Handbook of Micro Reactors 9783527315505, 3527315500

Table of contents :
Novel ProcessWindows......Page 5
Contents......Page 9
Motivation – Who should read the book!?......Page 19
Acknowledgments......Page 21
Abbreviations......Page 23
Nomenclature......Page 25
1.1 Prelude - Potential for Green Chemistry and Engineering......Page 29
1.2.1 12 Principles in Green Chemistry......Page 30
1.3.1 10 Key Research Areas in Green Engineering......Page 31
1.3.2 12 Principles in Chemical Product Design......Page 32
1.4.2 Microstructured Reactors......Page 34
1.6 Two Missing Links - Cross-Related......Page 37
References......Page 40
2.1 Transport Intensification - The Potential of Reaction Engineering......Page 43
2.2 Chemical Reactivity in Match or Mismatch to Intensified Engineering......Page 45
2.3 Chemical Intensification through Harsh Conditions - Novel Process Windows......Page 46
2.4 Flash Chemistry......Page 47
2.5 Process-Design Intensification......Page 49
References......Page 51
3.1 Length Scale......Page 53
3.2 Time Scale......Page 54
3.3 Length and Time Scale of Chemical Reactions......Page 56
3.3.1 Solution of Kinetic Equations......Page 57
3.3.2 Reaction Time and Reaction Classification......Page 59
3.3.3 Example for Reaction Time and Residence Time......Page 60
3.4.2 New Temperature Windows......Page 61
3.4.3 Reaction Rate - Arrhenius Equation......Page 63
3.5.1 Reaction Rate - Activation Volumes......Page 64
3.5.2 Equilibrium......Page 65
3.5.4 Material Properties......Page 66
3.5.5 Mixture Properties......Page 68
3.5.6 Illustration of Pressure Effect on Selected Chemical Reactions......Page 69
References......Page 70
4.1.1 Tetrazole Formation......Page 73
4.1.3 Phosgene Chemistry......Page 75
4.1.4 Diazomethane Synthesis......Page 76
4.1.5 Ozonolysis......Page 78
4.1.6 Organic Peroxide Formation......Page 79
4.2.2 Hydrogen Peroxide Synthesis......Page 80
4.2.4 Ionic Liquid Synthesis......Page 81
4.2.5 Moffatt-Swern Oxidation......Page 82
4.2.7 Nitration of Toluene......Page 83
4.2.9 Decarboxylative Trichloromethylation of Aromatic Aldehydes......Page 84
References......Page 86
5.1.1 Fluorination with Elemental Fluorine......Page 89
5.1.3 Direct Aryllithiums Route......Page 90
5.1.4 C-O Bond Formation by a Direct α-C-H Bond Activation......Page 91
5.1.5 Direct Adipic Acid Route from Cyclohexene......Page 93
5.1.6 New Biocatalytic Pathways without Protecting Groups - Inter-Glycosidic Condensation......Page 96
5.2.1 "Odor-Sealed" Isocyanide Formation......Page 97
5.3 Multistep One-Flow Syntheses......Page 98
5.4.1 Multistep Synthesis of [18F]-Radiolabeled Molecular Imaging Probe......Page 101
5.4.3 Two-Step Strecker Reaction......Page 103
5.5.1 Chlorohydrination of Allyl Chloride......Page 104
5.5.4 Ring-Closing Metathesis and Heck Reaction......Page 105
5.5.6 Suzuki-Miyaura Cross-Coupling/Hydrogenation......Page 106
5.5.8 Murahashi Coupling/Br-Li Exchange......Page 107
5.5.9 5′-Deoxyribonucleoside Glycosylation......Page 108
5.5.11 Low-Pressure Carbonylations with Acids as CO Precursors......Page 109
5.5.13 Synthesis of TAC-101 Analogs......Page 110
5.5.15 Multistep Enzymatic Synthesis to δ-d-Gluconolactone......Page 111
References......Page 115
Chapter 6 Activate - High-T Processing......Page 119
6.1.3 Modular Thermal Platform for High-Temperature Flow Reactions......Page 121
6.2.1 Synthesis of Triflates for the Heck Alkenylation......Page 122
6.2.2 Enantioselective 1,4-Addition of Enones......Page 124
6.2.3 Swern-Moffatt Oxidation of Benzyl Alcohol......Page 125
6.2.4 Tf2NH-Catalyzed [2+2] Cycloaddition......Page 126
6.3.2 C-F Bond Formation......Page 127
6.3.5 Nucleophilic Substitution of Difluoro-benzenes......Page 128
6.3.6 Aminolysis of Epoxides......Page 129
6.3.8 2-Methylbenzimidazole Formation, 3,5-Dimethyl-1-Phenylpyrazole Formation, and Diels-Alder Cycloaddition - Benchmarking High-p,t Flow to Microwave......Page 130
6.3.10 Thermal Hydrolysis of Triglycerides......Page 131
6.3.11 Chlorodehydroxylation to n-Alkyl Chlorides......Page 132
6.3.12 1,3,4-Oxadiazoles via N-Acylation of 5-Substituted Tetrazoles......Page 133
6.3.13 Cobalt-Catalyzed Borohydride Reduction of Tetralone......Page 134
6.3.14 Dimethylcarbonate Methylation......Page 135
6.3.15 Selective Aerobic Oxidation of Benzyl Alcohol Using Iron Oxide Nano-/TEMPO Catalyst......Page 136
6.3.16 Rufinamide Synthesis......Page 138
6.3.18 Click Chemistry......Page 139
6.3.19 4-(Pyrazol-1-yl) Carboxanilide Multistep Synthesis......Page 140
6.3.20 4-Hydroxy-2-cyclopentenone Synthesis......Page 141
6.3.22 Tetrahydroisoquinoline Synthesis......Page 142
6.4.2 Intramolecular Thermal Cyclization and Benzannulation......Page 144
6.4.4 Aminolysis of Epoxides......Page 145
6.5.1 Palladium-Catalyzed Aminocarbonylation......Page 146
6.5.3 Flash Flow Pyrolysis......Page 147
6.5.4 Indium Phosphide Nanocrystal......Page 148
6.5.5 Quantum Dot Synthesis......Page 149
6.5.6 High-T Flow Cycloaddition to Fullerene Derivatives......Page 151
References......Page 153
7.1.2 In-Plane Fiber-Based Interfaced Microreactor......Page 157
7.2.2 Carbamic Acid Formation......Page 158
7.2.4 Catalytic Hydrogenation of Acetone......Page 159
7.2.7 Hydrogen Gas Liquefication in Guaiacol Conversion (hydroprocessing)......Page 160
7.3.2 Nucleophilic Aromatic Substitution of Three p-Halonitrobenzenes......Page 161
7.3.3 Diels-Alder Reaction with Furylmethanols and Cyclopentadiene......Page 162
7.3.5 Esterification of Phthalic Anhydride......Page 164
7.4 Pressure for Advanced Fluidic Studies - to be Used for Shaping Materials and More......Page 165
References......Page 166
8.1.1 Polypropylene and Polycarbonate Polymerizations......Page 169
8.2.3 Claisen Rearrangement of Substituted Phenyl Phenols......Page 170
8.2.4 Michael Addition......Page 171
8.2.6 Beckmann Rearrangement (High-c)......Page 172
8.2.7 [2+2] Photocycloaddition of a Chiral Cyclohexenone (High-c)......Page 173
8.2.10 Synthesis of Nitro Herbicides (High-c, Solvent-Free)......Page 176
8.2.12 Enzyme and Coenzyme (High-c in Bioprocessing)......Page 178
8.2.13 Enzyme and Coenzyme (High-c in Bioprocessing)......Page 179
8.3 Supercritical Fluids to Combine the Former Separated - Mass Transfer Boost......Page 180
8.3.2 Supercritical Hydrogenations of Double and Triple Bounds......Page 183
8.3.5 Near-scCO2 Enzymatic Biodiesel Synthesis......Page 185
8.3.6 Supercritical Water, Non-Catalytic Beckmann Rearrangement......Page 186
8.3.8 Supercritical Water Oxidation......Page 187
8.3.10 Self-Optimizing Continuous Reactions under Supercritical Conditions......Page 190
References......Page 191
9.1.1 Integrated Micro-Steam Reformer-Catalytic Combustor for Methane Fuel Processing......Page 193
9.1.2 Integrated Microburner/Thermoelectric Device for System Start-Up......Page 194
9.1.5 Integrated Enzyme Microreactor-Extractor......Page 195
9.1.6 Integrated Membrane Microreactor for Knoevenagel Reaction......Page 196
9.1.8 Coupling of the Hydroxylation of Progesterone Using Rhizopus Nigricans with Flow Extraction......Page 197
9.1.9 Coupling of the Esterification to Isoamyl Acetate Using Lipase B with Flow Extraction......Page 198
9.2.2 Integrated Sensing, Catalyst, and Heating for Ammonia Oxidation......Page 199
9.2.3 Integration of the Esterification to Ethyl Oleate Using Lipase B with Photoionization Mass Spectrometry......Page 200
9.2.4 Integration of the Intramolecular Friedel-Crafts Addition with Ultra-High-Pressure Liquid Chromatography......Page 201
9.2.5 Integration of Pyrane Flow Reaction and Synchrotron-Based IR and X-Ray Beam Analysis......Page 202
9.2.6 Integration of Multistep Organic Transformations Catalyzed by Au Nanoclusters......Page 204
9.3.1 Thermal Integration of a Methanol Micro-fuel Processor/Fuel Cell......Page 205
9.4.2 Chassis-Type Unit Racking and System Automation......Page 206
9.5.1 Adipic Acid Large-Scale Manufacture......Page 207
References......Page 211
10.1.2 Bromination of Toluene......Page 213
10.2 Simplifying Separation......Page 214
10.2.1 Phenyl Boronic Acid Synthesis......Page 215
10.2.3 Trans-1,2-Cyclohexanediol - Exothermic Steps Done "All-in-Once"......Page 216
10.2.4 Olefin Autoxidation......Page 217
References......Page 218
11.1 Introduction......Page 219
11.3 Evaluation Methods......Page 220
11.3.1 Single Metrics......Page 221
11.3.2 Holistic Approaches......Page 222
11.3.3 Life Cycle Costing......Page 229
11.3.3.1 Cradle-to-Gate Approach in LCC......Page 230
11.3.3.2 Calculation of Costs......Page 232
11.4 Evaluation of the NPW Concept Impact on Sustainability......Page 234
11.4.1.2 Avoidance of Waste Products......Page 235
11.4.2.1 p-Xylene Partial Oxidation under Harsh Conditions......Page 236
11.4.2.2 Process Intensification of Biodiesel Generation......Page 237
11.4.2.3 Generation of Carbon Nanotubes at High Temperatures......Page 239
11.4.2.4 Activation of Carbon Dioxide......Page 241
11.4.2.5 Impact of Harsh Process Conditions of Catalyst Deactivation......Page 243
11.4.2.6 Methane Decomposition at High Reaction Temperatures......Page 244
11.4.2.7 Synthesis of Fullerenes by Pyrolysis versus Plasma......Page 245
11.4.2.8 Epoxidation Reaction at Accelerated Temperature......Page 247
11.4.2.9 Trade-Off Designs for a Hydrocarbon Biorefinery......Page 248
11.4.3.1 Sildenafil Citrate Process......Page 250
11.4.3.2 Suzuki-Miyaura Cross-Coupling......Page 252
11.4.3.3 Synthesis and Application of Ionic Liquids......Page 253
11.4.4.1 Acceleration of Multiphase Reactions via Ultrasound......Page 254
11.4.4.2 Process Optimization of a Pharmaceutical Synthesis......Page 257
11.4.4.3 Process Simplification of Adipidic Acid Synthesis......Page 259
11.4.4.4 Process Enhancement via Phase Transfer Catalysis......Page 260
11.4.4.6 Process Integration in Terms of Costs......Page 262
11.4.4.7 The Effects of Modular Plants......Page 263
11.5 Future Environmental and Economic Sustainability Evaluation in the Context of Flow-Chemistry under NPW Conditions......Page 264
References......Page 266
12.1 Reactor Types......Page 269
12.2 Scale-Up Parameters......Page 271
12.2.1 Geometry......Page 272
12.3.2 External Numbering-Up......Page 274
12.4.1 Fluid Dynamics in a Rectangular Channel......Page 275
12.4.2 Mean Residence Time and Its Distribution......Page 276
12.4.3 Pressure Loss and Mixing......Page 280
12.4.4 Heat Transfer in Channel Reactors......Page 282
12.4.5 Parametric Sensitivity and Reactor Thermal Stability......Page 284
12.5 Methodology for Continuous-Flow Process Development......Page 286
12.5.1 General Chemical Plants and Modular Setup......Page 289
12.5.2 Platform Concept and Scalability......Page 292
12.5.3 Modular Process Development......Page 295
12.5.3.1 Module A: Feasibility Study - Milestone Decision......Page 296
12.5.3.2 Module B: Process Synthesis and Optimization - Milestone Decision......Page 298
12.5.3.3 Module C: Process Robustness and Economy - Milestone Decision......Page 301
12.5.3.4 Module D: Pilot Production and Commercial Manufacturing......Page 302
12.6 Conclusions......Page 306
References......Page 308
13.1.2 NPW-Route Classification and Experimental Demonstration (2009)......Page 311
13.1.5 Flow Chemistry High-T Overview - Superheated Processing (2010)......Page 312
13.1.7 NPW-Methodology and Intensification Considerations for High-T Flow Reactions (2011)......Page 313
13.3 Funding Agency Initiatives......Page 314
13.3.1 DBU Cluster Novel Process Windows......Page 315
13.3.2 EU Funding......Page 317
References......Page 319
14.1.1 Search Methodology and Limitations......Page 321
14.1.2 High-Temperature, High-Pressure, High-Concentration in Chemistry and Chemical Engineering Literature......Page 322
14.1.3 High-T, -p, -c in Microreactor Literature......Page 323
14.2 Literature Share for Process-Design Intensification......Page 326
References......Page 328
15.1.2 Process-Automated Application Example......Page 329
15.2.1 Photoactivation......Page 331
15.2.2 Activation by Instable Catalyst Precursors......Page 333
References......Page 335
Index......Page 337
EULA......Page 343

Citation preview

Volker Hessel, Dana Kralisch, and Norbert Kockmann Novel Process Windows

Related Titles Hessel, V., Renken, A., Schouten, J.C., Yoshida, J. (eds.)

Reniers, G.L., Sörensen, K., Vrancken, K. (eds.)

Micro Process Engineering

Management Principles of Sustainable Industrial Chemistry

A Comprehensive Handbook 2009 Print ISBN: 978-3-527-31550-5 (Also available in a variety of digitial formats)

Theories, Concepts and Industrial Examples for Achieving Sustainable Chemical Products and Processes from a Non-Technological Viewpoint

Kashid, M.N., Renken, A., Kiwi-Minsker, L.

Microstructured Devices for Chemical Processing 2015 Print ISBN: 978-3-527-33128-4 (Also available in a variety of digitial formats)

2013 Print ISBN: 978-3-527-33099-7 (Also available in a variety of digitial formats)

Sanchez Marcano, J.G., Tsotsis, T.T.

Catalytic Membranes and Membrane Reactors Second Edition 2009 Print ISBN: 978-3-527-32362-3 (Also available in a variety of digitial formats)

Volker Hessel, Dana Kralisch, and Norbert Kockmann

Novel Process Windows Innovative Gates to Intensified and Sustainable Chemical Processes

The Authors Prof. Dr. Volker Hessel

Technische Universiteit Eindhoven Den Dolech 2 5600 Eindhoven Netherlands Dr. Dana Kralisch

Friedrich Schiller University Institut für Technische Chemie & Umweltchemie Lessingstr. 12 07743 Jena Germany

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.

Prof. Dr. Norbert Kockmann

TU Dortmund Fakultät Bio- und Chemieingenieurwesen Emil-Figge-Str. 68 44227 Dortmund Germany

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 . © 2015 Wiley-VCH Verlag GmbH & Co. KGaA, 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: 978-3-527-32858-1 ePDF ISBN: 978-3-527-65485-7 ePub ISBN: 978-3-527-65484-0 Mobi ISBN: 978-3-527-65483-3 oBook ISBN: 978-3-527-65482-6 Cover-Design Adam Design, Weinheim,

Germany Typesetting Laserwords Private Limited, Chennai, India Printing and Binding Markono Print Media Pte Ltd, Singapore

Printed on acid-free paper

V

Windows Provide Panoramas Novel Process Windows are a chance to explore new horizons for processing industry. Glass windows of churches initially were Romanic style, being small and protective, as long as constructional engineering was hampered by raw material manufacture, heat losses, and mechanical constraints. With improvements on the engineering side, they turned to the large, ornament-style gothic windows exposing the interior with sunlight giving rise to metaphysical illumination.

