AI-Guided Design and Property Prediction for Zeolites and Nanoporous Materials E-Book

• Herausgeber: John Wiley & Sons
• Kategorie: Wissenschaft und neue Technologien
• Sprache: Englisch

Seitenzahl: 1079

Veröffentlichungsjahr: 2023

Beschreibung

AI-Guided Design and Property Prediction for Zeolites and Nanoporous Materials A cohesive and insightful compilation of resources explaining the latest discoveries and methods in the field of nanoporous materials In Artificial Intelligence for Zeolites and Nanoporous Materials: Design, Synthesis and Properties Prediction a team of distinguished researchers delivers a robust compilation of the latest knowledge and most recent developments in computational chemistry, synthetic chemistry, and artificial intelligence as it applies to zeolites, porous molecular materials, covalent organic frameworks and metal-organic frameworks. The book presents a common language that unifies these fields of research and advances the discovery of new nanoporous materials. The editors have included resources that describe strategies to synthesize new nanoporous materials, construct databases of materials, structure directing agents, and synthesis conditions, and explain computational methods to generate new materials. They also offer material that discusses AI and machine learning algorithms, as well as other, similar approaches to the field. Readers will also find a comprehensive approach to artificial intelligence applied to and written in the language of materials chemistry, guiding the reader through the fundamental questions on how far computer algorithms and numerical representations can drive our search of new nanoporous materials for specific applications. Designed for academic researchers and industry professionals with an interest in synthetic nanoporous materials chemistry, Artificial Intelligence for Zeolites and Nanoporous Materials: Design, Synthesis and Properties Prediction will also earn a place in the libraries of professionals working in large energy, chemical, and biochemical companies with responsibilities related to the design of new nanoporous materials.

Leseprobe

AI-Guided Design and Property Prediction for Zeolites and Nanoporous Materials

Edited by

German Sastre Instituto de Tecnología Química UPV-CSIC Universidad Politecnica de Valencia Valencia, Spain

Frits Daeyaert SynopsisDeNovoDesign Beerse, Belgium

This edition first published 2023

© 2023 John Wiley & Sons Ltd

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to obtain permission to reuse material from this title is available at http://www.wiley.com/go/permissions.

The right of German Sastre and Frits Daeyaert to be identified as the authors of the editorial material in this work has been asserted in accordance with law.

Registered Offices

John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA

John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK

Editorial Office

The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK

For details of our global editorial offices, customer services, and more information about Wiley products visit us at www.wiley.com.

Wiley also publishes its books in a variety of electronic formats and by print-on-demand. Some content that appears in standard print versions of this book may not be available in other formats.

Limit of Liability/Disclaimer of Warranty

In view of ongoing research, equipment modifications, changes in governmental regulations, and the constant flow of information relating to the use of experimental reagents, equipment, and devices, the reader is urged to review and evaluate the information provided in the package insert or instructions for each chemical, piece of equipment, reagent, or device for, among other things, any changes in the instructions or indication of usage and for added warnings and precautions. While the publisher and authors have used their best efforts in preparing this work, they make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives, written sales materials or promotional statements for this work. The fact that an organization, website, or product is referred to in this work as a citation and/or potential source of further information does not mean that the publisher and authors endorse the information or services the organization, website, or product may provide or recommendations it may make. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for your situation. You should consult with a specialist where appropriate. Further, readers should be aware that websites listed in this work may have changed or disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.

Library of Congress Cataloging-in-Publication Data

Names: Sastre, German, editor. | Daeyaert, Frits, editor.

Title: AI-guided design and property prediction for zeolites and nanoporous materials / edited by German Sastre, Frits Daeyaert.

Description: Hoboken : John Wiley and Sons Ltd, 2023. | Includes bibliographical references and index.

Identifiers: LCCN 2022042565 (print) | LCCN 2022042566 (ebook) | ISBN 9781119819752 (hardback) | ISBN 9781119819769 (pdf) | ISBN 9781119819776 (epub) | ISBN 9781119819783 (ebook)

Subjects: LCSH: Zeolites--Analysis | Nanostructured materials--Analysis | Molecules--Models | Artificial intelligence

Classification: LCC QE391.Z5 A44 2023 (print) | LCC QE391.Z5 (ebook) | DDC 549/.68--dc23/eng20221121

LC record available at https://lccn.loc.gov/2022042565

LC ebook record available at https://lccn.loc.gov/2022042566

Cover design by Wiley

Cover Images: Binary Code: © fotograzia/Getty Images; Lightning: © John Sirlin/Getty Images; Zeolite image: Courtesy of Frits Daeyaert; Server: © cybrain/Getty Images; Parthenon: © Cliff Wassmann/Getty Images

Set in 9.5/12.5pt STIXTwoText by Integra Software Services Pvt. Ltd, Pondicherry, India

Contents

Cover

Title page

Copyright

List of Contributors

Preface

About the Cover

Acknowledgments

1 The Confluence of Organo-Cations, Inorganic Species, and Molecular Modeling on the Discovery of New Zeolite Structures and Compositions

1.1 Introduction

1.2 Inorganic Studies

1.3 Organic Structure-Directing Agents (OSDAs)

1.3.1 Purpose and Important Properties

1.3.2 Classes of Ammonium-based OSDAs

1.3.3 Methods of Making

1.4 OSDA–Zeolite Energetics and Rational Synthesis

1.5 Role of High Throughput and Automation

1.6 Cataloguing, Archiving, Harvesting, and Mining Years of Historical Data

1.7 Concluding Remarks

References

2 De Novo Design of Organic Structure Directing Agents for the Synthesis of Zeolites

2.1 Introduction

2.2 De Novo Design

2.2.1 Molecular Structure Generator

2.2.2 Scoring Function

2.2.3 Optimization Algorithm

2.2.4 Practical Implementation

2.3 Scoring Functions for OSDAs

2.3.1 Stabilization Energy

2.3.2 Other Constraints

2.3.3 Multiple Objectives

2.4 Applications

2.4.1 From Drug Design to the Design of OSDAs for Zeolites

2.4.2 Experimental Confirmation: Pure Silica STW

2.4.3 Experimental Confirmation: Zeolite AEI

2.4.4 Practical Application: SSZ-52 (SFW)

2.4.5 Design of Chiral OSDAs to Direct the Synthesis of Chiral STW

2.4.6 Design of Selective OSDAs Directed Toward BEA vs. BEB

2.4.7 Design of OSDAs for Chiral Zeolite BEA

2.4.8 Application of a Machine-Learning Scoring Function in the De Novo Design of OSDAs for Zeolite Beta

2.4.9 Design of OSDAs for Zeolites for Gas Adsorption and Separation

2.4.9.1 Carbon Capture and Storage: WEI, JBW, GIS, SIV, DAC, 8124767, 8277563

2.4.9.2 Carbon Dioxide/Methane Separation: GIS, ABW, 8186909, 8198030

2.4.9.3 Separation of Ethylene-Ethane: DFT, ACO, NAT, JRY

2.4.10 Design of MOFs for Methane Storage and Delivery

2.4.11 Multi-Objective De Novo Design of OSDAs for Zeolites Using an Ant Colony Optimization Algorithm

2.5 Conclusions and Outlook

References

3 Machine Learning Search for Suitable Structure Directing Agents for the Synthesis of Beta (BEA) Zeolite Using Molecular Topology and Monte Carlo Techniques

