CH2O–

M6L12

M12L24

M24L48

M30L60

M60L120

Figure 7.3 Self-assembly of some representative MOPs based on inorganic nodes with organic spacers, and the family of Mn L2n polyhedra. Source: Zhang et al. [20]; Sun et al. [66].

7.4.3.2

Goldberg Polyhedra

Platonic and Archimedean solids have vastly been prepared through self-assembly, but the report of Goldberg polyhedra (their topologies have been predicted using graph theory) remains relatively little. Fujita and coworkers have made a structure consisting of a combination of 8 triangles and 24 squares, and having the symmetry of a tetravalent Goldberg polyhedron [68]. The spherical structure contains 30 palladium ions and 60 bent ligands. Notably, a small difference in the bend angles (𝜃)

409

410

7 Structural Design and Rational Synthesis

of two organic bipyridyl ligands can critically switch the resultant self-assembled structure, as has been observed for the M12 L24 -to-M24 L48 transition at around 𝜃 = 131–134∘ . 7.4.3.3

Stellated Polyhedra

Stellated polyhedra are a unique group of polyhedra with concave surfaces, formed by extending the faces of a polyhedron until they intersect. Complexation of tris(4-pyridyl) ligands (L) with Pd2+ ions results in the construction of a Pd12 L24 cuboctahedral core with a Pd:L ratio of 1 : 2. Further addition of Pd2+ ions allows for converting the cuboctahedron into a Pd18 L24 stellated cuboctahedron. The square windows of the Pd12 L24 cuboctahedron are closed by the stellation [69].

7.4.4 7.4.4.1

Linkage Modes Linkage with Pd/Pt–Nitrogen

Fujita and coworkers have used an ethylenediamine “capping” ligand to enforce the 90∘ cis geometry around square-planar coordinated Pd2+ and Pt2+ ions. The octahedral cage self-assembles by simply mixing end-capped Pd2+ /Pt2+ ions with the tridentate, triangular ligand, 1,3,5-tris(4-pyridyl)triazine, in a 6 : 4 ratio, with the six Pd2+ ions located at the corners of the octahedron. Substitution of the end-capping ligand by N,N,N ′ ,N ′ -tetramethylethylenediamine or 2,2′ -bipyridine leads to similar cages, but alters the solubility and crystallinity [70]. An inert coordination bond which becomes labile by external stimuli can be termed as a “molecular lock” since a thermodynamic equilibrium structure can be trapped (or locked) to a kinetically stable form by turning off the stimuli [71]. A thermally switchable molecular lock was developed by exploiting the dual character of a Pt2+ -pyridine coordination bonding which is inert but temporally becomes labile by thermal stimuli. 7.4.4.2

Linkage with Ga/Fe–Catechol

Raymond and coworkers developed a tetrahedral capsule by using four octahedral metal ions (e.g. Ga3+ or Fe3+ ) and six naphthalene-linked bis(bidentate) catechol ligands [72]. All the environment around the metal centers is of the same chirality, and homochiral cages are formed exclusively (ΔΔΔΔ or ΛΛΛΛ) [70]. 7.4.4.3

Linkage with Ln–Tridentate Ligand

Coordination-directed self-assembly has become a well-established technique for the construction of polyhedral structures. By contrast to the most often exploited transition metals, lanthanide cations (Ln3+ ) have been less used in the design of self-assembled structures. Sun and coworkers reported the europium coordination with the pyridine-2,6-dicarboxamide coordination unit for the construction of tetrahedrally shaped cages [73]. 7.4.4.4

Linkage with Metal–Carboxylate

The possible structures in which square units are bridged by identical angled linkers are derived from four-connected networks in which there is a planar vertex

7.5 Metal–Organic Frameworks

arrangement. In the case of polyhedral architectures, the augmentation process is usually called truncation [74]. Based on the square paddle-wheel cluster (adopted by Cu2 (CO2 )4 ), the isophthalic acid with 120∘ between carboxyl groups is ideal for building a finite truncated cuboctahedral structure. Yaghi et al. have used this design principle toward the synthesis of large discrete molecular cages (termed MOP-1). 7.4.4.5

Nodes with Transition Metal–Calixarene

Using metal–calixarene subunits as vertexes, some extra-large multicomponent cages have been constructed by Liao and coworkers [75]. A cyclic polyphenol, p-tert-butylthiacalix[4]arene (H4 TC4A), has demonstrated the ability in constructing polynuclear complexes with transition metals. The M4 /TC4A (M = Zn2+ , Co2+ , Ni2+ , Fe2+ ) shuttlecock-like moiety can be joined together to form bigger entities or extended structures by using other ligands such as carboxylates. Through a [6+8] condensation, six Co4 /TC4A units can be bridged by eight tricarboxylates into a nanosized {Co24 M6 } sphere. Analogous metal–calixarene-based cages can be generated by replacing TC4A with PTC4A (PTC4A p-tert-phenylthiacalix[4]arene). Substituting the trimesic acid with an augmented ligand has led to the formation of the discrete calixarene-based cages having large periphery diameters (up to 4.7 nm) and internal voids (up to 1.7 nm).

7.5

Metal–Organic Frameworks

MOFs are an intriguing class of hybrid materials, which exist as infinite crystalline lattices comprising inorganic vertices (metal ions or clusters) and organic struts, connected by coordination bonds of moderate strength, their structure are determined by X-ray diffraction. MOFs have attracted tremendous scientific interest due to their high crystallinity, exceptional porosity, tailorable modularity, and diverse functionality. Under the instruction of topology principle, MOFs can be represented as periodic nets, in which vertices correspond to the symmetrical interlinked molecular modules (viz SBUs). It should be noted that the initial studies of MOFs’ fabrication were mostly done by trial-and-error procedure, but the uncertainty in metal–complex chemistry could be sometimes obviated by sophisticated design of multi-dentate organic linkers. Accordingly, operation of targeting products has been well developed in the reticular chemistry of MOFs as in the organic chemistry that virtually any reasonable object can be designed and made with high precision. Much effort has been devoted to constructing ultra-stable MOFs, which are classified into two categories: (high-valency) metal–carboxylate frameworks and (low-valency) metal–azolate frameworks, which is in line with Pearson’s hard/soft acid/base concept. MOF structures are composed of two or more SBUs, which are of two general types: (i) metal-containing units that range from having single metal atoms to infinite groups (like rods) and (ii) polytopic organic spacers that may themselves incorporate metal atoms (as in metalloporphyrin). In this section, we will provide an overview of inorganic SBUs (mainly of metal carboxylates) frequently

411

412

7 Structural Design and Rational Synthesis

encountered in MOF chemistry and classify them based on their geometry and number of metals, and the linkers bearing other binding groups, such as pyridines, pyrazolates, tetrazolates, imidazolates, catecholates (CATs), pyrogallate, and their representative examples are also enumerated. Some effective synthetic strategies, which are frequently implemented to construct MOFs, including modulated synthesis, isoreticular expansion, topology-guided design, are also highlighted. In addition, concept of MTV has also been applied to the design of MOFs, resulting in the successful incorporation of many distinct linker functionalities or many metal types or different inorganic SBU geometries within one single-crystal phase. MOFs are synthesized following a large variety of synthetic strategies, including solvo/hydro/ionothermal methods (in which ILs/deep eutectic solvent [DES] could also be used as solvents), slow diffusion, and microwave-assisted process, coupled with high-throughput protocol.

7.5.1

Reticulating Chemistry (Scale Chemistry)

The knowledge on the structural topology plays an essential role in the simplification and systematization and even subsequent simulation of MOFs. The crystal structure can be reduced to an irreducible network topology in which nodes representing the inorganic vertices/clusters are linked by lines representing the organic struts. One could be aided by the computer programs, e.g. “Systre/3dt” (http://gavrog .org) and “ToposPro” (https://topospro.com), and the related databases, e.g. “RCSR” (http://rcsr.anu.edu.au), IZA-SC (http://www.iza-structure.org), and EPINET (http://epinet.anu.edu.au). The orientation of organic ligands would lead to the construction of MOFs with predetermined structural topologies. And vice versa, the geometry of inorganic SBUs would also result in the formation of MOFs with underlying periodic lattices. Overall, it is the combination of both SBUs (as connectors) and organic ligands (as linkers) that determines the final architecture. For example, a diamondoid network can be built of four-connected tetrahedral clusters and ditopic linear linkers, and a cubic net can be constructed from six-connected octahedral clusters and ditopic linear linkers (Figure 7.4) [76]. The extendibility and modifiability of ligands provide a great deal of opportunity to create isoreticular MOFs with predefined topologies and functionalities. Based on MOF-5 of a prototype constructed from oxide-centered {Zn4 O} tetrahedral clusters and linear carboxylate ligands, its three-dimensional porous system can be functionalized with the organic groups –Br, –NH2 , –OC3 H7 , –OC5 H11 , –C2 H4 , and –C4 H4 and that its pore size can also be expanded with the extended molecular struts, biphenyl, tetrahydropyrene, pyrene, and terphenyl [77]. Yaghi and coworkers have also synthesized another isoreticular series (with the same network topology) based on the skeleton of MOF-74 from its original link of 1 phenylene ring to 2, 3, 4, 5, 6, 7, 9, and 11, termed IRMOF-74-I to XI, with pore apertures ranging from 14 to 98 Å. Unlike IRMOF-1 to 16 (in which expansion of carboxylic linkers leads to interpenetration), all members of this series have non-interpenetrating structures and exhibit robust architectures [78].

7.5 Metal–Organic Frameworks

SBU

Node

Linker

MOF

MOF-5

HKUST-1

PCN-222

Figure 7.4 Graphical illustration of the construction of some representative MOFs from SBUs and rigid linkers (MOF-5, HKUST-1, PCN-222). Source: Lu et al. [76]. © 2014 Royal Society of Chemistry.

413

414

7 Structural Design and Rational Synthesis

7.5.2 7.5.2.1

Inorganic Nodes (SBUs) Mononuclear Units

ZIFs are constructed by using individual divalent and tetrahedral metal cations (e.g. Zn2+ , Co2+ ) with angular anions of imidazolates, featuring neutral zeolite-like open frameworks [79, 80]. Generated mainly by using solvothermal methods (in an amide solvent such as DMF, heated at 85–150 ∘ C, for 24–96 hours), ZIFs are based on the nets of silicate zeolites: the tetrahedral Si and the bridging O are replaced with transition metal and imidazolate link, respectively. Deprotonated IM can make an M–IM–M angle, close to 145∘ , which is coincident with the Si–O–Si angle in zeolites. Bu and Zhang and coworkers introduced presynthesized boron imidazolate complexes into zeotype systems, allowing the use of ultralight chemical elements (i.e. Li+ and B3+ ) as framework nodes, thus further increasing the diversity of ZIFs [81]. By combining MoO4 or WO4 tetrahedra with zinc–imidazolate units, a class of hybrid zeolitic frameworks (denoted HZIFs) are also developed, integrating compositional and structural features of zeolites and ZIFs [82]. HZIF-1(Mo/W) were synthesized by the self-assembly of Zn2+ cations, MoO4 2− or WO4 2− anions, and 2-methylimidazolate under solvothermal conditions. 7.5.2.2

Dinuclear Units

A highly porous cuprous–tricarboxylate coordination polymer, [Cu3 (TMA)2 (H2 O)3 ]n (also referred to HKUST-1, where TMA is benzene-1,3,5-tricarboxylate), was formed under a hydrothermal condition [83]. HKUST-1 has square paddle-wheel-shaped units, dimeric [Cu2 (O2 C)4 ], which are interconnected by tritopic TMA spacers to form a three-dimensional network with a pore size of 1 nm and an accessible porosity of about 40% in the solid. 7.5.2.3

Trinuclear Units

Oxo-bridged trimers with the formula [M3 O(O2 C)6 X3 ]n+ (X is for coordination atom, O/N/F) are common in transition metal coordination chemistry. One of the simulation-assisted chemical structures, MIL-100 [84], is composed of trinuclear metal clusters (chelated by three carboxylic functions, having trigonal-prismatic geometry). The combination of chromium salts with benzene-1,3,5-tricarboxylate under hydrothermal conditions (in the presence of fluorhydric acid, at 220 ∘ C, for eight hours) leads to a microcrystalline green solid of MIL-100. By targeted chemistry and computational design, the trimers of Cr3+ octahedra are bridged by tritopic carboxylates to be a supertetrahedral cage, which are further connected into a zeotype structure of a huge unit cell (∼702 000 Å3 ) and a hierarchy of extra-large pore sizes (∼30–34 Å). 7.5.2.4

Tetranuclear Units

Combination of Zn2+ and the carboxylic group could lead to the formation of oxide-centered clusters as a distinct and well-defined tetranuclear unit, which can be linked via appropriate multi-carboxylate spacers to form a net, as in MOF-5 [85]. Diffusion of triethylamine into a solution of zinc nitrate and terephthalic acid

7.5 Metal–Organic Frameworks

(H2 BDC) in DMF/chlorobenzene resulted in the deprotonation of H2 BDC and its reaction with Zn2+ ions. A small amount of hydrogen peroxide was also added to the reaction mixture in order to facilitate the formation of O2− expected at the center of the tetrahedral unit. 7.5.2.5

Hexanuclear Units

Octahedrally shaped cluster of {Zr6 O4 (OH)4 } is the most frequently observed in Zr-SBUs, in which the six vertices of an octahedron are occupied by Zr4+ centers and eight triangular faces are capped by eight O2− /OH− groups. Each Zr4+ is eight-coordinated by O atoms in a square-antiprismatic coordination geometry, of which one square is formed by four O atoms supplied by carboxylate groups. When the {Zr6 O4 (OH)4 } cluster is fully coordinated by 12 carboxylate groups, an SBU of {Zr6 O4 (OH)4 (CO2 )12 } is generated in an Oh -symmetry, which is observed in UiO-66–68 by Lillerud and coworkers [86]. Products of UiO-66–68 crystallize as cubic microcrystals which are too small for structure determination by single-crystal X-ray diffraction. By varying the ratio of reactants and the amount of modulating reagents, Zhou and coworkers have obtained a series of porphyrinic zirconium MOFs, denoted PCN-222–224 based on a {Zr6 } cluster but with different symmetry (from Oh to D4h and then to D3d ) and reduced connectivity (from 12 to 8 and then to 6) [87]. Among them, PCN-222 contains open hexagonal channels, with a diameter of up to 3.7 nm and remains robust even in concentrated hydrochloric acid. PCN-222 can be synthesized by a solvothermal reaction of tetrakis(carboxylphenyl)porphyrin, ZrCl4 , and benzoic acid in DMF for 48 hours at 120 ∘ C [88]. Rare-earth metal ions (RE3+ ) can also form octahedral {RE6 } with the same structure and geometry as {Zr6 }. Under the guidance of thermodynamics, mixed-component RE-MOFs have been developed in the solvothermal reactions with the presence of 2-fluorobenzoic acid [89]. 7.5.2.6

Octanuclear Units

Using an appropriate choice of solvent mixtures (DMF and methanol), white solid (denoted MIL-125) [90] was isolated under the solvothermal reaction of titanium tetraisopropoxide and 1,4-benzene dicarboxylates (BDC) at 150 ∘ C. The quasi-cubic tetragonal structure can be viewed as an augmented version of the centered cubic structure in which the nodes are substituted by cyclic octamers (wheel-like {Ti8 O8 (OH)4 }). Linear dicarboxylate (BDC) replace the lines between atoms, link the octamers, and provide a 3D network. 7.5.2.7

Mixing SBUs

MOFs with two or more different SBUs are relatively scarce, and most of them reported involve dissymmetric multitopic linkers. The high-symmetry trimesic (BTC) linker indicates to be quite versatile in the combination with multiple SBUs in a single MOF. From C3h to C2v , biphenyl-3,4′ ,5-tricarboxylate (bhtc) of less symmetrical linkers [91] could drive hetero-SBUs to satisfy the constraints imposed by reduced symmetric linker geometries. UMCM-150 is the forerunner of such an MOF family with two distinct SBUs: a dicopper {Cu2 (CO2 )4 } paddle-wheel

415

416

7 Structural Design and Rational Synthesis

and a tricopper {Cu3 (CO2 )6 } trigonal-prism with four and six points of extension, respectively [92]. The two SBUs are interconnected by a C2v symmetry linker, bhtc, in which three carboxylate groups are not related symmetrically with C3h . The carboxylates at the isophthalate moiety of the linker form dicopper {Cu2 (CO2 )4 } clusters and the carboxylates at the benzoate moiety form tricopper {Cu3 (CO2 )6 } cluster. Accordingly, two different SBUs coexist in a crystal, and the underlying topology of UMCM-150 is an unusual (3,4,6)-connected net. 7.5.2.8

Rod-Shaped Chains

Some basic nets can be derived from linking chains (or helices and ladders). Rod-shaped metal–carboxylate SBUs provide means to accessing MOFs of which the nets do not interpenetrate due to the intrinsic arrangement of such rods in the crystal, as in MOF-74 with C3 -helical SBUs [93]. The rods of MOF-74 consist of edge-sharing transition metal octahedra alternately opposite and next-neighbor to form either a 31 or a 32 helix. A series of porphyrin-based zirconium–phenolate frameworks, featuring infinite ZrIV -oxo chains and eclipse-arranged porphyrin macrocycles, are also successfully constructed through a top–down process from simulation to synthesis. These are the unusual examples of Zr-MOFs (or MOFs in general) based on porphyrinic phenolates, instead of commonly known carboxylate-based types [94].

7.5.3

Organic Linkers with Carboxylates

The geometry of a ligand dictates the structure of the resulting outcome. Adjustments of the connectivity, length, ratio, and orientation of the ligand can tune the size, shape, and internal surface feature of an MOF. Monocarboxylates can be used to build Zr4+ -MOFs, not being as linkers but as modulators to control crystal growth in solvothermal reaction. 7.5.3.1

Dicarboxylates

Ditopic carboxylate linkers have been well studied due partly to their ready accessibility and partly to their easily predictable structures in combination with different SBUs. Reaction between zinc nitrate and terephthalate (BDC) gives single crystals of MOF-5 (with the formula of {[Zn4 O(BDC)3 ]}) under solvothermal conditions. A hydrothermal reaction of BDC with chromium nitrate, fluorhydric acid, and water at 220 ∘ C can produce green microcrystals of MIL-101 with the formula {Cr3 O(BDC)3 F(H2 O)2 }. UiO-66 is prepared under solvothermal conditions using ZrCl4 , BDC, and DMF as the solvent. Besides UiOs, MIL-125 is another known highly porous MOF constructed from 12-connected metal–oxo–hydroxo clusters ({Ti8 O8 (OH)4 }) and carboxylate linkers (BDC). Tetra-anionic 2,5-dioxido-1,4-benzene-dicarboxylate (DOBDC) is known to be the organic spacer for MOF-74, in which both the aryloxide and carboxylate groups are bonded to the metal sites [76].

7.5 Metal–Organic Frameworks

7.5.3.2

Tricarboxylates

Among tritopic ligands, BTC3− (viz deprotonated 1,3,5-benzenetricarboxylic acid) is widely used in MOF construction, such as HKUST-1 (Cu-BTC), CPM-5 (In-BTC), MIL-100 (Cr/Fe-BTC). BTC3− is also applicable for coordination with Zr4+ , resulting in the formation of a (3,6)-connected structure with spn topology, MOF-808 with the formula of {Zr6 (O)4 (OH)4 (BTC)2 (O2 CH)6 }. MOF-808 crystallizes in the cubic space group, in which each {Zr6 } SBU is connected to six BTC3− linkers and each BTC3− linker coordinates with three {Zr6 } SBUs [95]. 7.5.3.3

Tetracarboxylates

Compared with di- and tricarboxylate linkers, tetracarboxylate linkers have more coordination sites and richer coordination modes. More importantly, tetracarboxylic ligands of large-space steric hindrance could effectively avoid net interpenetration, being favorable for constructing substantially porous MOFs. Among all tetracarboxylic ligands for constructing MOFs, tetrakis(4-carboxylphenyl)porphyrin (H4 TCPP) is quite common, as in the well-known MOF-525/PCN-222 [88]. Tetratopic carboxylate linkers, with tetrahedral T d geometry, are also favored for the construction of 3D frameworks with wide channels and/or large pores [95]. 7.5.3.4

Hexacarboxylates

By using hexacarboxylate linkers in a C3 -symmetry, e.g. 3,3′ ,3′′ ,5,5′ ,5′′ -benzene-1, 3,5-triyl-hexabenzoate (BHB), and dicopper paddle-wheel SBUs, a (3,24)-connected network of rht topology (UTSA-20, if serving the cuboctahedral cages, {[dicopper]24 [isophthalate]48 }, as a 24-connection node) is made. The resulting structure of UTSA-20 (Cu-BHB) can also be viewed as a (3,3,4)-connected zyg-a (where “a” is for augmented version) topology, with the branching points on the BHB linker as nodes [96]. 7.5.3.5

Octacarboxylates

It has been well known that the networks based on linkers with long arms tend to constitute interpenetrated structures, but polytopic linkers could effectively prevent the interpenetration possibly due to their high connectivity. The reaction of a octacarboxylic acid (with the porphyrin ring as the core branching out 4 V-shaped arms) and dizinc paddle-wheel results into the isolation of UNLPF-1 [97]. UNLPF-1 displays the scu topology by considering the ligands as an eight-connected node and the dicopper paddle-wheel clusters as a four-connected node. 7.5.3.6

Mixing Ligands

Multivariate metal–organic frameworks (MTV-MOFs) are constructed from ligands with different pendant functional groups whose orientation, position, number, and ratio are controlled by virtue of its unaltered connectivity. Different interactions between linkers can lead to different distribution of functional groups [98]. Two linkers that are coordinatively identical but distinct in shape could build unique MOF structures. BDC yields MOF-5, while the extended BTB affords MOF-177 under essentially same synthetic environments. Combining BDC and BTB in the

417

418

7 Structural Design and Rational Synthesis

presence of Zn2+ results in a completely different MOF, UMCM-1. A systematic study (ditopic linker plus tritopic linker plus {ZnO4 (CO2 )6 } cluster) indicated that both the geometric factor (length ratio between two linkers) and the statistical factor (mole fraction of two linkers) play key roles in determining the underlying MOF structures. Slight changes in these two factors could lead to significantly different MOF structures [99].

7.5.4 7.5.4.1

Organic Linkers with Other Functional Groups Other Oxygen-Containing Ligands

Phosphonates can form stronger bonds than carboxylates with metal atoms. Low-valent metal phosphonates are soluble, as the valence of the metal increases, the solubility decreases. Mono- and divalent metal phosphonates are soluble enough that single crystals can be obtained by solvothermal techniques, whereas trivalent and tetravalent metal phosphonates are highly insoluble and tend to precipitate as powder products [100]. Compared with phosphonates, phenolates have the more specific coordination modes, which could also be used to construct MOFs. These MOFs are much more stable because of the higher pK a of the phenolate group than that of the carboxylate-based MOFs. With ditopic phenolic acid, H6 TzGal, as the starting ligand resource, Devic and coworkers fabricated a highly stable porous Zr-MOF, Zr(H2 TzGal) (MIL-163 built of 1D coordination chain constituted from Zr4+ and 1,2,3-trioxobenzene moieties) [101]. Yaghi and coworkers also developed a series of 3D metal–CATs of srs net. As a representative example, Fe-CAT-5 was prepared by dissolving iron sulfate, H6 THO (= triphenylene-2,3,6,7,10,11-hexakis(olate)), tetrabutylammonium nitrate, and amylamine in a solvent mixture of DMF/water/methanol in a Teflon vessel, at 180 ∘ C for 48 hours [102]. 7.5.4.2

Nitrogen–Heterocyclic Linkers

It is well known that the bonding strength of transition metal–nitrogen is greater than that of transition metal–oxygen in solution. Organic linkers containing one or more nitrogen donors, such as pyridine and azole derivatives, have been extensively studied aiming to achieve MOFs. Di- or tetrapyridyl linkers have been largely employed as spacers to connect 1D transition metal–silicofluoride chains into 3D frameworks [103]. Similar to carboxylic acids, imidazole, pyrazole, triazole, and tetrazole are usually deprotonated to coordinate with the metal cations [104]. Being one of the well-known classes of MOFs, transition metal–ZIFs have been extensively explored as the fact that the bridging angle between imidazolate and two bonding metals is nearly equal to that of Si–O–Si in zeolites. Pyrazolate has a bridging angle (c. 70∘ ), of which the coordination geometry is quite similar to the bidentate fashion of carboxylates, which is the basis of paddle-wheel-shaped, trigonal-prismatic, and octahedral SBUs commonly encountered in MOFs. There are also some typical SBUs based on pyrazolates, as octahedral [M4 (O)(pyrazolate)6 ], cubic [M8 (OH)6 (pyrazolate)12 ], triangular [M3 (O)(pyrazolate)3 ], double/triple zigzag chains. From a topological point of view, the cubic [Ni8 (OH)6 (pyrazolate)12 ]

7.5 Metal–Organic Frameworks

cluster is similar to the well-known {Zr6 O4 (OH)4 (O2 C)12 } cluster, which can also be a 12-connected node. PCN-601 and 602 are composed of {Ni8 }-clusters and pyrazolate-based porphyrinic ligands, adopting an ftw-a topology featuring large cubic cages within the frameworks [105, 106]. 7.5.4.3

Mixing Hetero-Linkers

Pillared-layer MOFs are one of the well-studied mixed-ligand structures. Polycarboxylates coordinate with {Zn2 } or {Cu2 } paddle-wheel-shaped units to be 2D sheets, and the pyridyl derivatives (mainly linear bitopic) are often used as pillars to link 2D sheets into layered structures, allowing diverse functionalities to be incorporated into MOFs [107].

7.5.5 7.5.5.1

Synthetic Strategies Solvothermal Reactions

Solvothermal reactions (which involve loading organic ligands, metal salts, and solvents in a Teflon-lined stainless-steel Parr autoclave, or a thick-wall glass tube, heating at elevated temperatures) have proven to be an effective synthetic strategy for the formation of MOFs with high crystallinity. The solvothermal synthesis is actually a Lewis acid–base reaction, in which the ligands deprotonated serve as Lewis bases, while the metal ions act as the Lewis acids. Modifying the solvothermal reaction conditions for matching the rate of ligand deprotonation with that of the coordination bond formation is crucial. Solvents such as DMF, diethyl formamide (DEF), and dimethyl acetamide (DMA) tend to undergo hydrolysis in the temperature range of 60–85 ∘ C, discharging amines capable of deprotonating the carboxylic acid to facilitate reaction. Without being decomposed, dimethyl sulfoxide (DMSO) can be used at higher temperature (100 ∘ C or above), which helps to overcome the energy barrier for some demanding MOFs. If necessary, the pH and polarity of the reaction solvents can be adjusted by adding other species, like inorganic acid or organic amine [108]. 7.5.5.2

Template-Directed Synthesis

Templates can provide control over the both the structure and the functionality of metal–organic materials. There are many kinds of templates, according to their nature, including (i) solvents (e.g. low-boiling-point solvents, such as diethyl ether, dichloromethane, and acetone, which are easily evacuated for use in activating samples), (ii) organic compounds (e.g. serving as SDAs, organic amines, and ILs, which have been utilized as solvents/templates in zeolites and some MOFs and metal–chalcogenide systems), (iii) coordination complexes, (iv) inorganic compounds (e.g. POMs, of which incorporation in MOFs to form POMs@MOFs hybrid), (v) gas molecules, and (vi) surfactants [109]. 7.5.5.3

Seed-Mediated Approach

Zr-MOFs are all nearly obtained from ZrCl4 and carboxylate under a solvothermal condition, often accompanying other phases throughout the synthesis procedure. Typical examples of syn-crystallized MOF pairs are PCN-222 and PCN-224, PCN-222

419

420

7 Structural Design and Rational Synthesis

and PCN-225. These Zr-MOFs have similar density and thus are difficultly separated via the solvent-induced method. As stated by Zhou and coworkers, all these two pairs of MOFs can be prepared in phase-pure forms via a seed-mediated approach. Introducing crystal seeds of a targeted MOF in the reaction system brings significant purity improvement in the resultant product. 7.5.5.4

Microwave-Assisted Synthesis

The crystallization of most MOFs through liquid diffusion or solvothermal condition need to take more than one day or even several weeks. Accordingly, microwave-assisted processes have been used to produce MOFs, allowing the synthesis to be completed in minutes. Typically, the reaction is carried out in a Pyrex sample vial, placed inside a hood behind a blast shield and heated by a microwave synthesizer [110]. Although the microwave heating method usually results in nanosized particles of MOFs, their crystal size and shape can be well controlled and the synthesis cycle can be largely shortened. 7.5.5.5

High-Throughput Methods

Traditional synthesis used to discover the original crystalline solids was tedious, unpredictable, time-consuming, and often a waste of solvents and reagents. High-throughput methods could be applied in screening for reaction conditions for the construction of new crystalline coordination compounds. For a multicomponent chemical organization, such as porous metal–organic network, it would be reasonable to assume that the most energetically favored structure would result [111].

7.5.6 7.5.6.1

Others Tailor-Made Approach

Automated assembly of SBUs is adapted to MOFs, exploring how an inorganic moiety and an organic linker, or even predefined synthons, may interconnect in 3D space to form periodic lattices. A virtual library of hypothetical frameworks has been produced, along with their crystallographic data (space group, cell parameters, atomic coordinates) and their simulated X-ray diffraction patterns. The comparison of the experimental pattern with the simulated one from each candidate structure determines the targeted structure without any recourse to single crystals. The final structure is refined with the Rietveld method from powder data. Based on structure–property insights, Snurr and coworkers identified and synthesized a MOF, NOTT-107, which is predicted to have better methane-storage capacity at 35 bar than that of PCN-14 [112, 113]. 7.5.6.2

Design Based on MOPs

Lots of MOPs are targetable and potentially can be utilized as SBUs to constitute three-periodic MOFs. As exemplified by MIL-100/101, the utilization of supertetrahedral synthons for building MOFs can give access to a zeotype of MTN topology.

References

7.5.6.3

Merged Nets Approach

Eddaoudi and coworkers introduced a concept of merged nets based on merging two edge-transitive nets into a minimal edge-transitive net for the deliberate construction of complex mixed-linker MOFs. By combining two edge-transitive nets, a (3,6)-coordinated spn and 6-coordinated hxg, a (3,6,12)-coordinated merged net is formed of sph topology [114].

7.6

Conclusion

We have given a brief survey of recent progress in the design and synthesis of POM/chalcogenidometalate-based clusters and open frameworks, coordination polyhedra, and metal–organic frameworks to show the knowledge on their assembly mechanism and procedures. First, the synthetic approaches to the many POM types have been presented, including iso/hetero-POMs and POM–organic hybrids. Next, the isolation strategies of crystallographically defined supertetrahedral chalcogenidometalate clusters have been summarized, most of which are allowed to be fabricated by the charge-balancing principle. Then, the developments of coordination-driven self-assembly of discrete metal–organic architectures have been outlined through well-defined symmetry-adapted and directional-bonding assembly routes. Finally, some synthetic strategies for building periodic MOFs have been provided, including modulated and topology-guided design (based SBUs of inorganic nodes and organic linkers, whose geometry dictates the structure of the resulting outcome), isoreticular expansion, MTV chemistry, etc.

Acknowledgments We thank the supports from the National Science Foundation of China (Grant No. 21501028), the Strategic Priority Research Program of Chinese Academy of Sciences (Grant No. XDB20000000), the Hundred-Talent Program of Fujian Province and the Hundred-Talent Program of Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences.

References 1 Long, D.-L., Abbas, H., Kogerler, P., and Cronin, L. (2004). J. Am. Chem. Soc. 126: 13880. 2 Muller, A., Beckmann, E., Bogge, H. et al. (2002). Angew. Chem. Int. Ed. 41: 1162. 3 Kortz, U., Savelieff, M.G., Bassil, B.S., and Dickman, M.H. (2001). Angew. Chem. Int. Ed. 40: 3384. 4 Li, Z., Lin, L.-D., Yu, H. et al. (2018). Angew. Chem. Int. Ed. 57: 15777. 5 Zang, H.-Y., Miras, H.N., Long, D.-L. et al. (2013). Angew. Chem. Int. Ed. 52: 6903.

421

422

7 Structural Design and Rational Synthesis

6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42

Jin, L., Zhu, Z.-K., Wu, Y.-L. et al. (2017). Angew. Chem. Int. Ed. 56: 16288. Han, X.-B., Zhang, Z.-M., Zhang, T. et al. (2014). J. Am. Chem. Soc. 136: 5359. Mitchell, S.G., Streb, C., Miras, H.N. et al. (2010). Nat. Chem. 2: 308. Miras, H.N., Vilà-Nadal, L., and Cronin, L. (2014). Chem. Soc. Rev. 43: 5679. AlDamen, M.A., Clemente-Juan, J.M., Coronado, E. et al. (2008). J. Am. Chem. Soc. 130: 8874. Ritchie, C., Moore, E.G., Speldrich, M. et al. (2010). Angew. Chem. Int. Ed. 49: 7702. Hussain, F., Conrad, F., and Patzke, G.R. (2009). Angew. Chem. Int. Ed. 48: 9088. Jin, L., Li, X.-X., Qi, Y.-J. et al. (2016). Angew. Chem. Int. Ed. 55: 13793. Ibrahim, M., Mereacre, V., Leblanc, N. et al. (2015). Angew. Chem. Int. Ed. 54: 15574. Xu, B.B., Wei, Y.G., Barnes, C.L., and Peng, Z.H. (2001). Angew. Chem. Int. Ed. 40: 2290. Xiao, F., Hao, J., Zhang, J. et al. (2010). J. Am. Chem. Soc. 132: 5956. Zhu, Y., Yin, P., Xiao, F. et al. (2013). J. Am. Chem. Soc. 135: 17155. Huang, Q., Wei, T., Zhang, M. et al. (2017). J. Mater. Chem. A 5: 8477. Zheng, S.-T., Zhang, J., Li, X.-X. et al. (2010). J. Am. Chem. Soc. 132: 15102. Zhang, Y., Gan, H., Qin, C. et al. (2018). J. Am. Chem. Soc. 140: 17365. Gong, Y., Zhang, Y., Qin, C. et al. (2019). Angew. Chem. Int. Ed. 58: 780. Xu, N., Gan, H., Qin, C. et al. (2019). Angew. Chem. Int. Ed. 58: 4649. An, H.-Y., Wang, E.-B., Xiao, D.-R. et al. (2006). Angew. Chem. Int. Ed. 45: 904. Zheng, S.-T., Zhang, J., and Yang, G.-Y. (2008). Angew. Chem. Int. Ed. 47: 3909. Qin, J.-S., Du, D.-Y., Guan, W. et al. (2015). J. Am. Chem. Soc. 137: 7169. Wang, X.-L., Qin, C., Wang, E.-B. et al. (2006). Angew. Chem. Int. Ed. 45: 7411. Tsunashima, R., Long, D.-L., Miras, H.N. et al. (2010). Angew. Chem. Int. Ed. 49: 113. Huang, P., Qin, C., Su, Z.-M. et al. (2012). J. Am. Chem. Soc. 134: 14004. Miras, H.N., Cooper, G.J.T., Long, D.-L. et al. (2010). Science 327: 72. Zheng, S.-T., Zhang, J., Clemente-Juan, J.M. et al. (2009). Angew. Chem. Int. Ed. 48: 7176. Li, Z., Li, X.-X., Yang, T. et al. (2017). Angew. Chem. Int. Ed. 56: 2664. de la Oliva, A.R., Sans, V., Miras, H.N. et al. (2012). Angew. Chem. Int. Ed. 51: 12759. Wu, T., Wang, X., Bu, X. et al. (2009). Angew. Chem. Int. Ed. 48: 7204. Xu, X.F., Wang, W., Liu, D.L. et al. (2018). J. Am. Chem. Soc. 140: 888. Zheng, N., Bu, X., Lu, H. et al. (2005). J. Am. Chem. Soc. 127: 11963. Zhang, Q.C., Liu, Y., Bu, X.H. et al. (2008). Angew. Chem. Int. Ed. 47: 113. Wu, T., Bu, X., Liao, P. et al. (2012). J. Am. Chem. Soc. 134: 3619. Xiong, W.-W., Li, J.-R., Hu, B. et al. (2012). Chem. Sci. 3: 1200. Zheng, N.F., Bu, X.H., and Feng, P.Y. (2004). Angew. Chem. Int. Ed. 43: 4753. Wang, L., Wu, T., Zuo, F. et al. (2010). J. Am. Chem. Soc. 132: 3283. Zheng, N., Haiwei Lu, X.B., and Feng, P. (2006). J. Am. Chem. Soc. 128: 4528. Feng, P., Bu, X., and Zheng, N. (2005). Acc. Chem. Res. 38: 293.

References

43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78

Zheng, N.F., Bu, X.G., Wang, B., and Feng, P.Y. (2002). Science 298: 2366. Lin, Q., Bu, X., Mao, C. et al. (2015). J. Am. Chem. Soc. 137: 6184. Wu, T., Wang, L., Bu, X.H. et al. (2010). J. Am. Chem. Soc. 132: 10823. Santner, S., Heine, J., and Dehnen, S. (2016). Angew. Chem. Int. Ed. 55: 876. Xiong, W.W., Athresh, E.U., Ng, Y.T. et al. (2013). J. Am. Chem. Soc. 135: 1256. Xiong, W.W., Miao, J.W., Ye, K.Q. et al. (2015). Angew. Chem. Int. Ed. 54: 546. Xiong, W.-W. and Zhang, Q. (2015). Angew. Chem. Int. Ed. 54: 11616. Zhang, X.-M., Sarma, D., Wu, Y.-Q. et al. (2016). J. Am. Chem. Soc. 138: 5543. Huang, S.L., He, L., Chen, E.X. et al. (2019). Chem. Commun. 55: 11083. Yang, H., Zhang, J., Luo, M. et al. (2018). J. Am. Chem. Soc. 140: 11189. Wu, T., Khazhakyan, R., Wang, L. et al. (2011). Angew. Chem. Int. Ed. 50: 2536. Zhang, Q.C., Bu, X.H., Zhang, J. et al. (2007). J. Am. Chem. Soc. 129: 8412. Lin, J., Zhang, Q., Wang, L. et al. (2014). J. Am. Chem. Soc. 136: 4769. He, Y.-P., Yuan, L.-B., Chen, G.-H. et al. (2017). J. Am. Chem. Soc. 139: 16845. Liu, C.-L., Zhang, R.-L., Lin, C.-S. et al. (2017). J. Am. Chem. Soc. 139: 12474. Meng, W., Breiner, B., Rissanen, K. et al. (2011). Angew. Chem. Int. Ed. 50: 3479. Qu, H., Wang, Y., Li, Z. et al. (2017). J. Am. Chem. Soc. 139: 18142. Su, K., Wu, M., Yuan, D., and Hong, M. (2018). Nat. Commun. 9: 4941. Tan, C., Jiao, J., Li, Z. et al. (2018). Angew. Chem. Int. Ed. 57: 2085. Olenyuk, B., Levin, M.D., Whiteford, J.A. et al. (1999). J. Am. Chem. Soc. 121: 10434. Bilbeisi, R.A., Ronson, T.K., and Nitschke, J.R. (2013). Angew. Chem. Int. Ed. 52: 9027. Olenyuk, B., Whiteford, J.A., Fechtenkötter, A., and Stang, P.J. (1999). Nature 398: 796. Leininger, S., Fan, J., Schmitz, M., and Stang, P.J. (2000). Proc. Natl. Acad. Sci. U.S.A. 97: 1380. Sun, Q.-F., Iwasa, J., Ogawa, D. et al. (2010). Science 328: 1144. Kumazawa, K., Biradha, K., Kusukawa, T. et al. (2003). Angew. Chem. Int. Ed. 42: 3909. Fujita, D., Ueda, Y., Sato, S. et al. (2016). Nature 540: 563. Sun, Q.-F., Sato, S., and Fujita, M. (2012). Nat. Chem. 4: 330. Yoshizawa, M., Klosterman, J.K., and Fujita, M. (2009). Angew. Chem. Int. Ed. 48: 3418. Ibukuro, F., Kusukawa, T., and Fujita, M. (1998). J. Am. Chem. Soc. 120: 8561. Caulder, D.L., Powers, R.E., Parac, T.N., and Raymond, K.N. (1998). Angew. Chem. Int. Ed. 37: 1840. Yan, L.-L., Tan, C.-H., Zhang, G.-L. et al. (2015). J. Am. Chem. Soc. 137: 8550. Eddaoudi, M., Kim, J., Wachter, J.B. et al. (2001). J. Am. Chem. Soc. 123: 4368. Liu, M., Liao, W., Hu, C. et al. (2012). Angew. Chem. Int. Ed. 51: 1585. Lu, W., Wei, Z., Gu, Z.-Y. et al. (2014). Chem. Soc. Rev. 43: 5561. Eddaoudi, M., Kim, J., Rosi, N. et al. (2002). Science 295: 469. Deng, H., Grunder, S., Cordova, K.E. et al. (2012). Science 336: 1018.

423

424

7 Structural Design and Rational Synthesis

79 Huang, X.-C., Lin, Y.-Y., Zhang, J.-P., and Chen, X.-M. (2006). Angew. Chem. Int. Ed. 45: 1557. 80 Park, K.S., Ni, Z., Côté, A.P. et al. (2006). Proc. Natl. Acad. Sci. U.S.A. 103: 10186. 81 Zhang, J., Wu, T., Zhou, C. et al. (2009). Angew. Chem. Int. Ed. 48: 2542. 82 Wang, F., Liu, Z.S., Yang, H. et al. (2011). Angew. Chem. Int. Ed. 50: 450. 83 Chui, S.S.Y., Lo, S.M.F., Charmant, J.P.H. et al. (1999). Science 283: 1148. 84 Férey, G., Mellot-Draznieks, C., Serre, C. et al. (2005). Science 309: 2040. 85 Li, H., Eddaoudi, M., O’Keeffe, M., and Yaghi, O.M. (1999). Nature 402: 276. 86 Cavka, J.H., Jakobsen, S., Olsbye, U. et al. (2008). J. Am. Chem. Soc. 130: 13850. 87 Feng, D., Chung, W.-C., Wei, Z. et al. (2013). J. Am. Chem. Soc. 135: 17105. 88 Feng, D., Gu, Z.-Y., Li, J.-R. et al. (2012). Angew. Chem. Int. Ed. 51: 10307. 89 Zhang, L., Yuan, S., Feng, L. et al. (2018). Angew. Chem. Int. Ed. 57: 5095. 90 Dan-Hardi, M., Serre, C., Frot, T. et al. (2009). J. Am. Chem. Soc. 131: 10857. 91 Schnobrich, J.K., Lebel, O., Cychosz, K.A. et al. (2010). J. Am. Chem. Soc. 132: 13941. 92 Wong-Foy, A.G., Lebel, O., and Matzger, A.J. (2007). J. Am. Chem. Soc. 129: 15740. 93 Rosi, N.L., Kim, J., Eddaoudi, M. et al. (2005). J. Am. Chem. Soc. 127: 1504. 94 Chen, E.-X., Qiu, M., Zhang, Y.-F. et al. (2018). Adv. Mater. 30: 1704388. 95 Bai, Y., Dou, Y., Xie, L.-H. et al. (2016). Chem. Soc. Rev. 45: 2327. 96 Guo, Z., Wu, H., Srinivas, G. et al. (2011). Angew. Chem. Int. Ed. 50: 3178. 97 Johnson, J.A., Lin, Q., Wu, L.C. et al. (2013). Chem. Commun. 49: 2828. 98 Deng, H., Doonan, C.J., Furukawa, H. et al. (2010). Science 327: 846. 99 Koh, K., Wong-Foy, A.G., and Matzger, A.J. (2010). J. Am. Chem. Soc. 132: 15005. 100 Gagnon, K.J., Perry, H.P., and Clearfield, A. (2012). Chem. Rev. 112: 1034. 101 Mouchaham, G., Cooper, L., Guillou, N. et al. (2015). Angew. Chem. Int. Ed. 54: 13495. 102 Nguyen, N.T.T., Furukawa, H., Gándara, F. et al. (2015). J. Am. Chem. Soc. 137: 15394. 103 Nugent, P., Belmabkhout, Y., Burd, S.D. et al. (2013). Nature 495: 80. 104 Yuan, S., Feng, L., Wang, K. et al. (2018). Adv. Mater. 30: 1704303. 105 Wang, K., Lv, X.-L., Feng, D. et al. (2016). J. Am. Chem. Soc. 138: 914. 106 Lv, X.-L., Wang, K., Wang, B. et al. (2017). J. Am. Chem. Soc. 139: 211. 107 Farha, O.K., Malliakas, C.D., Kanatzidis, M.G., and Hupp, J.T. (2010). J. Am. Chem. Soc. 132: 950. 108 Zhao, D., Timmons, D.J., Yuan, D., and Zhou, H.-C. (2011). Acc. Chem. Res. 44: 123. 109 Zhang, Z. and Zaworotko, M.J. (2014). Chem. Soc. Rev. 43: 5444. 110 Ni, Z. and Masel, R.I. (2006). J. Am. Chem. Soc. 128: 12394. 111 Banerjee, R., Phan, A., Wang, B. et al. (2008). Science 319: 939. 112 Wilmer, C.E., Leaf, M., Lee, C.Y. et al. (2011). Nat. Chem. 4: 83. 113 Ma, S., Sun, D., Simmons, J.M. et al. (2008). J. Am. Chem. Soc. 130: 1012. 114 Jiang, H., Jia, J., Shkurenko, A. et al. (2018). J. Am. Chem. Soc. 140: 8858.

425

8 Structural Topologies and Interpenetration in the Coordination Polymers Fei Wang, Hai-Xia Zhang, and Jian Zhang State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, 155 Yangqiao Road West, Fuzhou, 350002, Fujian, P.R. China

8.1 Introduction Over the past decade, crystal engineering based on supramolecular chemistry and metal coordination polymers has developed rapidly with the increase of the number of structures included in the small molecular crystal structure database (CSD, Cambridge Structure Data Base) and biological macromolecules CSD (Protein Data Bank, PDB). They have provided new concepts, methods, and paths for chemical science and made it possible to self-assemble and construct metal complexes with predefined structures and even expected functions, and even molecular-based materials [1–6]. Coordination chemistry has gradually gone beyond the scope of inorganic chemistry and began to infiltrate into other chemical branches, as well as physical science, life science, material science, environmental science, and other fields. Molecular-based metal functional complexes, which are easy to design and assemble by molecular tailoring, show potential applications in molecular and ion exchange, adsorption and selective catalysis, optoelectronics and magnetic materials, etc., and have become one of the hotspots of chemists, physicists, and biologists [1].

8.2 Supramolecular Assembly and Reticular Synthesis J.-M. Lehn defines supramolecular chemistry as “chemistry beyond the molecule.” It is the science of studying complex and orderly molecular aggregates with specific functions formed by intermolecular interactions. These molecular aggregates are referred to as supramolecules. Molecular recognition is the core concept of supramolecular chemistry [5]. The so-called molecular recognition is a process in which the organism (or receptor) selectively binds to the guest (or substrate) and produces a specific function. One molecule can be identified by the shape and size of another molecule or by chemical factors such as hydrogen bond formation, stacking

Advanced Structural Chemistry: Tailoring Properties of Inorganic Materials and their Applications, First Edition. Edited by Rong Cao. © 2021 WILEY-VCH GmbH. Published 2021 by WILEY-VCH GmbH.

426

8 Structural Topologies and Interpenetration in the Coordination Polymers

interaction, and electrostatic force. Supramolecules represent the complexity of matter at the next level after elementary particles, nuclei, atoms, and molecules. Self-assembly and self-organization are important tools in supramolecular research. With a large number of reports on crystal structure and in-depth theoretical research, the understanding of intermolecular interactions has shifted from the traditional strong hydrogen bonding (X—H⋅⋅⋅Y), various types of van der Waals force to weak hydrogen bonding (e.g. C—H⋅⋅⋅X, X—H⋅⋅⋅M), π–π stacking, C—H⋅⋅⋅π interaction, etc. Understanding and controlling these new intermolecular interactions is undoubtedly an indispensable part of the molecular design process. In recent years, with the use of a series of soft chemical synthesis methods, such as hydrothermal (solvothermal) methods, metal–organic coordination polymers have developed vigorously. The design and synthesis of metal–organic frameworks (MOFs) have gradually tended to be practical and theoretical [7–11]. In 2003, Yaghi et al. first put forward the concept of “reticular synthesis” in Nature [12]. The so-called “reticular synthesis” refers to the process of assembling well-designed rigid molecular building units into predetermined regular structures (or networks) through strong bonding. It is different from supramolecular self-assembly because it connects molecular building units by strong bonding. “Reticular synthesis” provides a reasonable way to synthesize rigid solid materials with high stability, extended structure, predesigned molecular building units, and special properties; therefore, it is also a branch of crystal engineering. Two important aspects of “reticular synthesis” are the rational design of secondary building units (SBUs) and framework topologies. SBUs can be derived from inorganic or organic components with different geometric structures, as shown in Figure 8.1. For the SBUs of the same geometry, many different topologies may be induced. For example, the same tetrahedral building unit can be connected in different ways to form up to 100 different topologies. As far as topology is concerned, Wells has done the most thorough and original work in this field as early as 1975 [13–15]. The following is a detailed description of the framework topology.

8.3 Geometric Basis of Crystallization Chemistry From 1954 to 1972, Wells introduced the geometric basis of crystallization chemistry in 11 parts of Acta Cryst [15–25]. He deduced a series of three-dimensional network from two-dimensional network using mathematical geometry method; a network composed of nodes and linkers between nodes. These networks are represented by symbols (n, p), where p denotes the number of nodes connected to a node and n denotes the number of nodes in the shortest ring around a node. The basic requirement for building three-dimensional structures is that each building element needs six directions (up, down, left, right, front, and back). Therefore, for different connecting nodes, the minimum number of nodes (Z t ) required for each building element is different, as shown in Figure 8.2. For example, to connect 3-connected nodes to form a three-dimensional network, each repetitive unit should contain at least four nodes (Z t = 4).

8.3 Geometric Basis of Crystallization Chemistry Inorganic units

SBUs

(a)

Organic units

SBUs

(f)

(b)

(g) (c)

(d)

(h)

(e)

(i)

Figure 8.1 Secondary building units (SBUs) with different geometric structures. Source: Reproduced with permission [12]. Copyright © 2003, Springer Nature.

P=

Zt =

Figure 8.2

3

3 and 4

4

6

4

3

2

1

Basic construction units in three-dimensional structures.

Wells classifies three-dimensional networks into two categories: uniform nets and nonuniform nets, as shown in Figure 8.3. Uniform nets also include Platonic networks and Catalan networks: platonic networks refer to all nodes that are p-connected, and all the shortest rings are n-edges, represented by symbols (n, p); Catalan networks refer to the inclusion of different connecting nodes p-, q-, and

427

428

8 Structural Topologies and Interpenetration in the Coordination Polymers

all shortest circuits n-gons Platonic (n,p) all points p-connected Uniform (n,p) nets p Catalan (n, ) q

all shortest circuits n-gons Points p–, q– and so on, connected

Shortest circuits of more than one kind Non-uniform (n,p) nets m ( n , p) Archimedean all points p-connected

Figure 8.3

Classification of uniform nets and nonuniform nets.

(

) p . q Nonuniform nets, also known as Archimedean networks, refer to all nodes that are p-connected, the shortest ring is more than one type, which is represented by ( but ) m the symbol ,p . n Wells lists a series of typical uniform nets and nonuniform nets. For example, the 3-connected Platonic network consists of five types, 73 , 83 , 93 , 103 , and 123 (superscript 3 indicates that the shortest rings in three directions of the 3-connected node are the same); at the same time, the 3-connected node also generates the Archimedean network 3.202 (superscript shows that the shortest ring in one direction of each 3-connected node is 3, while it is 20 in the other angular directions). so on, but all the shortest rings are still n-edges, represented by symbols

Uniform nets

Number

(12,3)

1

(10,3)

11

(9,3)

3

(8,3)

15

(7,3)

4

n,

For Platonic networks, the same identifying network also produces different structural types, as shown in the right table, for example, the network of (9,3) has three different structural types; the network of (8,3) has 15 different structural types. Wells uses (n, p)-a, (n, p)-b to distinguish these networks. Fischer and Koch use p/n/f # to differentiate these networks [26], where f represents the highest symmetry in the structure (cubic (c), hexagonal/trigonal (h), tetragonal (t), orthorhombic (o), monoclinic (o), and triclinic (a, for anorthic)). If the values of p, n, and f are the same, they are differentiated by number#. O’Keeffe also proposed a method of using extended long Schläfli symbols to identify [27]. For example, the long Schläfli symbols of (10,3)-a network are 105 .105 .105 , and subscript 5 indicates that there are five 10-member shortest rings in one direction of the 3-connection node. For (10,3)

8.3 Geometric Basis of Crystallization Chemistry

Table 8.1

Symbols of several (10,3) networks.

Wells

Zt

Fischer and Koch

Vertex symbol

(10,3)-a

4

3/10/c1

105 105 105

3/10/t1 3/10/t3 (10,3)-b

4

3/10/t4

102 104 104

(10,3)-c

6

2/10/h1

10 102 102

(10,3)-d

8

3/10/o1

102 104 104

(10,3)-e

12

—

103 103 105

(10,3)-f

16

—

(10,3)-g

20

—

103 103 103 102 102 104 102 102 104 104 104 104 104 104 104 103 104 105 103 104 105 —

16

3/10/t2

10 10 103

—

16

3/10/t5

102 104 104

—

16

3/10/t6

10 10 103

—

32

3/10/t7

10 10 103

networks, there are as many as 11 different structural types. Wells, Fischer and Koch, and O’Keeffe are used to identify the network, as shown in Table 8.1. O’Keeffe classifies the types of networks in more detail [28] and defines regular net and quasi-regular net, respectively. Regular net means that all nodes, edge lengths, and angles are identical, and the regular network must be uniform, for example, the diamond structure is the only regular network in 4-connected three-dimensional networks. Quasi-regular net means that all nodes and edge lengths are identical. In Table 8.2, some simple 4-connection network names and their types are listed. O’Keeffe standardized the nomenclature of the topological network and proposed to identify the topological network by three letters plus one extended character [29] (http://www.iza-structure.org/databases). This method is also closely combined with the nomenclature of zeolite molecular sieve network. For example, (10,3)-a network has a variety of naming methods before, such as “Laves net,” “Y*,” “3/10/c1,” “SrSi2 net,” and “labyrinth graph of the gyroid surface,” O’Keeffe standardized this kind of naming method and defined (10,3)-a network directly with the three letters of “srs.” He defines the corresponding symbols for some known topological networks and organizes them on the network as a database (http://rcsr.anu.edu.au).

429

430

8 Structural Topologies and Interpenetration in the Coordination Polymers

Table 8.2

The names and types of some simple 4-connected nets.

Net

Z

Symbol

Regular

Uniform

Diamond

2

62 .62 .62 .62 .62 .62

Yes

Yes

CdSO4

2

6.6.6.6. 62 .∞

No

Yes

Quartz

3

6.6.62 .62 .87 .87

Quasi

No

NbO

3

62 .62 .62 .62 .82 .82

Quasi

No

W2

3

3.7.7.7.72 .72

No

No

Dense

3

72 .∞.73 .73 .73 .73

No

Yes

Sodalite

6

4.4.6.6.6.6

Quasi

No

(a)

Figure 8.4

(b)

Self-dual net of cds network (a) and rutile network (b).

8.4 Dual Nets The new network formed by connecting the polyhedron (or polygon) centers in the network is called the dual nets of the network. If the dual nets is consistent with the original network structure, the original network is self-dual nets. For example, the dual net of the planar (4,4) network is still (4,4) network; the dual net of the simple cubic network (pcu) is also pcu network; therefore, both of them are self-dual nets. Similarly, the 4-connected cds network and (3,6)-connected rutile (TiO2 ) network are also self-coupling networks (Figure 8.4). However, the natural tiles for the quartz net are tetrahedra, indeed topologically the same as the CdSO4 tiles. However, the 4-connected quartz network is not consistent with its dual net, so the quartz structure is not self-dual (Figure 8.5). MOFs with self-dual nets often exhibit a twofold interpenetration structure [30].

8.5 Topological Network of Metal–Organic Framework In 1989, Professor Robson and coworker reported one metal–organic coordination polymer [CuI C(C5 H4 CN)4 ]n n+ [31] and extended Wells network topology work to organic, metal–organic compounds, and coordination polymers. In recent years, the

8.5 Topological Network of Metal–Organic Framework

Figure 8.5

Quartz networks and its dual net.

study of crystal structure topology has gradually become one of the hotspots in the field of crystal engineering because understanding the construction of a structure at the molecular and atomic levels is the basis of material design and synthesis. Enumerating novel network topologies (e.g. 84 ) and identifying previously unrecognized networks (e.g. CdSO4 , NbO, SrAl2 , etc.) have always been an important task in the field of crystal engineering. Next, we review some reported MOFs with typical topological structures.

8.5.1

Two-Dimensional Network

Wells summarizes some two-dimensional planar topological networks of 3-connection, 4-connection, and other hybrid connections in his classical works by mathematical methods. Figures 8.6–8.8 illustrate some typical examples. In recent years, a series of two-dimensional planar topological networks have been confirmed by the synthesis of a large number of two-dimensional MOFs. Here we introduce some special examples. 8.5.1.1 3-Connected Topology

Champness and coworkers reported the complexes [Er2 (L)3 (NO3 )6 ]⋅2CH3 OH [32], which were assembled by N,N ′ -dioxide-4,4′ -bipyridine ligands and rare-earth metals. Although the metal Er is 9-coordinated, only three metal Er are connected by three L-ligands, resulting in a twofold interpenetration fes topology network with the point symbol of 4.82 . Hong and coworkers reported one 2D graphite-like structure sheets with [Cu6 (CN)6 ] hexagonal units. In the structure, each Cu is coordinated to three CN− , and each CN− links two Cu to produce the lamellar structure with hcb topology [33]. In 2010, Zhang and coworkers reported a self-catenated coordination polymer (Figure 8.9) built by V-shaped 4-(4-carboxyphenylamino)-3,5-dinitrobenzoic acid and a flexible exo-bidentate co-ligand 1,3-bis(4-pyridyl)propane [34]. From a topological view, such framework can be considered as 2D 66 nets, with the vertex symbol of 62 ⋅62 ⋅62 ⋅62 ⋅62 ⋅62 . In 2012, they obtained a series of frameworks with 2D → 3D n-fold Borromean entangled topology [35]. The flexible

431

432

8 Structural Topologies and Interpenetration in the Coordination Polymers

Net 1, N = 2

Net 2a, N = 12

Net 2b, N = 8

Net 3, N = 4

Net 4, N = 4

Net 5, N = 6

Net 6, N = 6

Net 7, N = 6

Net 8, N = 6

Figure 8.6 Some typical 3-connected 2D planar topological networks, N represents the number of lattice point.

ligand 1,1′ -bis(4-carboxybenzyl)-4,4′ -bipyridinium and Co(II) ions generate a 2D honeycomb-like (6,3) layer and further stack in a –ABCABC– manner along the c-axis, resulting in a 3D porous network containing distorted butterfly-like window. 8.5.1.2 4-Connected Topology

The common 4-connected 2D topology is an sql topology. However, usual 4-connected network with Kagomé lattice (kgm) has been reported in MOFs. One example is Chungs reported compound [Co(L1 )2 (NCS)2 ]n (1-methyl-1′ -(3-pyridyl)2-(4-pyrimidyl)ethene) [36]. The asymmetric ligand L1 bridges four-linked Co atoms to form three-member and six-member ring structures, respectively, to form slightly distorted Kagomé lattice structure. In addition, they successfully obtained the complex [Co(L2 )2 (NCS)2 ]n [37], with larger pore size Kagomé lattice structure, using the long-chain rigid ligand L2 (trans-4-styrylpyridyl) benzene. Eddaoudi and coworkers also successfully constructed a metal–organic coordination framework with two-dimensional Kagomé lattice structure using

8.5 Topological Network of Metal–Organic Framework

Net 1. N = 1

Net 2a. N = 4

Net 3. N = 9

Net 2b. N = 4

Net 3. N = 6

Net 4a. N = 3

Net 4b. N = 3

Net 5. N = 12

Net 6. N = 4

Figure 8.7 Some typical 4-connected 2D planar topological networks, N represents the number of lattice point. Figure 8.8 Some typical (3,4)-connected (a, b) and (3,6)-connected (c) 2D planar topological networks. (a)

(b)

(c)

433

434

8 Structural Topologies and Interpenetration in the Coordination Polymers

(a)

Figure-eight knot

Trefoil knot

(b)

a c

b

Figure 8.9 (a) The transformation of trefoil-type knot (middle) to the standard trefoil knot (right) without destroying the crossings. (b) The trefoil knots catenate with each other and periodically extend along the ab plane. Source: Sun et al. [34]. © 2010 Royal Society of Chemistry.

2,4-pyridinedicarboxylic acid and In3+ , as shown in Figure 8.10 [38]. Four 2,4-pyridinedicarboxylic acid coordination one In atoms form a 4-connected node with a planar configuration. The sequential connection of the four-junction unit forms a three-member ring and a six-member ring structure, similar to the Kagomé lattice type.

8.5.1.3 (3,4)-Connected Topology

Zaworotko and coworkers reported for the first time a two-dimensional metal–organic coordination framework constructed from five-membered rings, assembled from hexamethylenetetramine ligands and Cu(Ac)2 units [39]. In the structure, hexamethylenetetramine ligands adopt tridentate and tetradentate coordination modes, showing 3-connection and 4-connection nodes, respectively; thus, the whole framework has a (3,4)-connected topological network. Bu and coworkers synthesized ZBIF-1 with (3,4)-connected topology built from two types of tetrahedral units BN4 and ZnN3 Cl (Figure 8.11). The tetrahedral ZnN3 Cl unit serves as a 3-connected node and is bonded by two 4-connected B(im)4 − ligands, one terminal Cl− anion and one μ2 -im ligand. Thus, B and Zn sites are 3- and 4-connected nodes, respectively, with the im ligands as linkers and they form the final topological network with the symbol of (5,3 4 ) [40].

8.5 Topological Network of Metal–Organic Framework

(a)

(b)

100 µm (c)

(d)

Figure 8.10 Metal–organic coordination framework with two-dimensional Kagomé lattice structure. Source: (a) Reproduced with permission Liu et al.[38]. Copyright 2005, The Aerican Chemistry Society, (b–d) Liu et al. [38]. © 2005 American Chemical Society.

b c

(a)

a

(b)

Figure 8.11 (a) The mcm net in ZBIF-1, the im ligands are highlighted as red lines; (b) the eclipsed stacking of the mcm-type layers, generating pentagonal channels (one is highlighted as a red pentagon). Source: Chen et al. [40]. © 2010 Royal Society of Chemistry.

8.5.1.4 6-Connected Topology

In 2016, Yaghi and coworkers reported a two-dimensional titanium–organic framework (MOF-901) that combines the chemistry of MOFs and covalent–organic frameworks (COFs). In MOF-901, a trigonal prismatic Ti6 O6 inner core was generated in situ as an SBU, which was surrounded by six terminal 4-aminobenzoate ligands.

435

436

8 Structural Topologies and Interpenetration in the Coordination Polymers

O

(a)

H2N

+

Ti(OiPr)4 (OiPr = Isopropoxide)

OH 4-Aminobenzoic acid (H-AB)

O O Benzene-1,4-dialdehyde (BDA)

(b)

(c)

III

Ti6O6(OCH3)6(AB)6

MOF-901 Ti6O6(OCH3)6L3 (L = C20H14N2(CO2)2)

Figure 8.12 The crystal structure of MOF-901 with 6-connected hxl topology. Source: Nguyen et al. [41]. © 2016 American Chemical Society.

These Ti6 O6 cores were further linked together through imine condensation reactions with benzene-1,4-dialdehyde to form a hexagonal porous layer with a 36 faced hxl topology (Figure 8.12) [41]. 8.5.1.5 (3,6)-Connected Topology

A family of two-dimensional materials with (3,6)-connected kgd layers have been designed and synthesized by Hong group. Through a cation-induced synthetic strategy, an hourglass-shaped anionic Cu6 I7 − cluster was constructed. Six organic ligand tppa is coordinated with Each Cu6 I7 − cluster acts as a 6-connected node which is further linked by tripodal organic ligands tppa from six directions to form a (3,6)-connected layer structure with a Kagomé dual (kgd) topology. The Schläfli symbol of which is (43 )2 (46 ⋅66 ⋅83 ) (Figure 8.13) [42].

8.5.2

3-Connected Three-Dimensional Topological Network

Fischer and Koch listed 54 kinds of 3-connected three-dimensional topological networks, in which the srs (SrSi2 ) topological network is the only uniform network, that is, the nodes have position symmetry and contain three rotating axes, so that all nodes, side lengths, and angles are the same. The minimum number of nodes in 3-connected three-dimensional topological networks is 4, and only two kinds

8.5 Topological Network of Metal–Organic Framework

N N

+

O HN

P NH HN

N

+

Figure 8.13 The construction of (3,6)-connected kgd layer. Source: Yu et al. [42]. © 2017 American Chemical Society.

Table 8.3 Name

Two basic 3-connected three-dimensional topological networks. Space group

Wyckoff

Point group

Z

Vertex symbol

Space group (AB)

SrSi2

I41 32

a or b

32 (D3 )

4

105 105 105

P21 3

ThSi2

I41 /amd

e

mm2 (C2v )

4

102 104 104

I41 md

Note: Z is the number of vertices in the topological unit, and space group (AB) is the space group for the ordered AB structure with two kinds of vertex.

of topological networks, srs(SrSi2 ) and ths(ThSi2 ), have this characteristic. Therefore, these two kinds of topological networks are often considered as the most basic 3-connected three-dimensional topological networks. The properties of these two kinds of topological networks are shown in Table 8.3. Next, we introduce some MOFs with 3-connected three-dimensional topological networks.

8.5.2.1 srs(SrSi2 ) Topological Networks

The main feature of this network is that there are multiple helical chains in all axes and all helical directions are identical, so the whole frame is chiral. The complex [Zn2 (BTC) (NO3 )](H2 O)(C2 H5 OH)5 (BTC = 1,3,5-benzenetricarboxylate) has this kind of topological network [43]. The BTC ligand bridges 3-connected binuclear zinc units to form a 10-member ring structure. The framework has 31 helical chains in one axis.

437

438

8 Structural Topologies and Interpenetration in the Coordination Polymers

8.5.2.2 ths(ThSi2 ) Topological Networks

The main feature of this network is that the adjacent layers of the zigzag chains are vertically connected to each other. At the same time, it can also be considered as the basic framework formed by T-shaped building units. The complex [Ag(4,4-bipyridine)NO3 ] has this kind of topological network [44]. Ag atoms form T-shaped building units through the Ag–Ag interaction which is not supported by the chain and the coordinated 4,4-bipyridine groups, which are sequentially connected to form ths topological network.

8.5.2.3 utp((10,3)-d) Topological Networks

Similar to the srs network, the utp network has 41 helical chains in one axis, but there are both left and right helical chains in the network and the number of them is equal. Therefore, the network shows no chirality. Complex [Co(NO3 )(1,4-bis(3-pyridyl)-2,3-diaza-1,3-butadiene)1.5 ]⋅H2 O has this kind of topological network [45]. Co atoms are connected with three Co atoms through ligands, and the whole framework is ultimately a strongly distorted utp topological network with twofold interpenetration.

8.5.2.4 Other 3-Connected Topological Networks

Wu and coworkers reported the compounds with a 3-connected 82 10 topological network for the first time, [Ph3 PCH2 Ph][Cd(tp)⋅Cl]⋅2H2 O [46]. Due to the presence of large guest molecules, the compound exhibits a high porous structure. Cd atoms are connected with three Cd atoms by two terephthalic acid ligands and two bridged Cl ions, forming a 3-connected unit. Eight Cd atoms are bridged by terephthalic acid ligands to form 8-rings. Finally, the whole framework is shown as a twofold interpenetration 82 10 topological network. Li and coworkers reported two metal complexes with (8,3) and 82 10 topological structures obtained by self-assembly of tetrazole ligands and CuI compounds, respectively, as shown in Figure 8.14 [47]. The N atoms in the tetrazole ligands of the two complexes are all involved in the coordination. The Cu atoms are in a four-coordination mode but only connected with the three adjacent Cu atoms. Therefore, the Cu atoms are regarded as 3-connected nodes. The three-dimensional framework of the two complexes is simplified to 3-connected (8,3) and 82 10 networks, respectively. There are 31 helix chains in (8,3) network and 41 helix chains in 82 10 network. Cao and coworkers designed and synthesized a twofold interpenetrating anionic 3-connected networks through assembling between a flexible ligand 1,3,5-tris[4-(carboxyphenyl)oxamethyl]-2,4,6-trimethylbenzene (TBDC) and Zn(NO3 )2 under solvothermal reaction [48]. In the structure, each TBTC3− ligand acts as a tri-bridging ligand that links three metal centers to a three-dimensional topological network. Then, two of these identical networks resulted in an overall twofold interpenetration in the framework.

8.5 Topological Network of Metal–Organic Framework

(a)

(b)

Figure 8.14 Two metal complexes with (8,3) and 82 10 topological structures. Source: Wu et al. [47]. © 2005 American Chemical Society.

Table 8.4

Some important 4-connected topological networks.

Symbol

Name

Space group

Wyckoff

Point group

Z

Vertex symbol

D

Diamond

Fd3m

a

43m (T d )

2

62 ⋅62 ⋅62 ⋅62 ⋅62 ⋅62

J*

NbO

Im3m

b

4/mmm (D4h )

3

62 ⋅62 ⋅62 ⋅62 ⋅82 ⋅82

Q

Quartz

P62 22

a

222 (D2 )

3

6⋅6⋅62 ⋅62 ⋅87 ⋅87

W*

Sodalite

Im3m

d

4m2 (D2d )

6

4⋅4⋅6⋅6⋅6⋅6

V

I41 32

c

222 (D2 )

6

3⋅3⋅102 ⋅102 ⋅103 ⋅103

S*

Ia3d

d

4 (S4 )

12

6⋅6⋅62 ⋅62 ⋅62 ⋅62

Lonsdaleite

P63 /mmc

f

3m (C3v )

4

62 ⋅62 ⋅62 ⋅62 ⋅62

CdSO4

P42 /mmc

a

mmm (D2h )

2

6⋅6⋅6⋅6⋅62 ⋅*

CrB4

I4/mmm

g

mm2 (C2v )

4

4⋅62 ⋅6⋅6⋅6⋅6

SrAl2

Imma

i

m

4

4⋅6⋅4⋅6⋅6⋅82 ⋅62

Mogantie

Cmmm

PtS

8.5.3

P42 /mmc

a

mmm

1

4⋅4⋅62 ⋅62 ⋅84 ⋅84

h

mm2

2

4⋅86 ⋅6⋅6⋅6⋅6

c

mmm (D2d )

2

4⋅4⋅82 ⋅82 ⋅82 ⋅82

f

4m2 (D2d )

2

4⋅4⋅87 ⋅87 ⋅87 ⋅87

4-Connected Three-Dimensional Topological Network

Table 8.4 lists some important 4-connected topological networks, in which diamond network belongs to uniform network and the other five networks belong to the nonuniform network, which are represented by symbols, respectively. The latter networks contain two types of nodes and links. Next, we introduce some MOFs with 4-connected three-dimensional topological networks.

439

440

8 Structural Topologies and Interpenetration in the Coordination Polymers

8.5.3.1 Dia (= Diamond) Topological Network

Yao and coworkers successfully obtained a complex [Cu10 Cl10 (4-S–C5 H4 NH)4 ]n with non-interpenetrating diamond network using 10-core copper clusters as building units [49]. Ten-core Cu clusters are simplified to 4-connected nodes. Because of their compact stacking structure, no interpenetration occurs. This is a very common and very important topology, which often occurs in interpenetration. Ciani and coworkers obtained a series of diamond network structures from 4-fold to 10-fold interpenetration by self-assembly of long-chain dicyanoalkane ligands with silver salts containing different anions [50]. Compounds [Ag(ddn)2 ]PF6 and [Ag(ddn)2 ]AsF6 (ddn = 1,12-dodecanedinitrile) exhibited 8-fold interpenetration diamond networks, while compound [Ag(ddn)2 ]NO3 showed 10-fold interpenetration. The anion size plays a decisive role in the interpenetration of diamond networks. In 2012, Zhang and coworkers synthesized one extraordinary neutral heterometallic–organic framework CdII CuI 2 (pybz)4 (Hpybz = 4-(pyridin-4-yl)benzoic acid) with 25-fold interpenetrating dia topology by assembling of nanosized Cu(pybz)2 metalloligands and tetrahedral [Cd(COO)4 ] units [51]. Generally, only anionic MOFs can be gained through cross-linking aromatic polycarboxylate ligands and 4-connected [MII (COO− )4 ] units. In their work, metalloligands were employed to effectively tune the charge-balance of the framework using −1 charged metalloligand to replace −2 charged organic dicarboxylate ligands. Due to the perfect linear coordination of Cu+ with two pyridine nitrogen atoms, a long Cu(pybz)2 metalloligand was formed. Then, each metalloligand chelated two Cd2+ centers leading to a dia topology network with a large diamondoid cage with an edged Cd· · ·Cd distance of 26.5 Å. Such long distance induced 25-fold interpenetration of the frameworks along the b-axis. 8.5.3.2 cds (= CdSO4 ) Topological Network

The minimum number of nodes to be included in the repetitive unit in the 4-connected three-dimensional topological network is 2. The cds topological network is the second network with this characteristic (the first is dia network), and the space group of the ideal network is P42 /mmc. Champness and coworkers reported the non-interpenetration cds topology network constructed by nonplanar 4-connection nodes [52]. The ideal cds topological network is constructed by planar 4-connected nodes. In this compound, the 4-connected metal center has a near tetrahedral configuration, which results in a strongly distorted 65 .8 cds topological network. Zaworotko and coworkers reported two types of MOFs constructed by isophthalic acid ligands, with 4-connected cds (CdSO4 ) topology network and a new 65 .8 topology network (USF-1) [53]. The ideal USF-1 topological network has R3c symmetry, a/c = (8/3)1/2 , and the nodes are in the 18d position, while the ideal cds topological network has P42 /mmc symmetry, c/a = 2, and the nodes are in the 2a position. Because of the existence of large void space and self-coupling behavior, cds topological network often leads to twofold or even multiple-fold interpenetration. For the

8.5 Topological Network of Metal–Organic Framework

Y

X Z

(a)

(b)

Figure 8.15 Cds topological network with fivefold interpenetration. Source: Bhogala et al. [54]. © 2004 American Chemical Society.

first time, Nangia and coworkers reported a fivefold interpenetration cds topology [54], as shown in Figure 8.15. The framework is constructed by hydrogen bond interaction. There are 16 × 17 Å one-dimensional pore structure in the single framework, which eventually leads to the fivefold interpenetration of the framework.

8.5.3.3 NbO Topological Network

Maverick and coworkers reported an MOF with 4-connection NbO topological network [55], which is connected by a planar tetragonal construction unit, Cu(Pyac)2 (PyacH = 3-(4-pyridyl)pentane-2,4-dione). In this compound, each metal center is connected with four metal centers by four straight-chain ligands in a planar position, which results in the formation of NbO topological network. Because of the existence of large void spaces and the self-dual behavior of NbO topological network, two identical frames are interpenetrated to form the final three-dimensional structure. The compound [Cu(ptz)]n (Hptz = 3,5-dipropyl-1,2,4-triazole) reported by Chen and coworkers [56], as shown in Figure 8.16, if the binuclear copper unit is considered as a single node, the whole network will behave as a 4-connected NbO topological network; if the one copper as a node and the ligand is simplified as a 3-connected node, the whole network will behave as a 3-connected 4.122 network. This also shows that the 3-connected network and the 4-connected network can be converted to each other.

8.5.3.4 Quartz Topological Network

Tong and coworker successfully constructed a 64 .82 quartz structure using a Cu4 cluster as a 4-connection building unit linked through the flexible bidentate ligands [57], as shown in Figure 8.28. In this compound, tetranuclear copper clusters are simplified to distorted tetrahedral nodes. The large chiral holes resulted in threefold interpenetration 64 .82 quartz structures.

441

442

8 Structural Topologies and Interpenetration in the Coordination Polymers

(a)

(b)

(c)

Figure 8.16 (a) Definitions of the 3- and 4-connected nodes, (b) the 4-connected NbO net, and (c) the 3-connected 4.122 net. Source: Zhang et al. [56]. © 2004 John Wiley & Sons.

8.5.3.5 Moganite and PtS Topological Network

zur Loye and coworkers successfully constructed a non-interpenetrating moganite topological network with Cu2+ salts and rigid pyridine carboxylic acid ligands [59], as shown in Figure 8.17. There are two crystallographically independent copper centers in the complex, which are connected with the four copper centers via the ligand, respectively. A non-interpenetrating moganite topological network is constructed by these two types of 4-connection nodes. The environment around Cu1 atom is 4.86 .6.6.6.6 and that around Cu2 atom is 4.4.62 .62 .84 .84 , and their locations proportion is 2 : 1, so the whole network symbol is abbreviated as (42 .62 .82 )(4.64 .8)2 . Tong and coworkers reported two MOFs with 4-connected moganite and PtS topological networks, respectively, constructed from tetrapyridine ligands generated in situ, as shown in Figure 8.18 [60]. Among these two compounds, tetrapyridine ligands have a planar quadrilateral configuration, and tetrahedral metal centers have

8.5 Topological Network of Metal–Organic Framework

N1O3 N3 CU1

N2A C5A

N4 O1

N2 CU2 O5 5A

N6 N5

b a

(a)

(b)

Cu1

b

Cu2

c

a

b

a

c

(c)

(d)

Figure 8.17 Non-interpolating moganite topological network constructed by two kinds of nodes. Source: Su et al. [59]. © 2004 American Chemical Society.

tetrahedral configuration. Moganite and PtS topological networks are obtained by assembling these two types of units. 8.5.3.6 86 (= (8,4)) Topological Network

Wells once predicted that diamond network (66 ) was the only uniform net in the 4-connected network, and other networks could not have the symbol of n6 ; Tong et al. broke this prediction and successfully constructed a complex Ni(pyara)2 (H2 O)2 [61], with 86 topological network. The long Schläfli symbol of the network is 83 .83 .83 .83 .82 .82 ; the highest symmetry is the orthogonal space group Pnna; interestingly, there are parallel single-helix and triple-helix structures in the network, and there are interpenetrating phenomena between eight-rings, showing that this is also a self-interpenetrating network. In the same period, Ma et al. also reported the complex [Cu(1,1′ -(1,4-butanediyl)bis(imidazole))(m-phthalate)(H2 O)]⋅5H2 O [62], which has the same 86 topology network. 8.5.3.7 Zeolite-Type Topology Networks

The inherent structural feature of inorganic zeolite is the 4-connected network with the corner-sharing tetrahedral TO4 as primary building unit in which T–O–T angle is about 145∘ . The strategy to create stable and rigid zeolite-type MOFs is based on that the bridging angle in the M–Im–M fragment (where M is the tetrahedral coordinated metal ion and Im is imidazolate) is coincident with that of T–O–T angle in zeolite. Chen, You, and Zhao successfully synthesized several MOFs with zeolite SOD, ANA, RHO, and GIS topologies by the assembling of zinc or cobalt ions with imidazolate ligands [63–66]. Soon after, Yaghi and coworkers have effectively expanding these

443

444

8 Structural Topologies and Interpenetration in the Coordination Polymers

(a)

(c)

(b)

(d)

Figure 8.18 Mogonite (b) and PtS (d) topological networks constructed by different assemblies of the same ligand and metal. Source: Hu et al. [60]. © 2005 John Wiley & Sons.

zeolite-type MOFs by the high-throughput synthesis method and called these materials as zeolitic imidazolate frameworks (ZIFs). So far, over 100 new ZIF structures have been reported [67, 68]. To integrate compositional and structural features of zeolites and ZIFs, Zhang group successfully developed a new kind of hybrid zeolitic imidazolate frameworks (denoted HZIFs) [69]. Two iso-structural compounds, HZIF-1Mo and HZIF-1W, were obtained by the mixing of Zn(CH3 COO)2 , H2 MoO4 or H2 WO4 , and 2-mim in N,N ′ -dimethylformamide (DMF) at 160 ∘ C for six days. The prominent structural feature of HZIF-1 was that the truncated octahedral cages of [Zn24 (2-mim)36 ] was interconnected by the inorganic MoO4 or WO4 units (Figure 8.19). The large [Zn24 (2-mim)36 ] cage was just the same as the subunit in ZIF-8. More interestingly, this cage was surrounded by the inorganic MoO4 or WO4 unit, which resulted in the connectivity between inorganic TO4 units and ZIF cages. Thus, the obtained HZIFs combined catalytic active building blocks and high thermal stability, which presents a new class of porous materials filling the gap between zeolites and ZIFs. In addition, boron imidazolate frameworks (BIFs) are a series of lightweight ZIF analogs based on predetermined tetrahedral boron imidazolate complexes. In 2009, Zhang et al. first reported the system of zeolitic BIFs [70]. BIF-1-Li and BIF-1-Cu ([LiB(im)4 ] and [CuB(im)4 ], im = imidazolate) exhibit a 4-connected 3D zni-type framework containing the SiO4 -like tetrahedron Li(im)4 and B(im)4 alternated through corner sharing. The first porous zeolitic BIFs are BIF-3-Li and

8.5 Topological Network of Metal–Organic Framework

(a)

W O

(b)

Figure 8.19 The connection between TO4 units and truncated octahedral cages in HZIF-1. Source: Wang et al. [69]. © 2011 John Wiley & Sons.

(a)

(b)

Figure 8.20 The SOD topology of BIF-3 (a) and RHO topology of BIF-9 (b). Source: Zhang et al. [71]. © 2016 Elsevier.

BIF-3-Cu ([LiB(mim)4 ] and [CuB(mim)4 ]), which possess the neutral sodalite (SOD) network structure synthesized via solvothermal method (Figure 8.20). Later, they extended the dense SOD network to a more open RHO topology [72], BIF-9-Li and BIF-9-Cu (LiB(4-mim)4 and CuB(4-mim)4 , 4-mim = 4-methyimidazolate), which represent the lightest zeolite RHOs known to date, by the selection of a 4-methyl imidazolate-based boron imidazolate ligand (HB(4-mim)4 ) (Figure 8.34). One advantage of the design strategy for BIFs is that both tetradentate and tridentate boron imidazolate ligands can be presynthesized prior to solvothermal synthesis. Because the 3-connected center of tridentate boron imidazolate ligand has a tetrahedral geometry configuration rather than a planar structure, it can serve as building blocks for constructing 3-connected interrupted zeolitic networks by assembling with tetrahedral metal node. In 2011, the first interrupted Zeolite A

445

446

8 Structural Topologies and Interpenetration in the Coordination Polymers

BIF-20 (a)

(b)

Figure 8.21 The interrupted LTA topology of BIF-20. Source: Zhang et al. [74]. © 2011 American Chemical Society.

(LTA) BIF-20 has been experimentally realized. There are both tetrahedrally coordinated Zn and B nodes realized in BIF-20, but each Zn node is only bonded by three tridentate boron imidazolate ligands with the fourth site covered by an oxygen atom from carboxylate ligand and each B node is bonded to three mim. Thus, an unusual 3-connected tetrahedral network with vertex symbol of (6.8.10)(6.82 )(82 .10)(83 ) (denoted as fjz) is formed by the alternate linking between Zn and B nodes via 2-methyimidazolate. This net is derived from the well-known LTA topology by omitting one linker from the tetrahedral node, in which both debonded α cages and β cages are found (Figure 8.21) [74]. In the early stage, tetrahedral metal centers with mono-positive charged Li and Cu were usually chosen for constructing neutral zeolitic BIFs. In 2012, Zhang group used a new approach to go beyond these limits for the synthesis of neutral zeolitic BIFs from octahedral metal centers with distorted tetrahedral building nodes. The +1 charged distorted tetrahedral metal node can be obtained by the combination of octahedral divalent metal center with terminally chelating monocarboxylate ligand. Two zeolitic BIFs with ACO (BIF-22) and ABW (BIF-23) topologies were successfully constructed through the cross-linking of tetrahedral boron imidazolate ligands B(im)4 − with [M(ac)]+ units (ac = acetate) [75]. Furthermore, they used dicarboxylate ligands to replace the terminal monocarboxylate ligands to bridging the adjacent two metal centers, and the pore space partition was induced in the zeolitic BIFs. Two zeolitic structures BIF-41 (GIS topology) and BIF-42 (ABW topology) with interesting pore space partition inside their host network were constructed by employing flexible adipic acid (H2 ad) and rigid 2,6-naphthalene dicarboxylic acid (H2 nda) as bridging ligands, respectively [76]. Furthermore, besides individual tetrahedrally coordinated atoms, it is highly desirable to use composite tetrahedral clusters for constructing zeolite frameworks because of the possibilities created by the clusters to introduce diversity in compositions, large pore size, and additional functionality. In 2012, Zhang group developed a synthetic method to create an extraordinary photoluminescent cluster–organic framework with MTN-type zeolitic topology (COZ-1, [Cu4 I4 (dabco)2 ]n ) using copper iodide clusters (Cu4 I4 ) as pseudotetrahedral units. A prominent structural

8.5 Topological Network of Metal–Organic Framework

11.293

(a)

(b)

9.918

(c)

(d)

Figure 8.22 The tetrahedral Cu4 I4 clusters, 64 512 cage and 512 cage, and topological representation of COZ-1. Source: Kang et al. [77]. © 2012 American Chemical Society.

feature of COZ-1 is that two types of giant cages are presented in the structure. Both of the giant cages are built from tetrahedral Cu4 I4 clusters. The larger ones are 64 512 cages that contain 28 Cu4 I4 clusters and 42 dabco ligands and have an inner diameter of 2.6 nm with a calculated pore volume of 9.2 nm3 . The smaller ones are 512 cages containing 20 Cu4 I4 clusters and 30 dabco linkers. These giant 64 512 cages and 512 cages via sharing faces give rise to a three-dimensional framework with MTN-type zeolitic net (Figure 8.22). With this procedure, it exhibited perfect integration of porosity and photoluminescent properties from both the cluster and the framework [77]. Champness and coworkers designed and synthesized a tetradentate planar ligand 2,3,4,5-tetra(4-pyridyl)thiophene and synthesized a novel 4-connected MOF by self-assembly with transition metals. The cationic framework has zeolite structure, and the topological network symbol 42 .84 which is the same as EDI and THO zeolites, and has the same cell symmetry as NAT zeolites [78]. 8.5.3.8 Other 4-Connected Topological Networks

Chen and coworkers reported that the complex Cu2 (imidazole)3 with mixed valence copper has a rare 4.85 topological network structure [79]. There is a crossover phenomenon among the 8-rings, that is to say, it is a self-interpolating 4-connection network. Metal complexes with planar 4-connected node topology rarely produce

447

448

8 Structural Topologies and Interpenetration in the Coordination Polymers

(a)

(b)

Figure 8.23 The structure illustration of FJI-H14 with the usf topology. Source: Liang et al. [81]. Licensed under CC BY 4.0.

more than fourfold interpolation structures. Wang et al. reported for the first time the complex [Cu(oba)(H2 O)]2 ⋅0.5H2 O [80], which not only has an lvt topology network constructed by planar 4-connected nodes (4.4.84 .84 .88 .88 ) but also exhibits a fivefold interpenetration structure. If 4-connected nodes are replaced by square lattices, the lvt network will be transformed into 3-connected 4.8.20 network. Hong and coworkers reported a crystal of FJI-H14 ([Cu(BTTA)H2 O]n ⋅6nH2 O) with Kagomé like usf topology constructed by 2,5-di(1H-1,2,4-triazol-1-yl) terephthalic acid (H2 BTTA) (Figure 8.23) [81]. In the structure, each Cu(II) ion has a square pyramidal coordination geometry and act as a planar 4-connected node, further linked by four tetradentate BTTA2− ligands into a 4-connected three-dimensional network.

8.5.4

(3,4)-Connected Three-Dimensional Topological Network

Zaworotko and coworkers reported two MOFs (USF-3 and USF-4) [82], constructed by pyromellitic acid ligands, with (3,4)-connected topological networks, respectively. Among these two compounds, pyromellitic acid ligands act as 3-connected planes, and carboxyl-bridged metal units form tetrahedron or quadrilateral planes. They are interconnected to form a unique (3,4)-connected topological network with macroporous structure. It is undoubtedly the most ideal way to construct (3,4)-connected topological network by assembling triangular ligands with metal centers with tetrahedral coordination geometry. Kim and coworkers reported the compound [Cu3 (tpt)4 ] (BF4 )3 ⋅(tpt)2/3 ⋅5H2 O (Him = 2,4,6-tris(4-pyridyl)-1,3,5-triazine) [83]. The ligand TPT has a triangular shape and is constructed with four coordinated copper centers to form a double interpolated (3,4)-connected cubic-C3 N4 topological network. Zhang group synthesized 3-connected tridentate boron imidazolate ligands and assembled these ligands with tetrahedral metal ions to form a series of (3,4)-connected topological networks. The assembly between triangular

8.5 Topological Network of Metal–Organic Framework

(a)

BIF-24 (b)

Figure 8.24 The ctn topology network of BIF-24. Source: Zhang et al. [84]. © 2013 John Wiley & Sons.

KBH(mim)3 units and tetrahedral Zn centers leads to the formation of a highly symmetrical ctn-type material BIF-24 ([Zn3 (BHmim3 )4 ]⋅(NO3 )2 ) (Figure 8.24) [84].

8.5.5

5-Connected Three-Dimensional Networks

There are two simple 5-connected three-dimensional networks, BN network and BCT network, whose repetitive units contain the least number of connecting nodes 2. The nodes in BN network have triangular bipyramidal coordination mode, and the nodes in BCT network have quadrangular pyramidal coordination mode. There are relatively few MOFs with 5-linked topological networks. At present, there are two reported complexes Cu(4,4′ -bpy)1.5 Cr2 O7 ⋅H2 O and [Mn(tcm)2 (bpeado)]n (tcm = tricyanomethanide, bpeado = 1,2-bis(4-pyridyl)ethane-N,N ′ -dioxide) [85, 86]. They are all shown as twisted twofold interpenetrated 46 64 BN topology networks.

8.5.6

6-Connected Three-Dimensional Networks

Long and coworkers reported the use of the Zn8 (SiO4 ) nucleus as a construction unit, which bridges 6 adjacent nuclei through 12 terephthalic acid ligands, eventually forming a highly stable twofold interpenetration MOF [87]. Zn8 (SiO4 ) core can be simplified as a node, which is connected with only 6 adjacent cores via 12 organic chains. Therefore, the MOF can be regarded as a 12-connected pcu topological network with twofold interpenetration. For the first time, Yao and coworkers combined rigid and flexible ligands to construct a non-interpolated 12-connected pcu topology network [88]. Tetranuclear Cd clusters are linked to six adjacent clusters through eight isophthalic acid ligands and four trimethylene bipyridine ligands. Each tetranuclear Cd cluster is simplified to 12-connected nodes. Interestingly, no interpenetration occurs. Hong and coworkers reported one compound based on CoII ion, an ox2− anion, and 1,2-bis(4-pyridyl)hydrazine (bphy) ligand in situ formation from azopyridine.

449

450

8 Structural Topologies and Interpenetration in the Coordination Polymers

b

a

Figure 8.25 View of the framework of compound showing Star of David catenanes and hierarchical structural assembly. Source: Huang et al. [89]. © 2017 John Wiley & Sons.

The main feature of this compound can be considered as fused Star of David catenanes and its hierarchical structural assembly. Six bphy ligands alternately bridge [CoII 3 (ox)2 ] helicate giving rise to a side-twisted triangular metallacycle or [CoII 2 (ox)] dimer giving rise to a vertex-truncated triangular metallacycle. Two triangular metallacycles are triply intertwined to form a Star of David catenane. Each Star of David catenane links its six neighbors to form a twofold intercatenated porous MOF. From a topological viewpoint, each Star of David catenane can be considered as a 6-connected node; this compound can be simplified as an hxg network with a point symbol of (46 ⋅69 ) (Figure 8.25) [89]. Assembly of silver(I) ions, 1,2,4-triazole, and the [A-PW9 O34 ]9− polyoxometalate species leads to the formation of 0D to 3D polycatenated supramolecular system. In the superstructure, the adamantane-like nanocages behave as basic building blocks and interlock with each other from its six vertices to generate a sextuple intercatenation of discrete adamantane-like cages. The robust and infinitely mechanical interlocked adamantane-like cages form an unprecedented 3D polycatenated supramolecular architecture with an NaCl-type α-Po topology in which the link modes are mechanically interlocking rather than the traditional covalent or coordination bonds interactions. Moreover, this polycatenated supramolecular architecture further twofold interpenetrated each other to form a complex supramolecular superstructure, in which the polyoxometalate species behaves as guest motif to balance the charge of the whole structure (Figure 8.26) [90].

8.5 Topological Network of Metal–Organic Framework

(a)

(d)

(b)

(c)

(e)

Figure 8.26 Schematic representation of the overall structure from a discrete [Ag24 (trz)18 ]6+ nanocage to three-dimensional infinite polycatenation. Source: Kuang et al. [90]. © 2010 Springer Nature.

8.5.7

8-Connected Three-Dimensional Topological Networks

Due to the limitation of steric hindrance and metal coordination number, high coordination rare-earth ions or large second construction units are usually chosen as nodes in the construction of metal–organic complexes with high connection topology network. Champness and coworkers reported an 8-connected topological network formed by self-assembly of N,N ′ -dioxide-4,4′ -bipyridine ligands with lanthanide metals [91]. The 8-connection (35 414 59 )(35 413 510 )2 topological network contains two kinds of nodes, which can be considered to be formed by the intersection of (4,4) lattices and SrAl2 frameworks. Wang et al. reported complexes with self-interpenetrating 8-connected topological networks constructed by terephthalic acid and flexible diimine ligands with Cd salts [92]. In this complex, the trinuclear Cd unit is connected with eight trinuclear Cd units by six terephthalic acid and two diimine ligands. As shown in Figure 8.27, the whole framework can be represented by a symbol 420 68 . In this network, there is self-interpenetration among the shortest 4-rings. In addition, they constructed a pentanuclear zinc cluster unit using terephthalic acid, which connected to eight cluster units around it by phthalic acid to form a new complex

451

452

8 Structural Topologies and Interpenetration in the Coordination Polymers

III

II

II

Cd3

I

I

III

(a)

(b)

(c)

Figure 8.27 Self-interpenetrating 420 68 topology network. Source: Wang et al. [92]. © 2006 John Wiley & Sons.

[Zn5 (μ3 -OH)2 (bdc)4 (phen)2 ] [93]. The complex exhibited a rare 8-connected 424 .5.63 topological network, in which self-interpenetration existed between two 4-rings. In the case of interpenetration, the network is also a self-interpenetrating 8-connected topology network. Lu and coworkers reported for the first time the complex [Cu3 Cl2 (4-ptz)4 (H2 O)2 ]⋅ 3DMF⋅5H2 O with 8-linked body-centered cubic (bcu) structure based on trinuclear Cu clusters, as shown in Figure 8.28 [58]. The trinuclear Cu clusters connect with eight cluster units around it through 4-ptz ligands, forming a bcu structure. The framework shows a very large void space in three directions. Because the dual net

Cu1

Cu2A O1A

Cu2

Cl1A

Cl1 N1

N2

O1 N5 C4 C3

N3 C5 C6

C2

C1 N4

(a)

(b)

b c

(c)

a

(d)

Figure 8.28 The 8-connected bcu topology network based on trinuclear Cu clusters. Source: Luo et al. [58]. © 2005 John Wiley & Sons.

8.5 Topological Network of Metal–Organic Framework

= (a)

(b)

(c)

Figure 8.29 [94].

(d)

One anatase-type (3,6)-connected porous MOF. Source: Based on Xiang et al.

of the bcu topological network is a 4-connected NbO topological network, it is not easy to form a multi-interpenetrating structure.

8.5.8

(3,6)-Connected Three-Dimensional Topological Networks

Wu and coworkers synthesized one anatase-type porous MOF, [KCo3 (C6 H4 O7 ) (C6 H5 O7 )(H2 O)2 ]⋅8H2 O, by a hydrothermal reaction of cobalt acetate, citric acid, and KOH at 120 ∘ C. In the structure, the Co4 core is encapsulated by four symmetry-related citrates via carboxylate groups in an anti-syn bridging mode, which are further linked to six tetrahedral Co centers at the 8c site, while each tetrahedral cobalt (as trigonal node) bonds to three Co4 citrate clusters (as octahedral linkers), resulting in an infinite 3D (3,6)-connected anatase net (Figure 8.29) by ignoring the ionic bond (four K+ ions occupying the 8d position can also be present with the clusters) [94]. Zhou and coworkers synthesized a porous coordination network PCN-53 which is composed of benzo-tris-thiophene carboxylate (BTTC) ligands and Fe3 O clusters. In the structure, the metal clusters of Fe3 O serve as 6-connected nodes and BTTC ligands as 3-connected nodes, which can be simplified as a (3,6)-connected net with a

453

454

8 Structural Topologies and Interpenetration in the Coordination Polymers

+

(a)

(b)

(c)

(d)

(e)

(f)

Figure 8.30 Polyhedra 3D packing in PCN-53 with a (3,6)-connected net. Source: Yuan et al. [95]. © 2012 Royal Society of Chemistry.

topology symbol of (42 ⋅6)2 (44 ⋅64 ⋅86 ⋅10). The prominent structural feature of PCN-53 is the presence of three types of cages constructed from these nodes resulting in micro- and meso-pores in the framework (Figure 8.30) [95]. Robson and coworkers reported the interpenetrating rutile (TiO2 ) topological network [96], which was assembled by planar triangular ligand C(CN)3− and a series of metal ions. Transition metals with a six-coordination mode are connected by planar triangular ligand to form the (3,6)-connected TiO2 network. The existence of pore space leads to a twofold interpenetrating structure.

8.5 Topological Network of Metal–Organic Framework

=

=

(b)

(a) c a

b

F

Ca

(c)

(d)

Figure 8.31 CaF2 topological network formed by the combination of tetragonal cylindrical unit and tetrahedral unit. Source: Chun et al. [98]. © 2004 John Wiley & Sons.

Wang and coworkers reported a series of lanthanide metal complexes with (3,6)connected rutile or rutile-like topological networks [97]. 2,5-Pyridinedicarboxylic acid ligands form two kinds of metal complexes which are formed by cross-linking 2,5-pyridinedicarboxylic acid ligands with mononuclear or binuclear lanthanide metals. Mononuclear or binuclear lanthanide metals act as 6-connected nodes, while 2,5-pyridinedicarboxylic acid ligands act as 3-connected nodes, thus forming twofold interpenetrating and non-interpenetrating rutile topological networks, respectively.

8.5.9

(4,8)-Connected Three-Dimensional Topological Networks

Kim and coworkers reported that an MOF with CaF2 topological network was obtained for the first time by assembling tetra(4-carboxybenzene) methane ligand with Cd(NO3 )2 ⋅4H2 O, as shown in Figure 8.31 [98]. Carboxyl bridges tetranuclear Cd to form 8-connected tetragonal cylindrical construction units, and tetra(4-carboxybenzene)methane ligands have 4-connected configuration. The two types of construction units are connected to form (4,8)-connected CaF2 topological network. Hong and coworkers synthesized one MOF (FJI-H8) based on a new robust multi-carboxylate ligand 3,3′ ,5,5′ -tetra(3,5-dicarboxyphenyl)-4,4′ -dimethoxybiphenyl (H8 tddb). In FJI-H8, each tddb ligand coordinates to eight dicopper(II)

455

456

8 Structural Topologies and Interpenetration in the Coordination Polymers

Figure 8.32 Structural representations of FJI-H8 containing three types of cages. Source: Pang et al. [99]. Licensed under CC BY 4.0.

paddle-wheel SBUs and each Cu2 SBU links to four tddb ligands. The Cu2 SBUs can be treated as 4-connected nodes and the tddb ligands as 8-connected nodes; in this way, FJI-H8 adopts the rare (4,4,8)-c urj network with the topological point symbol of 414 .612 .82 (Figure 8.32) [99]. Yuan and Zhou reported a series of porous coordination polymers PCN-608 with the (4,8)-connected csq network, containing coordinatively unsaturated Zr6 O4 (OH)8 (H2 O)4 clusters as 8-connected nodes and TPCB4 − ligands as planar 4-connected nodes. Interestingly, the coordination vacancies on the Zr6 clusters can be occupied by O atoms from linear dicarboxylate linkers, such as BDC, MPDC, and SBDC, by immersing PCN-608-NH2 into a solution of the dicarboxylate linkers. After linker installation, the connection number of the Zr6 cluster increased to 10-connected and the framework changed to a (4,10)-connected network with a topological point symbol of (32 ⋅42 ⋅52 )2 (38 ⋅416 ⋅58 ⋅613 ). The ionic linker of MPDC and SBDC installation afford ionic to the neutral framework resulting the precise regulation of pore environments in MOFs (Figure 8.33) [100]. Flexible conformations of ligands not only added to the structural diversity of MOFs but also complicated the structural control. A series of (4,8)-connected Zr-tetracarboxylate-based MOFs with three types of topologies were constructed from seven tetratopic carboxylate ligands with the same backbones. The formation of the three different types of structures with flu, scu, and csq topologies depends on the steric hindrance effect by altering the substituents on different positions of organic linkers. Additionally, controllable formation of different structures was successfully implemented by a combination of linkers with different steric effects at specific positions [101].

8.5.10 12-Connected Three-Dimensional Topological Networks Constructing metal complexes with higher connectivity to three-dimensional topological networks has always been a challenge for chemists. Recently,

8.6 Rod-Packing Topology Networks Position 2 –

OOC

COO–

R2 R1

Position 1 R1

–OOC

R2

COO–

+

flu

scu

csq

Figure 8.33 Porous coordination polymers PCN-608 with the (4,8)-connected csq network. Source: Based on Pang et al. [100].

Zhang et al. and Li have reported almost simultaneously metal complexes with 12-connected face-centered cubic topological networks (long Schläfli symbol: 324 436 56 ) based on high-core copper clusters [Cu3 (4-pyridinethiolate)2 (CN)] and [Cu12 (μ4 -SCH3 )6 (CN)6 ]n ⋅2H2 O [73]. As shown in Figure 8.34, the network structure of the complex [Cu3 (4-pyridinethiolate)2 (CN)] is based on Cu6 S4 cluster, which connects 12 Cu6 S4 clusters around it through eight 4-pyridinethiolate ligands and four CN ligands, forming a one face-centered cubic unit. The network structure of complex [Cu12 (μ4 -SCH3 )6 (CN)6 ]n ⋅2H2 O is based on Cu12 (μ4 -SCH3 )6 clusters, which are connected by 12 CN ligands to 12 surrounding clusters to form a face-centered cubic unit. This work provides an idea for constructing highly connected three-dimensional topological networks using high-core metal clusters as nodes. Du and coworkers chose In3+ as central metal ions, curved 2,5-thiophenedicarboxylic acid (H2 thb) with a bending angle of c. 156∘ as the wheel-formation agent, and 4,4′ -biphenyldicarboxylic acid (H2 pbdc) as the linker to achieve a 3D anionic porous metalloring organic framework [Me2 NH2 ][In(thb)(pbdc)]⋅2.5DMF⋅3.5H2 O (MROF-1). MROF-1 possesses striking topological features as follows: unprecedented sextuply interlocked nanocages, a 3D self-catenated framework, and an original 12-c topology net with the Schläfli symbol (460 ⋅66 ) (Figure 8.35) [102].

8.6 Rod-Packing Topology Networks The structure and performance of MOFs formed by SBUs have been extensively studied. When describing their network topology in detail, isolated SBUs are often regarded as nodes and organic ligands as connecting links. However, in many coordination polymers formed by metal–carboxylic acids, one-dimensional infinite rod-shaped construction units are often formed. O’Keffee and coworkers define these compounds as having a rod-packing topological network [103]. The connection of the points on a one-dimensional lattice forms simple rods, and

457

458

8 Structural Topologies and Interpenetration in the Coordination Polymers

(a)

(b)

(c)

Figure 8.34 12-Connected face-centered cubic topological network based on Cu6 S4 cluster. Source: Zhang et al. [73]. © 2005 American Chemical Society.

40°

(a)

(d)

(b)

(c)

Figure 8.35 MROF-1 possesses 12-connected striking topological feature. Source: Han et al. [102]. © 2016 Royal Society of Chemistry.

(e)

the accumulation of linear rods shows 13 topological structures as shown in Figure 8.36. In addition, the alternatives of simple rods are helix chains and double chains (ladders), which will lead to 3-connection and 4-connection networks, respectively. For example, in MOF-74-Zn, carboxylic acid and hydroxy group bridge Zn atoms to form a Zn–O–C helix chain. The rods consist of edge-sharing ZnO6 octahedra alternately opposite and next-neighbor to form either a 31 or a 32 helix. Then these rods are linked by the benzene units of the ligands to produce 3D framework (Figure 8.37) with bnn topology (Figure 36-1).

8.7 Conclusion

1

2

3

4

5

6

7

8

9

10

11

12

13

Figure 8.36 Invariant rod packings and their associated nets (1, bnn; 2, pcu; 3, kag; 4, hex; 5, ths-z; 6, cds; 7, bto-z; 8, qzd; 9, bmn; 10, nbo; 11, sgn; 12, gan; 13, utb-z). Source: Rosi et al. [103]. © 2005 American Chemical Society.

(a)

(b)

MOF-74 (c)

Figure 8.37 The rod-stacking topological network of MOF-74. Source: Rosi et al. [103]. © 2005 American Chemical Society.

8.7 Conclusion In the past 20 years, coordination polymer crystal engineering has experienced exponential development, showing unprecedented potential in the development of molecular magnets, microporous solids, heterogeneous catalysis, optics, conductivity, and chiral materials, and its rich and colorful topological network structure fully demonstrates the beauty of natural assembly. Nowadays, in the field of synthesis and material chemistry, the research focuses are exploring and utilizing

459

460

8 Structural Topologies and Interpenetration in the Coordination Polymers

many factors to influence the structure of complexes, constructing various new topological structures, and studying the relationship between the structure and physical properties of compounds. The ultimate goal is to assemble metal complexes with predefined topological structures and even expected functions to obtain the required molecular-based materials. To rationally design and synthesize functional metal complexes from ideal to reality, we can use the existing basic theory to explore the mechanism, structure, and micro-process of designated functions and the self-assembly method of molecular architecture, so as to find out the regularity of the structure and properties at the molecular level. Under the guidance of the research, according to the functional requirements, the molecular building blocks of the synthesized complexes were designed and assembled orderly to achieve the goal of directional construction.

References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

J. Becher, K. Schaumburg, 1995, Springer, Dordrecht. O. Kahn, 1996, Springer, Netherlands. Leonard V. Interrante, M. J. Hampden-Smith, 1997, Wiley-VCH, New York, NY. Kitagawa, S., Kitaura, R., and Noro, S. (2004). Angew. Chem. Int. Ed. 43: 2334. Lehn, J.-M. (1988). Angew. Chem. Int. Ed. 27: 89. Choudhury, A., Neeraj, S., Natarajan, S., and Rao, C.N.R. (2000). Angew. Chem. Int. Ed. 39: 3091. Yaghi, O.M., Li, H.L., Davis, C. et al. (1998). Acc. Chem. Res. 31: 474. Batten, S.R. and Robson, R. (1998). Angew. Chem. Int. Ed. 37: 1460. Férey, G. (2000). J. Solid State Chem. 152: 37. Kitagawa, S. and Kondo, M. (1998). Bull. Chem. Soc. Jpn. 71: 1739. Zaworotko, M.J. (1994). Chem. Soc. Rev. 23: 283. Yaghi, O.M., O’Keeffe, M., Ockwig, N.W. et al. (2003). Nature 423: 705. A. F. Wells, 1977, Wiley, New York, NY. A. F. Wells, 1975, Oxford University Press, Oxford. Wells, A.F. (1954). Acta Crystallogr. 7: 545. Wells, A.F. (1972). Acta Crystallogr. B 28: 711. Wells, A.F. (1969). Acta Crystallogr. B 25: 1711. Wells, A.F. (1968). Acta Crystallogr. B 24: 50. Wells, A.F. (1965). Acta Crystallogr. 18: 894. Wells, A.F. and Sharpe, R.R. (1963). Acta Crystallogr. 16: 857. Wells, A.F. (1956). Acta Crystallogr. 9: 23. Wells, A.F. (1955). Acta Crystallogr. 8: 32. Wells, A. (1954). Acta Crystallogr. 7: 535. Wells, A.F. (1954). Acta Crystallogr. 7: 842. Wells, A.F. (1954). Acta Crystallogr. 7: 849. Koch, E. and Fischer, W. (1995). Z. Kristallogr. 210: 407. O’Keeffe, M. (1991). Z. Kristallogr. 196: 21.

References

28 M. O’Keeffe, B. G. Hyde, 1996, Mineralogical Society of America: Washington, DC 29 Delgado-Friedrichs, O., O’Keeffe, M., and Yaghi, O.M. (2003). Acta Crystallogr. A 59: 22. 30 Delgado Friedrichs, O., O’Keeffe, M., and Yaghi, O.M. (2003). Solid State Sci. 5: 73. 31 Hoskins, B.F. and Robson, R. (1989). J. Am. Chem. Soc. 111: 5962. 32 Long, D.-L., Blake, A.J., Champness, N.R., and Schröder, M. (2000). Chem. Commun.: 1369. 33 Zhao, Y., Hong, M., Su, W. et al. (2000). J. Chem. Soc., Dalton Trans.: 1685. 34 Sun, J.K., Yao, Q.X., Ju, Z.F., and Zhang, J. (2010). CrystEngComm 12: 1709. 35 Sun, J.K., Yao, Q.X., Tian, Y.Y. et al. (2012). Chem. Eur. J. 18: 1924. 36 Shin, D.M., Lee, I.S., Chung, Y.K., and Lah, M.S. (2003). Inorg. Chem. 42: 5459. 37 Shin, D.M., Lee, I.S., and Chung, Y.K. (2003). Inorg. Chem. 42: 8838. 38 Liu, Y., Kravtsov, V.C., Beauchamp, D.A. et al. (2005). J. Am. Chem. Soc. 127: 7266. 39 Moulton, B., Lu, J., and Zaworotko, M.J. (2001). J. Am. Chem. Soc. 123: 9224. 40 Chen, S., Zhang, J., Wu, T. et al. (2010). Dalton Trans. 39: 697. 41 Nguyen, H.L., Gandara, F., Furukawa, H. et al. (2016). J. Am. Chem. Soc. 138: 4330. 42 Yu, M., Chen, L., Jiang, F. et al. (2017). Chem. Mater. 29: 8093. 43 Yaghi, O.M., Davis, C.E., Li, G.M., and Li, H.L. (1997). J. Am. Chem. Soc. 119: 2861. 44 Robinson, F. and Zaworotko, M.J. (1995). J. Chem. Soc., Chem. Commun.: 2413. 45 Dong, Y.-B., Smith, M.D., and zur Loye, H.-C. (2000). Inorg. Chem. 39: 4927. 46 Dai, J.-C., Wu, X.-T., Fu, Z.-Y. et al. (2002). Chem. Commun.: 12. 47 Wu, T., Yi, B.H., and Li, D. (2005). Inorg. Chem. 44: 4130. 48 Li, W.J., Liu, J., Sun, Z.H. et al. (2016). Nat. Commun. 7: 11830. 49 Cheng, J.K., Chen, Y.B., Wu, L. et al. (2005). Inorg. Chem. 44: 3386. 50 Carlucci, L., Ciani, G., Proserpio, D.M., and Rizzato, S. (2002). Chem. Eur. J. 8: 1519. 51 He, Y.P., Tan, Y.X., and Zhang, J. (2012). CrystEngComm 14: 6359. 52 Barnett, S.A., Blake, A.J., Champness, N.R., and Wilson, C. (2002). Chem. Commun.: 1640. 53 Moulton, B., Abourahma, H., Bradner, M.W. et al. (2003). Chem. Commun.: 1342. 54 Bhogala, B.R., Thallapally, P.K., and Nangia, A. (2004). Cryst. Growth Des. 4: 215. 55 Chen, B., Fronczek, F.R., and Maverick, A.W. (2003). Chem. Commun.: 2166. 56 Zhang, J.P., Zheng, S.L., Huang, X.C., and Chen, X.M. (2004). Angew. Chem. Int. Ed. 43: 206. 57 Hu, S. and Tong, M.-L. (2005). Dalton Trans.: 1165. 58 Luo, T.-T., Tsai, H.-L., Yang, S.-L. et al. (2005). Angew. Chem. Int. Ed. 44: 6063. 59 Su, C.-Y., Smith, M.D., Goforth, A.M., and zur Loye, H.-C. (2004). Inorg. Chem. 43: 6881.

461

462

8 Structural Topologies and Interpenetration in the Coordination Polymers

60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97

Hu, S., Chen, J.-C., Tong, M.-L. et al. (2005). Angew. Chem. Int. Ed. 44: 5471. Tong, M.-L., Chen, X.-M., and Batten, S.R. (2003). J. Am. Chem. Soc. 125: 16170. Ma, J.-F., Yang, J., Zheng, G.-L., and Li, J.-F.L. (2003). Inorg. Chem. 42: 7531. Tian, Y.Q., Cai, C.X., Ren, X.M. et al. (2003). Chem. Eur. J. 9: 5673. Tian, Y.Q., Zhao, Y.M., Chen, Z.X. et al. (2007). Chem. Eur. J. 13: 4146. Zhang, J.P. and Chen, X.M. (2006). Chem. Commun.: 1689. Huang, X.C., Lin, Y.Y., Zhang, J.P., and Chen, X.M. (2006). Angew. Chem. Int. Ed. 45: 1557. Phan, A., Doonan, C.J., Uribe-Romo, F.J. et al. (2010). Acc. Chem. Res. 43: 58. Banerjee, R., Phan, A., Wang, B. et al. (2008). Science 319: 939. Wang, F., Liu, Z.S., Yang, H. et al. (2011). Angew. Chem. Int. Ed. 50: 450. Zhang, J., Wu, T., Zhou, C. et al. (2009). Angew. Chem. Int. Ed. 48: 2542. Zhang, H.-X., Liu, M., Wen, T., and Zhang, J. (2016). Coord. Chem. Rev. 307: 255. Wu, T., Zhang, J., Zhou, C. et al. (2009). J. Am. Chem. Soc. 131: 6111. Zhang, X.-M., Fang, R.-Q., and Wu, H.-S. (2005). J. Am. Chem. Soc. 127: 7670. Zhang, H.X., Wang, F., Yang, H. et al. (2011). J. Am. Chem. Soc. 133: 11884. Wang, F., Shu, Y.-B., Bu, X., and Zhang, J. (2012). Chem. Eur. J. 18: 11876. Zhang, H.-X., Liu, M., Xu, G. et al. (2016). Chem. Commun. 52: 3552. Kang, Y., Wang, F., Zhang, J., and Bu, X. (2012). J. Am. Chem. Soc. 134: 17881. Dolomanov, O.V., Cordes, D.B., Champness, N.R. et al. (2004). Chem. Commun.: 642. Huang, X.C., Zhang, J.P., Lin, Y.Y. et al. (2004). Chem. Commun.: 1100. Wang, X.-L., Qin, C., Wang, E.-B. et al. (2005). Chem. Commun.: 5450. Liang, L., Liu, C., Jiang, F. et al. (2017). Nat. Commun. 8: 1233. Wang, Z., Kravtsov, V.C., and Zaworotko, M.J. (2005). Angew. Chem. Int. Ed. 44: 2877. Dybtsev, D.N., Chun, H., and Kim, K. (2004). Chem. Commun.: 1594. Zhang, H.-X., Fu, H.-R., Li, H.-Y. et al. (2013). Chem. Eur. J. 19: 11527. Pan, L., Ching, N., Huang, X., and Li, J. (2001). Chem. Commun.: 1064. Sun, H.-L., Ma, B.-Q., Gao, S., and Batten, S.R. (2005). Cryst. Growth Des. 5: 1331. Yang, S.Y., Long, L.S., Jiang, Y.B. et al. (2002). Chem. Mater. 14: 3229. Wen, Y.-H., Zhang, J., Wang, X.-Q. et al. (2005). New J. Chem. 29: 995. Huang, Y.G., Wu, S.Q., Deng, W.H. et al. (2017). Adv. Mater. 29. Kuang, X.F., Wu, X.Y., Yu, R.M. et al. (2010). Nat. Chem. 2: 461. Long, D.-L., Hill, R.J., Blake, A.J. et al. (2004). Angew. Chem. Int. Ed. 43: 1851. Wang, X.-L., Qin, C., Wang, E.-B., and Su, Z.-M. (2006). Chem. Eur. J. 12: 2680. Wang, X.-L., Qin, C., Wang, E.-B. et al. (2005). Chem. Commun.: 4789. Xiang, S.C., Hu, S.M., Sheng, T.L. et al. (2007). J. Am. Chem. Soc. 129: 15144. Yuan, D., Getman, R.B., Wei, Z. et al. (2012). Chem. Commun. 48: 3297. Batten, S.R., Hoskins, B.F., Moubaraki, B. et al. (1999). J. Chem. Soc., Dalton Trans.: 2977. Qin, C., Wang, X.-L., Wang, E.-B., and Su, Z.-M. (2005). Inorg. Chem. 44: 7122.

References

98 Chun, H., Kim, D., Dybtsev, D.N., and Kim, K. (2004). Angew. Chem. Int. Ed. 43: 971. 99 Pang, J., Jiang, F., Wu, M. et al. (2015). Nat. Commun. 6: 7575. 100 Pang, J., Yuan, S., Qin, J.S. et al. (2019). J. Am. Chem. Soc. 141: 3129. 101 Pang, J., Yuan, S., Qin, J. et al. (2017). J. Am. Chem. Soc. 139: 16939. 102 Han, Y.-H., Ye, Y., Tian, C. et al. (2016). J. Mater. Chem. A 4: 18742. 103 Rosi, N.L., Kim, J., Eddaoudi, M. et al. (2005). J. Am. Chem. Soc. 127: 1504.

463

465

9 Inorganic Chalcogenides: From Zero-Dimensional Clusters to Three-Dimensional Frameworks Hua Lin, Xin-Tao Wu, and Qi-Long Zhu Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, State Key Laboratory of Structural Chemistry, 155 Yangqiao Road West, Fuzhou 350002, China

9.1 Introduction Inorganic chalcogenides, an extensively studied class of functional materials that have unique electronic features and fascinating physical properties, are being evaluated for several applications, such as nonlinear optics (NLO), thermoelectricity (TE), magnetism, fluorescence, superconductivity, and photochemical catalysis. Among the various synthetic strategies, utilizing so-called “structural scissors,” i.e. alkali metals (A), alkaline-earth (AE) metals, rare-earth (RE) metals, or halogen elements (X), for constructing such materials has proven to be one of the most effective routes. From a structural point of view, these inorganic chalcogenides consist of two structural moieties, one of which is the negative moiety, anionic substructure constructed by the [MQn ] (M = metals; Q = chalcogen) polyhedra basic building units (BBUs) and the other is the positive moiety, discrete counteractive cations. In this chapter, selected 60 classical examples from the literature on a wide range of inorganic chalcogenides classes are presented, including centrosymmetric (CS) and non-centrosymmetric (NCS) crystal structures. These chalcogenides display a rich structure diversity based on the BBUs and can be divided into five classes according to the dimensional features: (i) the zero-dimensional (0D) discrete clusters, e.g. Ba12 In4 S19 [1], Ba23 Ga8 Sb2 S38 [2], Ba3 (BS3 )1.5 (MS3 )0.5 (M = Sb, Bi) [3], Ba3 (BQ3 )(SbQ3 ) (Q = S, Se) [3], (Cs6 Cl)2 Cs5 [Ga15 Ge9 Se48 ] [4], BaHgSe2 [5], Ba12 Sn4 S23 [6], Ba7 Sn3 S13 [6], Ba8 Sn4 S15 [7], and Ba4 M3 Q9 Cl2 (M = Si, Ge; Q = S, Se) [8] (see Section 9.2.1); (ii) one-dimensional (1D) chains, e.g. Ba4 M2 S8 (M = Ga, In) [1], Ln4 GaSbS9 (Ln = Pr, Nd, Sm, Gd—Ho) [9], La4 InSbS9 [10], Ba3 La4 Ga2 Sb2 S15 [11], BaLa3 GaSb2 S10 [11], Ba8 Zn4 Ga2 S15 [12], A4 Ge4 Se12 (A = Rb, Cs) [13], BaGeOSe2 [14], Ba5 In4 Te4 S7 [15], Ba8 Ga2 Sn7 Se18 [16], and Ba10 Ga2 Sn9 Se22 [16] (see Section 9.2.2); (iii) two-dimensional (2D) layers, e.g. La4 FeSb2 Q10 (Q = S, Se) [17], RE2 Mn3 Sb4 S12 (RE = Pr, Nd, Sm, Gd) [18], CsRECdTe3 (RE = La, Pr, Nd, Sm, Gd—Tm, and Lu) [19], (Cs6 Cl)6 Cs3 [Ga53 Se96 ] [20], Cs2 [Mn2 Ga3 S7 Cl] [21], Ba2 Cr4 GeSe10 [22], A2 Ge4 Se10 (A = Rb, Cs) [23], Na6 Zn3 III2 Q9 (III = Ga, In; Q = S, Advanced Structural Chemistry: Tailoring Properties of Inorganic Materials and their Applications, First Edition. Edited by Rong Cao. © 2021 WILEY-VCH GmbH. Published 2021 by WILEY-VCH GmbH.

466

9 Inorganic Chalcogenides: From Zero-Dimensional Clusters to Three-Dimensional Frameworks

Se) [24], RECuTe2 (RE = Tb–Tm) [25], CsMnInTe3 [26], Ba4 F4 MGa2 S6 (M = Cr, Mn, Fe) [27], Ba4 F4 MnIn2S6 [27] (see Section 9.2.3); (iv) three-dimensional (3D) frameworks, e.g. PbMnIn2 S5 [28], Ax RE2 Cu6−x Te6 (A = K–Cs; RE = La–Nd) [29], CsRE2 Ag3 Te5 (RE = Pr, Nd, Sm, Gd–Er) [30], RbRE2 Ag3 Te5 (RE = Sm, Gd–Dy) [30], Cs[RE9 Mn4 Se18 ] (RE = Ho–Lu) [31], Cs[RE9 Cd4 Se18 ] (RE = Tb–Tm) [32], Cs3 Cu20 Te13 [33], A–II4 –III5 –Q12 (A = K, Rb, Cs; II = Mn, Zn, Cd, Hg; III = Ga, In; Q = S, Se, Te) [34], A–III–Sn2 –Se6 (A = Rb, Cs; III = Ga, In) [35], Ba3 AGa5 Se10 Cl2 (A = K, Rb, Cs) [36], Ba4 MGa4 Se10 Cl2 (M = Mn, Zn, Cd, Cu/Ga, Co, and Fe) [37], Ba6 Li2 CdSn4 S16 [38], La2 CuSbS5 [39], Ba6 Zn7 Ga2 S16 [40], Na2 In4 SSe6 [41], NaGaIn2 Se5 [41], NaIn3 Se5 [41], PbGa2 MSe6 (M = Si, Ge) [42], SnGa4 Q7 (Q = S, Se) [43], Na2 ZnGe2 S6 [44], Na2 BaSnS4 [45], Ba–Na2 –IV–Q4 (IV = Ge, Sn; Q = S, Se) [46], Ba–Li2 –IV–Q4 (IV = Ge, Sn; Q = S, Se) [46], CsAg5 Te3 [47], Ba5 Cu8 In2 S12 [48], Cs[Lu7 Q11 ] (Q = S, Se) [49], (ClCs6 )[RE21 Q34 ] (RE = Dy, Ho; Q = S, Se, Te) [49], Yb6 Ga4 S15 [50], Lu5 GaS9 [50], BaAg2 MS4 (M = Ge, Sn) [51], Sr5 ZnGa6 S15 [52], Cd4 GeS6 [53], Cs2 [RE8 InS14 ] (RE = Ho–Lu) [54], Na7 IISb5 S12 (II = Zn, Cd, Hg) [55], Cs2 Ge3 M6 Te14 (M = Ga, In) [56], Na2 Ga2 MQ6 (M = Ge, Sn; Q = S, Se) [57], and Sn2 Ga2 S5 [58] (see Section 9.2.4); and (v) mixed-dimensional (MD) structures, e.g. RE3 M0.5 (Ge0.5 /M0.5 )S7 (RE = La, Sm; M = Ga, In) and RE3 In0.33 GeS7 (RE = La, Sm, Gd) [59] and CsCu5 S3 [60] (see Section 9.2.5). Summary of the selected inorganic chalcogenides from 0D to MD is listed in Table 9.1. In this chapter, we focus on the unit cell, space group, and dimensional change as well as the structural assembly of selected chalcogenides.

9.2 Inorganic Chalcogenides 9.2.1

Zero-Dimensional (0D) Cluster Chalcogenides

9.2.1.1 Ba12 In4 S19

Ba12 In4 S19 belongs the trigonal CS space group R3 (No. 148) with a = 9.6182(5) Å, b = 9.6182(5) Å, c = 75.393(7) Å and features a unique long period stacking structure of a combination of monometallic [InS4 ] tetrahedra, linear dimeric [In2 S7 ] polyhedra, disulfide S2 2− anions, and isolated S2− anions, which are surrounded by Ba cations (Figure 9.1a) [1]. The long stacking period of Ba12 In4 S19 is characterized by the very large c parameter of about 75 Å, which is about eight times larger than the other two unit cell parameters (a and b). Such an anionic repeat unit is also described as a “column,” as shown in Figure 9.1b by a green or a pink cylinder. The linear dimeric [In2 S7 ] polyhedra are constructed by the central base sharing In2/In3 bi-tetrahedra that are fused with [In4S4 ] tetrahedra on vertexes of S6, respectively. In3 and In2 are disordered over two crystallographic independent sites with occupancies of 79.7% and 20.3%, respectively. The In2–In3 bond length is 1.062(8) Å. Note that the coordinates of the principal axial atoms in the linear dimeric [In2 S7 ] anion and [InS4 ] anion (namely, In1, S4, In2, In3, S6, and In4 atoms) differ only in c-axis, which indicate that they are lined up in a straight line.

9.2 Inorganic Chalcogenides

Table 9.1

467

Summary of the selected inorganic chalcogenides from 0D to MD.

Dimension

Figure number

Chemical formula

Unit cell

Space group

Ba12 In4 S19

Trigonal

R3 (No. 148)

0D

Figure 9.1

Ba23 Ga8 Sb2 S38

Orthorhombic

Cmc21 (No. 36)

0D

Figure 9.2

Ba3 (BS3 )1.5 (MS3 )0.5

Monoclinic

P21 /c (No. 14)

0D

Figure 9.3

Ba3 (BQ3 )(SbQ3 )

Hexagonal

P62 m (No. 189)

0D

Figure 9.3

(Cs6 Cl)2 Cs5 [Ga15 Ge9 Se48 ]

Tetragonal

I4/m (No. 87)

0D

Figure 9.4

BaHgSe2

Orthorhombic

Pmc21 (No. 26)

0D

Figure 9.5

Ba12 Sn4 S23

Monoclinic

P21 /c (No. 14)

0D

Figure 9.6

Ba7 Sn3 S13

Orthorhombic

Pnma (No. 62)

0D

Figure 9.6

Ba8 Sn4 S15

Orthorhombic

Pca21 (No. 29)

0D

Figure 9.7

Ba4 M3 Q9 Cl2

Hexagonal

P63 (No. 173)

0D

Figure 9.8

Ba4 In2 S8

Triclinic

P1 (No. 2)

1D

Figure 9.9

Ba4 Ga2 S8

Monoclinic

P21 /c (No. 14)

1D

Figure 9.9

RE4 GaSbS9

Orthorhombic

Aba2 (No. 41)

1D

Figure 9.10

La4 InSbS9

Tetragonal

P43 21 2 (No. 96)

1D

Figure 9.11

Ba3 La4 Ga2 Sb2 S15

Orthorhombic

Ibam (No. 72)

1D

Figure 9.12

BaLa3 GaSb2 S10

Monoclinic

P21 /m (No. 11)

1D

Figure 9.12

Ba8 Zn4 Ga2 S15

Monoclinic

P21 /n (No. 11)

1D

Figure 9.13

A4 Ge4 Se12

Orthorhombic

Pna21 (No. 33)

1D

Figure 9.14

BaGeOSe2

Orthorhombic

P21 21 21 (No. 19)

1D

Figure 9.15

Ba5 In4 Te4 S7

Orthorhombic

Imm2 (No. 44)

1D

Figure 9.16

Ba8 Ga2 Sn7 Se18

Orthorhombic

Pnma (No. 62)

1D

Figure 9.17

Ba10 Ga2 Sn9 Se22

Orthorhombic

Cmc21 (No. 36)

1D

Figure 9.17

La4 FeSb2 Q10

Orthorhombic

Pbcm (No. 57)

2D

Figure 9.18

RE2 Mn3 Sb4 S12

Monoclinic

C2/m (No. 12)

2D

Figure 9.19

CsRECdTe3

Orthorhombic

Cmcm (No. 63)

2D

Figure 9.20

(Cs6 Cl)6 Cs3 [Ga53 Se96 ]

Trigonal

R3m (No. 166)

2D

Figure 9.21

Cs2 [Mn2 Ga3 S7 Cl]

Orthorhombic

Pnma (No. 62)

2D

Figure 9.22

Ba2 Cr4 GeSe10

Triclinic

P1 (No. 2)

2D

Figure 9.23

A2 Ge4 Se10

Monoclinic

Pc (No. 7)

2D

Figure 9.24

Na6 Zn3 III2 Q9

Monoclinic

C2/c (No. 15)

2D

Figure 9.25

RECuTe2

Trigonal

P3m1 (No. 164)

2D

Figure 9.26

CsMnInTe3

Monoclinic

C2/c (No. 15)

2D

Figure 9.27

Ba4 F4 XGa2 S6 and Ba4 F4 MnIn2 S6

Orthorhombic

Pnma (No. 62)

2D

Figure 9.28

Ba4 F4 MnGa2 S6

Orthorhombic

Cmca (No. 64)

2D

Figure 9.28

PbMnIn2 S5

Orthorhombic

Cmcm (No. 63)

3D

Figure 9.29

Ax RE2 Cu6−x Te6

Hexagonal

P63 /m (No. 176)

3D

Figure 9.30

CsRE2 Ag3 Te5 and RbRE2 Ag3 Te5

Orthorhombic

Cmcm (No. 63)

3D

Figure 9.31 (Continued)

468

9 Inorganic Chalcogenides: From Zero-Dimensional Clusters to Three-Dimensional Frameworks

Table 9.1

(Continued)

Space group

Dimension

Figure number

Trigonal

R3 (No. 148)

3D

Figure 9.32

Trigonal

R3 (No. 148)

3D

Figure 9.33

Chemical formula

Unit cell

Cs[RE9 Mn4 Se18 ] Cs[RE9 Cd4 Se18 ] Cs3 Cu20 Te13

Cubic

Fm3 (No. 202)

3D

Figure 9.34

A–II4 –III5 –Q12 type

Trigonal

R3 (No. 146)

3D

Figure 9.35

A–III–Sn2 –Se6 type

Trigonal

R3 (No. 146)

3D

Figure 9.36

Ba3 AGa5 Se10 Cl2 and its derivatives

Tetragonal

I4 (No. 82)

3D

Figure 9.37

Ba6 Li2 CdSn4 S16

Cubic

I43d (No. 220)

3D

Figure 9.38

La2 CuSbS5

Orthorhombic

Ima2 (No. 46)

3D

Figure 9.39

Ba6 Zn7 Ga2 S16

Trigonal

R3 (No. 146)

3D

Figure 9.40

Na2 In4 SSe6

Orthorhombic

Pca21 (No. 29)

3D

Figure 9.41

NaGaIn2 Se5

Trigonal

R32 (No. 155)

3D

Figure 9.41

NaIn3 Se5

Trigonal

P32 (No. 145)

3D

Figure 9.41

PbGa2 MSe6

Monoclinic

Cc (No. 9)

3D

Figure 9.42

SnGa4 Q7

Monoclinic

Pc (No. 7)

3D

Figure 9.43

Na2 ZnGe2 S6

Monoclinic

Cc (No. 9)

3D

Figure 9.44

Na2 BaSnS4

Tetragonal

I42d (No. 112)

3D

Figure 9.45

Ba–Na2 –IV–Q4 type

Trigonal

R3c (No. 161)

3D

Figure 9.45

Ba–Li2 –IV–Q4 type

Tetragonal

I42m (No. 121)

3D

Figure 9.46

CsAg5 Te3

Tetragonal

P42 /mcm (No. 146)

3D

Figure 9.47

Ba5 Cu8 In2 S12

Monoclinic

C2/c (No. 15)

3D

Figure 9.48

Cs[Lu7 Q11 ]

Orthorhombic

Cmca (No. 64)

3D

Figure 9.49

(ClCs6 )[RE21 Q34 ]

Monoclinic

C2/m (No. 12)

3D

Figure 9.49

Yb6 Ga4 S15

Monoclinic

C2/m (No. 12)

3D

Figure 9.50

Lu5 GaS9

Triclinic

P1 (No. 2)

3D

Figure 9.50

BaAg2 GeS4

Tetragonal

I42m (No. 121)

3D

Figure 9.51

BaAg2 SnS4

Tetragonal

I222 (No. 23)

3D

Figure 9.51

Sr5 ZnGa6 S15

Orthorhombic

Ama2 (No. 40)

3D

Figure 9.52

Cd4 GeS6

Monoclinic

Cc (No. 9)

3D

Figure 9.53

Cs2 [RE8 InS14 ]

Orthorhombic

Cmca (No. 64)

3D

Figure 9.54

Na7 IISb5 S12

Trigonal

R3 (No. 148)

3D

Figure 9.55

Cs2 Ge3 M6 Te14

Trigonal

P3m1 (No. 164)

3D

Figure 9.56

Na2 Ga2 MQ6

Orthorhombic

Fdd2 (No. 43)

3D

Figure 9.57

Sn2 Ga2 S5

Monoclinic

Cc (No. 9)

3D

Figure 9.58

Ln3 M0.5 (Ge0.5 /M0.5 )S7 and Ln3 In0.33 GeS7

Hexagonal

P63 (No. 173)

MD

Figure 9.59

o-CsCu5 S3

Orthorhombic

Pmma (No. 51)

MD

Figure 9.60

t-CsCu5 S3

Tetragonal

P421 c (No. 114)

MD

Figure 9.60

S10

In5

In4

In3

In1

S22–

S12

S11

S22–

S12

S10

9.2 Inorganic Chalcogenides

b a

(a)

c = 75.393(7) Å

(b)

Figure 9.1 (a) Structure of Ba12 In4 S19 viewed down the ab-plane with the unit cell marked. (b) Side view of “column” in Ba12 In4 S19 . Source: Liu et al. [1]. © 2011 American Chemical Society.

c b a

Figure 9.2 Structure of Ba23 Ga8 Sb2 S38 with the unit cell marked. The Ba—S bonds are omitted for a better view. Pink, Ba; dark-cyan, Ga; orange, Sb; yellow, S; light-blue tetrahedra: [GaS4 ]. Source: Chen et al. [2]. © 2012 American Chemical Society.

9.2.1.2 Ba23 Ga8 Sb2 S38

Ba23 Ga8 Sb2 S38 crystallizes in the orthorhombic NCS space group Cmc21 (No. 36) with a = 9.6111(3) Å, b = 32.25(2) Å, c = 12.883(4) Å and features an unique 0D structure containing totally discrete [GaS4 ] tetrahedra and isolated [SbS3 ] pyramids with Ba cations located between them (Figure 9.2) [2]. Such an NCS building motif containing a matrix of isolated [GaS4 ] tetrahedra has never been reported in the ternary A(AE)/Ga/S compounds. In addition, partial occupancies of the Wyckoff sites 8b (Ba8), 4a (Ba9), and 8b (Sb) are observed in the title compound. The distortion of the [SbS3 ] pyramid is evidenced by deviations of both the Sb—S bonds (2.42–2.47 Å) and the S–Sb–S angles (95–99∘ ), which are attributable to the stereo-chemically active lone pair (SCALP) on Sb. In spite of its local asymmetry, the sum polarization of [SbS3 ] units coming from the lone pairs is zero because the c glide plane at (x, 0, z) correlates with the [SbS3 ] units and the lone pairs of neighboring [SbS3 ] units are restricted to be oppositely oriented. In contrast, all of the distorted [GaS4 ] tetrahedra on which the 21 screw axis operates generate a net dipole moment along the c-axis, as expected from the Cmc21 space group symmetry. Although the [SbS3 ] pyramid is a minor component in quantity and hardly contributes to the overall polarity, it helps to disconnect the [GaS4 ] building units, breaking the symmetry and leading to the NCS packing.

469

470

9 Inorganic Chalcogenides: From Zero-Dimensional Clusters to Three-Dimensional Frameworks

CS

NCS

(BS3):(SbS3) = 3 : 1

(BS3):(SbS3) = 2 : 2

a

Ba3(BS3)1.5(SbS3)0.5

(a)

Figure 9.3 Structures of (a) Ba3 (BS3 )1.5 (SbS3 )0.5 ; (b) Ba3 (BS3 )(SbS3 ) with unit cell marked. Blue, Ba; purple, Sb; bright cyan, B; yellow, S. Source: Li et al. [3]. © 2015 American Chemical Society.

b

b

c

Ba B Sb S

c

a

Ba3(BS3)(SbS3)

(b)

9.2.1.3 Ba3 (BS3 )1.5 (MS3 )0.5 (M = Sb, Bi) and Ba3 (BQ3 )(SbQ3 ) (Q = S, Se)

Ba3 (BS3 )1.5 (MS3 )0.5 (M = Sb, Bi) adopts the monoclinic CS space group P21 /c (No. 14) with a Pearson’s symbol of mP88. Take Ba3 (BS3 )1.5 (SbS3 )0.5 as an example, it possesses a 0D structure formed by discrete [BS3 ] trigonal planes and isolated [SbS3 ] pyramids with Ba cations around them (Figure 9.3a) [3]. Each B atom has a trigonal plane coordination with three S atoms, and the distortions of the [BS3 ] trigonal planes are evidenced by deviations of both B—S bonds and S–B–S angles. Characteristically, these polarized [BS3 ] trigonal planes exhibit delocalized π-bonding formed by the nonbonding pz -orbitals of every S atom that are perpendicular to the trigonal plane. Nevertheless, the overall polarities of [BS3 ] units are canceled by the inversion center symmetry operation. The discrete [SbS3 ] pyramid is also distorted due to the SCALP electrons on Sb atom. The Ba1 , Ba2 , Ba4 , Ba6 ions are eightfold coordinated to S atoms, and Ba3 and Ba5 are ninefold coordinated to S atoms, respectively. Ba3 (BQ3 )(SbQ3 ) (Q = S, Se) belongs to the hexagonal NCS P62 m (No. 189). They are also a 0D structure constructed by discrete [BQ3 ] trigonal planes and isolated [SbQ3 ] pyramids. As shown in Figure 9.3b, every two isolated [BS3 ] or [Sb3S3 ] pyramids along the c-axis are anti-aligned in compound Ba3 (BS3 )(SbS3 ). The B1 located at 3g site and B2, B3 located at 3f site with m2m site symmetry are in typical trigonal plane coordination with three S atoms. The Sb1 (Wyckoff site 3g; occupancies: 0.830(2)) atom has a pyramidal geometry which is coordinated to two S2 and one S1 atoms. The Ba atoms show typical eight- or ninefold coordination with normal Ba—S bond length. 9.2.1.4 (Cs6 Cl)2 Cs5 [Ga15 Ge9 Se48 ]

(Cs6 Cl)2 Cs5 [Ga15 Ge9 Se48 ] belongs to the tetragonal CS space group I4/m (No. 87) with a = b 14.387(5) Å, c = 28.012(2) Å and possesses unprecedented isolated [M24 Se48 ]15− (M = Ga/Ge) supercubooctahedra, which are well cut by [Cs6 Cl]5+ complex cations and discrete Cs cations (Figure 9.4a) [4]. Such [M24 Se48 ]15− supercubooctahedra consists of 72 atoms, has an edge length of about 10.8 Å (Figure 9.4b), and contains 17 atoms more than the T4-[M20 Q35 ] cluster. Each of the eight vertices is in fact an [M3 Se9 ] secondary building unit (SBU), which is linked to the neighboring vertex through sharing Se–Se edges, and such an [M3 Se9 ] unit is commonly regarded as a T2 cluster without a corner. In complete contrast to the normal, Tn, Pn, or Cn clusters, such a cluster shows a novel type of supertetrahedral

9.2 Inorganic Chalcogenides a c

(b) a c

(a)

(c)

Figure 9.4 (a) Structure of (Cs6 Cl)2 Cs5 [Ga15 Ge9 Se48 ] viewed slightly along the ac-plane with the unit cell marked. (b) The supercubooctahedron [M24 Se48 ]15− (M = Ga/Ge) with the [M3 Se9 ] secondary building unit highlighted in blue. Light blue: Cs1 atom, green: Cs4 atoms, orange: Cs5 atom. (c) Interstitial [Cs6 Cl]5+ complex cation. Source: Huangfu et al. [4]. © 2015 John Wiley & Sons.

connectivity. The center of the [MSe4 ] tetrahedron is a disordered Ga/Ge atom with occupancies of about 0.62/0.38. (Cs6 Cl)2 Cs5 [Ga15 Ge9 Se48 ] contains five crystallographically unique Cs atoms. According to their coordination environment, the Cs atoms can be classified into three categories: Cs5 centering (001) and [001], Cs4 distributed around Cs5 on the faces parallel to (001) at c = 0, 0.5, and Cs1, Cs2, and Cs3, which form octahedral ([Cs6 Cl]5+ ) around the interstitial Cl anion (Figure 9.4b,c). 9.2.1.5 BaHgSe2

BaHgSe2 crystallizes in the orthorhombic NCS space group Pmc21 (No. 26) with unit cell parameters a = 4.3580(9) Å, b = 14.881(3) Å, c = 7.5900(15) Å, and Z = 4 [5]. In the asymmetric unit, there are two crystallographically independent Ba atoms (Wyckoff site, 2a), two Hg atoms (Wyckoff site, 2b), and four Se atoms (Wyckoff sites, 2a and 2b). Both Ba cations are surrounded by seven Se atoms with the Ba–Se distances between 3.274(3) and 3.427(3) Å. Hg1 cations are coordinated to three Se atoms with Hg—Se bond lengths ranging from 2.555(3) to 2.642(2) Å and a fourth Se at a much longer distance of 3.133(3) Å (Figure 9.5a). It is not obvious if the coordination geometry around Hg1 should be considered as trigonal pyramidal, which was also inappropriately described as “tetrahedral” (coordination number [CN] = 4) in a redetermination of the analogous sulfide BaHgS2 . The evidence suggests that a trigonal planar (CN = 3) geometry is more appropriate, according to the analysis of the crystal orbital Hamilton population (COHP) curves. As shown in Figure 9.5d, the integrated crystal orbital Hamilton population (–ICOHP) values for the three shorter Hg–Se contacts are far greater than the value for the longer one (2.1 eV/bond versus 0.3 eV/bond). Hg-centered trigonal planar units

471

9 Inorganic Chalcogenides: From Zero-Dimensional Clusters to Three-Dimensional Frameworks Se4

c

4

3.133(3)

Se4

(a)

0

–2

–4

Hg1–Se4(Long)

2

2.555(3)

Hg1–Se4(Short)

a

2.642(2)

Hg1–Se2

Se2

b

Hg2–Se

Se2

2.642(2)

Ba–Se

Hg1

Energy (eV)

472

–6 Se1 2.470(3)

(b)

Hg2

Hg Ba Se

Se3 2.467(3)

(c)

–8 –6 –3

0

3

6 –2 –1 0 1 2–2 –1 0 1 2 –1.0–0.5 0.0 0.5 1.0

–COHP (cell–1)

(d)

Figure 9.5 The coordination geometry of (a) Hg1 atom, (b) Hg2 atom, (c) the crystal structure of BaHgSe2 , and (d) the COHP curves for Ba—Se and Hg—Se bonds. Source: Li et al. [5]. © 2016 American Chemical Society.

are frequently seen in metal complexes, and a very recent study further proves its stable existence in inorganic hydrates. The Hg2 atom is surrounded at distances of 2.467(3)–2.470(3) Å (Figure 9.5b), which corresponds to strong bonds as confirmed by an –ICOHP value of 3.0 eV/bond. 9.2.1.6 Ba12 Sn4 S23 and Ba7 Sn3 S13

Ba12 Sn4 S23 crystallizes in the monoclinic CS space group P21 /c (No. 14) with a = 13.406(5) Å, b = 12.423(6) Å, c = 26.85(2) Å, 𝛽 = 95.820(3)∘ , and Z = 4 [6]. And Ba7 Sn3 S13 crystallizes in the orthorhombic CS space group Pnma (No. 62) with a = 12.386(5) Å, b = 24.17(2) Å, c = 8.872(4) Å and Z = 4. As shown in Figure 9.6, both of them feature 0D structures which are constructed by isolated [SnS4 ] tetrahedra with discrete Ba2+ cations and S2− anions (isolated S2 in Ba12 Sn4 S23 and S20 in Ba7 Sn3 S13 ) filling in them. However, in compound Ba12 Sn4 S23 (Figure 9.6a), there are S—S bonds existing which are built by S1–S2 (2.104(5) Å), S4–S18 (2.090(5) Å) and S10–S13 (2.087(5) Å). Then, the formula of Ba12 Sn4 S23 and Ba7 Sn3 S13 can be written as {(Ba2+ )12 [(SnS4 )4− ]4 [(S2 )2− ]3 (S2− )} and {(Ba2+ )7 [(SnS4 )4− ]3 (S2− )}. The lengths of S—S bonds or isolated S2− anions in the title compounds are commonly observed in other chalcogenides. In compound Ba12 Sn4 S23 , there are 4 crystallographically Sn atoms, 23 S atoms, and 12 Ba atoms. Among them, the Ba8 and Ba12 sites are deficient with the occupancies of 87% and 91%, and they are all disordered with the Ba8′ and Ba12′ with 13% and 9% occupancies, respectively. Interestingly, for Ba atoms, there are four different coordination environments [BaS6 ], [BaS7 ], [BaS8 ], and [BaS9 ]. And the distances of Ba–S are ranging from 3.026(4) to 3.644(6) Å. The Sn atoms are surrounded by four S atoms to form [SnS4 ] tetrahedra with the lengths of Sn–S ranging from 2.331(4) to 2.398(3) Å. In compound Ba7 Sn3 S13 , there are two crystallographically Sn, eight S atoms, and four Ba atoms. Among them, the Ba2 and Ba4 sites are disordered with Ba2′ and Ba4′ and the occupancies are 89%, 93%, 11%, and 6%, respectively. The coordination environments for Ba atoms in Ba7 Sn3 S13 are [BaS6 ], [BaS8 ] and [BaS9 ] with the lengths of Ba–S ranging from 3.067(4) to 3.669(7) Å. In addition, [SnS4 ] tetrahedra

9.2 Inorganic Chalcogenides

Ba

Sns4

S

S–S

a c (a)

a c (b)

b

Figure 9.6 View of the 0D structures of (a) Ba12 Sn4 S23 and (b) Ba7 Sn3 S13 . Source: Duan et al. [6]. © 2017 Royal Society of Chemistry.

are built by Sn connected with four S atoms with the lengths of bond Sn—S varying from 2.338(4) to 2.410(5) Å. 9.2.1.7 Ba8 Sn4 S15

Ba8 Sn4 S15 crystallizes in the polar space group Pca21 (No. 29) and features a unusual NCS 0D structure based on the completely isolated [SnS4 ] tetrahedra and [SnS3 ] pyramids (these are considered as first-order basic structural units) with Ba2+ cations inserting between them for charge balance (Figure 9.7a) [7]. For a clearly illustration, the secondary basic structure unit of Ba8 Sn4 S15 is regarded as a coin-like structure (Figure 9.7b). Interestingly, the coin-like basic structure units form the 2D honeycomb structure by alternate arrangement in parallel. The skeleton of the honeycomb is formed by 6 [SnS3 ] pyramidal units and 18 Ba atoms in ordered arrangement on the edge. And isolated nine [SnS4 ] tetrahedral units and eight Ba atoms fill in the cells of the honeycomb (Figure 9.7c). In the asymmetric unit, there are 16 crystallographically unique Ba atoms, 8 Sn atoms, and 30 S atoms. Ba atoms in the unit

473

474

9 Inorganic Chalcogenides: From Zero-Dimensional Clusters to Three-Dimensional Frameworks

b c a (a)

c

a c (b)

(c)

Figure 9.7 (a) View of the honeycomb structure of Ba8 Sn4 S15 along the ac-plane. (b) The coin-like basic structure unit (a) viewed down the ac-plane, and the center of coin (b) viewed down the bc-plane. The Ba—S bonds have been omitted for clarity. Black, Ba; red, Sn; yellow, S; red tetrahedron, [SnS4 ]. Source: Luo et al. [7]. © 2014 American Chemical Society.

cell present three types of coordination environment: Ba1, Ba2, Ba4, Ba5, Ba6, Ba7, Ba9, and Ba11 atoms are eightfold ligated by S atoms; Ba8, Ba12, Ba14, and Ba16 are sevenfold encompassed by S atoms; and Ba3, Ba10, Ba13, and Ba15 atoms are sixfold linked by S atoms. There are two types of coordination environment of Sn atoms in Ba8 Sn4 S15 : tetravalent Sn1–Sn6 atoms are surrounded to the tetrahedron by four S atoms, while the divalent Sn7 and Sn8 atoms are coordinated by three S atoms in a pyramidal geometry with the 5s2 electron lone-pair pointing toward the apex of pyramid. The Sn–S distances range from 2.3373(16) to 2.4124(16) Å in the [SnS4 ] tetrahedra and 2.5336(15) to 2.6817(16) Å in the [SnS3 ] pyramids. 9.2.1.8 Ba4 M3 Q9 Cl2 (M = Si, Ge; Q = S, Se)

All of the title compounds crystallize in the polar hexagonal space group P63 (No. 173) with a Pearson’s symbol of hP36 and cell parameters of a = b = 9.8200(4)– 10.0915(4) Å and c = 12.0544(12)–12.5091(8) Å [8]. For simplicity, the crystal structure of Ba4 Ge3 S9 Cl2 will be discussed in detail as a representative. There are two crystallographically unique Ba atoms (Ba1, site symmetry 6c; Ba2, site symmetry 2b), two crystallographically independent Cl atoms (Cl1, site symmetry 2b; Cl2, site symmetry 2a), one independent Ge atom (Ge, site symmetry 6c), and three different S atoms (S1, S2, S3, site symmetry 6c) in one asymmetric unit (Figure 9.8a,b). Ba1 atoms are fivefold coordinated to S and Cl atoms, whereas Ba2 atoms are sevenfold coordinated to S1, S2, and Cl1 atoms, respectively. Each Ge atom is coordinated in a slightly distorted [GeS4 ]4− tetrahedra with Ge—S bonds ranging from 2.1609(12) to 2.2674(8) Å and S–Ge–S angles ranging from 102.45(4)∘ to 114.52(4)∘ . The isolated [Ge3 S9 ]6− ring is formed by Q corner-sharing [GeS4 ]4− tetrahedra, and two [Ge3 S9 ]6− rings in a unit cell are further arranged around a 63 screw axis with all Ge—S1 bonds parallel to the c-axis (Figure 9.8c). Therefore, the [Ge3 S9 ]6− rings could be viewed as pseudo-layers which are stacked through the 63 screw axis to build up the structure with Ba2+ and Cl− atoms occupying the interspaces.

9.2 Inorganic Chalcogenides

C11

Ba Ge S Cl

Ba1 C12

Ba2 S1 c

S2

S3

a

(a)

(b)

b

a

b

c 63 axis

(c)

Figure 9.8 View of the structure of Ba4 Ge3 S9 Cl2 viewed along (a) [110] direction and (b) [001] direction. (c) The stack schematic of isolated [Ge3 S9 ]6− rings along c-axis. The 63 screw axis is visualized as a thick olive line. Black, Ba; pink, Ge; yellow, S; blue, Cl. Source: Liu et al. [8]. © 2017 Royal Society of Chemistry.

9.2.2

One-Dimensional (1D) Chain Chalcogenides

9.2.2.1 Ba4 In2 S8 and Ba4 Ga2 S8

Ba4 M2 S8 (M = Ga, In) represents the first 1D chain structure example in the ternary Ba/M/S systems. Ba4 M2 S8 exhibits the same stoichiometry, but it crystallizes in different space groups, P1 (for M = Ga) versus P21 /c (M = In) [1]. Ba4 In2 S8 crystallizes in the triclinic space group P1 with a = 6.236(2) Å, b = 10.014(4) Å, c = 13.033(5) Å, 𝛼 = 104.234(6), 𝛽 = 90.412(4), 𝛾 = 91.052(6), and Z = 2 (Figure 9.9a), while Ba4 Ga2 S8 crystallizes in the monoclinic space group P21 /c with a = 12.739(5) Å, b = 6.201(2) Å, c = 19.830(8) Å, 𝛽 = 104.254(6), and Z = 4 (Figure 9.9b). Nevertheless, they have similar unique 1D wavy-like chains of [MS4 ] tetrahedra that differ only in propagation directions (a in Ba4 In2 S8 versus b in Ba4 Ga2 S8 ) and wavy-like M1–M2–M1 angles (101.8∘ in Ba4 In2 S8 versus 104.8∘ in Ba4 Ga2 S8 ). Similar as the known Ba/M/S (M = Ga, In) compounds which adopt either 3D, 2D, or 0D structures, the primary building units in Ba4 M2 S8 are also [MS4 ] tetrahedra; their new 1D chain motif may be owed to the involvement of the disulfide S2 2− anion in the crystal structures. The isolated disulfide S2 2− anions are parallel to the 1D chain with S–S distances of 2.117(4) Å in Ba4 In2 S8 and 2.115(2) Å in Ba4 Ga2 S8 . The charge balanced formula of Ba4 M2 S8 can be described as Ba4 M2 S8 ≡ (Ba2+ )4 (M3+ )2 (S2− )6 (S2 )2− . 9.2.2.2 Ln4 GaSbS9 (Ln = Pr, Nd, Sm, Gd, Tb, Dy, Ho)

Ln4 GaSbS9 (Ln = Pr, Nd, Sm, Gd, Tb, Dy, Ho) belongs to the orthorhombic NCS space group, Aba2 (No. 41) [9]. The structure of Sm4 GaSbS9 viewed along the

475

476

9 Inorganic Chalcogenides: From Zero-Dimensional Clusters to Three-Dimensional Frameworks

S22–

S22–

S22–

S22–

S22–

b

S22–

a c

c (a)

(b)

Figure 9.9 Structure of (a) Ba4 In2 S8 and (b) Ba4 Ga2 S8 with unit cell outlined. Ba—S bonds are omitted for clarity. Black, Ba; blue, In; red, disulfide S2 2− anion with S–S = 2.117(4) Å; purple and red tetrahedra, [In1S4 ] and [In2S4 ]. Source: Liu et al. [1]. © 2011 American Chemical Society. b c

a

b b

(a)

c a

(b)

(c)

Figure 9.10 (a) Structure of Sm4 GaSbS9 viewed along the bc-plane with unit cell marked. (b) The NCS packing of the [(Ga2 S6 )(Sb2 S5 )]10− chains. (c) A pseudo-layer of separated infinite [(Ga2 S6 )(Sb2 S5 )]10− chains at 0.5 < c < 1.0. The Sm—S bonds are omitted for the sake of clarity. Blue: Ga; orange: Sb; yellow: S; pink: interlayer S7–S10; black: Sm; and blue tetrahedron: [GaS4 ] tetrahedron. Source: Chen et al. [9]. © 2011 American Chemical Society.

bc-plane with unit cell marked is shown in Figure 9.10a, it possesses infinite single anionic 1D chains of [(Ga2 S6 )(Sb2 S5 )]10− formed by dimeric [Sb2 S5 ] units linked with corner-sharing dimeric [Ga2 S6 ] polyhedra. Such infinite chains are parallel and separated by Sm3+ cations and interlayer S2− anions (namely, S7–S10). Figure 9.10b exhibits the packing of these 1D [(Ga2 S6 )(Sb2 S5 )]10− chains arranged in a NCS pseudo-layer motif parallel to the ab-plane. Another view along the ab-plane displaying the overlay of two such pseudo-layers is in Figure 9.10c. This NCS packing of the chains generates the in-phase alignment of the dipoles of both [Sb2 S5 ] and [Ga2 S6 ] asymmetric units. Each chain is construct of [Sb2 S5 ] units linked with [Ga2 S6 ] via 𝜇 2 -S3 atoms. The [GaS4 ] tetrahedra are dimerized via linking vertex (S5), while the [SbS4 ] units via S2–S2 edge. On the one hand, the small distortion of [GaS4 ] tetrahedron is indicated by the small deviations of both the Ga—S bond lengths (2.27–2.31 Å) and the S–Ga–S angles (105–112∘ ). On the other hand, the Sb atom is fourfold coordinated in a teeter-totter-shaped geometry

9.2 Inorganic Chalcogenides

constructed by three short length Sb—S bonds (2.45–2.67 Å) and a long Sb—S3 bond (2.84 Å). The Sm3+ cations and interlayer S2− anions (S7–S10) appear between the [(Ga2 S6 )(Sb2 S5 )]10− chains. The Sm atoms are coordinated with six or seven S atoms in normal environments, such as seven-coordinated mono-capped octahedra for Sm1 and distorted six-coordinated octahedra for Sm2–Sm4. All Sm—S bond distances are normal, varying from 2.729(5) to 3.052(6) Å. The size of the Ln metal cation does not affect the packing of the anionic chains in the title compounds, only a reduction of the unit cell parameters as their atomic numbers increase on going from Pr to Ho. 9.2.2.3 La4 InSbS9

La4 InSbS9 belongs to the tetragonal NCS space group P43 21 2 (No. 96) and is characterized by unusual 1D [(In2 S6 )(Sb2 S5 )]10− infinite helical chains propagating along the ab-plane and separated by isolated La3+ and S2− [10]. These 1D chains are further packed around the 41 helical axes (Figure 9.11a). Such chains are built from dimeric teeter-totter [Sb2 S5 ] polyhedra and dual-apex-shared [In2 S6 ] tetrahedra (Figure 9.11b); they are reminiscent of the chains in Ln4 GaSbS9 but differ in that [Sb2 S5 ] and [In2 S6 ] are arranged around a twofold screw axis, which means that neighboring dimers are oriented in opposite directions (Figure 9.11c), whereas the neighboring [Sb2 S5 ] or [Ga2 S6 ] in Ln4 GaSbS9 is arranged in-phase [9]. The [SbS4 ] polyhedron is remarkably distorted as a consequence of the SCALP electrons of Sb atom, with Sb—S bond distances of 2.44–2.91 Å. The [InS4 ] tetrahedron is less distorted, with In–S distances of 2.45–2.49 Å and S–In–S angles ranging from 104∘ to 111∘ . La3+ cations and isolated S2− anions (namely, S7–S10) occur between the infinite helical chains. The cationic La1 and La2 exhibit normal [LaS6 ] trigonal-prismatic coordination, and La3 and La4 are found in [LaS7 ] monocapped trigonal prisms with La–S lengths of 2.84–3.45 Å. Thus, the formula can be written as (La3+ )8 ([In2 Sb2 S11 ]10− )(S2− )7 .