VII

Contents Motivation – Who should read the book!? XVII Acknowledgments XIX Abbreviations XXI Nomenclature XXIII 1

From Green Chemistry to Green Engineering – Fostered by Novel Process Windows Explored in Micro-Process Engineering/Flow Chemistry 1

1.1 1.2 1.2.1 1.3 1.3.1 1.3.2 1.4 1.4.1 1.4.2 1.5 1.5.1 1.6

Prelude – Potential for Green Chemistry and Engineering Green Chemistry 2 12 Principles in Green Chemistry 2 Green Engineering 3 10 Key Research Areas in Green Engineering 3 12 Principles in Chemical Product Design 4 Micro- and Milli-Process Technologies 6 Microreactors 6 Microstructured Reactors 6 Flow Chemistry 9 10 Key Research Areas in Flow Chemistry 9 Two Missing Links – Cross-Related 9 References 12

2

Novel Process Windows 15 Transport Intensification – The Potential of Reaction Engineering 15 Chemical Reactivity in Match or Mismatch to Intensified Engineering 17 Chemical Intensification through Harsh Conditions – Novel Process Windows 18 Flash Chemistry 19 Process-Design Intensification 21 References 23

2.1 2.2 2.3 2.4 2.5

1

VIII

Contents

3

3.1 3.2 3.3 3.3.1 3.3.2 3.3.3 3.4 3.4.1 3.4.2 3.4.3 3.5 3.5.1 3.5.2 3.5.3 3.5.4 3.5.5 3.5.6

Chemical Intensification – Fundamentals 25 Length Scale 25 Time Scale 26 Length and Time Scale of Chemical Reactions 28 Solution of Kinetic Equations 29 Reaction Time and Reaction Classification 31 Example for Reaction Time and Residence Time 32 Temperature Intensification 33 Harsh Process Conditions 33 New Temperature Windows 33 Reaction Rate – Arrhenius Equation 35 Pressure Intensification 36 Reaction Rate – Activation Volumes 36 Equilibrium 37 Electron Kinetic Energy 38 Material Properties 38 Mixture Properties 40 Illustration of Pressure Effect on Selected Chemical Reactions 41 References 42

4

Making Use of the “Forbidden” – Ex-Regime/High Safety Processing 45

4.1 4.1.1 4.1.2 4.1.3 4.1.4 4.1.5 4.1.6 4.2 4.2.1 4.2.2 4.2.3 4.2.4 4.2.5 4.2.6 4.2.7 4.2.8 4.2.9 4.2.10

Hazardous Reactants and Intermediates 45 Tetrazole Formation 45 Strecker Synthesis 47 Phosgene Chemistry 47 Diazomethane Synthesis 48 Ozonolysis 50 Organic Peroxide Formation 51 Ex-Regime and Thermal Runaway Processing 52 Oxidation 52 Hydrogen Peroxide Synthesis 52 Direct Fluorination 53 Ionic Liquid Synthesis 53 Moffatt–Swern Oxidation 54 Reaction Between Cyclohexanecarboxylic Acid and Oleum 55 Nitration of Toluene 55 Aromatic Amidoxime Formation 56 Decarboxylative Trichloromethylation of Aromatic Aldehydes 56 Dihydroxylation Reactions with Nanobrush-Immobilized OsO4 58 References 58

5

Exploring New Paths – New Chemical Transformations 61

5.1

Direct Syntheses via One Step 61

Contents

5.1.1 5.1.2 5.1.3 5.1.4 5.1.5 5.1.6 5.2 5.2.1 5.3 5.4 5.4.1 5.4.2 5.4.3 5.5 5.5.1 5.5.2 5.5.3 5.5.4 5.5.5 5.5.6 5.5.7 5.5.8 5.5.9 5.5.10 5.5.11 5.5.12 5.5.13 5.5.14 5.5.15 5.5.16

6

6.1 6.1.1 6.1.2 6.1.3 6.2 6.2.1

Fluorination with Elemental Fluorine 61 Hydrogen Peroxide Synthesis out of the Elements 62 Direct Aryllithiums Route 62 C–O Bond Formation by a Direct α-C–H Bond Activation 63 Direct Adipic Acid Route from Cyclohexene 65 New Biocatalytic Pathways without Protecting Groups – Inter-Glycosidic Condensation 68 Direct Syntheses via Multicomponent Reactions 69 “Odor-Sealed” Isocyanide Formation 69 Multistep One-Flow Syntheses 70 Multistep Syntheses in One Microreactor/Chip 73 Multistep Synthesis of [18F]-Radiolabeled Molecular Imaging Probe 73 Combining Asymmetric Organocatalysis and Analysis on a Single Microchip 75 Two-Step Strecker Reaction 75 Multistep Syntheses in Coupled Microreactors/Chips 76 Chlorohydrination of Allyl Chloride 76 Lithiation/Borylation/Suzuki–Miyaura Cross-Coupling 77 Suzuki–Miyaura Cross-Coupling-Phenols-Aryl Triflates-Biaryls 77 Ring-Closing Metathesis and Heck Reaction 77 Imidazo[1,2-a]pyridine-2-carboxylic Acids in Two Steps 78 Suzuki–Miyaura Cross-Coupling/Hydrogenation 78 Sodium Nitrotetrazolate – Diazonium Ion Formation/Sandmeyer Reaction 79 Murahashi Coupling/Br–Li Exchange 79 5′ -Deoxyribonucleoside Glycosylation 80 Two-Carbon Homologation of Esters to α,β-Unsaturated Esters 81 Low-Pressure Carbonylations with Acids as CO Precursors 81 Coupled Microreactor-Purification-Analytics for δ-Opioid Receptor Agonist 82 Synthesis of TAC-101 Analogs 82 Multistep Enzymatic Synthesis to 2-Amino-1,3,4-Butanetriol 83 Multistep Enzymatic Synthesis to δ-D-Gluconolactone 83 Diarylethene Synthesis in Two Steps 87 References 87 91 Tailored High-T Microreactor Design and Fabrication 93 Glass Capillary Coil in Ceramic Housing 93 Modularly Packaged Silicon Microreactor 93 Modular Thermal Platform for High-Temperature Flow Reactions 93 Cryogenic to Ambient – Allowing Fast Reactions to be Fast Synthesis of Triflates for the Heck Alkenylation 94 Activate – High-T Processing

94

IX

X

Contents

6.2.2 6.2.3 6.2.4 6.3 6.3.1 6.3.2 6.3.3 6.3.4 6.3.5 6.3.6 6.3.7 6.3.8

6.3.9 6.3.10 6.3.11 6.3.12 6.3.13 6.3.14 6.3.15 6.3.16 6.3.17 6.3.18 6.3.19 6.3.20 6.3.21 6.3.22 6.4 6.4.1 6.4.2 6.4.3 6.4.4 6.5 6.5.1 6.5.2 6.5.3 6.5.4 6.5.5 6.5.6 6.6

Enantioselective 1,4-Addition of Enones 96 Swern–Moffatt Oxidation of Benzyl Alcohol 97 Tf2 NH-Catalyzed [2+2] Cycloaddition 98 From Reflux to Superheated – Speeding-Up Reactions 99 Kolbe–Schmitt Reaction 99 C–F Bond Formation 99 NMP Radical Polymerization of Styrene 100 Noncatalytic Claisen Rearrangement 100 Nucleophilic Substitution of Difluoro-benzenes 100 Aminolysis of Epoxides 101 Synthesis of 2,4,5-Trisubstituted Imidazoles 102 2-Methylbenzimidazole Formation, 3,5-Dimethyl-1-Phenylpyrazole Formation, and Diels–Alder Cycloaddition – Benchmarking High-p,t Flow to Microwave 102 Fischer Indole Synthesis of Tetrahydrocarbazole 103 Thermal Hydrolysis of Triglycerides 103 Chlorodehydroxylation to n-Alkyl Chlorides 104 1,3,4-Oxadiazoles via N-Acylation of 5-Substituted Tetrazoles 105 Cobalt-Catalyzed Borohydride Reduction of Tetralone 106 Dimethylcarbonate Methylation 107 Selective Aerobic Oxidation of Benzyl Alcohol Using Iron Oxide Nano-/TEMPO Catalyst 108 Rufinamide Synthesis 110 Several High-T, High-p Processes 111 Click Chemistry 111 4-(Pyrazol-1-yl) Carboxanilide Multistep Synthesis 112 4-Hydroxy-2-cyclopentenone Synthesis 113 Hydrothermal Treatment of Glucose 114 Tetrahydroisoquinoline Synthesis 114 Solvent-Scope Widening by Virtue of Pressurizing Existing High-T Reactions 116 Nucleophilic Aromatic Substitution of 2-Halopyridines 116 Intramolecular Thermal Cyclization and Benzannulation 116 Catalyst-Free Transesterification and Esterification of Aliphatic and Aromatic Acids 117 Aminolysis of Epoxides 117 New Temperature Field for Product and Material Control 118 Palladium-Catalyzed Aminocarbonylation 118 Aminolysis of Epoxides 119 Flash Flow Pyrolysis 119 Indium Phosphide Nanocrystal 120 Quantum Dot Synthesis 121 High-T Flow Cycloaddition to Fullerene Derivatives 123 Energy Activation Other than Temperature – Photo, Electrochemical, Plasma 125

Contents

6.6.1 6.6.2

Photo-Oxygenation of Dimethylsulfide 125 Microwave Flow Reactor for Stable High-p,T Operation References 125

7

129 Tailored High-p Microreactor Design and Fabrication 129 Solder-Based Silicon Microsystem 129 In-Plane Fiber-Based Interfaced Microreactor 129 High Pressure to Intensify Interfacial Transport in Gas–Liquid Reactions 130 Hydrogenation of Cyclohexane 130 Carbamic Acid Formation 130 Intramolecular Aldol Condensation to 1-Methyl-1-cyclopenten-3-one 131 Catalytic Hydrogenation of Acetone 131 Propylene Oxide Synthesis 132 Asymmetric Amino-2-indanol Hydrogenation 132 Hydrogen Gas Liquefication in Guaiacol Conversion (hydroprocessing) 132 Pressure as Direct Means – Activation Volume Effects and More 133 Claisen Rearrangement 133 Nucleophilic Aromatic Substitution of Three p-Halonitrobenzenes 133 Diels–Alder Reaction with Furylmethanols and Cyclopentadiene 134 Aza Diels–Alder Reaction 136 Esterification of Phthalic Anhydride 136 Pressure for Advanced Fluidic Studies – to be Used for Shaping Materials and More 137 scCO2 Droplets or Jets in Liquid Water 138 References 138

7.1 7.1.1 7.1.2 7.2 7.2.1 7.2.2 7.2.3 7.2.4 7.2.5 7.2.6 7.2.7 7.3 7.3.1 7.3.2 7.3.3 7.3.4 7.3.5 7.4 7.4.1

8

8.1 8.1.1 8.1.2 8.2 8.2.1 8.2.2 8.2.3 8.2.4 8.2.5 8.2.6

125

Press – High-p Processing

Collide and Slide – High-c and Tailored-Solvent Processing 141 Batch Process-Based Inspirations for High-c Flow Processes 141 Polypropylene and Polycarbonate Polymerizations 141 Enantioselective Thermal and Photochemical Solid-State Reaction 142 Solvent-Free or Solvent-Less Operation – “Highest-c” 142 Bromination of 3-Bromo-imidazo[1,2-a]Pyridine 142 Thiophene Bromination 142 Claisen Rearrangement of Substituted Phenyl Phenols 142 Michael Addition 143 Peroxidation of Methyl Ethyl Ketone 144 Beckmann Rearrangement (High-c) 144

XI

XII

Contents

8.2.7 8.2.8 8.2.9 8.2.10 8.2.11 8.2.12 8.2.13 8.3 8.3.1 8.3.2 8.3.3 8.3.4 8.3.5 8.3.6 8.3.7 8.3.8 8.3.9 8.3.10

9

9.1 9.1.1 9.1.2 9.1.3 9.1.4 9.1.5 9.1.6 9.1.7 9.1.8 9.1.9 9.2 9.2.1 9.2.2

[2+2] Photocycloaddition of a Chiral Cyclohexenone (High-c) 145 Bromination of Toluene (Solvent-Free) 148 Sulfonation of Nitrobenzene (Solvent-Free) 148 Synthesis of Nitro Herbicides (High-c, Solvent-Free) 148 Suzuki–Miyaura Reaction over Sol–Gel Entrapped Catalyst SiliaCat DPP-Pd 150 Enzyme and Coenzyme (High-c in Bioprocessing) 150 Enzyme and Coenzyme (High-c in Bioprocessing) 151 Supercritical Fluids to Combine the Former Separated – Mass Transfer Boost 152 Supercritical Hydrogenation of Cyclohexene 155 Supercritical Hydrogenations of Double and Triple Bounds 155 Ascaridole Synthesis under Photo-Supercritical Conditions 157 Citronellol Oxidation under Photo-Supercritical Conditions 157 Near-scCO2 Enzymatic Biodiesel Synthesis 157 Supercritical Water, Non-Catalytic Beckmann Rearrangement 158 Supercritical Water, Non-Catalytic Pinacol Rearrangement 159 Supercritical Water Oxidation 159 Supercritical-Acetonitrile, Nitriles from Carboxylic Acids 162 Self-Optimizing Continuous Reactions under Supercritical Conditions 162 References 163 Doing More by Combining – Process Integration 165 Integration of Reaction and Cooling/Heating, Separation, or Other 165 Integrated Micro-Steam Reformer-Catalytic Combustor for Methane Fuel Processing 165 Integrated Microburner/Thermoelectric Device for System Start-Up 166 Integrated Micro Reactor–Evaporative Cooler 167 Integrated Microwave–Microreactor 167 Integrated Enzyme Microreactor–Extractor 167 Integrated Membrane Microreactor for Knoevenagel Reaction 168 Continuous Multiple Liquid–Liquid Separation: Diazotization of Amino Acids 169 Coupling of the Hydroxylation of Progesterone Using Rhizopus Nigricans with Flow Extraction 169 Coupling of the Esterification to Isoamyl Acetate Using Lipase B with Flow Extraction 170 Integration of Process Control and Sensing 171 Integrated Process Control for Methanol Steam Reforming 171 Integrated Sensing, Catalyst, and Heating for Ammonia Oxidation 171

Contents

9.2.3 9.2.4 9.2.5 9.2.6 9.3 9.3.1 9.4 9.4.1 9.4.2 9.4.3 9.4.4 9.5 9.5.1

Integration of the Esterification to Ethyl Oleate Using Lipase B with Photoionization Mass Spectrometry 172 Integration of the Intramolecular Friedel–Crafts Addition with Ultra-High-Pressure Liquid Chromatography 173 Integration of Pyrane Flow Reaction and Synchrotron-Based IR and X-Ray Beam Analysis 174 Integration of Multistep Organic Transformations Catalyzed by Au Nanoclusters 176 Thermal Integration on a Process Level 177 Thermal Integration of a Methanol Micro-fuel Processor/Fuel Cell 177 Integration of Units on Racks, Backbones, Frames, Interfaces, or Similar Level 178 Integrated Circuit Socket for Fluidic-Electric Interface 178 Chassis-Type Unit Racking and System Automation 178 Modularization of (Flow) Plant Equipment 179 Modular Gas–Liquid Microreactor 179 Fully Intensified/Flow Process Development 179 Adipic Acid Large-Scale Manufacture 179 References 183

10

Doing the Same with Less – Process Simplification 185

10.1 10.1.1 10.1.2 10.1.3 10.2 10.2.1 10.2.2 10.2.3

Omitting the Use of a Catalyst 185 Air Oxidation of Cyclohexane 185 Bromination of Toluene 185 Tetrazole Click Chemistry 186 Simplifying Separation 186 Phenyl Boronic Acid Synthesis 187 Simplifying Operation 188 Trans-1,2-Cyclohexanediol – Exothermic Steps Done “All-in-Once” 188 Olefin Autoxidation 189 References 190

10.2.4

11

11.1 11.2

11.3 11.3.1 11.3.2 11.3.3 11.3.3.1 11.3.3.2

Implications of NPW to Green and Cost Efficient Processing 191 Introduction 191 Knowledge-Based Design of Future Chemistry – Coupling the Implementation of NPW with Evaluation and Decision Support Tools 192 Evaluation Methods 192 Single Metrics 193 Holistic Approaches 194 Life Cycle Costing 201 Cradle-to-Gate Approach in LCC 202 Calculation of Costs 204

XIII

XIV

Contents

11.3.3.3 11.3.4 11.4 11.4.1 11.4.1.1 11.4.1.2 11.4.2 11.4.2.1 11.4.2.2 11.4.2.3 11.4.2.4 11.4.2.5 11.4.2.6 11.4.2.7 11.4.2.8 11.4.2.9 11.4.3 11.4.3.1 11.4.3.2 11.4.3.3 11.4.4 11.4.4.1 11.4.4.2 11.4.4.3 11.4.4.4 11.4.4.5 11.4.4.6 11.4.4.7 11.5

Separation of Costs in Variable and Fixed Costs 206 Summary of Life Cycle-Based Evaluation Methods 206 Evaluation of the NPW Concept Impact on Sustainability 206 Evaluation of New Chemical Transformations 207 One-Pot Multistep Synthesis of Fine Chemicals 207 Avoidance of Waste Products 207 Evaluation of High-Temperature, High-Pressure Operation 208 p-Xylene Partial Oxidation under Harsh Conditions 208 Process Intensification of Biodiesel Generation 209 Generation of Carbon Nanotubes at High Temperatures 211 Activation of Carbon Dioxide 213 Impact of Harsh Process Conditions of Catalyst Deactivation 215 Methane Decomposition at High Reaction Temperatures 216 Synthesis of Fullerenes by Pyrolysis versus Plasma 217 Epoxidation Reaction at Accelerated Temperature 219 Trade-Off Designs for a Hydrocarbon Biorefinery 220 Reduction and Replacement of Solvents 222 Sildenafil Citrate Process 222 Suzuki–Miyaura Cross-Coupling 224 Synthesis and Application of Ionic Liquids 225 Evaluation of Process Integration 226 Acceleration of Multiphase Reactions via Ultrasound 226 Process Optimization of a Pharmaceutical Synthesis 229 Process Simplification of Adipidic Acid Synthesis 231 Process Enhancement via Phase Transfer Catalysis 232 Separate Step versus Coproduction for Methanol Production 234 Process Integration in Terms of Costs 234 The Effects of Modular Plants 235 Future Environmental and Economic Sustainability Evaluation in the Context of Flow-Chemistry under NPW Conditions 236 References 238

12

From Milligrams to Kilograms – Scale-Up in Modular Flow Reactors 241

12.1 12.2 12.2.1 12.3 12.3.1 12.3.2 12.4 12.4.1 12.4.2 12.4.3 12.4.4

Reactor Types 241 Scale-Up Parameters 243 Geometry 244 Numbering-Up 246 Internal Numbering-Up 246 External Numbering-Up 246 Single-Channel Operation 247 Fluid Dynamics in a Rectangular Channel 247 Mean Residence Time and Its Distribution 248 Pressure Loss and Mixing 252 Heat Transfer in Channel Reactors 254

Contents

12.4.5 12.5 12.5.1 12.5.2 12.5.3 12.5.3.1 12.5.3.2 12.5.3.3 12.5.3.4 12.6

13

13.1 13.1.1 13.1.2 13.1.3 13.1.4 13.1.5 13.1.6 13.1.7 13.2 13.3 13.3.1 13.3.2

14

14.1 14.1.1 14.1.2 14.1.3 14.1.4 14.2

Parametric Sensitivity and Reactor Thermal Stability 256 Methodology for Continuous-Flow Process Development 258 General Chemical Plants and Modular Setup 261 Platform Concept and Scalability 264 Modular Process Development 267 Module A: Feasibility Study – Milestone Decision 268 Module B: Process Synthesis and Optimization – Milestone Decision 270 Module C: Process Robustness and Economy – Milestone Decision 273 Module D: Pilot Production and Commercial Manufacturing 274 Conclusions 278 References 280 283 Multifaceted Novel Process Windows: Evolution 283 Novel Chemistry – Liberation of Chemical Potential (2005) 283 NPW-Route Classification and Experimental Demonstration (2009) 283 NPW-Reaction Compilation Out of One Source – 1 (2010) 284 NPW-Reaction Compilation Out of One Source – 2 (2010) 284 Flow Chemistry High-T Overview – Superheated Processing (2010) 284 Flow Chemistry High-T Overview – from-Cryo-to-Ambient Processing 285 NPW-Methodology and Intensification Considerations for High-T Flow Reactions (2011) 285 High-p,T Commercial Flow Chemistry Equipment 286 Funding Agency Initiatives 286 DBU Cluster Novel Process Windows 287 EU Funding 289 References 291 Evolution of Novel Process Windows

Scientific Dissemination of Novel Process Windows 293 Literature Share for Chemical Intensification 293 Search Methodology and Limitations 293 High-Temperature, High-Pressure, High-Concentration in Chemistry and Chemical Engineering Literature 294 High-T, -p, -c in Microreactor Literature 295 High-(T, p, c) and NPW Importance in Current Microreactor and Overall Literature 298 Literature Share for Process-Design Intensification 298 References 300

XV

XVI

Contents

15

15.1 15.1.1 15.1.2 15.2 15.2.1 15.2.2

Outlook 301 Process Automation 301 Computer-Controlled Flow Processing 301 Process-Automated Application Example 301 Means of Activation Other than High-Temperature, High-Pressure, High-Concentration, and High-Solvent 303 Photoactivation 303 Activation by Instable Catalyst Precursors 305 References 307 Index 309

XVII

Motivation – Who Should Read the Book!? This book may be approached by two kinds of readers

• with background knowledge in microreactors and process intensification in general which want a further deeping in a central enabling function in for these areas • without such background knowledge and primary interest in flow processing, yet with interest in green chemistry and green engineering. The aim is to get to know the key enabling function of flow processing and to transfer that knowledge to advanced green batch processing and synthesis. Novel Process Windows as defined in this book need microreactors to be performed in the best way. Yet, the general approach – in the same way as given for process intensification, as defined below – can provide learning also for other advanced chemical synthesis and processing, including with other advanced chemical apparatus. For the second class of readers, a short summary is given below on microreactors, for the first class on process intensification and green processing. The references help to find more comprehensive information in case more deeper and widespread information is wanted. Both classes of readers may check the summaries they are familiar with as well, for consistency and completeness. There are several books on micro process engineering, including those written or edited by the authors of this book. These books provide enabling the basic engineering enabling functions for flow chemical synthesis and processing such as the increased mass and heat transfer. Yet, knowing this without having considered the information here is like studying formula-1 racing cars, their construction principles and capabilities, however, to operate them still with normal tankstation fuel. In terms of such a picture, this means to run expensive, much advanced equipment under much reduced performance. Microreactors as used under such boundary conditions still have many advantages as given below. Yet, they have one prime disadvantage (apart from blocking sensitivity) which is their short contact time as an intrinsic feature. Organic reactions as we know are typically processed on much longer times. One the one side, there is seconds scale, and on the other side, often hours and days scale – seemingly a misfit.