3.1 Introduction

3.2 Artificial Neural Networks for Modeling Zeolite-SDA van der Waals Energy Applied to BEA Zeolite

3.3 Virtual Screening: Identifying Novel SDA with Favorable EZEO-SDA for the Synthesis of BEA Zeolite

3.4 Zeo-SDA Energy Calculation Using Atomic Models

3.5 Comparing Zeo-SDA Energy Calculation Using MLR, ANN, and Atomic Models

3.6 Conclusions

Acknowledgments

References

4 Generating, Managing, and Mining Big Data in Zeolite Simulations

4.1 Introduction

4.1.1 Computational Materials Databases

4.1.2 Zeolite Databases

4.2 Database of OSDAs for Zeolites

4.2.1 Developing a Docking Algorithm

4.2.2 Calibrating Binding Energy Predictions

4.2.3 Performing and Analyzing High-Throughput Screening Calculations

4.2.4 Recalling Synthesis Outcomes from the Literature

4.2.5 Proposing OSDA Descriptors

4.2.6 Designing with Interactivity

4.3 Outlook

References

5 Co-templating in the Designed Synthesis of Small-pore Zeolite Catalysts

5.1 Introduction

5.1.1 Definitions: Templates and Structure Directing Agents; Co-templating; Dual Templating; Mixed Templating

5.2 SAPO Zeotypes: “Model” Systems for Co-templating

5.2.1 The CHA-AEI-SAV-KFI System

5.2.2 Development of a Retrosynthetic Co-templating Approach for ABC-6 Structure Types

5.3 Co-templating Aluminosilicate Zeolites

5.3.1 Inorganic/Organic Co-templates

5.3.1.1 Targeting new phases in the RHO family using divalent cations

5.3.1.2 Designed synthesis of the aluminosilicate SWY, STA-30

5.3.1.3 Co-templating and the charge density mismatch approach

5.3.2 Two Organic Templates in Zeolite Synthesis

5.3.2.1 Applications of Dual/Mixed Organic Templating

5.4 Intergrowth Zeolite Structures as Co-templated Materials

5.5 Discussion

5.6 Conclusions

Acknowledgments

References

6 Computer Generation of Hypothetical Zeolites

6.1 Introduction

6.2 Genetic Algorithms

6.2.1 Codification of Genetic Algorithms

6.2.2 Selection Operators for Genetic Algorithms

6.2.3 Crossover Operators for Genetic Algorithms

6.2.4 Mutation Operators for Genetic Algorithms

6.3 Algorithms for Zeolite Structure Determination and Prediction

6.3.1 ZEFSAII

6.3.2 FraGen (Framework Generator)

6.3.3 SCIBS (Symmetry-Constrained Intersite Bonding Search)

6.3.4 TTL GRINSP (Geometrically Restrained Inorganic Structure Prediction)

6.3.5 EZs (Exclusive Zones)

6.3.6 P-GHAZ (Parallel Genetic Hybrid Algorithm for Zeolites)

6.3.7 zeoGAsolver

6.4 zeoGAsolver: A Specific Example of Genetic Algorithm for ZSD

6.4.1 Setting Up and Coding Scheme

6.4.2 Initialization

6.4.3 Fitness Evaluation

6.4.4 Crossover

6.4.5 Population Reduction and Termination Criterion

6.5 Graphics Processing Units in Zeolite Structure Determination and Prediction

6.5.1 Quick Presentation of GPU Cards

6.5.2 Efficient Parallelization of Evolutionary Algorithms on GPUs

6.5.3 Genetic Algorithms on GPUs for Zeolite Structures Problem

6.5.4 GPUs in Island Model for Interrupted Zeolitic Frameworks

6.6 Conclusions

Acknowledgments

References

7 Numerical Representations of Chemical Data for Structure-Based Machine Learning

7.1 Machine Readable Data Formats

7.1.1 Feature Vectors

7.1.2 Matrices

7.1.3 Mathematical Graphs

7.2 Graph-based Molecular Representations

7.2.1 Chemical Representations of Molecular Structures

7.2.2 Molecular Graphs

7.2.3 XYZ File to Molecular Graph

7.2.4 SMILES to Molecular Graph

7.2.5 Multiple Molecular Graph

7.3 Machine Learning with Molecular Graphs

7.3.1 General Architecture of Graph Neural Networks

7.3.2 Graph Convolutional Network

7.3.3 Graph Attention Network

7.3.4 Continuous Kernel-based Convolutional Network

7.3.5 Crystal Graph Convolutional Neural Network

7.4 Graph-based Machine Learning for Molecular Interactions

7.4.1 Vector Concatenation Approach to Prediction Molecule-to-Molecule Interactions

7.4.2 Attention Map Approach for Interpretable Prediction of Molecule-to-Molecule Interactions

7.5 Representation Learning from Molecular Graphs

7.5.1 Unsupervised Representation Learning

7.5.2 Supervised Representation Learning

7.6 Python Implementations

7.6.1 Data Conversion: Molecular Structures to Molecular Graphs

7.6.2 Machine Learning: Deep Learning Frameworks for Graph Neural Networks

7.6.3 Pymatgen for Crystal Structures

7.7 Graph-based Machine Learning for Chemical Applications

7.7.1 Message Passing Neural Network to Predict Physical Properties of Molecules

7.7.2 Scale-Aware Prediction of Molecular Properties

7.7.3 Prediction of Optimal Properties From Chromophore-Solvent Interactions

7.7.4 Drug Discovery with Reinforcement Learning

7.7.5 Graph Neural Networks for Crystal Structures

7.8 Conclusion

References

8 Extracting Metal-Organic Frameworks Data from the Cambridge Structural Database

8.1 Introduction

8.2 Building the CSD MOF Subset

8.2.1 What Is a MOF?

8.2.2 ConQuest

8.3 The CSD MOF Subset

8.3.1 Removing Solvents With the CSD Python API

8.3.2 Adding Missing Hydrogens

8.4 Textural Properties of MOFs and Their Evolution

8.5 Classification of MOFs

8.5.1 Identification of Target MOF Families

8.5.2 Identification of Surface Functionalities in MOFs

8.5.3 Identification of Chiral MOFs

8.5.4 Porous Network Connectivity and Framework Dimensionality

8.5.5 An Insight into Crystal Quality of Different MOF Families

8.6 The CSD MOF Subset Among All the MOF Databases

8.7 Conclusions

Acknowledgments

References

9 Data-Driven Approach for Rational Synthesis of Zeolites and Other Nanoporous Materials

9.1 Introduction

9.2 Rationalization of the Synthesis–Structure Relationship in Zeolite Synthesis: Application Machine Learning and Graph Theory to Zeolite Synthesis

9.3 Extraction of the Structure–Property Relationship in Nanoporous Nitrogen-Doped Carbons: Dealing with the Missing Values in Literature Data