21 axis

(a)

(b)

(c)

b

c

2.9 0

(d)

2.46

2.6

9

2.69

2.44

a

2.44

Figure 9.11 (a) View of the structure of La4 InSbS9 viewed down the ab-plane with unit cell marked, the La—S bonds are omitted for the sake of clarity. (b) and (c) Helical [(In2 S6 )(Sb2 S5 )]10− chains viewed down the c-axis. (d) Local coordination of [Sb2 S6 ] polyhedron with Sb–S distances marked. Green, La; pink, In; orange, Sb; yellow, S. Source: Zhao et al. [10]. © 2012 American Chemical Society.

6 2.90 2.4

477

478

9 Inorganic Chalcogenides: From Zero-Dimensional Clusters to Three-Dimensional Frameworks

9.2.2.4 Ba3 La4 Ga2 Sb2 S15 and BaLa3 GaSb2 S10

Ba3 La4 Ga2 Sb2 S15 represents a new structure type and crystallizes in the orthorhombic CS space group Ibam (No. 72) with a = 23.628(2), b = 7.8863(5), c = 13.5905(4) Å, and Z = 4 [11]. As shown in Figure 9.12a, Ba3 La4 Ga2 Sb2 S15 characterizes isolated [Sb2 S7 ] dimers that are well separated by Ba2+ , La3+ cations and isolated [GaS4 ] tetrahedra. This structure can be viewed as a stacking of slabs of La–Sb–S ([La4 Sb2 S7 ] slab) and Ba–Ga–S ([Ba3 Ga2 S8 ] slab) along the c-direction. It is worth mentioning that the isolated dimer of the teeter-totter [Sb2 S7 ] polyhedron is stabilized for the first time. More specifically, the [Ba3 Ga2 S8 ] slab includes two crystallographically unique Ba atoms (Wyckoff sites, 4a, 8f ), one Ga atom (Wyckoff site, 8f ), and two S atoms (Wyckoff site, 16k). And the slab is constituted of isolated [GaS4 ] tetrahedra and Ba2+ cations. Such two Ba atoms in Ba3 La4 Ga2 Sb2 S15 are all surrounded by 10 S atoms with the lengths of Ba–S varying from 3.182(3) to 3.424(3) Å. On the other hand, the [La4 Sb2 S7 ] slab includes two crystallographically La (Wyckoff site, 8j), one Sb (Wyckoff site, 8j), and three S atoms (Wyckoff sites, 8j, 16k, and 4c). In addition, Sb atoms are coordinated by four S atoms with two short Sb—S bonds (2.443(3) Å), and two long Sb—S bonds (2.727(4)–2.932(2) Å) to form teeter-totter polyhedrons [SbS4 ]. Then, the uncommon isolated dimer [Sb2 S7 ] is formed by two [SbS4 ] units via corner sharing with the long Sb—S bonds. BaLa3 GaSb2 S10 adopts the monoclinic CS space group P21 /m (No. 11) with a = 7.748(4), b = 13.507(7), c = 15.410(8) Å, 𝛽 = 90.05∘ , and Z = 4. As shown in Figure 9.12b, this compound features 1D structure in which [Ba–La–Ga–S] and [La–Sb–S] slabs are propagating along ac plane. And it is similar to La4 FeSb2 S10 which characters [La–Fe–S] and [La–Sb–S] slabs (Figure 9.12c). The [Ba–La–Ga–S] a

[La4 Sb2 S7]

c

[Ba3 Ga2 S8] (a) [BaLaGaS4] [La2FeS4]

[La2Sb2S6] b c

(b)

c a

(c)

Figure 9.12 (a) View the structures of (a) Ba3 La4 Ga2 Sb2 S15 , (b) BaLa3 GaSb2 S10 , and (c) La4 FeSb2 S10 with the unit cell outlined. Legend: black, Ba; red, La; blue, Sb; yellow, S; pink tetrahedron, [GaS4 ]; green tetrahedron, [FeS4 ]. Source: Duan et al. [11]. © 2017 Royal Society of Chemistry.

9.2 Inorganic Chalcogenides

slab in BaLa3 GaSb2 S10 features a centric 0D structure which contains isolated [GaS4 ] tetrahedra with Ba2+ and La3+ cations locating between them. Comparing with slab of [La–Fe–S] slab in La4 FeSb2 S10 [17], [GaS4 ] tetrahedron in BaLa3 GaSb2 S10 replaces [FeS4 ] tetrahedron. Meanwhile, to balance the valence, Ba atoms in BaLa3 GaSb2 S10 substitute La atoms in La4 FeSb2 S10 with the similar geometry configurations of [BaS10 ] versus [LaS10 ] On the other hand, the [La–Sb–S] slab exhibits 1D (SbS4 )n branched chains along a-axis with La3+ cations filling in them, which is comparable with [La–Sb–S] slab in La4 FeSb2 S10 . In the [La–Sb–S] slab, there are four La atoms (Wyckoff site, 2e), four Sb (Wyckoff site, 2e), and eight S atoms (Wyckoff sites, 4f , 2e, 4f , 2e, 4f , 4f , 2e, 2e). In addition, guided by [Ba3 Ga2 S6 ], [BaLaGaS4 ], and [La2 FeS4 ] slabs in compounds Ba3 La4 Ga2 Sb2 S15 , BaLa3 GaSb2 S10, and La4 FeSb2 S10 , the extra Ba atoms and [GaS4 ] tetrahedra in Ba3 La4 Ga2 Sb2 S15 enlarge the unit cell a from 15.066(4), 15.410(8) to 23.628(8) Å. Moreover, extra La in [La2 Sb2 S6 ] slabs of Ba3 La4 Ga2 Sb2 S15 also cannot be ignored to enlarge the unit cell a. Besides, b and c of the unit cell in Ba3 La4 Ga2 Sb2 S15 do not show obvious change comparable with BaLa3 GaSb2 S10 and La4 FeSb2 S10 . 9.2.2.5 Ba8 Zn4 Ga2 S15

Ba8 Zn4 Ga2 S15 belongs to a new structure type based on the single-crystal structure analysis and crystallizes in the CS monoclinic system with space group of P21 /n (No. 11) (crystallographic parameters: a = 13.328(2), b = 9.0097(4) Å, c = 13.332(2) Å, 𝛽 = 109.52(0)∘ and Z = 2) [12]. There are four unique Ba atoms, two unique Zn atoms, one unique Ga atom, and eight unique S atoms in the asymmetric units, and all atoms are in Wyckoff site 4e (site symmetry: 1) except for S5 in 2b (site symmetry: 1). In the structure, there are two BBUs: [Zn2 S6 ] dimers and [GaS4 ] tetrahedra. The [Zn1 S4 ] unit is connected with the [Zn2 S4 ] unit via the S2 atom as bridging sulfur to form the [Zn2 S6 ] dimers (Figure 9.13a). Two of such dimers are further interconnected into a [Zn4 S10 ] cluster (Figure 9.13b), which is first found in X/Zn/Ga/Q systems as far as we know. Then, these [Zn4 S10 ] clusters are further connected by sharing S atoms and isolated [GaS4 ] tetrahedra (Figure 9.13c) form a unique 1D [Zn4 Ga2 S15 ]16− chain (Figure 9.13d). Finally, the structure consists of 1D anionic chains that extend along the ac-plane and are separated by Ba2+ cations (Figure 9.13e). Moreover, there are no S—S or metal–metal bonds in Ba8 Zn4 Ga2 S15 ; therefore, the formal oxidation states of Ba/Zn/Ga/S can be expressed as 2+/2+/3+/2–, respectively. More interestingly, the relationships between structural dimension and [X/(Zn + Ga)] ratio in quaternary X/Zn/Ga/Q system have been shown in Figure 9.13f [25, 34, 40, 52]. When [X/(Zn + Ga)] ≤ 0.71, 3D framework structure will be taken; when [X/(Zn + Ga)] ≥ 1.20, low-dimensional structures (2D and 1D) will be adopted, respectively. Accordingly, guided by the above analysis, we can design and synthesize new chalcogenides with different dimensions by controlling the [X/(Zn + Ga)] value. 9.2.2.6 A4 Ge4 Se12 (A = Rb, Cs)

Both polyselenides A4 Ge4 Se12 (A = Rb, Cs) belong to the polar orthorhombic space group Pna21 (No. 33) [13]. As they are isostructural, the following discussion

479

480

9 Inorganic Chalcogenides: From Zero-Dimensional Clusters to Three-Dimensional Frameworks

(a) S3

S5

Zn1

Zn1

×2

S2

S1

S2

a

(e)

S6

c S1

Ga

+

Zn1

S2

S8 S3

S3

Zn2

S1

(c)

Zn2

S3

Zn2

S2

S4

(b) S5

S4

S7

S5

S4

b

(d)

Ba Zn

a

Ga S

3D

(f) AZn4Ga5S12

2D

Ba6Zn7Ga2S16

X5ZnGa6S15

a

Na6Zn3Ga2S9

a

c

a

b

c

A Zn/Ga S

c

1D Ba8Zn4Ga2S15

a

0.11

0.67

Zn/Ga S Na

Sr or Pb Zn Ga S

Ba Zn Ga S

b

0.71

[X/(Zn + Ga)] ratio

1.20

1.33

Figure 9.13 Crystal structure of Ba8 Zn4 Ga2 S15 : (a) [Zn2 S6 ] dimer; (b) [Zn4 S10 ] cluster; (c) [GaS4 ] tetrahedron; (d) 1D [Zn4 Ga2 S15 ]16− chain; (e) the structure consists of 1D anionic chains (blue and green polyhedron) that extends along the ac-plane and are separated by Ba2+ cations. (f) Sketch map of the empirical [X/(Zn + Ga)] ratio–structure relationships in quaternary X/Zn/Ga/Q system. Source: Li et al. [12]. © 2019 Royal Society of Chemistry.

will mainly focus on Cs4 Ge4 Se12 . As shown in Figure 9.14a, its structure features infinite 1D polyanionic [Ge4 Se12 ]4− chains extending along the ac-plane, separated by the counter Cs+ cations. Two [GeSe4 ] tetrahedra are coupled by sharing a Se atom and a diselenide Se2 linkage to form a dimeric [Ge2 Se4 (μ-Se2 )]4− unit; the [Ge2 Se4 (μ-Se2 )]4− units are further linked by sharing Se4 and Se10 atoms with alternating [Ge(1)Ge(4)Se5 (μ-Se2 )]4− and [Ge(2)Ge(3)Se5 (μ-Se2 )]4− to generate the 1D [Ge4 Se12 ]4− chains (Figure 9.14b). The asymmetrically bridged ditetrahedral [Ge2 Se4 (μ-Se2 )]4− polyanions are firstly discovered in solid-state inorganic compounds. All Ge atoms in both compounds are tetrahedrally coordinated by three bridged Se atoms with Ge—Se bond lengths of 2.379(1)–2.410(1) Å and one terminal Se atom with Ge—Se bond lengths of 2.243(1)–2.257(2) Å. The [GeSe4 ] tetrahedra moieties are distorted from the ideal geometry evidenced by the Se–Ge–Se angles ranging from 95.68(7)∘ to 119.08(8)∘ . 9.2.2.7 BaGeOSe2

BaGeOSe2 crystallizes in the NCS space group of orthorhombic P21 21 21 (No. 19) [14]. As shown in Figure 9.15a, it features parallel 1D chains of the infinite [GeOSe2 ]2− anion along the bc-plane. These chains are built from corner sharing [GeO2 Se2 ] tetrahedra, in which Ge4+ ion is coordinated by two terminal atoms Se and two bridging atoms O in a tetrahedral geometry (Figure 9.15b), presenting a novel coordination environment for germanium in the solid state. The large Ba atoms are bonded

9.2 Inorganic Chalcogenides

Figure 9.14 (a) Crystal structure of Cs4 Ge4 Se12 and (b) a single 1D [Ge4 Se12 ]4− chain. Source: Liu et al. [13]. © 2017 American Chemical Society.

Cs Ge Se

c a (a) Se1 Se3 Se9 Se4 Se10 Ge4 Ge2 Ge3 Ge1 Se12 Se6 Se2 Se5 Se8 Se11 Se7

Se4

c

(b)

b

Ba Be O c

Se b

(a)

b

a c

(b)

Figure 9.15 (a) Structure of BaGeOSe2 viewed slightly skewed from the bc-plane. (b) A single 1D [GeOSe2 ]2− chain. Source: Liu et al. [14]. © 2015 American Chemical Society.

to six Se atoms and one O atom to form asymmetric polyhedra, and a further combination with the asymmetric [GeOSe2 ]2− chains results in the crystallographic asymmetry of compound. The results of bond valence calculations (Ba, 1.96; Ge, 4.10) indicate that the Ba and Ge atoms are in oxidation states of +2, and +4, respectively. It is interesting to compare the structural features in BaGeOSe2 and BaGeO3 [61] to illustrate their structural evolution. The overall structure of BaGeOSe2 is similar to the known BaGeO3 that belongs to pyroxene-type structure type. It also can be described that BaGeOSe2 derive from the substitution of Se2− for terminal O2− anion in BaGeO3 . Nevertheless, distinct from BaGeO3 structure, except the cell

481

482

9 Inorganic Chalcogenides: From Zero-Dimensional Clusters to Three-Dimensional Frameworks

constants and cell volume greatly increase in BaGeOSe2 due to the different size of O2− and Se2− (ionic radius: 1.21 Å versus 1.84 Å) anions, the main differences in the structure are [GeO4 ] units being replaced by [GeO2 Se2 ] units, and the [BaO7 ] units replaced by [BaOSe6 ] units. Specially, the Ge centered tetrahedra in BaGeOSe2 are distorted, as evidenced by the Ge—Se bond lengths (2.291(1)–2.305(1) Å) and Ge—O bond distances (1.789(4)–1.790(4) Å) as well as the ∠Se—Ge—Se, ∠O—Ge—Se, and ∠O—Ge—O bond angles (124.77∘ , 99.50–111.97∘ , and 104.46∘ , respectively). And the [BaOQ6 ] polyhedra are also distorted with Ba—Se bond lengths ranging from 3.283(1) to 3.616(1) Å and Ba—O bond length of 2.608(4) Å. The inequilateral coordination in these mixed-anion BBUs is inclined to result in a NCS space group and is also beneficial for enhancing the dipole moment of individual polyhedra. 9.2.2.8 Ba5 In4 Te4 S7

Ba5 In4 Te4 S7 crystallizes in orthorhombic space group Imm2 (No. 44), with the unit cell parameters a = 39.110(3) Å, b = 4.3763(3) Å, c = 7.3452(6) Å, and Z = 2 [15]. The major structural characteristics of Ba5 In4 Te4 S7 are composed of infinite 1D [InS2 Te2 ]5− and 1∞[In2 S3 Te2 ]4− anionic chains with Ba2+ cations inserting between them for charge balance, and the two chains are parallel to b-axis (Figure 9.16). 1D [InS2 Te2 ]5− chain is composed of a [InS2 Te2 ] tetragonal-pyramid chain, and 1D [In2 S3 Te2 ]4− chain is composed of two [InS2 Te2 ] tetragonal-pyramid chains via edge sharing. In the asymmetric unit, there are three crystallographically unique

b a Ba In Te S

(a)

a c Ba In Te S

(b)

Figure 9.16 Structure of Ba5 In4 Te4 S7 viewed from the (a) ab-plane and (b) ac-plane. Source: Tan et al. [15]. © 2015 Royal Society of Chemistry.

9.2 Inorganic Chalcogenides

Ba atoms, two In atoms, four S atoms, and two Te atoms. Ba atoms in the unit cell present three types of coordination environment: Ba1 are sixfold coordinate by S atoms, Ba2 are eightfold coordinate by four S atoms and four Te atoms, Ba3 are eightfold coordinate by six S atoms and two Te atoms. The In atoms in the unit cell present one type of coordination environment: In1 and In2 are fourfold coordinate by two S atoms and two Te atoms. 9.2.2.9 Ba8 Ga2 Sn7 Se18 and Ba10 Ga2 Sn9 Se22

Orthorhombic Ba8 Ga2 Sn7 Se18 crystallizes in the CS Pnma (No. 62) with a = 12.466(9) Å, b = 9.358(6) Å, c = 18.14(2) Å and Z = 2 [16]. There are two crystallographically independent Ba atoms, one independent Ga atom, four independent Sn atoms and seven independent Se atoms. All Ba atoms, two of the Se atoms are at the Wyckoff 8d sites, and the other atoms occupy the Wyckoff 4c sites. The occupancies of all atoms are 100% except for Sn1 (50%). As illustrated in Figure 9.17a, the major structure motif of Ba8 Ga2 Sn7 Se18 is the infinite 1D [Ga2 Sn7 Se18 ]16− ladder chain that is built by alternatively jointed [GaSe4 ] tetrahedron and [SnSe4 ] tetragonal pyramid. To each of such polyhedra, an accessorial Aa

A

Aa

A

Ba Ga Sn Se

b c

(a) Ba8Ga2Sn7Se18, 1 (CS structure) 2Ba2+

[Sn2Se4]4–

(b) Ba8Ga2Sn7Se18(Ba2Sn2Se4) 2 (NCS structure)

a b

(c)

Ac

A

At

10

A21 10

9

9

c

Ac

10 9

9

10

A

Ba Ga Sn Sn Se

9

21 axis

10

Figure 9.17 The packing of the ladder chain in (a) Ba8 Ga2 Sn7 Se18 and (b) Ba10 Ga2 Sn9 Se22 . The asymmetric unit of the ladder chain is marked by atom numbers. Each ladder chain is combined by Chain 1 (orange) and Chain 2 (blue) via sharing of common Se atoms. (c) The [Sn2 Se4 ] units in Ba10 Ga2 Sn9 Se22 viewed from a direction, and 21 screw axis at (1/4, 1/4, z) is visualized as a light blue line. Source: Li et al. [16]. Licensed under CC BY 3.0.

483

484

9 Inorganic Chalcogenides: From Zero-Dimensional Clusters to Three-Dimensional Frameworks

[SnSe3 ] (Sn = Sn2) pyramid or [Sn2 Se4 ] (Sn = Sn3, Sn4) dimer is attached via a common edge or an apex, respectively. Orthorhombic Ba10 Ga2 Sn9 Se22 crystallizes in the NCS space group Cmc21 (No. 36) with a = 9.384(2) Å, b = 44.834(6) Å, c = 12.416(2) Å, and Z = 4. There are 5 crystallographically independent Ba atoms, 2 Ga atoms, 10 Sn atoms, and 17 Se atoms. All Ba atoms and five of the Se atoms are sitting at the Wyckoff 8b sites, and the rest of atoms occupy the Wyckoff 4a sites. The occupancies of all atoms are 100% except for Sn1 (50%) and Sn5 (50%). Ba10 Ga2 Sn9 Se22 is also a ladder chain structure; surprisingly, the ladder chains in this compound are nearly identical to those in Ba8 Ga2 Sn7 Se18 . The two short unit cell axes of two compounds are nearly the same, only the long axis is expanded owing to the embedding of [Sn2 Se4 ] dimers (Sn = Sn9, Sn10). As shown in Figure 9.17b, every two ladder chains are separated at the light blue moiety by the isolated [Sn2 Se4 ] dimers. The ladder chain in Ba10 Ga2 Sn9 Se22 is also built by alternatively jointed [GaSe4 ] tetrahedra (Ga = Ga(1) and Ga(2)) and [SnSe4 ] tetragonal pyramids (Sn = Sn1 and Sn5). To each of such polyhedra, an accessorial [SnSe3 ] (Sn = Sn2 or Sn6) pyramid or [Sn2 Se4 ] (Sn = Sn3, Sn4 or Sn7, Sn8) dimer is attached.

9.2.3

Two-Dimensional (2D) Layer Chalcogenides

9.2.3.1 La4 FeSb2 Q10 (Q = S, Se)

The structures of La4 FeSb2 Q10 (Q = S, Se) feature a new structure type and crystallize in the orthorhombic CS space group Pbcm (No. 57) with a = 15.066(4)–15.596(5) Å, b = 7.590(2)–7.869(2) Å, c = 13.341(4)–13.960(4) Å, and Z = 4 [17]. The projection of La4 FeSb2 S10 is shown in Figure 9.18a as an example with all La—S bonds omitted for the better view. The detailed compositions of the La/Sb/S and La/Fe/S slabs are displayed in Figure 9.18b,c, respectively. The unique teeter-totter (SbS4 )n chains are made of [SbS3 ] trigonal pyramids (Sb–S < 2.60 Å) and a relatively weak Sb–S interaction (Sb–S 2.90 Å), which is based on the computational results of COHP. These chains are interconnected by [La1S8 ] and [La2S8 ] in a layered motif that parallels the ab-plane, namely, the La/Sb/S slab (Figure 9.18b). Between these La/Sb/S slabs, there are La3, La4 cations and [FeS4 ] tetrahedra that are arranged in a layered motif (namely, the La/Fe/S slab) as shown in Figure 9.18c. The adjacent La/Sb/S and La/Fe/S slabs are further connected by [LaS8 ] and [LaS10 ] polyhedra a c 1 4

1

2

1

2 4

La/Sb/S slab La/Fe/S slab

b

b

a

a

(a)

(b)

(c)

Figure 9.18 (a) Structure of La4 FeSb2 S10 viewed approximately along the ac-plane. (b) The La/Sb/S slab and (c) the La/Fe/S slab viewed along the ab-plane. Light blue, La1 and La2; light green, La3 and La4; orange, Sb; pink, Fe; yellow, S. The La—S bonds are omitted for the better view. Source: Zhao et al. [18]. © 2010 American Chemical Society.

9.2 Inorganic Chalcogenides

via the shared S–S edges and apexes. The formula of the title compounds can be written as La4 [(FeS4 )(SbS3 )2 ]. There are two crystallographically independent Sb atoms. Each Sb atom has three short and one long Sb—S bonds, in the normal range of the Sb—S bonds. Both La1 and La2 exhibit a normal bicapped trigonal prismatic coordination sphere with eight La—S bonds of 2.90–3.06 Å. Differently, the other two crystallographically independent La sites, La3 and La4, with La–S contacts in the range of 3.08–3.34 Å have approximate bicapped square antiprismatic environments. All these La–S distances are comparable to the common range as in the related rare-earth chalcogenides. The Fe atom is tetrahedrally coordinated by S atoms with an average Fe–S distance of 2.30 Å and Fe–S–Fe angles ranging from 106.06∘ to 111.43∘ , such coordination environment is normal in Fe-based chalcogenides. 9.2.3.2 Ln2 Mn3 Sb4 S12 (Ln = Pr, Nd, Sm, Gd)

The four isostructural Ln2 Mn3 Sb4 S12 (where Ln = Pr, Nd, Sm, Gd) represent a new structure type and crystallize in the monoclinic CS space group C2/m (No. 12) with a = 19.928(2)–19.9672(6) Å, b = 3.9323(4)–3.8803(2) Å, c = 14.921(2)–14.9011(1) Å, and Z = 2 on going from Pr to Gd [18]. The decrease of the unit cell volume reflects the lanthanide contraction. For simplicity, the structure of Sm2 Mn3 Sb4 S12 will be discussed in detail as a representative. As shown in Figure 9.19a, Sm2 Mn3 Sb4 S12 displays a new wavy-like [MnS6 ] octahedron layer decorated on both sides by chains of an [SbS5 ] square pyramid. The Sm3+ cations are located between such layers with normal Sm–S distances ( M2 > M3. In case of ACd4 In5 Se12 with smaller r(II)/r(III) = 1.25 (Cd2+ CN = 4 , 0.78 Å, In3+ CN = 4 , 0.62 Å), II distributes nearly evenly at M1 ≈ M2 ≈ M3. Differently, the r(II)/r(III) = 1.07 (Mn2+ CN = 4 , 0.66 Å, In3+ CN = 4 , 0.62 Å) of AMn4 In5 Se12 is smaller than that of ACd4 Ga5 Q12 , and the occupancy of II follows the same trend of M1 > M2 > M3 as in ACd4 Ga5 Q12 . Despite the different distribution of II, the overall stoichiometry of (A+ )(II)4 (III)5 (Se2− )12 is fixed because of the charge balance requirement. The cationic A centers the Q12 cuboctahedron and the identity of the alkali metal on going from K+ , Rb+ , to Cs+ does not affect significantly the occupancy on the 9b site. 9.2.4.8 A–III–Sn2 –Se6 Type

Single crystal XRD data reveal that compounds with a general formula A–III–Sn2 –Se6 (A = Rb, Cs; III = Ga, In) [35] crystallize in the trigonal space group R3 (Person symbol hR22) with a = 10.4697(2)–10.6044(8) Å, c = 9.476(2)–9.660(2) Å, and Z = 3 for Rb-members, and a = 10.518(8)–10.625(2) Å, c = 9.539(2)–9.688(8) Å, and Z = 3 for Cs-members, which belong to the BaGa2 SnSe6 -structure type [68]. In the asymmetric unit, there is 1 crystallographically A atom (Wyckoff site, 3a), 1 M positions (Wyckoff site, 9b) randomly occupied by both Ga/In and Sn in the molar ratio of 1 : 2, and 2 Se atoms (Wyckoff sites, 9b and 9b). The remarkable structures of the 3D DLFs are formed by the tri-nuclear secondary basic structure unit [M3 Se9 ] (M = III/Sn), which is constructed with three vertex-sharing [MSe4 ] tetrahedra (Figure 9.36a,b). Furthermore, the A atom fills with 3D DLFs and centers the Se12 cuboctahedron (Figure 9.36c).

501

502

9 Inorganic Chalcogenides: From Zero-Dimensional Clusters to Three-Dimensional Frameworks

a c

A M = X/Sn Se

(a)

(b)

(c)

Figure 9.36 (a) 3D DLF structure of A–III–Sn2 –Se6 viewed down the ac-plane with the unit cell marked. (b) The tri-nuclear secondary basic structure unit [M3 Se9 ] (M = III/Sn), with the Se atom numbers marked, is outlined by green shadow in (a). (c) View of the A-centered Se12 cuboctahedron. Source: Lin et al. [35]. © 2017 Royal Society of Chemistry.

9.2.4.9 Ba3 AGa5 Se10 Cl2 and Its Derivatives

As the first inorganic open-framework compounds, Ba3 AGa5 Se10 Cl2 (A = Cs, 1; Rb, 2 and K, 3) represent a new type of NCS structure that is constructed by supertetrahedral clusters in the tetragonal space group I4 with a = b = 8.7348(6)–8.6341(7) Å, c = 15.697(3)–15.644(2) Å on going from Cs to K [36]. The projection of Ba3 CsGa5 Se10 Cl2 is shown in Figure 9.37a,b as an example, the 3D anionic ∞ covalent framework of 3 [Ga5 Se10 ]5− is interpenetrated by Ba2+ /Cs+ cations and Cl− anions. Such an anionic framework crystallizes in a zinc-blende topological structure in which all the Zn and S sites are alternately substituted by supertetrahedral [Ga4 Se10 ]8− cluster (T2) and tetrahedron [GaSe4 ]5− (T1) (Figure 9.37e). Four aligned [Ga1Se4 ] tetrahedra constitute a T2 supertetrahedral cluster via corner sharing (Se2, Se3) that is further linked with T1 tetrahedron [Ga2 Se4 ]5− by common Se1 corner (Figure 9.37a,c,d). The slightly distorted [Ga2 Se4 ] tetrahedron (T1) has mirror symmetry with Ga2—Se1 bond of 2.4111(5) Å and Se–Ga2–Se angles ranging from 103.5∘ to 122.1∘ . In comparison, the distortion of [Ga1Se4 ] tetrahedron of T2 is distinct with Ga1—Se bond lengths ranging from 2.3858(8) to 2.4083(7) Å and Se–Ga2 –Se angles varying from 103.2∘ to 112.5∘ . The T1 cluster is nonpolar because

9.2 Inorganic Chalcogenides

(b)

(a)

c

C11

Ba/Cs

C12 Ga2

a c

b

a c

(c)

(d)

(e)

Figure 9.37 (a) Unit cell of Ba3 CsGa5 Se10 Cl2 with atom outlined. (b)–(e) Structure viewed of the T1 –T2 hybrid covalent framework. Light blue: Cl anion, dark blue: Cs/Ba cation. Red: T1 tetrahedron [Ga2 Se4 ]5− and Ga2 , green: T2 supertetrahedral cluster [Ga14 Se10 ]8− and Ga1 . Source: Yu et al. [36]. © 2012 American Chemical Society.

of its mirror symmetry; however, the T2 supertetrahedral cluster is polar, because its constituent [Ga1Se4 ] tetrahedron as well as the whole T2 cluster itself only shows C1 symmetry. Then, guided by a dual-ion substitution strategy, six isomorphic compounds Ba4 MGa4 Se10 Cl2 (M = Zn, 4; M = Cd, 5; M = Mn, 6; M = Cu/Ga, 7; M = Co, 8 and M = Fe, 9) free of alkali metal are discovered [37]. Their compounds differs from 1 to 3 in that the center of each T1 tetrahedron in the latter can be either totally replaced by bivalent transition metals (Zn2+ , 4; Cd2+ , 5; Mn2+ , 6; Co2+ , 8; Fe2+ , 9), or partially substituted by monovalent metal (Cu+ , 4). Importantly, the alkali metal A+ cations in 1–3 have to be simultaneously substituted by alkali earth metal Ba2+ , otherwise, any synthetic attempt of the transition metal analog would be in vain. Such a dual-ion synergy substitution (Mn+ versus Ga3+ together with Ba2+ versus Cs+ ) is crucial to balance the charge, without which, the formation of 4–9 is impossible. Case 4 is of special. Let us consider, if one Cu+ substitutes one Ga3+ at the T1 as expected leaving the anionic moiety two electrons insufficient; by replacing one Cs+ with one Ba2+ only denotes one electron to the anionic moiety, thus, the resulting “[(Ba2+ )4 (Cu+ )(Ga3+ )4 (Se2− )10 (Cl− )2 ]− ” carries one electron without charge balance. To balance this, Ga3+ ion is welcome to co-share the T1 center with Cu+ leading to 4 with half–half Cu/Ga occupancies.

503

504

9 Inorganic Chalcogenides: From Zero-Dimensional Clusters to Three-Dimensional Frameworks

Ba

(d)

SnS4

μ1–S1

Li/CdS4

μ2–S2

0

2.40

2.358

4 2.

80

(a)

Three fold axes

b c (c)

a (b)

Figure 9.38 (a) View of the structure of Ba6 Li2 CdSn4 S16 along [111]; (b) view of the SnM3 S13 units (M = Li/Cd); (c) the connection style of Sn(Li/Cd)3 S13 units; and (d) the details of the SnM3 S13 unit. Source: Duan et al. [38]. © 2017 Royal Society of Chemistry.

9.2.4.10 Ba6 Li2 CdSn4 S16

Compound Ba6 Li2 CdSn4 S16 crystallized in the NCS cubic space group I43d (No. 220) with a = 14.6678 Å and Z = 4 [38]. In the asymmetric unit, there is 1 crystallographically independent Ba atom, 1 Sn atom, 2 S atoms, and a Wyckoff 12a site (noted M) that is co-shared by Li and Cd atom with the occupancies of 67% and 33%. As shown in Figure 9.38a, the structure is a 3D framework constructed by [SnS4 ] and [MS4 ] tetrahedra with the channels filled with Ba cations. As shown in Figure 9.38b,d, the basic unit of [SnM3 S13 ] lying in the threefold axis is constructed by 1 [SnS4 ] tetrahedron and three distorted [MS4 ] tetrahedra via sharing μ2 -S2 atoms and leaving μt -S1 as terminal atoms. The 3D framework of Ba6 Li2 CdSn4 S16 is constructed by [SnM3 S13 ] units that are connected with each other via [SnS4 ] tetrahedra, as shown in Figure 9.38c. 9.2.4.11 La2 CuSbS5

Single-crystal XRD data indicated that La2 CuSbS5 adopts the NCS orthorhombic system (space group: Ima2 [No. 46]; Pearson symbol: oI36; idealized Wyckoff sequence: c2 b3 a2 ) [39]. The structure of La2 CuSbS5 consists of the following distinct units: [La1S10 ] and [La2S8 ] polyhedra, [CuS4 ] tetrahedral, and [SbS4 ] pyramid. In a symmetric unit, La1 and Cu atoms are at 4a special sites with a local symmetry of C2, La2, Sb, and S1 atoms are at 4b special sites with a local symmetry of m, S2, and S3 atoms are at general positions (8c). The projection of La2 CuSbS5 viewed down the c-axis is depicted in Figure 9.39a. The 3D framework was constructed from two types of 2D slabs (La/Cu/S and La/Sb/S) through the shared S–S apexes and edges. The detailed views of the La/Cu/S slab are shown in Figure 9.39b,c. The 2D La–S layer is made of distorted square antiprismatic [La1S10 ] polyhedra that are connected via sharing vertexes (S3 atoms). Such layer is further filled with discrete

(c)

(a)

(b)

b

(f)

La/Cu/S slab

+ La/Sb/S slab

a

+

InS6

a

SbS4

c

c

a

“SALP” Induction approach La1 La2 Cu Sb S Sb–S3 = 2.467Å Sb–S1 = 2.475Å Sb–S3 = 2.952Å Sb–S2 = 3.312Å Sb–S1 = 3.538Å

3

0

–3

–1

0

1–1

0

1–1

0

–COHP

1–1

0

1 –1

0

1

b

La2CuInS5 (CS)

(g) 1 4

1.5

La2CuSbS5 (NCS)

La1 1 4

La2

In

Sb

Cu

S

1 4

1.2 0.9

Symmetry

0.6

Breaking

0.3 0.0

–6

a

(e) 1.8 –ICOHP (eV/bond)

Energy (eV)

(d) 6

c

2.4

2.6

2.8

3.0

3.2

Sb–S distance (Å)

3.4

3.6

1 4

1 4

Space group Pnma (no. 62)

Space group Ima2 (no. 46)

Figure 9.39 (a)–(c) The crystal structure, (d)–(e) COHP–ICOHP curves, and (f)–(g) structural evolution of La2 CuSbS5 . La—S bonds < 3.4 Å, Cu—S bonds < 2.5 Å, and Sb—S bonds < 3.0 Å. Source: Lin et al. [39]. © 2019 Royal Society of Chemistry.

506

9 Inorganic Chalcogenides: From Zero-Dimensional Clusters to Three-Dimensional Frameworks

[CuS4 ] tetrahedra, forming a dense layered pattern that perpendiculars to the a-axis, namely, the La/Cu/S slab (Figure 9.39b). Between these La/Cu/S slabs, there are [La2S8 ] bicapped-trigonal-prisms and [SbS4 ] pyramids that are also stacked in a layered pattern (namely, the La/Sb/S slab) as displayed in Figure 9.39c. The unique 1D teeter-totter (SbS4 )n chain is built of [SbS4 ] pyramids (a SbS3 pyramid as well as a relatively weak Sb—S bond), according to the analysis of the COHP–ICOHP curves (Figure 9.39d,e). Subsequently, these chains are corner share with 1D wave-like La–S chains to further create the La/Sb/S slab along the bc-plane. Interestingly, La2 CuSbS5 was obtained by the SCALP induction strategy by taking the cue from the known CS La2 CuInS5 [69]. Although both of them exhibit the identical stoichiometry 2–1–1–5 and crystallize in the same orthorhombic system, there are some significantly different characteristics in their crystal structures: (i) the coordination environments of the two crystallographically unique La atoms in La2 CuSbS5 are different (CN = 10 for La1 and CN = 8 for La2), but the same in La2 CuInS5 (CN = 8 for La1 and La2) with La—S bond distances 3.45 Å), which confirms that the geometry environment around Sn atoms is certainly fourfold coordination as mentioned above.

9.2.5

Mixed-Dimensional (MD) framework Chalcogenides

9.2.5.1 Ln3 M0.5 (Ge0.5 /M0.5 )S7 (Ln = La, Sm; M = Ga, In) and Ln3 In0.33 GeS7 (Ln = La, Sm, Gd)

Ln3 M0.5 (Ge0.5 /M0.5 )S7 (Ln = La, Sm; M = Ga, In) and Ln3 In0.33 GeS7 (Ln = La, Sm, Gd) are isostructural to the well-known Ln6 ZnSi2 S14 and crystallize in the polar NCS hexagonal space group P63 (No. 173) [59]. And Ln3 M0.5 (Ge0.5 /M0.5 )S7 (Ln = La, Sm; M = Ga, In) are the first Ga members in this family. Their structures feature chains of face-sharing [GaS6 ] (or [InS6 ]) octahedra that are surrounded by both discrete Ln3+ cations and isolated [TS4 ] tetrahedra (Figure 9.59a). The center of each [TS4 ] tetrahedron is disordered by Ge14+ /Ga13+ (or Ge14+ /In13+ ) in the cases of b a 2 1

1

2 2

2

2 2

2 2

1

2 2 1

2

2

b c

(a)

(b)

Figure 9.59 (a) Structure of La3 Ga0.5 (Ge0.5 /Ga0.5 )S7 viewed down the ab-plane. The La—S bonds were omitted for the sake of clarity. (b) View of chain of face-sharing [GaS6 ] octahedra and the arrangement of [(Ga/Ge)S4 ] tetrahedra along the c-axis. Rose: La; yellow: S; pink octahedron: [GaS6 ]; aqua tetrahedron: [TS4 ] (T = disordered Ga/Ge). Source: Shi et al. [59]. © 2015 American Chemical Society.

9.2 Inorganic Chalcogenides

Ln3 M0.5 (Ge0.5 /M0.5 )S7 (Ln = La, Sm; M = Ga, In) but solely occupied by Ge4+ in Ln3 In0.33 GeS7 (Ln = La, Sm, Gd). Figure 9.59b shows the extending of the chains of [GaS6 ] (or [InS6 ]) octahedra and the arrangements of the discrete [TS4 ] tetrahedra along the c-direction. The chains of [GaS6 ] octahedra are running parallel to the 63 axis, and the isolated [TS4 ] tetrahedra are operated by the threefold rotation axis. The distortion of the [TS4 ] tetrahedron is seen by the T—S bond difference of 0.04–0.07 Å for the title compounds. In comparison, [GaS6 ] or [InS6 ] octahedra are less distorted. The Ln atom is coordinated to eight S atoms in a distorted bicapped trigonal prism, with normal Ln–S distances. 9.2.5.2 CsCu5 S3

Single crystal diffraction data reveal that o-CsCu5 S3 crystallizes in Pmma (No. 51) with a = 9.6343(14) Å, b = 3.9590(7) Å, c = 8.9592(13) Å, and Z = 2, and t-CsCu5 S3 crystallizes in P421 c (No. 114) with a = b = 13.0550(17) Å, c = 7.8836(10) Å, and Z = 8 [60]. Markedly different from cubic Cu2 S with a high degree of disorder, in which eight Cu atoms are distributed over 204 Wyckoff sites showing a liquid-like behavior, all Cu atoms in both structures of CsCu5 S3 are fully localized on the corresponding Wyckoff sites with 100% occupancy. Both o- and t-CsCu5 S3 structures are built from the same [Cu4 S4 ] columnar structure building unit, which is constructed by threefold coordinated Cu1 and Cu2 atoms (Figure 9.60). Because the c parameter of t-CsCu5 S3 (c = 7.88 Å) is nearly two times the b value of o-CsCu5 S3 (b = 3.96 Å), the [Cu4 S4 ] columns in two structures are thus propagating with similar periodicity along different directions, [001] in the former versus [010] in the latter. However, the structure differences are also obvious. In o-CsCu5 S3 , the [Cu4 S4 ] column is extending at two opposite sides via the twofold coordinated Cu3 (Cu3–S = 2.21 Å) atoms into a wavy layer. Further, these layers are stacking along the ab-plane; in between, an array of Cs+ cations is accommodated (Figure 9.60a). The structural anisotropy is t–CsCu5S3

o–CsCu5S3

a

b

c

b a 22 11 11 22

(a)

(b)

Figure 9.60 Single crystal structures of (a) o-CsCu5 S3 and (b) t-CsCu5 S3 that are constructed by a common building unit, the [Cu4 S4 ] column, which is extending at two opposite sides into a waved 2D layer in part a, or simultaneously at four apexes into a 3D network in part b. Black, Cs; yellow, S; red, Cu1–2; blue, Cu3 in part a and Cu3–5 in part b. Source: Ma et al. [60]. © 2019 American Chemical Society.

527

528

9 Inorganic Chalcogenides: From Zero-Dimensional Clusters to Three-Dimensional Frameworks

roughly represented by the Cu/S atomic aggregation density (D) as follows: Db (column) > Da (intralayer) > Dc (interlayer). Dissimilarly, in the t-CsCu5 S3 structure, the [Cu4 S4 ] column is extending simultaneously at four apexes via Cu3 (threefold coordinated) and Cu4,5 (twofold coordinated) (blue ball, Figure 9.60b) into a 3D network that defines large channels running parallel to the [001].

9.3 Conclusion In this chapter, we summarized the recent developments of inorganic chalcogenides, including pure chalcogenides, mixed-anion chalcogenides, and mixed-anion chalcohalides. We focused on the unit cell, space group, and dimensional change as well as the structural assembly of selected chalcogenides. Although many materials have been mentioned in this chapter of inorganic chalcogenides, numerous additional compounds remain to be discovered and studied. It is obvious that with current synthetic techniques, such as the use of reactive fluxes in the conventional solid-state reaction, new phases can be isolated. Moreover, the combination of two or more types of asymmetric building units in a structure would have a high possibility to form an NCS structure.

Acknowledgments This research was supported by the National Natural Science Foundation of China (21771179, and 21301175), the Strategic Priority Research Program of the Chinese Academy of Sciences (No. XDB20000000), the support from “Chunmiao Projects” of Haixi Institute of Chinese Academy of Sciences, the Natural Science Foundation of Fujian Province (2019J01133), and the One Thousand Young Talents Program under the Recruitment Program of Global Youth Experts.

References Liu, J.W., Wang, P., and Chen, L. (2011). Inorg. Chem. 50: 5706–5713. Chen, M.C., Wu, L.M., Lin, H. et al. (2012). J. Am. Chem. Soc. 134: 6058–6060. Li, Y.Y., Li, B.Y., Zhang, G. et al. (2015). Inorg. Chem. 54: 4761–4767. Huangfu, S.X., Shen, J.N., Lin, H. et al. (2015). Chem. Eur. J. 21: 9809–9815. Li, C., Yin, W.L., Gong, P.F. et al. (2016). J. Am. Chem. Soc. 138: 6135–6138. Duan, R.H., Liu, P.F., Lin, H. et al. (2017). Dalton Trans. 46: 14771–14778. Luo, Z.Z., Lin, C.S., Zhang, W.L. et al. (2014). Chem. Mater. 26: 1093–1099. Liu, P.F., Li, Y.Y., Zheng, Y.J. et al. (2017). Dalton Trans. 46: 2715–2721. Chen, M.C., Li, L.H., Chen, Y.B., and Chen, L. (2011). J. Am. Chem. Soc. 133: 4617–4624. 10 Zhao, H.J., Zhang, Y.F., and Chen, L. (2012). J. Am. Chem. Soc. 134: 1993–1995. 11 Duan, R.H., Shen, J.N., Lin, C.S. et al. (2017). Inorg. Chem. Front. 4: 123–130. 1 2 3 4 5 6 7 8 9

References

12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34

35 36 37 38 39 40 41 42 43 44

Li, Y.Y., Liu, P.F., Lin, H. et al. (2019). Chem. Commun. 55: 7942–7945. Liu, B.W., Zhang, M.Y., Jiang, X.M. et al. (2017). Chem. Mater. 29: 9200–9207. Liu, B.W., Jiang, X.M., Wang, G.E. et al. (2015). Chem. Mater. 27: 8189–8192. Tan, D.M., Lin, C.S., Luo, Z.Z. et al. (2015). Dalton Trans. 44: 7673–7678. Li, Y.Y., Wang, J.Q., Liu, P.F. et al. (2017). RSC Adv. 7: 8082–8089. Zhao, H.J., Li, L.H., Wu, L.M., and Chen, L. (2009). Inorg. Chem. 48: 11518–11524. Zhao, H.J., Li, L.H., Wu, L.M., and Chen, L. (2010). Inorg. Chem. 49: 5811–5817. Liu, Y., Chen, L., and Wu, L.M. (2008). Inorg. Chem. 47: 855–862. Lin, H., Chen, H., Lin, Z.X. et al. (2016). Inorg. Chem. 55: 1014–1016. Zheng, Y.J., Shi, Y.F., Tian, C.B. et al. (2019). Chem. Commun. 55: 79–82. Chen, H., Chen, Y.K., Lin, H. et al. (2018). Inorg. Chem. 57: 916–920. Liu, B.W., Hu, C.L., Zeng, H.Y. et al. (2018). Adv. Opt. Mater. 6: 1800156. Abudurusuli, A., Wu, K., Rouzhahong, Y. et al. (2018). Inorg. Chem. Front. 5: 1415–1422. (a) Lin, H., Chen, H., Shen, J.N. et al. (2014). Chem. Eur. J. 20: 15401–15408. (b) Lin, H., Chen, H., Ma, N. et al. (2017). Inorg. Chem. Front. 4: 1273–1280. Lin, H., Shen, J.N., Chen, L., and Wu, L.M. (2013). Inorg. Chem. 52: 10726–10728. Luo, Z.Z., Lin, C.S., Cheng, W.D. et al. (2013). Dalton Trans. 42: 9938–9945. Yu, P., Wu, L.M., and Chen, L. (2013). Inorg. Chem. 52: 724–728. Meng, C.Y., Chen, H., Wang, P., and Chen, L. (2011). Chem. Mater. 23: 4910–4919. Meng, C.Y., Chen, H., and Wang, P. (2014). Inorg. Chem. 53: 6893–6903. Lin, H., Shen, J.N., Shi, Y.F. et al. (2015). Inorg. Chem. Front. 2: 298–305. Lin, H., Chen, H., Liu, P.F. et al. (2016). Dalton Trans. 45: 5775–5782. Huai, W.J., Shen, J.N., Lin, H. et al. (2014). Inorg. Chem. 53: 5575–5580. (a) Lin, H., Zhou, L.J., and Chen, L. (2012). Chem. Mater. 24: 3406–3414. (b) Lin, H., Chen, L., Zhou, L.J., and Wu, L.M. (2013). J. Am. Chem. Soc. 135: 12914–12921. (c) Lin, H., Liu, Y., Zhou, L.J. et al. (2016). Inorg. Chem. 55: 4470–4475. (d) Lin, H., Chen, H., Zheng, Y.J. et al. (2016). Dalton Trans. 45: 17606–17609. (e) Lin, H., Zheng, Y.J., Chen, H. et al. (2017). Chem. Asian J. 12: 453–458. (f) Lin, H., Chen, H., Zheng, Y.J. et al. (2017). Chem. Eur. J. 23: 10407–10412. (a) Lin, H., Chen, L., Yu, J.S. et al. (2017). Chem. Mater. 29: 499–503. (b) Lin, H., Chen, H., Zheng, Y.J. et al. (2017). Dalton Trans. 46: 7714–7721. Yu, P., Zhou, L.J., and Chen, L. (2012). J. Am. Chem. Soc. 134: 2227–2235. Li, Y.Y., Liu, P.F., Hu, L. et al. (2015). Adv. Opt. Mater. 3: 957–966. Duan, R.H., Liu, P.F., Lin, H. et al. (2017). J. Mater. Chem. C 5: 7067–7074. Lin, H., Li, Y.Y., Li, M.Y. et al. (2019). J. Mater. Chem. C 7: 4638–4643. Li, Y.Y., Liu, P.F., and Wu, L.M. (2017). Chem. Mater. 29: 5259–5266. Li, S.F., Jiang, X.M., Liu, B.W. et al. (2017). Chem. Mater. 29: 1796–1804. Luo, Z.Z., Lin, C.S., Cui, H.H. et al. (2015). Chem. Mater. 27: 914–922. Luo, Z.Z., Lin, C.S., Cui, H.H. et al. (2014). Chem. Mater. 26: 2743–2749. Li, G.M., Wu, K., Liu, Q. et al. (2016). J. Am. Chem. Soc. 138: 7422–7428.

529

530

9 Inorganic Chalcogenides: From Zero-Dimensional Clusters to Three-Dimensional Frameworks

45 Wu, K., Yang, Z.H., and Pan, S.L. (2016). Angew. Chem. Int. Ed. 55: 6713–6715. 46 Wu, K., Zhang, B., Yang, Z., and Pan, S. (2017). J. Am. Chem. Soc. 139: 14885–14888. 47 Lin, H., Tan, G., Shen, J.N. et al. (2016). Angew. Chem. Int. Ed. 55: 11431–11436. 48 Lin, H., Chen, H., Zheng, Y.J. et al. (2017). Chem. Commun. 53: 2590–2593. 49 Lin, H., Li, L.H., and Chen, L. (2012). Inorg. Chem. 51: 4588–4596. 50 Lin, H., Shen, J.N., Zhu, W.W. et al. (2017). Dalton Trans. 46: 13731–13738. 51 Chen, H., Liu, P.F., Li, B.X. et al. (2018). Dalton Trans. 47: 429–437. 52 Lin, H., Li, B.X., Chen, H. et al. (2018). Inorg. Chem. Front. 5: 1458–1462. 53 Li, M.Y., Li, B.X., Lin, H. et al. (2018). Inorg. Chem. 57: 8730–8734. 54 Chen, H., Liu, P.F., Lin, H. et al. (2019). Cryst. Growth Des. 19: 444–452. 55 Wu, K., Zhang, B., Yang, Z., and Pan, S. (2018). Chem. Commun. 54: 8269–8272. 56 Zhang, C.Y., Zhou, L.J., and Chen, L. (2012). Inorg. Chem. 51: 7007–7009. 57 Li, S.F., Jiang, X.-M., Fan, Y.H. et al. (2018). Chem. Sci. 9: 5700–5708. 58 Li, M.Y., Li, B.X., Lin, H. et al. (2019). Chem. Mater. 31: 6268–6275. 59 Shi, Y.F., Chen, Y.K., Chen, M.C. et al. (2015). Chem. Mater. 27: 1876–1884. 60 Ma, N., Jia, F., Xiong, L. et al. (2019). Inorg. Chem. 58: 1371–1376. 61 Hilmer, W. (1962). Acta Cryst 15: 1101–1105. 62 Mansuetto, M.F., Keane, P.M., and Ibers, J.A. (1992). J. Solid State Chem. 101: 257–264. 63 Aydemir, U., Pöhls, J.-H., Zhu, H. et al. (2016). J. Mater. Chem. A 4: 2461–2472. 64 Li, H., Malliakas, C.D., Peters, J.A. et al. (2013). Chem. Mater. 25: 2089–2099. 65 Berastegui, P., Eriksson, S., Hull, S. et al. (2004). J. Solid State Sci. 6: 433–441. 66 (a) Iwasaki, K., Takizawa, H., Uheda, K. et al. (2001). Sol. State Chem. 158: 61–67. (b) Iwasaki, K., Takizawa, H., Yamane, H. et al. (2003). Mater. Res. Bull. 38: 141–148. 67 Richardson, J.W., Pluth, J.J., Smith, J.V., and Dytrych, W.J. (1988). J. Phys. Chem. 92: 243–247. 68 Li, X.S., Li, C., Gong, P.F. et al. (2015). J. Mater. Chem. C 3: 10998–11004. 69 Huch, M.R., Gulay, L.D., Olekseyuk, I.D., and Pietraszko, A. (2006). J. Alloys Compd. 425: 230–234. 70 (a) Okorogu, A.O., Mirov, S.B., Lee, W. et al. (1998). Opt. Commun. 155: 307–312. (b) Boyd, G.D., Storz, F.G., McFee, J.H., and Kasper, H.M. (1972). IEEE J. Quantum Electron. 8: 900–908.

531

10 Relationship Between Structure and Electroluminescent, Photochromic, or Second-Order Nonlinear Optical Property Ming-Sheng Wang, San-Gen Zhao, Qian Wang, and Zhong-Ning Chen Chinese Academy of Sciences, State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, 155 Yangqiao Road West, Fuzhou, Fujian 350002, P.R. China

10.1 Introduction Optical materials refer to materials that may modify parameters of input light (such as phase, intensity, and frequency), or switch reciprocally optical and nonoptical signals (such as electric, heat, and sound). In this chapter, we wish to present three kinds of optical materials: electroluminescent materials, photochromic materials, and second-order nonlinear optical (NLO) materials. Species usually exist at ground state but, after simulated by some stimuli, such as light and electric, may convert to excited states. Decay of excited states is a complex process. It may undergo radiation (as light)/non-radiation (as heat, lattice vibration) transition, energy/electron transfer, or chemical reactions. Electroluminescent materials are materials that can be excited to excited states and then return to ground state with the release of light. Sometimes, they display emission profiles similar to those produced by light excitation. So, the first step to construct electroluminescent devices is usually to find one luminescent media with suitable photoluminescent spectrum. Photochromic materials are bistable materials that may be switched between two ground states with at least one direction being excited by light. The most common photochromic reactions include isomerization, dissociation, and electron transfer. Unlike photochromic materials, second-order NLO materials require nonlinear light, commonly referring to laser, to act as input light. A new beam of light with frequency of 2𝜔 will be generated when one beam of light with frequency of 𝜔 is imported. This so-called “second-harmonic generation (SHG)” effect is strongly related with total intensity of polarization of the media, which is proportional to second-order polarizability. It is just related with ground states. Optical materials are widely used in daily life. For example, electroluminescent materials have been widely applied to LEDs or organic light-emitting diodes (OLEDs), which have brought revolution of both lighting and display fields. The current aim in the electroluminescence (EL) field is to achieve new materials

Advanced Structural Chemistry: Tailoring Properties of Inorganic Materials and their Applications, First Edition. Edited by Rong Cao. © 2021 WILEY-VCH GmbH. Published 2021 by WILEY-VCH GmbH.

532

10 Relationship Between Structure and Electroluminescent, Photochromic

with lower cost, higher efficiency, customizable color, and ability for large-area fabrication. Photochromic materials have been used for glassware (such as sun glasses, automobile rearview mirror, and building window) and chromic ink (for anti-fake, cloth decorating, etc.) in the market. Other applications, such as switching, memory, bioimaging, and radiation monitoring, have also been demonstrated in lab. One aim in the photochromic field is to find new materials with richer colors, higher fatigue-resistance, better cyclability, faster color change, and higher conversion rate. The other one is to explore new applications. As for the second-order NLO materials, the aims include achieving new crystals with high laser-damaged threshold, high SHG, and fast growth methods for high-quality large-size crystals. In this chapter, electroluminescent materials, photochromic materials, and second-order NLO materials are separately introduced in Sections 10.1–10.4. The selected examples were basically published in the last 10 years. To fulfill the above aims, it is of importance to summarize relations between structure and optical properties for established materials. Thus, structure–optical property relations will be highlighted.

10.2 Organic Electroluminescent Materials–d8 /d10 Heteronuclear Metal Complexes Electroluminescence (EL) is an electro-optical conversion phenomenon driven by an electric current or a strong electric field. Organic electroluminescence is typically generated by means of a thin film device, i.e. so-called OLED having a sandwich structure composed of anode, light-emitting film, and cathode. As the core of the third-generation display technology, OLED is envisioned as one of the most competitive alternatives for the future solid-state display and lighting due to intrinsic advantages such as self-luminescence, high efficiency, high color rendering index, high contrast, ultrathin thickness, and flat and flexibility. To increase the mobility of carriers and improve the efficiency of electroluminescence, OLED is usually designed as a multilayer thin film device, but the core of the device is the light-emitting material. The luminescent emitters in OLEDs can be classified as fluorescent, thermally activated delayed fluorescent (TADF) and phosphorescent materials. According to the exciton statistical rule in OLED, while 100% energy can be utilized by phosphorescent and TADF materials, the energy utilization is only 25% for fluorescent materials. At present, the commercial OLED green and red-emitting materials are mainly phosphorescent cyclometallated Iridium(III) complexes, which are usually regarded as the second-generation light-emitting materials. Compared with phosphorescent mononuclear Iridium(III) complexes commercially used in OLEDs, polynuclear metal complexes display unique advantages as phosphorescent emitters. (i) They can be more easily prepared at room temperature in much higher yields so that the synthesis process is more simple and energy saving; (ii) intersystem crossing from singlet to triplet state is more facile due to

10.2 Organic Electroluminescent Materials–d8 /d10 Heteronuclear Metal Complexes

stronger spin–orbit coupling induced by multiple metal atoms, thus enhancing phosphorescent quantum yields; (iii) the polynuclear structure is more rigid, which can effectively suppress nonradiative relaxation by thermal vibration in triplet excited state to improve phosphorescent efficiency; (iv) the metal cluster structure exhibits higher stability to resist the photothermal radiation and improve the device lifetime; (v) the intermetallic contact comparable to hydrogen bond is not only beneficial for improving the stability to resist light and heat but also the participation of metal orbitals (d/p/s) in triplet excited state exerts a crucial influence on emissive energy and color. This section aims to focus on a family of d8 –d10 or d10 –d10 heteropolynuclear complexes and/or multicomponent complexes based on alkynyl ligands and their applications in OLEDs to achieve high-efficiency electroluminescence.

10.2.1 The Photoluminescence of d8 /d10 Heteronuclear Metal Complexes Luminescent complexes with d8 and d10 metal centers have been widely stimulated by their attractive spectroscopic and optical properties. A remarkable characteristic of d8 and d10 metal complexes differing from others is the capability to establish intense metal–metal contacts in clusters, whose energy is comparable with that of hydrogen bonds [1]. The d8 and d10 metal ions are accessible to form a large variety of coordination compounds that exhibit intriguing spectroscopic and optoelectronic properties with incorporation of various anionic or neutral ligands [2–5]. Additionally, d8 and d10 metal complexes generally exhibit manifold emissive origins arising from different structural features and intrinsic characteristics of the metal ions and the ligands [6–14]. Self-assembly is one of the most efficient approaches for the design of luminescent metal complexes with various molecular motifs by aggregating individual metal components into highly ordered oligo- and polymeric species. Thus, numerous mono-, di-, and trinuclear d8 and d10 metal complexes as well as metal clusters with higher nuclearities have been prepared using a varieties of ligands including halides, chalcogenides, thiolates, phosphines, acetylides, and N-heterocycles to reveal the correlation between structures and luminescent properties [2–14]. Among these functionalized ligands, thiolate, acetylide, and phosphine are easier to trigger the formation of ligand-linked metal cluster arrays through metal–metal interactions. It is noticeable that the metal–metal contacts make a critical difference in determining the spectroscopic properties and the emission features of these d8 and d10 polynuclear clusters [2–14]. Furthermore, the introduction of electron-donating or electron-withdrawing substituents to modify the electronic effect of the ligands is considered as one of the most effective pathways to regulate emission energy, lifetime, and quantum efficiency of the metal clusters. Alkynes are a large class of universal ligands to obtain metal alkynyl complexes with numerous nuclearities and spatial structural diversities depending on their bonding richness through σ- and/or π-coordination [14–21]. The metal-alkynyl species usually exhibit intense low-energy absorptions in the UV-vis region

533

534

10 Relationship Between Structure and Electroluminescent, Photochromic

mainly arising from alkynyl-centered π → π* (C≡CR) transition as well as charge transfer transitions between alkynyl ligands and metal ions or ancillary ligands [14, 17, 18]. More significantly, the intriguing luminescence with manifold emissive origins frequently occurs in d8 /d10 metal alkynyl complexes based on the structural features as well as the natures of the metal ions and the alkynyl ligands [14, 21]. Such multifunctional coordination systems are brilliantly qualified for exploring intramolecular triplet energy transfer between organometallic subunits as well as obtaining long-lived luminescent materials with high quantum efficiency. Compared with numerous homonuclear alkynyl complexes with d8 or d10 metal ions [14–21], the chemistry of d8 –d10 [22] or d10 –d10 [23] heteropolynuclear alkynyl complexes receives much less attention owing to the difficulties in controlling the heterometallic arrays. In view of the potentially bridging C donors in σ-bonding metal acetylides, one of the feasible synthetic strategies for attaining d8 –d10 or d10 –d10 heteropolynuclear alkynyl arrays is to incorporate d8 or d10 metal alkynl subunits with d10 coinage metal components. 10.2.1.1 d8 –d10 Heteronuclear Alkynyl Complexes

Among various available d8 metal alkynyl precursor species, the tetraalkynylplatinate(II) complexes [Pt(C≡CR)4 ]2− are particularly beneficial for the fabrication of heterometallic alkynyl cluster complexes due to their potential bridging character of the four alkynyl ligands through σ and/or π coordination [24]. Reactions of [Pt(C≡CC6 H4 R-p)4 ]2− (R = H, CH3 ) with [Ag2 (μ-PPh2 NHPPh2 )2 (MeCN)2 ]2+ caused isolation of neutral heterohexanuclear complexes Pt2 Ag4 (μ-Ph2 PNPPh2 )4 (C≡CC6 H4 R-p)4 (R = H, 1; CH3 , 2) [25]. These neutral Pt2 Ag4 clusters are likely considered as the combination of one anionic component [Pt2 (μ-Ph2 PNPPh2 )2 (C≡CC6 H4 R-p)4 ]2− with two cationic fragments [Ag2 (μ-Ph2 PNPPh2 )]+ instead of direct incorporation of [Pt(C≡CC6 H4 R-p)4 ]2− with [Ag2 (μ-PPh2 NHPPh2 )2 (MeCN)2 ]2+ . The formation of Pt2 Ag4 cluster structure is a direct consequence of facile deprotonating character of the PPh2 NHPPh2 . Ph2P

N

PPh2

Ag Ph2 P C C Pt C C Ph2P N PPh N 2 C C Pt C C P Ph2 Ag Ag Ag

R

R

Ph2P

N

PPh2

R = H 1, CH3 2

R

R

10.2 Organic Electroluminescent Materials–d8 /d10 Heteronuclear Metal Complexes

The anionic diplatinum(II) moiety [Pt2 (μ-Ph2 PNPPh2 )2 (C≡CC6 H4 R-p)4 ]2− adopts a face-to-face arrangement with the two square-planar PtII coordination planes oriented in parallel type [25]. The Pt atom is surrounded by two trans-arranged acetylide C atoms through η1 (σ) bonding and two trans-oriented P donors from deprotonated Ph2 PNPPh2 . Stronger intermetallic Pt–Pt contacts exist in the clusters as the Pt–Pt distances (3.15 Å for 1 and 3.11 Å for 2) are much shorter than those found in face-to-face diplatinum(II) complexes Pt2 (μ-Ph2 PCH2 PPh2 )2 (C≡CC6 H4 R-p)4 (3.25–3.44 Å) [26]. Such a consequence arises likely from direct interaction of the [Ag2 (Ph2 PNPPh2 )]+ subunit with [Pt2 (μ-Ph2 PNPPh2 )2 (C≡CC6 H4 R-p)4 ]2− moiety, which pulls the two PtII atoms into close proximity as a result of the reduced donating capability of the acetylides upon π-coordination with AgI centers. In addition to Pt–Pt interactions, other strong intermetallic contacts are reflected by the short Ag–Ag (3.18–3.33 Å) and Ag–Pt distances (2.90–2.94 Å) as well. The UV-vis electronic spectra of Pt2 Ag4 complexes show two low-energy absorption bands at 340–400 and c. 485 nm in acetonitrile solutions apart from ligand-centered high-energy absorptions below 300 nm. Upon irradiation of 1 and 2 at 𝜆ex > 320 nm, both solids and solutions of the Pt2 Ag4 species exhibit intense room-temperature luminescence with the lifetimes in the microsecond range. A remarkable blue-shifted emission from 1 (711 nm) to 2 (682 nm) in fluid acetonitrile solutions coincides with the higher π* energy level in the 4-methylphenylacetylide (2) relative to that in the phenylacetylide (1). The emissive origin is thus ascribed to the Pt2 Ag4 cluster to acetylide 3 [Pt2 Ag4 → RC≡C− ] MMLCT (metal–metal to ligand charge transfer) triplet state modified by strong intermetallic interaction [26]. A series of dppm-supported (dppm = bis(diphenylphosphino)methane) PtM, PtM2 , and Pt2 M3 (M = Cu, Ag, Au) heteronuclear alkynyl complexes [27] were synthesized by the reactions of the tetraalkynylplatinate(II) complexes [Pt(C≡CR)4 ]2− (R = But , C6 H5 , C6 H4 CH3 -4, C6 H4 But -4, C6 H4 OCH3 -4) with [M2 (dppm)2 ]2+ . Depending on the solvents and solution concentrations, PtM, PtM2 , and Pt2 M3 complexes with different nuclearities and structures could be isolated. The rather short Pt· · ·M distances (2.7–3.0 Å) suggest the presence of strong Pt–M interaction in these Pt–M heteronuclear complexes. They display weak to moderate luminescence in fluid solutions, but emit strongly in the solid states with a microsecond range of lifetimes. The phosphorescence is ascribed to substantial alkynyl-to-cluster [RC≡C → PtM/PtM2 /Pt2 M3 ] LMMCT (ligand to metal–metal charge transfer) transitions in view of the significant Pt–M interactions. Reactions of equivalent Ag(tht)(ClO4 ) (tht = tetrahydrothiophene) and bis(diphenyl-phosphinomethyl)phenylphosphine (dpmp) followed by the addition of 0.5 equiv. trans-Pt(PPh3 )2 (C≡CR)2 give rise to the isolation of a series of dpmp-supported [PtAg2 (dpmp)2 (C≡CR)2 ](ClO4 )2 complexes (3–10, R = aryl) in high yields [28]. Considerable short Pt⋅⋅⋅Ag (2.9–3.1 Å) distances indicate the presence of significant Pt–Ag interaction [26], which contributes significantly to the emissive properties.

535

536

10 Relationship Between Structure and Electroluminescent, Photochromic

R Rʹ Ph P

Ph2P Ag

Pt

Ph2P Rʹ R

P Ph

PPh2 Ag

2+

2ClO4–

PPh2

R=

Rʹ =

H CF3 CF3 But OMe OMe NH2 NMe2

3 H 4 H 5 CF3 6 H 7 H OMe 8 9 H 10 H

According to time-dependent density functional theory (TD-DFT) computational studies (Figure 10.1), the highest occupied molecular orbital (HOMO) of PtAg2 complex 3 is predominantly populated at phenylacetylide (66.9%) and PtAg2 atoms (27.5%, 5d(Pt) and 4d(Ag)), and the lowest unoccupied molecular orbital (LUMO) is uniformly resident on dpmp (43.2%), PtAg2 (34.0%, 6p(Pt) and 5p(Ag)) and phenylacetylide (22.8%) [28]. Therefore, electronic transitions due to HOMO → LUMO can be ascribed to significant [π(C≡CR) → π* (dpmp)] LLCT and PtAg2 cluster centered [d → p] states mixed with moderate [π → π* (C≡CR)] IL (intraligand) character. The UV-vis absorption spectra of 3–10 exhibit intense UV absorption bands at 380 nm. As depicted in Figure 10.2a, compared with the low-energy band of 3 at 402 nm, those of 4 (397 nm) and 5 (382 nm) are gradually blue shifted upon the introduction of one and two trifluoromethyl groups, respectively. In contrast, the low-energy absorption bands of 6 (411 nm), 7 (423 nm), 8 (438 nm), 9 (448 nm), and 10 (479 nm) with one or two But , OMe, NH2 , or NMe2 exhibit progressive red-shift relative to that of 3 (402 nm) because electron-donating substituents create the opposite effect.

HOMO

LUMO

Figure 10.1 Spatial plots of the HOMO and LUMO of complex 3. Source: Zhang et al. [28]. Copyright 2013, American Chemical Society.

10.2 Organic Electroluminescent Materials–d8 /d10 Heteronuclear Metal Complexes

ε ⨯ 105 (M/cm)

0.8 0.6 0.4 0.2

red shift

R 5 CF3 4 CF3 3 H t 6 Bu 7 OMe 8 OMe 9 NH2 10 NMe2

Rʹ CF3 H H H H OMe H H

red-shift Normalized intensity

1.0

5 3 6 7 8 9 10

red shift

R

Rʹ

CF3 H But OMe OMe NH2 NMe2

CF3 H H H OMe H H

0.0 250 300 350 400 450 500 550 Wavelength (nm)

(a)

500

600

700

800

900

Wavelength (nm)

(b)

Figure 10.2 (a) The UV-vis absorption spectra of PtAg2 complexes in CH2 Cl2 at ambient temperature. (b) The emission spectra (irradiation at 365 nm) of PtAg2 complexes (2.0 × 10−5 M) in fluid CH2 Cl2 solutions. Source: Zhang et al. [28]. Copyright 2013, American Chemical Society.

As shown in Figure 10.2b, upon irradiation at 𝜆ex > 300 nm, 3–10 display brilliant luminescence in both solid states and fluid CH2 Cl2 solutions at ambient temperature. The phosphorescence of this series of PtAg2 complexes was perfectly modulated by modifying the substituent in aromatic acetylide to realize a wide range of chrominance adjustment. The emission peaks exhibit a striking red-shift following 463 nm (5) → 476 nm (4) → 490 nm (3) → 502 nm (6) → 524 nm (7) → 552 nm (8) → 604 nm (9) → 625 nm (10), corresponding to phosphorescence color change of deep blue (5) → blue (4) → cyan (3) → green (6) → yellow-green (7) → yellow (8) → orange (9) → red (10). The red-shift of the emission coincides with the gradually increased electron-donating capability of acetylide ligands following CF3 → H → But → OMe → NH2 → NMe2 , ascribed to the increasingly reduced HOMO-LUMO gap. 10.2.1.2 d10 –d10 Heteronuclear Alkynyl Complexes

The chemistry in coinage metal acetylide complexes has experienced a great blossom, which is highly stimulated by their attractive optical properties and potential applications as optoelectronic materials [14]. A large amount of group 11 homonuclear cluster complexes were prepared and characterized structurally due to the bonding versatility and structural complexity of the coinage metal acetylides [14, 19, 20]. However, the synthetic difficulties in precise design of heteropolynuclear arrays bring about less attention to the chemistry of d10 –d10 heteropolynuclear acetylide complexes. To achieve luminescent group 11 heteropolynuclear alkynyl complexes, a variable approach has been established by utilizing d10 metal diphosphine component [M2 (μ-Ph2 PXPPh2 )2 (MeCN)2 ]2+ (M = Cu, Ag, Au; X = NH or CH2 ) to depolymerize polymeric d10 metal alkynyl species (M′ C≡CR)n (M′ = Cu, Ag, Au; R = alkyl, aryl). In this way, a series of highly phosphorescent d10 –d10 heteropolynuclear alkynyl cluster complexes have been prepared.

537

538

10 Relationship Between Structure and Electroluminescent, Photochromic

Au1 P1 Au3A P4 P5

P2A

Ag1 Ag2A

Au2

P3

Au4

Au2

P6 Ag1

Au2A

Au1

Ag2

P3

Au4A Ag1A

P2

P1

P2

P3A

Ag2

Au3

P1A

Au1A

(a)

(b)

Figure 10.3 ORTEP drawings of 12 (a) and 18 (b) showing 30% thermal ellipsoids. Phenyl groups on the P atoms are removed for clarity. Phenyl rings on phenylacetylides are also omitted for 18. Source: Xu et al. [29]. Copyright 2013, American Chemical Society.

The reactions of dpmp, Ag(tht)ClO4 , and polymeric (AuC≡CC6 H3 R′ R-2,4)n in 1 : 1 : 1 molar ratio gave a series of heterotetranuclear complexes [Au2 Ag2 (dpmp)2 (C≡CC6 H3 R′ R-2,4)2 ]2+ (11–16) [29]. Nevertheless, the heterododecanuclear clusters [Au8 Ag4 (dpmp)2 (C≡CC6 H3 R′ R-2,4)10 ]2+ (17–22) were obtained in 72–85% yield while the reactions were carried out in 1 : 1 : 3 stoichiometric ratio [29]. The Au2 Ag2 complex structure (Figure 10.3a) can be deemed as incorporating anionic [Au(C≡CC6 H5 )2 ]− into cationic [AuAg2 (dpmp)2 ]3+ through Ag–acetylide bonding and Au–Au (3.0581(14) Å) interaction between the two components. The Au8 Ag4 heterododecanuclear cluster 18 (Figure 10.3b) is viewed as a dimer of Au4 Ag2 species, in which two Au4 Ag2 moieties are bound to each other via ligand-unsupported Au–Ag (dAu4A–Ag1 = 3.1391(9) Å) contact. The two types of cluster structures are sufficiently stable in solutions without degradation or interconversion at room temperature.

Ph2P Rʹ

Ag

R

Ph P PPh2 Au

Ag

R

Rʹ

Au

PPh2 P Ph 13 14 15

11

12 H H

2+

Rʹ

Rʹ

R Rʹ

Ph2P Au

R

Ag Rʹ R

Ph2P

R = CF3 Rʹ = H

R

2+

16

But OMe OMe NMe2 H H OMe H

PhP

Au

Ph2P

Au

Au' Ag R R

R Rʹ Au Ag R Rʹ

R Rʹ

17 R = CF3 Rʹ = H

R 18 H H

19

Au

PPh2

Au

PPh

Au

PPh2

Ag Rʹ

R 20

21

22

But OMe OMe NMe2 H H OMe H

10.2 Organic Electroluminescent Materials–d8 /d10 Heteronuclear Metal Complexes

The UV-vis absorption spectra of complexes 11–22 in dichloromethane solutions display high-energy absorptions at 330 nm tailing to 470 nm, arising primarily from 1 LLCT and Au2 Ag2 metal-centered [d → p] transitions for 11–16, whereas Au8 Ag4 centered [d → p] and 1 IL (C≡CC6 H3 R′ R-2,4) states for 17–22. Upon irradiation at 𝜆ex > 300 nm, bright luminescence occurs in both solid states and fluid dichloromethane solutions at ambient temperature. Unlike intense phosphorescence in Au2 Ag2 complexes 11–16, the emission is much weaker in Au8 Ag4 complexes 17–22. Rapid association/dissociation between two Au4 Ag2 units in 17–22 through ligand-unsupported Au–Ag contact is responsible for the lower phosphorescence efficiency. The phosphorescence emission over the whole visible region can be systematically modulated. The introduction of electron-donating But , OMe, or NMe2 to phenylacetylides results in red-shift of the emission while it exhibits remarkable blue-shift upon introducing electronwithdrawing CF3 .

10.2.2 The Electroluminescence of d8 /d10 Heteronuclear Metal Complexes In 1998, phosphorescent platinum(II) porphyrin complex was utilized by Baldo et al. [30] as emissive dopants to fabricate OLEDs, generating saturated red emission with the peak external quantum efficiency (EQE) of 4% and internal quantum efficiency of 23% [30]. Shortly afterward, Baldo et al discovered highly phosphorescent cyclometalated iridium(III) complexes to achieve high-efficiency OLEDs with the peak EQE of over 8% [31]. Since then, a large number of cyclometalated iridium(III) species with various structures and emission colors have been synthesized to attain high-efficiency OLEDs with EQE exceeding 30% to date [32–34]. Meanwhile, a variety of highly phosphorescent platinum(II) complexes have been elaborately developed to achieve excellent OLEDs with EQE over 30% [35]. Compared with mononuclear metal complexes, d8 /d10 heteronuclear metal complexes exhibit great advantages as phosphorescent emitters in OLEDs. A quantity of phosphorescent d8 /d10 heteronuclear metal complexes with manifold nuclearities have been reported with various emission colors over the entire visible region [28, 29, 36], but the applications of heterometallic metal cluster complexes as phosphorescent emitters in OLEDs are scarce [37–46]. In recent years, our group reported the use of a series of d8 /d10 heterometallic complexes (23–58) to achieve high-efficiency OLEDs through solution process. The photoluminescence and electroluminescence are systematically modulated by regulating metal cluster structures with manifold nuclearities as well as modifying bridging and ancillary ligands. The electroluminescent data of 23–58 are summarized in Table 10.1. 10.2.2.1 The Electroluminescence of d8 –d10 Heteronuclear Complexes

The formation of aggregated structures through intramolecular metallophilic contacts can not only stabilize the metal cluster frameworks but also promote singlet to triplet intersystem crossing and dramatically prohibit non-radiative deactivation.

539

540

10 Relationship Between Structure and Electroluminescent, Photochromic

Table 10.1 The performance data of OLEDs based on phosphorescent d8 /d10 heterometallic complexes.

Complex

𝝀EL (nm)

23 24

V on (V)

Lmax (cd/m2 )

CEmax (cd/A)

PEmax (lm/W)

EQEmax (%)

499

4.2

2418

7.7

4.5

2.7

506

3.6

11 270

15.5

7.1

4.9

25

520

6.1

4503

29.1

12.2

8.1

26

548

4.0

138 584

34.0

23.8

9.8

27

556

4.1

13 808

51.7

26.4

14.5

28

532

4.7

24 408

78.2

26.4

21.5

29

547

4.5

32 915

78.3

40.4

20.2

30

517

5.0

7356

30.0

12.5

10.1

31

605

4.5

10 446

35.4

16.8

18.7

32

568

4.95

8975

21.7

9.7

8.0

33

527

4.7

21 975

67.4

33.0

17.4

34

547

5.7

20 447

58.4

24.0

15.3

35

526

6.5

12 363

25.2

9.7

6.8

36

547

5.1

4747

12.9

6.0

3.3

37

556

4.4

14 241

50.0

29.0

14.5

38

568

5.7

15 081

64.6

26.1

18.3

39

486

4.8

1703

27.2

13.3

11.1

40

527

4.8

7764

61.0

30.9

18.1

41

537

4.7

6652

57.0

28.7

16.6

42

616

4.6

1898

19.8

9.9

12.4

43

572

3.9

2336

30.7

18.8

10.4

44

527

3.0

10 415

38.7

22.9

10.3

45

539

3.4

17 079

55.6

35.1

14.8

46

565

3.5

14 755

45.5

28.3

14.1

47

541

5.2

10 911

58.3

26.1

14.9

48

556

5.5

12 711

62.8

25.1

17.7

49

556

4.2

6539

62.2

30.3

16.6

50

568

5.4

12 702

57.1

20.1

13.1

51

588

4.9

19 308

45.2

20.3

18.1

52

482

6.6

7981

7.9

2.7

3.9

53

490

6.7

6468

12.1

4.0

5.3

54

518

8.4

17 160

21.8

5.6

6.5

55

539a

4.6

8804

24.1

11.6

7.0

56

550

7.7

17 651

20.9

6.0

6.1

57

565

3.5

1300

31.8

20.8

10.4

58

573

3.4

3117

42.5

30.0

13.9

10.2 Organic Electroluminescent Materials–d8 /d10 Heteronuclear Metal Complexes

Although many mononuclear platinum(II) precursor complexes are nonluminescent or weakly emissive at ambient condition, the corresponding polynuclear Pt–M (M = Cu, Ag, Au) aggregates are highly phosphorescent when mononuclear platinum(II) precursors are bound to copper(I), silver(I), or gold(I) ions through d10 metal-acetylide bonding. 10.2.2.1.1 Triphosphine-Supported Pt–M Heteronuclear Complexes

A class of triphosphine-supported PtAu2 complexes [PtAu2 (dpmp)2 (C≡CR)2 ]2+ (23–32, R = aryl) [37–39] were isolated by the reactions of Pt(PPh3 )2 (C≡CR)2 , [Au(tht)2 ]ClO4 and dpmp in 1 : 2 : 2 stoichiometric ratio. These PtAu2 complexes are extremely stabilized by four five-membered coordination rings and significant Pt–Au interactions. They show intense phosphorescence in fluid solutions, crystals, powders, and doping films with the highest quantum yield of over 90%. The variable emission energy was realized by adjusting substitutional groups or π-conjugated aromatic systems in acetylide ligands to obtain manifold emission colors. The phosphorescent emission is mainly ascribed to 3 [π (C≡CR) → π* (dpmp)] 3 LLCT and 3 [π (C≡CR) → s/p (PtAu )] 3 LMCT triplet excited states for carbazole2 or phenothiazine-containing complexes as well as noticeable PtAu2 centered 3 [d → s/p] parentage for other complexes.

Ph2P Au Ph2P R2 R1 =

Ph P Pt P Ph

R1

2+

PPh2 2ClO4–

Au PPh2

R1 =

R2 =

R2 =

23 But

But

O

O

NN

NN

O NN

N

24

N

Et

Ph

25 N

30 N

Ph

Ph

26 Et S

N

27 N

29 N

Ph

N

S

31

N Et Et

N

N

Ph

Ph

Et

N

28 S

N S

32

541

10 Relationship Between Structure and Electroluminescent, Photochromic

120

104 Current density Luminance

100

103

80 60

102

40

101

20

Luminance (cd/m2)

Current density (mA/cm2)

Accessibility at mild conditions, excellent phosphorescent quantum yield, and favorable solubility in organic solvents enable these cationic PtAu2 heterotrinuclear complexes as ideal phosphorescent dopants for the fabrication of OLEDs through orthogonal solution process. When 8% PtAu2 species were doped to blended host materials consisting of hole-transport TCTA/mCP and electron-transport OXD-7 as light-emitting layers, the devices display highly efficient electroluminescence with the highest current efficiency (CE) close to 80 cd/A and EQE over 21%. Relative to those of 23 (𝜂 EQE = 2.7%) and 24 (𝜂 EQE = 4.9%) with phenylacetylide ligands, the devices of other PtAu2 complexes 25–32 (𝜂 EQE = 8.0–21.5%) with various functional aromatic acetylide exhibit much better performance. In most cases, the introduction of hole-transport groups to the heteronuclear system is beneficial to accomplish superior electroluminescent efficiency. The electroluminescent spectra of isomeric complexes 28–30 [38] with acetylide groups in different positions of functionalized carbazole-acetylide ligands exhibit a stepwise blue-shift following 547 nm (29) → 532 nm (28) → 517 nm (30) to obtain green, yellow-green, and bluish green emission, respectively. The phenomenon makes it possible to modulate electroluminescence and realize various emission colors of PtAu2 complexes by taking advantage of different positioned

100

0 0

2

4

6 8 10 Voltage (V)

12

14

16

(a) 100 20

80

15

60 40

5

20 0 2000

(b)

10

Current efficiency EQE

4000

6000

0 8000 10000

Luminance (cd/m2)

EQE (%)

Current efficiency (cd/A)

542

Figure 10.4 (a) Current density–voltage–luminance characteristics. (b) Current efficiency/external quantum efficiency versus luminance for OLED based on complex 28. Source: Xu et al. [38]. Copyright 2016, American Chemical Society.

10.2 Organic Electroluminescent Materials–d8 /d10 Heteronuclear Metal Complexes

9-phenylcarbazole-acetylide ligands. The peak CE and EQE of the devices were 78.2 cd/A and 21.5% for 28, 78.3 cd/A and 20.2% for 29, and 30.0 cd/A and 10.1% for 30 with a slow efficiency roll-off in practical brightness range of 500–5000 cd/m2 (Figure 10.4). For 31 and 32 [39] with phenothiazine-functionalized acetylide groups, solution-processed OLEDs achieve orange-red (𝜆EL = 605 nm) and yellow (𝜆EL = 568 nm) electroluminescence with the peak CE and EQE of 35.4 cd/A and 18.7% for 31 and 21.7 cd/A and 8.0% for 32. Analogously, the PtAg2 heterotrinuclear complexes 33–35 [40] supported by dpmp show rigid molecular structures through Pt/Ag-acetylide linkages with the acetylides in μ-η1 , η1 bonding mode. The PtAg2 complexes are highly phosphorescent in both dichloromethane solutions and doping films at ambient temperature, ascribable to 3 [carbazole-acetylides → dpmp/PtAg2 ] 3 LLCT/3 LMCT and metal cluster centered 3 [d → s/p] triplet states. When 8% PtAg2 complexes were doped to blended host material of TCTA and OXD-7 or mCP and OXD-7 in a 1 : 1 ratio as phosphorescent-emitting layers, solution-processed OLEDs give high-efficiency electroluminescence with the highest power efficiency (PE), CE, and EQE of 33.0 lm/W, 67.4 cd/A, and 17.4% for 33, 24.0 lm/W, 58.4 cd/A, and 15.3% for 34, and 9.7 lm/W, 25.2 cd/A, and 6.8% for 35, respectively.

Ph2P Ag Ph2P R

R=

Et

33

2+

PPh2 Ag

Pt P Ph

Et

N

R

Ph P

2ClO4–

PPh2

N

Et

34

N

35

Recently, a new class of dpmp-supported Pt2 Au heterotrinuclear complexes 36–38 as light emitters [41] was reported to reveal the relationship between structural rigidity and electroluminescence performance. The reaction of Au(tht)Cl, dpmp, and Pt(PPh3 )2 (C≡CPh)2 in 1 : 2 : 2 molar ratio nearly isolated Pt2 Au heterotrinuclear complex 36. While diacetylide-linked binuclear complexes [Pt2 (PPh3 )4 (DEBf)(C≡CPh)2 ] and [Pt2 (PPh3 )4 (DECz)(C≡CPh)2 ] (DEBf = dibenzofuran-4,6-diacetylide; DECz = 3,6-di-tert-butylcarbazole-1,8-diacetylide) were used as the precursors in place of Pt(PPh3 )2 (C≡CPh)2 , the reactions gave Pt2 Au complexes 37 and 38, respectively.

543

544

10 Relationship Between Structure and Electroluminescent, Photochromic

Ph2P Pt Ph2P

PPh PPh2 Au

+

Ph2P

Pt ClO4–

PPh PPh2

Pt Ph2P

O PPh PPh2 Au

+

Ph2P

– Pt ClO4

PPh PPh2

Pt Ph2P

Au

+

– Pt ClO4

PPh PPh2 38

37

36

N H PPh PPh2

Interestingly, as demonstrated in Table 10.2, the relatively flexible structure of Pt2 Au complex 36 (𝜆em = 503 nm, Φem < 0.1%) results in weak phosphorescence in fluid dichloromethane solution. However, much more intense photoluminescence is exhibited for complexes 37 (𝜆em = 585 nm, Φem = 4.9%) and 38 (𝜆em = 589 nm, Φem = 3.2%) with rigid conformation. The use of a rigid diacetylide ligand in place of two phenylacetylide ligands to fasten the Pt2 Au framework enhances the molecular rigidity so that phosphorescence is dramatically promoted because of the effective suppression of non-radiative vibrational deactivation. As a result, the emissive quantum yield in doping films is up to 89% for 37 and 93% for 38. By means of Pt2 Au complexes as phosphorescent emitters, solution-processed OLEDs exhibit relatively inferior device performance (EQE < 9.5%) by the use of commercially available PEDOT:PSS as hole-injection layer (HIL). Nevertheless, the electroluminescent efficiencies are greatly increased upon the addition of PSS-Na to PEDOT:PSS to afford m-PEDOT as HILs, which optimizes the work function and film morphology. The solution-processed devices based on m-PEDOT gave the peak CE and EQE of 50.0 cd/A and 14.5% for 37 and 67.7 cd/A and 19.0% for 38, respectively. Table 10.2

The luminescence data of Pt2 Au complexes 36–38. 𝝀em (nm)/𝝉 em (𝛍s)/𝚽em (%) 36

37

38

503/0.90/ ZnBr4 2− (−6.074 eV) > ZnCl4 2− (−6.484 eV). That is to say, electron-transfer efficiency should be in the order: (HNH)[ZnI4 ]⋅H2 O > (HNH)[ZnBr4 ]⋅H2 O > (HNH)[ZnCl4 ]⋅H2 O. Experimental results demonstrated this point. (HNH)[ZnI4 ]⋅H2 O showed eye-detectable color

559

560

10 Relationship Between Structure and Electroluminescent, Photochromic

hv

hv

/dark 3.075(15)

4.464(7)

b

a c

b a

(MCMP)[AgI2]

(MCMP)[Ag3I4]

Figure 10.17 Crystal structures and photochromism of (MCMP)[AgI2 ] and (MCMP)[Ag3 I4 ]. Source: Edited with permission Yu et al. [73]. Copyright 2016, The Royal Society of Chemistry.

change upon irradiation, while the other two had no color change even if they also generated some radical products (Figure 10.18).

10.3.2 Optical Applications of Viologens and their Analogs The electron-transfer behavior of viologens and their analogs offers a great chance to explore new optical applications. Here presents typical examples on opto-optical switching, radiation detection, and photocatalysis, respectively. 10.3.2.1 Opto-Optical Switching 10.3.2.1.1 Photoswitching of Luminescence

Photoswitching luminescence has potential application on bioimaging and anti-fake. The electron-transfer behavior of viologens and their analogs brings two results. One is that newly generated electron absorption bands for the radical products cover luminescent emission bands. This so-called “self-absorption” phenomenon weakens the luminescence, and thus an ON–OFF type luminescence switch can be formed [75]. One typical example is (MV)4 Bi6 Cl26 ⋅2H2 O (MV2+ = N,N-dimethyl viologen; Figure 10.19) [75]. This compound displayed interesting partial photochromic phenomenon, that is, photoinduced coloration is clear in one part and unclear in the other part. The as-prepared sample had

1.4

0.6 0.4 0.2

1.0

Absorbance (a.u.)

0.8

Absorbance (a.u.)

Absorbance (a.u.)

1.0

2 2P

0.8 0.6 0.4 0.2 0.0

0.0 200

1.4

1.2 1 1P Decolored

1.2

300

400 500 600 Wavelength/(nm)

700

–0.2 800 200

LUMO

1.2

400 500 600 Wavelength/(nm)

700

800

LUMO

0.6 0.4 0.2 0.0

300

LUMO

3 3P

1.0 0.8

200

–0.684 300

400 500 600 Wavelength/(nm)

700

–1.110

800

–1.485

hv

Energy (ev)

hv hv

Dark 1

2

3

–3.623

–5.589

–7.629

1.92

1.96

2.00 g

2.04

2.08

(HNH)[ZnI4].H2O(1) (a)

1.92

1.96

2.00 g

2.04

–7.575

3P

Decolored

Decolored

Decolored

–6.074

–6.484 –7.638

2P

1P

–3.612

–3.644

2.08

(HNH)[ZnBr4].H2O(2)

1.92

1.96

2.00 g

2.04

2.08

HNH2+

(HNH)[ZnCl4].H2O(3)

ZnCl42– HOMO

HNH2+

ZnBr42– HOMO

HNH2+

ZnI42– HOMO

(b)

Figure 10.18 Comparison of photochromic behavior (left: top, UV–vis spectra; middle, color change; bottom, ESR data) and frontier orbitals (right) for (HNH)[ZnX4 ]⋅H2 O (X = Cl, Br, I). Calculations were performed at the PBE/TZP level. Source: Based on Shen et al. [72]. Copyright 2017, The Royal Society of Chemistry.

130 °C 400

500

600

700 Before irradiation After irradiation

Absorbance (a.u.)

0.45

800 0.4

0.3

0.2 0.40

Monochromator change 0.1

0.35

180.0k

Before irradiation

160.0k

After irradiation

140.0k 120.0k 100.0k

uv

∆

80.0k 60.0k 40.0k

New absorption 0.0

20.0k 0.0

0.30 300 UV irradiation

200.0k

Intencity (cps)

300 0.50

400

500 600 Wavelength (nm)

700

800

450

500

550 600 Wavelength (nm)

650

700

Figure 10.19 Photochromic phenomenon and changes of absorption and luminescence for (MV)4 Bi6 Cl26 ⋅2H2 O. Source: Xu et al. [75]. Copyright 2010, The Royal Society of Chemistry.

10.3 Photochromic Materials–Viologen Compounds and their Analogs 1.0

R (%)

Excitation (reading)

70

40

Neutral form

O2

400 600 Wavelength (nm)

Excitation (reading) 365 nm coloration (writing) O2 bleaching (erasing)

UV

616 nm emission “On” state

30 200

465 nm

10 min

60

465 nm

800

Reduction form

Emission quenched “Off” state

Relative intensity

0 min

80

50

1.0 0 min

90

0.8

Normolized intensity

100

0.8 7

0.6 0.4

0.6 0.4

10 min

F0-5D6

7

F0-5D2

7

F0-5D2 7

F

0.2

0.1

-5D

1

0.0 250 300 350 400 450 500 550 Wavelength (nm)

0.2 0.0 550

600 650 Wavelength (nm)

700

Figure 10.20 Photochromism and luminescence switch of {[Eu(BA)(Bpybc)1.5 (H2 O)] (NO3 )2 ⋅5H2 O}n . Source: Sun et al. [76]. Copyright 2011, The Royal Society of Chemistry.

an emission band peaked at approximately 580 nm. After coloration, the yielded MV+⋅ radical generated a broad electron absorption band ranging from 450 nm to more than 800 nm. This absorption band overlapped the emission band, resulting in a significant luminescence drop. The observed switching ratio for luminescence reached 7.5 times. The other result is that energy levels of light absorber change and thus the so-called “antenna effect” for energy transfer disappears. One typical example is {[EuIII (BA)(Bpybc)1.5 (H2 O)](NO3 )2 ⋅5H2 O}n , (HBA = benzoicacid; H2 BpybcCl2 = 1,1′ -bis(4-carboxybenzyl)-4,4′ -bipyridiniumdichloride; Figure 10.20) [76]. Under excitation of 465 nm light, the as-prepared sample displayed intense red light emission owing to the presence of EuIII ions. The emission resulted from energy transfer from the ligands to EuIII . After irradiation by 365 nm light, this compound changed its color from yellow to dark blue. Energy levels of the ligands in the yielded radical product did not match the need for energy transfer, resulting in complete disappearance of luminescence. For the above examples, excitation bands for coloration and luminescence largely overlapped. The information written by coloration may be perturbed when reading through luminescence. Thus, one important issue toward applications is how to avoid the overlap. An effective method is to combine photochromophore and luminophore with different absorption bands in one compound. The readers can find related examples from the reviews written by Raymo and Tomasulo [77] and Fukaminato [78]. 10.3.2.1.2 Photoswitching of NLO Properties

Materials with switchable NLO properties, such as SHG effect, have some potential applications. For example, they provide an in situ, reversible, and controllable method for regulating NLO properties. This method may be applied in the currently emerging nanolaser field to adjust laser intensity. In addition, the switchable NLO properties can behave as readout signals for sensing, data storage, and other applications. For instance, they can be applied in non-destructive data storage of photochromic materials, because light of long wavelengths outside of the absorption band does not induce photoreaction. There have been many approaches to switch the SHG effect. Ion exchange [79] and gas adsorption/desorption [80] methods have been found to be effective for porous materials. Compared with these chemical methods, physical methods such as irradiation and heating are more convenient and applicable to more

563

10 Relationship Between Structure and Electroluminescent, Photochromic

material systems. As the most studied physical approach, SHG photoswitching has been almost realized in liquid or film photochromic materials with the presence of external optical [81]/electric [82] fields or special supported media [83]. For example, the Guerchais group reported photo-induced SHG-switching properties of some photochromic dithienylethene-based platinum(II) complexes in solution measured by the EFISH technique under an electric field [82]. Second-order NLO crystals do not require the above severe conditions to switch the SHG effects owing to their intrinsic polarities. For instance, Sliwa et al. found that photochromic anil molecules in the crystalline state exhibited SHG effects and SHG-switching properties without the induction of any external stimuli [84]. However, in an early research effort, photoswitching of the SHG effects in NLO crystals was fulfilled through large molecular isomerization reactions of organic photochromic components [49, 85]. The large structural change needs sufficient steric space and is usually prohibited in the crystal lattice. Electron-transfer photochromic processes are only accompanied by minor structural variation, which is more adaptable to a constrained medium [53]. These processes result in the rearrangement of charges and thus change the permanent dipole moment, which is closely related to the SHG effect of a crystal. First, Li et al. used electron-transfer photochromism to switch SHG of NLO crystals. They found that the photochromic compound [ZnBr2 (CEbpy)]⋅3H2 O (CEbpy = N-carboxyethyl-4,4′ -bipyridinium, Figure 10.21) had a SHG-switching contrast of 3.3 times [86], which was larger than those of previously reported photo-switchable NLO crystals (1.1–2.5 times). Thereafter, the Zang group reported that electron-transfer photochromism was also rather effective in switching the SHG efficiency of a crystalline viologen-functionalized chiral Eu-MOF, giving a SHG-switching contrast of 2.5 times [87]. Even so, the relative studies are just starting out, and it is still desirable to achieve NLO crystals with higher SHG-switching contrasts. An applicable method is to increase the ability to stabilize received electrons and the change of permanent molecular dipole moment. MQ+ (N-methyl-4,4′ -bipyridinium) has a more negative difference in Gibbs-free energy and a larger change in permanent dipole moment than those of the abovementioned CEbpy ligand after receiving one electron (ΔG: −5.42 eV for MQ+ , −2.26 eV for CEbpy; change of permanent dipole moment: −2.62 Debye for MQ+ , −0.48 Debye for CEbpy; calculated at the B3LYP/6-31+ G(d,p) level), which is indicative of a 200 SHG intensity (a.u.)

564

(Dipole moment) e–

e– –

+

0 min 10 min 20 min 35 min

150

Figure 10.21

150 100 50

100

3.3 times

Colored 0

1

2 3 Cycle times

4

50 0

– +

945

(a)

Decolored

200

(b)

950

955 𝜆 (nm)

960

965

SHG switching (right) of [ZnBr2 (CEbpy)]⋅3H2 O through electron transfer (left).

10.3 Photochromic Materials–Viologen Compounds and their Analogs

better ability to stabilize received electrons after electron transfer and a higher probability to obtain a large SHG-switching contrast. Using D(+)-camphoric acid as an acentric inducer, Xing et al. synthesized a new photochromic and thermochromic bifunctional compound, β-[(MQ)ZnCl3 ] [88]. This compound had a SHG-switching contrast of about eight times after laser irradiation or two times after thermally annealing. The former value is more than twice that of the highest-known record (∼3.3 times) for photoswitchable NLO crystals [86]. β-[(MQ)ZnCl3 ] also represents the first thermoswitchable NLO crystal that can maintain the expected second-order NLO intensity without any heat source. In the above examples, the SHG effects were all reduced after photo/heat-induced coloration. In addition, self-absorption might cause damage of the crystals although it favors the enhancement of SHG-switching ratio. Exploring suitable structural design strategies is highly desirable in the future. 10.3.2.2 Radiation Detection

X-rays are common in medical services (diagnosis and cancer therapy), crack detection, security checks, lab analysis, archeology, astronomy, and nuclear development. A series of detectors (such as scintillation counters, ionization gauges, and semiconductor detectors) and silver salt-based radiographic films have been developed and widely used to detect X-rays. However, it is difficult to detect low-energy X-rays using these, especially those with energies less than 2 keV, because of their low spatial resolution and low sensitivity. In recent years, radiochromic materials have been investigated extensively because of their greater X-ray sensitivity over silver salt-based radiographic films and higher spatial resolution for the detection of low-energy X-rays than commercial instruments [89]. X-ray-induced photochromic (XP) compounds, being an important subgroup of radiochromic materials, can be found in Prussian blue analogs, FeII spin-crossover complexes, cobalt dioxolene complexes, oxometallates with impurities (such as LiNbO3 :Cu), BiIII salt–nylon composites, scintillator–spiropyran composites, N-substituted bipyridinium-containing compounds, metal complexes bearing nonionic N-heterocyclic aromatic ligands, and NDI-based compounds [62]. Unlike other radiochromic materials and also the silver salt-based radiographic films, they can display color change when exposed to X-rays and restore the initial color after standing in the dark, heating or illumination. Such a reversible feature means that they may be used many times. However, only a few of them respond to hard (𝜆 < 1 Å) and soft (𝜆 > 1 Å) X-rays at room temperature. In 2012, Wang et al. reported the first photochromic compound that may response to both hard and soft X-rays at room temperature, [Zn(bpp)Br2 ] (XP-1; bpp = 4,4′ -bipyridinium-N-propionate; Figure 10.22) [90]. This compound showed color change from yellow to blue after irradiation by Mo K α (𝜆 = 0.7107 Å), Cu K α (𝜆 = 1.5406 Å), or Al K α (𝜆 = 8.357 Å) X-rays at room temperature. The color change was ascribed to X-ray-induced electron transfer from Br− to bpp, which resulted in the formation of a radical product with an ESR signal around 2.00 and electron absorption bands similar to those of the MV+⋅ radical. The coloration became more faster when the wavelength of X-ray decreased. The radical product

565

10 Relationship Between Structure and Electroluminescent, Photochromic

10mm

1mm 4mm

+ N

N

COO

O

C

–

Cu kα

Zn N

Mo kα

Br

[Zn(bpp)Br2]

bpp

30 s

+ N

N

5 min

AI kα

C N Zn Br

5 min

[(MQ)ZnBr3]

MQ+ (a)

5 min

(b)

Figure 10.22 (a) Molecular structures for bpp, [Zn(bpp)Br2 ], MQ+ , and [(MQ)ZnBr3 ]; (b) X-ray-induced photochromism of [Zn(bpp)Br2 ]. Source: Edited with permission Wang et al. [90]. Copyright 2012, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

could be retained more than half a year in dark at ambient temperature, but easily returned to the initial state by thermal annealing at 150 ∘ C for one hour. The whole coloration and discoloration process could be cycled more than 10 times. Similar to [Zn(bpp)Br2 ], [(MQ)ZnBr3 ] [91] (Figure 10.22) also displayed color change after exposing to Mo/Cu/Al K α X-rays at room temperature, but the radical product easily faded in air. A possible reason for this observation is that the bridging coordination to ZnBr2 enhances the rigidity of the bpp ligand in [Zn(bpp)Br2 ] which favors the stabilization of radicals. Differ from [Zn(bpp)Br2 ] and [(MQ)ZnBr3 ], the hydrated bpp ligand did not response to Mo K α or Cu K α X-ray, implying that the presence of atoms with high atomic numbers benefits X-ray response. After the discovery of XP-1, several groups found that some viologen compounds also displayed X-ray-induced photochromism at room temperature [92]. The Zhang group [92] found that ([Co(Bpybc)1.5 (H2 O)3 ]NO3 (OH)⋅11H2 O (blm-Co; Figure 10.23) was highly sensitive to X-rays. Different from previously reported XP materials, in which the information of structural transformation is too subtle to be detected after X-ray irradiation, blm-Co showed clear structural change after exposing to X-ray. This phenomenon is unusual. In addition, the coloration 110 100

Blank 20 s 40 s 1 min 5 min 10 min 15 min 20 min 25 min

90

R(%)

566

80 70 60

40 min 60 min 120 min 180 min 300 min

Twist angle 15.1(7)° e

c

50 40 30 20

∠N-C-C = 111.5(5)°

O...N distance 4.166(5) A

X-ray

b a Blank

500 (a)

∠N-C-C = 111.0(5)°

d

Twist angle 16.9(5)° O...N distance 4.288(6) A

1 min

550

10 min

15 min

60 min

600 650 700 Wavelength (nm)

300 min

750

800

(1 0 1) (1 0 –2) (2 0 –1) (2 –1 3) (3 0 0)

5 (b)

10

15

20 2θ (°)

25

blm-Co

30

blm-Co-X

(c)

Figure 10.23 For blm-Co: (a) time-dependent color change upon irradiation of Cu K α X-ray; (b,c) structural change monitored by powder and single-crystal X-ray diffraction analysis. Note: a and b, calculated and experimental powder patterns of the freshly prepared sample; c and d, calculated and experimental powder patterns of X-ray-irradiated sample; e, experimental powder pattern of the X-ray treated sample after standing in air for one hour. Source: Edited with permission Chen et al. [92]. Copyright 2017, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

10.3 Photochromic Materials–Viologen Compounds and their Analogs

time was only one minute when blm-Co was exposed to Cu K α X-ray from an X-ray generator with power of 40 W. This time was shorter than ever found XP compounds. Guo et al. [92] reported that introducing more atoms with high atomic number favored the shortening of coloration time. They found that [Zn2 (N,N ′ -dipropionate-4,4′ -bipyridinium)Br4 ]n (XP-2) showed color change when irradiated by Al K α X-ray for only one second. This is the fastest coloration time observed in known XP materials. Zhang and Wu [93] and Su et al. [62], respectively, found that metalloviologen compounds also might response to hard and soft X-rays at room temperature. By incorporating a radiosensitive glycolate ligand into an electron-transfer photochromic metalloviologen system, Su et al. obtained the first radiochromic material with energy-dependent X-ray-induced photochromism, [Zn2 (Dg)2 (4,4′ -bipy) (H2 O)4 ]⋅2H2 O (XP-3; Dg = diglycolate; Figure 10.24) [62]. The as-prepared pale-yellow crystalline sample of XP-3 displayed dark purple and purple when exposing to Al K α X-ray and Mo/Cu K α X-ray, respectively. Except for the colors, electron absorption spectra and ESR spectra were significantly different. The reason for these phenomena is that an electron-transfer process and a proton elimination process dominated when Al K α X-ray and Mo/Cu K α X-ray were used, respectively. It is very important to distinguish the energies of X-rays. X-ray detectors in synchrotron radiation facilities may serve this purpose. They use a set of single crystals to divide the X-rays according to the Bragg equation. But beyond that, commercial X-ray detectors are usually applied to detect the dosage of X-rays instead of the energies (or wavelengths). The established radiochromic or radiographic materials can display only one color when exposed to X-rays. The

O4

O5

Energy dependent

O2 O6 (Dg) O7

Zn1

Mo-Kα 0 h

1h

2h

X-Ray RADIATION

O1 Al-Kα 0 s

O4

(a)

300 s

(b) 0.6

Absorbance (a.u.)

0.4

XP-3

Al-Kα Cu-Kα As-synthesized

0.5 355

H2Dg As-synthesized As-purchased Al-Kα Decolored Al-Kα

0.3 0.2

376

Cu-Kα Decolored Cu-Kα

508

Mo-Kα

562

0.1

Decolored

0.0 300

(c)

10 s Time dependent

400

500

600

Wavelength (nm)

700

1.99 2.00 2.01 2.02

800

g-factor

1.99 2.00 2.01 2.02

g-factor

(d)

Figure 10.24 For XP-3: (a) molecular structure; (b) energy-dependent X-ray photochromism; (c) changes of electron absorption bands; (d) changes of ESR signals. Source: Edited with permission Su et al. [62]. Copyright 2018, the Royal Society of Chemistry.

567

10 Relationship Between Structure and Electroluminescent, Photochromic

unprecedented energy-dependent XP behavior makes XP-3 a potential candidate to qualitatively distinguish X-rays in a visual manner. Toward real application, one urgent issue is to further improve the coloration time when exposing to hard X-rays. The shortest observed time was more than one minutes [92]. For the real application, it is best to achieve a coloration time closing to the range of ms to s. Increasing the delocalized degree of the electron acceptors and the numbers of atoms with high X-ray absorption ability is an applicable route to address this issue. 10.3.2.3 Photocatalysis

Viologen cations were often used as electron mediators or oxidants for photochemical reactions, owing to their high electron-accepting ability. Reduced viologen species, usually existing as viologen radical monocations, can reduce protons to give H2 . Thus, H2 generation by solar energy conversion has been realized through photoreduction of viologens and photooxidation of EDTA with tri(bipyridyl)ruthenium(II) as a photocatalyst [94]. In this kind of reactions, the photocatalyst, instead of the viologen precursor, acted as a photosensitizer because absorption range of the viologen precursor located in the UV region. Recently, the He group proposed a very innovative method to reduce the reaction components for H2 generation [95]. They used chalcogen (E)-bridged viologens as both a photosensitizer and an electron mediator for their good absorption ability of visible light (Figure 10.25). Electron absorption spectrum of E-BnV2+ (Figure 10.25a) is strongly related to the type of E. It showed a clear red shift with E changing from S to Se and further to Te, resulting in strong absorption of visible light (Figure 10.25b). A hydrogen evolution system was built with E-BnV2+ both a photosensitizer and an electron mediator, EDTA as a sacrificial electron donor, and colloidal platinum E

+ N

Bn

+ N

Bn

2 OTF E = S (5a), Se (5b), Te (5c)

(c) EDTA

hv

E Bn N

e

5b/5c E = Se, Te E

EDTAox

H2

Bn N

N Bn

3 2

(d) 14

(e) 14

12

12

1

10

2.0 2.2 2.4 2.6 2.8 3.0 3.2 hv (eV)

2

5b 5c

8 6

2

4

0

H (μmol)

6

4

0 650

6

2 0

2

4

6 8 Times (h)

10

12

Bnv 5a 5b 5c

13.03

8

0 350 400 450 500 550 600 Wavelength (nm)

0h

10 1 min

4

2

H 2O

PVP–PT

H (μmol) 2

2 –1

5a 5b 5c

2

8

(ahv) / (nm -eV)

ɛ (103-l/mol/cm)

0

H2 H2

Pt

5bʹ/5cʹ Eg = 3.00 Eg = 2.78 Eg = 2.33

e– e –

–

N Bn

e– e –

(a)

e–

568

0

2h

4.73

7h 0 Bnv

0.12 5a 5b Active components

5c

Figure 10.25 (a) Molecular structures for E-BnV2+ . (b) Absorption spectra for 5a/5b/5c. (c) Photoinduced hydrogen production from water. (d) Time-dependent hydrogen generation from aqueous solution under a xenon lamp with 5b/5c as a photosensitizer. (e) Total hydrogen generation of active components; changed colors as the time go by under xenon lamp are shown as insets. Source: Edited with permission Li et al. [95]. Copyright 2017, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

10.3 Photochromic Materials–Viologen Compounds and their Analogs

particles stabilized by polyvinylpyrrolidone (colloids of PVP–Pt) as a catalyst. As shown in Figure 10.25c, the E-BnV2+ group was firstly excited to excited states (E-BnV2+* ) under visible light, which accepted one electron from EDTA and changed to radical cation E-BnV+⋅ . Then H+ was reduced to hydrogen on the surface of PVP–Pt and E-BnV+⋅ was oxidized to E-BnV2+ to finish the catalytic cycle. The Se-BnV2+ group was more effective than other two Se-BnV2+ groups for H2 evolution. This phenomenon was assigned to its lower HOMO level, which implies a stronger electron-accepting ability at the excitation state. 10.3.2.4 Electrical Applications of Viologens and their Analogs

Viologens and their analogs are strong electron acceptors. This electron-accepting ability may be applied to modify carrier concentration of electrical materials. In addition, strong radical· · ·π and cation· · ·π interactions between viologens and their analogs also favor the enhancement of carrier mobility. If one electrical material contains photo/thermo-active viologens or their analogs, the material will show modifiable electrical properties upon stimulation of photo or heat. This physical method differs from the traditional chemical methods, such as doping, surface functionalization, and defect engineering, which need to change compositions or structures. Several research groups have achieved great progress on design and synthesis of semiconductors with photo/thermo-switchable electrical properties based on viologens or their analogs. Leblanc et al. reported a photochromic inorganic–organic hybrid compound (MV)4 [Bi6 Cl26 ] (Figure 10.26) [96]. After UV irradiation, the color of this compound changed from white to brown. The single-crystal conductivity before and after UV irradiation was measured by a two-point configuration, hypothesizing that the current flowed in the full thickness of the sample. The relationship of linear natural logarithm of conductivity (𝜎) versus 1/T (Arrhenius law) 180–300 K range showed that the calculated activation energies (Ea ) decreased from 0.50 to 0.17 eV after coloration. After coloration, the 𝜎 value reduced at RT, but increased at lower temperature. The above work did not clarify the mechanism for conductivity change. However, it indicated that trapping carriers by discrete viologen cations is an effective approach to realize dropping of conductivity. If this phenomenon can be triggered

C110 C12 N1

N4

N2

a

C18 c b

Conductivity (S/m)

–7.5

–8.5

After UV irradiation

–9 –9.5 –10

C11 C110

(a)

Before UV irradiation

–8

(b)

–10.5 3.3

3.8

4.3

4.8

5.3

1000/T (K–1)

Figure 10.26 For (MV)4 [Bi6 Cl26 ]: (a) crystal structure; (b) conductivity of the same single crystal before and after UV irradiation, as a function of the inverse of the temperature. Source: Edited with permission Leblanc et al. [96]. Copyright 2010, American Chemical Society.

569

10 Relationship Between Structure and Electroluminescent, Photochromic –4 –5 log(𝜎 (S/cm))

570

3a Ea = 0.869 eV

–6 –7

Ea = 1.061 eV

–8 84.3% –9 2.4

(a)

2.6

2.8 3.0 1000/T (K−1)

3.2

3.4

(b)

Figure 10.27 For {(MV)2 [Pb7 Br18 ]}n : (a) crystal structure; (b) conductivity and color change of a single crystal after UV irradiation. Source: Edited with permission Sun et al. [97]. Copyright 2017, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

at high temperature, then the corresponding semiconductor may be used an over-temperature color indicator or a circuit-overload protector. Sun et al. have prepared a viologen-modified haloplumbate semiconductor {(MV)2 [Pb7 Br18 ]}n with a novel three-dimensional (3D) inorganic open framework by an in situ synthetic method and succeeded in observing an electron-transfer thermochromic phenomenon and thermoswitching of conductance (Figure 10.27) [97]. The color of the compound changed from yellow (3a) to brown (3b) upon annealing at 220 ∘ C. After annealing at 220 ∘ C for two hours and cooling to 300 K, the 𝜎 value fell by 84.3% from 1.0 × 10−8 to 1.7 × 10−9 S/cm for a single crystal sample by the two probe method using gold paste. They found that the localized electron state of spatially bound MV2+ only contributes hardly to the electron transport. After thermally induced electron transfer from Br to MV2+ cation ET, a hole in the inorganic skeleton is tied around MV⋅+ cation radical by strong Coulomb interaction to form a Frenkel exciton, which hardly contributes to the conductivity of 3B. This results in a significant reduction in the number of carriers and the conductivity of semiconductors. Roy et al. found that infinite π-stacking of viologen cations could yield semiconductors. They synthesized a viologen-based semiconductor through acidization of a bipyridyl molecule by an aromatic acid (Figure 10.28) [55]. This compound underwent photoinduced electron transfer from carboxylate to the protonated bipyridyl with color changing from yellow to green. The current–voltage (I–V) characteristics of specially prepared microcrystalline powders of the yellow and green samples were measured using conducting probe atomic force microscopy (CP-AFM), by uniformly dispersing the samples on thermally evaporated Au-coated SiO2 /p-Si substrates. The conductivity (𝜎) values for the yellow and green samples were calculated to be 0.10 ± 0.05 S/cm and 1.55 ± 0.20 S/cm, respectively. From these values, it can be found that, after coloration, conductivity of the sample increased nearly 16 times. Broad absorption, long-lived photogenerated carriers, high conductance, and high stability are all required for a light absorber toward its real application on solar cells. Inorganic–organic hybrid lead-halide materials have shown tremendous potential for applications in solar cells. Sun et al. offered a new design strategy

10.3 Photochromic Materials–Viologen Compounds and their Analogs

400 COOH

HOOC

COOH

O

N H41

300 N

BPCH

Current (nA)

HOOC

200

2a

2b

100

0

–100 –1.0

–0.5

0.0

0.5

1.0

Surface bias (V)

(a)

(b)

Figure 10.28 (a) Structures for a viologen compound and its precursors. (b) I–V curves and color change (inset) before and after irradiation. Source: Edited with permission Roy et al. [55]. Copyright 2012, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim. –12

In(𝜎) (S/cm)

–14 –16 Ea = 0.29 eV

–18 –20

Ea = 0.58 eV

–22

4A 4B

–24 –26

b c

(a)

(b)

36

37

38

39 39 41 1/kB T (eV–1)

42

43

Figure 10.29 For [Pb3 Cl6 (CV)]⋅H2 O]}n : (a) crystal structure; (b) conductivity change of a single crystal after UV irradiation. Source: Edited with permission Sun et al. [98]. Copyright 2018, American Chemical Society.

to improve the absorption range, conductance, photoconductance, and stability of these materials. They synthesized a new photochromic lead-chloride semiconductor, {[Pb3 Cl6 (CV)]⋅H2 O]}n (Figure 10.29), by incorporating a photoactive viologen zwitterion into a lead-chloride system in the coordinating mode [98]. This semiconductor has a novel inorganic–organic hybrid structure, where one dimensional (1D) semiconducting inorganic lead-chloride nanoribbons covalently bond to 1D semi-conducting organic π -aggregates. It shows high stability against light, heat, and moisture. After photoinduced electron transfer, it yielded a long-lived charge-separated state with a broad absorption band covering the 200–900 nm region while increasing its conductance and photoconductance. This work is the first to modify the photoconductance of semiconductors by photoinduced electron transfer. The observed increasing times of conductivity reached 3 orders of magnitude, which represents a record for photoswitchable semiconductors. The increasing photocurrent came mainly from the semiconducting organic

571

10 Relationship Between Structure and Electroluminescent, Photochromic 30

Photoinduced coloration Heating up Heat-induced decoloration Cooling down

5b(373 K)

25

Cation-π 3.40 Å H5A...O1 H1B...O1 2.85 Å H4A...O3 1.69 Å 2.50 Å H2A...O3 2.25 Å H2B...O4 2.40 Å H3C...O2 H2A...O4 2.32 Å 2.85 Å

σ (nS/cm)

572

20 15 10 5b(298 K) 5a(298 K)

0 0

(a)

(b)

5a(373 K)

Irr. off

5

2

4

6 8 Time (h)

5a(298 K) 10 12

Figure 10.30 For {(MV)2 [Pb7 Br18 ]}n : (a) crystal structure and interactions; (b) conductivity change for a pellet sample under a constant bias of 5 V during a successive treating process: irradiation at 298 K, heat treatment without irradiation, cooling to 298 K without irradiation. The saturation of coloration was reached after two hours. Source: Edited with permission Sun et al. [99]. Copyright 2019, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

π-aggregates, which indicates a chance to improve the photocurrent by modifying the organic component. These findings contribute to the exploration of light absorbers for solar cells. Breaking the intrinsic rule of semiconductors that conductivity increases with increase of temperature and realizing a dramatic dropping of conductivity at high temperature may arouse new intriguing applications. Sun et al. reported a T-type photochromic organic compound, (H2 bipy)(Hox)2 (H2 bipy = 4,4′ -bipyridin-1,1′ -dium; ox = oxalate; Figure 10.30) [99]. This compound displayed a clear color change from colorless to purple after irradiation, but restored the original color around 100 ∘ C. The 𝜎 value increased from 7.26 × 10−11 to 1.86 × 10−9 S/cm after photoinduced coloration at room temperature and further increased to 25.76 nS/cm when the temperature reached 100 ∘ C. After annealing at 100 ∘ C for several minutes, the 𝜎 value fell by 80.9%. This photochromic semiconductor can also be used as an over-temperature color indicator or a circuit-overload protector. 10.3.2.5 Molecular Recognition Applications of Viologens and their Analogs

Molecular recognition is an important theme since the growing up of supramolecular chemistry. Following this theme, new concepts, such as molecular motors, and new functions, such as medicine delivers, have been made and found, respectively. A key challenge to construct molecular motors or medicine delivers is to control molecular interactions. Viologen cations and their radicals may interact with neighboring delocalized species because of strong radical· · ·π and cation· · ·π interactions. This character has often been used to construct viologen-based molecular motors and medicine delivers.

10.3 Photochromic Materials–Viologen Compounds and their Analogs

10.3.2.5.1 Molecular Motors

We need to be able to switch the interactions between different parts of the molecule “ON” and “OFF,” to make the motor move to capture or release pharmaceutical molecule. Radical· · ·π interactions are stronger than cation· · ·π interactions [98], and thus changing the viologen component between cation radical and cation is an applicable to switch interactions. Stoddart and coworkers reported recently a tristable [2]rotaxane composed of a cyclobis(paraquat-p-phenylene) ring and a dumbbell with tetrathiafulvalene, dioxynaphthalene, and bipyridinium recognition sites (Figure 10.31) [100]. In this molecular motor, the position of the ring can be switched. On oxidation, it moves from the tetrathiafulvalene to the dioxynaphthalene, and on reduction, to the bipyridinium radical cation, provided the ring is also reduced simultaneously to the diradical dication. 10.3.2.5.2 Medicine Delivers

Based on the strong cation· · ·π interactions, Samanta et al. designed a supramolecule that may deliver doxorubicin (DOX) prodrug to cancer cells. The medicine deliver was assembled by capping a Fujita-type metal–organic polyhedron (MOP) with cucurbit[8]urils through molecular recognition (Figure 10.32) [101]. When a DOX prodrug was bonded to cucurbit[8]urils in the medicine deliver, the yielded MOP Stopper

Stopper

Ring N

O

O

O

O

O

N N N

Station 1 (dioxynaphthalene) Ring goes here when TTF is oxidized

N

N N N

N

N S

O

O

S

O

S N

N

Station 2 (viologen) Ring goes here when viologen is reduced

O

O

S

Station 3 (tetrathiafulvalene) Ring goes here unless TTF is oxidized or viologen is reduced

In D2O

X DO

X DO

+

CB8

N

DO X

DOX

DOX

Figure 10.31 Structure of an electrochemically driven tri-stable rotaxane. Source: Based on Trabolsi et al. [100]. Copyright 2010, Macmillan Publishers Limited.

O HN

–

2NO3

2

DO X

1

N

DOX

DOX X DO

HO OH

DO X

DO X

72+

[Pd12(1.CB8.2)24]

N

OH O

X

O

DOX

[Pd12124]72+

N

O Dox 2

HO + NH3

OH O

O

X=O H N X=N O

O 4

Figure 10.32 Self-assembly of a DOX-bearing metal–organic polyhedron. Source: Samanta et al. [101]. Copyright 2016, American Chemical Society.

573

574

10 Relationship Between Structure and Electroluminescent, Photochromic

is 10-fold more cytotoxic toward HeLa cells than equimolar quantities of DOX prodrug. The enhanced cytotoxicity can be traced to a combination of enhanced cellular uptake of the MOP and DOX release. This work provided a potent new platform for drug delivery application.

10.3.3 Prospective A large progress has been achieved since the discovery of viologens [102]. However, photochromic performance and mechanism of viologen compounds have not been well understood. Some issues remained are listed as below: (i) detailed photophysical and photochemical processes have not been well revealed. Yoon et al. [103] and Santos et al. [104] have analyzed dynamic processes of viologen-included zeolite and viologen borate by time-resolved spectroscopy techniques, respectively. Even so, some questions, such as which component (electron donor or acceptor) is firstly excited, what excited states are included, and how does charge separation and other processes compete, have not been answered. (ii) Reduction of viologens and formation of viologen radicals are clearly established, but the form of electron donors are still unclear after donating one electron [105]. For example, in MVCl2 , the Cl− ion was regarded as an electron donor. So, what is its form after donating one electron to MV2+ ? Supposing the formation of the Cl⋅ radical, it is highly reactive and should be unstable under the atmosphere. If Cl2 is formed through diffusion and combination of Cl⋅ radicals, then how to explain good cyclability observed in some photochromic viologen compounds? (iii) Knowledge on decoloration processes is scarce. Many reported viologen compounds require O2 to realize decoloration, but the role of O2 is unclear. Is it just an electron relay, or an oxidant? If it acts as an oxidant, then the “reversibility” condition required for the definition of “photochromism” is not fulfilled because side reactions will happen. In addition, for many crystal samples, interior crystal lattices have no chance to contact with O2 . In this case, how to understand the decoloration process? Elucidating these questions is highly important for improving fatigue resistance of viologen compounds and extending application range of viologen compounds to closed devices without O2 . Besides these questions, it is still a great challenge to realize photobleaching of the radical product. Photobleaching is common for other types of photochromic species, but has not been realized for viologen compounds. (iv) It is still difficult to control the competition between light and heat to trigger the forward electron-transfer process. For some viologen compounds, heat may induce electron transfer and form the same radical product as light does [106]. Response ability to multiple stimuli is important for multi-functional materials, but also brings some problems. For example, photoreactions may be disturbed by ambient heat sources, and, in some cases, the heat-induced product cannot be bleached through heat [107]. It is necessary to reveal the key structural factor that judge the selective occurrence of light or heat-induced process. To address the above four issues, thorough structural chemistry studies coupled with time-resolved spectroscopy studies and theoretical calculations are still highly desirable in the future.

10.4 NLO Materials

10.4 NLO Materials NLO materials, as the core devices of solid-state laser systems, can efficiently expand frequency range of common laser sources, which play a significant role in the field of laser-related science and technology, such as semiconductor manufacturing, photolithography, optical storage, and high-capacity communication networks. To be optically applicable, a NLO material must satisfy several fundamental but rigorous requirements on the structure-directing optical properties: wide transparent window, large SHG response, sufficient birefringence to achieve phase matching. To explore NLO materials which meet the above requirements, anionic group theory was proposed by Chen [108]. The main points of this theory include (i) anionic groups have significant contribution to the SHG effect; (ii) the total SHG coefficient can be calculated by summing up the second-order polarizabilities of the constituent anionic groups; and (iii) cationic contribution to the SHG coefficients can be neglected. This theory played a significant role in the discovery of many NLO materials, such as the notable NLO crystals discovered by our institute: KBe2 BO3 F2 (KBBF) [109], β-BaB2 O4 (BBO) [110], and LiB3 O5 (LBO) [111]. In this section, we will review the recent developments and underlying structure–property relationships in NLO materials.

10.4.1 KBBF Family Traditionally, the search for deep-UV (wavelengths below 200 nm) NLO materials mainly focused on borate systems owing to their deep-UV transparency. Till now, KBe2 BO3 F2 (KBBF) [112] is the sole material that can practically generate deep-UV coherent light by direct SHG process. KBBF and its analogous molecular formula of ABe2 (BO3 )F2 (A = Na, Rb, Cs, Tl) [109, 113] are named as KBBF family and listed in Table 10.4. ABe2 (BO3 )F2 (A = K, Rb, Cs) crystallizes in the same rhombohedral space group of R32, belonging to the uniaxial class. Although NaBe2 (BO3 )F2 has a different space group, it has a similar layered structure as those of ABe2 (BO3 )F2 (A = K, Rb, Cs). Their structures feature [Be2 BO3 F2 ]∞ layers, which extend infinitely in the ab plane Table 10.4

KBBF family.

Compounds

Space group

SHG efficiency (@1064 nm)

Absorption edge

NaBe2 (BO3 )F2

C2

155 nm

KBe2 (BO3 )F2

R32

a) d eff b) d 11

RbBe2 (BO3 )F2

R32

d11 = 0.45 ± 0.01 pm/V

CsBe2 (BO3 )F2

R32

d11 = 0.5 pm/V

151 nm

TlBe2 (BO3 )F2

R32

T c , the real part of low-frequency dielectric permittivity obeys the Curie–Weiss law, which suggests the potential type of phase transition. 11.1.1.4.3 P–E Hysteresis Loop

Under external electric field, the relationship between Ps and the electric field E is nonlinear and forms the P–E hysteresis loop (Figure 11.4). If external field is weak, the polarization P has a linear relationship with the change of E. When the external field E is enhanced, the P–E curve might become nonlinear. After point 3, the P tends to be saturated (point 4) under higher external field. If E reduces to zero (point 5), the polarization is not zero and the value is called residual polarization Pr . In order to remove the residual polarization of ferroelectrics, the reverse electric field is needed. When the reverse electric field reaches Ec (point 7), ferroelectric polarization reduces to zero. The Ec is called coercive field. With reverse electric

611

Table 11.3

Eighty-eight ferroelectric phase transitions with their number of states derived by Aizu.

Ferroelectric species

Number of states

Ferroelectric species

Number of states

Trigonal

3F1

3/2

Ferroelectric species

1F1

2

Monoclinic

2F1

2/2

3F1

6

23F2

6

mF1

2/2

3F3

2

23F3

8

2/mF1

4

32F1

6/2

m3F1

24

2/mF2

2

32F2

3/2

m3Fm

12

2/mFm

2

32F3

2

m3Fmm2

6

222F1

4/2

3mF1

6/2

m3F3

8

222F2

2

3mFm

3/2

432F1

24/2 12

Orthorhombic

Tetragonal

Cubic

23F1

Number of states

T

12/2

mm2F1

4/2

3mF1

12

432F2(s)

mm2Fm

4/2

3mF2

6

432F4

6

mmmF1

8

3mFm

6

432F3

8

mmmFm

4

3mF3m

2

43mF1

24/2

MmmFmm2

2

6F1

6/2

43mFm

12/2

4F1

4/2

6F1

6/2

43mFmm2

6

4F1

4/2

6Fm

3/2

43mF3m

4/2

4F2

2

6F3

2

m3mF1

48

4/mF1

8

6/mF1

12

m3mFm(s)

24

4/mFm

4

6/mFm

6

m3mFm(p)

24

Hexagonal

4/mF4

2

6/mF6

2

m3mFmm2(ps)

12

422F1

8/2

622F1

12/2

m3mF4mm

6

422F2(s)

4

622F2(s)

6

422F4

2

622F6

2

4mmF1

8/2

6mmF1

12/2

4mmFm

4/2

6mmFm

6/2

42mF1

8/2

6m2F1

12/2

42mF2(s)

4

6m2Fm(s)

6/2

42mFm

4/2

6m2Fm(p)

6/2

42mFmm2

2

6m2Fmm2

3/2

4/mmmF1

16

6m2F3m

2

4/mmmFm(s)

8

6/mmmF1

24

4/mmmFm(p)

8

6/mmmFm(s)

12

4/mmmFmm2(s)

4

6/mmmFm(p)

12

4/mmmF4mm

2

6/mmmFmm2(s)

6

6/mmmF6mm

2

m3mF3m 8 1. The number of states in which the denominator is 2 indicates that the ferroelectric can only be reoriented and cannot be reversed by 180∘ . 2. All ferroelectrics that can only be reoriented are ferroelastics.

614

11 Relationship Between Structure and Ferroelectric Properties

P

P

4 5

6

3

E 2

7 8

0

E

1

9 (a)

(b)

Figure 11.4 (a) The linear P–E curve of dielectric. (b) The nonlinear P–E hysteresis loop for ferroelectrics.

field increasing, the polarization is reversed and gradually reaches reverse saturation (point 9). With the constant change of electric field, the P–E hysteresis loop is obtained.

11.1.2 Symmetry Breaking of Ferroelectrics Symmetry breaking is an indispensable characteristic for most ferroelectrics (only few exceptions) [22, 23]. Using the Curie principle, we can analyze the symmetry relationship. Symmetry breaking involves the disappearance of symmetry elements during the phase transition from high-temperature PEP to low-temperature FEP. It is known that there are 32 point groups and 230 space groups to describe crystal symmetry, of which 20 asymmetric point groups enable piezoelectric effect. Further, only 10 polar point groups endow electric polarization (Table 11.4): 1(C1 ), 2(C2 ), m(Cs ), mm2(C2v ), 4(C4 ), 4mm(C4v ), 3(C3 ), 3m(C3v ), 6(C6 ), and 6mm(C6v ). Pyroelectric effect only exists in the 10 polar point groups, and ferroelectrics belong to a subgroup of pyroelectrics. From the viewpoint of Curie principle, crystal symmetry of low-temperature FEP should be a maximal subgroup of the high-temperature PEP. Aizu has summarized the possible 88 species of PEP-to-FEP transitions based on the Curie symmetry principle (Table 11.4). Currently, a variety of methods have been developed to obtain the information of crystal structures and symmetry breaking, such as single-crystal X-ray diffraction, neutron diffraction, and solid-state nuclear magnetic resonance (NMR). For instance, the hidden pseudo-symmetry can be detected by temperature-dependent second harmonic generation (SHG) and pyroelectric measurements, which act as strong evidence of non-centrosymmetric and polar phases.

11.1.3 Classification of Ferroelectrics The common classification of ferroelectric materials is based on the nature of structural changes during phase transitions, which correspond two types including the displacive and order–disorder type. Substantial inorganic oxide ferroelectrics such as

11.1 Concepts and Fundamentals

Table 11.4

Sixty-eight ferroelectric space groups belonging to the 10 polar point groups.

Point group

Space group

C1

P1

C2

P2, P21 , C2

C1h

Pm, Pc, Cm, Cc

C2v

Pmm2, Pmc21 , Pcc2, Pma2, Pca21 , Pnc2, Pmn21 , Pba2, Pna21 , Pnn2, Cmm2, Cmc21 , Ccc2, Amm2, Abm2, Ama2, Aba2, Fmm2, Fdd2, Imm2, Iba2, Ima2

C4

P4, P41 , P42, P43 , I4, I41

C4v

P4mm, P4bm, P42 cm, P42 nm, P4cc, P4nc, P42 mc, P42 bc, I4mm, I4cm, I41 md, I41 cd

C3

P3, P31 , P32 , R3

C3v

P3m1, P31m, P3c1, P31c, R3m, R3c

C6

P6, P61 , P65 , P62 , P64 , P63

C6v

P6mm, P6cc, P63 cm, P63 mc

BaTiO3 belong to the displacive type, in which the relative displacement of the ions creates electric polarization. In contrast, NaNO2 is a representative ferroelectric of order–disorder type, in which the reorientation of the dipolar NO2 − ions generates ferroelectricity. Actually, these two types of ferroelectricity origin are not mutually exclusive. Many ferroelectric materials often demonstrate both the displacive and order–disorder characteristics. Based on the magnitude of Curie constant (C), the types of ferroelectric compounds can be determined qualitatively. For instance, the majority of ferroelectric materials that show C value of ∼105 K belong to displacive type. Ferroelectric materials with the C value of ∼103 K exhibit the characteristic of order–disorder type. Moreover, the ferroelectrics with the C value of ∼10 K are called improper or extrinsic ferroelectrics in which the FEP is caused by some physical quantities instead of electric polarization. Another classification of ferroelectric materials is based on crystal symmetry, particularly on account of the group-to-subgroup relationship determined by Curie principle [24]. During phase transition from high-symmetry PEP to the low-symmetry FEP, symmetry breaking happens with losing some symmetry elements and hence Ps emerges. According to the Curie principle, the ferroelectric point group should be a subgroup of both paraelectric point group and symmetry group of Ps . That is, the paraelectric point group normally has several isomorphic subgroups, of which the polar direction corresponds to the Ps direction. The maximal one comprising the most symmetry elements along the Ps direction is the ferroelectric point group. Given a known point group at PEP, the number of polarization directions at FEP phase can be determined by n = N p /N f , where N p and N f are the respective orders of paraelectric and ferroelectric point groups, known as the sum of symmetry elements. As a result, when n = 2, the ferroelectrics are uniaxial

615

616

11 Relationship Between Structure and Ferroelectric Properties

with only two opposite Ps directions, and certainly multiaxial ferroelectrics are characterized by more than two Ps directions. For instance, the paraelectric space group and ferroelectric space group of LiNbO3 are R3c and R3c, respectively, their lattice symmetries are so close leading to the uniaxial feature. BaTiO3 undergoes successive phase transitions from the cubic Pm3m (N p = 48) to tetragonal P4mm (N f = 8) at 393 K, to orthorhombic Amm2 (N f = 4) at 278 K, and further to R3m (N f = 6) at 183 K. These complex phase transitions lead to the multiaxial ferroelectric merits for BTO, accompanying with 6, 12, and 8 equiv polarization directions at three FEPs, respectively. Thus, BaTiO3 is a typical multiaxial ferroelectric material.

11.1.4 Characterization of Ferroelectrics 11.1.4.1 Dielectric and Dielectric Switch

To deeply understand the switching process, we can rely on measuring other physical properties, such as dielectric, optical, and mechanical activities. It is known that dielectric substance has polarization-constrained charge that responds to external stimuli (e.g. electric field, temperature, pressure, and light). The degree of polarization under external electric field is described by the real part (𝜀′ ) of the complex dielectric permittivity. Generally, microscopic polarization has three intrinsic origins: dipole orientation, ionic displacement, and electron displacement. The dipole orientation plays an important role in the 𝜀′ value in molecular material. For phase transition material, the step-like dielectric anomalies occurring near the T tr suggest switchable dielectric characteristics (Figure 11.5). Due to the dipole reorientation between the ordered low-temperature phase (LTP) and disordered high-temperature phase (HTP), the polar component often affords driving force to the temperature-driven dielectric response. It is worth noting that most structural phase transitions are accompanied by dielectric changes near the T tr . 𝜀′ (𝜔) is a function of electric field frequency (𝜔), and the value of 𝜀′ decreases as 𝜔 increases. Thus, the dielectric dispersion behavior can respond to an applied electric field and reflect the dipole dynamics. Dielectric constant is a second-order tensor, and the correlation between crystal axis and dielectric component relates to the local motion of dipoles in the crystal lattice. For instance, the high-dielectric state corresponds to disordered phase, in which the molecules often undergo rotational or flipping motion. In terms of bulk properties, the spin transition exhibits a change in magnetic susceptibility between two states, while dielectric transition reflects the ɛʹ

Figure 11.5 Schematic illustration of a thermal-driven dielectric transition.

High state (on)

Low state (off) T

11.1 Concepts and Fundamentals

Spin dipole

Electric dipole

eg

High state

t2g

Motional

Un pairing

T Frozen

Pairing eg Low state

Dielectric transition

Switching

t2g

Spin transition

Figure 11.6 Analogies between dielectric transition and spin transition (taking Fe(II) for example). Note: The antiparallel alignment of permanent dipoles in the frozen state of dielectric transition does not mean antiferroelectricity.

change of dielectric constant (Figure 11.6). Below T tr , the electric dipoles are frozen while magnetic dipoles disappear or have smaller values due to spin pairing. Hence, dielectric responses resemble that of electric counterpart to spin transition, like the pair of ferroelectricity and ferromagnetism. 11.1.4.2 NLO and NLO Switch

Nonlinear optical (NLO) phenomena were reported in the 1960s and many NLO-related properties, such as SHG, stimulated Raman scattering, and stimulated Brillouin scattering, were also discovered at this time [25–30]. Nonlinear optics is one of the most important branches of optics that describes the behavior of laser light in media where the dielectric polarization responds nonlinearly to electric field component of light. For the molecule, if a resonant electromagnetic field is applied, a transient displacement of the electron density might occur, namely the polarization phenomenon. The polarization of light field (E) generated in the medium closely relates to P with the following relationship: P = 𝜀0 𝜒E

(11.13)

𝜒 is the polarizability, and 𝜀0 is the vacuum dielectric constant. However, when the incident light is laser, the polarization response exhibits the nonlinear characteristics. Therefore, the relationship between polarization and high-power laser is described as: P = 𝜀0 𝜒 (1) E + 𝜀0 𝜒 (2) EE + 𝜀0 𝜒 (3) EEE + · · · 𝜒 (2)

𝜒 (3)

(11.14)

and are the second-order and third-order nonlinear polarizabilities, where respectively. In general, 𝜒 (2) is a function of the frequency of the applied light wave and the frequency of polarization. When optical dispersion is neglected, it is almost independent of the frequency and becomes constant, called the nonlinear coefficient. It is noteworthy that the second-order polarization will result in typical second-order nonlinear effects, such as SHG (optical frequency conversion from 𝜔 to the double frequency 2𝜔). Parameter 𝜒 (2) represents the high nonlinear

617

618

11 Relationship Between Structure and Ferroelectric Properties

susceptibility of materials on the macroscopic scale, and 𝛽 denotes the quadratic molecular hyperpolarizability; both are responsible for NLO properties. In fact, extending molecular NLO properties to the material’s scale remains a challenge for the following reasons: SHG efficiency of a material is closely related to the material’s bulk quadratic susceptibility (𝜒 (2) ). However, the molecule arranged in a centrosymmetric crystal structure shows no 𝜒 (2) , even if the molecule has the high 𝛽 value. For instance, the majority of photochromic molecules that crystallize in the centrosymmetric structure eliminate SHG properties in the solid crystal state. With the advance of NLO field, the application prospects of NLO materials are booming in the optical information technology, laser technology, material analysis, nanophotonic technology, etc. Among them, the NLO switches can be used as the basic elements for electronic and photonic devices, which have shown a wide range of potential applications in optical communication and optical storage. In addition to NLO materials, researchers have paid substantial enthusiasm to NLO switching materials. In contrast to high-performance NLO-switching properties achieved in solution, the current results demonstrate that NLO switching can also be obtained in bulk solid-state materials. For instance, photoswitching suggests that the molecule has different structures and physical properties in different conditions, corresponding to two/more states named “on” and “off,” respectively. These different states can be reversibly switched under the external stimuli, fulfilling the purpose of switching control in device. Within this portfolio, a large number of NLO-switching materials with large 𝜒 (2) values have been synthesized in the past few decades. As mentioned in the review by Coe [31], most NLO-switching molecules adopt D–π–A structures with polarization (D is electron donor group, π is conjugated bridge, and A is electron acceptor group). The mechanisms for NLO-switching process can be divided into redox action and photochromism [32], and the azobenzene and stilbene groups are the light-inducing units [33, 34]. Some previous reviews have summarized their structures, NLO-switching properties, and perspectives. Nevertheless, the efficient and reversible NLO-switching properties in the solid state remain a challenge, particularly for the second-order NLO response because the macroscopic non-centrosymmetric alignment of NLO moieties must be reversibly controlled. Ferroelectric materials have developed as one of the most important branches for the condensed matter physics and have also been used as basic elements for diverse applications. Ferroelectric is characterized by the spontaneous polarization (Ps ) in a certain temperature range, of which the direction can be switched by the reversal of external electric field (E) direction. Structurally, ferroelectrics change from the high -temperature PEP (high symmetry) to the low-temperature FEP (low symmetry), called PEP-to-FEP transition. Symmetry breaking related to subtle structure changes occurs during the phase transition, which leads to dramatic variations of physical properties. As an effective method, SHG has been widely used to verify symmetry breaking in ferroelectrics, since all ferroelectric polar phases are SHG-active and their temperature-dependent SHG signals can solidly confirm the occurrence of symmetry breaking. This feature makes it possible to achieve the NLO-switching candidates in the family of ferroelectrics.

11.1 Concepts and Fundamentals

11.1.4.3 Pyroelectric Properties

The pyroelectric effect has been discovered by Theophrastus for a long-history time, which refers to the release of charge generated in the material due to the change of temperature. The requirement of the polar point symmetry suggests that materials with 10 polar point groups may have pyroelectric effect. Obviously, all ferroelectric materials are pyroelectric, but only pyroelectric materials that can switch polarization by external field belong to the class of ferroelectrics. As shown in Figure 11.7, the asymmetric environment experienced by the electrically charged species produces Ps within the crystal structure of the material [35]. The asymmetry of the potential energy of the cation along the x1 –x2 line is shown in Figure 11.8. Figure 11.7 Schematic diagram for the two-dimensional electrically polar lattice.

–

–

–

+

+

–

–

+

X1

0

–

+

X2

–

–

–

−

Ps

Potential energy

Figure 11.8 Potential energy of cation in lattice of Figure 11.7 along the line xl –x2 , E 1 and E n represent the quantized energy levels for the cation, and the locus A–B is the change in its equilibrium position with the change of energy.

B En E1 x1

A Distance

0 x2

619

620

11 Relationship Between Structure and Ferroelectric Properties

With the lattice temperature increasing, the quantized energy level inside the wells will change from E1 to En . The average equilibrium position in the crystal lattice is also changed, thereby affecting the total electric dipole moments. This macroscopic performance is called the pyroelectric effect. As an important figure of merit to quantitatively describe pyroelectric effect, the pyroelectric coefficient (p) can be given by the change rate of Ps : (11.15)

ΔPs = pΔT

which is dependent on the temperature (T). For a thin crystal piece of pyroelectric material that has been polarized, as shown in Figure 11.9, the charges can be detected by measuring the current through an external circuit: iP = Ap′ dT∕dt

(11.16)

dT/dt is the rate of change of the material temperature. The pyroelectric device can interact with the input energy to produce the temperature change. The energy can be either electromagnetic radiation absorbed by the material or heat generated by the reaction of the substance on the surface of the component. For most cases, a large pyroelectric effect is observed in ferroelectric materials, wherein the direction of the dipole moments can be switched by applying an electric field. In ferroelectrics, the electric polarization usually decreases with the increasing temperature but exhibits dramatic change in the vicinity of T c . Consequently, the pyroelectric coefficient is highly dependent on the temperature in ferroelectrics. For polar materials, the relationship between the electric bias field and dielectric displacement can be expressed by: D = 𝜀0 E + P + Ps = 𝜀E + Ps

(11.17)

The following expression can be obtained from the derivation: p′i = Ej

𝜕𝜀ij 𝜕T

(11.18)

+ pi External circuit Energy input

ip

Pyroelectrical material P′

Figure 11.9

Electroded surfaces

The electrode of pyroelectric element for measuring pyroelectric current.

11.1 Concepts and Fundamentals

It is clear that even a non-pyroelectric material can also enable the pyroelectric-like response, if polarized by the bias electric field. From a viewpoint of symmetry, the applied electric bias field makes a contribution to the polar symmetry of crystals. In ferroelectric materials, the electrical bias field can increase the transition temperature but lower the dielectric constant and flatten the dielectric peak. Besides, in the vicinity of T c , the electric polarization will decrease sharply. The combination of these effects accounts for the maximum pyroelectric responsiveness. A considerable number of ferroelectric materials have been used as pyroelectric candidates, including TGS, lithium niobate (LiNbO3 ), and their doped derivatives. In addition to the crystal form, the advantage for these ferroelectrics is that they can also be used in the polycrystalline form. The electric bias field can control the polar axes of crystallites and further generate bulk spontaneous polarization. The polarization directions are related to the number of polar axes derived from the high-symmetry PEP. Pyroelectric materials have been widely used in the electromagnetic radiation sensors to date. Compared with the semiconductor counterparts, pyroelectrics would work over a wide range of wavelengths and the radiation modulation frequency covers a broad range from a few hertz to many gigahertz region. 11.1.4.4 Polarization Switching – Ferroelectricity

The spontaneous polarization reversal (or switching) by electric field is the most important characteristic of ferroelectrics. The most direct proof for this process is the observation of polarization versus electric field (P–E) hysteresis loops (Figure 11.10), which originate from the conversion of domain walls. Here we will focus on the relationship between P–E hysteresis loop and polarization switching. Generally, the P–E hysteresis loop can be observed by the Sawyer–Tower circuit [36]. The

50 PS

40 30 Polarization (µC/cm2)

D C

E PR

20 10

F

0

B A

– EC

–10

+ EC

–20 –30 –40

– PR

G

–50 –300

–200

–100

0

100

200

300

Electric field (kV/cm)

Figure 11.10 Ferroelectric P–E hysteresis loop represents the polarization state of the material at the indicated fields. The symbols are explained in the text.

621

622

11 Relationship Between Structure and Ferroelectric Properties

polarization (Pi ) induced by an applied electric field vector (Ei ) in the insulating polarized material (dielectric) is given by Pi = 𝜒ij Ej

(11.19)

When the AC electric field is small, the polarization increases linearly with the field amplitude, corresponding to the AB segment. With the field increasing, the polarization of the domains with an unfavorable polarization direction will begin to switch along the electric field direction, causing the measurable charge density to increase rapidly, corresponding to the BC segment. Meanwhile, the polarization exhibits strong nonlinear response. At the point C, all the domains are arranged in the same direction, and ferroelectricity is restored to linear (i.e. the segment CD). If the electric field strength decreases, some domains will be reversibly switched. However, the polarization is nonzero at the zero field, as shown by the point E, of which the polarization value is called the remanent polarization (Pr ). If the polarization restores to the zero state, an inversion electric field must be applied (the point F). The electric field required to make the polarization zero is called the coercive field (Ec ). With the inversion electric field further increasing, the same behavior of polarization and the saturation of dipole moments will be observed (the segment FG). Subsequently, the electric field strength continues to decrease to zero and reverses again, completing the entire cycle. Typically, the Ps size is the extrapolation of the linear segment CD to the intercept of polarization axis. However, it should be mentioned that the Ec is not an absolute threshold field based on the intercept of P–E hysteresis loops. If a low electric field is applied to material for a sufficiently long time, the polarization can also be switched. Ideally, the P–E hysteresis loop has the symmetrical shape, i.e. +Ec = −Ec and +Pr = −Pr . However, the Ec , the size of Ps and Pr , and the shape of hysteresis loop are related to many factors, such as the thickness of sample, the presence of charged defects, mechanical stress, heat treatment, etc. Although the mechanism of polarization switching in ferroelectrics has been studied, it seems to lack a general mechanism to explain the polarization reversal of all ferroelectrics. Generally, the electric switching of polarization closely involves with the formation of existing antiparallel domains, domain wall motion, and nucleation and growth of new antiparallel domains. The relationship between the typical applied voltage waveform of ferroelectrics and the switching current i is presented in Figure 11.11. In terms of i ∝ dP/dt, the field applied antiparallel to the polarization will switch polarization from the −Pr state to +Pr state. The total current can be divided into two parts: the first spike is caused by the linear response of dielectric, while the bell-shaped curve relates to polarization switching. The total area under the curve is equal to ts

∫0

i(t)dt = 𝜀0 𝜀EA + 2PR A

(11.20)

where A is the area of electrode. After the polarization is switched, the same pulse parallel to the polarization is applied again and a transient spike of dielectric response can be obtained. Therefore, the kinetics for polarization reversal can be described by the switching time (ts ) of the different amplitude E. ts usually refers to

11.1 Concepts and Fundamentals

(a)

(b)

(c)

(d)

Figure 11.11 Probable sequence of polarization switching in ferroelectrics: (a) nucleation of oppositely oriented domains, (b) growth of oppositely oriented domains, (c) sidewise motion of domain walls, and (d) coalescence of domains. Black and white areas have different orientation of polarization.

E

tpulse (a) i imax

Switching current Nonswitching current

0.1imax tmax (b)

ts

Figure 11.12 Switching and transient current during the polarization switching. (a) Electrical field and (b) the current as a function of time.

the value that reduces the current to 0.1imax (Figure 11.12). That is, ts = t∞ e𝛼∕E

(11.21)

𝛼 is called an activation field. At very high field, this relation changes to a power law of the form ts ∝ E−n , where n depends on the material. The maximum switching current can be described by imax = i∞ e−𝛼∕E

(11.22)

where 𝛼, n, t∞ , i∞ are the temperature-dependent constants, and the switching time decreases as it approaches the Curie point. Recently, it has been suggested that detailed information on the nucleation and growth of domains can be obtained from the switch current data based on the classical Kolmogorov–Avrami crystal theory.

623

624

11 Relationship Between Structure and Ferroelectric Properties

11.1.4.5 Domain Motions

The Ps in ferroelectric crystals is generally not uniformly aligned throughout the whole region along the same direction. At the FEP, the P–E hysteresis loop is a typical feature of ferroelectric crystals, which is actually a macroscopic reflection of the domain motion or electric polarization under an external field. The electric domains are the region of crystal where the Ps direction is uniformly oriented. The area between two domains is called the domain wall. If domain wall is separated by the regions of oppositely oriented polarization, it refers to 180∘ wall; the domain wall that is separated from regions with vertically polarization is called the 90∘ wall. Microscopically, ferroelectric is initially composed of equivalent number of positive and negative domain compositions, corresponding to no net polarization in the entire crystal. The low electric field applied in the forward direction cannot induce the reversal of Ps , which corresponds to the linear process without domain switching. When the electric field strength approaches Ec , most of the negative domains switch into the positive direction and the polarization increases rapidly. Under higher electric field, all the domains are aligned as a single-phase domain in the positive direction, reaching the saturation state. While the field strength is reversed to be zero, some domains still remain aligned in the positive direction and the polarization is not zero. When Ec is applied in the negative direction, the polarization is eliminated totally. The further increasing of electric field leads to the negative reversal of polarization, as depicted in Figure 11.10. As exemplified by PbTiO3 , it changes from the cubic PEP to the ferroelectric tetragonal phase at 490 ∘ C. As shown in Figure 11.13, the Ps emerges along the ct -axis direction of the tetragonal unit cell. There are 6 equiv directions along the three ap -axes of its cubic unit cell. With temperature cooling below its T c , the Ps can be generated with equal probabilities along any of them (Figure 11.14). Since the ct and at -axes directions are different in the tetragonal phase, the angle between polarization directions on both sides of 90∘ domain wall is slightly smaller than 90∘ .

Tetragonal ferroelectric phase Cubic paraelectric phase

Pb cT

aC O Ti PS

aC

aC

aT PS ≠ 0

PS = 0 (a)

Figure 11.13

aT

(b)

Perovskite structure for PbTiO3 at (a) PEP and (b) FEP.

11.1 Concepts and Fundamentals

Ps

– Ps

+ Ps Ps

– Ps (a)

+ Ps

Domain wall region (b)

Domain wall region

Figure 11.14 Illustration of (a) 180∘ and (b) 90∘ ferroelectric domains and domain wall regions in the tetragonal perovskite. The change of polarization across the domain wall is shown for a 180∘ wall (a), and tetragonal distortion in (b) is exaggerated. Figure 11.15 Formation of 90∘ and 180∘ ferroelectric domain walls in a tetragonal perovskite (e.g. PbTiO3 ). Due to the formation of 90∘ walls, the deformation inside the domain wall region is exaggerated for the sake of picture clarity.