XVIII

Motivation – Who Should Read the Book!?

Indeed, it was prognosed from industrial microreactor experts that only 20% of all organic reactions are suitable for flow under such transfer-intensified operation (see citation [1] in Chapter 2). Using the approach provided in this book gives a further enabling chemicalsided push so that the authors roughly estimate that >70% of all organic reactions may become suitable for flow that is, in a timescale which is appropriate to microreactors. The price for this is a demanding and laborious reinvention of chemistry beyond the green chemistry needs. This must honestly be stated. Yet, there will be payback for this. The second process-sided push does allow – on an industrial scale – to truly benefit from such an advantage in application expansion. For both reasons, not considering this book’s knowledge does lead to an incomplete picture first of all in modern flow chemistry and micro processing engineering – yet finally also in green chemistry, green engineering, and process intensification. Thus, the authors believe to hopefully provide one more essential brick in the whole scenario of modern advanced and sustainable processing. Reference 1. Roberge, D.M., Ducry, L., Bieler, N.,

Cretton, P., and Zimmermann, B. (2005) Chem. Eng. Technol., 28, 318–323.

XIX

Acknowledgments Volker Hessel gratefully acknowledges funding provided by the Advanced European Research Council Grant “Novel Process Windows – Boosted Micro Process Technology” under grant agreement number 267443. He also acknowledges financial support from the Deutsche Bundesstiftung Umwelt. He likes to thank Tobias Illg for assistance with the preparation of Sections 3.1 and 3.2. Volker Hessel likes to thank IMM (Mainz, Germany) and Eindhoven University of Technology (The Netherlands) for offering the opportunities to work on the field of Novel Process Windows. This involves in particular Holger Loewe, Josef Heun, Patrick Loeb, Gunther Kolb, and Frank Hainel (all from IMM) and Jaap Schouten, Hans Niemantsverdriet, and Laurent Nelissen (all from TU Eindhoven). Dana Kralisch gratefully acknowledges the support of Ina Sell in all aspects of Life Cycle Cost analyses. Norbert Kockmann likes to acknowledge the help and support of Lonza colleagues, notably Dominique Roberge, for fruitful discussions on micro reactor development and application. He wants also to thank Christian Bramsiepe from TU Dortmund University, now INVITE GmbH in Leverkusen, for intensive discussions on modularization.

XXI

Abbreviations AP BOC CED DMP EDL EHS EP GWP HTP LCA LCC LCIA LCI LU NMP NP NPV ODP POCP RA RTD SLCA THF

acidification potential chemical protecting group, tert-butyloxycarbonyl cumulative energy demand dimethoxypropane electron diffusion layer environment, health and safety ecotoxicity potential global warming potential human toxicity potential life cycle assessment life cycle costing life cycle impact assessment life cycle inventory land use n-methylpyrrolidone eutrophication potential net present value (stratospheric) ozone depletion potential (tropospheric) photochemical ozone creation potential risk assessment residence time distribution simplified life cycle assessment tetrahydrofuran

XXIII

Nomenclature AC Aref AS a Bo b Cm c c0 cp Dm Dax Dchar De d dh di Ea −ΔHR h Kn k k0 kB kR lK ΔL L Lm m m

Cross-sectional area (m2 ) Reference area for heat transfer (m2 ) Wetted surface area (m2 ) Temperature conductivity = 𝜆∕𝜌 cp (m2 s−1 ) Bodenstein number (–) Channel width (m) Mixing coefficient (–) Molar concentration, volume based (mol l−1 ) Starting concentration (mol l−1 ) Heat capacity (J kg−1 K−1 ) Molecular diffusion coefficient (m2 s−1 ) Axial diffusion coefficient (m2 s−1 ) Characteristic diameter (m) Dean number (–) Diameter (m) Hydraulic diameter (m) Inner diameter (m) Activation energy (J mol−1 ) Molar reaction enthalpy (J mol−1 ) Channel height (m) Knudsen number (–) Overall heat transfer coefficient (W m−2 K−1 ) Reaction rate constant, unit depends on reaction order, for m = 1 (s−1 ) Boltzmann constant (J kg−1 K−1 ) Reaction rate coefficient, unit depends on reaction order, for m = 1 (s−1 ) Kolmogorov length scale (m) Length difference (m) Channel length (m) Length of mixing channel (m) Reagent mass in the reactor (kg) Reaction order (–)

XXIV

Nomenclature

ṁ N n n ṅ P Pe p Δp Q̇ Q̇ R q̇ h q̇ reac R RC R Re r r Sc T TC ΔT t tm tP tR UV U w V V̇ Vm VS ΔVi ≠ ΔVi 0 X Ẋ x xs Y z

Mass flow rate (kg s−1 ) Number of vessels (–) Exponent (–) Molar amount (mol) Molar flow rate (mol s−1 ) Wetted perimeter (m) Peclét number (–) Pressure (Pa) Pressure difference (Pa) Transferred heat (W) Reaction heat (W) Heat flux (W m−2 ) Heat generated from chemical reaction (W m−3 ) Universal gas constant (J mol−1 K−1 ) Channel or tube bend radius (m) Reaction component (–) Reynolds number (–) Channel or tube radius (m) Reaction rate (s−1 ) Schmidt number (–) Temperature (K) Cooling temperature (K) Temperature difference (K) Time (s) Mixing time scale (s) Characteristic process time, residence time (s) Characteristic reaction time (s) Volumetric heat transfer coefficient (W m−3 K−1 ) Conversion (–) Mean (fluid) flow velocity (m s−1 ) Device volume (m−3 ) Volumetric flow rate (m3 s−1 ) Internal mixing channel volume (m3 ) Wetted volume or channel volume (m3 ) Difference of partial molar volumes (m3 ) Standard difference of partial molar volumes (m3 ) Reaction progress or conversion (mol) Reaction progress rate (mol s−1 ) Length scale (m) Solubility (–) Yield (–) Coordinate in channel length (m)

Nomenclature

Greek Symbols

E 𝜀p 𝜂 𝜃 𝜆m 𝜇 𝜈 νi P ρm 𝜎 𝜎𝜃2 𝜏

Energy dissipation rate (m2 s−3 ) Dielectric constant (F m−1 ) Dynamic viscosity (N s m−2 ) Contact angle (rad) Friction factor in the mixing channel (–) Dipole moment (C m) Kinematic viscosity (m2 s−1 ) Stoichiometric coefficient (–) Density (kg m−3 ) Molar density (mol l−1 ) Surface tension (N m−1 ) Variance of dispersion (–) Time (s)

Note: Here, the ISO notation is used for international clarity, such as 𝜂 for the dynamic viscosity, 𝜆 for heat conductivity, 𝛼 for heat transfer coefficient, or k for the overall heat transfer coefficient, see also [Hesselgreaves, J.E. (2001) Compact Heat Exchangers, Elsevier, Amsterdam].

XXV

1

1 From Green Chemistry to Green Engineering – Fostered by Novel Process Windows Explored in Micro-Process Engineering/Flow Chemistry 1.1 Prelude – Potential for Green Chemistry and Engineering

Green Chemistry is since about 20 years an approach which is meanwhile quite established in chemical research and education [1]. Experts predict a fast-paced growth of the market for Green Chemistry-type processing, from $2.8 billion in 2011 to $98.5 billion in 2020 (Pike Research study [2]). Finally – yet probably not before the next 20 years – experts expect this to eliminate the need for Green Chemistry as an own approach, since it will be identical to the chemistry in the future. Anastas and Kirchhoff [3] bring this to the point in “Origins, Current Status, and Future Challenges of Green Chemistry” as follows. The revolution of one day becomes the new orthodoxy of the next. Green Engineering followed somewhat later and just recently came out the shadow of “its big brother.” The implementation of that idea in the chemical industry proceeds now steadily, yet for reasons of complexity of the chemical processes, unavoidably at a slow rate. Today, only 10% of the current process technologies employed on industrial scale can be considered environmentally benign. It is estimated that another 25% could be made so. That leaves room for exploring and discovering the residual 65% of industrial process technology and to render them sustainable [4]. That means that there is still a considerable need to improve the enabling technologies which render chemical synthesis and chemical processes green. Under the umbrella of process intensification, microreaction technology and flow chemistry are prime enablers on the reactor and process side (see later in this chapter for citations). They help to improve current Green Chemistry approaches and in addition even give opportunities to develop new Green Chemistry concepts, which are not possible with conventional equipment. On top of that, micro- and milli-continuous processing provides a more straightforward way to upscale new green ideas. Seeing the last paragraph and the achieved 10% penetration, this is obviously still an open issue.

Novel Process Windows: Innovative Gates to Intensified and Sustainable Chemical Processes, First Edition. Volker Hessel, Dana Kralisch and Norbert Kockmann. © 2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

2

1 From Green Chemistry to Microreactors

In continuation of the above given aphorism, this book shall open a window from Green Chemistry to Green Engineering as follows. The revolution in the chemical laboratory needs to stimulate and bridge to the sustainability evolution on the full-production scale. [5] 1.2 Green Chemistry

Driven by political as well as societal demands, sustainability aspects gain increasing importance in all areas of human beings. Chemical production of compounds, for example, textiles, construction, ingredients in food and cosmetics, packaging, pharmaceuticals, and so on, covers more or less all aspects of human needs. The resulting extensive impact on our environment and consumption of depletable resources distinctly demands for the most efficient use of raw materials and energy. Pollution has to be prevented or at least minimized at the source to avoid end-of-pipe treatments. New concepts have to come off with significant benefits, for example, in yield, selectivity, heat management, waste reduction, to become an environmentally benign alternative to the state of the art. Also, the environmental burdens of any reaction component, auxiliaries, and energies, obtained during upstream processes, as well as all downstream processes involved have to be taken into account. All this has stimulated an on-going and total rethinking how to change the elemental pathways of chemical synthesis design, which has become a large movement and created an own scientific filed and society known as Green Chemistry. While processes in the past were guided by economic, technical, and safety criteria, it is now becoming increasingly obvious and a to-do-must to have considered environmental criteria from the very beginning of the process development – which is the creative intuition of the organic chemist how to conceptually approach synthetic chemistry. 1.2.1 12 Principles in Green Chemistry

In one sentence, Green Chemistry was defined as follows [2]. Green chemistry is the utilization of a set of principles that reduces or eliminates the use or generation of hazardous substances in the design, manufacture, and application of chemical products. In kind of tabellaric goal definition, Green Chemistry was defined as follows [1b, p. 30]. 1) 2) 3) 4)

Prevention Atom economy Less hazardous chemical syntheses Designing safer chemicals

1.3

5) 6) 7) 8) 9) 10) 11) 12)

Green Engineering

Safer solvents and auxiliaries Design for energy efficiency Use of renewable feedstocks Reduce derivatives Catalysis Design for degradation Real-time analysis for pollution prevention Inherently safer chemistry for accident prevention.

Essentially, one can reduce that to three major incentives which are to optimize the type of feedstock, its efficiency in conversion, and the safety while doing so (derived from own thoughts and [2]). 1) Feedstock: A shift to renewable (non-petroleum) feedstocks 2) Efficiency: (i) make maximal use of starting materials (reactants) and minimize waste; (ii) minimize solvent load; and (iii) minimize energy efficiency 3) Safety: have maximal process safety and minimize toxicity (to human). Ideally, supposed-to-be nongreen reagents just vanish from the chemical protocol by using a new chemical route such as given for GSK (Glaxo-SmithKline)’s green Friedel–Crafts alkylations [6]. Manifold applications have been demonstrated with respect to modern synthetic strategies, alternative solvents, renewable resources, catalysis, and environmental-friendly enzymatic catalysis in flow [1c-e, 7]. 1.3 Green Engineering 1.3.1 10 Key Research Areas in Green Engineering

In 2005, the American Chemical Society (ACS) Green Chemistry Institute (GCI) and global pharmaceutical companies established the ACS GCI Pharmaceutical Round-table to motivate for integration of Green Chemistry and Engineering into the pharmaceutical industry [8]. This Roundtable developed a list of key research areas in green chemistry in 2007. In 2010, the Roundtable companies have identified a list of the key green engineering research areas that is intended to be the required companion of the first list. The companies involved were Boehringer Ingelheim Pharmaceuticals, Pfizer, Eli Lilly, GlaxoSmithKline, Dutch State Mines/De Staats Mijnen (DSM), Johnson & Johnson, AstraZeneca, and Merck (US). Ten key green engineering research areas were ranked in relevance (see Table 1.1). The issues 6–10 match with what is understood under process-design intensification in this book – (6) life-cycle analysis, (7) integration of chemistry and engineering, (8) scale-up, (9) process energy intensity, and (10) mass and energy integration. The key areas 1–5 in Table 1.1 refer partly to chemical intensification.

3

4

1 From Green Chemistry to Microreactors

Table 1.1 Ten prime green engineering research areas as identified by ACS GCI Pharmaceutical Round Table (reproduced with permission). Rank

Main key areas

Sub-areas/aspects

1

Continuous processing

2

Bioprocesses

3

Separation and reaction technologies Solvent selection, recycle, and optimization

Primary, secondary, Semi-continuous, and so on Biotechnology, fermentation, biocatalysis, GMOs Membranes, crystallizations, and so on

4

5 6

Process intensification Integration of life cycle assessment (LCA)

7 8 9

Integration of chemistry and engineering Scale-up aspects Process energy intensity

10

Mass and energy integration

Property modeling, volume optimization, recycling technologies, in process recycle, regulatory aspects, and so on Technology, process, hybrid systems, and so on Life cycle thinking, total cost assessment, carbon/eco-foot printing, social LCA, stream lines tools Business strategy, links with education, and so on Mass and energy transfer, kinetics, and others Baseline for pharmaceuticals, estimation, energy optimization Process integration, process synthesis, combined heat and power, and so on

Adapted with permission from [8]. Copyright 2012 American Chemical Society.

1.3.2 12 Principles in Chemical Product Design

A product-design view is provided by the 12 Principles of Green Engineering which were proposed by Anastas and Zimmerman [9]. Sustainability is here approached in a hierarchical crossover between the molecular, product, process, and system levels. 1) Inherent rather than circumstantial Designs of chemical processes shall be so much efficient and nonhazardous as possible. Example is a process to synthesize organic solvents from sugars, which has replaced many more hazardous solvents such as methylene chloride (Argonne National Laboratory). The very low energy input, high efficiency, elimination of large volumes of salt waste allows to reduce pollution and emissions. 2) Prevention instead of treatment The production of waste shall be prevented rather than planning process including waste treatment.

1.3

Green Engineering

3) Design for separation The quest of energy and material efficiency is not only to be put on reaction but rather on separation and purification processes as well. Example is the use of supercritical CO2 , nearcritical water, and CO2 -expanded liquids, which in addition comprise nontoxic substitutes for conventional solvents. 4) Maximize efficiency It shall be aimed at maximum efficiency in terms of mass, energy, space, and time. Example is a highly efficient family of catalysts to synthesize highperformance polymers from CO and CO2 . 5) Output-pulled versus input-pushed Outputs (“pull”) should be removed from the system, rather than adding more input stresses (“push”) to minimize the energy and material consumption. 6) Conserve complexity Different from the system design issue given in (5), the product design shall be complex. Idea is that the products then can have longer reuse times through better recycling than given for less complex products. 7) Durability rather than immortality Product design shall guarantee for the product lifetime, but no longer to avoid environmental problems. Example is cellulose acetate, which is used for the filters of cigarettes. Their decomposition after use took years in the past. This could be substantially improved by incorporation of weak organic acids in the material. These are released with rain water and degrade the cellulose acetate much quicker. 8) Meet need, minimize excess The production amount should be set just to meet the needs and not to result in over-production that creates wastes and is costly. 9) Minimize material diversity Recycling and reusing is much facilitated when fewer materials compose a product. Example is a “unibody” piece of aluminum laptop frame that considerably reduced the product’s weight and allowed for easy recycling (Apple Inc). 10) Integrate material and energy flows The utilization of waste energy and material flows can be used to improve the efficiency of another part of the production process. Example is the use of CO2 to replace traditional blowing agents for the production of polystyrene foam sheets. The CO2 used came from existing commercial processes as a waste by-product or from natural sources (Dow Chemical Company). 11) Design for commercial “afterlife” Selection of materials or components of a process, product, or system should head for reusability and keeping high value after fulfilling their initial product function.