9.4 Acceleration of Experimental Exploration of Nanoporous Metal Alloys: An Active Learning Approach

9.5 Summary

Acknowledgments

References

10 Porous Molecular Materials: Exploring Structure and Property Space with Software and Artificial Intelligence

10.1 Introduction

10.2 Computational Modeling of Porous Molecular Materials

10.2.1 Structure Prediction

10.2.2 Modeling Porosity

10.2.3 Amorphous and Liquid Phase Simulations

10.3 Data-Driven Discovery: Applying Artificial Intelligence Methods to Materials Discovery

10.3.1 Training Data Generation

10.3.1.1 Hypothetical Structure Datasets

10.3.1.2 Experimental Structure Datasets

10.3.1.3 Extraction of Data From Scientific Literature

10.3.1.4 Data Augmentation and Transfer Learning

10.3.2 Descriptor Construction and Selection

10.3.2.1 Local Environment Descriptors

10.3.2.2 Global Environment Descriptors

10.4 Efficient Traversal of the Chemical Space of Porous Materials

10.4.1 Evolutionary Algorithms

10.4.2 Reducing the Number of Experiments: Bayesian Optimization and Active Learning

10.4.3 Chemical Space Exploration with Deep Learning

10.5 Considering Synthetic Accessibility

10.6 Closing the Loop: How Can High-Throughput Experimentation Feed Back into Computation?

10.6.1 High-Throughput and Autonomous Experimentation

10.7 Conclusions

References

11 Machine Learning-Aided Discovery of Nanoporous Materials for Energy- and Environmental-Related Applications

11.1 Introduction

11.1.1 Nanoporous Materials

11.1.2 History and Development

11.1.3 Gas Separation and Storage Applications

11.1.4 Large-Scale Computational Screening for Gas Separation and Storage

11.2 Concepts and Background for Data-Driven Approaches

11.2.1 Dimensionality Reduction

11.2.2 Machine Learning Models

11.2.2.1 Linear Models

11.2.2.2 Decision Trees and Random Forests

11.2.2.3 Support Vector Machine

11.2.2.4 Neural Networks

11.2.2.5 Unsupervised Learning

11.3 Data-Driven Approaches

11.3.1 Nanoporous Structure Datasets

11.3.2 Identifying Feature Space of Materials to Screen

11.3.3 Methods to Search for Optimal Structures

11.3.4 Modeling Interatomic and Intermolecular Interactions

11.4 Case Studies

11.4.1 Post-Combustion CO2 Capture

11.4.2 Methane Storage

11.4.3 Hydrogen Storage

11.5 Summary and Outlook

References

12 Big Data Science in Nanoporous Materials: Datasets and Descriptors

12.1 Introduction

12.2 Repositories of Nanoporous Material Structures

12.2.1 Experimental Crystal Structures

12.2.2 Predicted Crystal Structures

12.3 Descriptors

12.3.1 Handcrafted Descriptors

12.3.2 Toward Automatically Generated and Multi-Scale Descriptors

12.4 Properties

12.5 Data Analysis

12.5.1 Material Similarity and Distance Measures

12.5.1.1 Diversity Selection

12.5.1.2 Cluster Analysis

12.6 Machine Learning Models of Structure–Property Relationships

12.7 Current and Future Applications

References

13 Efficient Data Utilization in Training Machine Learning Models for Nanoporous Materials Screening

13.1 Descriptor Selection

13.1.1 Engineering of Advanced Features

13.1.2 Engineering of Simpler Features

13.2 Material Selection

13.3 Model Selection

13.3.1 Linear Regression

13.3.2 Supported Vector Regressors

13.3.3 Decision Tree-based Regressors

13.3.4 Artificial Neural Networks

13.4 Data Usage Strategies

13.4.1 Transfer Learning

13.4.2 Multipurpose Models

13.4.3 Material Recommendation Systems

13.4.4 Active Learning

13.4.5 Machine Learning to Speed Up Data Generation

13.5 Summary and Outlook

References

14 Machine Learning and Digital Manufacturing Approaches for Solid-State Materials Development

14.1 Introduction

14.2 The Development of MOF Databases

14.3 Natural Language Processing

14.4 An Overview of Machine Learning Models

14.5 Machine Learning for Synthesis and Investigation of Solid State Materials

14.6 Machine Learning in Design and Discovery of MOFs

14.7 Current Limitations of Machine Learning for MOFs

14.8 Automated Synthesis and Digital Manufacturing

14.9 Digital Manufacturing of MOFs

14.10 The Future of Digital Manufacturing

References

15 Overview of AI in the Understanding and Design of Nanoporous Materials

15.1 Introduction

15.2 Databases

15.2.1 Structural Databases

15.2.2 Databases of Material Properties

15.2.3 Databases of Synthesis Protocols

15.3 Big-Data Science for Nanoporous Materials Design and Discovery

15.3.1 Representations of Chemical Data

15.3.2 Learning Algorithms

15.4 Applications

15.5 Zeolite Synthesis and OSDAs

15.6 Conclusion

References

Index

End User License Agreement

List of Tables

CHAPTER 01

Table 1.1 A non-exhaustive listing...

Table 1.2 Zeolite timeline. Source: Adapted...

Table 1.3 Table of discoveries, cations...

Table 1.4 Examples of rigid polycyclic...

Table 1.5 SSZ materials that use...

Table 1.6 SSZ materials that use...

Table 1.7 Calculated vdW interaction energies...

CHAPTER 02

Table 2.1 Two reactions available to...

Table 2.2 Score types that can...

Table 2.3 Mutation and combination genetic...

Table 2.4 Molecular properties and constraints...

Table 2.5 Overview of de novo...

Table 2.6 Components of the score...

CHAPTER 03

Table 3.1 Summary of predictive models...

Table 3.2 Features of ANN build...

Table 3.3 Topo-chemical descriptors used...

Table 3.4 Parameters resulting of validation...

Table 3.5 EBEA_SDA (kJ/mol Si...

Table 3.6 Zeo-OSDA van der...

Table 3.7 Zeo-OSDA van der...

CHAPTER 05

Table 5.1 Definitions of commonly used...

Table 5.2 Binding energies of the...

Table 5.3 Calculated interaction energies of...

Table 5.4 OSDAs used in JMZ...

Table 5.5 Probability matrix for Reichweite...

Table 5.6 Binding energies calculated using...

Table 5.7 Materials synthesized through co...

Table 5.8 Non-bonded energies for...

Table 5.9 Stabilization energy of template...

CHAPTER 06

Table 6.1 General comparison of computer...

Table 6.2 Comparison of GA-based...

Table 6.3 Configurations of unit cells...

Table 6.4 Target zeolites evaluated by...

CHAPTER 08

Table 8.1 Summary of the number...

CHAPTER 09

Table 9.1 Test accuracy of machine...

Table 9.2 Input and output features...

CHAPTER 12

Table 12.1 The most commonly used...

Table 12.2 List of possible linkage...

List of Illustrations

CHAPTER 01

Figure 1.1 Zeolites made from...

Figure 1.2 The graphic shows...

Figure 1.3 Magadiite/mordenite diagrams...

Figure 1.4 Four different zeolite...

Figure 1.5 OSDA and structure...

Figure 1.6 The arrangement of...

Figure 1.7 General formula for...

Figure 1.8 Predicted energy-minimized...

Figure 1.9 The PXRD pattern...

Figure 1.10 Predicted energy-minimized...

Figure 1.11 Web of Science...

Scheme 1.1 Two OSDAs made using...

Scheme 1.2 An example of using...

Scheme 1.3 An example of using...

Scheme 1.4 An example of using...

CHAPTER 02

Figure 2.1 Example of a...

Figure 2.2 Example of a...

Figure 2.3 Evolutionary cycle of...

Figure 2.4 Schematic illustration of...

Figure 2.5 Schematic illustration of...

Figure 2.6 Schematic illustration of...

Figure 2.7 Schematic illustration of...

Figure 2.8 Schematic illustration of...

Figure 2.9 Schematic illustration of...

Figure 2.10 Raw output as...

Figure 2.11 First three Pareto...

Figure 2.12 First three Pareto...

Figure 2.13 Monoquaternary imidazolium and...

CHAPTER 03

Figure 3.1 Types of layers...

Figure 3.2 Example of transforming...

Figure 3.3 R2e descriptor values...

Figure 3.4 Layout of neural...

Figure 3.5 Scatterplots of the...

Figure 3.6 Potential OSDAs for...

Figure 3.7 Channels (blue shaded...

Figure 3.8 Projection along [100...

Figure 3.9 Projection along [010...

Figure 3.10 Two views of...

CHAPTER 04

Figure 4.1 Naming convention we...

Figure 4.2 (a) Example of...

Figure 4.3 (a) Schematic on...