Stress

Creation of 90° walls aT aT cT

cT

aC

cT

Ed

Ps

aT Ps

cT aT + – + –

aT + + + +

Cubic phase aC

Ps

PS

PS

cT

– + – + – – – – Creation of 180° walls

Ferroelectric domain is the region with the minimum electrostatic energy of depolarizing fields and the elastic energy associated with mechanical constraints, to which ferroelectric material is subjected during the PEP-to-FEP transition [37]. The onset of Ps results in the formation of surface charges at T c . The electric field generated by this surface charge is the depolarizing field Ed , which is oriented opposite to the Ps (Figure 11.15). The nonuniform distribution of Ps will form quite strong depolarizing field (on the order of MV/m), making the single domain state of ferroelectrics to be unfavorable in energy. However, the electrostatic energy associated with the depolarizing field may be minimized: (i) if the ferroelectric splits into domains with the oppositely oriented polarization (Figure 11.15) or (ii) if the depolarizing charge is compensated by electrical conduction through the crystal or by charges from material. Due to ferroelectric domains, the grown ferroelectric crystals usually exhibit the reduced or even zero thermoelectric and piezoelectric effects. Due to the influence of mechanical stress, the splitting of ferroelectric domain walls may also occur. It is assumed that a portion of PbTiO3 crystal is mechanically

625

626

11 Relationship Between Structure and Ferroelectric Properties

compressed in the (100) cubic direction, while being cooled by the phase transition temperature. In order to minimize the elastic energy, its ct -axis of the tetragonal cells will develop perpendicular to the stress. In the unstressed portion, the short at -axis is perpendicular to the stress and the polarization remains parallel to the stress direction. Therefore, both domain walls of 90∘ and 180∘ in PbTiO3 can reduce the influence of depolarizing field, while only the 90∘ wall can reduce the elastic energy. With the temperature decreasing below T c , the combination of electrical and elastic boundary conditions results in the complex domain structures. Ferroelectric materials possess the domain walls with different directions from that of the Ps vectors, and domain walls also exhibit different orientations from the spontaneous strain tensors, namely ferroelastic domain walls. In PbTiO3 , the 180∘ walls coincide with the ferroelectric attribute, while the 90∘ walls are ferroelectric and ferroelastic, of which the direction differs from the polarization vector and the spontaneous strain tensor. The type of domain walls in ferroelectrics depends on the symmetry of both non-FEP and FEP [38]. Fousek and Janovec have proposed the criteria that can be used to determine possible types of domain walls in ferroelectrics. Observation by the transmission electron microscope revealed that the domain walls in ferroelectric thin films are approximately 1–10 nm.

11.2 Recent Advance of Molecular Ferroelectrics 11.2.1 Organic Ferroelectrics Organic molecular ferroelectrics are an important branch of ferroelectric materials, including the low-molecular-mass organic compounds, the polymer ferroelectrics, organic salts, and vinylidene fluoride oligomers in thin-film form [39, 40]. From the viewpoint of chemical compositions, ferroelectric liquid crystals also fall into this category [41]. Although organic compounds occasionally crystallize in polar structures, their dielectric properties and potential ferroelectric properties are easily overlooked. In comparison with inorganic counterpart, ferroelectric properties are rarely found in the organic family. It is notable that the first ferroelectric of RS (potassium sodium tartrate tetrahydrate, KNaC4 H4 O6 ⋅4H2 O), discovered in 1920, also belongs to the organic salt, which opens the door for ferroelectric materials. Thiourea is a typical example of the low-molecular-mass organic ferroelectric, which exhibits the complex structural changes along with successive commensurate and incommensurate states of superstructures between its PEPs and FEPs [42, 43]. Subsequently, 2,2,6,6-tetramethyl-1-tube methoxy (TEMPO) was reported as the stable molecular ferroelectric with the room-temperature polarization [44]. The diacetylene monomer of 1,6-bis(2,4-dinitrophenoxy)-2,4-hexadiyne was also discovered as organic molecular ferroelectric [45–47], which can exist in the form of polymer and single crystal. The concept proposed by Choudhury and Chitra has been verified by some ferroelectric materials [48]. If conversion between different degeneration directions of ferroelectrics is to be achieved, the molecular symmetry needs to be higher than C1 and more than 30% of organic structures

626

11 Relationship Between Structure and Ferroelectric Properties

compressed in the (100) cubic direction, while being cooled by the phase transition temperature. In order to minimize the elastic energy, its ct -axis of the tetragonal cells will develop perpendicular to the stress. In the unstressed portion, the short at -axis is perpendicular to the stress and the polarization remains parallel to the stress direction. Therefore, both domain walls of 90∘ and 180∘ in PbTiO3 can reduce the influence of depolarizing field, while only the 90∘ wall can reduce the elastic energy. With the temperature decreasing below T c , the combination of electrical and elastic boundary conditions results in the complex domain structures. Ferroelectric materials possess the domain walls with different directions from that of the Ps vectors, and domain walls also exhibit different orientations from the spontaneous strain tensors, namely ferroelastic domain walls. In PbTiO3 , the 180∘ walls coincide with the ferroelectric attribute, while the 90∘ walls are ferroelectric and ferroelastic, of which the direction differs from the polarization vector and the spontaneous strain tensor. The type of domain walls in ferroelectrics depends on the symmetry of both non-FEP and FEP [38]. Fousek and Janovec have proposed the criteria that can be used to determine possible types of domain walls in ferroelectrics. Observation by the transmission electron microscope revealed that the domain walls in ferroelectric thin films are approximately 1–10 nm.

11.2 Recent Advance of Molecular Ferroelectrics 11.2.1 Organic Ferroelectrics Organic molecular ferroelectrics are an important branch of ferroelectric materials, including the low-molecular-mass organic compounds, the polymer ferroelectrics, organic salts, and vinylidene fluoride oligomers in thin-film form [39, 40]. From the viewpoint of chemical compositions, ferroelectric liquid crystals also fall into this category [41]. Although organic compounds occasionally crystallize in polar structures, their dielectric properties and potential ferroelectric properties are easily overlooked. In comparison with inorganic counterpart, ferroelectric properties are rarely found in the organic family. It is notable that the first ferroelectric of RS (potassium sodium tartrate tetrahydrate, KNaC4 H4 O6 ⋅4H2 O), discovered in 1920, also belongs to the organic salt, which opens the door for ferroelectric materials. Thiourea is a typical example of the low-molecular-mass organic ferroelectric, which exhibits the complex structural changes along with successive commensurate and incommensurate states of superstructures between its PEPs and FEPs [42, 43]. Subsequently, 2,2,6,6-tetramethyl-1-tube methoxy (TEMPO) was reported as the stable molecular ferroelectric with the room-temperature polarization [44]. The diacetylene monomer of 1,6-bis(2,4-dinitrophenoxy)-2,4-hexadiyne was also discovered as organic molecular ferroelectric [45–47], which can exist in the form of polymer and single crystal. The concept proposed by Choudhury and Chitra has been verified by some ferroelectric materials [48]. If conversion between different degeneration directions of ferroelectrics is to be achieved, the molecular symmetry needs to be higher than C1 and more than 30% of organic structures

11.2 Recent Advance of Molecular Ferroelectrics

meet this basic requirement. Depending on the Cambridge Structural Database, the discovery of hidden “pseudo-symmetry” is a potential pathway to exploit undisclosed ferroelectric candidates [49]. Hence, Zikmund and coworkers had proposed more than a dozen potential ferroelectric molecules using this method, and weak ferroelectricity was successfully verified in the organic molecular compound, cyclohexane-1,1′ -diacetic acid [50]. However, this experimental strategy still has the underlying limitations, since the interactions between the dipoles tend to cancel out the neighboring molecular dipoles in the crystalline state. It remains challenging to achieve the crystallization of non-centrosymmetric molecules into a polar structure with large dipole moments. Another hint is that ferroelectric materials should possess sufficiently low energy barrier for the reversal switching of molecular moieties in the crystalline solids. The known polymer ferroelectrics usually to have quite large coercive field. Table 11.5 lists the parameters of organic ferroelectrics and a comparison with other inorganic and inorganic–organic hybrid ferroelectrics. Except for thiourea, the majority of organic compounds have low dielectric constants and small Ps values. This is probably because their ferroelectricity is a side effect of an extrinsic origin, such as a ferroelastic or structural instability other than dipole–dipole interactions. A potential design strategy for exploring new organic ferroelectrics is to assemble the analogs to ferroelectric oxides by introducing ionic moieties, of which the atomic displacement would trigger ferroelectricity. Such ferroelectrics usually adopt unstable structures that spontaneously cause the lattice deformation into polar structural states. This requires a delicate balance between the electrostatic attraction and short-range counter-interaction of ionic moieties. In this sense, organic ferroelectrics can be achieved by introducing “multi-component” moieties to trigger electric polarization. In addition to dipole molecules and electrical displacement, dynamic protons on hydrogen bonds may also trigger ferroelectric order of the lattice. For example, in the KH2 PO4 (KDP) family, the collective site-to-site transfer of protons in the O—H· · ·O bond leads to the switching of electric polarization (Figure 11.16a). Such proton transfer-type ferroelectrics may simultaneously exhibit the characteristics of order–disorder and atomic displacement. For instance, the simultaneous displacive deformation of PO4 3− ions also contributes to Ps alongside the protonic order–disorder phenomena [51]. The protons of such ferroelectrics transfer within hydrogen bonds, which greatly reduces the steric difficulties of ferroelectrics, corresponding to small coercive field. Recently, ferroelectricity has been discovered in the organic tricyclohexylmethanol (TCHM) molecule [52], which is a hydrogen-bonded dimer capable of displaying the thermoelectric charge at low temperature. However, the origin of its ferroelectricity is the dipole reorientation that differs from the KDP-type ferroelectric. The O—H· · ·O hydrogen bonds connecting two TCHM molecules are simultaneously broken, while the two OH groups are redirected around the CO bonds [53, 54]. Organic squaric acid is a bipolar molecule carrying two protons (Figure 11.16b) [55]. The strong intermolecular hydrogen bonds form a sheet-like polar network, and the interlayer ordering counteracts the volume polarization. When the polarity of molecular layers is altered by proton transfer, the material

627

628

11 Relationship Between Structure and Ferroelectric Properties

Table 11.5 Properties of molecular ferroelectrics compared with those of typical ferroelectrics of polymeric, inorganic, and inorganic–organic hybrid compounds. 𝜺′

T c (K) Tc

T cD

𝜿 RT

𝜿 max

Thiourea

169

185

30

TEMPO

287

288

10

CDA

397

—

—

TCAA

355

—

4.5

Benzil

84

88

—

DNP

46

—

4.0

TCHM

104

—

VDF oligomer

—

—

Materials

C (×103 k)

P s (𝛍c/cm2 ); T

E c (kV/cm)

104

3.7

3.2; 120 K

0.2

16

—

0.5

—

25

—

—

—

6.5

0.0076

0.2; RT

4

2.7

—

3.6 × 10−3 ; 70 Ka)

—

22

0.026

0.24; 10 Ka)

—

9.6

100

2.7

6 × 10−3 ; 96 Ka)

—

6

—

—

13; RT

1200

Single component

CT complexes TTF–CA

81

84

40

500

5.7

—

—

TTF–BA

50

—

—

20

—

—

—

304↑

110

3 × 103

5.0

1.8; 160 Ka)

0.8

4.0

0.8; 105 Ka)

0.5

H-bonded supramolecules Phz–H2 ca

253

Phz–H2 ba

138

204↑

30

1.7 × 103

[H-55dmbp][Hia]

269

338↑

250

900

14

4.2a ; 110 Ka)

2

63.7

—

—

220

—

6 × 10−3 ; 25 Ka)

—

VDF0.65 –TrFE0.35

363

—

20

50

—

8; RT

500

Nylon-11

—

—

—

4

—

5; RT

600

437

—

—

4 × 103

4.7

10; 140 K

5

104

150

26; RT

10

Clathrate β-Quinol-methanol Polymers

Inorganic compounds NaNO2 BaTiO3

381

—

5 × 103

PbTiO3

763

—

210

9 × 103

410

75; RT

7

SbSI

293

—

—

6 × 104

233

20; 270 K

—

Organic–inorganic compounds KH2 PO4 (KDP)

123

213

30

2 × 104

2.9

5.0

0.1

HdabcoReO4

374

—

6

22

—

16; RTa

>30

TGS

323

333

45

2 × 103

3.2

3.8; 220 K

0.9

TSCC

127

—

5

80

Rochelle salt

297

308

—

4 × 103

—

0.27; 80 K

3

2.24

0.25; 276 K

0.2

𝛽-Quinol = hydroquinone; CDA = cyclohexan-1,1′ -diacetic acid; dabco = diazabicyclo[2.2.2]octane; DNP = 1,6-bis(2,4-dinitrophenoxy)-2,4-hexadiyne; TCAA = trichloroacetamide; TCHM = tricyclohexylmethanol; TEMPO = 2,2,6,6-tetramethyl-1-piperidinyloxy (tanane); TSCC = tris-sarcosine calcium chloride; VDF oligomer = CF3 (CH2 CF2 )17 I; VDF0.65 –TrFE0.35 = vinylidene fluoride-trifluoroethylene ((CH2 CF2 )0.65 (CHF-CF2 )0.35 )n ; T c D = transition point for deuterated compound; 𝜅 RT = room-temperature dielectric constant; 𝜅 max = maximum dielectric constant; C = paraelectric Curie constant. a) Pyroelectric charge measurement.

11.2 Recent Advance of Molecular Ferroelectrics BaTiO3

NaNO2

KH2PO4(KDP)

B O2–

O H P

A

Disorder Order N

P P

(a)

Dipolar ion/molecule Single component p

H+ transfer

Displacement CT complexes

P

P

Hydrogen-bonded compounds H O O H O

N N+ H

O

Thiourea

Polymer e– p

C F

(b)

PVDF

TTF–CA

M2P–DMTCMQ

Squaric acid

Hdabco+

Figure 11.16 Conventional types of ferroelectrics and the origin of their dipole moment or polarization P (open arrows). (a) Typical examples of inorganic ferroelectrics. (b) Organic ferroelectric or antiferroelectric substances.

has an in-plane dielectric constant of ∼300 above the ordered–disorder phase transition temperature (373 K) [56]. The calculation reveals that it has the large in-plane polarization [57]. Structurally, squaric acid contains β-diketone enol (enolone, HO—C=C—C=O) units with the strong coupling between proton and the π-electron system, which can drive tautomerization between the keto and enol forms: —C=O· · ·HO—C= ↔ =C—OH· · ·O=C— (the resonance-assisted hydrogen bonding) [58, 59]. Many other organic compounds with the similar tautomeric structures such as β-diketoalkane [60], hydroxyphenalenone, and bisquaric acid were investigated [61–63], demonstrating the possibility of quantum paraelectric, isotope effects and/or phase transitions caused by the ruthenium substitution. Similarly, the diazabicyclo[2.2.2]octane (dabco) salts often adopt the tetrahedral ionic structure, and the monovalent Hdabco+ cation is linearly linked through strong N− H+ · · ·N hydrogen bonds between the basic nitrogen and the adjacent moiety. The direction of this linear hydrogen-bonding chain is bistable and can be reversed by a collective proton transfer process, which triggers ferroelectric polarization as high as 16 μC/cm2 (Figure 11.16b). The similar dielectric behavior is also observed in the monoprotonated pyrazinium salts [64]. It is concluded that the KDP-type ferroelectrics require the formation of hydrogen bonds in chains, sheets, or three-dimensional (3D) networks, and their interactions must be strong enough to induce site-to-site proton motions. In contrast, the bistable nature of ferroelectrics requires the preservation of chemical properties of the molecule and

629

630

11 Relationship Between Structure and Ferroelectric Properties

the number of collective protons during the transfer process. For example, KDP retains two protons on each PO4 unit that are critical for polarization reversal. The mechanisms of “displacement” and “proton transfer” in systems with two or more components have been successfully applied to the design of organic ferroelectrics. Charge transfer (CT) mechanism provides a pathway for the design of organic ferroelectrics. For crystals, there are many molecular charges that can be divided into electron donor D+𝜌 and acceptor A−𝜌 , where 𝜌 is a finite CT degree (0 < 𝜌 < 1). As shown in Figure 11.17, the stacking with alternating D+𝜌 and A−𝜌 molecules contains the initially nonpolar · · ·donor and acceptor (DA)⋅DA· · · arrays, forming the bipolar DA dimers in the polar chain. The characteristics for one-dimensional (1D) ferroelectrics were firstly discovered during the dielectric measurements of several CT complexes. For the complex of tetrathiafulvalene (TTF) with the p-bromanil (tetrabromo-p-benzoquinone; BA) and p-chloranil (tetrachloro-p-benzoquinone; CA) [65, 66], the peak-like anomalies of the dielectric constants indicate the occurrence of phase transition. Structural studies on their LTP confirm that molecular displacement leads to symmetry breaking of polar structures [67, 68]. In particular, the ferroelectric ordering of TTF–CA crystals is accompanied by the transformation called neutral ion (NI) phase transition [69, 70]. Figure 11.17a displays the phase transitions for TTF–CA and the monobromine-substituted analog TTF–QBrCl3 (2-bromo-3,5,6-trichloro-p-benzoquinone) [71]. The structure changes from the monoclinic lattice (space group P21 /n) to the polar lattice (space group Pn). The (0k0) X-ray reflections (where k is an odd number) below the NI transition temperature (T NI ) demonstrate ferroelectric ordering with a loss of twofold screw axis (Figure 11.17b). The dielectric constants in the PEP follow the Curie–Weiss law and increase sharply to 500 (Figure 11.17c), similar to the characteristics of the first-order FEP transition. Just above the T c , in the stack-axis polarized reflectivity in the far-infrared region [72], the soft mode is displayed as a Drude-like high-reflectance band. Meanwhile, the derivatives of dimethyl-substituted TTF (a series of dimethyl-substituted TTF [DMTTF–CA]) were also found to undergo the NI transformation with antiferroelectric ordering [73]. In addition, the actual dielectric constant is accompanied by soliton-like domain walls, resulting in a large deviation in the Curie–Weiss law at FEP (Figure 11.17c). The NI transition system can be regarded as a special system, with ferroelectricity is caused by intermolecular electron transfer and rearrangement of molecular charge distribution. Ferroelectric behaviors have been observed in some organic CT salts, such as (TMTTF)2 X and (ET)2 RbZn(SCN)4 (TMTTF = tetramethyl-TTF; ET = bis(ethylenedithio)TTF, and X stands for monovalent ions, such as PF6 and ReO4 ) [74]. In addition to the NI transition, the 1D S = 1/2 spin system produces a nonmagnetic (spin singlet) of D+ A− pairs. Many ion paramagnetic CT complexes belong to this type [75]. The spontaneous bending of molecular component forms a bipolar DA chain, which is potential for the design of organic ferroelectric (Figure 11.16b) [76]. It is notable that multicomponent molecular components play a key role for the design of ferroelectrics. The dielectric constant of CT complexes possesses very large 𝜅, while the dielectric loss is also large. These compounds tend to be semiconductor

11.2 Recent Advance of Molecular Ferroelectrics

D+ A–

D0 A0

0.6

Ionicity (p)

TTF–CA TTF–QBrCl3

Ionic

0.5 0.4

Neutral

0.3 0.2 (a) (030) intensity (a.u.)

2

Dimerization

1

Paraelectric (P21/n)

Ferroelectric (Pn)

0 (b) 30 kHz

800

a

k

600

400

O

b

200

0 (c)

0

50

100

150

200

250

300

Temperature (K)

Figure 11.17 Neutral-ionic phase transitions for TTF–CA and TTF–QBrCl3 crystals. (a) Molecular ionicity obtained from the frequency shift of the charge-sensitive intramolecular C=O stretch vibrational mode. (b) The long-range ferroelectric order monitored by the X-ray (030) reflection intensity. (c) Dielectric constant measured at f = 30 kHz parallel to the DA stack. Inset: molecular stacking viewed along the crystallographic c-direction.

or incomplete insulators. The CT complexes also undergo the current-induced resistance switching [77], which will prevent the occurrence of ferroelectricity. Current leakage will reduce Ps and thus hinder various ferroelectric applications. The assembly of supramolecular structures by replacing CT interactions with the hydrogen-bonding interactions might reduce dielectric loss. Therefore, the

631

632

11 Relationship Between Structure and Ferroelectric Properties

acid–base combination is a direct alternative to form the binary or multicomponent molecules by intermolecular hydrogen bonds. The typical cases are the neutral supramolecular system of simple hydrogen-bonded nonpolar molecules [78], and the supramolecular system with proton transfer [79, 80]. Both types of ferroelectrics consist of alternating linear chains of acid–base molecules. As an acidic component, the dissociation constant pK 1 value for 2,5-dihalo-3,6-dihydroxy-p-benzoquinones (H2 xa) is smaller than the bipyridine bases. When H2 xa is used as the proton donor, two protons are released at a time (Figure 11.18), thus forming ionic DA supramolecular structures with various bipyridine bases [81]. In contrast, chloranilic acid (H2 ca) and bromanilic acid (H2 ba) maintain two protons to react with the weaker base of phenazine (Phz, Figure 11.18). These dibasic acid or base molecules locate two proton-donating O–H groups or proton-accepting nitrogen atoms on both sides and are thus suitable for forming hydrogen bonds in an infinite DA alternating chain. The other important feature is that their neutral and divalent forms are symmetrical, while the monovalent species adopt asymmetric motif, 2,5-Dihydroxy-p-benzoquinones O O OH –H+ X X HO

X O H2xa

+H+

HO

O X

O Hxa–

pK1

O

–

+H+

–

O

X

–H+

–O

X O xa2–

pK2

X = F Cl Br I H2fa H2ca H2ba H2ia pK1 1.40 0.73 0.80 pK2 3.30 3.08 3.10 (a) Phenazine (Phz) N

+H+

N

+

+H+ N –H+

R

N

H

pK2 +H+ H –H+

R

pK1 = 1.20

N

R

pK1

N

R

+

–H+

2,2-Bipyridines R

N

H N

+

N+

N

H

R = H CH3 22bpy 55dmbp pK1 4.44 4.2 pK2 –0.52

R

(b)

Figure 11.18 Deprotonation/protonation processes and acidic dissociation constants of binary molecules. (a) Anilic acids with two proton-donating hydroxy groups. (b) Base molecules with two proton-accepting nitrogen atoms.

11.2 Recent Advance of Molecular Ferroelectrics

as shown in 2,2′ -bipyridines (22bpy) and 5,5′ -dimethyl-2,2′ -bipyridine (55dmbp) (Figure 11.18). For the ferroelectric of Phz–H2 xa, the components are completely symmetrical at room temperature and its crystal structure is shown in Figure 11.19a. At the Curie temperature (T c = 253 K), the dielectric constants (𝜅) increase rapidly to 2000–3000. Below T c , the emergence of thermoelectric charge, reversible polarity, and P–E hysteresis loops confirms its ferroelectric state (Figure 11.20c). Deuterium substitution of acidic protons increased its T c by more than ∼50 K, revealing the important role of the proton for its phase transition (Figure 11.20). In contrast, there was no significant change of T c when the Phz molecule was completely deuterated. This remarkable isotope effect is due to the fact that the O–D vibration mode is inherently softer

b

O

c

(a) P PE phase

FE phase C N

O

H

C

O

(b)

N

N

H

O

O H

N

N

H O

O H

N

N

(c)

Figure 11.19 Crystal structure of Phz–H2 xa. (a) Molecular packing and displacement (yellow arrows) viewed along the a-axis direction. (b) Molecular structures of H2 ba in PEP and FEP. (c) Schematic drawing of the displacement of hydrogens (blue arrows) and molecules (yellow arrows) at FE phase.

633

11 Relationship Between Structure and Ferroelectric Properties 3,000 D2ca

H2ca

Phz–H2xa/D2xa

D2ba

2,000 H

P‖b

D

k

H2ba

304 K

1,000

Phz–D2ca

300 K

0

0

100

(a)

200 Temperature (K)

300

II IC

Tc

P

Tc

294 K

2 Ps (μC/cm2)

634

D2ca H2ca

1 μC/cm2 D2ba

1

288 K H2ba I Tc

0

(b)

0

100

200 Temperature (K)

300

(c)

–2

0 E (kV/cm2)

2

Figure 11.20 Ferroelectric and related physical properties of Phz–H2 xa cocrystals. (a) Temperature-dependent dielectric constant. (b) P s obtained by pyroelectric currents. Successive ferroelectric-to-incommensurate (T c IC ) and incommensurate-to-second ferroelectric (T c II ) transitions (arrows) are observed. (c) P–E hysteresis loops. The electric field E was applied parallel to the crystal b-axis (see the picture).

than the O–H vibration mode. The room-temperature ferroelectricity (T c = 304 K) of the deuterated Phz–D2 ca crystal is observed, as shown in Figure 11.20c. Its spontaneous polarization is as high as 0.7–0.8 μC/cm2 at room temperature and about 2 μC/cm2 at low temperature (Figure 11.20b). The measurement of heat capacity demonstrates small entropy change at T c compared with that expected for configurational order–disorder mechanism, suggesting a displacement mechanism. There are two other successive phase transitions at lower temperature that still retain the ferroelectric state (Figure 11.20b). The symmetry breaking of Phz–H2 xa (from P21 /n to P21 ) results in the Ps along the b-axis. At the low-temperature FEP, the acidic proton retains the neutral O–H⋅⋅⋅N form, rather than the site proton transfer between O atom and N atom. X-ray structure analysis shows relative displacement between the D and A molecules (Figure 11.19a). Further neutron diffraction demonstrates that the microscopic origin of molecular symmetry breaking is the displacement of hydrogen nucleus (Figure 11.19b) [82], as shown by the schematic in Figure 11.19c. This sort of proton migration is an incipient phenomenon of the acid–base reaction to the proton-transferred monovalent form. Proton transfer can be achieved by matching the proton affinities of N and O, even for the heteronuclear asymmetric bonds. The pK 1 values of Phz and H2 xa are almost the same, thus causing this abnormal hydrogen displacements. The electrodeization theory based on Berry-phase picture can demonstrate that the ionic consistency plays an important role for the displacive-type ferroelectrics. Therefore, combined with computational analysis, the formation of strong hydrogen bonds

11.2 Recent Advance of Molecular Ferroelectrics CH3 X

H

O

O–

O

O

H

X

N+

O–

O N

H

X

O

O X

P

CH3 CH3

b

X O

O H

–O

(a)

P

X

N

O +N

O X

O H

O

H –O

X

(b)

CH3

N –

H

+

N

O

O

c

O

H

N –

H

+

N

O

O

H

N –

H

+

O H

O

N

(c)

Figure 11.21 Crystal structure of [H-55dmbp][Hia] at FEP. (a) Molecular structure and hydrogen bonds (broken lines) with two opposite polarizations (open arrows), which can be reversed by proton transfer (curved arrows). (b) Crystal structure at 50 K viewed along the a-axis. (c) Schematic draw of proton ordering and collective proton transfer processes (curved arrows) during the polarization reversal.

between molecules with well-matched proton affinities can be used as a strategy for designing molecular ferroelectrics. Another type of ferroelectrics is the monovalent ionic compound with 55dmbp as the base. For [H-55dmbp][Hca], an above-room-temperature phase transition (T c = 318 K) and high dielectric constants (𝜅 ≈ 140) are observed. Its supramolecular crystal structure in Figure 11.21 reflects the proton-ordered form below T c . As all the protons are transferred simultaneously, the chain is reversed in polarity without losing its chemical identity. This collective proton transfer process under an external electric field is suggestive of ferroelectricity. Above T c , the hydrogen bond chain in the 55dmbp salt resumes the antisymmetry, and the protons are disordered in the O–H⋅⋅⋅N and N− H + ⋅⋅⋅O− forms. Importantly, the geometry of the π electrons closely relates to the protonation state of two molecules. The ordering of proton is confirmed by the large dielectric response observed along the direction of hydrogen bonds. Ferroelectricity of [H-55dmbp][Hia] has been confirmed by the occurrence of temperature-dependent dielectric behavior, P–E hysteresis loop, and pyroelectric activities (Figure 11.22). The ferroelectric properties of these two compounds are derived from the parallel chain polarity. The Ps of [H-55dmbp][Hia] is much larger than Phz–H2 xa. The application of hydrostatic pressure inhibits the FEP transition with atomic displacement (Figure 11.22). At the pressure of 0.8 GPa, the FEP disappears, which relates to the deuteration of hydrogen bonds. As in Phz–H2 xa, T c is significantly elevated by hydrogenation. In addition, [D-55dmbp][Dia] is a room-temperature ferroelectric with thermoelectricity and large thermoelectric coefficient (≈400 μC/cm2 /K) [83].

635

11 Relationship Between Structure and Ferroelectric Properties

Paraelectric Tc

200

0

0

0.69

[H-55dmbp][Hia]

2

0.53

0 GPa

2

1000

D-salt

0 –2

500

297 K

2

0.85

[D-55dmbp][Dia]

3

0.5 1.0 Pressure (GPa) 0.76

0.90

[H-55dmbp][Hia]

4

P (µC/m /K)

1,000

500

[D-55dmbp][Dia]

Ferroelectric

100

P (µC/cm2)

1,500

2 Ps (µC/cm )

T(K)

300

k

636

0.80 –5 1

0 E (kV/cm)

5

0 0

(a)

0 0

100

200 Temperature (K)

300

(b)

0

100

200

300

Temperature (K)

Figure 11.22 Ferroelectric and related physical properties of [H-55dmbp][Hia] and [D-55dmbp][Dia]. (a) Temperature dependence of dielectric constants. Inset: phase diagram. (b) Temperature variation of P s (solid curves) and pyroelectric coefficient (broken curves). The dc poling field of 2–3 kV/cm was applied before cooling from above T c . Inset: P–E hysteresis loop at room temperature measured along its c-axis.

11.2.2 Binary Molecular Ferroelectrics Ferroelectricity in binary oxides is of scientific interest and has been theoretically predicted for alkaline oxides. Binary ferroelectric materials are often used as potential candidates for electromechanical actuators and ultrasonic transducers. As early as 1997, piezoelectric properties of ferroelectric crystal Pb(Zn1/3 Nb2/3 )O3 –PbTiO3 (PZNT) and Pb(Mg1/3 Nb2/3 )O3 –PbTiO3 (PMNT) were reported [84–89], of which the piezoelectric coefficients reach up to 2500 pC/N. Besides, high electromechanical coupling of 90% and low dielectric loss of 1% make these materials for high-performance solid-state actuators. In 2000, single crystals of (1 − x)Pb(Sc1/2 Nb1/2 )O3 –xPbTiO3 (PSN–PT) within the morphotropic phase boundary (MPB) region were synthesized [90–92], which show excellent dielectric and ferroelectric properties with the maximum dielectric constant of 60 000 (at 1 kHz, T c = 213 ∘ C) and remnant polarization of 25 μC/cm2 . The electromechanical coupling factor of the pre-poled crystals reaches 78%. Both T c and T MPB of PSN–PT crystals are higher than those of PMNT and PZNT piezocrystals, making them potential high-T c and high-performance piezoelectric materials. In 2002, Pb(Yb1/2 Nb1/2 )O3 –PbTiO3 (PYNT) crystal with large piezoelectric coefficient [93, 94] and high T c (>250 ∘ C) was reported. Electrical properties of PYNT (60/40) along the (001), (011), and (111) pseudo-cubic directions reveal the Pr values of 26.5, 27.6, and 41.6 μC/cm2 , while coercive fields were 10.1, 9.4, and 11.5 kV/cm. Meanwhile, piezoelectric coefficient of the (001) oriented sample was found to be ∼1200 pC/N, and the strain reached 0.54% at 100 kV/cm. 70Pb(Mg1/3 Nb2/3 )O3 –30PbTiO3 crystals show the piezoelectric coefficient of 1240 pC/N, with a strain level reaching 0.12% at 12 kV/cm. Lately, a series of (1 − x)Pb(Lu1/2 Nb1/2 )O3 –xPbTiO3 (PLN–xPT) binary ferroelectrics with high Curie temperature were obtained (Figure 11.23) [95], which have excellent piezoelectric properties. For example, PLN–0.49PT crystal has high Curie temperature of 360 ∘ C and high rhombohedral–tetragonal phase transition temperature at 110 ∘ C. The fascinating properties of piezoelectric coefficient (1630 pC/N), electromechanical

11.2 Recent Advance of Molecular Ferroelectrics

Figure 11.23 P–E hysteresis loops for the (001) oriented PLN-xPT (x = 0.47 and x = 0.49) crystals (f = 2 Hz).

PLN–0.47PT PLN–0.49PT

Polarization (μC/cm2)

30 20 10 0 –10 –20

E = 20kV/cm

–30 –20

–10

0

10

20

Electric field (kV/cm)

coupling coefficient (81%), coercive field (13.8 kV/cm), and remanent polarization (26.6 μC/cm2 ) suggest its potentials for high-performance multifunctional electric device applications. Moreover, the transition metal oxides ZrO2 and HfO2 as well as their solid solutions are widely studies [96], for which the ferroelectricity in thin films might be unexpected. Grazing incidence X-ray diffraction (GI-XRD) structure analysis reveals a composition-dependent tetragonal to orthorhombic to monoclinic phase transition in HfO2 –ZrO2 thin films. The temperature-dependent nonlinear dielectric activities in pure ZrO2 and its solid solutions with HfO2 confirm the existence of phase transition and also suggest the assumption of ferroelectricity. The field-driven and temperature-dependent ferroelectric transition was observed in pure ZrO2 and the Zr-rich part of the phase diagram, revealing component-related structure transformation (Figure 11.24).

P (μC/cm2)

30 15 0 –15 –30

HfO2

Hf0.6Zr0.4O2

Hf0.7Zr0.3O2

Hf0.5Zr0.5O2

Hf0.3Zr0.7O2

ZrO2

ɛr

50 40 30 20 –4 –2

0

2 ±4 –2 0

2 ±4 –2

0

2 ±4 –2

0

2 ±4 –2

0

2

4 –2

0

2

Electric field (MV/cm)

Figure 11.24 P–E hysteresis loops at 1 kHz and small signal capacitance-voltage (CV) hysteresis at 10 kHz for HfO2 –ZrO2 thin films. An evolution of paraelectric HfO2 , ferroelectric HfO2 –ZrO2 , and antiferroelectric-like ZrO2 can be observed in Polarization-voltage (PV) and CV results.

4

637

11 Relationship Between Structure and Ferroelectric Properties

11.2.3 Organic–Inorganic Hybrid Ferroelectrics Organic–inorganic hybrid compounds represent a large family of ferroelectric materials, which have attracted great attention in recent years. Here, we mainly focus on several typical compounds with intriguing and clear-cut structure changes to discuss the structure–property relationship. In the class of [(CH3 )4 N]2 [MX4 ] [97–99], [(CH3 )4 N]2 [FeCl4 ], [(CH3 )4 N]2 [MnCl4 ], and [(CH3 )4 N]2 [CuCl4 ] undergo the phase transitions with the change of temperature and pressure. When the organic cation is replaced by dimethylammonium, two hybrid compounds of [(CH3 )2 NH2 ]2 [CoCl4 ] and [(CH3 )2 NH2 ]2 [ZnCl4 ] were found to show bulk ferroelectricity. [(CH3 )2 NH2 ]2 [CoCl4 ] was grown from the aqueous solution as dark blue crystals [100]. At room temperature, it crystallizes in the monoclinic space group P21 /n (Figure 11.25) and undergoes several phase transitions in the temperature range of 200–430 K. The similar compound [(CH3 )2 NH2 ]2 [ZnCl4 ] is crystallized in the monoclinic space group P21 /n at room temperature. Its phase sequence has not been established [101]. Organic–inorganic hybrid ferroelectric [(CH3 )2 NH2 ]3 [CuCl4 ]Cl [102, 103] adopts the disordered chloride ions and anions at room temperature. Dielectric measurements exhibit successive first-order phase transitions at 279 and 253 K with the thermal hysteresis of 6 and 7 K, respectively (Figure 11.26). The step-like dielectric Figure 11.25 Crystal structure of [(CH3 )2 NH2 ]2 [CoCl4 ] at 293 K. Green dotted lines represent hydrogen bonds. b c

Cl Co N C

Figure 11.26 Temperature dependence of dielectric constant of [(CH3 )2 NH2 ]3 [CuCl4 ]Cl.

300

200

f = 1.2 kHz (100)

εʹ

638

100

0 240

260

280 T / (K)

300

11.2 Recent Advance of Molecular Ferroelectrics

constant anomalies suggest its improper ferroelectricity. At high-temperature PEP, [(CH3 )2 NH2 ]3 [CuCl4 ]Cl belongs to the orthorhombic space group of Pnam, and the intermediate-temperature phase is ferroelectric with the Ps of 3.4 μC/cm2 . Hybrid compound [CH3 NH3 ]2 [ZnCl4 ] crystallizes in the monoclinic system with space group of P21 /c at room temperature [104, 105]. DSC and dielectric studies reveal the presence of successive phase transitions at 483 and 555 K. The high values of the dielectric constants indicate its potential ferroelectricity [106]. [C2 H5 NH3 ]2 [CuCl4 ] belongs to the family of [Cn H2n+1 NH3 ]2 [MCl4 ] where M is divalent Mn, Cd, Fe, or Cu [107–110]. Depending on the length of organic chains, different structures and chain packings can be achieved. Such compounds recently have attracted attention because of their rich structural varieties and magnetic phases, which bring up an opportunity to explore multiferroic properties. [C2 H5 NH3 ]2 [CuCl4 ] shows a perovskite-type layer structure, containing infinite layers of corner-sharing distorted CuCl6 octahedra and ethylammonium layers (Figure 11.27) [111]. The hydrogen-bonding interactions between organic and inorganic parts exert an effect on the temperature evolution of structural phase transitions. [C2 H5 NH3 ]2 [CuCl4 ] undergoes a series of phase transitions, and sharp dielectric anomaly is found around 247 K. The broad shape and the maximum value of 𝜀′ at 247 K indicate the characteristics for an improper ferroelectric (Figure 11.28a). The Ps below 247 K is measured to be 18 μC/cm2 at 200 K (Figure 11.28b), which might be ascribed to ordering of the organic [C2 H5 NH3 ]+ . For [Cn H2n+1 NH3 ]2 [MX4 ] perovskites, it is common that the NH3 group of organic

a

c

b Cu N

C

Figure 11.27

Cl

Crystal structure of [C2 H5 NH3 ]2 [CuCl4 ] at 293 K.

20 f = 1 kHz (100)

15 Ps (μC/cm2)

16

20

εʹ

12 8

5

4 T4

0

(a)

(100)

10

100

150

200 T (K)

T5 250

300

(b)

0 120

160

200 T (K)

240

280

Figure 11.28 (a) Temperature dependence of dielectric constant, and (b) P s versus T curve of [C2 H5 NH3 ]2 [CuCl4 ].

639

11 Relationship Between Structure and Ferroelectric Properties

chain flips among 4 equiv orientations inside the cavity formed by the CuCl6 octahedra. Interestingly, [C2 H5 NH3 ]2 [MCl4 ] shows ferromagnetic interaction with a magnetic phase transition at 10.2 K. In a wide temperature range, it shows characteristics of the two-dimensional (2D) Heisenberg ferromagnet with a dominant intralayer exchange coupling J/kB = 18.6 K for spins at the nearest-neighbor Cu sites [112, 113]. N-(methylpyrrolidinium)3 Sb2 Br9 is an organic–inorganic hybrid ferroelectric that has the zero-dimensional (0D) structure and notable semiconducting properties [114]. At 293 K, N-(methylpyrrolidinium)3 Sb2 Br9 crystallizes in the polar space group R3c. The [Sb2 Br9 ]3− cluster adopts the twisted octahedron and each antimony atom is connected to six bromide atoms (Figure 11.29a). The difference in Sb—Br bond length is caused by the displacement or relative motion of bridged Br atoms, which triggers an asymmetric structure. With temperature increasing above 355 K, the [Sb2 Br9 ]3− cluster becomes symmetrical and the bridge atoms are evenly distributed (Figure 11.29b). The organic cation and the inorganic moiety undergo an order–disorder transformation, which leads to the PEP-to-FEP transition. In order to detect the symmetry breaking and phase transitions, DSC and SHG measurements were performed on N-(methylpyrrolidinium)3 Sb2 Br9 (Figure 11.30a). The temperature-dependent SHG response shows that a sharp change of SHG signals near T c , indicating that N-(methylpyrrolidinium)3 Sb2 Br9 b

b c

c

Sb Br N C

(a)

(b)

0.3

600

Cooling

1/ɛʹ Fitted with Curie–Weiss law f = 500 kHz

0.04

0.0

450 c-axis Heating

–0.3 1.5

ɛʹ

Heating flow (mW)

Figure 11.29 Projection of crystal structures of N-(methylpyrrolidinium)3 Sb2 Br9 at (a) 293 K and (b) 355 K.

SHG (a.u.)

640

1.0 0.5

300

150 Cooling

0.00 315

320 325 330 Temperature (K)

335

Heating 0

0.0 300 (a)

0.02

1 KHz 50 KHz 100 KHz 500 KHz 1000 KHz

320 340 Temperature (K)

310

360 (b)

320 330 Temperature (K)

340

Figure 11.30 (a) DSC and SHG results for N-(methylpyrrolidinium)3 Sb2 Br9 . (b) The temperature-dependence of 𝜀′ measured along the c-axis direction.

11.2 Recent Advance of Molecular Ferroelectrics

changes from the noncentral symmetry to central symmetry, as confirmed by thermal anomalies in the DSC trace. Further, temperature dependence of 𝜀′ revealed that N-(methylpyrrolidinium)3 Sb2 Br9 shows a distinct abnormal peak at T c (Figure 11.30b). This follows the Curie–Weiss law and confirms the PEP-to-FEP transition. Most importantly, the P–E hysteresis loop measured at 306 K affords the Ps value of ∼7.6 μC/cm2 , higher than that of the most hybrid ferroelectrics. Besides, photoconductive measurements indicate that N-(methylpyrrolidinium)3 Sb2 Br9 is a semiconductor ferroelectric. (C4 H10 N)CdCl3 is the ferroelectric material used as solid-state NLO switch with “off-on-off” states in cooling process [115]. Such interesting NLO changes relate to the arrangement and orientation of polar cations during its successive phase transitions (Figure 11.31a). At HTP (above 298 K), the organic cations locate at 2 mm symmetry sites with four dynamically disordered sites. Between 298 and 215 K (intermediate-temperature phase [ITP]), organic cations locate around the m planes of the vertical a-axis with two disordered sites. The N atoms adopts one orientation along c-axis direction, leading to its net electric polarization. At LTP (below 215 K), it transforms to a centrosymmetric space group C2/c, and the interchain at ab plane changes greatly with the a-axis expanding by 20.7% and the b-axis decreasing by 17.7%. The chain slides along the c-axis, and the 𝛽 angle changes to 95.29∘ . The organic cations show obvious displacement and situate in the narrow space between inorganic chains. Thus, (C4 H10 N)CdCl3 exhibits the “off-on-off” NLO-switching states during its successive phase transitions (Figure 11.31b). The large bandgap of inorganic ferroelectrics often limits their optoelectronic device applications, which inspired the exploration of ferroelectrics with the narrow bandgap. Organic–inorganic hybrid ferroelectric of (2-(aminomethyl) pyridinium)SbI5 , which consists of the zigzag chains of SbI6 octahedra, was recently reported to show large Ps and narrow bandgap of 2.03 eV [116]. At room-temperature FEP, it crystallizes in a polar space group of Pb21 a, which changes to Pbca at PEP (above 360 K). The alignment of organic cations results in its ferroelectricity (Figure 11.32). SHG, P–E hysteresis loop, and piezoresponse force microscopy measurements confirm its bulk ferroelectricity (Figure 11.33).

a SHG intensity (a.u.)

SHG-on state c

0.4 Cooling 0.2 SHG-off state

SHG-off state Heating

0.0 140 LTP

(a)

ITP

HTP

160

180

200

220

240

260

280

300

T(K)

(b)

Figure 11.31 (a) Crystal structures of (C4 H10 N)CdCl3 at HTP, ITP, and LTP. (b) The variable-temperature SHG results for (C4 H10 N)CdCl3 .

641

11 Relationship Between Structure and Ferroelectric Properties b

a

III

IV

III A

IV A

I

II

IA

II A

III

IV

III A

IV A

I

II

IA

II A

C N I Sb

Figure 11.32 Projection of crystal structures of (2-(aminomethyl) pyridinium)SbI5 at (a) PEP and (b) FEP, respectively.

c

(a) b

Sb I C N

(b)

Figure 11.33 Ferroelectric P–E hysteresis loops of (2-(aminomethyl)pyridinium)SbI5 .

4 Polarization (μC/cm2)

642

E // c-axis 2 0 338 K 345 K 350 K 363 K

–2 –4 –15

–10

–5

0

5

10

15

E (kV/cm) 1 – 4

– 1: 0 – 6:1 – 4

– 1: 0 – 6:1 – 4

Symmetry breaking

Symmetry restoration – 1: 0 – 6:1 – 4

– 1: 0 – 6:1 – 4

P63/m (No.176)

Figure 11.34

P61 (No.169)

Symmetry change of [(CH3 )4 N]CdBr3 during its phase transition.

[(CH3 )4 N]CdBr3 undergoes an ordered–disordered phase transition at 156 K and adopts the ferroelectric space group P61 [117, 118]. The change of symmetry does not abide by the Curie principle because P61 is not one of the maximal nonisomorphic subgroups of the PEP P63 /m (Figure 11.34). Temperature dependence of the dielectric constants for [(CH3 )4 N]CdBr3 shows a 𝜆-type peak along the c-axis direction. The Ps value is about 0.1 μC/cm2 at 135 K (Figure 11.35), stemming from the orientational ordering of organic cations. Similarly, the mercury-containing hybrid

11.2 Recent Advance of Molecular Ferroelectrics 10 0.012

εʹ

Ps (μC/cm2)

(001)

8

6

f = 100 kHz (100)

(a)

200 T (K)

0.004 0.000 120

4 100

0.008

300

130

140

150

160

T (K)

(b)

Figure 11.35 (a) Temperature dependence of dielectric constant and (b) P s versus T curve for [(CH3 )4 N]CdBr3 .

Figure 11.36 293 K.

Crystal structure of [(CH3 )4 P][HgCl3 ] at

Hg

b

N

Cl C

c

counterparts, including [(CH3 )4 N]HgX3 (X = Cl, Br, and I) and [(CH3 )4 P]HgBr3 , show ferroelectricity before their decomposition temperatures [119, 120]. They crystallize in polar space groups of P21 or Pb21 m at room temperature. Structurally, the 1D HgX3 chains are separated by the discrete [(CH3 )4 N]+ or [(CH3 )4 P]+ ions (Figure 11.36), and the ordering of cations results in polarization (Ps ) of 1–3 μC/cm2 [121]. Compound [MV][BiBr5 ] (MV = methylviologen dication) is the semiconducting ferroelectric synthesized by solvothermal method [122], which belongs to monoclinic space group P21 /c at room temperature but turns to polar P21 at 233 K. Structurally, the transconnected 1D BiBr5 regular chain propagates along the a-axis (Figure 11.37), separated by the planar MV2+ cations. Dielectric constants exhibit anomalies at 243 K (Figure 11.38), confirming the occurrence of a phase transition. However, further evidence is needed to confirm the real ferroelectric polarization in this compound. Ferroelectric of [(CH3 )2 NH2 ]3 Sb2 Cl9 crystallizes in the monoclinic space group P21 /c at room temperature and adopts a ferroelectric space group Pc below 242 K [123]. In the crystal structure, the corner-sharing [SbCl6 ]3− octahedra form 2D anionic layers parallel to the bc plane (Figure 11.39). There are two nonequivalent [(CH3 )2 NH2 ]+ cations in the asymmetric unit cell, that is, one at the symmetry center inside the polyanionic cavities and the other at the general position. At LTP,

643

11 Relationship Between Structure and Ferroelectric Properties

Figure 11.37

Crystal structure of [MV][BiBr5 ] at 233 K.

b a

Br Bi C N

Figure 11.38 Temperature-dependent dielectric constants of [MV][BiBr5 ].

200

f = 5 kHz

160 εʹ

644

120

80 200

225

250

275

Figure 11.39 Crystal structure of [(CH3 )2 NH2 ]3 Sb2 Cl9 at 100 K.

C Sb

a

N

b

Cl

c

c

(a)

(b)

distinct deformation of hydrogen bonds and polyanionic layers are observed, which probably account for the ferroelectricity in [(CH3 )2 NH2 ]3 Sb2 Cl9 . Dielectric constants of [(CH3 )2 NH2 ]3 Sb2 Cl9 display the critical slowing down characteristic for second-order phase transition (Figure 11.40) [124]. Similarly, the analog of [(CH3 )2 NH2 ]3 Sb2 Br9 undergoes first-order FEP transition at 164 K [125], resulting from the freezing of [(CH3 )2 NH2 ]+ cations. The ferroelectric of [(CH3 )3 NH]3 Sb2 Cl9 undergoes complex phase transitions, which adopts the 2D layer structure (Figure 11.41) [126]. The ordering of organic cations leads to PEP-to-FEP transition at 363 K.

11.2 Recent Advance of Molecular Ferroelectrics 0.8

1000 800 Ps (μC/cm2)

0.6 f = 1 kHz (100)

εʹ

600 400

0.4 (100) 0.2

200 0 238

240

242

244

246

T (K)

(a)

0.0 210

248

220

230

240

T (K)

(b)

Figure 11.40 (a) Temperature dependence of dielectric constant and (b) P s versus T curve for [(CH3 )2 NH2 ]3 Sb2 Cl9 . Figure 11.41 Crystal structure of [(CH3 )2 NH2 ]3 Sb2 Cl9 at 165 K.

c b

N Cl

C

Sb

750 2.0 600

x=0

εʹ

450

Ps (μC/cm2)

f = 1 kHz (001)

300

0 340 (a)

1.0 0.5

150

350 T (K)

360

x = 0.42

1.5

[(CH3)3NH]3[Sb2CI9-3xBr3x]

0.0 320

370 (b)

330

340

350

360

370

T (K)

Figure 11.42 (a) Temperature dependence of dielectric constant of [(CH3 )2 NH2 ]3 Sb2 Cl9 and (b) P s versus T curve for [(CH3 )3 NH]3 Sb2 Cl9−3x Br3x .

Two dielectric anomalies are observed at 363 and 364 K for [(CH3 )2 NH2 ]3 Sb2 Cl9 , and the latter is independent of temperature change while the former exhibits an appreciable thermal hysteresis (Figure 11.42a). The Ps versus T curve shows typical feature for the first-order phase transition. The halogen element replacement causes a reduction of Ps value in [(CH3 )3 NH]3 Sb2 Cl9 , i.e. from ∼2 to 1.5 μC/cm2 for [(CH3 )3 NH]3 Sb2 Cl9−3x Br3x (x = 0.42) (Figure 11.42b) [127]. For [(CH3 )3 NH]3 Sb2 Br9 , the structure consists of discrete Sb2 Br9 bioctahedra and disordered organic cations, and the 2D polyanionic layers are synergetic for the ferroelectricity.

645

11 Relationship Between Structure and Ferroelectric Properties

[CH3 NH3 ]5 Bi2 Cl11 and [CH3 NH3 ]5 Bi2 Br11 are isomorphous ferroelectrics with the bioctahedral Bi2 Cl11 or Bi2 Br11 anionic clusters [128, 129]. The second-order phase transition results from the ordering of organic cations and the distinct distortion of the [Bi2 Cl11 ]5− moiety (Figure 11.43). Dielectric constants of [CH3 NH3 ]5 [Bi2 Cl11 ] along the c-axis exhibit sharp anomalies with the peak value of ∼5 × 103 (Figure 11.44). According to the Curie–Weiss law, the Cpara and Cferro are estimated to be 1.38 × 103 and 0.36 × 103 K, giving the Cpara /Cferro ratio of 3.83, which suggests the second-order phase transition type. [CH3 NH3 ]5 Bi2 Br11 demonstrates the similar dielectric behaviors to that of [CH3 NH3 ]5 [Bi2 Cl11 ] and the phase transition is also of order–disorder type [130]. Recently, 2D hybrid perovskite ferroelectric of (BA)2 (MA)2 Pb3 Br10 was used for assembling photodetectors (Figure 11.45) [131], of which the phase transition results from the distortion of inorganic moieties and synergic ordering of organic moieties. Due to the presence of ferroelectricity, dark currents (∼10−12 A) and excellent photo-detecting performances were achieved, including large on/off current ratios (∼2.5 × 103 ) and faster response rate (150 μs). Significantly, this 2D perovskite is environmentally stable under ambient condition and exhibits a high thermal stability up to 230 ∘ C. These results highlight the promise of 2D layered perovskite ferroelectrics for high-performance optoelectronic applications. Figure 11.43 Crystal structure of [CH3 NH3 ]5 [Bi2 Cl11 ] at 293 K.

b C Cl

Bi

c

N

1.00

6000

f = 1 kHz (001)

2000

Ps (μC/cm2)

0.75 4000 εʹ

646

0.50 0.25

0 303 (a)

306

309 T (K)

0.00 280

312 (b)

300

290

310

T (K)

Figure 11.44 (a) Temperature dependence of dielectric constant of and (b) P s versus T curve of [CH3 NH3 ]5 Bi2 Cl11 .

11.2 Recent Advance of Molecular Ferroelectrics Pb Br C N

Current (nA)

40

MAPbBr3

Tailoring

EQE (%)

60

20

40

5V 3V 2V

20 0 400 500 600 Wavelength (nm)

0

700

365 nm 420 nm 475 nm 520 nm

–20 –40

600 nm 650 nm 700 nm Dark

–60 –4

(a)

0

2

BA2MA2Pb3Br10 ON

ON

ON

ON

0.4

ON

90

tr~ 150 µs

60

tr~ 570 µs

0.2 off

OFF

0 0

OFF

50

OFF

100

OFF

150

Times (s)

OFF

200

OFF

0.0 46.2

250

0.3 0.2

0.1

30

0.4

Iph (nA)

Iph (nA)

0.3

0.1 544

(c)

4

VSD (V)

120

Current (nA)

–2

(b)

46.3

(d)

46.4

546

548 Time (ms)

46.5 46.6 Time (s)

550

46.7

552

46.8

Figure 11.45 (a) Scheme for the design of 2D hybrid perovskite structure. (b) Variation of photoresponse under different incident wavelengths. (c) Recyclable switchable operation of photo-responses, showing highly stable detection after a long-time illumination. (d) The rise and decay times during on–off light switching.

11.2.4 Metal–Organic Framework Ferroelectrics Ferroelectric metal–organic frameworks (MOFs) fill the gap between pure inorganic and organic ferroelectrics. As a branch of hybrid systems, they benefit from structural variability and tunability of both inorganic and organic building blocks, such as metal ions, ligands, and templates. Such advantages could break through the bottleneck of the fundamental study of ferroelectrics, that is, serendipitous synthesis and limited number of ferroelectrics. MOFs discussed here cover a wide range of compounds, including conventional MOFs in which the metal ions or clusters are coordinately bridged by organic linkers to form 1D, 2D, or 3D structures and 0D ionic compounds. The reason why we classify the hydrogen-bonded ionic compounds into MOFs is that the hydrogen bond bears features of covalent bond and van der Waals, ionic interactions, making it unique among supramolecular interactions. Particularly, dynamics of proton transfer through hydrogen bonds can trigger FEP transitions in so-called hydrogen-bond ferroelectrics. The ferroelectric MOFs are organized by anionic or organic parts. Emphases are put on their structural analysis, dielectric and ferroelectric properties. Potassium sodium tartrate tetrahydrate, [KNa(C4 H4 O6 )]⋅4H2 O, usually called RS, is a typical hydrogen-bonding MOF [132], which is the pioneer in the field of ferroelectricity. Structure analysis indicates that unit cell of RS includes one tartrate, four water molecules, one Na cation, and one half of K(1) and K(2) cations (Figure 11.46)

647

11 Relationship Between Structure and Ferroelectric Properties

Figure 11.46 A view of the unit cell of RS at PEP. Green dotted line denotes hydrogen bond.

K1 H

O1 C

O8

O2

O7

O4

Na1 O5 O6 O3 O10 O9

a b

K2

[133, 134]. K and Na ions are surrounded by oxygen atoms of tartrate and water molecules. K(1) and Na(1) ions act as bridge ions between two tartrates, K(2) ions as bridge ions between four tartrates. Tartrate is tightly connected to molecules along the b-axis by O(8) water atoms. There is a row of tartaric acid ions parallel to the a-axis, connected by K ions alternating along the b-axis and Na ions. Two Curie points at 255 and 297 K were observed for RS, and between these two temperatures RS crystallizes in the monoclinic ferroelectric space group of P21 . The symmetry breaking of 222F2 suggests that the symmetry elements change from 4 (E, C2 , 2C′ 2 ) to 2 (E, C2 ). The ferroelectric P21 belongs to one of the maximal nonisomorphic subgroups of the paraelectric space group P21 21 2 (Figure 11.47). Two anomalies of dielectric constants at 255 and 297 K confirm its successive phase transitions (Figure 11.48a). At FEP (278 K), the Ps of RS is 0.25 μC/cm2 (Figure 11.48b). Notably, the deuterated RS, [NaK(C4 H2 D2 O6 )]⋅4D2 O [135], exhibits two Curie points at 251 and 308 K with the increasing Ps of 0.35 μC/cm2 . Structure analysis of ferroelectric [(NH4 )Li(C4 H4 O6 )]⋅H2 O shows that each Li center is coordinated by one water molecule, three carboxyl groups, and one hydroxyl group (Figure 11.49a). The NH4 + ions exist in the gaps formed by the planes that are formed by the hydrogen bond between NH4 + ions and the oxygen atoms of tartaric acids (Figure 11.49b) [136]. Dielectric constants of [MLi(C4 H4 O6 )]⋅H2 O P21212 a

O Symmetry breaking

P21212

648

Symmetry restoration c P21212 (No.18)

P21 (No.4)

Figure 11.47 Symmetry breaking for RS with the change of symmetry elements from 4 (E, C 2 , 2C ′ 2 at PEP) to 2 (E, C 2 at FEP).

11.2 Recent Advance of Molecular Ferroelectrics 0.4 3 Ps (μC/cm2)

ɛʹ (103)

Deuterated RS

0.3

f = 1 kHz 2 RS Deuterated RS

RS

0.2

0.1

1 0.0 100

150

(a)

200

250

300

255

Figure 11.48 of RS.

270

(b)

T (K)

285

300

315

T (K)

(a) Temperature dependence of dielectric constant, and (b) P s versus T curve

N H

a

O C

Li

b b a

(a)

(b)

Figure 11.49 Structures of [(NH4 )Li(C4 H4 O6 )]⋅H2 O at 293 K. (a) Unit cell along c-axis and (b) layer structure in the ab plane. Green dotted lines denote hydrogen bonds. 4 150

εʹ (103)

εʹ

90 b-axis

60

(a)

2 a-axis 1

30 0 80

f = 10 kHz

3

120

b-axis 0 120

160

200 T (K)

240

280

0

(b)

25

50

75

100

T (K)

Figure 11.50 Temperature dependence of dielectric constant of (a) [(NH4 )Li(C4 H4 O6 )]⋅H2 O and (b) [LiTl(C4 H4 O6 )]⋅H2 O.

exhibit the anomalies at 106 K with the peak value of 140 (Figure 11.50a). Fitting the Curie–Weiss law gives a T 0 of 93.8 K and a Cpara of 37 K [137]. [TlLi(C4 H4 O6 )]⋅H2 O has isomorphous structure to that of [(NH4 )Li(C4 H4 O6 )]⋅ H2 O, which undergoes the second-order phase transition at 11 K (Figure 11.50b) [138], and the Ps value is parallel to the a-axis direction (Figure 11.51). The dielectric constants along the a-axis of [LiTl(C4 H4 O6 )]⋅H2 O have the peak value of 5000, followed by a slight decrease in the temperature of liquid helium. This is similar to the dielectric behaviors of general ferroelectrics and explained by the contribution of domain wall motion [139].

649

650

11 Relationship Between Structure and Ferroelectric Properties

TI Li

H O C

a b

Figure 11.51 View of the unit cell of [LiTl(C4 H4 O6 )]⋅H2 O at 293 K. Green dotted lines represent hydrogen bonds.

b a

a c

(a)

Mn o c

(b)

Figure 11.52 (a) Crystal structure of [Mn3 (HCOO)6 ]⋅C2 H5 OH, and (b) the arrangement of ethanol guest molecules in the channel along the b-axis at 190 K.

[Mn3 (HCOO)6 ]⋅C2 H5 OH adopts NaCl-type structure and contains guest molecules in the cavity [140], which sheds light on the exploration of metal–formate ferroelectrics. Structurally, there is a porous channel that contains different objects (Figure 11.52) and thus modulates the 3D remote magnetic sorting [141]. The magnetic properties come from Mn2+ cation (s = 5/2) incorporated in the host lattice. [Mn3 (HCOO)6 ] is not affected by the direction of electric field, and its ′ dielectric constant (𝜀 = 5) is very small and almost independent of temperature. When the polar solvent is compatible with the channel, the compounds will have an electrical activity, indicating the possibility of collective freezing of guest molecules in the channel [142]. Dielectric constants of [Mn3 (HCOO)6 ]⋅C2 H5 OH have sharp peak-like anomalies at 165 K, which are indicative of the FEP transition (Figure 11.53). The P–E hysteresis loop is only recorded in the temperature range of 145–166 K, and the ferroelectric properties are mainly derived from the guest ethanol molecule. The deuterated

11.2 Recent Advance of Molecular Ferroelectrics

60 C2H5OD 40 εʹ

(100) 20

C2H5OH 0 100

200

300

T (K)

Figure 11.53

Dielectric constants of [Mn3 (HCOO)6 ]⋅C2 H5 OH measured along a-axis.

[Mn3 (HCOO)6 ]⋅C2 H5 OD exhibited similar dielectric behaviors of the sharp peaks at 164 K (Figure 11.53). The family of metal–formate perovskites, [(CH3 )2 NH2 ][M(HCOO)3 ] (M = Mn, Fe, Co, Ni, and Zn) were reported as the promising multiferroic candidates with the coupling of electric and magnetic orderings. For instance, [(CH3 )2 NH2 ][M(HCOO)3 ] crystallizes in the centrosymmetric space group of R3c at PEP [143, 144], and M2+ centers are coordinated by six O atoms to form 3D perovskite frameworks (Figure 11.54). The (CH3 )2 NH3 + cations are located inside the cage with three different orientations. The order–disorder of organic cation results in phase transitions, as verified by dielectric anomalies with large thermal hysteresis of 10 K (Figure 11.55). This temperature dependence of dielectric constants might suggest the improper ferroelectric properties. Further, temperature-dependent SHG effects were measured to confirm the occurrence of their PEP-to-FEP transitions [145]. According to the phenomenological Landau theory, if ignoring high-order terms, there is the Figure 11.54 Perovskite structure of [(CH3 )2 NH2 ][M(HCOO)3 ]. The green dotted lines represent hydrogen bonds.

O

C M

N

c b a

651

11 Relationship Between Structure and Ferroelectric Properties

30

f = 1 kHz

εʹ

20

10

160

200

180

220

T (K)

Figure 11.55 Temperature dependence of the dielectric constant for the family of metal–formate perovskites [(CH3 )2 NH2 ][Mn(HCOO)3 ].

0.1

SHG intensity (a.u)

SHG intensity (a.u)

0.2

0.1

531 532 533 Wavelength (nm)

531 532 533 Wavelength (nm)

0.0 0

50

100

150 T (K)

200

250

300

0.4

χ(2) (pm/V)

120 K 150 K 180 K 181 K 182 K

0

50

0.3 0.2

100

150 T (K)

200

250

300

10 K 40 K 100 K 140 K 152 K 154 K

SHG intensity (a.u)

0.0

10 K 125 K 133 K 135 K 137 K 300 K

χ(2) (pm/V)

0.2 0.3 χ(2) (pm/V)

652

531 532 533 Wavelength (nm)

0.1 0.0 0

50

100

150 200 T (K)

250

300

Figure 11.56 Temperature dependence of SHG effect for (a) [(CH3 )2 NH2 ][Co(HCOO)3 ], (b) [(CH3 )2 NH2 ][Mn(HCOO)3 ], and (c) [(CH3 )2 NH2 ][Zn(HCOO)3 ]. Inset: intensity of SHG signals as a function of wavelength at different temperatures.

linear relationship between SHG coefficient (𝜒 (2) ) and Ps , namely, 𝜒 (2) = 6𝜀0 𝛽Ps . For [(CH3 )2 NH2 ][M(HCOO)3 ], variable-temperature SHG signals exhibit clear variations in the vicinity of T c , indicating the symmetry breaking during their phase transitions (Figure 11.56). The SHG-active properties disclose that metal–formate hybrid perovskites should adopt non-centrosymmetric structures [146]. Structure analysis reveals that their

11.2 Recent Advance of Molecular Ferroelectrics

¼ 0

a

Symmetry breaking Symmetry restoration c

–

R3c (No.167)

Cc (No. 9)

Figure 11.57 Symmetry transformation of [(CH3 )2 NH2 ][M(HCOO)3 ] from PEP to FEP with the change of symmetry elements from 12 (E, 2C 3 , 3C 2 , i, 2S 6 , 3𝜎 v ) to 2(E, 𝜎 h ).

ferroelectric space group Cc is a subgroup of the paraelectric space group R3c (No. 167) because the largest nonhomogeneous subgroup of R3c includes R3c, R32, R31, and C2/c, and Cc is a subgroup of C2/c, while there are also C2 and P1 (Figure 11.57). A series of [MII (HCOO)3 ]-based MOFs have been successfully synthesized, and the organic cations exert strong template effect on their structure topology [147, 148]. Using the smallest NH4 + cation, the (NH4 )[M(HCOO)3 ] perovskites adopt the 49 ⋅66 topology. Recently, ferroelectric properties were reported in (NH4 )[Zn(HCOO)3 ] [149], which crystallizes in the paraelectric space group of P63 22 at 290 K and transforms to the ferroelectric space group of P63 at 110 K. The organic NH4 + cation plays the dominant role for its symmetry-breaking phase transition (Figure 11.58). The NH4 + cation links to metal–formate framework through N–H⋅⋅⋅O with the N⋅⋅⋅O distances of 2.972 Å at 290 K and 2.83–3.12 Å at 110 K. Strikingly, the NH4 + cation in the channel shifts about 0.40 Å along the c-axis direction at 110 K. It is the ordering and atomic displacement of organic NH4 + cation that contribute to the ferroelectricity. Temperature-dependent dielectric constants of (NH4 )[Zn(HCOO)3 ] show the sharp dielectric anomalies along its c-axis (Figure 11.59). By fitting the Curie–Weiss law, the T 0 and Cpara values are estimated to be 181 and 5.4 × 103 K, which suggest the typical order–disorder-type FEP transition. The P–E hysteresis loop affords the Ps value of ∼1.0 μC/cm2 at 163 K (the inset in Figure 11.59). b

b

a

Zn OC

(a)

Figure 11.58

N

H

(b)

Crystal structures of (NH4 )[Zn(HCOO)3 ] at (a) 110 and (b) 290 K.

a

653

11 Relationship Between Structure and Ferroelectric Properties

600

450

P (μC/cm2)

1.2 0.6

163K

f = 100 kHz

0.0

–0.6

εʹ

654

300

–1.2

–20

–10

0

10

20

E (KV/cm)

150

0 50

100

150

200

250

T (K)

Figure 11.59 Temperature dependence of dielectric constant of (NH4 )[Zn(HCOO)3 ] measured along the c-axis. Inset: the P–E hysteresis loop at 163 K.

Ferroelectricity has been well established in the glycine-containing compounds, such as TGS, triglycine selenate, triglycine fluorobenzoate, and glycine phosphite [150–152]. It is interesting that the zwitterions glycine can be used to assemble the ferroelectric MOFs, such as [Ag(NH3 CH2 COO)(NO3 )]. It shows the symmetry-breaking phase transition with PEP P21 /a to the FEP P21 , following the Curie principle (Figure 11.60). Structurally, the Ag+ ions are coordinated by oxygen atoms of glycines to form 2D layers in the ac plane, which are linked together through hydrogen bonds. During its phase transition, the atomic displacement of Ag ions from the coplane with oxygen atoms results in the polarization (Ps = 0.60 μC/cm2 ) [153]. Dielectric constants along the b-direction show a large peak at 218 K (Figure 11.61), and the fitted Curie–Weiss law affords T 0 and Cpara of 218 and 446 K, respectively. Similarly, [Ca(CH3 NH2 CH2 COO)3 Cl2 ] is the ferroelectric that undergoes phase transition from the PEP Pnma to the FEP Pn21 a at 127 K (Figure 11.62). The calcium ions are octahedral coordinated by six oxygen atoms of six sarcosine moieties (Figure 11.63) [154]. Each sarcosine links two Ca2+ ions and forms a 1D chain along 1 – 4

b

0

b

0 Symmetry breaking

Symmetry restoration a

a

P21/a (No. 14)

P21 (No. 4)

Figure 11.60 Symmetry breaking of [Ag(NH3 CH2 COO)(NO3 )] with the change of symmetry elements from 4 (E, C 2 , i, 𝜎 h ) to 2 (E, C 2 ).

11.2 Recent Advance of Molecular Ferroelectrics

3.0 2.5 f = 10 kHz

lg (εʹ)

2.0 1.5 (010) 1.0

(100)

0.5

(001) 50

100

(a)

150

200

250

T (K) 0.7 0.6

Ps (μC/cm2)

0.5 0.4 0.3 0.2 0.1 0.0 90

120

150

(b)

180

210

240

T (K)

Figure 11.61 (a) Temperature dependence of dielectric constants and (b) P s versus T curve of [Ag(NH3 CH2 COO)(NO3 )]. 2121 21 – P –n a– m ¼

Pna21

¼ Symmetry breaking

Pbn21

1 – 4

2121 21 – P –b n– m

Figure 11.62 Symmetry breaking of [Ca(CH3 NH2 CH2 COO)3 Cl2 ] with the change of symmetry elements from 8 (E, C 2 , 2C ′ 2 , i, 𝜎 h , 2𝜎 v ) to 4 (E, C 2 , 2𝜎 v ).

Symmetry restoration ¼

¼

Pnma (No. 62)

Pna21 (No. 33)

the a-axis. The strong N—H· · ·Cl hydrogen bonds are formed between Cl and N atoms with each N atom involving two hydrogen bonds. Notably, the pressure has a great effect on dielectric and ferroelectric properties of [Ca(CH3 Nd2 CH2 COO)3 Cl2 ] and a new phase appears at 5 kbar [155, 156]. When the pressure increases, the dielectric anomaly also increases under higher temperature.

655

656

11 Relationship Between Structure and Ferroelectric Properties

Figure 11.63 Crystal structure of [Ca(CH3 NH2 CH2 COO)3 Cl2 ] at FEP. Green dotted lines represent hydrogen bonds.

N

C

Cl

O Ca

b a

Halogen substitution of Cl by Br/I greatly influences the phase transition properties. For instance, [Ca(CH3 NH2 CH2 COO)3 Br2 ] is an initial ferroelectric but becomes the quantum ferroelectric under hydrostatic pressure [157]. With the increasing of the I− concentration, the T c value also decreases, and the 𝜀0 and Ps peak values decrease rapidly. The effect on dielectric properties for the I-to-Cl substitution is much more significant than that for Br-to-Cl substitution [158]. The hydrate of [Ca{(CH3 )3 NCH2 COO}(H2 O)2 Cl2 ] undergoes sequential structural phase transition [159], which crystallizes in the orthogonal space group Pnma at HTP (>164 K). The Ca2+ ion coordinates with two Cl ions, two water molecules, and two oxygen atoms of betaines to form distorted octahedra (Figure 11.64). The symmetry breaking can be described with an Aizu notation of mmmFmm2, that is, changing from PEP Pnma to FEP P21 ca (Figure 11.65). A series of incommensurate and commensurate phases were observed between 164 and 46 K [160, 161],

H O C

Figure 11.64 Crystal structure of [Ca{(CH3 )3 NCH2 COO}(H2 O)2 Cl2 ] at the PEP. Green dotted lines represent hydrogen bonds.

Ca Cl

N

a c

11.2 Recent Advance of Molecular Ferroelectrics 2121 21 –– P –n m a ¼

Pca21

¼

2121 21 – n– – Pm b

Symmetry breaking

Pbc21

1 – 4

Symmetry restoration

¼

¼

Pnma (No. 62)

Pca21 (No. 29)

Figure 11.65 Symmetry breaking for [Ca{(CH3 )3 NCH2 COO}(H2 O)2 Cl2 ] with the change of symmetry elements from 8 (E, C 2 , 2C ′ 2 , i, 𝜎 h , 2𝜎 v ) to 4 (E, C 2 , 2𝜎 v ).

b

C O

Sr1 Ca1

(a)

a

b a

(b)

Figure 11.66 (a) Crystal structure of Ca2 Sr(CH3 CH2 COO)6 at PEP. The disordered bonds are depicted as gray dotted lines. (b) Coordination geometry of the metal ions. The ethyl group of the propionate is omitted for clarity.

which show the 1D structural modulation. Below 46 K, this compound undergoes an improper FEP transition with the Ps value of 2.5 μC/cm2 . PEP-to-FEP transition from P41 21 2 to P41 was observed in Ca2 Sr(CH3 CH2 COO)6 at 282.6 K. In its paraelectric structure, four disordered methyl groups occupy two equilibrium positions, which become partially ordered along with the atomic displacement of Sr, Ca, and O atoms below T c (Figure 11.66). Temperature dependence of dielectric constants exhibits sharp anomalies along the c-axis of Ca2 Sr(CH3 CH2 COO)6 (Figure 11.67), of which the fitted curve affords T 0 and Cpara values of 278 and 73 K, respectively. Structure analyses reveal that displacement of CaO6 octahedra relative to Sr ion, as well as the movement of propionic acid ions, triggers the PEP-to-FEP transition of Ca2 Sr(CH3 CH2 COO)6 . Ca2 Ba(CH3 CH2 COO)6 undergoes phase transition at 267 K, with the symmetry changing from paraelectric P41 21 2 to ferroelectric P41 . Structurally, six propionate ligands surround Ba2+ ions with the O atoms linking to Ba2+ ions to form a triangular reverse prism. Structural analysis shows that the propionate moieties are disordered, of which the spatial interaction affords the driving force to its phase transitions. The high pressure (9.0 × 107 Pa) can induce the phase transition

657

11 Relationship Between Structure and Ferroelectric Properties

0.3 D

f = 1 kHz

H Ps (μC/cm2)

90

εʹ

60 H 30

0 250

260

270

(a)

280

290

0.2

0.1

0.0 230

300

D

240

250

(b)

T (K)

260

270

280

290

T (K)

Figure 11.67 (a) Temperature dependence of dielectric constant, and (b) P s versus T curve of Ca2 Sr(CH3 CH2 COO)6 (H) and it deuterated form (D).

0.08 120 f = 1 kHz 90 MPa

Pr (μC/cm2)

(100)

80 εʹ

658

40

0 240

(a)

85 MPa

0.06

0.04

0.02

250

260

270 T (K)

280

0.00 230

290

(b)

240

250

260

270

T (K)

Figure 11.68 (a) Temperature dependence of dielectric constant, and (b) P s versus T curve of Ca2 Ba(CH3 CH2 COO)6 .

of Ca2 Ba(CH3 CH2 COO)6 , as verified by the temperature-dependent dielectric constants (Figure 11.68). The isomorphous analog, Ca2 Pb(CH3 CH2 COO)6 , undergoes the second-order phase transition from the paraelectric space group P41 21 2 to ferroelectric space group P41 at 333 K. The symmetry breaking process follows the Curie symmetry principle, that is, P41 belongs to one of the subgroups of its paraelectric space group P41 21 2. In the crystal structures, the ordering of methyl groups accounts for its phase transition as well as the Ps along the c-direction (Figures 11.69 and 11.70). A series of ferroelectrics with the general formula [C(NH2 )3 ][M(H2 O)6 ](XO4 )2 , where C(NH2 )3 is guanidinium, M is trivalent Al, V, Cr, or Ga, and X is S or Se, were reported, lacking the Curie points. For instance, [C(NH2 )3 ][Al(H2 O)6 ](SO4 )2 (GASH) belongs to the triangular space group P31 m at room temperature. The planar guanidinium cation is perpendicular to the c-axis, and Al3+ ion coordinated by six water molecules locates on the threefold axis (Figure 11.71). Water molecules are linked to the sulfate tetrahedra by hydrogen bonding. Three different ions are stacked layer by layer along the c-axis in the order of –Al(H2 O)6 –SO4 –C(NH2 )3 –SO4 –. Dielectric constants of GASH are independent of temperature with the value of ∼6 along the c-axis. The Ps value along the c-axis is 0.35 μC/cm2 . It is interesting

11.2 Recent Advance of Molecular Ferroelectrics

b

b

c

c

Ca O C Pb

(a)

(b)

Figure 11.69 (a) Crystal structure of Ca2 Pb(CH3 CH2 COO)6 at PEP. The disorder bonds are depicted as gray dotted lines. (b) Polyhedra of the metal ions. The ethyl group of the propionate is omitted for clarity. 40 0.16 Annealed

(001)

εʹ

20

f = 10 kHz

Ps (μC/cm2)

30

As-grown 10

0.12 0.08 0.04

0 0.00 150

(a)

200

250

300

T (K)

350

240

260

280

(b)

300

320

T (K)

Figure 11.70 (a) Temperature dependence of dielectric constant of annealed and as-grown crystal and (b) P s versus T curve of Ca2 Pb(CH3 CH2 COO)6 . Figure 11.71 Crystal structure of [C(NH2 )3 ][Al(H2 O)6 ](SO4 )2 at FEP.

a c

N O

C S

Al

that the FEP-to-PEP transition does not occur, until the decomposition temperature of 200 ∘ C. Subsequently, FEP transition of [(CH3 )2 NH2 ][Al(H2 O)6 ](SO4 )2 was reported at 152 K, with the space group changing from P21 /n (PEP) to Pn (FEP). Structurally, the regular octahedron is established by the coordination of Al3+ cation with six water molecules (Figure 11.72). The channel formed by Al(H2 O)6 –SO4 sublattice is occupied by the dimethylammonium cation. Due to the rotation around the axis defined by two carbon atoms, the disordering of organic cation results in four equilibrium

659

11 Relationship Between Structure and Ferroelectric Properties

Figure 11.72 Crystal structure of [(CH3 )2 NH2 ][Al(H2 O)6 ](SO4 )2 at FEP.

c b

S O C

AI

N

1.6 3.0 1.2

f = 1 MHz Ps (μC/cm2)

2.5 lg (εʹ)

660

2.0 1.5

0.8

0.4 1.0 0.5 100

(a)

150

200 T (K)

250

0.0 110

300

(b)

120

130

140

150

T (K)

Figure 11.73 (a) Temperature dependence of dielectric constant and (b) P s versus T curve of [(CH3 )2 NH2 ][Al(H2 O)6 ](SO4 )2 .

positions for NH2 group. This dynamic order–disorder accounts for the phase transition. At 1 MHz, the peak value of dielectric constant is 700 at 160 K (Figure 11.73a) and the Ps value is 1.4 μC/cm2 at 120 K (Figure 11.73b). NMR studies reveal that origin of FEP transition in [(CH3 )2 NH2 ][Al(H2 O)6 ](SO4 )2 is the rotational freezing of dimethylammonium cation, along with the rotation and reorientation of H2 O molecules. The hydrate of [CH3 NH3 ][Al(H2 O)6 ](SO4 )2 ⋅6H2 O undergoes phase transition at 177 K, coupling with the change from paraelectric space group P21 3 to ferroelectric space group P21 . According to the Curie symmetry principle, the paraelectric space group P21 3 contains two maximal nonisomorphic subgroups, R3 and P21 21 21 , while the subgroup of P21 21 21 includes P21 . Hence, the Curie symmetry principle is obeyed (Figure 11.74). Structurally, six water molecules surround the Al3+ ion to form a closely octahedral arrangement (Figure 11.75). Sulfate tetrahedron, methylammonium, and the Al(H2 O)6 octahedron connect with water molecules. It exhibits a discontinuity of the dielectric constant at T c , suggesting the first-order phase transition (Figure 11.76a). The Curie–Weiss law is fitted to a narrower temperature range (6 K), resulting in T 0

11.2 Recent Advance of Molecular Ferroelectrics

Figure 11.74 The proposed symmetry breaking for [CH3 NH3 ] [M(H2 O)6 ](SO4 )2 ⋅6H2 O with the change of symmetry elements from (a) 12(E, 8C 3 , 3C 2 ) to 2(E, C 2 ).

1 – 4

1 – 4

a

0 Symmetry breaking 1 – 4

1 – 4

Symmetry restoration c 1 – 4

1 – 4

P213 (No. 198)

P21 (No. 4)

a

b c

S

O N C

AI

Figure 11.75

Unit cell of [CH3 NH3 ][Al(H2 O)6 ](SO4 )2 ⋅6H2 O at PEP. 1.2

60

f = 10 kHz (100)

0.9 Ps (μC/cm2)

εʹ

45

30

15

0.3

0 100

(a)

0.6

150

200 T (K)

250

0.0 160

300

(b)

165

170

175

T (K)

Figure 11.76 (a) Temperature dependence of dielectric constant and (b) P s versus T curve of [CH3 NH3 ][Al(H2 O)6 ](SO4 )2 ⋅6H2 O.

and Cpara of 168.5 and 500 K, respectively. The Ec increases rapidly with the temperature decreasing, and the saturated P–E hysteresis loop can only be observed in a temperature range of 15 K (Figure 11.76b). The hybrid compounds of [H2 dbco][Cu(X2 O)6 ](SeO4 )2 (H2 dbco = diprotonated 1,4-diazabicyclo[2.2.2]octane; X = H or D) were reported as new ferroelectrics. At PEP, the crystal belongs to the nonpolar space group P21 /c, and the H2 dbco cations

661

662

11 Relationship Between Structure and Ferroelectric Properties

b a

Se O N

b C Cu

(a)

c

(b)

Figure 11.77 Unit cell of [H2 dbco][Cu(H2 O)6 ](SeO4 )2 at (a) FEP and (b) PEP. Green dotted lines represent hydrogen bonds.

are severely disordered (Figure 11.77). Upon cooling down to 133 K, the ordering of H2 dbco cations leads to the disappearance of mirror plane, as well as the FEP with polar space group P21 . The measured P–E hysteresis loop affords the Ps value of 1.51 μC/cm2 along the b-axis direction, solidly confirming its ferroelectricity.

11.3 Conclusion and Perspective The exploration and design of ferroelectrics are of booming importance not only in the academic science, but also in the practical technological device applications. The assembly of multicomponent molecular compounds involves more opportunities for exploring ferroelectrics than the single-component counterparts. The subtle balance between molecules is an essential issue for crystal engineering, ultimately focusing on structural phase transitions, dielectric and ferroelectric attributes. However, it is challenging to predict the occurrence of phase transitions and ferroelectricity at this stage, due to the lack of in-depth understanding on ferroelectricity origin. In this context, the exploration of distinguishing ferroelectrics still needs multidisciplinary studies. It is known that the Curie symmetry principle establishes the group–subgroup relationship between PEP and FEP, while the Neumann principle defines relationship between macrophysical properties and macroscopic symmetry. Therefore, based on the Curie symmetry principle and Neumann principle, ferroelectrics can be expected by combining the Cambridge Structural Database. For instance, if a compound crystallizes in one of the 10 polar point groups without dielectric anomalies, it can identified as piezoelectric or pyroelectric rather than ferroelectric. Otherwise, in the case of sharp dielectric anomaly, the possibility of a PEP-to-FEP transition should not be ignored. Besides, temperature dependence of SHG effect would be effective to confirm symmetry breaking and ferroelectricity. If phase transition belongs to 1 of 88 paraelectric–ferroelectric types, furthermore, physical measurements, e.g. the P–E hysteresis loop and domain motions, are

References

needed to confirm the ferroelectricity. Of course, there are exceptions to the aforementioned rule. For example, the aforementioned procedure is no longer used for ferroelectrics without the Curie temperature. Fortunately, depending on the piezoelectric response force microscope (PFM), the polarization reversal and ferroelectric domains can be observed directly. Once the target compound proves to be a ferroelectric, it can be employed as the template to explore analog ferroelectrics through the molecular design or crystal engineering. It is expected that this targeted research for molecular ferroelectrics will possibly promote their advance in molecular recognition, domain engineering, molecular catalysis, and biocompatible electronic devices. The unique advantages of molecular ferroelectrics, including lightweight, structural tunability, environmental friendliness, mechanical flexibility, ease of low-temperature processing, and biocompatibility, make molecular systems as potential alternatives or supplements to inorganic counterparts. It is no doubt that, with the ongoing advance of molecular ferroelectric, it is possible to envision their future use as indispensable promising candidates.

References 1 (a) Zyss, J. (1994). Molecular Nonlinear Optics: Materials, Physics, and Devices. San Diego, CA, USA: Academic Press. (b) Agullo-Lopez, F., Cabrera, J.M., and Agullo-Ra, F. (1994). Electrooptics: Phenomena, Materials and Applications. London: Academic Press. (c) Desiraju, G.R. (1989). Crystal Engineering: the Design of Organic Solids. Amsterdam, The Netherlands: Elsevier. (d) Lines, M.E. and Glass, A.M. (2001). Principles and Applications of Ferroelectrics and Related Materials. Clarendon, Oxford: Oxford University Press. (e) Uchino, K. (2000). Ferroelectric Devices. New York, NY: Marcel Dekker. (f) Dawber, M., Rabe, K.M., and Scott, J.F. (2005). Rev. Mod. Phys. 77: 1083. (g) Piecha, A., ´ Bialonska, A., and Jakubas, R. (2012). J. Mater. Chem. 22: 333. 2 Clark, J.B., Hastie, J.W., Kihlborg, L.H.E. et al. (1994). Pure Appl. Chem. 66: 577. 3 Izyumov, Y.A. and Syromyatnikov, V.N. (1990). Phase Transitions and Crystal Symmetry. Dordrecht, The Netherlands: Kluwer Publishers. 4 Landau, L.D., Lifshitz, E.M., and Reichl, L.E. (1981). Phys. Today 34: 74. 5 Tokunaga, M. and Mitsui, T. (1976). Ferroelectrics 11: 451. 6 Mitsui, T., Nakamnura, E., and Tokunaga, M. (1973). Ferroelectrics 5: 185. 7 Haas, C. (1965). Phys. Rev. 140: A863. 8 Cross, L.E. (1987). Ferroelectrics 76: 241. 9 Cochran, W. (1959). Phys. Rev. Lett. 3: 412. 10 Cochran, W. (1960). Adv. Phys. 9: 387. 11 Ma, M.M., Guo, L., Anderson, D.G., and Langer, R. (2013). Science 339: 186. 12 Fu, D.W., Cai, H.L., Li, S.H. et al. (2013). Phys. Rev. Lett. 110: 257601. 13 de Gennes, P.G. (1963). Solid State Commun. 1: 132. 14 Brout, R., Muller, K.A., and Thomas, H. (1966). Solid State Commun. 4: 507. 15 Gills, N.S. (1975). Phys. Rev. B 11: 309.

663

664

11 Relationship Between Structure and Ferroelectric Properties

16 Aubry, S.J. (1975). J. Chem. Phys. 62: 3217. 17 Stamenkovic, S., Plakide, N.M., Aksienov, V.L., and Siklos, T. (1976). Ferroelectrics 14: 655. ´ 18 Katrusiak, A. and Szafranski, M. (1999). Phys. Rev. Lett. 82: 576. 19 Fu, D.W., Song, Y.M., Wang, G.X. et al. (2007). J. Am. Chem. Soc. 129: 5346. 20 Wuttig, M. and Yamada, N. (2007). Nat. Mater. 6: 824. 21 (a) Lock, P.J. (1971). Appl. Phys. Lett. 19: 390. (b) Horiuchi, S., Kumai, R., and Tokura, Y. (2005). J. Am. Chem. Soc. 127: 5010. 22 Li, S.G., Luo, J.H., Sun, Z.H. et al. (2013). Cryst. Growth Des. 13: 2675. 23 Shi, P.P., Tang, Y.Y., Li, P.F. et al. (2016). Chem. Soc. Rev. 45: 3811. 24 Tagantsev, A.K., Cross, L.E., and Fousek, J. (2010). Domains in Ferroic Crystals and Thin Films. New York, NY: Springer. 25 Franken, P.A., Hill, A.E., Peters, C.W., and Weinreich, G. (1961). Phys. Rev. Lett. 7: 118. 26 Eckhardt, G., Hellwarth, R.W., McClung, F.J. et al. (1962). Phys. Rev. Lett. 9: 455. 27 Garmire, E., Townes, F., and Pandarese, C.H. (1963). Phys. Rev. Lett. 11: 160. 28 Chiao, R.Y., Stoicheff, B.P., and Townes, C.H. (1964). Phys. Rev. Lett. 12: 592. 29 Szoke, A. and Javan, A. (1963). Phys. Rev. Lett. 10: 521. 30 Kosonocky, W.F. and Harrison, S.E. (1966). J. Appl. Phys. 37: 4789. 31 Coe, B.J. (1999). Chem. Eur. J. 5: 2464. 32 Castet, F., Rodriguez, V., Pozzo, J.L. et al. (2013). Accounts Chem. Res. 46: 2656. 33 Hugel, T. (2002). Science 296: 1103. 34 Waldeck, D.H. (1991). Chem. Rev. 91: 415. 35 Whatmore, R.W. (1986). Rep. Prog. Phys. 49: 1335. 36 Sawyer, C.B. and Tower, C.H. (1930). Phys. Rev. 35: 269. 37 Arlt, G. (1990). J. Mater. Sci. 25: 2655. 38 Fousek, J. and Janovec, V. (1969). J. Appl. Phys. 40: 135. 39 Furukawa, T., Date, M., and Fukada, E. (1980). J. Appl. Phys. 51: 1135. 40 Noda, K., Ishida, K., Kubono, A. et al. (2003). J. Appl. Phys. 93: 2866. 41 Largerwall, S.T. (2004). Ferroelectrics 301: 15. 42 Solomon, A.L. (1956). Phys. Rev. 104: 1191. 43 Goldsmith, G.J. and White, J.G. (1959). J. Chem. Phys. 31: 1175. 44 Bordeaux, D., Bornarel, J., Capiomont, A. et al. (1973). Phys. Rev. Lett. 31: 314. 45 Lipscomb, G.F., Garito, A.F., and Wei, T.S. (1980). Ferroelectrics 23: 161. 46 Schultes, H., Strohriegl, P., and Dormann, E. (1986). Ferroelectrics 70: 161. 47 Gruner-Bauer, P. and Dormann, E. (1992). J. Phys. Condens. Matter 4: 5599. 48 Choudhury, R.R. and Chitra, R. (2006). Cryst. Res. Technol. 41: 1045. 49 Zikmund, Z., Vanˇek, P., Havránková, M. et al. (1994). Ferroelectrics 158: 223. 50 Kroupa, J., Vanˇek, P., Krupková, R., and Zikmund, Z. (1997). Ferroelectrics 202: 229. 51 Koval, S., Kohanoff, J., Lasave, J. et al. (2005). Phys. Rev. B 71: 184102. 52 Szklarz, P. and Bator, G. (2005). J. Phys. Chem. Solids 66: 121. 53 Yamamura, Y., Saitoh, H., Sumita, M., and Saito, K. (2007). J. Phys. Condens. Matter 19: 176219.

References

54 Bator, G., Jakubas, R., and Malarski, Z. (1986). J. Phys. C: Solid State Phys. 19: 2799. 55 Semmingsen, D. and Feder, J. (1974). Solid State Commun. 15: 1369. 56 Feder, J. (1976). Ferroelectrics 12: 71. 57 Ishii, F., Nagaosa, N., Tokura, Y., and Tearkura, K. (2006). Phys. Rev. B 73: 212105. 58 Sugawara, T. and Takasu, I. (1999). Adv. Phys. Org. Chem. 32: 219. 59 Gilli, G., Bellucci, F., Ferretti, V., and Bertolasi, V. (1989). J. Am. Chem. Soc. 111: 1023. 60 Katrusiak, A. (1992). J. Mol. Struct. 269: 329. 61 Sugawara, T., Mochida, T., Miyazaki, A. et al. (1992). Solid State Commun. 83: 665. 62 Mochida, T., Izuoka, A., Sugawara, T. et al. (1994). J. Chem. Phys. 101: 7971. 63 Takasu, I., Izuoka, A., Sugawara, T., and Mochida, T. (2004). J. Phys. Chem. B 108: 5527. ´ 64 Katrusiak, A. and Szafranski, M. (2006). J. Am. Chem. Soc. 128: 15775. 65 Tokura, Y., Koshihara, S., Iwasawa, N., and Saito, G. (1989). Phys. Rev. Lett. 63: 2405. 66 Okamoto, H., Mitani, T., Tokura, Y. et al. (1991). Phys. Rev. B 43: 8224. 67 Le Cointe, M., Lemee-Cailleau, M.H., Cailleau, H. et al. (1995). Phys. Rev. B 51: 3374. 68 Garcia, P., Dahaoui, S., Fertey, P. et al. (2005). Phys. Rev. B 72: 104115. 69 Torrance, J.B., Vazquez, J.E., Mayerle, J.J., and Lee, V.Y. (1981). Phys. Rev. Lett. 46: 253. 70 Torrance, J.B., Girlando, A., Mayerle, J.J. et al. (1981). Phys. Rev. Lett. 47: 1747. 71 Horiuchi, S., Okimoto, Y., Kumai, R., and Tokura, Y. (2000). J. Phys. Soc. Jpn. 69: 1302. 72 Okimoto, Y., Horiuchi, S., Saitoh, E. et al. (2001). Phys. Rev. Lett. 87: 187401. 73 Horiuchi, S., Okimoto, Y., Kumai, R., and Tokura, Y. (2003). Science 299: 229. 74 Nad, F. and Monceau, P. (2006). J. Phys. Soc. Jpn. 75: 051005. 75 Girlando, A., Pecile, C., and Torrance, J.B. (1985). Solid State Commun. 54: 753. 76 Horiuchi, S., Kumai, R., Okimoto, Y., and Tokura, Y. (1999). J. Am. Chem. Soc. 121: 6757. 77 Tokura, Y., Okamoto, H., Koda, T. et al. (1988). Phys. Rev. B 38: 2215. 78 Horiuchi, S., Ishii, F., Kumai, R. et al. (2005). Nat. Mater. 4: 163. 79 Kumai, R., Horiuchi, S., Okimoto, Y., and Tokura, Y. (2006). J. Chem. Phys. 125: 084715. 80 Horiuchi, S., Kumai, R., and Tokura, Y. (2007). Angew. Chem. Int. Edit. 46: 3497. 81 Zaman, M.B., Tomura, M., and Yamashita, Y. (2001). J. Org. Chem. 66: 5987. 82 Asaji, T., Seliger, J., Zagar, V. et al. (2007). J. Phys. Condens. Matter 19: 226203. 83 Cohen, R.E. (1992). Nature 358: 136. 84 Saitoh, S., Kobayashi, T., Harada, K. et al. (1998). IEEE Trans. Ultrason. Ferroelectr. Freq. Control 45: 1071. 85 Iwata, M., Araki, T., Maeda, M. et al. (2002). Jpn. J. Appl. Phys. 41: 7003.

665

666

11 Relationship Between Structure and Ferroelectric Properties

86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120

Lim, L.C. and Rajan, K.K. (2004). J. Cryst. Growth 271: 435. Park, S.E. and Shrout, T.R. (1997). J. Appl. Phys. 82: 1804. Dong, M. and Ye, Z.G. (2000). J. Cryst. Growth 209: 81. Long, X. and Ye, Z.G. (2007). Acta Mater. 55: 6507. Yamashita, Y. and Shimanuki, S. (1996). Mater. Res. Bull. 31: 887. Bing, Y.H. and Ye, Z.G. (2003). J. Cryst. Growth 250: 118. Bing, Y.H. and Ye, Z.G. (2006). J. Cryst. Growth 287: 326. Yasuda, N., Ohwa, H., Kume, M. et al. (2001). Jpn. J. Appl. Phys. 40: 5664. Zhang, S., Rehrig, P.W., Randall, C., and Shrout, T.R. (2002). J. Cryst. Growth 234: 415. Liu, Y., Li, X.Z., Wang, Z.J. et al. (2013). CrystEngComm 15: 1643. Müller, J., Böscke, T.S., Schröder, U. et al. (2012). Nano Lett. 12: 4318. Zhang, W. and Xiong, R.G. (2012). Chem. Rev. 112: 1163. Sawada, A., Sugiyama, J., Wada, M., and Ishibashi, Y. (1980). J. Phys. Soc. Jpn. 48: 1773. Gesi, K. and Iizumi, M. (1980). J. Phys. Soc. Jpn. 48: 775. Vasil’ev, V.E., Rudyak, V.M., Bobrova, Z.A., and Varikash, V.M. (1987). Fiz. Tverd. Tela 29: 1539. Bobrova, Z.A. and Varikash, V.M. (1986). Dokl. Akad. Nauk BSSR 30: 510. Bobrova, Z.A., Varikash, V.M., Baranov, A.I., and Shuvalov, L.A. (1987). Kristallografiya 32: 255. Czapla, Z., Eliyashevskyy, Y., and Dacko, S. (2006). Ferroelectr. Lett. Sect. 33: 1. Daoud, A. (1977). J. Appl. Crystallogr. 10: 133. Perez-Mato, J.M., Manes, J.L., Fernandez, J. et al. (1981). Phys. Status Solidi A 68: 29. Priya, R., Krishnan, S., Bhagavannarayana, G., and Jerome Das, S. (2011). Physica B 406: 1345. Knorr, K., Jahn, I.R., and Heger, G. (1974). Solid State Commun. 15: 231. Kind, R., Blinc, R., and Zeks, B. (1979). Phys. Rev. B 19: 3743. Chapuis, G., Arend, H., and Kind, R. (1975). Phys. Status Solidi A 31: 449. Steadman, J.P. and Willett, R.D. (1970). Inorg. Chim. Acta 4: 367. Kundys, B., Lappas, A., Viret, M. et al. (2010). Phys. Rev. B 81: 224434. De Jongh, L.J., Botterman, A.C., de Boer, F.R., and Miedema, A.R. (1969). J. Appl. Phys. 40: 1363. De Jongh, L.J., Van Amstel, W.D., and Miedema, A.R. (1972). Physica 58: 277. Sun, Z., Zeb, A., Liu, S. et al. (2016). Angew. Chem. Int. Ed. 55: 11854. Xu, W.J., He, C.T., Ji, C.M. et al. (2016). Adv. Mater. 28: 5886. Li, P.F., Tang, Y.Y., Liao, W.Q. et al. (2017). NPG Asia Mater. 9: e342. Aguirre-Zamalloa, G., Madriage, G., Couzi, M., and Breczewski, T. (1993). Acta Crystallogr., Sect. B: Struct. Sci. 49: 691. Asahi, T., Hasebe, K., and Gesi, K. (1991). Acta Crystallogr., Sect. C: Cryst. Struct. Commun. 47: 1208. Fatuzzo, E. (1960). Proc. Phys. Soc. 76: 797. Fatuzzo, E., Nitsche, R., Roetschi, H., and Zingg, S. (1962). Phys. Rev. 125: 514.

References

121 Arend, H., Ehrensperger, M., Muralt, P. et al. (1982). Ferroelectronics Lett. 44: 147. 122 Bi, W., Leblanc, N., Mercier, N. et al. (2009). Chem. Mater. 21: 4099. 123 Jakubas, R. (1986). Solid State Commun. 60: 389. 124 Latanowicz, L., Medycki, W., and Jakubas, R. (2005). J. Phys. Chem. A 109: 3097. 125 Zaleski, J., Pawlaczyk, C., Jakubas, R., and Unruh, H.G. (2000). J. Phys. Condens. Matter 12: 7509. 126 Jakubas, R., Czapla, Z., Galewski, Z., and Sobczyk, L. (1986). Ferroelectr. Lett. 5: 143. 127 Wojtas, M., Bator, G., Jakubas, R., and Zaleski, J. (2003). J. Phys. Condens. Matter 15: 5765. 128 Jakubas, R., Sobczyk, L., and Lefebvre, J. (1989). Ferroelectrics 100: 143. 129 Jakubas, R. (1989). Solid State Commun. 69: 267. 130 Szklarz, P., Galazka, M., Zielinski, P., and Bator, G. (2006). Phys. Rev. B 74: 184111. 131 Li, L., Sun, Z., Wang, P. et al. (2017). Angew. Chem. Int. Ed. 56: 12150. 132 Valasek, J. (1921). Phys. Rev. 17: 475. 133 Solans, X., Gonzalez-Silgo, C., and Ruiz-Pérez, C. (1997). J. Solid State Chem. 131: 350. 134 Kikuta, T., Shimizu, K., Nozaki, R., and Shiozaki, Y. (1997). Ferroelectrics 190: 51. 135 Jona, F. and Pepinsky, R. (1953). Phys. Rev. 92: 1577. 136 Kamba, S., Brezina, B., Petzelt, J., and Schaack, G. (1996). J. Phys. Condens. Matter 8: 8669. 137 Abe, R. and Matsuda, M. (1974). J. Phys. Soc. Jpn. 37: 437. 138 Matthias, B.T. and Hulm, J.K. (1951). Phys. Rev. 82: 108. 139 Fousek, J., Cross, L.E., and Seely, K. (1970). Ferroelectrics 1: 63. 140 Cornia, A., Caneschi, A., Dapporto, P. et al. (1999). Angew. Chem. Int. Ed. 38: 1780. 141 Wang, Z., Zhang, B., Fujiwara, H. et al. (2004). Chem. Commun. (4): 416. 142 Cui, H.B., Wang, Z., Takahashi, K. et al. (2006). J. Am. Chem. Soc. 128: 15074. 143 Jain, P., Ramachandran, V., Clark, R.J. et al. (2009). J. Am. Chem. Soc. 131: 13625. 144 Jain, P., Dalal, N.S., Toby, B.H. et al. (2008). J. Am. Chem. Soc. 130: 10450. 145 Lee, J.H., Fang, L., Vlahos, E. et al. (2010). Nature 466: 954. 146 Sánchez-Andújar, M., Presedo, S., Yáez-Vilar, S. et al. (2010). Inorg. Chem. 49: 1510. 147 Wang, X.Y., Gan, L., Zhang, S.W., and Gao, S. (2004). Inorg. Chem. 43: 4615. 148 Wang, Z., Zhang, B., Otsuka, T. et al. (2004). Dalton Trans. 4: 2209. 149 Xu, G.C., Ma, X.M., Zhang, L. et al. (2010). J. Am. Chem. Soc. 132: 9588. 150 Pepinsky, R., Vedam, K., Hoshino, S., and Okaya, Y. (1958). Phys. Rev. 111: 430. 151 Pepinsky, R., Okaya, Y., Eastman, D.P., and Mitsui, T. (1957). Phys. Rev. 107: 1538. 152 Brennan, J.C. (1992). Ferroelectrics 132: 245.

667

668

11 Relationship Between Structure and Ferroelectric Properties

153 Rao, J.K.M. and Viswamitra, M.A. (1972). Acta Crystallogr., B 28: 1481. 154 Mishima, N., Itoh, K., and Nakamura, E. (1984). Acta Crystallogr., Sect. C: Cryst. Struct. Commun. 40: 1824. 155 Fujimoto, S., Yasuda, N., and Kameyama, S. (1980). J. Phys. D: Appl. Phys. 13: L95. 156 Fujimoto, S., Yasuda, N., Takagi, K. et al. (1980). J. Phys. D: Appl. Phys. 13: L217. 157 Hikita, T. and Maruyama, T. (1992). J. Phys. Soc. Jpn. 61: 2840. 158 Fujimoto, S., Yasuda, N., and Funado, S. (2000). J. Phys. D: Appl. Phys. 15: 1469. 159 Rother, H.J., Albers, J., and Klöpperpieper, A. (1984). Ferroelectrics 54: 107. 160 Chaves, M.R., Almeida, A., Simeo Carvalho, P. et al. (1991). Phys. Rev. B 43: 11162. 161 Brill, W., Gmelin, E., and Ehses, K.H. (1990). Ferroelectrics 103: 25.

669

12 The Relationship Between Structure and Electric Property Guan-E Wang 1 , Yonggang Zhen 2 , Guo-Dong Wu 1 , Huanli Dong 2 , and Gang Xu 1 1 Chinese Academy of Sciences, Fujian Institute of Research on the Structure of Matter, State Key Laboratory of Structural Chemistry, 155 Yangqiao West Road, Fuzhou, Fujian 350002, P.R. China 2 Chinese Academy of Sciences, Institute of Chemistry, Beijing National Laboratory for Molecular Sciences, Key Laboratory of Organic Solids, Beijing 100190, P.R. China

12.1 Introduction Conductive materials are available as adhesives, compounds, encapsulants, gels, thermal pads, and gap fillers. All of these materials are available in a wide range of delivery formats, viscosities, cure chemistries, and thermal and mechanical properties to help draw out heat in a broad variety of industries and applications. Electrically conductive materials combine exceptional mechanical and conductive properties shielding for today’s most demanding applications in these markets. Electronic devices based on conductive materials, such as integrated circuits and solar cells, are closely related to life. Electrically conductive materials, such as metals, have delocalized electrons that can easily move along a potential gradient. Electrical conductivity in metals is a result of the movement of electrically charged particles. The atoms of metal elements are characterized by the presence of valence electrons, which are electrons in the outer shell of an atom that are free to move about. It is these “free electrons” that allow metals to conduct an electric current. Because valence electrons are free to move they can travel through the lattice that forms the physical structure of a metal. Under an electric field, free electrons move through the metal much like billiard balls knocking against each other, passing an electric charge as they move. Electrons in the covalent bonds of organic molecules such as polypropylene must remain near their host atoms and are not free to move through the material. As a result, they are poor conductors of electricity. As for conductive materials, they can be divided into inorganic conductive materials, organic conductive materials, and inorganic–organic hybrid conductive materials. The objectives of this part are to summarize the recent literature describing the properties of conductive materials and the relationship between structure and conductivity, to elaborate their use in electronic devices and to assess the current status of the research concerning their integration with other materials Advanced Structural Chemistry: Tailoring Properties of Inorganic Materials and their Applications, First Edition. Edited by Rong Cao. © 2021 WILEY-VCH GmbH. Published 2021 by WILEY-VCH GmbH.

670

12 The Relationship Between Structure and Electric Property

to create devices with novel capabilities. In this chapter, the relationship between structure and electrical conductivity will be discussed according to the composition of elements, including the inorganic type, the organic type, and inorganic–organic hybrids.

12.2 Structure and Electrical Properties of Inorganic Conductive Materials Inorganic conducting materials are the earliest emerging conductive materials and now become one of the most crucial functional materials. Classic electronic theories are generally based on the research of inorganic conducting materials which will be introduced in this part, but metal materials are not included, for the metal material itself is a large class of the material system and thus a variety of related books can be found. Of course, this section also does not comprise carbon materials, such as graphene and carbon nanotube. Although these novel materials attract great attention and have a wide range of applications, there is also some special literature on carbon materials. Here, inorganic conducting materials mainly involve conventional inorganic semiconductor materials which play an important role in the theoretical construction of electronic materials and devices. By studying the relationship between the structure and properties of these inorganic semiconductor materials, we can gain a deeper understanding of the electrically conducting mechanism. The electronic conduction mechanism lays a solid theoretical foundation for designing semiconductor devices with better performance. These inorganic conductive materials incorporate elemental conducting materials, transparent conducting oxides (TCOs), nitride conducting materials, carbide conducting materials, and sulfide conductive materials, which will be described in detail below.

12.2.1 Elemental Conducting Materials 12.2.1.1 General Introduction of Elemental Conducting Materials

In terms of the scope, element conductive materials basically include metal elements, carbon materials (graphene, carbon nanotube, the fullerene, etc.), and elemental semiconductor materials. Here, we merely focus on elemental semiconductor materials. The nature and structure of elemental semiconductor materials are more susceptible to regulation, and thus they can better represent the structure–activity relationship between the structure and properties of element conductive materials. As is known to all, silicon (Si) and germanium (Ge) are the most typical inorganic elemental semiconductors with Si accounting for more than 90% of the industrial semiconductor materials. Si and Ge materials are widely used in electronics, military, aerospace, and other significant fields. Especially, Si is almost the cornerstone of the modern electronic information industry. The single crystals of Si and Ge are both members of the simplest threedimensional (3D) lattice system referred to as the cubic lattice system. Any lattice

12.2 Structure and Electrical Properties of Inorganic Conductive Materials

Figure 12.1

The unit cell and diamond lattice structure for Si and Ge.

system having a cubic volume as a unit cell belongs to the cubic family, as shown in Figure 12.1. The length dimension of the unit cell defines the lattice constant (a). Si and Ge have a cubic lattice structure known as the diamond lattice structure, and the unit cell is actually two interpenetrating face-centered cubic (fcc) lattices separated by a/4 along each axis of the cell. Si and Ge can also be formed into an alloy of Si–Ge with a molecular formula of the form Si1−x Gex . The single structure of Si1−x Gex has the identical diamond lattice structure with Si and Ge. Si–Ge can serve as a semiconductor in integrated circuits for heterojunction bipolar transistors or as a strain-inducing layer for complementary metal oxide semiconductor (CMOS) transistors. The electrically conducting behavior of Si and Ge can be regulated with a wide range by doping strategy. The addition of a small amount of foreign atoms in the crystal lattice of Si or Ge can easily produce dramatic changes in their electrical properties and thus form n-type and p-type semiconductors. For instance, the addition of pentavalent impurities, such as antimony, arsenic, or phosphorus, contributes the form of free electrons, greatly enhancing the conductivity of the intrinsic Si or Ge semiconductor, while the adding of trivalent impurities, such as boron, aluminum, or gallium, can result in deficiencies of valence electrons, called “holes,” as shown in Figure 12.2. The band theory shows that extra levels have been added by the doping of impurities. In n-type Si or Ge materials, there are electron energy levels near the top of the bandgap so that they can be easily excited into the conduction band. In p-type Si or Ge materials, extra holes in the bandgap allow excitation of valence band electrons, leaving mobile holes in the valence band.

671

672

12 The Relationship Between Structure and Electric Property

Donor impurity contributes free electrons

Si

Si

Sb

Si

Si

Figure 12.2

B

Si

Si

Si (a)

Si

Acceptor impurity creates a hole

(b)

The doping mechanism for Si single crystal.

Ge is the material used in the first transistor invented by Bardeen, Brattain, and Shockley at Bell Laboratories in 1947, a historical event in the semiconductor world. With this as a landmark, the semiconductor era started. Nevertheless, Si was soon preferred to Ge for its practical superiority and intrinsic properties and thus became the main material in the semiconductor industry. Si now has taken the lead in the miniaturization and large-scale integration of modern electronic devices. After being ignored for a long period of time, Ge has again become the focus of keen interest since the 1990s for its applications in ultrafast electronic devices and its favorable basic properties of intrinsic carrier mobility. Now, Ge crystals were used more in the limited fields of ray detectors and infrared optics in the basic research community. A number of groups around the world have redeveloped Ge as a material for mainstream semiconductors [1]. 12.2.1.2 Structure and the Property of Elemental Conducting Materials

The industrial processes of single-crystalline Si or Ge and their doping strategy are now very mature. Currently, research focused more on the regulation of the relationship between Si or Ge nano/micro structures and their properties. For example, compared with bulk Si, nanostructured Si possesses unique optical and electrical properties because of its quantum confinement effect. The lack of transitional symmetry in nanostructured Si gives rise to relaxation of momentum conservation and supra indirect-gap absorption. These extraordinary properties endow nanostructured Si with new opportunities for its application in the flexible and portable electronics industry as it is a way of electric power generation in the future. In addition, nanostructured Si shows promising thermoelectric and piezoelectric properties via smart surface modification, providing an alternative avenue of electricity generation through thermal and mechanical energy, respectively. In this respect, Sun et al. have summarized the utilization of nanostructured Si in a range of fields, including flexible solar cells, thermal electricity, and piezoelectric generators [2], as shown in Figure 12.3. Apart from the well-known Si quantum confinement effect, hot luminescence from indirect bandgap Si provides a brand new and promising approach to realize

12.2 Structure and Electrical Properties of Inorganic Conductive Materials

Figure 12.3 Schemes of nanostructured Si used for flexible solar cells, thermal electricity, and piezoelectric generators. Source: Sun et al. [2].

Solar cell

10 mm

150 nm

After first etch

Nanostructured silicon 500 nm

3 μm

Thermoelectric generator

Pizoelectric generator Current (nA)

8 Current versus vibration Sound fork

4 0

–4 –8

0

60 120 180 240 300 Time (s)

monolithically integrated Si optoelectronics thanks to phonon-assisted light emission. Wang and coworkers reported multiband hot photoluminescence generated from Si nanowire arrays by introducing trapezoid-shaped nanocavities that support hybrid photonic–plasmonic modes. By continuously adjusting the geometric parameters of the Si nanowires with trapezoidal nanocavities, they figure out that the multiband hot photoluminescence can be tuned in the range from visible to near infrared regions independent of the excitation of the laser wavelength, as shown in Figure 12.4. The highly tunable wavelength bands and concomitant compatibility with Si-integrated electronics enable tailoring of silicon-based light sources to fit next-generation optoelectronics devices [3]. Besides, geometric and compositional modulations are the essential parameters of control to tailor the band profile in the Ge/Si nanostructure system as well, which has been achieved mainly by alternating the feeding precursors during uniaxial growth of Ge/Si nanowires (NWs). Yu and coworkers reported self-automated growth of Ge/Si hetero island-chain nanowires (hiNWs) (as shown in Figure 12.5). The surface running in droplets enforces circulative hydrodynamics that can modulate the absorption depth into the amorphous bilayer and enable single-run growth of superlattice-like hiNWs without any external manipulation. This simple self-assembly growth and modulation dynamics can assist in establishing a powerful new concept or strategy to tailor and program the geometric and compositional profiles of more intricate hetero nanowire structures, which can be viewed as promising building blocks to develop advanced nanoelectronics or optoelectronics (Figure 12.5) [4].

12.2.2 Transparent Conducting Oxides (TCOs) Traditionally, electrical conductivity and the transparency of materials are inherently contradictory. Transparent materials are normally insulators like SiO2 glasses. Furthermore, highly electrically conductive materials generally reveal the opaque color. For instance, Ag and Cu metals show high-reflectivity metallic colors. However, TCOs are a unique type of semiconducting materials that own both properties

673

12 The Relationship Between Structure and Electric Property

(b)

Dielectrics

AI

Si

Wc

Ag

O

Ag

Wb

Si

Glass sub (a)

(c)

6 × 103

Intensity (a.u.)

674

Photonic Si NW Plasmonic Si NW

585 nm 650 nm

525 nm

525 nm 4 × 103

585 nm

650 nm

2 × 103

0

500

600 700 800 Wavelength (nm)

(d)

900 (e)

Figure 12.4 (a) The schematic of the trapezoidal nanocavity-embedded Si nanowire array, (b) cross-sectional scanning transmission electron microscopy (STEM), (c) enlarged STEM of the nanocavity-embedded Si nanowire array, (d) room-temperature photoluminescence (PL) spectra, and (e) optical and confocal laser fluorescence microscopy images of the Si nanowire array. Source: (e) Mu et al. [3]. Copyright 2017, American Chemical Society. (a) 200 nm

(b)

Si[002] Si[111]

Si[111] Si[002] Si[220]

Growth

2 1 nm

(e)

Viewed@ Spot splitting Si[1–10]

Si[11–1]Si[–1–11]

Si

d=3.22 Å

(c)

(d)

Ge Si[111]

d=3.14 Å

Ge

Si Si[331]

Si Ge

10 nm

Si Ge

10 nm

Si Ge

Growth direction along Si(220)

Figure 12.5 (a) A typical transmission electron microscope (TEM) characterization of an in-plane Ge/Si hiNW, (b) a selected area diffraction (SAD) pattern of the region, (c,d) the red and blue squares on the opposite sides of the phase-separated boundary with closer high-resolution TEM examinations of the lattice fringes in two distinct regions, and (e) the energy dispersive spectrometer (EDS) elemental mapping along the hiNW. Source: Zhao et al. [4]. Copyright 2018, American Chemical Society.

simultaneously. TCOs usually have an intrinsic energy gap larger than 3.0 eV that can be able to keep high optical transmission in the visible range and high electrical conductivity ranging from 10−1 to 104 S/cm [5]. Owing to these outstanding properties, TCOs can be widely used as transparent electrodes in flat panel displays, solar cells, and touch panels. To design for TCOs, it is of necessity to convert

12.2 Structure and Electrical Properties of Inorganic Conductive Materials

CB ns0 ns2

M

Electron

Antibonding (CB)

Eg

Eg

Nonbonding (VB top)

Bandgap

2p4

Bonding (VB bottom)

Hole

O VB 2p6

Figure 12.6 The schematic molecular orbital diagram and band structure of oxides. Source: Hosono and Ueda [6]. Copyright 2017, Springer, New York, NY.

transparent insulators into highly conducting semiconductors by increasing electrical conductivity without coloration. Therefore, understanding of the electronic structures of basic metal oxides (MOs) is inevitable for the materials design for TCOs. The representative electronic structure of MOs illustrated in the form of molecular orbitals and energy bands is shown in Figure 12.6. The conduction band is made up of empty metal ns0 orbitals, and the valence band consists of occupied oxygen 2p orbitals. The conduction band has antibonding features, and the bottom of the valence band has bonding traits as a counterpart. The top of the valence band (TVB) has nonbonding features because nonbonding oxygen 2p orbitals remain without varying the energies [6]. Up to now, most of the emerging TCOs have been n-type semiconductors. Compared with n-type TCOs, p-type TCOs are far behind the n-type material system in performances and applications. This is due to the electronic structure and energy band structure of MOs: metal atoms in MOs bond with oxygen atoms in the form of an ionic bond, and 2p (O) energy level is much lower than metal’s valence band electron level. By reason of the strong electronegativity of oxygen ions, the hole at the TVB has a strong localized binding effect. Hence, even if the hole is introduced at the TVB, it will form a deeply dominant energy level, which makes it hard for hole carriers to move in the materials [7]. 12.2.2.1 Materials Designed for n-Type TCOs

For n-type TCOs, a conduction band involving empty orbitals sets up a conduction path for electrons. To obtain the sufficient overlap of orbitals for the conduction path, cations’ orbitals should be extended as widely as possible and intercation distances should be as short as possible. In consequence, the essential principle for designing n-type TCOs is to choose suitable cations for widely extended orbitals, and at the same time to select a proper crystal structure for shorter intercation distances [8].

675

676

12 The Relationship Between Structure and Electric Property

Corner sharing octahedra

Long

Figure 12.7 Configurations of octahedra in crystal structures. Source: Hosono and Ueda [6]. Copyright 2017, Springer, New York, NY.

Edge sharing octahedra Intercation distance

Face sharing octahedra

Short

For the selection of cations, p-block atoms are given priority to form widely extended orbitals. It is known to all that radii of orbitals are proportional to the principal quantum number n. Although heavy atoms with large n are preferable, atoms with n = 6, such as Tl and Pb are inadequate because their orbitals are too widely extended to keep the transparency of the wide bandgap. Transition metal (TM) ions with dn open shells are also improper as the result of the coloration. In addition to the radii of orbitals, the shape of orbitals is also pivotal to maximize the orbital overlap. Isotropic orbitals, which are free from bonding direction, are more preferable than p- or d orbitals. Therefore, cations with the d10 ns0 (n = 4, 5) electronic configuration are favorable to form the conduction path for electrons in the conduction band. Actually, the selected cations include cations, such as Zn2+ , In3+ , and Sn4+ , found in major TCOs. In the choice of crystal structures, the configuration of polyhedra containing a cation at the center and oxygen ions at the apexes should be examined. In the case of the cations with d10 ns0 (n = 4, 5) like In3+ and Sn4+ , they usually form oxygen octahedra due to the size of the ions. Various configurations, such as isolated octahedra, corner-sharing octahedra, edge-sharing, and face-sharing octahedra, are seen in crystal structures. Among them, intercation distances are the shortest in face-sharing structures as shown in Figure 12.7. It is known, however, that such structures are generally unstable in oxides owing to the intercation coulomb repulsion [9] and hardly seen in many crystal structures. Therefore, the edge-sharing structure substantially gives the shortest intercation distances. Accordingly, with widely extended orbitals, edge-sharing structures can form conduction paths in n-type TCOs. 12.2.2.2 Typical n-Type TCOs and Their Electronic Structure

From the aforementioned design strategy, MOs, which possess the cations with d10 ns0 (n = 4, 5) electronic structure and the partial crystal structures with

12.2 Structure and Electrical Properties of Inorganic Conductive Materials

O

O M

O

O M

O

e– O

M O

O M

O

O M

O

O Spinel (Cdln2O4, Mgln2O4)

Rutile (SnO2, TiO2) (a)

(b)

Figure 12.8 Crystal structures of typical n-type TCOs. (a) SnO2 , TiO2 , (b) MgIn2 O4 , CdIn2 O4 . Source: Hosono and Ueda [6]. Copyright 2017, Springer, New York, NY.

edge-sharing octahedra, are leading candidate materials for n-type TCOs. As shown in Figure 12.8, SnO2 and In2 O3 , known as n-type TCOs for a long time, have rutile structure consisted of edge-sharing octahedra along the c-axis. Their doping forms, such as CdIn2 O4 and MgIn2 O4 , have spinel structure and three-dimensionally connected edge-sharing octahedra [10]. The effective ionic radii of In3+ and Sn4+ are large and the intercation distances in the oxides approach the interatomic distances in the metals to a large extent; it would be easily supposed that 5s0 orbitals between In3+ and Sn4+ ions in the oxides can largely overlap as in the metals. The large overlap of the orbitals in In2 O3 and SnO2 appears as large energy dispersion of the conduction band, where the valence band maximum (VBM) is set to zero in energy. The large energy dispersion contributes to the large curvature and mobility at the bottom of the conduction band. Since the Fermi energy Ef is taken as zero in energy, Ef is located at the bottom of the conduction band, indicating these materials are n-type TCOs. The large dispersion of the conduction band brings about a gentle increase of the density of states, which appears as a gradual increase in the photoemission intensity above Ef . The gradual increase at the bottom of the conduction band is in contrast to the steep increase at the TVB, especially in In2 O3 . The characteristics of the crystal and electronic structures preferable to n-type TCOs are commonly seen in other n-type TCOs, such as ZnO, Ga2 O3 , Cd2 SnO4 , and SrGeO3 . The edge-sharing octahedra in crystal structures are not always essential if the overlap of the orbitals for the conduction band is sufficiently large to form a conduction path. Therefore, even amorphous oxides like a-In2 O3 –ZnO and a-In2 O3 –Ga2 O3 –ZnO can be n-type TCOs because the conduction paths are formed by the orbitals’ overlap in these materials irrespective of the crystal structures [11]. Since TCOs in amorphous crystals can be formed at low temperatures and are

677

12 The Relationship Between Structure and Electric Property Ba

Ba0.92La0.08SnO3–δ films Sn

O a AO(0001)/1000(N2)

(200)

LAO(111)/1000(N2) (002)(002) (002) (002)

(110) (110) (110)

(001) (001) (001)

b

STO(0001)/1000(N2)

(111)

(001)

(002)

(110)

(002)

c

Intensity (a.u.)

678

LAO(110)/1000(N2) LAO(001)/1000(N2) LAO(001)/1000(air) LAO(001)/1000(O2) LAO(001)/800(air)

20

30

40

50 2θ (°)

60

70

80

Figure 12.9 X-ray diffraction (XRD) patterns for all solution-derived BLSO films, and a unit cell for Bismuth silicon oxide (BSO) given by the inset. Source: Wei et al. [13]. Copyright 2015, American Institute of Physics.

durable during bending, they can be deposited not only on rigid glasses but also in flexible substrates. Their advantages are utilized in growing applications, such as flexible electronics and printable electronics. Some TCOs like TiO2 and KTaO3 have the conduction bands that are not composed of s0 orbitals but d0 orbitals [12]. These materials are deemed to have conduction paths incorporating the d0 orbitals. As the d0 orbitals are not isotropic in shape, the electrical conductivity would be largely influenced by the symmetry of the crystals. As a type of perovskite TCOs , La-doped BaSnO3 is perceived as an all-important material to construct all transparent perovskite devices. Sun and coworkers have reported La doping and film dislocation density control in the n-type transparent perovskite BaSnO3 film (Figure 12.9). The chemical solution deposition of Ba0.92 La0.08 SnO3 (BLSO) films possesses enhanced carrier mobility (∼23 cm2 /V/s) and visible light transmittance over 80%. They put forward that the oxygen vacancy is the vital regulatory factors to determine carrier mobility. Furthermore, by Sb doping at the Sn sites, the carrier concentration of the BaSnO3 film has been enhanced, and the conductivity of the BaSnO3 film has increased remarkably. They further established the correlation between the growth mechanism of the BaSnO3 thin film solution method and photoelectric performance [13]. 12.2.2.3 Materials Designed for p-Type TCOs

In p-type TCOs, in contrast to n-type TCOs, it is the filled valence bands rather than the empty conduction bands that form the conduction paths for carriers, namely, positive holes. However, the formation of the hole conduction paths is not as easy

12.2 Structure and Electrical Properties of Inorganic Conductive Materials

as that of the electron conduction paths. This is why p-type TCOs are rare. The valence bands of oxides are composed of filled oxygen 2p orbitals in general. As a consequence, the oxygen 2p orbitals must be sufficiently overlapped with each other to form the conduction paths for holes. Nevertheless, the oxygen 2p orbitals at the VBM consist of nonbonding states, which indicate almost no interaction, that is, no orbitals overlap between oxygen ions. This makes it extremely difficult to form the hole conduction paths at the valence band, resulting in considerably small hole mobility even if the holes are doped into the valence band. Hole doping is also substantially difficult in oxides because oxygen has large electronegativity, indicating that O2− ions with closed shells are too stable to accept holes. Actually, the energy of oxygen 2p orbitals is quite deep from the vacuum level, giving oxygen the large electron affinity. Therefore, it is considered that the formation of a hole conduction path by utilizing simply oxygen 2p orbitals is almost impossible. To improve the low hole mobility and relieve difficulty in hole doping in oxides, chemical modulation of the valence band by hybridizing cations’ orbitals with oxygen 2p orbitals was proposed [14]. The hybridization of orbitals usually occurs between orbitals with similar energies. Therefore, the energy of the candidate cations’ orbitals must be close to that of oxygen 2p orbitals. As similar to In2 O3 and SnO2 , the oxygen 2p6 band consists of the valence band. Apart from the oxygen 2p valence band, Cd2+ 4d10 and In3+ 4d10 bands have higher binding energies with almost the same energy intervals. Because the Cd and In are located at adjacent positions, the periodicity predicts that the energy of Ag+ 4d10 or Cu+ 3d10 orbitals is very close to the oxygen 2p energy and they are easily hybridized with the oxygen 2p orbitals. From the viewpoint of chemical stability, Cu 3d10 orbitals are considered to be the most favorable for the chemical modulation of the valence band. Figure 12.10 shows the hybridization between 3d10 (Cu) and 2p (O) nonbonding orbitals, which can form Cu—O bonding and antibonding states. The Cu—O antibonding states thus become the TVB [15]. Therefore, Cu—O bonding band (A), M—O bonding band (B), Cu nonbonding band (C), and Cu—O antibonding band (D) basically construct the valence band. The chemical modulation by 3d10 (Cu) orbitals turns the nonbonding 2p (O) orbitals into the Cu—O antibonding band at the TVB and elevates the energy of the VBM simultaneously. Accordingly, the shallow Cu—O antibonding band is expected to increase the hole mobility to make hole doping easier. The larger modulation generates higher hole mobility via forming a steady hole conduction path at the VBM. The chemical modulation normally narrows the bandgap by the formation of the shallow antibonding band atop the valence band. Because the degree of the modulation usually rests with the dimension of crystal structures, lower-dimensional crystal structure offers moderate hybridization, namely, a moderately shallow antibonding band. Therefore, layered structures and chain structures can better maintain the wide bandgap and transparency. 12.2.2.4 Typical p-Type TCOs and Their Electronic Structure

For the material design of p-type TCOs, the Cu+ -incorporated oxides with layered structures or chain structures are dominant options. The representative example is CuAlO2 with delafossite structure, which is a layered structure composed of Cu+ and AlO6 octahedra layers (Figure 12.11) and considered as a prototype of p-type

679

680

12 The Relationship Between Structure and Electric Property

Antibonding (CB)

ns0 ns2

M

Eg

Cu–O hybrid (antibonding) D

C

(VB top) 2p4

Nonbonding Nonbonding Cu 3d

O

2p6

3d10

Cu (closed shell) A Cu–O hybrid bonding

B Bonding (VB bottom)

Figure 12.10 Chemical modulation of the valence band by Cu 3d10 orbitals in oxides shown in a schematic molecular orbital diagram. Source: Hosono and Ueda [6]. Copyright 2017, Springer, New York, NY.

S

S

S

Cu

Cu

h+

S

S

S

S

Cu

Cu

S

S

CuS tetrahedra La O Cu AIO6 octahedra

Delafossite (CuAIO2, etc.) LaCuOS

Figure 12.11 Crystal structures of typical p-type TCOs. Source: Hosono and Ueda [6]. Copyright 2017, Springer Nature.

12.2 Structure and Electrical Properties of Inorganic Conductive Materials

Metal cation

Metal cation CBM Energy

Energy

CBM

Eg

Eg VBM

VBM O (a)

2p6 (b)

O 2p6

M 3d10

Figure 12.12 Comparison between the band structure of (a) oxide and (b) Cu-doped oxide. Source: Based on Liu et al. [17].

TCOs [16]. CuAlO2 is almost transparent in the visible region, which vastly differs from p-type Cu2 O that shows the deep dark red color. This is attributed to an O2 ion in CuAlO2 bonds only having one Cu+ and three Al3+ ions, resulting in moderate Cu–O hybridization. Besides, Cu-doped NiO (Cu:NiO) also shows typical p-type conductivity. Shan and coworkers reported the fabrication of p-type Cu-doped NiO thin films with the low-temperature (LT) solution process ( 20 meV/atom ΔH2 > 0

*

Sufficiently large bandgap (Eg > 2.8 eV) Low hole effective mass

* (m*h < 1.2 m0) *Instrinsic good p-type conductivity

character of VBM ensures their optical transparency and intrinsic good p-type conductivity [20]. Ag-based TCOs is another emerging p-type TCO material. Sun and coworkers reported a new sort of p-type delafossite Ag-based TCOs, AgCrO2 (ACO) (Figure 12.14). They have investigated the evolution of microstructure, morphology, and optical properties in ACO with respect to annealing temperature. The stoichiometric ACO shows self-assembled c-axis orientation, and the 500 ∘ C-annealed ACO presents relatively high quality, the dense surface, and sound optical transmittance among all the derived thin films [21]. Furthermore, with Mg doping, p-type conductivity is enhanced for AgCr1−x Mgx O2 (0.04 ≤ x ≤ 0.20). The conductivity initially increases from 3.1 × 10−3 to 67.7 × 10−3 S/cm and from x = 0.04 to 0.12; then, it slightly decreases with further increasing of Mg concentration. Meanwhile, AgCr1−x Mgx O2 also shows high magnitude of optical transmittance near 60% in the visible region and wide optical bandgap (3.41–3.66 eV). There are also several controversial reports on novel p-type TCOs. The experimental determination of p- or n-type conduction is sometimes untoward when carrier concentration is extremely high/low or both electrons and holes are present in conductors.

12.2.3 Nitride Conducting Materials 12.2.3.1 General Introduction of Nitride Conducting Materials

Nitride conducting materials are an emerging category of functional materials. The typical nitride conducting materials include group III nitrides (such as GaN and InN) and transition metal nitrides (TMNs, such as TiN, ZrN, and TaN). Besides these mentioned nitrides, all TMNs of the group IV/V/VI in the periodic table of elements have significant technological importance, as shown in Figure 12.15. These nitrides can

683

Delafossite

Coating solution

Repeat

Spin coating

Chemical synthesis Crystallized film Precursor

AO (0012)

80 Visible 70

500 °C/a-SO 800 °C/AO 700 °C/AO 600 °C/AO 500 °C/AO

50

250 °C-baked/AO

10

JCPD3 No. 70-1703

AgCrO2 film

20

30

40

Amorphous film

(b)

50 60 2θ (°)

70

80

800 °C/AO 700 °C/AO 600 °C/AO 500 °C/AO 400 °C/AO 250 °C-baked/AO

30

Annealing

10

Infrared

40

20

400 °C/AO

Wet film Baking

60

Ultraviolet

b

a

90 (0013)

R3m

(0015)

c

(0012)

3R type

(006)

O

(003)

AgCrO2

Transmittance (%)

(a)

Ag Cr

AO (006) (009)

12 The Relationship Between Structure and Electric Property

Intensity (a.u., in sqrt scale)

684

90

100

(c)

AO substrate 0 200 400 600 800 1000 1200 1400 1600 1800 2000 Wavelength (nm)

(d)

Figure 12.14 (a) Structure of delafossite AgCrO2 , (b) the schematic diagram of AgCrO2 thin film preparation, (c) XRD patterns of AgCrO2 thin films synthesized under different conditions, and (d) optical transmittance results for AgCrO2 thin films with different processing conditions. Source: Based on Xu et al. [20]. II

IIIa

Va

Be

B

N 2s22p3

Mg

AI 3s23p1 AIN: w, zb Eg = 6.3 eV

P

IIIb

IVb

Vb

VIb

Ca

Sc 1 2 3d 4s ScN: rs a = 0.442 nm Eg = 1.58 eV

Ti 3d24s2 TiN: rs a = 0.424 nm Epu = 7.8 eV

V 3d34s2 VN: rs a = 0.414 nm

Cr 3d44s2 CrN: rs a = 0.416 nm Epu = 0 eV

Ga 4s24p1 GaN: w, zb Eg = 3.4 eV

As

Sr

Y 4d15s2 YN: rs a = 0.487 nm Eg = 0.85 eV

Zr 4d25s2 ZrN: rs a = 0.461 nm Epu = 6.98 eV

Nb 4d45s1 NbN: rs a = 0.439 nm Epu = 8.93 eV

Mo 5d54s1 MoN: rs a = 0.424 nm Epu = 9.57 eV

In 5s25p1 InN: w, zb Eg = 0.7 eV

Sb

Ba

La 5d16s2 LaN: rs a = 0.530 nm Eg = 0.75 eV

Hf 5d26s2 HfN: rs a = 0.452 nm Epu = 7.95 eV

Ta 5d36s2 TaN: rs a = 0.439 nm Epu = 9.33 eV

W 5d46s2 WN: rs a = 0.421 nm Epu = 10.5 eV

TI

Bi

rs: rock salt zb: zinc blende w: wurtzite

Epu: unscreened plasma energy of a conductor Eg: bandgap of semiconductor

Figure 12.15 An excerpt of the periodic table of elements where the usual constituents of the various conductive binary and ternary transition metal nitrides are shown. Source: Patsalas et al. [22]. Licensed under CC BY 4.0.

form cubic rocksalt-type crystals and become a class of very vital functional materials due to their excellent electrical conductivity, exceptional mechanical properties, high melting points, and ideal environmental tolerance [22]. Group III nitrides generally have three crystalline structures (Figure 12.16): the wurtzite, zinc blende, and rock salt. Under ambient conditions, the thermodynamically stable structure is wurtzite for bulk AlN, GaN, and InN. The zinc blende structure for GaN and InN has been stabilized by epitaxial growth of thin films on

12.2 Structure and Electrical Properties of Inorganic Conductive Materials

C View normal to [0001] and [111]

B

B

A

A

Ga Wurtzitic

N

Zinc blende

View along [0001] and [111]

(a)

(b)

Figure 12.16 A stick-and-ball stacking model of crystals with (a) wurtzitic and (b) zinc blende polytypes. Source: Morkoc and Ozgur [23]. Copyright 2008, John Wiley & Sons.

{0 1 1} crystal planes of cubic substrates, such as Si, SiC, MgO, and GaAs. In these cases, the intrinsic tendency to form the wurtzite structure is overcome by the topological compatibility. However, wurtzite structure could very likely be present at the extended defect sites. The form of rock salt is possible merely under high pressure and, therefore, is a laboratory form of exercise. Group IV nitrides of the periodic table of elements have four valence electrons with a configuration of d2 s2 for all the three metals ([Ar] 3d2 4s2 , [Kr] 4d2 5s2 , and [Xe] 4f14 5d2 6s2 for Ti, Zr, and Hf, respectively), and they form bonds with N atoms (valence electronic configuration 2s2 2p3 ). While various nitride phases such as the Me2 N and Me3 N4 (Me = Ti, Zr, and Hf) were reported, the most stable and durable nitrides of these metals are those in the cubic rocksalt B1-MeN arrangement called ◽-MeN (Me = Ti, Zr, and Hf) [24] as well. Among these, B1-TiN is the archetypical example of the conductive nitrides. Group V nitrides cover VN, NbN, and TaN. V, Nb, and Ta have one more valence electron than Ti, Zr, and Hf, respectively. Consequently, they have five valence electrons and their valence electron configurations are [Ar] 3d3 4s2 , [Kr] 4d4 5s1 , and [Xe] 5d3 6s2 , respectively. VN and NbN can form easily the B1 nitride phase, while B1-TaN is metastable. B1-VN, B1-NbN, and B1-TaN are paramagnetic, while they all have been reported to be superconductors [25]. Especially for NbN, it is a well-known superconductor whose superconducting properties have been under investigation since the early 1980s. Group VI nitrides include CrN, MoN, and WN. Cr, Mo, and W have one more electron than V, Nb, and Ta, respectively. Hence, they have six valence electrons and their valence electron configurations are [Ar] 3d4 4s2 , [Kr] 4d5 5s1 , and [Xe] 5d4 6s2 , respectively. The attention that B1-CrN received was due to its superior wear and

685

686

12 The Relationship Between Structure and Electric Property

anti-corrosion performance compared with B1-TiN. The magnetic properties of CrN are unique among all the TMN and clearly proven by Duan et al. who performed spin polarized calculations and predicted its semiconductive character when in the anti-ferromagnetic state. The interplay between the magnetic properties of CrN and its semiconductive behavior was experimentally confirmed by several groups and CrN was regarded as a candidate plasmonic conductor [26]. Ternary conductive nitrides: alloying the aforementioned TMN to form ternary (or more accurately pseudo-binary) films is an effective way to control their mechanical, optical, and electronic properties. The similar crystal structure (rock salt), local symmetry (octahedra), and the similar bonding to N (via hybridization of the metal’s d electrons with the nitrogen’s 2p electrons) make all the nitrides of the group IV, V, and VI metals completely soluble to each other in the B1 phase and in the entire compositional range; in fact, such alloying may cause more stable films with less structural defects in comparison with some metastable binary nitrides, with the most prominent example being the stable Ta-rich B1-Tix Ta1−x N and the metastable B1-TaN. The most widely used alloying phase for electronic applications is the B1-TiN, which is implemented as the basis for the formation of Tix Zr1−x N, Tix Hf1−x N, Tix V1−x N, Tix Nb1−x N, Tix Ta1−x N, Tix Cr1−x N, Tix Mo1−x N,and Tix W1−x N. Electronics-relevant Ta-based (Tax Zr1−x N and Tax W1−x N), Zr-based (Zrx Hf1−x N, Zrx Nb1−x N, Zrx Cr1−x N, and Zrx Y1−x N), and V-based (Vx W1−x N) ternary nitrides were also reported [27].

12.2.3.2 Structure and the Property in Typical Nitride Conducting Materials

GaN, which represents an extremely crucial semiconductor family at present, has a wide range of modern applications, including light-emitting diodes, laser diodes, and power electronics. Doping is a most efficient method for engineering new and improved properties of GaN. The unintentionally doped GaN is n-type in all cases, and the electron concentration of the best sample is about 4 × 1016 cm−3 . In general, the p-type samples prepared are highly compensated. With the successful control of the crystalline quality, it could start to work on p-type doping. In 1989, the Mg concentration was found to have good controllability in the growth of GaN by organometallic vapor phase epitaxy (OMVPE) using Cp2 Mg as a p-type dopant source [28]. Then, for the first time, distinctly p-type GaN with low resistivity was discovered by low-energy electron-beam irradiation (LEEBI) on high-quality Mg-doped GaN. P-type AlGaN was achieved in 1991 and p-type GaInN in 1994, respectively [29]. Later, Nakamura succeeded in making p-type GaN by thermal annealing in an N2 atmosphere of Mg-doped GaN using Cp2 Mg. Regarding n-type doping, researchers attempted to dope with SiH4 in 1986, but it was hard to control the conductivity on account of the high density of residual donors. Researchers succeeded in controlling the conductivity of n-type nitrides using high-quality GaN or AlGaN grown with the LT-buffer layer in combination with SiH4 doping. The electron concentration could be linearly controlled from an undoped level to 1019 cm3 without deterioration of surface morphology, when the SiH4 flow rate was varied.

12.2 Structure and Electrical Properties of Inorganic Conductive Materials

Besides, n-type and p-type GaN can be achieved by doping. As a core component of power electronic devices, semi-insulating GaN is also usually achieved by carbon (C) doping. However, the doping behaviors, especially the lattice site occupation of C in GaN, still remain in debate. As a group IV element, C is amphoteric in GaN, that is, it may occupy either the Ga site to form CGa or the N site to form CN, and they can also coexist with each other, individually or forming complexes with other defects. Nevertheless, there is so far no direct experimental evidence for the lattice site of C in GaN, partially owing to the insufficiency of valid characterization methods, despite the fact that the identification of the lattice location of Mg in GaN has made great progress recently. As a result, there are lots of controversies about C-related defects and their roles in electrical and optical properties. Therefore, Shen and coworkers clarify the lattice site of C in GaN using polarized Fourier-transform infrared and Raman spectroscopies, in combination with first-principle calculations [30]. Two local vibrational modes (LVMs) at 766 and 774 cm−1 in C-doped GaN are observed. The 766 cm−1 mode is assigned to the nondegenerate A1 mode vibrating along the c-axis, whereas the 774 cm−1 mode is ascribed to the doubly degenerate E mode confined in the plane perpendicular to the c-axis. The two LVMs are identified to originate from isolated C−N with local C3v symmetry. Experimental data and calculations are in agreement with both the positions and the intensity ratios of the LVMs tremendously. We thus provide unambiguous evidence of the substitutional C atoms occupying the N site with a −1-charge state in GaN and therefore bring essential information to a long-standing controversy (Figure 12.17). Besides doping strategy, reducing dimensions is also an effective means for integration between structure and the property. GaN with two-dimensional (2D)

–

CN with C3𝝂 symmetry Excited

E

state

A1 (ω1)

(ω2)

𝝂1

𝝂2

(ω3) 𝝂3

c⃗

Ground state Vibration//c⃗

Vibration⟂c⃗

Figure 12.17 The schematic energy level diagrams and fundamental dipole transitions in a C−N center with C3v symmetry. The calculated vibration directions of C atoms are also depicted. The larger brown spheres represent C atoms, while the smaller green spheres denote Ga atoms [30].

687

688

12 The Relationship Between Structure and Electric Property

c a b

a b (b)

1 nm

0.5 nm

1 nm

(c)

(0001) c 0.5 nm Ga

a

0.5 nm

1 nm

(e)

(1010) Ga (a)

N

1 nm

0.5 nm Overlayer 0.5 nm

(d)

(f)

Figure 12.18 Atomic structure of 2D GaN single crystals: (a) the scheme of the crystalline structure of GaN, (b,c) high-angle annular dark-field detector (HAADF) (b) and BF STEM (c) images of the 2D GaN crystal acquired along the [0001] zone axis, (d) atomic resolution EDS elemental mapping of Ga, N, and their combinations along the [0001] zone axis, and (e,f) HAADF (e) and BF STEM (f) images of the 2D GaN crystal acquired. Source: Chen et al. [31]. Copyright 2018, American Chemical Society.

structure has been highly anticipated because its quantum confinement effect enables desirable deep-ultraviolet emission, excitonic effect, and electronic transport properties to happen. However, the currently obtained 2D GaN can merely exist as intercalated layers of atomically thin quantum wells or nanometer-scale islands, limiting further exploration of its intrinsic traits. Fu and coworkers first reported the growth of micron-sized 2D GaN single crystals on liquid metals via a surface-confined nitridation reaction and demonstrated that the 2D GaN showed the uniformly incremental lattice, unique phonon modes, blue shifted photoluminescence emission, and improved internal quantum efficiency, providing direct evidence to the previous theoretical predictions [31]. The as-grown 2D GaN exhibited electronic mobility up to 160 cm2 /V/s. Their findings provide a fascinating way to potential optoelectronic applications of 2D GaN single crystals (Figure 12.18). In addition to experimentally synthesizing crystalline conducting nitrides, computational screening was also employed to search for promising nitride semiconductors. Hinuma et al. have used calculations to screen a set of compounds for potential semiconductor candidates [32]. The study identified 11 previously unreported materials, embracing the particularly promising compound calcium zinc nitride (CaZn2 N2 ). They reported computational screening of ternary zinc nitrides through a combination of the prototype-based and evolutionary algorithm structure searches. Twenty-one promising semiconductors with small carrier effective masses

12.2 Structure and Electrical Properties of Inorganic Conductive Materials

Band gap (eV)

5

Direct Indirect Experimental

4 3 2

0

(a) 2

IR

1

Effective mass (m0)

UV

6

LiZnN KZnN Be2ZnN2 Ca2ZnN2 Ba2ZnN2 ZnSiN2 ZnSnN2 ZnZrN2 Zn2PN3 Zn2NbN3 Zn3WN4 GaN NaZnN CaZn2N2 Mg2ZnN2 Sr2ZnN2 Zn3LaN3 ZnGeN2 ZnTiN2 ZnHfN2 Zn2VN3 Zn2TaN3 Zn3N2

Hole Electron

1

0

LiZnN KZnN Be2ZnN2 Ca2ZnN2 Ba2ZnN2 ZnSiN2 ZnSnN2 ZnZrN2 Zn2PN3 Zn2NbN3 Zn3WN4 GaN NaZnN CaZn2N2 Mg2ZnN2 Sr2ZnN2 Zn3LaN3 ZnGeN2 ZnTiN2 ZnHfN2 Zn2VN3 Zn2TaN3 Zn3N2

(b)

Figure 12.19 Electronic properties of theoretically identified ternary zinc nitrides: (a) bandgaps, (b) effective masses for holes and electrons normalized by the free-electron rest mass. Source: Hinuma et al. [32]. Licensed under CC BY 4.0.

are identified via systematic first-principle calculations of stability and electronic structure (Figure 12.19). The proposed nitrides include earlier unreported CaZn2 N2 with earth-abundant components and a direct bandgap, which is synthesized by high-pressure methods. Alloy calculations signify a bandgap capable of covering most of the visible light range. Native defects and dopant calculations also predict p- and n-type dopability for CaZn2 N2 and the related Ca2 ZnN2 . Foremost among the characteristics they sought in any new compound was high conductivity, which is crucial for most semiconductor applications. GaN and InN have spatially diffuse orbital traits, which result in the high electrical conductivity that has led to their success. The researchers considered materials with such properties similar to GaN and InN in order to find new industry-relevant semiconductors. Zinc nitride has these desirable properties but is difficult to be made. As a result, the researchers decided to screen for ternary compounds comprising zinc, nitrogen, and a third element. This approach brings about the identification of 11 new materials, covering the exciting CaZn2 N2 (Figure 12.20). The left is the Ca–Zn–N ternary phase diagram; the right is crystal structure of CaZn2 N2 . Only the Zn—N bonds are illustrated for easy visualization. Feng and coworkers have also used first-principle calculations to extensively study the phase stability, electronic structure, elastic and metallic properties of manganese

689

690

12 The Relationship Between Structure and Electric Property Zn CaZn13 CaZn11 CaZn5 CaZn2

Ca (Zn3N2)

CaZn

Zn

CaZn2N2

N Ca2ZnN2

Ca

(a)

Ca2N Ca3N2

1 N 2 2

(b)

Figure 12.20 Predictions on the stability and crystal structure of CaZn2 N2 . (a) Ca—Zn—N ternary phase diagram, (b) Crystal structure of CaZn2 N2 . Only the Zn—N bonds are illustrated for easy visualization.

nitrides (Mn4 N, Mn2 N0.86 , Mn3 N2 , and MnN) [33]. The negative values of cohesive energy and formation enthalpy show that these compounds are thermodynamically stable. The bonding of manganese nitrides is the combinations of covalent and metallic bonds (Figure 12.21). The bands are broader and bandgaps narrower in nitrides than in the corresponding oxides. Nitrides and oxynitrides containing fully oxidized transition metal ions are often brightly colored compared with the corresponding colorless oxides. In nitrides comprising incompletely oxidized transition metal ions, delocalization of the valence electrons is favored (Figure 12.22), and so these materials receive attention for their electronic properties. The binaries TiN, ZrN, VN, NbN, TaN, and MoN are all metallic and become superconducting at temperatures between 6 and 15 K. When the layered nitride chlorides b-ZrNCl and b-HfNCl are reduced by intercalation with lithium, they become superconducting at 13 and 25 K, respectively. Although plentiful nitride conducting materials have been synthesized, an increasing number of new nitrides can be combined and predicted. Nitrides break new ground for material discovery and design – so long as we have a rational understanding of the factors that drive stability in these relatively unexplored spaces. Sun et al. used computational material discovery and informatics tools to build a large stability map of the ternary metal nitride space. Their objective was not only to predict and synthesize new ternary metal nitrides but also to further visualize large-scale relationships between nitride chemistry and thermodynamic stability and to rationalize these trends from their deeper chemical origins. As it stands, the map is necessarily incomplete – it represents a current “upper-bound” on the ternary nitride stability landscape [34]. As new exotic structures and bonding motifs are discovered in the ternary metal nitrides, the procedures in this work can be iteratively reapplied to update and refine our understanding of this extended compositional space (Figure 12.23). From a broader perspective, our computational approach offers a systematic blueprint for mapping uncharted chemical spaces, providing synthetic chemists with guidance in their quests to continuously extend the frontier of solid-state chemistry.

12.2 Structure and Electrical Properties of Inorganic Conductive Materials

(a) Mn4N

(b) Mn2N0.86

(c) Mn3N2

(d) MnN

Mn

N

Figure 12.21 The crystal structure of manganese nitrides: Mn4 N (a), Mn2 N0.86 (b), Mn3 N2 (c), MnN (d). Source: Yu et al. [33]. Copyright 2015, The Royal Society of Chemistry.

12.2.4 Carbide Conducting Materials Carbide conducting materials are an important class of conductive materials that have broad applications. Examples of different types of carbides include the silicon carbide (SiC), the boron carbide (B4 C), and transition-metal carbides (TMCs). 12.2.4.1 The General Structure and the Property of SiC

SiC, as the representative of third-generation semiconductor materials, is a conducting material containing silicon and carbon elements. It exists in the nature as the extremely rare mineral moissanite. The high sublimation temperature of SiC makes it useful for bearings and furnace parts. SiC does not melt at any known temperature and is also highly inert chemically. Large-scale production of SiC powder used as an

691

692

12 The Relationship Between Structure and Electric Property

Inductive

Aδ+

Nδ–

e– Bδ+ e–

Nδ–

N-rich A

e–

N

+

B

e–

N2

N N-poor

Reductive

Mδ–

e– e–

Mδ–

Nδ+

Figure 12.22 Delocalization of the valence electrons in nitrides. Source: Sun et al. [34]. Copyright 2008, 2019 Springer Nature.

abradant has started since 1893. Grains of SiC can be bonded together by sintering to form adamantine ceramics widely used in fields requiring high endurance, such as car brakes, car clutches, and ceramic plates in bulletproof vests. As one of most promising wide-bandgap semiconductors, SiC has also imperative applications in the field of electronic components for its high thermal conductivity, electric field breakdown strength, and maximum current density. SiC also has a very low coefficient of thermal expansion and experiences no phase transitions that would cause discontinuities in thermal expansion. Electronic applications of SiC for high-powered devices have raised more and more attention. Large single crystals of SiC can be grown by the Lely method, and they can be cut into gems known as synthetic moissanite. There are more than 200 crystal forms of SiC. The polymorphism of SiC is characterized by a large family of similar crystalline structures called polytypes (Figure 12.24). They are variations of the same chemical compound, and they are identical in two dimensions but different in three dimensions. Thus, they can be viewed as layers stacked in a certain sequence. Among these structures, α-SiC with a hexagonal crystal structure is the most commonly encountered polymorph, which is formed at temperatures greater than 1700 ∘ C. However, the β modification (β-SiC), with a zinc blende crystal structure, is formed at temperatures below 1700 ∘ C. Until recently, the β form has had relatively few commercial uses, but there is now increasing interest in its use as a support for heterogeneous catalysts, owing to its higher surface area compared with the α form. 12.2.4.2 The Doping of Silicon Carbides and Its Electronic Property

SiC is a variety of typical semiconductors, which can be doped n-type by nitrogen or phosphorus and p-type by Be, B, Al, or Ga. Metallic conductivity has been achieved by heavy doping with boron, aluminum, or nitrogen. Superconductivity has been detected in 3C-SiC:Al, 3C-SiC:B, and 6H-SiC:B at the same temperature of 1.5 K. A crucial difference is, however, observed for the magnetic field behavior between aluminum and boron doping: SiC:Al is type-II, same as Si:B. On the contrary, SiC:B

Li Ba Sr Ca Mg Zn Na K Rb Cs V Nb Ta Hf Zr Ti Si Ge C B W Mo Cr Mn Fe Co Ni In Ga Al Bi Sb Sn Pt Pd Rh Ir Ru Os Re Au Ag Cd Pb Cu Te Se S Sc Y

12.2 Structure and Electrical Properties of Inorganic Conductive Materials

Li Ba Sr Ca Mg Zn Na K Rb Cs V Nb Ta Hf Zr Ti Si Ge C B W Mo Cr Mn Fe Co Ni In Ga Al Bi Sb Sn Pt Pd Rh Ir Ru Os Re Au Ag Cd Pb Cu Te Se S Sc Y ΔHq (eV/atom)

ΔHq (eV/atom)

New stable ternary space

Experimentally Stabilizable known in ICSD ΔμN2 < 1 eV/N

–2

–1 Stable

0

–2

–1 0 Metastable versus binaries

ΔHq (eV/atom)

0

1 2 Metastable versus elements

Figure 12.23 The map of the inorganic ternary metal nitrides . Source: Sun et al. [34]. Copyright 2008, 2019 Springer Nature.

(a)

(b)

(c)

Figure 12.24 Structure of major SiC polytypes: (a) β-3C-SiC, (b) 4H-SiC, and (c) α-6H-SiC. Source: Materialscientist, https://en.wikipedia.org/wiki/Silicon_carbide. CC BY-SA 3.0.

is type-I. In an attempt to explain this distinction, it was noted that Si sites are of more significance than carbon sites in terms of superconductivity in SiC, whereas boron substitutes carbon in SiC and Al substitutes Si sites. Therefore, Al and B “see” different environments that might explain disparate properties of SiC:Al and SiC:B [35].