5

6

1 From Green Chemistry to Microreactors

12) Renewable rather than depleting With a similar intention, renewable sources should be used for energy, materials, or reagents, wherever possible. Example is a water-based, catalytic process for producing gasoline, jet fuel, or diesel from biomass with little external energy consumption (Virent Energy Systems Inc). 1.4 Micro- and Milli-Process Technologies

Micro- and milli-process technologies refer to (chemical) processing with reactors and other equipment with open internals in the micro ( 10 Free molecular motion

Kn < 0.01 Continuum

105 Free surface

Thermal motion diffusion Molecules, cluster colloids

Quantum, decoherence

Liquid

Gas – liquid T, 𝜂

Electrons Photons

𝜎

100 μm

?

Inner circulation

𝜎, 𝜂

100 μm

Elliptic bubble

x =

Chaotic bubble

100 μm

Emulsions 𝜎W2

Brownian motion

Atoms, crystall lattice

Solid

Laminar – vortex – turbulent Spherical Scope bubble

Molecule

2Dt EDL

Solitons Phonons

L (m)

100

Macro

𝜗W

𝜎W

𝜗W

𝜎W

𝜗W

𝜎12

Liquid – solid

𝜎W1

Grain structure

Continuum

Liquid

Mechanical, acoustic

Continuum

Electric

Continuum, magnetic dipole

Magnetic Optic

Figure 3.1 Length scale and material properties together with phase behavior. (Kockmann [2]; reproduced by permission of Springer).

equation [2]: x2 = 2 D t

(3.1)

The transport length by diffusive mixing in gases (D = 10−5 to 10−6 m2 s−1 ) and in liquids with low diffusivity (D = 10−9 to 10−10 m2 s−1 ) is displayed over the corresponding time in Figure 3.2. Conventional equipment has typical geometries in the range of centimeters and produces fluid structures in the range from 100 μm to 1 mm. The corresponding diffusion time in gases is slower than approximately 1 ms and in liquids in the range of 1 s. Microstructured devices with typical length scales from 100 μm to 1 mm provide fluid structures with length scales of approximately 1 μm. These small fluid structures lead to mixing times shorter than 100 μs in gases and approximately 1 ms in liquids. This is the main reason for the enhanced selectivity and high yield of chemical reactions in microreactors besides good temperature control and shortening residence time to what is kinetically needed. The impact is larger for liquids than for gases due to their low diffusivity. 3.2 Time Scale

The length scale often coincides with a time scaling of the relevant processes. In general, the shorter the length, the shorter the characteristic time for transport

3.2

103

Time

Liquids

Electron density of a peptide

Gases Diffusion Wall turbulence

1s Re>180

10–3 10–6

Time Scale

Stirred vessel Static mixer

Micromixer

Processintensification

Catalytic surface

10–9 10–12 10–15 Electron density H2O and CO2

10–18

μm

nm

mm

m

Length

Figure 3.2 Length and time scale in transport phenomena, from molecular scale to macroscopic transport processes. (Adapted from Kockmann [2]).

processes will be, and the higher the transformation frequencies. The diffusion of a species in a surrounding fluid displays this process and is described in Eq. (3.1). The typical diffusion length within 1 s is approximately 7 mm in gases (air) and approximately 70 μm in liquids such as water. Similarly, the conduction length can be derived from the basic balance equation for the momentum: √ (3.2) xp = 2 ν t with the kinematic viscosity 𝜈 and the heat transfer √ xq = 2 a t

(3.3)

with the temperature conductivity a. The characteristic time is proportional to the square of the length variation and to the transport coefficient. For turbulent flow, the finest vortex length scale is given by the energy dissipation rate 𝜀 and the kinematic viscosity 𝜈: ( lK =

ν3 𝜀

)1 4

(3.4)

This length scale is also called Kolmogorov length scale and can be rearranged to a time scale with Eq. (3.2). Additionally, the typical length and time scales for fast chemical reactions such as neutralizations or slow chemical reactions such as complex formation or polymerization can be given as well as typical scales for micromixers to compare the processes. The mass transfer in micromixers acts on a length scale of a few microns within milliseconds or less. Different time scales

27

28

3 Chemical Intensification – Fundamentals

are typical for partial reactions in complex chemical reaction systems. With properly designed micromixers and an adjustment of the component concentration, the selectivity of a complex reaction can be increased dramatically. The mixing time is only one criterion for an optimal chemical reaction, another is the scale of fluids residence time within the device. Within small devices, the fluids rest only briefly (seconds or less), which can be detrimental to slow reactions. A good overview of residence time, backmixing, and heat transfer in microchannels is given by Knösche [3].

3.3 Length and Time Scale of Chemical Reactions

The macroscopic view of chemical reaction includes a typical length and time scale, the former given by the reaction space and related transport conditions. The time scale of a chemical reaction, if not governed by the transport conditions, is described by the reaction rate and its related kinetics. The chemical reaction is formally described by its stoichiometric equations: s ∑

νj Rj = 0

(3.5)

j=1

with R as reaction component and ν as stoichiometric coefficient. The reaction progress is the change of a key component j with time: X=

Δnj νj

=

nj − nj,0 νj

or for continuous processes with molar flow rate: Δṅ j ṅ j − ṅ j,0 Ẋ = = νj νj

(3.6)

(3.7)

Conversion of a reagent is the consumed amount related to the starting amount: ṅ j,0 − ṅ j (3.8) Uj = ṅ j,0 When all of the starting material can be converted into the product, its yield is defined as: ṅ YP = P (3.9) ṅ j,0 The amount of species converted is determined by the reaction rate – the change of species amount per time unit by chemical reaction: r=

d (cj V ) 1 1 d nj 1 1 = d τ νj V d τ νj V

(3.10)

Here, the concentration c is defined as the species amount per volume (e.g., moles per liter) and can be translated into other units such as weight-based units

3.3

Length and Time Scale of Chemical Reactions

(e.g., grams per liter or moles per gram total) by the molar density ρm . The reaction rate is described by the kinetic relation including the rate coefficient, the concentration, and the reaction order m: ṅ j − ṅ j,0 r= (3.11) = kR ⋅ cm j νj ⋅ ρm V The influence of the temperature on the reaction will be discussed in one of the next sections in more detail. The characteristic reaction time indicates the time scale for comparison with mixing and heat transfer characteristics: tR =

1 kR cm−1 0

(3.12)

This time scale gives an approximate estimation for the time scale necessary in the process space. Main influencing parameters are the starting conditions, temperature, and reaction order. The residence time within the reactor must be larger than the characteristic residence time, while the mixing time should be shorter for high yield in mixing sensitive reactions. 3.3.1 Solution of Kinetic Equations

The temporal development of the concentration of starting material and product can be described analytically for simple cases. The temperature influence on the reaction rate is discussed in Section 3.4. Table 3.1 shows the correlation of the temporal concentration profile for reaction order m = 1 and others. For m = 0–2, four profiles of starting material (educt) over the time are shown in Figure 3.3. The temporal development of a chemical reaction can be described by its characteristic time scale. In Figure 3.3, typical time scales are shown for the concentration profile for a reaction with m = 1. The exponential decay of the starting material does not theoretically allow for full conversion, but in practice nearly full conversion (99% or similar) is assumed for reaction completion. Taking 50% conversion as criteria, the characteristic time will be tR = ln(2)∕k for constant rate Table 3.1 Solution of kinetic equations – the second line allows for calculation of many reactions orders. Conversion equation

Reaction order

Differential equation

} A →P A+B

1

kR c = −

A→P

n (n ≠ 1)

kR cn = −

A+B→P

2

kR cA cB = −

dc dt

Integration

Unit of kR

c = −k t c0 ( ) 1 1 1 − = −k t 1 − n cn−1 cn−1 0 cA,0 cB 1 ln = −k t cB,0 cA cA,0 − cB,0

s−1

ln

dc dt dcA dt

mol1−n ln−1 s−1 l mol−1 s−1

29

3 Chemical Intensification – Fundamentals

0.1

m=0

m = 0.5

m=1

m=2

ge

n Ta

0.06

nt

Concentration (mol l−1)

0.08

50% 0.04 37% 0.02

0 0

20

40

60

80

100

120

140

160

180

200

Time (s)

Figure 3.3 Starting material concentration of model reaction with constant rate coefficient kR = 0.01 (unit varies from reaction order) and different reaction order c0 = 0.1 mol l−1 .

coefficient k. When the initial rate is known, a further characteristic time scale can be determined with Eq. (3.8). At this time, more than 63% of the starting material is already consumed. Both characteristic reaction times are displayed in Figure 3.3. The influence of the reaction order and the rate coefficient is displayed in Figure 3.4. Nearly complete conversion after 10 s needs a rate coefficient that is four orders of magnitude higher for second order reaction than for zeroth order reaction. 0.1 m=0

m = 0.5

m=1

m=2

0.08 Concentration (mol l−1)

30

0.06

0.04

0.02

0 0

1

2

3

4

5

6

7

8

9

10

Time (s)

Figure 3.4 Starting material concentration of model reaction with constant conversion X = 99% (end concentration of product of 99%); rate coefficients for m = 0, 0.5, 1, 2 are 0.01, 0.0568, 0.4543, 96.28, respectively. The units depend on the reaction rate.

3.3

Length and Time Scale of Chemical Reactions

Generally, there are three methods to determine chemical kinetics from varying inlet concentration. It depends on the process setup and analytical protocol, which is best suited in continuous-flow investigations: 1) Integral method: Assume an order and derive the correlations between conversion and residence or reaction time from the balance equations. Appropriate data display can proof the correct assumption (measured data are lying on a straight line, for example). 2) Differential method: The incline of the concentration-time or concentrationspace-time run over time at minimum two positions give the reaction order and the rate coefficient. 3) Method of half times: The concentration is measured at different times. The ratio of half-time to quarter-time values gives an estimate of the reaction order. The first and third methods are well suited for whole number orders. For broken number orders, the differential method is one of the choices. Many and accurate data points are very helpful. A simple proof of first order reaction is the variation of the inlet concentration. Variation of the conversion with inlet concentration excludes first order reaction. If the conversion is constant for varying inlet concentrations, the reaction is definitely of first order. 3.3.2 Reaction Time and Reaction Classification

To facilitate continuous-flow reactor design, Roberge et al. [4] proposed a reaction classification scheme with three types according to their typical kinetics and related reactor characteristics, such as mixing and residence time. Type A reactions are very rapid, mixing controlled reactions, even in fast mixing microstructured devices. Slower reactions of type B with kinetically controlled rates in the range of 10 s to few minutes demand larger residence time in the reactor with appropriate temperature control. Here, a flexible modular reactor setup is very helpful to rapidly develop an appropriate process and perform an efficient reaction with considerable heat release. The third reaction class type C is of hazardous or autocatalytic nature and demands for small volume and excellent temperature control. In a microstructured device, the reaction can be started under harsh process conditions and safely be performed within the reactor. In addition, a fourth class type D can be introduced for reactions outside of the former categories, which still can be performed in continuous-flow setup [5]. With this reaction classification, the reactor setup can be designed and assembled with appropriate mixing elements and residence time modules for adjusted temperature profile. Mean residence time is given by the internal volume divided by the volume flow rate and is also called the process time. This characteristic time should be two to five times larger than the typical reaction time. The following example will demonstrate the influence of the residence time on the reaction yield.

31

3 Chemical Intensification – Fundamentals

3.3.3 Example for Reaction Time and Residence Time

Rapid reaction optimization is demonstrated for the synthesis of N-Bocbenzylamine for BOC (tert-butyloxycarbonyl) protection of benzylamine [6]. In a Cytos lab system with inline IR system, feed 1 consist of benzylamine and triethylamine (each with 0.8 M THF (tetrahydrofuran)), and feed 2 of di-tert-butyl dicarbonate (0.8 M THF). In Figure 3.5, the yield of the BOC-protected molecule is displayed over the residence time. With constant flow rate of 4 ml min−1 , the reactor volume was increased stepwise to yield longer reaction time. While only four data points were measured, an estimation of the reaction order can be performed. Assuming isothermal conditions with no temperature influence on the reaction rate, a trend analysis in Excel gives first hints on the reaction order. A logarithmic trend line does not fit the data points, since it overestimates the first incline and does not represent the slow decay at the end. A least square estimation with second order correlation gives a better trend and a rate coefficient of k = 0.0759 s−1 for the yield given in percentage. With little effort, kinetic information can be gained out of raw data. In more detailed investigations, it has to be tested whether the enlargement of the reactor or the increase of flow rate, when changing the residence time, has an influence. Mixing effects have to be excluded during this analysis. The long reaction time of more than 10 min probably excludes strong mixing effects. A discussion of the residence time distribution by the same authors can be found in [7] together with the impact on a reaction library of a Grignard reaction system.

100 95 90 Yield (%)

32

Yield Y 2nd order Log. (Yield)

85 80 75 y = 9.2509ln(x) + 77.212 R2 = 0.9801

70 65 60 0

2

4

6

8

10

12

Residence time (min) Figure 3.5 Yield over residence time and logarithmic trend for BOC group protection. (Data from Schwalbe et al. [6]).

3.4

Temperature Intensification

3.4 Temperature Intensification

Besides the reaction order and initial concentration, the temperature is the major parameter for the reaction rate, described by the Arrhenius correlation. The following part is based on a chapter already published in [8]. 3.4.1 Harsh Process Conditions

Novel Process Windows are often located outside from conventional pressure, temperature, and concentration range with intensified physical processes and accelerated chemical reactions within safety limits [9]. Microstructured devices can provide harsh process conditions with narrow temperature and concentration range for accelerated reaction rates, better selectivity, or enhanced safety. Continuous-flow processes allow for high volumetric throughput with constant product quality. Both measures combined have huge potential for higher productivity, less solvent use or even neat processes, new pathways for clean reactions, use of more active catalyst, or milder solvents allowing higher reaction temperatures for easier separation step. Safety is always obligatory for production processes well covered by closed handling of hazardous reagents in continuous-flow reactors. The process temperature plays the dominant role due to its nonlinear influence on the reaction rate, heat release, and further temperature increase accompanied with increased reaction rate [10]. This behavior is called parametric sensitivity, and can lead to uncontrolled reaction progress and reaction runaway, bad selectivity, or hazardous accidents in the extreme cases. With high temperature, high pressure is often associated to avoid solvent boiling. Additionally, pressure has a beneficial effect on reaction parameters, which is described later. Process conditions describe the physical and chemical parameters in the vicinity of a molecule in a certain process space. Main parameters are the molecular kinetic energy, total molecular density, and species amount and type, expressed by temperature, pressure, and concentration as macroscopic parameters of a chemical system. 3.4.2 New Temperature Windows

For pure solvents, temperature and pressure are combined in the vapor curve as a limit for performing liquid-phase reactions. The vapor pressure is often not dramatically changed with reagents added to the solvent. Figure 3.6 shows the pressure–temperature diagram of two typical solvents, ethanol as low-boiling solvent and n-methylpyrrolidone (NMP) as high-boiling solvent, together with their typical application ranges. Starting from lower left side, the blue area indicates cryogenic process conditions with retarded reaction

33

3 Chemical Intensification – Fundamentals

1.E+05 Novel process windows

1.E+04 Pressure (kPa)

34

Supercritical conditions

1.E+03 Superheated conditions

1.E+02

Cryogenic conditions

“Standard” conditions

1.E+01

1.E+00 100

Temperature (K)

VP EtOH

KRP

VP NMP

KRP

Solid

Solid

1000

Figure 3.6 Pressure–temperature diagram with vapor pressure of ethanol (EtOH) as lowboiling solvent and n-methylpyrrolidone (NMP) as high-boiling solvent. (Hessel et al. [8]; reproduced by permission of Wiley-VCH).

conditions for better selectivity and reaction control. This range is operated with low-boiling point solvents. Rapid reactions are performed under cryogenic conditions to slow down their rate and to subdue side reactions. Due to better temperature control and rapid heat transfer as well as relatively ‘faster mixing’ (as compared to slowed-down reaction), many of formerly cryogenic processes could successfully be operated under elevated temperatures (ex-cryo). The green area represents typical process conditions around room temperature up to the boiling point of the solvent under reflux. Solvent change to high-boiling liquid may speed up the reaction rate. This rate increase can be well under control in structured continuous-flow reactors also due to the excellent temperature control. Additionally, the amount of produced material is quite small and can be consumed immediately if necessary. Many reactions, which are still slow under reflux conditions, can be accelerated under superheated conditions with increased pressure (orange area). A pressure control valve at the outlet of the flow reactor controls the pressure in closed tubing arrangement or containment. Finally, supercritical conditions are also a candidate for enhanced reaction rates and increased solubility, and also allow for novel process windows [11]. Besides the temperature level, extremely rapid temperature changes can lead to novel process conditions and produce unstable intermediates. In gas-phase flow, temperature rates higher than 106 K s−1 were achieved for producing nanoparticles and aerosols [12]. Unusual temperature conditions are also achieved in plasma environment [13]. Microplasmas (plasmas generated in microstructured devices) generate a favorable reaction environment because of high densities of reactive ions, radicals, and free electrons. The authors describe a system for reforming of methanol into a hydrogen-rich product stream, useful for energy applications such as combustion or fuel cells.

3.4

Temperature Intensification

3.4.3 Reaction Rate – Arrhenius Equation

The reaction rate varies with the concentration (mass action law), the temperature according to Arrhenius’ equation, and the pressure. The Arrhenius equation is determined by the activation energy Ea and the rate coefficient A: k = Ae−Ea ∕RT

(3.13)

This nonlinear correlation is based on the collision theory of molecules and has no analytical solution for the reaction rate and temperature development in a continuous-flow reactor. The dependency of the reaction rate constant is different with regard to the pressure and temperature changes indicating that there are both medium and chemical coordinate changes coupled, which should be treated and investigated separately. The temperature of the reagents along the reaction channel is determined by the reaction rate, the flow rate, the heat capacity of the reagents, and the heat transfer to the ambient. The reaction rate changes with the reaction temperature according to the following correlation: ( ( )) ( )) ( ) ( EA T − T0 kR EA 1 1 = kR,0 ⋅ exp kR = kR,0 − = kR,0 ⋅ exp kR,0 R ⋅ T ⋅ T0 R T0 T (3.14) A rule of the thumb of process chemists is to halve the reaction time by increasing the temperature by 10 ∘ C. This general rule is only roughly valid, but depends on the reaction temperature itself and on the activation energy. Reformulation of Eq. (3.14) gives a more accurate correlation: ( ) kR EA 1 R 1 1 1 ⇔ = − ln 2 = − ln (3.15) kR,0 R T0 T EA T0 T For small temperature differences ΔT compared to the reactor temperature T0 , the above correlation can be simplified to ( ) R T02 kR EA 10 ΔT = exp ln 2 or = (3.16) EA kR,0 R T02 The above two correlations allow the estimation of the temperature increase for doubling the reaction rate or the rate increase for a temperature rise of 10 ∘ C. In both cases, the activation energy has to be known. Both equations can also be used to give a rough estimation of the activation energy from measured data. The temperature course along the reactor length is discussed by Renken et al. [14] for the exothermic synthesis of ionic liquids. The entire reactor consists of the caterpillar mixer, a microstructured device R1 with parallel microchannels (0.65 × 1 mm2 , 10.8 ml), a 1/8′′ capillary tube R2 (di = 1.75 mm, 6.7 ml), and finally a 1/4′′ tube (di = 4.37 mm, 13.8 ml) (Figure 3.7). The simulated temperature and conversion profile is displayed under the tubes. At the entrance of R2, a small temperature increase is visible from the exothermic reaction. Nevertheless, the

35

36

3 Chemical Intensification – Fundamentals Thermostat 1

Thermostat 2

TIC 005

Pump

TIC 006

PIR 001

Bath 1 MiM P1 M1

TIR 001

μ

Micromixer Pump

TIR 002

MSR

Reactor temperature, T (K)

380

(a)

370

TIR 004

Capillary tube 1/4* HE1 R3

PIR 004

μ

Microheat exchanger

Tc = 368 K

360 350

Tc = 343 K

340 330 320 1.0

Tc = 323 K

(b)

exp. exp.