Figure 4.4 Simplified diagram of...

Figure 4.5 Schematic of binding...

Figure 4.6 Schematic of literature...

Figure 4.7 Different descriptors used...

Figure 4.8 (a) Relationship between...

Figure 4.9 Dependence of an...

Figure 4.10 Graphical interface of...

CHAPTER 05

Figure 5.1 Number of frameworks...

Figure 5.2 Left: the orientation...

Figure 5.3 Modeled positions of...

Figure 5.4 Sketches of the...

Figure 5.5 (Above) OSDAs selected...

Figure 5.6 The lowest-energy...

Figure 5.7 Framework structures of...

Figure 5.8 Energy-minimized location...

Figure 5.9 (Left) The relationship...

Figure 5.10 The fit of...

Figure 5.11 The fit of...

Figure 5.12 The left panel...

Figure 5.13 Example of JMZ...

Figure 5.14 Distribution of cavity...

Figure 5.15 Types of cavities...

CHAPTER 06

Figure 6.1 Evolutionary flow of...

Figure 6.2 Application of the...

Figure 6.3 Example of basic...

Figure 6.4 Zeolite Structure Determination...

Figure 6.5 Combinatorial explosion of...

Figure 6.6 Layered zeolite structure...

Figure 6.7 (a) Angle and...

Figure 6.8 Example of application...

Figure 6.9 (a) CPU (top...

Figure 6.10 (a) CPU (top...

Figure 6.11 Heterogeneity of machine...

Figure 6.12 (a) New [4664...

Figure 6.13 (a) Interrupted framework...

CHAPTER 07

Figure 7.1 Examples of the...

Figure 7.2 Image data of...

Figure 7.3 XYZ file format...

Figure 7.4 The overall process...

Figure 7.5 The general architecture...

Figure 7.6 The architecture of...

Figure 7.7 The architecture of...

Figure 7.8 t-SNE visualization...

Figure 7.9 Overview of the...

Figure 7.10 Selected elemental attributes...

CHAPTER 08

Figure 8.1 Metal-organic frameworks...

Figure 8.2 Evolution of the...

Figure 8.3 Example of two...

Figure 8.4 Summary of the...

Figure 8.5 Evolution of Criterion...

Figure 8.6 Schematic representation of...

Figure 8.7 Organization of the...

Figure 8.8 Example of a...

Figure 8.9 Schematic 2D illustration...

Figure 8.10 Histograms comparing geometric...

Figure 8.11 Criteria developed for...

Figure 8.12 (a to d...

Figure 8.13 (a to d...

Figure 8.14 Histograms showing the...

Figure 8.15 Criteria developed to...

Figure 8.16 Histograms showing geometric...

Figure 8.17 Histograms of the...

Figure 8.18 Schematic 2D illustration...

Figure 8.19 Analysis of metal...

CHAPTER 09

Figure 9.1 A computational workflow...

Figure 9.2 A decision tree...

Figure 9.3 A similarity network...

Figure 9.4 Correlations between the...

Figure 9.5 (a) Importance of...

Figure 9.6 An active learning...

Figure 9.7 (a) Ternary diagrams...

CHAPTER 10

Figure 10.1 Structures of molecules...

Figure 10.2 The topological and...

Figure 10.3 The qualitative computational...

Figure 10.4 Energy-structure-function...

Figure 10.5 CC2: (a) formed...

Figure 10.6 A closed-loop...

Figure 10.7 An overview of...

Figure 10.8 An overview of...

Figure 10.9 The workflow used...

CHAPTER 11

Figure 11.1 The number of...

Figure 11.2 Database dependence of...

Figure 11.3 Schematic illustration of...

Figure 11.4 Workflow of one...

Figure 11.5 Bayesian optimization approach...

Figure 11.6 mCBAC method. (a...

Figure 11.7 MPNN illustration. (a...

Figure 11.8 Relationship between Henry...

Figure 11.9 Persistent homology illustration...

Figure 11.10 SVM-predicted vs...

Figure 11.11 (a) SVM- and...

Figure 11.12 (a) Illustration of...

Figure 11.13 Barcodes in zeolite...

Figure 11.14 Deliverable energy of...

Figure 11.15 Illustration of LASSO...

Figure 11.16 Comparison of different...

CHAPTER 12

Figure 12.1 Example using rhr...

Figure 12.2 Building blocks for...

Figure 12.3 (A) An example...

Figure 12.4 A single decision...

CHAPTER 13

Figure 13.1 Example of an...

Figure 13.2 (a) Example Voronoi...

Figure 13.3 Left: a thiol...

Figure 13.4 Various distributions of...

Figure 13.5 (a) Example of...

Figure 13.6 (a) Parity plots...

Figure 13.7 (a) Visualization of...

Figure 13.8 (a) Schematics of...

Figure 13.9 (a) Schematic of...

Figure 13.10 Illustration of transfer...

Figure 13.11 (a) Methods to...

Figure 13.12 Illustration of 23...

Figure 13.13 (a) Parity plots...

Figure 13.14 Parity plots comparing...

Figure 13.15 In a material...

Figure 13.16 Evolution of GP...

Figure 13.17 Message passing neural...

CHAPTER 14

Figure 14.1 Schematic showing the...

Figure 14.2 A flow diagram...

Figure 14.3 A flow diagram...

Figure 14.4 A Web of...

Figure 14.5 Input topologies of...

Figure 14.6 A comparison of...

Figure 14.7 A schematic showing...

Figure 14.8 A universal system...

Figure 14.9 Skeletal structures of...

Figure 14.10 Schematic of the...

Figure 14.11 Process flow diagram...

Figure 14.12 Schematic of the...

Figure 14.13 Set-up for...

Figure 14.14 Process flow diagram...

Guide

Cover

Title page

Copyright

Table of Contents

List of Contributors

Preface

About the Cover

Acknowledgments

Begin Reading

Index

End User License Agreement

Pages

i

ii

iii

iv

v

vi

vii

viii

ix

x

xi

xii

xiii

xiv

xv

xvi

xvii

xviii

xix

xx

xxi

xxii

xxiii

xxiv

xxv

xxvi

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

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

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

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

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

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

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

400

401

402

403

404

405

406

407

408

409

410

411

412

413

414

415

416

417

418

419

420

421

422

423

424

425

426

427

428

429

430

431

List of Contributors

Estefania ArgenteValencian Research Institute for Artificial Intelligence (VRAIN), Universitat Politècnica de València, Valencia, Spain

Laurent BaumesExxonMobil, Clinton, New Jersey, USA

Steven BennettDepartment of Chemistry, Molecular Sciences Research Hub, Imperial College London, White City Campus, London, United Kingdom

Rocio Bueno-PerezAdsorption & Advanced Materials Laboratory (AAML), Department of Chemical Engineering & Biotechnology, University of Cambridge, Cambridge, United Kingdom

Watcharop ChaikittisilpResearch and Services Division of Materials Data and Integrated System (MaDIS), National Institute for Materials Science (NIMS), Tsukuba, Ibaraki, Japan

Ruxandra G. Chitac EaStCHEM School of Chemistry, University of St Andrews, St Andrews, United Kingdom

Yamil J. ColonDepartment of Chemical and Biomolecular Engineering, University of Notre Dame, Notre Dame, Indiana, USA

Paul A. CoxSchool of Pharmacy and Biomedical Sciences, University of Portsmouth, Portsmouth, United Kingdom

Frits DaeyaertSynopsisDeNovoDesign, Beerse, Belgium

Archit DatarWilliam G. Lowrie Department of Chemical and Biomolecular Engineering, The Ohio State University, Columbus, Ohio, USA