693

694

12 The Relationship Between Structure and Electric Property

Despite of the superb characteristics of pristine SiC, its features can be further enhanced by tuning the available properties by way of doping transition metals (TM). For instance, when electronic properties are considered, the introduction of TM elements in the wide bandgap SiC offers the tailoring of the band diagram, conduction, spin-related features, which expands the scope of use of the material, including applications of spintronics and electrochemical devices. In the condensed matter, the term doping basically refers to the introduction of an impurity in the pristine material for achievement of best innovative properties. The substitutional incorporation of impurity atoms in the host matrix may refer to two cases: alloying and doping. These terms are currently used side by side and can be differentiated on the basis of percent concentration of the foreign atoms. The concentration less than 10% is usually perceived as the limit of doping after which the material’s properties are drastically changed to develop new structural phases referred to as alloying. In the context of doping with a few percent of impurities, the host matrix usually does not lose its structural properties. There are various doping strategies which basically depend upon the requirements of the crystal grower. The doping types include chemical doping, conductivity doping, magnetic doping, neutron transmutation doping, modulation doping, doping in organic molecular semiconductors, etc. The magnetic doping is carried out by the substation of open shell atoms of 3d, 4d, 5d, 4f, and 5f in the host matrix. For high power electronics, such as n-type Si, neutron transmutation doping is introduced. 12.2.4.3 Materials Designed for New SiC

Although SiCs have been extensively studied for their excellent physical/chemical properties, searching for new-type SiC materials has also been brought into focus for enriching the material system. In view of rich structures of SiCs, predicting and searching for new SiCs by theoretical simulation are increasingly popular. Cao and coworkers reported a new-fashioned wide-bandgap SiC semiconductor, SiC4 , with low density and high elasticity [36]. The new material is designed and distinguished by ab initio molecular dynamics simulations and first-principle calculations. As shown in Figure 12.25, the d orbitals of Si and the π* orbitals of ethynyl moieties are conjugated in the through-space direction, which can largely enhance the photoconductivity. This new-style superlight and super flexible SiC semiconductor is predicted to have unique electronic, optical, and mechanical properties, and it is a quite promising material for the high-performance UV optoelectronic devices propitious for various practical demands in a complex environment. Although great efforts are devoted to the utility of SiC for the main reason of its wide bandgap property. In fact, it is desirable to see its broader applications in electronics. Especially given its 2D structures like graphene, we find that the bandgap of SiC is too large to restrict its application in nanoelectronics. To cope with this problem, a variety of approaches toward modulating the bandgap of SiC have been proposed and developed. Recent experimental and theoretical works have demonstrated that the covalent functionalization, such as hydrogenation (H-SiC) and fluorination (F-SiC), is an effective approach to tune the bandgap. In consideration of this, Zeng and coworkers conducted a systematic theoretical investigation of the atomic

12.2 Structure and Electrical Properties of Inorganic Conductive Materials

5

me = 0.230 m0

0.14 0.12

3 Ee = 5.02

2

–1

0.10 0.08 0.06 0.04

1 mh = 0.239 m0

0

(a)

0.16

Current (nA)

Energy (eV)

4

CBM

VBM R

0.02 0.00 0.0

0.4

0.8

1.2

1.6

2.0

Bias (V)

M

(b)

Figure 12.25 The bulk phase of the new-designed SiC4 material and its electrical performance. (a) Band structure and PDOS of SiC4 , (b) current–voltage (I–V ) curve of the material. Source: Sun et al. [36]. Copyright 2016, American Chemical Society.

and electronic structure of a fully hydrogenated/fluorinated SiC (H/F-SiC) heterobilayer [37], which has a quasi-metallic character in its most stable stacking pattern (Figure 12.26), to predict its electronic and optical properties. They demonstrate that a direct bandgap can be opened in the quasi-metallic H/F-SiC heterobilayer by applying an external electric field. Especially, when the strength of the field is altered, this system undergoes a transition from the quasi-metallic state to the semiconductor. The predicted mobility is rather high due to the low carrier effective mass and high Fermi velocity. Light absorption spectra indicate that the H/F-SiC heterobilayer has evident infrared light absorption, and complete electron–hole separation can enhance the photocatalytic efficiency. Their findings pave the way for experimental research on the development of 2D material science using weak interlayer interactions and indicate the great application potential of the SiC in future nanoelectronics and optoelectronics. 12.2.4.4 Materials Designed for Transition-Metal Carbides

Generally, TMCs are intrinsic metal, and many of them exhibit the unusual combination of conductivity and hydrophilicity. Y. Gogotsi’s group predicted a few TMCs, such as Mo2 CTx and Mo2 TiC2 Tx , showing semiconducting-like behaviors with appropriate termination Tx . However, most of these semiconducting TMCs are of indirect bandgaps. Further enriching the TMC family and introducing semiconducting members with direct bandgaps are critical for expanding the application of TMCs in optoelectronics. Thus, Huang and coworkers designed new 2D direct bandgap semiconducting TMCs, namely ScCx OH, with an experimentally measured bandgap approximated to 2.5 eV (Figure 12.27) [38]. The fabrication of this hydroxyl-functionalized and carbon-deficient scandium carbide is achieved by selective etching of a layered parent ScAl3 C3 compound. Furthermore, the ScCx OH-based device exhibited excellent photo-response in the ultraviolet-visible light region. This 2D ScCx OH direct-bandgap semiconductor has great application prospect in the area of visible-light detectors, photocatalytic chemistry, and optoelectronic devices.

695

12 The Relationship Between Structure and Electric Property

d

(a)

(b)

(c)

D

(d)

Figure 12.26 Side (top) and front (bottom) views of the most stable structure of the H/F-SiC heterobilayer for (a) C–H· · ·F–C, (b) Si–H· · ·F–Si, (c) C–H· · ·F–Si, and (d) Si–H· · ·F–C interaction, with the ground state structure favoring the C–H· · ·F–C bonding type. Source: Chen et al. [37]. Copyright 2016, The Royal Society of Chemistry. 10

Sc C O H Al

AI3C2 ScC

ScCx OH

Photocurrent (pA)

696

5

280 nm

40 μW/cm2

360 nm

160 μW/cm2

0

–5

TMAOH etching –10 –10

–5

550 nm

320 μW/cm2

600 nm Dark

274 μW/cm2

0

5

10

Applied voltage (V)

(a)

(b)

Figure 12.27 The structure of ScCx OH (a) and its photoelectric characterization (b). Source: Zhou et al. [38]. Copyright 2019, American Chemical Society.

In contrast to conventional preparation of MXene, the layered ternary Zr3 Al3 C5 material instead of MAX phases is used as the source for hydrofluoric acid treatment. The structural, mechanical, and electronic properties of the synthesized 2D carbide are investigated, combined with first-principle density functional calculations. A comparative study on the structural stability of our obtained 2D Zr3 C2 Tz and Ti3 C2 Tz MXenes at elevated temperatures is performed (Figure 12.28) [39]. The obtained 2D Zr3 C2 Tz exhibits relatively better ability to maintain 2D nature and

12.2 Structure and Electrical Properties of Inorganic Conductive Materials

M Γ

Zr

K

C (a)

(b) 3

Zr3C2O2

3

Zr3C2F2

2

1

1

1

0

0

–1

–2 –3

E (eV)

2

–1

(d)

(c)

2 E (eV)

E (eV)

3

T

M

K

Γ

–3

(e)

0

–1

–2 Γ

Zr3C2(OH)2

–2 Γ

M

K

Γ

–3

(f)

Γ

M

K

Γ

Figure 12.28 (a) Top views and (b) side views of the 2D Zr3 C2 T2 (T = O, F, and OH), (c) the Brillouin zone of the 2D hexagonal lattice, and (d–f) the electronic energy bands of Zr3 C2 O2 , Zr3 C2 F2 , and Zr3 C2 (OH)2 , respectively. Source: Zhou et al. [39]. Copyright 2016, John Wiley & Sons.

structural integrity compared with Ti-based MXene. The difference in structural stability under the high temperature condition is explained by a theoretical investigation on binding energy.

12.2.5 Sulfide Conductive Materials 12.2.5.1 General Introduction of Sulfide Conductive Materials

Sulfide conducting materials, mainly metal sulfides, are a key type of inorganic semiconducting materials for their practical applications in electronic devices, solar cells, photo catalysis, and energy storage. The bonding in metal sulfides is highly covalent, which gives rise to their semiconducting properties and in turn is in connection with the deep colors. Examples of metal sulfides incorporate CdS, PbS, ZnS, SnS, YS, Cu2 S, MoS2 , In2 S3 , Ni3 S2 , Ag2 S, Na2 S, Bi3 S2 , SeS2 , Sb2 S3 , TiS2 , FeS2 , WS2 , Ni2 FeS4 , NiCo2 S4 , Co–Mo–S, Ni–Fe–Co–S, and so forth. Among these various metal sulfides, their structure is very variable, from lamellar structures with weak interplanar interactions to 3D stronger and harder frameworks [40]. Commonly, these are the structures exhibited by a much larger group of crystalline solids, such as the rocksalt structure of the galena group (Figure 12.29a), the sphalerite and wurtzite forms of ZnS (Figure 12.29b,c). Sulfides, such as covellite (CuS) (Figure 12.29d) and molybdenite (MoS2 ), have layer structures, or linkage of metal-sulfur octahedra along the c-axis direction in pyrite (FeS2 ), marcasite (FeS2 ), and arsenopyrite (FeAsS) (Figure 12.29e) [41]. MoS2 is now a noted semiconducting metal sulfide, which forms layers with an interplanar distance of 3.49 Å and can lubricate until it is transformed to MoO3 in air above 500 ∘ C. The sliding properties of MoS2 are attributed to the weak interactions between the sulfur atoms of adjacent layers (Figure 12.30). In the MoS2 structure,

697

698

12 The Relationship Between Structure and Electric Property

(a)

(d)

Galena

Sphalerite

(b)

Covellite

Pyrite

(c)

Wurtzite

Marcasite Loellingite Arsenopyrite

(e)

Figure 12.29 Crystal structures of the major sulfides: (a) galena (PbS), (b) sphalerite (ZnS), (c) wurtzite (ZnS), (d) covellite (CuS, MoS2 ), and (e) pyrite (FeS2 ), marcasite (FeS2 ), loellingite (FeAs2 ), and arsenopyrite (FeAsS). Source: Vaughan and Corkhill [41]. Copyright 2017, Mineralogical Society of America.

Mo S 6.65 Å

3.49 Å

S Mo

Figure 12.30

The view of MoS2 layers. Source: Copyright 2015, Woodhead Publishing.

each Mo(IV) center occupies a trigonal prismatic coordination sphere, being coupled with six sulfide ligands. Tin sulfide is another vital sulfide semiconducting material. Thanks to the versatile ways of coordination of both Sn and S, tin sulfides have several structures. The Sn(IV) sites are octahedrally coordinated in the ribbon, while the Sn(II) sites occupy trigonal-pyramidal sites. This compound exhibits polymorphism and at least three disparate types have been identified: α-Sn2 S3 , β-Sn2 S3 , and γ-Sn2 S3 . The primary structure of this compound is shown in Figure 12.31. It is noteworthy to point out that Sn(II) is on the edge while Sn(IV) is in the middle [42]. Tin monosulfide, SnS, known as the mineral herzenbergite, can be prepared by stoichiometric quantities of S and Sn at temperatures between 600 and 750 ∘ C. The structure is composed of layers with a distance of 3.383 Å. Sn is coordinated in a very

12.2 Structure and Electrical Properties of Inorganic Conductive Materials

Figure 12.31 A representation of the ribbon structure of Sn2 S3 , the trigonal-pyramidal Sn(II) located at the edge and the octahedral Sn (IV) at the center of the ribbon with six coordinated sulfur atoms. Source: Based on Greyson et al. [42].

distorted octahedron with three short bonds of 2.7 Å and three long ones of nearly 3.4 Å with S [43]. Antimony trisulfide is currently the most effective metal sulfide in the friction industry. Sb2 S3 crystallizes in the orthorhombic system. Each Sb atom is combined with three oxygen atoms forming a trigonal pyramid in line with Figure 12.32. The material consists of ribbons of Sb2 S3 (four for each cell) parallel to the c-direction, which forms chains of ribbon pairs Sb4 S6 . Each pair is weakly linked to the neighbor chain, and this feature of forming weakly linked chains is what provides the lubricant properties. The structure of sulfide conducting materials is very rich. Various valence states of sulfur coordinated with metal exhibit versatile chemical structures which are closely related to the ratio between sulfur and metal. Therefore, the research on sulfide materials is of great significance for enriching conductive material systems and exploring the relationship between nanostructure and properties. 12.2.5.2 The Structure and the Property of Typical Sulfide Conductive Materials

MoS2 as a most typical sulfide semiconductor has drawn great attention by reason of interesting physical and chemical applications. MoS2 possesses rich crystal structures stacked by different MoS2 motifs, including 1H, 1T, 1T′ , and 1T′′′ MoS2 single layers (Figure 12.33). Among these, noncentrosymmetric MoS2 semiconductors own not only novel electronic structures of spin–orbit coupling and valley polarization but also remarkable nonlinear optical effects. 1T′′′ -MoS2 layers, a very interesting noncentrosymmetric structure, was predicted to be built up from [MoS6 ] octahedral motifs merely by theoreticians. For that reason, Huang and coworkers have successfully harvested 1T′′′ MoS2 single crystals by a topochemical method, which was determined from single-crystal X-ray diffraction [44]. The prepared 1T′′′ MoS2 was verified to be semiconducting and showed a bandgap of 0.65 eV, which was different from metallic nature of 1T or 1T′ MoS2 . More surprisingly, the 1T′′′ MoS2 have not

699

700

12 The Relationship Between Structure and Electric Property

Figure 12.32 Crystal structure of Sb2 S3 . Source: BenNasr et al. [43]. Copyright 2011, Elsevier.

Arrangement of Mo atoms

1T‴: √3 ×√3

Mo S

(a)

(b)

1T′: √3 × 1

(c)

1T: 1 × 1

(d)

1H: 1 × 1

Figure 12.33 The schematic structure for the four phases of MoS2 . (a) 1T′′′ MoS2 with a √ √ √ 3 × 3 superstructure, (b) 1T′ MoS2 with a 3 × 1 superstructure, (c) 1T MoS2 with a unit cell of 1 × 1, (d) 1H MoS2 with a unit cell of 1 × 1. Source: Fang et al. [44]. Copyright 2019, American Chemical Society.

12.2 Structure and Electrical Properties of Inorganic Conductive Materials

S

A Mo

2H-MoS2

b A

Mo4+ 4d2

dxz dyz dx2 – y2 d xy dz2

Trigonal prism (a) 1T-MoS2 A

Mo4+ dz2 dx2 – y2 4d2

b C

Octahedron

dxz dyz dxy

(b)

Figure 12.34 (a) The schematic diagram of the 2H-MoS2 crystal structure, (b) the schematic diagram of the crystal structure of 1T MoS2 . Source: Fang et al. [45]. Copyright 2018, John Wiley & Sons.

presented strong optical second-harmonic generation signals. Their work provided the first-layered noncentrosymmetric semiconductor of edge-sharing MoS6 octahedra for the research of nonlinear optics. Meanwhile, Huang and coworkers further successfully synthesized the 1T MoS2 single crystals and redetermined the crystal structure of 1T MoS2 from single-crystal X-ray diffraction [45]. The individual MoS2 layer is composed of the MoS6 octahedron sharing edges with each other (Figure 12.34). More surprisingly, a superconducting transition of T c = 4 K was observed in the bulk 1T MoS2 crystals, which is the first observation of superconductivity in the pure 1T MoS2 phase. Besides MoS2 , considerable progress has also been made in the fabrication of disparate layer-structured III–VI semiconductors in various forms. Among the III–VI group of semiconductor materials, GaS, as one of the most significant materials, has two different stoichiometric means, i.e. GaS and Ga2 S3 . As everyone knows, hexagonal GaS has layered structures with each layer consisting of an S–Ga–Ga–S repeating unit built by six-membered Ga3 S3 rings. Layered GaS nanosheets have been attracting increasing research interests due to their highly anisotropic structural, electrical, optical, and mechanical properties, which are conducive to many applications. Li and coworkers have first fabricated few-layer GaS nanosheets and reported few-layer GaS-based photodetectors (Figure 12.35). The GaS 2D photodetector showed different photo-responses in various gas environments with a fast and stable response [46]. A theoretical investigation illustrated that the charge transfer (CT) between the adsorbed gas molecules and the photodetector leads to the different photo-responses. Copper sulfides (Cu2−x S) are also a type of the pivotal p-type sulfide semiconductor. The bandgap energy of Cu2−x S relies on the atomic ratio between Cu and S, varying from ∼1.2 eV for chalcocite (x = 0) to ∼2.0 eV for covellite (x = 1). Owing to these unique properties, enormous efforts have been devoted to the synthesis of diverse Cu2 S and CuS nanocrystals and Cu2−x S-based ternary and

701

12 The Relationship Between Structure and Electric Property

Ga

S

(a)

(c) 0.0 μm

35.0 μm

6 5 4 3 2

4.50 nm

7.5 Å

Thickness (nm)

702

1 0

–1 11

(d)

(b)

Au

Au

13

15 17 19 21 Distance (μm) Light

23

25

Au

GaS Au SiO2 Si A

(e)

(f)

Figure 12.35 (a) The top view and (b) the side view of the GaS nanosheet, (c) atomic force microscope (AFM) image and (d) a height profile of the few-layer GaS nanosheet, (e) the optical image of the device, and (f) the schematic of the device operation. Source: Yang et al. [46]. Copyright 2014, The Royal Society of Chemistry.

quaternary structures (e.g. CuInS2 and CuInx Ga1−x S2 ). In view of this, Yu and coworkers designed Cu2−x S polymorphs (Cu1.94 S–CuS) for the first time [47]. Unprecedentedly, the formation of Cu1.94 S–CuS heterojunction in dumbbell-like architecture combined one-dimensional (1D) Cu1.94 S with 2D CuS (Figure 12.36). Thanks to the particular heterointerface constructed by intrinsic band alignment and the enhanced contact between high conductivity hexagonal planes and the working electrode, this self-coupled Cu1.94 S–CuS heterojunction structure exhibited significantly enhanced photoelectrochemical properties compared with the individual Cu1.94 S nanocrystals and CuS nanoplates. This designed structure anticipating emerging interfacial charge separation could provide useful hints for sulfide semiconductor applications in optoelectronic devices or photocatalysis. Among sulfide conductive materials, silver sulfide compounds are highly conducting semiconductors with excellent electrical properties. Chen and coworkers have in detail analyzed the chemical bonding in α-Ag2 S crystal structure and revealed systems of planes with relatively weak atomic interactions (Figure 12.37). Due to the silver diffusion, silver–silver and sulfur–silver bonds are irregularly distributed. Hence, their α-Ag2 S suppressed the cleavage of the material, and thus resulted in unprecedented ductility. Experiment demonstrated that this α-Ag2 S semiconductor

12.2 Structure and Electrical Properties of Inorganic Conductive Materials

200 Cu1.94S CuS

CuS

PDOS

150

100 Cu1.94S

50 S Cu

0 –2

–1

0

1

e– h+

2

E – Ef (eV)

(a)

(b)

Before contact 0

Vac

Vac

Φ = 4.56

–1

Energy (eV)

After contact

Φ = 5.24

–2 –3

Ec1 –4

Ec2

Eg = 0.77

Ef2

–6

CuS

Cu1.94S

(c)

–

–

+

+

+

Eg = 1.18

Ef1

–5

–

– – + – + +

–

–

–

+

+

+

CuS

Cu1.94S

Figure 12.36 (a) Calculated partial density of states for the Cu1.94 S–CuS heterojunction, (b) simulated charge distributions at the Cu1.94 S–CuS heterojunction interface, and (c) the schematic illustration of the band alignment and charge separation at the interface of the self-coupled Cu2−x S polymorphs. Source: Han et al. [47]. Copyright 2016, American Chemical Society. 102

Elongation (%)

Metals 101 Ag2S 100

–1

10

Fe3

Insulators Si3N4 AlN ZeO2 BaTiO3 Al2O3 BeO SiC

10–20

(a)

Fe Ti Pt Cu Mg

10–15

10–10

Si

GaAs Ge SiC a

ZnSe Semiconductors

100 10–5 Conductivity (S/m)

105

1010

S

Ag

b

(b)

Figure 12.37 The sulfide semiconductor α-Ag2 S: (a) elongation versus electrical conductivity for α-Ag2 S and various materials, (b) the perspective view of the α-Ag2 S crystal structure along the [001] direction. Source: Shi et al. [48]. Copyright 2018, Springer Nature.

703

704

12 The Relationship Between Structure and Electric Property

exhibited extraordinary metal-like ductility with high plastic deformation strains at room temperature [48]. Their work opened up the possibility of searching for ductile sulfide conductive materials for flexible electronic devices. In conclusion [48], these aforementioned inorganic conducting materials, as a batch of most vital semiconductors, have factually better electrical properties than organic materials, although they seem to be less diverse in structures. However, recent studies have shown that, with the deepening research of inorganic conducting materials from the bulk to 2D structure, inorganic materials reveal rich forms and bonding systems in structure, and they also show plenty of applications in functions with the transformation of structures. In the future, with the assistance of theoretical calculation and artificial intelligence, the systems of inorganic conductive materials will be further expanded. Many fresh and undiscovered physical and chemical characteristics will be revealed and verified.

12.3 Structure and Electrical Properties of Organic Conductive Materials Organic electronics is an interdisciplinary research field concerning the design, synthesis, characterization, and application of organic small molecules or polymers that show desirable electronic properties such as conductivity, semiconductivity, and even superconductivity. The first organic conductor can be traced back to 1954 when perylene–bromine complex was demonstrated to show electrical conductivity as high as 0.1 S/cm [49], totally changing the traditional view that organic materials cannot be utilized as conductors due to very low conductivity. After that, a lot of organic conductors have been achieved based on organic charge transfer cocrystals and doped conjugated polymers. It is noteworthy to refer to the discovery of tetracyanoquinodimethane (TCNQ) complexes and doped polyacetylene with a metallic conductivity up to 104 –105 S/cm in the 1970s. Generally, all CT complexes and conducting polymers are “salts” consisting of multiple components for two reasons; charge carriers are stabilized by the counter ions to maintain their charge neutrality, and they provide strong intermolecular interactions delocalizing conducting electrons. These findings suggested that appropriate molecular design and doping can tune the conductivity, which makes organic conjugated materials fascinating for potential applications [50]. Semiconductivity is another important feature of organic materials, which have attracted a lot of attentions in the recent three decades due to their promising applications in organic field-effect transistors (OFETs), organic light-emitting diodes (OLEDs), organic photovoltaic cells (OPVs), and their related devices/circuits. Among the various solid states, organic crystalline materials, especially organic single crystals, are particularly intriguing because of the absence of grain boundaries, long-range periodic order, as well as minimal traps and defects. Hence, organic single crystals provide a powerful tool for revealing the intrinsic properties, examining the structure–property relationships, demonstrating the important factors for high performance devices, and uncovering fundamental physics in organic conductors and semiconductors. Over the past decades, significant

704

12 The Relationship Between Structure and Electric Property

exhibited extraordinary metal-like ductility with high plastic deformation strains at room temperature [48]. Their work opened up the possibility of searching for ductile sulfide conductive materials for flexible electronic devices. In conclusion [48], these aforementioned inorganic conducting materials, as a batch of most vital semiconductors, have factually better electrical properties than organic materials, although they seem to be less diverse in structures. However, recent studies have shown that, with the deepening research of inorganic conducting materials from the bulk to 2D structure, inorganic materials reveal rich forms and bonding systems in structure, and they also show plenty of applications in functions with the transformation of structures. In the future, with the assistance of theoretical calculation and artificial intelligence, the systems of inorganic conductive materials will be further expanded. Many fresh and undiscovered physical and chemical characteristics will be revealed and verified.

12.3 Structure and Electrical Properties of Organic Conductive Materials Organic electronics is an interdisciplinary research field concerning the design, synthesis, characterization, and application of organic small molecules or polymers that show desirable electronic properties such as conductivity, semiconductivity, and even superconductivity. The first organic conductor can be traced back to 1954 when perylene–bromine complex was demonstrated to show electrical conductivity as high as 0.1 S/cm [49], totally changing the traditional view that organic materials cannot be utilized as conductors due to very low conductivity. After that, a lot of organic conductors have been achieved based on organic charge transfer cocrystals and doped conjugated polymers. It is noteworthy to refer to the discovery of tetracyanoquinodimethane (TCNQ) complexes and doped polyacetylene with a metallic conductivity up to 104 –105 S/cm in the 1970s. Generally, all CT complexes and conducting polymers are “salts” consisting of multiple components for two reasons; charge carriers are stabilized by the counter ions to maintain their charge neutrality, and they provide strong intermolecular interactions delocalizing conducting electrons. These findings suggested that appropriate molecular design and doping can tune the conductivity, which makes organic conjugated materials fascinating for potential applications [50]. Semiconductivity is another important feature of organic materials, which have attracted a lot of attentions in the recent three decades due to their promising applications in organic field-effect transistors (OFETs), organic light-emitting diodes (OLEDs), organic photovoltaic cells (OPVs), and their related devices/circuits. Among the various solid states, organic crystalline materials, especially organic single crystals, are particularly intriguing because of the absence of grain boundaries, long-range periodic order, as well as minimal traps and defects. Hence, organic single crystals provide a powerful tool for revealing the intrinsic properties, examining the structure–property relationships, demonstrating the important factors for high performance devices, and uncovering fundamental physics in organic conductors and semiconductors. Over the past decades, significant

12.3 Structure and Electrical Properties of Organic Conductive Materials

advances have been achieved in this field from aspects of controllable growth of high-quality organic single crystals, modulation of conductivity, investigation of structure–property relationship, as well as device applications [50, 51]. In this context, an overview of some representative fantastic advancements in terms of crystal growth methods, organic crystal conductors, and semiconductors as well as organic co-crystals is presented along with current challenges and future research directions provided finally.

12.3.1 Growth Methods and Packing Arrangements of Organic Single Crystals In contrast to inorganic materials that are connected together through strong ionic and chemical bonds, organic materials are connected through weak van der Waals intermolecular interactions, which makes them be difficult to grow into large-sized high quality single crystals. Moreover, the diversified molecular structures and intermolecular interactions also significantly affect the molecular crystallization behavior including crystal quality, shapes, sizes, molecular arrangements, as well as their resulting properties. Thus, how to grow high-quality organic single crystals by optimizing the various parameters in molecular crystallization for desirable habits is crucial to investigate the structure–property relationship and reveal the intrinsic charge transport behaviors. 12.3.1.1 Organic Small Molecule Crystals

Currently, physical vapor transport (PVT) technique and solution assembly process are two typical methods widely used for the growth of organic single crystals, each of which has their own advantages and disadvantages and application ranges. For instance, PVT method is usually used for the growth of organic small molecule crystals that are featured with low molecular weight and easy sublimation under vacuum and even atmospheric pressure, but cannot be used for polymer materials. The mechanism and steps of PVT are illustrated in Figure 12.38 [52]. First, the precursor powders are heated into vapors at center region and then the vapors are transported into a low temperature zone by gas and condensate on substrate. Once condensates are populated, transported vapors are likely to be saturated in the condensates and recrystallized into specific shapes. Adjusting processing parameters like carrier gas flow rate, precursor temperature (T p ), and sample collection temperature (T s ) can yield the different morphology and molecular orders. In general, the low deposition rate and high substrate temperature tend to decrease the nucleation density, thus achieve the high domain size because at higher substrate temperature, the a−1

Ar

High T zone (sublimation)

Low T zone (crystallization)

Pump

Figure 12.38 Schematic view of PVT apparatus. Source: Tang et al. [52] Copyright 2018, John Wiley & Sons.

705

706

12 The Relationship Between Structure and Electric Property

deposited molecules have sufficient energy to overcome substrate interactions and thereby prefer to stand on the substrate (edge-on orientation) and make it possible to self-assemble into large domains, while lower deposition rate provides enough time to the molecules to rearrange themselves and form condensed packing modes with high order [53]. Solution-processing methods open up another possibility to grow the crystals with the poor thermal stability and large molecular weights, which cannot be achieved from PVT method [54]. Developing from the traditional recrystallization method, scientists have exploited several modified methods to grow organic semiconductors crystals including drop-casting method, dip-coating method, solvent-exchange technique, solvent-vapor diffusion method, and solution-shearing method. There are three approaches: (i) slow evaporation of solvents; (ii) changing the temperature slowly to adjust the solubility of the organic semiconductor (normally, decreasing the temperature will result in lower solubility); (iii) slow introduction of poor solvents into the solutions to obtain crystals with high quality from solution [50]. The wetting issue between small molecules and substrates is also a big problem. Hence, the factors that influent crystallization process are very complicated. Different solvent, evaporation rate, temperature, pressure, and concentration will affect significantly packing arrangements, morphologies, as well as properties even if the molecular structures are the same. In this section, we illustrated several typical solution-processing methods involving drop-casting method, dip-coating method, and solution-shearing method and how different factors affect the qualities, morphologies, and properties of semiconductors. Drop-casting method (Figure 12.39a) is a solution-processing technique to grow crystals without external force, and therefore it is usually regarded as a self-assembly process. Generally, solvent with a high boiling point is favorable for the ordered molecular packing arrangements due to the slow self-assembly process. Meanwhile, high solubility may prevent the precipitation until only a small amount of solution is left, leading to polycrystals, while it is unlikely to obtain crystals or large crystals from solvents with low solubility. Furthermore, the concentration of solution is another important factor. Generally, dip-coating methods are used to produce thin films. As illustrated in Figure 12.39b, the substrate was first immersed into the solution and then pulled out slowly. Due to the introduction of external force, the thin films will be orientated in the same direction [55]. Too fast evaporation rate (high temperature, low boiling point) may result in the clusters along the contact line [50]. Too fast pulling speed may lead to discontinuous films while too slow pull speed will comprise the uniformity of crystals. The crystals yielded from such process have much poorer quality than that from drop-casting process because the external force facilitate the solvent evaporation [50]. As shown in Figure 12.39c, solution shearing is a highly versatile coating technique where a movable top shearing blade holds an organic semiconductor (OSC) solution droplet above a temperature-controlled substrate [57]. The solutions evaporation and the formation of thin films happen with blade moving relative to substrate at a fixed speed. The intermolecular π–π stacking distance can be tuned by moving speed and shape of blade, leading to dramatically enhanced charge transport.

12.3 Structure and Electrical Properties of Organic Conductive Materials Shearing direction

1 Solution

Precipitated crystals

Shearing plate

2

Semiconductor solution

3

Nucleation site Initial contact line Soluble acene crystal

PTS-treated SiO2/Si (stationary) Heated bottom substrate

Propogating crystaline film

(a)

(b)

(c)

Figure 12.39 Typical schema of drop-casting (a), dip-coating (b), and solution-shearing (c) methods. Source: (a) Wang et al. [50]. Copyright 2018, The Royal Society of Chemistry. (b) Jan et al. [55]. Copyright 2012, John Wiley & Sons. (c) Giri et al. [56]. Copyright 2012, Springer Nature.

12.3.1.2 Packing Arrangements in Organic Crystals

Not only molecular structures but also intermolecular interactions determine packing structures, which play an important role in optoelectronic properties by tailoring electronic couplings, band structures, as well as exciton behaviors. Two representative aryl· · ·aryl synthons involving aromatic C–H· · ·π and π–π interactions greatly govern the arrangements of organic conjugated molecules together with other geometrical and chemical recognition factors. Four typical packing motifs discovered for organic small molecule crystals are summarized in Figure 12.40 (i) the herringbone packing motif without π–π overlap dominated by C–H· · ·π interactions between neighboring molecules (e.g. pentacene); (ii) the cofacial herringbone packing motif controlled by both C–H· · ·π interactions and π–π interactions between neighboring molecules (e.g. rubrene); (iii) one-dimensional slipped stacking motif attributed to π–π interactions (e.g. C8-PTCDI); (iv) two-dimensional brickwork motif resulting from π–π interactions and substituent effects (e.g. tips-pentacene) [58]. Therefore, organic semiconducting molecules can manifest either one or both of aromatic C–H· · ·π and π–π interactions. The preferred packing arrangement depends on the weight of contribution of each interaction term (i.e. dispersion, electrostatic, induction, and exchange interactions) and the steric hindrance driven by the molecular structure [59]. Generally, the herringbone packing motif without π–π overlap usually give rise to 2D morphology, whereas the cofacial herringbone packing motif can lead to either 1D or 2D morphology depending on the balance between C–H· · ·π interactions and π–π interactions. Obviously, 1D slipped stacking motif tend to form 1D morphology in contrast to 2D brickwork motif with a preference to 2D morphology.

12.3.2 Structure–Property Relationship in Organic Conductor Crystals The requirements for high conductivity and even for superconductivity have been well reviewed in terms of molecular structures, packing structures, energy levels, and the degree of charge transfer in recent two decades [56]. Thus in this section, we just provide a brief summary of the recent important development in the field of organic conductors.

707

708

12 The Relationship Between Structure and Electric Property

B2 t1

B1

t1

t2

B1

C1

A

t2

B2

B2

t2

B1

(a)

t3

t1

A

t1

t3

C2

B1 t2

B2

(b) B2

t2

t1

B1

t2

B2

C2

t3

C1

A t3

t2

t2

B2 t1

B1

B2

(c) B1 B5

t5

B4

t1

t2

A

t4 t3

B2 t6

B6

B3

(d)

Figure 12.40 Schematic images of common molecular packing modes of organic molecules in solid state and the possible charge transport pathways in each case. (a) Typical herringbone packing with charge transport dominated by the aromatic edge-to-face interactions; (b) cofacial herringbone packing with strong face-to-face interactions between neighboring molecules, where a new effective pathway for charge transport is added compared to typical herringbone packing; (c) 1D slipped packing with -stacking, demonstrating strong anisotropic electron transport along the -stacking direction; (d) brickwork packing with efficient 2D network charge transport. Source: Dong et al. [53]. Copyright 2013, John Wiley & Sons.

12.3.2.1 Doped Systems

Doping is a general method to improve the conductivity of organic and polymeric materials, which can result in high conductive and superconductive materials. However, the control of doping process and the selection of dopants for different material systems are crucial [51]. In 2018, Zhen et al. reported that the benzoporphyrin molecules spontaneously form 1D nanofibers with cofacial stack in chloroform containing 1% trifluoroacetic acid (TFA), showing acid-responsive 1D conductivity as high as 1904 S/m (Figure 12.41) [60]. A small fraction (2.7%) of BP in the fiber exists in a cation radical state, and 1.5 equiv of TFA is located in an intercolumnar void. Dedoping and redoping of TFA with trimethylamine vapor results in 1300- to 2700-fold decreases and increases in the conductivity. The conductivity of the benzoporphyrin fiber also shows a correlation with the pK a of acid dopants. Although a variety of stable molecular p-dopants have been developed and successfully utilized in devices in the past decade, air-stable molecular n-dopants suitable for materials with low electron affinity are still elusive. A general approach to this issue involves air-stable precursors that liberate strongly reducing species during their insertion into the host, or through bond cleavage and/or formation accompanies electron transfer. The first method is exemplified by

12.3 Structure and Electrical Properties of Organic Conductive Materials

NH N

Relative conductivity ratio

Top view

Side view

BP

N HN NH N N HN

3.27 Å

+• NH N N HN NH N

14.54 Å

N HN

BP:TFA = 100 : 145 BP:BP+• = 100 : 2.7 High conductivity (1904 S/m)

BP+•

101 100

TFA

10–1 10–2 10–3 10–4 10–5

TEA 0

1 2 3 4 5 Cycle number

TFA = CF3COOH TFA– = CF3COO–

200 nm

Figure 12.41 Spontaneous formation of a nanofiber of tetrabenzoporphyrin that shows acid-responsive conductivity. Source: Zhen et al. [60]. Copyright 2018, American Chemical Society.

[RuCp* Mes]2

10–3

POPy2

POPy2 Ru H

2

Ru+

H Ru

–2.4

–2.0

–2.67 D+/D•

–2.04 D+/0.5D2

–1.6

[RuCp* Mes]+

–2.24 A/A•–

After irradiation, in dark

10–5 10–6 10–7

Befor irradiation, in dark

10–8 Ru/O = 0.20

10–9 –1.2

σ measured at constant 50 V σ extracted from IV-sweep

–1.10 D2•+ /D2

[RuCp* Mes]2

Irradiation

10–4 Conductivity (S/cm)

P O

E (V versus FeCp2*/FeCp2)

–2.8

–0.8

10–10 0

40

80 120 Time (h)

160

Figure 12.42 Molecular structure and electrochemical redox potentials of the host and the dopant that show conductivity after doping as a function of time, before, during, and after UV irradiation. Source: Lin et al. [53]. Copyright 2017, Springer Nature.

certain cationic dyes, from which the corresponding neutral radicals (and/or other highly reducing species) sublime when heated in vacuum. The second method involves several classes of materials including (i) molecules that can be regarded as hydride reduction products of stable organic cations; (ii) tetraalkylammonium salts of simple inorganic ions such as halides; and (iii) dimers formed by certain 19-electron organometallic sandwich compounds or organic radicals [61]. Lin et al. demonstrate that photoactivation of a cleavable air-stable dimeric dopant (pentamethylcyclopentadienyl)(1,3,5-trimethylbenzene)ruthenium dimer ([RuCp* Mes]2 ) can result in kinetically stable and efficient n-doping of host semiconductors, whose reduction potentials are beyond the thermodynamic reach of the dimer’s effective reducing strength (Figure 12.42) [61]. The initial increase in conductivity is fast although complete saturation to a photo-equilibrium level under the experimental conditions used here takes a few hours. Electron-transport layers doped in this manner are used to fabricate high-efficiency OLEDs. 12.3.2.2 Single-Component Systems

The design of single-component organic metals beyond the conventional salt framework has been a challenging subject. Kobayashi et al. report a single component electroactive molecule, zwitterionic tetrathiafulvalene (TTF)-extended

709

12 The Relationship Between Structure and Electric Property

dicarboxylate radical (technology, entertainment, design [TED]), which exhibits conductivities of 530 S/m at 300 K and 1000 S/m at 50 K with metal-like electronic properties as shown in Figure 12.43 [62]. Spectroscopic and theoretical investigations of the carrier-generation mechanism and the electronic states of this single molecular species reveal a unique electronic structure with a spin-density gradient in the extended TTF moieties that becomes, in itself, a metallic state only by an intermolecular interaction of single molecular species with s and p electrons. In 2018, Zhu and coworkers prepared a highly crystalline copper(II) benzenehexathiolate (Cu-BHT) coordination polymer with a two-dimensional kagome structure, exhibiting bulk superconductivity at about 0.25 K, as shown in Figure 12.44 [63]. Another diamagnetic transition at about 3 K suggested a second superconducting phase that may be associated with a single layer or few layers of Cu-BHT. It is the first time that superconductivity has been observed in a coordination polymer.

12.3.3 Structure–Property Relationship in Organic Semiconductor Crystals In this section, we will give a brief summary of this comprehensive overview of molecular structure, packing arrangement, and charge transport characteristics of –0.199 a.u.

3.0

+0.245 a.u.

2.5

O S

H

S

C C

C

C

H

C S

S

O C

C

C H

S

S

C

C

C S

S

O C O

7.12 D

(a)

𝜌(×10–3 Ω cm)

2.0 1.5 1.0 0.5 0.0

0

100

(b)

200

300

T (K)

Figure 12.43 Spin density, dipole moment (a), and temperature dependence of resistivity of tetrathiafulvalene-extended dicarboxylate (TED) (b). Source: Reproduced with permission Ref. [54]. Copyright 2017, Macmillan Publishers Limited, part of Springer Nature.

1.0

C S Cu

20 nm

0.8 0.2 0.0

(a)

(b)

(c)

1.0 0.8 0.6 0.4 0.2 0.0

R(T)/R(300 K)

7.38 Å

R(T)/R(300 K)

710

0.0 0.1 0.2 0.3 0.4 0.5 T(K)

0

50 100 150 200 250 300 T(K)

Figure 12.44 (a) Structure characterization of Cu-BHT, (b) HRTEM image of Cu-BHT nanosheet, the upper right inset is the FFT dashed box marked area in HRTEM image, lower right is the corresponding inverse FFT images, (c) temperature dependence of the normalized resistance R(T)/R(300) from 50 mK to 300 K. Inset: expanded scale for temperatures near the superconducting transition. Source: Reproduced with permission Jang et al. [55].

12.3 Structure and Electrical Properties of Organic Conductive Materials

semiconductive organic small molecule, which are expected to give a clear picture of the state-of-art status and guide future work in this area. 12.3.3.1 Linear-Shaped Molecules

Generally, the linear-shaped molecules without substituents like acenes, oligophenyls, and oligothiophenes tend to self-organize into a herringbone packing motif without π-overlap (Figure 12.45a) as a consequence of the minimization of exchange repulsion by reducing intermolecular π-overlap, while the total energy for the herringbone packing is stabilized mainly through C–H· · ·π interactions. In a herringbone packing motif, a molecule is surrounded by six electronically and structurally relevant adjacent molecules. Due to such well-balanced interactions, the linear molecules with herringbone packing motif are prone to forming a 2D or quasi 2D morphology. For example, Hu and coworkers prepared plate-like microcrystals based on 2,6-diphenylanthracene (2,6-DPA) by PVT method, which showed herringbone packing without π-overlap but with strong C–H· · ·π interactions, resulting into an excellent FET mobility as high as 34 cm2 /V/s [65]. Furthermore, they introduce naphthyl units at 2,6-positions of anthracene to achieve 2,6-di(2-naphthyl)anthracene (2,6-DNA), which also adopts herringbone packing without π-overlap although anthracene core is coplanar with end-substituted naphthyl units in contrast to 2,6-DPA with a twisted conformation. Single-crystal field-effect transistors show mobility up to 12.3 cm2 /V/s. Organic light-emitting transistors (OLETs) based on 2,6-DNA single crystals distribute outstanding balanced ambipolar charge transporting property (𝜇 h = 1.10 cm2 /V/s, 𝜇 e = 0.87 cm2 /V/s) and spatially controllable emission, which is one of the best performances for OLETs [66]. In heteroacenes with low C/H ratio, the C–H· · ·π interactions still is sufficient to balance exchange repulsion, such as [1]benzothineno[3,2-b]benzothiophene (BTBT), dibenzo[d,d′ ]thieno[3,2-b;4,5-b′ ]dithiophene (DBTDT), dinaphtho[2,3-b:2′ , S

Oligoacene H2n+1Cn

S S

n–1

n–1

Oligophenyl

n–1

CI S

Oligothiophene

S

BTBT

Rubrene

DCT

DNTT

CI

(a)

DNSS

Si

Pentacene

Si

(b)

S S

Si

TIPS-ADT

TES-PEN

CI

CI

F

TIPS-pentacene

TT1

MCT

Si

Si

Si

PTA

CI

TCT

S

Si

CI

MCT

PTA

S

(c)

S S

S

S

CI

S

CnH2n+1

S S

S

S

Si S

S

N

N

F N

Si

diF-TES ADT

N Si

TIPS-TAP

TIPS-pentacene

(d)

Figure 12.45 Representative linear molecules for typical herringbone packing (a), cofacial herringbone stacking (b), 1D slipped packing (c), and 2D brickwork packing (d). Source: Yu et al. [64]. Copyright 2019, Elsevier.

711

712

12 The Relationship Between Structure and Electric Property S CnH2n+1 S mono-Cn-BTBT

Intermolecular interaction energies BTBT–BTBT

BTBT–Chain

(kJ/mol) –40 e1 e6

e6

e1

e3

e4

e6

e6

0

e5

e3

–20

e2

e6

e3

e5

0

e5

e4

e3 e4

e5 e4

200 μm

200 μm n=3

–60 –40 –20 0

e1 e2

e3 e4

n=

(a)

n=2

Total

–40 e1 e2

–20

0

e2

e5

e2

–20

Chain–Chain

–40 e1

:0

:4

:9

200 μm n=4

n=9

n = 12

n = 14

(b)

Figure 12.46 The relationships between intermolecular interaction energy and length of alkyl chain, and corresponding photographs. (a) Molecular packing motif and calculated intermolecular interaction energies of mono-Cn-BTBT (n = 0, 4, 9). (b) Photographs of typical crystals of mono-Cn-BTBT obtained by recrystallization from solution from 1,2-dichlorobenzene. Source: Reproduced with permission Ref. [67]. Copyright 2017, American Chemical Society.

3′ -f ]thieno[3,2-b]thiohene (DNTT) [53, 67]. The increase of C/H ratio prevent the C–H· · ·π interactions and facilitate the transition of packing arrangements to a 1D slipped packing motif with π-overlap as shown in the case of PTA (pentathienoacene) as shown in Figure 12.45b [58]. Introduction of side groups will give rise to the steric effects, hindering the C–H· · ·π interactions. Besides, increased energy stabilization via direct substituent interaction between intrastack molecular pairs allows the offset face-to-face molecular arrangements possible (Figure 12.45c, d). One strategy to construct face-to-face packing motifs is to incorporate bulky groups at aromatic systems. Particularly, those groups with half diameter of backbone are most favorable for the formation of 2D brickwork packing arrangements (Figure 12.45d). If the length of the substitution is not suitable, a slipped π-stacking may end up being adopted (Figure 12.45c) [58]. Simultaneously, alky groups also can increase the intralayer interaction energy and lower the interlayer interaction energy, thus leading to a 2D morphology. For example, Minemawari et al. found that the intermolecular interaction energy of monolinear-alkyl substituted BTBTs increased with increasing chain length in 2017 (Figure 12.46a) [68]. Consequently, mono-Cn-BTBTs with n ≥ 4 tend to afford fairly thin flake crystal while mono-Cn-BTBTs with n = 2 or 3 provide to form 1D belt morphology as shown in Figure 12.46b. Incorporation of halogen atoms, cyanoimide, ester, or other substituents into the parent backbone can increase the C/H ratio, reducing C–H· · ·π interactions and then enhancing π–π interactions. Pei and coworkers modified the intermolecular displacements and distances of five benzodifurandione-based oligo(p-phenylenevinylene) (BDOPV)-based small molecules by tuning the

12.3 Structure and Electrical Properties of Organic Conductive Materials

S

S

S

S

S

S

S

S

S

S

S

S

TTF

DBTTF

DNTTF

Figure 12.47 Representative TTF derivatives and corresponding packing structures. Source: Yu et al. [64]. Copyright 2019, Elsevier.

amounts and/or the positions of the substituent fluorine atoms, resulting in different packing structures from slipped stacking to cofacial herringbone to antiparallel cofacial stacking. The electronic couplings for electron transfer can vary from 71 meV in a slipped stack to 201 meV in a nearly cofacial antiparallel stack, leading to an increase in the electron mobility of the BDOPV derivatives from 2.6 to 12.6 cm2 /V/s [69]. Li and coworkers found that 6,13-dichloropentacene (DCP) adopted slipped π–π stacking model with C–H· · ·Cl interactions. The single crystalline ribbon transistors displayed extremely high mobility up to 9.0 cm2 /V/s. As far as we know, this is one of the highest mobility reported for pentacene derivatives [70]. TTF derivatives show a different packing arrangement from acenes. With increasing number of side-fused phenyl rings, S· · ·π interactions together with C–H· · ·π interactions predominate over S· · ·S or C–H· · ·S interactions, altering packing arrangements from cofacial herringbone or twisted packing for TTF to displaced herringbone for dibenzotetrathiafulvalene (DBTTF) to typical herringbone motif without π-overlap for dinaphthotetrathiafulvalene (DNTTF) (Figure 12.47) [71]. The thin film transistor based on DNTTF showed a mobility up to 0.42 cm2 /V/s, six times higher than that of DBTTF [72]. Hu and coworkers obtained controllably the parent compound TTF crystals of α- and β-phases by drop-casting method with the selection of different solvent. The field-effect mobility of α-phase single crystals reached up to 1.2 cm2 /V/s, while β-phase single crystals showed the maximum mobility as low as 0.23 cm2 /V/s. The superior charge transport behavior of α-phase single crystals could be attributed to the strong π–π interactions along the short b-axis and the short contacts between S atoms [71a]. 12.3.3.2 Cylinder-Like or Disk-Like Molecules

In contrast to linear-shaped molecules, cylinder-like or disk-like molecules can adopt sandwich herringbone, cofacial herringbone, or 1D slipped stacking that mainly depend on the number and positioning of C and H atoms (Figure 12.48). Porphine and tetrathia[22]annulene[2,1,2,1] (TTA) exhibit sandwich herringbone packing, while pyrene and perylene can adopt either sandwich or cofacial herringbone packing depending on the balance between π–π and C–H· · ·π interactions. With increasing molecular size as well as C/H ratio, 2D molecules like HBC

713

714

12 The Relationship Between Structure and Electric Property

S S

HN

NH S

S

N

N

S

S

N

S S

N

Porphine

TTA

Perylene

Pyrene

PET

Se

DITT

TPBIQ S

S

S

S

PES TTA

Sandwich herringbone

Perylene

Cofacial herringbone

DPTTA

DITT

1D slipped packing

Figure 12.48 Transition from sandwich herringbone to cofacial herringbone to 1D slipped packing depending on the number and positioning of C and H atoms. Source: Yu et al. [73]. Copyright 2019, Elsevier Inc.

(hexabenzocoronene), TPBIQ (tetraphenylbis(indolo[1,2-a]quinoline), MPc (metallophthalocyanines), and DITT (diindenodithienothiophene) prefer to self-organize into face-to-face arrangements with cofacial herringbone or 1D slipped stacking, typically showing 1D natures in both electronic and morphological features due to the higher intrastack interaction energy than lateral interstack interactions [53, 58]. Similar to linear-shaped molecules, the introduction of imide groups, aryl units, alkyl, or alkoxy chains into the 2D aromatic backbone hinder C–H· · ·π interactions but promote π–π interactions, inducing the transition from edge-to-face to face-to-face arrangements. For example, perylo[1,12-b,c,d]thiophene (PET) and potential energy surface (PES) are two kinds of chalcogen annulated perylenes, which adopt cofacial herringbone stacking and promote the formation of double-channel superconstructure for charge transport by chalcogen· · ·chalcogen contacts. Such molecular packing is totally different from the sandwich herringbone packing of its parent compound perylene. As a result, the single-crystal FET devices based on PET and PES exhibited hole mobilities up to 0.8 and 2.6 cm2 /V/s, respectively [53]. In contrast to TTA with sandwich herringbone stacking, meso-diphenyl substituted TTA (DPTTA) exhibits 1D slipped stacking, significantly improving charge transport properties with the mobility up to 0.65 cm2 /V/s [72]. Introducing alkyl, alkoxy, ester chains into the cores of triphenylene, perylene, HBC, porphyrin, phthalocyanine, or other macrocyles can lead to the formation of parallel face-to-face packing with Colh , Colr , or other columnar liquid crystal phases, which displayed promising carrier mobilities as well as other optoelectronic performances [74]. Introducing polar substituents like cyano groups into the backbones can give rise to the dipole-induced interstack interactions, thus tuning the morphology from 1D to 2D as shown in the case of N,N ′ -bis(n-octyl)dicyanoperylene-3,4:9,10-bis(dicarboximide) (PDI8-CN2) and N,N ′ -bis-(2-ethylhexyl)-1,7-dicyanoperylene-3,4:9,10-bis (dicarboximide) (N1400) [67]. With the degree of halogenation increasing, the core distortion becomes more and more severe. For example, Cl8-PTCDI (Figure 12.49) containing eight Cl atoms make a large distorted backbone, leading to a 2D brick packing with the help of π–π and N–H· · ·O interactions [77].

12.3 Structure and Electrical Properties of Organic Conductive Materials

b

O

N

N

C8H17

1 (210)

a

O O C8H17

N

2 (100)

c N

O

z

PDI8-CN2

(a)

z

b y x

a

3 (110)

x

y

Face-to face arrangement CI

O

CI CI

NH

CI

O

O

CI

HN

CI O

Hydrogen bonding

CI

CI

CI8-PTCDI

(b)

Figure 12.49 Molecular structure and packing arrangement of PDI8-CN2 (a) and Cl8-PTCDI (b). Source: (a) Rivnay et al. [75]. Copyright 2009, Springer Nature. (b) Park et al. [67, 76]. Copyright 2018, John Wiley & Sons.

12.3.3.3 Bowl-Shaped Molecules

Incorporation of five-membered rings into sp2 carbon hexagonal π-systems enables formation of aromatic bowls, as exemplified in the cases of corannulene and sumanene, which are the smallest C3v - or C5v -symmetric fragment molecules for the structural motifs of fullerenes. The curvature of bowl-shaped molecules makes their solid-state packing very interesting but complex. The types of molecular packing are summarized in Figure 12.50. In type A and quasi A, all columns are oriented toward the same direction, leading to polar crystals. With the surface area and depth of the bowl increasing, 1D columnar stacking motif tend to be

(a) Corannulene

Sumanene

μ>0

(b)

μ>0

(c)

μ=0

(d)

μ=0

(e)

Figure 12.50 Typical molecular structures and packing arrangements for bowl-shaped polycyclic aromatic systems. (a) Schematic depiction for the generation of corannulene and sumanene. Source: Hou et al. [78] Copyright 2016 Elsevier. (b–e) Representative packing arrangements for bowl-shaped molecules. Source: Yu et al. [64]. Copyright 2019, Elsevier.

715

716

12 The Relationship Between Structure and Electric Property

formed because of stronger intermolecular interactions. The dipole moment of a buckybowl is an important factor to yield 1D columnar stacks. It is noteworthy that the periphery substituents can strongly affect crystal packing arrangements. Whereas the parent corannulene shows a highly disordered packing structure dominated by C–H· · ·π interactions, the introduction of imide, trifluoromethyl, cyano, or other groups at the rim positions tends to form a regular columnar stack based on bowl bowl-in-bowl π–π interactions [73]. Out-of-plane coordination with central heteroatoms of macrocycles provides another way to make the bowl-shaped structures, e.g. boron subphthalocyanine (subPc), boron subporphyrin, titanylphthalocyanine (TiOPc), and vanadylphthalocyanine (VOPc). Most metal phthalocyanies like CuPc are planar structure in a cofacial herringbone packing while TiOPc and VOPc exhibit a shallow square-pyramid structure with the oxygen atom out of the conjugated plane in a 2D brickwork packing motif. Due to the such close concave π–π distance (d1 (π–π) = 3.211 Å, d2 (π–π) = 3.145 Å) for both convex and concave pairs, TiOPc showed a high mobility of 26.8 cm2 /V/s in single crystal [76].

12.3.4 Structure–Property Relationship Beyond Single Component Compared with single-component organic materials, integrating two or more components together is an effective strategy to combine various electrical and optical properties synergistically, circumventing the issues of complex design and synthetic procedure. In this section, we focus on the properties of organic cocrystals and solid solutions as well as their applications in optoelectronic devices. 12.3.4.1 Organic Cocrystals

Cocrystals combine two or more different components in the same lattice through weak van der Waals forces, charge–transfer interactions, π–π interactions, hydrogen bonds, and halogen bonds, exhibiting novel and multifunctional properties such as ambipolar charge transporting, nonlinear optics, and white light emission [73]. In this subsection, we will give an introduction about factors which significantly determine the properties of donor–acceptor (D–A) cocrystals and then we will provide some representative examples to illustrate the relationships between packing and properties based on cocrystals, according to the classification of packing modes involving 1 : 1 mixed-stack cocrystals, 1 : 1 segregated stacking cocrystals, and other special cocrystals, e.g. 1 : 2 and 1 : 3. 12.3.4.1.1 Factors That Determine the Properties of D–A Cocrystals

The key factors relative to the properties of D–A complexes are extremely complicated because the precursors will not only introduce their intrinsic characteristics but also make a large difference in crystal stacking structures and degree of charge transfer as well as band structure. For common 1 : 1 cocrystals, there are two packing modes: mixed stacking and segregated stacking, as illustrated in Figure 12.51. In mixed-stack modes (· · ·DADA· · ·), the donors and acceptors alternate face to face, while donors and acceptors separately arranged in segregated-stack modes

12.3 Structure and Electrical Properties of Organic Conductive Materials

Figure 12.51 Schematic illustration of D–A cocrystals with different structural modes. Source: Zhang et al. [79]. Copyright 2017, American Chemical Society.

Mixed stacking Segregated stacking

1:1

2:1

3:1

Acceptor Donor

(DDDD· · ·AAAA). It is still difficult to predict which packing mode can be adopted although the molecular structures, molecular shape, and energy levels have been considered to be the important factors. We deduce if face-to-face intermolecular interactions between donors and acceptors predominate over those existed in donor or acceptor themselves, the mixed stack would be the favorable one. On the contrary, the segregated stack tends to be formed when homo-molecular face-to-face interactions surpass hetero-molecular interactions. Interestingly, some D–A pairs can take both packing modes, where the mixed stack might be the thermodynamically stable phase occurring more frequently and the segregated stack might be in a metastable state [64]. It is verified theoretically and experimentally that weak electronic coupling between the donor and acceptor leads to integer charge transfer (ion formation), while partial charge transfer originates from strong electronic coupling. The mixed stack between organic donors and acceptors usually results in a lower degree of charge transfer and poorer conductivity compared with the segregated stack [80]. Ratio rather than 1 : 1 is rare in cocrystals where extra donor or acceptor molecules exist either within the stack or lie within the interstitial space between stacks [80]. The final highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) of the D–A complexes can be tuned by stoichiometric variation between the donors and acceptors. Degree of charge transfer increases if more than one donor contributes charges to one acceptor, thus significantly influencing the optoelectronic properties [56a]. 12.3.4.1.2 1 : 1 Mixed-Stack Cocrystals

Generally, most 1 : 1 mixed-stack complexes tend to align into 1D single crystals due to their strong intermolecular D–A π–π interaction. Although the nearest-neighbor D (or A) molecules in the stacking column are too far away from each other to interact directly, super exchange effects result from the hybridization of the frontier orbitals of the two closest donor molecules with the orbitals of the effects resacceptor molecule, thus promoting both hole and electron transport along the alternatingly stacked DA columns. For example, Zhang et al. used a drop-casting

717

12 The Relationship Between Structure and Electric Property

tdirect

tsuper

tdirect

(a)

tsuper

(b) 5.0 × 10–9

2.0 × 10–4

0V 15 V

VD8 = –60V

1E–8

VD8 = +60V

1.6 × 10–4

8.0 ×

10–5

0V

ID (A)

1E–9

(IID|)1/2 (A1/2)

1.2 × 10–4

30 V

45 V 60 V

0.0

ID (A)

718

–5.0 × 10–9

–15 V –30 V

–45 V

4.0 × 10–5 1E–10

–60 V 0.0 –60

–40

–20

0

20

40

60

–1.0 × 10–8 –60

VG (V)

(c)

(d)

–40

–20

0

20

40

60

VD (V)

Figure 12.52 Electronic couplings (a,b) and electrical characteristics (c,d) of DPTTA-TCNQ cocrystals. Source: Zhang et al. [79]. Copyright 2017, American Chemical Society.

method to prepare microribbon-shaped DPTTA-TCNQ cocrystals, where DPTTA and TCNQ stack alternatingly into a 1D column structure along the a-axis. The intermolecular distance is approximately 3.4 Å, showing a strong π–π interaction between DPTTA and TCNQ (Figure 12.52). The micro-cocrystal-based device showed air-stable balanced hole- and electron-transport behaviors in ambient atmosphere with mobilities of 0.03 and 0.04 cm2 /V/s for the holes and electrons, respectively. TCNQ derivatives with different electron-deficient substituents including 2,5-difluoro-7,7,8,8-tetracyanoquinodimethane (F2TCNQ) were investigated to construct 1 : 1 complexes with DPTTA. The corresponding complexes all cocrystallized in a mixed-stack mode along the a-axis. The superexchange coupling increased and reached a value of 85–91 meV in the mixed stack of F2TCNQ-DPTTA cocrystals, which showed hole and electron mobilities as high as 1.57 and 0.47 cm2 /V/s, respectively. A clear tendency was observed where the superexchange-induced electronic coupling increased with the degree of intermolecular CT of the DA pairs. Park et al. developed a series of highly luminescent and ambipolar transporting CT complex cocrystals based on an isometric D–A approach using distyrylbenzene as the backbone structure. By introducing dipole-inducing –CN groups into both D and A molecules, strong interstack local dipole interaction was promoted, leading

12.3 Structure and Electrical Properties of Organic Conductive Materials

Mixed-stack

(i)

87°

3.2 Å 3.4 Å

b

(ii)

O C

a

(iii) VG 60 V, VD 100 V 25 μm

(a) ID EQE

ID

1.5%

1E–6

0.14%

EQE

0.01

0.1

1E–8 0.01

EQE (%)

1E–8

EQE (%) ID (A)

1E–7

–ID (A)

1

0.1

1E–7

1E–9 1E–3 1E–9 –80

(b)

–60

–40

–20

0

20

1E–10 –20

VG (V)

0

20

40

60

80

VG (V)

Figure 12.53 Molecular arrangement (a), device optical microscopy (b), and optoelectronic performances of distyrylbenzene-based CT cocrystal (c–d). Source: Reproduced with permission Kobayashi et al. [62]. Copyright 2017, WILEY-VCH.

to the formation of few-millimeter scaled 2D CT cocrystal with only a few hundred nanometers in thickness as depicted in Figure 12.53. This CT complex cocrystal showed unprecedentedly high luminescence efficiency and OLET performance [59]. In addition, Hu and coworkers prepared a series of mixed-stack cocrystals based on halogen bond or CT interactions by a simple drop-casting method, with promising applications in white light emission, nonlinear optics, as well as optical waveguides [81]. For example, the cocrystals based on Spe (4-styrylpyridine) and TCNB (1,2,4,5-tetracyanobenzene) exhibited two-photon absorption property in contrast to single component material Spe or TCNB [82]. This phenomenon was mainly contributed to the close bonding interactions in the cocrystals providing the electronic delocalization while the individual molecules cannot provide the adequate electronic delocalization due to their herringbone packing without π–π interaction. 12.3.4.1.3 1 : 1 Segregated-Stack Cocrystals

1 : 1 Segregated-stack complexes have exhibit a varieties of properties such as electrical conductance, ambipolar charge transport, optical waveguide, and photothermal conversion. High degree of charge transfer in segregated-stack cocrystals can lead to high conductivity, whereas low degree of charge transfer gives rise to semiconducting behaviors. The individual electron- and hole-transport pathways this motif provides transport networks and determine effective transfer

719

720

12 The Relationship Between Structure and Electric Property

N

Ha

N

O Hb N

NH a

N

N R

N

O

N H a

O

DP-P2P:NDI-81 (a)

Ha

O

R = C8H17

DP-P2P:NDI-CyHex

O Hb N O

O N Hb O

Ha N N Ha N

DP-P2P:NDI-H2

R = C6H11

(b)

Figure 12.54 Cocrystals based on DP-P2P and NDI. (a) H-bonded coassemblies of DP-P2P with NDI. (b) Packing arrangement of DP-P2P:NDI-81. p-type DP-P2P channels (green) and n-type NDI channels (blue) are highlighted. Source: Black et al. [82]. Copyright 2014, WILEY-VCH.

integrals, respectively. Zhang et al. grew the 2D segregated-stack cocrystals by a drop-casting method using DPTTA as the donor molecule and C60 and C70 as the acceptor molecules. The DPTTA conformation slightly deviated from planarity with half of the molecule bent up and the other half bent down to fit to the curved surface of the fullerenes. There are efficient π–π interactions along both DPTTA and fullerene columns. The plate-shaped cocrystals, especially for C70 -4,4′ ,4′′ tri(phenothiazinyl)triphenylamine (TPTTA), showed balanced ambipolar transport properties of 𝜇 e = 0.05 cm2 /V/s and 𝜇 h = 0.07 cm2 /V/s. Hydrogen bonding can be utilized to enhance ground state charge–transfer interactions in the solid state and tailor the D–A dual channel of cocrystals. For example, Perepichka and coworker introduced diphenyldipyrrolopyridine (DP-P2P) as a new donor semiconductor capable of complementary hydrogen bonding with naphthalenediimide (NDI) acceptors [82]. Altering the substituents on the NDI nitrogen controls the supramolecular arrangement, into either segregated or mixed stacks of the donor and acceptor components (Figure 12.54). The segregated-stack cocrystal based on DP-P2P–NDI showed relatively balanced ambipolar transport properties with hole and electron mobilities of 0.043 and 0.089 cm2 /V/s, respectively. Furthermore, segregated-stack cocrystals, where donors and acceptors form molecular-scale heterojunctions with nanoscale interpenetrating network while

12.3 Structure and Electrical Properties of Organic Conductive Materials

AI Si

J (mA/cm)

0.0

Au

/S

–0.2 Dark

iO

Light

–0.4

2

0.0

(a)

(b)

(c)

0.2

0.4

0.6

Voltage (V)

Figure 12.55 Organic photovoltaic device based on DPTTA-C60 cocrystals. (a) Segregated-stack arrangement existed in the cocrystal packing structure. (b) Schematic device configuration. (c) Photovoltaic characteristics of the device. Source: Zhang et al. [83]. Copyright 2016, John Wiley & Sons.

maintaining the continuous pathways of the hole and electron transport in their respective domains, can serve as an ideal model system for studies of charge separation and transport. Zhang et al. studied the photovoltaic (PV) properties of DPTTA-fullerene cocrystals, and an overall power conversion efficiency (PCE) of 0.27% for the C60 -DPTTA cocrystal solar cell was observed [83]. This was the first molecular-scale heterojunction solar cell based on segregated-stack cocrystals, which provides an ideal model system for quantum simulations of OPVs (Figure 12.55). In addition, Hu and coworkers prepared segregated-stack cocrystals based on halogen bond or CT interactions by a simple drop-casting method, with promising applications in light emission, optical waveguides, as well as photothermal conversion [81d, 84]. For example, DBTTF and TCNB are selected as electron donor and acceptor to self-assemble into new cocrystals with a NIR charge transfer absorption due to the strong D–A interactions. Under a NIR laser illumination, the temperature of the cocrystal sharply increases with high photothermal conversion efficiency of 18.8%, opening a avenue to the development of novel potential transformer (PT) materials based on organic cocrystals [84]. 12.3.4.1.4 Other Special Cocrystals

There are some investigations about the preparation of nonequal ratio complexes and their properties. For example, the P1T1 and P3T1 cocrystals also were synthesized through low evaporation of the mixed solution by precisely adjusting the concentration of perylene and TCNQ [85]. P1T1 possessed a mixed-stack arrangement along a-axis, whereas the molecular structure of P3T1 consists of donor–acceptor stacks of the form (…–A–D–D–A–D–D–A–…) along the c-axis with a perylene molecule inserted between the stacks. And the OFET devices showed an ambipolar charge transporting property with 𝜇 e = 2.1 × 10−5 cm2 /V/s and 𝜇 h = 0.03 cm2 /V/s for P3T1. The three cocrystals with different D/A ratio ranging from 1 : 1 to 3 : 1 based on perylene and TCNQ were also prepared by McNeil and coworkers through PVT method (Figure 12.56) [86]. A D–A complex in which the novel asymmetry donor DTDTP (2,7-di-tert-butyl-10,14-di(thiophen-2-yl)phenanthro[4,5-abc][1,2,5]thiadiazolo[3,4-

721

722

12 The Relationship Between Structure and Electric Property

c b

c a

b

b

P1T1 (a)

c

a

a

P3T1

P2T1 (b)

(c)

Figure 12.56 Molecular packing of perylene–TCNQ complexes with different stoichiometries. P1T1 with the 1 : 1 ratio of perylene to TCNQ (a), P2T1 with the 2 : 1 ratio of perylene to TCNQ (b), and P3T1 with the 3 : 1 ratio of perylene to TCNQ (c). Source: Vermeulen et al. [86]. Copyright 2017, American Chemical Society.

i]phenazine) co-crystallized with TCNQ in D/A ratio of 2 : 1 adopted a tight DDADDADDA mixed linear column structure [87]. Only the n-type behavior was observed during the characterization of the co-crystal devices due to the large transfer integral for the electrons along π–π interactions (19.58 meV) and direction with five-membered rings of DTPTP:DTPTP dimers approaching each other (25.76 meV). Benefiting from self-assembly induced solid-state “olefin metathesis” reaction, for the first time Pei and coworkers observed the cocrystallization of three BDOPV derivatives with antiparallel cofacial stacking or slipped stacking, just starting from two-component or monocomponent precursors, e.g. BDOPV/F6-BDOPV or F3-BDOPV. Consequently, as the total number of halogen atoms increased, energy levels of the cocrystals decreased monotonously in the range from −5.94 to −6.96 eV and −4.19 to −4.48 eV, respectively [88]. In summary, although such large developments have been achieved, there are several long-standing challenges in this emerging research. Second, the relationships between crystal structures and the resulting performance and the CT degree within cocrystals connecting to lattice stability, optoelectronic, and photophysical properties, as well as the mechanism of electronic interaction between building blocks remains elusive which greatly impedes the properties regulation of the cocrystals. 12.3.4.2 Solid Solution

Solid solutions that are used for solid phases that have the characteristics of a solution are widely known for a mixture of two metals neighboring on the periodic table, but they have seldom been explored for organic solids, inter alia, in organic electronics research probably due to the difficulty in designing a crystalline mixture of different molecules without changing the crystal lattice of host molecules. There are two common types of solid solutions: (i) interstitial solid solution where a guest molecule can be included in interstices within the structures of host molecule and (ii) substitutional solid solution where the guest molecules can randomly place the host molecules, as illustrated in Figure 12.57 [89]. The solid solutions remain the same crystal lattice as host molecules in contrast to cocrystals which generally will form a new crystal structure compared to the components. In crystalline solid solutions, stoichiometry affects both structural and physicochemical properties. This behavior is exemplified by Vagard’s law, which correlates composition and unit cell dimensions. Other features that vary regularly with

12.3 Structure and Electrical Properties of Organic Conductive Materials

(a)

(b)

Figure 12.57 Representation of common types of solid solutions. Interstitial solid solution (a) and substitutional solid solution (b). Source: Lusi [89]. Copyright 2018, The Royal Society of Chemistry. SiiPr3

SiiPr3 X N

X

N

X X

SiiPr3

1a: X = H 1b: X = CI

SiiPr3

2

(a)

Melting point (°C)

190 180 170 160 150 140 (b)

2

(2)0.85(1a)0.15

(c)

1a

Fluorescence quantum yield

14%

0.5 1 x in (2)x(1a)1–x

Ex: 535 nm

12%

Ex: 570 nm

10% 8% 6% 4% 2% 0% 0

(d)

0

0.2

0.4 0.6 0.8 x in (2)x(1a)1–x

1

Figure 12.58 A solid solution toward ternary-component photovoltaics. (a) Structures of 1 and 2; (b) a plot of melting point of (2)x (1a)1−x versus x; (c) crystals of 2, (2)0.85 (1a)0.15 and 1a under UV light (365 nm, left) and corresponding plot of fluorescence quantum yield of (2)x (1a)1−x versus x. Source: Copyright, The Royal Society of Chemistry 2014.

composition include thermal stability (melting point and inversion point between polymorphs, etc.), electronic, or optical properties such as light polarization, absorbance–emission, and fluorescence [89]. For example, solid solution 1a/2 with different ratio was synthesized by Miao and coworkers, and the molecular structure are illustrated in Figure 12.58 [90]. They found that the (2)x (1a)1−x solid solution has a melting point close to that of 2 when x > 0.5, while the (2)x (1a)1−x solid solution has a melting point close to that of 1a when x < 0.5 and adding c. 22% of 1a in 2 significantly increased the fluorescence quantum yield from 0.2% to 12.3%. Simultaneously, discontinuous variations are also possible; then, unexpected properties can emerge. These characteristics of crystalline solid solutions provide

723

12 The Relationship Between Structure and Electric Property CI CI Si

Si

NH N

O

Insoluble BP

0

AI

BP

ITO

–4 BP:CABP= 100 : 0 75 : 25 50 : 50 25 : 75 0 : 100

–6 –8 0

(b)

SIMEF2

BP:CABP:SIMEF2

–2

(e)

(f)

(g)

(h)

(i)

N HN

0.2 0.4 0.6 Voltage (V)

Normalized absorbance

(a)

(d) NH N

N HN

Current density (mA/cm2)

724

Insoluble CABP

1 75 : 25 50 : 50

0.8 0.6

25 : 75

0.4

BP:CABP= 100 : 0 0.2

0 : 100

ITO

0 680

0.8

BP:CABP BP PEDOY:PSS

700

720 740 760 Wavelength (nm)

780

800

(c)

Figure 12.59 A solid solutions based on benzoporphyrin derivatives toward ternary solar cell (a) chemical structures of BP, SIMEF2, and CABP. (b) J–V curves of ternary blend BHJ cells under amplitude modulation (AM) 1.5 illumination at 100 mW/cm2 . (c) Absorption spectra of the BP:CABP blend with The BP:CABP blend ratio ranging from 100 : 0 to 0 : 100; SEM images. (d–i) For the blend of BP and CABP on the BP layer after removal of SIMEF2 by washing with toluene. Top view images at the BP:CABP blend ratios of 100 : 0 (e), 75 : 25 (f), 50 : 50 (g), 25 : 75 (h), and 0 : 100 (i), and a cross-sectional view of the 75 : 25 blend (j). Scale bar is 200 nm. Source: Xu et al. [90] Copyright 2015, American Chemical Society.

an opportunity for crystal engineering and for the development of functional materials [89]. For example, Zhen et al. reported a photovoltaic cells based on ternary mixture of SIMEF2 (dimethyl(o-anisyl)silylmethyl)(dimethylphenylsilylmethyl)-[60] fullerene), BP (tetrabenzoporphyrin), and CABP (dichloroacenaphtho[q]tribenzo [b,g,l]porphyrin) [91]. The 75 : 25 BP:CABP blend showed a very smooth surface and glass-like interior reminiscence of a molten solid in contrast to other ratio of BP:CABP blends as shown in Figure 12.59. Besides, the red-shifted absorption toward 750 nm was observed with increasing the BP content due to intercalation of CABP into the BP matrix, where the acenaphthylene moiety of CABP enables J-aggregate with BP for red-shifted light absorption. Therefore, the OSC containing BP and CABP in a ratio of 75 : 25 showed the highest PCE of 2.7%, with a J sc of 8.0 mA/cm2 . 12.3.4.2.1 Conclusions and Outlook

In this part, we have discussed systematically how to design organic conductors and semiconductors from the perspective of single component and beyond single component. We have underlined the relationships among molecular structures, packing modes, and optoelectronic properties. We have emphasized the impact of structure modifications on intermolecular interactions, which can predominate the packing arrangement, thus influencing significantly charge transport behaviors and optical properties. In addition, we have present a review of important advancements on crystal engineering beyond single component such as organic cocrystals and organic solid solutions. Inspite of a lot of advancements on organic optoelectronic materials, there are still several key challenges in the future. First, the structure–property relationship needs

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials

to be further elucidated, particularly for n-type and ambipolar semiconductors, which far lag behind p-type counterpart in term of material varieties and electrical characteristics. Second, in-depth understanding of how to control the crystal nucleation and growth pathways is important to achieve the preferred crystal plane, polymorph, and morphology strongly associated with optoelectronic properties. Last, there is a plenty room to investigate crystal engineering beyond single component such as organic cocrystals, solid solutions, ternary or higher grade multicomponent molecular materials with multifunctional optoelectronic properties.

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials 12.4.1 Structure and Electrical Properties of MOFs The metal–organic framework (MOF), with periodic structure and highly regularized pores, is crystalline hybrid materials composed by organic linker bridged metal ions or inorganic clusters [92]. Flexible design and synthesis of the organic bridging ligand combining diverse metal and inorganic clusters provide a powerful tool to produce limitless new framework structure. Wide structure variety endows MOF with material versatility and rapid accelerating interest in applying this material to gas storage, separations, catalysis, nonlinear optics, luminescence, and magnetism. The realization of MOFs as electronic conductors is a highly sought-after goal. The design and synthesis of MOFs that exhibit through-framework conduction have been limited; however, people have a keen interest in them owing to the fascinating prospects for integrating multiple functions [93]. Conductive MOF materials could be split into protonic (ionic) and electronic conductive MOFs. In this chapter, we will focus on the electronic conductivity. MOFs as an electron conducting material have promising merits: (1) The crystal structure of MOF enables us to precisely regulate the structure from the atomic scale and affords us a good opportunity to study and comprehend the fundamental structure–conductivity relationship, which also can be optimized through theoretical calculation to obtain better conductivity; (2) The inherent electrical properties can easily be tuned via structure designing; (3) The combination of conducting nature with porosity may create novel conducting materials which are ultralight, ultra large on the surface area, allowing easy doping, nonnative functionality, and ionic electronic mixed conductivity to innovate the current research on Li ion battery, the fuel cell, the solar cell, the field effect transistor, and other nano-based electrical devices [93]. Carboxylates are the most frequently used ligands in MOFs, which have been broadly utilized to construct MOFs. However, the σ-bond nature between the carboxylate group and metal hampers strong electron communication among metals or inorganic clusters, which causes the majority of known MOF materials to be insulators or to show low conductivity. It is still a big challenge to obtain MOF materials with highly conducting nature.

725

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials

to be further elucidated, particularly for n-type and ambipolar semiconductors, which far lag behind p-type counterpart in term of material varieties and electrical characteristics. Second, in-depth understanding of how to control the crystal nucleation and growth pathways is important to achieve the preferred crystal plane, polymorph, and morphology strongly associated with optoelectronic properties. Last, there is a plenty room to investigate crystal engineering beyond single component such as organic cocrystals, solid solutions, ternary or higher grade multicomponent molecular materials with multifunctional optoelectronic properties.

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials 12.4.1 Structure and Electrical Properties of MOFs The metal–organic framework (MOF), with periodic structure and highly regularized pores, is crystalline hybrid materials composed by organic linker bridged metal ions or inorganic clusters [92]. Flexible design and synthesis of the organic bridging ligand combining diverse metal and inorganic clusters provide a powerful tool to produce limitless new framework structure. Wide structure variety endows MOF with material versatility and rapid accelerating interest in applying this material to gas storage, separations, catalysis, nonlinear optics, luminescence, and magnetism. The realization of MOFs as electronic conductors is a highly sought-after goal. The design and synthesis of MOFs that exhibit through-framework conduction have been limited; however, people have a keen interest in them owing to the fascinating prospects for integrating multiple functions [93]. Conductive MOF materials could be split into protonic (ionic) and electronic conductive MOFs. In this chapter, we will focus on the electronic conductivity. MOFs as an electron conducting material have promising merits: (1) The crystal structure of MOF enables us to precisely regulate the structure from the atomic scale and affords us a good opportunity to study and comprehend the fundamental structure–conductivity relationship, which also can be optimized through theoretical calculation to obtain better conductivity; (2) The inherent electrical properties can easily be tuned via structure designing; (3) The combination of conducting nature with porosity may create novel conducting materials which are ultralight, ultra large on the surface area, allowing easy doping, nonnative functionality, and ionic electronic mixed conductivity to innovate the current research on Li ion battery, the fuel cell, the solar cell, the field effect transistor, and other nano-based electrical devices [93]. Carboxylates are the most frequently used ligands in MOFs, which have been broadly utilized to construct MOFs. However, the σ-bond nature between the carboxylate group and metal hampers strong electron communication among metals or inorganic clusters, which causes the majority of known MOF materials to be insulators or to show low conductivity. It is still a big challenge to obtain MOF materials with highly conducting nature.

725

726

12 The Relationship Between Structure and Electric Property

In order to obtain high conductivity in MOF materials, the frontier orbitals of metal and ligands must be effectively overlapped, and the charge carrier among structure components needs to be efficaciously delocalized. To realize a greater degree of orbital overlap, the organic linker designed as the conjugated system with suitable connectors whose pπ orbitals can overlap with the dπ orbital of the metal center is extremely favored. Meanwhile, a redox-active bridging ligand contained in MOF structure is another prerequisite to obtain better electron transport properties. Moreover, in order to construct porous structure, organic linkers should feature high rigidity for supporting the framework structure, appropriate length for preventing interpenetration, multi topics for extending the dimension of the structure, and so on. To settle all of aforementioned problems, elaborate organic linker design and synthesis are required. Fortunately, at least to some extent, modern organic chemistry has built powerful synthesis methodology, which can rationally control inherent properties of MOFs through organic linker modification. In this chapter, the inorganic and organic components of the MOF materials are linked by coordination bonds, and we classify them in accordance with the dimensions of their structures.

12.4.1.1 Three-Dimensional (3D) Metal–Organic Frameworks Through Different Linkers 12.4.1.1.1 Carboxylates

MOF-5 is a very classic and famous carboxylic acid-based MOF material with 3D framework connected by terephthalate and Zn4 O13 clusters (as shown in Figure 12.60) [94]. This MOF is deemed as a semiconductor based on its photoluminescence spectrum with its bandgap of ∼2 eV [95]. Although the terephthalate ligand may “sensitize” the Zn4 O13 quantum dots in MOF-5 structure, these luminescent centers remain electrically isolated, and thus the crystal is an excellent insulator. M.D. Allendorf and coworkers found that ZnO nanoparticle may jointly form with MOF-5 during synthesis, which maybe the source of the anomalously low optical bandgap. Other factors, such as trapped solvent and partial lattice collapse, can also affect the spectrum of luminescence, indicating that other methods apart from optical spectroscopy are required to precisely determine the electrical properties of MOF-5 [96]. Conductivity measurement by drop casting a MOF-5 single crystal on an Au-coated slide and electrically contacting its top with a conductive probe inside a scanning electron microscope (Figure 12.61) shows its insulating nature [97]. Even the bias was over 150 V, no current (pA resolution) was measured, even after bombardment with a 1 kV electron beam. Notably, at such high bias, contact resistance does not contribute substantially to the observed transport characteristics. For the low conductivity in carboxylate-based MOFs, two main methods were used to improve this property. One is the “through-space” approach based on π-stacking interactions between electroactive moieties. The other is “through-bond” approach, where both symmetry and energy overlap between the covalently bonded components must exist to promote good charge transport. According to above

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials

Figure 12.60 The crystal structure of MOF 5 shown as ZnO4 tetrahedral joined by benzene dicarboxylate linkers. Source: Modified from Nathaniel et al. [94].

(a)

(b)

I/V MOF Au SANDIA 1.0 KV

X300

100 µ m

Figure 12.61 (a) Experimental configuration for measuring electrical conductivity of MOF single crystals, (b) the SEM image of MOF-5 crystal probed in SEM. Source: Allendorf et al. [97]. Copyright 2011, John Wiley & Sons.

two approaches, the conductivity of carboxylate-based MOF material has been significantly improved. Several interesting examples have been shown as follows: MOF Mn2 (DSBDC) [98] (H4 DOBDC = 2,5-dihydroxybenzene and 1,4-dicarboxylic acid) was acquired by M. Dinc˘a and coworkers using the “through-bond” method, which resembles M2 (DOBDC) (MOF-74) [99]. Different from the structure in MOF-74, the OH group in DOBDC was replaced by an –SH group (Figure 12.62a). Because sulfur can alleviate the orbital energy mismatch between metal and oxygen and improve the electrical properties of MOFs, in this structure, Mn2+ is coordinated by carboxylate oxygen atoms and sulfur atoms from thiophenoxide groups to form an infinite (–Mn–S–)∞ chain which is further bridged by DSBDC4− ligands to form a 3D framework. Notably, this infinite (–Mn–S–)∞ chain endows Mn2 (DSBDC) with high mobility for intrinsic charge transport [98]. The charge mobility of Mn2 (DSBDC) and methanol-exchanged and activated Mn2 (DSBDC) are 0.02 and 0.01 cm2 /V/s, respectively (Figure 12.62b,c). Another “through space” approach is found in [Zn2 TTFTB(H2 O)2 ]⋅H2 O⋅2DMF (TTFTB = tetrathiafulvalene tetrabenzoate) [100]. As a building block, TTF has its

727

12 The Relationship Between Structure and Electric Property

HO

C

O

O

OH

M

HO O OH H4DOBDC

ϕ∑μ (10–5 cm2/V/s)

1.6

Methanol-exchanged Activated

1.2

0.8

0.4 0.0

(b)

0

2

HO

O

6

8

C SH

HS O OH H4DSBDC

(a)

4 Time (μs)

Methanol-exchanged Activated

10–4

?

M

S

O

Current (A)

728

10–5

10–6

10–7 0

(c)

2

4 Time (μs)

6

8

Figure 12.62 (a) Conceptual design of MOFs containing (–M–S–)∞ chains obtained by replacing phenol groups in M2 (DOBDC) with thiophenol groups, (b) conductivity transients observed by FP TRMC, and (c) photocurrent transients observed by time of flight (TOF) for the methanol exchanged and activated sample. Source: Sun et al. [98]. Copyright 2017, American Chemical Society.

high propensity to π-stack that would balance well with the stronger driving force of the metal–ligand bond formation, thereby enforcing the construction of a porous material that will provide a pathway for efficient charge transport. Thereby, the material based on TTFTB exhibits both porosity and a pathway for efficient charge transport. [Zn2 TTFTB(H2 O)2 ]⋅H2 O⋅2DMF shows a permanently porous structure with an apparent surface area of 662 m2 /g. The closest intermolecular S· · ·S contact is 3.803(2) Å between neighboring TTF moieties (Figure 12.63a,b). The intrinsic charge mobility of [Zn2 TTFTB(H2 O)2 ]⋅H2 O⋅2DMF is 0.2 cm2 /V/s, which is higher than that found in those common organic conductors (polyphenylenevinylenes, 0.01–0.1 cm2 /V/s; polythiophenes, 0.015–0.075 cm2 /V/s) (Figure 12.63c,d). This work also suggests that if the molecular overlap between neighboring TTF cores can be engineered to either increase the number of S· · ·S contacts or shorten the length of the contact, MOFs with even higher charge mobility may be available. By using the identical ligand but varying the center metal cation (Mn, Co, Zn, or Cd), the authors found a striking inverse correlation between the size of metal cations and the shortest S· · ·S interaction between neighboring TTF cores in their further studies [101]. Larger cations will cause a severe pinching of S· · ·S contact, which further affects the orbital overlap between pz orbitals on neighboring S and C atoms. The result of density functional theory (DFT) calculations shows that pz orbitals on neighboring S and C atoms are critically involved in the valence band of these materials. Hence, the band dispersion and conductivity will be varied through modulating the S· · ·S distance. Among these four cations, the Cd, with the largest size and shortest S· · ·S contact, presents the highest electrical conductivity, 𝜎 = 2.86 (±0.53) × 10−4 S/cm (single crystal and two probes, Figure 12.63e), which is 72 times

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials 10–8

ϕ∑μ (m2/V s)

(a) 3.8 Å

10–9

10–10 2

(b)

(c)

2.86 × 10–4

10–3

(d)

8

10

10–6

10–5 Time (s)

Cd2(TTFTB)

2.0 × 10–4

8.64 × 10–5 1.49 × 10–5 0.0 3.95 × 10–6

10–4

10–5

2 4 6 Time (10–6 s)

4.0 × 10–4

σ(S/cm)

Current (A)

10–2

0

10–4

3.64 (e)

Mn2(TTFTB) Co2(TTFTB) Zn2(TTFTB)

3.68 3.72 3.76 S…S distance (Å)

3.80

Figure 12.63 (a) The infinite helical Zn carboxylate chains, (b) a side view of a helical TTF stack with a depiction of the shortest intermolecular S· · ·S contact, (c) conductivity transients observed by FP TRMC upon, (d) photocurrent transients observed for 20–26 μm thick solid films of materials, and (e) the correlation between S· · ·S distance and electrical conductivity in M2 (TTFTB). Source: Refs. [100, 101]. Copyright 2012 American Chemical Society; Copyright 2015, American Chemical Society.

higher than the average conductivity of Zn2 (TTFTB) (𝜎 = 3.95 (±0.56) × 10−6 S/cm). Mn2 (TTFTB) and Co2 (TTFTB), with intermediate S· · ·S distances between the four cations, show intermediate conductivity values of 8.64 (±1.21) × 10−5 and 1.49 (±0.29) × 10−5 S/cm, respectively. Another approach to modify the conductivity is introducing guest molecules into the pores of MOFs to construct the conductive path. [Zn3 (DLlac)2 (pybz)2 ]⋅2.5DMF (pybz = 4-pyridylbenzoate) [102] shows uptake and the release of iodine (Figure 12.64) reported by Zeng and coworkers I2 was imbedded into the channel and used as a conductive pathway for the charge carrier. Electrical conductivity measurement on single crystals of [Zn3 (DLlac)2 (pybz)2 ] ⊃ 3I2 was taken. The 𝜎|| and 𝜎⟂ values were found to be around 3.42 × 10−3 and 1.65 × 10−4 S/cm, respectively. These values were 440 times higher than the value for I2 (7.69 × 10−6 S/cm). The significant increase compared with solid I2 can be ascribed to the restriction of polyiodide ions within a well-regulated aromatic nanochannel, which induces a high efficiency of n to 𝜎 * charge transfer (CT). The introduction of TCNQ into the channel of Cu3 (BTC)2 (HKUST-1) is another extremely successful example. Cu3 (BTC)2 possesses a porous structure with

729

12 The Relationship Between Structure and Electric Property

O7

O10

O6

O9

a

O1 Zn2

O3

O4

11.2 Å

O5 Zn1

Zn3 O8

10.2 Å

O2

c

11.5 Å

730

N1

N2

(a)

(b)

b 11.5 Å

a

10.5 Å

(c)

I2

(d)

Figure 12.64 (a) Coordination environment of Zn atoms in [Zn3 (DLlac)2 (pybz)2 ]⋅2.5DMF, (b) perspective views of the 3D open framework with 1D channel in [Zn3 (DLlac)2 (pybz)2 ]⋅2.5DMF, the guest DMF molecules being shown in channels, (c) the completely desolvated framework [Zn3 (DLlac)2 (pybz)2 ], and (d) the sketch of I2 molecules diffusing in the channels of [Zn3 (DLlac)2 (pybz)2 ]. Source: Ming-Hua Zeng [102]. Copyright 2010, American Chemical Society.

binuclear copper paddle wheels connected by benzenetricarboxylate (Figure 12.65) [103]. The conductivity of Cu3 (BTC)2 is about 10−7 S/cm for the residual solvent and water present in the framework. When tested under the inert ambience, the conductivity is less than 10−10 S/cm [104]. M.D. Allendorf and coworkers significantly improved the conductivity of Cu3 (BTC)2 thin film by introducing TCNQ into the framework [105]. Through dipping the activated Cu3 (BTC)2 to TCNQ/CH2 Cl2 solution, the open metal site of the paddle-wheel structure in Cu3 (BTC)2 was connected by redox-active molecule TCNQ. Finally, a path for carrier transport was created between Cu3 (BTC)2 and TCNQ. The conductivity increases by 6 orders of magnitude and reaches ∼10−1 S/cm. Further analysis revealed that TCNQ integrates closely with two neighboring copper paddle wheels to create an infinite path for the charge carrier. The introduction of F4 -TCNQ with the higher electron affinity into Cu3 (BTC)2 would not get such high conductivity, because the high electron affinity of F4 -TCNQ inhibits electrons from moving. In addition, infiltration with H4 -TCNQ, which lacks a delocalized π-network, resulted in no measurable increase in conductivity.

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials

Figure 12.65 [Cu3 (BTC)2 (H2 O)3 ] viewed along the cell body diagonal [001].

12.4.1.1.2 2,3-Pyrazinedithiolate (pdt)

S. Takaishi et al. utilized pyrazinedithiolate (pdt) ligand as the building unit to construct a conductive porous MOF Cu [Cu (pdt)2 ] as shown in Figure 12.66a,b [106]. The 3D framework was formed by Cu ions connected by N and S atoms from pdt ligand and showed paramagnetic behavior in the entire temperature range. The size of 1D vacant space in this structure is 3.4 Å × 3.4 Å. The valence in this compound was +2 (CuII [CuII (pdt)2 ] form) confirmed from the distance of Cu—S bonds in structure analysis. There was a charge-transfer (CT) band from CuII [CuII (pdt)2 ] to CuI [CuIII (pdt)2 ] states which was verified by the broad peak at approximately 0.7 eV from the NIR spectrum. The conductivity of CuII [CuII (pdt)2 ] is 6 × 10−4 S/cm at 300 K. Nevertheless, this structure is not steady, and the framework will collapse under desolvation. Cu[Ni(pdt)2 ] (Figure 12.66a,b) using [NiIII (pdt)2 ]− instead of [CuIII (pdt)2 ]− synthesized by M.D. Allendorf and J.R. Long is more stable than Cu[Cu(pdt)2 ]. Cu[Ni(pdt)2 ], an isomorphic compound of Cu[Cu(pdt)2 ], is stabilized up to 120 ∘ C under dynamic vacuum, and has a Brunauer–Emmett–Teller (BET) surface area of 385 m2 /g. The conductivity of Cu[Ni(pdt)2 ] is 1 × 10−8 S/cm at indoor temperature, and this value can increase to 104 orders of magnitude (approximately 1 × 10−4 S/cm) when the activated Cu[Ni(pdt)2 ] film was treated with I2 vapor at 50 ∘ C (Figure 12.66c,d). The increase in conductivity after oxidative doping implies that Cu[Ni(pdt)2 ] is a p-type semiconductor. The conductivity will drop to 5 × 10−5 S/cm when exposed to air for 12 hours, suggesting that the doping can be partly reversed. When doping temperature is at 150 ∘ C, the electrical conductivity for the I2 adsorption–desorption equilibrium was almost not increased. The conduction occurs primarily through the MOF rather than I2 for the amount of I2 doping is very small. These results demonstrate that: (i) changing the coordination metal can tune the bandgap and (ii) doping guest molecule can also adjust the bandgap.

731

12 The Relationship Between Structure and Electric Property

Cu(2)

a b

Conductivity (S/cm)

Cu(1)

10–5 10–6

0

10–3

150 °C 50 °C

I2 on

10–7 500

1000 Time (s)

(c)

150100 50

0 °C

–50

2000

–100

Doped at 50 °C

10–5 10–7

1500

Doped at 150 °C Undoped

10–9

10–11

a

(b)

I2 on

10–4

c

(a)

Conductivity (S/cm)

732

b

3

(d)

4

1/T (K–1)

5

6 × 10–3

Figure 12.66 (a) Crystal structure of Cu[Cu(pdt)2 ], (b) the perspective view of the crystal structure of Cu[Cu(pdt)2 ], and (c,d) conductivity of Cu[Ni(pdt)2 ] film cast on Pt interdigitated electrodes. Source: Takaishi et al. [106]. Copyright 2009, American Chemical Society.

12.4.1.1.3 Cyanide (CN)

Prussian blue (PB) (Fe4 [Fe(CN)6 ]3 ), as the oldest coordination compound reported in the scientific literature, is transition metal hexacyanide (Figure 12.67), which can be synthesized by mixing ferric (ferrous) and hexacyanoferrate ions: Fe3+ and [FeII (CN)6 ]4− or Fe2+ and [FeIII (CN)6 ]3 . PB has a cubic lattice with the alternating iron(II) and iron(III) locating on a face-centered cubic lattice. The iron(III) ion with high spin is coordinated by six N ligands to form an octahedral configuration, while the iron(II) ion with low spin is coordinated by six C ligands. The conductivity of PB is very low under vacuum environment (∼10–9 S/cm, 303 K), which may derive from the hopping or small-polaron mechanism [108]. Because of the presence of donor and acceptor sites at well-defined positions, PB is an ideal material for studying hopping conduction. Water molecules presenting in the interstices tremendously enhance electronic conduction. Absorption of trace of water can increase the conductivity of PB by several orders of magnitude (the conductivity at 303 K, ∼3 × 10−5 S/cm) [108]. The electrical conductivity of the PB samples heat treated in vacuum at 300 and 350 ∘ C (∼6.10 × 10−7 S/cm for 300 ∘ C and 1.01 × 10−6 S/cm for 350 ∘ C) is higher than that of the samples heat treated at lower temperatures [109], which is probably due to a mixed-valence state created by the partial flipping of the cyano ligands. The partial flipping of cyano ligands can be best described as a linear bridge arrangement Fe(II)–CN–Fe(III)–CN–Fe(II) which is remarkable owing to the multiple mixed valency of the heat-treated samples. The electronic conductivity of PB can be modified by the electrochemical method. The

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials

Figure 12.67 The general crystal framework for PBAs (top) and schematic structure of Fe in PB. Source: Wojdel et al. [107]. Copyright 2019, The Royal Society of Chemistry.

Site B N C

K+

Site A C N Site B

Site A

Site A

Fe2+

Site B

Site B

C

Fe3+ N

N

C

oxidation or reduction processes during electrochemical treatment can be used to dope PB with either electrons or holes. The forms of redox are Prussian yellow and Everitt’s salt, and the conductivity of these two forms has a similar value of around 5 × 10−7 S/cm [110]. Prussian blue analogs (PBAs) have two sorts of cubic crystal structures (Figure 12.68): (i) AI MA II [MB III (CN)6 ], where MA and MB are transition metal ions and alkali ion. A alternatively exists in the interstitial sites, (ii) MA II [MB III (CN)6 ]2/3 ⋅zH2 O, which contains coordinated and noncoordinated water molecules [111]. These analogs have structural flexibility due to the stretching and vibrational modes of the cyano-bridging ligand and are classified as mixed-valence compounds because of the metal-to-metal charge transfer mediated by the cyano ligand. Rbx Mn[Fe(CN)6 ]y ⋅zH2 O, as the PBA, has the two valence-tautomeric forms (MnII FeIII and MnIII FeII ) which are functions of temperature (120–350 K) and frequency (10−2 –106 Hz) [112]. The high-temperature (HT) FeIII (S = 1/2)–CN–MnII (S = 5/2) form of this compound has a face-centered cubic structure (F43m), whereas in the LT FeII (S = 0)–CN–MnIII (S = 2) phase, the crystal structure is tetragonal (I4m2) and the MnN6 octahedra manifesting a Jahn–Teller-type distortion. As shown in Figure 12.69, the electrical conductivity is ∼10−7 S/m at 300 K, which is 2 orders of magnitude higher than that reported for vacuum-dried PB. 𝜎 dc is strongly thermally activated in the investigated temperature range and drops to 10−12 S/m at 180 K. The conductivity displays a large thermal hysteresis loop between ∼220 and 300 K. The hysteresis region on the HT side corresponds well to the magnetic data, but a significant variation is observed on the LT side. Between 210 and 170 K, where the HT phase transforms to the LT phase, the conductivity has a similar value independently if it is recorded in the heating or cooling cycles. In the first instance, one may therefore suppose that there is a crossover in the conduction

733

12 The Relationship Between Structure and Electric Property

Vacacncy-type (Fm3m)

1 : 1: 1-type (Fm3m)

MA||[MB|||(CN)6]2/3·zH2O

A|MA||[MB|||(CN)6]

H O HHH O O HHH O H H O O H H

O HH H O H H O H O HH HH H H O O

O HHH O HH H O (a)

MB|||

MB|||

C

C

N

N ||

MA||

MA

(b)

Figure 12.68 Schematic crystal structures of PBA: (a) MA II [MB III (CN)6 ]2/3 ⋅zH2 O, (b) AI MA II [MB III (CN)6 ]. Source: Tokoro and Ohkoshi [111]. Copyright 2011, The Royal Society of Chemistry. 5.6

Magnetic moment Conductivity (20 V) Conductivity (2 V)

10–7

5.4

10–9

5.2

10–11

5.0

10–13

150

200

250 T (K)

300

350

μeff (μB)

10–5

σdc (S/m)

734

Figure 12.69 Temperature dependence of the dc conductivity (measured at 2 and 20 V bias) and the effective magnetic moment for sample Rbx Mn[Fe(CN)6 ]y ⋅zH2 O. Source: Molnár et al. [112]. Copyright 2009, American Chemical Society.

4.8

mechanism around 220 K and the conductivity becomes rather insensitive to the actual electronic and crystallographic form below the crossover point. Jeffrey R. Long and coworkers reported another two PBAs: Fe4 [Ru(CN)6 ]3 ⋅18H2 O (FeIII RuII ) and K1.2 Ru3.6 [Ru(CN)6 ]3 ⋅16H2 O (RuIII RuII ) [113]. The dehydrated forms of these two samples are microporous with BET surface areas of 670 and 325 m2 /g, respectively. The rise in the degree of electronic localization for the unsymmetrical iron–ruthenium analog. Fe4 [Ru(CN)6 ]3 ⋅18H2 O is mirrored in a shift of the intervalence charge transfer (IVCT) band to the UV-Vis-NIR diffuse reflectance spectra with higher energies (Figure 12.70a). Consequently, it exhibits poor conductivity with 5.05 × 10−6 S/cm at 300 K. Inversely, the all-ruthenium analog K1.2 Ru3.6 [Ru(CN)6 ]3 ⋅16H2 O exhibits a lower-energy IVCT band (Figure 12.70a), as well as the highest electrical conductivity (5.7 × 10−3 S/cm at 300 K and increases to 1.9 × 10−2 S/cm at 350 K, Figure 12.70b), possibly owing to the combined effects of electronic delocalization and the presence of potassium ions. Unlike PB, the ruthenium and iron–ruthenium analogs show no magnetic ordering transition above 1.8 K.

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials 200

–1.5

160

F(R)

120 80 40

Data points Linear fit

–1.8

log σ(S/cm))

RU|||RU||(3) Fe|||RU|| (2) Fe|||Fe|| (1)

–2.1 –2.4

Y = –1084.6x + 1.3751 R2 = 0.9999

–2.7 –3.0 –3.3

0 5000

15 000

25 000

0.0025

35 000

(a)

0.0030

0.0035

0.0040

1/T (K–1)

Energy (cm–1)

(b)

Figure 12.70 (a) Diffuse reflectance spectra for Fe4 [Fe(CN)6 ]3 ⋅14H2 O, Fe4 [Ru(CN)6 ]3 ⋅18H2 O, and K1.2 Ru3.6 [Ru(CN)6 ]3 ⋅16H2 O, (b) temperature dependence of the electrical conductivity of K1.2 Ru3.6 [Ru(CN)6 ]3 ⋅16H2 O. Source: Jogendra et al. [113]. Copyright 2009, American Chemical Society.

The conductivity of PBAs can also be turned by the electric field. Sato et al. found a serious of PBAs with the electric field induced conductance transitions and thermally induced conductivity switching [114]: Na0.5 CoII 1.25 [FeIII (CN)6 ]⋅4.8H2 O, and Rb0.5 CoIII CoII 0.25 [FeII (CN)6 ]⋅5.9H2 O Na0.38 CoII 1.31 [FeIII (CN)6 ]⋅5.4H2 O, III-LS which are correspondingly the Fe CN–CoII-HS , FeIII-LS –CN–CoII-HS , and II-LS III-LS Fe –CN–Co moieties at room temperature and show relatively low conductivity (1.6 × 10−5 S/cm for Na0.5 CoII 1.25 [FeIII (CN)6 ]⋅4.8H2 O, 8 × 10−7 S/cm for Na0.38 CoII 1.31 [FeIII (CN)6 ]⋅5.4H2 O, and 4 × 10−8 S/cm for Rb0.5 CoIII CoII 0.25 [FeII (CN)6 ] ⋅5.9H2 O). The conductivity of these MOFs significantly relates to the transition between FeII-LS –CN–CoIII-LS and FeIII-LS –CN–CoII-HS by external perturbations. For Na0.5 CoII 1.25 [FeIII (CN)6 ]⋅4.8H2 O and Na0.38 CoII 1.31 [FeIII (CN)6 ]⋅5.4H2 O, this transition can happen by varying temperature, while for Rb0.5 CoIII CoII 0.25 [FeII (CN)6 ] ⋅5.9H2 O, it can happen by altering the electric field. For the first time, the authors found that the temperature for phase transition can be continuously tuned by varying the Co-to-Fe ratio in these MOFs, which is distinct from conventional switching materials. 12.4.1.1.4 Nitrogen Containing Heterocyclic Compounds

Nitrogen containing heterocyclic compounds comprise the nitro compound, the amine, quaternary ammonium salt, the diazo compound, the azo compound, the azide compound, etc. Zeolitic imidazolate frameworks (ZIFs) are typical nitrogen containing heterocyclic 3D frameworks which were synthesized under solvothermal conditions [115], but studies on the conductivity of ZIFs are rare. The resistance of ZIF-67 [116] and [Co(Im)2 ]n [117] (Im = imidazolate) powder, the coating on Al2 O3 substrate with Ag–Pd interdigitated electrodes are ∼107 Ω at 150 ∘ C and ∼108 Ω at 75 ∘ C, respectively. Triazole with one more N atom than imidazole, 1,2,3-triazoles, and 1,2,4-triazoles (Figure 12.71) broaden the diversity and utility of metal-triazolate MOFs and can bridge metal ions in five different

735

736

12 The Relationship Between Structure and Electric Property

HN

N

N

1,2,3-Triazoles

HN

N

N N Tetrazole

HN

N

N 1,2,4-Triazoles

MET-1 (Mg)

MET-2 (Mn)

MET-4 (Co)

MET-5 (Cu)

Figure 12.71 Molecular structures of 1,2,3-triazoles, 1,2,4-triazoles, and tetrazole.

MET-3 (Fe)

10 Å

MET-6 (Zn)

Figure 12.72 The illustration of the series of METs. Source: Gandara et al. [119]. Copyright 2012, John Wiley & Sons.

coordination modes [118], which provides infinite possibilities for the design and preparation of metal-triazolate MOFs. As a typical example, semiconducting metabolic equivalent of energy (METs) made from 1H-1,2,3-triazole with different metal can be obtained by a liquid phase method (Figure 12.72) [119]. Authors use the charge-flipping method to deal with the complex crystal structure of METs: all the metal ions are octahedrally coordinated with the nitrogen atoms of triazolate such that five metal centers are joined through bridging triazolate ions to form super-tetrahedral units that lie at the vertexes of a diamond-type structure. The variation in the size of metal ions across the series provides for precise control of pore apertures to a fraction of an Angstrom in the range from 4.5 to 6.1 Å. The electrical conductivity of Fe centered MET-3 was performed in a pressed pellet of the polycrystalline material. MET-3 shows an intrinsic conductivity value of 0.77 × 10−4 S/cm, which can increase to 1 × 10−3 S/cm after exposed to I2 . The increase in conductivity after exposure to iodine might be due to the fact that FeII is oxidized to FeIII , generating MV conductivity, such as Fe3 O4 found in oxides [119].

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials

Tetrazole unit exhibits more coordination modes [118], and various tetrazolebased MOFs have been successfully synthesized in the past 10 years. However, the conductivity in tetrazole-based MOF was not yet found. 12.4.1.2 Two-Dimensional (2D) Metal–Organic Frameworks Through Different Linkers 12.4.1.2.1 TCNQ and Its Derivatives

7,7,8,8-Tetracyanoquinodimethane, commonly referred as TCNQ, was first synthesized by Dupont researchers in the 1960s [120] and has attracted great attention of chemists and material scientists. It is a good electron acceptor molecule, which can easily accommodate extra electrons (one, two, or three) to form the open shell anion radical (TCNQ⋅− ), dianion (TCNQ2− ) (Figure 12.73), and even trianion (TCNQ3− ) [122]. TCNQ can be a tetradentate or bidentate bridging ligand to coordinate with metal, and their MOFs exhibit electrically conductive properties and electrical on/off switching characteristics [123]. After coordinated with metal, neutral TCNQ is reduced to TCNQ− , and this process could be monitored by infrared spectroscopy (IR) [124]. The most reported TCNQ compounds have a 2D structure (such as Cu (TCNQ), Ag (TCNQ)), and so we discuss TCNQ compounds in this part. Dunbar and coworkers reported that Cu (TCNQ) has two polymorphs, phase I and phase II [124], which exhibit interesting electrical properties; polymorphs can be prepared by the solution method, vapor deposition, chemical and electrochemical reduction, and spontaneous electrolysis. Two phases (phases I and II) can be converted to external electrical field stimuli. Phase I could be immediately obtained by adding CuI to a hot acetonitrile solution of TCNQ, or dissolving a quantity of [Bu4 N][TCNQ] and [Cu(MeCN)4 ][BF4 ] in the acetonitrile solution. The morphology of phase I has nanowire, nanorods, and nanotubes and can be used as the semiconductive component to fabricate the electrical nano devices [125], such as crossbar memory devices, and the FET [126]. The thin film of Cu(TCNQ) could display favorable photovoltaic effect [127], excellent field emission, and electrical switching properties [128] and has been applied to the diode humidity sensor [129] and the photo-response sensor [130]. Phase II is a result of the thermodynamically stable product of Cu (TCNQ). It can be obtained by using the same reaction as that for phase I but extending the reaction time to four days at room temperature or refluxing N

B4 B3

B5 N

B2

N

–

N

+1e–

+1e–

–1e–

–1e–

N

–

N

N

–

N

B1

N

N

N

·

N

Figure 12.73 Chemical structure of TCNQ, TCNQ⋅− , and TCNQ2− . Source: Tseng et al. [121]. Copyright 2010, Springer Nature.

737

12 The Relationship Between Structure and Electric Property Cu

Cu

N

N

N

C

Cu

C N

100

C

C

C

C N

Cu

C

Cu N

C N

C

N

C Cu N

Cu N

C

C

C

C

C N Cu

C

N

N Cu

N Cu

(c)

Cu N C

Cu

(b)

Cu

C

Cu

(a)

N

N

N

C N

C

C

C

C

C

N

C C

N N Cu NC N C N Cu

Cu C N C N Cu N C C N Cu

Conductivity (S/cm)

738

10–1 10–2 10–3

Phase 1 Phase 2

10–4 10–5 10–6 0.003

0.005

0.007

0.009

1/T (K–1)

N Cu

(d)

(e)

Figure 12.74 The crystal structures of Cu(TCNQ): (a) phase I and (b) phase II, views of the two interpenetrating networks in (c) phase I and (d) phase II, (e) plots of conductivity 𝜎 (S/cm) versus temperature for bulk Cu(TCNQ) phases I and II. Source: Heintz et al. [124]. Copyright 1999, American Chemical Society.

the reaction for several hours (the following eq. (12.1)) [131]. Cu + TCNQ → Cu (TCNQ) (phase I) → Cu (TCNQ) (phase II)

(12.1)

The crystal structures of phases I and II were determined by the single- crystal X-ray diffraction technique (Figure 12.74a,b). The biggest structural difference is the coordination configuration of Cu atoms. In phase I, Cu coordinates with four N atoms of four TCNQ to form a highly distorted tetrahedral structure with N–Cu–N angles of 92∘ and 142∘ , respectively. Neighboring TCNQ molecules in Cu(TCNQ) phase I are rotated 90∘ with respect to one another. The quinoid rings of the TCNQ units are engaged in interplanar stacking at a distance of 3.24 Å, which is smaller than the van der Waals distance of 3.4 Å for carbon atoms. Different from phase I, the metal geometry in phase II is much closer to tetrahedra. Coplanar TCNQ molecules of nearly tetrahedral structure are oriented in the same direction but in two perpendicular planes. The closest distance between TCNQ anions is ∼6.8 Å which is far from effective π–π interaction (Figure 12.74c,d). The conductivity measured on pressed pellets of Cu(TCNQ) phase I is 0.25 S/cm with a bandgap of 0.137 eV, whereas the conductivity of phase II is 1.3 × 10−5 S/cm with a bandgap of 0.332 eV (Figure 12.74e). The better conductivity of phase I can be attributed to much closer stacking of TCNQ molecules in phase I [132]. The strong π–π interactions among TCNQ ligands observably benefit to the electronic conductivity of phase I. Ag(TCNQ) with similar molecular formulas as Cu(TCNQ) also has two phases. But unlike the case of Cu(TCNQ), the two Ag(TCNQ) phases are not simply a matter of kinetic versus thermodynamic products [133]. The conductivity of Ag(TCNQ) phases I is not tested. Ag(TCNQ) phase II has alike structure as Cu(TCNQ) phase I, but the conductivity is 3.6 × 10−4 S/cm, which is lower than that of Cu(TCNQ) phase I. The reason may be that the mean interplanar distance within the TCNQ stacks is 3.50 Å [134], which is larger than that normally observed in the Cu(TCNQ) phase I. For the difference in the electron affinity between the Ag(TCNQ) and Cu(TCNQ),

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials a a

c b

(a)

100 c

10–1

(c) a

b

b c

a

c

Conductivity (S/cm)

b

2

10–2 10–3 10–4 10–5

1

–6

10

10–7

(b)

(d)

(e)

140 160 180 200 220 240 260 280 300 320 Average temperature (K)

Figure 12.75 A perspective view of the crystal structure along the short axis of Tl(TCNQ): (a) phase I and (c) phase II and side views of the crystal structures of (b) phase I and (d) phase II, and (e) conductivity measurements performed on pressed pellets of phase I (1) and phase II (2) of Tl(TCNQ). Source: Avendano et al. [136]. Copyright 2011, John Wiley & Sons.

the field emission of the Ag(TCNQ) nanowire is much better than that of Cu(TCNQ) nanowire [135]. Tl(TCNQ) also has two polymorphs as Cu(TCNQ) with radically different conducting properties [136]. Tl(TCNQ) phase I can be synthesized by slow diffusion of a methanol solution of Li(TCNQ) and an aqueous solution of TlPF6 . Phase II can be obtained through exposing Tl(TCNQ) phase I to ambient water vapor. In phase I, adjacent TCNQ stacks rotated by 90∘ around each other, while in phase II adjacent TCNQ stacks are in a parallel arrangement with respect to each other (Figure 12.75a,c). The distance of uneven spacing π-stacked TCNQ radicals is 3.16 and 3.35 Å in phase I, while the parallel arrangement of the TCNQ stacking distance of is 3.22 Å in phase II (Figure 12.75b,d). As a result, phase I has relatively poor conductivity, with room-temperature conductivity of merely 2.4 × 10−4 S/cm, while the room-temperature conductivity of phase II is 5.4 × 10−1 S/cm (Figure 12.75e). In phase I, the alternating distances between both TCNQ molecules and the adjacent Tl metal ions along the stacking generate a partial dimerization of the TCNQ radicals, which leads to the formation of a spin-Peierls insulator state and comparatively low conductivity. TCNQ can also coordinate with other metal ions to produce 1 : 2 M(TCNQ)2 ⋅(solv)x (M = Mn, Fe, Co, and Ni; solv = MeOH, H2 O) [137]. M(TCNQ)2 (M = Mn, Fe, Co, and Ni) [138] reported by Kim R. Dunbar et. al. shows typical semiconducting properties. The conductivity of Ni(TCNQ)2 and Mn(TCNQ)2 is 1.4 × 10−3 and 7.7 × 10−5 S/cm, respectively (Figure 12.76). It is interesting to note that these values are intermediate between the ones obtained previously for the two phases of Cu (TCNQ). The conductivity values of these MOFs follow the sequence of Ni > Co > Fe > Mn, which is correlated to distance from copper on the elemental table. Dunbar reported a range of isostructural 2D fishnet-type network compounds, [{Ru2 (O2 CCF3 )4 }2 − ⋅(TCNQRx )]⋅n(solv) (Rx = H4 , Br2 , Cl2 , F2 , and F4 ) [139], which were synthesized from the reactions of a paddlewheel diruthenium(II, II)

739

12 The Relationship Between Structure and Electric Property

Mn(TCNQ)2 Ni(TCNQ)2 Fe(TCNQ)2 Co(TCNQ)2

10–3

10–4 σ (S/cm)

10–5

10–6 0.003 0.0035 0.004 0.0045 0.005 0.0055 0.006 0.0065 I/T (K–1)

Figure 12.76 Semilogarithm plots of the conductivity 𝜎 versus 1/T for samples. Source: Clérac et al. [138]. Copyright 2003, American Chemical Society.

1011 Powder pellet

1010

V A

109 0

0

b

c

𝜌 (Ω·cm)

740

108 107 106

a

(a)

a

(b)

105

1 2 3 4 5

104 160 180 200 220 240 260 280 300 T (K)

(c)

Figure 12.77 Packing diagrams of [Ru2 (O2 CCF3 )42 (TCNQF4 )]⋅n(solv) projected along (a) the c-axis and (b) the b-axis, (c) temperature dependence of the resistivity (𝜌). Source: Miyasaka et al. [139]. Copyright 2010, American Chemical Society.

complex, [Ru2 II,II (O2 CCF3 )4 ], and neutral TCNQ derivatives (TCNQRx = 2,3,5,6or 2,5-halogensubstituted 7,7,8,8-tetracyanoquinodimethane) under anaerobic conditions (Figure 12.77a,b). According to the electron affinity of TCNQRx in connection with its first reduction potential, the Ru2 series has the requisite driving force for charge transfer from [Ru2 II,II (O2 CCF3 )4 ] to TCNQRx , which can form a mixed-valence state of [{Ru2 4.5+ }-(TCNQRx⋅ − )-{Ru2 4.5+ }] for the 2D network. Such a charge (or an electron) transfer results in magnetic exchange interactions between [Ru2 ] units (S = 1 for [Ru2 II,II ] and S = 3/2 for [Ru2 II,III ]+ ) via TCNQRx⋅ − S = 1/2 radicals, giving rise to long-range magnetic ordering in the layer. These MOFs show a typical semiconductive behavior with the electrical conductivity varying from 1.7 × 10−7 S/cm to 4.4 × 10−5 S/cm at room temperature (Figure 12.77c). The conductivity of TCNQRx MOFs follows the order of H4 < F2 < Br2 ∼ Cl2 , F4 , which is in agreement with the order of the electron affinity of TCNQRx .

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials 10–2 10–3 –4 Inσ (S/cm)

σ (S/cm)

10–4 10–5 10–6

a

c

–12 –16 3

10–7 120

b

(a)

–8

(b)

160

200

240

4 5 6 7 1/T (10–3 K–1) 280

8

320

T (K)

Figure 12.78 (a) The packing diagram of Cd2 (TCNQ)3.5 (H2 O)2 , (b) temperature dependence of the electron conductivity (𝜎) of a single crystal of the compound, and the inset being the Arrhenius plot of the conductivity (the red line shows the linear fit of the data). Source: Zhang et al. [140]. Copyright 2014, The Royal Society of Chemistry.

Besides 2D MOFs, TCNQ can also be applied to construct conductive MOFs with 3D structure. For instance, Dunbar and coworkers have reported another 3D MOF, Cd2 (TCNQ)3.5 (H2 O)2 [140] which exhibits the coexistence of μ2, μ3 and μ4 bridging modes of the TCNQ species in one framework (Figure 12.78a). The TCNQ anions are stacked in a slightly staggered fashion into 1D columns with seven TCNQ moieties, and the adjacent π-stacked columns are perpendicular to each other. Usually when the π-stacking distance is longer than 3.4 Å (twice the length of the van der Waals radius of carbon), the interaction between TCNQ rings is weak. Although this weak stacking in Cd2 (TCNQ)3.5 (H2 O)2 (with inter-planar distance varying from 3.287 to 3.687 Å) is crucial for maintaining the integrity of the structure, it is not eminent for effective magnetic interactions or electrical conductivity. The conductivity of this MOF being 5.8 × 10−3 S/cm is acquired by the two-probe method (Figure 12.78b). 12.4.1.2.2 Catechol and Its Derivatives

Catechol, referred to as pyrocatechol, 1,2-dihydroxybenzene, or benzene-1,2-diol, is an organic compound with the molecular formula C6 H4 (OH)2 (Figure 12.79). Its derivatives cover benzenehexathiol (BHT), 2,3,6,7,10,11-hexahydroxytriphenylene (HHTP), 2,3,6,7,10,11-hexaaminotriphenylene hexahydrochloride (HATP⋅6HCl), 2,3,6,7,10,11-triphenylenehexathiol (HTT), etc. The MOFs constructed by these molecules usually have 2D structure, and the conductivity of these molecule-based MOFs is generally high. Yaghi et al. had gain a highly conjugated 2D porous MOF, metal-catecholates (M-CATs) by tricatecholate (HHTP), and metal ions (CoII , NiII , and CuII ). M-CATs were prepared by combining 1 equiv of HHTP with 2 equiv of the respective metal(II) acetate hydrated in an aqueous solution and heating at 85 ∘ C for 24 hours to produce needle-shaped crystals. The crystal structure (the single crystal for Co-CAT-1, and powder for Ni-CAT-1) was solved in the trigonal space group, P3c1. There are two crystallographically independent metal atoms, each in an octahedral coordination environment. One of these metal atoms is coordinated with two adjacent deprotonated HHTP linkers and two water ligands to complete

741

742

12 The Relationship Between Structure and Electric Property

SH HS

OH OH

SH

HS

SH SH

Catechol OH

BHT SH

NH2 OH

NH2 H2N

HO xH2O

HO

6HCl

H2N

HS

OH OH HHTP

SH HS

NH2

SH

NH2

SH

HATP

HTT

Figure 12.79 Molecular structures of catechol, benzenehexathiol (BHT), 2,3,6,7,10,11-hexahydroxytriphenylene (HHTP), 2,3,6,7,10,11-hexaaminotriphenylene hexahydrochloride (HATP⋅6HCl), and 2,3,6,7,10,11-triphenylenehexathiol (HTT)

the octahedral coordination sphere, which forms an extended 2D framework. Conversely, the second metal atom is coordinated to merely one HHTP linker and four water ligands, giving rise to discrete complexes Co3 (HHTP)(H2 O)12 . As a result, the structure of Co-CAT-1 comprises two distinct types of alternatively stacked layers (Figure 12.80a). The first layer is an extended honeycomb structure with hexagonal pores (Figure 12.80b), and the second one is formed by the discrete units (Figure 12.80c) described above. The two axial water ligands involving in the formation of these discrete complexes are hydrogen bonded to oxygen atoms of the HHTP in the adjacent layers. These hydrogen bonds, accompanied by π–π interactions, create a distortion of the overall structure, leading to corrugated hexagonal layers (Figure 12.80d). The layers stack in an eclipsed way with the HHTP molecules in each layer rotated 60∘ around each other. A hexagonal array of 1D pore with a 12 Å diameter is thus formed. The surface areas for Co- and Ni-CAT-1 are calculated to be 490 and 425 m2 /g, respectively. HHTP is a redox-active linker that can undergo reversible interconversions among catecholate, semiquinonate, and quinone forms. On the basis of the charge balance, the oxidation state of the deprotonated HHTP in the covalently extended first type of layers (Co3 (HHTP)2 (H2 O)6 ) is −3, which suggests that each of the three dioxolene fragments is in the semiquinone oxidation level. An electron paramagnetic resonance (EPR) study on existence of these monoanionic semiquinonate units within the structure, which is also possibly found in Cu-CAT with similar structure, is carried out. Single crystal of Cu-CAT-1 showed good conductivity of ∼2 × 10−1 S/cm (four probe), possibly thanks to its ligand-centered monoradical and low defect in the crystal. The conductivity of Co and Ni-CAT-1 was not reported. The conductivity of M-CATs can be improved via replacing the –OH group of HHTP by –NH2 , or –SH. Reaction of 2,3,6,7,10,11-hexaaminotriphenylene (HATP)

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials

(a)

(b)

(c)

(d)

Figure 12.80 Space filling drawings of the single-crystal structure of Co-CAT-1: (a) the view of the Co-CAT-1 structure along the c-axis, (b) the extended layer of Co-CAT-1, (c) the layer formed by the trinuclear complexes Co3 (HTTP)(H2 O)12 , and (d) the view of the two extended corrugated layers along the [113] direction. Source: Hmadeh et al. [141]. Copyright 2012, American Chemical Society.

(six-NH2 ) with Ni2+ in the aqueous NH3 solution under aerobic conditions produces Ni3 (HITP)2 (HITP = 2,3,6,7,10,11-hexaiminotriphenylene), a new 2D porous MOF (Figure 12.81) [142]. From the powder X-ray diffraction (PXRD), extended X-ray absorption fine structure (EXAFS), and DFT data, Ni3 (HITP)2 reveals a hexagonal structure (space group: P6/mmm) with slipped-parallel (AB) stacking. PXRD analysis of Ni3 (HITP)2 revealed a crystalline structure with prominent peaks at 2𝜃 = 4.7∘ , 9.5∘ , 12.6∘ , and 16.5∘ , indicative of long-range order within the ab plane. An additional, weaker and broader peak when 2𝜃 = 27.3∘ , corresponding to the [001] reflections, denotes poorer long-range order along the c-direction, as expected for covalently linked layered materials. It was difficult to distinguish the AA (eclipsed) and AB (slipped-parallel) sequences from PXRD. Ni K-edge EXAFS analysis of a sample of Ni3 (HITP)2 revealed a spectrum that better agrees with a simulated spectrum of AB than that of AA. By the two-probe method and van der Pauw electrical measurements, the conductivity values of the bulk (the pellet) and the surface (the thin film) are 2 and 40 S/cm, respectively. The impressive electrical properties might be explained by the fact that metals can mediate very efficient 2D conjugation pathways between electroactive organic molecules. Diradical Ni-bisdiimine linkages, shown on the bottom of Figure 12.81, are likely to exist in the structure and make contribution to good conductivity. Besides –NH2 and –OH group, Xu and coworker use –SH group (HTT in Figure 12.82a) to coordinate with Pt to form an analogous 2D MOF Pt–HTT [143]. Pt–HTT was synthesized by the reaction of the hexaanion of HTT and Pt(CH3 CN)2 Cl2 in DMA (N,N-dimethylacetamide). Since the reaction of –SH with

743

744

12 The Relationship Between Structure and Electric Property

H N

Figure 12.81 The view of the porous structure of Ni3 (HITP)2 . Source: Sheberla et al. [142]. Copyright 2014, American Chemical Society.

H N Ni

N H

N H

HN NH HN Ni NH

NH NH Ni NH HN

~2 nm

NH HN Ni HN

NH

Ni3(HITP)2 H N

HN HN Ni NH NH

H N Ni

N H

N H

H N

H N

0

Ni N H

(a)

N H

(b)

(c)

Figure 12.82 (a) A schematic drawing of the honeycomb net of HTT–Pt, (b) a single net from a crystal structure model based on standard bonding geometries, and (c) the stacking of two neighboring sheets. Source: Cui and Xu [143]. Copyright 2014, The Royal Society of Chemistry.

Pt is ultrafast, only powder can be obtained. The structure of Pt–HTT was gained through simulation. From X-ray powder diffraction, we can see that both the positions and the intensity profile are consistent with a hexagonal grid modeled on standard bonding interactions between the tritopic HTT molecule and the square planar Pt(II) centers (e.g. Pt–S distance: 2.36 Å; see Figure 12.82 for the model), with a staggered stacking among the neighboring sheets (Figure 12.82c; interlayer distance: 3.46 Å). The 2D honeycomb model is also in keeping with the distinct laminae revealed by scanning electron microscopy. The conductivity of Pt–HTT is 10−6 S/cm, which is lower than that of –NH2 and –OH-based samples. The fundamental reason of the poor conductivity still needs further studies. The three benzene rings in HHTP, HATP, and HTT can be replaced by benzene to produce other π-conjugated graphene-like 2D MOFs. Nishihara and coworkers [144] synthesized π-conjugated nickel bis(dithiolene) complex nanosheets (micro-1) by a liquid–liquid interfacial reaction using BHT in the organic phase and nickel(II)

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials

23.29 Å (a)

(b)

(c)

Figure 12.83 The schematic illustration and chemical structure of the monolayer nickel bis(dithiolene) complex nanosheet. Source: Kambe et al. [144]. Copyright 2013, American Chemical Society.

acetate in the aqueous phase (Figure 12.83). 2D planar nanosheet has sixfold symmetry via the formation of the nickel bis(dithiolene) motif. PXRD analysis using high-energy synchrotron radiation produced several diffraction peaks. This diffraction pattern was reproduced by a crystal model with the P63 /mmc space group when a = b = 1.41 nm and c = 0.76 nm. The a- and b-axes are identical to the in-plane periodicity of the nanosheet. Furthermore, the c-axis is perpendicular to the nanosheet: the nanosheet layers stack in a staggered manner. Also, the nickel bis(dithiolene) complex shows reversible redox behavior (0, −1, and −2 are possible oxidation states). The conductivity of the prepared thin film was 0.15 S/cm, which decreased to 6.7 × 10−3 S/cm after reduction (two probe). They further investigated the conductivity of this compound (Figure 12.84) [145]. The average oxidation number is −3/4 for each complex unit in the as-prepared sample; oxidation or reduction can change this to 0 or −1, respectively. Refined electrical conductivity measurement, involving a single microflake sample measured by the van der Pauw method under scanning electron microscopy control, reveals conductivity of 1.6 × 102 S/cm, which is remarkably high for a coordination polymeric material. Conductivity also modulates with the change of the oxidation state. Both PES and the band calculation suggest that oxidized tris(4-bromophenyl) aminium hexachloroantimonate and as-prepared nickel bis(dithiolene) complex

745

12 The Relationship Between Structure and Electric Property Ni

(a)

S S Ni S S

S

–1 S

NaTCNQ

S

S

Ni S Red-1

Reduction

S

Ni S Ap-1

S

Ni

S

S

S

S

S Ni

S

S S

Ni

Ni

S

S

S

S

S

S

–3/4 (4-BrPh)3N SbCI6

S

S Ni S

S

S

Ni S

(c)

S S

S

Ni

S

S

Ni

S S

(b)

S

S

S

Ni

S S

Ni 0 S

S

30 μm

Ni S S OX-1

Oxidation

Figure 12.84 (a) The illustration of the chemical structure of the nickel bis(dithiolene) complex nanosheet, (b) the schematic illustration on redox control, (c) the SEM image for the van der Pauw measurement of ox-1. Source: Kambe et al. [145]. Copyright 2014, American Chemical Society.

nanosheets are metallic. As-prepared nickel bis(dithiolene) complex nanosheets contained Na+ as a counter-cation, which amplified the structural confusion and reduced the conductivity (2.8 S/cm). Zhu and coworkers reported another conductive organic 2D material Cu-BHT via a liquid–liquid interface reaction [146]. Cu-BHT has a hexagonal lattice with a = b = 8.45 Å and interlayer distance of 3.38 ± 0.03 Å through synchrotron radiation grazing incident X-ray diffractions (GIXRDs). As shown in Figure 12.85a, one BHT connects with other six BHTs through the shared Cu atoms, forming a lattice with sixfold symmetry. Each Cu atom coordinates with four S atoms in a square-planar manner, producing a dense topological structure without evident pores. If the six-membered carbon rings of phenyl moieties are overlooked, a continuous 2D Cu–S network could be observed. The interlayer distance is 3.38 Å in line with the PES scan and GIXRD data. Because of the minor energy difference (∼65 meV) for AA and AB stacking pattern, there probably exists a mixture of AA and AB stacking configuration. As shown in Figure 12.85b, a film sample (thickness of 400 nm) T (K) 75

150 300

T (K)

1,575

10

300 225 150

3.20

75

103.196

a

1,500

σ (S/cm)

b

1,425

103.19 103.18 103.17

Cu C S

(a)

1,350 0

(b)

75

150 T (K)

225

300

σ (S cm–1)

103.192

O

σ (S/cm)

746

103.186 103.180 103.174 0.24 0.27 0.30 0.33

Ea = 2.06 meV

1/T 0.25 (K–0.25)

Ea = 0.13 meV

0.025 0.020 0.015 0.010 0.005

1/T(K–1)

(c)

Figure 12.85 (a) The two-dimensional lattice of [Cu3 (C6 S6 )]n derived from the in-plane GIXRD pattern (a = b = 8.76 Å) and temperature dependence of electrical conductivity of a Cu-BHT film, (b) electrical conductivity (s) of a 150-nm film as a function of temperature ranging from 2 to 300 K, (c) plots of s (T) versus T −1 and (the inset in c) s (T) versus T −0.25 (this measurement was performed on a film with the thickness of 400 nm by the four-probe method). Source: Huang et al. [146]. Licensed under CC BY 4.0.

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials

with room temperature conductivity of 1580 S/cm maintains high conductivity (1360 S/cm) at 2 K. The ratio of 𝜎(300 K)/𝜎(2 K) is equal to 1.16. This sample shows a nonlinear relationship between the logarithm of the conductivity and the reciprocal of the absolute temperature (Figure 12.85c). The slope of the curve can be seen as Ea /k, where Ea is the activation energy of the electrical conductivity and k is the Boltzmann constant. The Ea continuously rises with increasing temperature and ranges from 0.12 meV at 40 K to 2.06 meV at 300 K, which is extremely small especially in low temperature range. As illustrated in the inset of Figure 12.85c, the plot of log10 𝜎 versus T −0.25 within the range of temperature from 40 to 250 K is well fitted to the 3D variable range hopping model. It implies that the hopping process between nanosheets should be the critical step for the electrical conducting process of the Cu-BHT film. Thermopower measurements were carried out under room temperature, as thermopower is much less susceptible to the resistive boundaries in a zero-electrical-current measurement. The thermopower of Cu-BHT films ranges from −4 to −10 μVK−1 . Such small values are typically observed in metal or highly conducting polymers with metallic behavior, which suggests the metallic nature of Cu-BHT films. Meanwhile, Cu-BHT displays ambipolar charge transport behavior and extremely high electron and hole mobility (99 cm2 /V/S for holes and 116 cm2 /V/S for electrons) under field-effect modulation. Beside –SH, –NH2 group in hexaaminobenzene (HAB) can also obtain 2D MOF with Ni, Cu, and Co through “bottom-up” techniques [147]. The structure of Ni–HAB is similar to the –SH group sample (Figure 12.86). In all three cases, devices were analogously resistive in the range of GΩ. 12.4.1.2.3 Nitrogen Containing Heterocyclic Compounds

Unlike the nitrogen containing heterocyclic 3D structure, the nitrogen incorporating the heterocycle with introduced –SH can produce conductive polymers with 2D structure. Hong, Cao, and coworkers utilized pyridine-2-thiolate (C5 H4 NS) (Figure 12.87a) and the coordinate Ag+ to obtain [Ag(C5 H4 NS)] the graphite-like array of distorted Ag6 hexagons. [Ag(C5 H4 NS)] shows low electrical semiconductivity of 2.04 × 10−5 S/cm at 298 K (the powder sample) [149]. They also use another similar ligand to construct another semiconductive polymer [Ni2 (C4 N2 H3 S)4 ]n (C4 N2 H3 S = pyrimidine-2-thiolate) [148]. Each ligand in [Ni2 (C4 N2 H3 S)4 ]n acts as a μ3 -bridge to link three nickel atoms through S and N atoms (Figure 12.87b–d). The product exhibited better conductivity (5 × 10−3 S/cm, the powder sample) than [Ag(C5 NH4 S)]n , which can be attributed to its characteristic structural feature of the interconnected array of nickel(II) with pyrimidine. 12.4.1.3 One-Dimensional (1D) Metal–Organic Frameworks Through Different Linkers 12.4.1.3.1 Halogens

MX and MMX (M is a metal cation; X is a halogen) are typical 1D conductive MOFs [150]. These materials exhibit the characteristic features of quasi-1D polymers, such as the spin density wave (SDW), the charge density wave (CDW), solitons, polarons, and bipolarons. Figure 12.88 shows these two sorts of MX and MMX chains. In MX

747

748

12 The Relationship Between Structure and Electric Property

M

NH2 H2N

NH2

H2N

NH2 NH2

(a)

M||(acac)

2 in

EtOAc

NH

HN

H N

HN

N H

M

M

NH

M2+ = Ni2+, Cu2+, Co2+

(b)

Figure 12.86 (a) General synthesis of CPs (b) the schematic illustration of the structure of the nanosheet (the tube model with hydrogen is omitted for clarity: gray, carbon; blue, nitrogen; red, metal ion), and the photograph of the reaction vessel highlighting the liquid/liquid (EtOAc-water) interfacial synthesis of Ni–HAB complexes showing the polymer as a black solid at the interface. Source: Lahiri et al. [147]. Copyright 2017, American Chemical Society.

chain, the combination of both dz 2 orbitals coming from the metal center with the pz orbitals of the halogen bridge makes these materials have the theoretically infinite 1D structure. Most of them can be schemed as follows: [ML2 ][ML2 X2 ]A4 (where M = Ni, Pd, and Pt; L = alkyl diamine; X = Cl, Br, I; and A = Cl, Br, I, ClO4 , BF4 , and SO4 ). Some heterobimetallic compounds have also been reported in the literature. Takaishi et al. synthesized [Pt(chxn)2 I]I2 single crystals [151] by the slow diffusion of I2 vapor into a methanol solution of Pt (chxn)2 I2 (chxn = 1R,2R-diaminocyclohexane). [Pt(chxn)2 I]I2 has a novel iodo-bridged linear chain with the quasi-2D CDW ground state. They discovered in the boundary region an anomalous valence state at which the CDW phase alternates in the crystal by electron spin resonance (ESR), X-ray diffuse scattering, scanning tunneling microscope (STM), and electrical resistivity. This anomalous state can be explained by the fast fluctuation between PtIV –I· · ·PtII and PtII · · ·I–PtIV in the double well potential. The conductivity value of [Pt(chxn)2 I]I2 single crystals is 10−4 S/cm with the four-probe method at room temperature. The activation energy varied at approximately 190 K, which was evaluated to be 0.16 and 0.28 eV below 180 K and above 200 K, respectively. These activation energies are smaller than energies of the optical gap, and the conduction may be considered to occur via impurity sites. They also synthesized another SDW-type MX chains: [Ni(S,S-bn)2 Br]Br2 [152], (S,S-bn = 2S,3S-diaminobutane). In this complex, an unpaired electron (S = 1/2) in the Ni3+ 3dz2 orbital is strongly localized on the Ni atom owing to the strong

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials C(3)

N

(a) S

N

(b) S

S(2) S(1) N1

C(2) N(2)

N

C(1)

N(8) S(2)

C(4)

N(1)

N(4A) N(A)

S(1)

N(3) C(5)N(3B)

C(8A) C(7A)

C(6)

N(3C) N(4)

N(3A) C(5A)

S(1A) N(1A)

C(7)

C(6A)

N

C(8)

C(4A)

S(2A)

N(C)

C(1A) N(2A) C(2A)

(c)

C(3A)

(d)

Figure 12.87 (a) Pyridine-2-thiolate (C5 NH4 S), (b) pyrimidine-2-thiol (C4 N2 H3 S), (c) the structure of the basic unit [Ni2 (C4 N2 H3 S)4 ], and (d) the view of the lamellar structure in [Ni2 (C4 N2 H3 S)4 ]n . Source: Zhao et al. [148]. Copyright 2001, The Royal Society of Chemistry.

X

MII

X

MIV

X

MII

X

MIV

X

MIII

X

MIII

X

MIII

X

MIII

X

(a)

X

(b)

Figure 12.88 Schematic representation of mixed-valence MX, the charge density wave (CDW) for Pd and Pt complexes (a), and the spin density wave (SDW) for Ni (b). Source: Givaja et al. [150]. Copyright 2012, The Royal Society of Chemistry.

electron correlation. Quite a strong antiferromagnetic interaction (J = 3600 K) acts between neighboring Ni atoms through the Br 4pz orbitals because of the superexchange interaction. The conductivity of this compound was not reported. If the MX chains were further bridged by a secondary ligand, MX chains can be used as the building block to create porous MOFs with large cavities in the structure. For example, Kitagawa and coworkers successfully synthesized [Pt(en)(bpy)I]4 (NO3 )8 ⋅16H2 O [153] in the light of a square-shaped building block [Pt(en)(bpy)]4 (NO3 )8 ⋅5H2 O and I2 bridge (Figure 12.89, en = ethylenediamine; bpy = 4,4′ -bipyridine). The window size of this 1D channel is 5.9 Å × 5.9 Å. The sorption properties of the dehydrated form indicate that it can adsorb water (H2 O)

749

750

12 The Relationship Between Structure and Electric Property

N N N Pt N N N Pt N N

Oxidative polymerization N N N Pt N

+ I2

N N N Pt N

[Pt(en)(bpy)]4 (NO3)8 Square-shaped building block

Halogen

Square-prism-shaped nanotube

Figure 12.89 The schematic of the fabrication of [Pt(en)(bpy)I]4 (NO3 )8 ⋅16H2 O nanotube. Source: Otsubo et al. [153]. Copyright 2011, Nature Publishing Group.

AV CP CDW ACP

M2.5+ M2.5+ X–

M2.5+ M2.5+

M2+ M3+

X–

M2+ M3+

X–

M2+ M2+

X–

M3+ M3+

X–

M2+

X–

M3+ M2+

X–

M3+

Figure 12.90 Possible electronic states in MMX chains. Source: Givaja et al. [150]. Copyright 2012, The Royal Society of Chemistry.

and alcohol vapors, whereas N2 and CO2 do not adhere to it. The amounts of adsorbates at P/P0 = 0.97 were about 16 (H2 O), 10 (methanol), and 7 (ethanol) molecules per formula unit. In general, the interchain electronic interaction of the MX-chain is reflected in the interchain correlation of the CDW state. Its semiconductive bandgap can be controlled by exchanging structural components and guest molecules. Nevertheless, the electrical conductivity of this compound was not reported. When two metal centers are bridged by four ligands with an appropriate bite angle, such as amidates or carboxylates, a MM dimeric unit is formed, and further connections of which by X can form MMX chain. The geometry of the building block of MMX has been called the “paddle-wheel” or “china lantern” complex, which is constructed by four five-membered rings, including the M–M interaction. An important consequence of passing from MX to MMX chains is the gain of internal degrees of freedom from the charge–spin–lattice point of view. Compared with MX chain, MMX chain shows internal degrees of freedom, which results in new valence ordered states: the average valence (AV), charge polarization (CP), the CDW, and alternate charge polarization (ACP) states (Figure 12.90) [150]. It is noteworthy that the ground state might be changed by applying pressure or replacing the halogen bridge, the counter ion and the ligand.

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials

C S Pt

S Pt

I

0.5 nm

Figure 12.91 Structure of a [Pt2 (dta)4 I]n (dta = dithioacetato) single fiber. Source: Hermosa et al. [154]. Copyright 2013, Springer Nature.

Zamora and coworkers reported a typical example of MMX chain [Pt2 (dta)4 I]n (dta = dithioacetato) [154]. The paddle-wheel-like dimetallic subunit was connected by the iodine atom as a bridge (Figure 12.91). The conductivity of a defect-free nanoribbon is 2 × 102 S/cm owing to the 1D nanostructure with low defect content (0.4% < c < 1%, c calculated based on conductivity along disparate directions), while the conductivity of the pellet is 8 × 10−1 S/cm. Zamora and coworkers reported a similar [Pt2 (n-pentylCS2 )4 I] MMX chain [155]. The crystal structure of Pt2 (n-PrCS2 )4 I at room temperature (RT) showed three well-differentiated Pt–I distances as well as alternatively disordered dta ligands. When temperature rises, a phase transition occurs at 359 K, giving rise to a pure AV state for the chain. DC electrical conductivity measurements carried out on crystals of [Pt2 (n-pentylCS2 )4 I] illustrated that under room temperature the range of conductivities is 0.3–1.4 S/cm. The crystals were cooled from 300 to 100 K and then warmed to 300 K. The cooling run reflects a metallic behavior with a rounded minimum at around 255–270 K, followed by a surge at nearly 200 K (Figure 12.92). The thermally activated regime follows the 1D variable range hopping model (𝜎 = 𝜎 0 * ⋅exp[−(T 0 /T)1/2 ]), to explain conductivity in multiple conducting polymers. This fact suggests that the conductivity in [Pt2 (n-pentylCS2 )4 I] is primarily 1D, as expected from its linear structure. However, this model fails to reproduce the conductivity in the intermediate temperature range 225 < T < 255 K, which implies the presence of an intermediate regime between the metallic and semiconducting regimes. In the intermediate regime, the coexistence of metallic and semiconducting regimes may be the result of a progressive distortion of the Pt–I distance along the chain. Apart from Pt and Pd, Ni can also form the similar MMX chain, [NiII/III 2 (RCS2 )4 I]∞ (R = Et, n-Pr, and n-Bu) (Figure 12.93) [156]. The reported room temperature electrical conductivity values are 1.6 × 10−3 , 7.6 × 10−4 , and 6.0 × 10−4 S/cm for R = Et, n-Pr, and n-Bu, respectively. Ni2 (EtCS2 )4 I derivatives present a CP state at indoor temperature. In the n-Pr derivative, Ni2 (n-PrCS2 )4 I, there are three markedly different Ni–I distances. The valence order can be regarded as –I· · ·Ni3+ –Ni2+ –I· · ·Ni2.5+ –Ni2.5+ –I· · ·Ni2+ –Ni3+ –I· · ·, showing an approximate AV ground state at RT, attributed to the entropy reservoir of the ligand alkyl chains. An alike behavior was observed in Ni2 (n-BuCS2 )4 I chains. The MMX chain can also form zigzag chains in (–Rh–Rh–X–)n . The reaction of 2,2-cis-[Rh2 (acam)4 (H2 O)2 ] ClO4 (Hacam = acetamide) [157] with sodium halides in the aqueous solution produces crystals of [Rh2 (acam)4 X]n ⋅mnH2 O (X = Cl, Br,

751

12 The Relationship Between Structure and Electric Property

Figure 12.92 (a) A single polymer chain of [Pt2 (n-pentylCS2 )4 I], (b) thermal variation of the electrical resistivity of [Pt2 (n-pentylCS2 )4 I] (the dashed line shows the behavior of a non-heated sample; the inset denotes the derivative of the resistivity versus temperature around the RT–HT transition). Source: Guijarro et al. [155]. Copyright 2010, John Wiley & Sons.

I2

Pt3

I1 Pt2

Pt1 I1 Pt1

Pt2 Pt3

(a) 20

0.05

δp/δT

0.00

15 ρ (Ω cm)

752

–0.05 –0.10

10 –0.15 300

310

320

330

340

350

T (K)

5

0 200 (b)

250

300 T (K)

350

400

and I). Their structures comprise zig–zag chains of repeating (–Rh–Rh–X–) units (Figure 12.94). As depicted in Figure 12.94, they are formed through hydrogen-bond formation between amidato-NH donors and amidato-O acceptors of adjacent units for the bromide chain and between amidato-O and interstitial water molecules for the chloride derivative. Pressed pellets of both compounds showed a conductivity value below 2 × 10−7 S/cm at room temperature.

12.4.2 Structure and Electrical Properties of Organic–Inorganic Hybrids Since the late nineteenth century, organic–inorganic perovskites have recently received extraordinary attention from the research community because of their unique physical properties, which make them promising candidate materials for application in photovoltaic (PV) and related optoelectronic devices. This prospect derives from the combined effect of several factors, including extremely high optical absorption, small effective masses for electrons and holes, dominant point defects that generate just shallow levels, and grain boundaries that are essentially benign. In organic–inorganic perovskites, when the inorganic part is transition-metal halides, the hybrids are generally good insulators (wide-gap semiconductors), but when the metal in metal halides is group 14 metal (Ge, Sn, and Pb), they are semiconductors with narrower bandgaps. In this chapter, the electron conductivity of organic–inorganic hybrids is focused on. Due to the critical role of the dimension

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials

I1 2.8642(5)Å Ni1 2.5336(8)Å Ni2 2.9446(6)Å I2 2.9374(6)Å Ni3 2.5502(11)Å Ni3ʹ I1 2.9186(6)Å Ni1 2.5479(7)Å Ni2 2.9085(6)Å I1ʹ

I1 2.8681(11)Å Ni1 2.5349(17)Å Ni2 2.9115(13)Å I2 2.9105(14)Å Ni3 2.550(2)Å Ni3ʹ

I2ʹ

I2ʹ

Ni2ʹ

Ni2ʹ

Ni1ʹ

Ni1ʹ

I1ʹ

(a)

(b)

I1ʹ

(c)

Figure 12.93 Room temperature solid-state structure and significant distances of Ni2 (EtCS2 )4 I (a), Ni2 (n-PrCS2 )4 I (b), and Ni2 (n-BuCS2 )4 I (c) (all hydrogen atoms are not displayed for clarity, and the disordered ligand chain position is shown). Source: Mitsumi et al. [156]. Copyright 2009, American Chemical Society.

N4

N2 CI1 N1

Rh2 N3

Rh1

O3

O1 O4 O2 O6

O5

(a)

N2

O1

Br1 Rh1

O2

N1

(b)

Figure 12.94 Crystal structure of [Rh2 (acam)4 Cl]n ⋅7nH2 O (a) and [Rh2 (acam)4 Br] (b) with the partial numbering scheme (hydrogen atoms are not shown and simply selected water solvent molecules are displayed. Short range contacts are depicted as red dotted lines). Source: Givaja et al. [150]. Copyright 2012 Royal Society of Chemistry.

753

754

12 The Relationship Between Structure and Electric Property

Figure 12.95 Combined ball-and-stick or skeletal (top) and shaded polyhedral (bottom) representations of the 3D cubic ABX3 perovskite structure (the structure is built upon corner-sharing BX6 octahedra and emphasized by cyan-shaded polyhedra, with A cations filling the 12-fold cuboctahedral voids. A, B, and X atoms are shown as gold, cyan, and red spheres, respectively). Source: Saparov et al. [158]. Copyright 2016, American Chemical Society.

of the inorganic anion played in electron conducting organic–inorganic hybrids, we are going to show the progress in this filed in the light of the different dimension of anions in organic–inorganic perovskites. 12.4.2.1 Three-Dimensional (3D) Organic–Inorganic Hybrids

Perovskite refers to the mineral CaTiO3 and the structure with the same 3D structural framework (Figure 12.95) as ABX3 . In the case of typically organic–inorganic hybrid perovskites, the “A” cation is organic, while “B” is metal and “X” is a halogen (Cl, Br, or I). Take CH3 NH3 PbI3 as an example, it employs methylammonium as “A” cations. The 3D perovskite (i.e. BX3 ) structure is constructed by the corner-sharing network of BX6 octahedra, while the “A” cations occupy 12-fold coordinated holes within the structure and counterbalance the charge of the BX3 extended anion. The ionic size of “A” in 3D perovskite structure has strict restraint.

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials

Table 12.2

Effective radii of molecular cations and anions.

Cation

Effective radius r A,eff (pm)

Ammonium [NH4 ]+

146 +

Hydroxylammonium [H3 NOH]

216

Methylammonium [CH3 NH3 ]+

217

+

Hydrazinium [H3 N–NH2 ]

217

Azetidinium [(CH2 )3 NH2 ]+

250 +

Formamidinium [NH2 (CH)NH2 ]

253

Imidazolium [C3 N2 H5 ]+

258 +

Dimethylammonium [(CH3 )2 NH2 ]

272

3-Pyrollinium [NC4 H8 ]+

272 +

Ethylammonium [C2 H5 NH3 ]

274

Guanidinium [C(NH2 )3 ]+

278 +

Tetramethylammoninm [(CH3 )4 N]

292

Thiazolium [C3 H4 NS]+

320 2+

322a

Piperazinium [C4 H12 N2 ] Tropylium [C7 H7 ]+

333

Dabconium [C6 Hl4 N2 ]2+

339a

Anion

Effective radius r x,eff (pm)

Fluoride

129

Chloride

181

Bromide

196

Iodide I−

220

Formate HCOO−

136

a

Radii calculated from the single crystal X-ray data. Source: Saparov and Mitzi [158]. Copyright 2016, American Chemical Society.

Treating all ions as rigid spheres and considering close packing, we yield the Gold√ schmidt’s Tolerance Factor concept, namely (RA + RX ) = t 2 (RB + RX ), where RA , RB , and RX are the ionic radii for the corresponding ions and the tolerance factor must satisfy t ≈ 1. It is found empirically that 0.8 ≤ t ≤ 1.0 applies to the tolerance factor for most 3D perovskites. In lead halide-based perovskites, the ionic radii of RPb are 1.19 Å and that radii of RI are 2.20 Å; the limit on RA is found to be approximately 2.6 Å (t = 1) for PbI3 − frameworks. The effective radii of molecular cations and anions are listed in Table 12.2. In PbI3 − or the SnI3 − 3D perovskite framework, methylammonium and formamadinium can be inserted to the anionic holes, whereas ethylammonium would not. Besides ionic radius constraints, charge balance must also be achieved. It means if the “A” cation is monovalent, then the “B” cation must be divalent; if “X” is a halogen, all sites are fully occupied. Divalent “B” cations include Ge2+ , Sn2+ , and Pb2+ ; divalent alkaline-earth and rare-earth ions, such as Ca2+ , Eu2+ , and Yb2+ can also

755

12 The Relationship Between Structure and Electric Property

Ttr,u

–5

–6 log10(l/𝜌)

756

–7

2.7

3.0

3.3 1000 / T (K–1)

3.6

Figure 12.96 Temperature dependence of the electrical resistivity 𝜌 (Ω cm) of CH3 NH3 PbI3 between 0 and 95 ∘ C. Source: Osvald Knop et al. [161].

be employed. Ge2+ , Sn2+ , Pb2+ , divalent alkaline-earth and rare-earth ions, such as Ca2+ , Eu2+ , and Yb2+ are usually utilized as divalent “B” cations. This valence assignment rule can also be extended considering that the +2 charge on the “B” cation can be achieved by employing equal numbers of +1 and +3 cations (on average yielding a +2 overall charge on this site), as is the case for the semiconducting inorganic halide perovskite CsAuI3 , wherein the Au is split into +1 and +3 states; it is to say, the compound can be expressed more specifically as Cs2 Au(I)Au(III)I6 . Both the electron and hole diffusion lengths are at least 100 nm in solution processed MAPbI3 [159]. The high photoconversion efficiency of these systems may stem from the comparable optical absorption and charge-carrier diffusion lengths. Using the electron and hole diffusion coefficients for MAPbI3 (0.036 and 0.022 cm2 /s) obtained to fit the decay dynamics of photoluminescence spectra to the diffusion model [159]. The calculated resistivity value for a single crystal of CH3 NH3 PbI3 , ignoring the nonlinear parts of the plot, is 51 MΩ cm at room temperature [160]. The resistivity of the pressed CH3 NH3 PbI3 compact decreased when heated from ∼1.3 × 107 Ω cm at 0 ∘ C to ∼2.6 × 105 Ω cm at 94 ∘ C (Figure 12.96), i.e. 50-fold over the 100 K interval, and was unaffected by the presence or absence of daylight. The Arrhenius plot is essentially linear if the problems with the sample are taken into account. The activation energy for the conduction process obtained from the plot is 0.39 eV [161]. Xiao et al. have demonstrated ion migration under the electric field in MAPbI3 [162]. In MAPbI3 , the ionic conductivity is found to be higher than the electronic

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials

σeon σion σtot

10–6

σ (S/cm)

Figure 12.97 The properties of mixed-cation lead iodide, MA1−x FAx PbI3 and changes of electronic, ionic, and total conductivity as a function of the fatty acid (FA) faction. Source: Yang et al. [163]. Copyright 2015 John Wiley & Sons.

10–7

10–8 0.0

0.2

0.4

0.6

0.8

1.0

X

conductivity (Figure 12.97). The large ionic transference number points toward high native ionic disorder and substantial trapping of the electronic carriers [163]. CH3 NH3 SnI3 has similar structure as CH3 NH3 PbI3 and the conductivity is 100 S/cm, which is much larger than that of CH3 NH3 PbI3 . The carrier concentration of CH3 NH3 SnI3 is 1020 cm−3 , corresponding to 0.003–0.005 carriers per Sn atom. NH2 CH=NH2 SnI3 has a cubic perovskite structure, and the lattice constant for NH2 CH=NH2 SnI3 is a = 6.316(1) Å, which is probably 1.2% larger than that for the isostructural compound CH3 NH3 SnI3 . The electrical resistivity of a pressed pellet of the new compound exhibits semimetallic temperature dependence from 10 to 300 K, with evidence of a structural transition at approximately 75 K [164]. Figure 12.98 shows the resistivity of NH2 CH=NH2 SnI3 . While the sample is being warmed from 10 to 300 K, the gradual increase of resistivity is accompanied by increasing temperature. At approximately 60 K, the resistivity rises sharply, reaching a peak at 75 K. Above 75 K, the resistivity decreases with increasing temperature, until around 200 K. Above this point, a small positive temperature derivative is again achieved, with room temperature resistivity of the order 10 mΩ cm. Some hysteresis is detected in the transition at 75 K during the warming and cooling process (Figure 12.98, the inset), suggesting that the transition is the first order. The unusual electrical conductivity in each of metal halide-based systems arises from the large dispersion of the Sn 5s band (hybridized with I 5p) along the (111) direction in the cubic Brillouin zone, leading to a marginal crossing of the Sn 5s and Sn 5p bands near the R point ([1/2, 1/2, and 1/2] 2π/a), with the Fermi energy falling roughly between the two bands. Presumably, the larger observed resistivity in the system (CH3 NH3 )1−x (NH2 CH=NH2 )x SnI3 , with x = 1 (relative to the x = 0 samples), is at least partially the result of the expansion of the cubic lattice, generating a narrowing of the bands and therefore a lower density of states at the Fermi energy. In addition, it is also possible that the substitution of the formamidinium cation for methylammonium might result in a different preferred concentration of defects or vacancies in the structure as a result of the disparate molecular shape and functional groups. Small concentrations of such defects could readily produce a substantial shift in the already small carrier densities in these materials. In addition to perovskite structure, inorganic metal halides can also be connected by edge-sharing or face-sharing to gain 3D open framework structure. Xu, Guo,

757

12 The Relationship Between Structure and Electric Property

14 Resistivity (mΩ cm)

10

12 Resistivity (mΩ cm)

758

10 8

9 8 7 6 5 25

50

75

100

125

150

Temperature (K)

6 (a)

4 2

(b)

0 0

100

200

300

Temperature (K)

Figure 12.98 Electrical resistivity measurements for pressed pellet samples of (CH3 NH3 )1−x (NH2 CH=NH2 )x SnI3 with (a) x = 1.0 and (b) x = 0.0 (measurements were made using four-point contact geometry. For clarity, only the warming curves are shown. The inset provides a more detailed look at the hysteresis of NH2 CH=NH2 SnI3 in the low-temperature resistive transition, namely in the warming and cooling process). Source: Mitzi and Liang [164]. Copyright 1997, Elsevier.

and coworkers reported two plumbichloride hybrids with 3D open framework structures [165]. The conductivity of (H2 DABCO)(Pb2 Cl6 ) with 3D open framework structure is 2.18 × 10−7 S/cm at 30 ∘ C, which gradually increases with the rise of the temperature and reaches 2.13 × 10−5 S/cm at 170 ∘ C. The variety of the conductivity of (H3 O)(Et2 BCO)8 (Pb21 Cl59 ) has a similar trend to that of (H2 DABCO)(Pb2 Cl6 ). The conductivity of (H2 DABCO)(Pb2 Cl6 ) is 2.83 × 10−7 S/cm at 30 ∘ C and 2.24 × 10−3 S/cm at 170 ∘ C. The test results indicate that the conductivity of the material is connected with the structure of the inorganic part. Guo, Wang, and coworkers found another moisture-resistant semiconductor {(MV)2 [Pb7 Br18 ]}n (MV2+ = methyl viologen cations) with the electron-transfer thermochromic property [166]. As shown in Figure 12.99a, the ribbon-like double chain extending along a-direction was constructed by block 1. Such chains are linked toward the b-direction by block 2 to form a 2D layer lying in the ab plane (Figure 12.99b). Finally, the 3D inorganic open framework is completed in block 3 by sharing edges and corners with the Pb4-centered monocapped octahedron and the Pb2-centered octahedron from the block 1, respectively (Figure 12.99c). The MV2+ cations, which occupy 40.8% of the crystal volume, locate at the channels of the open framework (Figure 12.99d). {(MV)2 [Pb7 Br18 ]}n (denoted as 1a) shows thermochromic behavior through heating (denoted as 1b) for thermally induced electron transfer between Br atoms in the inorganic framework and the MV2+ . The conductivity value measured along the crystallographic a-axis is about 1.0 × 10−8 S/cm. After annealing at 220 ∘ C for two hours and cooling to 300 K, it fell by 84.3% to about 1.7 × 10−9 S/cm (Figure 12.100a).