0.9 Conversion, X

PIR 002

R2

PIR 003

R1 Microstructured reactor

DES P2

Capillary tube 1/8*

Bath 2 TIR 003

Tc = 343 K Tc = 323 K Tc = 368 K

0.8 0.7 0.6

Tubular reactor 1: dt = 1.75 mm Tc = 323 K

Tubular reactor 2: dt = 4.75 mm

0.5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Reactor volume, Vr (cm3)

Figure 3.7 Reactor setup with feed pumps, micromixer, microstructured residence time module for enhanced heat transfer, 1/8′′ and 1/4′′ capillary tubes, together with the

simulated temperature and conversion profile in the capillary tubes. (Renken et al. [14]; reproduced by permission of Elsevier).

reaction temperature is well under control. Increasing the cooling temperature from 323 to 343 K and finally to 368 K leads to full conversion without losing the control over the reaction temperature.

3.5 Pressure Intensification 3.5.1 Reaction Rate – Activation Volumes

Reaction rate with rate coefficient k and its pressure dependence is expressed at constant temperature by

3.5

( RT

∂ ln k ∂p

)

= −ΔVi ≠

Pressure Intensification

(3.17)

T

where ΔVi ≠ is the difference of the partial molar volumes in the standard state between the transition state and the reactants. Thus, reactions with volume constriction are enhanced under higher pressure, while dissociative processes are restricted by increasing pressure. The volume difference between the partial molar volumes of reactants and products is usually determined experimentally from the pressure dependence of the rate constants. Most association reactions and bond cleavage mechanisms account for −10 and 10 cm3 mol−1 in volume change, respectively [15]. At high pressure, some reaction pathways are inhibited, while others are accelerated. This opens the way to a novel selectivity of chemical reactions. Overcoming the energy barriers in high-pressure transformations can, as usual, be facilitated by increasing the temperature. The joint control of temperature and pressure enhances the subtleties of the high-pressure phenomena. In fact, while high temperature favors dissociation, high pressure induces condensation, and the balance of the two parameters is an additional control resource. In general, the activation volume is the basic piece of information to identify the structure of the transition state and to clarify the reaction mechanism. This knowledge can be of great practical importance. Whenever the reaction can develop along different pathways associated to transition states with distinct activation volumes, the pressure variable offers the opportunity to control the selectivity of the reaction. This is also the case when the reaction proceeds as a multistep process and the pressure has a different effect on the various intermediates. Pressure increase will favor reaction pathways corresponding to more negative activation volumes. 3.5.2 Equilibrium

The equilibrium constant K of a chemical reaction is the ratio of the activities of the product to the reactants of the system at equilibrium, when Gibbs free energy is at its minimum. Chemical potential dependence on pressure at constant temperature can be related to the molar volume in the standard state through: (

∂ ln K RT ∂p

) = −ΔVi 0

(3.18)

T

where ΔVi 0 is the difference of the partial molar volumes in the standard state between the products and the reactants. Thus, for reactions involving decrease in volume along the reaction coordinate, equilibrium constants exponentially increase with increasing pressure at constant temperature.

37

38

3 Chemical Intensification – Fundamentals

3.5.3 Electron Kinetic Energy

Higher pressure leads to closer molecule contact and occasionally to electron shell overlap. Severe confinement and overlap of the electron clouds raise the kinetic energy of the electrons and bring the system into a highly repulsive region of the free energy surface. The response of the system to recover a novel free energy minimum will result in a number of phenomena including phase transformations, ionization, condensation, polymerization, amorphization, dissociation and, as an ultimate response to the most extreme conditions, atomization, and metallization [16]. It has been observed that at high pressure the elements tend to behave like the heavier elements of the same group of the periodic table. Due to a shift of the energy levels of electrons, the chemical reactions occurring at very high pressure can be different to familiar laboratory practice [17]. The π → π* transitions are generally the most sensitive to pressure. These transitions are usually the lowest in energy for unsaturated compounds; hence, these systems are particularly reactive at high pressure. This has been reported for several conjugated systems [18, 19]. 3.5.4 Material Properties

The freezing point is also affected by the pressure. Fluids, where under freezing condition the solid phase swims on the liquid phase, have negative compressibility. Here, high pressure leads to liquefaction and better mobility of the molecules (water as example). Normal liquids have a coefficient of 1.5 ± 2 K per 10 MPa [20–22]. In Figure 3.6, on the left side, the vertical lines depict the solidification line. The scale cannot display the incline of the solidification line, due to the small effect in this scale. Despite being counted as incompressible, liquid and solid molecular materials can be compressed with a volume reduction by up to one order of magnitude, a size contraction corresponding to shrinkage of the intermolecular distances by a factor of 2 [16]. One important point is that severe geometrical constraints are imposed on the molecules of any pressurized sample and, in the simplest approach, the reactions occurring with minimum molecular and atomic displacements will be favored [23]. Compressibility of liquids depends on the polarity of a molecule/solvent and the packing of the molecules. One of the most important effects of pressure on the reaction rate is related to the packing of the solvent molecules around the reactants and around the transition state complex. Pressurization can affect both the medium–medium and the medium–reactant interactions. If the transition state has a higher polarity or is ionic compared to the reactants, the packing of a polar solvent around the transition state will be more efficient (electrostriction) and the medium contribution to the activation volume can be larger or even of opposite sign with respect to the intrinsic contribution. A significant contribution to the activation volume is associated

3.5

Pressure Intensification

with steric hindrance or overcrowding in the transition state. Nonpolar solvents undergo more electrostriction and volume reduction than polar media due to the initially weaker intermolecular interactions. In general, effects of solvent polarity and high pressure can be used to draw conclusions about whether the activated complex interacts more with the solvent than the initial reactants. In this context, solvent-compressibility values and the dielectric constant are of importance [20]. The dielectric constant of supercritical water can be varied continuously from 1 (a typical value for nonpolar solvents) to 30. Additionally, in supercritical water, the ionic product can be increased by three orders of magnitude and the H+ concentration increased by 30 times compared to normal conditions. The dielectric constant ε of the liquid is a weak function of temperature, while being strongly dependent on pressure: [ ] 𝜀p0 B+p (3.19) = 1 − A ln 𝜀p B+1 with A and B as characteristic parameters for a liquid. Both do not depend on pressure, while B decreases with increasing temperature. Both constants are usually determined experimentally for specific combination of liquids. The dielectric constant determines the magnitude of the electrostriction induced by the solvent in the vicinity of charged species, thus playing a significant role in activation volume calculations. Laidler and Eyring [24] derived an equation relating rate constant to the solvent polarity and dipole moments of the reactants and transition state based on the Kirkwood electrostatic model: ( 2 ) μA μ2B μ2≠ 1 (3.20) + 3 + 3 ln k = ln k0 − kB T rA3 rB r≠ where k 0 refers to the rate constant at vacuum conditions, k B is the Boltzmann constant, and μ are the dipole moments of the two spherical molecules A, B, and the transition state. It has been shown that there is a high solvent dependency of the reaction rate constant of cycloaddition of tetracyanoethylene to enol ethers, for example, k(acetonitrile)/k(cyclohexane) = 29 000 for anethole, 10 800 for 1ethoxyisobutene, and 2600 for butyl vinyl ether, as well as k(acetonitrile)/k(carbon tetrachloride) = 17 000 for 2,3-dihydro-4H-pyrene [21, 25]. Higher pressure also increases the viscosity of a liquid, with approximately a factor of 2 every 100 MPa [26]. Increased viscosity at high pressures in this range can accelerate chemical processes such as Diels–Alder reactions, 1,3-dipolar cycloadditions, and Claisen rearrangements [27]. The opposite effect was also observed, where higher viscosity hinders reaction rate. It is generally agreed that transition state theory may be invalid in highly viscous media, where diffusion processes play an important part. Since pressure increases the viscosity coefficient exponentially, such effects apparently must be taken into account. As an illustrative example, in the Z–E isomerization of 4-(dimethylamino)-4-nitroazobenzene, the rate constant initially increases with pressure, then decreases at higher pressures, when the medium is a viscous silicone oil [28].

39

3 Chemical Intensification – Fundamentals

3.5.5 Mixture Properties

Solubility, one of the fundamental properties of a system with two or more components, is a function of both pressure and temperature. The pressure coefficient of the logarithm of solubility, xS , is thermodynamically related to the volume change, ΔV 0 , accompanying the dissolution, as shown in the following equation: ( ) ∂ ln xs RT = −ΔV 0 (3.21) ∂p T with ΔV 0 = V sat − V *. In case of toluene, where V * is the molar volume of a liquid hydrophobic solvent, ΔV 0 corresponds to a volume change for hydrophobic hydration. In the past, the sign of the volume change accompanying hydrophobic hydration has been believed to be negative at atmospheric pressure. However, using the equations above, it can be shown that the solubility maximum at each temperature means that the sign of ΔV 0 for hydrophobic hydration changes at this point, with increasing pressure, from negative to positive [29]. Above the critical point, the supercritical solvent shares the properties of the liquid and of the gas phase. From this, several advantages of the supercritical solvents are derived and among these are the increased solvating power compared to normal liquids, and the higher diffusivity approaching the gas-phase behavior. However, the main resource is that by changing the pressure and temperature, the solvent properties can be fine tuned between the liquid and gas phase extremes and adjusted for the reaction of interest. Additionally, mass transfer operations, such as solvent extraction, can be facilitated within supercritical conditions [30]. The impact of pressure-induced changes (up to 750 MPa) on the dielectric constant 𝜀, viscosity 𝜂, and reaction constant k is exemplified for Click Chemistry synthesis, namely the 1,3-dipolar Huisgen cycloaddition (Figure 3.8) P=750 MPa 1.4 P= 675 MPa

1.31 1.3

𝜀p/𝜀0

40

1.2 P= 525 MPa 1.1 1 20

P= 225 MPa 20 29.22

𝜂p /𝜂

0

10 0

0

100

Figure 3.8 (𝜀, 𝜂, k) Novel Process Windows at 1 bar and 25 ∘ C for a Click Chemistry reaction, here the 1,3-dipolar Huisgen cycloaddition, under assumptions as follows. Constants used in the underlying equations were

200

300

400

k p/k 0

500

600 574.2

derived from heptan-1-ol and the activation volume of [3+2] cycloadditions. (Borukhova et al. [31]; reproduced by permission of IMRET-12).

3.5

Pressure Intensification

[31]. Solvent polarity, expressed through the dielectric constant ε, has a weak dependence on temperature, while being strongly impacted by pressure. Moreover, the dielectric constant determines the magnitude of the electrostriction induced by the solvent in the vicinity of charged species, thus playing a significant role for the activation volume. Viscosity increases with pressure along with the reaction constant, which can cause viscosity-associated acceleration.

3.5.6 Illustration of Pressure Effect on Selected Chemical Reactions

The pressure effects on the physical media are in the following exemplified at the example of cycloaddition reactions. An intriguing example of the difference between low- and high-pressure conditions is the reaction between 2,3dimethylbuta-1,4-diene and {[(E)-2-nitroethenyl] seleno}benzene (Scheme 3.1). Under low-pressure conditions, the adducts endo-A and exo-B were obtained in only 16% overall yield in a ratio of 25 : 75. In contrast, at high pressure (1.2 GPa), the yield increased to 65%, and the endo/exo ratio changed to 46 : 54. Under even higher pressure (1.6 GPa), 70% yield and an endo/exo ratio of 66 : 21 were achieved. In the last case (1.6 GPa), the reaction, thus, mainly gave the endoadduct A, the smaller-activation-volume product, as well as the tautomerization product C (Scheme 3.1) [26].

SePh

SePh

Sealed tube, THF, 120 °C, 11 d SePh + NO2

1.2 GPa, CH2Cl2, 50 °C, 4 d 1.6 GPa, CH2Cl2, 25 °C, 1 d

A

NO2

B

NO2

SePh

C

NO2

16% yield A + B + C (25:75:0)

65% yield A + B + C (46:54:0)

70% yield A + B + C (66:21:13)

Scheme 3.1 The effects of pressure on the cycloaddition between 2,3-dimethylbuta-1,4diene and {[(E)-2-nitroethenyl]seleno}benzene. (McKeem Mortimer [26]; reproduced by permission of Royal Society of Chemistry).

Multicomponent reactions can benefit from high-pressure conditions, and even double multicomponent reactions [32]. For example, 3-[(E)-2nitroethenyl]pyridine (71), the benzyl vinyl ether (72), and methyl acrylate (73) underwent a three-component cycloaddition via tandem [4+2]/[3+2] reaction, to afford the bicyclic nitroso acetal (74) as a mixture of three diastereo isomers (see

41

42

3 Chemical Intensification – Fundamentals

Scheme 3.2). When N-phenylmaleimide (24) was used instead of 73, compound 75 was obtained as a single stereoisomer [33]. MeO

NO2

RO

O

O

O

O

1.0 GPa, CH2Cl2

O +

N

N

r.t., 18 h

+

74%

71

72

N

73 RO

Ph O

71

+

72

N

O

N

74 R = 4-MeO-Bn O O

O

+

24

Ph

90%

O N

75 Scheme 3.2 Three-component cycloaddition via tandem [4+2]/[3+2] reaction from 3-[(E)-2-nitroethenyl]pyridine (71), the benzyl vinyl ether (72), and methyl

acrylate (73) to bicyclic nitroso acetal (74); N-phenylmaleimide (24) as alternative component. (Kuster et al. [33]; reproduced by permission of Wiley-VCH).

References 1. Hessel, V., Vural-Gürsel, I., Wang, Q.,

2.

3. 4.

5.

6.

7.

8.

Noël, T., and Lang, J. (2012) Chem. Eng. Technol., 35 (7), 1184–1204. Kockmann, N. (2008) Transport Phenomena in Micro Process Engineering, Springer, Berlin, p. 14. Knösche, C.M. (2005) Chem. Ing. Tech., 77 (11), 1715–1722. Roberge, D.M., Ducry, L., Bieler, N., Cretton, P., and Zimmermann, B. (2005) Chem. Eng. Technol., 28 (3), 318–323. Kockmann, N., Gottsponer, M., and Roberge, D.M. (2011) Chem. Eng. J., 167 (2-3), 718–726. Schwalbe, T., Autze, V., Hohmann, M., and Stirner, W. (2004) Org. Process Res. Dev., 8 (3), 440–454. Golbig, K., Hohmann, M., Kursawe, A., and Schwalbe, T. (2004) Chem. Ing. Tech., 76 (5), 598–603. Hessel, V., Kralisch, D., Kockmann, N., Noël, T., and Wang, Q. (2013) ChemSusChem, 6 (5), 746–789.

9. Kockmann, N. (2008) Chem. Eng. Tech-

nol., 31 (8), 1188–1195. 10. Kockmann, N. and Roberge, D.M.

11. 12.

13. 14.

15. 16. 17.

(2009) Chem. Eng. Technol., 32 (11), 1682–1694. Marre, S., Roig, Y., and Aymonier, C. (2012) J. Supercrit. Fluids, 66, 251–264. Kockmann, N., Dreher, S., Engler, M., and Woias, P. (2008) Chem. Eng. J., 135S (S1), 121–125. Lindner, P.J. and Besser, R.S. (2012) Chem. Eng. Technol., 35 (7), 1249–1256. Renken, A., Hessel, V., Löb, P., Miszczuk, R., Uerdingen, M., and Kiwi-Minsker, L. (2007) Chem. Eng. Process., 46 (9), 840–845. Asano, T. and Le Noble, W.J. (1978) Chem. Rev., 78 (4), 407–489. Schettino, V. and Bini, R. (2007) Chem. Soc. Rev., 36, 869–880. Drickamer, H.G. and Frank, C.W. (1973) Electronic Transitions and the High Pressure Chemistry and Physics of Solids, Chapman & Hall, London.

References 18. Wiederhorn, S. and Drickamer, H.G.

19. 20.

21.

22. 23. 24.

25. 26.

(1959) J. Phys. Chem. Solids, 9 (3-4), 330–334. Samara, G. and Drickamer, H.G. (1962) J. Chem. Phys., 37, 474–479. Isaacs, N.S. (1981) Liquid Phase High Pressure Chemistry, John Wiley & Sons, Inc., New York. Hammann, S.D. (1957) Physico-Chemical Effects of Pressure, Butterworth, London. Firestone, R.A. and Vitale, M.A. (1981) J. Org. Chem., 46 (10), 2160–2164. Cohen, M.D. and Schmidt, G.M.J. (1964) J. Chem. Soc., 1996–2000. Laidler, K.J. and Eyring, H. (1941) The Theory of Rate Processes, McGraw-Hill, New York. Steiner, G. and Huisgen, R. (1973) J. Am. Chem. Soc., 95 (15), 5056–5058. McKee, C. and Mortimer, M. (2002) in Chemical Kinetics and Mechanism: The Molecular World (eds M. Mortimer and P.G. Taylor), Royal Society of Chemistry, Cambridge, p. 86.

27. Swiss, K.A. and Firestone, R.A.J. (2000)

Phys. Chem. A, 104 (13), 3057–3063. 28. Shito, A., Takahashi, T., Ohga, Y.,

29.

30. 31.

32.

33.