Tracy M. DavisChevron Technical Center, Richmond, California, USA

Michael DeemCertus LLC, Houston, Texas, USA

Giulia Lo DicoIMDEA Materials Institute, Madrid, Spain Tolsa Group, Madrid, Spain

Saleh ElomariChevron Technical Center, Richmond, California, USA

David Fairen-JimenezAdsorption & Advanced Materials Laboratory (AAML), Department of Chemical Engineering & Biotechnology, University of Cambridge, Cambridge, United Kingdom

Lawson. T. GlasbyDepartment of Chemical and Biological Engineering, University of Sheffield, Sheffield, United Kingdom

Rafael Gómez-BombarelliDepartment of Materials Science and Engineering, Massachusetts Institute of Technology, Massachusetts, USA

Diego A. Gómez-GualdrónChemical and Biological Engineering, Colorado School of Mines, Colorado, USA

Maciej HaranczykIMDEA Materials Institute, Madrid, Spain

Kim E. JelfsDepartment of Chemistry, Molecular Sciences Research Hub, Imperial College London, White City Campus, London, United Kingdom

Christopher M. LewChevron Technical Center, Richmond, California, USA

Aurelia LiAdsorption & Advanced Materials Laboratory (AAML), Department of Chemical Engineering & Biotechnology, University of Cambridge, Cambridge, United Kingdom

Li-Chiang LinWilliam G. Lowrie Department of Chemical and Biomolecular Engineering, The Ohio State University, Columbus, Ohio, USA

Department of Chemical Engineering, National Taiwan University, Taipei, Taiwan

María Gálvez-Llompart Molecular Topology and Drug Design Unit, Department of Physical Chemistry, University of Valencia, Valencia, Spain

Instituto de Tecnologia Quimica (UPV-CSIC), Universidad Politecnica de Valencia, Valencia, Spain

Qiang Lyu School of Materials Science and Engineering, China University of Petroleum (East China), Qingdao, Shandong, China

James MattockSchool of Pharmacy and Biomedical Sciences, University of Portsmouth, Portsmouth, United Kingdom

Alechania MisturiniInstituto de Tecnología Química UPV-CSIC; Universitat Politècnica de València, Valencia, Spain

Peyman Z. MoghadamDepartment of Chemical and Biological Engineering, University of Sheffield, Sheffield, United Kingdom

Department of Chemical Engineering, University College London (UCL), London, United Kingdom

Seyed Mohamad Moosavi Department of Mathematics and Computer Science/Mathematics, Artificial Intelligence for the Sciences, Freie Universität Berlin, Berlin, Germany

Gyoung S. NaKorea Research Institute of Chemical Technology (KRICT). Daejeon, Republic of Korea

German SastreInstituto de Tecnologia Quimica (UPV-CSIC), Universidad Politecnica de Valencia, Valencia, Spain

Joel E. Schmidt Chevron Technical Center, Richmond, California, USA

Daniel Schwalbe-KodaDepartment of Materials Science and Engineering, Massachusetts Institute of Technology, Massachusetts, USA

Mervyn D. ShannonJohnson Matthey Technology Centre, Billingham, United Kingdom

Cory M. SimonSchool of Chemical, Biological, and Environmental Engineering, Oregon State University, Oregon, USA

Michael M.J. TreacyDepartment of Physics, Arizona State University, Tempe, Arizona, USA

Alessandro TurrinaJohnson Matthey Technology Centre, Billingham, United Kingdom

Soledad ValeroValencian Research Institute for Artificial Intelligence (VRAIN), Universitat Politècnica de València, Valencia, Spain

Emily H. WhaitesDepartment of Chemical and Biological Engineering, University of Sheffield, Sheffield, United Kingdom

Paul A. WrightEaStCHEM School of Chemistry, University of St Andrews, St Andrews, United Kingdom

Dan XieChevron Technical Center, Richmond, California, USA

Stacey I. ZonesChevron Technical Center, Richmond, California, USA

Preface

There will never be a better video game than Pac-Man.1 Well, yes, this is a matter of opinion, but it suggests a level of creativity that might never be reached by AI (Artificial Intelligence). But, if AI could ever reach a similar level, would we be happy with an AI equaling human’s creativity? Limits to AI are necessary. AI is becoming more and more integrated into an ever-increasing number of aspects of human life and ethical concerns must be discussed along with or even prior to these developments. The ethics applied to AI will mirror our own ethical beliefs, hence concerns about AI are also concerns about ourselves.2 Justice and transparency are perennial needs of our society and will be implemented in future developments of AI.3 But power also attracts particular interests, with the risk of a globally biased AI.4

With our task in this book being rather modest with respect to the big trends outlined above, it is nevertheless our duty not to produce useless data as well as to follow recommendations and protocols that will make information and results understandable, reproducible, and widely available.5 AI has made an enormous impact recently in the field of materials chemistry, and in particular nanoporous materials. Books may not contribute actively to impact, but they do provide an invaluable service to the scientific community by presenting a more careful approach to the foundations of our knowledge. This is our task with this book, with the help of the invaluable commitment of all the authors, who analyze how AI in its various flavors is pervading our way to do research and, hopefully, gain knowledge.

1  Molecular Modeling and New Zeolite Structures

The book opens with an account of the expertise collected by Stacey Zones and coworkers at the Chevron zeolite group in the development of new zeolites and the role therein of OSDAs (organic structure directing agents) in combination with other synthesis parameters. For more than 40 years this world leading group have been pioneers which, from the industry, have played an outstanding role in the elucidation and interpretation of the mechanisms of zeolite synthesis. The chapter describes in particular how molecular modeling techniques have provided insights into the working mechanism of OSDAs, and how these insights are becoming instrumental in the search for new zeolites. The role of high-throughput and automation of collecting, validating, and analyzing data in order to further develop machine learning and AI tools is discussed. The ultimate goal is to gain a deeper insight into the zeolite synthesis, as well as to use these tools to open up the large structural space of hypothetical zeolites, paving the way for future prediction of on-demand synthesis of zeolite materials for target applications.

2  De Novo Design of OSDAs

Given the progress in molecular modeling and machine learning techniques to accurately model the templating effect of OSDAs (organic structure directing agents), the second chapter by Frits Daeyaert and Michael Deem describes de novo design of OSDAs as an effective approach to exploit this ability. De novo design originated in drug design as an automated tool to generate molecules interacting with a biological target. In the zeolite-OSDA setting, the aim is to design molecules that closely interact with a target zeolite structure. It is discussed how issues like the synthesizability of the designed compounds and the simultaneous optimization of multiple additional properties and constraints of OSDAs can be addressed. The application to several OSDA design projects of a de novo algorithm developed by the authors is described.

3  ML Search for SDAs Using Molecular Topology and MC Techniques

In Chapter 3, María Gálvez-Llompart and German Sastre present the application of 3D-QSAR algorithms to the modeling and design of SDAs for zeolite synthesis. 3D-QSAR algorithms have proven their merit in drug design. They rely on the principle that the activity of a drug molecule relies on its binding energy with a biological target, and that this binding energy can be modeled using a combination of well-chosen molecular descriptors and appropriate machine learning algorithms. This situation resembles zeolite synthesis with the aid of SDAs, where the interaction energy between an SDA and its target zeolite is a determining factor in the outcome of a zeolite synthesis. As in drug design, accurate prediction of this interaction energy using atomistic modeling is very computationally expensive. It is shown how the use of molecular topology descriptors in combination with appropriate statistical and machine learning techniques allows an orders of magnitude speed-up in predicting zeolite–SDA interaction energies. This in turn enables virtual screening of large numbers of available molecules for use as OSDAs for zeolite synthesis. The method is demonstrated for BEA zeolite and, importantly, uses carefully selected experimental data.