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials

= Block 1

a

(a)

Pb2 Pb4 Pb1 Pb3 Br

Block 2

b

Block 3 ab plane

a (b)

c

c

b

b (c)

(d)

Figure 12.99 Crystal structure of {(MV)2 [Pb7 Br18 ]}n : (a–c) structural evolution of the 3D inorganic open framework, (d) a packing diagram with the inorganic open framework and MV2+ cations being shown in space filling and wireframe models, respectively (H atoms are omitted for clarity). Source: Sun et al. [166]. Copyright 2017, John Wiley & Sons.

The relationship of the linear natural logarithm of s versus 1/T (Arrhenius law) shows that before and after heating, activation energies (Ea ) of the compound behaving as the intrinsic semiconductor are calculated to be 0.896 (1a) and 1.061 eV (1b), respectively. Here, the temperature range of the single-crystal sample is 300–420 K (Figure 12.100a). A comparison of valence band X-ray photoelectron spectroscopy (XPS) spectra data (Figure 12.100c) with the calculated partial density of state (PDOS) (Figure 12.100b) illustrates that the density of states near the TVB, which has almost the same proportion with the inorganic skeleton, decreased from 81.28% to 77.57% after coloration. The carriers from intrinsic thermal excitation originate mainly from electron states near the TVB. This indicates that the carrier density in the semiconducting inorganic skeleton of 1b was reduced, bringing about a decrease of the conductivity value of {(MV)2 [Pb7 Br18 ]}n at 300 K after coloration. 12.4.2.2 Two-Dimensional (2D) Organic–Inorganic Hybrids

The 2D perovskite structure is obtained by cutting the 3D perovskite in different directions, and the inorganic layer and the organic layer are stacked alternately to show multiple quantum wells, where the inorganic layers are potential “wells”

759

12 The Relationship Between Structure and Electric Property

6

–4

5

–6 –7

1b

Ea = 0.896 eV

Ea = 1.061 eV

1a

–8 –9 2.4 (a)

2.6

2.8 3.0 1000/T (K−1)

3.2

3

x Γc

2 1

x Γb

0 x Γa

–1

84.3%

Pb Br MV2+

x Γe x Γd

4 E − EF (eV)

log(σ) (S/cm)

–5

–2 –3

3.4 (b)

Γ

L

PDOS

Z

CB

CB

–

–

–

1a 1a Intensity (a.u.)

760

MV2+

+ – + – + –

81.28% 18.72%

CB

1b 1b

77.57% 22.43% 8

6

+

Free hole

–2

intrinsic thermal excitation – +

(d)

LUMO

VB CB

LUMO

VB

2 0 4 Binding energy (eV)

– Free electron

(c)

LUMO

– MV–+

+ – + + –

+

+ –

–

– MV–+

+

+

+

On

Off

Frenkel excition model

+

LUMO

VB

MV2+

– +

Figure 12.100 (a) Temperature-dependent electrical conductivity of 1a and 1b, measured in air using a single crystal (the inset: photographs before and after coloration; straight lines represent the fit of the data to the Arrhenius law), (b) energy band structures (left) and partial density of states (PDOSs) for all Pb atoms, Br atoms, and MV2+ cations (right) (circles in red, blue, and green correspond to Pb, Br, and MV2+ , respectively; the Fermi level is set to zero), (c) valence band XPS spectra for 1a and 1b (the circle, experimental data; gray lines, resolved peaks; blue/red solid lines, the sum of the resolved peaks), (d) a schematic diagram showing why carriers are reduced after coloration (VB, the valence band; CB, the conductance band; LUMO, lowest unoccupied molecular orbital). Source: Sun et al. [166]. Copyright 2017, John Wiley & Sons.

and the organic layers are potential “barriers.” The general formula of anions of (100)-oriented layered perovskites is Bn X3n+1 −n−1 ; the formula of the (110)-oriented family is Bm X3m+2 −m−2 ; along the (111) direction, the formula of the 3D perovskites is Bq X3q+3 −q−3 . As shown in Figure 12.101, the 2D dimension perovskite is reduced by removing the metal B," which results in an increase in purely ionic interactions between “A” and “X,” two components with a very different set of electronegativity that also brings about a larger bandgap.

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials

761

n=3 m=3 (110)

(100)

q=3 (111) q=2

m=2

n=2

q=1

m=1 n=1

B2X73– NH+3

Aʹ+

R

(RNH3)2AB2X7 (n = 2)

(b)

B2X93– Aʹ+

A+

A+ (a)

B2X84–

A– Aʹ2A2B2X8 (m = 2)

(c)

Aʹ2AB2X9 (q = 2)

Figure 12.101 Schematic representation of the derivation of the 2D dimensional organic–inorganic perovskites (lower sections) from different cuts of the 3D perovskite structure (top sections): (a) the family of (100)-oriented layered perovskites with a general formula of (RNH3 )2 An−1 Bn X3n+1 obtained by taking n layers along the (100) direction of the parent structure, (b) cuts along the (110) direction of the 3D perovskite structure producing the (110)-oriented family, A′ 2 Am Bm X3m+2 , which includes 1D chain (m = 1) and 2D layered (m > 1) members, (c) the (111)-oriented family, A′ 2 Aq−1 Bq X3q+3 , formed by cutting along the (111) direction of the 3D parent and featured by 0D isolated clusters (q = 1) and 2D layered (q > 1) members. In each of these layered structures, the perovskite framework is separated by a layer of typically larger organic cations. Source: Saparov and Mitzi [158]. Copyright 2016, American Chemical Society.

Although there is no definitive restriction for the length of the interlayer organic cations for 2D perovskite frameworks, there is a limit for the width of the organic cations, which must be fitted into an area defined by the terminal halides from four adjacent corner-sharing octahedra. This is significant from the perspective of both hydrogen bonding and electron counting schemes. The charge balance requirement dictates the presence of a certain concentration of cations. Too large organic molecules may cause the steric hindrance with adjacent organic molecules, which will render it impossible to fit specific cations into the targeted perovskite framework. However, if they are much smaller than the area provided by the inorganic substructure, the organic molecules can recline to adapt to inorganic structures. Both the organic cations and the anionic inorganic framework have the templating influence on each other, allowing a certain degree of control over the final structures and properties. A family of conducting layered organic–inorganic perovskites, (C4 H9 NH3 )2 (CH3 NH3 )n−1 Snn I3n+1 (n = 1–5), was found with (100)-terminated CH3 NH3 SnI3

762

12 The Relationship Between Structure and Electric Property

CH3NH+3

Sn NH+3 C4H9

b

(101) n=1

n=2

n=3

n→∞

Figure 12.102 Schematic representation of the n = 1 to ∞ (the cubic perovskite) compounds (the long unit-cell dimension b perpendicular to the perovskite sheets increases in the order 27.576(2), 39.395(5), 51.870(3), 64.29(2), and 76.79(2) Å at room temperature as n increases from 1 to 5). Source: Mitzi et al. [167]. Copyright 1994, Springer Nature.

perovskite sheets alternated by butylammonium bilayers [167]. These systems are natural self-assembling analogs to semiconductor quantum-well multilayers, with the perovskite layers acting as wells and the longer-chain alkyl layers forming barrier layers. The thickness of the layer increases with the augmentation of n (Figure 12.102). The organic fragments show significant disorder within the structure, with the methylammonium cation apparently rotating or significantly disordered as in the cubic CH3 NH3 PbI3 perovskite structure. The nitrogen atoms attached to the butyl chain are situated off the plane formed by the outer apical iodine and are displaced toward the Sn planes. Transport measurements of n = 1–5 samples and CH3 NH3 SnI3 were tested using four-point geometry on pressed pellets (Figure 12.103). Increasing n results in a marked drop in resistivity and a transition from the semiconductor (n < 3) to metal (n > 3). When n = 3, the resistivity decreases with the rise of temperature which is below 75 K, presumably due to localization of carriers or gain boundary effects in the pressed pellet. The materials with n = 4 and 5 become progressively more metallic, with merely a very slight upturn in resistivity below 20 K. Room-temperature Hall effect measurements on pressed pellets of n = 3 and 5 materials indicated the carriers are holes. Nevertheless, even when n = 5, the Hall carrier concentration is simply 7 × 1018 cm−3 . In contrast to the (100)-oriented (C4 H9 NH3 )2 (CH3 NH3 )n−l Snn I3n+1 family, [NH2 C(I)=NH2 ]2 (CH3 NH3 )m Snm I3m+2 with layered organic–inorganic perovskites was prepared by an aqueous solution growth technique. The structure determination reveals an unusual structural class with sets of m (110)-oriented CH3 NH3 Snl3 perovskite sheets separated by iodoforma-midinium cations [168]. While the m = 2 compound is semiconducting with a bandgap of 0.33 ± 0.05 eV, increasing m leads to more metallic characters. For m = 2, 3, and 4, the monoclinic lattice constants are a = 6.2649(4) Å, b = 8.6624(5) Å, c = 14.787(2) Å, and 𝛽 = 92.960(8)∘ ; a = 6.2671(3) Å, b = 8.7183(4) Å, c = 38.434(2) Å, and 𝛽 = 92.34(2)∘ ; a = 6.2485(7) Å, b = 8.7162(10) Å, c = 23.653(4) Å, and 𝛽 = 91.83(1)∘ , respectively. In some crystals examined, weak superstructure reflections indicate a doubling or quadrupling of the basic unit cell in the

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials

106 n=1 104

ρ (Ω cm)

102 n=2 n=3

100

n=4 n=5

10–2

10–4

n=∞

0

100

200

300

400

Temperature (K)

Figure 12.103 Resistivity as a function of temperature for pressed-pellet samples of (C4 H9 NH3 )2 (CH3 NH3 )n−1 Snn I3n+1 (n = 1–5) and the cubic (at room temperature) perovskite (n → ∞) CH3 NH3 SnI3 . Source: Mitzi et al. [167]. Copyright 1994, Springer Nature.

amount of increase along the c-axis, perhaps as a result of partial ordering of the iodoformamidinium cations. Electrical resistivity measurements were performed in a displex (closed-cycle refrigeration) system with an airtight cell comprising a calibrated four-point probe and a pressed-pellet sample. Ohmic electrical contacts were made with spring-activated pins directly contacting the sample. The m = 2 sample is semiconducting with room-temperature resistivity of ∼700 ohm cm. Increasing the thickness of the perovskite sheets (larger m) dramatically lowers the resistivity and induces a transition to more metal-like transport properties. This phenomenon has also been observed in the (100)-oriented perovskites (C4 H9 NH3 )2 (CH3 NH3 )n−1 Snn I3n+1 which undergo a semiconductor–metal transition with increasing n. Extended Htickel band structure calculations were performed on the m = 2 and m → ∞ (cubic perovskites) systems to comprehend the nature of the semiconductor–metal transition in these uncommon conducting halides with a basis set consisting of valence orbitals with Sn 5s and 5p as well as I 5s and 5p. The metallic character of CH3 NH3 SnI3 results from the large dispersion of the Sn 5s band (hybridized with I 5p) along the (111) direction of the cubic Brillouin zone, leading to a marginal crossing of the Sn 5s and Sn 5p bands near the R point ([1/2, 1/2, and 1/2] 2π/a), with the Fermi energy falling roughly between the two bands. Low concentrations of defects or vacancies in the structure are expected to influence the observed carrier densities. In terms of the m = 2 structure, the layering of the perovskite structure results in the formation of terminal Sn–I interactions, which raises the Sn 5p antibonding (Sn–I) orbitals to higher energy as well as reduces the band widths, generating an apparent bandgap (of order 0.5 eV). Increasing m leads to progressive stabilization of the Sn 5p antibonding orbitals and broadening of

763

HOOC

+

NH3

PEA

TRA +

A

NH3

C4

HOOC

I

106

+

GABA

+

NH3 +

N–H

I

IPy

NH3

A2

+

NH3

H3 N C5di

2

IPy

1

Compressed powder of C4

00

100

200

300

T (K)

104

AEPI

IC4

GABA

IC4 +

+

108

+

NH3

σ(T)/σ(295 K)

12 The Relationship Between Structure and Electric Property

ρ (Ω cm)

764

NH3 N+ H AEPi

102

TRA

C4

PEA C5di

100 0

100

T (K)

200

300

Figure 12.104 Organic cations used in the synthesis of A2 SnI4 and electrical resistivity (𝜌) of A2 SnI4 compounds as a function of temperature (the inset: temperature dependence of the normalized conductivity of (PEA)2 SnI4 ). Source: Takahashi et al. [169]. Copyright 2007, American Chemical Society.

the Sn 5s and Sn 5p band widths, thereby effectively decreasing the bandgap and producing more metallic characters. In 2D layered perovskites, organic layers that act as a template for the structure and are able to tune the band structure as well. A series of A2 SnI4 (A = organic ammonium) were synthesized using different organic amines (Figure 12.104a) [169]. The inorganic layer is similar as that in (100)-oriented perovskites (C4 H9 NH3 )2 (CH3 NH3 )n−1 Snn I3n+1 . The electron conductivity test showed that the resistivity relies strongly on the type of organic species: the temperature dependence of the resistivity for the A2 SnI4 crystals having the highest resistance (A = IPy and A2 = AEPi) corresponds to normal, thermally activated semiconducting behavior. In contrast, the resistivity of the crystals having the lowest resistance (A = tralomethrin [TRA], 2-phenylethanol [PEA], and C5di) decreases when cooled in the high-temperature region. The structure of the tin iodide perovskite layer in the low-resistance compound (C5 di) SnI4 considerably differs from that in the high-resistance compound (IPy)2 SnI4 . This distinction is prominently reflected in the calculated widths of the valence band, and the width in (C5 di)SnI4 is nearly twice the width in (IPy)2 SnI4 (Figure 12.105). As the valence band (W) decreases, the bandgap (EG) inevitably becomes larger. The major charge carriers for A2 SnI4 are holes. Consequently, the mechanism of hole conduction is facilitated by the excitation of electrons from the valence band to acceptor levels in the bandgap. The acceptor levels (0.01–0.02 eV) are thought to be formed above the TVB. When W is small, the hole mobility should also be small. Therefore, the high resistivity of the compounds with small W is thought to be caused by small hole mobility and large activation energy. The chief components of the valence band are occupied by Sn 5s and I 5p orbitals. Hence, the vacant acceptor levels close to the TVB may be formed by oxidation of the SnI4 2− unit. The electrical resistivity measurement showed that the resistivity of the different percent of SnI4 -doped (PEA)2 SnI4 crystals reduced as the percent of doping increased [169].

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials

Conduction band

3

10

ρ (Ω cm)

Acceptor levels Valence band

As grown

1

10

W

A B

0

10

Low–resistance A2Sni4 (a)

102

High–resistance A2Sni4

–1

10 0

100

T (K)

200

300

(b)

Figure 12.105 Schematic representation of the m = 2 − 4 and ¥ (Cubic perovskites) compounds. Source: Takahashi et al. [169]. Copyright 2007, American Chemical Society.

The electrical properties of a material greatly rest on its electronic structure. T. Umebayashi et al. investigated the electron structures of 3D CH3 NH3 PbI3 and 2D C4 H9 NH3 PbI4 crystals by photoelectron spectroscopy and band calculations using the linear combination of atomic orbitals within the density-functional theory [170]. The calculated bandgap energies are 1.45 eV for the 3D CH3 NH3 PbI3 crystal and 2.24 eV for the 2D C4 H9 NH3 PbI4 crystal. As shown in Figure 12.106a, the lowest unoccupied states consist of Pb 6p–I 5s σ-antibonding and Pb 6p–I 5p π-antibonding orbitals for the [PbI6 ]−4 cluster zero-dimensional (0D) system. The higher occupied states can be decomposed into three parts: a Pb 6s–I 5p σ-antibonding orbital in the top of the states, I 5p orbitals in the middle energy region, and Pb 6p–I 5s s-bonding and Pb 6p–I 5p π-bonding orbitals in the bottom of the state. The Pb 6s–I 5p σ-bonding orbital is situated in the lower-energy region. The local bonding scheme of the 3D and 2D crystals reflects the traits of the [PbI6 ]−4 molecule. Figure 12.106b is the bonding diagram about the TVB and the bottom of the conductive band (BCB) for the 3D and 2D crystals. The TVB for the 3D crystal consists of the Pb 6s–I 5p σ-antibonding orbital, while the BCB is composed of the Pb 6p–I 5s σ-antibonding and Pb 6p–I 5p π-antibonding orbitals. In the 2D crystal, the I1 and I2 electronic orbitals are split into 2D ligand fields. The TVB is made up of the Pb 6s–I1 and I2 5p σ-antibonding orbitals. The BCB of Pb 6p–I1 5s σ-antibonding and Pb 6p–I1 5p π-antibonding orbitals exhibits greater dispersion than that of the Pb 6p–I2 5s σ-antibonding band. (BAESBT)PbI4 (BAESBT = 5,5′ -bis(ammoniumethylsulfanyl)-2,2′ -bithiophene) adopts a triclinic cell (P-1) with the lattice parameters as follows: a = 8.4741(5) Å, b = 8.9255(6) Å, c = 16.876(1) Å, 𝛼 = 88.328(5)∘ , 𝛽 = 81.806(4)∘ , 𝛾 = 88.864(5)∘ , and Z = 2 [171]. The structure of PbI4 2− perovskite sheets resembles that in SnI4 2− -based hybrids. PbI4 2− perovskite sheets and diammonium cation layers alternate along c-axis. Large crystals of (BAESBT)PbI4 were obtained, enabling measurement of the electrical conductivity. Figure 12.107b shows the electrical conductivity, 𝜎, of one (BAESBT) PbI4 single crystal under the electric field of 16 kV/cm. As shown in the inset, d(ln(𝜎))/d(1/T) is roughly temperature independent; conduction in

765

12 The Relationship Between Structure and Electric Property

(i) Pb 6p–I 5s *

(a)

Pb 6p–I 5p * Ef

Energy (a.u.)

Pb 6p

(ii) Pb 6s–I 5p Pb 6s

(iii) I 5p orbitals

I 5p

(iv) Pb 6p–I 5s Pb 6p–I 5p

I 5s

(v) Pb 6p–I 5p

(b)

0D

3D

Pb 6p–I 5s *

Ef

2D

Pb 6p–I1 5s *

Pb 6p–I 5p * Energy (a.u.)

766

Pb 6p–I2 5s * BCB TVB

Pb 6s–I 5p *

Figure 12.106 The bonding diagram of (a) the [PbI6 ]−4 cluster (the zero-dimensional system), (b) 3D crystal CH3 NH3 PbI3 , and (b) 2D crystal (C4 H9 NH3 )2 PbI4 at the top of the valence band and the bottom of the conduction band. Source: Umebayashi et al. [170].

(BAESBT)PbI4 is then dominated by thermally activated processes. The apparent activation energy is 2.535 eV corresponding to the characteristic wavelength of 490 nm. This value is similar to the exciton energy deduced from UV-vis absorption spectra. This similarity is in favor of the constraint of conduction along the inorganic PbI4 2− layers. Finally, the transport within the planes is investigated in Figure 12.107c for a (BAESBT)PbI4 single crystal in the dark at ambient temperature. At the low electric field, an Ohmic behavior is acquired. A further increase of the applied current leads to a crossover to quadratic dependence of the current on the applied electric field. This behavior of the current versus the applied electric field is commonly observed in semiconductors using a standard semiconductor

a

b

1E–8 1E–9

10

–20 3.4

3.5

3.6

3.7 3.8 3.9 1000/T (K−1)

1E–10

(DAESBT)PbI4

1

1E–11 1E–12 3.4

(a)

0 –10 Current (nA)

c

Conductivity (S/cm)

1E–7

d(In(σ)/d(1/T)

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials

3.6

3.8

4.0 1000/T (K−1)

(b)

0.1 1

10 Electric field (kV/cm)

(c)

Figure 12.107 (a) Layered crystal structure of (BAESBT)PbI4 , (b) evolution of conductivity of a (BAESBT)PbI4 single crystal as a function of reciprocal absolute temperature (the inset: the plot of d(ln(𝜎))/d(1/T) to check the activated behavior of the conductivity. Deviations from the horizontal line are due to electrical noise amplified by the calculation of the derivative), (c) evolution of the dark current of a (BAESBT)PbI4 single crystal as a function of the applied electric field, applied along the layers, at room temperature (the dashed line is a fit to an Ohmic behavior where the current is proportional to the electric field, but the solid line denotes a fit to F 2 dependence). Source: Zhu et al. [171]. Copyright 2003, American Chemical Society.

o

a

Current (nA)

120

ON

ON

ON

ON

ON

90

60

30

c

BA+ (a)

0

MA+ (b)

OFF 0

OFF

50

OFF

OFF

100

150

OFF

200

OFF

250

Times (s)

Figure 12.108 (a) Crystal structures of (C4 H9 NH3 )2 –(CH3 NH3 )2 Pb3 Br10 , (b) recyclable switching operation of photocurrent response, showing no obvious attenuation after long-time illumination. Source: Li et al. [172]. Copyright 2017, John Wiley & Sons.

model and is normally interpreted in terms of space-charge limited conduction at intermediate electric fields. Luo and coworkers reported the first hybrid ferroelectric with the 2D multilayered perovskite framework, (C4 H9 NH3 )2 –(CH3 NH3 )2 Pb3 Br10 [172], which was constructed by the tailored alloying of the mixed organic cations into the 3D prototype of CH3 NH3 PbBr3 . (Figure 12.108) As reported in the photodetector of MoS2 /poly (vinylidene fluoride-trifluoroethylene), large ferroelectric polarization enables an ultrahigh electrostatic field (Ei ∼ 109 V/m), which greatly depresses dark currents. With respect to (C4 H9 NH3 )2 –(CH3 NH3 )2 Pb3 Br10 , the Ei value can be approximately estimated from 𝜎 = 𝜀r 𝜀0 𝜀i , where 𝜎 is the charge density at electrode surfaces that is related to the remnant polarization (Pr ); 𝜀r is the dielectric constant; 𝜀0 is the vacuum permittivity. The device of this compound exhibits I ph /I dark (I ph = photocurrent) ratios at least ∼2.5 × 103 from the light-on to light-off.

767

768

12 The Relationship Between Structure and Electric Property

(BdA)PbI4

(BdA)Pb2I6

(b)

(a)

(b)

(a)

(c) c.

(c) Pb I N C

Pb I C N

O

(A)

One-dimensional structure

(B)

Two-dimensional structure

Figure 12.109 Crystal structure of (A) solvated (BdA)Pb2 I6 and (B) (BdA)PbI4 along three axes. Source: Safdari et al. [174]. Copyright 2017, The Royal Society of Chemistry.

12.4.2.3 One-Dimensional (1D) Organic–Inorganic Hybrids

In addition to the perovskite common vertex link, the halide octahedron link has various forms, such as edge-sharing and face-sharing [173]. The 1D inorganic chains in the (BdA)Pb2 I6 chemical structure comprise face-sharing lead iodide octahedral units. The organic dications and solvent in the structure were disordered, while the structure of (BdA)PbI4 is the 2D layer (Figure 12.109) [174]. The conductivity of (BdA)Pb2 I6 was recorded as 5.3 × 10−6 S/cm compared with 1.3 × 10−5 S/cm for (BdA)PbI4 . These conductivity values rely highly on the chemical structure and dimensionalities of the two materials. In (BdA)PbI4 , charge conduction occurs through the lead iodide planes, while for (BdA)Pb2 I6 connectivity and charge transport are solely along lead iodide chains. Also, the I–V curve measurements for the single crystals of [Cu(2,2′ -bipy)2 I] (Pb2 I6 ) and [(H2 EDAB)](Pb2 I6 ) ([(H2 EDAB)2+ ] = Et2 HNC6 H4 C6 H4 NHEt2 ) with a two-terminal-probe direct current method along the extension direction of face-sharing (Pb2 I6 )2− chains showed that the conductivity of [Cu(2,2′ -bipy)2 I] (Pb2 I6 ) and [(H2 EDAB)](Pb2 I6 ) is 2.50 × 10−10 and 1.30 × 10−10 S/cm, respectively. These values are very alike due to their similar band structures [175]. 1D lead halide can also be linked in a variety of ways, such as nanotubes. Xu and coworkers reported a metal halide-based crystalline nanotube array constructed from an unprecedented giant [PbII 18 I54 (I2 )9 ] wheel cluster (Figure 12.110a) [176]. The electrical properties of this single crystal were studied to have typical semiconductivity and highly anisotropic conductivity (Figure 12.110b–e). Conductivity along and perpendicular to the nanotube has positive temperature effect, demonstrating its semiconductive features. Under room temperature, the nanotube has conductivity of 0.9 × 10−10 and 0.8 × 10−9 S/cm along and perpendicular to the nanotube, respectively. Nevertheless, when the temperature is 180 ∘ C, the conductivity reaches their highest values of 1.5 × 10−8 and 0.9 × 10−6 S/cm correspondingly along and perpendicular to the nanotube. The activation energy of conductivity along and perpendicular to the nanotube is 0.28 and 0.54 eV, respectively.

12.4 Structure and Electrical Properties of Organic–Inorganic Hybrid Conductive Materials

C

I–I covalent bond 2.834 Å

(a) 2.8×10-9

140 120 100 80 60 40

1.4×10-9

180 °C 160 140 120 100 80 60 40

6.0×10-7

I (A)

2.1×10-9

I (A)

8.0×10-7

180 °C 160

4.0×10-7 2.0×10-7

7.0×10-10

0.0

0.0 0

1

2

3

4

5

1

2

3

U (V)

5

U (V)

(b)

(c)

–7.6

–5.6

–8.0

Ea = 0.28 eV

log(σ) (S/cm)

log(σ) (S/cm)

4

–8.4 –8.8

Ea = 0.54 eV

–6.4 –7.2 –8.0

–9.2 –8.8 –9.6 –9.6 –10.0 2.2

2.4

2.6

2.8

3.0

3.2

3.4

2.2

1000/T (K–1) (d)

2.4

2.6

2.8

3.0

3.2

3.4

1000/T (K–1) (e)

Figure 12.110 The 1D nanotube of [PbII 18 I54 (I2 )9 ]: (a) temperature-dependent I–V curves for the nanotube along and perpendicular to the nanotube, (c) Arrhenius plots of the nanotube along the a-direction (d) and c-direction (e) (E a is the activation energy). Source: Wang et al. [176]. Copyright 2016, John Wiley & Sons.

(MV)4 [Bi6 Cl26 ] results from the combination of three (MV)[Bi2 Cl8 ] subunits and one (MV)Cl2 , and the inorganic networks in this compound are segments of the 1D infinite inorganic anionic chains (Figure 12.111). Hence, in every three subunits, one methylviologen dichloride unit is inserted. (MV)4 [Bi6 Cl26 ] transforms into black upon UV irradiation. This photochromic property is probably on account of the photoinduced electron transfer from the anionic part to the methylviologen dications. The single-crystal electrical conductivity measurements of (MV)4 [Bi6 Cl26 ] before and after irradiation show that both of them are semiconductors with weak room

769

12 The Relationship Between Structure and Electric Property –7.5 Before UV irradiation

–8

b c CI N

c

Conductivity (S/m–1)

770

–8.5 After UV irradiation

–9 –9.5

C

–10

H BI

–10.5 3.3

(a)

3.8

4.3

4.8

5.3

1000/T (K–1)

(b)

Figure 12.111 (a) The general view along the a-axis of the structure of (MV)4 [Bi6 Cl26 ], (b) conductivity of a single crystal of (MV)4 [Bi6 Cl26 ] before and after UV irradiation, as a function of the inverse of the temperature (dotted lines represent the fit of the data by an Arrhenius law (ln 𝜎 ∝ E g /2k B T) shifted from the data for clarity). Source: Leblanc et al. [177]. Copyright 2010, American Chemical Society.

temperature conductivity and that the bandgap of the irradiated crystal (0.35 eV) is much smaller than that of the original hybrid (1.0 eV) [177].

References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Carolan, D. (2017). Prog. Mater. Sci. 90: 128–158. Sun, B., Shao, M., and Lee, S. (2016). Adv. Mater. 28: 10539–10547. Mu, Z., Yu, H., Zhang, M. et al. (2017). Nano Lett. 17: 1552–1558. Zhao, Y., Ma, H., Dong, T. et al. (2018). Nano Lett. 18: 6931–6940. King, P.D. and Veal, T.D. (2011). J. Phys. Condens. Matter 23: 334214. Hosono, H. and Ueda, K. (2017). Handbook of Electronic and Photonic Materials, Chapter 58. Springer International Publishing AG. ISBN: 978-3-319-48933-9. Kawazoe, H., Yanagi, H., Ueda, K., and Hosono, H. (2000). MRS Bull. 25: 28–36. Kawazoe, H., Ueda, N., Un’no, H. et al. (1994). J. Appl. Phys. 76: 7935–7941. Pauling, L. (1960). The Nature of the Chemical Bond. Cornell University Press. ISBN-13: 978-0801403330. Shannon, R.D., Gillson, J.L., and Bouchard, R.J. (1977). J. Phys. Chem. Solids 38: 877–881. Sasabayashi, T., Ito, N., Nishimura, E. et al. (2003). Thin Solid Films 445: 219–223. Furubayashi, Y., Hitosugi, T., Yamamoto, Y. et al. (2005). Appl. Phys. Lett. 86: 252101. Wei, R.H., Tang, X.W., Hui, Z.Z. et al. (2015). Appl. Phys. Lett. 106: 101906. Kawazoe, H., Yasukawa, M., Hyodo, H. et al. (1997). Nature 389: 939–942. Yanagi, H., Kawazoe, H., Kudo, A. et al. (2000). J. Electroceram. 4: 407–414. Yanagi, H., Inoue, S.-i., Ueda, K. et al. (2000). J. Appl. Phys. 88: 4159. Liu, A., Zhu, H., Guo, Z. et al. (2017). Adv. Mater. 29: 1701599.

References

18 Mizoguchi, H., Hirano, M., Fujitsu, S. et al. (2002). Appl. Phys. Lett. 80: 1207–1209. 19 Hu, L., Wei, R., Yan, J. et al. (2018). Adv. Electron. Mater. 4: 1700476. 20 Xu, J., Liu, J.-B., Wang, J. et al. (2018). Adv. Funct. Mater. 28: 1800332. 21 Wei, R., Tang, X., Hu, L. et al. (2017). J. Mater. Chem. C 5: 1885–1892. 22 Patsalas, P., Kalfagiannis, N., Kassavetis, S. et al. (2018). Mater. Sci. Eng., R 123: 1–55. 23 H. Morkoc, U. Ozgur, Zinc Oxide: Fundamentals Materials and Device Technology, 2009. https://doi.org/10.1002/9783527623945. 24 Matenoglou, G.M., Koutsokeras, L.E., and Patsalas, P. (2009). Appl. Phys. Lett. 94: 152108. 25 Nigro, A., Nobile, G., Rubino, M.G., and Vaglio, R. (1988). Phys. Rev. B 37: 3970–3972. 26 Duan, X.F., Mi, W.B., Guo, Z.B., and Bai, H.L. (2013). J. Appl. Phys. 113: 023701. 27 Metaxa, C., Kassavetis, S., Pierson, J.F. et al. (2017). ACS Appl. Mater. Interfaces 9: 10825–10834. 28 Pampili, P. and Parbrook, P.J. (2017). Mater. Sci. Semicond. Process. 62: 180–191. 29 Akasaki, I. (2002). J. Cryst. Growth 237–239: 905–911. 30 Wu, S., Yang, X., Zhang, H. et al. (2018). Phys. Rev. Lett. 121: 145505. 31 Chen, Y., Liu, K., Liu, J. et al. (2018). J. Am. Chem. Soc. 140: 16392–16395. 32 Hinuma, Y., Hatakeyama, T., Kumagai, Y. et al. (2016). Nat. Commun. 7: 11962. 33 Yu, R., Chong, X., Jiang, Y. et al. (2015). RSC Adv. 5: 1620–1627. 34 Sun, W., Bartel, C.J., Arca, E. et al. (2019). Nat. Mater. 18: 732–739. 35 Yanase, Y. and Yorozu, N. (2008). Sci. Technol. Adv. Mater. 9: 044201. 36 Sun, M.J., Cao, X., and Cao, Z. (2016). ACS Appl. Mater. Interfaces 8: 16551–16554. 37 Chen, X., Jiang, J., Liang, Q. et al. (2016). J. Mater. Chem. C 4: 7406–7414. 38 Zhou, J., Zha, X.H., Yildizhan, M. et al. (2019). ACS Nano 13: 1195–1203. 39 Zhou, J., Zha, X., Chen, F.Y. et al. (2016). Angew. Chem. Int. Ed. 55: 5008–5013. 40 Makovicky, E. (2006). Rev. Mineral. Geochem. 61: 7–125. 41 Vaughan, D.J. and Corkhill, C.L. (2017). Elements 13: 81–87. 42 Greyson, E.C., Barton, J.E., and Odom, T.W. (2006). Small 2: 368–371. 43 BenNasr, T., Maghraoui-Meherzi, H., BenAbdallah, H., and Bennaceur, R. (2011). Physica B 406: 287–292. 44 Fang, Y., Hu, X., Zhao, W. et al. (2019). J. Am. Chem. Soc. 141: 790–793. 45 Fang, Y., Pan, J., He, J. et al. (2018). Angew. Chem. Int. Ed. 57: 1232–1235. 46 Yang, S., Li, Y., Wang, X. et al. (2014). Nanoscale 6: 2582–2587. 47 Han, S.K., Gu, C., Zhao, S. et al. (2016). J. Am. Chem. Soc. 138: 12913–12919. 48 Shi, X., Chen, H., Hao, F. et al. (2018). Nat. Mater. 17: 421–426. 49 Akamatu, H., Inokuchi, H., and Matsunaga, Y. (1954). Nature 173: 168–169. 50 Wang, C., Dong, H., Jiang, L., and Hu, W. (2018). Chem. Soc. Rev. 47: 422–500. 51 Hu, W. and Jiang, H. (2020). Angew. Chem. Int. Ed. 59: 1408–1428. 52 Park, C., Park, J.E., and Choi, H.C. (2014). Acc. Chem. Res. 47: 2353–2364.

771

772

12 The Relationship Between Structure and Electric Property

53 (a) Shao, W., Dong, H.L., Jiang, L., and Hu, W.P. (2011). Chem. Sci. 2: 590–600. (b) Dong, H., Fu, X., Liu, J. et al. (2013). Adv. Mater. 25: 6158–6182. 54 (a) Lim, J.A., Liu, F., Ferdous, S. et al. (2010). Mater. Today 13: 14–24. (b) Kim, D.H., Park, Y.D., Jang, Y. et al. (2005). Macromol. Rapid Commun. 26: 834–839. (c) Ihn, K.J., Moulton, J., and Smith, P. (1993). J. Polym. Sci., Part B: Polym. Phys. 31: 735–742. 55 Jang, J., Nam, S., Im, K. et al. (2012). Adv. Funct. Mater. 22: 1005–1014. 56 (a) Jiang, H., Hu, P., Ye, J. et al. (2018). J. Mater. Chem. C 6: 1884–1902. (b) Mori, T. (2004). Chem. Rev. 104: 4947–4970. (c) Yamada, J.-i., Akutsu, H., Nishikawa, H., and Kikuchi, K. (2004). Chem. Rev. 104: 5057–5084. 57 (a) Giri, G., Verploegen, E., Mannsfeld, S.C.B. et al. (2011). Nature 480: 504–U124. (b) Liu, Z., Becerril, H.A., Roberts, M.E. et al. (2009). IEEE Trans. Electron Dev. 56: 176–185. 58 Wang, C., Dong, H., Hu, W. et al. (2012). Chem. Rev. 112: 2208–2267. 59 Park, S.K., Kim, J.H., and Park, S.Y. (2018). Adv. Mater. 30: 1704759. 60 Zhen, Y., Inoue, K., Wang, Z. et al. (2018). J. Am. Chem. Soc. 140: 62–65. 61 Lin, X., Wegner, B., Lee, K.M. et al. (2017). Nat. Mater. 16: 1209–1217. 62 Kobayashi, Y., Terauchi, T., Sumi, S., and Matsushita, Y. (2016). Nat. Mater. 16: 109–114. 63 Huang, X., Zhang, S., Liu, L. et al. (2018). Angew. Chem. Int. Ed. 57: 146–150. 64 Yu, P., Zhen, Y., Dong, H., and Hu, W. (2019). Chem 5 (11): 2814–2853. 65 Liu, J., Zhang, H., Dong, H. et al. (2015). Nat. Commun. 6: 10032. 66 Li, J., Zhou, K., Liu, J. et al. (2017). J. Am. Chem. Soc. 139: 17261–17264. 67 Li, R., Jiang, L., Meng, Q. et al. (2009). Adv. Mater. 21: 4492–4495. 68 Minemawari, H., Tanaka, M., Tsuzuki, S. et al. (2017). Chem. Mater. 29: 1245–1254. 69 Dou, J.-H., Zheng, Y.-Q., Yao, Z.-F. et al. (2015). J. Am. Chem. Soc. 137: 15947–15956. 70 Wang, M., Li, J., Zhao, G. et al. (2013). Adv. Mater. 25: 2229–2233. 71 (a) Jiang, H., Yang, X., Cui, Z. et al. (2007). Appl. Phys. Lett. 91: 123505. (b) Mas-Torrent, M., Hadley, P., Bromley, S.T. et al. (2005). Appl. Phys. Lett. 86: 012110. (c) Naraso, J., Nishida, I., Ando, S. et al. (2005). J. Am. Chem. Soc. 127: 10142–10143. 72 Singh, K., Sharma, A., Zhang, J. et al. (2011). Chem. Commun. 47: 905–907. 73 Wang, Y., Zhu, W.G., Dong, H.L. et al. (2016). Top. Curr. Chem. 374: 83. 74 Wohrle, T., Wurzbach, I., Kirres, J. et al. (2016). Chem. Rev. 116: 1139–1241. 75 Rivnay, J., Jimison, L.H., Northrup, J.E. et al. (2009). Nat. Mater. 8: 952. 76 Zhang, Z.P., Jiang, L., Cheng, C.L. et al. (2016). Angew. Chem. Int. Ed. 55: 5206–5209. 77 Liu, H., Cao, X., Wu, Y. et al. (2014). Chem. Commun. 50: 4620–4623. 78 Hou, X.-Q., Sun, Y.-T., Liu, L. et al. (2016). Chin. Chem. Lett. 27: 1166–1174. 79 Zhang, J., Xu, W., Sheng, P. et al. (2017). Organic Donor–Acceptor Complexes as Novel Organic Semiconductors. Acc. Chem. Res. 50 (7): 1654–1662. 80 Goetz, K.P., Vermeulen, D., Payne, M.E. et al. (2014). J. Mater. Chem. C 2: 3065–3076.

References

81 (a) Sun, L., Zhu, W., Wang, W. et al. (2017). Angew. Chem. Int. Ed. 56: 7831–7835. (b) Zhu, W., Zheng, R., Fu, X. et al. (2015). Angew. Chem. Int. Ed. 54: 6785–6789. (c) Zhu, W., Zhu, L., Sun, L. et al. (2016). Angew. Chem. Int. Ed. 55: 14023–14027. (d) Zhu, W., Zheng, R., Zhen, Y. et al. (2015). J. Am. Chem. Soc. 137: 11038–11046. 82 Black, H.T. and Perepichka, D.F. (2014). Angew. Chem. Int. Ed. 53: 2138–2142. 83 Zhang, H., Jiang, L., Zhen, Y. et al. (2016). Adv. Electron. Mater. 2: 1500423. 84 Wang, Y., Zhu, W., Du, W. et al. (2018). Angew. Chem. Int. Ed. 57: 3963–3967. 85 Zhu, W., Yi, Y., Zhen, Y., and Hu, W. (2015). Small 11: 2150–2156. 86 Vermeulen, D., Zhu, L.Y., Goetz, K.P. et al. (2014). J. Phys. Chem. C 118: 24688–24696. 87 Zhang, J., Gu, P., Long, G. et al. (2016). Chem. Sci. 7: 3851–3856. 88 Dou, J.-H., Yu, Z.-A., Zhang, J. et al. (2019). J. Am. Chem. Soc. 141: 6561–6568. 89 Lusi, M. (2018). CrystEngComm 20: 7042–7052. 90 Xu, X.M., Shan, B.W., Kalytchuk, S. et al. (2014). Chem. Commun. 50: 12828–12831. 91 Zhen, Y., Tanaka, H., Harano, K. et al. (2015). J. Am. Chem. Soc. 137: 2247–2252. 92 Ferey, G. (2008). Chem. Soc. Rev. 37: 191–214. 93 D’Alessandro, D.M., Kanga, J.R.R., and Caddy, J.S. (2011). Aust. J. Chem. 64: 718–722. 94 Rosi, N.L., Eckert, J., Eddaoudi, M. et al. (2003). Science 30: 1127–1129. 95 Silva, C.G., Corma, A., and García, H. (2010). J. Mater. Chem. 20: 3141–3156. 96 Feng, P.L., Perry, J.J., Nikodemski, S. et al. (2010). J. Am. Chem. Soc. 132: 15487–15489. 97 Allendorf, M.D., Schwartzberg, A., Stavila, V., and Talin, A.A. (2011). Chem 17: 11372–11388. 98 Sun, L., Miyakai, T., Seki, S., and Dinca, M. (2013). J. Am. Chem. Soc. 135: 8185–8188. 99 Nathaniel, J.K., Rosi, L., Eddaoudi, M. et al. (2005). J. Am. Chem. Soc. 127: 1504–1518. 100 Narayan, T.C., Miyakai, T., Seki, S., and Dinca, M. (2012). J. Am. Chem. Soc. 134: 12932–12935. 101 Park, S.S., Hontz, E.R., Sun, L. et al. (2015). J. Am. Chem. Soc. 137: 1774–1777. 102 Wang, Q.-X., Zeng, M.-H., Tan, Y.-X. et al. (2010). J. Am. Chem. Soc. 132: 2561–2563. 103 Chui, S.S.-Y., Lo, S.M.-F., Charmant, J.P.H. et al. (1999). Science 283: 1148–1150. 104 Stavila, V., Talin, A.A., and Allendorf, M.D. (2014). Chem. Soc. Rev. 43: 5994–6010. 105 Talin, A.A., Centrone, A., Ford, A.C. et al. (2014). Science 343: 66–69. 106 Takaishi, S., Hosoda, M., Kajiwara, T. et al. (2009). Inorg. Chem. 48: 9048–9050. 107 Wojdel, J.C., Moreira, I.d.P.R., Bromley, S.T., and Illas, F. (2019). Prediction of half-metallic conductivity in Prussian Blue derivatives. J. Mater. Chem. 19 (14): 2032.

773

774

12 The Relationship Between Structure and Electric Property

108 Tennakone, K. and Dharmaratne, W.G.D. (1983). J. Phys. C: Solid State Phys. 16: 5633–5639. 109 Inoue, H., Nakazawa, T., Mitsuhashi, T. et al. (1989). Hyperfine Interact. 46: 725–731. 110 Xidis, A. and Neff, V.D. (1991). J. Electrochem. Soc. 138: 3637–3642. 111 Tokoro, H. and Ohkoshi, S. (2011). Dalton Trans. 40: 6825–6833. 112 Molnár, G., Cobo, S., Mahfoud, T. et al. (2009). J. Phys. Chem. C 113: 2586–2593. 113 Behera, J.N., D’Alessandro, D.M., Soheilnia, N., and Long, J.R. (2009). Chem. Mater. 21: 1922–1926. 114 Sato, O., Kawakami, T., Kimura, M. et al. (2004). J. Am. Chem. Soc. 126: 13176–13177. 115 (a) Pimentel, B.R., Parulkar, A., Zhou, E.K. et al. (2014). ChemSusChem 7: 3202–3240. (b) Stock, N. and Biswas, S. (2012). Chem. Rev. 112: 933–969. 116 Chen, E.X., Yang, H., and Zhang, J. (2014). Inorg. Chem. 53: 5411–5413. 117 Chen, E.X., Fu, H.R., Lin, R. et al. (2014). ACS Appl. Mater. Interfaces 6: 22871–22875. 118 Aromí, G., Barrios, L.A., Roubeau, O., and Gamez, P. (2011). Coord. Chem. Rev. 255: 485–546. 119 Gandara, F., Uribe-Romo, F.J., Britt, D.K. et al. (2012). Chem 18: 10595–10601. 120 Acker, D.S., Harder, R.J., Hertler, W.R. et al. (1960). J. Am. Chem. Soc. 82: 6408–6409. 121 Tseng, T., Urban, C., Wang, Y. et al. (2010). Charge-transfer-induced structural rearrangements at both sides of organic/metal interfaces. Nat. Chem. 2: 374–379. 122 Khatkale, M.S. and Devlin, J.P. (1979). J. Chem. Phys. 70: 1851–1859. 123 Melby, L.R., Harder, R.J., Hertler, W.R. et al. (1962). J. Am. Chem. Soc. 84: 3374–3387. 124 Heintz, R.A., Zhao, H., Ouyang, X. et al. (1999). Inorg. Chem. 38: 144–156. 125 Liu, Y., He, M., Meng, Q. et al. (2012). Small 8: 557–560, 478. 126 Müller, R., Genoe, J., and Heremans, P. (2006). Appl. Phys. Lett.: 88. 127 Qu, N., Lu, W., Pang, S.M. et al. (1994). Thin Solid Films 243: 468–471. 128 Liu, H., Liu, Z., Qian, X. et al. (2010). Cryst. Growth. Des. 10: 237–243. 129 Wang, K., Qian, X., Zhang, L. et al. (2013). ACS Appl. Mater. Interfaces 5: 5825–5831. 130 Basori, R., Das, K., Kumar, P. et al. (2013). Appl. Phys. Lett. 102: 061111. 131 Liu, Y., Ji, Z., Li, H. et al. (2009). Nano Res. 2: 630–637. 132 Ji, H.X., Hu, J.S., Guo, Y.G. et al. (2008). Adv. Mater. 20: 4879–4882. 133 O’Kane, S.A., Clérac, R., Zhao, H. et al. (2000). J. Solid State Chem. 152: 159–173. 134 Shield, L. (1985). J. Chem. Soc., Faraday Trans. 81: 1–9. 135 Liu, H., Zhao, Q., Li, Y. et al. (2005). J. Am. Chem. Soc. 127: 1120–1121. 136 Avendano, C., Zhang, Z., Ota, A. et al. (2011). Angew. Chem. Int. Ed. 50: 6543–6547.

References

137 Zhao, H., Heintz, R.A., Ouyang, X., and Dunbar, K.R. (1999). Chem. Mater. 11: 736–746. 138 Clérac, R., O’Kane, S., Cowen, J. et al. (2003). Chem. Mater. 15: 1840–1850. 139 Miyasaka, H., Motokawa, N., Matsunaga, S. et al. (2010). J. Am. Chem. Soc. 132: 1532–1544. 140 Zhang, X., Zhang, Z., Zhao, H. et al. (2014). Chem. Commun. 50: 1429–1431. 141 Hmadeh, M., Lu, Z., Liu, Z. et al. (2012). Chem. Mater. 24: 3511–3513. 142 Sheberla, D., Sun, L., Blood-Forsythe, M.A. et al. (2014). J. Am. Chem. Soc. 136: 8859–8862. 143 Cui, J. and Xu, Z. (2014). Chem. Commun. 50: 3986–3988. 144 Kambe, T., Sakamoto, R., Hoshiko, K. et al. (2013). J. Am. Chem. Soc. 135: 2462–2465. 145 Kambe, T., Sakamoto, R., Kusamoto, T. et al. (2014). J. Am. Chem. Soc. 136: 14357–14360. 146 Huang, X., Sheng, P., Tu, Z. et al. (2015). Nat. Commun. 6: 7408. 147 Lahiri, N., Lotfizadeh, N., Tsuchikawa, R. et al. (2017). J. Am. Chem. Soc. 139: 19–22. 148 Zhao, Y., Hong, M., Liang, Y. et al. (2001). Chem. Commun.: 1020–1021. 149 Su, W., Hong, M., Weng, J. et al. (2000). Angew. Chem. Int. Ed. 39: 2911–2914. 150 Givaja, G., Amo-Ochoa, P., Gomez-Garcia, C.J., and Zamora, F. (2012). Chem. Soc. Rev. 41: 115–147. 151 Takaishi, S., Kawakami, D., Yamashita, M. et al. (2006). J. Am. Chem. Soc. 128: 6420–6425. 152 Takaishi, S., Tobu, Y., Kitagawa, H. et al. (2004). J. Am. Chem. Soc. 126: 1614–1615. 153 Otsubo, K., Wakabayashi, Y., Ohara, J. et al. (2011). Nat. Mater. 10: 291–295. 154 Hermosa, C., Vicente Alvarez, J., Azani, M.R. et al. (2013). Nat. Commun. 4: 1709. 155 Guijarro, A., Castillo, O., Welte, L. et al. (2010). Adv. Funct. Mater. 20: 1451–1457. 156 Mitsumi, M., Yoshida, Y., Kohyama, A. et al. (2009). Inorg. Chem. 48: 6680–6691. 157 Yang, Z., Ebihara, M., Kawamura, T. et al. (2001). Inorg. Chem. 321: 97. 158 Saparov, B. and Mitzi, D.B. (2016). Chem. Rev. 116: 4558–4596. 159 Xing, G., Mathews, N., Sun, S. et al. (2013). Science 342: 344–347. 160 Stoumpos, C.C., Malliakas, C.D., and Kanatzidis, M.G. (2013). Inorg. Chem. 52: 9019–9038. 161 Knop, O., Wasylishen, R.E., White, M.A. et al. (1989). Can. J. Chem. 68: 412–421. 162 Xiao, Z., Yuan, Y., Shao, Y. et al. (2015). Nat. Mater. 14: 193–198. 163 Yang, T.Y., Gregori, G., Pellet, N. et al. (2015). Angew. Chem. Int. Ed. 54: 7905–7910. 164 Mitzi, D.B. and Liang, K. (1997). J. Solid State Chem. 134: 376–381. 165 Wang, G.E., Xu, G., Wang, M.S. et al. (2015). Chem. Sci. 6: 7222–7226.

775

776

12 The Relationship Between Structure and Electric Property

166 Sun, C., Wang, M.S., Li, P.X., and Guo, G.C. (2017). Angew. Chem. Int. Ed. 56: 554–558. 167 Mitzi, D.B., Feild, C.A., Harrison, W.T.A., and Guloy, A.M. (1994). Nature 369: 467–469. 168 Mitzi, D.B., Wang, S., Feild, C.A. et al. (1995). Science 267: 1473–1476. 169 Takahashi, Y., Obara, R., Nakagawa, K. et al. (2007). Chem. Mater. 19: 6312–6316. 170 Umebayashi, T., Asai, K., Kondo, T., and Nakao, A. (2003). Phys. Rev. B: 67. 171 Zhu, X.-H., Mercier, N., Frère, P. et al. Inorg. Chem. 42: 5330–5339. 172 Li, L., Sun, Z., Wang, P. et al. (2017). Angew. Chem. Int. Ed. 56: 12150–12154. 173 Wang, G.-E., Jiang, X.-M., Zhang, M.-J. et al. (2013). CrystEngComm: 15. 174 Safdari, M., Svensson, P.H., Hoang, M.T. et al. (2016). J. Mater. Chem. A 4: 15638–15646. 175 Wang, G.-E., Yao, M.-S., Cai, M.-L. et al. (2017). J. Mater. Chem. A 5: 18409–18413. 176 Wang, G.E., Xu, G., Liu, B.W. et al. (2016). Angew. Chem. Int. Ed. 55: 514–518. 177 Leblanc, N., Bi, W., Mercier, N. et al. (2010). Inorg. Chem. 49: 5824–5833.

777

13 Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets Zhu Zhuo 1,2 , Guo-Ling Li 1,2 , and You-Gui Huang 1,2 1 Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, CAS Key Laboratory of Design and Assembly of Functional Nanostructures, and Fujian Provincial Key Laboratory of Nanomaterials, 155 Yangqiao Road West, Fuzhou 350002, P.R. China 2 Xiamen Institute of Rare Earth Materials, Haixi Institutes, Chinese Academy of Sciences, Xiamen 361021, P.R. China

13.1 Introduction As one of the important functionalities of materials, magnetism has been a basic research subject in physics, chemistry, and material science [1–3]. Traditional magnetism known as “atom-based magnetism” has been explored mainly for inorganic substances in which electron spins on atomic orbitals contribute to the magnetism [4]. Since the discovery of bulk ferromagnetism in molecular complexes such as [FeIII (C5 (CH3 )5 )2 ][TCNE] and p-nitrophenyl nitronyl-nitroxide, magnetism exploration has been extended to coordination complexes and pure organic complexes in which the magnetism originates from unpaired electrons on delocalized molecular orbitals [5, 6]. These new magnetic complexes have been termed “molecular magnets” by Kahn [7]. A shift from atom-based inorganic traditional magnets to molecular magnets may enable the fine modulation of the magnetic properties by chemical tuning, and the ease process ability may be improved by the use of methods applicable to organic materials [8–10]. Molecular magnets may also provide the technological basis for new magnetic materials or devices and complement the traditional magnets. A chemically stable molecular magnet may be potentially used for various practical applications, including permanent bulk magnets, magnetic, and magneto-optic recording. In contrast to traditional magnets, molecular materials are usually insulating. This should lead to a variety of optical properties that may find technological utilities such as electro- or photomagnetic switch and polarized light manipulation in integrated optical devices. In addition, new phenomena and mechanisms for cooperative magnetic phenomena may emerge. Potential areas encompass liquid magnets, magnetic ink, layered Langmuir–Blodgett (LB)-based thin film and multilayered magnets, magnetostrictive sensors, microwave materials, magnetic bubbles, soft magnetic materials with low coercive fields for ac motors, generators, Advanced Structural Chemistry: Tailoring Properties of Inorganic Materials and their Applications, First Edition. Edited by Rong Cao. © 2021 WILEY-VCH GmbH. Published 2021 by WILEY-VCH GmbH.

778

13 Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets

and transformers, as well as biocompatibility-induced magnetic imaging and transducers for medical implants. These potential applications are highly depending on the magnetic properties which are further closely correlated with molecular structures. According to the structural dimensionalities, molecule-based magnetic materials can be classified into discrete molecules, clusters, and extended structures [11, 12]. The last of which is presented by coordination polymers featuring single-metal ions or recognizable building blocks of metal complexes or clusters bridged by organic ligands into one-dimensional (1D), two-dimensional (2D), and three-dimensional (3D) architectures. Commensurate with the structural complexity, their magnetic properties are usually complex and thus not readily understood. In the courses of the studies of molecular magnetism, materials exhibiting magnetic phenomena with potential applications such as slow magnetic relaxation, spin transition, spin frustration, and magnetic ordering have been mainly focused. To correlate the sophisticated molecular structures with their magnetic phenomena, several general approaches have been proposed. Representative examples include a symbol approach focusing on the relationship between the dimension of the overall structure and the overall magnetic properties [11], an approach mainly focusing on the effect of the magnetic anisotropy of spin carrier on single-ion magnet (SIM) behavior [13], an approach focusing on the influence of structure on spin transition [14], and a symmetric approach to understand the correlation between framework topologies and magnetic phenomena [15]. This chapter is therefore not intended to be a comprehensive collection of these materials. Rather, according to the origin of the magnetic phenomena, an overview of magneto-structural correlation concerning magnetic behaviors of spin transition, SIM, single-molecule magnet (SMM), and single-chain magnet (SCM) is presented to interpret the seemingly complex magneto-structural correlations exhibited by the large number of reported molecular magnets. In addition, the effects of bridging modes of some special ligands such as oxalate and azido on the magnetic exchanges they mediate and the impacts of the framework topologies on spin frustration and magnetic ordering are discussed. Significantly, the in-depth understanding of magneto-structural correlation may guide the future design of molecular materials with desired magnetic phenomena. The first part is dedicated to the correlation between the structures of magnetic molecules and their spin-crossover properties. In these materials, paramagnetic centers undergo a spin transition from high-spin state to low-spin state enabling them to be good candidates for application as memory devices and molecular switches. This kind of transition can be thermally and/or photoinduced. Several structural features including molecular shape and intermolecular interactions have great influence on their spin crossover properties. In the second part, we will discuss the influences of molecular structures on slow magnetic relaxation behaviors exhibited by SMMs, SIMs, and SCMs. Magnetic anisotropy, ground states, and magnetic couplings between spin centers are supposed to be important for these low-dimensional magnetic behaviors. The third part presents the effects of structural features of some particular ligands such as oxalate and azido on the magnetic coupling mediated by them. Many structural parameters such as coordination

13.2 Magneto-Structural Correlations in Spin-Crossover Compounds

modes, bridging angles, bond distances, dihedral angles, magnetic orbitals of spin carriers, etc. are major factors in determining the nature and magnitude of exchange interactions. The diversities of the magnetic exchanges between spin carriers lead to various magnetic behaviors including various long-range magnetic ordering (presenting magnetic cooperativity), spin crossover (presenting elastic cooperativity), and SMM phenomena (no cooperative properties). At last, we will introduce the relationship between the framework topologies and the cooperative magnetic properties of magnetic frameworks. Magnetic lattices containing odd-sided polyhedral with antiferromagnetic coupling have attracted great interest due to their intrinsic spin frustration resulting from competing magnetic couplings. Spin frustration in combination with metal ions possessing different g values may produce complex magnetic phase behaviors. The examples presented here will show that these geometrically frustrated magnets provide an ideal platform to create new functional materials exhibiting quantum behaviors.

13.2 Magneto-Structural Correlations in Spin-Crossover Compounds Spin crossover also called spin transition or spin equilibrium behavior is a phenomenon that involves the rearrangement of electrons in a metal ion, from a high-spin state to a low-spin state under the external stimuli such as temperature, photo, and pressure [16]. This phenomenon is an entropic effect, occurring at a temperature where the enthalpy of the stronger metal–ligand (M—L) bonds in the low-spin state is overcome by the higher configurational and vibrational entropy of the high-spin form [17]. These two different states correspond to the different distributions of electrons within the orbital energy levels of metal ions that possess the maximum and minimum number of unpaired electrons, respectively. For example, the conversion between the low-spin and high-spin states of a Fe2+ spin-crossover complex can be described in terms of the “1D” potential energy diagram (Figure 13.1). Although this phenomenon is often studied in the solid state, it can occur in any phase of matter. A large number of physical properties of a solid material, for example, magnetic moment, dielectric constant, color, etc., may be 5

1

T2

A1

Potential energy, E

Figure 13.1 One-dimensional potential energy diagram for ΔS = 2 spin-crossover transition in a d6 transition metal complex.

High spin ΔE Low spin Reaction coordinate

779

780

13 Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets

influenced by spin crossover. Because of their switching properties, spin-crossover materials have been applied in several areas. They include: display and memory devices which make use of a spin-crossover material whose color or dielectric constant is switchable; electrical and electroluminescent devices where changes in the electrical resistance of a spin-crossover material can be used to quench light emission; and temperature-sensitive magnetic resonance imaging (MRI) contrast agents making use of the switchable paramagnetism of a spin-crossover material. This phenomenon is particularly prevalent in iron chemistry, but also has been observed in some Co2+ , Mn3+ , Cr2+ , and Ni2+ compounds with dn (n = 4–7) electron configuration. These spin-crossover molecules have various structures ranging from mononuclear compound with different coordination numbers to 3D coordination frameworks. It occurs accompanying with two principle possibilities: the spin crossover with or without a change of the coordination number at the metal center. The latter is typical for octahedral complexes, particularly octahedral Fe2+ (n = 6) complexes. In these complexes, the occupation of the antibonding eg orbitals in the high-spin state that were empty in the low-spin state (Figure 13.2) results in an elongation of M—L bond lengths and a decrease of the splitting of the d-orbitals. In the last decades, the spin crossover with a change of the coordination number has received increasing attention because it opens up the possibility of magnetic bistability of molecules in homogeneous solution at room temperature [18]. For example, an equilibrium between a low-spin monomeric species [Ni(β-ketoenolate)2 ] with a square planar coordination sphere and a high-spin trimeric species [Ni(β-ketoenolate)2 ]3 with an octahedral coordination sphere exists in nonpolar solvents. In polar solvents, a different equilibrium between the low-spin monomeric species and a high-spin octahedral complex with two solvent molecules in the axial positions was observed [19]. The splitting of the d-orbitals changed with the change of the coordination number leading to their different occupation (Figure 13.3). Regarding spin-crossover solid materials, the challenge is the rational design of a new spin-crossover molecule and controlling the form of the spin transition. That is: whether it occurs or not; whether it occurs gradually or abruptly, in a single step or in stepwise mode, with or without hysteresis, and whether or not it occurs to full

eg T,P hv

eg T2g

T2g Figure 13.2 Scheme illustration of electron rearrangement of spin crossover without coordination number change, illustrated for a d6 metal center, e.g. Fe2+ . The energy differences between the eg and t2g orbitals change due to the spin transition.

13.2 Magneto-Structural Correlations in Spin-Crossover Compounds

L

(a) Ni

T

+2L L

(b) eg

t2g

eg

t2g

Figure 13.3 Example of a complex with the change of spin state due to coordination number change. (a) Structural change. (b) Spin-state change. Due to the change in coordination number and coordination geometry, the splitting of the d-orbitals of the metal center changes, resulting in different magnetic and optical properties.

completeness (Figure 13.4). To achieve this goal, it is important to understand the structural factors that govern spin transition. This has been comprehensively introduced by Halcrow in his reviews [14, 16, 20]. Herein, we summarize the structural factors that influence the occurrence and cooperativity of spin crossover, with an aim to complete the magneto-structural correlations about spin crossover reviewed by Halcrow.

13.2.1 Influence of Structure on the Occurrence of Spin Crossover Spin crossover is a cooperative magnetic behavior, so both the intrinsic molecular structure and the intermolecular interactions are important for its occurrence. 13.2.1.1 Molecular Structure

Generally, spin transition leads to large structure changes, which can be correlated with M—L bonds (Table 13.1) [21]. Spin crossover in O-, S-, and Cl-donor complexes usually generates much smaller structure changes than that in N- and P-donor complexes whose M—L bonds contract by 10–13% from high-spin to low-spin state. On the other hand, spin crossover in octahedral d6 (S = 2 → 0) and d5 (5/2 → 3/2) metal ions generally involves greater structural changes than for d4 (S = 2 → 1) or d7 (3/2 → 1/2) because it involves a greater depopulation of the antibonding eg d-orbital manifold. These structural changes involving spin crossover are mainly caused by changes in the spin center coordination geometry and/or ligand conformational rearrangement. 1. Metal Coordination Geometry Inducing Spin Crossover Correlating the presence or absence of spin crossover in a molecule with its structure has been reported in several studies. Appropriate structural flexibility caused

781

13 Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets

(b)

(c)

(d)

(e)

(f)

×T (cm3/mol K)

×T (cm3/mol K)

(a)

×T (cm3/mol K)

782

T

T

Figure 13.4 Six representative spin-crossover transitions demonstrating the different types of cooperativity that can be observed. (a–c) Increasing degrees of abruptness; (d) thermal hysteresis; (e) discontinuous transition; and (f) incomplete transition.

by coordination geometry and/or ligand conformation rearrangement appears to be beneficial for spin crossover. To describe the flexibility of coordination geom∑ etry of Fe2+ ions, the distortion parameters and Θ (Eqs. (13.1) and (13.2)) were introduced by Hendrickson and coworkers to measure small differences in the coordination geometries of high-spin Fe2+ complexes [22]. ∑

=

12 ∑

∣ 90 − 𝛼i ∣

(13.1)

i=1 24

Θ=

∑

∣ 60 − 𝜃j ∣

(13.2)

j=1

where 𝛼 i are the 12 cis-N–Fe–N angles about the Fe2+ ion and 𝜃 j are the 24 N–Fe–N angles measured on the projection of two triangular faces of the octahedron along ∑ their common pseudo-threefold axis (Figure 13.5). is a general measure of the deviation of a metal ion from an ideal octahedral geometry, whereas Θ specifically indicates its distortion from an octahedral toward a trigonal prismatic structure.

13.2 Magneto-Structural Correlations in Spin-Crossover Compounds

Table 13.1 Typical metal–ligand bond lengths (Å) in different spin states of octahedral spin-crossover complexes, with different combinations of metal ions and ligand donor atoms.

Fe2+ Fe3+ Co2+ Mn3+

High spin

N

O

P

S

Cl

2.2–2.3

2.0–2.1

2.6

—

2.4

Low spin

1.9–2.0

2.0

2.3

2.3

2.3

High spin

2.1–2.2

1.9–2.0

—

2.4

—

Low spin

1.9–2.0

1.9

—

2.2–2.3

—

High spin

2.0–2.2

—

—

2.4–2.6

—

Low spin

1.9–2.1

—

—

—

—

High spin

2.0–2.3

1.9–2.1

—

—

—

Low spin

2.0–2.2

1.9–2.0

—

—

—

Figure 13.5 Definitions of the angles used to calculate the ∑ distortion indices and Θ for a six-coordinate complex. (a) Scheme definition of 𝛼 i . (b) Scheme definition of 𝜃 j .

θj (a)

𝛼i

(b)

Spin crossover induced by change of metal coordination geometry has been mainly observed in Fe2+ complexes of 2,6-di(pyrazolyl)pyridines and their related ligands. To correlate spin crossover with molecular structures, crystallographic data for high-spin and low-spin state must be available. The Fe2+ /1-bpp series (1-bpp = 2,6-bis(pyrazol-1-yl)pyridine) sits in this situation. The [Fe(1-bpp)2 ]2+ salts exhibit variable anion-dependent behavior. [Fe(1-bpp)2 ](BF4 )2 undergoes an abrupt spin transition at 260 K with a narrow hysteresis loop; its [Ni(mnt)]− salt exhibits a similar transition at 180 K, but with a wider and unusually structured hysteresis loop, and the [Co(C2 B9 H11 )2 ]− salt exhibits a much more gradual spin transition centered above room temperature and only proceeds to 50% completeness [23–25]. In contrast, some solids composed of this cation always remain in high-spin state. Furthermore, thanks to synthetic routes, a large number of Fe2+ complexes exhibiting various behaviors of substituted 1-bpp derivatives are available. Therefore, magneto-structural correlation concerning their spin-state behavior has been established. In solution, most [Fe(1-bpp)2 ]2+ complexes exhibit spin crossover with T 1/2 lying between c. 230 and 270 K, but it is particularly common for crystalline [Fe(1-bpp)2 ]2+ complexes to remain high spin at all temperatures. An unusual twisted six-coordinate coordination geometry caused by a Jahn–Teller distortion of the 5 T high-spin manifold is usually visited by high-spin complexes. To

783

784

13 Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets

N N N N N

N

Fe

N N

𝜃

Figure 13.6 The two angles of measuring Jahn–Teller distortion in high-spin [Fe(1-bpp)2 ]2+ complexes (𝜃 ≤ 90∘ , 𝜙 ≤ 180∘ ).

𝝋 N

N

measure the deviation of a [Fe(1-bpp)2 ]2+ complex from its ideal D2d symmetry, two angles (Figure 13.6): a rotation of one ligand with respect to the other about the Fe2+ ion and a twist angle of the plane of one ligand about its Fe—N bond were introduced. In [Fe(1-bpp)2 ]2+ complexes, they are observed as ∑ 180∘ ≥ 𝜙 ≥ 154.5∘ and 90∘ ≥ 𝜃 ≥ 59.8∘ . The distortion parameters and Θ related to the local coordination geometry about the Fe2+ vary consistently with 𝜙 and 𝜃. Careful comparison of all these parameters for different compounds indicates that only [Fe(1-bpp)2 ]2+ complexes with weak Jahn–Teller distortion undergo spin transition. The structural change that would be required is too great to be accommodated for high-spin structures with strong Jahn–Teller distortion [20]. 2. Ligand Conformation Flexibility Inducing Spin Crossover Ligand flexibility in a coordination molecule would also provide the structural flexibility required for accommodation of the structural change occurred during spin transition. Therefore, spin crossover may be benefited from flexible ligands. These have been observed in several series of complexes including Fe2+ complexes of scorpionate ligands and iron complexes of polydentate Schiff base ligands. Several Fe2+ complexes of tris-pyrazolyborate and tris-pyrazolymethane derivatives, known as “scorpionate” ligands, have been synthesized, and their spin-crossover behaviors have been widely studied. At room temperature, both [Fe(pz3 BH)2 ] and [Fe(pz3 CH)2 ]2+ are low spin (pz3 BH = tris-pyrazolyborate; pz3 CH = tris-pyrazolymethane), but the former undergoes an abrupt transition to high spin at about 400 K [26]. Bearing substituents at the pyrazole rings and/or at the B or bridgehead C atom of the ligand gives rise to derivatives with several of which exhibit spin crossover below room temperature. Particularly, different spin-transition behaviors were observed in [Fe(pz* 3 CH)2 ]I2 derivatives (Hpz* = 3,5-dimethylpyrazole). The solvent-free form [Fe(pz* 3 CH)2 ]I2 undergoes an abrupt spin transition at 203 K with a 15 K hysteresis loop, but [Fe(pz* 3 CH)2 ]I2 ⋅CH2 Cl2 remains at high-spin state at all temperatures and the BF4 − salt [Fe(pz* 3 CH)2 ][BF4 ]2 undergoes an abrupt spin transition at 203 K proceeding to 50% completeness. In order to explain the spin-transition diversities, the average of the torsion angles Fe–N2–N1–C5 was introduced to measure the tilting of the pyrazoyl rings which is related to the spin state of these complexes (Figure 13.7). The average Fe–N2–N1–C5 torsion angle in [Fe(pz* 3 CH)2 ]I2 and high-spin [Fe(pz* 3 CH)2 ]I2 ⋅CH2 Cl2 is 171.2(5)∘ and 162.1(3)∘ , respectively. That

13.2 Magneto-Structural Correlations in Spin-Crossover Compounds

C5 N

N

N2 X

N N

N N

N

N N

Fe

N

N

N

N1

N

N X

X

N N

N N

N

Fe

N

X

N

N

N

N (a)

(b)

Figure 13.7 The two different torsions that have been used to express the degree of pyrazolyl ring twist in [Fe(pz3 BH)2 ] and [Fe(pz3 CH)2 ]2+ and their derivatives. (a) The torsion angle of Fe–N2–N1–C5. (b) The torsion angle of Fe–N2–N1–X. (a)

(b)

𝛼 = 76.6(3)°

𝛼 = 121.80(9)°

Figure 13.8 Crystal structures of two salts of [Fe(saltrien)]+ , (a) [Fe(saltrien)]ClO4 high-spin state, (b) [Fe(saltrien)]BPh4 ⋅0.5C2 H4 Cl2 high-spin state, showing the range of ligand conformations adopted by the compound in its high-spin state.

in high-spin [Fe(pz* 3 CH)2 ][BF4 ]2 is 168.0(3)∘ , while in the low-temperature phase, it is 161.7(3)∘ for the high-spin iron site and 179.0(3)∘ for the low-spin site. Alternatively, the average Fe–N2–N1–X (X = bridgehead C or B) angle was also introduced to measure the pyrazolyl twist. In perfect C3v symmetry, the torsion angle Fe–N2–N1–C5 and Fe–N2–N1–X is equal to 180∘ and 0∘ , respectively (Figure 13.8). It has been confirmed that high-spin compounds whose Fe–N2–N1–C5 and Fe–N2–N1–X torsions differ by greater than c. 11∘ from their ideal values generally do not undergo spin crossover upon cooling [27]. However, different from [Fe(1-bpp)2 ]2+ complexes, the pyrazolyl twist in these compounds reflects a librational vibration of the coordinated ligand about the Fe⋅vector rather than a distortion of the metal coordination sphere. Schiff bases are mostly derived from condensation reactions of salicylaldehydes or acetonylacetone with aliphatic or heterocyclic primary amines. Iron complexes of polydentate Schiff bases often exhibit spin crossover. Iron saltrien complexes and iron salen complexes are the two types of complexes whose spin crossover is strongly related to Schiff base ligand conformation. The [Fe(saltrien)]+ complexes exhibit anion and/or solvents dependent spin-crossover behaviors. The solid Cl− , Br− , I− , and NO3 − salt hydrates are all low spin; the unsolvated

785

786

13 Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets

BF4 − and BF6 − salts and one polymorph of the [Ni(dmit)2 ]− salt are high spin; the other [Ni(dmit)2 ]− polymorph exhibits an abrupt spin transition at 243 K, with a 30 K hysteresis loop; and the ClO4 − , BPh4 − , and [MnCr(C2 O4 )3 ]− salts exhibit more gradual spin crossover below room temperature to varying degrees of completeness [28, 29]. Similar variability is also shown by derivatives of [Fe(saltrien)]+ -bearing substituents. The variable spin-crossover behavior is not apparently related to the coordi∑ nation geometry of the high-spin state of these compounds as expressed by and Θ; however, the conformation flexibility of the saltrien ligand appears to be more relevant. The dihedral angle 𝛼 has been introduced to measure the relative disposition of the phenoxy “arms” in the molecule [30]. Low-spin complexes have 61.8∘ ≤ 𝛼 ≤ 73.6∘ , but high-spin ones show a wider range of conformations with 76.6∘ ≤ 𝛼 ≤ 125.0∘ (Figure 13.8). High-spin [Fe(saltrien)]+ complexes with 𝛼 > 90∘ mostly do not undergo spin crossover, although there are some exceptions. Presumably, this may be because a solid lattice is too rigid to accommodate the conformation rearrangement associated with the contraction in 𝛼 that would be required. Uncharged Fe2+ complexes with this ligand type can be isolated by nitration of the saltrien phenoxy groups. The observed values of 𝛼 in both spin states of these Fe2+ complexes span a narrower range, 84.9∘ ≤ 𝛼 ≤ 115.1∘ , but no apparent correlation between 𝛼 and the existence of spin crossover can be obtained. A comparable magneto-structural correlation has also been established for complexes of type [Fe(salen)L2 ]+ , where L is a monodentate N-donor base. [Fe(salen)(Im)2 ]+ complexes (Im = imidazole) which have different counterions exhibit salen2− ligand-dependent spin-crossover behaviors. When the salen2− ligands are in an envelope conformation, the [Fe(salen)(Im)2 ]+ complexes exhibit thermal spin crossover and the envelope conformation is preserved in the low-spin state. When the salen2− ligands are in either meso conformation or umbrella conformation, the [Fe(salen)(Im)2 ]+ complexes are high spin at all temperature. Therefore, it was proposed that only high-spin compounds with envelope salen2− conformation undergo spin crossover, since it is the stereochemistry preferred by the low-spin state of this class of material. High-spin compounds may not undergo spin crossover if its molecular structure deviates too strongly from what would be expected for its low-spin state [31]. This ligand conformation-dependent spin-crossover phenomenon was also observed for [Fe(pap)2 ]+ -, [Fe(qsal)2 ]+ -, and Jäger-type complexes (Figure 13.9) (pap = (E)-2-((pyridin-2-ylmethylene)amino)phenol and qsal = (E)-2-((quinolin-7-ylmethylene)amino)phenol) [32, 33].

13.2.1.2 Crystal Packing and Intermolecular Steric Contacts

Beside molecular shape, the intermolecular interactions can also greatly influence the occurrence of spin crossover. In general, crystal packing and intermolecular steric contacts such as hydrogen bonding and π–π interactions govern the intermolecular interactions.

13.2 Magneto-Structural Correlations in Spin-Crossover Compounds

(a)

(b)

Envelop conformation

Umbrella conformation

Figure 13.9 Crystal structures of two complexes of Jäger-type chelates, showing the different conformations adopted by the tetradentate Jäger-type chelates. (a) The envelop conformation. (b) The umbrella conformation.

1. Crystal packing: Where the structural differences between the two states of a spin-crossover material are ignorable, spin crossover can be controlled by crystal packing. A particular dense packing with many short intermolecular contacts can inhibit the structural changes associated with a spin-state change in either direction. There are several compounds in two or more polymorphs, where one crystal form undergoes spin crossover, while the others are high spin as summarized by Halcrow (Table 13.2) [20]. Although the sample is small, it is apparent that the density of the high-spin polymorph is higher than that for the spin-crossover polymorph(s) in all the examples. It can be rationalized that a more compact environment cannot readily accommodate the structural rearrangement associated with a thermal spin transition [34]. Obviously, this apparent correlation Table 13.2 Calculated densities of the high-spin forms of compounds crystallizing in spin-crossover and high-spin polymorphs.

[Fe(NCS)2 (btz)2 ]

𝝆calc , spin-crossover polymorph (g/cm3 )

𝝆calc , high-spin polymorph (g/cm3 )

1.525

1.646 1.499a)

[Fe(NCS)2 (abpt)2 ]

1.496, 1.498,

[Fe(NCSe)2 (abpt)2 ]

1.686

1.698

[Fe(NCNCN)2 (abpt)2 ]

1.480

1.553

[Fe(IC6 H4 BpzMe 3 )2 ]

1.662 (200 K)

1.687 (150 K)

1.519

The data were measured at room temperature unless otherwise stated (btz = 2,2′ -bithiazine; abpt = 4-amino-3,5-bis(pyrid-2-yl)1,2,4-triazole; and HpzMe = 3-methylpyrazole). a) Three spin-crossover polymorphs and one high-spin polymorph of this compound has been identified.

787

788

13 Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets

between crystal packing and the occurrence of spin crossover is only valid where the molecular structure in the two crystal forms is essentially the same. 2. Intermolecular steric contacts: For a few systems, the inhibition of spin crossover in a material needs to be attributed to specific intermolecular steric contacts. Examples include [Fe(Pz3 tren)]2+ [35], [Fe(NCS)2 (dppa)] [36], and [Fe(pic)3 ]2+ [37] type compounds (Pz3 tren = (E)-N-((1H-pyrazol-3-yl)methylene)-N,Nbis(((E)-((1H-pyrazol-3-yl)methylene)amino)methyl)methanediamine; dppa = 3-aminopropyl-bis(2-pyridylmethyl)amine; and pic = 2-(aminomethyl)pyridine). The nitrate and triflate salts of [Fe(Pz3 tren)]2+ exhibit spin crossover centered near 140 K, while the hydrated BF4 − and ClO4 − salts are high spin. During spin crossover, the alkyl “legs” of the ligand associate with a 0.3–0.5 Å movement of the bridgehead N atom away from the iron atom. In the BF4 − and ClO4 − crystals, this part is in a crowded environment with short C–H· · ·X (X = F or O) contacts, but it occupies a more open site in the spin-crossover nitrate salt. Thus, spin crossover in the BF4 − and ClO4 − crystals is possibly inhibited by crowded intermolecular steric contacts. For [Fe(NCS)2 (dppa)], the intermolecular steric contacts causing the twisted pyridyl groups in the high-spin polymorph were proposed to inhibit its spin transition. For [Fe(pic)3 ]2+ type compounds, the trans-N(pyridyl)–Fe–N(pyridyl) angle is c. 7∘ wider in the low-spin state and this molecular rearrangement on steric grounds was inhibited by bulky alcohol solvents thus preventing spin crossover in n-propyl and tert-butyl alcohol solvates.

13.2.2 Influence of Structure on the Cooperativity of Spin Crossover In addition to above influences of structure on the occurrence of spin crossover, molecular structure also plays an important role on the abruptness, hysteresis, and step of a spin transition. The cooperativity in the solid lattice, the efficiency with which structural changes at individual spin-crossover metal centers propagate through the bulk material has been demonstrated to be the determining factor. Spin centers may be linked by coordination bonds in polynuclear complexes or coordination polymers or by weaker hydrogen bonding, π–π interactions, and simple mechanical contacts. But it is not necessary so that the strongest bonding between spin sites will simply give rise to the most cooperative transition. Cooperativity associating with changes in the solid energy between the spin states is a sum of all the different interactions and steric contacts in the crystal lattice. It has been clarified that the interaction between spin states causes various types of cooperative phenomena between the low-spin and high-spin phases [38]. To consider the cooperativity in the spin-crossover transition, Wajnflasz and Pick (WP) proposed an Ising model [39], in which the spin state is described by a fictitious spin (𝜎 = –1(1) for the low-spin [high-spin] state) and the short-range interactions J between)the ∑ ∑ ( spin states are introduced in the form H = −J 𝜎i 𝜎j + i Δ − 12 kB Tlng 𝜎i , where Δ and g denote the energy difference and the degeneracy ratio between the high-spin and low-spin states, respectively. The change between a gradual crossover and a first-order transition has been well explained as a function of the parameters,

13.2 Magneto-Structural Correlations in Spin-Crossover Compounds

J, Δ, and g with this model. With increase of the strength of the intermolecular interactions, the temperature dependence of the high-spin component changes from a gradual crossover to a first-order transition [40]. In general, molecular structure and intermolecular stacking are the most important factors for cooperativity in a molecular material; thus herein, we collect the structural data from various complexes, with an aim to correlate cooperative spin crossover with structural changes. 13.2.2.1 Influence of Molecular Structure on the Cooperative Behavior

The conformational changes at ligands and crystallographic disorder in a crystalline solid have been shown to be important for cooperative spin transition. In [Fe(1-bpp)2 ]2+ series, two complexes exhibit unusual and cooperative spin crossover. [Fe(1-bpp)2 ][Ni(mnt)2 ]2 ⋅CH3 NO2 (mntH2 = maleonitrile-1,2-dithiol) undergoes an abrupt transition in cooling mode but exhibits four discrete steps in warming mode. The other example [Fe((3′ -Me)2 -1-bpp)2 [BF4 ]2 ⋅xH2 O ((3′ -Me)2 -1-bpp = 2,6-bis(3-methylpyrazol-1-yl)pyridine; x = 0 or 1/3) undergoes x-dependent spin crossover in two steps, both with hysteresis [41]. The two materials exhibit Δ𝜃 = 4.5∘ and 5.5∘ between their high- and low-spin structures, respectively, which are larger than those of more typical spin-crossover Fe(1-bpp)2 ]2+ complexes (Δ𝜃 ≤ 0.6∘ ) (Figure 13.10). The high value of Δ𝜃 reflects significant twisting in the ligands, so the highly cooperative spin transition appears to be driven by changes at the ligands. The cooperativity in [Fe(saltrien)][Ni(dmit)2 ] also appears to be induced by large ligand conformational change between its high- and low-spin states (Δ𝛼 = 34∘ ). [Fe(CN)2 (L)]⋅H2 O (L = 2,13-dimethyl-6,9-dioxa-3,12,18-triazabicyclo[12.3.1]octadeca-1(18),2,12,14,16pentaene) undergoes a complete spin transition with T 1/2 ↓ = 155 K on cooling, but the spin crossover occurs in abrupt two steps at T 1/2 ↑ = 170 and 212 K in warming mode [42]. The spin transition in this complex involves a change in coordination number induced by ligand conformational change. Cooperative spin transition may also be induced by changes in crystallographic disorder of ligand, anion, or solvent. [Fe(HPy-dapp)][BF4 ]2 exhibits a discontinuous spin transition mediated by changes in disorder of its two 4-azapentyl chains. Changes in conformational disorder in the alkyl chains induce abrupt and hysteretic low → high spin transitions in some [Co(terpy)]2+ complexes. Changes in solvent disorder result in multiple phase changes

Figure 13.10 Overlays of the high-spin (blue) and low-spin (pink) structures of the complex in [Fe(1-bpp)2 ][BF4 ]2 and [Fe({3′ -Me}2 -1-bpp)2 [BF4 ]2 emphasizing the changes in ligand conformation during spin crossover. (a) Structural overlay of [Fe(1-bpp)2 ]2+ . (b) Structural overlay of [Fe({3′ -Me}2 -1-bpp)2 ] 2+ .

(a)

(b)

789

790

13 Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets

during the progress of some [Fe(pic)3 ]Cl2 ⋅ROH complexes, and the strong cooperativity of a 40 K hysteric abrupt spin crossover in [Fe(L)2 ][ClO4 ]2 (L = 2,6-bis[5-{2-hydroxyphenyl}pyrazol-3-yl]pyridine) was attributed to an order/ disorder transition [43]. 13.2.2.2 Influence of Crystal Packing on the Cooperative Behavior

Crystal packing governed by intermolecular contacts and interactions plays an important role on the cooperative spin transition. An unambiguous relationship between intermolecular contact and cooperativity was noted by Marchivie et al., in [Fe(NCS)2 (PM-R)2 ] (R = aryl) series [44]. If “R” is not too large, these complexes readily undergo spin crossover with widely varying degrees of cooperativity. A good linear relationship was found between the width of the spin transition and the length of the crystallographic intermolecular contact in the high-spin state of the molecules. A crystal contact index (CCI) has been introduced by Weber and coworkers, to quantify the number of short intermolecular contacts in Fe2+ Jäger Schiff base complexes, and it can be concluded that a higher CCI often correlates with more cooperative spin crossover [45]. Hydrogen bonding and π–π stacking as representative intermolecular interactions play an important role on cooperativity. In the case of hydrogen bonding, the strongest evidence is provided by [Fe(μ-Rtz)3 ]X2 ⋅nH2 O complexes [46]. They commonly undergo spin crossover with a 20–40 K hysteresis loop. This strong cooperativity appears to be transmitted purely via the anions and lattice water because there are no hydrogen bonds or other steric contacts directly linking adjacent polymer chains. In the case of π–π stacking, an observation was made by Real et al., concerning the four related complexes [Fe(NCS)2 L2 ] whose cooperativity increases along the series L = btz < bipy or phen < dppz [47]. The strength of these inter-ligand interactions perfectly parallels the cooperativity of spin crossover in these complexes. Lone pair–π interactions and halogen bonding apparently enhanced the cooperative spin-crossover properties of a family of Fe2+ spin-crossover complexes of dipyridylamino-substituted triazines [48]. 13.2.2.3 Cooperative Spin-Crossover Coordination Polymers

Unlike discrete spin-crossover molecules, the spin centers in coordination polymers are linked by covalent and coordination bonds. The cooperative properties of coordination polymers are thus mainly influenced by the framework structures as well as host–guest interactions. In a series of 3D metal–organic framework polymorphs [Fe2 (H0.67 bdt)3 ]⋅xH2 O (H2 bdt = 5,5′ -(1,4-phenylene)bis(1H-tetrazole)), which are constructed from similar Fe2+ –tetrazole rod secondary building units, their diverse spin-crossover behaviors are regulated by the inter-chain cooperativity [49]. The cooperative properties of a series of porous Hoffman-like metal–organic frameworks could be tuned by host–guest interactions. A series of fascinating spin-crossover behaviors with abrupt, stepwise, and hysteretic features were obtained by exchange with a range of protic solvents in [Fe(2,5-bpp){Au(CN)2 }2 ]⋅xSolv (2,5-bpp = 2,5-bis(pyrid-4-yl)pyridine) complexes [50]. These complexes provide the representative example of host–guest hydrogen

13.3 Magneto-Structural Correlations of Low-Dimensional Magnets

bonding primarily responsible for pronounced cooperativity of spin-crossover behaviors. Introducing the naphthalene guest into the pores of the Hoffman-like metal–organic framework [Fe(dpb){Au(CN)2 }2 ] (dpb = 1,4-di(pyrid-4-yl)benzene) resulted in a thermal hysteresis loop of 73 K [51]. It has been found that guest removal in a 2D Hoffman-like material [Fe(proptrz)2 M(CN)4 ] (M = Pd or Pt, proptrz = (E)-3-phenyl-N-(4H-1,2,4-triazol-4-yl)prop-2-yn-1-imine) greatly increased its spin-crossover cooperativity [52]. Furthermore, guest programmable multistep spin crossover has been achieved in a porous material [Fe(bztrz)2 Pd(CN)4 ] (bztrz = (E)-1-phenyl-N-(1,2,4-triazol-4-yl)methanimine), in which guest-depended one-step, two-step, and three-step spin crossover has been observed in the same framework [53]. In conclusion, a high-spin molecule will not undergo spin crossover in the solid state if its shape differs too strongly from that preferred by its low-spin state. The required molecular rearrangements in such a case cannot be accommodated by a rigid solid lattice, and the material remains trapped in its high-spin state at all temperatures. A particularly high-density lattice will be less likely to accommodate the structural rearrangements required for spin crossover. Regarding cooperativity, strong, direct intermolecular contacts between spin centers can lead to strong cooperativity. These contacts might arise through hydrogen bonding, π–π interactions, host–guest interactions, or simply van der Waals forces.

13.3 Magneto-Structural Correlations of Low-Dimensional Magnets Low-dimensional magnets represent one of the most significant and promising magnetic materials. Low-dimensional molecular magnets display slow magnetization relaxation behavior below their certain temperature, which endues them with the application of high-density information storage medium, quantum computation, and molecular spintronics [54, 55]. Unlike traditional high-dimensional magnetic materials whose slow magnetization relaxation behaviors depend on long-range magnetic ordering, the magnetization of low-dimensional molecular magnet is purely molecular origin. According to their structural and magnetic features, low-dimensional molecular magnets are generally branched out into SMMs, SIMs, and SCMs. SMMs characterize superparamagnetic and exhibit magnetic hysteresis below their blocking temperature. These materials can be magnetized under an external magnetic field and can maintain the magnetic moment after the applied magnetic field being removed. This is because of the existence of reversal energy barrier between two spin ground states +M s and −M s for SMMs. The spin reversal energy barrier is generated by magnetic anisotropy which originates from zero-field splitting (ZFS) of the metal centers. When the spin multiplicities (2S+1) for ground states are beyond triplets, ZFS may occur if the molecular symmetry is lower than cubic. For these molecules, their excited states communicate with the ground state via spin–orbit coupling, leading to the ground state to further separate into

791

13 Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets

Ms = 0

Ms = 0

Virtual Energy

792

–Ms (a)

–Ms

+Ms U = S2|D| D61 K)

–∆, fast +∆

–∆, slow

LSβ D3d symmetry

HSβ* Single-molecule magnet

+∆ (120 K)

+∆ 505 nm

D3d symmetry

850 nm or +∆ (> 56 K) LSα Ci symmetry

HSα*

Figure 13.23 Conversion between five distinct states of complex [FeII (ptz)6 ](BF4 )2 . Source: From Feng et al. [82]. Copyright 2013, American Chemical Society.

̂ = −2J ∑L 𝜎i 𝜎i+1 − g𝜇B H ∑L 𝜎i , where J is ferromagnetic exchange constant H i=1 i=1 between spin units 𝜎, g, and 𝜇 B stands for gyromagnetic factor and Bohr magneton [83]. Gatteschi confirmed Glauber’s predication via observation of slow magnetic relaxation in an alternating Co(II)-radical chain [[Co(hfac)2 (NITPhOMe)] (hfac = hexafluoroacetylacetonate; NITPhOMe = 49-methoxy-phenyl-4,4,5,5tetramethylimidazoline-1-oxyl-3-oxide) in 2001 [84]. However, this Co(II)-radical chain cannot be analyzed by the Gluaber model due to non-collinear ferrimagnetic feature. After that, lots of SCMs, including Mn(III)–Ni(II) ferromagnetic chains, Fe(III)–Co(II) ferromagnetic chains, and homo-spin Co(II) ferromagnetic have been reported, which were introduced in a review paper [85].

13.4 Influence of Ligand Bridging Modes on the Mediated Magnetic Coupling In the last decades, molecule-based magnetic materials have attracted great interest for the development of new functional molecule-based materials and the fundamental research of magnetic interactions and magneto-structural correlations [86]. Using bridging ligands that can effectively mediate the magnetic coupling between the local spin carriers such as CuII , NiII , CoII , MnII–IV , FeII/III to bridge metal ions into larger architectures is a prevalent strategy for designing molecule-based magnets. Generally, the shorter and the more conjugated the bridges are, the more efficient the transmitted magnetic coupling will be. Therefore, two magnetic

13.4 Influence of Ligand Bridging Modes on the Mediated Magnetic Coupling

systems that are metal oxides with the oxygen atom as a single-atom bridge and cyanide-bridged molecular magnets including some room-temperature molecular magnets are widely explored and broadly used [87]. Although short bridges are efficient for magnetic coupling, they are somewhat lacking on the diversity of bridging modes. Longer conjugated bridges with various bridging modes have been widely used for molecular magnets. Depending on the geometry characters of bridging modes, different magnetic couplings and therefore various magnetic behaviors may be obtained. However, it is a great challenge to build the relationship between the magnetic coupling and the ligand bridging modes. Among numerous bridging ligands, N3− , monocarboxyl groups, and C2 O4 2− have been particularly widely investigated, and relatively systematic magneto-structural correlations about above ligands have been illustrated [88–90]. These ligands lie in the middle for magnetic coupling and possess diverse coordination modes, thus justify themselves as good candidates to construct molecular magnets. Herein, we will give a brief review mainly about the relationship between the magnetic coupling mediated by above ligands and their bridging modes.

13.4.1 Azido-Mediated Systems As an excellent ligand with multiple coordination sites, azido can readily bridge two or more paramagnetic metal ions, which will effectively mediate the magnetic couplings between the spin carriers and thus lead to molecular magnets showing diverse magnetic behaviors. This characteristic has attracted much attention of magneto-chemists, and azido becomes one of the most prevalent ligands to construct molecular magnets. In the last few decades, a large number of coordination compounds of azido which shows diverse bridging modes and different types of structures such as 1D chains, 2D layers, and 3D networks have been reported. Bu and coworkers have summarized the reported 1D chain modes (Figure 13.24) and 2D layer modes (Figure 13.25) based on azido bridges in their tutorial review [91]. Among these structures, eight azido bridging modes (Figure 13.26) have been observed with the end-end (EE) mode (μ2 -1,3-N3 ) and the end-on (EO) (μ2 -1,1-N3 ) mode as the most common modes. For the EE-azido system, a series of compounds of pillared (4,4) layer structure can be obtained with different co-ligands: the terminal ligand brt [Mn(N3 )2 (brt)2 ], the magnetic inactive bridging ligand bpg [Mn(N3 )2 (bpg)], and the magnetic active bridging ligand pzdo [Mn(N3 )2 (pzdo)] [92, 93]. In all these compounds, Mn2+ ions are connected into (4,4) layers by azido in EE mode which are further pillared forming 3D structures. On the other hand, a disk-like heptanuclear cluster [Co7 (bzp)6 (N3 )9 (CH3 O)3 ](ClO4 )2 (H2 O)2 represents the EO-azido system [94]. In this compound, the azido ligands adopt EO mode to bridge Co2+ ions. The structure is formed by a closest-packing arrangement of donor N/O and Co2+ ions with a local S6 symmetry. Generally, azido bridges in EE mode usually mediate antiferromagnetic coupling, while the EO-azido leads to ferromagnetic interactions. In addition, the magnetic coupling also depends on other detailed structural parameters such as the

805

806

13 Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets

N N N M

N N N M

N N N M

M

N N N

N N N

N M

N N N

M

N

M

M

N

N N N

M

M

N N N N N N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

M

M

M

N

M

N N N M

M

M

N

N

N

N

N

N

N

N

N

N N N

N N N M

N N N

N

M

M

M

M

N

N N N

N N

N N N M

M

N N N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

M

M

M N

N

N N

N

M

M

M

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

M

M

M

M

M N

N

N

N

N

N

N

N

N

N

N

N

Figure 13.24

N N N

N

N N N

N

N

N N N

M

M

N

N M

N N N M

M

N N N N N M M M M N N N N N N N N

N N N

N M

N

N N

M

N

N

M

N M

N N

N N N N N N N N M M M N N N N N N N N

N N N N N N

N N N

N

M

M

N N N

N

N N N

N N N

N

M N N N

M

NNN

N NNN N N

NNN M

NNN

NNN M

M

M N

NNN

N N

The 1D chains reported with azido bridges.

(a)

(b)

(c)

Figure 13.25 Some kinds of azido-bridged 2D layer complexes. (a) The square layer. (b) The honeycomb layer. (c) The Kagomé layer.

M–Nazido –M angles and the dihedral angle between the mean planes M–N–N–N and N–N–N–M′ (Figure 13.27). The rich bridging modes of azido combined with the resulted different magnetic couplings lead to abundant magnetic behaviors. Ferromagnetism, antiferromagnetism, ferrimagnetism, spin-canting, spin-flop, single-molecular magnets, and SCMs have been observed in the azido systems. The next we will give an overview of the azido-mediated molecular magnets according to their different behaviors. 13.4.1.1 Ferromagnets

As EO-azido bridges usually mediate ferromagnetic couplings, this feature has been well utilized for constructing ferromagnets, and thus many ferromagnetic

13.4 Influence of Ligand Bridging Modes on the Mediated Magnetic Coupling

Figure 13.26 The bridging modes reported for azido anions.

M M N N N

N N

N

N N

N

M

M

M N N

N

M

M

M M N N M

M N N

M

M M N N M M

N

M N

M

M

M M N N M

M

N

M N M M

𝛿 Mʹ M τ (a)

𝜃

τ

Mʹ M

M (b)

Mʹ

(c)

Figure 13.27 The structural parameters for EO and EE azido bridges. (a) The M–Nazido –M angle 𝜃 and the angle 𝜏 defining the out of plane displacement of EO azido. (b) The dihedral angle 𝛿defining the out of plane displacement of double EE azido bridges. (c) The M–NNN–M dihedral angle 𝜏.

compounds containing EO-azido bridges have been obtained. Many of these ferromagnetic compounds are based on Cu2+ ions, and usually a co-ligand was used to fulfill the coordination geometry. [Cu6 (N3 )12 (N-eten)2 ] and [Cu9 (N3 )18 (1,2-pn)4 ]⋅H2 O (N-eten = N-ethylethylene-diamine; 1,2-pn = 1,2-diaminopropane) were prepared by a self-assembly process [95]. [Cu6 (N3 )12 (N-eten)2 ] is a 3D framework based on EO-azido-bridged hexanuclear Cu2+ clusters (Figure 13.28a), while the 3D framework of [Cu9 (N3 )18 (1,2-pn)4 ]⋅H2 O was composed of EE-azido-bridged Cu8 clusters and [Cu(diamine)2]2+ cations (Figure 13.28b). All magnetic exchanges between adjacent Cu2+ ions in [Cu6 (N3 )12 (N-eten)2 ] are ferromagnetic resulting in ferromagnetic ordering at 3.5 K, while the magnetic exchange between ferromagnetic coupled Cu8 cluster is antiferromagnetic and no ordering is observed in [Cu9 (N3 )18 (1,2-pn)4 ]⋅H2 O. Although carboxylate (especially for syn–syn mode) usually mediates antiferromagnetic couplings, moderately strong ferromagnetic coupling can be obtained through the mixed COO− /EO-N3 − bridges. Nicotinate and its derivatives were utilized to construct EO-azido-based ferromagnets. [Cu1.5 (N3 )2 (isonicotinate)] was synthesized under hydrothermal conditions [96]. In [Cu1.5 (N3 )2 (isonicotinate)], linear syn–syn COO− /EO-N3 − -bridged Cu3 clusters were connected by μ1,1,3 -N3 − forming 2D sheets which are further built into a

807

808

13 Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets

(a)

(b)

Figure 13.28 The hexanuclear cluster in [Cu6 (N3 )12 (N-eten)2 ] (a) and the Cu8 cluster in [Cu9 (N3 )18 (1,2-pn)4 ]⋅H2 O (b).

3D structure by pyridine carboxylate ligands. Moderate strong ferromagnetic couplings with J = 40 ± 2.55 cm−1 were observed in the trinuclear cluster and long-range ordering occurs at 6 K. An 1D complex [Cu1.5 (N3 )2 (hnta)(H2 O)] and a trinuclear molecule [Cu3 (N3 )2 (hnta)4 (H2 O)3 ] (hnta = 6-hydroxylnicotinate) were reported by Bu and coworkers [97]. The chain structure of [Cu1.5 (N3 )2 (hnta)(H2 O)] is composed of double EO-azido linked Cu3 entities which are bridged by COO− /EO-N3 − bridges. [Cu3 (N3 )2 (hnta)4 (H2 O)3 ] is a trinuclear molecule based on syn–syn COO− and EO-azido (Figure 13.29). Overall strong ferromagnetic interactions were observed on both compounds with J = +99.6 cm−1 (EO-azido) and J = +44.5 cm−1 (COO− /EO-N3 − ) for [Cu1.5 (N3 )2 (hnta)(H2 O)], and J = +34.8 cm−1 for [Cu3 (N3 )2 (hnta)4 (H2 O)3 ] as demonstrated by density functional theory (DFT) calculation. However, long-range ordering was only observed on [Cu3 (N3 )2 (hnta)4 (H2 O)3 ]. Besides Cu2+ ferromagnets based on EO-azido, ferromagnetic compounds based on other transitional metal ions have also been reported, although relatively rare. An overall ferromagnetic interaction occurs in a double chain structure

(a)

(b)

Figure 13.29 (a) The 1D alternated chain of [Cu1.5 (N3 )2 (hnta)(H2 O)] and (b) the discrete trinuclear molecule of [Cu1.5 (N3 )2 (hnta)(H2 O)].

13.4 Influence of Ligand Bridging Modes on the Mediated Magnetic Coupling

[N(C2 H5 )4 ]⋅[Mn2 (N3 )2 (H2 O)] with μ1,1 -N3 − , μ1,1,1 -N3 − , and terminal N3 − [98]. 2-bzpy (2-benzoyl-pyridine) was utilized as a co-ligand, and a uniform chain structure based on EO-azido was synthesized in which the magnetic coupling J was determined to be +0.40 cm−1 [99]. Nicotinate was also used as a co-ligand to construct an EO-azido-bridged Ni2+ ferromagnet [Ni1.5 (N3 )(nic)2 (Hnic)] which is a 3D framework based on COO− /EO-N3 − -bridged linear Ni3 trinuclear clusters [100]. The intra-cluster and inter-cluster magnetic couplings were determined to be +66.2 ± 1 cm−1 and −0.60 cm−1 , respectively. Although the EE-azido usually mediates antiferromagnetic coupling, there are some exceptions in which ferromagnetic interactions are induced by EE-azido bridges. For example, asymmetric EE-azido bridges transmit ferromagnetic interactions in [Cu(14ane)Cu(N3 )4 ] (14ane = 1,4,8,11-tetraazacyclo-tetradecane) which contains single EE-azido chains formed by Cu(14ane)2+ and Cu(N3 )4 2− [101]. Ferromagnetic interactions mediated by EE-azido were also observed in a 3D diamond framework [Fe(N3 )2 (4,4′ -bpy)] (4,4′ -bpy = 4,4′ -bipyridine) which shows a critical temperature T c = 20 K [102]. 13.4.1.2 Ferrimagnets

The spontaneous magnetization of the uncompensated magnetic moment in mixed ferromagnetic and antiferromagnetic coupled spin systems would generate ferrimagnets. Due to the different preferences for ferromagnetic and antiferromagnetic couplings of EE- and EO-azido, the strategy of employing mixed bridges of EE- and EO-azido has usually been used to construct ferrimagnets. A series of 1D complexes of azido bridges with pyridine derivatives as co-ligands [Cu(3-Clpy)2 (N3 )2 ], [Mn(3-Mepy)2 (N3 )2 ], [Mn(Menic)2 (N3 )2 ], and [Mn(3,4Dmepy)2 (N3 )2 ] have been prepared (3-Clpy = 3-chloropyridine; 3-Mepy = 3-methylpyridine; Menic = 3-methylnicotinate; 3,4-Dmepy = 3,4-dimethyl-pyridine) [103]. The chains of [Mn(3-Mepy)2 (N3 )2 ] and [Mn(3,4-Dmepy)2 (N3 )2 ] are based on (–EO–EE–EE–) azido sequence, while those of [Cu(3-Clpy)2 (N3 )2 ] and [Mn(Menic)2 (N3 )2 ] have an arrangement of (–EE/EO–EE/EO–EO–EO–) and (–EE–EO–EO–EO–EO–), respectively. Both [Cu(3-Clpy)2 (N3 )2 ] and [Mn(Menic)2 (N3 )2 ] exhibit ferromagnetic-like behavior with a ground-state S = 1 per Cu4 unit and S = 3/2 per Mn2+ ion, respectively. Similarly, the ground state of S = 5/2 per Mn3 unit was determined by the magnetic study for [Mn(3-Mepy)2 (N3 )2 ] and [Mn(3,4-Dmepy)2 (N3 )2 ]. Several 2D ferrimagnets are also prepared from azido bridges and pyridine derivatives. [Co(4,4′ -bipy)(N3 )2 ] and [Ni(4,4′ -bipy)(N3 )2 ] are 2D arrays formed by 4,4′ -bipy connected azido chains of (–EE/EO–EE/EO–EO–) sequences. Ferromagnetic behaviors with S = 3/2 and S = 1 per Co3 and Ni3 unit are observed [104]. With 4-N3 py as a co-ligand, a 2D (4,4) network [Mn(4-N3 py)2 (N3 )2 ] in which double EO-azido-bridged [Mn(N3 )2 Mn]2 units were linked to four Mn2+ ions via four EE-azido has been synthesized. In [Mn(4-N3 py)2 (N3 )2 ], a long-range ferromagnetic ordering being characteristic of antiferromagnetically coupled S = 5/2 and 10/2 spin system was substantiated [105].

809

810

13 Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets

13.4.1.3 Antiferromagnets

Anti-parallel coupling often requires a lower energy than that of parallel exchange; therefore, most of the magnetic compounds show antiferromagnetic properties. EE-azido favor antiferromagnetic couplings and have been widely used as bridges for designing antiferromagnets based on various paramagnetic ions. The structures of these EE-azido-bridged antiferromagnets are greatly diverse for a lot of co-ligands which can be utilized. A series of Ni2+ compounds [Ni(L1 )2 (N3 )](ClO4 )⋅H2 O, [Ni(L2 )2 (N3 )](PF6 ), [Ni(L3 )2 (N3 )](ClO4 ), and [Ni(L3 )2 (N3 )](PF6 ) with 1D structure were prepared (L1 = 2-(aminoethyl)pyridine, L2 = 1,2-diamino-2-methylpropane, L3 = N-ethylethylenediamine). For all the four compounds, Ni2+ ions are singly bridged by EE-azido forming a chain and global antiferromagnetic interactions were determined [106]. A 3D structure [Co1.5 (N3− )(OH)(Isonic)] was prepared under hydrothermal reactions, in which μ1,1,1 -N3 − /μ3 -OH layers were pillared by tridentate isonicotinate [107]. Treating the 3D network as 2D entities from a magnetic point of view, a set of parameters: A + B = 3.53 cm−1 (Curie constant), E1 /k = 127 ± 2 K (spin–orbital coupling), and E2 /k = 4.5 K (antiferromagnetic interaction) were obtained according to the Rueff procedure. As an exception, antiferromagnetic couplings were mediated by EO-azido bridges in [Cu(H2 O)6 ][Cu2 (N3 )4/3 (OH)(pta)]6 (pta = phthalate) which features as a 2D inorganic azido–Cu2+ –hydroxyl layer reinforced by pta [108]. Herein, Cu(H2 O)6 2+ not only serves as a counterion but also be a template in the assembly of the [Cu24 ] macrocycle via hydrogen bonds (Figure 13.30). 13.4.1.4 Other Magnetic Behaviors

The magnetic couplings mediated by azido bridges collaborate with other structural factors influencing magnetism would result in various novel magnet behaviors. Spin canting can be induced by employing azido to bridge ions with high magnetic anisotropy mediating asymmetric exchange. An 1D EE-azido chain compound

(a)

(b)

Figure 13.30 (a) The kagóme topology for the inorganic sheet of [Cu(H2 O)6 ][Cu2 (N3 )4/3 (OH)(pta)]6 and (b) Cu(H2 O)6 2+ serves as countercation and template for the assembly of {Cu24 } macrocycle.

13.4 Influence of Ligand Bridging Modes on the Mediated Magnetic Coupling

[Co(L)4 (N3 )](ClO4 )⋅H2 O (L = 5-methylpyrazole) has been prepared with a capping co-ligand [109]. The M–H plot of this compound indicates spin canting with a saturation value of 2.64 N𝛽 lower than the expected value of 3 N𝛽. If the intermolecular antiferromagnetic interactions of EO-azido-bridged ferromagnets can be overcome by external field, metamagnets can be expected and the field required to switch an antiferromagnetic to a ferromagnetic interaction is defined to be critical field H c . For example, a 2D complex [Fe(4,4′ -bipy)(N3 )2 ] contains ferromagnetic EO-azido-bridged chains, and the interchain antiferromagnetic interactions can be overcome by a field H c = 25 kOe [102]. Surprisingly, multiple phase transitions were observed at low field with T N1 = 20 K and T N2 = 5 K. A series of EE-azido-bridged perovskite-type compounds were synthesized by Wang and coworkers, and these materials show cation-dependent magnetic ordering at up to 92 K and magnetic bistability near room temperature [110]. Low-dimensional magnetic properties can also be anticipated if highly anisotropic paramagnetic ions are connected into polynuclear clusters or well-isolated chains. A [Mn3+ ]2 complex bridged with EO-azido was reported in which a ground state of S = 4 was accomplished and slow relaxation of magnetization could be observed [111]. In the context of SCM, a double EO-azido-bridged chain complex [Co(bt)(N3 )2 ] has been synthesized on which a blocking temperature of 5 K, energy barrier of 94 K, and relaxation time of 𝜏 0 = 3.4 × 10−12 s were determined (Figure 13.31) [112].

13.4.2 Monocarboxylate-Mediated Systems Similar to azido, the carboxyl group as an excellent multidentate ligand could also readily bridge paramagnetic ions into various complicated structures showing interesting magnetic properties. Since only the COO− moiety involves magnetic coupling, carboxyl group is actually a three-atom bridge as N3 − . Multiple bridging modes have been observed on it (Figure 13.32), and interestingly all of the bridging modes can find their corresponding analog in azido-bridged systems. Depending on the geometries of the carboxyl bridge (syn–syn, anti–anti, or syn–anti), different magnetic couplings can be mediated.

Figure 13.31 The double EO-azido-bridged chain complex [Co(bt)(N3 )2 ] showing SCM behavior. Source: Based on Liu et al. [112].

811

812

13 Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets

O

O

M

O

O

O

M

M

M

2.21 syn–syn

2.20 M

Figure 13.32 Bridging modes for monocarboxylate.

M

M

M

O O 2.21 ant–anti

2.21 syn–anti M

M O

O

O

M 3.12

M

M O

O

4.22 M

M

Due to the differences between carboxyl and azido in coordination atom (O or N), electronic structures, and the geometric characteristics, the resulting structures and magnetic properties can be greatly different. Although numerous carboxyl-bridged magnetic materials have been reported, the magneto-structural correlations involving carbonyl group are strangely lacking. As the smallest carboxylate, formato has been widely used to construct magnetic metal–formate frameworks. Employing protonated amine cations of different sizes in non-coordinating solvent, a series of anionic metal–formate frameworks can be obtained whose structures are depending on the size and shape of the cations. NH4 + resulted in a chiral magnetic framework [NH4 ][M(HCOO)3 ] with the 49 ⋅69 topology consisting of octahedral metal centers connected by 2.11-anti–anti formato [113]. In the cases of the mid-sized monoammonium cations CH3 NH3 + , (CH3 )2 NH2 + , CH3 CH2 NH3 + , and (CH2 )3 NH2 + , the product is a series of perovskite compounds [AmineH+ ][M(HCOO)3 ] also with 2.11-anti–anti format linkages between metal sites [114]. These metal–formate frameworks usually show 3D long-range antiferromagnetic ordering with weak ferromagnetism arising from the antisymmetric exchange via the formato linkages. However, bulky cations such as (CH3 CH2 )3 NH+ (CH3 CH2 )2 NH2 + , and CH3 CH2 CH2 NH3 + led to the porous [M3 (HCOO)6 ] family possessing dia frameworks consisting of the apx-sharing M-centered MM4 tetrahedron nodes [115]. This family of materials shows metal ion-dependent magnetic behaviors. Mn and Fe members are ferrimagnets with T c = 8.0 and 16.1 K, respectively, the Co analog is a spin-canting antiferromagnet, and the Ni compound shows 3D long-range ferromagnetic ordering at 2.7 K. In addition, the distance between paramagnetic metal ions has also been demonstrated to be important for the magnetism of monocarboxyl-mediated systems. In compounds with Gd3+ pairs solely bridged by two μO:κ 2 OO carboxylates, a ferromagnetic coupling is observed in most cases and the magnitude of the interaction correlates with the Gd· · ·Gd distance. At a critical value of 4.12 Å, the sign of the coupling changes [116].

13.4.3 Oxalate-Mediated Systems The dicarboxylate ligands have been extensively studied in the field of molecular magnetism, and various magnetic behaviors have been observed on the molecular magnets involving such ligands. Among the different dicarboxylates, major work has been done with oxalate, coordinating in a bisbidentate chelating fashion. For the great diversity of the coordination modes of ox2− , a large number of coordination

13.4 Influence of Ligand Bridging Modes on the Mediated Magnetic Coupling

b a

Figure 13.33 View of the framework featuring fused Star-of-David catenanes [CoII (ox)(bphy)⋅0.2(DMF)]n . Source: Based on Huang et al. [117].

complexes based on ox2− have been reported. Recently, a metal–organic framework featuring fused Star-of-David catenanes [CoII (ox)(bphy)⋅0.2(DMF)]n (ox = oxalate; bphy = 1,2-bis(4-pyridyl)hydrazine) has been obtained employing oxalate and azopyridine (Figure 13.33) [117]. In this material, Co2+ ions are bridged by ox2− into helical chains which are further connected into an interpenetrated 3D framework by bphy. Dominant antiferromagnetic coupling between Co2+ ions was mediated by ox2− , resulting in coexistence of slow magnetic relaxation and long-range ordering. Due to the large number of magnets based on ox2− , a few magneto-structural correlations involving ox2− -mediated magnetic couplings have been established. For example, μ-oxalate-bridged dinuclear complexes of Cu2+ containing linear polyamine terminal ligands have been thoroughly studied and it has been found that the magnetic interaction between two Cu2+ ions is antiferromagnetic in almost all the cases. Antiferromagnetic interactions are also prevalent in homometallic extended structures constructed by ox2− . A ladder-like 1D complex formed by a squarate ligand linked Cu(ox) chains and an alternating chain complex of Cu2+ with oxalate and pyrazine repeated alternately were reported, and weak antiferromagnetic interaction and strong antiferromagnetic interaction were observed on the former and latter, respectively [118]. Antiferromagnetic interactions were also observed in a series of heterometallic honeycomb structures [NBu4 ][MII Mn(ox)3 ] (MII = Fe, Co, and Ni) that give rise to ferromagnetic or weak ferromagnetic ordering [119]. For the different electronic structures of different magnetic centers, the couplings mediated by ox2− may be influenced by metal centers. A series of compounds

813

814

13 Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets

containing ox2− -bridged cations [{M(phen)2 }2 (μ-ox)][Cr(phen)(ox)2 ]2 ⋅4H2 O (M = Mn2+ , Co2+ , Ni2+ , Cu2+ , phen = bis(phenanthroline)) were synthesized [120]. In these compounds, M2+ in homodinuclear cations is coordinated with two phen molecules and one bridging ox2− ; Cr3+ in anions is coordinated with one phen ligand and two bidentate ox2− . Oxalate ligands mediate ferromagnetic coupling in the Cu2+ compound, whereas antiferromagnetic interactions are observed in other compounds. Two double ox2− -bridged zigzag Sr–Cr paramagnetic chains were synthesized; the observed paramagnetism indicated the magnetic coupling between Cr3+ ions mediated by Sr(ox)2 moieties were broken by diamagnetic Sr2+ ions [121]. In conclusion, the correlations involving the magnetic couplings mediated by N3 − , ox2− , and monocarboxyl groups have been briefly reviewed. In general, EE-azido bridges usually mediate antiferromagnetic coupling, while the EO-azido leads to ferromagnetic interactions. The factors influencing the magnetic couplings mediated by monocarboxyl groups at least include the coordination mode of carboxylate, metal ions, as well as the distance between them. Antiferromagnetic interaction is prevalent in ox2− -mediated systems, but the magnetic exchange is also depending on the metal ions and may be broken by diamagnetic ions such as Sr2+ .

13.5 Geometrically Frustrated Molecular Magnetic Materials 13.5.1 Geometric Magnetic Frustration The concept of magnetic frustration, a term applied to the situation wherein a large fraction of magnetic sites in a lattice is subject to competing or contradictory constraints, was first proposed by Wannier in 1950 [122] and Anderson in 1956 [123], who realized that the Ising model on an antiferromagnetic polychlore lattice maps onto Pauling’s model of proton disorder in ice. When frustration arises purely from the geometry of topology of the lattice, it is called geometric frustration. Spin frustration occurs when the geometric constraints of a crystal lattice inhibit the antiparallel alignment of electron spins on different sites into magnetically ordered arrays. As the Hamiltonian for the interaction between any two spins can be written as H ex = − 2JS1 S2 , the energy is minimized for collinear (parallel or antiparallel) spin alignments. Under the conditions that J < 0 which favors the antiferromagnetic interaction and that J is equal for all n.n. (nearest neighbor) pairs, it is obvious that only two of the three spin constraints can be satisfied simultaneously in a triangular plaquette (Figure 13.34). Any lattice based on triangular or tetrahedral arrangements of magnetic centers sits in this situation, and the spin frustration inherent in these materials enhances the classical spin degeneracy and offers the possibility for nonordered spin liquid states which have potential applications on quantum memory and quantum computation [124]. Therefore, geometrically frustrated antiferromagnetic materials have attracted much attention over the past few decades.

13.5 Geometrically Frustrated Molecular Magnetic Materials

?

? ? Jnn ~ Jnnn

?

? (a)

(b)

(c)

Figure 13.34 Frustrated “plaquettes” : (a) the equilateral triangle, (b) the tetrahedron, (c) square plane with Jnn ∼ Jnnn .

13.5.2 Evaluation of Geometric Frustration Magnetic interactions are expressed on an energy scale set by the exchange energy, H ex ∼ − 2JS2 ∼kT, where T ≫ 0. The manifestation of spin frustration can be evaluated by the Weiss constant, 𝜃 c , which appears in the Curie–Weiss law, 𝜒 = C/(T − 𝜃 c ). ∑ zn J n , From the mean field theory [125], 𝜃 c can be expressed as that 𝜃c = 2S(S+1) 3k where n is the nth neighbor and J n is the corresponding exchange constant, i.e. 𝜃 c is the algebraic sum of all of the exchange interactions and thus sets the energy scale for the magnetic interactions. In the cases of frustration absence, the onset of deviations from Curie–Weiss law is at T∼|𝜃 c | and the establishment of a long-range ordered state is also near |𝜃 c |. For ferromagnetic order, this is nearly idealized, as |𝜃 c |/T c ∼1. For antiferromagnetic order, typical values of |𝜃 c |/T c for non-frustrated lattices are in the range of 2–4 or 5. Although it is somewhat arbitrary, |𝜃 c |/T c > 10 has been taken as a criterion for the presence of frustration [126].

13.5.3 Ground States of Frustrated Magnets For a system without spin frustration, the possibility that a long-range ordered state can be realized is determined by the dimensionality of both the lattice, d, and the spin, D. D is equal to the number of spin components which must be considered in the expanded form of Eq. (13.1), H = −2J(Six Sjx + Siy Sjy + Siz Sjz ). The case of D = 1 corresponds to Ising model, D = 2 is the X–Y model, and D = 3 is the isotropic Heisenberg model. Only systems with d = 3 will undergo long-range order for a system without frustration, except the 2d Ising model, while the presence of a frustrated lattice complicates the ground states of magnetic materials. Various ground states may exist for frustrated systems including long-range order, spin glass, spin liquid, and spin ice. Most frustrated magnets which undergo long-range order adopt non-collinear spin configurations. To rank some of the common lattice in terms of the level of geometric frustration, Lacorre [127] proposed a constraint function: ∑ ∑ Fc = − EE = − k Jk Sk1 Sk2 / k ∣ J k ||Sk1 ||Sk2 ∣, where Eb is the basis energy for a b non-frustrated pair-wise interaction and E is the energy of the frustrated spin pair. Eb can be written as Eb = |J|S(S + 1)p(p − 1) and Eb /p |J| = eb = S(S + 1)(p − 1),

815

816

13 Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets

where p is the number of spins in a cluster and eb is the normalized energy per cluster. Thus, F c = e/eb = −1/(p − 1) and F c = −1 define a non-frustrated system, while F c becomes more positive as the frustration level increases. However, there are many materials with frustrated lattices that do not undergo long-range order. From a microscopic point of view, such lattices posses an enormous degeneracy. The essentially infinite degeneracy can be perturbed by inclusion of other complicating factors which can work to select a ground state or set of states. For example, spin glass state is a configuration of spins frozen into a random pattern, in spin liquid state the spins remain dynamic down to the lowest temperatures, and spin ice is a special case of spin liquid formed by Ising spins maps exactly onto the distribution of protons around O atoms in solid. Greedan has briefly reviewed geometrically frustrated inorganic framework materials [128]. Herein, we will briefly review the selected geometrically frustrated molecular magnets. These materials can be single molecules or infinite structures, and the discussion will begin with isolated molecules and progress to 3D lattices.

13.5.4 Molecular Materials Showing Spin Frustration 13.5.4.1 Isolated Molecules

The simplest approach to achieve spin frustration is provided by attempting to place three antiferromagnetically coupled spins on a triangular lattice. Although most research to date concerning this strategy has focused on inorganic or coordination crystals where the rigid lattice allows strict control of magnetic exchange interaction and long-range order, comparatively few examples of spin frustration in crystals of organic radicals have been reported. 1. Geometrically frustrated organic molecular triangle: Generally, the three spins can be placed either on three sites in a single trigonal molecule or on three independent molecules arranged into a triangular lattice. Representative trigonal organic radicals are a series of π-conjugated systems in which the intramolecular magnetic coupling can be controlled by oxidative modulation (Figure 13.35). Trinitroxide-substituted triphenylamine (TPA) synthesized by Iwamura and coworkers was in a doublet ground state with intramolecular coupling of J/kB ≈ −135 K in its neutral state [129]. Its cyclic voltammogram shows a reversible one-electron redox wave owing to the oxidation of TPA leading to the oxidized cationic species which is expected to have a triplet ground state. Recently, an oxidative spin-state conversion from a triradical doublet ground state to diradical cation triplet ground state was also achieved on N-tert-butylnitroxide-substituted trioxytriphenylamine (NTO) by Okada and coworkers [130]. Unlike the unstable oxidized cationic species of trinitroxide-substituted TPA, the crystal structures of both the doublet and triplet NTO were determined and the magnetic coupling was switched from strong antiferromagnetic to ferromagnetic. Different from open trigonal molecules, a closed spin-frustrated molecular triangle with three naphthalene-1,4:5,8-bis-(dicarboximide)s [(+)-NDI-Δ3(−⋅) (CoCp2 + )3 ] was

13.5 Geometrically Frustrated Molecular Magnetic Materials

tBu

tBu

O

N

O–

N+

l e– oxidation l e– reduction

N O

N tBu

tBu

N O

N+

O

N tBu

N O [TPA]+

TPA

tBu

tBu

O

N

l e– oxidation O

O

N tBu

N O

O

O–

O N+

O

tBu

N+

O

l e– reduction

N O

tBu

N tBu

N O

O

tBu

[NTO]+

NTO

Figure 13.35 Switching of the π-conjugation networks of trinitroxide-substituted triphenylamine (TPA) and N-tert-butylnitroxide-substituted trioxytriphenylamine (NTO).

O

N

OO

N

O

O

N

OO

N

O

3e– ON O O N O

Figure 13.36

ON O N O

O

ON O O N O

ON O N O

O

Chemical structures of (+)-NDI-Δ and (+)-NDI-Δ3(−⋅) .

obtained by Wasielewski and coworkers through the reaction of (+)-NDI-Δ and CoCp2 [131]. This compound can be conveniently described as a triangular triradical trianion intimately associated with three (CoCp2 + ) cations by means of electrostatic forces and intermolecular interactions (Figure 13.36). Magnetic study shows ferromagnetic ordering below T C = 20 K with a value of f = 9.7. 2. Geometrically frustrated clusters: Although numerous cluster compounds have been available in literature, attentions on spin frustration were mainly paid on those with odd-sided polyhedral core of isotropic spin centers, particularly Cu3 triangular clusters. A {Cu3 Cl} triangular cluster compound ([Cl(CuCl2 tachH)3 ]Cl2 , tach = cis,trans-1,3,5-triamino-cyclohexane) that is formed by simultaneous coordinative and hydrogen-bonded interactions

817

818

13 Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets

Figure 13.37 Representations of the triangular {Cu3 Cl} cluster.

NH2 H2N +H

3N

tachH

facilitated by the utilization of the trans-tach ligand was reported by Cronin and coworkers [132] (Figure 13.37). Both the intra- and inter-triangle exchange interactions were observed to be antiferromagnetic (J intra /kB ≈ –9.2 K, J inter /kB ≈ −2.7 K), implying spin frustration in the {Cu3 Cl} triangular cluster. The mechanisms of the magnetic coupling interactions for trigonal-bipyramid {Cu3 X2 } complexes Cu3 (μ3 -X)2 (μ-pz)3 X3 and trigonal trinuclear {Cu3 X} complexes Cu3 (μ3 -X)-(μ-pz)3 X3 were investigated by the calculations based on density functional theory combined with broken-symmetry approach (DFT-BS) (X = Cl or Br and pz = pyrazolato anion). Based on the calculation on the magneto-structural correlation, the exchange interaction is sensitive to the Cu–(μ3 -X)–Cu angle. It is switched from ferromagnetic to antiferromagnetic with the Cu–(μ3 -X)–Cu angle changing from 76∘ to 120∘ , and spin frustration exists in these complexes [133]. A large number of larger clusters with triangle-based structures have been found to be spin-frustrated. The effects of both intramolecular dipolar interactions and geometric spin frustration were observed on a trigonal prismatic {Gd7 } nanomagnet (i Pr2 NH2 )6 [Gd7 (μ3 -OH)3 (CO3 )6 (O2 Ct Bu)12 ] with C3h symmetry for antiferromagnetic coupling (Figure 13.38a) [134]. Furthermore, there are also some heterometallic cluster compounds been found to be spin-frustrated. A series of disk-like (Mn2+ )4 (Mn3+ )3 compounds of azide in combination with N-methyldiethanolamine (mdaH2 ) or triethanolamine (teaH3 ) were obtained. For those with mda2− as ligand, a ground state of S = 11 was determined for the competing antiferromagnetic interactions that frustrate perfectly antiparallel spin alignments [135]. A fan-shaped 3d–4f {Gd6 Cu12 } amino acid cluster with a hollow {Gd6 } octahedron encapsulated by six {Cu}2 dimers was reported by Wu and coworkers (Figure 13.38b). The magnetic coupling for Cu–Cu and Cu–Gd was determined to be 857.3 and 9.6 cm−1 , respectively, therefore spin frustration occurred within the {GdCu2} triangle [136].

13.5 Geometrically Frustrated Molecular Magnetic Materials

(a)

(b)

Figure 13.38 Crystal structures of (i Pr2 NH2 )6 [Gd7 (μ3 -OH)3 (CO3 )6 (O2 Ct Bu)12 ] view along S3 axis (a) and the [Gd6 Cu12 ] cluster cation (b) with an axial-fan shape viewed along the c-axis.

13.5.4.2

𝚫-Chain Lattice

By sharing corners, edges, and even faces of triangles, more complicated frustrated patterns can be predicted, among which Δ-chain structures and 2D systems are particularly interesting. Δ-chain spin lattices have been the focus of intense study from both theoretical and experimental of view because the interplay of geometric frustration and quantum fluctuations, and eventually broken translation invariance, can give rise to novel magnetic phases. Magnetic isotropic S = 1/2 Cu2+ Δ-chain systems are interesting because quantum effects suppressing magnetic orders are strong, and spin frustration has been observed in these systems. Lindgrenite Cu3 (OH)2 (MoO4)2 is composed of alternating corner- and edge-sharing Δ-chains of CuO4 (OH)2 octahedra comprising two crystallographically distinct Cu2+ ions. Although lindgrenite is a ferromagnet, the incorporation of organic pillaring 4,4′ -dipyridyl disrupts ferromagnetic exchange to unveil a spin-frustrated antiferromagnet with T N = 3.1 K and f = 19.4 [137]. Δ-chain systems of cis-edge- or corner-sharing triangular spin-1/2 “CuV2 ” fragments were synthesized recently, and one of these compounds [enH2 ]Cu(H2 O)2 [V2 O2 F8 ] (enH2 = ethylenediammonium) exhibits spin frustration with f > 9 and a plateau in its M–H curve at 2 K (Figure 13.39). With the H increasing at 2 K, the material transits from a singlet ground state to a triplet excited state of V dimmers and finally saturates [138]. In combination with magnetic anisotropic metal ions, Δ-chain spin systems may produce complex magnetic phase behavior of great theoretical and experimental importance. Particularly, Co2+ Δ-chain structures have attracted much attention and a large number of coordination polymers based on edge- or corner-sharing {Co2+ }3 triangular plaquettes have been reported. Hydrothermal reaction of CoCl2 with pyridine-2,5-dicarboxylic acid and KOH in water afforded [Co3 (NC5 H3 (CO2 )2 -2,5)2 (μ-OH)-(OH2 )2 ] which is based on a chain of OH-bridged scalene triangles that share edges and vertexes [139]. Multiple areas of magnetic bistability were observed in this topological

819

820

13 Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets

Triangular fragment

V

Cu V

Figure 13.39

(a)

Representation of the “CuV2 ” triangular fragment in [enH2 ]Cu(H2 O)2 [V2 O2 F8 ].

(b)

(c)

Figure 13.40 Suggested arrangement for the spins on the cobalt ions for (a) H < 10 kOe, (b) 10 kOe < H < 15 kOe, and (c) H > 15 kOe.

ferrimagnet. In small fields below 30 K, the complex shows canted antiferromagnetism. Between 5 and 12 kOe, a spin flip phase is responsible for a second area of bistability. At H > 15 kOe, a third area of bistability observed may be due to a field-induced reorientation of the spins (Figure 13.40). Utilizing pyridine-2,4-dicarboxylic acid (2,4-pydc), our group obtained two porous coordination polymers [Co3 (2,4-pydc)2 (μ-OH)2 (H2 O)]n ⋅7nH2 O and [Co3 (2,4-pydc)2 (μ-OH)2 ]n ⋅5nH2 O also based on corner- and edge-sharing {Co2+ }3 triangles but with different symmetry [140]. Regarding their Co2+ Δ-chains, one is formed by one type of triangles and the other contains two different types of triangles (Figure 13.41). The former is a typical antiferromagnet, while spin-canting and field-induced spin flop were observed in the latter.

13.5 Geometrically Frustrated Molecular Magnetic Materials

Figure 13.41 Co2+ Δ-chains composed of corner- and edge-sharing triangles in (a) [Co3 (2,4-pydc)2 (μ-OH)2 (H2 O)]n ⋅7nH2 O and (b) [Co3 (2,4-pydc)2 (μ-OH)2 ]n ⋅5nH2 O.

(a)

(b)

13.5.4.3 2D Lattice

As summarized by Richter et al., there are 11 uniform Archimedean lattices with regular polygons in 2D systems. Among them, the famous Kagomé lattice and the so-called “star” lattice are theoretically potential for the observation of quantum spin liquid behavior. In the Kagomé lattice, the triangles are corner-sharing, but those in the star lattice are separated by a dimer. As a result, the next nearest neighboring exchange interactions are different in these two lattices. All the exchange pathways in the Kagomé lattice are equivalent, but the intra-triangular exchange pathway J T is different from the dimerized pathway J D in the star lattice. Kagomé lattice has attracted much attention for the structural beauty as well as potential quantum paramagnetism. Electro-oxidation of the quinoidal bisdithiazole (BT) in the presence of [Bu4 N][GaBr4 ] afforded radical ion salt in which the radical cations pack forming an organic Kagomé structure [141]. Magnetic studies suggest strongly frustrated antiferromagnetic interactions, but reveal no magnetic order down to 30 mK. S = 1/2 Kagomé coordination polymers have attracted much attention as its intrinsic spin frustration is possibly responsible for superconductivity. Hydrothermal treatment Cu(OH)2 with isophthalic acid (1,3-bdcH2 ) led to the formation of a metal–organic Kagomé net composed of {Cu3 (OCO)3 } triangles [142]. A ferromagnetic ordering was observed at 2 K in this material with a strong frustration (f = 16.5). Remarkably, superconductivity has been observed in a Kagomé coordination polymer [Cu3 (C6 S6 )]n (Cu-BHT) which is confirmed by the zero resistivity, ac magnetic susceptibility, and specific heat measurements [143]. The second superconductivity phase (T c ≈ 3 K) with a small volume fraction was also detected in this material. Some research groups in our institute have focused on inorganic Kagomé lattice compounds aiming at the observation of quantum spin liquid. Two isostructural compounds BiOCu2 (XO3 )(SO4 )(OH)⋅H2 O [X = Te or Se] with a distorted octa-Kagomé lattice were obtained by He and coworkers [144]. BiOCu2 (TeO3 )(SO4 )(OH)⋅H2 O exhibits a gapped ground state (Figure 13.42), while BiOCu2 (SeO3 )(SO4 )(OH)⋅H2 O possesses a typical antiferromagnetic ordering at ∼24 K. Such different magnetic behaviors proposed originate from a slightly structural modification induced by nonmagnetic XO3 anionic groups. Furthermore, the origin of gapped ground state may be due to the dimerization of Cu2+ ions as suggested by theoretical simulation. A rare earth-based material Eu9 MgS2 B20 O41 with triple-Kagomé-layer slabs made up of three Kagomé layers was reported by Deng and coworkers (Figure 13.43) [145]. The magnetic susceptibility of this material shows a smooth kink the height of which is reduced gradually with increasing probe magnetic field which is very similar to that observed for ZnCu3 (OH)6 Cl2 showing quantum spin liquid behavior.

821

822

13 Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets

6 5 4 3 2 1 0 0 (a)

10

20

30 H(T)

40

50

60

(b)

Figure 13.42 Structural illustration (a) and M–H plot (b) of the octa-Kagomé lattice BiOCu2 (SeO3 )(SO4 )(OH)⋅H2 O. Source: Based on Tang et al. [144]. Copyright 2017, American Chemical Society. Eu S Eu

(a)

Eu

(b)

Figure 13.43 Structure (a) and temperature-dependent magnetic susceptibilities (b) of the triple-Kagomé-layer slabs. Source: Based on Chi et al. [145]. Copyright 2019, American Chemical Society.

The “star” compounds are relatively rare in literature. Using acetic acid to link triangular clusters, a “star” compound [Fe(μ3 -O)(μ-OAc)6 (H2 O)3 ]Fe3 (μ3 -O) (μ-OAc)7.5 ]2 ⋅7H2 O was synthesized [146]. Magnetic studies showed coexistence of strong spin-frustration effect with f = 129 and long-range magnetic ordering with T c ≈ 4.5 K in this complex. 13.5.4.4 3D Lattice

By sharing corners of tetrahedral plaquettes, the pyrochlore A2 B2 O7 and spinel AB2 O4 lattice may be realized which have played a critical role in the development of current ideas about spin frustration. These lattices were realized in a number of inorganic oxides such as Gd2 Ti2 O7 , Ho2 Ti2 O7 , and LiMn2 O4 in some of which spin ice behavior has been observed [147]. However, metal–organic frameworks with these lattices are still elusive. In conclusion, the current state of knowledge in the area of geometrically frustrated molecular magnets has been described to provide at least a hint of the relationship between spin frustration and molecular structure. Triangle molecules of antiferromagnetically coupled spins and structures formed by corner- and

References

edge-sharing triangle or tetrahedral plaquettes probably show strong spin frustration effect and thus lead to exotic phenomena such as spin liquid, spin ice, and spin glasses behavior.

Acknowledgments This work was supported by the Recruitment Program of Global Youth Experts, the Xiamen Recruitment Program of Experts, NSF of China (21871262), and NSF of Fujian Province (2019J01130).

References 1 Sorai, M., Nakano, M., and Miyazaki, Y. (2006). Calorimetric investigation of phase transitions occurring in molecule-based magnets. Chem. Rev. 106: 976–1031. 2 Yin, L., Ouyang, Z.W., Wang, J.F. et al. (2019). Anisotropic magnetization plateaus in Seff = 1/2 skew-chain single-crystal Co2 V2 O7 . Phys. Rev. B 99: 134434. 3 Liu, P.P., Cheng, A.L., Liu, N. et al. (2007). Toward tuning the bulk magnetic behaviors of metal-azido materials by organic pillars. Chem. Mater. 19: 2724–2726. 4 Ihoh, K. and Kinoshita, M. (eds.) (2000). Molecular Magnetism: New Magnetic Materials. Tokyo: Kodansha, Gordon and Breach Science Publishers. 5 Turnbull, M.M., Sugimoto, T., and Tompson, L.K. (eds.) (1996). Molecule-Based Magnetic Materials: Theory, Techniques, and Applications, ACS Symposium Series 644. Washington, DC: American Chemical Society. 6 Lahti, P.M. (ed.) (1999). Magnetic Properties of Organic Materials. New York: Marcel Dekker. 7 Kahn, O. (1993). Molecular Magnetism. New York: Wiley-VCH. 8 Miller, J.S. (1994). Molecular/organic magenets-potential applications. Adv. Mater. 6: 322–324. 9 Schweinfurth, D., Krzystek, J., Atanasov, M. et al. (2017). Tuning magnetic anisotropy through ligand substitution in five-coordinate Co(II) complexes. Inorg. Chem. 56: 5253–5265. 10 Ara, F., Oka, H., Sainoo, Y. et al. (2019). Spin properties of single-molecule magnet of double-decker Tb(III)-phthalocyanine (TbPc2 ) on ferromagnetic Co film characterized by spin polarized STM (SP-STM). J. Appl. Phys. 125: 183901. 11 Zheng, Y.Z., Zheng, Z.P., and Chen, X.M. (2014). A symbol approach for classification of molecule-based magnetic materials exemplified by coordination polymers of metal carboxylates. Coord. Chem. Rev. 258: 1–15. 12 Huang, Y.G., Jiang, F.L., and Hong, M.C. (2014). Magnetic lanthanide-transition-metal organic-inorganic hybrid materials: from discrete clusters to extended frameworks. Coord. Chem. Rev. 253: 2814–2834. 13 Meng, Y.S., Jiang, S.D., Wang, B.W., and Gao, S. (2016). Understanding the magnetic anisotropy toward single-ion magnets. Acc. Chem. Res. 49: 2381–2389.

823

824

13 Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets

14 Halcrow, M.A. (2009). Iron(II) complexes of 2,6-di(pyrazol-1-yl)pyridines – a versatile system for spin-crossover research. Coord. Chem. Rev. 253: 2493–2514. 15 Nisoli, C., Kapaklis, V., and Schiffer, P. (2017). Deliberate exotic magnetism via frustration and topology. Nat. Phys. 13: 200–207. 16 Halcrow, M.A. (ed.) (2013). Spin-Crossover Materials: Properties and Application. West Sussex: Wiley. 17 König, E., Ritter, G., and Kulshreshtha, S.K. (1985). The nature of spin-state transitions in solid complexes of iron(II) and the interpretation of some associated phenomena. Chem. Rev. 85: 219–234. 18 Venkataramani, S., Jana, U., Dommaschk, M. et al. (2011). Magnetic bistability of molecules in homogeneous solution at room temperature. Science 331: 445–448. 19 Holleman, A.F., Wiberg, E., and Wiberg, N. (2007). Lehrbuch der anorganischen Chemie. Berlin: de Gruyter. 20 Halcrow, M.A. (2011). Structure: function relationships in molecular spin-crossover complexes. Chem. Soc. Rev. 40: 4119–4142. 21 König, E. (1987). Structural changes accompanying continuous and discontinuous spin-state transitions. Prog. Inorg. Chem. 35: 527–622. 22 McCusker, J.K., Rheingold, A.L., and Hendrickson, D.N. (1996). Variable-temperature studies of laser-initiated 5 T2 → 1 A1 intersystem crossing in spin-crossover complexes: empirical correlations between activation parameters and ligand structure in a series of polypyridyl ferrous complexes. Inorg. Chem. 35: 2100–2112. 23 Holland, J.M., McAllister, J.A., Kilner, C.A. et al. (2002). Stereochemical effects on the spin-state transition shown by salts of [FeL2 ]2+ [L = 2,6-di(pyrazol-1-yl)pyridine]. J. Chem. Soc., Dalton Trans.: 548–554. 24 Holland, J.M., McAllister, J.A., Lu, Z. et al. (2001). An unusual abrupt thermal spin-state transition in [FeL2 ][BF4 ]2 [L = 2,6-di(pyrazol-1-yl)pyridine]. Chem. Commun.: 577–578. 25 Elhaïk, J., Kilner, C.A., and Halcrow, M.A. (2006). Structural diversity in iron(II) complexes of 2,6-di(pyrazol-1-yl)pyridine and 2,6-di(3-methylpyrazol-1-yl)pyridine. Dalton Trans.: 823–830. 26 Salmon, L., Molnár, G., Cobo, S. et al. (2009). Re-investigation of the spin crossover phenomenon in the ferrous complex [Fe(HB(pz)3 )2 ]. New J. Chem. 33: 1283–1289. 27 Reger, D.L., Gardinier, J.R., Elgin, J.D. et al. (2006). Structure-function correlations in iron(II) tris(pyrazolyl)borate spin-state crossover complexes. Inorg. Chem. 45: 8862–8875. 28 Faulmann, C., Szilágyi, P.A., Jacob, K. et al. (2009). Polymorphism and its effects on the magnetic behaviour of the [Fe(sal2 -trien)][Ni(dmit)2 ] spin-crossover complex. New J. Chem. 33: 1268–1276. 29 Clemente-León, M., Coronado, E., López-Jordà, M. et al. (2010). Multifunctional magnetic materials obtained by insertion of a spin-crossover FeIII complex into bimetallic oxalate-based ferromagnets. Chem. Eur. J. 16: 2207–2219.

References

30 Pritchard, R., Barrett, S.A., Kilner, C.A., and Halcrow, M.A. (2008). The influence of ligand conformation on the thermal spin transitions in iron(III) saltrien complexes. Dalton Trans.: 3159–3168. 31 Ross, T.M., Neville, S.M., Innes, D.S. et al. (2010). Spin crossover in iron(III) Schiff-base 1-D chain complexes. Dalton Trans. 39: 149–159. 32 Li, X.M., Ji, Y., Wang, C.F. et al. (2009). Synthesis, structures and magnetic properties of nickel bis(dithiolene) complexes with [Fe(qsal)2]+ . J. Coord. Chem. 62: 1544–1552. 33 Bauer, W. and Weber, B. (2009). X-ray structure and magnetic properties of dinuclear and polymer iron(II) complexes. Inorg. Chim. Acta 362: 2341–2346. 34 Desiraju, G.R. (2007). Crystal engineering: a holistic view. Angew. Chem. Int. Ed. 46: 8342–8356. 35 Lazar, H.Z., Forestier, T., Barrett, S.A. et al. (2007). Thermal and light-induced spin-crossover in salts of the heptadentate complex [tris(4-{pyrazol-3-yl}-3-aza-3-butenyl)amine]iron(II). Dalton Trans.: 4276–4285. 36 Zein, S., Matouzenko, G.S., and Borshch, S.A. (2005). Quantum chemical study of three polymorphs of the mononuclear spin-transition complex [Fe(DPPA)(NCS)2 ]. J. Phys. Chem. A 109: 8568–8571. 37 Chernyshow, D., Vandal, B., Törnroos, K.W., and Bürgi, H.B. (2009). Chemical disorder and spin crossover in a mixed ethanol-2-propanol solvate of FeII tris(2-picolylamine) dichloride. New J. Chem. 33: 1277–1282. 38 Sorai, M. and Seki, S. (1974). Phonon coupled cooperative low-spin 1 A1 high-spin 5 T2 transition in [Fe(phen)2 (NCS)2 ] and [Fe(phen)2 (NCSe)2 ] crystals. Phys. J. Chem. Solids 35: 555–570. 39 Wajnflasz, J. and Pick, R. (1971). J. Phys. Colloq. (Paris) 32: C191. 40 Nishino, M., Boukheddaden, K., Konishi, Y., and Miyashita, S. (2007). Simple two-dimensional model for the elastic origin of cooperativity among spin states of spin-crossover complexes. Phys. Rev. Lett. 98: 247203. 41 Money, V.A., Carbonera, C., Elhaïk, J. et al. (2007). Interplay between kinetically slow thermal spin-crossover and metastable high-spin state relaxation in an iron(II) complex with similar T1/2 and T(LIESST). Chem. Eur. J. 13: 5503–5514. 42 Guionneau, P., Le Gac, F., Kaiba, A. et al. (2007). A reversible metal-ligand bond break associated to a spin-crossover. Chem. Commun.: 3723–3725. 43 Money, V.A., Elhaïk, J., Evans, I.R. et al. (2004). A study of the thermal and light induced spin transition in [FeL2 ](BF4 )2 and [FeL2 ](ClO4 )2 L = 2,6-di(3-methylpyrazol-1-yl)pyrazine. Dalton Trans.: 65–69. 44 Marchivie, M., Guionneau, P., Létard, J.F., and Chasseau, D. (2003). Towards direct correlations between spin-crossover and structural features in iron(II) complexes. Acta Crystallogr., Sect. B: Struct. Sci. 59: 479–486. 45 Pfaffeneder, T.M., Thallmair, S., Bauer, W., and Weber, B. (2011). Complete and incomplete spin transitions in 1D chain iron(II) compounds. New J. Chem. 35: 691–700.

825

826

13 Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets

46 AromíG, L., Barriosa, A., Roubeau, O., and Gamez, P. (2011). Triazoles and tetrazoles: prime ligands to generate remarkable coordination materials. Coord. Chem. Rev. 255: 485–546. 47 Real, J., Gallois, B., Granier, T. et al. (1992). Comparative investigation of the spin-crossover compounds Fe(btz)2 (NCS)2 and Fe(phen)2 (NCS)2 (where btz = 2,2′ -bi-4,5-dihydrothiazine and phen = 1,10-phenanthroline). Magnetic properties and thermal dilatation behaviour and crystal structures of Fe(btz)2 (NCS)2 at 293 K and 130 K. Inorg. Chem. 31: 4972–4979. 48 Nassirinia, N., Amani, S., Teat, S.J. et al. (2014). Enhancement of spin-crossover cooperativity mediated by lone pair-π interactions and halogen bonding. Chem. Commun. 50: 1003–1005. 49 Yan, Z., Li, M., Gao, H.L. et al. (2012). High-spin versus spin-crossover versus low-spin: geometry intervention in cooperativity in a 3D polymorphic iron(II)–tetrazole MOFs system. Chem. Commun. 48: 3960–3962. 50 Li, J.Y., Chen, Y.C., Zhang, Z.M. et al. (2015). Tuning the spin-crossover behaviour of a hydrogen-accepting porous coordination polymer by hydrogen-donating guests. Chem. Eur. J. 21: 1645–1651. 51 Li, J.Y., He, C.T., Chen, Y.C. et al. (2015). Tunable cooperativity in a spin-crossover Hoffman-like metal-organic framework material by aromatic guests. J. Mater. Chem. C 3: 7830–7835. 52 Zenere, K.A., Duyker, S.G., Trzop, E. et al. (2018). Increasing spin crossover cooperativity in 2D Hofmann-type materials with guest molecule removal. Chem. Sci. 9: 5623–5629. 53 Murphy, M.J., Zenere, K.A., Ragon, F. et al. (2017). Guest programmable multistep spin crossover in a porous 2-D Hofmann-type material. J. Am. Chem. Soc. 139: 1330–1335. 54 Gatteschi, D. and Sessoli, R. (2003). Quantum tunneling of magnetization and related phenomena in molecular materials. Angew. Chem. Int. Ed. 42: 268–297. 55 Bogani, L. and Wernsdorfer, W. (2009). Molecular spintronics using single-molecule magnets. Nat. Mater. 7: 179–186. 56 Frost, J.M., Harriman, K.L., and Murugesu, M. (2016). The rise of 3-D single-ion magnets in molecular magnetism: towards materials from molecules? Chem. Sci. 7: 2470–2491. 57 Sessoli, R., Gatteschi, D., Caneschi, A., and Novak, M.A. (1993). Magnetic bistability in a metal-ion cluster. Nature 365: 141–143. 58 Sessoli, R., Tsai, H.L., Schake, A.R. et al. (1993). High-spin molecules: [Mn12 O12 (O2 CR)16 (H2 O)4 ]. J. Am. Chem. Soc. 115: 1804–1816. 59 Bogani, L. and Wernsdorfer, W. (2008). Molecular spintronics using single-molecule magnets. Nat. Mater. 7: 179–186. 60 Atkinson, J.H., Fournet, A.D., Bhaskaran, L. et al. (2017). Effects of uniaxial pressure on the quantum tunneling of magnetization in a high-symmetry Mn12 single-molecule magnet. Phys. Rev. B 95: 184403. 61 Aromi, G. and Brechin, E.K. (2006). Synthesis of 3D metallic single-molecule magnets. Struct. Bond. 122: 1–67.

References

62 Milios, C.J., Vinslava, A., Wernsdorfer, W. et al. (2007). A record anisotropy barrier for a single-molecule magnet. J. Am. Chem. Soc. 129: 2754–2755. 63 Waldamann, O. (2007). A criterion for the anisotropy barrier in single-molecule magnets. Inorg. Chem. 46: 10035–10037. 64 Hill, S., Datta, S., Liu, J. et al. (2010). Magnetic quantum tunneling: insights from simple molecule-based magnets. Daltron Trans. 39: 4693–4707. 65 Gao, S. (2015). Molecular Nanomagnets and Related Phenomena. Heidelberg, New York, Dordrech, London: Springer-Verlag. 66 Ishikawa, N., Sugita, M., Ishikawa, T. et al. (2003). Lanthanide double-decker complexes functioning as magnets at the single-molecular level. J. Am. Chem. Soc. 125: 8694–8695. 67 Ishikawa, N., Sugita, M., and Wernsdorfer, W. (2005). Quantum tunneling of magnetization in lanthanide single-molecule magnets: bis(phthalocyaninato) terbium and bis(phthalocyaninato) dysprosium anions. Angew. Chem. Int. Ed. 44: 2931–2935. 68 Woodruff, D.N., Winpenny, R.E.P., and Layfield, R.A. (2013). Lanthanide single-molecule magnets. Chem. Rev. 113: 5110–5148. 69 Takamatsu, S., Ishikawa, T., Koshihara, S., and Ishikawa, N. (2007). Significant increase of the barrier energy for magnetization reversal of a single-4f-ionic single-molecule magnet by a longitudinal contraction of the coordination space. Inorg. Chem. 46: 7250–7252. 70 Liu, J.L., Chen, Y.C., and Tong, M.L. (2018). Symmetry strategies for high performance lanthanide-based single-molecule magnets. Chem. Soc. Rev. 47: 2431–2453. 71 Uugur, L., Le Roy, J.J., Korobkov, I. et al. (2014). Fine-tuning the local symmetry to attain record blocking temperature and magnetic remanence in a single-ion magnet. Angew. Chem. Int. Ed. 53: 4413–4417. 72 Liu, J., Chen, Y.C., Liu, J.L. et al. (2016). A stable pentagonal bipyramidal Dy(III) single-ion magnet with a record magnetization reversal barrier over 1000 K. J. Am. Chem. Soc. 138: 5441–5450. 73 Ding, Y.S., Chilton, N.F., Winpenny, R.E.P., and Zheng, Y.Z. (2016). On approaching the limit of molecular magnetic anisotropy: a near-perfect pentagonal bipyramidal dysprosium(III) single-molecule magnet. Angew. Chem. Int. Ed. 55: 16071–16074. 74 Goodwin, C.A.P., Ortu, F., Reta, D. et al. (2017). Molecular magnetic hysteresis at 60 kelvin in dysprosocenium. Nature 548: 439–442. 75 Guo, F.S., Day, B.M., Chen, Y.C. et al. (2017). A dysprosium metallocene single-molecule magnet functioning at the axial limit. Angew. Chem. Int. Ed. 56: 11445–11449. 76 Guo, F.S., Day, B.M., Chen, Y.C. et al. (2018). Magnetic hysteresis up to 80 kelvin in a dysprosium metallocene single-molecule magnet. Science 362: 1400–1403. 77 Donati, F., Rusponi, S., Stepanow, S. et al. (2016). Magnetic remanence in single atoms. Science 352: 318–321.

827

828

13 Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets

78 Natterer, F.D., Yang, K., Paul, W. et al. (2017). Reading and writing single-atom magnets. Nature 543: 226–228. 79 Freedman, D.E., Harman, W.H., Harris, T.D. et al. (2010). Slow magnetic relaxation in a high-spin iron(II) complex. J. Am. Chem. Soc. 132: 1224–1225. 80 Harman, W.H., Harris, T.D., Freedman, D.E. et al. (2010). Slow magnetic relaxation in a family of trigonal pyramidal iron(II) pyrrolide complexes. J. Am. Chem. Soc. 132: 18115–18126. 81 Zadrozny, J.M., Xiao, D.J., Atanasov, M. et al. (2013). Magnetic blocking in a linear iron(I) complex. Nat. Chem. 5: 577–581. 82 Feng, X., Mathonière, C., Jeon, I.R. et al. (2013). Tristability in a light-actuated single-molecule magnet. J. Am. Chem. Soc. 135: 15880–15884. 83 Glauber, R. (1963). Time-dependent statistics of the ising model. J. Math. Phy. 4: 294–307. 84 Caneschi, A., Gatteschi, D., Lalioti, N. et al. (2001). Cobalt(II)-nitronyl nitroxide chains as molecular magnetic nanowires. Angew. Chem. Int. Ed. 40: 1760–1763. 85 Zhang, W.X., Ishikawa, R., Breedlove, B., and Yamashita, M. (2013). Single-chain magnets: beyond the Glauber model. RSC Adv. 3: 3772–3798. 86 Sato, O., Hayami, S., Einaga, Y., and Gu, Z.Z. (2003). Control of the magnetic and optical properties in molecular compounds by electrochemical, photochemical and chemical methods. Bull. Chem. Soc. Jpn. 76: 443–470. 87 Tanase, S. and Reedijk, J. (2006). Chemistry and magnetism of cyanido-bridged d–f assemblies. Coord. Chem. Rev. 250: 2501–2510. 88 Wang, Z.M., Zhang, B., Fujiwara, H. et al. (2004). Mn3 (HCOO)6 : a 3D porous magnet of diamond framework with nodes of Mn-centered MnMn4 tetrahedron and guest-modulated ordering temperature. Chem. Commun.: 416–417. 89 Wang, X.Y., Wang, Z.M., and Gao, S. (2008). Constructing magnetic molecular solids by employing three-atom ligands as bridges. Chem. Commun.: 281–294. 90 Wang, X.Y., Wang, Z.M., and Gao, S. (2006). Solvent-tuned azido-bridged Co2+ layers: square, honeycomb, and Kagomé. J. Am. Chem. Soc. 128: 674–675. 91 Zeng, Y.F., Hu, X., Liu, F.C., and Bu, X.H. (2009). Azido-mediated systems showing different magnetic behaviors. Chem. Soc. Rev. 38: 469–480. 92 Wang, X.Y., Wang, L., Wang, Z.M. et al. (2005). Coexistence of spin-canting, metamagnetism, and spin-flop in a (4,4) layered manganese azide polymer. Chem. Mater. 17: 6369–6380. 93 Ma, B.Q., Sun, H.L., Gao, S., and Su, G. (2001). A novel azido and pyrazine-dioxide bridged three-dimensional manganese(II) network with antiferromagnetic ordering (TN = 62 K) and a spin flop state. Chem. Mater. 13: 1946–1948. 94 Zhang, Y.Z., Wernsdorfer, W., Pan, F. et al. (2006). An azido-bridged disc-like heptanuclear cobalt(II) cluster: towards a single-molecule magnet. Chem. Commun.: 3302–3304. 95 Gu, Z.G., Zuo, J.L., and You, X.Z. (2007). A three-dimensional ferromagnet based on linked copper-azido clusters. Dalton Trans.: 4067–4072.