Asano, T., Saito, H., Matsuo, K., and Luedemann, H.D. (2000) Z. Naturforsch., A, 55, 616–622. Sawamura, S., Nagaoka, K., and Machikawa, T. (2001) J. Phys. Chem. B, 105 (12), 2429–2436. Luther, S.K. and Braeuer, A. (2012) J. Supercrit. Fluids, 65 (5), 78–86. Borukhova, S., Varas, A.C., Hessel, V., Wang, Q., Watts, P., and Wiles, C. (2012) Safe reaction conditions and their scaleup in microstructured reactors. Book of Abstracts, IMRET-12, International Conference on Microreaction Technology, Lyon, France. Matsumoto, K., Kim, J.-C., Iida, H., Hamana, H., Kumamoto, K., Kotsuki, H., and Jenner, G. (2005) Helv. Chim. Acta, 88 (7), 1734–1753. Kuster, G.J.T., Steeghs, R.H.J., and Scheeren, H.W. (2001) Eur. J. Org. Chem., 3, 553–560.

43

45

4 Making Use of the “Forbidden” – Ex-Regime/High Safety Processing First, an overview of the coming chapters and subchapters shall be given. The chapters are defined by the six Novel Process Window(NPW) routes given on the Figure 4.1. The subchapters are defined to exemplify the three main motifs of this review, which are “enable, speed-up, and uplift.” These bridge to and are derived from the Process Intensification definition of Górak and Stankiewicz as given in Chapter 2 “Novel Process Windows – Definition.” Two of the four motives refer to chemical and process-design intensification.

4.1 Hazardous Reactants and Intermediates

Microreactors reduce significantly the risks associated with the handling of certain toxic and explosive hazardous compounds. Such reactants can be generated in situ in microreactors and react immediately away in subsequent transformations, hereby eliminating the need for storage. Even when the reactor fails, the amount of hazardous reactants spilled in the environment remains limited. On a production scale, reactor failure can be detected by the use of control units and, subsequently, can be isolated and replaced by another microreactor. As such, the synthesis of hazardous reactants can be done on a production scale within microstructured reactors without compromising public safety. 4.1.1 Tetrazole Formation

Five-substituted 1H-tetrazoles are very common motives used in a wide range of applications, such as explosives, catalysis, and medicinal chemistry. However, their synthesis involves the use of HN3 , which is extremely toxic (toxicity comparable with HCN) and explosive and has a low boiling point, enhancing its exposure risk. Kappe, Roberge et al. [1]. have developed a continuous-flow setup in which HN3 is generated in situ from NaN3 and acetic acid (Scheme 4.1).

Novel Process Windows: Innovative Gates to Intensified and Sustainable Chemical Processes, First Edition. Volker Hessel, Dana Kralisch and Norbert Kockmann. © 2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

46

4 Making Use of the “Forbidden” – Ex-Regime/High Safety Processing

Maximize effectiveness of intramolecular and intermolecular events SPEED-UP

6.2

6.3 6.4

6.5

6.1

6.6

8.1

7.1

8.2 8.3

7.2

Activate

7.3

Collide and slide

Press 7.4

high T

Novel process windows

Biocatalysis Tandem catalysis

LE

E

B NA

Explore new paths

5.5 5.4

5.3 5.2 5.1

Make use of the forbidden

4.2

9.1

Make more by combining make same with less

9.2 9.3

10.2 10.1 9.5

9.4

UP

LIF

T

4.1

Maximize the synergistic effects of partial processes Figure 4.1 Overview of the coming Chapters: 5.1: Hazardous reactants and intermediates; 5.2: Ex-regime and thermal runaway; 6.1: Direct syntheses; 6.2: Multistep one-flow; 7.1: High-T specialty microreactors; 7.2: Cryogenic to ambient; 7.3: Solvent-scope widening; 7.4: Product and material control; 8.1: High-p speciality microreactors; 8.2: Interfacial transport; 8.3: Activation volume; 8.4: Solvent mastering 8.5: Materials shaping; 9.1:

Solvent-less and solvent-free; 9.2: Designer solvents; 10.1: Reaction and separation; 10.2: Process and sensing; 10.3: Thermal integration; 10.4: Full-unit racking; and 10.5: Fullflow process development. The opening of process windows by allowing safe processing under otherwise hardly accessible or even ex-regime conditions was formerly considered also as a novel process window and still is so here.

HN3 is subsequently consumed by addition of nitrile to the corresponding five-substituted 1H-tetrazoles at high temperatures (220–260 ∘ C) and a pressure of 36 bar in a silica-coated stainless steel coil (Sulfinert). On exiting the microreactor, the reaction stream is cooled down in a heat exchanger and the residual HN3 is decomposed by quenching with aqueous NaNO2 . Excellent yields (68–98% isolated yield) are obtained for a broad range of five-substituted 1H-tetrazoles within a 10–15 min time frame.

4.1

Hazardous Reactants and Intermediates

NaN3 + AcOH

R-CN + HN3

NMP, H2O

R

N N N N H

- 68–98 % yield - T = 220–260 °C - p = 34 bar - t = 10–15 min - Tubular microreactor (Flowsyn) Scheme 4.1 Handling of HN3 in the continuous-flow synthesis of five-substituted 1Htetrazoles [1].

4.1.2 Strecker Synthesis

The Strecker synthesis involves the use of highly toxic HCN and is therefore hazardous on a batch scale. Stevens et al. [2] investigated the use of microreactor technology as a safe and reliable way to generate in situ small quantities of HCN from KCN and HOAc (Scheme 4.2). HCN is immediately consumed in the reaction to yield 3,4-diamino-1H-isochromen-1-ones in good yields (49–66% yield). In order to circumvent irreproducible residence times, the reaction temperature was kept low (50 ∘ C) to prevent flashing of the solvent. Low concentrations of the reagents (0.1–0.2 M) were required to prevent clogging of the system. KCN + AcOH O H OH O

HCN Aryl-NH2 MeOH

NHAr NH2 O O - 49–75 % yield - T = 50 °C - p = 1 bar - t = 39 min - Tubular microreactor (CYTOS)

Scheme 4.2 Handling of HCN for the 3,4-diamino-1H-isochromen-1-ones synthesis in continuous flow [2].

4.1.3 Phosgene Chemistry

Another very toxic yet useful compound for organic synthesis is phosgene. A safe protocol for the generation/handling of phosgene was realized in a

47

48

4 Making Use of the “Forbidden” – Ex-Regime/High Safety Processing

microflow reactor by mixing relatively safe triphosgene with Hünig’s base [3]. The phosgene was immediately consumed by reaction with a carboxylic acid to yield an acid chloride. Epimerization of the acid chloride could be efficiently suppressed by reduction of the residence time (1.5 s) in the first reactor. The reaction stream containing the acid chloride was next merged with a stream of amine to yield amide in good to excellent yields (89–95% overall yield). Notably, better-isolated yields and less epimerization were observed in flow when compared to the corresponding batch reactions (see Scheme 4.3 and Table 4.1 for comparison). 1. Triphosgene i Pr2NEt H N

O O

O

O Cl

OH OBn

Cl

H N

O

2. R2NH, CH2Cl2

O

O NR2 OBn

- 70–93 % yield - T = 20 °C - p = 1 bar - t = 1.5 s first reactor 3.7 s second reactor - Tubular microreactor Scheme 4.3 Handling of phosgene for the amide synthesis in continuous flow [3].

4.1.4 Diazomethane Synthesis

Diazomethane is a common methylating agent for organic synthetic chemistry. However, its toxic, carcinogenic, and explosive nature renders it difficult to use in Table 4.1 Comparison between batch and flow for the amide synthesis [3].

H N

O O

O OH OBn

Entry

Reactor type

1 2 3

Flow Flow Batch

1. Triphosgene, iPr2NEt 2. nOctyl-NH2, CH2Cl2

H N

O O

O NHnOctyl OBn

Reaction time (s)

Yield (%)

Ee (%)

3.7 7.4 20

93 91 76

95 92 80

4.1

Hazardous Reactants and Intermediates

both laboratory and industry. Intermediate diazomethane generation eliminates the need to store, transport, or handle this dangerous material. A continuous generation and consumption of diazomethane, while keeping the maximum inventory low, can minimize the risks associated with this compound on both laboratory and industrial scale [4]. Diazomethane can be synthesized safely in a microreactor by treating N-methyl-N-nitroso-p-toluene sulfonamide (Diazald) with aqueous KOH [5]. Subsequent introduction of benzoic acid ensures a complete consumption of the hazardous diazomethane, hereby yielding methyl benzoate in 75% yield. Kim et al. [6] reported on the continuous in situ generation, separation, and reaction of diazomethane in a dual-channel polydimethylsiloxane (PDMS) membrane reactor (Figure 4.2). Diazomethane was generated in one channel by reaction of Diazald with aqueous KOH. Next, diazomethane was selectively transferred from the aqueous channel to the organic channel where the subsequent reaction takes place. Due to this integrated separation step in the microfluidic device, basic decomposition of diazomethane could be circumvented and high yields (81–99% yield) were obtained. The use and synthesis of other diazo compounds in microstructured devices have also been reported [7, 8]. This has been utilized for a flow multistep synthesis of α-halo ketones starting from N-protected amino acids [9]. α-Halo ketones are chiral building blocks for the synthesis of HIV protease inhibitors such as atazanavir and darunavir. The first step of the multistep synthesis is the formation of a mixed anhydride in a tubular reactor (Scheme 4.4).

O

O R OH

O R

OH

O R

O

O R

O

N2

R

O Cl

R

H

Reactions Diazald

KOH

Reactant + CH2N2

Product

Separation

Diazald + KOH

CH2N2

Generation Waste

Figure 4.2 Continuous in situ generation, separation, and reaction of diazomethane in a dual-channel PDMS membrane reactor. (Maurya et al. [6]; reproduced by permission of Wiley-VCH).

49

50

4 Making Use of the “Forbidden” – Ex-Regime/High Safety Processing

O PGHN

OH + ClCH3O2R

THF,Bu3N rt,6.7min

O PGHN

OCO2R rt, 26.7 min

R

R O

MeOH/H2O O KOH S N H2C N N N O CH3

O PGHN

OCO2R R

Diethyl ether, HCl 0° C, 13.3 min

Tube-in-tube reactor (hydrophobic membrane)

O PGHN

CH2Cl

55–82% yield

R

PG = protecting group Scheme 4.4 Continuous in situ generation and reaction of anhydrous diazomethane in a tube-in-tube reactor, thereby producing α-halo ketones from N-protected amino acids [9].

Then, the anhydride is reacted with anhydrous diazomethane in a tube-in-tube reactor, having a gas-permeable, hydrophobic inner tube, which is encased by a thick-walled, impermeable outer tube. Diazomethane is generated in aqueous phase in the inner tube, which diffuses as anhydrous diazomethane through the permeable membrane into the outer channel. As a result of such in situ processing, an α-diazo ketone is formed, which finally is converted to the halo ketone with anhydrous ethereal hydrogen halide. In this way, α-halo ketone building blocks were made in one run without racemization in excellent yields (87% yield). The laboratory process gave 1.84 g product in 4.5 h. 4.1.5 Ozonolysis

Ozone (O3 ) can be considered as a clean and atom-efficient oxidation reagent. Nevertheless, the presence of ozonides, which are unstable and explosive intermediates in the ozonolysis reaction, complicates the use of ozone on a large scale [10]. In addition, the final work-up step (reductive or oxidative quenching) is an exothermic step, which requires an efficient temperature control to minimize sideproduct formation. For these reasons, microreactors can be employed as safe and reliable devices for ozonolysis reactions. Kappe et al. [11] utilized a commercially available flow ozonolysis reactor (O-Cube, Thalesnano) for the oxidation of a wide range of compounds, such as aromatic and aliphatic olefins, alkynes, nitro compounds, and thioethers. A semipermeable membrane reactor for ozonolysis reactions was developed by Ley and coworkers [12]. This allows for the distribution and dosage of ozone through a gas-permeable membrane and also maximizes the contact area between gaseous O3 and liquid phase. The reactor system consists of a gas-permeable membrane tubing (i.e., Teflon AF-2400), in which the liquid phase flows. The tubing was placed in a glass container in which a flow

4.1

Hazardous Reactants and Intermediates

of O3 was introduced. The ozonolysis of several aromatic and aliphatic olefins to the corresponding ketone products (57–95% yield) could be performed in flow (Scheme 4.5). Hübner and Jähnisch [13] investigated four microstructured reactors, including falling-film microreactors, cyclone micromixer and five- and onechannel micromixers, for the ozonolysis of intermediates of vitamin D analogs. Surprisingly, the best results were obtained with a five-channel micromixer; excellent conversion of the substrate was observed for a broad range of flow rates for the substrate flow. It was surmised that, in this design, an annular flow regime was present in the microchannels. This flow regime increases the interfacial area and enhances the reaction rate of the ozonolysis. R4

R3 R3

R4

R1

R2

O

O3

+

Methanol

O 1

R R4 O

R3 R2

O R1

R2

R4 O

R3

O

O R2

O R1

- 57–95% yield - T = RT - p = 1 bar -t=1h - Gas permeable tubular microreactor

Scheme 4.5 Continuous-flow ozonolysis in gas-permeable membrane to enhance gas–liquid contact [12].

4.1.6 Organic Peroxide Formation

Organic peroxides are widely used in the chemical industry, for example, as initiators for radical polymerizations. However, these compounds are thermally instable, which can lead to explosions. Illg et al. [14] developed a continuous-flow synthesis of tert-butyl peroxypivalate in an orifice microreactor. The synthesis involved an initial deprotonation of tert-butyl hydroperoxide with KOH and subsequent reaction with pivaloyl chloride to yield the desired product (Scheme 4.6). In order to increase the interfacial area of the biphasic reactor, a microreactor equipped with orifices was employed. Shortening the distance between the orifices lowers the required residence time for full conversion but results in a less efficient heat removal. A space-time yield of 420 000 g l−1 h−1 was obtained in a microreactor setup with nine orifices and a reaction temperature of 40 ∘ C. This was significantly higher than the space-time yield of 190 g l−1 h−1 , which was obtained with a triple-cascaded batch reactor process.

51

52

4 Making Use of the “Forbidden” – Ex-Regime/High Safety Processing

1. aq. KOH OOH

2. Pivaloyl chloride

O O O - 420 000 g L–1 h–1 - T = 40 °C - p = 1 bar - Single-channel orifice microreactor

Scheme 4.6 Continuous-flow synthesis of tert-butyl peroxypivalate in a single-channel orifice microreactor [14].

Methyl ethyl ketone peroxidation with hydrogen peroxide can be carried out in a microreactor in the absence of organic solvents [15]. It was found that careful temperature control was crucial to obtain good yields, since higher reaction temperatures lead to decomposition of the peroxide product. Taking advantage of the excellent heat dissipation in microreactors, high throughput production of methyl ethyl ketone peroxide is feasible. The synthesis of peracids can also be done efficiently in a microreactor system [16]. The reaction could be accelerated with sulfuric acid as an acid catalyst. The microreactor consisted out of an integrated reaction-heat exchange, which avoided the formation of hot spots, and has a capacity to produce up to 100 t/a performic acid and up to 170 t/a peracetic acid.

4.2 Ex-Regime and Thermal Runaway Processing

Highly reactive, exothermic reactions are difficult to carry out in conventional batch equipment. However, on a microscale, heat can be efficiently dissipated due to the high surface-to-volume ratio reducing hereby the risks associated with such reactions. Reactions that present such challenges are typically hydrogenations, oxidations, and fluorinations. 4.2.1 Oxidation

The gas–liquid oxidation of cyclohexane using pure oxygen was performed at high temperatures (>200 ∘ C) and pressures (up to 25 bar) in a microreactor [17]. Such explosive conditions could be safely handled in microflow and good selectivity (88%) and productivity (2.84 mol l−1 h−1 ) were obtained. 4.2.2 Hydrogen Peroxide Synthesis

The direct synthesis of hydrogen peroxide (H2 O2 ) starting from H2 and O2 over a palladium catalyst is an environmentally benign and atom-efficient method

4.2

Ex-Regime and Thermal Runaway Processing

[18]. However, it involves the handling of an explosive gas mixture, making it a hazardous operation on a macroscale. Jensen et al. [19] developed a multichannel microchemical reactor filled with Pd/C as a catalyst, which could selectively yield H2 O2 at pressures of 2–3 MPa. It was found that increases in yield could be obtained by placing microreactors in series. 4.2.3 Direct Fluorination

The direct fluorination of aromatic substrates is highly desired, since it avoids the use of expensive and hazardous fluorinating reagents. However, direct fluorination is rarely employed on a conventional batch scale due to large heat release, formation of hydrogen fluoride, and lack of selectivity. Gas–liquid microreactors allow for the direct fluorination of toluene using F2 [20, 21]. Conversions of up to 76% were obtained in a falling-film microreactor, and a selectivity of 28% for the monofluorinated products was achieved (Scheme 4.7) [20]. It was found that conversions could be enhanced with higher fluorine-to-toluene ratios and higher reaction temperatures. Also, the solvent choice was crucial; acetonitrile gave the best selectivities. Me

Me 10% F2 in N2 F

CH3CN

- 76% conversion - 28% monofluorination - T = –15 °C - Falling film microreactor Scheme 4.7 Direct fluorination of toluene [20].

A safe concept for the direct fluorination of carbon monoxide in microreactors was developed by Navarrini et al. [22]. The reaction is highly exothermic and therefore difficult to control. They successfully circumvented a thermal runaway by using a stainless steel parallel channel microreactor (surface/volume ratio ∼1 × 104 m−1 , residence time 𝜏 ∼ 0.1 s). 4.2.4 Ionic Liquid Synthesis

The synthesis of ionic liquids is known for its instantaneous and extreme heat release, leading to side-product formation. Therefore, careful temperature control is necessary to provide high-quality ionic liquids [23]. Toward this end, heat pipes were used for dynamic cooling of microstructured reactors by Löwe et al. [24, 25] in the synthesis of 1-ethyl-3-methylimidazolium trifluoromethanesulfonate (Scheme 4.8). With this reactor setup, the cooling rate could be

53

54

4 Making Use of the “Forbidden” – Ex-Regime/High Safety Processing

automatically adjusted depending on the released heat, hereby avoiding the formation of hot spots and thermal runaways. Production rates of 1.24 kg h−1 of the ionic liquid were feasible with this device; this corresponds with a space-time yield of 4270 t h−1 m−3 . Et N + N

MeSO3CF3

Et N

CF3SO3 N Me - Quantitative yields - up to 1.24 kg h–1 - t = 1–23 s - Tubular microreactor + heat pipe cooling

Scheme 4.8 Continuous-flow synthesis of 1-ethyl-3-methylimidazolium trifluoromethanesulfonate with heat pipe dynamic cooling [25].