4  Big Data in Zeolite Simulations

Application of data science methods to zeolite discovery has been hindered by the diversity of synthesis routes and conditions and by the absence of machine-readable experimental data. In Chapter 4, Daniel Schwalbe-Koda and Rafael Gómez-Bombarelli discuss how combining simulations and data-driven methods lead to the understanding and prediction of the role of OSDAs in zeolite synthesis. The concept of the FAIR, findable–accessible–interoperable–reusable, principle in the context of computational material databases is discussed. An overview is given of existing zeolite databases of experimental and predicted structures, and of efforts to create curated datasets of zeolite syntheses based upon the FAIR principle. The creation of a web-based platform to interactively explore OSDA-templated synthesis routes by simultaneously mining both experimental and computational data is described. Details of how the computational data have been compiled are discussed. These include efficient algorithms for OSDA-zeolite docking and binding energy calculations, the set-up and analysis of high-throughput computational screening, and consideration of additional OSDA descriptors such as synthetic accessibility, volume, and shape and charge descriptors.

5  Co-Templating for Small-Pore Zeolite Catalysts

Small-pore zeolites, which are of importance for their use in catalytic reactions and selective adsorption of small molecules, often contain more than one type of cavity. Thus, improved synthesis of these materials might be envisaged by the use of more than one template. This is discussed in Chapter 5 by Alessandro Turrina and coworkers forming a solid cluster of collaborators, including several UK universities and the leadership of Johnson Matthey who has worked on zeolite synthesis for more than three decades since the pioneering work by Paul Wright, John Casci, and Paul Cox. Synthesis of nanoporous solids using co-templating, aided by computational simulations, has been described for SAPO zeotypes. For Al-Si zeolites, the situation is complicated by the presence of metal cations that interact with both the crystallizing frameworks and OSDAs that are present in reaction mixtures. In the case of mixed inorganic/organic templating, it is discussed how computational modeling is applied in both synthesis design and structure elucidation. Co-templating has been successfully exploited in the synthesis and structural control of disordered intergrowth zeolite structures containing more than one cavity type.

6  Computer Generation of Hypothetical Zeolites

In Chapter 6, German Sastre and coworkers discuss the algorithmically closely related fields of computer generation of hypothetical zeolite structure prediction (ZSP) and determination (ZSD). Both ZSP and ZSD present complex, combinatorial problems, and one approach to solve these is the use of genetic algorithms. Thus, a separate paragraph is devoted to this subject. An overview of ZSP and ZSD algorithms and their software implementation is given and one particular method, the zeoGAsolver program, is discussed in detail. ZSP and ZSD are very computationally expensive, and therefore can benefit from the advent of high-speed GPUs, requiring efficient parallelization of existing algorithms, as has been implemented by Laurent Baumes and coworkers in the code Parallel Genetic Hybrid Algorithm for Zeolites. Also, the groundbreaking algorithms included in the SCIBS (Symmetry-Constrained Intersite Bonding Search) software are described in detail by Michael Treacy. This method enumerates all possible four-valent networks within each space group given the number of unique tetrahedral atoms. Using this program, in collaboration with other authors, the first database of hypothetical zeolites was created and has been freely available since 2004.

7  Numerical Representations of Chemical Data for Machine Learning

In Chapter 7, Gyoung Na discusses how molecular structures are represented in a way that can be input into machine learning algorithms. In chemistry, a natural representation of molecules is as molecular graphs, which can be input into graph neural networks. The general architecture of graph neural networks is presented, and practical implementations including graph convolutional networks, graph attention networks, message passing neural networks, and crystal graph convolutional neural networks are summarized. Many problems in chemistry involve interactions between multiple molecules, and therefore the adaptation of graph neural networks to these problems is also discussed. While graph-based representations of molecular structures have proven to be very useful, they do not always capture all aspects of molecular structure. Representation learning can be applied to generate latent representations of molecular graphs that can subsequently be used as input for classification and regression algorithms. In a separate paragraph, Python implementations for generating molecular graphs and using these as input in the graph-based machine learning algorithms presented in this chapter are discussed. Finally, the application of graph-based machine learning for a number of chemical applications is demonstrated.

8  Extracting Metal–Organic Framework Data from the Cambridge Structural Database

To successfully apply data-based learning to the design and property prediction of nanoporous materials, data have to be collected and carefully curated. Chapter 8, by Aurelia Li, Rocio Bueno-Perez, and David Fairen-Jimenez, presents how the MOF subset of the Cambridge Structural Database, CSD MOF, has been extracted. It describes how the experimental data have been curated in terms of disorders, solvent removal, and addition of missing hydrogens. It is shown how the CSD MOF subset can be used to analyze the textural properties of MOFs and to classify MOFs in terms of chemical families, functional groups, chirality, porosity, and crystal quality.

9  Data-Driven Approach for Rational Synthesis of Zeolites and Other Nanoporous Materials

Searching for and optimizing nanoporous materials using traditional trial and error experiments is time-consuming and cost-inefficient. In Chapter 9, Watcharop Chaikittisilp illustrates through a number of example studies how data-driven techniques can significantly speed up these processes at different stages. In one study, it is shown how organic free zeolite syntheses can be modeled and predicted by applying machine learning on data collected from the literature. Property prediction of nanoporous materials is illustrated by setting up and validating models to predict the electrochemical catalytic oxygen reduction reaction of doped nanoporous carbons. This particular case is characterized by the occurrence of missing data, and appropriate ways to address this issue are discussed. Optimization of experimental conditions to generate better performing materials is illustrated by the use of active learning to design nanoporous metal alloys for electrochemical applications. It is suggested that data science techniques as discussed in this chapter can be combined with automated high-throughput experiments to significantly enhance the discovery of new and better performing nanoporous materials.

10  Porous Molecular Materials: Exploring Structure and Property Space With Software and Artificial Intelligence

Porous molecular materials, or PMMs, differ from other materials such as zeolites or MOFs in that they are not formed by direct covalent or coordination bonds, but instead consist of molecular units that are held together by non-covalent interactions. In Chapter 10, Steve Bennett and Kim Jelfs describe how computational methods are playing an increasing role in rationalizing the structure and property of PMMS and in the discovery of new PMMs. The chapter starts with an overview of computational modeling of PMMs. Next, the generation of data and the construction of appropriate descriptors required for the application of machine learning techniques is discussed. It is then shown how these ML techniques can be combined with generative algorithms to explore chemical space in search for novel PMM materials. The importance of combining these efforts with the consideration of synthetic accessibility is emphasized.

11  ML-Aided Discovery of Nanoporous Materials for Energy- and Environmental-Related Applications

An important application of nanoporous materials involves gas adsorption for gas storage and separation. Traditionally, computational methods based upon molecular simulations and first principle calculations have been deployed for the discovery of novel materials in this field. In Chapter 11, Archit Datar, Wiang Lyu, and Li-Chiang Lin describe how data-driven approaches have led to significant improvements in this area, especially in terms of computational efficiency. First, an introduction to commonly used statistical and machine learning techniques is provided. It is then described how these techniques are deployed in data-driven approaches for studying gas adsorption and separation in nanoporous materials. Topics discussed are the modeling of interatomic and intermolecular interactions, the development of reliable nanoporous structure datasets, the selection of appropriate molecular descriptors, and methods to search chemical space for the discovery of optimal materials. Deployment of the methods described is illustrated in a number of case studies, including the screening of materials for post-combustion CO2 capture as well as methane and hydrogen storage.