References

96 Zeng, Y.F., Liu, F.C., Zhao, J.P. et al. (2006). An azido-metal-isonicotinate complex showing long-range ordered ferromagnetic interaction: synthesis, structure and magnetic properties. Chem. Commun.: 2227–2229. 97 Zeng, Y.F., Zhao, J.P., Hu, B.W. et al. (2007). Structures with tunable strong ferromagnetic coupling: from unordered (1D) to ordered (discrete). Chem. Eur. J. 13: 9924–9930. 98 Goher, M.A.S., Cano, J., Journaux, Y. et al. (2000). Synthesis, structural characterisation, and Monte Carlo simulation of the magnetic properties of the 3D-stacked honeycomb Csn [{Mn(N3 )3 }n ] and the irregular double chain [{N(C2 N5 )4 }n ][{Mn2 (N3 )5 (H2 O)}n ]. Chem. Eur. J. 6: 778–784. 99 Abu-Youssef, M.A.M., Escuer, A., Gatteschi, D. et al. (1999). Synthesis, structural characterization, magnetic behavior, and single crystal EPR spectra of three new one-dimensional manganese azido systems with FM, alternating FM-AF, and AF coupling. Inorg. Chem. 38: 5716–5723. 100 Liu, F.C., Zeng, Y.F., Li, J.R. et al. (2005). Novel 3-D framework nickel(II) complex with azide, nicotinic acid, and nicotinate(1-) as coligands: hydrothermal synthesis, structure, and magnetic properties. Inorg. Chem. 44: 7298–7300. 101 Woodard, B., Willett, R.D., Haddad, S. et al. (2004). Structure of two azide salts of a copper(II) macrocycle and magnetic properties of Cu(14ane)Cu(N3 )4 . Inorg. Chem. 43: 1822–1824. 102 Fu, A.H., Huang, X.Y., Li, J. et al. (2002). Controlled synthesis and magnetic properties of 2D and 3D iron azide networks [Fe(N3 )2 (4,4′ -bpy)] and [Fe(N3 )2 (4,4′ -bpy)]. Chem. Eur. J. 8: 2339–2247. 103 Abu-Youssef, M.A.M., Escuer, A., Goher, M.A.S. et al. (2000). Can a homometallic chain be ferromagnetic? Angew. Chem. Int. Ed. 39: 1624–1626. 104 Martin, S., Barandika, M.G., Lezama, L. et al. (2001). Weak M(II)-Azide-4,4′ -bipy ferromagnets based on unusual diamondoid (M = Mn) and 2D arrays (M = Co, Ni). Inorg. Chem. 40: 4109–4115. 105 Escuer, A., Mautner, F.A., Goher, M.A.S. et al. (2005). A two-dimensional azido-based topologic ferrimagnet. Chem. Commun.: 605–607. 106 Monfort, M., Resino, I., Fallah, M.S.E. et al. (2001). Synthesis, structure, and magnetic properties of three new one-dimensional nickel(II) complexes: new magnetic model for the first one-dimensional S = 1 complex with alternating ferro-ferromagnetic coupling. Chem. Eur. J. 7: 280–287. 107 Liu, F.C., Zeng, Y.F., Jiao, J. et al. (2006). First metal azide complex with isonicotinate as a bridging ligand showing new net topology: hydrothermal synthesis, structure, and magnetic properties. Inorg. Chem. 45: 2776–2778. 108 Zeng, Y.F., Hu, X., Zhao, J.P. et al. (2008). Partial substitution of hydroxyl by azide: an unprecedented 2D azido-copper-hydroxyl compound with a [Cu24 ] macrocycle in the presence of [Cu(H2 O)6 ]2+ . Chem. Eur. J. 14: 7127–7130. 109 Hong, C.S., Koo, J.E., Son, S.K. et al. (2001). Unusual ferromagnetic couplings in single end-to-end azide-bridged cobalt(II) and nickel(II) chain systems. Chem. Eur. J. 7: 4243–4252.

829

830

13 Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets

110 Zhao, X.H., Huang, X.C., Zhang, S.L. et al. (2013). Cation-dependent magnetic ordering and room-temperature bistability in azido-bridged perovskite-type compounds. J. Am. Chem. Soc. 135: 16006–16009. 111 Ge, C.H., Cui, A.L., Ni, Z.H. et al. (2006). μ1,1 -Azide-bridged ferromagnetic MnIII dimer with slow relaxation of magnetization. Inorg. Chem. 45: 4883–4885. 112 Liu, T.F., Fu, D., Gao, S. et al. (2003). An azide-bridged homospin single-chain magnet: [Co(2,2′ -bithiazoline)(N3 )2 ]n . J. Am. Chem. Soc. 125: 13976–13977. 113 Wang, Z.M., Zhang, B., Inoue, K. et al. (2007). Occurrence of a rare 49 ⋅66 structural topology, chirality, and weak ferromagnetism in the [NH4 ][MII (HCOO)3 ] (M = Mn, Co, Ni) frameworks. Inorg. Chem. 46: 437–445. 114 Wang, X.Y., Gan, L., Zhang, S.W., and Gao, S. (2004). Perovskite-like metal formates with weak ferromagnetism and as precursors to amorphous materials. Inorg. Chem. 43: 4615–4625. 115 Dybtsev, D.N., Chun, H., Yoon, S.H. et al. (2004). Microporous manganese formate: a simple metal-organic porous material with high framework stability and highly selective gas sorption properties. J. Am. Chem. Soc. 126: 32–33. 116 Roubeau, O., Lorusso, G., Teat, S.J., and Evangelisti, M. (2014). Cryogenic magneto-caloric effect and magneto-structural correlations in carboxylate-bridged Gd(III) compounds. Dalton Trans. 43: 11502–11509. 117 Huang, Y.G., Wu, S.Q., Deng, W.H. et al. (2017). Selective CO2 capture and high proton conductivity of a functional star-of-david catenane metal–organic framework. Adv. Mater. 29: 1703301. 118 Mukherjee, P.S., Konar, S., Zangrando, E. et al. (2002). Synthesis, crystal structure and magneto-structural correlation of two bi-bridging 1D copper(II) chains. J. Chem. Soc., Dalton Trans.: 3471–3476. 119 Coronado, E., Galán-Mascarós, J.R., and Martí-Gastaldo, C. (2006). Oxalate-based 2D magnets: the series [NBu4 ][MII MnIII (ox)3 ] (MII = Fe, Co, Ni, Zn; ox = oxalate dianion). J. Mater. Chem. 16: 2685–2689. ´ M., Toric, ´ F., and Pajic, ´ D. (2017). The influence of 120 Dubraja, L.A., Juric, metal centres on the exchange interaction in heterometallic complexes with oxalate-bridged cations. Dalton Trans. 46: 11748–11756. ´ M., Popovic, ´ J. et al. (2018). Magneto-structural correla121 Dubraja, L.A., Juric, tions in oxalate-bridged Sr(II)Cr(III) coordination polymers: structure, magnetization, X-band, and high-field ESR studies. Dalton Trans. 47: 3992–4000. 122 Wannier, G.H. (1950). Antiferromagnetism. The triangular ising net. Phys. Rev. 79: 357–364. 123 Anderson, P.W. (1956). Ordering and antiferromagnetism in ferrites. Phys. Rev. 102: 1008–1014. 124 Schoutens, K. (2009). Wormholes in quantum matter. Nat. Phys. 5: 784–785. 125 Smart, J.S. (1966). Effect Field Theories of Magnetism. W. B. Saunders. 126 Schiffer, P. and Ramirez, A.P. (1996). Commun. Condens. Matter Phys. 10: 21. 127 Lacorre, P. (1987). The constraint functions: an attempt to evaluate the constraint rate inside structures that undergo ordered magnetic frustration. J. Phys. C 20: L775–L781.

References

128 Greedan, J.E. (2001). Geometrically frustrated magnetic materials. J. Mater. Chem. 11: 37–53. 129 Itoh, T., Matsuda, K., and Iwamura, H. (1999). A triphenylamine derivative with three p-(N-tert-butyl-N-oxylamino)phenyl radical units and yet a doublet ground state. Angew. Chem. Int. Ed. 38: 1791–1793. 130 Suzuki, S., Nagata, A., Kuratsu, M. et al. (2012). Trinitroxidetrioxytriphenylamine: spin-state conversion from triradical doublet to diradical cation triplet by oxidative modulation of a π-conjugated system. Angew. Chem. Int. Ed. 51: 3193–3197. 131 Wu, Y.L., Krzyaniak, M.D., Stoddart, J.F., and Wasielewski, M.R. (2017). Spin frustration in the triradical trianion of a naphthalenediimide molecular triangle. J. Am. Chem. Soc. 139: 2948–2951. 132 Seeber, G., Kögerler, P., Kariuki, B.M., and Cronin, L. (2004). Supramolecular assembly of ligand-directed triangular {CuII 3 Cl} clusters with spin frustration and spin-chain behaviour. Chem. Commun.: 1580–1581. 133 Wang, L.L., Sun, Y.M., Yu, Z.Y. et al. (2009). Theoretical investigation on triagonal symmetry copper trimers: magneto-structural correlation and spin frustration. J. Phys. Chem. A 113: 10534–10539. 134 Pineda, E.M., Lorusso, G., Zangana, K.H. et al. (2016). Observation of the influence of dipolar and spin frustration effects on the magnetocaloric properties of a trigonal prismatic {Gd7 } molecular nanomagnet. Chem. Sci. 7: 4891–4895. 135 Stamatatos, T.C., Foguet-Albiol, D., Poole, K.M. et al. (2009). Spin maximization from S = 11 to S = 16 in Mn7 disk-like clusters: spin frustration effects and their computational rationalization. Inorg. Chem. 48: 9831–9845. 136 Xiang, S.C., Hu, S.M., Sheng, T.L. et al. (2007). A fan-shaped polynuclear Gd6 Cu12 amino acid cluster: a “hollow” and ferromagnetic [Gd6 (μ3 -OH)8 ] octahedral core encapsulated by six [Cu2 ] glycinato blade fragments. J. Am. Chem. Soc. 129: 15144–15146. 137 Shores, M.P., Bartlett, B.M., and Nocera, D.G. (2005). Spin-frustrated organic-inorganic hybrids of lindgrenite. J. Am. Chem. Soc. 127: 17986–17987. 138 Gautier, R., Oka, K., Kihara, T. et al. (2013). Spin frustration from cis-edge or -corner sharing metal-centered octahedra. J. Am. Chem. Soc. 135: 19268–19274. 139 Humphrey, S.M. and Wood, P.T. (2004). Multiple areas of magnetic bistability in the topological ferrimagnet [Co3 (NC5 H3 (CO2 )2 -2,5)2 (μ3 -OH)2 (OH2 )2 ]. J. Am. Chem. Soc. 126: 13236–13237. 140 Huang, Y.G., Yuan, D.Q., Pan, L. et al. (2007). A 3D porous cobalt-organic framework exhibiting spin-canted antiferromagnetism and field-induced spin-flop transition. Inorg. Chem. 46: 9609–9615. 141 Postulka, L., Winter, S.M., Mihailov, A.G. et al. (2016). Spin frustration in an organic radical ion salt based on a Kagome-coupled chain structure. J. Am. Chem. Soc. 138: 10738–10741. 142 Nytko, E.A., Helton, J.S., Müller, P., and Nocera, D.G. (2008). A structurally perfect S = 1 /2 metal–organic hybrid Kagomé antiferromagnet. J. Am. Chem. Soc. 130: 2922–2923.

831

832

13 Understanding Magneto-Structural Correlations Toward Design of Molecular Magnets

143 Huang, X., Zhang, S., Liu, L.Y. et al. (2018). Superconductivity in a copper(II)-based coordination polymer with perfect Kagome structure. Angew. Chem. Int. Ed. 57: 146–150. 144 Tang, Y.Y., Peng, C., Guo, W.B. et al. (2017). Octa-Kagomé lattice compounds showing quantum critical behaviors: spin gap ground state versus antiferromagnetic ordering. J. Am. Chem. Soc. 139: 14057–14060. 145 Chi, Y., Xu, J., Xue, H.G. et al. (2019). Triple-Kagomé-layer slabs of mixed-valence rare-earth ions exhibiting quantum spin liquid behaviors: synthesis and characterization of Eu9 MgS2 B20 O41 . J. Am. Chem. Soc. 141: 9533–9536. 146 Zheng, Y.Z., Tong, M.L., Xue, W. et al. (2007). A “star” antiferromagnet: a polymeric iron(III) acetate that exhibits both spin frustration and long-range magnetic ordering. Angew. Chem. Int. Ed. 46: 6076–6080. 147 Bramwell, S.T. and Gingras, M.J.P. (2001). Spin ice state in frustrated magnetic pyrochlore materials. Science 294: 1495–1501.

833

14 Relationship Between MOF Structures and Gas Absorption Properties Qi Yin and Tian-Fu Liu Chinese Academy of Sciences, Fujian Institute of Research on the Structure of Matter, 155 Yangqiao Road West, Fuzhou, Fujian 350002, P.R. China

14.1 Introduction Metal–organic frameworks (MOFs), self-assembled by organic building blocks and metal ions/clusters, is a class of porous materials with high crystallinity, high surface area (SA), and tunable structure. These merits draw tremendous research interests to explore their potential application in gas storage and separation, catalysis, sensing, and so on. In term of gas storage, a next generation of adsorbents is required to meet high volumetric capacity for gases but without moving to higher adsorption pressures or lower temperatures, both of which would add significant complexity and cost. And MOFs provide us a promising platform that utilizes the intrinsic high surface area for targeting desire adsorption behaviors through rational structural design and functional tailoring. Therefore, MOFs have received significant attention as a new class of adsorbents. In this book chapter, we will introduce this field from the following six parts: (i) the basic conception about gas adsorption. (ii) The correlation between the type of adsorption isotherms and MOF structures. (iii) The pore sizes of MOFs fall into the range from several angstroms to several nanometers. These different structures give rise to different adsorption isotherms. From this part we will know some usually adsorption behaviors being observed in MOFs. (iv) Defects in MOFs and their influence on adsorption properties. (v) The factors effecting on gas adsorption and some representative MOFs with high surface area. (vi) The adsorption enthalpy of MOF materials including definition, influence factor, and effect on the adsorption capacity and working capacity of the practical application.

14.2 Basic Conceptions About Gas Absorption 14.2.1 Physical Adsorption and Chemical Adsorption As a thermodynamic function, Gibbs-free energy reflects the process of physical or chemical reactions. It refers to the reduced internal energy of a system that can Advanced Structural Chemistry: Tailoring Properties of Inorganic Materials and their Applications, First Edition. Edited by Rong Cao. © 2021 WILEY-VCH GmbH. Published 2021 by WILEY-VCH GmbH.

834

14 Relationship Between MOF Structures and Gas Absorption Properties

Table 14.1 The comparison between physical adsorption and chemical adsorption. Physical adsorption

Chemical adsorption

Effect force

van der Waals force

Chemical bond

Molecular layer

Monolayer or multilayers

Monolayer

Selectivity

Bad

Good

Adsorption heat

Low

High

Adsorption rate

Quick

Slow

be converted into external situation [1]. For reducing the Gibbs-free energy, the porous solids tend to absorb more gas molecules on their abundant surfaces. This phenomenon that the relative aggregation of gas molecules onto solid surfaces is named as “gas adsorption.” On the basis of Langmuir’s monolayer adsorption theory [2], Stephen Brunauer, Paul Hugh Emmett, and Edward Teller developed the multimolecular layer adsorption theory (called Brunauer–Emmett–Teller [BET] theory). The usual adsorbents are nitrogen and krypton for small surface area. The interference of chemical adsorption can be avoided by adsorption in the low temperature of liquid nitrogen. Gas adsorption theory has been used to study various porous materials, such as zeolite, porous silicon dioxide, porous polymer, and MOFs, for different application, such as chromatographic analysis, gas separation and purification, heterogeneous catalysis, medical detection, and other fields [3]. Adsorption effects can be categorized into physical adsorption and chemical adsorption owing to different forces between solid and gas molecules. For physical adsorption, van der Waals force is the main contributor, while chemical adsorption means to form new chemical bonds between adsorbate and adsorbent (Table 14.1). Both physical adsorption and chemical adsorption often occur simultaneously in the actual gas adsorption process [4].

14.2.2 Adsorption Curves At a certain temperature and pressure, when the rates of adsorption and desorption are equal, the adsorption equilibrium will be established. At this time, the amount of adsorbate taken up by the adsorbent per unit mass (or volume) of the adsorbent is called adsorption capacity (expressed as a). The adsorption isotherms of porous materials are usually used to study the process of adsorption. At the constant temperature, the plot reflecting the direct relationship between the adsorption equilibrium pressure p and the adsorption volume a is named adsorption isotherm. It is crystal clear that, for effective design of adsorption system and improvement of adsorption mechanism pathways, understanding and interpretation of adsorption isotherms are critical. Following, we list six common adsorption isotherms (Figure 14.1) [5].

14.2 Basic Conceptions About Gas Absorption

Amount adsorbed

Figure 14.1 Types of physisorption isotherms. Source: From Sing et al. [5a]. © IUPAC.

I

II

III

IV

V

VI

Relative pressure

14.2.3 Langmuir Monolayer Adsorption Isotherms Irving Langmuir presented his gas adsorption model of solid surface in 1916. In 1932, Langmuir was awarded the Nobel Prize for his research on surface chemistry. He hypothesized that a given surface has a certain number of equivalent sites to which a species can “stick,” either by physisorption or chemisorption [6]. There are six types of adsorption isotherms: (i) type-I (monolayer adsorption): gas molecules are adsorbed at a monolayer on the surface of the solid. Type-I models are presented when the adsorbed amount can be expressed only as a surface excess. While they attack the absorbing surface of the adsorbent, absorption will not happen. (ii) Type-II (multilayer): there is no effect between the adsorbed gas molecules. Thus, the adsorbed gas molecules could desorb without limitation of adjacent adsorbed gas molecules. (iii) Type-III (pore filling): the interaction between the adsorbent and the adsorbate is less than the interaction between the adsorbates. In low-pressure part, the adsorption capacity is small, and the effort between adsorbent and gas is very weak. In high-pressure part, high adsorption capacity indicates the presence of stacking pores [7]. (iv) Type-IV, types of physisorption isotherms: on account of capillary condensation, adsorption hysteresis occurs in these adsorbents. This phenomenon of adsorption hysteresis is related to the shape and size of the pores, so the theoretic size distribution of the pores can be known by analyzing the isotherm of adsorption–desorption. This is a special manifestation of mesoporous materials. (v) Type-V: a combination of types III and IV. (vi) Type-VI: multilayer adsorption on a solid uniform surface. And only a few materials have such adsorption curves. According to hypothesis type-I, the adsorption rate (r ads ) is proportional to the pressure of the gas (p) and (1 − 𝜃). rads = k1 (1 − 𝜃)p

(14.1)

where 𝜃 and (1 − 𝜃) are the percentage of the solid surface area occupied by the gas molecules is called the coverage degree and the percentage of uncovered solid surfaces, respectively.

835

836

14 Relationship Between MOF Structures and Gas Absorption Properties

As the hypothesis type-II, the desorption rate (r d ) is proportional to 𝜃. rd = k2 𝜃

(14.2)

When the adsorption equilibrium is reached, the adsorption rate is equal to the desorption rate. k1 (1 − 𝜃)p = k2 𝜃

(14.3)

Obviously, the adsorption volume a on the solid surface is proportional to 𝜃. kbp a = k𝜃 = (14.4) 1 + bp (b = k1 ∕k2 ) (a) When the pressure is very low, 1 + bp ≈ 1. a = kbp

(14.5)

(b) When the pressure is very high, 1 + bp ≈ bp. a=k

(14.6)

(c) Now, we rewrite this equation. p 1 + bp p 1 = = + a kb kb k

(14.7)

As shown in Figure 14.1, plotting a line using p/a in terms of p, the slope of curve gives the value of k, while the intercept on the y-axis gives the value of b [8]. It should be pointed out that the basic assumptions of the Langmuir equation can be only suitable for predicting adsorption on the ideal monolayer adsorption, namely the type-I isotherms. Although the Langmuir equation has certain limitations, the Langmuir equation is still a meaningful adsorption formula and its derivation provides important information for other adsorption studies.

14.2.4 BET Multilayer Adsorption Isotherms The BET theory was first published in 1938 by Stephen Brunauer et al. [9]. The BET theory extensively applies to determine the number of molecules of gas required to establish systems of multilayer adsorption and generally utilizes exploring gases that do not chemically react with the surface of the material to quantify particular surface area. Because most of gases and solids interact weakly to each other, the solid material must be cooled. Typically, standard BET analysis often directed at the boiling point of N2 (77 K). There are three hypotheses of BET theory: (i) gas molecules physically can be adsorbed to the surface of material by many layers. (ii) There are no interaction forces between each layer. (iii) Each layer adsorbent can fit the Langmuir theory.

14.3 The Correlation Between Physical Adsorption Isotherms and MOFs Structures

The BET equation has been derived: V=

Vm Cp (p∗ − p)[1 + (C − 1)p∕p∗ ]

(14.8)

where V

the volume of adsorbed gas at standard conditions, when the pressure is p

Vm

the volume of adsorbed gas at standard conditions, when the gas is absorbed on the monolayer of the surface

p*

saturated vapor pressure of the gas

C

a constant associated with gas adsorption

Here, we can rewrite the formula: p 1 C − 1• p = + V(p∗ − p) Vm C Vm C p∗

(14.9)

Apparently, if we plot a line using p/V(p* − p) in terms of p/p* , the slope is (C − 1)/V m C and intercept are 1/V m C. Vm =

1 slope + intercept

(14.10)

Thus, we can acquire the specific surface area S of a porous material, as long as we know the area of one gas molecule A. S=

Vm L A • 22 400 W

(14.11)

The unit of V m should be cm3 . 22 400

the molar volume of gas in the standard state

L

Avogadro number is 6.022 × 1023

W

the weight of the porous material

Based on these gas adsorption theories, reasonable design of porous adsorbents is great demands in heterogeneous catalysis, chromatographic analysis, gas separation and purification, medical detection, and other fields [10].

14.3 The Correlation Between Physical Adsorption Isotherms and MOFs Structures Nitrogen sorption isotherms are usually used to characterize the specific surface area and pore sizes of MOFs. Depending on the size of their pores or channels, porous materials like MOFs can be classified to three types, named micropores (less than 2 nm), mesopores (2–50 nm), and macropores (greater than 50 nm). Based on pore sharp, they also can be grouped into channel and cage. The appearance of N2 isotherms of MOFs at 77 K is related to the pore size and shapes, and we will discuss in detail as follows [11].

837

14 Relationship Between MOF Structures and Gas Absorption Properties

The majority of physisorption isotherms can be grouped into the six types as shown in Figure 14.1. Generally, MOFs exhibited type-I and/or type-IV isotherms, which are associated with microporous/mesoporous adsorption. In general, the characteristics of the type-I isotherms are that gas adsorption amount increases rapidly at the lower relative pressure area (p/p0 , where p0 is the saturation pressure of the adsorptive), and there is almost no obvious N2 adsorption after the micropores full filled, since the factor to limit uptake is mainly governed by the accessible pore volume of MOFs rather than by the internal surface area. When the saturation pressure reached, adsorbent condensation due to the metastability of the adsorbed multilayer may occur, and the hysteresis loop associated with strong interaction occasionally occurs during desorption. The type-IV isotherm commonly happened in mesoporous MOFs materials. It is usually characterized by the hysteresis phenomenon of sorption isotherms. A platform can be observed in areas with high p/p0 values, and sometimes desorption isotherms end up with higher level than adsorption isotherms (not closed).

14.3.1 Effect of Pore Size on N2 Adsorption Isotherms 14.3.1.1 The Microporous MOFs Sorption Isotherms

The type-I isotherm is the typical form of isotherm achieved from microporous MOFs. For example, Yaghi and coworkers reported a microporous MOF (MOF-5) based on Zn2+ and 1,4-benzene-dicarboxylate (BDC), which is stable when it is heated up to 300 ∘ C, in 1999 [12]. It is constructed by eight Zn4 (O)(CO2 )6 clusters with linear BDC linkers into a network constitute a unit cell (UC). The sorption of nitrogen revealed a reversible type I isotherm and shows no hysteresis upon desorption of gases from the pores. In 2008, Lillerud and coworkers use BDC as a linker to prepare Zr-MOF University of Oslo (UiO-66) [13]. An extremely stable framework is formed by Zr—O coordinate bonds. The reversible nitrogen adsorption isotherm of UiO-66 is shown as the standard type-I isotherm (Figure 14.2). 14.3.1.2 The Mesoporous MOFs Sorption Isotherms

There are two types of prototypical adsorption isotherms of mesoporous solids (types IV and V), but type IV isotherms are more common in various mesoporous MOFs 400

Uptake (cm3/g)

838

300

200

100

0

(a)

(b)

0

0.2

0.4

0.6

0.8

1.0

P/P0

Figure 14.2 (a) The crystalline structure of UiO-66. Source: From Wiersum et al. [14]. © John Wiley & Sons. (b) N2 isotherm of UiO-66 at 77 K. Source: Based on Furukawa et al. [15].

14.3 The Correlation Between Physical Adsorption Isotherms and MOFs Structures

materials. In 2012, a stable Zr-MOFs, porous coordination networks (PCN)-222, was reported by Zhou and coworkers Using highly stable Zr6 clusters as nodes, this MOFs based on photosensitizer TCPP (TCPP = tetrakis(4-carboxyphenyl) porphyrin) as linkers are candidates of photocatalysts and photo therapeutic agent [16]. The structure is shown in Figure 14.3. The TCPP is connected to four Zr6 clusters with a coordinate bond. The pore diameter of its meso-channel is 3.7 nm. Nitrogen adsorption isotherm of PCN-222 is shown in Figure 14.3, which has typical characteristics of the type-IV isotherm. The isotherm demonstrates a sharp rise at the point of P/P0 = 0.3, revealing the microporous nature of the MOFs. Yaghi and coworkers reported the mesoporous MOF (MOF-200) [17] by using of 4,4′ ,4-[benzene-1,3,5-triyl-tris(benzene-4,1-diyl)] tribenzoate (BBC) as linkers and Zn4 O(CO2 )6 as nodes. The cage size for MOF-200 is 18 with an internal diameter of 28 Å, which is on the border of micropore and mesopore. Because of the micropore filling at P/P0 = 0.12, the predicted isotherms just show a step. The type of adsorption isotherm obtained from the experiment is similar to the calculated one, and the value is slightly smaller, which conforms the characteristics of sorption isotherm of mesoporous materials. The maximum nitrogen uptake capacities of MOF-200 at 77 K is 2340 cm3 /g. 14.3.1.2.1 The Macroporous MOFs Sorption Isotherms

With the studies reported so far, the macroporous materials prepared on the basis of microporous MOF still conform to the adsorption isotherm of microporous materials. For example, Li and coworkers synthesized a new three-dimensional (3D) crystalline MOF, named single-crystal ordered macropore (SOM–zeolitic imidazolate framework [ZIF-8]) with the macro-micropores and applied the ZIF-8 structure as a proof of concept (Figure 14.4) [19]. The adsorption isotherm of SOM–ZIF-8 conforms to the microporous MOF isotherm, with standard type-I isotherms. From the above, we can get a conclusion that N2 gas sorption is a basic characterization of MOFs materials, and the shape of the sorption isotherms can provide valuable information about MOFs materials.

3.7 nm

(a)

(b)

(c)

N2 adsorbed quantity (cm3/g)

1000

800

600

400

PCN-222 (Fe) adsorption PCN-222 (Fe) desorption

200

0 0.0

(d)

(e)

(f)

0.2

0.4

0.6

0.8

1.0

Pressure (P/P0)

Figure 14.3 (a–e) Crystal structure and underlying network topology of PCN-222(Fe). (Zr, black spheres; C, gray; O, red; N, blue; Fe, orange). (f) N2 sorption isotherms for PCN-222(Fe) at 77 K, 1 atm. Source: From Feng et al. [16]. © 2012, John Wiley & Sons.

839

14 Relationship Between MOF Structures and Gas Absorption Properties

Removing PS

THF

PS@MOF

Macro@MOF

Figure 14.4 A strategy of designing microporous MOFs materials. Source: From Zhao et al. [18]. © 2019 American Chemical Society.

14.3.2 Effect of Pore Shape on Sorption Isotherms 14.3.2.1 The Gas Sorption Isotherms for MOFs with Channels

MOFs with channels often adsorb minor gas molecules below the phase change pressure. A steep slope with a sharp increase often occurs at the beginning of the adsorption isotherms of MOFs material as a kind of characteristic. However, the crystal structure of MOFs material can be easily destroyed by multiple gas exchange. This phenomenon will restrict the practical application of MOFs on storage capacities and enhanced selectivity. The pores of the channels in the MOFs are sometimes deformed, and this flexibility often leads to the step-like shape of adsorption isotherm. As an example, Long and coworkers studied the gas adsorption property and structural transformation based on a new family of flexible MOFs Co(bdp) (bdp = 1,4-benzenedipyrazolate) [20]. As shown in Figure 14.5, these MOFs are all low-porosity, collapsed phase to high-porosity, expanded phase as the gas pressure increases. Due to the flexibility of the structure, the adsorption isotherm at low pressure presents an unconventional step-like shape.

30 77 K 25 +CH4

–CH4

N2 adsorbed (mmol/g)

840

20 15 10 5 0 1E–4

(a)

(b)

1E–3

0.01

0.1

1

P(bar)

Figure 14.5 (a) Co(p-F2 -bdp) powder X-ray diffraction structures in vacuum (top) and 20 bar CH4 (bottom). (b) Low-pressure N2 adsorption of Co(p-F2 -bdp). Source: (a) From Taylor et al. [20]. © 2016, American Chemical Society, (b) Based on Taylor et al. [20].

14.3 The Correlation Between Physical Adsorption Isotherms and MOFs Structures

O

OH

MIL-100 O

HO

T ~9.7Å

OH O 1,3,5-BTC MIL-101 O C H2O,F Cr

T ~12.6Å

Trimers of chromium octahedra O OH HO

O 1,4-BDC Tetrahedra MIL-101/MIL-101

Cage A 𝜙~ 25–39Å

20 T

28 T Cage B 𝜙~ 25–34Å

~5 Å

~10 Å

Figure 14.6 Top: Trimers of chromium octahedra and 1,4-benzenedicarboxylate moieties or 1,3,5-benzenetricarboxylate groups constructed the tetrahedra (T) in MOFs structures. Bottom: The small cages, large cages, and basic unit cell of MOFs. Source: From Horcajada et al. [22]. © 2006, John Wiley & Sons.

14.3.2.2 The Gas Sorption Isotherms of MOFs with Cage Structures

The mesoporous MOF with cage pore may show the characteristics of micropore sorption under the influence of windows. It is worth mentioning that some MOFs with large cavities (>2 nm) but small connecting channels ( 2800 m2 /g). N2 uptakea) (cm3 /g)

SABET b) (m2 /g)

Pore volumes (cm3 /g)

Al-soc-MOF-1

∼1500

5585

2.3

[73]

Be-BTB

∼1000

4030

—

[74]

Beijing University of Technology (BUT-12)

982

3387

1.52

[75]

BUT-13

1422

3948

2.20

[75]

BUT-30

∼1300

3940

1.55

[76]

Cu2 (bpdc)2 (ted)

∼1100c)

3265e)

1.18

[77]

Cu2 (sdc)2 (ted)

∼1000c)

3129e)

1.07

[77]

Cu3 (btb)

∼1180

3288

1.77

[78]

Cu3 (tatb)

∼1250

3360

1.91

[78]

Cu6 O(TZI)3

∼750

2847

1.01

[79]

MOFs

DUT-6, MOF-205

References

1460

4810

2.26

[80]

1420

4460

2.16

[17]

DUT-13

1279

—

1.98

[81]

DUT-23(Co)

∼1350

4850

2.03

[82]

DUT-25

1442

4670

2.22

[83]

DUT-49

1880

5476

2.91

[84]

Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences (FJI)-H5

1240

4255

1.92

[85]

FJI-H6

1393

5007

2.16

[86]

FJI-H7

1029

3831

—

[86]

IRMOF-8

∼1320

4461

1.827

[87]

IRMOF-20

—

3409

1.53

[88]

Mg2 (dobpdc)

∼

3270

—

[89]

MIL-100(Cr)

∼800

3100d)

1.16

[90]

MIL-100(Fe)

∼600

>2800d)

—

[91]

MIL-101(Cr)

∼1260

4100

2.0

[21]

MOF-5

996.8

3800

—

[92]

MOF-177

∼1360

4740

1.59

[69, 93]

MOF-200

2340

4530

3.59

[80]

MOF-210

2330

6240

3.60

[80]

Nanjing University Bai’s group (NJU-Bai10)

∼720

2883

1.11

[94]

NOTT-101

∼710

2805

1.080

[95]

NOTT-102

∼800

3342

1.268

[95] (Continued)

14.5 Representative MOFs with Ultrahigh Surface Area and the Factors Effecting on Surface Area

Table 14.3

(Continued)

MOFs

N2 uptakea) (cm3 /g)

SABET b) (m2 /g)

Pore volumes (cm3 /g)

References

NOTT-103

∼750

2958

1.157

[95]

NOTT-110

790

2960

1.22

[96]

NOTT-111

768

2930

1.19

[96]

NOTT-112

∼1050

3800

1.62

[97]

NOTT-116

∼1500

4664

2.17e)

[98]

NOTT-119, PCN-69

1522

4118

2.35

[99]

∼1400

3989

2.31

[100]

NU-100

∼2000

6143

2.82

[101]

NU-109

∼2500

7010

3.75

[39]

NU-110

∼2800

7140

4.40

[39]

NU-111

∼1500

5000

2.38

[102]

NU-125

∼860

3120

1.29

[103]

NU-140

1275

4300

1.974

[104]

NU-1100

∼900

4020

1.53

[105]

NU-1101

∼1100

4422

1.72

[106]

NU-1102

∼1200

4712

2.00

[106]

NU-1103

∼1800

5646

2.91

[106]

NU-1104

∼1780

5290

2.79

[106]

PCN-6, CuTATB-60

∼1200

3811

1.55

[107]

PCN-20

∼1020

3525

1.59

[108]

PCN-56

—

3741

—

[109]

PCN-61

∼890

3000

1.36

[110]

PCN-66

∼1050

4000

1.63

[110]

PCN-68

∼1400

5109

2.13

[111]

PCN-80

∼940

3850

1.47

[112]

PCN-88

—

3308

—

[113]

PCN-88

′

—

3042

—

[113]

PCN-94

∼850

3377

1.32

[114]

PCN-111

>1200

4825

—

[115]

PCN-228

1245

4510

—

[116]

PCN-229

1455

4619

—

[116]

PCN-230

1085

4455

—

[116]

PCN-333(Al)

2490

4000

3.81

[117]

PCN-426-Cr(III)

∼875

3155

—

[118]

PCN-521

∼700

3411

—

[119] (Continued)

857

858

14 Relationship Between MOF Structures and Gas Absorption Properties

Table 14.3

(Continued)

MOFs

N2 uptakea) (cm3 /g)

SABET b) (m2 /g)

Pore volumes (cm3 /g)

References

PCN-888(Al)

2770

>3700

4.0

[120]

Seoul National University (SNU)-70′

∼1350

5290

2.17

[121]

SNU-77H

1050

3670

1.52

[122]

UiO-67

—

3000d)

—

[13]

UiO-68

—

4170d)

—

[13]

University of Michigan Crystalline Material (UMCM-1)

∼1400

4100

2.141

[123]

UMCM-2

1500

5200

—

[3]

UTSA-28a-Cu

∼850

3179

1.33

[124]

UTSA-76a

698.2

2820

1.09

[125]

UTSA-100a

∼800

3241

1.263

[126]

ZJU-5

694.5

2823

1.074

[127]

ZJU-32

958.3

3831

1.482

[128]

ZJU-35a

747

2899

1.159

[129]

ZJU-36a

1033

4014

1.599

[129]

Zr-4,4 -azobenzenedicarboxylic acid (ABDC)

∼1000c)

3000e)

—

[130]

Zr-4,4′ ,4′′ ,4′′′ -([1,1′ -biphenyl]3,3′ ,5,5′ -tetrayltetrakis(ethyne2,1-diyl)) tetrabenzoic acid (BTBA)

∼1080

4342

1.68

[131]

Zr-4,4′ ,4′′ ,4′′′ -(pyrene-1,3,6,8tetrayltetrakis(ethyne-2,1-diyl)) tetrabenzoic acid (PTBA)

∼1000

4116

1.55

[131]

NU-1301

∼3100c)

4750e)

3.9

[132]

′

a) b) c) d) e)

N2 uptake at 77 K. Calculated from N2 sorption isotherms at 77 K. Ar uptake at 87 K. Langmuir surface area. Calculated from the Ar sorption isotherm.

In 2007, Zhou and coworkers reported porous MOF, named PCN-6′ [133], generated by the copper paddlewheel second building unit (SBU) and 4,4′ ,4′′ -s-triazine2,4,6-triyl-tribenzoate (TATB) ligands, which is isostructural with HKUST-1, generated by Cu clusters and BTC ligands [134], of twist boracite topology [107]. In addition, both HKUST-1 and PCN-6′ crystallize in the cubic space group F432. In structure, each dicopper tetra carboxylate paddlewheel SBU connects four ligands to assemble a Td-octahedron. Because of the ligand length extension from BTC to TATB, square channel size, and Langmuir surface area enlarge from 8.0 Å and 1800 m2 /g (HKUST-1) to 15.16 Å and 2700 m2 /g (PCN-6′ ). In 2013, Wu and

14.5 Representative MOFs with Ultrahigh Surface Area and the Factors Effecting on Surface Area

Cu metal ao = 3.6 A

(a)

(b)

(c)

(d)

Figure 14.17 (a) One-unit cell of copper drawn to scale with: (b) Zr-MOF with 1,4-benzene-dicarboxylate (BDC) as linker, UiO-66, (c) Zr-MOF with 4,4′ -biphenyl-dicarboxylate (BPDC) as linker, UiO-67, (d) Zr-MOF with terphenyl dicarboxylate (TPDC) as linker, UiO-68. Zirconium, oxygen, carbon, and hydrogen atoms are red, blue, gray, and white, respectively. Source: From Cavka et al. [13]. © 2008, American Chemical Society.

coworkers reported two isoreticular crystals with HKUST-1, named Zhejiang University (ZJU)-35a and ZJU-36a, respectively [129, 134]. ZJU-35a and ZJU-36a are constructed from Cu clusters and ligands of L1 and L2, respectively. The results revealed that although L1 and L2 are apparently less symmetric than BTC, ZJU-35a, and ZJU-36a have the same topology (tbo) with HKUST-1 as well. Because of extending ligands, both ZJU-35a and ZJU-36a have larger pore sizes than those of HKUST-1: the smallest cage diameters are 5.3, 6.4, and 7.5 Å, and the largest diameters are 10.8, 14.4, and 16.5 Å in HKUST-1, ZJU-35a, and ZJU-36a, respectively, considering the influence of van der Waals radius. The BET measurements demonstrated that the surface areas of HKUST-1, ZJU-35a, and ZJU-36a are 1800, 2899, and 4014 m2 /g, respectively. From 2011 to 2016, Wang, Zhou, and Schröder teams reported three isoreticular crystals, named Guangdong Medical University (GDMU-2), University of Nottingham (NOTT-112), and PCN-69/NOTT-119, respectively [97, 99, 100, 135]. They are constructed from Cu clusters and ligands of 3,3′ ,3′′ ,5,5′ ,5′′ -pyridyl1,3,5-triylhexabenzoic acid (PHB), 1,3,5-tris(3′ ,5′ -dicarboxy[1,1′ -biphenyl]-4-yl-) benzene (BBCC), and 5,5′ ,5′′ -(benzene-1,3,5-triyl-tris(biphenyl-4,4′ -diyl)) triisophthalic acid (BTTI), and the BET surface area of them are 2758, 3800, and 3989/4118 m2 /g, respectively. Just as described above, introducing of ethynyl units in the structure can improve the surface area for MOFs with given network topology and similar ligand length. 14.5.2.2 Influence of Interpenetration

Interpenetration can be described as constructing an interpenetrated polymer comprising two or more networks are at least interlaced on a polymer scale but not covalently bonded to each other. As the most investigated types of entanglement, structural interpenetration (or catenation) is the common form in MOFs. MOFs with 2- to 54-fold interpenetration has been widely investigated till now. Generally speaking, the interpenetration in MOFs leads to decreasing of surface area through occupying void space, increasing density, and narrowing pore sizes. Though decreasing the empty space of the frameworks and improving intermolecular interactions

859

14 Relationship Between MOF Structures and Gas Absorption Properties

between frameworks, structural interpenetration can enhance the stability of the frameworks in MOFs through filling void space and constructing intermolecular weak interactions. In 2007, Long and coworkers investigated that MOF-5, reported by Yaghi and coworkers in 1999 [12], has a BET surface area of 3800 m2 /g after a full evaluation [92]. But there are different surface areas (700–3800 m2 /g) in reported papers that are shown the crucial role of preparing procedure of MOF-5. Lillerud and coworkers demonstrated that materials with low surface area exist two different classes of crystals through single-crystal X-ray diffraction (XRD) analysis [136]. One of two phases consisted of twofold interpenetrated MOF-5 networks, inducing empty space and surface area of the frameworks. In 2007, Zhou and coworkers introduced a pair of crystals with interpenetration and noninterpenetration constructed from Cu clusters and TATB ligands, named PCN-6 and PCN-6′ , respectively [133]. PCN-6 and PCN-6′ are the first interpenetration isomers with the same single net. PCN-6 can be represented by two identical interpenetrated nets of PCN-6′ . Interestingly, PCN-6 with interpenetration has a higher Langmuir surface area (3800 m2 /g) than PCN-6′ without interpenetration (2700 m2 /g), although PCN-6′ (86%) has a greater solvent-accessible volume than PCN-6 (74%) calculated by program PLATON (Figure 14.18). Herein, catenation led to 41% increasing of Langmuir surface area. Comparison with PCN-6′ , PCN-6 has smaller pores and larger adsorption sites from catenations, strengthening the overall interaction between gas molecules and the pore walls to increase the surface area. In 2011, Ma and coworkers reported 54-fold interpenetrated 3D MOF with the 103 -srs topology, constructed from Ag atoms and tri(4-imidazolylphenyl)amine (TIPA) ligands [137]. From a topological point of view, the structure consists of 54 interpenetrating nets with the three-connected 103 -srs topology (Figure 14.19). Because of the remarkable and unprecedented interpenetration, the potential solvent-accessible volume of this MOF only occupied 10.4% of total volume. But they did not test explore the surface area of this material. In 2012, Kim and coworkers reported two crystals constructed from Zn clusters and 2-nitrobenzene-1,4-dicarboxylate (NDC)–NO2 ligands, named MOF-123 and 2.0

1000 PCN-6

800 600

PCN-6ʹ

400 200 0 0.0

Hydrogen uptake (wt%)

1200 Vads/(cm3/g (STP))

860

1.5 1.0 0.5 PCN-6 PCN-6ʹ

0.0 0.2

0.4

0.6 P/P0

0.8

1.0

0

200

400

600

800

P(Torr)

Figure 14.18 Gas sorption isotherms (77 K) of PCN-6′ and PCN-6 activated at 50 ∘ C for N2 (a) and H2 (b). Source: From Zhao et al. [110]. © 2009, American Chemical Society.

14.5 Representative MOFs with Ultrahigh Surface Area and the Factors Effecting on Surface Area

(a)

(b)

Figure 14.19 (a) Interpenetration of four nets of one handedness (green/blue nets) with one of the opposite handedness (red). (b) Interpenetration of all 54 networks. Source: From Wu et al. [110]. © 2009, American Chemical Society.

MOF-246, respectively [138]. For MOF-123, the central Zn atom is a six-coordinated with 2-nitrobenzene-1,4-dicarboxylate (NDB) ligands and N,N-dimethylformamide (DMF) molecules. As for MOF-246, it was a four-coordinated between Zn atoms and ligands. In MOF-123, two DMF molecules are the terminal ligands on Zn clusters that protrude into the pores. The framework shows the fqr topology, which is in turn derived from decoration of the primitive cubic (pu) topology, being particularly favorable for the formation of interpenetrating structures. After heating MOF-123 to 320 ∘ C, MOF-246 with twofold interpenetrated networks can be achieved. The experimental BET surface area of the MOF-123 exhibits a pretty value (1200 m2 /g), while MOF-246 has shown only rare N2 uptake at 77 K and surface area. 14.5.2.3 Influence of Metal Moieties

Through adjusting synthetic conditions, we can get isoreticular MOFs when using the same ligands and different metals. In their structures, there are only a spot of differences, such as coordination bond strength and length between ligands and metal atoms. There are similar simulated powdery X-ray diffraction (PXRD) patterns owing to similar structures. On the one hand, different coordination bond strengths between ligands and metal clusters lead to different robustness of MOFs, influencing the activation of MOFs. On the other hand, different metal in structure influence the density of MOFs. Both of them influence the change of specific surface area. From 2005 to 2012, Yaghi, Dietzel, Matzger, and Long teams reported a series of isoreticular MOFs constructed by 2,5-dihydroxy-1,4-benzenedicarboxylic acid (DOBDC) ligands and different metal clusters, named M-MOF-74 (also called as coordination polymer of oslo [CPO]-27-M, M = Mg/Co/Fe/Ni/Zn) [139]. We know that surface area is defined as an area per given mass, and there is a trend that the lighter metal in a series of isostructural materials has a higher surface area without

861

862

14 Relationship Between MOF Structures and Gas Absorption Properties

considering other factors. The BET surface area of Mg-MOF-74 is 1495 m2 /g, which is higher than those of Co-MOF-74 (1080 m2 /g), Fe-MOF-74 (920 m2 /g), Ni (1070 m2 /g), and Zn-MOF-74 (816 m2 /g). Matzger teams then investigate the number of N2 molecules per UC of Mg/Co/Ni/Zn-MOF-74, showing two significant different results. For Mg/Co/Ni-MOF-74, every UC can load similar numbers (35–38) of molecules, indicating that there is not any specific interaction between the metal and the adsorbed gas and surface area differences mainly depend on variation in the density of the structure. As for Zn-MOF-74, every UC has a much lower loading value (26 molecules) caused by improper activation route bringing pore blockage and/or partial collapse of the structure. 14.5.2.4 Influence of Functionality

Generally speaking, for MOFs with fully evaluation, introducing functional group into MOFs will decrease the specific surface area of materials, owing to occupying free void, blocking part of pores and cages, and improving density. In this section, we give some examples to discuss the influence of functionality in MOFs. In 2002, Yaghi and coworkers reported seven isoreticular MOFs based on the MOF-5 structure from BDC with different functionalities (BDC, BDC-Br, BDC-NH2 , BDC-R4 , BDC-R5 , BDC-R6 , and BDC-R7 ), named isoreticular metal-organic framework (IRMOF-1) to IRMOF-7 (Figure 14.20) [140]. Through Table 14.4, we can find that IRMOF-1 has the largest free volume and free diameter among those crystals and except for IRMOF-2, free volumes decrease as increasing of molecule weight of ligands for other IRMOF while the tendency of crystal density shows an opposite appearance, indicating functionality of MOFs can occupy free space, decrease free volume, and improve density. Then they studied adsorption property of IRMOF-6. The result shows that IRMOF-6 has a Langmuir surface area of 2630 m2 /g, which is clearly lower than that of IRMOF-1 (3800 m2 /g), approving the conclusion mentioned above that functionality of the MOFs decrease surface area. In 2010, Lillerud and coworkers synthesized a family of isoreticular MOFs based on the UiO-66 structure from BDC, BDC-NH2 , BDC-NO2 , and BDC-Br, named UiO-66, UiO-66-NH2 , UiO-66-NO2 , and UiO-66-Br, which retain porosity despite the presence of different functional groups on the linkers and have the Langmuir surface areas of 1300, 1250, 856, 899 m2 /g, respectively [141]. This phenomenon can be

L IRMOF-1

L

HOOC

IRMOF-1

L

O

COOH

IRMOF-5

Br

L

HOOC

COOH

O COOH

HOOC

NH2

IRMOF-1

L

HOOC

IRMOF-1

L

HOOC

COOH

IRMOF-6

L

HOOC

COOH

IRMOF-7

L

HOOC

COOH

COOH

O

O

Figure 14.20 Single-crystal X-ray structures of IRMOF-1 to IRMOF-7, respectively. Color scheme is as follows: Zn clusters, blue polyhedral; O atoms, red; C atoms, black; Br atoms, green; N atoms, blue [140].

14.6 Adsorption Enthalpy of MOF Materials

Table 14.4 The comparison of molecular weight of ligand, the calculated percent free volume, crystal densities, free diameter, and fixed diameter for IRMOF-1 to IRMOF-7.

MOFs

Molecular weight of ligand

Free volume

Crystal density

Free diameter

Fixed diameter

IRMOF-1

166

79.2

0.61

11.2

18.5

IRMOF-2

245

78.4

0.78

7.5

16.4

IRMOF-3

181

78.7

0.63

9.6

18.6

IRMOF-4

282

64.6

0.86

5.4

14.1

IRMOF-5

338

55.8

1

3.8

12.8

IRMOF-6

192

77.5

0.65

9.3

18.6

IRMOF-7

216

76.6

0.71

5.5

13.6

ascribed to the result of both reducing available free space and increasing the overall weight of the new MOFs owing to introducing large and heavy functional groups.

14.6 Adsorption Enthalpy of MOF Materials 14.6.1 Adsorption Enthalpy Adsorption enthalpy (ΔH) or isosteric heat of adsorption (Qst , ΔH = −Qst ) is one of the most important thermodynamic quantities to understand the adsorption heat effects. It can be used to compare the strength of interaction of various adsorbents with adsorbates and evaluate the uniformity that the arrangement of the adsorbates on the porous surface of the adsorbents.

14.6.2 Determination of Adsorption Enthalpy The adsorption enthalpy is usually calculated by the Clausius–Clapeyron equation or obtained by microcalorimetry directly. Until now, several different methods were used to obtain the adsorption enthalpy of MOFs for gases as follows [142]. 14.6.2.1 Method 1

To fit the gas sorption data, Eq. (14.12) usually was a reference. And ai , bi , and T are independent of temperature. In Eq. (14.12), P, N, and T are followed by pressure, the amount of adsorbed gas, and temperature. Concluding, by use of Eq. (14.13), the values of the isosteric heat of adsorption (adsorption enthalpy) can be calculated, where R is the universal gas constant [143]. To estimate the values of the isosteric heat of adsorption (adsorption enthalpy), Eq. (14.13) is applied, where R is the universal gas constant [143]. ∑ 1∑ ai N i + bi N i T i=0 i=0 m

ln P = ln N +

m ∑ Qst = −R ai N i i=0

m

(14.12) (14.13)

863

864

14 Relationship Between MOF Structures and Gas Absorption Properties

14.6.2.2 Method 2

The following virial-type equation (Eq. (14.14)) is used to fit the adsorption data at a fixed temperature. ln ( N∕P) = A0 + A1 N + A2 N 2 + A3 N 3 + Qst = R ln

P1 T1 T2 P2 T2 − T1

(14.14) (14.15)

A function of surface coverage of the adsorbed Qst is calculated by using the Clausius–Krapperon equation (Eq. (14.15)), and R is equal to the gas constant. 14.6.2.3 Method 3

In order to fix the temperature, use Eq. (14.16) to calculate and fit. And in Eq. (14.17), N m is amount of adsorbed gas at saturation, B and t are the constants [144]. BP(1∕t) N = Nm 1 + BP(1∕t) ) ( N∕Nm P= B + BN∕Nm

(14.16) (14.17)

The Langmuir–Freundlich equation can be rearranged to Eq. (14.17), and the isosteric heat of adsorption is obtained by using Eq. (14.15).

14.6.3 Definition of Adsorption Amount and Usable Adsorption Amount There are a number of criterion for evaluating gas adsorption capacities. For example, in describing high-pressure adsorption capabilities, the terms “over”, ”total,” and “absolute” often appear. The abuse of these words may lead to unnecessary uncertainty in the comparison of gas adsorption [145]. At the basic level, gas will present the greater molecular density when it is adsorbed on the surface by gravity, compared with the free one under the constant temperature T and pressure P. When gas is adsorbed on a two-dimensional (2D) surface, the interaction strength between the gas molecules and the surface diminish with the raising of distance, until the surface’s gravitational force is negligible. At this distance, a hypothetical line, called the Gibbs interface, can be drawn and divided the total free volume into an adsorption zone from a free zone. The absolute amount of adsorption is simply defined as the total number of molecules in the adsorption zone (Figure 14.21) [146]. Regrettably, absolute adsorption cannot be measured directly because that is not practicable to empirically determine the location of the Gibbs interface or the size of the adsorption zone. Total adsorption, including all gas molecules in the adsorbent pore, is usually used as an approximation for absolute adsorption (Figure 14.21b). ntot = nex + Vp 𝜌bulk (P, T)

(14.18)

The total adsorption (ntot ) amount can be calculated from the adsorption amount (nex ) and the total pore volume V p measured experimentally (Eq. (14.18)). The pore

14.6 Adsorption Enthalpy of MOF Materials

Adsorbed Bulk +

(a)

=

Excess

Bulk

+

(b)

Excess

Absolute

=

Pore volume × bulk density

Total

Figure 14.21 As shown in (a), when adsorbed on a 2D surface (rectangular), there are two parts which can be seen, the green molecules are in an adsorbed state and the blue molecules are in a large amount status. And those two parts were divided by the Gibbs interface (the red line) from the free volume. The total number of gas molecules in the surface-free adsorption regions and excess adsorption measured in the experiment including the entire gas molecule in the adsorption state is called absolute adsorption. (b) For porous materials, all gas molecules in the pore volume of the material are called total adsorption, which includes excess adsorbed pore volume and unadsorbed pore volume. However, for micropores, the estimated absolute adsorption amount is considered to be the total adsorption amount. Source: From Mason et al. [145]. © 2014, Royal Society of chemistry. Licenced Under CC BY 3.0.

volume is usually conditioned by the N2 adsorption isotherm at 77 K, all wells are completely filled with condensed N2 at very large P/P0 [5a]. 14.6.3.1 The Definition of Usable Adsorption

As we know, the gas adsorption capacity of the materials can be compared by the total adsorption capacity, but for industrial applications, the total adsorption capacity cannot accurately reflect the working adsorption capacity of different materials. For example, it is useful to compare with the 35 bar CH4 capacity of different adsorbents for initial evaluation, but not all of these capacities are available in practice since gas containers are required to keep the lowest intake pressure to maintain the regular delivery of natural gas to an engine. Thus, the usable CH4 capacity [68] is defined as the amount of CH4 that can be transported when the filling or adsorption pressure is reduced to a specific desorption pressure (Figure 14.22) [56]. 14.6.3.2 Effect of Adsorption Enthalpy on Usable Adsorption

It is well known that with the increasing adsorption enthalpy, it takes more energy to release the gas from the porous materials in desorption route. It is different for usable adsorption which means adsorption at a given pressure for industrial requirement. If the adsorption enthalpy is too high, it means high capacities for gas storage,

865

14 Relationship Between MOF Structures and Gas Absorption Properties

Total CH4 adsorbed (cm3 (STP)/g)

500

AX-21

25 °C

MOF-5

400

HKUST-1 PCN-14 Mg2(dobdc)

300

Ni2(dobdc) Co2(dobdc)

200

100

0 0

20

40

60

80

100

P (bar)

(a)

Total CH4 adsorbed (cm3(STP)/cm3)

866

HKUST-1 Ni2(dobdc) PCN-14 Co2(dobdc) MOF-5 Mg2(dobdc) AX-21

25 °C 250 200 150

Pure CH4

100 50 0 0

(b)

20

40

60

80

100

P (bar)

Figure 14.22 (a) The volume of CH4 gas that can be released under the specified adsorption temperature, adsorption pressure, desorption temperature, and desorption pressure can be called usable capacity. (b) Because it is a function of adsorption temperature and desorption temperature, the volumetric usable CH4 capacity can be used for adsorption at 35 bar, and desorption at 5 bar and 25 ∘ C. Source: From Mason et al. [119]. © 2014, John Wiley & Sons. Licenced Under CC BY 3.0.

but high-energy consumption and ineffectiveness for gas desorption. When adsorption enthalpy is too low, it means low-performance for materials and only a small amount of gas can be adsorbed on material at given pressure. So materials with modest adsorption enthalpy are good candidates for unable adsorption for industrial requirement. The optimal adsorption enthalpy can guarantee the usable adsorption. It is worth noting that the optimal desorption conditions determine the optimal solution for the adsorption enthalpy. Figure 14.23 shows the percentage of saturated capacity available according to different adsorption enthalpy and desorption temperatures [147].

14.6 Adsorption Enthalpy of MOF Materials

% Usable of saturation capacity

80

35 bar adsorption 5 bar desorption 145 °C

60

40

Desorption temperature

Optimal Qst

20 25 °C 10

12

14

16

18

20

22

24

–Qst (kJ/mol)

Figure 14.23 Assuming a single-site Langmuir isotherm, the percentage of the saturation capacity that is usable is plotted for isosteric heats of adsorption, Qst , ranging from −10 to −25 kJ/mol and desorption temperatures from 25 to 145 ∘ C, with adsorption at 35 bar, desorption at 5 bar, and a molar entropy of adsorption of −9.5R. As the desorption temperature increases, the optimal Qst and usable capacity also increase. Source: From Mason et al. [145]. © 2014, Royal Society of chemistry. Licenced Under CC BY 3.0.

The graphs reveal that, when the desorption temperature increases or the desorption pressure decreases, the optimal adsorption enthalpy increases. Furthermore, it reveals that reach the climax of usable capacity is strikingly enclosed by higher desorption temperatures and optimized adsorption enthalpy [148].

14.6.4 Factors Affecting MOFs Gas Adsorption As a consequence, the different interactions between gas molecules and MOFs cause the same MOF materials to have different adsorption enthalpies and show various adsorption capacity for different gases, which reflected their abilities of selectivity in gas adsorption. For example, the main reason for the greenhouse effect is excessive carbon dioxide emissions, so how to selectively adsorb carbon dioxide gas has been an urgent problem. As a promising porous material for gas adsorption and separation, MOFs have been developed as great candidates for selective adsorption of carbon dioxide gas. Some representative MOFs and their CO2 uptake and gas separation capacities are shown in Table 14.5. MOFs containing different metal sites or different ligands with functional groups exhibit noticeable differences in gas adsorption property and selectivity. In the previous researches, the metal sites and the functional ligands had an impressive effect on adsorption of MOFs gas and play crucial roles. 14.6.4.1 MOFs Containing Open Metal Sites

As we all know, coal-fired power is an overwhelming majority of generating electricity in the four sides of the world nowadays, which could pay for the largest part

867

Table 14.5 CO2 capture and separation properties of selected metal–organic frameworks categorized as containing open metal sites, interpenetrated, flexible, and functionalized. Separation properties

Materials

Functionalized

Containing open metal sites

Flexible

CO2 uptake

Separation application, selectivity

References

[Zn4 O(btb)2 ] (MOF-177)

33.5 mmol/g (298 K, 32 bar)

CO2 /CH4

[149]

[Zn4 O(bdc)3 ] (MOF-5)

2.10 mmol/g, 9.24 wt% (295 K, 1 atm)

CO2 /CH4

[150]

[Zn4 O(NH2bdc)3 ]

—

CO2 /CH4

[150]

[Zn4 O(fma)3 ]

69 wt% CO2 versus 8.6 wt% CH4 (300 K, 28 bar)

CO2 /CH4

[150]

[Cu3 (btc)2 ]

0.9 mmol/g (298 K, 6 bar),

CO2 /CH4 , CO2 /N2

[150]

[Cr3 F(H2O)2O(btc)]3 (MIL-100)

18 mmol/g versus 7.5 mmol/g CH4 (303 K, 48.7 bar)

CO2 /CH4

[151]

[Cr3 F(H2 O)2 O(bdc)]3 (MIL-101)

40 mmol/g versus 12 mmol/g CH4 (303 K, 48.7 bar)

CO2 /CH4

[151]

[Zn(bdc)(bpy)0.5 ] (MOF-508b)

6 mmol/g, 26.0 wt% versus 5.5 wt% N2 and 3.2 wt% CH4 (303 K, 4.5 bar)

CO2 /CH4 , CO2 /N2

[152]

[Ni(cyclam)2 (mtb)]

2.53 mmol/g, 11.2 wt% (195 K, 1 atm)

CO2 /CH4 , CO2 /N2

[153]

[Co(F-pymo)2 ]

7 mmol/g (273 K, 20 bar)

CO2 /CH4

[153]

[Zn(F-pymo)2 ]

8 mmol/g (273 K, 20 bar)

CO2 /CH4

[153]

[(Ni2 L1)(bptc)] (ethyl-bridged)

9.3 wt% (298 K, 1 atm), 15 wt% (15 bar)

CO2 /CH4 , CO2 /H2 , CO2 /N2

[153]

[(Ni2 L2)(bptc)] (butyl-bridged)

0 wt% (298 K, 1 atm), 21 wt% (15 bar)

CO2 /CH4 , CO2 /H2 , CO2 /N2

[153]

14.6 Adsorption Enthalpy of MOF Materials

of carbon dioxide emissions in our earth. Generally, the effluent CO2 concentration was about less than 15% of flue gas at one atmosphere. Therefore, designing porous material such as MOFs for capturing low-concentration carbon dioxide has attracted widespread attention and research. Metal sites, as a basic brick of MOFs, are the first factor considered to design MOFs for gas adsorption. For example, Matzger and coworkers investigated the CO2 adsorption capacity of a series of isostructural MOFs composed of transition metals (Zn, Mg, Ni, Co) and 2,5-dihydroxytetraacetic acid as a ligand [138c]. According to X-ray single-crystal diffraction analysis, as-synthesized MOFs contained a typical honeycomb one-dimensional (1D) channels which are lined with H2 O or DMF coordinated to the transition metals (Figure 14.24a). After activation of the sample, the amount of unsaturated metal sites has been produced. Further, they mainly studied these isostructural MOFs’ ability of adsorption of CO2 and found that the amount of carbon dioxide adsorbed by Mg/DOBDC is the highest at 296 K among the four isostructural MOFs (Figure 14.24b–d). Furthermore, as shown in an inset of Figure 14.24d, owing to the increased ionic character of the Mg—O bond, the carbon dioxide uptake value of Mg/DOBDC was double the value of the other materials in the M/DOBDC (M = Zn, Mg, Ni, Co) series at 0.1 atm [154]. In brief, their work indicated metal sites or clusters played a crucial role in the CO2 uptake especially in the low concentration of CO2 uptake. 14.6.4.2 MOFs with Functional Organic Ligand

In the past decades, both the effect of organic ligand with functional groups/sites and the role of open metal sites in MOFs have been studied for gas storage gas separation. Although open metal sites have a crucial role in MOF’s gas adsorption, there

180 160

(a) 450

Uptake volume (ml/g)

140

(b)

Intensity (a.u)

400 350 300

100 80

120 100

60

250 40

200 150

Ni/DOBDC

100

Mg/DOBDC

50

Co/DOBDC

0

(c)

120

20

30

40 2θ (°)

50

60

70

80

Ni/DOBDC

60

Zn/DOBDC

40 20

Mg/DOBDC

20

0 0

0.025

0.05

0.075

0.1

0 0

Zn/DOBDC 10

Co/DOBDC

(d)

0.2

0.4

0.6

0.8

1

Pressure (atm)

Figure 14.24 (a) One-dimensional channel structure view of M/DOBDC, with solvent molecules removed. (b) The space-filling structure model of Mg/DOBDC unit cell pores (Mg, green; C, gray; O, red; H atoms and solvent molecules have been removed from the model). There are 12 CO2 molecules (blue molecule) in the channel of the unit cell, which refers to adsorption at 296 K and 0.1 atm. (c) The PXRD data of the M/DOBDC series show that the MOFs of this series are isomorphic. (d) Difference in CO2 adsorption isotherms of M/DOBDC series from 0 to 1 atm at 296 K, inset: enlarged view of low pressure area from 0 to 0.1 atm (filled, adsorption; open, desorption). Source: Reprinted with the permission from Caskey et al. [138c]. © 2008, American Chemical Society.

869

14 Relationship Between MOF Structures and Gas Absorption Properties

are some limitations in this application. For example, these open metal sites contribute significantly to the high methane storage capacity at a low pressure about 5 bar; therefore, the amount of adsorption can be improved by a functional organic group decorated in the framework. Two methods for functionalizing the pores of the NOTT-101 framework have been introduced by Banglin Chen and coworkers, consisting of synthesis of new functional ligands using Lewis basic nitrogen sites and incorporation of linkers with the different functionalities [154]. In short, the article shows that adding appropriate functional components to the MOF can significantly increase the storage capacity and working capacity of methane, which provides a new and promising method for increasing the adsorption of methane gas (Figure 14.25). During this year, a new microporous MOF University of Texas at San Antonio (UTSA)-100a was reported by their group [155], which contained a suitable pore/cage space that preferentially absorbed more acetylene than ethylene, owing to the functional amine groups on the pore/cage surface enhance their interaction with the acetylene molecule. According to the X-ray single-crystal diffraction analysis (Figure 14.26), UTSA-100 had a 3D framework, a diamond-shaped open zigzag nanochannel, and the amino- and tetrazole-functionalized walls operated in the c-direction. In addition, the breakthrough experiment showed that the separation of C2 H2 /C2 H4 (1 : 99, v/v) is the first example of porous materials (Figure 14.27), which makes it (UTSA-100a) a potential material for practical industrial ethylene purification applications. The results from Figure 14.25 show that these MOFs with functional groups/sites have a surprisingly high CH4 storage working capacity of ∼188–197 cm3 (STP)/cm3 . (a)

(b)

(c)

COOH HOOC

HOOC

HOOC

(A)

COOH HOOC

COOH HOOC

N N

COOH

N

COOH HOOC

COOH HOOC

H4L2

H4L1

(d)

COOH HOOC

N

H4L3

N

Enhanced volumetric CH4 uptake

257

UTSA-76a

250

ZJU-5a

240 NOTT-101a

237

230 0.8

(C)

0.9

1.0

1.1

L2

L3

L4

(B)

197 cm3/cm3 Enhanced CH4 working capacity

UTSA-76a

L1

COOH

H4L4

260 3 Total CH4 uptake [cm3(STP)/cm ]

870

181 cm3/cm3

1.2

1.3

NOTT-101

ZJU-5

UTSA-75

UTSA-76

Pore volume (cm3/g)

Figure 14.25 (A) Schematic structure of the organic ligands H4 L1 –H4 L4 that serve as linkers in NOTT-101 (a), ZJU-5 (b), UTSA-75 (c), and UTSA-76 (d), respectively. (B) Crystal structures of NOTT-101 (gray), ZJU-5 (green), UTSA-75 (yellow), and UTSA-76 (red). Spheres denote pores within the frameworks. (C) Total CH4 uptake capacities of NOTT-101a (gray), ZJU-5a (green), UTSA-75a (yellow), and UTSA-76a (red) at 65 bar and normal temperature. It shows that the absorption of methane can indeed be achieved by combining various functional groups into NOTT-101a. Source: From Li et al. [154]. © 2015, Royal Society of chemistry.

14.6 Adsorption Enthalpy of MOF Materials

(a)

(b)

a b (c)

c

(d)

Figure 14.26 (a) The coordination mode of ATBDC2- and Cu(II). (b) The framework topology of apo-type (3,6)-connected networks with Schlfli symbol. (c) A 1D diamond channel with a diameter of approximately 4.3 Å can be seen in the direction of the c-axis. (d) The cage has a diameter of approximately 4.0 Å and a window opening of 3.3 Å. Color scheme: Cu, bright green; O, red; N, blue; C, orchid; and H, yellow. Source: From Hu et al. [155]. © 2015, Springer Nature. Licensed Under CC BY 4.0.

Further, the computational analysis demonstrates that the significant increase in methane volume storage and working capacity of functionalized MOFs is mainly due to the dynamic freedom of Lewis basic nitrogen sites and functionalized linking groups. Nowadays, a new microporous MOF UTSA-100a was reported by their research group [155], since the functional amine groups on the pore/cage surface enhance their interaction with acetylene molecules, it contains suitable pore/cage spaces which tend to absorb more acetylene in ethylene and acetylene systems. According to the X-ray single-crystal diffraction analysis (Figure 14.26), UTSA-100 had a 3D framework, a diamond-shaped opening zigzag nanochannel, and there are functionalized walls of amino and tetrazole along c-direction. In addition, the breakthrough experiment showed that the separation of ethylene/acetylene (1 : 99, v/v) is the first example of porous materials (Figure 14.27), which makes it (UTSA-100a) a potential material for practical industrial ethylene purification applications.

14.6.5 Effect of the Guest Molecules into the MOFs on Gas Adsorption Although open metal active sites and functional organic ligands can improve the gas adsorption capacity of MOFs, there are other methods, such as introducing

871

14 Relationship Between MOF Structures and Gas Absorption Properties 2

0

10

10–1 –2

10

C2H2 C2H4 0

50

100

150

140

1000

100

10

200

Dimensionless time, τ=t u / 𝜺L

(a)

PpmC2H2 at outlet of adsorber

1

10

10–3

(c)

10 000

100

150

200

0.8

MgMOF-74

0.6 FeMOF-74

CoMOF-74 80

60

80

0.4 0.2

, M MOF-3a NOTT-300 60

50

Dimensionless time, τ = t u / 𝜺L

1.0

UTSA-100a

120

100

0

(b)

F/F0

Concentration at outlet, ci / mol/m3

10

C2H2 captured during 0 – τbreak (mmol/l)

C2H4 C2H2

0.0 120

100

Dimensionless break through time, τbreak

0

5

10

15

20

25

Time (min)

(d)

Figure 14.27 (a) The curve of transient breakthrough with UTSA-100a at ethylene and acetylene mixture systems (containing 1% C2 H2 ). (b) The data of the outlet gas of an adsorber bed which is packed with various MOFs: M’MOF-3a (blue), MgMOF-74 (pink), CoMOF-74 (green), FeMOF-74 (black dash), NOTT-300 (black), and UTSA-100a (red). (c) Data of C2 H2 uptake. (d) The curve of experimental column breakthrough of UTSA-100a at ethylene and acetylene mixture systems (containing 1% C2 H2 ) at 296 K (FeMOF-74 at 318 K and NOTT-300 at 293 K). Source: From Hu et al. [155]. © 2015, Springer Nature. Licensed Under CC BY 4.0. 3.00 PEI-silica Diamine-silica

2.50

Adsorption capacity (mmol/g)

872

Mg/DOBDC ED–Mg/DOBDC

2.00

1.50

1.00

0.50

0.00 0

(a)

(b)

1

2

3

4

5

Cycle number

Figure 14.28 (a) From the cell structures of Mg/DOBDC, there are six EDs in per unit cell. (b) The change of CO2 adsorption amount of different composite adsorbents with the number of adsorption–desorption cycles. Source: From Choi et al. [156]. © 2012, American Chemical Society.

References

guests into the rich pores or cages of MOFs, to affect the adsorption of MOFs gas. For instance, Christopher W. Jones’s group has introduced ethylene diamine (ED) into the MOF (Mg/DOBDC) micropores by binding ED molecules into open metal coordination sites [156]. As shown in Figure 14.28, although the conventional impregnated amine adsorbent (polyethylenimine [PEI]/silica) has the highest adsorption capacity in the first cycle, there is a significant loss in its adsorption capacity in continuous operation. However, ED–Mg/DOBDC still has an amazing CO2 adsorption capacity under the super-diluted CO2 partial pressure, increases the stability/regeneration at the same time, and can maintain excellent gas adsorption after four cycles.

Acknowledgments The authors wish to thank the following scholars from the Liu research group, who contributed selflessly to proofreading of the manuscript: Li Tao, Liu Bai-Tong, Zhang Xiao-Meng, Huang Ge, Zhang An-An, Li Lan, and Hu Xiao-Jing (equal contribution).

References 1 Marcus, Y. (1991). J. Chem. Soc., Faraday Trans. 87: 2995–2999. 2 Krause, S., Bon, V., Senkovska, I. et al. (2016). Nature 532: 348–352. 3 Koh, K., Wong-Foy, A.G., and Matzger, A.J. (2009). J. Am. Chem. Soc. 131: 4184–4185. 4 Myers, A.L. and Monson, P.A. (2002). Langmuir 18: 10261–10273. 5 (a) Sing, K.S.W., Everett, D.H., Haul, R.A.W. et al. (1985). Pure Appl. Chem. 57: 603–619. (b) Sing, K.S.W. (1982). Pure Appl. Chem. 54: 2201–2218. 6 Langmuir, I. (1932). Nature 130, 768-768. 7 Haynie, D.T., Cho, E., and Waduge, P. (2011). Langmuir 27: 5700–5704. 8 Levan, M.D. and Vermeulen, T. (1981). J. Phys. Chem. 85: 3247–3250. 9 Brunauer, S., Emmett, P.H., and Teller, E. (1938). J. Am. Chem. Soc. 60: 309–319. 10 Davis, B.H. (2017). The Posthumous Nobel Prize in Chemistry, Volume 1. Correcting the Errors and Oversights of the Nobel Prize Committee, vol. 1262, 165–206. Washington, DC, USA: ACS. 11 Li, W., Xia, X., Cao, M., and Li, S. (2019). J. Mater. Chem. A 7: 7470–7479. 12 Li, H., Eddaoudi, M., O’Keeffe, M., and Yaghi, O.M. (1999). Nature 402: 276–279. 13 Cavka, J.H., Jakobsen, S., Olsbye, U. et al. (2008). J. Am. Chem. Soc. 130: 13850–13851. 14 Wiersum, A.D., Soubeyrand-Lenoir, E., Yang, Q. et al. (2011). Chem. Asian J. 6: 3270–3280. 15 Furukawa, H., Gandara, F., Zhang, Y.-B. et al. (2014). J. Am. Chem. Soc. 136: 4369–4381.

873

874

14 Relationship Between MOF Structures and Gas Absorption Properties

16 Feng, D., Gu, Z.-Y., Li, J.-R. et al. (2012). Angew. Chem. Int. Ed. 51: 10307–10310. 17 Furukawa, H., Ko, N., Go, Y.B. et al. (2010). Science 329: 424–428. 18 Zhao, X., Pachfule, P., Li, S. et al. (2019). J. Am. Chem. Soc. 141: 6623–6630. 19 Shen, K., Zhang, L., Chen, X. et al. (2018). Science 359: 206–210. 20 Taylor, M.K., Runcevski, T., Oktawiec, J. et al. (2016). J. Am. Chem. Soc. 138: 15019–15026. 21 Ferey, G., Mellot-Draznieks, C., Serre, C. et al. (2005). Science 309: 2040–2042. 22 Horcajada, P., Serre, C., Vallet-Regi, M. et al. (2006). Angew. Chem. Int. Ed. 45: 5974–5978. 23 Klimakow, M., Klobes, P., Rademann, K., and Emmerling, F. (2012). Microporous Mesoporous Mater. 154: 113–118. 24 Howarth, A.J., Peters, A.W., Vermeulen, N.A. et al. (2017). Chem. Mater. 29: 26–39. 25 Mittemeijer, E.J. (2010). Fundamentals of Materials Science, Chapter 5, 201–244. Berlin, Heidelberg: Springer-Verlag. 26 Park, J., Wang, Z.U., Sun, L.-B. et al. (2012). J. Am. Chem. Soc. 134: 20110–20116. 27 Shoaee, M., Anderson, M.W., and Attfield, M.P. (2008). Angew. Chem. Int. Ed. 47: 8525–8528. 28 Bordiga, S., Regli, L., Bonino, F. et al. (2007). Phys. Chem. Chem. Phys. 9: 2676–2685. 29 Polozij, M., Rubes, M., Cejka, J., and Nachtigall, P. (2014). ChemCatChem 6: 2821–2824. 30 Petkov, P.St., Vayssilov, G.N., Liu, J. et al. (2012). ChemPhysChem 13: 2025–2029. 31 (a) Fang, Z., îrholt, J.P., Kauer, M. et al. (2014). J. Am. Chem. Soc. 136: 9627–9636. (b) Marx, S., Kleist, W., and Baiker, A. (2011). J. Catal. 281: 76–87. 32 Barin, G., Krungleviciute, V., Gutov, O. et al. (2014). Inorg. Chem. 53: 6914–6919. 33 Kozachuk, O., Luz, I., Llabrés i Xamena, F.X. et al. (2014). Angew. Chem. In. Ed. 53: 7058–7062. 34 Ameloot, R., Vermoortele, F., Hofkens, J. et al. (2013). Angew. Chem. Int. Ed. 52: 401–405. 35 Choi, J.-S., Son, W.-J., Kim, J., and Ahn, W.-S. (2008). Microporous Mesoporous Mater. 116: 727–731. 36 Cubillas, P., Anderson, M.W., and Attfield, M.P. (2012). Chem. Eur. J. 18: 15406–15415. 37 Liu, M., Wong-Foy, A.G., Vallery, R.S. et al. (2010). Adv. Mater. 22: 1598–1601. 38 (a) C. S. Tsao, M.-S. Yu, C.-Y. Wang, P.-Y. Liao, H.-L. Chen, U.-S. Jeng, Y.-R. Tzeng, T.-Y. Chung, H.-C. Wu, J. Am. Chem. Soc. 2009, 131, 1404-1406; (b) Li, Y.W. and Yang, R.T. (2006). J. Am. Chem. Soc. 128: 8136–8137. (c) C. S. Tsao, M.-S. Yu, T.-Y. Chung, H.-C. Wu, C.-Y. Wang, K.-S. Chang, H.-L. Chen, J. Am. Chem. Soc. 2007, 129, 15997–16004; (d) Ravon, U., Domine, M.E., Gaudillere, C. et al. (2008). New J. Chem. 32: 937–940.

References

39 (a) Ravon, U., Savonnet, M., Aguado, S. et al. (2010). Microporous Mesoporous Mater. 129: 319–329. (b) Llabres i Xamena, F.X., Cirujano, F.G., and Corma, A. (2012). Microporous Mesoporous Mater. 157: 112–117. 40 Park, T.-H., Hickman, A.J., Koh, K. et al. (2011). J. Am. Chem. Soc. 133: 20138–20141. 41 Choi, K.M., Jeon, H.J., Kang, J.K., and Yaghi, O.M. (2011). J. Am. Chem. Soc. 133: 11920–11923. 42 (a) Huang, B.L., McGaughey, A.J.H., and Kaviany, M. (2007). Int. J. Heat Mass Transfer 50: 393–404. (b) Huang, B.L., Ni, Z., Millward, A. et al. (2007). Int. J. Heat Mass Transfer 50: 405–411. 43 Gadipelli, S. and Guo, Z. (2014). Chem. Mater. 26: 6333–6338. 44 Tu, B., Pang, Q., Wu, D. et al. (2014). J. Am. Chem. Soc. 136: 14465–14471. 45 (a) Llabres i Xamena, F.X., Cirujano, F.G., and Corma, A. (2012). Microporous Mesoporous Mater. 157: 112–117. (b) Øien, S., Wragg, D., Reinsch, H. et al. (2014). Cryst. Growth Des. 14: 5370–5372. 46 Shearer, G.C., Chavan, S., Ethiraj, J. et al. (2014). Chem. Mater. 26: 4068–4071. 47 Vermoortele, F., Vandichel, M., Van de Voorde, B. et al. (2012). Angew. Chem. Int. Ed. 51: 4887–4890. 48 Wu, H., Chua, Y.S., Krungleviciute, V. et al. (2013). J. Am. Chem. Soc. 135: 10525–10532. 49 (a) Vermoortele, F., Bueken, B., Le Bars, G. et al. (2013). J. Am. Chem. Soc. 135: 11465–11468. (b) Vandichel, M., Hajek, J., Vermoortele, F. et al. (2015). CrystEngComm 17: 395–406. 50 Cliffe, M.J., Wan, W., Zou, X. et al. (2014). Nat. Commun. 5: 4176. 51 Bueken, B., Reinsch, H., Reimer, N. et al. (2014). Chem. Commun. 50: 10055–10058. 52 Guillerm, V., Ragon, F., Dan-Hardi, M. et al. (2012). Angew. Chem. Int. Ed. 51: 9267–9271. 53 (a) Chizallet, C., Lazare, S., Bazer-Bachi, D. et al. (2010). J. Am. Chem. Soc. 132: 12365–12377. (b) Nguyen, L.T.L., Le, K.K.A., Truong, H.X., and Phan, N.T.S. (2012). Catal. Sci. Technol. 2: 521–528. 54 (a) Leus, K., Muylaert, I., Vandichel, M. et al. (2010). Chem. Commun. 46: 5085–5087. (b) Leus, K., Vandichel, M., Liu, Y.-Y. et al. (2012). J. Catal. 285: 196–207. 55 Kozachuk, O., Meilikhov, M., Yusenko, K. et al. (2013). Eur. J. Inorg. Chem. 2013: 4546–4557. 56 Vermoortele, F., Ameloot, R., Alaerts, L. et al. (2012). J. Mater. Chem. 22: 10313–10321. 57 Feyand, M., Mugnaioli, E., Vermoortele, F. et al. (2012). Angew. Chem. Int. Ed. 51: 10373–10376. 58 Yang, S., Lin, X., Lewis, W. et al. (2012). Nat. Mater. 11: 710–716. 59 Shen, L., Yang, S.-W., Xiang, S. et al. (2012). J. Am. Chem. Soc. 134: 17286–17290. 60 Valvekens, P., Jonckheere, D., De Baerdemaeker, T. et al. (2014). Chem. Sci. 5: 4517–4524.

875

876

14 Relationship Between MOF Structures and Gas Absorption Properties

61 Ishizaki, M., Akiba, S., Ohtani, A. et al. (2013). Dalton Trans. 42: 16049–16055. 62 (a) Carlucci, L., Ciani, G., Moret, M. et al. (2000). Angew. Chem. Int. Ed. 39: 1506–1510. (b) Moret, M. and Rizzato, S. (2009). Cryst. Growth Des. 9: 5035–5042. 63 Whittington, C.L., Wojtas, L., and Larsen, R.W. (2014). Inorg. Chem. 53: 160–166. 64 Fang, Z., Bueken, B., Vo De Vos, D.E., and Fischer, R.A. (2015). Angew. Chem. Int. Ed. 54: 7234–7254. 65 Hou, X.-J., He, P., Li, H., and Wang, X. (2013). J. Phys. Chem. C 117: 2824–2834. 66 Cai, G. and Jiang, H.L. (2017). Angew. Chem. Int. Ed. 56: 563–567. 67 Zhang, Y., Zhang, X., Lyu, J. et al. (2018). J. Am. Chem. Soc. 140: 11179–11183. 68 Li, H., Wang, K., Sun, Y. et al. (2018). Mater. Today 21: 108–121. 69 Farha, O.K., Eryazici, I., Jeong, N.C. et al. (2012). J. Am. Chem. Soc. 134: 15016–15021. 70 Schnobrich, J.K., Koh, K., Sura, K.N., and Matzger, A.J. (2010). Langmuir 26: 5808–5814. 71 (a) Gardner, G.B., Venkataraman, D., Moore, J.S., and Lee, S. (1995). Nature 374: 792–795. (b) Moghadam, P.Z., Li, A., Wiggin, S.B. et al. (2017). Chem. Mater. 29: 2618–2625. 72 Duren, T., Millange, F., Ferey, G. et al. (2007). J. Phys. Chem. C 111: 15350–15356. 73 Alezi, D., Belmabkhout, Y., Suyetin, M. et al. (2015). J. Am. Chem. Soc. 137: 13308–13318. 74 Sumida, K., Hill, M.R., Horike, S. et al. (2009). J. Am. Chem. Soc. 131: 15120–15121. 75 Wang, B., Lv, X.L., Feng, D. et al. (2016). J. Am. Chem. Soc. 138: 6204–6216. 76 Lv, X.-L., Tong, M., Huang, H. et al. (2015). J. Solid State Chem. 223: 104–108. 77 Seki, K. (2001). Chem. Commun.: 1496–1497. 78 Zheng, B., Yang, Z., Bai, J. et al. (2012). Chem. Commun. 48: 7025–7027. 79 Nouar, F., Eubank, J.F., Bousquet, T. et al. (2008). J. Am. Chem. Soc. 130: 1833–1835. 80 Klein, N., Senkovska, I., Gedrich, K. et al. (2009). Angew. Chem. Int. Ed. 48: 9954–9957. 81 Grunker, R., Senkovska, I., Biedermann, R. et al. (2011). Chem. Commun. 47: 490–492. 82 Klein, N., Senkovska, I., Baburin, I.A. et al. (2011). Chem. Eur. J. 17: 13007–13016. 83 Grunker, R., Bon, V., Heerwig, A. et al. (2012). Chem. Eur. J. 18: 13299–13303. 84 Stoeck, U., Krause, S., Bon, V. et al. (2012). Chem. Commun. 48: 10841–10843. 85 Pang, J., Jiang, F., Wu, M. et al. (2014). Chem. Commun. 50: 2834–2836. 86 Zheng, J., Wu, M., Jiang, F. et al. (2015). Chem. Sci. 6: 3466–3470. 87 Feldblyum, J.I., Wong-Foy, A.G., and Matzger, A.J. (2012). Chem. Commun. 48: 9828–9830. 88 Rowsell, J.L. and Yaghi, O.M. (2006). J. Am. Chem. Soc. 128: 1304–1315.

References

89 McDonald, T.M., Lee, W.R., Mason, J.A. et al. (2012). J. Am. Chem. Soc. 134: 7056–7065. 90 Férey, G., Serre, C., Mellot-Draznieks, C. et al. (2004). Angew. Chem. Int. Ed. 116: 6456–6461. 91 Horcajada, P., Surble, S., Serre, C. et al. (2007). Chem. Commun.: 2820–2822. 92 Kaye, S.S., Dailly, A., Yaghi, O.M., and Long, J.R. (2007). J. Am. Chem. Soc. 129: 14176–14177. 93 Ojha, R.P., Lemieux, P.A., Dixon, P.K. et al. (2004). Nature 427: 521–523. 94 Lu, Z., Du, L., Tang, K., and Bai, J. (2013). Cryst. Growth Des. 13: 2252–2255. 95 He, Y., Zhou, W., Yildirim, T., and Chen, B. (2013). Energy Environ. Sci. 6: 2735–2744. 96 Yang, S., Lin, X., Dailly, A. et al. (2009). Chem. Eur. J. 15: 4829–4835. 97 Yan, Y., Lin, X., Yang, S. et al. (2009). Chem. Commun.: 1025–1027. 98 Yan, Y., Telepeni, I., Yang, S. et al. (2010). J. Am. Chem. Soc. 132: 4092–4094. 99 Yan, Y., Yang, S., Blake, A.J. et al. (2011). Chem. Commun. 47: 9995–9997. 100 Yuan, D., Zhao, D., and Zhou, H.C. (2011). Inorg. Chem. 50: 10528–10530. 101 Farha, O.K., Yazaydin, A.O., Eryazici, I. et al. (2010). Nat. Chem. 2: 944–948. 102 Farha, O.K., Wilmer, C.E., Eryazici, I. et al. (2012). J. Am. Chem. Soc. 134: 9860–9863. 103 Wilmer, C.E., Farha, O.K., Yildirim, T. et al. (2013). Energy Environ. Sci. 6: 1158–1163. 104 Barin, G., Krungleviciute, V., Gomez-Gualdron, D.A. et al. (2014). Chem. Mater. 26: 1912–1917. 105 Gutov, O.V., Bury, W., Gomez-Gualdron, D.A. et al. (2014). Chem. Eur. J. 20: 12389–12393. 106 Wang, T.C., Bury, W., Gomez-Gualdron, D.A. et al. (2015). J. Am. Chem. Soc. 137: 3585–3591. 107 Kim, J., Yang, S.-T., Choi, S.B. et al. (2011). J. Mater. Chem. 21: 3070–3076. 108 Wang, X.S., Ma, S., Yuan, D. et al. (2009). Inorg. Chem. 48: 7519–7521. 109 Jiang, H.L., Feng, D., Liu, T.F. et al. (2012). J. Am. Chem. Soc. 134: 14690–14693. 110 Zhao, D., Yuan, D., Sun, D., and Zhou, H.C. (2009). J. Am. Chem. Soc. 131: 9186–9188. 111 Yuan, D., Zhao, D., Sun, D., and Zhou, H.C. (2010). Angew. Chem. Int. Ed. 49: 5357–5361. 112 Lu, W., Yuan, D., Makal, T.A. et al. (2012). Angew. Chem. Int. Ed. 51: 1580–1584. 113 Majumder, M., Buckton, G., Rawlinson-Malone, C.F. et al. (2013). CrystEngComm 15: 4041–4044. 114 Wei, Z., Gu, Z.Y., Arvapally, R.K. et al. (2014). J. Am. Chem. Soc. 136: 8269–8276. 115 Stewart, L., Lu, W., Wei, Z.W. et al. (2017). Dalton Trans. 46: 14270–14276. 116 Liu, T.F., Feng, D., Chen, Y.P. et al. (2015). J. Am. Chem. Soc. 137: 413–419. 117 Feng, D., Liu, T.F., Su, J. et al. (2015). Nat. Commun. 6: 5979–5987. 118 Liu, T.F., Zou, L., Feng, D. et al. (2014). J. Am. Chem. Soc. 136: 7813–7816.

877

878

14 Relationship Between MOF Structures and Gas Absorption Properties

119 Zhang, M., Chen, Y.P., Bosch, M. et al. (2014). Angew. Chem. Int. Ed. 53: 815–818. 120 Lian, X., Chen, Y.P., Liu, T.F., and Zhou, H.C. (2016). Chem. Sci. 7: 6969–6973. 121 Prasad, T.K. and Suh, M.P. (2012). Chem. Eur. J. 18: 8673–8680. 122 Park, H.J., Lim, D.W., Yang, W.S. et al. (2011). Chem. Eur. J. 17: 7251–7260. 123 Mu, B., Schoenecker, P.M., and Walton, K.S. (2010). J. Phys. Chem. C 114: 6464–6471. 124 He, Y., Guo, Z., Xiang, S. et al. (2013). Inorg. Chem. 52: 11580–11584. 125 Li, B., Wen, H.M., Wang, H. et al. (2014). J. Am. Chem. Soc. 136: 6207–6210. 126 Wen, H.M., Li, B., Li, L. et al. (2018). Adv. Mater. 30: 1704792. 127 Rao, X., Cai, J., Yu, J. et al. (2013). Chem. Commun. 49: 6719–6721. 128 Cai, J., Rao, X., He, Y. et al. (2014). Chem. Commun. 50: 1552–1554. 129 Kong, G.Q., Han, Z.D., He, Y. et al. (2013). Chem. Eur. J. 19: 14886–14894. 130 Schaate, A., Dühnen, S., Platz, G. et al. (2012). Eur. J. Inorg. Chem. 2012: 790–796. 131 Kalidindi, S.B., Nayak, S., Briggs, M.E. et al. (2015). Angew. Chem. Int. Ed. 54: 221–226. 132 Li, P., Vermeulen, N.A., Malliakas, C.D. et al. (2017). Science 356: 624–627. 133 Ma, S., Sun, D., Ambrogio, M. et al. (2007). J. Am. Chem. Soc. 129: 1858–1859. 134 Chui, S.S., Lo, S.M., Charmant, J.P. et al. (1999). Science 283: 1148–1150. 135 Liu, J., Liu, G., Gu, C. et al. (2016). J. Mater. Chem. A 4: 11630–11634. 136 Hafizovic, J., Bjorgen, M., Olsbye, U. et al. (2007). J. Am. Chem. Soc. 129: 3612–3620. 137 Wu, H., Yang, J., Su, Z.M. et al. (2011). J. Am. Chem. Soc. 133: 11406–11409. 138 Choi, S.B., Furukawa, H., Nam, H.J. et al. (2012). Angew. Chem. Int. Ed. 51: 8791–8795. 139 (a) Dietzel, P.D., Morita, Y., Blom, R., and Fjellvag, H. (2005). Angew. Chem. Int. Ed. 44: 6354–6358. (b) Bloch, E.D., Queen, W.L., Krishna, R. et al. (2012). Science 335: 1606–1610. (c) Caskey, S.R., Wong-Foy, A.G., and Matzger, A.J. (2008). J. Am. Chem. Soc. 130: 10870–10871. (d) Dietzel, P.D., Panella, B., Hirscher, M. et al. (2006). Chem. Commun.: 959–961. (e) Bhattacharjee, S., Choi, J.S., Yang, S.T. et al. (2010). J. Nanosci. Nanotechnol. 10: 135–141. 140 Eddaoudi, M., Kim, J., Rosi, N. et al. (2002). Science 295: 469–472. 141 Kandiah, M., Nilsen, M.H., Usseglio, S. et al. (2010). Chem. Mater. 22: 6632–6640. 142 Suh, M.P., Park, H.J., Prasad, T.K., and Lim, D.W. (2012). Chem. Rev. 112: 782–835. 143 Dinca, M., Dailly, A., Liu, Y. et al. (2006). J. Am. Chem. Soc. 128: 16876–16883. 144 Rege, S.U. and Yang, R.T. (1997). Ind. Eng. Chem. Res. 36: 5358–5365. 145 Mason, J.A., Veenstra, M., and Long, J.R. (2014). Chem. Sci. 5: 32–51. 146 Gibbs, J.W. and Longley, W.R. (1957). The Collected Works of J. Willard Gibbs. London: Yale University Press. 147 Bhatia, S.K. and Myers, A.L. (2006). Langmuir 22: 1688–1700. 148 Garrone, E., Bonelli, B., and Areán, C.O. (2008). Chem. Phys. Lett. 456: 68–70. 149 Millward, A.R. and Yaghi, O.M. (2005). J. Am. Chem. Soc. 127: 17998–17999.

References

150 Xue, M., Liu, Y., Schaffino, R.M. et al. (2009). Inorg. Chem. 48: 4649–4651. 151 Llewellyn, P.L., Bourrelly, S., Serre, C. et al. (2008). Langmuir 24: 7245–7250. 152 Bastin, L., Barcia, P.S., Hurtado, E.J. et al. (2008). J. Phys. Chem. C 112: 1575–1581. 153 Galli, S., Masciocchi, N., Tagliabue, G. et al. (2008). Chem. Eur. J. 14: 9890–9901. 154 Li, B., Wen, H.-M., Wang, H. et al. (2015). Energy Environ. Sci. 8: 2504–2511. 155 Hu, T.-L., Wang, H., Li, B. et al. (2015). Nat. Commun. 6: 7328–7336. 156 Choi, S., Watanabe, T., Bae, T.-H. et al. (2012). J. Phys. Chem. Lett. 3: 1136–1141.

879

881

15 Relationship Between Structure and Separation Property Zhanfeng Ju 1 , El-Sayed M. El-Sayed 1,2,3 , and Daqiang Yuan 1,2 1 Chinese Academy of Sciences, State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, 155 Yangqiao Road West, Fujian, Fuzhou, 350002, China 2 University of the Chinese Academy of Sciences, No. 19A Yuquan Road, Beijing, 100049, P.R. China 3 Egyptian Petroleum Research Institute, Chemical Refining Laboratory, Refining Department, Ahmed El-Zomor Street - El Zohour Region, Nasr City, Cairo, 11727, Egypt

15.1 Introduction Separation has been a significant, old technology since ancient times in human development and life. The production of dyes, drugs, and natural metals, for example, has been carried out for thousands of years, encouraging human civilization and growth in a significant way. Separation is now a significant process in several applications, such as petroleum refining, mining, the manufacture of chemicals and pharmaceutical engineering, etc. Meanwhile, separation processes represent a significant proportion of global energy consumption, around 15%. The explanation is that the distillation is responsible for 90–95% of all separations in the chemical and petroleum refining industries. In particular, different molecules of industrial interest (olefins, paraffins, alkanes, rare gases, etc.), due to similarities in physicochemical properties of mixed molecules, are being extracted using (cryogenic) distillation. Therefore, alternative separation approaches with low energy demand and a better economy should be examined. In addition, the right emphasis is the separated molecules, which are typically different in size, shape, functionality, and polarizability. Such molecules can, therefore, be distinguished rationally on the basis of their structures and chemistry. In this regard, exceptional research efforts are focused on the use of adsorption-based solutions and the employment of porous materials as separation agents. It should be pointed out that natural and/or synthesized zeolites are the most industrially used porous materials. Inorganic components of Si and/or Al tetrahedral metal ions, bridged by oxygen atoms, are typically highly thermally and chemically stable. Due to their intrinsic structural characteristics, however, small pore-based zeolites typically discriminate between molecules with an average difference of a minimum of 1 Å, which are not capable of separating a similar size Advanced Structural Chemistry: Tailoring Properties of Inorganic Materials and their Applications, First Edition. Edited by Rong Cao. © 2021 WILEY-VCH GmbH. Published 2021 by WILEY-VCH GmbH.

882

15 Relationship Between Structure and Separation Property

for the majority of molecules. Porous organic adsorbents, including polymers and carbons, were evaluated as potential separating solid candidates for gas separation because of their simple manufacturing and relatively high stability. However, their spectrum of separation is somewhat limited by the wide distribution of the pores, which restricts their efficient use for separating molecules with very identical physical and chemical characteristics. Fortunately, for the past two decades, metal–organic frameworks (MOFs) with the required on-demand fine-tuning of the porous system (pore-aperture size, shape, and functionality) have been introduced as prospective materials to overcome these shortages and offer opportunities to tackle problems related to painful separations. MOFs are constructed by metal ions/clusters (secondary building units, SBUs) and organic linkers. This class of hybrid crystalline materials, with their many options of SBUs and connectors, provides an unparalleled range of structural and chemical diversity for the design and engineering of pore systems using various efficient design/assembly approaches. The MOF pore size extends from a few angstroms to a few nanometers, providing additional prospects for a vast array of separation applications. Furthermore, the definite structures and functionalities of the pore arising from SBUs and/or organic linkers of MOFs offer more microscopic mechanisms for separation, such as thermodynamic (enthalpy) separation, conformational-driven separation, kinetic separation, molecular sieving, etc. Obviously, MOFs not only exhibit potential powerful abilities in separations but also offer a unique platform to investigate the relationship between structure and separation. In this chapter, we will introduce CO2 capture and separation, separation of hydrocarbons, separation of noble gases, separation of hydrogen isotopes, and enantioselective separation based on MOFs. Meanwhile, the relationship between structure and separation during these processes will be discussed in detail.

15.2 CO2 Capture and Separation In the process of human industrialization, global warming has attracted more and more attention, and a larger emission of CO2 is considered as the main factor. Consequently, in the last decade, CO2 capture and separation have received much research focus. In industry, liquid amines are primarily employed as separating agents for CO2 capture and separation. However, tremendous energy is consumed during this process, and separation agents with less-energy consumption are demanded. MOFs, in particular, have drawn much interest to overcome this shortage due to their defined and facile tunable structure in CO2 capture and separation in recent years. Post-combustion flue gases are usually 15% CO2 and 85% N2 at power plants, which are essential sources of emissions of CO2 . The total worldwide disposal of CO2 from flue gas emissions from primary stationary sources is nearly 60% above the average proportion of other CO2 emission sources, including the transport and industrial uses. Hence, the separation between CO2 and N2 is particularly vital in practical, which is the critical problem in CO2 capture and separation. Many

15.2 CO2 Capture and Separation

strategies have been reported to enhance the capture and separation of CO2 from a mixture of CO2 and N2 . Here, we introduce some studies listed by diverse strategies.

15.2.1 Utilizing Open Metal Sites In specific MOFs, H2 O or other solvent molecules coordinate to the metal atoms partially, and unsaturated metal sites within MOF pores are generated when the coordinated solvents are eliminated. Such open metal sites are considered as Lewis acid sites where CO2 molecules are highly polarized, whereas N2 molecules are hard to polarize, and thus, these MOFs exhibit higher CO2 capture ability over N2 . M-MOF-74 (M = Mg, Mn, Fe, Co, Ni, Cu, Zn) is a series of MOFs as a representative example of open metal sites containing MOFs displaying high capacity and selectivity to adsorb CO2 . MOF-74, also known as M/DOBDC (DOBDC = 2,5-dioxido-1,4-benzene-dicarboxylate), is constructed by DOBDC and MII ions generating infinite linear metal–oxygen chains, which create the honey comb-type frameworks with one-dimensional (1D) hexagonal channel with a high exposed metal site density (shown in Figure 15.1a). In 2008, Matzger and coworkers recorded the remarkable tuning of the uptake of CO2 by modifying the type of metal on the M-MOF-74 platform [1]. N2 adsorptions have been performed to estimate the BET surface area for these MOFs, and the values were 1495, 1080, 1070, and 816 m2 /g for Mg-, Co-, Ni-, and Zn-MOF-74, respectively. The CO2 sorption isotherms were determined at 296 K and at a different pressure to clarify the influence of metal type on the uptake of these isostructural MOFs. Measurement of Zn-MOF-74 delivered an uptake of 5.8 wt% at 0.1 atm and 24.4 wt% at 1 atm. For Ni-MOF-74, an uptake of 11.6 wt% at 0.1 atm and 25.6 wt% at 1 atm was recorded, which is similar to that of Co-MOF-74 (11.7 wt% at 0.1 atm) but lower than that of Co-MOF-74 at 1 atm (30.6 wt%). The highest uptake for CO2 was founded of Mg-MOF-74 among these four MOFs with 23.6 wt% at 0.1 atm and 35.2 wt% at 1 atm. The sequence for the uptake of CO2 at 298 K is Mg > Co > Ni > Zn. Eliminating the weight effect of each metal, Mg-MOF-74 hosts ≈12 molecules of CO2 /unit cell at 0.1 atm (Figure 15.1b), while other metal-based MOF-74 take up lower CO2 /unit cell. The highest uptake of CO2 for Mg-MOF-74 could be ascribed to somewhat the increased ionic character of the Mg—O bond. In 2014, Queen et al. studied CO2 adsorption in the M-MOF-74 system (M = Mg, Mn, Fe, Co, Ni, Zn) [2]. Among these M-MOF-74s, Mg-MOF-74 remained the highest for CO2 uptake.

(a)

(b)

Figure 15.1 (a) View of 1D channels present in the MOF-74 structure (solvent omitted). (b) Space-filling model of the pore structure of one unit cell of Mg-MOF-74 occupied by 12 molecules of CO2 (blue) that signifies the sorption at 0.1 atm and 296 K. Source: From Caskey et al. [1].

883

15 Relationship Between Structure and Separation Property

Mg-MOF-74 is still a benchmark representative for solid adsorbents at low to moderate CO2 trapping pressure from flue gas with the highest known potential for CO2 adsorption. Dramatic tuning in the uptake of CO2 and selectivity of CO2 /N2 was reported in 2016 by Zhai et al. by systematic modification of components of metal in the CPM-200 system. Accordingly, a sequence of compounds was made of metal trimers and tetracarboxylated planar ligands of the soc topological system [3]. Unlike the reported single-type trivalent metal ion used in CPM-200, a heterometallic strategy was adopted in this work. The M2+ /M3+ heterometallic trimers were incorporated M3+ (OH)•(COO)6 ] and provided a CPM-200 into the structural platform as [M2+ 2 with eight various combinations (Mg/Sc, Mg/V, Mg/Fe, Mg/Ga, Mg/In, Mn/In, Co/In, and Ni/In) as shown in Figure 15.2a. In general, the Mg-containing compounds show greater isosteric heat and increased CO2 uptake, close to the MOF-74 model. When fixing a type of metal in combination with the other form, their uptake of CO2 at 273 K and 1 bar differ in a wide range of Fe/Mg > In/Mg > V/Mg > Sc/Mg, with a very high CO2 uptake value of 9.27 mmol/g in the Fe/Mg case. The ideal adsorbed solution theory (IAST) selectivity may also be tuned to the sequence of V/Mg (406) > Fe/Mg (201) > In/Mg (48) > In/Co (33) > Ga/Mg (24) for a binary mixture of 50/50 CO2 /N2 at 273 K. In addition, the isosteric heat value for V/Mg of −79.6 kJ/mol among Lewis acid sites MOFs is the highest (Figure 15.2b). The increase in isosteric heat was proposed to be associated with the charge-to-radius ratio of metal ions. MOF materials typically have inferior uptake capacities of N2 at ambient conditions and ideally adsorb CO2 over N2 . The absolute selectivity value of CO2 /N2 , therefore, becomes less essential than other separation applications. Alternatively, the uptake capacity (particularly capacity at ≈0.15 atm), working capacity, suitable working cycle, and the cost of energy in regeneration, along with the material stability (e.g. moisture survival), have the most important practical effects. One of the trends in the latest CO2 capture research is that the realistic industry processes for adsorption and regeneration are often considered based on specific characteristics M3+

e tiv n era tio op iza Co stall cry

M3+ Mg2+

(a)

Sc3+ V3+ Fe3+ Ga3+ In3+

M2+

M2+

Charge neutral

100

C cr oop ys er tal at liz ive at ion

90 80 70

[M2+2M3+ (OH) (COO)6]

M2+ Mg2+ Mn2+ Co2+ Ni2+

Q4(kJ/mol)

884

In3+

60 50

V/Mg In Fe/Mg In/Mg In/Ni In/Mn In/Co Ga/Mg Sc/Mg

40 30 20 10 0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 Co2 uptake (mmol/g)

(b)

Figure 15.2 (a) M2+ and M3+ combinations for CPM-200s; (b) isosteric heat for CO2 for CPM-200s. Source: From Zhai et al. [3]. © American Chemical Society.

15.2 CO2 Capture and Separation

of an MOF adsorbent (such as pressure swing adsorption [PSA], temperature swing adsorption [TSA], or vacuum swing adsorption [VSA]). In 2011, a systemic evaluation of MOFs on carbon dioxide capture after combustion under a temperature swing set was first reported by Long and coworkers. The potential availability for low-pressure CO2 capture in a power plant as an energy source for regeneration is thought particularly promising because a vacuum is hard to compress or to use at such great values as a low-pressurized gas stream. The evaluation of two distinct MOF-74 (Mg) and MOF-177 compared to zeolite NaX showed that MOF-74 (Mg) is significantly more efficient than MOF-177 and zeolite NaX in selectivity and working capacity with high binding sites density. Furthermore, MOF-74(Mg) reveals a dual-site adsorption activity and its participation by the open metal site is nearly stoichiometric as seen from the isosteric heat plot. The creation of open metal sites is a big technique for building strong sites that differ according to the type of metal, and the sequence of binding forces depends on the interaction between the metal and the guest. In 2008, the dramatic tuning of the CO2 uptake at 298 K was reported by Matzger et al. as of Mg > Co > Ni > Zn, while the CO2 uptake at 1 atm and 298 K by Mg/DOBC was still the highest. In comparison to the MOF-74, MOFs constructed of trigonal prismatic metal trimer clusters are a much wider platform of both inorganic and organic components. The spatial arrangement of these binding sites may have a significant role as well as the exposed metal sites type. Zhou and coworkers built a unique MOF from copper paddle wheels and a bent di-isophthalate-type linker, known as PCN-88 [4]. This compound has a small pocket in which the exposed Cu2+ sites are aligned to point to each other, creating a distinctive “single-molecule trap” (SMT) that is capable of precisely trapping a CO2 molecule within it. It was stressed that the carboxylate linker was carefully selected to predesign the pocket holding SMT. This is a great example of how MOF materials provide researchers with adequate space in material design (shown in Figure 15.3). The resulting materials can preferably adsorb CO2 over N2 and CH4 , although the zero-coverage isosteric heat and CO2 /N2 selectivity are moderate, primarily because of the low percentage of SMT among the overall framework. However, the single molecular trap remains a precious principle, and improved separation efficiency can be anticipated if high SMT density can be embedded.

15.2.2 Introducing Polar Functional Groups Strong binding to CO2 , for instance, can also be realized with the incorporation of highly polarized inorganic groups by other methods. Zaworotko and coworkers studied the CO2 sorption by SIFSIF MOFs [5]. Such compounds comprise square nets of divalent metal nodes and neutral bipyridine (or pyrazine) linkers, which are further pillared by SiF6 2− anions. Therefore, a strong electrostatic field is available within the porous frameworks, imitating that of zeolite (Figure 15.4). The authors showed the power of pore engineering to tune the adsorption and separation properties. The pores in this group of materials differ according to the linkers size and interpenetration, resulting in three representative materials. SIFSIF-Cu is nano-sized with a long bipyridine connector and no interpenetration and its CO2

885

886

15 Relationship Between Structure and Separation Property

CO2 Trapping

Co-ordination assembly SMT

(Two strategies to build SMTs into extended porous solides)

(a)

MOFs

Composites

(b)

Figure 15.3 Designing porous materials with SMTs for CO2 molecule capture. (a) Schematic representation of the design and construction of an SMT for CO2 adsorption. The organic linker is shown in orange, and the metal node with an open adsorption site is shown in purple and yellow; green spheres represent the cavity for CO2 adsorption. (b) Two proposed strategies for building a predesigned SMT into extended porous solids: assembly of the SMTs by ligand extension and accommodation of SMTs into other porous materials, such as mesoporous silica. Source: From Li et al. [4]. © Springer Nature.

uptake is moderate at 298 K and 1 bar. In contrast, there is a very narrower 5.1 Å pore and lower surface area in the doubly interpenetrated SIFSIF-2-Cu-i. The total uptake capacity of CO2 is 5.4 mmol/g, the highest of the three reported phases. SIXSIX-3-Zn has even narrower 3.8 Å pore and lowest surface area with its shorter pyrazine linker. The small size of the pores provides early saturation (approximately 0.3 atm), with sorption of CO2 , and a mild average uptake at 1 bar (Figure 15.4c). In fact, the narrowest pores have greater contact between the host and the guest, as can be seen from the elevated and steady isosteric heat of 45 kJ/mol. Consequently, the CO2 uptake is even stronger at 0.1 bar relative to the one-component sorption curves reported by SIFSIF-2-Cu-i as well as the direct-sorption mixture of 10/90 CO2 /N2 measured at 1 bar and the column breakthrough experiment (Figure 15.4d). The separate CO2 sorption mechanism offers these two substances advantages at multiple separation opportunities. In this regard, the SIXSIX-3-Zn resembles to a typical zeolite material which can partly be seen in comparison with zeolite 13X. Due to its high selectivity (240) for CO2 over H2 in a CO2 /H2 :30/70 mixture, SIXSIX-2-Cu-i is better designed for syngas separation and SIXSIX-3-Zn is better suited for capture of CO2 after combustion.

15.2 CO2 Capture and Separation

N

N

N

N

SIFSIX–2–Cu

SIFSIX–2–Cu–i

SIFSIX–3–Zn

SIFSIX–2–Cu–i

5 4 3 2

0

1.5

100 200 300 400 500 600 700 800 Pressure (torr)

Ads, at 258 K Ads, at 273 K Des, at 258 K Des, at 273 K

Ads, at 288 K Des, at 288 K

0.4

0.0 0

Ads, at 298 K Des, at 298 K

(c)

0.8

1.0

0.0 0

(b)

CO2: SIFSIX–3–Zn

0.5

1

N2: SIFSIX–2–Cu–i N2: SIFSIX–3–Zn

1.2

2.0 CA/CO

6

CO2: SIFSIX–2–Cu–i

1.6

3.5 Co2 uptake (mmol/g)

Co2 uptake (mmol/g)

7

SIFSIX–3–Zn

3.0

8

20

40

60 80 100 Pressure (torr)

0

120 140

Ads, at 298 K Ads, at 308 K Ads, at 318 K Des, at 298 K Des, at 308 K Des, at 318 K Ads, at 328 K Ads, at 338 K Des, at 328 K Des, at 338 K

1.000

2.000

3.000

4.000

Time (s)

(d)

Figure 15.4 (a) Crystal structures of SIFSIX-2-Cu, SIFSIX-2-Cu-i, and SIFSIX-3-Zn. (b,c) CO2 sorption isotherms under different temperatures for SIFSIX-2-Cu-i and SIFSIX-3-Zn. (d) Column breakthrough experiment for a CO2 /N2 :10/90 gas mixture at 298 K, 1 bar. Source: From Nugent et al. [5].

Eddaoudi and coworkers also found that the pore size could be identified through substituting Zn with Cu in SIXSIX-3-M [6]. This probably reflects a Jahn–Teller distortion from Cu2+ which leads to a distorted geometry in the octahedral coordination, with elongated Cu–F links and relatively shorter Cu—N bonds. Therefore, at room temperature, the obtained SIXSIX-3-Cu revealed even steeper adsorption curve for CO2 , implying a greater interaction with higher adsorption heat of −54 kJ/mol (Figure 15.5). This function potentially makes this material to be utilized for trace amounts of CO2 from air. The SIXSIX-3-Cu CO2 uptake is 1.24 mmol/g about 10 times the SIXSIX-3-Zn and 20-fold the SIXSIX-2-Cu-i CO2 uptake. Indeed, physisorption becomes, in this case, so strong by crystal engineering that it is analogous to the material that has chemical adsorption sites. Identically, Ni version can also be produced in a number of concentrations with pore size materials that are appropriate for CO2 capture. Zhang and coworkers were similarly studying CO2 capture on a two-dimensional (2D) coordination network called Qc-5-Cu-sql, a square lattice [7]. This compound has two forms, the solvated one (Qc-5-Cu-sql-α) has a pore size of 3.8 Å, while the desolvated form (Qc-5-Cu-sql-β) has a pore size of 3.3 Å. Since the kinetic diameters of CO2 , N2 , and CH4 are 3.3, 3.64, and 3.8 Å, respectively, there has been a significant selectivity difference of 40 000 for 0.15 bar CO2 /0.85 bar N2 mixture. This material

887

15 Relationship Between Structure and Separation Property

CO2 uptake cm3 (STP) cm3

120

SIFSIX–3–Cu 298 K SIFSIX–3–Zn 298 K Mg–DOBDC 313 K SIFSIX–2–Cu–i 298 K

100 80 60 40 20 0 0.00

0.01

0.02

0.03

0.04

0.05

Pressure (bar) (a) SIFSIX–3–Cu SIFSIX–3–Zn SIFSIX–2–Cu–i

60

50 Qst (kJ/mol)

888

40

30

20 0.0 (b)

0.5 1.0 CO2 update (mmol/g)

1.5

Figure 15.5 (a) CO2 volumetric uptake for SIFSIX-3-Cu at 298 K compared with SIFSIX-3-Zn, SIFSIX-2-Cu-I, and Mg-MOF-74. The adsorption results at very low pressure (400 ppm, 5%) for the SIFSIX-3-Cu showed that the Cu analog exhibits steeper adsorption isotherms at very low CO2 concentration in comparison with other materials. At 7.6 Torr (0.01 bar), SIFSIX-3-Cu uptakes 82.6 cm3 (STP)/cm3 versus 55 and 28 cm3 (STP)/cm3 for SIFSIX-3-Zn and Mg-MOF-74, respectively. (b) Isosteric heats of adsorption at low coverage for SIFSIX-3-Cu, SIFSIX-3-Zn, and SIFSIX-2-Cu-i. The Qst for SIFSIX-3-Cu was observed to be higher than the Zn analog and steady constant up to relatively higher recorded CO2 loadings. This is indicative of the presence of homogenous binding sites over the full range of CO2 loading for SIFSIX-3-Cu. Source: From Shekhah et al. [6].

is also stable in humidity and its relatively high CO2 /H2 O selectivity ensures the continued efficiency of its CO2 capture in the wet flue gas stream. Coordinated water may also often promote the CO2 capture, as shown in Figure 15.6, UTSA-16, a potassium cobalt citrate coordination compound with terminal water [8]. Different from unvarying performs in which metal sites containing coordinated solvent molecules are eliminated during activation, the coordinated

15.2 CO2 Capture and Separation

Co1 Co2 C H O K

O3W

O3W

K1 Co1

O5

O6

O4 O2 O3 O3W Co2

O7

O1

K1 O3W

(a)

(b)

(c) K1

O3W

2.971 O72

3.067 2.911 O71

3.3 × 5.4 Å2

2.906

O71 O6 2.936

(d)

(e)

(f)

Figure 15.6 The structure for CO2 -loaded UTSA-16. (a) The coordination mode of citrate ligand. Each citrate chelates one cubic (Co(2)4 O4 ) cluster, two Co(1) tetrahedral and two K ions. There is two water (O3 w) coordinated with each K ions. (b) Each cubic (Co(2)4 O4 ) cluster node is linked by four K polyhedral linkers to give a diamondoid cage (yellow ball of about 4.5 Å in diameter). There are four crystallographically independent O3 w from 4 K polyhedral toward the cage. For clarity, Co1, C, and H on citrate ligand are omitted. The linkages between the cubic clusters are present with the green sticks (not the real bond) to illustrate the framework topology. (c) The resulting dia network. (d) The diamondoid cage has a small window with the dimensions of 3.3 Å × 5.4 Å. All metal ions are present as the polyhedral with the colors same as those for the atoms in (a). (e) A couple of Co2 dimers are trapped within the cage. (f) The cooperative interactions between CO2 molecules and framework (O71(CO2 )…O71(CO2 ) and O71(CO2 )…O6 short contact; O72(CO2 )…H-O3w, O71(CO2 )…H-O3w, and O3w…H-O6 hydrogen-bonding interactions). The angle between the two linear CO2 molecules is of 64.93(2)∘ . Source: From Xiang et al. [8].

water in UTSA-16, after sample activation, was reserved. It is very strongly associated with CO2 with a steady Qst of around −32 kJ/mol. Due to its relatively high frame density of about 1.6 g/cm3 , its volumetric uptake capacity ranks second in ambient conditions and slightly below Mg-MOF-74. Its CO2 /N2 selectivity is even somewhat higher than Mg-MOF-74 but lower than mmen-CuBTTri with chemical adsorption. A fixed breakthrough experiment on the 15/85 CO2 /N2 mixture has shown that in terms of breakthrough time, the performance of UTSA-16 is identical to NaX, but slightly smaller than Mg-MOF-74.

15.2.3 Lewis Base Incorporation in Metal Sites The conventional process of CO2 capture comprises chemisorption using alkylamine-containing liquids, as Lewis bases, with a high affinity for CO2 . This process, however, results in higher energy regeneration costs and corrosion

889

15 Relationship Between Structure and Separation Property

O C O :

890

(1) en (2) CO2

H2N

N: H2

Figure 15.7 A portion of the structure of the sodalite-type framework of Cu-BTTri (1) showing surface functionalization of a coordinatively unsaturated CuII site with ethylenediamine, followed by the attack of an amino group on CO2 . Purple, green, gray, and blue spheres represent Cu, Cl, C, and N atoms, respectively; framework H atoms are omitted for clarity. Source: From Demessence et al. [9].

issues. This inspiration led Long and coworkers to develop a strategy for increasing CO2 binding through grafting alkylamine onto porous solids from exposed metal sites [9]. Particularly, as shown in Figure 15.7, on a sodalite-type triazolate-bridged framework, CuBTTri (H3 BTTri = 1,3,5-tris (1H-1,2,3-trazol-5-yl) benzene), ethylenediamine (en) was grafted onto the exposed Cu site to generate en-CuBTTri. The resulting material indicates a high affinity for CO2 binding with an isosteric heat of −78 kJ/mol. On using N,N′ -dimethylethylenediamine (mmen) as the grafted amine, this led to higher value (−96 kJ/mol). Such strong binding association suggests the chemisorption feature of the host–guest interaction relative to physisorption. Due to its high binding strength, the TGA-monitored gas cycling experiment has verified that the substance could be regenerated at moderate temperature (333 K) during a temperature-swing process. Nevertheless, N2 /CO2 mixture may again be required to provide pure N2 flow during regeneration. This strategy was further utilized by the same group on the M2 (dobpdc) platform (M = Mg2+ , Mn2+ , Fe2+ , Co2+ , Ni2+ , Zn2+ ), which is an extended version of M2 (dobdc) (MOF-74) [10]. Interestingly, the behavior of “phase change” has been found with unusual step-shaped CO2 -adsorption isotherms for CO2 sorption with mmen-grafted of M2 (dobpdc) (Figure 15.8). The spectroscopic, diffraction, and computational experiments revealed that the sharp adsorption step was generated by an unprecedented cooperative CO2 insertion process. The step-shaped isotherm generally has dramatic changes in the capacity

15.2 CO2 Capture and Separation

(a)

(b)

2.29(6) Å 2.10(2) Å

(c)

(d)

(e)

Figure 15.8 Powder X-ray diffraction structures of mmen-Mn2 (dobpdc). (a,b) Space-filling models of the solid-state structures of mmen-Mn2 (dobpdc) (a) and CO2 -mmen-Mn2 (dobpdc) (b) at 100 K. (c,d) Portions of the crystal structures for mmen-Mn2 (dobpdc) before (c) and after (d) CO2 adsorption, as determined from powder X-ray diffraction data. (e) A portion of the crystal structure for the final configuration of CO2 adsorbed within mmen-Mn2 (dobpdc), depicting the formation of an ammonium carbamate chain along the pore surface. Green, gray, red, blue, and white spheres represent Mn, C, O, N, H atoms, respectively; some H atoms are omitted for clarity. Source: From McDonald et al. [10].

around a certain threshold pressure, which in comparison to regular adsorption, can result in a high working capacity. For a fact, the threshold pressure varies dramatically, with temperature, the type of metal, and the grafted diamine type; thus, with minor temperature swings and low regeneration emerges, significant CO2 separation capacities can be attained. Such research suggests a new way of producing highly efficient adsorbents for CO2 from different gas mixtures. Hydroxide anion is a prominent species due to high affinity to CO2 producing HCO3 − from the reaction between OH− and CO2 . The reversible reaction is, however, challenging to accomplish and therefore unfit for efficient applications

891

892

15 Relationship Between Structure and Separation Property

(a) O

O

C O

N

Cl

O C O

Cl

MII N

N

N

O Cl

MII Cl

N

O

N O C O

M N

M Cl

N

N

Cl

III

Cl N

OH

H

N

III

Cl N

(b)

Figure 15.9 Comparison of the (a) local framework structures (C, gray; H, light gray; N, blue; O, red; M(II), purple; M(III), orange) and (b) CO2 adsorption mechanisms of 1/2 and 1′ /2′ . Source: From Liao et al. [11].

of CO2 capture and separation. Because monodentated hydroxide is a leading active site for carbonic anhydrase, which, without changing the thermodynamic equilibrium, can significantly accelerate the conversion of CO2 and HCO3 − in the aqueous system, it is a viable candidate for integrating strong CO2 affinity and reversibility. Nevertheless, very few monodentated hydroxide-containing MOFs have been reported, and most of these have poor CO2 capture performance, indicating that hydroxide is not really monodentate. In 2015, Zhang and coworkers reported a unique set of isostructural MOFs with MOF-74-like hexagonal-shaped 1D channels, with their formula as [MII 2 Cl2 (bbta)] and [MII MIII Cl2 (bbta)] (H2 bbta = 1H,5H-benzo(1,2-d:4,5-d′ )bis-triazole, MAF-X25 and MAF-X25ox, M = Mn; MAF-X27 and MAF-X27ox, M = Co) [11]. When all the metals are in low valence form, upon activation, the MII adopts the square-pyramidal coordination acting as a Lewis acid site for CO2 binding, resembling the case in MOF-74. It is interesting to note that the in situ oxidation of the MII MIII mixed-valence form in a 1 : 1 ratio during material synthesis can be statistically produced. The corresponding excess positive charge is balanced by the monodentate hydroxide groups at the MIII sites. The hydroxide is supposed to be tightly bound to CO2 in such a uniform way, that HCO3 − can be transformed to OH− without excessive energy costs (Figure 15.9). The MAF-X25ox isosteric zero-coverage isosteric heat is about three times higher than that of MAF-25. It clearly shows the chemisorption properties, as in the case of the cobalt version (MAF-27). Additionally, the high CO2 /N2 selectivity is shown by both single-component adsorption and breakthrough experiments. More specifically, even at high relative humidity, these materials can trap up to 4.1 mmol/cm3 or 13.4 wt% of CO2 , and rapidly liberate them at mild regeneration conditions, offering the best CO2 capture results recorded at this time.

15.2 CO2 Capture and Separation

15.2.4 Functionalization of Ligands Another method for improving CO2 adsorption and separation efficiency is the insertion of functional groups into the pores of MOF by ligand alteration. In 2014, Li and coworkers reported two UiO-67 analogs named BUT-10 (with formula [Zr6 O4 (OH)4 (FDCA)6 ]) and BUT-11 (with formula [Zr6 O4 (OH)4 (DTDAO)6 ]), (H2 FDCA = 9-fluorenone-2,7-dicarboxylic acid; H2 DTGAO = dibenzo[b,d] thiophene-3,7-dicarboxylic acid 5,5-dioxide) [12]. FDCA and DTDAO can be considered as the modified version of biphenyl-4,4′ -dicarboxylic acid, which was used in UiO-67. Without much alteration in the ligand length and with an exactly identical structure, BUT-10 and BUT-11 demonstrated large improvement in CO2 uptake capacity (from 22.9 to 50.6 and 53.5 cm3 /g at 298 K and 1 atm, in the order of UiO-67, BUT-10, and BUT-11) and CO2 /N2 selectivity (from 9.4 to 15.6 and 31.5). Applying the computational study shows that the sulfonate groups greatly improve the affinity to CO2 molecules by ligand modification. A subsequent study showed a drastic increase in CO2 uptake and CO2 /N2 selectivity by adding a pair of nitro or sulfonate groups on the Zr-NDC platform. The working capacity and adsorption selectivity at 0.1 and 1.0 bar of these materials follow the following trend, Zr-NDC < ZrNDC–2NO2 < ZrNDC–2NH2 < ZrNDC–2SO3 H. Yaghi and coworkers reported that the ligand of IRMOF-74-III can be modified and functionalized by covalent bonding with primary amine (IRMOF-74-III-CH2 NH2 ) to be utilized for CO2 selective capture [13]. By covalent ligand modification, Figure 15.10 shows six analogs with different functional groups (–CH3 , –NH2 , –CH2 NHBoc, –CH2 NMeBoc, –CH2 NH2 , and –CH2 NHMe) were realized. At 298 K, the measured CO2 adsorption isotherms imply that IRMOF-74-III-CH2 NH2 and -CH2 NHMe have the highest uptake capacities at low CO2 pressure range ( Co2+ > Mg2+ (Figure 15.18), which is consistent with the softness and hardness of metals and the ability to form feedback π bonds.

5

FeMOF-74 CoMOF-74

4 MgMOF-74

3 CuBTC Nax

2 1 0 0

PCN-16 NOTT-102 UTSA-20

NaETS-10

1 2 3 4 5 99%+ C2H6 produced during Δrads / (mol/l)

99.5%+ C2H4 produced during Δrads / (mol/l)

99.5%+ C2H4 produced during Δrdss / (mol/l)

15.3 Separation of Hydrocarbons 5

FeMOF-74 CoMOF-74

4 MgMOF-74

3 CuBTC

Nax

2 PCN-16

1 0 0

NaETS-10 NOTT-102 UTSA-20

6 12 2 4 8 10 14 C2H4/C2H6 adsorption selectivity. Sads

(a)

Figure 15.18 A comparison of various MOFs for the productivity of 99.5%+ pure C2 H4 expressed as a function of (a) productivity of 99%+ pure C2 H6 and (b) C2 H4 /C2 H6 adsorption selectivity. The conditions chosen are p1 = p2 = 50 kPa and T 1/4 296 K. The data for FeMOF-74 are at 318 K and therefore considered to be conservative. Source: From He et al. [25]. 5

Uptake (mmol/g)

Ag S O

AgBF4

C2H4 C2H6 C3H6 C3H8

4 3 2 1 0 0

(a)

(b)

20

40 60 80 Pressure (kPa)

100

Figure 15.19 (a) Immobilization of Ag(I) into an MOF with –SO3 H sites. (b) Single-component sorption isotherms for various hydrocarbons in (Cr)-MIL-101-SO3 Ag at 303 K. Source: From Chang et al. [26].

Besides the use of open metal sites for MOFs themselves, the introduction of Ag(I) or Cu(I) plasma, which is easy to form π coordination bond with olefins, will be of great support to the separation of ethylene and ethane. As shown in Figure 15.19, MOF [(Cr)-MIL-101-SO3 H] was synthesized by Bao, Chen, and coworkers, and then Ag (I) was introduced and bound to sulfonate in the MOF. The adsorption enthalpy and adsorption capacity of ethylene were greatly improved by Ag(I) modified MOF, and the separation of ethane and ethylene was better than that of unmodified MOF [26]. As aforementioned, the separation of ethylene and ethane by MOFs is achieved by conventional adsorbents trapping ethylene in over ethane, which means that several steps are needed in separation processes. A superior version is that if the adsorbent preferentially adsorbs ethane in preference to ethylene, and in turn, in one step, high-purity ethylene could be attained. In 2015, Zhang and coworkers reported an example of MOF, MAF-49, which was prepared by reaction of H2 batz (bis(5-amino-1H-1,2,4-triazol-3-yl)methane in addition to Zn(OH)2 in dilute aqueous ammonia and produced a porous metal–azolate framework [Zn(batz)]⋅0.5H2 O with a unique structure [27]. As shown in Figure 15.20a, a 3D coordination framework contains a narrow 1D zigzag channel (3.3 Å × 3.0 Å), a

903

904

15 Relationship Between Structure and Separation Property

(a)

(b)

Figure 15.20 X-ray crystal structure of MAF-49⋅H2 O. (a) Framework (Zn, purple; C, dark gray; H, light gray; N, blue) and pore surface (yellow/gray curved surface) structures. Guest molecules are omitted for clarity. (b) Local environment and hydrogen-bonding interactions of the narrowest channel neck (highlighted by green dashed lines). Source: From Liao et al. [27]. Licensed Under CC BY 4.0.

pair of methylene groups with a cis-configuration, which is accommodated by a guest H2 O molecule with two O—H…N and two C—H…O hydrogen bonds. Adsorption isotherms of single-component for C2 H6 and C2 H4 were calculated for activated MAF-49 at different temperatures. Notably, MAF-A adsorbs C2 H6 over C2 H4 , based on their different isotherm shapes, and it was found that the host–guest binding obeys C2 H6 > C2 H4 . The calculated gas adsorption enthalpies were 60 and 48 kJ/mol for C2 H6 and C2 H4 , respectively. Practically, for investigating the separation capacity of MAF-49, breakthrough experiments were performed, and a gas mixture of four components (CH4 /CO2 /C2 H4 /C2 H6 ) can be separated effectively (Figure 15.20b). On using ambient conditions, passing a typical cracked gas mixture (15 : 1 C2 H4 /C2 H6 ) through 1 l of this C2 H6 selective adsorbent leads to direct production of 56 l of C2 H4 with 99.95%+ purity at the outlet, with a single breakthrough operation, compared to other MOFs. To be aware of this outstanding performance for separation of C2 H6 /C2 H4 , host–guest structures and energy changes were calculated by GCMC simulation and further periodic density functional theory (DFT) optimization. Taking into account both the host–guest binding and the distortion of the host framework, total energies or adsorption enthalpies can be measured as −56.5 and −45.2 kJ/mol for C2 H6 and C2 H4 , respectively, which are in consistence with the experimental values. The C2 H6 creates three strong C—H…N hydrogen bonds and three weak C–H…N electrostatic interactions with MAF-49 (Figure 15.21a). For C2 H4 , only two less strong C—H…N hydrogen bonds and two fragile C–H…N electrostatic interactions were observed (Figure 15.21b). These findings show that in distinguishing between C2 H6 and C2 H4 with large adsorption enthalpy difference, both the electronegative nitrogen and electropositive methylene groups have significant roles. This finding is beneficial in removing a small amount of ethane from ethylene and obtaining high purity ethylene. In 2018, Zhou, Chen, and coworkers investigated the separation of ethylene/ ethane based on two reported ultra-microporous MOFs (Cu(ina)2 and Cu(Qc)2 ) [28]. These two MOFs were constructed by combining Cu2+ with ina−

15.3 Separation of Hydrocarbons

(a)

(d) N68

N8

N

H72

–

C6

N2

H8A

H63

N8A

H H

H H73

R N

N

H

N4C

H61

(g)

H

H

R

H

N

N

H C

C

H

H

H71

R

N

H

H N –

C3A

N

N3A

N1A

N

(b)

(h) N

N8

N

H

–

R N

H

N

H61

H

C6

C7

N

(e)

N3D

H72

–

N

N

N

N

N

–

H62

H C

H71

C

H

H

N2A

N

N N6A

N

–

–

(c)

N

N

N

(f)

(i)

N8

O1

H31 C3

C6 O1A

H31A

R N8A

H R

R

H

H N

C H

O

C3A

C

O

H

R C

N H

H

R

H

R

Figure 15.21 Host–guest fittings and interactions. Preferential adsorption sites for (a) C2 H6 and (b) C2 H4 in MAF-49 revealed by computational simulations (Zn purple, C dark grey, H light grey, N blue). Source: From Liao et al. [27]. Licensed Under CC BY 4.0.

(Hina = isonicotinic acid) and Qc− (HQc = quinoline-5-carboxylic acid), respectively. Figure 15.22a,b show that the two MOFs exhibit isoreticular framework with the diverse ligand. The result from the larger size of ligand, the pore size of Cu(Qc)2 is narrower than that of Cu(ina)2 (Figure 15.22c). Under ambient conditions, low-pressure sorption data of C2 H6 and C2 H4 show that Cu(ina)2 possesses selective sorption behavior for ethane to some degree, affording C2 H6 /C2 H4 uptake ratio of 105%. By comparison, Cu(Qc)2 exhibits smaller pore-aperture sizes and larger aromatic π systems offer more distinct ethane-selective sorption performance under identical conditions (Figure 15.22d). The uptake ratio (237%) of C2 H6 /C2 H4 for Cu(Qc)2 is significantly higher than Cu(ina)2 and is the highest in the recorded MOFs. Besides, the calculated selectivity of C2 H6 /C2 H4 for Cu(Qc)2 corresponding to a binary equimolar mixture is up to 3.4 at 298 K and 100 kPa, which is greater than other high-performance MOF materials of ethane selective at that time (Figure 15.22e,f). From a structural point of view, high-resolution NPD measurements demonstrate that the C2 D6 molecule locates in a rhombic cavity generated from the ligands of aromatic rings of the same-layered network. Multiple C–D…π interactions between each C2 D6 molecule and aromatic rings within the rhombic

905

15 Relationship Between Structure and Separation Property (a)

(b)

(C)

Pore size distribution

Cu(ina)2 6 4 Cu(Qc)2 2 0 0

C2H6

C2H4

C2H6

C2H4

(e)4

Cu(Qc)2

50

25 298 K

0 0.0

0.2

0.4 0.6 Pressire (bar)

50/50 C2H6/C2H4 ZIF-8 UTSA-33 NI(bdc)(ted)0.5 UTSA-35

0.8

1.0

3 MAF-49 ZIF-7 IRMOF-8

2

PCN-250

Cu(ina)2

1 0.0

0.2

0.4 0.6 Pressire (bar)

0.8

1.0

1

2 3 Pore size / (Å)

(f) C2H6/C2H4 selectivity

Cu(ina)2: Cu(Qc)2:

IAST selectivity

(d) 75 Uptakes (cm3 (STP)/cm3)

906

4

5

Cu(Qc)2 *

3.5 3.0 2.5

MAF-49

2.0

IRMOF-8 PCN-250 ZIF-8

1.5 1.0

ZIF–7

Ni(bdc)(ted)0.5

UTSA-35 UTSA-33 Cu(ina)2

C2H6 selective C2H4 selective

100 120 140 160 180 200 220 240

C2H6/C2H4 uptake ratio (%)

Figure 15.22 Comparison of crystal structures and channels between Cu(ina)2 (a) and Cu(Qc)2 (b); Cu, O, N, and C are represented by green, red, light blue, and gray, respectively, and guest molecules are omitted for clarity. (c) Pore size distribution for Cu(ina)2 and Cu(Qc)2 based on sphere probes. (d) The C2 H6 and C2 H4 sorption isotherms for Cu(ina)2 and Cu(Qc)2 at 298 K. (e) Pressure-dependent C2 H6 /C2 H4 selectivities for Cu(ina)2 and Cu(Qc)2 in comparison to the best materials reported to date. (f) The C2 H6 /C2 H4 selectivities/uptake ratios for Cu(ina)2 (blue diamond) and Cu(Qc)2 (red star) at 298 K and 1 bar. Source: From Lin et al. [28].

cavity for Cu(Qc)2 were found. An NPD for C2 D4 loaded MOFs, and other calculations were also performed in the article. These all structural messages well supported the high performance of Cu(Qc)2 for C2 H6 -selective separation of C2 H6 /C2 H4 . Influenced with natural metalloenzymes and synthetic compounds for alkane C–H activation in which M-peroxo, M-hydroperoxo, and M-oxo [M = Cu2+ , Co3+ , and Fe2+/3+ ] are active catalytic intermediates, in 2018, Li, Zhou, Chen, and coworkers introduced similar functional sites into MOFs to expect the MOFs with stronger binding of alkanes than alkenes [29]. After oxidizing the well-known Fe(II)-MOF-74 (also known as Fe2 (dobdc)), Fe2 (O2 )(dobdc) was obtained, while the framework of Fe(II)-MOF-74 was maintained. The separation performance for C2 H6 and C2 H4 mixtures have been examined, and Fe2 (O2 )(dobdc) displays C2 H6 favorable binding over C2 H4 . Due to its maintained framework and surface area, Fe2 (O2 )(dobdc) took up moderately high amounts of C2 H6 . For single-component sorption isotherms at 298 K and pressures up to 1 bar, the C2 H6 uptake amount is 74.3 cm3 /g, which is higher than that of C2 H4 . Measurements of C2 D6 -loaded and C2 H4 -loaded Fe2 (O2 )(dobdc) samples at 7 K were performed at high-resolution NPD for elucidating the adsorption mechanism of C2 H6 and C2 H4 by this MOF. C2 D6 molecules show preferential binding with the peroxo sites through C—D…O hydrogen bonds, which indicate a relatively strong interaction. Compared with other high performing MOFs, Fe2 (O2 )(dobdc) exhibits outstanding performance for separation of C2 H6 /C2 H4 and

15.3 Separation of Hydrocarbons

considered as a new benchmark for C2 H6 /C2 H4 adsorption selectivity with high recovery of pure C2 H4 . Further study shows that highly pure C2 H4 (≥99.99%) of the C2 H6 /C2 H4 mixtures can easily be obtained during the first breakthrough cycle, with reasonably high productivity and low energy costs. The consequence is that Fe2 (O2 )(dobdc) was at this time the most effective material to isolate C2 H6 from C2 H6 /C2 H4 mixtures. While [Fe2 (O2 )(dobdc)] shows good selectivity as well as ethane capacity, the energy penalty when regenerated (Qst = −66.8 kJ/mol) remains large. In fact, this material is air-unstable and needs careful treatment under the dry box. In 2019, Telfer and coworkers reported an MOF, MUF-15 ([Co3 (μ3 -OH)(ipa)2.5 (H2 O)], MUF = Massey University Framework), constructed from inexpensive isophthalic acid and cobalt acetate [30]. Through analyzing the structure of MUF-15, the framework defined three narrow zigzag 1D pores that intersect each other. Figure 15.23 shows that these orthogonal channels run along the a-, b-, and c-axes with pore-limiting windows of 8.5 Å × 3.5 Å, 7 Å × 3.8 Å, and 3.2 Å × 1.2 Å, respectively. The BET surface area of 1130 m2 /g and a pore volume of 0.51 cm3 /g were given by N2 adsorption isotherm at 77 K. An MUF-15 showed C2 H6 -selective separation performance because of its relatively high pore volume and exhibited the (a)

(b) t

t t

t

(c)

O

O OH

OH

H2ipa Co6(μ3–OH)2(RCO2)10

Ipa

(d)

(e)

Figure 15.23 (a) The SCXRD structure of MUF-15 comprises hexanuclear cobalt(II) clusters (cobalt, dark blue; oxygen, red; carbon, gray; hydrogen, pink [most omitted for clarity]). Sites occupied by terminal H2 O ligands are marked with a “t.” (b) Structure of the H2 ipa linker and its stick representation. (c,d) Cobalt(II) clusters and ipa ligands assemble into a network that defines a 3D array of channels. (e) Internal structure of the network of pores in MUF-15 illustrated by the Connolly surface in yellow (probe of diameter 1.0 Å). Source: (a) From Qazvini et al. [30].

907

15 Relationship Between Structure and Separation Property

highest C2 H6 uptake among top-performing ethane-selective MOFs. In addition, its relative small pores decorated by benzene rings contribute to close contacts between ethane and walls of the framework. First-principles of dispersion-corrected DFT calculations were performed for obtaining additional information into the preferential adsorption pathway of C2 H6 /C2 H4 in MUF-15. At its favored binding site, the measured static binding energy for C2 H6 is roughly −36.7 kJ/mol, while for C2 H4 it is −35.0 kJ/mol. The strong host–guest interactions with C2 H6 are based on the experimental results. Such strong interactions thanks to van der Waals interactions between the ethane and the neighboring π electron clouds. Based on DFT calculations, Figure 15.24 shows that the ethane molecules are bound in a pocket defined by four phenyl rings. An MUF-15 cavity complements the size of the C2 H6 molecule to permit C–H…π interactions between all six hydrogens of C2 H6 and three adjacent phenyl rings, while the C2 H4 molecule displays short contacts only with two parallel edges of the cavity. To verify the gas adsorption mechanism, adsorption isotherms of C2 H2 for MUF-15 were calculated, and the adsorption of C2 H2 was less than both C2 H4 and C2 H6 at low pressure. The calculated heat of adsorption of C2 H2 was less than that of C2 H4 and C2 H6 as well. Such lower C2 H2 affinity further demonstrated that the guest-binding mechanism is mainly based on van der Waals interactions instead of continuous (e) 1.25

(c) 1.4 1.2

1.00 C/C0

1.0

0.75

0.8 0.6

C2H6–exp C2H4–exp C2H6–sim C2H4–Sim

0.4 0.2 0.0 0 3.5

2.5 2.0 1.5

0.00 0 10 15 20 25 30 35 40 Time (min) (f)

Ni(bdc) (ted)0.5

ZIF–8 ZIF–7

100

200 300 400 Time (min)

500

600

1.0 0.8

Fe2(O2)(dobdc)

IRMOF–8 MIL–142–A PCN–245 ZIF–4

1.0 0.5

0.25

MUF–15 PCN–250

3.0 Ethylene seiective MOFs

1

Ethane upatakemix (mmol/g)

(d)

(b)

5

C2 H6 C2 H4

0.50

Cu(Qc)2 MAF–49

C/C0

3

C/C0

1

2

(a)

2

908

0.6 0.4

C2H6 C2H4

0.2 0.0

1

2 3 4 IAST selectivity (C2H6/C2H4)

0 5 10 15 20 25 30 35 40 45 50 Time (min)

Figure 15.24 Comparison of the preferential (a) C2 H6 and (b) C2 H4 adsorption sites (Co, blue; O, red; C, dark gray; H, white) observed by DFT-D3 calculations. (c) Simulated and experimental breakthrough curves for a 50/50 C2 H6 /C2 H4 mixture at 293 K and 1.1 bar in an adsorption column packed with MUF-15. (d) Ethane uptake from an equimolar mixture of C2 H6 /C2 H4 as a function of IAST selectivity for the best ethane-selective materials reported to date. (e) C2 H6 /C2 H4 separation cycles for a 25/75 C2 H6 /C2 H4 mixture lasting for 600 minutes. Each separation process was carried out at 293 K and 1.1 bar, and MUF-15 was regenerated by being kept under vacuum at ambient temperature for 20–30 minutes. (f) Simulated breakthrough curves for a mixture of 0.1/99.9 C2 H6 /C2 H4 at 293 K and 1.1 bar. Source: From Qazvini et al. [30].

15.3 Separation of Hydrocarbons

electric dipoles/quadrupoles. An MUF-15 was the first recorded material with selective adsorption of C2 H6 over both C2 H4 and C2 H2 . Breakthrough experiments were performed at room temperature in which C2 H6 /C2 H4 mixtures of 50/50, 25/75, 10/90, 1/99, and 0.1/99 were served as feeds to model a variety of industrial process conditions. All the mixtures can be separated effectively and gave a pure product of C2 H4 . The cycle experiment was conducted as well to examine the recyclability of MUF-15, and the results indicated that the C2 H6 adsorption and separation capacity for MUF-15 were not significantly reduced over 12 cycles. Furthermore, MUF-15 can be prepared in a straightforward approach utilizing low-cost precursors, robust and easily handled, and can be readily regenerated. Along with the “reverse” selectivity of C2 H6 /C2 H4 and high separation performance, MUF-15 is close to the industrial application for purifying C2 H4 . The isolation of acetylene and ethylene in the market is of considerable significance. Typical ethylene produced in steam crackers contains on the order of 1% of acetylene. During ethylene polymerizations, the amount of acetylene (>5 ppm) in ethylene can poison the Ziegler–Natta catalyst and can also reduce the product value of the obtained polymers. The isolation of acetylene and ethylene in the market is of considerable significance. In addition, the acetylenic compounds are also transformed into solid that obstructs the fluid stream and often result in an explosion. Thus, the development of novel alternative approaches to the separation of C2 H2 /C2 H4 is essential. Chen and cowokers studied the C2 H2 /C2 H4 separation of two flexible mixed metal–organic frameworks (M’MOFs) materials in 2011 [31]. The M’MOFs were constructed by combining metalloligand with metal ion and dicarboxylic acid. As shown in Figure 15.25, the reaction of Cu(SalPyCy) with Zn(NO3 )2 and 1,4-benzenedicarboxylic acid or 1,4-cyclohexanedicarboxylic acid gives two M’MOFs, termed as M’MOFs-2 and -3. At 273 and 295 K, the adsorption isotherms of C2 H2 and C2 H4 on activated M’MOF-2 and -3 were calculated. They displayed type I sorption isotherms with minimal hysteresis, and the selectivities toward C2 H2 /C2 H4 on M’MOF-2a at 273 and 295 K were 1.5 and 1.9, respectively. An M’MOF-3a showed improved C2 H2 /C2 H4 selectivities of 4.1 and 5.2 at 273 and 295 K, respectively, which are 2.5 times higher than the analogous values for M’MOF-2a. Later, they reported another MOF material UTSA-100 and related separation work through the synthesis of aminotetrazole derived ligand H2 atbdc [32]. The framework of UTSA-100 [Cu(atbdc)] is an apo topological and the 3D structure formed by the connection between Cu2 (-COO)4 unit and ligand atbdc2− (Figure 15.26). The 1D channel in the structure is formed by the connection of the cage with a diameter of 4.0 Å through the hole and window of 3.3 Å (Figure 15.27a). The specific surface areas of Langmuir and BET of UTSA-100 are 1098 and 970 m2 /g, respectively, and the pore volume is 0.399 cm3 /g. The adsorption capacities of C2 H2 and C2 H4 of UTSA-100 at room temperature and pressure are 95.6 and 37.2 cm3 /g, respectively. The uptake ratio of C2 H2 /C2 H4 (2.57) exceeded all other representative MOF materials except M’MOF-3a (4.75) at that time (Mg-MOF-74: 1.12, Co-MOF-74:

909

910

15 Relationship Between Structure and Separation Property

RR N

N Cu

N

O

O

N

+ OH

Zn(NO3)2–6H2O

O

O

OH

H HO

H O

1.6 O

HO

Zn3(BDC)3[Cu(SalPycy)].(G)x MʹMOF–2

Zn3(CDC)3[Cu(SalPycy)].(G)x MʹMOF–3

1.6

Enhanced separation selectivity of C2H2 / C2H4

21

Enhanced chiral recognition of 1-phenylethyl alcohol (ee)

25.5 64

Figure 15.25 Syntheses and separation capacities of M’MOFs-2 and -3. Source: From Xiang et al. [31].

(a)

(b)

Figure 15.26 X-ray crystal structure of UTSA-100. (a) The coordination environment of organic ligand ATBDC2− and Cu(II), and dinuclear copper(II) unit as a 6-connected node (purple balls) and ATBDC2− as a 3-connected node (orange ball). (b) The framework topology of apo-type (3,6)-connected network with Schläfli symbol [4.62 ]2 [42 .69 .84 ]. Source: From Hu et al. [32]. Licensed Under CC BY 4.0.

1.16, Fe-MOF-74: 1.11, NOTT-300: 1.48). Moreover, the C2 H2 adsorption heat of UTSA-100 is only 22 kJ/mol, which is lower than all other MOF materials (Mg-MOF-74: 41 kJ/mol, MOF-74: 45 kJ/mol, Fe-MOF-74: 46 kJ/mol, NOTT-300: 32 kJ/mol, M-MOF-3a: 25 kJ/mol). The low adsorption heat of C2 H2 shows that the material is easy to regenerate and is beneficial for practical applications. The selectivity calculation of the IAST model C2 H2 /C2 H4 (1 : 99, v) indicated that the C2 H2 /C2 H4 adsorption selectivity of UTSA-100 (10.72) was significantly higher than that of Mg-MOF-74 (2.18) and Co-MOF-74 (1.70).

15.3 Separation of Hydrocarbons

The finding demonstrated that the C2 H2 /C2 H4 adsorption selectivity of UTSA-100 was significantly higher than that of Mg-MOF-74 (2.18), Co-MOF-74 (1.70), Fe-MOF-74 (2.08), and NOTT-300 (2.17). However, it was lower than M’MOF-3a (24.03). The high C2 H2 /C2 H4 separation potential of UTSA-100 is further supported by simulated and measured gas penetration experiments. Theoretical calculation and single-crystal structure analysis show that this selective adsorption characteristic of UTSA-100 is attributed to the shape and size of the pore and the NH2 group on the pore wall (Figure 15.27b). The weak acid–base interaction between adsorbed C2 H2 and NH2 group is considered to have a significant function in the selective adsorption of C2 H4 /C2 H2 . In 2016, Cui et al. had successfully assembled Cu2+ , organic connectors with SiF6 2− and obtained a series of ternary MOFs. Through the effective control of the coordination polymer of the column layer structure, ethylene and acetylene were separated based on pore size and pore chemistry [33]. The SiF6 2− unit in the material has a strong hydrogen bond with acetylene. Combined with the pore size and the chemical distribution in the pore, the energies associated with the host–guest binding interactions result in high adsorption capacity (2.1 mmol/g at 0.025 bar), and the separation selectivity of acetylene reaches the highest value in the literature (from 39.7 to 44.8). At the same time, it was found that the adsorbed acetylene molecules have a specific synergistic effect with porous materials and acetylene molecules, which makes the materials possess high selectivity and high adsorption capacity simultaneously. The selective adsorption mechanism of acetylene on the obtained materials was also determined by neutron diffraction and computational simulation. The ultimate step for any separation is molecular sieving, which allows the full isolation of a component from other components on the basis of the size criterion and enables the achievement of infinite selectivity. In 2017, Chen, Zaworotko, and coworkers reported a new variant of SIFSIX-2-Cu-i, SIFSIX-14-Cu-i (also termed as UTSA) [34]. In which the organic linker was turned to 4,4′ -azopyridine (azpy, 9.0 Å)

b c

(a)

(b)

Figure 15.27 The pore structure of UTSA-100 and the C2 H2 binding site. (a) The pore structure showing the zigzag channels along the c-axis and the cage with a diameter of about 4.0 Å in the pore wall with window openings of 3.3 Å. (b) The acetylene sits right at the small cage connecting two adjacent channel pores. (Multiple-point interactions of the acetylene molecule with framework: d [O(–CO2 )…H(C2 H2 )] = 2.252 Å, d [H(–NH2 )…(C2 H2 )] = 2.856 Å.). Source: From Hu et al. [32]. Licensed Under CC BY 4.0.

911

912

15 Relationship Between Structure and Separation Property

(d) Small apertures

(b) (a)

Cavities

Loadting N

C2H2/C2H4 9.0 Å

N

b

N

azpy

UTSA–200a ⊃ C2H2

(c)

N

a UTSA–200a

a UTSA–200a ⊅ C2H4

b

c UTSA–200a ⊃ C2H2

Figure 15.28 Structure description of UTSA-200a. (a) The channel structure of UTSA-200a reveals a pores size of approximately 3.4 Å. (b) DFT-D-calculated C2 H2 adsorption models in UTSA-200a, revealing that this pore size enables the passage of C2 H2 molecules. (c) Simulated C2 H4 adsorption in UTSA-200a, indicating that the C2 H4 molecule is too large to pass through the pores. (d) Schematic illustration of ideal molecular sieves based on the structure of UTSA-200a ⊃ C2 H2 in which larger cavities suitable for strongly binding C2 H2 molecules are interconnected by narrow apertures that serve as sieves for C2 H4 but not for C2 H2 . The different nets are highlighted in gray and purple for clarity. Color code: Cu, turquoise; Si, dark green; F, red; N, blue; C, gray; and H, green spheres. Source: From Li et al. [34].

that is shorter than 4,4′ -dipyridylacetylene (dpa, 9.6 Å). A UTSA-200 possesses doubly interpenetrated nets isostructural to the SIFSIX-2-Cu-i nets (Figure 15.28). The pore sizes of desolvated structure (UTSA-200a) were reduced considerably to 3.4 Å, which were much smaller than the kinetic diameter of C2 H4 molecule (4.2 Å) but slightly larger than that of C2 H2 (3.3 Å), consistent with the potential for selective molecular sieving in C2 H2 /C2 H4 separations. At 298 K up to 1 bar, one-component equilibrium adsorption isotherms for C2 H2 and C2 H4 were calculated. The results indicated that UTSA-200a has a steep and high C2 H2 uptake of 116 cm3 /cm3 at 298 K and 1 bar, which offered a new benchmark for C2 H2 uptake at 0.01 bar even higher than Mg- and Fe-MOF-74. Moreover, the smaller static pore size of UTSA-200a can inhibit C2 H4 molecules from entering entirely below 0.2 bar and has minimal uptake (≈0.25 mmol/g) up to 0.7 bar. Even up to 1 bar, the C2 H4 uptake amounts of UTSA-200a were still very low. Such results showed that the UTSA-200a contracted pore size is capable of blocking C2 H4 molecules efficiently without sacrificing its high C2 H2 adsorption capacity, making UTSA-200a a perfect choice for C2 H2 /C2 H4 separation at ambient conditions. At 1 bar and 298 K, UTSA-200a showed an exceptionally IAST selectivity of over 6000 for binary C2 H2 /C2 H4 (1/99, v/v) mixtures, which is considerably better than the previous high-performance materials. The UTSA-200a is the first example of a porous material that completely overcomes the trade-off between selectivity and

15.3 Separation of Hydrocarbons

uptake capacity in which it shows not only the unprecedented high selectivity but also records the high uptake capacity in the context of C2 H2 /C2 H4 separation. As aforementioned, some MOFs exhibit high performance in the separation of C2 H6 /C2 H4 or C2 H2 /C2 H4 . However, in practical use, one-step removal of various impurities, particularly trace impurities, is more important. In 2019, Chen et al. developed a synergistic sorbent separation approach for the one-step production of polymer-grade C2 H4 with a series of physisorbents in a packed-bed geometry from ternary (C2 H2 /C2 H6 /C2 H4 ) or quaternary (CO2 /C2 H2 /C2 H6 /C2 H4 ) gas mixtures [35]. Contrast to present technology, this synergistic sorbent separation technology (SSST) has at least three advantages: one-step process, ambient temperature and pressure, and low energy spent for regeneration. First, the authors chose and synthesized ultra-selective microporous materials, which can be quickly regenerated, including Zn-atz-ipa (atz = 3-amino-1,2,4triazolate; ipa = isophthalate), TIFSIX-2-Cu-i (TIFSIX = TiF6 ; 2 = 4,4′ dipyridylacetylene; i = interpenetrated), and SIFSIX-3-Ni (SIFSIX = SiF6 2− ; 3 = pyrazine). Then by adjusting the ratio of these three MOFs and the order of the packing column, high purity of C2 H4 can be obtained from the mixture of C2 H2 /C2 H6 /C2 H4 or the mixture of CO2 /C2 H2 /C2 H6 /C2 H4 . This work takes benefit of the variations in pore geometry and pore chemistry of three ultra-microporous sorbents to tackle one-step C2 H4 purification using SSST. The selection of task-specific ultra-selective sorbents is unlikely to be restricted to the three sorbents or target gas examined in tandem-packed sorbent beds of the type employed herein. To maximize overall efficiency, sorbents with higher selectivity, higher uptake capacity, or both could probably be replaced. The SSST success in purifying gas mixtures of C2 and the growing availability of ultra-selective physisorbents indicates that the scope of SSST is possibly wide enough to handle many industrial commodity purifications with a high energy footprint.