4.2.5 Moffatt–Swern Oxidation

The Moffatt–Swern oxidation of several alcohols was also performed in microreactors [26]. The use of microstructured reactors avoids the accumulation of labile reactive intermediates. Hence, the exothermic Pummerer rearrangement of these intermediates could be circumvented, minimizing the risk for runaway reactions (Scheme 4.9).

Me O

Me

O S

O

O

Me

F3C

O

Me H

Step 2. testosterone

+ CF3

Step 3. i Pr2NEt

H

H

O - 95 % yield - up to 117 g L–1 h–1 - T = 0 °C - p = 1 bar - Tubular microreactor Scheme 4.9 Continuous-flow Moffatt–Swern oxidation of testosterone [26].

Consequently, higher temperatures were feasible in microflow when compared to batch conditions, that is, 0–20 ∘ C in flow compared to −70 ∘ C in batch. With this system, the Moffatt–Swern oxidation of testosterone could be performed for 1.5 h, achieving a productivity of up to 117 g l−1 h−1 .

4.2

Ex-Regime and Thermal Runaway Processing

4.2.6 Reaction Between Cyclohexanecarboxylic Acid and Oleum

Wang, Luo et al. [27] studied the kinetics of the fast exothermic multiphase reaction between cyclohexanecarboxylic acid and oleum (Scheme 4.10). Due to the excellent mixing efficiency of the micromixer, the reaction could be completed in 97% was observed. COOH

COOSO3H +

H2SO4

H2O - Quantitative conversion - > 97 % selectivity - T = 40–90 °C - p = 1 bar -t 1

Stepwise synthetic protocols Figure 5.6 Labor-intensive stepwise synthetic protocols versus one-pot multistep protocols for the decoration of polymers. (Lundberg et al. [15]; reproduced by permission of WileyVCH).

Such kind of synthesis is used nowadays in organic synthesis [16–18]. This requires synthetic strategies which are orthogonal and efficient. This matches very nicely with the chemistry control given in microreactors and their spatiotemporal separation opportunities besides providing a “one-flow” environment. Microreactors are also suited tools to face the main challenges of multistep one-flow syntheses, such as to increase the range of reactions without need for metal catalysis, to develop libraries of compatible reactions, and to transfer to an industrial scale. Such multistep syntheses are divided into non-tandem reactions (NTRs) and tandem reactions (TRs) (Figure 5.7) [15]. The first class comprises multicatalytic and domino reaction strategies. In multicatalytic reactions, all reagents and catalysts are added at different stages of the reaction; all within “one pot.” Domino reactions have mutually dependent serial reaction steps, with all materials being A. One-pot (multicatalytic) reaction

cat. A

cat. B

B. Domino (cascade) reaction

1 cat.

Figure 5.7 Schematic representation of (a) one-pot multicatalytic reactions and (b) domino (cascade) reactions for the decoration of polymers. (Lundberg et al. [15]; reproduced by permission of Wiley-VCH).

71

72

5 Exploring New Paths – New Chemical Transformations

present from the start. NTRs with at least three domino reactions are also named cascade reactions. In TRs, the catalysts are present from the start and two or more independent transformations are involved. Three classes of TRs are (i) orthogonal tandem reactions, (ii) auto-tandem reactions, and (iii) assisted tandem reactions. Orthogonal tandem reactions have several catalysts and reagents, whereas auto-tandem reactions comprise just a single catalyst which catalyzes several reactions. Assisted tandem reactions are somehow intermediate in the sense that one catalyst performs several reactions, but only after being externally “tuned” – by adding additional reagents or energies (heat or UV). TRs and NTRs with their requirements of high efficiency and orthogonality demand for the development of new chemical reactions and strategies. In this context, click chemistry is a preferred concept which besides orthogonality and almost quantitative yields provides high regio-specificity and produces only easily removable by-products [19]. This is mostly achieved via a high thermodynamic driving force, which typically is above 20 kcal mol−1 . Four types of click chemistries are known: 1) 2) 3) 4)

cycloadditions of unsaturated species, nucleophilic substitution chemistry with strained heterocyclic electrophiles, carbonyl chemistry of non-aldol type, and additions to carbon–carbon multiple bonds.

Several of these click chemistries were combined to a multicatalytic reaction in one pot – the Wittig olefination, the Knoevenagel condensation, the Diels-Alder cycloaddition, and the CuI-catalyzed alkyne reaction (Scheme 5.10) [20]. An excellent yield of >80% was achieved for a synthesis of a complex molecule. N

O

O

O O PPh3

X R

+

+

X R O

O

N N R

O i) L-proline, EtOH, 65°C, 3–12 h

X R O X R

ii) R–N3, CuSO4/Cu(s) RT, 15–48 h

O

O

R N N N

80–94% yield Scheme 5.10 Organo/Cu(I)-catalyzed four-component one-pot batch reaction through Wittig/Knoevenagel/Diels–Alder/Huisgen cycloaddition [20].

Such multistep protocols were also used for various polymerization techniques for making dendritic, hydrophilic, hydrophobic, amphiphilic, block-co-polymers,

5.4

Multistep Syntheses in One Microreactor/Chip

and Si particles-polymer hybrids. An orthogonal tandem approach was used for the synthesis of miktoarm star terpolymers (Scheme 5.11) [21]. Ph O N

O

O

O +

O

O O

n

O

O O

Ph Ph O N

OH

+

+ R N3

CuBr/PMDETA Sn(Oct)2,125 °C, 48 h

with R–N3 = N3

R N N N O

O

H n

O n

Scheme 5.11 Click-assisted one-pot multistep protocol for the synthesis of highly functionalized miktoarm star polymers in batch [21].

5.4 Multistep Syntheses in One Microreactor/Chip

Here, the complete multistep processing is done within one microreactor/chip with multiple consecutive additions of reactants at certain locations into the main flow on the microreactor/chip (Figure 5.8). 5.4.1 Multistep Synthesis of [18F]-Radiolabeled Molecular Imaging Probe

The synthesis of the radiolabeled molecular imaging probe 2-deoxy-2[18F]fluoro-D-glucose was achieved in an integrated microfluidic device (Figure 5.9) [22]. Five processes, that is, [18F]fluoride concentration, water evaporation, radiofluorination, solvent exchange, and hydrolytic deprotection, were orchestrated on a single polydimethylsiloxane (PDMS) microchip to afford the desired compound within 14 min (38% radiochemical purity, 97.6% radiochemical purity), which provided enough material for several mouse experiments.

Figure 5.8 Principle of multistep syntheses in one microreactor/chip.

73

74

5 Exploring New Paths – New Chemical Transformations

N

Dilute 1. Anion exchange Concentrated beads F– KF 2. K CO 2

O

OO

O

OO N

3

Anionexchange column

N O

AcO AcO

OAc OTf O

1

a: X = 18F b: X = 19F

ii. Solvent exchange

i. Concentration of fluoride Regular valve Regular valve for pump Sieve valve for column

A

Kryptofix (K222)

OO K+ OO

O OAc

N

OAc

F–

MeCN

O

AcO AcO

OAc

X

iii. Fluorination

2a,b HCl iv. Solvent (3.0 N) exchange v. Hydrolysis

Reaction loop

OH HO HO

O

OH

X

Fluoride concentration loop

FDG (3a,b)

B Circulating pump

Reaction loop Metering pump

2 mm

Fluoride concentration loop

Figure 5.9 Synthesis of the radiolabeled 2-deoxy-2-[18F]fluoro-D-glucose in an integrated microfluidic device, which integrated five sequential processes. (Lee et al. [22]; reproduced by permission of The American Association for the Advancement of Science).

5.4

Multistep Syntheses in One Microreactor/Chip

5.4.2 Combining Asymmetric Organocatalysis and Analysis on a Single Microchip

The enantioselective Brønsted-acid-catalyzed vinylogous Mannich reaction, the enantioselective separation process, and mass spectrometric detection were integrated on a single microfluidic chip (Figure 5.10) [23]. This chip allowed evaluation of several organocatalysts in a time-efficient manner. The results, obtained in micro-flow, were comparable with the analogous batch experiments. 5.4.3 Two-Step Strecker Reaction

An integrated chip microreactor enabled a two-step reaction, consisting of a homogeneous reaction followed by a heterogeneously catalyzed reaction [24]. The Strecker reaction is a three-component one-pot reaction between a carbonyl-containing compound, an amine, and a cyanide source to yield initially an aminonitrile which then can be hydrolyzed to give an amino acid. This one-pot procedure has disadvantages such as use of harsh process conditions, use of expensive Lewis-acid catalysts, side reactions, and variable yields. Thus, the potential of sequential reactant addition and temporal-spatial decoupling of the single reaction step was checked. A borosilicate glass chip with a footprint of 3.0 × 3.0 × 0.6 cm contained two T intersections for reactant entry with Enantioselective, BrØnsted acid catalyzed, vinylogous Mannich reaction OMe N

O O P O OH

H

1a +

HN

O

2

OMe

TBS

OEt

3a

OEt

tBuOH/iPrOH/2-Me-2-BuOH

Aqueous dilution

Injection

O

4a

NanoES

Separation

Makeup-flow

MS orifice

Figure 5.10 Schematic representation of a microchip which enables a organocatalytic reaction, enantiomer separation, and MS analysis on a single chip. (Belder et al. [23b]; reproduced by permission of Wiley-VCH).

75

76

5 Exploring New Paths – New Chemical Transformations

150 × 50 μm microchannels. After mixing the aldehyde and amine, the cyanide source, trimethylsilyl cyanide (TMSCN), was introduced for nucleophilic addition to the imine. An innovation key is the use of a solid-supported Lewis-acid catalyst, to lower the respective costs. In this two-step flow manner, five αaminonitriles were prepared with very high yield (>99.5%) and analytical purity (Scheme 5.12). O

R H

+ R

NH TMSCN, PS-RuCl3

CN

NH2 - >99.5% yield - T = RT - p = 1 bar - t = 1.1–2.2 s - Glass microreactor

Scheme 5.12 Synthesis of α-aminonitriles via a multicomponent Strecker reaction in flow [24].

In a follow-up work by the same group, a non-hydrolysable solid-supported Ga(OTf )3 was used as catalyst in a micro-flow fixed-bed reactor [25]. This enabled a flow synthesis of α-aminonitriles utilizing stoichiometric quantities of the cyanide source. A very large reduction in reaction time from 8 h to 1 min was achieved for ketones. Similarly, the synthesis using aldehydes was boosted down to 30 s. In both cases, yields higher than 99% have been achieved for all substrates tested (0.24 mmol g−1 h−1 ). No catalyst leaching is observed, as confirmed by inductively coupled plasma mass spectrometry (ICP-MS) analysis, whereas conventional batch synthesis relies on homogeneous catalysis at 400 ppm level.

5.5 Multistep Syntheses in Coupled Microreactors/Chips

Here the complete multistep processing is done within one flow, but passes through multiple coupled microreactors/chips (Figure 5.11). In each of these microreactors/chips, a new reactant can be added to initiate the next reaction. 5.5.1 Chlorohydrination of Allyl Chloride

A multistage strategy was investigated which connects several microchemical units in series for the chlorohydrination of allyl chloride with chlorine in water [26]. A dichloropropanol concentration higher than 6 wt% with a selectivity

5.5

Multistep Syntheses in Coupled Microreactors/Chips

Figure 5.11 Principle of multistep syntheses in coupled microreactors/chips.

higher than 96% is obtained using this strategy. Low temperature and high pressure highly improve the microreaction performance. In contrast to the conventional alternative, the microreaction process had the advantages of higher yield, higher dichloropropanol concentration, less water waste, and lower energy consumption. The new process made the reaction process utilizine chlorine more controllable and safe. 5.5.2 Lithiation/Borylation/Suzuki–Miyaura Cross-Coupling

A multistep synthesis with lithiation/borylation/Suzuki–Miyaura cross-coupling was conducted for the synthesis of aryl halides to biaryls in a microreactor [27]. Lithiation of aryl bromides and heteroaromatic precursors was favorably done under ambient conditions. Ultrasonic irradiation was used to avoid clogging by solids [28]. 5.5.3 Suzuki–Miyaura Cross-Coupling-Phenols-Aryl Triflates-Biaryls

The Suzuki–Miyaura cross-coupling from phenols to give various biaryls was performed in a microreactor at very high yield (83–99% yield over two steps) (Scheme 5.13) [29]. First, the phenols were reacted to aryl triflates. Next, impurities were removed in-line by a microfluidic extraction. Then, the triflates were converted to biaryls via a Suzuki–Miyaura cross-coupling reaction in a packed-bed reactor to enhance mixing of the biphasic mixture [30]. 5.5.4 Ring-Closing Metathesis and Heck Reaction

A two-step one-flow synthesis consisted of a Ru-catalyzed ring-closing metathesis and a Heck reaction; several individual reactions being tested [31]. The fast injection of small volumes of reagents and catalysts into the microreactor allowed

77

78

5 Exploring New Paths – New Chemical Transformations

OH R

Ar

1. Tf2O, Et3N 2. Microfluidic extraction 3. XPhos precatalyst, Ar-B(OH)2, TBAB, K3PO4

R

- 88–99% yield - T= –20 °C (triflate formation) = 90 °C (Suzuki–Miyaura coupling) - p = 1 bar - t = 152–203 s (triflate formation) = 240–320 s(Suzuki–Miyaura coupling) - Tubular microreactor + microextractor + micro packed-bed reactor Scheme 5.13 Continuous-flow synthesis of Pd(0)-catalyzed Suzuki–Miyaura cross-coupling starting from substituted phenolic substrates [29].

fast optimization with respect to catalyst choice, catalyst concentration, and type of heating (oil bath/microwave). However, this investigation also showed some limits in coupling of reactions in flow. Whereas the ring-closing metathesis was successful in all cases, the performance of the Heck reactions in this two-step procedure was not at the same level as for the single step coupling. It was not as widely applicable as used for stand-alone synthesis, because some substrates failed and in some cases an aromatization of the pyrolidene substituent, created by metathesis, introduced a second intermediate species. This led to a more undefined starting protocol for the Heck reaction with two substrates with different Heck reactivity. 5.5.5 Imidazo[1,2-a]pyridine-2-carboxylic Acids in Two Steps

A continuous-flow synthesis of imidazo[1,2-a]pyridine-2-carboxylic acids directly from 2-aminopyridines and bromopyruvic acid was achieved [32]. This is considered as a significant improvement when compared to batch processing. In this way, the multistep synthesis of imidazo[1,2-a]pyridine-2-carboxamides without isolation of intermediates was enabled using two microreactors. For the synthesis of a Mur ligase inhibitor, an antibacterial agent, a yield of 46% was found as opposed to 16% yield in batch (Scheme 5.14). 5.5.6 Suzuki–Miyaura Cross-Coupling/Hydrogenation

2-Amino-4′ -chlorobiphenyl was obtained in a two-step flow fashion with good yield and high selectivity [33]. First using homogeneous catalysis, the palladiumcatalyzed Suzuki–Miyaura cross-coupling was conducted at high temperature in a microtubular reactor. This is followed by in-line scavenging of the palladium

5.5

OH NH2

O

1. HO2C

Multistep Syntheses in Coupled Microreactors/Chips

OH

COOMe

Br

N

N

N

HN

NH

O

2. HOBT, EDC, DIPEA COOMe H2N

Mur ligase inhibitor

NH - 46% yield - T = 100 °C (reaction 1) = 75 °C (reaction 2) - p = 1 bar - t = 20 min (reaction 1) = 10 min (reaction 2) - Glass microchip (AFRICA)

Scheme 5.14 thesis [32].

Continuous-flow synthesis of a Mur ligase inhibitor via a two-step flow syn-

metal and a heterogeneous hydrogenation which was performed over platinumon-charcoal. 5.5.7 Sodium Nitrotetrazolate – Diazonium Ion Formation/Sandmeyer Reaction

A safe two-step synthesis of explosion-sensitive sodium nitrotetrazolate by Sandmeyer reaction was achieved in a micromixer and tube reactor [34]. This allowed production of 4.4 g h−1 of sodium nitrotetrazolate (Scheme 5.15) NH2 N NH N N

1. H+, NaNO2 2. NO2–

NO2 N N N N

Na

- 4.4 g h−1 - T = RT - p = 1 bar - t = 10 s - Tubular reactor + micromixer Scheme 5.15

Flow synthesis of sodium nitrotetrazolate [34].

5.5.8 Murahashi Coupling/Br–Li Exchange

The Murahashi coupling was coupled to the Br–Li exchange of Ar1Br with BuLi in a flow reactor (Scheme 5.16) [35]. To enable such reaction integration, the second step had to be accelerated through use of palladium catalysts bearing a carbene ligand.

79

80

5 Exploring New Paths – New Chemical Transformations

Ar

Br 1. nBuLi R

2. Ar-X, PEPPSI-SIPr

R

- 54–93% yield - T = 0 °C (lithium–halogen exchange) = 50 °C (Murahashi coupling) - p = 1 bar - t = 2.6 s (lithium–halogen exchange) = 94 s (Murahashi coupling) - Tubular microreactor Scheme 5.16 Continuous-flow two-step lithiation/Murahashi coupling [35].

5.5.9 5′ -Deoxyribonucleoside Glycosylation

A green and efficient Brønsted-acid-catalyzed glycosylation was performed as one-flow, multistep synthesis to yield 5′ -deoxyribonucleoside pharmaceuticals such as doxifluridine, galocitabine, and capecitabine [36]. Brønsted acids are considered as greener and more efficient catalyst as compared to the otherwise used Lewis acids which need (over-)stoichiometric amounts. The one-flow, multistep synthesis avoids purification of the intermediate products and produces unprotected 5′ -deoxyribonucleosides in a streamlined manner. For capecitabine, a one-flow three-step synthesis was developed (Figure 5.12). To have such flow synthesis running, special concerns have to be taken to ensure chemical stability of the various reactants and intermediates. Therefore, it has also to be checked if and where harsh conditions can be usefully applied. For example, an initially chosen silylated 5-fluorocytosine derivative was withdrawn, since its glycosylation resulted in notable decomposition. Only 50% yield was obtained. Therefore, an alternative injection scheme had to be developed using the reagent 6 m for the glycosylation. This reagent was glycosylated in 20 min O Me

OAc

O

Cl

Me Syringe 1 tBu

OAc OAc

+

N

tBu

C5H11

Syringe 3

Pyridine CH3CN

NaOH MeOH/H2O (1 M)

Syringe 4

HN F

100 PSI 120 μL PFA Tubing reactor

NH2 F 6m

130 °C tr = 20 min

N N

OTMS

CH3CN (0.48 M)

O

Syringe 5

Ice bath

– 5 H OTf (10 mol %)

4 CH3CN (0.4 M)

O

CH3CN

Syringe 2

PFA Tubing reactor

PFA Tubing reactor

rt tr = 30 min

rt tr = 2 min

Me

O

O N

N

O

OH OH Capecitabine 3 72% yield

Figure 5.12 One-flow three-step synthesis of capecitabine. (Shen and Jamison [36]; reproduced by permission of the American Chemical Society).