12  Big Data Science in Nanoporous Materials: Datasets and Descriptors

The digital revolution brought about by the widespread availability of digital computers and electronic databases has enabled the application of big data science to accelerate materials discovery. In Chapter 12, Maciej Haranczyk and Giulia Lo Dico describe this evolution in the context of nanoporous materials research. Topics discussed are repositories of experimental and predicted crystal structures, appropriate descriptors of these structures, the collection and representation of materials properties, and the prediction of properties from structure descriptors using machine learning models. The chapter concludes with an overview of current applications and a reflection on future possibilities.

13  Efficient Data Utilization in Training ML Models for Nanoporous Materials Screening

Machine learning approaches have become important tools in materials science and will remain so in the future. However, machine learning is a data hungry approach and available data in the field of nanoporous materials remains relatively scarce. It is therefore important to effectively use these available data. This is discussed in depth by Diego Gómez-Gualdrón, Cory Simon, and Yamil Colón in Chapter 13, which covers descriptor, material, and model selection for the prediction of material properties, and methods to effectively deploy these in the search of novel materials. While these topics are discussed with examples in the context of modeling of adsorption in MOFs, the conclusions and lessons learned are generally applicable in the prediction of other properties and of nanoporous materials.

14  ML and Digital Manufacturing Approaches for Solid-State Materials Development

In Chapter 14, Lawson Glasby, Emily Whaites, and Peyman Moghadam discuss how the introduction of data science and digital technologies paves the way to more efficient synthesis and development of novel solid materials. An important challenge here is the collection and use of appropriate data. This is illustrated in the case of MOFs by discussing existing databases and the application of natural language processing to extend and maintain these data, and by giving an overview of machine learning algorithms to mine these. The state-of-the -art in automated solid-state synthesis is illustrated through a number of examples. The emergence of open-source data repositories of consistent and reliable reports of synthesis results and conditions, including those of unsuccessful experiments, is identified as an important evolution that holds promise to overcome current limitations.

15  Overview of AI in the Understanding and Design of Nanoporous Materials

Throughout the chapters of this book, the contributing authors have described the latest developments and state-of-the-art of the application of AI methods in nanoporous materials development. Topics covered included the development and maintenance of databases of structures, properties, and synthesis protocols, and the mining of these using big data science. Current applications of machine learning were discussed, including generative algorithms and data-driven synthesis approaches, prediction of adsorption properties, and OSDA modeling. The authors of the concluding chapter of this book summarize these contributions, identify remaining challenges, and give their view on future developments.

Frits DaeyaertGerman Sastre

July 2022

Notes

1

Cohen, D.S. “‘Pac-Man’ – the Most Important and Iconic Video Game of All Time.”

https://www.lifewire.com/pac-man-video-game-729560

.

2

Lauer, D. (2021). You cannot have AI ethics without ethics.

AI Ethics

1: 21–25.

https://doi.org/10.1007/s43681-020-00013-4

.

3

Benjamins, R. (2021). A choices framework for the responsible use of AI.

AI Ethics

1: 49–53.

https://doi.org/10.1007/s43681-020-00012-5

.

4

Asimov, I. (1950).

I Robot

. Gnome Press.

5

Wang, A.Y.-T., Murdock, R.J., Kauwe, S.K. et al. (2020). Sparks: Machine learning for materials scientists: an introductory guide toward best practices.

Chem. Mater

. 32: 4954−4965.

About the Cover

The storm over the Parthenon symbolizes how the advent of artificial intelligence is having a profound impact on our traditional approach of building up human knowledge. STW is one of the few known zeolite frameworks exhibiting chirality, which opens up possibilities for enantioselective absorption, separation and catalysis. BHEHPI is an organic linker molecule used to build the metal-organic framework with the highest specific surface area known. The AI revolution is made possible by high performance computing hardware.

Acknowledgments

First of all we would like to thank Wiley for the call to identify an area of materials chemistry where a new book might be of wide and general interest, and in particular to Sarah Higginbotham who actively supported our initial suggestions in the months following September 2019 when Emma Strickland knocked at our door with her kind and challenging invitation. During this time of almost three years that it took to prepare this book, we have also received active support from other members of the team whom we also would like to thank. Gnanapriya Pattel, who took care of the CTA Agreements, and the other co-editors of the book, Sakeena Quraishi and Stefanie Volk, who showed a continued and enthusiastic support that contributed to keep the ball rolling almost without interruption during the hard pandemic times. And not to forget the hard work carried by Durgadevi Shanmugasundaram, Hemalathaa Krishnamoourthy, and Jo Egre, editing bits and pieces here and there, always with exemplary professionality and good sense of humor.

We were also happily surprised by an invitation by Sakeena to contribute to the cover design and were allowed to explore a huge in-house store of drawings and photographs. After hours, one of them immediately caught our attention, “Storm over Parthenon.” Seldom could a better symbol of human knowledge than the Parthenon be found. And then, the thunderstorm resembles how this traditional paradigm may be jeopardized by a new way of acquiring knowledge, which is Artificial Intelligence. Hopefully the storm will transform into water that will irrigate a new fertile ground for the benefit of mankind. Jennifer Cossham helped so much throughout this process of cover design and allowed us, incompetent amateurs, to put our two cents into the marvelous work made by the designers.

Well, this shows the atmosphere of working hand-in-hand with the Wiley team. We exchanged warm greetings at Christmas and felt almost part of the 100-years Celebration at Wiley-VCH1 in March 2021. E-mails indicated that the Wiley team was mostly located in Germany but during all this time we had a feeling that they were next door to us.

We are especially indebted to authors who took over replacements of chapters that were scheduled but could not be finished and very kindly adapted their schedule to facilitate a timely update. And hence we come to the most important acknowledgment, that due to the authors of this book. Thanks for contributing generously with the best of your research to build a coherent, complete. and overall high-quality view of how Artificial Intelligence is taking the science of materials chemistry by storm.

Finally, thanks to all the people who contributed by revising chapters and suggesting additions and improvements to the book, and also those who gave us confidence and support: Jeffrey D. Rimer (University of Houston), Wei Fan (University of Massachusetts), Rob G. Bell (University College London), Koki Muraoka (The University of Tokyo), Zach Jensen (MIT), Kenta Iyoki (The University of Tokyo), Takahiko Moteki (The University of Tokyo), Yi Li (Jilin University), Hiromasa Kaneko (Meiji University), Seda Keskin (Koç University), Angeles Pulido (Cambridge Crystallographic Data Centre), Dewi W. Lewis (University College London), Daniele Ongari (Swiss Federal Institute of Technology Lausanne), Aditi Krishnapriyan (Lawrence Berkeley National Laboratory), Andrew Rosen (Lawrence Berkeley National Laboratory), Anubhav Jain (Lawrence Berkeley National Laboratory), Teng Zhou (Max-Planck-Institut für Dynamik Komplexer Technischer Systeme), Andrew M. Beale (University College London), François-Xavier Coudert (Paris Sciences et Lettres University), Valeria Molinero (University of Utah), Johannes Hachmann (University at Buffalo), Steven K. Kauwe (University of Utah), Anthony Wang (University of Waterloo), Mark E. Davis (Caltech), Chris Wilmer (University of Pittsburgh), Keisuke Takahashi (Hokkaido University), and Amanda S. Barnard (Australian National University).