15.3.2 Separation of C3 Hydrocarbons The mechanism of separation for propane/propylene mixture is similar to that of ethane/ethylene based on MOF. The HKUST-1 MOF was first used for adsorption separation of propane and propylene, and subsequently, HKUST-1 was widely studied for this separation. Similar to its separation performance for ethylene/ethane, this material has more excellent bonding and ability to propylene in multiple pressure ranges than propane (at zero coverage of 41.8 and 28.5 kJ/mol) [36]. The GCMC simulation indicated that propane was mainly adsorbed in the octahedral pocket of the framework, while propylene mainly interacted with the open Cu site. UV-vis spectra can also prove this fact. Through the coordination of ethylene with Cu, the d–d transition of Cu moves from 540 nm to low energy band. Propane has little effect on UV-vis spectra, which indirectly proves that propane is mainly adsorbed at the octahedral cage site of the framework. Based on the above results, the successful separation of propane and propylene by HKUST-1 is very promising. Many researchers analyzed the specific properties and mechanical stability of this material

913

914

15 Relationship Between Structure and Separation Property

and used it as an industrial adsorbent for plastic pressing. The results showed that HKUST-1 can still maintain high activity after molding, pressing, and extrusion. The M-MOF-74 has a highly dense coordination unsaturated metal sites over HKUST-1. Most of the M-MOF-74 compounds have been used in the study of propane and propylene separation [25, 37]. These studies showed that the separation degree of propane and propylene is 3 : 20 for different equivalent M-MOF-74 (Figure 15.29). The separation degree of these materials increases with the increase of pressure mainly because the interaction between the adsorbed materials becomes highly significant. At 318 K and 1 bar, Fe-MOF-74 can separate the same amount of propylene/propane mixture into a single component with a purity of more than 99%. Another iron-based MOF, MIL-100 (Fe), can separate propylene from propane by reducing part of the Fe3+ center [38]. The presence of Fe2+ can increase the adsorption heat of ethylene from 30 to 70 kJ/mol. The ZIFs can also effectively separate propane from propylene by sieving mechanism, which is confirmed by the size exclusion study of ZIF-8 [39]. Similar to ethylene separation, ZIF-7 showed the opposite selectivity at 373 K, and the ability to adsorb propane was stronger than that of propylene, which was considered to be due to the “open door effect” [22]. This effect makes propylene prefers to gather on the outer surface of the ZIF-7, delaying the pressure required to “open the door.” Other MOFs also exhibit specific adsorption selectivity for propane and propylene because of “open door effect” and kinetic effect, shape selectivity and pore size limitation. The KAUST-7, a highly stable fluorinated MOF material, has been utilized to separate CO2 /N2 mixture as aforementioned. Cadiau et al. applied KAUST-7 to separate propylene and propane for the first time in 2016 [40]. The MOF is constructed by Ni2+ , pyrazine, and inorganic NbOF5 . In addition to serving as a pillar to connect the 2D grid of Ni-pyrazine, the larger volume of (NbOF5 )2− performs a more fundamental part in deciding the pore size and opening size of MOF. Compared with the previous MOF with (SiF6 )2− as the column, the pore size and maximum opening size of KAUST-7 were significantly smaller, whereas the effective pore size was 3.05 Å and the maximum opening size was 4.75 Å. Because of its small pore size, KAUST-7 has excellent selective adsorption of propylene and propane, and at 298 K and 1 bar, the material can only adsorb propylene as confirmed by the measured adsorption heat. Besides, it was found that KAUST-7 has high stability, adsorption, and desorption cycle. All these merits provide a basis for the separation of propane and propylene at room temperature and pressure by utilizing KAUST-7 as a separation material. The C3 H6 is a primary olefin feedstock material for petrochemical manufacturing, inside which trace C3 H4 (1000 or 10 000 ppm) are often mixed. Thus, separation C3 H4 from this raw material is one of the most essential processes for the production of polymer-grade C3 H6 gas (the C3 H4 impurity should be less than 5 ppm). However, separation of C3 H4 /C3 H6 is difficult and challenging due to their closer molecular sizes. The difference in kinetic diameter between C3 H4 and C3 H6 (4.2 and 4.6 Å) is much closer (nearly 0.4 Å) than C2 H2 /C2 H4 (Figure 15.30), which indicated that the separation would be more difficult.

15.3 Separation of Hydrocarbons

8

(a)

Gas uptake (mol/kg)

6

4

2

6 5 4 3 2 1 0 0.000

0.005

0.010

0.6

0.8

1.0

0 0.2

0.0

0.4

Pressure (bar) 50

(b)

C3H6/C3H8 selectivity

40 Co-MOF-74 Mn-MOF-74 30

Mg-MOF-74

20

10

0 0.0

0.2

0.4

0.6

0.8

1.0

Pressure (bar)

Figure 15.29 (a) Adsorption isotherms for C3 H6 (circles) and C3 H8 (triangles) in Co-MOF-74 at 298 K. The inset shows the low-pressure isotherms. (b) The IAST-predicted C3 H6 /C3 H8 selectivities for equimolar mixture adsorption in a series of M-MOF-74 materials at 298 K as a function of pressure. These were calculated using the experimental isotherms. Source: Based on Yoon et al. [38].

915

15 Relationship Between Structure and Separation Property

4.84 Å

3.28 Å

3.32 Å

C2H2

Increased separation complexity

4.18 Å

3.34 Å

5.70 Å

C2H4

Kinetic diameter difference: 0.9 Å 6.44 Å

4.01 Å

4.65 Å

6.51 Å

6Å 4.1

6Å 4.1

C3H4

Figure 15.30 Comparison of molecular size and kinetic diameter difference of C2 H2 /C2 H4 and C3 H4 /C3 H6 . Source: From Li et al. [41].

C3H6

Kinetic diameter difference: 0.4 Å

In 2018, Chen, Li, and coworkers investigated the capacity of diverse MOFs for separation of C3 H4 /C3 H6 mixture and found UTSA-200 that have been reported by Chen and coworkers in the separation of C2 H2 /C2 H4 exhibits the highest performance [41]. In comparison to other reported MOFs in their work, UTSA-200 possessed the narrowest pore size, which played a key role in the separation of C3 H4 /C3 H6 . As shown in Figure 15.31, the sequence of pore sizes for the MOFs are UTSA-200 (3.4 Å) < SIFSIX-3-Ni (4.2 Å) < SIFSIX-2-Cu (4.4 Å) < SIFSIX-1-Cu

Ultrafine tuning of pore size N

N

N N N N

N

N N

N

C3H4

50

C3H6 0.2

0.4 0.6 0.8 Pressure (bar)

1.0

(3.4 Å) UTSA-200 80

60

60

40

40 C3H4

20

C3H6 0 0.0

0.2

0.4 0.6 0.8 Pressure (bar)

1.0

20

C3H4 C3H6

0 0.0

0.2

0.4 0.6 0.8 Pressure (bar)

1.0

Adsorption (cm3/g)

100

Adsorption (cm3/g)

80

150

Adsorption (cm3/g)

200

0 0.0

(4.2 Å) SIFSIX-3-Ni

(4.4 Å) SIFSIX-2-Cu-i

(8.0 Å) SIFSIX-1-Cu

(a) Adsorption (cm3/g)

916

60 40 20 0 0.0

C3H4 C3H6 0.2

0.4 0.6 0.8 Pressure (bar)

(b)

Figure 15.31 (a) The pore-aperture and pore-chemistry of SIFSIX materials. (b) Their associated C3 H4 and C3 H6 adsorption isotherms at 298 K. Source: From Li et al. [41].

1.0

15.3 Separation of Hydrocarbons

(8.0 Å). All these MOFs exhibited the steep adsorption of C3 H4 at the low-pressure region over C3 H6 for their strong binding toward C3 H4 and led to the benchmark selectivity reported so far. Both C3 H4 and C3 H6 can pass these MOFs pores except UTSA-200 with small pore size, which limited their high gas selectivities and enhanced UTSA performance. The UTSA exhibits a high C3 H4 uptake but minimal C3 H6 adsorption at the low-pressure region, providing the opportunity to be the right candidate for the separation of C3 H4 /C3 H6 . At 0.01 bar, the capture capacity of C3 H4 is increased in the sequence of SIFSIX-1-Cu < SIFSIX-2-Cu-i < SIFSIX-3-Ni < UTSA-200. The UTSA-200 has set a new benchmark at 0.01 and 0.001 bar compared to other high-performance products, which renders it the most competitive substance for the elimination of trace C3 H4 . More significantly, it is possible to improve the sieving influence of C3 H6 by raising the temperature to 318 K, while retaining the capacity for the low-pressure C3 H4 uptake, signaling its clear potential to separate C3 H4 /C3 H6 at a broader operating temperature. Breakthrough tests confirmed the ability of UTSA-200 to entirely extract trace C3 H4 from 1/99 as well as 0.1/99.9 (v/v) mixtures, attaining 99.999 9% purity high record of C3 H6 production scale. The finding shows that the flexibility of the system in MOFs can be used to address some very demanding gas separations and meet in the future the potential of burgeoning microporous MOFs for gas separation.

15.3.3 Separation of Long-Chain Hydrocarbon Compared with short olefins and paraffins, the separation of long-chain hydrocarbons (the numbers of C ≥ 4) is more complicated because there are many factors, such as flexibility, geometry, hydrophobicity of long carbon chain and the interaction between carbon chain and the double bond π electron cloud. Nevertheless, the sieving mechanism can still play a role. Based on the potential separation capacity of ZIF-7, its adsorption of trans-2-butene is much smaller than that of n-butane, 1-butene, and cis-2-butene [42]. Because of the spatial flexibility and molecular surface area of long-chain alkanes, the adsorption capacity of long-chain alkanes is much larger than that of their corresponding olefins. This rule was revealed by an MOF without open metal sites, which indicated that the adsorption capacity of n-octane was much higher than that of 1-octene [43]. Other studies showed that the interaction between an MOF, isomorphic to MOF-5, and n-hexane is stronger than that of 1-butyl. Among them, the interaction between MOF-5 and n-hexane was the largest but gradually decreased with the length of the linkers. The correlation model can also show that the adsorption capacity of n-hexane in MOF-5 is slightly higher than that of 1-butyl. It was mentioned earlier that unsaturated metal sites guide a critical function in the separation of C2 and C3 light alkenes, and this effect also exists for the separation of long-chain alkanes from olefins. For example, the adsorption of pentene by HKUST-1 is larger than that of pentane, and this mechanism is also used for the separation of isobutane and isobutene. The adsorption heats of isobutane and isobutene are measured to be 42 and 46 kJ/mol, respectively [44].

917

918

15 Relationship Between Structure and Separation Property

The mechanism of selective adsorption of long-chain alkanes and olefins by MIL-53 and MIL-125 is complex and no longer depends on open metal sites. For example, the adsorption and separation of n-hexane and 1-hexene by amino-MIL-53 (Al) do not follow this model, and n-hexane has a stronger interaction with MOF [45]. Besides, n-butane has also been found to interact stronger than 1-butyl with the MOF but the reported experimental details are not very satisfactory. The unexpected results do not exist only in the MIL-53 MOF series but the selective capture of isoprene by MIL-125 and amino-MIL-125 is much than that of 2-methylbutane. In addition to the separation of olefins from corresponding alkanes, there are isomers in long-chain hydrocarbons, which do not need to be finely classified when they are used as fuels, but as unique chemical raw materials, and the guarantee of purity has a significant impact on the final product. Therefore, the separation of these isomers is particularly essential. Eddaoudi and cowokers designed and constructed a new rare earth RE-fcu-MOF using the strategy of co-linking chemistry [46]. Through careful selection of organic connectors, the pore size of MOF has adjusted accurately, and the window size of RE (Y3+ and Tb3+ ) fcu-MOF is 4.7 Å (Figure 15.32). The adsorption of butane and pentane isomers with side chains can be “cut off”: the adsorption capacity of these alkanes with branched chains is almost zero, while n-butane and n-pentane have excellent adsorption properties (Figure 15.33). Therefore, it has a very excellent separation effect on the isomers of this kind of alkanes. It was reported that butene could be adsorbed by a variety of materials and the amount of trans-butene adsorbed by compound Cu(hfipbb)⋅(H2 hfipbb)0.5 at 303 K is more than 40% of cis-butene due to its volatility. Whereas in ZIF-7, cis-butene is more suitable for entry into the pores. Interestingly, both isomers of butene are not

(b)

(a)

+

8.4 Å

6.9 Å

5.0 Å

(c)

~6 Å

Equilibrium –based adsorption

–5 Å

Kinetically driven separation

4.7 Å

Molecular exclusion

Figure 15.32 Schematic representation of the components of the RE fcu-MOFs platform, including three possible organic linkers of different lengths. (a) The RE fcu-MOFs consist of two types of cages one with octahedral shape and the other with tetrahedral shape. (b) Both of them accessible through triangular windows. (c) The judicious choice of the organic building block allowed to isolate the new fcu-MOF with the smallest aperture. Source: From Assen et al. [46].

15.3 Separation of Hydrocarbons

2.2 2.0

Adsorbed amount (mmol/g)

1.8 1.6 1.4 1.2 Y-fum pentane ads. @293 K Y-fum pentane des. @293 K Y-fum lsopentane ads. @293 K Y-fum lsopentane des. @293 K

1.0 0.8 0.6 0.4 0.2 0.0 0

50

100

150

200

250

300

350

400

450

Pressure (Torr) 2.2 2.0

Adsorbed amount (mmol/g)

1.8 1.6 1.4 1.2

Y-fum butane ads. @293 K Y-fum butane des. @293 K Y-fum lsobutane ads. @293 K Y-fum lsopentane des. @293 K

1.0 0.8 0.6 0.4 0.2 0.0 0

100

200

300

500 400 Pressure (Torr)

600

700

800

Figure 15.33 Adsorption isotherms for the MOF collected at 293 K. (a) n-Pentane and isopentane. (b) n-Butane and isobutane. Source: From Assen et al. [46].

readily adsorbed by MIL-53 (Al) or MIL-47, even though the pores of these two MOFs are wide enough to fit them as objects [47]. For n-hexane solution, HKUST-1 is more likely to adsorb cis-butene than trans-butene, and the separation coefficient is 1.9 : 1. cis-Butene is generally more easily adsorbed by molecular sieves because of the facile formation of π complexes with cations outside the framework, which is the same mechanism for HKUST-1.

919

920

15 Relationship Between Structure and Separation Property

After adsorbing one butene, every four Cu points to one pore, that is, every four Cu absorbs one butene molecule. These results are beneficial to such industrial separation problems, which are almost impossible to solve at present. In addition to the separation of cis–trans isomerized olefins, MOFs also have a specific separation effect on olefin isomers with double bonds at different positions. For example, the adsorption of 1-butene by ZIF-7 is 25% less than all isomers of 2-butene. These findings provide some help for the separation of these hydrocarbons in the future. Long and coworkers reported a highly stable MOF material, Fe2 (BDP)3 , which exhibited good shape selectivity for hexane isomers [48]. An Fe2 (BDP)3 was obtained in the form of black microcrystals from the reaction of iron acetylacetone(III) and ligand H2 BDP in DMF. In Fe2 (BDP)3 , each Fe (III) atom coordinates with six N atoms from six BDP2− ligands in an octahedral configuration. Two N atoms on each pyrazole five-membered ring on the ligand bridge two-connected Fe (III) atoms at the same time to form a 1D FeN6 chain. Through the interconnection of the backbone of the ligand BDP2− , the 3D frame structure of Fe2 (BDP)3 is formed. An Fe2 (BDP)3 showed high chemical and thermal stability. On boiling in aqueous solution from pH = 2 to 10 for two weeks or heating in air at 280 ∘ C, Fe2 (BDP)3 does not lose its crystalline state. After removing the object, the BET specific surface area of Fe2 (BDP)3 is 1230 m2 /g. The adsorption isotherms of five hexane isomers of Fe2 (BDP)3 were measured at high temperatures (130, 160, and 200 ∘ C). The results showed that although all the five isomers could be adsorbed by Fe2 (BDP)3 , and the adsorption amount is the same, the adsorption forces between the five isomers and the Fe2 (BDP)3 are different. The adsorption heat calculation showed that the adsorption force between straight-chain n-hexane and Fe2 (BDP)3 frame is the strongest. They believed that this is because n-hexane interacts with the triangular pore surface of Fe2 (BDP)3 to a greater extent than its other four isomers. The adsorption heat of the other four branched-chain isomers was significantly lower, especially the bi-branched 2,3-dimethylbutane and 2,2-dimethylbutane. Breakthrough experiments of five isomers based on Fe2 (BDP)3 were carried out at 160 ∘ C with N2 flow. The results showed that the bi-branched chains of 2,3-dimethylbutane and 2,2-dimethylbutane are the first to penetrate, followed by the isomer of a single branched-chain, and finally n-hexane. They believed that this form of separation is extremely advantageous in the application. The most penetrating branched-chain isomer has a high RON value and can be collected directly as a product. Then the penetrating n-hexane can be recycled into the catalytic isomerization system for further isomerization. Also, the separation process can be carried out at high temperatures, which can be connected to the catalytic isomerization reaction, eliminating the step of cooling and reheating and reducing the energy loss of the separation process.

15.3.4 Separation of Alkane with Different Carbon Atoms Alkanes with different carbon chains often exist together in crude oil or petroleum products. Effective separation of alkanes plays a vital role in further chemical

15.3 Separation of Hydrocarbons

production and application. The separation of different alkanes based on MOFs needs more extensive theoretical support and research. In MOF-5 and PCN-6′ , the aggregation effect below the critical temperature is found. If one component has an aggregation effect at the separation temperature and the other components do not, then this effect is beneficial for the separation. This trend is further found in strong interactions, accompanied by a decrease in degrees of freedom at the cost of entropy. The MOF-5 is often used as a research platform for the separation of short-chain alkanes because of its high specific surface area and good thermal stability. The theoretical study shows that because of the interaction with van der Waals of the pore wall, the corresponding adsorption heat and entropy will be more negative with the increase of alkane carbon chain [49]. The adsorption step in the adsorption curve will also be seen in short-chain alkanes with the rise of pressure, and the adsorption amount of adsorbed material will also increase (Figure 15.34). In the mixture, with the increase of the carbon chain, there will be strong adsorption before the selectivity reaches a maximum. In MOF-5, the adsorption heat of n-butane is almost twice as high as that of methane (−23.6 and 10.6 kJ/mol). The same is true in HKUST-1, where the adsorption heat of n-butane and methane is 29.6 and 12.0 kJ/mol, respectively. Similar conclusions have been obtained from many calculations of HKUST-1 for the separation of low alkanes. These differences show that for alkanes with different carbon chain lengths, MOFs can separate them well in theory. The adsorption capacity of MIL-47 for methane, ethane, propane, and butane increased with the increase of the carbon chain. Combined with theory and experiment, it was found that these alkanes did not show the primary action site at low loading. However, as the loading increases, methane and ethane tend to be close to μ2 -O groups, while n-butane is close to aromatic connectors. Propane and ethane can rotate freely without a particular spatial orientation, while butane is more fixed. The capacity of ZIF-8 for these alkanes is methane > ethane > propane > butane. In the case of ZIF-7, on the contrary, the “open door” pressures of ethane, propane, and butane at 298 K were 0.12, 0.012, and 0.008 bar, respectively. These results show that the adsorption capacity is inversely proportional to the door pressure. The compound Ni8 (5-bbdc)6 (μ3 -OH)4 (bbdc2− = 5-tert-butyl-1,3-phthalic acid) can Figure 15.34 Simulation of adsorption curves of different alkanes in MOF-5. Source: From Jiang and Sandler [49].

30 C1

ρ (mmol/g)

25

C2

20

C3 n C4

15

n C5

10 5 0 10–1

100

101

103 102 F (kPa)

104

105

921

922

15 Relationship Between Structure and Separation Property

also distinguish methane from ethylene or distinguish propane from butane by door-open effect through carefully controlling the temperature. This means it is likely to extract methane from ethane, propane, and butane [50]. In brief, some tendencies can be found from the adsorption of short-chain alkanes in MOFs, and adsorption steps can be observed, which correspond to different pressures during adsorption. Furthermore, the molar capacity of short-chain alkanes is relatively high, but the adsorption capacity is low, and the small size enables more molecules for entering the pores, but the smaller molecular surface area is not conducive to the interaction with the pore wall. Due to the existence of a variety of adsorption steps and a variety of adsorption thermodynamic behaviors, it can be assumed that MOFs can separate many linear alkane combinations, not only those reported in the literature. It is also important for olefin separation with different carbon atom numbers, because in the preparation of these olefins, catalytic dehydrogenation is often random, and the resulting products are often mixtures of different olefins. The separation of olefins with different carbon atomic numbers is also a difficult problem in the chemical industry. The adsorption of these hydrocarbons by MOFs can help solve this problem. For example, for HKUST-1, the specific adsorption capacity is cis-2-butene > 1-pentene > trans-2-pentene. It can be found that MOFs have a good adsorption and separation effect for short-chain hydrocarbon molecules, including not only the separation of saturated hydrocarbons but also the separation of unsaturated hydrocarbons and various isomers. Of course, the current MOF research is mainly in the laboratory stage, and perhaps more efforts need to be invested in industrial separation-related processes. In addition, the effects of water vapor and other complex components that need to be separated in actual production on the separation effect of MOFs should also be considered.

15.4 Separation of Noble Gases Noble gases have been applied to produce many critical industrial devices, for example, xenon is widely used in lighting, laser, and therapeutic areas in our daily lives, and krypton also plays a role in fluorescent lamps. These two noble gases are naturally accessible in the atmosphere, where the contents of Xe and Kr are very low. For example, concentrations of only 0.09 and 1 ppm are detectable in the air for Xe and Kr, respectively. The separation of these two noble gases is generally based on the cryogenic distillation method for their distinguished boiling points (165 K for Xe and 120 K for Kr), which is currently commercially used and serves as a high energy cost process. Besides, unclear powder has become an essential part of the global power system, whereas it should pay attention to the problem of off-gas treatment resulting from nuclear plants. Radioactive Kr and Xe can be produced during the fission of uranium and plutonium, with approximately 10 mol of Xe are created throughout fission for each mole of Kr. However, the half-life of radioactive Kr is much longer than that of Xe (10.8 years versus 36.3 days). Either to mitigate the volume of stored waste gas or to recover Xe for additional commercial

922

15 Relationship Between Structure and Separation Property

also distinguish methane from ethylene or distinguish propane from butane by door-open effect through carefully controlling the temperature. This means it is likely to extract methane from ethane, propane, and butane [50]. In brief, some tendencies can be found from the adsorption of short-chain alkanes in MOFs, and adsorption steps can be observed, which correspond to different pressures during adsorption. Furthermore, the molar capacity of short-chain alkanes is relatively high, but the adsorption capacity is low, and the small size enables more molecules for entering the pores, but the smaller molecular surface area is not conducive to the interaction with the pore wall. Due to the existence of a variety of adsorption steps and a variety of adsorption thermodynamic behaviors, it can be assumed that MOFs can separate many linear alkane combinations, not only those reported in the literature. It is also important for olefin separation with different carbon atom numbers, because in the preparation of these olefins, catalytic dehydrogenation is often random, and the resulting products are often mixtures of different olefins. The separation of olefins with different carbon atomic numbers is also a difficult problem in the chemical industry. The adsorption of these hydrocarbons by MOFs can help solve this problem. For example, for HKUST-1, the specific adsorption capacity is cis-2-butene > 1-pentene > trans-2-pentene. It can be found that MOFs have a good adsorption and separation effect for short-chain hydrocarbon molecules, including not only the separation of saturated hydrocarbons but also the separation of unsaturated hydrocarbons and various isomers. Of course, the current MOF research is mainly in the laboratory stage, and perhaps more efforts need to be invested in industrial separation-related processes. In addition, the effects of water vapor and other complex components that need to be separated in actual production on the separation effect of MOFs should also be considered.

15.4 Separation of Noble Gases Noble gases have been applied to produce many critical industrial devices, for example, xenon is widely used in lighting, laser, and therapeutic areas in our daily lives, and krypton also plays a role in fluorescent lamps. These two noble gases are naturally accessible in the atmosphere, where the contents of Xe and Kr are very low. For example, concentrations of only 0.09 and 1 ppm are detectable in the air for Xe and Kr, respectively. The separation of these two noble gases is generally based on the cryogenic distillation method for their distinguished boiling points (165 K for Xe and 120 K for Kr), which is currently commercially used and serves as a high energy cost process. Besides, unclear powder has become an essential part of the global power system, whereas it should pay attention to the problem of off-gas treatment resulting from nuclear plants. Radioactive Kr and Xe can be produced during the fission of uranium and plutonium, with approximately 10 mol of Xe are created throughout fission for each mole of Kr. However, the half-life of radioactive Kr is much longer than that of Xe (10.8 years versus 36.3 days). Either to mitigate the volume of stored waste gas or to recover Xe for additional commercial

15.4 Separation of Noble Gases

usage, Xe must be separated from Kr. For energy-saving and recycling purposes, it is desirable to develop alternative adsorption and separation methods utilizing porous solids at ambient temperature and pressure. Application of MOFs in the separation of Xe/Kr are investigated for more than 10 years, while it remains limited. Only a handful of cases are reported in the literature. Here, we briefly summarize the representative MOFs, and the relationship between separation behavior for Xe/Kr and their structure is discussed below.

15.4.1 HKUST-1 Two parameters can be modified for noble gas separation: pore size and interactions between porous host and noble gases. These parameters are selected on the basis that noble gases specially possess complex diameter and polarizability. The HKUST-1 is an ideal replacement for gas-adsorption applications for its structural nature. There are small and large cavities of internal diameter approximately 1.3, 1.1, and 0.5 nm, and plenty of available open Cu(II) metal sites found within its framework. Furthermore, it can be easy for preparation on an industrial scale from cheap, marketable starting materials. In 2006, Mueller et al. first used HKUST-1 as a filler for separating rare gas mixture through breakthrough experiments [51]. At 55 ∘ C and 40 bar, a mixture of Kr (≈94 mol%) and Xe (≈6 mol%) was continuously fed to an isothermal tubular reactor filled with HKUST-1. In the leaving stream, only 50 ppm level of Xe was recorded, which indicated the preferential adsorption of Xe by the packed bed of HKUST-1, while the majority of Kr can pass through. The estimated HKUST-1 capacity for Xe is over 60% (twice) of a high-surface activated carbon adsorbent under the same conditions. However, at room temperature and under 1 bar, the Xe capacity and the selectivity of Xe/Kr for HKUST-1 (when using three various gas compositions, i.e. Xe/Kr 20 : 80, 50 : 50, 80 : 20) are lower than those for activated carbon. This was reported by Liu et al. in 2012 through performing the dynamic breakthrough column measurement with HKUST-1 [52]. Furthermore, some analytical techniques, including GCMC simulations, 129 Xe NMR spectroscopy, and synchrotron neutron and X-ray diffraction, were conducted to locate the favored adsorption sites for Xe in HKUST-1 [53]. Figure 15.35 shows that the findings revealed the preferential Xe adsorption happened either inside the small pockets or at the surrounding windows toward the cavity. It differs from other normal gases whose primary site is located at the accessible unsaturated Cu(II) center.

15.4.2 The MOF-74 (M-DOBDC) As noted above, the MOF-74 series, as a class of powerful MOFs, was used for capture and separation of various adsorbates owing to their highly dense open metal sites. In 2012, Thallapally and coworkers carried out an adsorption experiment on NiMOF-1 and discovered Xe adsorption capacity of 55 wt% at 100 kPa and 298 K, which is analogous to that of activated carbon. However, the Xe selectivity over Kr (∼5–6) is as twice as that of charcoal [54]. Later, Perry et al. conducted a systemic

923

924

15 Relationship Between Structure and Separation Property

Figure 15.35 View of HKUST-1 showing the strongest binding site for Xe (orange atoms) at the pocket center, with two views of the small pocket, isolated (top left and bottom left) and view of one unit cell down the c-axis (right). Source: From Hulvey et al. [53].

study on the effect of metal types tuning in MOF-74 for Xe/Kr selectivity [55]. In their experimental research, Xe/Kr selectivity exhibited a decreasing trend as follows Co > Ni > Mg > Zn. However, in calculated results, the adsorption performance and selectivity were highly unchanged by metal center variation, as the formal charge, in addition to the nature of the interaction across the series, stays the same. These results indicated that the interaction between noble gases and the porous host is relatively weak and did not display different topologies and pore sizes, even for the powerful MOF-74. To find higher adsorption performance and selectivity of noble gases based on MOF, computational study could help design and construct MOFs with a rational aperture structure.

15.4.3 SBMOF-1 Calcium-based MOF, originally produced for the separation of CO2 /N2 , is SBMOF-1 (also called CaSDB, SDB = 4,4-sulfonyldibenzoate). In 2016, Benerjee et al. predicted Xe/Kr selectivity of approximately 5000 existing MOFs and approximately 120 000 hypothetical MOFs by high-throughput computational screening. Among their calculated results of existing MOFs, SBMOF-1 possesses the highest thermodynamic selectivity of approximately 70.6 [56]. It should be noted that the above selectivity is feasible at dilute conditions applicable to used nuclear fuel reprocessing off-gas. Measured single component equilibrium adsorption, kinetic adsorption experiment, as well as column breakthrough were conducted, respectively, to test the separation performance of SBMOF-1. Figure 15.36 shows that the pure component adsorption of Xe and Kr revealed large Xe Henry coefficient and high Xe/Kr selectivity (≈16) exhibited by SBMOF-1 under dilute conditions among all recorded Xe and Kr adsorption isotherms before that work. The authors subsequently performed single-column breakthrough experimentation with a representative gas mixture (400 ppm Xe, 40 ppm Kr, 78.1% N2 , 20.9% O2 ,

Xe/Kr selectivity

1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0

Kr Xe 0.0

0.2

(a)

0.4 0.6 Pressure (bar)

0.8

1.0

0.8

40 39

0.6

38

0.4

37 0.2

36 0

10

20 30 40 Time (min)

50

60

40

SBMOF-1 CC3 Co-formate SBMOF-2 HKUST-1 MOF-505++ PCN-14++ Ni-MOF-74 Zinc tetrazolate IRMOF-1++ FMOF-Cu+

1.4 Xe uptake (mmol/g)

Weight (mg)

5 10 15 20 25 30 35 Xe Henry coeffcient (mmol/g/bar)

(b)

41

(c)

0

1.0

42

35

18 16 14 12 10 8 6 4 2 0

Pressure (bar)

Gas uptake (mmol/g)

15.4 Separation of Noble Gases

0.0

1.2 1.0 0.8 0.6 0.4 0.2 0.0 0

(d)

1

2

3

4

5 6 Cycle

7

8

9

10

Figure 15.36 Experimental characterization of Xe and Kr adsorption in SBMOF-1. (a) Experimental Xe and Kr adsorption isotherms. Horizontal line indicates one atom per pore segment. (b) Survey of thermodynamic Xe/Kr separation performance in top-performing materials. Henry coefficients are extracted from pure-component Xe and Kr adsorption isotherms reported in the literature. Data at 298 K, exceptions denoted by a dagger (w) for 297 K and a double dagger (ww) for 292 K. (c) Xe adsorption kinetics experiments. The blue curve shows the pressure drop in a chamber feeding Xe to an initially evacuated chamber with the SBMOF-1 sample; the red curve shows the corresponding weight increase due to Xe adsorption. (d) Xe adsorption/desorption cycling data; a sinusoidal curve is superimposed on the data. (a,c,d) Data for SBMOF-1 at 298 K. Source: From Banerjee et al. [56]. Licensed Under CC BY 4.0.

0.03% CO2 , and 0.9% Ar) to highlight the practical implementation of SBMOF-1 to Xe capture from nuclear fuel reprocessing off-gas. The findings revealed that Xe was held for more than one hour in the column, while all other gasses separated in minutes. The Xe adsorption capacity (13.2 mmol Xe/kg) is higher than that of the benchmark materials, and even with moisture, it is more necessary to achieve comparable capacities. From the point of structural view, the pore size of 4.2 Å for SBMOF-1 is just slightly larger than a Xe atom, approximately 4.1 Å. Such a tailored pore diameter was found to be necessary for realizing a highly selective material for Xe. Furthermore, the dense atoms wall contributed to constructing a binding site for Xe imparting high affinity. Such characteristics render SBMOF-1 potentially useful to extract Xe and Kr from nuclear-reprocessing facilities with a far lower energy demand than cryogenic distillation.

15.4.4 CROFOUR-1-Ni and CROFOUR-2-Ni The isostructural materials, CROFOUR-1-Ni and CROFOUR-2-Ni, as aforementioned, a family of hybrid ultra-microporous materials, are sustained by pillaring square grid sheets (Ni⋅(bpe)2+ , bpe = 1,2-bis(4-pyridyl)ethylene or

925

15 Relationship Between Structure and Separation Property

Ni(azp)2+ , azp = 4,4′ -azopyridine) by angular inorganic pillars CrO4 2− and presented efficient capacity to separate of CO2 and N2 . In 2016, Thallapally and coworkers studied the separation ability of Xe/Kr based on these MOFs [57]. The adsorption isotherms of Xe and Kr were collected at 298, 288, and 278 K for CROFOUR-1-Ni and CROFOUR-2-Ni. The isotherms indicated that Xe uptakes were 39.6 cm3 /g (47.1 cm3 /cm3 ) and 36 cm3 /g (47.1 cm3 /cm3 ) for CROFOUR-1-Ni and CROFOUR-2-Ni, respectively and Kr uptakes of 11 cm3 /g (13.1 cm3 /cm3 ) and 11.5 cm3 /g (14.6 cm3 /cm3 ) at 298 K and 1 bar, respectively. As shown in Figure 15.37, the steep Xe uptakes at low pressure show that the Xe high affinity over Kr is more applicable to the trace and low-Xe concentrations and separation of both materials. Clausius Clapeyron and Langmuir Freundlich methods utilizing single adsorption isotherms determined the isosteric heats of adsorption (Qst ) of Xe and Kr gases. CROFOUR-1-Ni and CROFOUR-2-Ni exhibit the high value for Xe Qst at low loading: 37.4 and 30.5 kJ/mol, in which, CROFOUR-1-Ni shows the highest value among most other kinds of porous materials, such as MOMs and activated carbon at that time. Column breakthrough and single adsorption experiments are also investigated, the selectivity for a 50 : 50 Xe/Kr binary gas mixture at 298 K and 1 bar were calculated to be 26 and 16 for CROFOUR-1-Ni and CROFOUR-2-Ni, Ni2+ +

CrO42– +

Xe Kr

40

1,2-Bis(4-pyridyl)ethene

Uptake cm3 (STP)/g

35

4,4ʹ-Azopyridine

30 25 20 15 10 5 0 0

20

40

60

80

100

80

100

Pressure (kPa) 40

Xe Kr

35

CROFOUR-1-Ni

Uptake cm3 (STP)/g

926

CROFOUR-2-Ni

30 25 20 15 10 5 0 0

HC

(a)

CH

CrO42–

HC

CH

N

N

CrO42–

+ N N

20

40

60

Pressure (kPa)

(b)

Figure 15.37 (a) The formation and structure of mmo nets; CROFOUR-1-Ni and CROFOUR-2-Ni, based on the SMCs and CrO4 2− pillars. The bottom row shows the two types of pores in each net: lined by the functionalized organic linker (C=C from bpe or N=N from azp) or the oxygen atoms from the inorganic linkers (two from each CrO4 2− ). (b) The formation and structure of mmo nets; CROFOUR-1-Ni and CROFOUR-2-Ni, based on the SMCs and CrO4 2− pillars. The bottom row shows the two types of pores in each net: lined by the functionalized organic linker (C=C from bpe or N=N from azp) or the oxygen atoms from the inorganic linkers (two from each CrO4 2− ). Source: From Mohamed et al. [57].

15.4 Separation of Noble Gases

CROFOUR-1-Ni (a)

CROFOUR-2-Ni (b)

Figure 15.38 Illustration of the adsorbed Xe atom (violet) at the primary binding site in CROFOUR-1-Ni (a) and CROFOUR-2-Ni (b) as determined from simulation. The primary binding site is located within the cage that is enclosed by three CrO4 2− ions, where the adsorbate interacts with six terminal oxygen atoms (two from three different CrO4 2− moieties) simultaneously. Atom colors: C cyan, H white, N blue, O red, Cr yellow, Ni silver. Source: From Mohamed et al. [57].

respectively, 22 and 15.5 for a 20 : 80 Xe/Kr mixture, respectively, and 21.5 and 15 for 10 : 90 Xe/Kr binary mixture, respectively. Such values also outperform the present benchmark porous material at that time. Molecular simulations indicated that the primary adsorption site for Xe in both the materials is located in the cage that involves three CrO4 2− ions in proximity to each other. The atoms of Xe at the same time interact with six terminal oxygen atoms (Figure 15.38). Nevertheless, interactions at this site in both materials between Kr and moieties are weak, probably because of the smaller polarizability of Kr.

15.4.5 [Co3 (C4 O4 )2 (OH)2 ]⋅3H2 O The record maintained by CROFOUR-1-Ni for selectivity of Xe/Kr was covered in 2019. Bao, Li, and coworkers reported a robust squarate-based MOF, which presents a record of both high affinity and selectivity for Xe over Kr [58]. The MOF was constructed by squaric acid and Co2+ , with formula as [Co3 (C4 O4 )2 (OH)2 ]⋅3H2 O(C4 O4 2− = squarate), and given perfect pore size (4.1 Å × 4.3 Å) comparable with the kinetic diameter Xe (4.047 Å). Furthermore, the surface of this ultra-micropore was decorated with very polar hydroxyl groups (Figure 15.39), making it strongly polar. At 273 K, the dehydrated framework permanent porosity was verified through CO2 adsorption isotherm, offering an apparent surface area of 95 m2 /g of the MOF. At 298 and 313 K, adsorption isotherms for a single component of Xe, Kr, N2 , Ar, and O2 were collected. The findings suggest the outstanding efficiency of the MOF at low pressures for Xe capturing. The MOF takes up 66.1 and 58.9 cm3 /cm3 at 298 and 313 K, respectively that corresponds to the gas occupancy of 1.18 and 1.05 Xe atoms per unit cell and under identical circumstances considerably higher than all previous recorded state-of-the-art MOFs. The Xe Qst at nearly zero-loading for the MOF is evaluated to be 43.6 kJ/mol, considerably greater than those of other measured MOFs. This highest value of the MOF, in comparison with recorded materials,

927

(a)

c

60

60 Xe-313 K Kr-313 K Ar-313 K N2-313 K O2-313 K

50 40 30 20 10

50 Xe uptake cm3/cm3

ba

Volumetric uptake cm3/cm3

15 Relationship Between Structure and Separation Property

40 30 20 10 0

0

0

(c)

20

40 60 80 Pressure (kPa)

CROFOUR-1-Ni CROFOUR-2-Ni Cu-BTC IRMOF-1 FMOFCu

100

0

Co3(HCOO)6 MOF-Cu-H

1

(e)

2 3 Pressure (kPa)

SB-MOF-2 1a

Ni-MOF-74

4

5

Ag@NiMOF-74

60 50

5.651

a c

(b)

O C Co

b H

Henry Xe selectivity

100 Xe/Kr IAST selectivity

928

10

1a

40

CROFOUR-1-Ni

20

CROFOUR-1-Ni PAF-45S

SBMOF-1 CC3

MOF-Cu-H

Co3(HCOO)6

10

IRMOF-1

SBMOF-2 Cu-BTC MOF-505 Ni-MOF-74

1 0

(d)

20

40 60 80 Pressure (kPa)

100

0

FMOF-Cu

0

10 20 30 40 200 Xe Henry coefficient mmol/g/bar

(f)

Figure 15.39 (a) The coordination environment of squaric acid ligand. (b) Perspective view of the structure showing the –OH groups decorated in the rhombohedral channels. (c) Experimental single-component adsorption isotherms of different gases at 313 K. (d) Comparison of the IAST selectivity of 1a versus those of previously reported best-performing materials for Xe/Kr (20/80, v/v) mixtures at varying pressures at room temperature. (e) Xe adsorption isotherms of 1a at low pressure (0–0.05 bar) at 298 K and comparison with other materials. (f) Survey of thermodynamic Xe/Kr separation performance in reported top-performing materials. Henry coefficients are extracted from single-component Xe and Kr adsorption isotherms at 298 K, except FMOF-Cu at 297 K, MOF-505 at 292 K. Source: From Li et al. [58].

demonstrated that the polar functional groups’ insertion, as well as the perfect pore size formulation within the framework, are the critical features for the high effective Xe adsorption. The calculated IAST selectivity for 20 : 80 Xe/Kr binary gas mixture is approximately 69.7 at 298 K and 100 kPa that exceeds any current benchmark porous material. The DFT-D (dispersion-corrected density functional theory) calculations are carried out to explore Xe/Kr gas mixture interaction nature within this material. Within the simulations, two separate adsorption positions are correlated with Xe or Kr, Site-I and Site-II, as shown in Figure 15.40. Higher polarization in Xe compared to Kr efficiently improved the binding energy for Xe. The optimal pore size is equally important, and the short distance confirms the close contact of the gas-host, and accordingly, Xe atoms are tightly trapped within the pore space of the MOF (Figure 15.40). Besides, short Xe…O distance indicated a strong interaction between Xe and O atom in the pore, whereas these values are all longer for Kr compared with Xe. Suitable pore size and polar –OH groups synergistically donate to the maximization of the interaction between Xe atoms and framework. Furthermore, the breakthrough experiments, cycle experiments, and chemical stability measurements are performed, and the results demonstrate that the MOF has extraordinary separation presentation, outstanding framework stability, and full regeneration capability, which makes this material a really auspicious applicant for the separation of inert gases by adsorption.

15.4 Separation of Noble Gases

2.825

2.825

(a)

(c)

3.702

3.702

3.702

3.702

(e)

2.876

2.876

3.653

3.653

3.653

3.653

a b

(b)

(d)

(f)

Figure 15.40 Illustration of the adsorbed Xe atom (glaucous) at the primary binding sites in 1a as determined from simulation. The primary binding sites are located within the cage. (a) Top view of the Xe adsorbed packing diagram. (b) Side view of the Xe atoms in the 1D rhombohedral channels with framework omitted. (c,e) The Xe atom interacts with H (Xe…H, 2.825 Å) atoms from hydroxyl group (Site I) and O (Xe…O, 3.702 Å) atoms from organic ligands (Site II). (d,f) The Kr atom interacts with H (Kr…H, 2.876 Å) atoms from hydroxyl group (Site I) and O (Kr…O, 3.653 Å) atoms from organic ligands (Site II). Source: From Li et al. [58].

15.4.6 SCU-11 Besides normal traditional MOFs, in 2018, Wang et al. employed an unsaturated Th4+ site in a new 3D porous thorium–organic framework (SCU-11), which comprises a series of cages with an efficient size of approximately 21 Å × 24 Å [59]. The Th4+ in SCU is 10-coordinated with a unique bicapped square prism coordination geometry, which has never been recorded previously for any metal cation complexes. Two coordinated water molecules occupied the bicapped position, which can be eliminated to give a unique open Th4+ site. The Th4+ site can be confirmed by X-ray structure analysis, thermogravimetry, and spectroscopy (shown in Figure 15.41). The BET of the degassed material (SCU-11-A) is about 1272 m2 /g, which is among recorded actinide materials of the highest values. The SCU-11-A was stable to water vapor and can absorb Kr and Xe with uptake capacities of 0.77 and 3.17 mmol/g and give a Xe/Kr selectivity of 5.7. This work demonstrates that actinides can contribute to developing unique MOF materials that cannot be achieved in the transition metal and lanthanide system. In other words, for the noble gas separation, there is a wide range to consider for the group of metals.

15.4.7 FMOF-Cu For the separation of Xe and Kr, most recorded MOFs selectively adsorb Xe over Kr, and these MOFs usually have a larger pore size than the Xe size. Xe is preferentially

929

930

15 Relationship Between Structure and Separation Property

(a)

Figure 15.41 (a) The coordination mode of Th(IV) occurs in SCU-11. (b) The thorium dimer in SCU-11. (c) The molecular cage in SCU-11. (d) Simplify of the cage in SCU-11. (e) The 3D framework of SCU-11. (f) The simplify of the 3D structure with cages. Atom color codes: Th, orange; O, cyan; C, blue. Source: From Wang et al. [59].

(b)

(c)

(d)

(e)

(f)

adsorbed because of the stronger van der Waals interaction between host and Xe arising from the more optimal pore or stronger polarizability of Xe. Additionally, if the pore size of MOF is between the kinetic diameters of Xe and Kr, Kr may be selectively adsorbed by the molecular sieving effect over Xe (reverse selectivity). The first example for the separation of Kr/Xe by utilizing partially fluorinated Cu derived MOF (FMOF-Cu) was reported by Fernandez et al. [60]. This material has tubular cavities (≈0.51 nm × 0.51 nm) with bottleneck windows possessing approximately a dimension of 0.35 nm × 0.35 nm. Therefore, only Kr (≈0.36 nm) is capable of entering the pore window in suitable circumstances but Xe (≈0.39 nm) is obstructed. It is worth mentioning that the adsorption behavior is temperature-dependent. Based on that, the preferential adsorption to Kr was observed only under 0 ∘ C as demonstrated in Figure 15.42, and the trend was reversed above 0 ∘ C. This pattern can be recognized when the pores narrow to stop the entrance of Xe while the pores widen to grant Xe to diffuse in the pores at a high temperature.

15.5 Separation of Hydrogen Isotopes Deuterium is a possible energy supply for nuclear fusion reactors of the next generation. It is commonly used in a range of uses such as the heavy-water nuclear reactor acting as a neutron moderator, nonradioactive isotopic tracing, and neutron scattering systems. Nevertheless, deuterium is exceptionally small in total abundance, accounting for just 0.0184% of the available hydrogen on earth

15.5 Separation of Hydrogen Isotopes Krypton-70C Krypton-40C Krypton-20C

2

Krypton-0C Krypton-24C Krypton-40C

Xenon-20C Xenon 0C 0.45 mmol/g

mmol/g

1.5

1

0.5

0

(a)

Xenon-70C Xenon-40C

0.6

Xenon 24C Xenon 40C

0.3

0.15

0 0

0.2

0.4

0.6

0.8

0

1

Pressure (bar)

(b)

0.2

0.4

0.6

0.8

1

Pressure (bar)

Figure 15.42 Kr (a) and Xe (b) sorption isotherms for FMOF-Cu at various temperatures. Note the decrease in Xe uptake below 0 ∘ C at all pressures. Source: From Fernandez et al. [60].

and approximately 156.25 ppm in the ocean. Therefore, it is crucial to enrich deuterium from the isotope mixtures for further use. However, the separation of D2 /H2 is a daunting job due to the very close features in terms of size, shape, and thermodynamic characteristics. The existing deuterium industrialization primarily depends on cryogenic distillation or heavy-water electrolysis from the Girdler sulfide process. Such methods, however, take a long time and are high in energy. Porous solids, in particular MOFs, could instead be employed for helping in the separation of H2 /D2, which, by tackling the said mentioned costs and performance problems, may become a viable alternate solution. Because of the complexity, the H2 /D2 separation studies using MOF are not as detailed and limited in comparison to other gas separations. The separation was usually often accomplished by size confinement or chemical affinity, taking into account the quantum effect. The recorded studies can be loosely divided into two groups based on the current mechanism of separation: kinetic quantum sieving (KQS) as well as chemical affinity quantum sieving (CAQS).

15.5.1 Kinetic Quantum Sieving Beenakker et al. presented the KQS method for isotope separation in 1995 [61]. They suggested that separating isotopes in nanopores is feasible when the difference between pore size (d) and molecular size (𝜎) is comparable to the de Broglie wavelength of the molecules (𝜆) if in their transverse motion molecules are limited in a particular space. Because D2 includes a de Broglie wavelength less than H2 , D2 has a significantly smaller effective particle size than H2 . Taking advantage of this slight disparity between H2 and D2 with a small enough gap, greater mobility of D2 in porous mediums can be detected with small apertures. Such faster diffusion of D2 over H2 contributes to the isolation of isotopes. The D2 adsorption was explored by several research groups after 2006 when the hydrogen storage was a hot subject, and [Cu2 (L)(H2 O)2 ] (H4 L = terphenyl-3,3′ ,5,5′ tetracarboxylic acid) and HKUST-1 have been the first finders of the distinction

931

15 Relationship Between Structure and Separation Property –3.5

–4.0 –4.5

–4.0 –5.0 –5.5

–4.5

–6.0

–5.0

–6.5 k2D2 (77.3 K) k2D2 (77.3 K) k2H2 (77.3 K) k2H2 (77.3 K) Linear regression k2D2 Linear regression k2D2 Linear regression k2H2

–7.0 –7.5 –8.0 –8.5 0.0

In(k) In(sʹ)

In(k) In(sʹ)

932

k2D2 (87.3K) k2H2 (87.3K) k2H2 (87.3K) Linear regression k2H2

–6.0

Linear regression k2H2 Linear regression k2D2

Linear regression k2H2

0.5

1.0

1.5

2.0

2.5

3.0

–6.5 0.0

Amount adsorbed mmol/g

(a)

k2D2 (87.3K)

–5.5

(b)

Linear regression k2D2

0.5

1.0

1.5

2.0

2.5

Amount adsorbed mmol/g

Figure 15.43 The variation of ln(k 1 ) and ln(k 2 ) obtained from the double exponential model with amount adsorbed for H2 and D2 adsorption on M’MOF1 (Zn3 (BDC)3 Cu(Pyen)) at (a) 77.3 K and (b) 87.3 K. Source: From Chen et al. [63].

between H2 and D2 sorption, respectively [62]. The measured D2 /H2 ratio at 78 K, nevertheless, was generally less than 1.2 equiv to porous carbon and zeolite NaA. The first systematic quantum-sieving experimental study was performed by Chen et al. in 2008 using M’MOF-1 as a mixed MOF representative material [63]. In the c-axis direction, the M’MOF-1 has a curved pore of 5.6 Å × 12 Å. As Figure 15.43 shows, the D2 adsorption rates are rapid relative to the respective H2 profiles and the increased adsorption rates can be realized by increasing the adsorbed amount. By utilizing H2 and D2 isotherm data through virial analysis, D2 -surface interactions were marginally higher than those of H2 -surface interactions, whereas D2 –D2 interactions were lower than H2 –H2 interactions. The former is due to variations in the quantum statistical mass effect on the normal vibrational energy levels to the surface as well as the dispersion energy, whereas the latter is related to the high H2 zero-point energy. Noguchi et al. have simultaneously researched the separation of D2 /H2 on the CuBOTf ionic MOF [64]. Such MOF incorporates a triflate counter anion cationic structure positioned in the channels of the pores and its 1D channel has an entrance size of 0.87 nm × 0.87 nm and 0.2 nm × 0.2 nm, respectively (Figure 15.44a). At 40 and 77 K, experimental and IAST calculations were conducted for selectivity of D2 /H2 . As shown in Figure 15.44b, the quantum H2 blocking effect is more dominant than quantum D2 , while D2 selectivity over H2 was in the range of 2.6–5.8 at 40 K. Nevertheless, the temperature range is far too small for enhanced selectivity and this is not ideal for practical implementations. In a laboratory evaluation of four microporous materials of various pore sizes, comprising two MOFs, ZIF-7 (3.0 Å) and ZIF-8 (3.4 Å) and two covalent–organic frameworks (COFs), COF-1 (9.0 Å) and COF-102 (12.0 Å), Oh et al. investigated the relationship between effective quantum screening and pore size [65]. The ZIF-8 showed a very high selectivity (≈11), at exceptionally low pressure and 20 K, while ZIF-7 displayed no noticeable H2 uptake. Based on such findings, the anticipated optimal pore aperture for achieving quantum cryo-sieving was estimated to be between 3.0 and 3.4 Å.

15.5 Separation of Hydrogen Isotopes

(a)

(b) Larger pore entrance (0.87 × 0.87 nm)

b a

P42nm a = b = 1.99335 nm c = 0.96455 nm

S(D2/H2)

Smaller pore entrance (0.20 × 0.20 nm)

Exp. CuBOTf. Exp. CuBOTf.

6

(40 K) (40 K) (77 K) (77 K)

4

2

0 10–5

10–4 Pressure (MPa)

10–3

Figure 15.44 (a) Schematic view of the frameworks of CuBOTf. Effective diameters of the pore entrances are shown with large spheres (Brown: Cu; gray: C; blue: N: red; O: yellow; S: yellow green: F; white: H). (b) S(D2/H2) in CuBOTf calculated by IAST over the pressure range 10−5 to 10−3 MPa at 77 and 40 K. Experimental values are shown in circles with line, and simulated values from CuBOTf model are shown in lines. Source: From Noguchi et al. [64].

In 2013, Teufel et al. researched MFD-4, a highly thermally and hydrolytically stable Zn-based MOF, for the separation of H2 /D2 [66]. As shown in Figure 15.45a, the structure possesses alternately next to each other organized two separate pores, which resemble the MOF-5 structure. They have a cavity size of 3.88 and 11.94 Å in diameter, linked by a small square opening of 2.52 Å gated with the atoms of Cl. Adsorption isotherms typically show reduced gas uptake at increasing temperature of adsorption. Nevertheless, the rising uptake of hydrogen isotopes with increasing temperatures was demonstrated in this research (Figure 15.45b). It was assumed that such an anomalous pattern is caused by the temperature-induced opening of the aperture, which minimizes the kinetic barrier and promotes the uptake of gas. This means that the intact crystal lattice has a high diffusion barrier suppressing the diffusion of hydrogen beyond the opening temperature of the gate because of the Pauli repulsion between H2 and the Cl atoms; nevertheless, the greater lattice vibration and the higher energy of the gas molecule enable the possibility of diffusion at higher temperatures. This reminded us of the reverse selectivity in the case of Kr/Xe separation because of the effect of molecular sieving in Section 15.4.7. An identical scenario is probably to arise when the gas molecule dimension and aperture of the pores are very proximate to each other. The identification of individual isotopes of hydrogen from their mixture is difficult, owing to analogous properties and very low molecular weight. Before this research, the selectivity of D2 /H2 is usually achieved through measurement based on single-component adsorption isotherms. In the current study, Teufel and coworkers first directly calculated the selectivity (with a 1 : 1 D2 /H2 mixture) using a customized cryogenic thermal desorption spectroscopy (TDS) device. Utilizing this method, the temperature and time-dependent selectivity have been investigated and for short exposure times of approximately 15 minutes and a maximum selectivity of up to 7.5 at 60 K and 10 mbar has been recorded.

933

15 Relationship Between Structure and Separation Property

10

50 K H2

250

50 K D2 8

60 K H2 60 K D2

200

70 K H2 6

70 K D2

150

4

100

2

50

0

Adsorption (molecules per unit cell)

(a)

Adsorption mmol/g

934

0 1E–3

0.01

0.1 Pressure (bar)

1

(b)

Figure 15.45 (a) Crystal-structure packing model of the MFU-4 framework showing the arrangement of interconnected large and small cavities (inner diameters of 11.94 and 3.88 Å) separated by narrow apertures of 2.52 Å. (b) Low-temperature adsorption isotherms for pure hydrogen and pure deuterium measured up to 1 bar at 50 K (black), 60 K (red), and 70 K (blue). The open symbols refer to hydrogen and the closed symbols to deuterium. Source: From Teufel et al. [66].

15.5.2 Chemical Affinity Quantum Sieving In comparison to the examples described previously, the leading cause of separation is the regulation of pore sizes, and CAQS is another method, primarily based on strong interactions between gas molecules and the pore surface, given by characteristics of open metal sites.

15.5 Separation of Hydrogen Isotopes

Fe-MOF-74

S(D2/H2)

5

Co-MOF-74

Ni-MOF-74

4

3

2

1 0.01

1

100

0.01

1

100

0.01

1

100

Total pressure (mbar)

Figure 15.46 Selectivity (solid curves) of D2 over H2 as determined by an IAST analysis of the experimental isotherms at 77 K (black), 87 K (red), 100 K (green), 120 K (blue), and 150 K (brown). The dashed curves show the na, D2 /na, H2 ratio at the same temperatures as determined by the two-site Langmuir fit. Source: From FitzGerald et al. [67].

It has been established that MOF-74 provides some of the highest enthalpies of hydrogen adsorption thanks to the highly dense exposed metal sites. In 2013, MOF-74-M (M = Fe, Co, and Ni) was first explored for quantum sieving by Fitz-Gerald et al. [67]. D2 and H2 sorption were measured in a wide range of temperatures from 77 to 150 K. D2 displays a considerably higher initial adsorption heater than H2 , while 1.4 kJ/mol (Ni-MOF-74) is the main difference. In the IAST measurement, the selectivity of D2 /H2 is as high as 5.0 at 77 K (Figure 15.46), and this is substantially higher than hitherto recorded MOFs. Ironically, a strong correlation has been established through an infrared analysis between selectivity and the center-of-mass translational frequencies of the adsorbed molecule, which suggests that the selectivity mainly stems from the distinction in the zero-point energies in the adsorbed isotopes of the high binding sites. Oh et al. specifically detected an excellent separation of the hydrogen isotopes at the exposed open metal sites by TDS study in further analysis on CPO-27 (Co) (MOF-74-Co) [68]. The selectivity of 11.8 was demonstrated by the efficacy of the separation of an equimolar mixture of D2 /H2 at 60 K and 30 mbar. At an even high temperature of 80 K, the selectivity still displayed a reasonably high value of 6.3 (Figure 15.47). The authors subsequently established a three-cycle temperature swing deuterium enrichment system thanks to the high selectivity of MOF-74-Co. A final mixture with a substantially enriched D2 content (95% D2 + 5% H2 ) can be achieved after only three temperature swing cycles at operating temperatures between 80 and 100 K, based on the isotope mixture of low deuterium (5% D2 + 95% H2 ). The high performance and relatively lower operating temperature indicate the feasibility of this method as a fruitful solution for commercial industries. Weinrauch et al. investigated in 2017 the separation isotopes using MFU-41 (extended structure of MFU-4) by employing a linear ditriazolate type ligand

935

15 Relationship Between Structure and Separation Property

20

D2:H2 1 : 1 total D2:H2 quantum effect

16

D2:H2 thermal effect

Selectivity

936

12 8 4 1 0 40

(a)

50

60

(b)

70 80 90 100 Temperature (K)

120

Figure 15.47 (a) Structure of finite model of CPO-27-Co as optimized by DFT and corresponding equilibrium position of the hydrogen molecule adsorbed at the central cobalt atom (blue, cobalt; red, oxygen; silver, carbon; gray, lithium; white, hydrogen; black, adsorbed hydrogen molecule). (b) Theoretically predicted selectivity of D2 /H2 for an equimolar 1 : 1 mixture (dark red curve, filled squares) adsorbed in CPO-27-Co. The contribution of quantum (red filled triangles) and thermal (empty red triangles) effects in the isotope separation is shown. Source: From Oh et al. [68].

called bis(1H-1,2,3-triazolo-[4,5-b],[4,5-i])dibenzo[1,4]dioxin (H2 -BTDD) [69]. The Cu(I)-MFU-4I was obtained by partially substituting the Zn–Cl terminal sites with Cu(I) metal sites (Figure 15.48), showing surprisingly high hydrogen adsorption enthalpies with marked isotope effects, even at temperatures over 100 K. At various loading temperatures, TDS study, in addition to the inelastic neutron scattering (INS) analysis, show a powerful isotope exchange mechanism allowing H N C O CI Cu Zn

𝝓

r –

R

(a)

(b)

Figure 15.48 Structural model of Cu(I)-MFU-4l. H2 adsorption in Cu(I)-MFU-4l: cluster showing a full pore (a), computational model (b). R is the distance between Cu and H2 center, r is the H–H distance of adsorbed H2 , 𝜙 is the angle of rotation of adsorbed H2 in the plane normal to R, and Θ is the angle of out-of-plane rotation. Source: From Weinrauch et al. [69].

15.5 Separation of Hydrogen Isotopes

enriching heavy isotopologs from low-concentration phases. An experimental selectivity of D2 /H2 (for 1 : 1 mix) was achieved at 100 K, and greater selectivities for thermodynamic equilibrium settings dependent on calculations of first principles can be anticipated. The high Cu(I)-MFU-4I selectivity renders it a very appealing commodity for low-hydrogen gas mixtures, which could greatly boost radioactive wastes management and simultaneously generate valuable 3 He and tritium fusion reactor fuel. Forster and coworkers identified a high affinity that leads to high-temperature selectivity of D2 /H2 [70]. VSB-5 showed that the initial adsorption heat is higher than 16 kJ/mol for H2 and is the highest measured, and D2 still at 1.5 kJ/mol, and this is high heat of adsorption over H2 at low loadings. Such discrepancy implies that any substance with temperatures above 120 K has the best calculated initial D2 /H2 selectivity.

15.5.3 Special Quantum Sieving In 2017, Hirscher, Moon, Oh, and coworkers employed a strategy based on MIL-53(Al), a representation of flexible MOFs, to effectively separate hydrogen isotopes by dynamic pore change based on the breathing of the employed MOF [71]. Direct experiments for the separation of D2 over H2 with 1 : 1 D2 –H2 mixture reveal that selectivity is closely related to MIL-53(Al) porous structure (Figure 15.49). As a feature of the dynamic structural change from np, MIL-53(Al) can continually produce productive pores opening by optimizing the quantum sieving influence, leading to a high selection efficiency (S⋅D2 /H2 = 10.5) with 2.3 mmol/g at 40 K. Improved separation efficiency was accomplished by consistently tuning the breathing process by exposure pressure and time attaining a high selectivity, S⋅D2 /H2 = 13.6, with high separation capacity of 2.9 mmol/g (Figure 15.50). The Figure 15.49 Schematic view of D2 separation in breathing propagation along the 1D channel in MIL-53(Al). (a) before H2/D2 enter the narrow pore, (b) optimizing the quantum sieving influence, (c) H2/D2 in the larger pore. Source: From Kim et al. [71].

H2 D2 2.6 Å

np (a) 8.5 Å 2.6 Å

Ip

tum

n Qua

(b) 8.5 Å

Ip (c)

ing siev

np

Breathing propagation

937

0 10–5 10–4 10–3 10–2 10–1 100 Pressure (bar) 25

12

20

9

15

6

10

3

5

0

0

(b)

10 mbar 77 K

Breathing

20

15 10 5 0 10–5 10–4 10–3 10–2 10–1 100 Pressure (bar) 40 K, 10 min

15

10

12

8

9

6

6

4

3

2

0

0 20 40 60 80 100 Pexp (mbar)

0

15 10 5 0 10–5 10–4 10–3 10–2 10–1 100 10 Pressure (bar)

(c)

15

10 mbar, 40 K 3.5

12

2.8

9

2.1

6

1.4

3

0.7

0 0

0.0 30 60 90 120 150 texp (min)

Amount of adsorbed D2 mmol/g

5

20 30 40 50 60 70 80 Texp (K)

25

Breathing

Selectivity (nD2/nH2)

10

Amount of adsorbed H2 mmol/g

15

10 mbar, 10 min

15

40 K

20

Selectivity (nD2/nH2)

(a)

10 mbar

25

20 K Breathing

Amount of adsorbed D2 mmol/g

20

10 mbar

Amount of adsorbed H2 mmol/g

25

Amount of adsorbed D2 mmol/g

Amount of adsorbed H2 mmol/g

15 Relationship Between Structure and Separation Property

Selectivity (nD2/nH2)

938

(d)

Figure 15.50 (a) Isotherms of H2 on MIL-53 at 20, 40, and 77 K. The green vertical dotted lines indicate the points at which the pressure was 10 mbar, which was the exposuremixture pressure for the TDS measurements. The selectivity (black symbols) and the corresponding amount of adsorbed D2 (red symbols) as functions of (b) exposure temperature (T exp ), (c) exposure pressure (P exp ), and (d) exposure time (t exp ). Source: From Kim et al. [71].

authors anticipate this research to offer an impression of producing a productive model with a high degree of selectivity as well as separation capacity to separate gas mixtures of atoms/molecules of comparable dimensions and shape. In 2017, Moon and coworkers implemented KQS as well as CAQS in one system, and D2 could be efficiently isolated and kept in the MOF pore [72]. In their work, imidazole molecules, by employing a post-modification strategy, were introduced into the channels with a high density of MOF-74-Ni open metal sites (Figure 15.51). Figure 15.51 Imidazole molecules were introduced into the channels of MOF-74-Ni (denoted as MOF-74-IM). Source: From Kim et al. [72].

IM H2 D2 OMS

H2

D2

MOF-74-IM

15.6 Enantioselective Separation

This easily accessible method eliminated the new MOF synthesis for optimizing the aperture size as well as raising the internal bonding energy. In this method, the high dense open metal sites of MOF-74-Ni acting as CAQS were maintained in the channels. Meanwhile, by optimization of the coordinated imidazole amount, the synergistic effect of KQS was produced. These combined two quantum sieving yielded remarkably high selectivity of approximately 26 with high D2 uptake (2.84 mmol/g) even at high temperatures (77 K). Additionally, the obtained material was even applied to other gas mixtures separation, including tritium, deuterium, and hydrogen mixture or He isotopic mixtures. This strategy of integrating KQS and CAQS in one system would offer new prospects for the intelligent design of porous materials, and this led to the establishment of a new highly reliable isotope as well as gas separation systems.

15.6 Enantioselective Separation Racemic mixture isolation for supplying enantiopure compounds is essential for the development of medicinal and fine chemicals. Nevertheless, enantioselective separation is a serious issue, owing to the almost similar chemical and physical characteristics of the racemates. Chiral chromatography is a popular technique of chiral separation utilizing a chiral stationary phase (CSP). The commonly used CSP is derived from oligosaccharides, including cellulose or cyclodextrin (CD). Nevertheless, there are many prerequisites for chiral separation and problems of low efficiency, which cannot be met by CSP material currently in use, and this supports the demanding development of new CSP materials. Porous chiral solids can be perfect applicants for CSP materials as they incorporate chirality and sieving effect into one integrated system. For enantioselective separation, chiral zeolite materials have been examined, even if their constraints are many. For instance, the chiral topology of zeolites has limited types, and homochiral zeolite materials are not simple to prepare. Furthermore, zeolites pore sizes are typically small and are therefore not appropriate for large organic molecules, which are characteristic for pharmaceutical molecules. The MOFs are a particular kind of porous materials, which can possibly be made homochiral with the size of pores suited to separation applications. An enantiopure ligand can be incorporated, for instance, to easily produce homochiral MOF. Lately, certain studies have reported that homochiral MOFs can only be generated with achiral components in the existence of sufficient chiral induction agents [73]. The accomplishment in the homochiral MOFs synthesis has motivated many studies on their utilization for enantioselective separation implementations [74]. The first homochiral MOF was identified in 1999 by Aoyama and coworkers [75], several chiral MOFs have since been published. Thanks to the adaptable homochiral structures and structural diversities, MOFs are attractive candidates for enantioselective separation. The MOFs are mainly used as CSPs for HPLC, GC or as a solution and solid-phase extraction (SPE) absorbents, despite the use of many chiral separations approaches. Most chiral MOF absorbents were examined for enantioseparation, including amino acids, sulfoxides, amines, alcohols, and therapeutic

939

940

15 Relationship Between Structure and Separation Property

molecules such as naproxen. Although numerous reports have been conducted in enantioseparation relied on MOFs, the analyte categories are still few, and the majority of analytes identified are molecules with small kinetic diameters of less than 10 Å. Subsequent research should also concentrate on separating large enantiomers and utilizing different analytes.

15.6.1 Chiral MOF as Absorbent in SPE and Solutions Kim, Fedin, and coworkers reported, in 2006, a homochiral MOF, [Zn2 (bdc)(L-lac)(DMF)]⋅DMF, in which the SBU was formed by Zn2+ and L-lactic acid (L-H2 lac), while 1,4-benzendicarboxylic acid acted as an organic linker and connected the SBU to offer the homochiral MOF with permanent porosity, size, and enantioselective sorption features, as well as catalytic activity (Figure 15.52) [76]. The MOF in a CH2 Cl2 solution comprising a racemic mixture of sulfoxides was stirred for 16 hours and then collected, washed with methanol for extracting the adsorbed guest molecules. The MOF displayed a remarkable sorption capability toward the sulfoxides with smaller substituents. Although the enantiomeric excess (ee) values for the adsorbed guests were modest, it still suggests that the MOF may be beneficial to separate or purify sulfoxides.

b a (a)

c a

c

b

b (b)

(c)

Figure 15.52 (a) A view of the 1D chiral chains (as both ball-and-stick/polyhedra and wire models; polyhedra represent the Zn coordination environments) in the structure of the MOF. (b) Perspective view of the structure of the MOF along the a-axis. (c) Projection of the structure of the MOF in the (110) plane. Hydrogen atoms and guest molecules were omitted for clarity. Zn, green; N, blue; O, red; C, gray; chiral C atoms of the lactic acid ligand were shown in white. Source: From Dybtsev et al. [76].

15.6 Enantioselective Separation

With the use of a metallo-salen derived bipyridine ligand and a dicarboxylate ligand, Chen and coworkers established a mixed MOF approach that led to the synthesis of two MOFs: M’MOF-2 (using benzenedicarboxylate [BDC]) and M’MOF-3 (using cyclohexane-dicarboxylate [CDC]). Both compounds are homochiral because of the Cu-salen enantiopure ligand. These compounds were also utilized to the encapsulation by liquid-phase adsorption for the racemic mixture of (R)- and (S)-1-phenylethy alcohol (PEA). The M’MOF-3 was found to solely take up S-PEA affording M’MOF-3S-PEA. The HPLC chiral analysis of the desorbed PEA from the PEA-included M’MOF-2 provided an ee value of 21.1%. With a much higher ee-value of 64%, the smaller pores of the enantiopure M’MOF-3 have greatly increased its enantioselectivity to separate R/S-PEA. The structure series has been extended using the analogous approach by the same research group [77]. Because of its narrower but more selective pore that is limited by bulky tert-butyl groups toward the chiral cavity, M’MOF-7 displays an even higher ee value of 82.4 for the separation of 1-PEA. Nonselective items can be expelled by the occupation of the chiral pore space by the organic struts. In 2014, two chiral MOFs generated from enantiopure ligands and Mn2+ were produced and described structurally by Cui and coworkers [78]. As shown in Figure 15.53, the as-synthesized MOF involves 1D channels amended by chiral dihydroxyl groups (L1 ) or chiral dimethoxy groups, and the nanoscale aperture window size is approximately 1.5 nm ×1.0 nm. (R)-L1 -MOF adsorption preference for S-enantiomers is overcoming racemic counterparts and vice versa. This is due to various orientations and strength of binding between racemic constituents in MOF channels. The D-histidine was inserted in the ZIF-8 system in 2015 by Wang and coworkers via a one-pot ligand substitution approach [79]. With ee values of 78.52% and 79.44% for alanine and glutamic acid, respectively, the synthesized D-his-ZIF-8 accomplished enantioseparation. The pores of chiral ZIF can pick up S-alanine and S-glutamic acids, while ZIF-8 cannot distinguish between these enantiomers. Such capability is because of the stereoselectivity and N–H…O noncovalent bonding between the amino acid and the D-his ligand. After three cycles, D-his-ZIF-8 is stable, without any significant reduction in the racemic separation. Cui and coworkers published in 2016 a newly synthesized chiral MOF for enantioselective sulfoxide adsorption [80]. The as-synthesized MOF is formed from Zn2+ , BDC ligand, and DHIP-based ligand (L). Along the b-axis, L-Zn-L establishes spiral chains, and along the c-axis, the size of 1D channel is approximately 4.7 Å × 5.4 Å. For 3-methoxyphenyl methyl sulfoxide, 37.1 (R) ee% is the best-obtained performance. For this work also, cadmium chiral MOF was formed without racemic recognition and separation potential. In 2017, a method reported by Zhang and coworkers utilized an achiral SURMOF template for the preparation of chiral polymer films to be employed for enantiomer separation [81]. HKUST-1 was primarily grown on QCM substrate, and L-DOPA was then added to MOF pores, as displayed in Figure 15.54. After the self-polymerization of L-DOPA, acid etching was used to eliminate the MOF

941

15 Relationship Between Structure and Separation Property

(a)

(b)

(c)

(d)

Figure 15.53 X-ray structure of (S)-1a. (a) The coordination mode of the L1 ligand. (b) The infinite Mn carboxylate chain. (c) Four metal-carboxylate chains linked by the L1 ligands forming a 41 helical channel. (d) View of the 3D porous structure of (S)-1a along the c-axis (the Mn atoms are shown in polyhedra) (Mn, green; O, red; N, purple; C, blue). Source: From Peng et al. [78].

2.5

L-DOPA

2.0

R-Naproxen S-Naproxen

Uptake (μg/cm2)

942

1.5

L-DOPA@MOF

MOF thin film

thin film

O2 polymerization

R-Naproxen S-Naproxen

1.0 0.5 0.0

Etching

Porous poly(L-DOPA) thin film

(b)

0

50

100 150 Times (s)

200

250

Poly(L-DOPA)@MOF thin film

(a)

Figure 15.54 (a) The growth process from MOF thin film, L-DOPA loaded MOF thin film to chiral poly(L-DOPA) thin film. (b) The uptake of R-naproxen and S-naproxen adsorption in chiral poly(L-DOPA) thin film. Source: From Gu et al. [81].

15.6 Enantioselective Separation

(+)–EP

(–)–EP

Before After

(+)–EP

(–)–EP MOF bed

HPLC analysis

(b) 10

(a)

11

12

13

14

(c)

Figure 15.55 (a) SPE separation of EP in hexane:EtOH 75 : 25 using Cu(GHG) as a chiral bed. (b) HPLC chromatograms of EP racemate before (dashed line) and after (solid line) passing through the MOF bed. (c) The 1D channels in Cu(GHG) are surrounded by functional sites prone to establish supramolecular interactions, well fitted for chiral recognition and discrimination. Source: From Navarro-Sanchez et al. [82].

template. For enantioselective naproxen separation, the as-synthesized chiral porous polymer film reached a maximum value of approximately 32% (R). For enantioselective separation between EP and MA by SPE, a homochiral MOF Cu(GHC) was employed by Martí-Gastaldo and coworkers [82]. In this MOF, Cu2+ is tetrahedrally coordinated by amino, methylamino, carboxylate group, and imidazole from ligand GHG, respectively. A fourfold spiral chain is formed by the linkage of carboxylic acids with Cu-GHG-Cu, as displayed in Figure 15.55. An MC simulation shows, thanks to H links between the reagents and the ligands on the framework, that (+)-MA and (+)-EP are more desirable for stereoselective separation from the enantiomer mixture. Furthermore, the interaction between (+)-EP-Cu(GHG) is strengthened by extra H-bonds. A four minutes SPE/MOF separation is used to separate (54 ± 2)% of (+)-EP from 1 : 1 racemic mixture, and a 44% (+)-EP is extracted by direct soaking adsorption by chiral MOF.

15.6.2 Chiral MOFs as Stationary Phase in HPLC Fedin and coworkers investigated in 2007 the behavior of [Zn2 (bdc)(L-lac) (DMF)]⋅DMF for the column chromatographic separation [83]. Also, for this MOF,

943

15 Relationship Between Structure and Separation Property

𝜌-MePhSOMe Concentration / (a.u.)

Concentration / (a.u.)

PhSOMe R

S 0

5

10

15

20

Eluate volume /

25

30

(cm3

)

S 0

5

10 15 20 25 30 35 40 Eluate volume / (cm3)

(b) 𝜌-BrPhSOMe

5

S

10 15 20 25 Eluate volume / (cm3)

PhSOiPr Concentration / (a.u.)

R

0

(c)

R

35

(a) Concentration / (a.u.)

944

R

0

S

20 40 60 Eluate volume / (cm3)

80

(d)

Figure 15.56 Separation of alkyl aryl sulfoxides using the MOF as the chiral stationary phase. Eluents: (a,b) 12 cm3 of 0.01 M DMF solution in CH2 Cl2 then 1% DMF in CH2 Cl2 ; (c,d) a 20 cm3 of 0.01 M DMF solution in CH2 Cl2 , then 1% DMF in CH2 Cl2 . Elution rate: 2 cm3 /h. Source: From Nuzhdin et al. [83].

the solution based enantioselective sorption has been estimated. For their new work, a suspension of the MOF (14 g) in a 10% solution of DMF in CH2 Cl2 was loaded into a glass tube with 8 mm inner diameter for the production of a 33 cm high column. The sulfoxide of each probe (c. 0.15 mmol) was present in 0.2 cm3 of CH2 Cl2 . In Figure 15.56, for one of the sulfoxides (PhSOMe), the experimental chromatograms show a clear peak resolution enabling complete enantiomers separation. This research represents the first semipreparative scale chromatographic separation of chiral sulfoxides and the first recorded utilization of MOFs for preparative liquid chromatography as CSPs. As previously stated, Cui and coworkers used enantiopure ligands and Mn2+ for the synthesis of two chiral MOFs [78]. In Figure 15.57, Using L1 -MOF in HPLC as CSP allows for the separation of benzoylated 1-PEA and derivatives reaching selective factor (𝛼) and chromatographic resolution (Rs ) as high as 1.4 and 2.7, respectively, within 0.5 hours. Thermostable, as well as robust frameworks, make it possible to recycle and reuse material, but without precise performance decreases. The findings have shown different orientations and binding energies of the two enantiomers in the microenvironment of MOF to be related to the inherent chiral recognition and separation.

15.6 Enantioselective Separation NHCOPh

NHCOPh

9

11

13 15 17 19 Retention time (min)

21

23

6

8

12 14 16 10 Retention time (min)

18

NHCOPh (c)

(a)

20 NHCOPh

F

12 (b)

14

16 18 20 22 Retention time (min)

24

26

10

12

(d)

18 14 16 20 Retention time (min)

22

24

Figure 15.57 The HPLC separation of racemic amines on the 1a-packed column. (a) 1-phenylethylamine, (b) 1-(4-fluorophenyl)ethylamine, (c) 1-phenylpropylamine, (d) 1-naphthalen-1-yl-ethylamine after benzoylation using hexane/isopropanol of 90/10 (v/v) as the mobile phase at 23 ∘ C, monitored with an ultraviolet detector at 254 nm. Source: From Peng et al. [78].

In 2015, Liu and coworkers were using CSP based on chiral MOF for the separation of racemic ibuprofen utilizing HPLC [84]. The enantiopure ligand D-camphorate (D-cam) and planar Mn4 O cluster contributed to a homochiral structure, (Me2 NH2 )2 [Mn4 O(D-cam)4 ]⋅(H2 O)5 , containing 1D channel with approximately 4.8 Å × 4.8 Å square-aperture size built by [Mn4 O(D-cam)4 ]2− . To separate racemic ibuprofen using HPLC, the selectivity factor and resolution were found to be 6.48 and 2.02, respectively. CD-MOFs were utilized by Stoddart and coworkers in 2016 as stationary-phase materials for different HPLC separation of organic compounds [85]. For (R)-(+)-limonene and (S)-(-)-1-phenylethanol, it was found that they outperformed their enantiomers by selective factors as 1.72 and 2.26, respectively. Because of the strong interaction between exocyclic double bond and CD-MOF-1, the α-, β-pinene separation, and their racemic counterparts show remarkable performance. Furthermore, the geometry orientation of substrates in openings and the interaction of hydrogen bonding between analytes and frameworks are also recorded in this work for the separation of alkyl-, vinyl-benzenes, and haloaromatics, as seen in Figure 15.58.

15.6.3 Chiral MOFs as Stationary Phase in GC The Yuan group first identified chiral MOF (H2 sala = N-(2-hydroxylbenzyl)-Lalanine)-coated capillary column utilized in GC to separate enantiomers [86]. After that, many other chiral MOFs were also used in GC as stationary chiral phases. As a stationary phase for GC separation, a chiral MOF Ni(D-cam)(H2 O)2

945

946

15 Relationship Between Structure and Separation Property

(a)

(b)

(c)

Figure 15.58 The grand canonical Monte Carlo (GCMC) simulations of benzene within CD-MOF-2 determined at 1 kPa. (a,b) Representations of a (γ-CD)6 cube portion of CD-MOF that show benzene residing in the large 1.7 nm diameter pore (center of image) and in the transverse pores located between adjacent (γ-CD)6 cubes (stick representations, black), as well as a single benzene molecule located in each of the eight triangular-shaped pores (space-filling representations, black). (c) View of the triangular-shaped pore along the (111) direction, revealing the presence of free primary hydroxyl groups (ball-and-stick representation) that may act as sites to hydrogen bond with molecules of methanol from the crystallization procedure, which could effectively block the entrance of this pore and prevent the passage of benzene and toluene, limiting their retention within the framework. Source: From Hartlieb et al. [85].

with surprising full integration of molecular chirality, absolute helicity, and 3D intrinsic chiral net was selected [87]. The dynamic coating method demonstrates strong isomer and linear alkane separation, particularly for racemates, in open tubular column. The group of Yuan also recorded the development of a 3D chiral nanoporous MOF, [(CH3 )2 NH2 ][Cd(bpdc)1.5 ]⋅2DMA, which was coated as capillary column (2 m long × 75 μm i.d.) using a dynamic coating process for the separation of racemates by GC [88]. As can be seen in Figure 15.59, the racemates comprise aldehyde, ketone, organic acid, amino acid, and alcohol, and all enantioseparations are retained for a short period and can be analyzed easily. The mechanism of chiral recognition was projected, even though the chiral microenvironment had a complex effect on the chiral characteristics of chromatographic systems. The MOFs were made with wide chiral pore appropriate for the chiral guests reachable to the chiral channel for chiral recognition. Also, certain interactions such as hydrogen bonding, van der Waals forces, dipole–dipole, and dispersion, between racemates and 3D chiral nanoporous MOFs had participated in the chiral recognition. The MOFs-coated capillary column displayed good reproducibility with RSDs of 0.23% and 2.1% for retention time and peak area, respectively, and excellent separation performance with theoretical plates of 2180 plates/m. CDs and their derivatives are among the chiral selectors that are most used and intriguing. The Yuan group initially attempted to incorporate chiral MOF Co(D-Cam)1/2 (bdc)(tmdpy) with permethylated β-CD to examine if chiral MOF may improve the enantiomer separation by GC on a stationary CD phase [89]. The dynamic coating method, combined with a static coating, was used to produce a composite capillary column. The accompanying single coated column, MOF column, as well as a permethylated β-CD column, were also produced and used for comparison. The findings showed increased resolution and high performance of the

15.6 Enantioselective Separation

(–)

Solvent dichloromethane

(+)

Solvent dichloromethane (–) (+) 0.5 mv

D

0.5 mv 1 mv

L

0.5

D

0.5

1 Time (min)

1

1

0

(b)

Solvent dichloromethane

(c)

D

(+)

(–)

1 0.5 Time (min)

Time (min)

Time (min)

(a)

L

Solvent dichloromethane

(d)

D

0.5 mv

0

L

1 mv

L

1 mv

0

1

Time (min)

(e)

1 mv

Time (min)

1

D L

0.5 1 Time (min)

0.5 1 Time (min) (g)

(f) Solvent dichloromethane

1 mv

(h)

Time(min)

Solvent rthanol L (+) D 1 mv

0.5 mv

(+) (–)

(–)

0.25 mv

(i)

0.5 1 Time (min)

1

0 (j)

Time (min)

1

0 (k)

Time (min)

Figure 15.59 GC chromatograms on the MOF-coated open tubular column D (2 m long × 75 μm i.d.) for the separation of racemates. (a) Isoleucine derivative. (b) Citronellal. (c) 1-Phenyl-1,2-ethandiol. (d) Leucine derivative. (e) Phenyl-succinic acid. (f) 1-Phenyl-ethanol. (g) Valine derivative. (h) Alanine derivative. (i) Proline derivative. (j) Camphor. (k) 2-Methyl-1-butanol derivative. Source: From Xie et al. [88].

947

15 Relationship Between Structure and Separation Property

column on the current chiral system, owing to the intrinsic features of the employed chiral MOF, such as helical channels and an indefinite expansion to the supramolecular reticular structure in three dimensions. Subsequently, InH(D-C10 H14 O4 )2 and [Cd(LTP)2 ]n has also been included in a permethylated β-CD attaining columns of the chiral capillary to be used for the racemates separation by GC [90]. The composite chiral MOFs-permethylated β-CD columns displayed a greater selectivity and higher resolution ability as opposed to the permethylated β-CD column for the recognition of the racemates. Zhang and coworkers employed a layer-by-layer LPE system in 2016 for growing homochiral SURMOFs on functionalized GC capillary columns for the enantioseparation of methyl lactate [91]. Figure 15.60 shows that the chiral MOF Cu2 (D-cam)2 P (P = dabco and bipy) grown on the surface of poly(L-DOPA) functionalized column is along the 110 direction. The reason for that is because the column surface-exposed COOH groups substitute acetate groups found in the Cu-paddlewheels. Cu2 (D-cam)2 (dabco)@poly(L-DOPA) decorated columns are strongly oriented and homogeneous, leading to significantly improving the separation of enantioselective methyl lactate compared to modified columns of Cu2 (D-cam)2 (dabco) or poly(L-DOPA), and retention times for the three columns are 2.63/3.61, 2.91/3.12, and 2.95/2.95, respectively. Zhang and coworkers developed homochiral MOF in 2016 with unusual ligand-unsupported Cu–Cu interactions and has the formula of ([CuI2 CuII(L21)2 (CN)(H2 O)](NO3 )DMF(54) (L21 = 5-eatz = (1S)-1-(5-tetrazolyl) ethylamine) [92]. For the enantioselective separation of different racemic alcohols, the prepared MOF was employed, and it was observed that MOF(54) is feasible for the separation of

Capillary column

Gas

L/D Poly(L-DOPA)

T

On/off (a)

Relative intensity

948

L-DOPA

O2 pH=8.5

(b) Cu(OAc)2 EtOH Dcam/P

L-DOPA

o o Cu o o o o Cu o o

E HO

NH2 COOH

HO = HO

H N

HO

Poly(L-DOPA)

o o Cu o o o o Cu o o

Cu2(Dcam)2dabco@poly(L-DOPA) D-MeLt

N

N o o o o Cu o

o o Cu o o o o Cu o o

L-MeLt

N

Cu

o

D Cu2(Dcam)2dabco

Cu

o o Cu o o o o Cu o o

o o Cu o o o o Cu o o

COOH

(c)

o o o o Cu o

L

P

N

2.0 (e)

2.5

3.0

3.5 Time (min)

4.0

4.5

Cu2(Dcam)2P

(d)

Figure 15.60 (a) The setup of MOF and poly(L-DOPA) thin-film grown in capillary column using layer-by-layer approach; (b) the layer-by-layer preparation procedure of poly(L-DOPA) functionalized capillary column and MOF thin film; (c) structure of poly(L-DOPA); (d) and homochiral MOF Cu2 (D-cam)2 P (P = dabco and bipy); (e) the retention time of a mixture of enantiomers (L- and D-methyl lactate) in Cu2 (D-cam)2 (dabco)@poly(L-DOPA) grown in capillary column using a gas chromatographic method. Source: From Gu et al. [91].

References

OH

OH *

R,S–

R

*

R

R–

OH

OH

R,S–

R,S– *

R

*

R

Figure 15.61 Possible interactions between the chiral framework and the chiral aromatic alcohols. Source: From Liu et al. [92].

only aromatic alcohols instead of aliphatic alcohols. The authors provided two explanations for this extraordinary MOF(54) action: first, it was hard for large molecules to enter through the 1D narrow network. The enantioselective separation was probably present on the surface of the chiral MOF, where the separation of aromatic alcohols is easier relative to aliphatic alcohols. Secondly, it was possible to enantiomerically separate aromatic alcohols because of three interaction types: π–π interactions between the surface of the MOF and aromatic alcohols, and H-bonding and stereochemical interactions with the chiral L20 ligands on the surface of MOF. However, for aliphatic alcohols, there were only H-bonds that did not result in enantioselective separation. The enantioselective separation process of the chiral MOF is shown in Figure 15.61.

References 1 Caskey, S.R., Wong-Foy, A.G., and Matzger, A.J. (2008). J. Am. Chem. Soc. 130: 10870–10871. 2 Queen, W.L., Hudson, M.R., Bloch, E.D. et al. (2014). Chem. Sci. 5: 4569–4581. 3 Zhai, Q.G., Bu, X.H., Mao, C.Y. et al. (2016). J. Am. Chem. Soc. 138: 2524–2527.

949

950

15 Relationship Between Structure and Separation Property

4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

Li, J.R., Yu, J.M., Lu, W.G. et al. (2013). Nat. Commun. 4: 1538. Nugent, P., Belmabkhout, Y., Burd, S.D. et al. (2013). Nature 495: 80–84. Shekhah, O., Belmabkhout, Y., Chen, Z.J. et al. (2014). Nat. Commun. 5: 4228. Chen, K.J., Madden, D.G., Pham, T. et al. (2016). Angew. Chem. Int. Ed. 55: 10268–10272. Xiang, S.C., He, Y.B., Zhang, Z.J. et al. (2012). Nat. Commun. 3: 954. Demessence, A., D’Alessandro, D.M., Foo, M.L., and Long, J.R. (2009). J. Am. Chem. Soc. 131: 8784–8786. McDonald, T.M., Mason, J.A., Kong, X.Q. et al. (2015). Nature 519: 303–308. Liao, P.Q., Chen, H.Y., Zhou, D.D. et al. (2015). Energy Environ. Sci. 8: 1011–1016. Wang, B., Huang, H.L., Lv, X.L. et al. (2014). Inorg. Chem. 53: 9254–9259. Fracaroli, A.M., Furukawa, H., Suzuki, M. et al. (2014). J. Am. Chem. Soc. 136: 8863–8866. Zheng, B.S., Bai, J.F., Duan, J.G. et al. (2011). J. Am. Chem. Soc. 133: 748–751. Zheng, B.S., Yang, Z., Bai, J.F. et al. (2012). Chem. Commun. 48: 7025–7027. Bastin, L., Barcia, P.S., Hurtado, E.J. et al. (2008). J. Phys. Chem. C 112: 1575–1581. Mohamed, M.H., Elsaidi, S.K., Wojtas, L. et al. (2012). J. Am. Chem. Soc. 134: 19556–19559. Scott, H.S., Ogiwara, N., Chen, K.J. et al. (2016). Chem. Sci. 7: 5470–5476. Serre, C., Groves, J.A., Lightfoot, P. et al. (2006). Chem. Mater. 18: 1451–1457. Wriedt, M., Sculley, J.P., Yakovenko, A.A. et al. (2012). Angew. Chem. Int. Ed. 51: 9804–9808. Bhatt, P.M., Belmabkhout, Y., Cadiau, A. et al. (2016). J. Am. Chem. Soc. 138: 9301–9307. Gucuyener, C., van den Bergh, J., Gascon, J., and Kapteijn, F. (2010). J. Am. Chem. Soc. 132: 17704–17706. Wang, Q.M., Shen, D.M., Bulow, M. et al. (2002). Microporous Mesoporous Mater. 55: 217–230. Bao, Z.B., Alnemrat, S., Yu, L. et al. (2011). Langmuir 27: 13554–13562. He, Y.B., Krishna, R., and Chen, B.L. (2012). Energy Environ. Sci. 5: 9107–9120. Chang, G.G., Huang, M.H., Su, Y. et al. (2015). Chem. Commun. 51: 2859–2862. Liao, P.Q., Zhang, W.X., Zhang, J.P., and Chen, X.M. (2015). Nat. Commun. 6: 8697. Lin, R.B., Wu, H., Li, L.B. et al. (2018). J. Am. Chem. Soc. 140: 12940–12946. Li, L.B., Lin, R.B., Krishna, R. et al. (2018). Science 362: 443–446. Qazvini, O.T., Babarao, R., Shi, Z.L. et al. (2019). J. Am. Chem. Soc. 141: 5014–5020. Xiang, S.C., Zhang, Z.J., Zhao, C.G. et al. (2011). Nat. Commun. 2: 204. Hu, T.L., Wang, H.L., Li, B. et al. (2015). Nat. Commun. 6: 7328. Cui, X.L., Chen, K.J., Xing, H.B. et al. (2016). Science 353: 141–144. Li, B., Cui, X.L., O’Nolan, D. et al. (2017). Adv. Mater. 29: 1704210. Chen, K.J., Madden, D.G., Mukherjee, S. et al. (2019). Science 366: 241. Lamia, N., Jorge, M., Granato, M.A. et al. (2009). Chem. Eng. Sci. 64: 3246–3259.

References

37 Bae, Y.S., Lee, C.Y., Kim, K.C. et al. (2012). Angew. Chem. Int. Ed. 51: 1857–1860. 38 Yoon, J.W., Seo, Y.K., Hwang, Y.K. et al. (2010). Angew. Chem. Int. Ed. 49: 5949–5952. 39 Li, K.H., Olson, D.H., Seidel, J. et al. (2009). J. Am. Chem. Soc. 131: 10368–10369. 40 Cadiau, A., Adil, K., Bhatt, P.M. et al. (2016). Science 353: 137–140. 41 Li, L.B., Wen, H.M., He, C.H. et al. (2018). Angew. Chem. Int. Ed. 57: 15183–15188. 42 van den Bergh, J., Gucuyener, C., Pidko, E.A. et al. (2011). Chem. Eur. J. 17: 8832–8840. 43 Peralta, D., Chaplais, G., Simon-Masseron, A. et al. (2012). J. Am. Chem. Soc. 134: 8115–8126. 44 Hartmann, M., Kunz, S., Himsl, D. et al. (2008). Langmuir 24: 8634–8642. 45 Couck, S., Remy, T., Baron, G.V. et al. (2010). Phys. Chem. Chem. Phys. 12: 9413–9418. 46 Assen, A.H., Belmabkhout, Y., Adil, K. et al. (2015). Angew. Chem. Int. Ed. 54: 14353–14358. 47 Alaerts, L., Maes, M., van der Veen, M.A. et al. (2009). Phys. Chem. Chem. Phys. 11: 2903–2911. 48 Herm, Z.R., Wiers, B.M., Mason, J.A. et al. (2013). Science 340: 960–964. 49 Jiang, J.W. and Sandler, S.I. (2006). Langmuir 22: 5702–5707. 50 Ma, S.Q., Sun, D.F., Wang, X.S., and Zhou, H.C. (2007). Angew. Chem. Int. Ed. 46: 2458–2462. 51 Mueller, U., Schubert, M., Teich, F. et al. (2006). J. Mater. Chem. 16: 626–636. 52 Liu, J., Thallapally, P.K., and Strachan, D. (2012). Langmuir 28: 11584–11589. 53 Hulvey, Z., Lawler, K.V., Qiao, Z.W. et al. (2013). J. Phys. Chem. C 117: 20116–20126. 54 Thallapally, P.K., Grate, J.W., and Motkuri, R.K. (2012). Chem. Commun. 48: 347–349. 55 Perry, J.J., Teich-McGoldrick, S.L., Meek, S.T. et al. (2014). J. Phys. Chem. C 118: 11685–11698. 56 Banerjee, D., Simon, C.M., Plonka, A.M. et al. (2016). Nat. Commun. 7: 11831. 57 Mohamed, M.H., Elsaidi, S.K., Pham, T. et al. (2016). Angew. Chem. Int. Ed. 55: 8285–8289. 58 Li, L.Y., Guo, L.D., Zhang, Z.G. et al. (2019). J. Am. Chem. Soc. 141: 9358–9364. 59 Wang, Y.L., Liu, W., Bai, Z.L. et al. (2018). Angew. Chem. Int. Ed. 57: 5783–5787. 60 Fernandez, C.A., Liu, J., Thallapally, P.K., and Strachan, D.M. (2012). J. Am. Chem. Soc. 134: 9046–9049. 61 Beenakker, J.J.M., Borman, V.D., and Krylov, S.Y. (1995). Chem. Phys. Lett. 232: 379–382. 62 (a) Lin, X., Jia, J.H., Zhao, X.B. et al. (2006). Angew. Chem. Int. Ed. 45: 7358–7364. (b) Jia, J.H., Lin, X., Wilson, C. et al. (2007). Chem. Commun.: 840–842. (c) Xiao, B., Wheatley, P.S., Zhao, X.B. et al. (2007). J. Am. Chem. Soc. 129: 1203–1209. 63 Chen, B.L., Zhao, X., Putkham, A. et al. (2008). J. Am. Chem. Soc. 130: 6411–6423.

951

952

15 Relationship Between Structure and Separation Property

64 Noguchi, D., Tanaka, H., Kondo, A. et al. (2008). J. Am. Chem. Soc. 130: 6367–6372. 65 Oh, H., Park, K.S., Kalidindi, S.B. et al. (2013). J. Mater. Chem. A 1: 3244–3248. 66 Teufel, J., Oh, H., Hirscher, M. et al. (2013). Adv. Mater. 25: 635–639. 67 FitzGerald, S.A., Pierce, C.J., Rowsell, J.L.C. et al. (2013). J. Am. Chem. Soc. 135: 9458–9464. 68 Oh, H., Savchenko, I., Mavrandonakis, A. et al. (2014). ACS Nano 8: 761–770. 69 Weinrauch, I., Savchenko, I., Denysenko, D. et al. (2017). Nat. Commun. 8: 14496. 70 Sharma, A., Lawler, K.V., Wolffis, J.J. et al. (2017). Langmuir 33: 14586–14591. 71 Kim, J.Y., Zhang, L.D., Balderas-Xicohtencatl, R. et al. (2017). J. Am. Chem. Soc. 139: 17743–17746. 72 Kim, J.Y., Balderas-Xicohtencatl, R., Zhang, L.D. et al. (2017). J. Am. Chem. Soc. 139: 15135–15141. 73 (a) Morris, R.E. and Bu, X.H. (2010). Nat. Chem. 2: 353–361. (b) Zhang, J., Chen, S.M., Wu, T. et al. (2008). J. Am. Chem. Soc. 130: 12882–12883. (c) Zhang, J., Chen, S.M., Nieto, R.A. et al. (2010). Angew. Chem. Int. Ed. 49: 1267–1270. (d) Zhang, S.Y., Li, D., Guo, D. et al. (2015). J. Am. Chem. Soc. 137: 15406–15409. 74 (a) Seo, J.S., Whang, D., Lee, H. et al. (2000). Nature 404: 982–986. (b) Zhang, S.Y., Yang, C.X., Shi, W. et al. (2017). Chem 3: 281–289. 75 Ezuhara, T., Endo, K., and Aoyama, Y. (1999). J. Am. Chem. Soc. 121: 3279–3283. 76 Dybtsev, D.N., Nuzhdin, A.L., Chun, H. et al. (2006). Angew. Chem. Int. Ed. 45: 916–920. 77 Das, M.C., Guo, Q.S., He, Y.B. et al. (2012). J. Am. Chem. Soc. 134: 8703–8710. 78 Peng, Y.W., Gong, T.F., Zhang, K. et al. (2014). Nat. Commun. 5: 4406. 79 Zhao, J.S., Li, H.W., Han, Y.Z. et al. (2015). J. Mater. Chem. A 3: 12145–12148. 80 Zhang, J., Li, Z.J., Gong, W. et al. (2016). Inorg. Chem. 55: 7229–7232. 81 Gu, Z.G., Fu, W.Q., Liu, M., and Zhang, J. (2017). Chem. Commun. 53: 1470–1473. 82 Navarro-Sanchez, J., Argente-Garcia, A.I., Moliner-Martinez, Y. et al. (2017). J. Am. Chem. Soc. 139: 4294–4297. 83 Nuzhdin, A.L., Dybtsev, D.N., Bryliakov, K.P. et al. (2007). J. Am. Chem. Soc. 129: 12958–12959. 84 Hailili, R., Wang, L., Qy, J.Z. et al. (2015). Inorg. Chem. 54: 3713–3715. 85 Hartlieb, K.J., Holcroft, J.M., Moghadam, P.Z. et al. (2016). J. Am. Chem. Soc. 138: 2292–2301. 86 Xie, S.M., Zhang, Z.J., Wang, Z.Y., and Yuan, L.M. (2011). J. Am. Chem. Soc. 133: 11892–11895. 87 Xie, S.M., Wang, B.J., Zhang, X.H. et al. (2014). Chirality 26: 27–32. 88 Xie, S.M., Zhang, X.H., Wang, B.J. et al. (2014). Chromatographia 77: 1359–1365. 89 Liu, H., Xie, S.M., Ai, P. et al. (2014). ChemPlusChem 79: 1103–1108. 90 (a) Yang, J.R., Xie, S.M., Liu, H. et al. (2015). Chromatographia 78: 557–564. (b) Yang, J.R., Xie, S.M., Zhang, J.H. et al. (2016). J. Chromatogr. Sci. 54: 1467–1474. 91 Gu, Z.G., Fu, W.Q., Wu, X., and Zhang, J. (2016). Chem. Commun. 52: 772–775. 92 Liu, J., Wang, F., Ding, Q.R., and Zhang, J. (2016). Inorg. Chem. 55: 12520–12522.

953

16 Catalysis in Metal–Organic Frameworks: Relationship Between Activities and Structures Yuan-Biao Huang 1,2 , Teng Zhang 1,2 , and Rong Cao 1,2 1 Chinese Academy of Sciences, State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, 155 Yangqiao Road West, Fuzhou, Fujian 350002, P.R. China 2 University of Chinese Academy of Sciences, No.19(A) Yuquan Road, Shijingshan District, Beijing 100049, P.R. China

16.1 Introduction Catalysis is one of the earliest proposed and explored applications of metal–organic frameworks (MOFs). When Robson and coworker pioneered the research of “infinite polymeric frameworks” constructed from coordination bonds in early 1990s, they proposed catalysis as one of the potential applications for such material [1, 2]. Later in 1994, Fujita et al. reported the catalytic activity of a square coordination network Cd(4,4′ -bipyridine)2 (NO3 )2 [3], which is widely recognized as the first report of catalytic active MOF. Compared to other porous catalytic materials such as zeolites or mesoporous silica, MOFs have unique advantages. First, MOFs are constructed from molecular organic linkers with well-defined structure, resulting in uniform local environment for catalytic active sites. Such structural uniformity assures single-site catalysis and better selectivity. Second, the high diversity of organic linkers leads to MOF structure with much larger variety, thus enabling systematic tuning of pore sizes, channel hydrophilicity/hydrophobicity, chirality, and other properties that affect catalysis. Especially, incorporation of catalytic active metal sites is more convenient in MOFs than in zeolites or mesoporous silica because the MOF structural diversity allows pre-installation of different metal binding sites, while only oxo/hydroxyl groups are available in zeolites and mesoporous silica. More importantly, the molecular nature of MOF catalysts allows for systematic engineering of the active sites and/or open channel sizes. In this way, it is possible to fine-tune the orientation and distance of catalytic centers to achieve catalytic site isolation for single-site catalysis or cooperative/synergistic catalysis with adjacent sites, which is difficult to achieve in molecular, homogeneous systems. Despite the advantages, MOF catalysts still suffer several concerns. Most MOF catalysts are not stable under harsh catalytic conditions due to their molecular nature. Diffusion of substrates and products through the channels/pores is another Advanced Structural Chemistry: Tailoring Properties of Inorganic Materials and their Applications, First Edition. Edited by Rong Cao. © 2021 WILEY-VCH GmbH. Published 2021 by WILEY-VCH GmbH.

954

16 Catalysis in Metal–Organic Frameworks: Relationship Between Activities and Structures

potential issue, while in some cases, diffusion is found to be the rate-determining factor. In the last 20 years, chemists have developed multiple strategies to account for these concerns and significant success was achieved. However, it is still not close to a perfect solution of these issues for practical application purpose. Chemists classify MOF catalysis in different ways. Hupp and coworkers divided MOF catalysis into two major classes: the “opportunistic” catalysis by defects or labile coordination bonds on metal nodes or by degradation products of MOFs and the “designed” catalysis by a well-defined site on metal nodes, linkers, or encapsulated species [4]. In this chapter, we will follow a similar strategy to give a brief but broad scope overview of the MOF catalysis area. We will first introduce the catalysis on metal nodes (secondary building units [SBUs]) and modified SBUs, then organic linkers, then pore-encapsulated species. Three major approaches are discussed next to show the unique advantage of MOFs compared to traditional heterogeneous and homogeneous catalytic systems: heterogeneous asymmetric catalysis, cooperative or sequential catalysis that involves more than one active sites, and synergistic photocatalysis.

16.2 Catalysis by Metal Nodes 16.2.1 Lewis Acid Catalysis MOFs are constructed by metal nodes, which could be acted as coordinatively unsaturated metal sites (CUMs) to catalyze many reactions, such as Lewis acid catalysis, oxidation, and Suzuki reactions [5]. Lewis acid catalysis is a typical reaction for MOFs using metal nodes as active CUMs. A wide range of metal ions (e.g. Cr3+ , Al3+ , Fe3+ , Sc3+ , Mn2+ , Co2+ , Cu2+/1+ , Zn2+ , and Zr4+ ) can be acted as CUMs to promote Lewis acid catalysis. In general, CUMs could be accessible for substrates after removal of weakly coordinated solvent molecules by vacuum heating treatment. MIL-101(Cr) (Cr3 (F,OH)(H2 O)2 O[(O2 C)-C6 H4 -(CO2 )]3 ⋅nH2 O (n ≈ 25), MIL = Matérial Institut Lavoisier) [6] consisted of μ3 -oxo bridged chromium(III)-trimers that crosslinked terephthalate groups. The chromium ions possess a pseudo-octahedra coordination that are composed by the μ3 -oxo-O atom in the middle of the chromium trimers and four oxygen atoms from the terephthalate linkers (Figure 16.1). Each chromium is coordinated by a water molecule, a fluorine atom, or a hydroxyl group. Thus, the Cr3+ sites in the mesoporous MIL-101(Cr) can be served as CUMs after removal of the terminal coordinated water for various Lewis acid catalysis reactions. Cyanosilylation of aldehydes is a typical reaction to examine the Lewis acidity catalysis of CUMs (Figure 16.2). Compared with other MOFs such as Cu3 (BTC)2 (BTC = 1,3,5-benzenetricarboxylate), the chromium(III) in MIL-101 as CUMs were found to be highly active in the cyanosilylation of aldehydes reaction with a very low catalyst loading of only 0.5 mol% [7]. In a typical reaction, 4 mmol benzaldehyde in heptane can be completely converted in the presence of 15 mg MIL-101 at 313 K after three hours with a high product yield of 98.5%. The higher activity is due to the substrates that can be easily contacted with

16.2 Catalysis by Metal Nodes

∼14.7 Å ∼12 Å

∼16 Å

Hexagonal windows

Pentagonal windows

(a)

(b)

∼34 Å

∼29 Å

Cr O C

Mesoporous cages

(c)

(d)

Figure 16.1 (a,b) Ball-and-stick view and free dimensions (Å) of the pentagonal and hexagonal windows. (c,d) Ball-and-stick view of the two cages. Chromium octahedra, oxygen, fluorine, and carbon atoms are in green, red, red, and blue, respectively. Source: Reproduced with permission from ref. [6].

OSiMe3

O

MIL-101(Cr) +

Figure 16.2

Si

CN

CN

Addition of trimethylsilylcyanide to benzaldehyde by Cr CUMs of MIL-101(Cr).

the open Cr(III) sites through the mesoporous cages (2.9 and 3.4 nm) and accessible microporous windows (1.2 and 1.4 nm) (Figure 16.1). After cyanosilylation for 15 minutes reaction, the catalyst was filtered and the filtrate was continued stirring at the same reaction temperature. Neither additional benzaldehyde is converted nor is product formed, suggesting that the reaction is a heterogeneous mechanism. Moreover, MIL-101 is very stable and chromium(III) cannot be reduced by the benzaldehyde at high reaction temperatures. Although the product yield decreased from 98.5% to 91.3% in the third run, the powder X-ray diffraction (PXRD) pattern revealed that the framework of MIL-101 is intact after using three times.

955

956

16 Catalysis in Metal–Organic Frameworks: Relationship Between Activities and Structures

16.2.2 Oxidation Reaction Besides Lewis acid catalysis, the open metal sites in the metal nodes in MOFs could promote oxidation reactions such as the oxidation of tetralin. After removal of terminal water molecules, the atomically distributed Cr(III) sites (Figure 16.1) in mesoporous MIL-101(Cr) showed high activity in the oxidation of tetralin to produce 1-tetralone, which can be used as an additive to enhance the cetane number in diesel fuels [8]. When using tert-butyl hydroperoxide (t-BuOOH) or in situ generated acylperoxy radicals through trimethylacetaldehyde and O2 as oxidants, the tetralin oxidation reaction could yield the major product 1-tetralone with 73% conversion and c. 86% selectivity at 100 ∘ C. Compared with the water-coordinated Cr(III) sites with 36% conversion and 68% selectivity to 1-tetralone, the dried MIL-101 showed significantly higher activity with 68% conversion and 86% selectivity to 1-tetralone. This is because the reagent molecule could coordinate to the Cr(III) sites to form the active intermediate {t-BuO–O–Cr} or {–CO2 –O–Cr} in the presence of a non-coordinating solvent. Thus, when using t-BuOOH as an oxidant, the conversion decreased in the following order: C6 H5 Cl > C6 H6 ≫ CH3 CN > THF. The heterogeneous catalyst MIL-101(Cr) exhibited very stable in liquid phase oxidation of tetralin at a fixed temperature (80 ∘ C). The leached Cr content in the filtrate solution was only 0.12 ppm, and the catalyst can be used many times and the conversion and product selectivity after using five cycles were virtually identical with those of the first time. As an important subclass of MOFs, zeolitic imidazolate frameworks (ZIFs) are constructed by mimicking the structure feature of zeolites by the tetrahedrally coordinated divalent cations (M2+ = Zn2+ or Co2+ ) and the imidazolate ligands (im− ) [9, 10]. Similar coordination ability between O and N atoms to divalent metal centers gives a chance to combine TO4 (T = Mo6+ or W6+ ) and Mim4 (M = Zn or Co) tetrahedral units together. Based on this consideration, Zhang’s group reported such kind of hybrid ZIFs (denoted hybrid ZIF (HZIFs); general formula (Figure 16.3): O

O (a)

T

R2

mi (b)

O

M N

im M

M

O

X

N M

im im

M

im im

R2

O

O

X

N

N

M

O R1

M

R1

O

(c)

Figure 16.3 Construction of HZIFs (c) by self-assembly of TO4 (T = Mo6+ or W6+ , (a) and Mim4 (M = Zn or Co, (b) tetrahedral units. Source: From Wang et al. [11] with permission from Wiley-VCH.

16.2 Catalysis by Metal Nodes

M4 (im)6 TO4 ) with catalytic active TO4 (T = Mo6+ or W6+ ) building blocks, which presents a new class of porous materials filling the gap between zeolites and ZIFs [11]. Two HZIFs (HZIF-1Mo and HZIF-1W) were successfully synthesized by the self-assembly of Zn2+ cation, MoO4 2− or WO4 2− anions, and 2-methylimidazole (2-mim) under solvothermal conditions. The introduction of TO4 units generates three merits: (1) Unique framework with sdt topology: The Zn atoms were linked by 2-mim ligands to form large [Zn24 (2-mim)36 ] cage with an effective diameter of 12.5 Å and a pore aperture of 3.3 Å in HZIF-1W just the same as the subunit in ZIF-8 metal-azolate framework (MAF-4). Then each Zn vertex of this cage was linked by another T (T = W) node via the Zn–O–T connectivity. Although it was surrounded by the inorganic TO4 units, each window of the cage remained open and faces next to symmetry-related window of the adjacent cage (Figure 16.4).

(a)

(b)

(c)

(d)

Figure 16.4 Each WO4 2− anion links four truncated octahedral cages of [Zn24 (2-mim)36 ] in the structure of HZIF-1W (a), and each cage is surrounded by the WO4 2− anions (b). The resulting 3D framework with free voids (green balls) in the cages (c), and its natural tiling showing the packing of the cages (red tiles) (d). Source: From Wang et al. [11] with permission from Wiley-VCH.

957

16 Catalysis in Metal–Organic Frameworks: Relationship Between Activities and Structures

Different to ZIF-8 with sodalite (SOD) topology, the whole framework topology is identified as the 4-connected net with the symbol sdt. (2) Exceptional thermal and acid–base stability: The PXRD patterns at 550 ∘ C were coincident with the corresponding patterns simulated from single-crystal X-ray diffraction (XRD) structures, confirming that the structural and thermal stabilities of HZIFs were significantly greater than those of ZIFs and many other porous MOFs. More importantly, recent results also proved that HZIF-1 has high acid–base stability (with pH values of 1–14). (3) Remarkable catalytic properties: Different to traditional SiO4 or AlO4 units in aluminosilicate zeolites, the framework of HZIF-1 contained catalytic active MoO4 or WO4 units. Selective oxidation of benzyl alcohol to benzaldehyde was chosen as the demo reaction to examine their catalytic properties. The results show the 100% selectivity to benzyl aldehyde (no trace of benzoic acid) for both catalytic materials, and the conversion of benzyl alcohol is 51.4% and 72.2% for HZIF-1Mo and HZIF-1W, respectively, under the temperature of 353 K for six hours in solution of water.

16.2.3 Suzuki–Miyaura Coupling Reaction Palladium reagents are typical catalysts for Suzuki–Miyaura coupling reaction. In 2015, starting from dimetallic PdII clip 1 (Figure 16.5) and organic ligand 2 (Figure 16.5), Sun and coworkers have prepared segmental and continuous hexagonal-packed mesoporous Pd-based metal–organic nanotubes (MMONTs, 4a and 5, respectively; Figure 16.5) with outside diameters of up to 4.5 nm and channel sizes of 2.4 nm [12]. The two novel nanotubular materials (4a, 5) have excellent catalytic performance for Suzuki–Miyaura carbon–carbon coupling reaction with as low as 0.45 mmol% loading. Desirable products were obtained in high yields with 5 as the catalyst, whereas only moderate yields were observed using 4a instead. Besides, kinetic studies revealed that the catalysis by 5 was over 10-fold faster than 4a. Therefore, a new way is provided for the controllable preparation and application of novel metal–organic nanotubular materials.

N N

N N

Pd

O O N O

N N

Pd

N

O

+2

HN

III

N Pd

N

N N

(i)

III

Ar

Pdll

III

Pd

(ii)

N

(HO) B 2

(iii)

III

OR

2

Me

0.45 mol% loading

III 1

I +

N N

N

N

O

O

3 Ar

III

958

(a)

Me

(b) N

N Pd

N

N Pd

4

4a

5

Figure 16.5 Assembly of segmental and continuous hexagonal-packed mesoporous Pd-based metal–organic nanotubes (MMONTs) 4a and 5, respectively. Source: From Zhang et al. [12] with permission from Wiley-VCH.

16.3 Catalysis by Functionalized Linkers

16.2.4 Other Reactions In addition, CUMs in the metal nodes of MOFs can effectively catalyze other reactions, such as Friedel–Crafts benzylation reaction and ring-opening reaction. [Fe(BTC)] (BTC = 1,3,5-benzenetricarboxylate), also named MIL-100(Fe), is constituted with trimers of iron octahedra connected by vertex μ3 -O linked by the BTC moieties, which form two types of mesoporous cages with apertures of 2.5 and 2.9 nm, accessible through microporous windows of 5.5 and 8.6 nm (Figure 16.6) [13]. After removal of coordinated solvent molecules, the substrate can be activated by the Fe ions. For example, MIL-100(Fe) could efficiently promote the Friedel–Crafts benzylation reaction of benzene with benzyl chloride to produce diphenylmethane (DPM) with 100% conversion and nearly 100% DPM selectivity at 70 ∘ C after a short induction period (five minutes) [13]. The high benzylation activity of MIL-100(Fe) may be ascribed to the redox property of Fe CUMs (Fe3+ + e2 ↔ Fe2+ ), which plays an important role in activating both the reactants. In addition, [Fe(BTC)] can promote the ring opening of styrene oxide with alcohols (Figure 16.7) and aniline [14]. The catalysis mechanism was proposed (Figure 16.8): initially, styrene oxide is coordinated to the Fe(III) ions of [Fe(BTC)] via an acid–base interaction to increase electrophilic nature of the carbon attached to the phenyl group or the one with more alkyl substituents. Subsequently, the positive charged carbon is attacked by methanol to give the product.

16.3 Catalysis by Functionalized Linkers One of the unique features for MOFs is that MOFs can be constructed from almost infinite choices of organic linkers. Modifying linkers with catalytic active functionalities represents an effective and widely used strategy to develop MOF catalysts. In this section, we will briefly discuss the incorporation of different catalytic active sites on MOF linkers and the resulting catalysts. Introducing basicity to MOFs with amino, pyridyl, or other basic functional groups is among the earliest efforts to make functionalized MOFs. These MOF basic catalysts have been employed in Knoevenagel reaction, transesterification, and epoxide cycloaddition reactions. Post-synthetic modification (PSM) of NH2 -UiO-66 with 2-methylaridine led to a primary-amine-functionalized MOF, which catalyzed Knoevenagel reaction with high activity [15]. Carbon dioxide-epoxide cycloaddition is another reaction that is widely explored in basic MOF catalysis. Amino-functionalized MOFs such as isoreticular metal-organic framework (IRMOF-3) or NH2 -UiO-66 have shown good catalytic activity for the reaction [16]. Bronsted acid-functionalized MOFs are less common. Jiang and coworkers reported a sulfonic acid-functionalized MOF, MIL-101-SO3 H, for catalytic alcoholysis of epoxides. The MOF was synthesized via solvothermal process followed by post-synthetic proton exchange. It catalyzes epoxide ring opening under ambient conditions, while the sodium salt form has reduced activity [17]. Porphyrinic MOFs are a big family of MOFs that are widely explored as catalysts and photocatalysts. Wu and coworker used Mn(III) porphyrin-based MOF CZJ-4

959

960

16 Catalysis in Metal–Organic Frameworks: Relationship Between Activities and Structures

O

OH

O

HO O

OH

(a)

(b)

25 Å

~4.7-5.5 Å

~8.6 Å

29 Å

(c)

(d)

Figure 16.6 Structure of MIL-100(Fe). (a) A trimer of iron octahedra and trimesic acid. (b) Schematic view of one unit cell of MIL-100(Fe). (c) The two types of cages in polyhedral mode. (d) Pentagonal and hexagonal windows in balls and sticks (Fe, gray; O, red; C, black). Source: From Horcajada et al. [13] with permission from Royal Society of Chemistry.

(Figure 16.9) for olefin epoxidation and C–H oxidative activation [18]. Zhou et al. first reported porphyrinic MOFs based on water-stable zirconium–carboxylate metal nodes. The PCN-222 (porous coordination network) (Figure 16.9) MOF series constructed from Zr6 clusters and tetrakis(4-carboxyphenyl)-porphyrin linkers were stable under boiling water or strong acidic conditions. The PCN-222(Fe) catalyzed selective oxidation of pyrogallol, mimicking peroxidase functionality [19].

16.3 Catalysis by Functionalized Linkers

OCH3

O

OH CH3OH MIL-100(Fe)

Figure 16.7

Ring opening of styrene oxide with methanol by MIL-100(Fe).

O

Fe

Fe

δ–

Fe

δ+

O

represents [Fe(BTC)]

+

Fe

Fe O

CH3OH

OCH3 OH

Figure 16.8 Proposed mechanism for the ring opening of styrene oxide with methanol catalyzed by MIL-100(Fe). Source: From Dhakshinamoorthy et al. [14] with permission from Wiley-VCH.

(a)

(b)

(c)

Figure 16.9 The structure of CZJ-4 (left), PCN-222 (middle), and PCN-224 (right). Source: From Yang et al. [18]; Feng et al. [19]; Feng et al. [20] with permission from American Chemical Society and Wiley-VCH.

Constructed from the same building blocks but with reduced connection number, PCN-224 (Figure 16.9) MOF series exhibit higher surface area and better catalytic activities. The PCN-224(Co) catalyzes CO2 /epoxide cycloaddition, and substituted PCN-224(Fe) catalyzes aerobic oxidation of 3-methylpentane [20, 21]. Catalytic active transition metal centers can also be incorporated through PSM. An intensively studied example is the bipyridine-Universitetet i Oslo

961

16 Catalysis in Metal–Organic Frameworks: Relationship Between Activities and Structures

(bpy-UiO) MOF series built from 2,2′ -bipyridine-derived linkers and Zr6 clusters. The bipyridine site is left open during MOF synthesis due to its low affinity toward hard Lewis acidic Zr(IV), which allows further post-synthetic metallation. Cohen and coworker first reported post-synthetic metallation of bpy-UiO by Pd(MeCN)2 Cl2 and Suzuki–Miyaura coupling catalyzed by the incorporated bpy-Pd centers [22]. Lin and coworkers introduced Ir centers to these bipyridine-derived MOFs and obtained excellent catalysts for broad scope C–H activation reactions (Figure 16.10). Impressively, these MOFs are more active than their homogeneous counterpart with turnover number (TON) enhancement as high as 1250 times and they can be recycled and reused for more than 15 times without significant activity loss [23]. In this case, the dramatic enhanced performance is attributed to the site isolation effect. In homogeneous, molecule-based transition metal catalysis, intermolecular decomposition is a major pathway of catalyst deactivation. Structurally bulky, multidentate ligands have been designed to protect the transition metal sites from intermolecular decomposition. It has been a widely applied strategy in transition metal catalysis; yet it often compromises catalytic activity against stability, not to mention the increasing costs to make such elaborately designed ligands. In MOF catalysts, such intermolecular decomposition is eliminated since catalytic sites are generally too far to interact after incorporated into the walls of channels and cavities. Therefore, MOF catalysts can achieve higher activity with more open metal sites without worrying about stability issues. Isolated sites of first-row transition metals show even better performance. Lin and coworkers studied the catalytic activity of single-site Co center incorporated in bipyridine-derived MOFs. Through post-synthetic metallation followed by reduction, they can obtain formally Co(0) species with low oxidation state and high

O

R2 R2

R1

R1

O Si Et2

+ Et2SiH2

50

0

R

+ B2(pin)2

Homogeneous catalyst (5% lr) MOF catalyst (0,1% lr)

0

Bpin

R

6

3 Time (d)

100

Et2SiH2 + R

GC conversion (%)

100

NHMc

R

NMc Si Et2

GC conversion (%)

962

50

0 5

10 Runs

15

Figure 16.10 MOF-incorporated bipyridine–Ir species for broad scope C–H activation. Source: From Manna et al. [23] with permission from American Chemical Society.

16.4 Catalysis by Species in Pores

TON~2 500 000 R

R Bpin R

N

THF

TON~22 000

Co THF

N

R Ar

Bpin

TON~1000 Ar

H

Figure 16.11 MOF-incorporated bipyridine–Co(0) species for olefin hydrogenation, hydroboration, and C–H borylation. Source: From Zhang et al. [24] with permission from American Chemical Society.

catalytic activity for olefin hydrogenation and hydroboration (Figure 16.11). This species is unstable in solution and quickly decomposes into Co nanoparticles (NPs) at catalysis, yet remains stable and active in MOFs. This work shows powerful potential of site-isolated MOF catalysts in transition metal catalysis [24]. Yuan and coworkers reported incorporation of Zn(II) Lewis acidic site in the bpy-UiO MOF. The MOF–Zn complex is an active catalyst for intramolecular hydroamination of alkynes, while the bipyridine/Zn2+ homogeneous counterpart gives alkyne hydration product. The different selectivity was attributed to microenvironment differences of MOF pores and homogeneous solvents [25].

16.4 Catalysis by Species in Pores The highly porous MOFs with well-defined pore structures could make them promising candidates for loading catalytically active guest species, including metal nanoparticles (MNPs), polyoxometalates, metal complexes, and enzyme. These active species were confined in the pores of MOFs and could not leach during the catalysis. The substrate can be easily accessible to the active species through the pores.

16.4.1 MNPs Supported in MOFs for Catalysis MNPs have shown highly efficient catalytic performances due to the high density of active atoms that could be exposed to the substrates [26]. However, these active MNPs are easily aggregated to large bulks and resulted in loss of catalytic properties. Therefore, suitable stabilizers and/or supports are generally required to stabilize

963

964

16 Catalysis in Metal–Organic Frameworks: Relationship Between Activities and Structures

MNPs. Highly porous MOFs have tunable pores that can limit MNPs’ growth to the same size as the pore width. Moreover, the uncoordinated organic functional groups in the linkers could stabilize the MNPs; thus, it is no need to add any surfactant during synthesis. Now, several effective methods have been developed for the immobilization of MNPs inside the pores of MOFs, including solution infiltration, incipient-wetness impregnation, ion-exchange method, chemical vapor deposition (CVD), and double-solvent method. MIL-101(Cr) has two types of mesoporous cages (2.9 and 3.4 nm) accessible through microporous windows (1.2 and 1.4 nm) (Figure 16.1) [6]. Thus, the metal salts can easily enter into the mesocages via the large windows, followed by reduction with H2 or NaBH4 to obtain highly dispersed MNPs inside the cages. Due to the smaller windows than those of cages, the MNPs can be confined in the mesoporous cages that restrict their growth during the reduction process. Therefore, very small MNPs with the sizes between the mesocages and the window width can be immobilized in MIL-101. Using this solution infiltration method, highly dispersed Pd NPs with the average size of 2.6 nm were immobilized into the cages of MIL-101(Cr) [27]. Because of the very small Pd NPs dispersed in the cages, only 0.1 mol% Pd catalyst (Pd@MIL-101(Cr)) can promote the direct C2 arylation of substituted indoles with various iodine benzenes with high conversion and selectivity (Figure 16.12). The Pd@MIL-101(Cr) also exhibited high efficiency in the direct C2 arylation of indoles with aryl boronic acids in the presence of O2 or 2,2,6,6-tetramethylpiperidine N-oxyradical as an external oxidant in CH3 COOH (Figure 16.13) [28]. Moreover, the halogen (–F, –Cl, –Br, –I) group-functionalized indoles can be tolerated and C2-arylated products in high yields and excellent selectivity can be obtained. The Suzuki reaction does not happened due to the smaller windows of MIL-101(Cr) than that of the products of Suzuki reaction and the acidic media inhibited the side reaction. Both of the two C–H activations are in organic solvent, which is not environmentally benign compared with that using water as reaction medium. However, the reactants are difficult to contact with the active sites of heterogeneous catalysts that dispersed in water. To solve the problem, ultrafine Pd NPs were encapsulated with the hydrophobic perfluoroalkane-functionalized

R2

Pd

Pd

R2

H Pd

Pd Pd

R1

Pd

3

N R1

Pd

R I

R3

Pd

N

Pd

Pd

Figure 16.12 Direct C2 arylation of substituted indoles with various iodine benzenes catalyzed by Pd NPs encapsulated in MIL-101(Cr). Source: From Huang et al. [27] with permission from Wiley-VCH.

16.4 Catalysis by Species in Pores

R1 ,R2 = electron-donating substituents O2 R2

Pd

Pd

H N +

R2

Pd

R3

Pd

R1

Pd Pd

N

R3 (HO)2B

Pd

R1

Pd

Pd

R1 ,R2 = electron-withdrawing substituents TEMPO

Figure 16.13 Direct C2 arylation of substituted indoles with aryl boronic acids catalyzed by Pd NPs encapsulated in MIL-101(Cr) in the presence of O2 or 2,2,6,6-tetramethylpiperidine N-oxyradical as an external oxidant in CH3 COOH. Source: From Huang et al. [28] with permission from Wiley-VCH.

Pd(acac)2

SALI method RCOOH, DMF O C H Zn

O C H F Zn

R=

F2 C F3C

C F2

F2 C

CF2

F9-NU-1000

;

F3C

F2 C

F2 C C F2

C F2

F15-NU-1000

(1) Incipient-wetness impregnation method (2)H2, 200°C

C

F2 C CF2 ; F3C

O H F Zn

F2 C

F2 C C F2

C F2

F2 C C F2

CF2

F19-NU-1000

Figure 16.14 Schematic representation of introduction of perfluoroalkane chains into the pores of NU-1000 and immobilization of the Pd NPs into the pores of perfluoroalkane-functionalized NU-1000 using incipient-wetness impregnation method. Source: From Huang et al. [29] with permission from Elsevier Inc.

mesoporous NU-1000 using incipient-wetness impregnation method, followed by reduction with H2 at 200 ∘ C for three hours (Figure 16.14) [29]. Thus, the organic substrates can be easily contacted with the hydrophobic groups-modified Pd active sites in water. The resultant Pd NPs supported on the perfluoroalkane exhibited high activity and regioselectivity in the direct C–H arylation of indoles with iodine benzenes in water. The yield for the C2 arylation of indole was improved c. seven times than that of the Pd NPs supported on the unmodified MOF (Pd@NU-1000). The amine groups-stabilized Pd NPs can be prepared using an anion-exchange method [30]. The amine groups of MIL-101(Cr)–NH2 were first ionized with an aqueous HCl solution. Then, the chloride anions in the mesocages of MIL-101(Cr)–NH2 were exchanged with [PdCl4 ]2− and followed by reduction using NaBH4 to obtain the highly dispersed Pd NPs immobilized in MIL-101(Cr)–NH2

965

966

16 Catalysis in Metal–Organic Frameworks: Relationship Between Activities and Structures

Pd

Cl Pd

HCOONH4

Pd

Pd

Pd Pd Pd

Pd Pd

H2O, 25°C R

Pd

Pd

Pd Pd

Pd Pd

R

Pd

Figure 16.15 Dehalogenation of aryl chlorides catalyzed by Pd/MIL-101(Cr)–NH2 using HCOONH4 as hydrogen source in water at room temperature. Source: From Huang et al. [30] with permission from Elsevier Inc.

(1) Double solvents method H2PtCl6/H2O (small amount) Hexane (large amount)

(2) H2/He reduction 200°C, 5h

MIL-101

Pt@MIL-101

Figure 16.16 Schematic representation of fabrication of Pt NPs inside the MIL-101(Cr) matrix using double-solvent method. Source: From Aijaz et al. [32] with permission from American Chemical Society.

with small size (2.5 nm; denoted as Pd/MIL-101(Cr)–NH2 ). The resulted Pd catalysts showed extremely high activity in the dehalogenation of aryl chlorides using HCOONH4 as hydrogen source in water (Figure 16.15). To completely immobilize MNPs in the pores of MOF, CVD and double–solvent methods have been developed [31, 32]. Using CVD method [31], the volatile [Ru(cod)(cot)] molecules were first successfully loaded in the pores of MOF-5, followed by hydrogenolysis to result Ru@MOF-5, in which the Ru NPs with 1.5–1.7 nm are embedded in the pores. The obtained Ru@MOF-5 could catalyze the hydrogenation of benzene to produce cyclohexane with 25% yield. Using a double-solvent strategy followed by H2 reduction, ultrafine Pt NPs were successfully encapsulated in the mesocages of MIL-101(Cr) [32]. The double-solvent method involves the hydrophilic H2 PtCl6 aqueous solution that can enter the hydrophobic pores of MIL-101(Cr) dispersed in hexane (Figure 16.16). The key point is that the volume of H2 PtCl6 solution should be less than or equal to the pore volume of MIL-101(Cr). The transmission electron microscopy (TEM) and electron tomographic measurements clearly revealed that the ultrafine Pt NPs with an average size of 1.8 ± 0.2 nm were uniformly distributed in the interior cavities of MIL-101(Cr). The obtained Pt@MIL-101(Cr) showed high activities in the hydrolysis of liquid-phase ammonia borane and gas-phase CO oxidation.

16.4 Catalysis by Species in Pores

16.4.2 Polyoxometalates Encapsulated in MOFs for Catalysis Polyoxometalates, as a unique class of metal-oxide clusters, have shown highly catalytic performance due to their acid–base and redox properties that can be tuned easily by simply changing polyanion chemical composition [33]. However, their utilization for industrial application had been limited due to the low special surface area (1–10 m2 /g) as heterogeneous catalysts or difficult to be separated from catalytic system as homogeneous catalysts. Considerable efforts have been directed toward their heterogenization onto various solid supports via dative, covalent, or electrostatic binding, such as silica, active carbon, cucurbit[n]uril, mesoporous molecular sieves, ion-exchanged resin, especially the recent emerging MOFs. MIL-101(Cr) is one of the excellent MOF materials due to its high special surface area, pore volume, and thermal (up to 300 ∘ C) and chemical stability to water and common organic solvents [6]. With those advantages, MIL-101(Cr) had been used as support to encapsulate phosphotungstic acid (PTA, one of polyoxometalates) for oxidative desulfurization [34, 35]. However, the PTA anions (1.3–1.4 nm diameters) would leach out from the MIL-101(Cr) due to the large windows (1.6 nm diameter) of the mesocages in MIL-101(Cr). To solve this problem, two strategies had been applied, one is the decrease of window size of MOFs. For example, PTA were encapsulated into MIL-100(Fe), which is also a stable MOF with high special area and pore volume but with smaller window size (0.86 nm diameter) in mesocages [36]. The relatively smaller window compared with the PTA ensures that the restricted PTA is not leached from the mesocages of MIL-100(Fe). The other one is the increase of interaction force between the support MOF and PTA anions, such as electrostatic interaction. PTA was loaded into the cages of amine-functionalized MIL-101(Cr), in which the amino group can increase the electrostatic interaction between MOF and PTA [37]. In the 0.1 M HCl aqueous solution, the protonated amino groups have strong interaction with PTA anions. The encapsulated PTA in MIL-101(Cr)–NH2 show high efficiency for oxidative desulfurization and good recycle ability (Figure 16.17). After six consecutive cycles, only 4% PTA lost and 99% conversion rate was retained, which

Figure 16.17 Oxidative desulfurization catalyzed by PTA@MIL-101-(Cr)–NH2 . Source: From Wang et al. [37] with permission from Royal Society of Chemistry.

S

Model oil MeCN

Fast-diffusion H2O2 S

O

O S

Cr

Cr

O

O C NH3 + Electrostatic interaction C

O

O

Cr

Cr

PTA@MIL-101(Cr)–NH2

967

968

16 Catalysis in Metal–Organic Frameworks: Relationship Between Activities and Structures

can be attributed to the protection from amino groups in MIL-101(Cr)–NH2 . In contrast, without the amino group protection, the catalyst PTA@MIL-101(Cr) lost 45% PTA only after three runs prepared by impregnation method.

16.4.3 Metal Complexes Trapped in MOFs for Catalysis The homogeneous metal complexes catalyst usually shows high activity and selectivity. However, an efficient recovery of the metal complexes is required to make this system sustainable. Moreover, the encapsulation of metal complexes into the pores of porous materials could prevent its decomposition. Metal–organic polyhedral (MOP) is a newly developed class of inorganic–organic discrete coordination complexes with applications limited to mild conditions due to its instability [38]. Encapsulation of MOP into the cavity of a host with windows smaller than the MOP is a way to circumvent this problem. Li’s group developed a new and efficient hydrophilicity-directed approach (HDA) to encapsulate MOP molecules beyond the aperture size limitation in the nanospace of MOFs, as exemplified by the self-assembly of M6 L4 into MIL-101(Cr) (Figure 16.18) [39]. The confined M6 L4 exhibited significantly improved catalytic efficiency and stability compared with the pristine MOP material in the selective oxidation of benzyl alcohol to benzaldehyde. As compared to M6 L4 , the benzaldehyde yield of M6 L4 ⊂ MIL-101 was enhanced by a factor of c. 3.5 at the first run. More impressively, the benzaldehyde yield obtained on the M6 L4 ⊂ MIL-101 system was nearly 20 times higher than that of M6 L4 after being reused for five times. The HDA strategy requires neither particular

N

N

4

N

+

NH2 6

NO3 Pd

NH2

N

N

NO3

N (denoted as:

)

(denoted as:

)

(denoted as:

)

(a)

(b)

MIL-101

MIL-101 filled with H2O

n-Hexane

Figure 16.18 Preparation of M6 L4 (a) and M6 L4 ⊂ MIL-101 (b) by double-solvent method. Source: From Qiu et al. [39] with permission from American Chemical Society.

16.4 Catalysis by Species in Pores

types of MOFs nor destruction of the MOFs, thus offering a versatile approach to encapsulate a broad range of MOP complexes into MOFs with significantly enhanced performances for various applications.

16.4.4 Enzyme Trapped in MOFs for Catalysis Enzymes have great potential in chemical manufacturing such as pharmaceuticals and food processing. However, the low operational stability hampered their reuse. Therefore, engineering of enzymes is highly desirable. MOFs are one class of promising alternatives to encapsulate enzymes for catalysis. However, the immobilization of enzymes is challenging for MOFs mainly due to the small pore size, which cannot accommodate the bulky enzyme molecules. Although some MOFs have enough pore size, the one-dimensional (1D) channel-like structures usually cause severe enzyme leaching and aggregation. Zhou and coworkers developed a series of stable MOFs (designated as PCN-332 and PCN-333) with rationally designed ultra-large mesoporous cages for enzyme encapsulation [40]. PCN-333(Al) exhibits the largest cage (5.5 nm) among all reported MOFs and shows high stability in aqueous solutions with pH values ranging from 3 to 9, making it an extraordinary candidate for enzyme encapsulation (Figure 16.19) [40]. Three enzymes with different sizes, horseradish peroxidase (HRP), cytochrome c (Cyt c), and microperoxidase-11 (MP-11), have been successfully encapsulated into PCN-333(Al) with extremely high loading amounts among all reported solid supports. These immobilized enzymes were applied for oxidation of o-phenylenediamine (by HRP) and 2,2′ -azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) (by Cyt c and MP-11). The encapsulated enzymes either maintain or surpass their catalytic activities over the free enzymes and exhibit smaller K m and better catalytic performance in organic solvents. Most importantly, these immobilized enzymes show almost no leaching during catalysis and recycling and maintain high catalytic activity which cannot be achieved by the materials with 1D channel-like structure such as SBA-15.

HRP Before enzyme loading

Cyt C

MP-11 After enzyme loading

Figure 16.19 Color variations of PCN-333(Al) when loaded with different enzymes at different concentrations. Source: Reproduced from Feng et al. [40] with permission from Nature Publishing Group.

969

970

16 Catalysis in Metal–Organic Frameworks: Relationship Between Activities and Structures

The work presented in this report enables the potential application of enzymes in fundamental processes such as nitrogen fixation, hydrogen production, and the conversion of natural gas into liquid fuels of high energy density with unprecedented efficiency.

16.5 Asymmetric Catalysis in MOFs Asymmetric catalysis is one of the most attractive applications of MOFs. After practically important asymmetric catalysis first achieved in 1960s, many chiral ligands, organocatalysts, and organometallic catalysts have been synthesized and studied. Industrial production of many enantiopure compounds relies on those catalysts nowadays, yet the difficulty in recycling the chiral catalysts increases the cost of the whole process. Thus, it is of practical importance to heterogenize such asymmetric catalysts and facilitate the recycle and reuse properties. While asymmetric catalysts can be readily incorporated into MOFs via direct (pre)catalyst incorporation or PSM, there is no other effective way to achieve asymmetric catalysis with traditional solid catalysts. The first example of MOF-based asymmetric catalysis was reported by Kim and coworkers in 2000. The reported MOF, POST-1, was built from tartaric acid derivative ligands and Zn3 O(RCOO)6 clusters. It catalyzes transesterification reaction with a modest enantiomeric excess (ee) value of 8% [41]. While such a low ee was generally viewed as nonselective, Lin and coworkers reported in 2005 the first asymmetric catalysis with a practically meaningful ee of ∼93% [42]. Just like the case of homogeneous asymmetric catalysis, MOF-based asymmetric catalysts also largely depend on “privileged” chiral ligands or compounds. The term “privileged” here means that such ligands can create effective asymmetric environments and control the enantioselectivity for reactions with unrelated mechanisms [43]. Some examples of privileged ligands are shown in Figure 16.20.

OH OH

PPh2 P

PPh2

BINOL

BINAP

N

MeDuPhos

O

N

O t-Bu

O

t-Bu

t-Bu

N t-Bu

t-Bu

Metal-salen complexes

Figure 16.20

O N

M t-Bu

P

Examples of privileged ligands.

bis(oxazoline)

16.5 Asymmetric Catalysis in MOFs

Metallosalen complexes were one of the extensively investigated families in MOF-based asymmetric catalysis. In 2006, Hupp and coworkers reported a simple Mn(salen)-derived MOF for asymmetric olefin epoxidation [44]. Lin and coworkers performed a detailed and systematic study in a similar system [45]. They synthesized a series of Mn(salen)-derived dicarboxylic acid ligands (Figure 16.21) with different lengths and built IRMOF-type pcu networks using zinc salt and these ligands. Interpenetration of the networks can be tuned by using different formamide solvents, and a series of MOFs, bearing the same catalytic active site but with different channel sizes, were obtained. The time-dependent conversion curves for the MOFs showed that the apparent reaction rates increase with the channel sizes, suggesting that the diffusion of substrates and/or products can be a significant factor for MOF catalysis. Since metallosalen catalysts are capable for a broad range of different reactions, it is possible to integrate different metallosalen complexes in a single MOF solid for sequential or tandem catalysis. Cui and coworkers reported systematic synthesis of multivariate metallosalen MOFs containing up to three different components [46]. Among those MOFs, the ternary Cu–Mn–Cr and Cu–Mn–Co ones catalyze sequential asymmetric epoxidation/ring-opening reactions. Some metallosalen-catalyzed reactions occur through cooperative bimetallic activation pathways. Such reactions are not easy to transfer to MOFs, for the orientation of metal centers needs careful designing. Cui and coworkers reported the first example of MOF-based cooperative asymmetric catalysis [47]. The MOF was constructed from cadmium–carboxylate 1D chains and Co(salen)-derived ligands (Figure 16.22). Chiral Co(salen) complex was known to be a good catalyst for hydrolytic kinetic resolution of epoxides via cooperative pathways, in which one Co(salen) center activates the epoxide and the other activates H2 O or other nucleophiles. The shortest Co–Co distance in the reported MOF is 6.28 Å, and this close proximity as well as appropriate spatial orientation of Co(salen) units allow cooperative activation of the reactants in MOF channels, resulting in a highly active catalyst.

CMOF-1

(a)

CMOF-3

(b)

CMOF-5

Conversion

1.0

0.5 L3-Me2

(c)

CMOF-1 CMOF-2 CMOF-3 CMOF-4 CMOF-5

0.0 0

(d)

CMOF-2

(e)

CMOF-4

3

6

9

Time (h)

Figure 16.21 The cavity sizes and time-dependent conversion curves for the Mn(salen)-derived MOFs CMOF 1–5. The sizes are shown as red balls: (a) CMOF-1 (1.4 nm), (b) CMOF-3 (2.0 nm), (c) CMOF-5 (1.8 nm), (d) CMOF-2 (2.6 nm), and (e) CMOF-4 (3.2 nm). Source: From Song et al. [45] with permission from American Chemical Society.

971

972

16 Catalysis in Metal–Organic Frameworks: Relationship Between Activities and Structures

Ni3

Figure 16.22 The building block (a) and framework structure (b) of the cadmium-metallosalen MOFs. Source: From Zhu et al. [47] with permission from American Chemical Society.

Ni3A

Ni1A

Ni1 Cd1 Cd2

Cd2A Cd3

Ni4 Ni2

Ni4A Ni2A

(a)

(b)

SiMe3 O N V O

N

N O

O

N

Cd

(a)

C O H O O V N N O

Cd

N O

N C – SiMe3

Cd H

(b)

Cu

Cd

O

O O V N N O

Figure 16.23 Proposed mechanism for the turning on (a)/off (b) of bimolecular activation. Source: From Zhu et al. [48] with permission from American Chemical Society.

Cui and coworkers later showed another example in which such cooperative effects can be tuned on or off by tuning the metal content [48]. The reported MOF bears VO(salen) moiety and is connected by 1D cadmium–carboxylate chains (Figure 16.23). The MOF was an active catalyst for asymmetric cyanosilylation of aldehydes with high conversions (up to 99%) and high ee’s (up to 99%). However, when some of the VO(salen) motif was replaced by Cu(salen), both the conversion and enantioselectivity decrease. This phenomenon was attributed to the absence of VO–VO cooperative activation pathway in the binary MOF. Binaphthyl-derived ligands (BINOL and BINAP) represent another family for MOF-based asymmetric catalysis. The first asymmetric catalysis with practically meaningful ee was achieved by BINOL-derived MOF catalyst [42]. Incorporation of

16.5 Asymmetric Catalysis in MOFs

O Ar

H + Zn(Et)R

OiPr O Ti O OiPr

OH *

Ar

R

Up to 91% ee

Figure 16.24 Schematic representation of asymmetric diethylzinc additions catalyzed by the MOF/Ti-BINOLate catalyst within large open channels. Source: From Ma et al. [49] with permission from Nature Publishing Group.

Ti(IV) to the MOF leads to a catalyst for asymmetric diethylzinc addition of aldehydes (Figure 16.24) [49]. BINAP-derived MOFs were found to be active catalysts for asymmetric hydrogenation, 1,4-addition, reductive cyclization, Alder-ene reaction, and Pauson–Khand reaction after Ru or Rh incorporation. These works showed the possibility that chiral MOFs are as capable for broad scope organic catalysis as molecular privileged chiral ligands [50, 51]. Other privileged ligands have also been utilized in MOF-based asymmetric catalysis. Cui and coworkers reported 2,2′ -biphenol phosphate-derived MOFs as metal-binding platforms for asymmetric Lewis acid catalysis [52]. Chiral phosphoric acid with 1,1′ -spirobiindane-7,7′ -diol (SPINOL) backbone was also introduced to MOF-based systems for heterogeneous asymmetric acid catalysis such as acetalization or Friedel–Crafts reaction. The Bronsted acidity of MOF catalysts surpasses their homogeneous counterpart [53]. While point-chiral and axial-chiral moieties are widely used in MOF-based asymmetric catalysis, not much success has been achieved in utilizing framework chirality for such application. One of the few examples was reported by Duan and coworkers. They synthesized a chiral Ag-TPHA [TPHA = tris(4-(1-(2-pyridin-2-ylhydrazono) ethyl)phenyl) amine] framework with the addition of cinchonine as the chirality inducer (Figure 16.25). Ag-TPHA catalyzes asymmetric 1,3-dipolar addition with ee’s up to 82% [54]. Lin et al. reported in another unusual case that chiral MOF catalyst can show reverse enantioselectivity against its homogeneous counterpart. This unusual phenomenon was observed in a copper–carboxylate MOF built from 3,3′ ,6,6′ -substituted BINOL phosphate-derived linkers. Quantum mechanical/molecular mechanical (QM/MM) studies indicate that chirality of the cavity dominates the transition state conformation and flips the product handedness (Figure 16.26) [55].

973

974

16 Catalysis in Metal–Organic Frameworks: Relationship Between Activities and Structures

(a)

(b)

Figure 16.25 The coordination environment (left) and framework structure (right) of Ag-TPHA. Source: From Jing et al. [54] with permission from Wiley-VCH.

16.6 Photocatalysis in MOFs Photochemistry studies the chemistry of electronic excited states. Photo-induced radicals, photochemical pericyclic reactions, and photoredox reactions are an important section of modern organic methodology; yet the most important photochemical reaction is natural photosynthesis, in which plants and algae harvest solar light and use that harvested energy to convert CO2 and water into carbohydrates. Typically, a photocatalytic system consists of several components: a photocatalyst or photosensitizer, one or more co-catalysts, and the substrates. Co-catalysts are not always necessary in the system; yet in many cases, they serve to quench the excited photosensitizer and transfer photo-induced redox equivalents (electron–hole pairs) to the substrates. This step usually plays important roles for charge separation and enables photo-induced redox chemistry. The ligand-to-metal charge transfer (LMCT) and ligand π–π* excitation of most MOFs falls in UV (ultraviolet) and visible regions. Thus, MOFs have potential application in photocatalysis. Moreover, MOFs have unique advantages as photocatalytic systems. The well-defined and highly ordered structure makes it possible to integrate multiple catalytic components, i.e. the photosensitizer and co-catalyst, into a single solid. This kind of hierarchical organization is ideal for highly efficient energy/electron transfer between those components and leads to synergistic functions of the system.

16.6.1 Photo-Degradation of Pollutants Photocatalytic degradation of organic pollutants is one of the intensely studied topics in environmental science. The degradation usually involves a photosensitization step of dioxygen into active oxidative species such as singlet oxygen, superoxide, or hydroxyl radical. Organic dyes, such as methylene blue, rhodamine B, or methyl orange, are often used as model compounds for such organic pollutants, for the dye degradation can be monitored easily by UV-vis spectroscopy.

16.6 Photocatalysis in MOFs

0.8nm

2.6 nm

Homo

CMOF O

Cu

OO

Cu

O

OH O P O O

OH O P O O

O OO O

OH O P O O

OH O P O O O

O

Cu

O

O

Cu

O

(R)-3,3ʹ-Diphenyl-1,1ʹbinaphthyl phosphate

OO

O

CMOF-1

O O

OH O P O O

O H

H

O

N

O O P O O

H N

H

H

S

N

H

H H N

S-product

R-product TS–1

TS–3

O H O

O O P O O H

S

N S O

O H

H

O O P O O N

H

H

N

H

R-product

S-product TS–2

TS–4

Home–TS–1 : 13.5 kcal/mola

CMOF–TS–1 : >30 kcal/mol

Home–TS–2 : 14.7 kcal/mol Home–TS–3: 16.0 kcal/mol

CMOF–TS–2 : 16.9 kcal/mol CMOF–TS–3 : 16.3 kcal/mol

Home–TS–4: 14.8 kcal/mol

CMOF–TS–4 : >30 kcal/mol

a transition

H

N

state energy barrier

Figure 16.26 The crystal structure and QM/MM calculations showing reverse enantioselectivities in MOF and homogeneous catalysis. Source: From Zheng et al. [55] with permission from Royal Society of Chemistry.

975

16 Catalysis in Metal–Organic Frameworks: Relationship Between Activities and Structures

Cd C N O

B C

(a)

(b)

Figure 16.27 The framework structure (left) and topological network representation (right) of [Cd(btec)0.5 (bimb)0.5 ]n . Source: From Wen et al. [56] with permission from American Chemical Society.

Not much effort has been devoted to find out mechanisms of such oxidative degradation. One of the several examples was carried out by Li and coworkers [56]. The MOF [Cd(btec)0.5 (bimb)0.5 ]n they reported were good photocatalysts for the degradation of the organic dye X3 B (Figure 16.27). They used terephthalic acid to probe the mechanism, showing the formation and existence of hydroxyl radical in the photocatalytic reaction. Photoreduction of highly toxic Cr(VI) into less toxic Cr(III) is one of the major strategies for decontamination of Cr(VI). Huang, Cao, and coworkers reported a multivariate MOF [57], containing porphyrin and imidazolium-derived ligand, for photo-driven reduction of Cr(VI). The photoactive porphyrin derivative serves as the photosensitizer, and cationic imidazolium group enhances adsorption of anionic Cr(IV) effectively, increasing the reduction activity (Figure 16.28). 1.0 Dark light 0.8

Without catalyst H 2TCPP ⊂ (I–)Meim-UiO-66 in dark Im-UiO-66 H 2TCPP ⊂ Im-UiO-66

0.6

H 2TCPP ⊂ (I–)Meim-UiO-66

Ct/C0

976

0.4 Heat

CH I 3

Mixed-ligand strategy

0.2

PSM strategy 0.0 –60 H2TCPP⊂ Im-UiO-66

(a)

H TCPP⊂ (I–)Meim-UiO-66 2

0

10 Time (min)

20

30

(b)

Figure 16.28 The synthesis (left) and bichromate removal activity (right) of H2TCPP⊂(I-)Meim-UiO-66. Source: From Wang et al. [57] with permission from Elsevier Inc.

16.6 Photocatalysis in MOFs

16.6.2 Organic Photocatalysis Besides simple dye degradation, chemists also show great interest in more selective photo-assisted organic transformations in recent years. Many of such transformations belong to the reaction type of oxidation or oxidative coupling (Scheme 16.1). aza-Henry reaction CH3NO2 Catalyst, hv

N

N

R

R

NO2

Oxidative amine coupling NH2

R

Air Catalyst, hv

R

N

R

Oxidation of sulfide R1

S

R2

Scheme 16.1

Air Catalyst, hv

R1

O S

R2

Photocatalytic oxidation and oxidative coupling of organic compounds.

Lin and coworkers first introduced organometallic dyes such as Ru(bpy)3 2+ or Ir(ppy)2 (bpy)+ into MOF skeletons to make photoactive MOFs and investigated their photocatalytic activity with those model reactions [58]. These multivariate UiO-type MOFs exhibit good activity and stability upon several reuse cycles. However, size selectivity was not observed in those photocatalysis. Jiang and coworkers reported photocatalytic oxidative coupling coupled with hydrogen evolution with Pt/PCN-777 as the photocatalyst (Figure 16.29). Instead of using singlet oxygen as the oxidant, they deposited Pt NPs in the MOF channels, which catalyze hydrogen evolution [59]. In the proposed mechanism, an electron–hole pair was generated upon irradiation, then the electron transferred to Pt NPs to enable hydrogen evolution and the hole was used to oxidize the amine substrate. It represents one of the rare examples that selective oxidation and reduction are simultaneously promoted by light. MOF photocatalysis can also be achieved through two-photon excitation. Duan and coworkers reported aza-Henry and amine oxidative coupling reactions catalyzed by a photoactive MOF ZJU-56-0.2, which is built from Zn2 (RCOO)4 paddle-wheel nodes and two carboxylate ligands, 2,5-bis(3,5-dicarboxyphenyl)pyridine and 2,5-bis(3,5-dicarboxyphenyl)methylpyridium (Figure 16.30) [60]. ZJU-56-0.2 exhibits two-photon excitation at 672 nm and emits at 470 nm. It also shows photocatalytic activities upon 660 nm excitation. This two-photon excitation property was attributed to the high spatial density of chromophores in MOF structure. Photocatalytic oxidation of other organic compounds was also explored. Wu and coworkers reported a tin porphyrin-derived MOF that enables photocatalytic oxidation of 1,5-dihydroxynaphthalene [61]. Huang and coworkers used

977

978

16 Catalysis in Metal–Organic Frameworks: Relationship Between Activities and Structures

Oxidation end

NH

H2

2

e–

e–

h+

Charge separation

N

H2O Reduction end

Figure 16.29 Illustration of simultaneous proton reduction and benzylamine oxidation over Pt/PCN-777 by photocatalysis. Source: From Liu et al. [59] with permission from Wiley-VCH.