Frits DaeyaertGerman Sastre

July 2022

Notes

1

The Wiley Network; Celebrate with Wiley-VCH – 100 Years of Growing Knowledge;

https://www.wiley.com/network/latest-content/celebrate-with-wiley-vch-100-years-of-growing-knowledge

1 The Confluence of Organo-Cations, Inorganic Species, and Molecular Modeling on the Discovery of New Zeolite Structures and Compositions

Christopher M. Lew, Dan Xie, Joel E. Schmidt, Saleh Elomari, Tracy M. Davis, and Stacey I. Zones

Chevron Technical Center, Richmond, California, USA

1.1 Introduction

The use of quaternary ammonium cations as organic structure-directing agents (OSDAs) in zeolite synthesis dates back to well over half a century. They are perhaps the greatest driver in discovering new zeolite frameworks and compositions, but at the same time are limited by their cost, which can prevent economic commercial-scale production of new materials [1]. There are already many fine reviews on zeolite synthesis and structure direction, and Table 1.1 contains a part of a non-exhaustive list of these. The purpose of this review is to highlight a subset of zeolite synthesis reactions where subtle changes in the synthesis variables lead to new structures and compositions. Moreover, we elucidate the role of molecular modeling to help us understand the results and to guide new experiments. While many materials found with these minor perturbations tend to be difficult to access synthetically, their unique properties can give them superior performance in numerous applications, which can catapult them from a laboratory curiosity to a high-demand commercial product. Perhaps there is no better modern example of this than the use of SSZ-13 as an automotive deNOx catalyst, and the story of that material is well-known in the zeolite community (the SSZ-designation is used for novel materials prepared by Chevron) [2].

Table 1.1 A non-exhaustive listing of review papers covering zeolite synthesis and structure direction.

Title

Authors

Year

References

Zeolite and Molecular Sieve Synthesis

Davis, M.E. and Lobo, R.F.

1992

[

35

]

Synthesis of Porous Silicates

Helmkamp, M.M. and Davis, M.E.

1995

[

127

]

Searching for New High-Silica Zeolites Through a Synergy of Organic Templates and Novel Inorganic Conditions

Zones, S.I., Nakagawa, Y., Lee, G.S. et al.

1998

[

16

]

Synthesis of All-Silica and High-Silica Molecular Sieves in Fluoride Media

Camblor, M.A., Villaescusa, L.A., and Diaz-Cabanas, M.J.

1999

[

128

]

Ordered Porous Materials for Emerging Applications

Davis, M.E.

2002

[

129

]

Towards the Rational Design of Zeolite Frameworks

Wagner, P. and Davis, M.E.

2002

[

130

]

The Hydrothermal Synthesis of Zeolites: History and Development from the Earliest Days to the Present Time

Cundy, C.S. and Cox, P.A

2003

[

3

]

The Hydrothermal Synthesis of Zeolites: Precursors, Intermediates, and Reaction Mechanism

Cundy, C.S. and Cox, P.A.

2005

[

131

]

The Chemistry of Phase Selectivity in the Synthesis of High-Silica Zeolites

Burton, A.W., Zones, S.I., and Elomari, S.

2005

[

80

]

The Fluoride-Based Route to All-Silica Molecular Sieves: A Strategy for Synthesis of New Materials Based Upon Close-Packing of Guest–Host Products

Zones, S.I., Hwang, S.J., Elomari, S. et al.

2005

[

23

]

Zeolite Molecular Sieves: Preparation and Scale-Up

Casci, J.L.

2005

[

132

]

Organic Molecules in Zeolite Synthesis: Their Preparation and Structure-Directing Effects

Burton, A.W. and Zones, S. I.

2007

[

31

]

Present and Future Synthesis Challenges for Zeolites

Coronas, J.

2010

[

133

]

Inorganic Molecular Sieves: Preparation, Modification and Industrial Application in Catalytic Processes

Martínez, C. and Corma, A.

2011

[

134

]

Zeolites: From Curiosity to Cornerstone

Masters, A.F. and Maschmeyer, T.

2011

[

135

]

Needs and Trends in Rational Synthesis of Zeolitic Materials

Wang, Z., Yu, J., and Xu, R.

2012

[

136

]

Towards the Rational Design of Efficient Organic Structure‐Directing Agents for Zeolite Synthesis

Moliner, M., Rey, F., and Corma, A.

2013

[

137

]

New Trends in the Synthesis of Crystalline Microporous Materials

Bellussi, G., Carati, A., Rizzo, C. et al.

2013

[

138

]

New Stories of Zeolite Structures: Their Descriptions, Determinations, Predictions, and Evaluations

Li, Y. and Yu, J.

2014

[

139

]

Zeolites From a Materials Chemistry Perspective

Davis, M.E.

2014

[

140

]

Synthesis Strategies for Preparing Useful Small Pore Zeolites and Zeotypes for Gas Separations and Catalysis

Moliner, M., Martinez, C., and Corma, A.

2014

[

141

]

Synthesis of New Zeolite Structures

Li, J., Corma, A., and Yu, J.

2015

[

32

]

Introduction to the Zeolite Structure-Directing Phenomenon by Organic Species: General Aspects

Gómez-Hortigüela, L. and Camblor, M.A.

2017

[

142

]

Small-Pore Zeolites: Synthesis and Catalysis

Dusselier, M. and Davis, M.E.

2018

[

2

]

A timeline of zeolite development is given in Table 1.2, and the first synthetic zeolites date back to around 1950 with the pioneering inorganic work of Richard Barrer and Robert Milton, the founding fathers of synthetic zeolites, and a little later with Donald Breck and Edith Flanigen [3]. Over 20 synthetic zeolites were quickly discovered, but the first zeolite patent was not granted until 1959, even though the application was filed in 1953 [4]. The first reported organics for zeolite synthesis were simple methylammoniums used to produce higher silica versions of zeolites A and X, as was first reported in 1961 and followed later by patents [5–9]. Once the genie was out of the bottle with regard to this innovation, it was quickly followed by reports of zeolite beta in 1967 [10] and ZSM-5 in 1972 [11]. To this day, these zeolites remain two of the high-silica zeolites with the largest industrial applications and have been the subject of innumerable patents and research papers. Some of the most notable early organizations in the zeolite field were Linde, Mobil, Grace, and Union Carbide, and their labs discovered and commercialized a remarkable number of materials in a short time. A great amount of progress has been made in the field in the past 50 years, and herein we will give an overview of zeolite synthesis variables, with special attention paid to OSDAs, to highlight how subtle perturbations to these variables can lead to significant differences in the crystalline product.

Table 1.2 Zeolite timeline.

Year

Discovery

Antiquity

Used in Roman aqueducts and Mayan filtration systems

1756

Axel Fredrick Cronstedt coins the term “zeolite,” meaning boiling stone

1905

Used commercially to soften water

1930s

Pioneering work of Richard Barrer on adsorption and synthesis (

MOR

,

KFI

)

1950s

Milton and Breck discover A, X, Y

1954

Union Carbide commercializes synthetic zeolites for separations

1962

Mobil Oil uses zeolite X for cracking

1967–1969

Mobil reports beta and ZSM-5

1969

Grace reports USY

1974

Henkel introduces zeolite A in detergents

1977

Union Carbide commercializes ion-exchange separations

1980s

AlPOs discovered by Union Carbide. Heteroatoms: Fe, Ga, Ti, Ge. Secondary modifications

1990s

MeAPOs, MOFs, mesoporous materials

2000s

Charge density mismatch aluminosilicates, germanosilicate compositions, fluoride media

2010s

ADOR, 2D materials, chirality, dry gel conversion, computational advances, small-pore deNO

x

2020s

Machine learning, data mining of historical data, high-throughput experiments

Source:

Adapted with permission from [

143

].

1.2 Inorganic Studies

In the preface of his recent monograph, “Insights into the Chemistry of Organic-Structure Directing Agents in the Synthesis of Zeolite Materials,” Luis Gomez-Hortiguela states that the activity in this area has largely been based upon empirical observations and then decisions, but that the field is emerging into new opportunities with advanced computer modeling [12