O2 R

ZJU–56–0.2*

NH2 1O

N hv = 660 nm EnT

R

N

R

2

R

PET hv = 660 nm N O2N

HOOC

R

COOH N COOH

HOOC

Zn2+

H4L2 HOOC

COOH +

HOOC

N OH– H4L1 OH

COOH

Figure 16.30 Representation of the two-photon responsive ZJU-56-0.2 for the photocatalytic oxidative aza-Henry and amine coupling reactions. Source: From Li et al. [60] with permission from American Chemical Society.

16.6 Photocatalysis in MOFs

bismuth–carboxylate MOFs for photocatalytic oxidation of benzyl alcohols to benzaldehydes. Scavenger experiments indicated that while singlet oxygen (1 O2 ) is the major photogenerated species, both 1 O2 and H2 O2 are the major oxidants for the reaction [62]. Combining photoactive MOFs with asymmetric catalysts can lead to enantioselective photocatalysis. Duan and coworkers reported a pair of MOF enantiomers, Zn-PYI1 and Zn-PYI2, as asymmetric photocatalysts for α-alkylation of aliphatic aldehydes [63]. The MOFs were constructed from a photoactive ligand 4,4′ ,4′′ -nitrilotrisbenzoic acid, Zn2+ metal ion, and a secondary chiral amine (Lor D-pyrrolidine-2-ylimidazole ([PYI]). The secondary amine, in its Boc-protected form during MOF synthesis, serves as a template to build the chiral framework, which is confirmed via single-crystal crystallography as well as circular dichroism (CD) spectroscopy. Remarkably high enantioselectivities (ee up to 92%) were achieved. MOFs with opposite chirality gave opposite enantioselectivities, showing that MOF is the true photocatalyst. The reaction mechanism was proposed as similar to the same reaction triggered by organometallic dyes, yet new phenomena were also observed. When using achiral MOFs bearing the same triphenylamine photoactive moiety, Ho-TCA (tris(4-carboxylatephenyl)amine) or MOF-150, as the photosensitizer and the same chiral amine as co-catalyst, much lower ee values were obtained. It indicates that the framework chirality also plays an important role in enantioselectivity. Yields for more bulky aldehydes 1d (Figure 16.31) were much

O

O OH

+ Br

H

+

N

N

N

O N N

N

Boc

Boc

O

N

N

O

O

Zn-BCIP2

73(–81)

61(–78)

85(–89)

90(+21)

93(+20) 95(–20)

Ho–TCA/D–PYIb

85(–21)

90(–20)

MOF–150/L–PYIb

67(+21)

78(+24)

80(+20)

MOF–150/D–PYIb

62(–22)

73(–22)

80(–21)

1a

Molecular size 10.4 Å O H

1b

12.9 Å 4.6 Å

O

Deprotection

H 13.5 Å

1c

(a)

O

O

N

N

H 17.4 Å

1d

N N

HN

O

O

O O

O H

O

O

13.8 Å

NH

O

Zn-PYI1

O O N

N

5.2 Å

O

O O

O

Et

84(+92)

86(+20)

O

O

Me

5

65(+86)

Ho–TCA/L–PYIb

Substrate O

74(+92)

Zn–PYI2a

O O N

N

N

Zn-BCIP1

Boc Self-assembly

O H

H

6.1 Å

O O

D-BCIP

N

N

Boc

O

H ph

Zn–PYI1a

N

N

Catalyst

CO Et 2 R

O

O

Entry

CO2Et

H

0.5 mol% catalyst

CO Et 2

Zn

O OH

HO O L-BCIP

R

2+

O

Fliorescent lamp

CO2Et

Zn-PYI2

(b)

Figure 16.31 Schematic representation for the synthesis of Zn-PYI1 and Zn-PYI2 (a) and the photoredox catalysis (b). Source: From Wu et al. [63] with permission from American Chemical Society.

979

980

16 Catalysis in Metal–Organic Frameworks: Relationship Between Activities and Structures c

O

HO

b a

N OH

O OH

O

Zn2+

(a) Photoredox catalyst N N

HN

Chiral organocatalyst

(c)

(e)

N2 N8

O16 O8 Zn1 N3 N5

O6 O2 O4

A°

Zn2 O12

N1 Zn C N O

8.4

N6

6.8

6.8

A°

A°

14.1 A°

O6A Zn2A

(b)

(d)

Intra-framework distances

(f)

Intra-framework distances

Figure 16.32 The components (a) and coordination environments (b) of InP-1. The isolated (c) and two-fold interpenetrated (e) framework structures of InP-1. The intra-(d) and inter-(f) framework distances of the catalytic sites. Source: From Xia et al. [64]. Licensed Under CC BY 4.0.

lower than the other substrates. This size selectivity proves that catalysis takes place in the MOF channels other than the external surface. The same group later reported another pair of MOF enantiomers, InP-1 and InP-2, bearing the same components but assembling in a different twofold interpenetrate structure (Figure 16.32) [64]. InP-1 and InP-2 catalyze β-arylation of aldehydes under irradiation and show moderate enantioselectivity (ee 30–55%). Interestingly, when Zn-PYI MOFs were used as the photocatalyst, lower conversion and negligible enantioselectivity were observed. Therefore, the chiral framework structure is more crucial for the enantioselectivity than the “local” chirality at the amine center. This phenomenon is unique in MOF catalysis and has not been reported elsewhere.

16.6.3 Photocatalytic Water Splitting and Artificial Photosynthesis Photo-driven water splitting and artificial photosynthesis are among the most promising scenarios for solar energy conversion. These energy uphill reactions are nearly ideal for storing solar energy in the form of chemical fuels. MOFs have also been intensively studied as potential materials for such conversion. Simple MOFs like UiO-66 have been tested as photocatalyst for hydrogen evolution. In a methanol–water mix solution, UiO-66 showed modest catalytic activity, which was significantly improved with platinum NPs added as a co-catalyst [65]. Enhanced photocatalytic activities can be achieved by tuning the MOF absorption band. MIL-125, an MOF built from Ti8 O8 (OH)4 (RCOO)12 cluster and terephthalic acid, is a white solid with no visible absorption. Introducing amino group to the terephthalic acid moiety leads to NH2 -MIL-125, which exhibits a visible absorption

16.6 Photocatalysis in MOFs CO2

F(R)

HCOO–

e–

(a)

Ti

Ti

(a)

O

(b)

Ti

O

H2N

(b)

3+

Ti O

Visible light

*

O–

H2N e–

200 300 400 500 600 700 800 𝜆 / (nm) (b) (a)

NH2–MIL–125(Ti)

.–

TEOA

NH2–MIL–125(Ti) TEOA

Figure 16.33 (A) Absorption spectra of MIL-125(Ti) (line a) and NH2 -MIL-125(Ti) (line b). (B) Proposed mechanism for the photocatalytic CO2 reduction. Source: Reproduced from Fu et al. [66] with permission from Wiley-VCH.

band to about 550 nm. NH2 -MIL-125 thus showed enhanced photocatalytic activity for the reduction of CO2 to formate. Triethanolamine was used as the sacrificial electron donor in this reaction (Figure 16.33) [66]. To make sufficient use of visible light energy, well-defined chromophores such as porphyrin were used to construct photoactive MOFs. Huang and coworkers reported an MOF constructed from aluminum and copper–porphyrinic tetrakiscarboxylic acid. They found that the MOF was able to convert CO2 into methanol under visible light irradiation. The presence of Cu(II) in the center of porphyrin ring promotes CO2 chemical adsorption and activation, thus enhancing methanol production efficiency [67]. MOF-constrained metalloporphyrin motif can also be synthesized via post-synthetic metallation of free base porphyrin-containing MOFs. Using this strategy, Jiang and coworkers prepared a Pt–porphyrin MOF and used it for photocatalytic hydrogen production [68]. From a similar approach, Wang and coworkers co-deposited both Ir and Pt atoms in free base porphyrin-derived MOF PCN-222. With iridium–porphyrin as the photosensitizer and platinum–porphyrin as the hydrogen evolution catalyst, the resulting material showed a photocatalytic hydrogen production rate more than 200 μmol/g/h [69]. The stability of MOFs is important for their catalytic applications. Reported by Zhang, Lin, and coworkers, Zr(IV)-pyrogallic porphyrin MOFs, built from polyphenolate and 1D Zr-oxo chain, are very stable in a wide pH range from 0.1 M HCl to saturated NaOH (Figure 16.34). The cobalt–porphyrin derivative showed good catalytic activity for photo-driven CO2 reduction with CO as the main product [70]. In most cases, two components, a photosensitizer and a co-catalyst, need to perform synergistically to achieve good activity for hydrogen production or CO2 reduction. MOF provides an ideal platform to integrate such multiple components in a single solid material. Lin and coworkers reported in situ deposition of platinum NPs in MOFs containing Ir(ppy)2 (bpy)+ chromophores (Figure 16.35). The close proximity of the photosensitizer (molecular iridium complex) and the co-catalyst (platinum NPs) facilitates electron transfer between the two and leads to highly active catalyst for hydrogen evolution [71].

981

982

16 Catalysis in Metal–Organic Frameworks: Relationship Between Activities and Structures

THPP

THBPP

(b) (a)

(d)

(c)

Figure 16.34 (a) The coordination structure of Zr-pyrogallate chain. (b) Structure of ZrPP-1. (c) Structure of ZrPP-2. (d) Topological network representation of ZrPP-n (n = 1,2) structures. Source: From Chen et al. [70] with permission from Wiley-VCH.

+

CO2H

CO2H

N

+

N Ir

N

N

CO2H

CO2H

H2bpdc

H2L3

ZrCl4 HOAc

K2PtCl4

DMF 100 °C

TEA hv

3 Zr6(O)4(OH)4(bpdc)4.94(L1)0.06

Pt@3

CO2H

+ N

N

ZrCl4 TFA

K2PtCl4

Ir N N

DMF 100 °C

CO2H

H2L4

Zr6(O)4(OH)4(carboxylate)12

TEA hv

4 Zr6(O)4(OH)4(L4)6·64DMF

bpdc

L1

Pt@4

L2

Figure 16.35 Synthesis of phosphorescent zirconium-carboxylate MOFs and subsequent loading of Pt NPs inside MOFs 3 and 4. Source: From Wang et al. [71] with permission from American Chemical Society.

The same group later used a one-pot approach to synthesize a POM@MOF composite with Ru(bpy)3 2+ -derived bridging ligand and polyoxometalate P2 W18 O62 6− in the MOF cavity (Figure 16.36). The POM@MOF assembly allows multi-electron injection from photosensitizers to the encapsulated redox-active POMs, which is almost impossible in homogeneous systems. Therefore, efficient photo-driven hydrogen evolution can be achieved [72].

16.7 Multi-Component Catalysis Using MOFs

t

2H+

UV

ligh

Visible light

e– H2

+

Integration

2H+

H2

Figure 16.36 Schematic representation of the one-pot synthesis of POM@MOF composite and the photocatalytic hydrogen evolution. Source: From Zhang et al. [72] with permission from American Chemical Society.

Li and coworkers incorporated Ru(bpy)(CO)2 Cl2 and Ru(bpy)3 2+ functional components in a single MOF solid by post-synthetic metallation of MOF-253, an MOF containing open bipyridine coordination sites. The dual-functionalized MOF showed high efficiency for photo-driven CO2 reduction with triethanolamine as the sacrificial reducing agent, giving formate as the major product and CO as minor product. Interestingly, if only Ru(bpy)(CO)2 Cl2 moiety was incorporated in the MOF, the production of formate was greatly suppressed and CO became the major product [73]. Huang and coworkers reported photocatalytic overall water splitting by an aluminum-based MOF. Ni(II) cations were introduced to Al-ATA (2-aminoterephthalate) MOF built from aluminum(III) and 2-aminoterephthalate [74]. With a xenon lamp light source, the Ni-incorporated MOF catalyzed water splitting, while H2 and O2 were produced in the stoichiometric ratio of 2 : 1. The authors believed that Ni(II) cation here serves as the hydrogen evolution center, yet the detailed mechanism remains unclear.

16.7 Multi-Component Catalysis Using MOFs MOFs are composed of metal nodes and functional organic linkers, which can be employed as active sites for catalysis. In addition, many active guest species including MNPs, metal complexes, and polyoxometalates can be introduced in their pores for catalytic performances. Thus, the combination of these multiple active sites in MOFs can be used for synergistic catalysis and tandem reactions [75]. Synergistic catalysis involves that the reactants are simultaneously activated by two or more distinct active sites in MOFs [76]. In synergistic catalysis, the cooperative action of the multiple active sites results in a decrease in the LUMO (lowest

983

984

16 Catalysis in Metal–Organic Frameworks: Relationship Between Activities and Structures

unoccupied molecular orbital) energy level of one reactant and an increase in the HOMO (highest occupied molecular orbital) energy level of another reactant. Therefore, the activity and selectivity of existing reactions could be improved, and some new chemical transformations may be realized. One-pot tandem reactions consist of two or more successive reactions that usually catalyzed by two catalysts or a bifunctional catalyst with two or more types of active sites [77]. In tandem reactions, the reactants are catalyzed by one active site (catalyst) to obtain an intermediate, which is further activated by another active site (catalyst) to produce the final product. Therefore, it is no need to separate and purify the intermediates, which results in reducing the reaction steps, energy consumption, and waste [78]. The advantages make synergistic catalysis and tandem reactions sustainable green processes that illustrate the concepts of efficiency and atom economy. According to the type of active center, these multifunctional MOFs can be categorized as follows: (i) CUMs and functional organic linkers in the MOF backbone and (ii) active guest sites and active sites in the MOF backbone.

16.7.1 16.7.1.1

CUMs and Functional Organic Linkers in the MOF Backbone Mixed Metal Centers as Active Sites for Synergistic Catalysis

Mixed metals in metal nodes would occupy equivalent crystallographic framework sites or affect their electronic properties each other, which thus exhibit enhanced catalytic activities and/or produce different products. A series of porphyrin-based MOFs composed of (M3 O)2 (TCPPM)3 and M3 O (M = Mg, Mn, Co, Ni, Fe) trigonal SBUs and mixed-metal SBUs including Mnx Fe3−x O, Nix Fe3−x O, Cox Ni3−x O, Mnx Co3−x O, Mnx Mg3−x O, and Mnx Ni3−x O were designed and prepared (Figure 16.37) [79]. The metal spatial arrangement were found to exist in the form of well-mixed metals in their SBUs of these MOFs rather than the metal nodes having one kind of metal but different from one SBU to another. Thus, the MOFs with mixed-metal SBUs showed higher activity than the correspond MOFs with single metal node in the photo-oxidation of 1,5-dihydroxynaphthalene. 16.7.1.2

Mixed Linkers as Bifunctional Active Sites for Tandem Reactions

Incorporating different functional groups into one linker or using mixed linkers containing different active sites to assemble an MOF could endow them a multifunctional platform for synergistic catalysis and/or tandem reactions. For example, bifunctional Brønsted acid–base site-isolated catalyst, MIL-101(Cr)–NH2 –SO3 H, was synthesized using –NO2 and –SO3 Na functionalized terephthalate linkers followed by reducing with SnCl2 (Figure 16.38) [80]. The resulting MIL-101(Cr)–NH2 –SO3 H contain isolated Brønsted acid (–SO3 H)–base (–NH2 ) sites, which showed highly efficient catalytic performance in one-pot tandem deacetalization–nitroaldol reaction converting benzaldehyde dimethyl acetal to trans-1-nitro-2-phenylethylene (Figure 16.39). However, the homogeneous catalysts that contain acid and base sites could not efficiently promote

16.7 Multi-Component Catalysis Using MOFs

Variation of metals in porphyrin linker OH

O

O

OH OH

O N H N

H N

N

OH

O

N

N

N Mg N

N Co N N

N O

HO O

O

HO

HO

O

O

HO

HO

TCPP-H2 OH

O

TCPP-Mg

O

TCPP-Co OH

O

HO

OH

O

OH OH

O

(M3O)2(TCPP-M)3 stp topology

O

OH OH

O

O OH

O

N

N

N

N Ni N

N Cu N

N Zn N

N

N

HO

O O

N

HO

HO

TCPP-Ni

O O

HO

TCPP-Cu

HO

O O

HO

TCPP-Zn

Variation of metals in second building units (SBU)

Mn3O(CO2)6

Mn

Fe3O(CO2)6

Co3O(CO2)6

Fe

Mg3O(CO2)6

Co

Mn Fe (MnxFe3−xO)2(TCPP-Ni)3

Ni3O(CO2)6

Mg

Ni

Fe Ni (NixFe3−xO)2(TCPP-Co)3

Mn Co (MnxCo3−xO)2(TCPP-Ni)3

Co Ni (MixCo3−xO)2(TCPP-Ni)3

Mn Mg (MnxMg3−xO)2(TCPP-Ni)3 Mn Ni (MnxNi3−xO)2(TCPP-Ni)3

Figure 16.37 Single-component MOF series, (M3 O)2 (TCPP-M)3 , constructed by five different SBUs and six different porphyrin linkers, and corresponding multivariate (MTV-MOFs) with mixed-metal SBUs. Source: From Liu et al. [79] with permission from American Chemical Society.

985

986

16 Catalysis in Metal–Organic Frameworks: Relationship Between Activities and Structures NO2

+ NO2

SO3Na COOH

NH2

COOH

COOH

CrO3

SnCl2 / EtOH

H2O / HCI

343 K, 6h

COOH

SO3H MIL-101–NO2-SO3H

SO3H MIL-101–NO2–SO3H

Figure 16.38 Preparation of the site-isolated, bifunctional Brønsted acid–base catalyst MIL-101(Cr)–NH2 –SO3 H. Source: From Lee et al. [80] with permission from Royal Society of Chemistry. OCH3 OCH3

O Acid H2O

H

Base CH3NO2

NO2

Figure 16.39 One-pot tandem deacetalization–nitroaldol reaction of benzaldehyde dimethyl acetal to trans-1-nitro-2-phenylethylene. Source: From Lee et al. [80] with permission from Royal Society of Chemistry.

this reaction, which highlights the isolating bifunction active sites in this unique MOF for tandem reactions. 16.7.1.3 Metal Nodes and Linkers as Bifunctional Active Sites for Synergistic Catalysis

The multiple active sites in metal nodes and linkers make MOFs very attractive to be designed as heterogeneous catalysts for synergistic catalysis [75]. Currently, carbon dioxide capture and utilization (CCU) technologies, whereby CO2 is captured from a specific source and utilized as a green C1 chemical feedstock to produce valuable chemicals, are seen as sustainable chemical approaches to reduce CO2 from the atmosphere [81]. Cyclic organic carbonates can be directly synthesized from gas phase CO2 and various epoxides in the presence of Lewis acid catalyst and tetraalkyl ammonium halide as a cocatalyst (Figure 16.40) [82]. Up to now, various kinds of MOFs containing active metal sites could catalyze the cycloaddition reaction, but a Lewis basic co-catalyst is usually required to make these systems to be optimally effective (faster, higher selectivity) under mild conditions. MOF can be designed to integrate active acid sites and Lewis basic linkers acting as the permanent co-catalyst. For example, Cao, Huang, and coworkers designed a bifunctional imidazolium functionalized Zr-MOF (Figure 16.41), called (I− )Meim-UiO-66 for the synergistic catalysis of this reaction [82]. First, a 2-(imidazol-1-yl)-terephthalic acid (Im-H2 BDC) linker and in situ formed zirconium metal oxide clusters (Zr6 ) were coordinated to obtain the imidazole-functionalized UiO-66 derivative (Im-UiO-66) with fcu topology, in which each Zr6 node should be connected with other 12 Zr6 clusters through 12 Im-BDC2− linkers in principle. However, rich Brønsted acidic sites, such as Zr-OH and Zr-H2 O, were formed due to bulky imidazole group on terephthalic acid (BDC) backbone, which led to the decreased number

16.7 Multi-Component Catalysis Using MOFs O O

+

Catalyst

CO2

O

O

1 atm

R

R

Epoxides

Cyclic carbonates

(a)

(1) ZrCl4, DMF 120° C, 48 h

(2) CH3I, CH3CN 80° C, 48 h

I N Zr O C H

Im-H2-BDC

(I–)Meim-UiO-66

Im-UiO-66

(b)

Figure 16.40 (a) Catalytic cycloaddition of CO2 with epoxides to form cyclic carbonates. (b) Syntheses of Im-UiO-66 and (I− )Meim-UiO-66 via reticular chemistry and PSM method, respectively. Green polyhedra, Zr6 clusters; yellow ball, micropores. Source: From Liang et al. [82] with permission from Royal Society of Chemistry.

I N I

N

R

N O

N

R

H

H O H

O H O Zr6

Zr6

R

I

O

O

IV

II R

O I

I

N N

N

O O

N

O H O H Zr6

O C O

III

O H O H Zr6

Figure 16.41 Plausible mechanism of (I− )Meim-UiO-66 catalyzed cycloaddition reaction of CO2 with epoxides. Source: From Liang et al. [82] with permission from Royal Society of Chemistry.

987

988

16 Catalysis in Metal–Organic Frameworks: Relationship Between Activities and Structures

of linkers on each Zr node. The acid sites on defect Zr6 nodes was indicated by NH3 -TPD (temperature-programmed desorption) analysis. Second, the initially formed Im-UiO-66 could be converted to (I− )Meim-UiO-66 through a PSM method. The catalyst with Brønsted acidic sites (Zr-OH and Zr-H2 O) and I− nucleophiles could synergistically catalyze the cycloaddition of various epoxies and CO2 under 1 atm and mild conditions. The proposed acid/base synergistic catalysis process could be described as shown in Figure 16.41. First, the epoxide ring is adsorbed onto (I− )Meim-UiO-66 and stabilized by non-covalent interactions (I). Then the ring-opening step occurs due to the synergistic activation effect of the Brønsted acid (Zr-OH/Zr-OH2 ) and the nucleophilic I− (II). Immediately, CO2 inserts into the anionic intermediate species to form an acyclic ester (III), which is transformed to a carbonate by intramolecular cyclization, releasing the original catalyst for the next catalytic cycle (IV).

16.7.2

Active Guest sites and Active Sites in the MOF Backbone

The well-defined tunable pores of MOFs could make them promising candidates for loading active guest species, such as MNPs and polyoxometalate (POM). Therefore, the combination of active guest species and the sites (open metal sites and/or functional groups) of the backbone in the confined space could synergistically promote catalysis and catalyze tandem reactions. 16.7.2.1 MNP@MOF Composites for Synergistic Catalysis and Tandem Reactions

The functional group in MOFs not only can stabilize MNPs but also could assist MNPs to co-activate substrates and synergistically enhance reaction rates. PtPd alloy NPs supported on an acid–base bifunctional micro-mesoporous MOF, MM-MIL-53, could synergistically promote oxidant-free alcohol dehydrogenation reaction (Figure 16.42) [83]. As well known, the conversion of alcohols to carbonyl compounds is generally required using dangerous and costly oxygen as the oxidant [84]. Thus, the O2 -free strategy is an atom-efficient, safe, and promising method. In the metal node of MM-MIL-53, Al(III) centers can be acted as Lewis acid site, while –OH in the interconnected chains of corner-sharing AlO4 (OH)2 octahedra can be served as base active sites. The isolated acid–base active sites can activate the hydroxyl of alcohol, while the MOF-supported PtPd alloy could activate the C—H bond of alcohol (Figure 16.42). Thus, no need of O2 , the PtPd/MM-MIL-53(Al) composite can smoothly catalyze alcohol dehydrogenation reaction, although high reaction temperature (140–170 ∘ C) and long reaction time (48–72 hours) are needed. Several secondary alcohols, such as 2-octanol, can be converted to their corresponding ketones with moderate yields in an N2 atmosphere. In comparison, the PtPd supported on the SBA-15 that only contain acid sites cannot promote this reaction under the same conditions. So, the proposed mechanism is that the proton of alcohol was abstracted by a basic site of MM-MIL-53 to yield an alkoxide group on the Al(III) Lewis acid site. At the same time, the C—H bond was dissociated by a PtPd NP to form PtPd-H and obtain the final ketone product with the release of H2 gas.

16.7 Multi-Component Catalysis Using MOFs

Mesopore Micropore

A O C

(a)

H R1

O

R1

O

O H

R2 (b)

AI

AI

OH

R1

R2

R2

Figure 16.42 (a) The walls of hierarchically micro- and mesoporous MM-MIL-53(Al) were formed by microporous MIL-53(Al) crystals. (b) The proposed mechanism for the oxygen-free alcohol dehydrogenation reaction catalyzed by the PtPd/MM-MIL-53(Al) composite. Source: From Huang et al. [83] with permission from Elsevier Inc.

The close proximity of the active sites in the backbone and the encapsulated MNPs could make the composite a promising reactor for improving tandem reactions. A uniform core–shell Pd@IRMOF-3 composite was fabricated [85], in which Pd NPs (35 nm) acted as core that was coated with the amine-functionalized IRMOF-3. The Pd NPs were first prepared and then added to a mixture of N,N-dimethylformamide (DMF)–ethanol solution containing Zn(NO3 )2 , 2-aminoterephthalic acid (NH2 -H2 bdc) and polyvinylpyrrolidone (PVP), and a solvothermal method was subsequently applied to assemble IRMOF-3 around Pd species (Figure 16.43). The core–shell Pd@IRMOF-3 composite was then used in a one-pot tandem Knoevenagel condensation–hydrogenation reaction. The amino groups of the IRMOF-3 shell can catalyze the Knoevenagel condensation of 4-nitrobenzaldehyde (A) with malononitrile to form 2-(4-nitrobenzylidene)malononitrile (B). Then, the –NO2 group of the intermediate (B) was reduced to amine by the near Pd NP cores with assistant of H2 (Figure 16.43). 16.7.2.2 POM@MOF Composites for Synergistic Catalysis and Tandem Reactions

The confined POMs in the pores of MOFs can act as multiple active sites, including protons, oxygen atoms, and metal sites. Therefore, the combination of POMs and the active sites in the backbones of MOFs could synergistically enhance catalysis and catalyze tandem reactions. For example, the chiral POM-based MOFs named as Ni-PYI1 and Ni-PYI2, respectively, were assembled from the Keggin-type [BW12 O40 ]5− and the chiral L- and D-pyrrolidin-2-ylimidazole (PYI) ligands (Figure 16.44) [86]. With the help of H2 O2 , both Ni-PYI1 and Ni-PYI2

989

990

16 Catalysis in Metal–Organic Frameworks: Relationship Between Activities and Structures

NH2

HO +

NH2

O NH2

Zn(NO3)2 +

NH2 NH2

O

OH

NH2 NH2

NH2

Pd NPs

NH2 NH 2

NH2

NH2

NH2

NH2-H2BDC

: Zn-O cluster

Pd@IRMOF-3

(a) CN

O

CN

IRMOF-3 NC

O2 N

(b)

CN

CN

Pd NPs H2

O2N

A

CN

H2N

B

C

Figure 16.43 (a) Synthesis of the core–shell Pd@IRMOF-3 and (b) catalyze one-pot tandem reaction involving Knoevenagel condensation and subsequent selective hydrogenation. Source: From Zhao et al. [85] with permission from American Chemical Society.

OH NiCl2

N

N

H

OH C H2

+ N

N

R

R

N Boc

Asymmetric dihydroxylation

130 °C, 72 h

Guest exchange

Figure 16.44 Assembly of Ni-PYI1 and catalysis of the asymmetric olefin dihydroxylation reaction. Source: From Han et al. [86] with permission from American Chemical Society.

showed moderate activity and high ee values (>95%) in the asymmetric aryl olefin dihydroxylation (Figure 16.44). This is because the [BW12 O40 ]5− anion was responsible for the dihydroxylation, whereas the chiral PYI part and the chiral channel environment induce chiral transformation. Furthermore, the similar homochiral POM-based ZnW-PYI1 were also prepared, which can be applied in the

References

Polyoxometalate oxidant catalyst N

N

Chiral organocatalyst N N HN

One

TBHP

NH2

Lewis acids catalyst

O

HN

Zn2+ Bridge ligand for CO2 adsorption

Pot

CO2 N N H2N

O

O O

Figure 16.45 Synthesis of the homochiral ZnW-PYI1 and catalysis of the tandem reaction for asymmetric cyclic carbonate synthesis from olefins and carbon dioxide. Source: From Han et al. [87]. Licensed Under CC BY 4.0.

one-pot tandem asymmetric catalysis for the conversion of olefins into value-added enantiomerically pure cyclic carbonates under CO2 atmosphere (Figure 16.45) [87]. The amino-functionalized homochiral ZnW-PYI1 was synthesized from the assembly of ZnW12 O40 6− Keggin anion and the chiral PYI organocatalyst (Figure 16.45). The asymmetric epoxidation of the styrene was first catalyzed by the synergistic effect of the ZnW12 O40 6− Keggin anion and PYI of MOF in the presence of t-butylhydroperoxide (TBHP). Then, the epoxide intermediate and CO2 were converted to enantiomerical cyclic carbonates by the active Zn Lewis acid sites in ZnW-PYI1 in the presence of enantiomerically pure cyclic carbonates. The high ee values may be ascribed to the hydrogen-bonding interactions between the ZnW12 O40 6− and the chiral PYI organocatalyst and the π–π interactions between the PYI imidazole ring and styrene oxide benzene ring. The chiral directing agent PYI moiety and the electrophilic ZnW12 O40 6− oxidant also controlled the orientation of the substrates, leading to stereoselective catalysis. Although this one-pot asymmetric catalytic reaction needs four days, the use of tandem epoxidation–cycloaddition strategy by a chiral MOF to produce useful chiral compounds from CO2 is an environmentally friendly synthesis method.

References 1 Hoskins, B.F. and Robson, R. (1989). J. Am. Chem. Soc. 111: 5962. 2 Hoskins, B.F. and Robson, R. (1990). J. Am. Chem. Soc. 112: 1546. 3 Fujita, M., Kwon, Y.J., Washizu, S., and Ogura, K. (1994). J. Am. Chem. Soc. 116: 1151. 4 Lee, J., Farha, O.K., Roberts, J. et al. (2009). Chem. Soc. Rev. 38: 1450. 5 Liu, J., Chen, L., Cui, H. et al. (2014). Chem. Soc. Rev. 43: 6011. 6 Férey, G., Mellot-Draznieks, C., Serre, C. et al. (2005). Science 309: 2040. 7 Henschel, A., Gedrich, K., Kraehnert, R., and Kaskel, S. (2008). Chem. Commun.: 4192.

991

992

16 Catalysis in Metal–Organic Frameworks: Relationship Between Activities and Structures

8 Kim, J., Bhattacharjee, S., Jeong, K.-E. et al. (2009). Chem. Commun.: 3904. 9 Huang, X.-C., Lin, Y.-Y., Zhang, J.-P., and Chen, X.-M. (2006). Angew. Chem. Int. Ed. 45: 1557. 10 Banerjee, R., Phan, A., Wang, B. et al. (2008). Science 319: 939. 11 Wang, F., Liu, Z.-S., Yang, H. et al. (2011). Angew. Chem. Int. Ed. 50: 450. 12 Zhang, G.-L., Zhou, L.-P., Yuan, D.-Q., and Sun, Q.-F. (2015). Angew. Chem. Int. Ed. 54: 9844. 13 Horcajada, P., Surble, S., Serre, C. et al. (2007). Chem. Commun.: 2820. 14 Dhakshinamoorthy, A., Alvaro, M., and Garcia, H. (2010). Chem. Eur. J. 16: 8530. 15 Luan, Y., Qi, Y., Gao, H. et al. (2015). J. Mater. Chem. A 3: 17320. 16 Kim, J., Kim, S.-N., Jang, H.-G. et al. (2013). Appl. Catal., A 453: 175. 17 Zhou, Y.-X., Chen, Y.-Z., Hu, Y. et al. (2014). Chem. Eur. J. 20: 14976. 18 Yang, X.-L. and Wu, C.-D. (2014). Inorg. Chem. 53: 4797. 19 Feng, D.W., Gu, Z.Y., Li, J.R. et al. (2012). Angew. Chem. Int. Ed. 51: 10307. 20 Feng, D., Chung, W.-C., Wei, Z. et al. (2013). J. Am. Chem. Soc. 135: 17105. 21 Huang, N., Yuan, S., Drake, H. et al. (2017). J. Am. Chem. Soc. 139: 18590. 22 Fei, H., Cohen, S.M., and Robust, A. (2014). Chem. Commun. 50: 4810. 23 Manna, K., Zhang, T., and Lin, W. (2014). J. Am. Chem. Soc. 136: 6566. 24 Zhang, T., Manna, K., and Lin, W. (2016). J. Am. Chem. Soc. 138: 3241. 25 Li, B., Ju, Z., Zhou, M. et al. (2019). Angew. Chem. Int. Ed. 58: 7687. 26 Roucoux, A., Schulz, J., and Patin, H. (2002). Chem. Rev. 102: 3757. 27 Huang, Y., Lin, Z., and Cao, R. (2011). Chem. Eur. J. 17: 12706. 28 Huang, Y., Ma, T., Huang, P. et al. (2013). ChemCatChem 5: 1877. 29 Huang, Y.-B., Shen, M., Wang, X. et al. (2016). J. Catal. 333: 1–7. 30 Huang, Y., Liu, S., Lin, Z. et al. (2012). J. Catal. 292: 111. 31 Schroder, F., Esken, D., Cokoja, M. et al. (2008). J. Am. Chem. Soc. 130: 6119. 32 Aijaz, A., Karkamkar, A., Choi, Y.J. et al. (2012). J. Am. Chem. Soc. 134: 13926. 33 Wang, S.S. and Yang, G.Y. (2015). Chem. Rev. 115: 4893–4962. 34 Ribeiro, S., Barbosa, A.D.S., Gomes, A.C. et al. (2013). Fuel Process. Technol. 116: 350. 35 Hu, X., Lu, Y., Dai, F. et al. (2013). Microporous Mesoporous Mater. 170: 36. 36 Wang, X.-S., Li, L., Liang, J. et al. (2017). ChemCatChem 9: 971. 37 Wang, X.-S., Lin, Z.J., Huang, Y.-B., and Cao, R. (2014). Dalton Trans. 43: 11950. 38 Ahmad, N., Younus, H.A., Chughtai, A.H., and Verpoort, F. (2015). Chem. Soc. Rev. 44: 9. 39 Qiu, X., Zhong, W., and Li, Y.W. (2016). J. Am. Chem. Soc. 138: 1138. 40 Feng, D., Liu, T.-F., Su, J. et al. (2015). Nat. Commun. 6: 5979. 41 Seo, J.S., Whang, D., Lee, H. et al. (2000). Nature 404: 982. 42 Wu, C.D., Hu, A., Zhang, L., and Lin, W. (2005). J. Am. Chem. Soc. 127: 8940. 43 Yoon, T.P. and Jacobsen, E.N. (2003). Science 299: 1691. 44 Cho, S.-H., Ma, B., Nguyen, S.T. et al. (2006). Chem. Commun.: 2563. 45 Song, F., Wang, C., Falkowski, J.M. et al. (2010). J. Am. Chem.Soc. 132: 15390. 46 Xia, Q., Li, Z., Tan, C. et al. (2017). J. Am. Chem.Soc. 139: 8259. 47 Zhu, C., Yuan, G., Chen, X. et al. (2012). J. Am. Chem. Soc. 134: 8058. 48 Zhu, C., Xia, Q., Chen, X. et al. (2016). ACS Catal. 6: 7590.

References

49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87

Ma, L., Falkowski, J.M., Abney, C., and Lin, W. (2010). Nat. Chem. 2: 838. Falkowski, J.M., Sawano, T., Zhang, T. et al. (2014). J. Am. Chem.Soc. 136: 5213. Sawano, T., Thacker, N.C., Lin, Z. et al. (2015). J. Am. Chem. Soc. 137: 12241. Chen, X., Peng, Y., Han, X. et al. (2017). Nat. Commun. 8: 2171. Gong, W., Chen, X., Jiang, H. et al. (2019). J. Am. Chem. Soc. 141: 7498. Jing, X., He, C., Dong, D. et al. (2012). Angew. Chem. Int. Ed. 51: 10127. Zheng, M., Liu, Y., Wang, C. et al. (2012). Chem. Sci. 3: 2623. Wen, L., Zhao, J., Lv, K. et al. (2012). Cryst. Growth Des. 12: 1603. Wang, X.-S., Chen, C.-H., Ichihara, F. et al. (2019). Appl. Catal., B 253: 323. Wang, C., Xie, Z.G., deKrafft, K.E., and Lin, W.L. (2011). J. Am. Chem. Soc. 133: 13445. Liu, H., Xu, C., Li, D., and Jiang, H.-L. (2018). Angew. Chem. Int. Ed. 57: 5379. Li, H., Yang, Y., He, C. et al. (2019). ACS Catal. 9: 422. Xie, M.-H., Yang, X.-L., Zou, C., and Wu, C.-D. (2011). Inorg. Chem. 50: 5318. Zhang, R., Liu, Y., Wang, Z. et al. (2019). Appl. Catal., B 254: 463. Wu, P., He, C., Wang, J. et al. (2012). J. Am. Chem.Soc. 134: 14991. Xia, Z., He, C., Wang, X., and Duan, C. (2017). Nat. Commun. 8: 361. Silva, C.G., Luz, I., Llabrés i Xamena, F.X. et al. (2010). Chem. Eur. J. 16: 11133. Fu, Y., Sun, D., Chen, Y. et al. (2012). Angew. Chem. Int. Ed. 51: 3364. Liu, Y., Yang, Y., Sun, Q. et al. (2013). ACS Appl. Mater. Interfaces 5: 7654. Fang, X., Shang, Q., Wang, Y. et al. (2018). Adv. Mater. 30: 1705112. He, T., Chen, S., Ni, B. et al. (2018). Angew. Chem. Int. Ed. 57: 3493. Chen, E.-X., Qiu, M., Zhang, Y.-F. et al. (2018). Adv. Mater. 30: 1704388. Wang, C., deKrafft, K.E., and Lin, W. (2012). J. Am. Chem. Soc. 134: 7211. Zhang, Z.-M., Zhang, T., Wang, C. et al. (2015). J. Am. Chem.Soc. 137: 3197. Sun, D., Gao, Y., Fu, J. et al. (2015). Chem. Commun. 51: 2645. An, Y., Liu, Y., An, P. et al. (2017). Angew. Chem. Int. Ed. 56: 3036. Huang, Y.-B., Liang, J., Wang, X.-S., and Cao, R. (2017). Chem. Soc. Rev. 46: 126. Allen, A.E. and MacMillan, D.W.C. (2012). Chem. Sci. 3: 633. Felpin, F.-X. and Fouquet, E. (2008). ChemSusChem 1: 718. Wasilke, J.-C., Obrey, S.J., Baker, R.T., and Bazan, G.C. (2005). Chem. Rev. 105: 1001. Liu, Q., Cong, H., and Deng, H. (2016). J. Am. Chem. Soc. 138: 13822. Lee, Y.-R., Chung, Y.-M., and Ahn, W.-S. (2014). RSC Adv. 4: 23064. Liang, J., Huang, Y.-B., and Cao, R. (2019). Coord. Chem. Rev. 378: 32. Liang, J., Chen, R.-P., Wang, X.-Y. et al. (2017). Chem. Sci. 8: 1570. Huang, Y.-B., Shen, M., Wang, X. et al. (2015). J. Catal. 330: 452. Shimizu, K., Kon, K., Seto, M. et al. (2013). J. Catal. 300: 242. Zhao, M., Deng, K., He, L. et al. (2014). J. Am. Chem. Soc. 136: 1738. Han, Q., He, C., Zhao, M. et al. (2013). J. Am. Chem. Soc. 135: 10186. Han, Q., Qi, B., Ren, W. et al. (2015). Nat. Commun. 6: 10007.

993

995

Index a ABCO3 F 588–590 AB4 O6 F 584–586 absolute adsorption 864, 865 absolute adsorption amount 865 acetonitrile 10, 38, 41, 50, 171, 181, 301, 395, 535, 737 achiral component 21, 300, 939 achiral Lindqvist hexamolybdate 395 actinide triple helices 297, 298 acyl-transfer reaction 13 adamantane-based terpyridine ligand 72 adamantine 50, 692 adsorption capacity 833–835, 865, 867, 869, 871, 873, 894, 901–903, 911, 912, 917, 918, 921–923, 925 adsorption curves 834–835, 902, 921 adsorption enthalpy determination of 863–864 usable asorption on 865–867 adsorption isotherms activating solvent effects on 842 BET multilayers 836–837 effect of pore size on N2 838–840 Langmuir monolayer 835–836 Xe and Kr of 926 adsorption, physical and chemical 834 Ag-based TCOs 683 Ag2 (bga)2 (pzdc)(H2 O) subunits 293 [Ag(4,4-bipyridine)NO3 ] 438 [Ag(ddn)2 ]PF6 and [Ag(ddn)2 ]AsF6 440 A2 Ge4 Se10 465, 490

aggregation-induced emission (AIE) 17, 25, 32, 42 Ag+ heterometallic chalcogenidostannates [Ag4 SnIV SnII S8 ]n 6n– ribbon 231 Ag–Sn–Q (Q = S, Se) 234, 235 [AgSn3 Se8 ]n 3n– anionic chain 232 [Bmmim]7 [AgSn12 Se28 ] 234 [CH3 NH3 ]2 Ag4 SnIV 2SnII S8 229 [CH3 NH3 ]6 Ag12 Sn6 S2 229 [CH3 NH3 ]6 Ag12 Sn6 S21 228 [(Me)2 NH2 ]0.75 [Ag1.25 SnSe3 ] 231 (NH4 )4 Ag12 Sn7 Se22 234 SBUs 228 [SnAg6 Se10 ]n 10n– 233 SnSe4 units 231 3D-[CH3 NH3 ]6 Ag12 Sn6 S21 228 3D-[(Me)2 NH2 ]0.75 [Ag1.25 SnSe3 ] 231 Ag(NH3 CH2 COO)(NO3 ) 654 [Ag(pyppt)(NO2 )](CHCl3 ) 287 Ag–Sn–Q (Q = S, Se) compounds 234, 235 [AgSn3 Se8 ]3– 231 [Ag12 Sn7 Se22 ]4– 231 [AgSn3 Se8 ]n 3n– anionic chains 233 α-Ag2 S semiconductor 703 Ag3 (tpst)2 304 Ag7 (tpst)4 (ClO4 )2 (NO3 )5 (DMF)2 304 A–II4 –III5 –Q12 type 500–501 A–III–Sn2 –Se6 type 501 air-and moisture-stable coordinatively unsaturated organometallic complexes 14

Advanced Structural Chemistry: Tailoring Properties of Inorganic Materials and their Applications, First Edition. Edited by Rong Cao. © 2021 WILEY-VCH GmbH. Published 2021 by WILEY-VCH GmbH.

996

Index

air–liquid interfacial growth method 341 air-stable dimeric dopant, photoactivation of 709 air-stable precursors 708 S-alanine 941 A4 L4 anion cage’s host–guest chemistry 41 Al(H2 O)6 -SO4 sublattice 659 alkaline beryllium borates 577 alkene-linked dipyridiniums 552 alkoxopolyoxovanadate-based SBU 46 alkoxo-polyoxovanadate [V6 O6 (OCH3 )9 -(μ6 -SO4 )(COO)3 ]2 47 2-8, alkyl-monoamines 200 alkynylplatinum(II)-containing donor ligand 18 alloying 686, 694, 767 all-ruthenium analog 734 all-trans conformation 204, 205 Al30 supramolecular compounds 121 aluminium oxo clusters hydrolysis and condensation aqueous synthetic routes 120–122 organic solutions 119 aluminoborates (ABOs) 115 aluminum hydroxide 120, 121 amide-containing tridentate chelating 44 amido-pyridine-triazole (APT) 40 amine-based solvothermal reaction 401–402 amino acid ligands 104, 142 (2-(aminomethyl)pyridinium)SbI5 641 3-ammonium-1-hydroxypropylidene1,1-diphosphonic acid 375 Anderson-type POM (NBu)3 [MnMo6 O18 (tri)2 ] 379 anion-coordination-induced emission (ACIE) 42 anion-exchange method 965 anionic sulfur donors 44 anion-templating effect 36 antenna effect 563

anthracene-bridged benzimidazole ligands 35 anticancer activity 10 antiferroelectric transition 607 antiferromagnetic coupling 779, 794, 805, 807, 809, 810, 813, 814, 818 antiferromagnets 810 antimony trisulfide 699 Archimedean networks 335, 428 Archimedean polyhedral design rhombicuboctahedra 408 truncated tetrahedra 408 argentophilic interactions 294, 295 aromatic D–A–π interactions 13 artificial photosynthesis 980–983 asymmetric catalysis 8, 954, 970–973, 991 asymmetric 1,3-dipolar addition 973 atom-based magnetism 777 aurophilicity-directed self-assembly 21 auxiliary ion–dipole interactions 22 A3 VO(O2 )2 CO3 590–593 Ax RE2 Cu6–x Te6 (A = K–Cs; RE = La–Nd) 495–496 4,4′ -azanediyldibenzoate units 351 azido-mediated systems 805–806 azole-containing ligands 87 azopyridine (azpy) 449, 813, 896

b Ba3 AGa5 Se10 Cl2 502–504 Ba3 (BQ3 )(SbQ3 ) (Q = S, Se) 470 Ba3 (BS3 )1.5 (MS3 )0.5 (M = Sb, Bi) 470 Ba2 Cr4 GeSe10 489 Ba5 Cu8 In2 S12 513–515 Ba4 F4 CrGa2 S6 493 Ba4 F4 MnGa2 S6 493 Ba4 F4 XGa2 S6 493–494 Ba4 Ga2 S8 475 Ba23 Ga8 Sb2 S38 469 Ba10 Ga2 Sn9 Se22 484 BaGeOSe2 480–482 BaHgSe2 471–472 [Ba(H2 O)3 ](UO2 )3 [C6 H4 (PO3 )2 ]2 (O)⋅5(H2 O) (BaUbbp) 374

Index

Ba4 In2 S8 475 Ba12 In4 S19 466–469 Ba5 In4 Te4 S7 482–483 Baker–Figgis isomers 120 Baker–Figgis–Keggin isomers 120 BaLa3 GaSb2 S10 478–479 Ba3 La4 Ga2 Sb2 S15 478–479 Ba6 Li2 CdSn4 S16 504 Ba–Li2 –IV–Q4 512–513 (BA)2 (MA)2 Pb3 Br10 646 Ba–Na2 –IV–Q4 type 511–512 band edge excitonic emission 263 133 Ba2+ , 63 Ni2+ , and 60 Co2+ 273 barium titanate 610 barrel-shaped cages calixarene constructed 67–68 dimetallic clips-constructed 68–70 BaSnO3 thin film 678 Ba7 Sn3 S13 472–473 Ba8 Sn4 S15 473–474 Ba12 Sn4 S23 472–473 Ba6 Zn7 Ga2 S16 506–507 Ba8 Zn4 Ga2 S15 465, 479, 480 [Be2 BO3 F2 ]∞ 577 1,4-benzenebisphosphonic acid 374 m-benzenedicarboxylate 555 benzenedicarboxylates (BDC) 85, 90, 351, 941 1,4-benzene dicarboxylates (BDC) 90, 415, 841 1,3-benzenedicarboxylate units 351 1,3-benzenedicarboxylic acid 68 1,4-benzenedicarboxylic acid 58, 68, 310, 861, 909 benzene-1,2-diol 741 1,3,5-benzenetricarboxylate (btc) 56, 347, 395, 437, 841, 954, 959 1,3,5-benzene tricarboxylic acid tris[N-(4-pyridyl)amide] (4-btapa) 357 4,4’,4-benzene-1,3,5-triyl-tribenzoic acid (BTB) 56 4,4’,4’’-benzene-1,3,5-triyl-tribenzoicacid (H3 btb) 348 benzimidazole-based ligand 35

benzodifurandione-based oligo(p-phenylenevinylene) (BDOPV)-based small molecules 712 benzoporphyrin fiber, conductivity 708 benzoxazinones 20 2-benzoylpyridyl (bzp) 301 beryllium-free borates 577 BiCd4 O(BO3 )3 586 bidentate dithiocarbamate units 21 bilayer-network MOFs binodal (3,6)-connected net 336, 337 Co(II) and flexible tricarboxylate ligand 330 interdigitated and interpenetrated bipillared-bilayer 327 interpenetration and interdigitation 329 kgd network 337 470-MOF 331 471-MOF 331 μ2 -OH bridges 329, 330 proline derivatives and bipyridines 331 2D + 2D → 2D interpenetrated 327, 328 2D double-layered network 325 2D ladder-like 332 2D MOF nanosheets 340, 341 2D structures 325 Zn(II) displays 327, 328 Zn-TCPP nanosheets 341 Bi2 (mdhbqdc)(ox)2 (DMF)4 324 binaphthol (BINOL) 362 binaphthyl-derived ligands 972 binary molecular ferroelectrics 636–637 binodal (3,6)-connected net 336, 337 binuclear half-sandwich metal (Rh or Ir) molecular clips 12, 14 biphenyl-4,4′ -bis[carbonyl-N-(proline)] (BPBP) 332 biphenyl-3,4’,5-tricarboxylate (bhtc) 415 4,4′ -bipy ligand 555, 556 4,4′ -bipyridine 346 N,N’-bipyridine-based ligands 313

997

998

Index

bipyridine (bpy) ligands 311 4,4′ -bis(imidazol-1-methyl)-biphenyl (bimb) 305 bisdithiazole (BT) 821 1,2-bis(4-pyridyl)ethane (bpe) 332 1,2-bi-s(4-pyridyl)hydrazine (bphy) ligand 449 bispyridyl-based Schiff base ligand 305 1,4-bis(pyrid-4-yl)benzene (bpb) 312 1,4-bis[2-(4-pyridyl)ethenyl]benzene (BPEB) 332 bisquaric acid 629 9,9-bis[(S)-2-methylbutyl]-2,7bis(4-pyridylethynyl)fluorene (bmbpyf) 313 bisterpyridine aligands 11 bisthienylethene-based donor 17 bisthienylethene (BET) hexagons 17 tris(bisurea) ligand 41 bitopic triazole 4-(2-pyridinyl)1,2,4-triazole 306 [Bmim]4 [Sn9 Se20 ] 219 [Bmmim]7 [AgSn12 Se28 ] 234 [Bmmim]1.5 [detaH]0.5 [Ni(deta)2 ][Sn4 Se9 ]2 226 [Bmmim]6 [Mn(deta)2 ]2 [Sn15 Se35 ] 224 [Bmmim]8 [Ni2 (teta)2 (μ-teta)][Sn18 Se42 ] 224 [Bmmim]3 [Ni(en)3 ]2 [Sn9 Se21 ]Cl 226 [Bmmim]4 [Sn9 Se19 (Se2 )0.9 Se0.1 ] 219 Bolin’s model 169 borates synthesis inorganic templated borates 112–113 NLO 111 organic templated borates 113 porous aluminoborates (ABOs) 115 TMC-templated 113–115 borogermanates 116–118 boron imidazolate frameworks (BIFs) 444 Borromean rings (BR) 13, 14 bowl-shaped molecules 715–716 breathing action 269 brick-like PBUs 195 brick-wall networks 320, 323

bridging-ligand-substitution reaction 73 μ6 -Bridging sulfide ligand (μ6 -S) 166 μ2 -Bridging sulfido ligand (μ2 -S) 164 μ3 -Bridging sulfido ligand (μ3 -S) 164 μ4 -Bridging sulfido ligand (μ4 -S) 164 μ5 -Bridging sulfido ligand (μ5 -S) 164 Brillouin scattering 617 bromanilic acid (H2 ba) 632 Bronsted acidity 973 Brown’s equal valence rule 403 Brunauer–Emmett–Teller (BET) multilayers adsorption isotherms 836–837 theory 834, 836 BTB-bridged cages 58 BTC-bridged cages 58 building unit (BU) 5, 42, 70, 82, 85, 87, 98, 119, 137, 172–175, 178, 189, 195, 283, 284, 286, 306, 350, 392, 393, 396, 407, 426, 427, 438, 440, 441, 443, 469–471, 475, 486, 488, 491, 499, 507, 527, 528, 581, 594, 731, 790, 858, 882, 954, 985 1,4-butanediamine (bda) 205 tert-butoxycarbonyl (Boc) 324 p-tert-butylthiacalix[4]arene (H4 TC4A) 411

c Ca2 Ba(CH3 CH2 COO)6 657 [Ca(CH3 )3 NCH2 COO(H2 O)2 Cl2 ] 656, 657 [Ca(CH3 NH2 CH2 COO)3 Br2 ] 656 Ca(CH3 NH2 CH2 COO)3 Cl2 654, 655 cadmium-based hybrids 208 cadmium(II) MOFs 302 cage-like Gd38 (ClO4 )6 cluster 124, 125 calcium zinc nitride (CaZn2 N2 ) 688 p- and n-type dopability 689 stability and crystal structure 690 calixarenes 47, 56, 98, 136, 137, 285, 406 constructed barrel-shaped cages 67–68 macrocycles 98 calix[4]resorcinarene 68

Index

camphoric acid (H2 cam) 346 Ca2 Pb(CH3 CH2 COO)6 658, 659 carbide conducting materials 691 silicon carbide (SiC) 691–692 transition metal carbides 695–697 carbonate NLO materials 588 carbon dioxide capture and utilization (CCU) technologies 986 tetra(4-carboxybenzene) methane ligand 455 1,1′ -bis(4-carboxybenzyl)-4,4′ bipyridinium 432 carboxylate-based MOFs 418, 726, 727 carboxylate-containing ligands 85 carboxylate ligands 42, 56, 73, 85–88, 98, 123, 126, 127, 131, 137, 321, 343–347, 350, 376, 377, 412, 446, 455, 456, 808, 977 carboxylate linkers ditopic carboxylate ligands 343–347 hexatopic or octatopic carboxylate 351–353 tetratopic 350–351 tritopic 347–350 carboxylate supported POTis 85 2,4-bis(4-carboxyphenylamino)-6bis(carboxymethyl)amino1,3,5-triazine) 317 4-(4-carboxyphenylamino)-3,5dinitrobenzoic acid 431 1,3,5-tris[4-(carboxyphenyl)oxamethyl]2,4,6-trimethylbenzene (TBDC) 438 (C4AS)2 Ag3 (μ-2,2′ -bpy)2 (2,2′ -bpy)2 285 Ca2 Sr(CH3 CH2 COO)6 657, 658 C4AS-trisilver block 285 Catalan networks 335, 427 catalytic active transition metal centers 961 catechol 85, 363, 364, 391, 405, 410, 741–747 catecholates (CATs) 412 cationic M2 L4 cage 35 [Cd(abt)2 (ClO4 )2 ]⋅H2 O 555

[Cd(bppd)(NO3 )2 (CH3 OH)2 ⋅Cd2 (bppd)3 (NO3 )4 ]⋅(4(HCCl3 )⋅2H2 O)n 307 Cd(4-btapa)2 (NO3 )2 ⋅6H2 O⋅2DMFn 357 [Cd(3,3′ -DPA)(NO3 )2 (H2 O)2 ]n 289 [Cd(envp)2 (ClO4 )2 ]⋅11EtOH⋅6H2 O 301 Cd4 GeS6 520–521 Cd(imb)(pda)] 304 [Cd2 (nbpy4)3 (NO3 )4 ], metal centers 308 [Cd2 ((R)-Hdmpa)2 ((S)-Hdmpa)2 (dpe)3 ]n 306 cds (= CdSO4 ) Topological Network 440–441 Cd2 (TCNQ)3.5 (H2 O)2 741 [Cd(BCbpy)(bpdc)0.5 X]⋅7H2 O 319 Ce4 (H2 TTS)4 tetrahedron 44 central pyridine unit 25 centrosymmetric (CS) 5, 41, 117, 127, 204, 218, 228, 230, 231, 237, 242, 249, 250, 337, 465, 618, 641, 651 cerium-based neutral molecular tetrahedron 44 chain breakage-reunion process 186 Δ-chain lattice 819–821 chalcogenide coordination polymers 189 chalcogenidoantimonates crystallographic data for 238–240 group 12(II) ions and antimony(III) 236–245 group 13(III) ions and antimony(III) Ga–Sb–S compounds 246–248 In–Sb–Q (Q = S, Se) compounds 248–254 SBUs and TBUs 246 group 14(IV) ions and antimony(III) Ge–SbS compounds 254–259 Sn–Sb–S compounds 259–260 thioantimonates 235 chalcogenidoantimonates(III) 196, 249, 267 chalcogenidometalate 4, 198, 209, 212, 218, 236–246, 254–260, 267–269, 271, 273–276, 391, 398–403, 421 chalcogenidometalate-based clusters 4

999

1000

Index

chalcogenidometalate superlattices alkali metals or transition/lanthanide metal complexes 398 amine-based solvothermal reaction 401–402 anion insertion 404 hydrothermal/solvothermal conditions 398, 399 ionothermal approach or surfactants 398 ionothermal synthesis 403 metal complexes 400–401 mimicking zeolites 401 mixing metals for valance–balance 398–399 organic quaternary ammonium or phosphonium cations 398 organic superbases 400 phase transformation 404 postsynthetic insertion of heterometals 404 pyridines/imidazoles linkage 404 reaction parameters 404 surface-capping ligands 399 surfactants 403 synthetic and design approaches 398 chalcogenidostannate Ag+ heterometallic chalcogenidostannates 228 crystalline selenidostannates 222 crystallographic data 220, 221 ILs 219 1D-[Bmmim]2 [Sn3 Se7 ] 223 In8 Sn8 Se34 -T2,2 227 SDAs 224 SnS5 trigonal bipyramids 219 3D frameworks 225 2D-[Bmmim]7 [AgSn12 Se28 ] 226 Chan–Kim–Rees (CKR) 169 charge-neutral mixed-valence cerium(III/IV) clusters 126 charge transfer (CT) mechanism 630 1,4-C6 H4 (CH(OH)(PO3 H2 ))2 375 chelate side-chain auxiliary ligands 292

chemical affinity quantum sieving (CAQS) 931, 934–937 chemical synthetic methods 197 chemical vs. physical adsorption 834 chiral NH-functionalized cyclic structures 14 chloranilic acid 632 chloroacetic acid 124 [(CH3 )4 N]CdBr3 642, 643 (C4 H10 N)CdCl3 641 [CH3 NH3 ]2 Ag4 SnIV 2SnII S8 229, 230 [CH3 NH3 ]6 Ag12 Sn6 S21 221, 228, 229, 232 [(CH3 )2 NH2 ][Al(H2 O)6 ](SO4 )2 659–661 [CH3 NH3 ]5 Bi2 Br11 646 [CH3 NH3 ]5 Bi2 Cl11 646 [(CH3 )2 NH2 ]2 [CoCl4 ] 638 [C2 H5 NH3 ]2 [CuCl4 ] 639 [(CH3 )2 NH2 ]3 [CuCl4 ]Cl 638, 639 [CH3 NH3 ]4 Ga4 SbS9 S0.28 O0.72 H 274 [CH3 NH3 ]2 [H3 O]Ag5 Sn4 Se12 ⋅EtOH 231 [(CH3 )2 NH2 ][M(HCOO)]3 651, 652 [C2 H5 NH3 ]2 [MCl4 ] 640 [CH3 NH3 ][M(H2 O)6 ](SO4)2 ⋅6H2 O 661 [(CH3 )2 NH2 ]3 Sb2 Cl9 643–645 [CH3 NH3 ]2 [ZnCl4 ] 639 choline-binding protein ChoX 36 chondroitin-4-sulfate (Chs) 32 [(CH3 )4 P][HgCl3 ] 643 C3 hydrocarbons separation C3 H6 914 HKUST-1 MOF 913 KAUST-7 914 M-MOF-74 914 open door effect 914 SIFSIX materials 916 UTSA performance 917 UV-vis spectra 913 circular dichroism (CD) spectroscopy 36, 979 Clausius–Clapeyron equation 863, 926 (ClCs6 )[RE21 Q34 ] 515 cluster-surface analogy 163 Cn-alkylpyrogallol[4]arenes 406 [C(NH2 )3 ][Al(H2 O)6 ](SO4 )2 658, 659 [C(NH2 )3 ][Al(H2 O)6 ](SO4 )2 (GASH) 658

Index

[C(NH2 )3 ][M(H2 O)6 ](XO4 )2 658 CO2 -adsorption isotherms 890 Co(II) and flexible tricarboxylate ligand 330 cobalt-based catalytic sites 44 CO2 capture and separation functionalization of ligands 893 hydroxide anion 891 monodentated hydroxide 892 open metal sites 883 polar functional groups 885 pore size and function control 895 spectroscopic, diffraction and computational experiments 890 cocrystals 716 donor–acceptor (D–A) properties of 716–717 structural modes 717 1 : 1 mixed-stack 717–719 P1T1 and P3T1 721 1 : 1 segregated-stack 719–721 (Co(ebic)2 )⋅H2 O)n 288 [Co(1,4-bis(4-pyridyl)butadiyne) (NO3 )2 (H2 O)2 ⋅2H2 O]n 284 CO2 -induced thermoresponsive behaviour 34 [Co(en)3 ]3 (en)[In6 Sb6 S21 ]⋅H2 O 253 CO2 isotherms 899 [Co2 (6-mna)2 ]⋅3H2 O (CUK-2) 365 [Co2 (nbpy4)3 (NO3 )4 ] solvents 308 conducting layered organic-inorganic perovskites 761 conductive materials 6, 7, 669–770 4-connected 2D topology 432, 433 (3,6)-connected kgd layer 436, 437 3-connected second building units (SBUs) 47 5-connected three-dimensional networks 449 (3,4)-connected three-dimensional topological network 448 8-connected three-dimensional topological network 451, 452 (3,6)-connected three-dimensional topological networks 453–455

(4,8)-connected three-dimensional topological networks 455, 456 12-connected three-dimensional topological networks 456, 457 (3,4)-connected topological network 434 3-connected topology 431, 432 3,4-connected topology 433 (3,6)-connected topology 436 4-connected topology network 433 consecutive condensation reaction 26 constructive reactivity 164, 165 controlled hydrolysis 119, 133 conventional MOFs 647 cooperative behavior influence of crystal packing 790 of molecular structure 789, 790 cooperative spin-crossover, coordination polymers 790, 791 coordinated water 49, 121, 289, 326, 330, 338, 349, 374, 888, 929, 954 coordination bonds 32, 38, 61, 200, 236, 283, 338, 410, 411, 419, 450, 726, 788–790, 861, 903, 953, 954 coordination chemistry 40–42, 92, 122, 168, 276, 373, 414, 425 coordination-driven self-assembly 9, 12, 17, 19, 20, 27, 28, 32, 34, 50, 391, 421 coordinatively unsaturated metal sites (CUMs) 954 [Co3 (2,4-pdc)2 (μ3 -OH)2 ]⋅9H2 O(CUK-1) 365 copper(II) benzenehexathiolate coordination polymer (Cu-BHT) 710 copper ions 91, 336 copper–pseudohalide components 38 copper sulfides (Cu2–x S) 701 bandgap energy of 701 copper(I) thiolates 178 CO2 -promoted hydrogel formation 34 CO2 -responsiveness 34 corner-to-corner packing 47 CO2 -triggered morphology transition 34 coumarin 20, 90

1001

1002

Index

counter-anion 10, 17, 71, 123, 179, 932 CPM-132 ([(TPyP-Zn)2 F2 Zn2 (SiF6 )]n 373 Cp* Rh-based dynamic conformational binuclear building block 13 [Cr7 S8 Cl2 (NH3 )14.5 (H2 O)1.5 ]Cl3 ⋅H2 O 218 crystal field method 796 crystalline conductive materials 6 crystalline group 11–15 metal chalcogenides 195, 196 crystalline metal sulfides 273 crystalline solid solutions 722, 723 crystallization chemistry 426–430 crystal packing 1, 76, 147, 330–332, 339, 501, 507, 594, 716, 721, 786–788, 790 crystal structure BaCuFSe of 681 (BdA)Pb2 I6 of 768 (BdA)PbI4 of 768 (C4 H9 NH3 )2 –(CH3 NH3 )2 Pb3 Br10 of 767 LaCuOSe of 681 (MV)2 [Pb7 Br18 ]n of 759 Pt2 (n-PrCS2 )4 I 751 [Rh2 (acam)4 Br] 753 [Rh2 (acam)4 Cl]n ⋅7nH2 O 753 SrCu2 O2 of 681 crystal structure database (CSD) 425 CsAg5 Te3 513 CsBCO3 F 588 (Cs6 Cl)2 Cs5 [Ga15 Ge9 Se48 ] 470–471 (Cs6 Cl)6 Cs3 [Ga53 Se96 ] 486–487 CsCu5 S3 527–528 Cs3 Cu20 Te13 499–500 Cs2 Ge3 M6 Te14 523–524 CsLnCdTe3 486 Cs2 [Mn2 Ga3 S7 Cl] 487–488 CsMnInTe3 491–493 CsPbCO3 F 589 CsRE2 Ag3 Te5 496–497 Cs[RE9 Cd4 Se18 ] 498–499 Cs2 [RE8 InS14 ] 521–522 Cs[RE9 Mn4 Se18 ] 497–498

CS space group 466, 469–472, 475, 477, 478, 480, 482, 484–486, 491, 494, 496, 497, 499, 507–511, 513, 516, 523, 524, 584, 590 C3 symmetric trisbis(urea) ligand 42 cubane clusters 167, 170–172, 179, 190, 191 cubane-like heterometallic sulfide clusters 175 + Cu -based p-type TCOs electrical properties of 682 cubes, metallacages 53 Cu-BHT film 746, 747 cubic lattice system 670 Cu-bipy chains 306 [Cu(bipy)]2 [HPMo12 O40 ] 306 [Cu(4,4′ -bpy)2 (H2 O)][Cu(2-pySO3 )3 ] (NO3 )⋅H2 O 311 Cu3 (btb)2 (H2 O)3 ⋅(DMF)9 (H2 O)2 (MOF-14) 348 Cu3 (BTC)2 HKUST-1 729 thin film 730 Cu-CAT-1 single crystal 742, 743 cucurbit[6]uril-based pseudorotaxanes 321 [Cu(dbpyf)2 (NO3 )2 ] 313 [Cu3 (ddbbe)4 (DMF)6 (H2 O)3 (ClO4 )] [ClO4 ]5 ⋅10DMF⋅10EtOH⋅7H2 O 308 Cu+ 3d10 orbitals 679, 680 [Cu(dpa)2 (SiF6 )]n (SIFSIX-2-Cu) 372 [Cu(dpds)2 (H2 O)⋅2NO3 ⋅3H2 O]n 300 Cu-dsoa 376, 377 [Cu(hfac)2(3,3′ -DPA)1.5 (NO3 )2 ]n 289 [CuIC(C5 H4 CN)4 ]n n+ 430 [Cu4 I4 (dabco)2 ]n (COZ-1; dabco = 1,4-diazabicyclo-[2.2.2]octane) 358 Cu2 (imidazole)3 447 [Cu5 (N3 )10 (bzp)2 ]n 301 [Cu4 (N3 )8 (CH3 CN)3 (bzp)2 ]n 301 [Cu(ptz)]n (Hptz = 3,5-dipropyl-1,2,4-triazole) 441 ([Cu6 (pybz)8 (OH)2 ]⋅I5 –⋅I7 –)n 327

Index

[Cu(Pz)]3 and [Cu2 (Bpz)]n 11 Curie–Weiss law 603, 611, 630, 641, 646, 649, 653, 654, 660, 815 Cu-salen enantiopure ligand 941 Cu1.94 S–CuS heterojunction 702–703 CuTEI cage 65 [Cu3 (TMA)2 (H2 O)3 ]n 414 cyanide (CN) 732–735 cyclobutane products 12 cyclodextrin (CD) 946 1,2-cyclohexanediamine 360 cyclohexyl groups 15 cyclopentadienyl ligands 140, 142 cycloreversion 36 cylinder-like molecules 713–715

d D2 adsorption 931, 932, 934 D-[Cu4 [N(CN)2 ]2 (hmp)4 (CH3 COO)2 ⋅ CH3 CN]∞ 296 ddbbe = (R)-6,6′ -dichloro-2,2′ -diethoxy1,1′ -binaphthyl-4,4′ bis(p-ethynylpyridine) 308 defects definition 843 effect on adsorption 844–852 types of 843–844 delafossite AgCrO2 TCOs 684 dendronized organoplatinum(II) metallacyclic polymers 28 dendronized polymers (DPs) 28 dendronized rhomboidal organoplatinum (II) metallacyclic polymers 28 density function theoretical (DFT) 36, 41, 95, 262, 536, 554, 556, 585, 728, 743, 808, 818, 852, 904, 908, 912, 928, 936 density of coordination sites (DOCS) 25, 27 deprotonation 364, 415, 419, 632 Devonshire theory 606 D-histidine 941 2,6-di(2-naphthyl)anthracene (2,6-DNA) 711 dia topological network 440

diazabicyclo[2.2.2]octane (dabco) salts 629 di(bromomethyl)benzene derivatives 49 N,N’-dibenzyl-4,4′ -di(4-pyridyl)butadiene 552 4,4′ -dicarboxybiphenyl sulfone (H2 dbsf) 297 ′ 3,3’,5,5 -tetra(3,5-dicarboxyphenyl)-4,4′ dimethoxy-biphenyl) 352 2,5-bis(3,5-dicarboxyphenyl) methylpyridium 977 2,5-bis(3,5-dicarboxyphenyl)pyridine 977 6,13-dichloropentacene (DCP) 713 dielectric constant 611, 616–617, 621, 627, 629–631, 633–636, 638–639, 642–646, 648–655, 657–660, 779–780 1,2-di(4-pyridyl)ethane 346 7-(diethylamino)-coumarin (donor) decorated dipyridyl ligand 19 9,9-diethyl-2,7-bis(4-pyridylethynyl) fluorene (dbpyf) 313 diethylformamide (DEF) 187, 310, 315, 316, 319, 348, 352, 358, 359, 419 differential scanning calorimetry (DSC) 611, 612, 639–641 diffuse reflectance spectra for iron-ruthenium analogue 734, 735 1,6-dihydro-2-methyl-6-oxo-(3,4′ bipyridine)-5-carbonitrile 299 1,2-dihydroxybenzene 741 6,11-dihydroxy-5,12-naphthacenedione 12 2,3-dihyroquinazolinones 50 diiridium precursor 13 β-diketoalkane 629 dimerization 36, 63, 179, 739, 821 dimetallic clips constructed barrel-shaped cages 68, 70 di(1H-naphtho[2,3-d]imidazol1-yl)methane (L) 10 dimethoxyphenyl substituents 25

1003

1004

Index

N,N-dimethylaminoethyl methacrylate (DMAEMA) 34 5,5′ -dimethyl-2,2′ -bipyridine 633 N,N’-dimethyl-4,4′ -bipyridinium 552 N,N’-dimethylethylenediamine 890 N,N’-dimethylformamide (DMF) 231, 395, 444, 861, 989 3,5-dimethylpyrazole 12 dimethylpyridine-2,6-dicarboxylate 50 1,6-bis(2,4-dinitrophenoxy)-2,4-hexadiyne 626 dinuclear cluster 167, 171, 172 dinuclear metal-carbene organometallic clips 12 dinuclear metallamacrocycles 10 dinuclear thiomolybdates and thiotungstates 175–178 dinuclear triple-stranded europium helicate 36 N, N’-dioxide-4,4′ -bipyridine ligands 431 dip-coating methods 706, 707 2,6-diphenylanthracene (2,6-DPA) 711 2,2-diphenylbenzopyran (DP) 313 diphenyldipyrrolopyridine (DP-P2P)-naphthalenediimide cocrystals 720 bis(diphenylphosphino)ferrocene 408 bis(diphenylphosphino)methane) PtM 535 bis(diphenyl-phosphinomethyl) phenylphosphine 535 diplatinum (II) acceptors 17, 18 dipole orientation 616 diprotonated 2,6-bis(4-pyridyl methylidene) cyclohexanone 552 2,2′ -dipyridyl 128 4,4′ -dipyridylacetylene (dpa) 372 diquat (N,N’-ethylene-2,2′ -bipyridinium) 552 dirhodium precursor 13 discrete calixarene-based coordination cages 56 discrete chalcogenido Tn clusters 209–216 discrete clusters-based compounds 214

discrete metal-organic metallasupramolecular structures 3 discrete organoplatinum(II) metallacycles 15, 16, 34 discrete Tn (T3 to T5) chalcogenides 216 disk-like molecules 713–715 disk-shape ligand 54 disodium-2,2′ -disulfonate-4,4′ oxydibenzoic acid (Na2 H2 dsoa) 376 dispersion-corrected density functional theory 928 displacive 610, 614, 615, 627, 634 displex (closed-cycle refrigeration) system 763 distyrylbenzene-based CT cocrystal 719 N, N’-disubstituted-bipyridinium salts 551 disulfido ligand 164, 167 4,4′ -disulfo-2,2′ -bipyridine-N,N’-dioxide (dsbpdo) 376 ditopic bridging ligand 71 ditopic carboxylate ligands 4,4′ -bipyridine 346 camphoric acid (H2 cam) 346 IRMOFs 343, 344 M-O-C metal clusters 343 terephthalate (bdc) 343–347 tetraanoinc 345 ditrigon structures 22 divalent Ca(DMSO)6 2+ cation 186 DMAEMA 34 dodecahedron 3, 61–65, 73, 376, 405, 407, 512, 513 dodecanuclear cage-like Mo-Cu(Ag)-S clusters 166 domain motion 624–626, 662 domain wall 621–626, 630, 649 donor-acceptor (D-A) cocrystals 716 properties of 716–717 structural modes 716–717 donor-acceptor-donor stacking 25 donor-acceptor π⋅⋅⋅π stacking interactions 50

Index

donor-type hydrogen bonds 54 doping mechanism 671, 672 doping of silicon carbide (SiC) 685, 691–696, 703 organic conductor crystals 707–708 double chains, MOF 234, 284, 300–304, 327, 497 doxorubicin (DOX) prodrug 573, 574 [dpaH]5 [In5 Sb6 S19 ]⋅1.45H2 O 253 dpe = 1,2-di(4-pyridyl)ethylene) 306 DPTTA-fullerene cocrystals 721 DPTTA-TCNQ cocrystals 718 drop-casting method 706, 707, 713, 719–721 DPTTA-TCNQ cocrystals 718 Drude-like high-reflectance band 630 dual nets 430 Dubinin–Radushkevich method 849, 850 Dubinin–Radushkevich surface 376 Dy26 cores 132 [Dy(H2 O)4 Q[5 ]](C6 H6 O2 )(NO3 )3 ⋅7H2 O 291

e EDTA 568, 569 EFISH technique 564 electrical conductivity (MV)2 [Pb7 Br18 ]n 758, 760 [Pt2 (n-pentylCS2 )4 I] crystals 751 electrically conductive materials 669, 673 electrical properties of Cu+ -based p-type TCOs 682 electrical resistivity CH3 NH3 PbI3 756 CH3 NH3 SnI3 762, 763 (CH3 NH3 )1-x (NH2 CH=NH2 )x SnI3 757 displex (closed-cycle refrigeration) system 763 SnI4 -doped (PEA)2 SnI4 crystals 764 electroluminescence (EL) d8-d10 heteronuclear complexes

tetraphosphine-supported PtAg2 acetylide cluster complexes 545, 546 triphosphine-supported Pt-M heteronuclear complexes 541–544 d10 -d10 heteronuclear complexes 548–549 definition 532–533 OLEDs 532–533, 548–549 photoluminescence of d8 /d10 heteronuclear metal complexes 533–534 electroluminescent materials 5, 531–551 electronic conduction mechanism 670 18-electron M2 L3 -typed cylinders 14 16-electron M2 L2 -type metallacycles 14 electron structures of 2D C4 H9 NH3 PbI4 crystals 765 of 3D CH3 NH3 PbI3 crystals 765 electron-transfer photochromic compounds 557–559 processes 564 electron-transfer photochromism 555, 556, 564 elemental conducting materials semiconductors 670–672 structure and property 672–673 elevatorplatform 264, 265 embonic acid (H4 L) 44, 406 [Emim]2 [Sn2 As2 S4 (S2 )2 Br2 .43 Cl1.56 ] 218 S-enantiomers 941 enantiomer separation 941, 946 enantiopure D and g-MOCs-16 cages 63 and gPd6 (RuL3 )8 cages 62 enantiopure ligands 50, 371, 941, 944 enantiopure proline 396, 397 enantiopure pyrene-functionalized C2 symmetrical bis(tridentate) ligands 36 enantioselective separation chiral MOFs in GC 945–949 in HPLC 943–945 MOFs 939–949

1005

1006

Index

enantioselective separation (contd.) naproxen 943 SPE and solutions 940–943 enantioselective sulfoxide adsorption 941 energy transfer (ET) 4, 19, 20, 39, 534, 555, 556, 563 Er60 complex 130, 131 ethylammonium layers 639 ethylenediamine 107, 109, 116, 182, 183, 240, 253, 306, 361, 395, 410, 749, 798, 810, 890 [Eu(dsbpdo)3 (H2 O)2 ]3– 376 europium 36, 38, 269, 410 Eu/Tb mixed tetrahedron 39–40

f face-to-face π–π interactions 13 Fe(bpb)2 (NCS)2 312 [Fe(1,10-phenanthroline)3 ]2+ controls 400 Fe(NCS)2 ⋅[Fe(bpb)2 (NCS)2 ]⋅2.5 H2 O 312 Fermi energy 677, 681, 757, 763 ferrimagnets 809 ferroelectric domain 625, 663 ferroelectricity 601, 602, 605, 609, 610, 615, 617, 621–623, 627, 630, 631, 634–639, 641, 643–647, 653, 654, 662, 663 ferroelectric phase transition 606, 611–612 ferroelectric properties 6 characterization of dielectric and dielectric switch 616–617 domain motion 624–626 NLO and NLO switch 617–618 polarization switching 621–623 pyroelectric effect 619 classification of 614–616 MOFs 647 molecular 626–662 structural phase transitions dielectric constant 611 DSC 611, 612

microscopic theory 609–611 P-E hysteresis loop 611–614 solid-state substances 601 thermodynamic theory 602–609 symmetry breaking 614 ferromagnets 806–811 Fe4 [Ru (CN)6 ]3 ⋅18H2 O 734 field effect transistors (FETs) 342, 681, 704, 711 fine-tuning conformational motion 10 fixed-bed adsorption process 896 FJI-H14 ([Cu(BTTA)H2 O]n ⋅6nH2 O) 365, 448 FJSM-SnS 222, 271–274 flexible bis(imidazole) ligands 293 flexible polydentate bridging dithiocarbamate ligands 74 fluorescence-resonance energy transfer (FRET) 19, 20 4-formylimidazole 61 Forster energy transfer efficiency (FRET) 19, 20 Frechet-type dendrons 28 Friedel–Crafts benzylation reaction of benzene 959 Friedel–Crafts reaction 973 frustrated magnets, ground states of 815–816 functionalized linkers 893, 959–963 organic 893 functional substances 1, 2 function-directed structural design 1

g gadolinium-based metal-organic octahedron 55 gallium thioantimonates 247, 275 GaN carbon doping 687 2D structure 687 n-type doping 686 gas adsorption material defects 842–852 gas adsorption theory 834 Ga-Sb-S compounds 246–248 GaS 2D photodetector 701

Index

GaS nanosheets 701, 702 GaSnS-1 274, 275 gas sorption isotherms, MOFs for with cage structures 841 with channels 840–841 gas uptake properties MOFs of 7 gate-opening effect 901 geometric frustration 814 evaluation of 815 magnetic 814, 815 germanium (Ge) 671 cubic lattice system 670 diamond lattice structure 671 electrically conducting behavior 671 unit cell 671 germinates 116–118 GeSb2 S7 cluster 255 Ge-Sb-S compound 254, 255, 259, 269, 274 [GeSb2 S6 ]n 2n– 256 Ge/Si hetero island-chain nanowires (hiNWs) 673, 674 giant nested supramolecules 26 Gibbs free energy 564, 602, 604, 833, 834 glamorous enzyme-like reactivity 54 Glauber model 803 S-glutamic acids 941 glycine-containing compounds 654 [G3]-metallodendrimer 28 Goldberg MOPs 396 Goldberg polyhedra 409–410 Goldschmidt’s Tolerance Factor concept 755 Gram-positive bacteria MRSA 26 grand-canonical Monte Carlo (GCMC) simulations 902, 913, 923, 946 grazing incidence X-ray diffraction (GI-XRD) structure 637 group 11(I) metal ions 196 group 12(II) Hg2+ ion 196 group 12(II) ions and antimony(III) 236–245 group 12(II) metal chalcogenides 196 group 13 metal ion indium(III) 212 group III nitrides 683, 684

group IV nitrides 685 group V nitrides 685 group VI nitrides 685 guest-free supramolecular framework 136 guest-induced interactions 13 guest molecules effect, on MOFs 873 gyroidal MOFs 54

h half-sandwich iridium/rhodium 12 halogen atoms 13, 712, 722 halogens 740, 747–752 hard and soft acids and bases principle (HSAB) 123 hazardous metal ions 4, 283 hazardous non-actinide isotopes 269 [H2 [Cu(en)2 H2 O]8 [Cu(en)2 ]3 [(α-SiW11 O39 )Ce(H2 O)(η2 ,m-1,1)CH3 COO]4 ]⋅22H2 O (en = 1,2-ethylenediamine) 306 [H2 dbco][Cu(H2 O)6 ](SeO4 )2 662 [H2 dbco][Cu(X2 O)6 ](SeO4 )2 661 [H-55dmbp][Hca] 635 [H-55dmbp][Hia] 635 H2 dmpa = 6,6′ -dimethyl-1,1′ -biphenyl2,2′ -dicarboxylic acid 306 head-to-head (HH) isomers 11 helicates 3, 35–38, 81 heparin 32 heptaniobate 396, 397 herringbone networks 319–325 herringbone packing linear-shaped molecule 711–713 heteroacenes 711 N-heterocyclic aromatic ligand 557 N-heterocyclic linkers imidazolates 353 pyridine 357–358 heteroleptic structure 25 heterometallic coordination cages 62 heterometallic coordination octahedrons 55 heterometallic 3d-4f clusters 142–149

1007

1008

Index

heterometallic Mo-Fe-S cubane-like clusters 170 heterometallic strategy 884 hexacarboxylate linkers 417, 895 hexadecahedrons 3, 65–67 hexadentate phosphonate 375 hexagonal metallacycles 14, 17–18, 32, 34 2,3,6,7,10,11-hexaiminotriphenylenesemiquinonate (HITP) 341, 342 hexameric metallacyclic complex 15 n-hexane 917–920 1,6-hexanediamine (hda) 205 hexatopic carboxylate linkers 351–353 HgSb2 Q5 -I 242, 243 HgSb2 Q5 -II 242, 243 HgSb2 Q5 -III 242 HgSbQ3 -I 243 HgSbQ3 -II 243 HgSbQ3 -III 243 HgSbQ3 -IV 245 HHTP linkers 741, 742 highly oriented pyrolytic graphite (HOPG) 25 high-nuclearity cupric-niobite clusters 393 high-nuclearity lanthanide clusters 123 calixarenes 136–138 chalcogen elements 139 N-donor ligands 127–129 N,O-donor ligands 129–136 O-donor ligands 123–127 polychalcogen species 140 high-power laser 617 high sublimation temperature 691 high temperature phase (HTP) 491, 492, 616, 641, 656 1,3-di(1H-imidazol-4-yl)benzene (imb) 303 HKUST-1 114, 348, 374, 414, 729, 842, 858–859, 902, 913–914, 919, 921–924, 931, 941 2H-MoS2 crystal structure 701

homo and heteroligand poly NHC metal assemblies 50 homochiral 21 auxiliary ligands 292 1D-helical coordination polymer 292 honeycomb networks 319, 320 host-guest charge-transfer systems 47, 49 host-guest chemistry 9, 35, 41, 50, 77, 405 host-guest Coulombian interaction 400 host-guest interactions 54 H2 pidc– anions 288 HP-MOFs, synthesis of 852, 853 HPrz+ -based compounds 558 2-(4H-pyrazol-3-yl)pyridine 68 H2 pzdc (H2 pzdc = pyrazine-2,3-dicarboxylic acid) 293 H3 tri (2-(hydroxymethyl)-2-(pyridin4-yl)-1,3-propanediol) 379 bis(1H-1,2,3-triazolo-[4,5-b],[4,5-i]) dibenzo[1,4]dioxin (H2-BTDD) 936 Hund’s Rule 794 hyaluronic acid (HA) 32 hybrid zeolitic imidazolate frameworks (HZIFs) 355, 414, 444, 956–958 hydrocarbons separation acetylene and ethylene isolation 909 adsorption separation 900 C2 H6 and C2 H4 mixtures 906 C2 H2 binding site 911 chemical productions 900 China’s energy-saving and low-carbon economy 900 Fe-MOF-74 902 frozen distillation 900 gate-opening effect 901 host-guest fittings and interactions 904–905 kinetic and thermodynamic processes 900 metalloenzymes and synthetic compounds 906

Index

mixed metal-organic frameworks (M’MOF) materials 909, 910 M-MOF-74 series 902 UTSA-100 909 hydrodesulfurization process (HDS) 163 hydrogenated/fluorinated SiC (H/F-SiC) heterobilayer 695, 696 hydrogen-bonded frameworks (HOF) 58 hydrogen-bonded ionic compounds 647 hydrogen-bonding interactions 10, 35, 36, 44, 207, 296, 299, 304, 340, 631, 639, 889, 904, 991 hydrogen isotopes chemical affinity quantum sieving 934–937 kinetic quantum sieving 931–934 special quantum sieving 937–939 1-hydroxyl-2-(3-pyridyl)ethylidene-1,1diphosphonic acid 374 2-(((2-hydroxy-3-methoxyphenyl) methylene)amino)-2(hydroxymethyl)-1,3-propa-nediol 135 1,1,1-tris-(hydroxymethyl)ethane 126 4-(2-hydroxy-1-naphthyl)-1,2napthoquinones 63 hydroxyphenalenone 629

i ideal adsorbed solution theory (IAST) selectivity 884, 894–896, 902, 908, 910, 912, 915, 928, 932–933, 935 III–VI semiconductors 701 II–VI based inorganic–organic hybrid compounds crystallinity, colour and habit 207 crystallographic data for 201–202 crystal structures of 200, 203 di-and mono-amine molecules 199 mono-and di-amine molecules 208 nanostructures 199 optimal range of 208 imidazolates based MOFs 353–355 imidazole ligand 62, 354

imidazolium-based ionic liquids 222, 283 2,5-bis(4′ -(imidazol-1-yl)benzyl)-3,4diaza-2,4-hexadiene (ImBNN) 312 2-(imidazol-1-yl)-terephthalic acid (Im-H2 BDC) 986 iminodiacetic acid (H2 IDA) 143–144 incipient-wetness impregnation method 965 inelastic neutron scattering analysis (INS) 936 infinite polymeric frameworks 953 inorganic chalcogenides 0D to MD 467–468 mixed-dimensional (MD) framework CsCu5 S3 527 Ln3 M0.5 (Ge0.5 /M0.5 )S7 526, 527 nanoclusters 212 one-dimensional (1D) chain chalcogenides Ba4 Ga2 S8 475 Ba8 Ga2 Sn7 Se18 483–484 Ba10 Ga2 Sn9 Se22 484 BaGeOSe2 480–482 Ba4 In2 S8 475 Ba5 In4 Te4 S7 482–483 Ba8 Zn4 Ga2 S15 479, 480 La4 InSbS9 477, 478 Ln4 GaSbS9 475–477 three-dimensional (3D) framework A-II4 -III5 -Q12 500–501 A-III-Sn2 -Se6 type 501 Ax RE2 Cu6-x Te6 495–496 Ba3 AGa5 Se10 Cl2 502–503 BaAg2 GeS4 517–519 BaAg2 SnS4 517–519 Ba5 Cu8 In2 S12 513–514 Ba6 Li2 CdSn4 S16 504 Ba-Li2 -IV-Q2 type 512, 513 Ba-Na2 -IV-Q4 type 511 Ba6 Zn7 Ga2 S16 506–507 Cd4 GeS6 520–521 CsAg5 Te3 513 Cs3 Cu20 Te13 499–500

1009

1010

Index

inorganic chalcogenides (contd.) Cs2 Ge3 M6 Te14 523–524 (ClCs6 )[RE21 Q34 ] 515 Cs[Lu7 Q11 ] 515–516 CsRE2 Ag3 Te5 496–497 Cs[RE9 Cd4 Se18 ] 498–499 Cs2 [RE8 InS14 ] 521–522 Cs[RE9 Mn4 Se18 ] 497–498 La2 CuSbS5 504–506 Lu5 GaS9 516–517 NaGaIn2 Se5 507–508 Na2 Ga2 MQ6 524, 525 Na7 IISb5 S12 522–523 NaIn3 Se5 507–508 Na2 In4 SSe6 507–508 Na2 ZnGe2 S6 510–511 PbGa2 MSe6 508–509 PbMnIn2 S5 494–495 RbRE2 Ag3 Te5 496–497 SnGa4 Q7 509–510 Sn2 Ga2 S5 525–526 Sr5 ZnGa6 S15 519–520 Yb6 Ga4 S15 516–517 two-dimensional (2D) layer A2 Ge4 Se10 490 Ba2 Cr4 GeSe10 489 Ba4 F4 CrGa2 S6 493–494 Ba4 F4 MnGa2 S6 493–494 Ba4 F4 XGa2 S6 493 (Cs6 Cl)6 Cs3 [Ga53 Se96 ] 486–487 CsLnCdTe3 486 Cs2 [Mn2 Ga3 S7 Cl] 487–488 CsMnInTe3 491–492 La4 FeSb2 S10 484–485 Ln2 Mn3 Sb4 S12 485–486 Na6 Zn3 III2 Q9 490–491 RECuTe2 491 zero-dimensional (0D) cluster 466 inorganic conducting materials 670, 704 inorganic conductive materials 6, 669–704 inorganic elemental semiconductors 670 inorganic ligands bridged POTis 88, 89 inorganic metal chalcogenides 196–197

inorganic/organic crystalline conductors 6 inorganic–organic hybrid conductive materials 6, 669 inorganic–organic hybrid frameworks 397 inorganic–organic hybrid metal chalcogenides optical and electric properties 261–263 photocatalytic property ion-exchange property 269–275 photocatalytic hydrogen production 265–267 photodegradation of organic dye molecules 267–268 thermal expansion behavior 263–264 inorganic–organic hybrid TMSPs hybrid TMSPs via synergistic effect 109 organic frameworks 109 polyoxotantalates (POTas) 109 via same lacunary 107–108 supramolecular POTs-templated MOFs 109 via synergistic effect 109 inorganic oxide ferroelectrics 614 inorganic templated borates 112–113 inorganic ternary metal nitrides 690, 693 In-Sb-Q (Q = S, Se) compounds 248–254 [In4 SbS9 SH]n 4n– 253 interconvertible metallarectangles 13, 14 interfacial electron transfer (IET) 89 interlocked Borromean rings (6a-BRs) complexes 13, 14, 431 interlocked quintuple-helices 299 interlocking 1D ladders 309 intermolecular decomposition 962 intermolecular steric contacts 786–788 interpenetrated microporous MOF 895 iodo-functionalized hexamolybdates 395 1,3,5-tris(4′ -iodophenyl)ethynylbenzene 407 ion-exchange chalcogenidometalates 269

Index

ion-exchange property 269–275 ionic-liquid assisted precursor method 213 ionic liquids (ILs) 84, 209 ionic or covalent bonding modes 2 ionothermal synthesis 88, 198, 209, 276, 403 ionothermal techniques 209 iridium(III) complexes 532, 539, 549 iron-molybdenum cofactor (FeMoco) crystallographic structure of 168 double cubane cluster 167 electron transfer pathway 168 [MoFe3 S3 C] fragment of 173 in nitrogenases 167 single cubanes 172 spectroscopic and structural analysis 168 structure of 168 irregular metallacycles 20–22 isolated molecules 816–819 5-(isonicotinamido) isophthalate (inaip) 365 isonicotinic acid (HIN) 131, 132, 377, 378, 905 isoreticular MOFs (IRMOFs) 343, 344, 346, 348, 412, 855, 861, 862 isosteric heat of adsorption 863, 864 isostructural chalcogenidometalate compounds 212 isostructural two-dimensional (2-D) fishnet-type network compounds 739–740

j Jahn–Teller distortion 783, 784, 887 Jahn–Teller effect 793

k kagomé dual (kgd) topology 436 kagomé lattice 821 Keggin-type [BW12 O40 ]5- anions 397 Keggin-type [PMo12 O40 ]3- 306 Keggin-type polyaluminum cations 121 KH2 PO4 (KDP) family 627

kinetically closed pores 65 kinetic quantum sieving method 931, 934 Kolmogorov–Avrami crystal theory 623 4K polyhedral linkers 889

l LaCuOS transparent p-type material 681 La2 CuSbS5 504–506 La/Cu/S slabs 506 ladder-like chains 294, 304–309 La-doped BaSnO3 678 La4 FeSb2 S10 484, 485 La4 InSbS9 477, 478 Landau phase transition theory 606–607 Langmuir–Blodgett (LB)-based thin film 777 Langmuir–Freundlich methods 926 Langmuir monolayer adsorption isotherms 835–836 lanthanide heterometallic 3d-4f clusters 142–149 high-nuclearity 123–140 calixarenes 136–138 chalcogen elements 139, 140 O-donor ligands 123–127 polychalcogen species 140 monometallic lanthanide based single-molecule magnets 140–142 SMMs 122 lanthanide-based single-molecule magnets (Ln-SMMs) 122, 140–142 lanthanide-germanate clusters 118 lanthanide–organic complexes 40 lanthanide phosphonates 375 lanthanide single-ion magnets 794–801 late-transition-metal palladium-oxo polyanions 82 layer-by-layer preparation procedure of poly(L-DOPA) 941, 942, 948 L -[Cu4 [N(CN)2 ]2 (hmp)4 (CH3 COO)2 ⋅ CH3 CN]∞ 296 Lewis acid catalysis 954–956, 973

1011

1012

Index

Lewis acidic ionic liquid 209 Lewis acid sites 883, 884, 991 ligand 1,2-bis(3-pyridyl)ethyne (3,3′ -DPA) 289 ligands squaric acid 332 ligand truncation method 849, 851 Lindqvist-type polyanion [Nb6 O19 ]8- 102 Lindqvist-type polyoxovanadate clusters 380 Lindqvist-type POM SBUs 380 linear chains 183, 186, 284–287, 632 linear dicarboxylate ligand 68, 345 linear pyridine ligand 13, 317 linear-shaped molecules 711–714 liquid phase epitaxy (LPE) approach 92, 93, 948 Li4 Sr(BO3 )2 580 [LiTl(C4 H4 O6 )]⋅H2 O 649 Ln-𝛽-diketone cages 127 Ln4 GaSbS9 475, 477 Ln3 M0.5 (Ge0.5 /M0.5 )S7 526, 527 Ln2 Mn3 Sb4 S12 485, 486 Ln2n L3n 50 local density approximation (LDA) 262 local vibrational modes (LVMs), in C doped GaN 687 lone electron pair (LEP) 106, 196 long-chain hydrocarbon separation alkanes and olefins 918 butane and pentane isomers 918 cis-butene 919 chemical and thermal stability 920 2,3-dimethylbutane and 2,2-dimethylbutane 920 MOF-5 and n-hexane 917 n-pentane and isopentane 918–919 sieving mechanism 917 low dimensional materials 309 lower critical solution temperature (LCST) behavior 15, 17, 32, 34 low-temperature ferroelectric phase (FEP) 602 low temperature phase (LTP) 491, 616, 785 low valent metal phosphonates 418

m macroporous MOFs sorption isotherms 839, 840 MAF-49⋅H2 O 904 MAF-4, Zn(mIM)2 ) 354 magnetic frustration 814 geometric 814, 815 magnetocaloric effect (MCE) 122, 123, 125, 126, 133, 144 magneto-structural correlations 7 of low-dimensional magnets 791–804 in spin-crossover compounds 779–791 [maH]4 [In4 SbS9 SH] 253 main-group elements aluminium 119–122 borates synthesis ABO 115–116 inorganic templated borates 112–113 NLO 111 organic templated borates 113 TMCs templated 113–115 borogermanates 116–118 description 111 germinates 116–118 manganese nitrides 690 bonding of 690 crystal structure of 691 material defects, on gas adsorption 842 maximum switching current 623 [M4 (calix)(μ4 -Y)] 58 MCMP 558–560 mediated magnetic coupling 804–814 azido-mediated systems 805–811 monocarboxylate-mediated systems 811–812 oxalate-mediated systems 812–814 [(Me)2 NH2 ]0.75 [Ag1.25 SnSe3 ] 270 [(Me)2 NH2 ]2 [Ga2 Sb2 S7 ] 246 [(Me)2 NH2 ]4/3 [(Me)3 NH]2/3 [Sn3 S7 ]⋅ 1.25H2 O 219 6-mercapto-3-pyridinecarboxylate (6-mna) anions 365 meso-helix MOFs 293 mesoporous MOFs materials 838

Index

mesoporous MOFs sorption isotherms 838–839 mesoporous MONTs (MMONTs) 70, 958 mesoporous supramolecule 53 metal building units 283 metal-catecholates (M-CATs) 741, 742 metal chalcogenides 4 chalcogenidoantimonates 238 chalcogenidostannates 218–234 group 13(III) and 14(IV) 196 group 15(III) metal ion 196 group 11-15 metal ions 195 group 11(I) metal ions 196 hydrothermal and solvothermal methods 197 II–VI compound semiconductors 198–209 inorganic–organic hybrid 260–264 in ionic liquids cationic [Sb7 S8 Br2 ]3+ clusters 216 cations and anions of 212 charge-balanced [AlCl4 ]- anions 216 crystallographic data 210, 211 definition of 209 discrete chalcogenido Tn clusters 209, 212–216 discrete metal chalcogenides 209 (Emim)Br-AlCl3 216 [Emim]2 [Sn2 As2 S4 (S2 )2 Br2.43 Cl1.56 ] 218 ionothermal synthesis 209 [Sn32.5 Ge27.5 Se132 ]24- 218 organic and inorganic components 198 organic-inorganic hybrid architectures 198 PBUs 195 structural chemistry 196 synthesis of 199 metal-coordinate complexes 224 metal-doped POTis 91 metal-doped POVs 97–98 metal-formate hybrid perovskites 652 metal-formate perovskites 651, 652 metal-imidazolate cages 54, 72

metal-imidazolate coordination cubes 53 metal ions spanning 39 metallacages barrel-shaped cages 67–70 cubes 52–54 cuboctahedrons 65 dodecahedron 61–65 helicates 35–38 hexadecahedrons 65–67 metallosupramolecular cages 74 multiple structural cages 71–74 octahedron 54–61 tetrahedron 38–46 three-dimensional (3D) structures 35 triangular prism 47–50 truncated tetrahedron 46–47 metallacycles dinuclear metallamacrocycles 10 hexagons 14–20 irregular 20–22 multi-layered metallacycles 22–28 polygon-based polymers 28–34 rectangular 12–14 responsive dynamic metallacycles 34–35 supramolecular triangles 10 two-dimensional (2D) structures 9 metallasupramolecular structures 3, 20 metal–ligand (M–L) bonds 19, 77, 100, 405, 728, 779–781, 783, 800 metallodendrimers 18–19, 28 metalloligand RuL3 62 metallo-linkers BINAP 363, 364 BINOL 362–363 bipyridyl 363 catechol 363 N-heterocyclic carbenes 363 porphyrins 359–360 salen 360–362 metallo-porphyrin MOF [Co(TCPP)Co1.5 ] (PIZA-1) 359 metallosalen-catalyzed reactions 971

1013

1014

Index

metallo-salen derived bipyridine ligand 941 metallosupramolecular cages 74 metalloviologen compounds 553–557 metal nodes Friedel–Crafts benzylation reanction 959 Lewis acid catalysis 954–955 oxidation reaction 956–958 ring-opening reaction 959 Suzuki–Miyaura coupling reaction 958 metal-organic coordination metallacages barrel-shaped cages 67–70 cubes 52–54 cuboctahedrons 65 dodecahedron 61–65 helicates 35–38 hexadecahedrons 65–67 metallosupramolecular cages 74 multiple structural cages 71–74 octahedron 54–61 tetrahedron 38–46 three-dimensional (3D) structures 35 triangular prism 47–50 truncated tetrahedron 46–47 metallacycles dinuclear metallamacrocycles 10 hexagons 14–20 irregular 20–22 multi-layered 22–28 polygon-based polymers 28–34 rectangular 12–14 responsive dynamic metallacycles 34–35 supramolecular triangles 10 two-dimensional (2D) structures 9 metal-organic frameworks (MOFs) asymmetric catalysis 970–972 carboxylates 725 CO2 capture and separation properties 868

crystal structure 725 definition 283, 411 design and synthesis 725 as electron conducting material 725 enzymes 969–970 ferroelectrics 647–662 functionalized linkers 959–963 with functional organic ligand 869–871 gas adsorption factors affecting 867–871 guest molecules effect 871–873 high-throughput methods 420 history of 283 inorganic nodes (SBUs) dinuclear units 414 hexanuclear units 415 mixing SBUs 415–416 mononuclear units 414 octanuclear units 415 rod-shaped chains 416 tetrahedral unit 415 trinuclear units 414–415 materials, adsorption enthalpy of 863–873 merged nets approach 421 metal complexes 968–969 metal nodes 954–959 microwave-assisted synthesis 420 molecular constituents used for 853, 854 MOPs 4207 one dimensional (1D) 284–309, 747–752 open metal sites 867, 869 organic linkers with carboxylates 416–418 photocatalysis in 974–983 polyoxometalates 967–968 rate-determining factor 954 redox-active bridging ligand 726 reticulating chemistry 412, 413 seed-mediated approach 419–420 solvothermal reactions 419 structures of 283

Index

surface areas functionality influence 862–863 interpenetration influence 859–861 ligand length influence 855, 858–859 metal moieties effect 861–862 tailor-made approach 420 template-directed synthesis 419 three-dimensional (3D) 342–380, 726–737 topological network 430–457 two-dimensional (2D) 309–342, 737–747 with ultra-high surface area 853–855 UTSA-100a 870–872 metal–organic hosts 47 metal–organic macrocycles 3, 9–77 metal–organic nanocapsules (MONCs) 73 metal–organic nanotube (MONT) 70, 364, 958 metal–organic polyhedra (MOPs) 47, 54, 55, 73–74, 98, 395, 396, 409, 420, 968 metal–organic tetrahedron 44, 46 metal–organic zeolites 353–354 metal oxides (MOs) electronic structure of 675 energy band structure of 675 molecular orbital diagram and band structure 675 metal-oxo clusters 3, 81–149 metal sulfides bonding in 697 crystal structures of 698 examples 697 metal-triazolate MOFs 735–736 metal-variable isostructural octahedral nanocages 54 4’,4’’,4’’’,4’’’’-methanetetrayltetrabiphenyl4-carboxylate (mtbc) 350 methylamine molecules 61, 62, 208, 231 1-methyl-3,5-bis[3-(pyrid-2-yl)-1,2,4triazolyl]pyridone 128 N-methyldiethanolamine 126

methylenebis(3,5-dimethylpyrazole) (H2 MDP) 322 methylphosphonic acid (MePO3 H2 ) 137 N-(methylpyrrolidinium)3 Sb2 Br9 640 Mg-MOF-74 844, 849, 862, 883–884, 888, 889, 902, 909–911 microporous chalcogenides 196, 218 microporous MOFs sorption isotherms 838 microscopic theory 609–610 [MII (HCOO)3 ]-based MOFs 653 MIL-91(Ti) 897–898 MIL-100(Cr) 347 MIL-101(Cr) 344–345, 954–956, 964–968, 986 M(1,2-dap)3 InSb3 S7 252 1 : 1 mixed-stack cocrystals 717–719 [MLi(C4 H4 O6 )]⋅H2 O 648 M8 L12 -type supramolecular cubes 50 M3 [(M4 Cl)3 (BTT)8 ]2 357 Mn2 (api)Sb2 S5 236–237, 264–265 [Mn3 (HCOO)6 ]⋅C2 H5 OH 650 Mn(ImBNN)2 (NO3 )2 312 Mn-linked cubic framework 393 M3n L2n 71 Mn2 S6 octahedron single chain 485 Mobil Thirty-Nine (MTN) 347, 358, 420, 446–447 Δ-MOC-16 62 M-O-C metal clusters 343 modified pillaring strategy 896 MOF-5 343, 367, 412–414, 416–417, 726–727, 838, 860, 862, 917, 921, 933, 966 MOF-74 345–346, 416, 459, 844, 883–885, 890, 892 MOF-74 (M-DOBDC) 923–924 MOF-901 82, 435–436 MOF [Cu(3,3′ -DPA)(CH3 OH)(NO3 )2 ]n 289 Mo-Fe-S clusters 170, 171 MOF-74-Zn 458 moganite topological network 442–443 molecular-based metal functional complexes 425

1015

1016

Index

molecular ferroelectrics binary 636–637 organic 626–636 organic–inorganic hybrid compounds 638–647 molecular magnets 7, 122, 125–126, 459, 777–823 molecular organometallic compounds 12 molecular packing modes, of organic molecules 708 molecular recognition 3, 9, 35, 47, 54, 96, 377, 425, 572–573, 663 properties 54 molecular simulations 927 molecular square 9 molybdenum blue (MB) 87, 104 molybdenum sulfide (MoS2 ) 697, 698 crystal structures 699 phases, structure of 700 sliding properties 697–698 monocarboxylates 416, 4467 mediated systems 811–812 monocyclopentadienyl (CP) ligand 140 monolayer nickel bis (dithiolene) complex nanosheet 745 monolinear-alkyl substituted BTBTs 712 monometallic lanthanide based single-molecule magnets 140–142 monothiolate ligands 180 monovacant Keggin-type polyanions, lanthanide complexes 306 Monte Carlo simulation 897, 904 Mott-Hubbard insulator, of LaVO3 682 Mo(W)-Cu(Ag)-S clusters multiple cubane-like structures 178–182 thiomolybdates or thiotungstates building units 173–175 tri-and dinuclear thiomolybdates and thiotungstates 175–178 Mo(W)-Cu-S cubane-type clusters 179 (MQ)ZnBr3 566 β-[(MQ)ZnCl3 ] 565

[MSbQ3 ]n n– moieties 243 M2 S island 164 multi-bisthienylethene hexagons 17, 18 multi-layered metallacycles 22–28 multiple structural cages 71–74 multivariate (MTV) MOFs 367, 417–418, 971, 976, 985 [MV][BiBr5 ] 643 (MV)4 [Bi6 Cl26 ] 569 (MV)4 Bi6 Cl26 ⋅2H2 O 560, 562 (MV)2 [Pb7 Br18 ]n 570

n Na2 BaSnS4 511 N2 adsorption 883 of HKUST-1 842 of MOF-14 842 of NU-1000 843 PCN-222 839 pore size effect 838 single-crystal ordered macropore (SOM-ZIF-8) 839 UiO-66 838 naked sulfide anions 178 NaNO2 615 nanostructured Si utilization 672, 673 nanotube fabrication, [Pt(en) (bpy)I]4 (NO3 )8 .16H2 O 749 naphthalene diimides (NDIs) 551 1,5-naphthalenedisulfonate (1,5-nds) 376 naphthanoimidazolium moieties 10 narcissistic 50 self-sorting 38 Na3 (2,4,6-trihydroxy-1,3,5benzenetrisulfonate) (β-PCMOF2) 375 Nature’s chemical armory 14 Na2 ZnGe2 S6 510, 511 Na6 Zn3 III2 Q9 490, 491 [Nb31 O93 (CO3 )]23– 396 NbOFFIVE-1-Ni framework 899 NbO topological network 441, 453 NCG with benzoyl group (NBzG) 369

Index

NCS space group 469, 471, 475, 477, 480, 482, 484, 507, 511 N-donor adsorption sites 898 N-donor 1,2-di(4-pyridyl)ethane 346 N-donor ligands participated POTis 87–89 networked reactor system 397 neutral amine molecules 283 New Fuzhou model 168, 169 (NH4 )4 Ag12 Sn7 Se22 233, 234 (NH4 )[M(HCOO)3 ] 653 (NH4 )[Zn(HCOO)3 ] 653 [Ni4 (bpp)2 (OH-bdc)2 ] 307 nickel bis (dithiolene) complex nanosheet 746 [Ni18 Cl6 (L1)6 (MNA)6 ] 67 nicotinamide adenine dinucleotide (NADH) model 20 nicotinic acid (HNIC) 131 [Ni(deta)2 ]1.5 [In3 Sb2 S9 ]⋅H2 O 252 [Ni(en)3 ]3 (en)[In6 Sb6 S21 ] 253 [Ni(en)3 ][InSbS4 ] 249 Ni3 (HITP)2 crystalline structure 743 porous structure 743, 744 [Ni(imb)(pda)(H2 O)] 303 Ni8 L12 X4 61, 62 Ni(NO3 )2 311 [Ni(phen)3 ]2 Sb18 S29 235 [Ni(α-RR-hmta)](ClO4 )2 294 [Ni(α-SS-hmta)](ClO4 )2 294 N-isopropylacrylamide (NIPAAM) 32, 34 nitrate anions 10, 35, 54, 55, 313, 318 nitric oxide (NO) 44 nitride conducting materials 683, 690 cubic rocksalt-type crystals 684 group III nitrides 684 structure and property 686 4,4’,4’’-nitrilotrisbenzoic acid 979 nitroalkylated imidazole 400 nitrogen alkylation 391, 395 nitrogen containing heterocyclic compounds 735

two-dimensional metal organic frameworks 747 nitrogen sorption isotherms 837 nitronyl nitroxide (PTIO) 44 p-nitrophenyl-acetylacetonate 90 p-nitrophenyl acetylacetone ligands 89 noble gas separation [Co3 (C4 O4 )2 (OH)2 ]⋅3H2 O 927 CROFOUR-1-Ni and CROFOUR-2-Ni 925 FMOFCu 929 HKUST-1 923 MOF-74 (M-DOBDC) 923, 924 SBMOF-1 924 SCU-11 929 noncentrosymmetric (NCS) 465 crystal structures 5 MoS2 semiconductors 699 nonlinear coefficient 617 nonlinear optical (NLO)-active MOFs 317 nonlinear optical (NLO) effects 111 nonlinear optical (NLO) materials ABCO3 F family 588–590 apatite-like borates 586–588 KBBF 575–582 nonlinear optical (NLO) phenomena 617 non-uniform nets 427, 428 N2 O tridentate chelating units 54 N-p-tolylsulfonyl-L-glutamate (NTsG) 369 N-substituted monocyclic aromatic ion-templated compounds 558–560 NTE 264 nth-order phase transition 602 nut-like hexagonal bismetallo-architecture 27

o octahedral cages 58, 63–65, 73, 345, 347, 356 octahedral nanocages 54 octatopic carboxylate linkers 351–353

1017

1018

Index

O-donor organic ligands 85 olefin-functionalized bridging ligands 12 oligomeric aluminum oxo clusters 119 1D-α and β-[ZnTe(hdz)2 ] 207, 208 1D azido–copper chains 301 1D-[Bmmim]2 [Sn3 Se7 ] 222, 223 1D chiral chains 940 1D double chains 302, 303 1D infinite non-interpenetrated molecular ladder 305 1D ladder-like polyoxometalate-based MOFs 306 1D [M(ebic)]n zigzag chains 288, 289 1D molecular-ladder MOFs 306 1D polymeric array 287 1D polymeric phosphates 119 1D-[Prmmim]2 [Sn3 Se7 ] 222 1D ribbon-like structures 249 1D zigzag Cu(I)-MOF 289 1D-[ZnQ(pda)] compounds 261 one-dimensional (1D) metal–organic frameworks 747 [Co(1,4-bis(4-pyridyl)butadiyne) (NO3 )2 (H2 O)2 2H2 O]n 284, 285 double chains 300–304 helical chains actinide triple helices 297, 298 AgPF6 299 bga ligand 293 biopolymers 291 [Cd(bpea)(phen)2 ] 300, 301 chelate side-chain auxiliary ligands 292 chiral or achiral building blocks 291 cucurbit[5]uril reaction 291 Cu[N(CN)2 ]2 (Hhmp) 297 1,6-dihydro-2-methyl-6-oxo-(3,4′ bipyridine)-5-carbonitrile 299 D- or L-configuration 292, 293 formation of 293 homochiral crystallization 292 μ1 -(η2 -N,O)-pzc ligands 293 nanotube corners 299 Σ-shaped ligands 294

three left-handed helical chains 297 three right-handed helical chains 297, 298 V-shaped bridging ligand dicyanamide 292 ladder-like chains 304–309 linear chains 284–287 [Ni4 (4,4′ -bpy)4 ] 285 polyoxometalates (POMs) 285 SC-SC transformation 287 zigzag type chain 287 one-dimensional (1D) organic–inorganic hybrids 768 one-pot self-assembly reactions 171, 191 open door effect 914 optical limiting effects 94, 163 optical materials 5 definition 531 electroluminescence (EL) 532 NLO materials 575–594 optical transverse modes 610 opto-optical switching NLO properties 563–565 photoswitching luminescence 560–563 Orbach process 792, 796 order-disorder phase transition 610 order-disorder type 610, 614, 615, 646, 653 organic amine molecules 283 organic conductive materials 669, 704–724 organic conductor crystals 707 doped systems 708–709 single-component systems 710 organic electronics 704, 722 organic ferroelectrics 6, 626–636, 647 organic–inorganic hybrids 754 compounds 638 conductive material 7 one-dimensional 768–770 perovskites 754 three-dimensional 754–759 two-dimensional 759–768

Index

organic-ligand-free discrete T3 cluster 212 organic light-emitting diode (OLED) 532, 704 organic linkers with carboxylates dicarboxylates 416 hexacarboxylate 417 multivariate 417 octacarboxylates 417 tetracarboxylates 417 tricarboxylates 417 organic linkers with functional groups mixing hetero-linkers 419 nitrogen-heterocyclic linkers 418–419 oxygen-containing ligands 418 organic photocatalysis 977–980 organic photochromic compounds 551 organic semiconductor crystals 710 bowl-shaped molecules 715–716 cylinder-like or disk-like molecules 713–715 drop-casting method 706 growth 706 linear-shaped molecules 711–713 organic single crystals 704, 705 packing arrangements 707 physical vapor transport (PVT) technique 705 solution assembly process 705 organic superbases 399 organic templated borates 113 organometallic capsule 14 organoplatinum(II) acceptors 17 oxadiazole-bridging ligands 294 oxalate-mediated systems 812–814 oxazolinebased bis(tridentate) ligand 39 oxidation reaction 7, 101, 175, 554, 956–958 oxidative kinetic resolution (OKR) 64 oxido-centered chromium trimers 347 oxime ligands 87 oxo-bridged trimers 414 oxo clusters lanthanides 122–149 main-group elements

aluminium 119 borates synthesis 111–116 borogermanates 116–118 description 111 germinates 116–118 of transition metal 81–102 oxothiomolybdate dimer [Et4 N]2 [Mo2 O2 S2 (edt)2 ] 177 oxotrithiomolybdate MoOS3 2– 174 oxygen-and nitrogen-supported POVs 98–99 oxygen-containing ligands 315, 418

p packing arrangements, in organic crystals 707–708 [paH]3 [In3 Sb2 S9 ] 252 paraelectric phase (PEP) 602–604, 607, 609, 610, 614, 615, 618, 621, 624, 625, 630, 633, 639–642, 644, 648, 651, 653, 654, 656, 657, 659, 661, 662 paramagnetic Gd3+ 55 Pauling’s electrostatic valence rule 398, 399, 403 Pauli Principle 794 [Pb(bimb)1.5 (NO3 )2 ](DMF)n 305 Pb2 BO3 Cl 581 Pb2 (BO3 )(NO3 ) 582 p-benzenedicarboxylate 555 PbGa2 GeSe6 509 PbGa2 SiSe6 508, 509 PbMnIn2 S5 494–495 PbO6 F2 polyhedron 589, 590 PbTiO3 624–626, 628 Pb(Yb1/2 Nb1/2 )O3 –PbTiO3 (PYNT) crystal 636 PCN-224 359, 419, 961 PDMAEMA 34 P-E hysteresis loop 611, 614, 621, 622, 624, 633–637, 641, 642, 650, 653, 654, 661, 662 pentadecanuclear Eu(III) complex 130 pentathienoacene (PTA) 712 perfect reaction chemistry 1

1019

1020

Index

peripheral pyridyl groups 15 perovskite-type transparent conducting oxides 682 perylene-TCNQ complexes 721, 722 perylo[1,12-b,c,d]selenophene (PES) 714 perylo[1,12-b,c,d]thiophene (PET) 714 PF6 -anions 54, 299 1,10-phenanthroline 128, 374 1,10-phenanthroline-2,9-dicarbal-dehyde dioxime (H2 phendox) 136 1,10-phenanthroline-2,9-dicarboxylic acid (H2 phenda) 136 phen ligand 297, 299, 300, 303, 814 ortho-phenylene-bridged oligourea ligands 41 1,4-phenylenediacetic acid (H2 pda) 303 o-phenylenediamine 969 phenylphosphonic acid (PhPO3 H2 ) ligands 137 phonon-assisted light emission 673 phophonates 85, 87 phosphate/phosphite 555 phosphite-stabilized Ti3 (μ3 -O) 86 phosphonate (POTis) 85, 86, 88, 90, 100, 315, 316, 373–375, 418 phosphonate linkers 373–375 1,3,5,7-tetrakis(4-phosphonatophenyl) adamantane 373 phosphorescence with high quantum yields (PLQY) 546 phosphorus-, arsenic-and antimony-containing ligands (POVs) 98 phosphotungstic acid (PTA) 967, 968 photocatalysis CUMs and functional organic linkers active guest sites and active site 984, 988–991 metal nodes and linkers 986–988 mixed linkers 984–986 mixed metal centres 984 organic photocatalysis 977–980 photo-degradation of pollutants 974, 976

photo-driven water splitting and artificial photosynthesis 980 photochromic bisthienylethene units 18 photochromic compounds, viologen NN’-dihydro-4,4′ -bipyridiniums 552 diquat 552 NN’-disubstituted-bipyridinium salts 551 electrical applications 569–572 electron-accepting ability 552 metallo 553–557 molecular recognition 572–574 opto-optical switching 560, 562–565 Paraquat 552 photocatalysis 568–569 radiation detection 565–568 N-substituted monocyclic aromatic ion-templated compounds 558–560 Tony blue 552 photocyclization 36 photodegradation of organic dye molecules 267–268 photoluminescence of d8/d10 heteronuclear metal complexes 8 10 d -d heteronuclear alkynyl complexes 534–537 d10 -d10 heteronuclear alkynyl complexes 537–539 photolysis 12, 268 photoswitching luminescence 560, 562–563 p-hydroxybenzoic acid 292 physical vapor transport (PVT) technique 705 physical vs. chemical adsorption 833–834 physisorption isotherms 835, 838 Phz–H2 xa 633–635 pillared-layer MOFs 419 π–π interaction 949 planar clusters 120, 350 Platonic network 427, 428 Platonic polyhedron cubic systems 406

Index

dodecahedra 407 icosahedra 407 octahedra 406 tetrahedra 406 PLN-0.49PT crystal 636 plumbichloride hybrids 758 PM6 semiempirical molecular orbital analysis 28, 32 PNIPAAM 32, 34 polarization switching 621–623 polyalcohols 127 polyaluminum clusters 120 polyanion-induced hierarchical self-assembly 32 polycatenated 3D molecular ladders 308 polycyclic supramolecules 28 polyethylene glycol (PEG) 65, 403 polygonal/polyhedral complexes Archimedean polyhedral design cuboctahedra 407–408 rhombicuboctahedra 408 truncated tetrahedra 408 Ga/Fe-catechol linkage 410 Goldberg polyhedra 409–410 Ln–tridentate ligand 410 metal–calixarene subunits 411 metal-carboxylate linkage 410–411 Pd/Pt-nitrogen linkage 410 Platonic polyhedron cubic systems 406 dodecahedra 407 icosahedra 407 octahedra 406 tetrahedra 406 prismlike systems 408 Stellated polyhedra 410 polygon-based polymers metallacycles 28 polyhedral projection labeling 73 polymeric W(Mo)-Ag-S anions 182 polymorphism 205, 207, 692, 698 polynuclear single-molecule magnets 792–794 polyoxometalate (POM) 4, 7, 81, 82, 104–105, 107, 110, 287, 307, 377,

391, 419, 421, 450, 551, 963, 967, 982, 983, 988, 989 charge-balancing by complexes 397 cubic molecular cage 395 inorganic anions/cations inclusion 392–393 MOF assembly 982 multivariate metal mixing 393–395 networked reactor system 397 nitrogen alkylation 395 1D MOF 285 organic anions/cations inclusions 393 organic ligands coordination POM-based cubic molecular cage 395–396 POMs-based nets 396 POM-based nets 396 reaction pH control 397 single-crystal to single-crystal transformations 397 templating by clusters 397 polyoxometalate-based inorganic-organic hybrid compounds 306 polyoxoniobate cluster anions of (Nb27 O76 )16- 396 polyoxoniobates (PONbs) 102–104, 396, 397 polyoxopalladates (POPs) 82, 105–106 polyoxotantalates (POTas) 109–111 polyoxothiometalates 393 polyoxotitanates (POTis) 82 bandgap engineering and photo-related activities heterometallic doping 91–92 ligand modification 89–91 diverse structures of carboxylate surpported POTis 85 inorganic ligands bridged 88–89 N-donor ligands participated 87–88 phosphite-stabilized 86 phosphonate ligands 86 potential application of 92–95 polyoxotungstates (POTs) 107–109 inorganic–organic hybrid 107–109 TMSPs 107

1021

1022

Index

polyoxovanadate-based octahedrons 55 polyoxovanadates (POVs) 56 carboxylate derivative 101 chemistry of 95 ligands-participated oxygen-and nitrogen-containing ligands 98 phosphorus-, arsenic- and antimony-containing ligands 100 silicon- and germanium-containing ligands 101 metal-doped 97–98 MOPs 396 organic-inorganic hybrids 101 organic polyhedra 47, 396 structure and catalytic property 101, 102 template-effect-determined 96, 97 poly(ethylene glycol) (PEG)–selenidostannate composite 403 poly(L -DOPA) thin-film 948 POM. see polyoxometalate (POM) POM-organic frameworks (POMOFs) 377 pore-size regulation 897 porous coordination cages (PCCs) 58 porous coordination polymers (PCPs) 283, 348, 456, 457, 820 porphyrin 316, 359–360, 406, 416, 417, 419, 714, 976, 981, 985 positive thermal expansion (PTE) 264 post-combustion flue gases 882 post-synthetic ligand exchange (PSE) 364 post-synthetic modification (PSM) method 358, 959, 987, 988 potassium sodium tartrate tetrahydrate, (KNaC4 H4 O6 ⋅4H2 O) 626, 647 powder X-ray diffraction 56, 182, 208, 340, 743, 745, 840, 869, 891, 897, 899, 955 primary building unit (PBU) 195, 245, 276, 443, 475, 488

2-propylimidazole-4,5-dicarboxylate (H3 pidc) ligand 288 protein-sized windmill-like cluster 103 proton transfer 163, 627, 629, 630, 632, 634, 635, 647 Prussian blue (PB) (Fe4 [Fe(CN)6 ]3 ) 732 crystal framework for 733 electrical conductivity of 732 electronic conductivity of 732 Prussian blue analogs (PBA) crystal structures 733, 734 iron-ruthenium analog 734 Rbx Mn[Fe(CN)6 ]y ⋅zH2 O 733, 734 pseudo-intramolecular pathway 44 pseudo-octahedral CdII center 25 pseudo-pentadentate ligand 25 pseudospin theory 610 PTA. see phosphotungstic acid (PTA) [Pt(chxn)2 I]I2 single crystals 748 p-tert-butylthiacalix[4]arene (H4BTC4A) 136, 147, 411 Pt-HTT metal organic frameworks 743, 744 [Pt2 (μ-Ph2 PNPPh2 )2 (C≡CC6 H4 R-p)4 ]2– 535 PtS topological networks 442–443 p-type Cu-doped NiO thin film fabrication 681 p-type transparent conducting oxides (TCOs) crystal structures 680 electronic structure 679 materials designed for 678 puckered layers of 2D [Ag3 Te5 ], 497 PXRD patterns 208, 209, 271, 861, 955, 958 pybz ligands 328 pyrazine (Prz) compounds 556 2,3-pyrazinedithiolate (pdt) 731 pyrazine spacing ligands 12 pyrazole based ligands 11 pyridine-2,6-dicarboxamide coordination 38, 410 2,4-pyridinedicarboxylate (2,4-pdc) 365 2-pyridine-sulfonate 311

Index

pyridine-2-thiolate (C5 H4 NS) 747, 749 4-pyridinethiolate ligands 457 3,5-bis(pyridine-2-yl)-1,2,4-trizole) 128 5-(pyridin-4-yl)isophthalate (PIP) ligand 65 2,6-bis(4-pyridinylmethyl)-benzo[1,2-c:4 5-c′ ]dipyrrole-1,3,5,7(2H,6H)tetrone), 290 N,N’-bis-(4-pyridinylmethylene)-1,5naphthalenediamine (nbpy4) 308 4-pyridyl benzoate (pybz) 327 4,4′ -bis(4-pyridyl)biphenyl ligand 311 4-pyridyl-CH2 N(CH2 COOH)(CH2 PO3 H2 ) 374 pyridyl ligands 12 tris2-[(4′ -pyridylmethyl)amino] ethylamine (hsb-1) 297 N,N’-bis(4-pyridylmethyl)-pyromellitic diimide) 307 2,4,6-tris(4-pyridyl)pyridine (pytpy) 357 3-(2-pyridyl)-4-(4-pyridyl)-5-(3-pyridyl)4H-1,2,4-triazole (pyppt) 287 1,3,5-tris(4-pyridyl)triazine 410 pyrocatechol 741 pyroelectric effect 614, 619, 620

RECuTe2 491, 492 recyclable enantioseparation 64 redox-active metal 20 redox-active molecular flask 20 Ree’s model 169 remanent polarization 622, 637 reorientable ferroelectric 607 responsive dynamic metallacycles 34–35 Reticular synthesis 425–426 reversible gel-to-sol transition 34 reversible photochromism 36, 88 rhodamine 20 rhodamine (acceptor) functionalized diplatinum (P) unit 19 rhomboid metallacycle 17 Rietveld method 420 Rietveld refinements 208 R-naproxen 942 Rochelle salt (RS) 609, 626, 647–649 rod-packing topology networks 457–459 room/low temperature solution methods 197 Ru(II)-organic building blocks (ROBBs) 28

s q Qc-5-Cu-sql 887 quantum dots (QDs) 209, 262–264, 266, 399, 726 quantum tunneling of magnetization (QTM) 792, 796–798, 801–803 quartz topological network 441 quasi-regular net 429 quinhydrone (QHQ) cofactor 47 quinone monoimine 58 quinoxalinones 20

r racemic macrocyclic compounds 294 radiation detection 560, 565–568 Raman process 792, 797 RbCaCO3 F 588, 589 reaction mechanism 226, 979 rectangular metallacycles 12

saturated hydrocarbons 922 SBMOF-1 924–925 [Sb7 S8 Br2 ](AlCl4 )3 209, 211, 216, 217 SBUs. see second building units (SBUs) Schiff bases 129, 135, 785 Schläfli symbols 314, 317, 331, 335, 337, 350, 370, 372, 428, 436, 443, 457, 910 Schläfli topology 309, 311, 319 second building units (SBUs) 47, 117, 195, 196, 242, 246 Anderson-type POM 379 automated assembly of 420 binuclear-Ni 372 B3 O7 112 Co3 331 condensation of PBUs 276 construction of multidimensional chalcogenides 212

1023

1024

Index

second building units (SBUs) (contd.) copper paddle-wheel 362 Cp3 Zr3 O(OH)3 42 Cr-based trimeric 344 Cu2 (COO)4 352 Cu3 (II) dinuclear 286 1D 287 dicopper paddle-wheel 347, 417 with different geometric structures 427 discrete 286 heterometallic 228, 255 Hg2 Sb2 Q10 244 infinite inorganic rod-type 345 inorganic 392, 411, 412 inorganic nodes 414–416 isolated 457 kinds of 224 Lindqvist-type POM 380 mixed-metal 984, 985 modified 7, 954 (M2 SbS8 )-1 243 open metal nodes 7 paddle-wheel 337, 456 pentanuclear Ge–Sb–S clusters 258 POM clusters as 377 POVs 98 rational design of 5, 426 In2 Sb2 S9 249 In2 Sb2 S10 251, 253 In2 Sb2 S12 252 In2 SbS8 248 Sn–Q 196 symmetrically compatible 350 symmetric polyatomic 98 tetranuclear 256 trigonal 984 trinuclear Zn3 (COO)6 367 Zn2 (COO)3 320 Zn4 O(CO2 )6 352 [Zn4 (μ4 -O)(O2 CR)6 ] 361 Zn(II) paddle-wheel 371 Zr6 417 Zr6 O4 (OH)4 (O2 CR)12 cluster 364

second-harmonic generation (SHG) effect 531, 563–565, 575, 579, 581, 585, 589, 590, 651, 652, 662 1 : 1 segregated-stack cocrystals 719–721 selenidostannates 209, 218, 219, 222, 224–226, 276 self-absorption phenomenon 560 self-assembly 5, 9, 11 amphiphilic discrete organoplatinum 16 anion-binding metal–organic macrocycles 34 artificial multicomponent 408 aurophilicity-directed 21 bipyrazolate ligands 68 cartoon presentation of 19 of cationic tetranuclear-metal building blocks 58 of CdII ions 25 chiral luminescent europium coordination tetrahedral cages 38 of chiral NH-controlled supramolecular metallacycles 15 coordination-directed 410 coordination-driven 12, 17, 27, 34, 50, 391 of copper(I) cuboctahedral coordination cages 71 in crystal engineering 358, 405 with 120∘ diplatinum (II) acceptor 19 direct anion-adaptive 35 of discrete hexagonal nanobarrels 70 of discrete metallacycle 32 discrete supramolecular fractals 28 DNA-mimic homochiral 1D helical Cu(II) chain 297 of DOX-bearing metal–organic polyhedron 573 hetero-chiral 21 of hierarchical 31, 53, 70 of long-chain dicyanoalkane ligands 440 of metal cations and sulfide anions 163

Index

metal–imidazolate coordination cubes 53 metal–imidazolate tetartoids and cubes 73 mixed-metal 39 M3n L2n family 72 molecular 145 narcissistic self-sorting 38 novel supramolecular systems 40 of octahedral metal centers 896 octanuclear Zn8 L|b6 cages 50 one-pot 27, 83, 149, 171, 191, 376 rational coordination directed 408 of rectangle, molecular 13 of ring-in-ring structures 23 of silver(I) salts 306 of simple inorganic salts 170 of snowflakes 27 solvent effects in 20 spontaneous 61 stepwise 26, 28, 68 stereocontrolled 36, 37 subcomponent 52, 54, 407 supramolecular 26, 426 tetratopic pyridyl ligand 27 of tetrazole ligands and CuI compounds 438 TMSPs via 107 with transition metals 447 triple-anion helicate 36 unprecedented 39 water-soluble and ultra-stable Ti4 L6 tetrahedron 46 Zn2+ cations 414 self-selective ligation 22 separation processes alkane separation, with different carbon atoms 920–922 CO2 capture 882–899 global energy consumption 881 metal-organic frameworks (MOFs) 882 porous organic adsorbents 882 SHG-switching contrast 564, 565 short-chain alkanes 921, 922

Sierpinski triangle 28, 29 Σ-shaped organic spacers 294 silicon (Si) cubic lattice system 670 diamond lattice structure 671 doping mechanism 672 electrically conducting behavior 671 nanowire arrays 673, 674 unit cell 671 silicon- and germanium-containing ligands stabilized POVs 101 silicon carbide (SiC) applications 692 doping of 692–694 electronic property 692–694 materials designed for 694–695 polymorphism 692 polytypes 692, 693 silver sulfide compounds 702 silver-tin-selenide compound 231 single-chain magnets (SCMs) 778, 791, 803–804, 806 single-component systems 710 single-crystal ordered macropore (SOM-ZIF-8) 839 single-crystal X-ray diffraction (SXRD) 56, 148, 208, 415, 487, 554, 566, 614, 699, 701, 738, 860, 958 single-ion magnet (SIM) 140, 778, 794, 796–803 single-M(salen) linked MOFs 64 single-molecule cell thermometer 40 single-molecule magnetism (SMMs) 122, 140, 142, 778, 791–794, 796, 797, 802, 803 single-molecule magnets (SMMs) 122, 140–142, 149, 395, 778, 792–794 single-molecule trap (SMT) 885, 886 S-naproxen 942 SnGa4 Q7 509–510 Sn2 Ga2 S5 525–526 [Sn32.5 Ge27.5 Se132 ]24- 218 Sn-Sb-S compounds 259–260 [Sn3 Se7 ]n 2n– double-chain 223 social self-sorting 50

1025

1026

Index

soft mode theory 609, 610 solid solutions 722–725 based on benzoporphyrin derivatives 724 crystalline 722, 723 ternary-component photovoltaics 723 types of 722, 723 solid-state substances 601 solution assembly process 705 solution-derived BLSO films 678 solution-processing methods 706 solution-shearing method 706 solvothermal reactions 12, 288, 289, 315, 319, 359, 366, 415, 419 sorption isotherms activation effect 841–842 pore shape effect 840–841 space group 5, 41, 127, 200, 204, 207, 223, 228, 229, 231, 233, 247, 249, 256, 293, 317, 319, 320, 361, 417, 420, 440, 443, 466, 469–475, 477–480, 482, 484–487, 490, 491, 493–495, 499–502, 504, 506–513, 515–517, 519–526, 528, 575, 579, 583, 584, 588–590, 606, 607, 610, 614–616, 630, 638–643, 648, 651, 653, 656, 658–662, 741, 743, 745, 855, 858 spiderweb structure 22 spin-crossover cooperativity of 781, 788–791 ligand conformation flexibility inducing 784, 786 magneto-structural correlations in 779–791 metal coordination geometry inducing 781–784 molecular structure 781–786, 788 spin equilibrium behavior 779 spin multiplicities 791 spin reversal energy barrier 791, 793, 794, 797, 803 spin transition 601, 616, 617, 778–781, 783, 784, 786–790, 803

spontaneous polarization 6, 601–603, 608, 618, 621, 634 5-[(2S)-pyrrolidine-2-yl]-1H-tetrazole ((S)-HPTZ) 313 square-grid (or rhombohedral) networks bipyridine 311 bipyridine (bpy) ligands 311 bulky elongated organic ligands 313 [Cd(ImBNN)2 (NO3 )2 ]∞ 312 [Cd2 Cl2 (DP)3 ]⋅(CHCl3 )4 (DMF)4 EtOH 313 [Cd-(ImBNN)2 (CF3 SO3 )2 ]⊃ guestn (guest=C7 H8 ) 313 [Cd(4,4′ -bpy)2 ](NO3 )2n 311, 312 dicarboxylate ligand 316 dinuclear Cu2 (AcO)4 moiety 314 mixed-ligand strategy 317 mixed rigid and flexible ligands 318, 319 oxygen-containing ligands 315 Schläfli topology 311 triazine-based carboxylic ligands 316 2D layers stack 314 square-planar palladium(II) ions 35 [SrBO3 ]∞ 581 [Sr(H2 O)3 ](UO2 )2 [C6 H4 (PO3 )2 ](OH)2 (H2 O)⋅3(H2 O) (SrUbbp) 374 srs (SrSi2 ) topological network 436, 437 Sr5 ZnGa6 S15 519–520 S4 symmetric octahedron 54 stable layered double perovskites 682, 683 star lattice 821 star-shaped pentadeca-palladate 106 star-shaped supramolecular copolymers 34 Stellated polyhedra 410 stereo-chemically active lone pair (SCALP) 469, 470, 477, 495, 506 stick-and-ball stacking model of crystals 685 stimulated Raman scattering 617 stimuli-responsive polymer gels 32 string-bag units 169

Index

structural chemistry definition of 1 issues of 2 Structure Commission of International Zeolite Association (IZA-SC) 401 structure-directed precise synthesis 1 structure directing agents (SDAs) 84, 107, 116, 197, 198, 224, 225, 228, 229, 234, 235, 269, 354, 397–402, 419 styrene oxide 959, 961, 991 (4,4′ -di(substituent)oxybiphenyl-3,3,’5,5′ tetra-(phenyl-4-carboxylic acid)) 350 sulfide conducting materials 697–704 antimony trisulfide 699 copper sulfides 701 metal sulfides 697 MoS2 697–701 silver sulfide compounds 702 structure of 699–704 tin monosulfide 698 tin sulfide 698 sulfide ligands 165, 698 sulfide S2- ligand 164 sulfido ligand 164 5-sulfoisophthalic acid ligand 376 sulfonate-carboxylate ligands 376 sulfonate linkers 343, 373–377 bis(2-sulfonatostyryl) biphenyl 339 sulfurated calixarenes 136 super snowflakes complexes 26 supertetrahedra (ST) 347 supertetrahedral clusters 128, 196, 209, 226, 227, 398, 399, 401, 404, 502, 503, 507, 508 supramolecular-assisted synthetic methodology 12 supramolecular chemistry 26, 42, 405, 425, 572 supramolecular fractal 28 supramolecular hexagram 27 supramolecular hydrogels 34 supramolecular metal-organic nanoribbons (SMON) 25

supramolecular pentagram 27 supramolecular triangles, cancer cells transpotation 10, 11 surface-capping ligands 399 surface functionalization 569, 890 surface modification 132, 133 surfactants 398, 403 Suzuki–Miyaura coupling reaction 7, 958, 962 Suzuki reaction 954, 964 symmetry-breaking crystal structures 6 symmetry-breaking ferroelectrics 614–615 synchrotron neutron 923 synergistic sorbent separation technology (SSST) 913 synthetic moissanite 692 synthetic processes 1

t [TAEAH][TAEAH2]0.6 Ga2.2 Sn1.8 S8 ⋅H2 O 274 t-butylhydroperoxide (TBHP) 991 template-effect-determined POVs 96, 97 templating effects 3, 35, 36, 123, 134, 405 terephthalate (bdc) 343, 416, 726, 954, 984 terminal Mo atom 168 terminal sulfido ligand (term-S) 164 ternary conductive nitrides 686 ternary metal nitrides 690, 693 ternary zinc nitrides computational screening of 688 electronic properties of 689 L-N-tertbutoxycarbonyl-2-(imidazole)1-pyrrolidine (L-BCIP) 323 tert-butyl hydroperoxide (t-BuOOH) 956 tertiary building units (TBU) 195, 237, 241–243, 245–246, 249–253, 255–256, 276 tetrabenzoporphyrin nanofiber formation 709 tetrabutylammonium cations 377 tetracarboxylates 334, 350, 362, 417

1027

1028

Index

7,7,8,8-tetracyanoquinodimethane (TCNQ) 737 tetracyanoquinodimethane (TCNQ) complexes 704, 718, 722, 729–730, 737–741 tetrahedral phosphonate ligand 373 tetrahedron metallacages 38–46 tetrahydrothiophene 535 1,2,4,5-tetrakisphosphonomethylbenzene 373 3,3’,5,5′ -tetramethyl-4,4′ -bipyrazole (H2 Bpz) 12 N,N,N’,N’-tetramethylethylenediamine 182, 410 tetranuclear complexes 12, 71 tetranuclear metallarectangle 13 tetranuclear [Rh4 ] pyramidal frustums 14 tetraphenylborate 10 tetraphenylethylene (TPE) 17, 25, 32, 50, 406 tetrapolyoxometalate square 395 tetrathiafulvalene (TTF) derivatives 573, 630, 710, 713, 727 tetrathiomolybdate MoS4 2– 173 tetratopic carboxylate linkers 350–351 tetratopic pyridyl ligand 27 tetratopic terpyridine ligands 25 tetrazole 137, 357, 418, 438, 736–737, 870–871 thermal desorption spectroscopy (TDS) 933, 935–936, 938 thermal-driven dielectric transition 616 thermally activated delayed fluorescent (TADF) 532 thermochromic behavior, of (MV)2 [Pb7 Br18 ]n 758 thermodynamic theory antiferroelectric transition 607 Curie–Weiss Law 603 first-order and second-order types 602 first-order phase transition 603–604 Landau phase transition theory 606–607 PEP-to-FEP transition 607

phase transition in 602 second-order phase transition 604–606 thermosensitive amphiphilic hexagonal metallacycle 17 thiacalix[4]arenes 47 thiacalixarene scaffold 47 thiacalix[n]arenes 136 thioantimonates 234–236, 247, 275 thiocatecholate 364 2,5-thiophenedicarboxylic acid (H2 TDC) 47 third-generation semiconductor materials 691 3d-block single-ion magnets 801–803 3D-[Bmim]4 [Sn9 Se20 ] 219 3D-[Bmmim]2 [Ni(1,2-pda)3 266 3D-[Bmmim]4 [Sn9 Se19 (Se2 )0.9 Se0.1 ] 219 3D chiral nanoporous MOFs 946 3D-[CH3 NH3 ]2 Ag4 SnIV 2SnII S8 228 3D-[CH3 NH3 ]6 Ag12 Sn6 S21 228 3D drumlike metallacages 17 3D lattice 822, 823 3D-[(Me)2 NH2 ]0.75 [Ag1.25 SnSe3 ] 231, 270 3D-[(Me)2 NH2 ]2 GeSb2 S6 269–270 3D metal-organic assembly 18 3D-(MQ)(diamine)0.5 204 3D polyimidazolium 49 3D pseudopolyrotaxane isomer 323 3D thorium MOF 349 3D-α-ZnTe(bda)0.5 205 3D-α-ZnTe(diamine)0.5 204, 205 3D-α-ZnTe(en)0.5 200, 208 3D-α-ZnTe(hda)0.5 205 3D-α-ZnTe(hdz)0.5 208 3D-β-ZnTe(en)0.5 208 3D-ZnTe(diamine)0.5 205 three-dimensional (3D) MOFs carboxylate 726–731 ditopic 343–347 hexatopic/octatopic 351–353 tetratopic 350–351 tritopic 347–350 crystal structures of 342 cyanide (CN) 732–735

Index

definition 342 metallo-linkers 358–365 mixed N-/O-donors ligands with mixed N- and O-donor atoms 365–366 mixed carboxylate linkers 367 mixed-metal–organic framework 367–369 mixed N-donated linkers and O-donated linkers 369–373 N-heterocyclic linkers imidazolates 353–355 pyridine 357–358 nitrogen containing heterocyclic compounds 735–737 phosphonate linkers 373–375 polyoxometalate (POM) 377–380 2,3-pyrazinedithiolate (pdt) 731–732 sulfonate linkers 375–377 three-dimensional networks 8-connected three-dimensional 451, 452 two-dimensional network 3-connected 431–432 3,4-connected 433 (3,6)-connected 436 4-connected 432–435 6-connected 435–436 (3,4)-connected topological network 434, 435 three-dimensional (3-D) organic–inorganic hybrids CH3 NH3 PbI3 754–756 CH3 NH3 SnI3 757 plumbichloride hybrids 758 three-dimensional (3D) polyimidazolium cages 49 through-bond approach 726 through-space approach 726–728 ths(ThSi2 ) topological networks 438 Ti4 L6 -Co3 cage 44 Ti4 L6 -Ln cage 44 time-dependent density function theoretical (TDDFT) method 536, 556

tin monosulfide (SnS) 698 tin sulfide 698 Ti6 P2 cluster 90, 92 titania nanoparticles 85–86 titanium oxo-carboxo-alkoxides 85 titanium tetraisopropoxide 415 titanylphthalocyanine (TiOPc) 716 [TlLi(C4 H4 O6 )]⋅H2 O 649 TMC templated borates 113–115 1T MoS2 , crystal structure of 701 TMTA ligands 320–321 N-p-tolylsulfonyl-L -glutamate (NTsG) 369 (4,4) topological 2D MOFs 318 86(= (8,4)) topological network 443 topological network of metal organic framework (3,6)-connected three-dimensional 453–455 4-connected three-dimensional 86(= (8,4)) 443 cds (= CdSO4 ) 440–441 6-connected three-dimensional networks 449–451 Cu2 (imidazole)3 447 dia 440 mooganite 442–444 NbO 441 PTS 442–444 quartz 441 zeolite 443–447 12-connected three-dimensional 456–457 3-connected three-dimensional network 82 10 438 Srs(SrSi2 ) 437 ths(ThSi2 ) 438 utp((10,3)-d) 438 5-connected three-dimensional networks 449 (3,4)-connected three-dimensional topological network 448–449 (4,8)-connected three-dimensional topological networks 455–456

1029

1030

Index

82 10 topological structures 438 ToposPro 412 TPE-based dipyridyl ligand 17 TPE-derived tetraamines 50 tpst = 2,4,6-tris[(4-pyridyl)methylsulfanyl]-1,3,5-triazine 304 trans-conformation 10 transitional metal complexes (TMCs) template borates 111 transition metal carbides 691, 695–697 transition metal nitrides 683–684 transition metal-oxo clusters conventional synthesis 83 definition 81 hydro(solvo)thermal method 84 inothermal and eutectic solvent synthesis 84 one-pot self-assembly pathway 83 polyoxometalates (POMs) 81, 82 polyoxomolybdates (POMos) 104–105 polyoxoniobates (PONbs) 102–104 polyoxopalladates (POPs) 82, 105–106 polyoxotitanates (POTis) 82, 84–95 polyoxovanadates (POVs) 95–102 step-by-step approach 83 strict inert condition synthesis 83 transition-metals-substituted-POTs (TMSPs) 107, 109 transparent conducting oxides (TCOs) 673 in amorphous crystals 677 intrinsic energy gap 673 n-type crystal structures of 677 electronic structure 676–678 materials designed for 675–676 perovskite 678 p-type crystal structures 680 electronic structure 679–683 materials designed for 678 triangle-grid networks 309–310 triangular metal cluster 164, 167 triangular metallacycles 11, 450

triangular prism 3, 47–50, 56, 324, 514, 520, 549 triazine-based carboxylic ligands 316–317 triazole-pyridine-amido (L2) chelating moiety 39 1,2,3-triazoles 736 1,2,4-triazoles 736 tricarboxlate ligands 65 tricarboxylate ligands 56, 58, 330, 371, 377 tricyclohexylmethanol (TCHM) molecule 627 trideca-palladate derivatives 106 bis(tridentate) 38 tris(tridentate) 38, 40 bis(trieth-ylphosphino) 408 triglycine sulfate (TGS) 610 trigonal bipyramidal 72, 196, 237, 324 trigonal prismatic coordination cage 47, 49 triisopropylsilyl (TIPS) 65 5-((triisopropylsilyl)ethynyl) isophthalic acid (TEI) 65 trilayered polythreading 2D MOF 332 4,4’,4’’-(2,4,6-trimethylbenzene-1,3,5triyl)tribenzoic acid (H3 TMTA) 319 4,4-trimethylenedipyridine 296 trimethylphosphine 340 trinuclear butterfly species 178 trinuclear [In2 Sb8 ] 248 trinuclear [M3 (μ2 -S)3 (μ3 -S)4 ]4+ cluster units 176 trinuclear silver(I) assemblies 50 tri-nuclear zinc(II) building block 310 trinuclear zirconocene cluster 42 triphenylamine (TPA) 41, 816, 817 fragment 44 triphenyltriazine core 40 triphenyltriazine plane 42 triphosphine-supported Pt-M heteronuclear complexes 541–544 tris-CTAs metallacycles 32

Index

tritopic carboxylate linkers 347–350 tritopic ligands 25, 65, 72, 417, 855 2,4,6-tri(4-pyridinyl)-1,3,5-triazine (tpt) 372 truncated tetrahedron 46–48, 71, 99, 360, 405 T-shaped building blocks 304 tunneling effect 803 2D-[Bmmim]7 [AgSn12 Se28 ] 226 2D-[Bmmim]3 [Mn(en)3 ]2 [Sn9 Se21 ]Cl 225 2D-[Bmmim]2 [Ni(teta)(en)][Sn3 Se7 ]2 226 2D-[Bmmim]2[Sn3 Se7 ] 222 2D-[CH3 NH3 ]20 Ge10 Sb28 S72 ⋅7H2 O 274 2D GaN single crystals atomic structure 688 electronic mobility 688 growth on liquid metals 688 2D-[Ga2 Sb2 S7 ]n 2n– network 246 2D-[GeSb2 S6 ]n 2n– 255 2D-[In3 Sb2 S9 ]n 3n– layers 252 2D lattice 821, 822 2D layered MOFs 338 2D-[(Me)2 NH2 ]4/3 [(Me)3 NH]2/3 [Sn3 S7 ]⋅ 1.25H2 O 271 2D metal-organic assembly 18 2D MOFs [Zn2 M(bpdc)3 (DMF)2 ]⋅4DMF 310, 311 2D-(MQ)(monoamine) 200 2D-MQ(monomine) 208 2D-M 2 Q2 (monomine) 208 2D non-interpenetrated square-grid layers 315 2D-[Pmmim]4 [Sn17 Se38 ] 219 2D ScCx OH direct-bandgap semiconductor 695 2D-ZnSe(ba) 208 2D Zr3 C2 T2 697 2D Zr3 C2 Tz 696 two-dimensional (2D) MOFs bilayer-network 325–333 brick-wall networks 319–325 catechol and derivatives 741–747 definition 309

herringbone networks 319–325 honeycomb networks 319 nitrogen containing heterocyclic compounds 747 square-grid (or rhombohedral) networks 311–319 TCNQ and derivatives 737–741 triangle-grid networks 309 two-dimensional (2-D) organic–inorganic hybrids A2 SnI4 (A = organic ammonium) 764, 765 (BAESBT) PbI4 single crystal 765, 766 (C4 H9 NH3 )2 -(CH3 NH3 )2 Pb3 Br10 767 CH3 NH3 PbI3 762 conducting layered 761–762 displex (closed-cycle refrigeration) system 763 schematic representation 759–761 two-dimensional (2D) rhombus star-shaped supramolecules 26 two double chains MOFs 303 two-over/two-under interwoven 2D network, [HgI2 (bpmbpt)] 291 n-type transparent conducting oxides (TCOs) crystal structures of 677 and electronic structure 676–678 materials designed for 675

u UCR-20GaGeS-TAEA 400 UiO-66 crystalline structure 838 UiO-66 samples defect density 849 ultra-microporous MOFs 895, 896, 904 ultra-stable Ti4 L6 tetrahedron 44 UMCM-150 415, 416 uniform nets 427, 428 unit cell 466, 471, 474 germanium 671 silicon 671 unit construction approach 179, 191 unit construction strategy 3, 164, 167, 171–182, 191

1031

1032

Index

UNLPF-1 417 UO2 (dbsf)(phen) 297 [UO2 (dbsf) (phen)]⋅H2 O 297 UO2 Np(H2 O)2 [CH2 (PO3 )(PO3 H)]2 (UNpC1P2-1) 374 uranium 269, 272, 275, 374, 922 uranium phosphonates 374 2-ureido-4-pyrimidinone (UPy) 28 usable adsorption 7, 864–867 usable CH4 capacity 865, 866 utp((10,3)-d) topological networks 438

triphenylphosphine or halide anions 182 wave-like chain 189 W-Ag-S linear polymeric chain 182 wave-like Mo-Ag-S chain 189 white-light emission systems 40 wide-bandgap semiconductors 692 würtzite-type structure 200 Wyckoff sequence 487, 504, 520, 521

x 129

v Vagard’s law 722 valence band chemical modulation of 679, 680 valence band maximum (VBM) antibonding character of 683 valence-shell electrons 794 vanadium-oxygen clusters 56 vanadylphthalocyanine (VOPc) 716 van der Waals force 834 van der Waals interactions 198, 296, 908 vapochromic materials 18 vapochromic metallacycle 18 VIII-containing POVs 100 VMOP-α 47, 396 V-MOP-β 396 V-shaped bridging ligand dicyanamide 292

w W(Mo)-Ag-S coordination polymers anionic 188 cation-induced structural diversity of 188 cations and solvents 188 conductivities and energy gaps of 190 configuration of 182 divalent Ca(DMSO)6 2+ cation 186 semiconducting properties 190 structural variations in 188 structure-directing agent 187 structure types of 184, 186 synthetic routes 182

Xe NMR spectroscopy 923 X-ray diffraction techniques 54, 56, 148, 182, 202, 208, 227, 408, 411, 415, 420, 487, 554, 566, 614, 637, 699, 701, 738, 743, 746, 840, 860, 861, 891, 897, 923, 955, 958

y Yb6 Ga4 S15

516

z Zeoball [Sn36 Ge24 Se132 ]24– 218 zeolite sodalite-type borogermanate 118 zeolite-type topology networks 443–447 zeolitic imidazolate frameworks (ZIFs) 343, 354–355, 404, 414, 418, 444, 735, 914, 956–958 zeolitic MOFs 377 zero-dimensional (0D) cluster chalcogenides Ba3 (BQ3 )(SbQ3 ) (Q = S, Se) 470 Ba3 (BS3 )1.5 (MS3 )0.5 (M = Sb, Bi) 470 Ba23 Ga8 Sb2 S38 469 BaHgSe2 471–472 Ba12 In4 S19 466 Ba7 Sn3 S13 472–473 Ba8 Sn4 S15 473–474 Ba12 Sn4 S23 472–473 (Cs6 Cl)2 Cs5 [Ga15 Ge9 Se48 ] 470–471 zero dimensional ionic compounds 647 zero-field splitting (ZFS) 100, 147, 791, 792, 796, 802 zero/negative thermal expansion 283 zigzag type chain 287–291

Index

zigzag W-Ag-S anionic chain 188 zinc azolate system 898 zinc salen complexes 14 zirconium–phenolate frameworks 416 zirconium–porphyrin frameworks 359 Zn4 (μ4 -O)(O2 CR)6 361 [Zn(HCOO)2 (4,4′ -bipy)] 554 Zn(bac)2 (bpp)]⋅1.5H2 O (bac = 1-benzoylacetone bpp = 1,3-bis(4-pyridyl)propane) 290 [Zn(bpp)Br]2 566 [Zn(ebic)2]⋅EtOHn 288 Zn-e-Keggin cluster 378 [Zn2 (H2 O)2 (malonate)2 (Prz)]n 558 [Zn(hsb-2)(H2 O)2 ⋅G⋅2H2 O]n (G = naphthalene-2,7-disulfonate) 357 [Zn(hsb-2)(NO3 )2 ⋅2H2 O]n 315 Zn(H2 tib)(UO2 )2 (EDP)(HEDP) (H2 EDP)0.5 ⋅3H2 O 374 [Zn(ImBNN)]n 312 [Zn2 I4 (tmdp)2 ]n ⋅[Zn2 I4 (tmdp)2 ]n 295 Zn8 L6 cages 50 Zn2+ metal ion 979 Zn2 (mtba)(H2 O)2 ⋅(DMF)6 (H2 O)5 (MOF-36) 350 Zn(NO3 )2 310, 315, 438, 909 Zn(NO3 )2 ⋅6H2 O 208, 319, 352 Zn4 O(btb)2 ⋅(DEF)15 (H2 O)3 (MOF-177) 348

[Zn2 (OH)(tcpa)]⋅2DMF⋅2H2 O 349 Zn(phen)(sdc) (phen = 1,10-phenathroline; H2sdc = trans-stilbene-4,4′ -dicarboxylic acid) 290 [Zn(PR64)0.5 (m-bdc)Cl]⋅3H2 O 321 Zn4 -p-tert-butylsulfonylcalix[4]arene clusters 406 δ-ZnTe(hda)0.5 205 ZnTe and MnSe based hybrids 208 Zn-tetra-(4-carboxyphenyl)porphyrin (ZnTCPP) 316 Zn2 (TMTA)(H2 O)2 ⋅NO3 ⋅2H2 O⋅0.5DMA (β-1) 319 Zn2 (TMTA)(H2 O)2 ⋅NO3 ⋅6H2 O⋅DEF (α-1) 319 Zn8 (TPC4R)2 (Ln)4 nanocages 68 [Zn2 TTFTB(H2 O)2 ]⋅H2 O⋅2DMF 727, 728 Zr-MOFs 351, 418–420, 838–839, 859, 986 Zr6 O4 (OH)4 415 ZrSQU, structural and topological representation 850 ZTE 263, 264 zwitterionic tetrathiafulvalene (TTF)-extended dicarboxylate radical (TED) 710

1033