5.5

Multistep Syntheses in Coupled Microreactors/Chips

at 130 ∘ C, followed by carbamate formation which needs 30 min at ambient temperature. It was tried to speed up the latter reaction by heating, but without success. As a third example of chemical stability, it was found that fast mixing and on-the-spot consumption of the intermediate formed from chloroformate and pyridine at 0 ∘ C is mandatory. Otherwise, decomposition occurs. Capecitabine is generated in flow in 72% yield in less than 1 h. The flow process is claimed to be greener and more efficient as compared to the state-of-the-art syntheses. 5.5.10 Two-Carbon Homologation of Esters to 𝛂,𝛃-Unsaturated Esters

The two-carbon homologation of esters to α,β-unsaturated esters faces major synthetic difficulties which are given for its two existing synthetic routes. The first route comprises two synthetic steps. The first is to produce an aldehyde by the partial reduction of an ester using diisobutylaluminum hydride (DIBALH). This aldehyde is then reacted by the Horner–Wadsworth–Emmons olefination. In a negative way, this route has become well known for its capriciousness in the DIBALH reduction step which has led to fewer use of it. A multistep alternative provided by the reduction of the ester to a primary alcohol followed by an oxidation to the aldehyde and by its olefination has also its problems. Additional waste is generated and, moreover, it requires the isolation and purification of aldehydes, which are difficult to handle. A flow synthesis for the two-carbon homologation of esters to α,βunsaturated esters was developed [37]. This homologation telescopes three reaction steps – ester reduction, phosphonate deprotonation, and Horner– Wadsworth–Emmons olefination – into a single, uninterrupted system. Thereby, isolation or purification of the aldehyde intermediates is not needed anymore. The homologated products are obtained in high yield and selectivity. 5.5.11 Low-Pressure Carbonylations with Acids as CO Precursors

Several carbonylation reactions, such as the Heck reaction, Sonogashira reaction, and radical carbonylations, were performed in a two-channel coaxial-flow membrane reactor [38]. In ex situ fashion, carbon monoxide is generated by the Morgan reaction, which is the dehydration of formic acid by sulfuric acid (Figure 5.13). This is introduced into a micro-flow system using a tube-in-tube reactor with a gas-permeable inner tube which allows gas transfer into the liquid flow in the outer tube. In particular, the Heck aminocarbonylation of aryl iodide using triethylamine as base was done concurrently with the formic acid/sulfuric acid decomposition (Figure 5.14).

81

82

5 Exploring New Paths – New Chemical Transformations

P Decarbonylation chamber

Pressure monitor Carbonylation chamber

HCOOH Solution A

H2SO4

Figure 5.13 CO generation from formic acid in a two-chamber reactor. (Brancour et al. [38]; reproduced by permission of the American Chemical Society).

H2SO4

HCOOH

10.0 μl h−1

Hastelloy mixer PTFE tube Stainless steel tube AF 2400 tube

3.6 μl h−1

80 °C Tube-in-tube reactor

1, 2 (2 equiv) Pd(dba)2, PPh3 Et3N

0.5 ml h−1 3 H2SO4/H2O

Figure 5.14 Micro-flow system for carbonylation reaction with formic acid/sulfuric acid using “tube-in-tube” reactor. (Brancour et al. [38]; reproduced by permission of the American Chemical Society).

5.5.12 Coupled Microreactor-Purification-Analytics for 𝛅-Opioid Receptor Agonist

A series of microreactors was connected to flow-purification devices and an own-developed in-line IR flow cell for synthesis of N,N-diethyl-4-(3fluorophenylpiperidin-4-ylidenemethyl)benzamide, a potent δ-opioid receptor agonist developed by AstraZeneca [39]. The purification is based on solidsupported reagents. The in-line IR analysis is used for monitoring of the last input flow stream. 5.5.13 Synthesis of TAC-101 Analogs

1,3,5-Tribromobenzene was decorated by a sequence of three combinations of a Br/Li exchange reaction and subsequent quenching reaction with an electrophile (Figure 5.15) [40]. The microreactor system consisted of six micromixers and

5.5

Multistep Syntheses in Coupled Microreactors/Chips

Br 0 °C

Br

Br BuLi

M1

R1 0.22 s R2 M2

0 °C

2.2 s M3

R3SiX

R3 0.14 s R4

BuLi

M4

0 °C

6.7 s M5

R′3SiX BuLi OCN

R5 M6

tR5 s

SiR3

R6 1.3 s

H N

R′3Si O

CO2Me

CO2Me

Figure 5.15 Schematic representation of the microfluidic setup for the synthesis of TAC-101 analogs. (Nagaki et al. [40]; reproduced by permission of the Royal Society of Chemistry).

six micro tubular reactors. Analogs of TAC-101, a synthetic retinoid, could be obtained within 13 s in good yield (49–77% yield). 5.5.14 Multistep Enzymatic Synthesis to 2-Amino-1,3,4-Butanetriol

De novo metabolic engineering offers chances for multistep enzymatic synthesis to give in one step complex molecules [41]. The pharmaceutical building block and diastereoisomer 2-amino-1,3,4-butanetriol is synthesized from simple achiral substrates using two enzymes, transketolase (TK) and transaminase (TAm). His6-tagged TK and Tam were immobilized onto Ni-nitrilotriacetic acid (NTA) agarose beads, which were packed into a tube to each give an enzymatic micro-flow reactor. The two latter microreactors were cascaded in series. In first step, l-erythrulose was synthesized from lithium-hydroxypyruvate and glycolaldehyde and thereafter converted to 2-amino-1,3,4-butanetriol using (S)-methylbenzylamine as amine donor (Scheme 5.17). Eighty-three percent conversion was reached in 20 min at 60 mM concentration. A kinetic investigation yielded the reaction constant. 5.5.15 Multistep Enzymatic Synthesis to 𝛅-D-Gluconolactone

Microfluidic glass chips were functionalized with poly(2-hydroxyethyl methacrylate) polymer brushes as anchors for co-immobilization of the enzymes glucose-oxidase and horseradish peroxidase [42]. In this cascade reaction, glucose-oxidase converts D-glucose into δ-D-gluconolactone and hydrogen peroxide (Scheme 5.18).

83

84

5 Exploring New Paths – New Chemical Transformations

O

O HO

OH

HOOC

Transketolase O

Hydroxypyruvate (HPA)

OH

HO

TPP, Mg2+

CO2

OH

Glycolaldehyde (GA)

L-Erythrulose (ERY)

NH2

Transaminase (S)-alpha Methylbenzylamine (MBA) PLP O

NH2

Acetophenone

OH

HO OH

2-Amino-1, 3, 4-butanetriol = ABT Scheme 5.17 Two step pathway for the synthesis of 2-Amino-1, 3, 4-butanetriol (ABT) [41].

H OH OH

OH HO

H H OH H

OH

H

GOx

O2

OH OH

D-Glu

HO

O S

N

S

O

N N ABTS

N O

S

S O

OH

1/2H2O2

HRP

HO

OH H O

H2O2

O

H H OH GDL

S O

O S

N N

N

N S

O S OH O

ABTS+

Scheme 5.18 Multienzymatic cascade reaction using glucose-oxidase and horseradish peroxidase [42].

Atomic force microscopy, Fourier transform infrared (FTIR) spectroscopy, and field emission scanning microscopy were used for characterization of such supramolecular architecture. The enzyme-functionalized glass chips performed a bi-enzymatic cascade reaction to measure glucose in human blood samples with high selectivity and reproducibility (Figure 5.16). A fast analysis was reached in 20 s with a detection limit of 60 mM and the results agreed with analysis common to conventional hospital laboratory.

5.5

Multistep Syntheses in Coupled Microreactors/Chips

O Br O Si

CH2OH CH2

O

O

O

+ O

Atom transfer radical polymerization

O

Br

O

O Si

OH

O

O O =C CH2

CH2OH CH2 O O =C

CH2

C CH3

C n CH3

O Br

O

O Si O

CH2

O =C O CH2 CH2OH

O Br

O

O Si

Initiator monolayer

O

O

O C

O O =C CH2 [

C

CH2

CH3

CH2 [

CH3 CH2 C

O=C O CH2 CH2OH

C ]n CH3 CH3 C ]n

O= C O

1.

O

O

O

CH2 [ OH

2. O

N

CH2 CH2OH

O O C

GOX

O CH2 [

CH2 CH2OH

PHEMA

CH2 C CH3 CH3 C

C ]n CH3

C

HRP

O O C

O C (CH2)2 O C O

(CH2)2 O C O (CH2)2 (CH2)2 O O O= C O= C

GOX + HRP

CH2

[

CH3 CH2

H N

HN O

O

(CH2)2 (CH2)2 O C O C O O (CH2)2 (CH2)2 O O O= C O=C

CH2OH CH2 O O =C

O= C O

N O

Br O

CH3 C n

PHEMA

N O

CH2OH CH2

CH3 CH2 C

O OH HO Si OH

+

]n

CH2

[

]

CH2 C C n CH3 CH3 CH3

CH3 CH2

C

C

]n

C= O O

C= O O

C =O O

C=O O

(CH2)2 O C O (CH2)2

(CH2)2 O C O (CH2)2

(CH2)2 O C O (CH2)2

(CH2)2 O C O (CH2)2

C O O O

O N

O

C O

C O N

O

GOX

O NH

HRP

C O

O NH

O

PHEMA functionalized with NHS

PHEMA with co-immobilized GOX and HRP

Figure 5.16 Enzyme immobilization onto glass surface. From the first monolayer via various chemical functionalizations toward the co-immobilization of the two enzymes glucose-oxidase and horseradish peroxidase. (Costantini et al. [42]; redrawn from said reference).

85

5 Exploring New Paths – New Chemical Transformations

Br/Li exchange reaction Ar–Br + BuLi Ar–Br

F F + 2LiF

F F

F

Ar

Ar

tR2 s R2 M3

T°C

F F

F F

F F

F F

F

R1 M2

T°C F F

F

F F F 2Ar–Li + F

tR1 s M1

BuLi

Reaction with octafluorocyclopentene

Ar–Li + BuBr

F F

R3

F F

MeOH

Ar

F

F F Ar

Figure 5.17 Schematic representation of the microfluidic setup for the synthesis of the diarylethenes in two steps. (Asai et al. [43]; reproduced by permission of Wiley-VCH). (a) Conversion of 100

96100 97 94

95

96 93 82

Br 75 74

Me

S

81 94 97 97 94 96 91 100 78 83 89 100 98 100 88 75 84 95 83100

79

75

95 97 98

(b) Yield of 59

Me 63 S

Me

70

67 70

70

70

24 19

44

3

0

0

0 0

70 61 56 38 21 4 0 72 69 77 81 81 79 66

73 76 74 86

83 85 86

(c) Yield of 6 SMe

0

1

7 0 0

0

0

42 54 71 74

24

0 0

0 11 14 22 34 51 67 76 −20 0 1 3 5 9 0

0

10−1.5 10−1.0 10−0.5

A

10 tR1/s

Br

Li

BuLi

Me X (X = S, O)

X

20

78 73 67

°C

19 Me

0

0

0

0.5

10

0 10

Tl

86

−40

1.0

Ring-opening reaction

Me

E: Electrophile

Me X

Li

E: Electrophile

E Me B

X

Me

Figure 5.18 A: Temperature–residence time map for the Br/Li exchange reaction with nBuLi followed by the reaction with methanol. (a) Contour plot for conversion of reactant, (b) contour plot of the yield of

X

E

2-methylbenzofuran, and (c) contour plot of the yield of 1-hydroxy-2-(prop-1-ynyl)benzene. B: underlying reaction scheme showing the ring-opening step. (Asai et al. [43]; reproduced by permission of Wiley-VCH).

References

5.5.16 Diarylethene Synthesis in Two Steps

Diarylethenes were synthesized in an integrated flow microreactor system via intermediate heteroaryllithium formation and reacted with octafluorocyclopentene in a follow-up step [43] (Figure 5.17). Process simplification is achieved, since the usually applied cryogenic conditions (100 ∘ C) to high boiling points such as toluene give good conversion. Heptane was added as cosolvent, both to enable higher catalyst activity and to lower pressure drop due its lower viscosity. The latter has the effect of allowing lower pressures for the oxygen feed which has a positive effect on the catalyst lifetime. Only by virtue of adding TEMPO or phthalimide-N-oxyl (PINO formed in situ from N-hydroxyphthalimide (NHPI)) the conversion could be increased to quantitative amounts. The reason is that stable nitroxyl radicals are generated that act as effective hydrogen shuttles in the catalytic cycle and are also potent oxidation inhibitors, that is, prevent the overoxidation of the benzylalcohol to the corresponding carboxylic acid. A 2 : 1 mixture of n-heptane/dioxane gives the best results (42% conversion to benzaldehyde, selectivity >99%, traces of benzyl formate 99% enantiomeric excess are achieved. Kinetic investigations revealed pseudo-first-order kinetics with ΔH = 18 ± 2 kcal mol−1 and ΔS = −38 ± 3 cal mol−1 K−1 .

O

1 M N-methyl pyrrolidinone 0.03 M acetic acid in water

OH

O

OH

PhCH2COOH DCC-DMAP, DCM

Microreactor (240 °C, 200 bar, 4-Hydroxy-2-cyclopentenone Furfuryl alcohol (FA) residence time 1.5 min) (R,S)-1 98%

O

O O

80%

CH2Ph

(R,S)-2

Immobilized CAL B t-BuOH, 1% water

OH (R)-1

Immobilized Pen G acylase Diisopropyl ether E > 200

O

O

O

+ O O (R)-2

CH2Ph

O OH (S)-1

t-BuMe2SiCl, (C2H3)3N, DMAP, dry. THF

OTBDMS (S)-3

Scheme 6.14

Conversion of furfuryl alcohol to 4-hydroxy-2-cyclopentenone [34].

6 Activate – High-T Processing

6.3.21 Hydrothermal Treatment of Glucose

The alkaline hydrothermal treatment of glucose and biomass, such as cellulose and straw, yields lactic acid (LA) in a simple way [35]. The choice of base is central and when screening different bases, hydroxides of alkali earth metals were found to show the largest activation. Initial batch processing was transferred into flow processing. Best results were found here when using barium hydroxide for the reaction with glucose. Yields of up to 57% were achieved in only 3 min, which are among the highest reported performances. High-temperature processing was done up to 280 ∘ C (Figure 6.14). This method is applicable also to raw cellulosic biomass and thus can compete with established biotechnological fermentation processes. Including side products, different commodity chemicals such as LA, formic acid (FA), acetic acid (AA), and glycolic acid are formed, which shows the potential of the new hydrothermal-based process. 6.3.22 Tetrahydroisoquinoline Synthesis

Tetrahydroisoquinoline (THIQ) core represents are found in natural products and biologically active compounds [36]. A flow synthesis of THIQs was achieved starting from laterally lithiated aziridines by a thermally induced isomerization reaction. This reaction sequence was achieved in a microreactor system consisting of two T-shaped micromixers, two tube reactors, and three precooling units (P1, P2, and P3) (Figure 6.15). This allowed for precise temperature control, which

35 30

Yield (%)

114

25

LA

20

FA

15

AA

10 5 0 150

170

190

210 230 T (°C)

250

270

290

Figure 6.14 Yield of lactic acid (LA), formic acid (FA), and acetic acid (AA) as a function of temperature in the hydrothermal treatment of glucose. (Esposito and Antonietti [35]; reproduced with permission from Wiley-VCH).

6.3

From Reflux to Superheated – Speeding-Up Reactions

is crucial to decide which product is formed. Depending on the temperature chosen, two different products can be formed from entirely the same molecules. At a temperature 350 ∘ C) and a short reaction time to minimize side reactions. This is accomplished using high-boiling point solvents, such as tetraglyme and diphenyl ether. Alternatively, the reaction can also be carried out using solvent-free conditions, which, however, often leads to complicated work-up due to unwanted precipitation. The easy handling of high-p,T flow systems widens much the scope in solvent choice. For the flow operation, tetrahydrofuran was used under superheated conditions, although it is a typical low-boiling point solvent. The maximal used

6.4

Solvent-Scope Widening by Virtue of Pressurizing Existing High-T Reactions

temperature of 360 ∘ C is almost six times higher than its boiling point, which nicely illustrates the novel process window. In this way, pyridopyrimidinones and hydroxylquinolines were synthesized in moderate to high yields (60–95%) (Scheme 6.16). An easy work-up followed the reaction.

N

NH2 +

Ph2O

O COOEt

OH - 92% yield - T = 360–370 °C - p = 130 bar - t = 45 s - X Cube flash Scheme 6.16

Intramolecular thermal benzannulation [37].

6.4.3 Catalyst-Free Transesterification and Esterification of Aliphatic and Aromatic Acids

The reaction conditions of (trans-) esterifications can be easily transferred to supercritical conditions in flow (MeOH: T c = 239 ∘ C; pc = 81 bar; EtOH: 268 ∘ C; 61 bar) [22]. Under such supercritical conditions, the transesterification of 2-phenyl acetic acid ethyl ester with methanol and the esterification of benzoic acid with ethanol could be performed as a catalyst-free process. While almost no reaction is observed below 200 ∘ C, completion of reaction can be fast achieved under supercritical conditions in flow. The methanol-based transesterification was performed much above the critical temperature at 350 ∘ C and quantitative conversion was achieved in 8 min. The ethanol-based esterification could not be operated at high temperature (300 ∘ C), since the reactant was thermally sensitive. Lowering the temperature gave rise to a longer reaction time of 12 min. The use of (low boiling) alcohols facilitates in both cases the work-up due to easy evaporation. 6.4.4 Aminolysis of Epoxides

For the aminolysis of epoxides, the addition of a small amount of a polar protic solvent to a polar aprotic solvent (mixtures of water/ethanol and acetonitrile) could accelerate the aminolysis reaction [18]. Such dedicated solvent mix could not be used under normal reflux conditions, unless accepting the drawbacks: extended reaction times due to the lowering of the reaction temperature to prevent solvent boiling.

117

118

6 Activate – High-T Processing

6.5 New Temperature Field for Product and Material Control 6.5.1 Palladium-Catalyzed Aminocarbonylation

The Pd-catalyzed aminocarbonylation can be shifted to two kinds of products in a microreactor depending on the pressure and the temperature (Scheme 6.17, Figure 6.16) [38]. On one hand, the yield of amide increases at higher temperature and lower CO pressures. On the other hand, α-ketoamide can be stimulated at high CO pressure and lower temperatures. The key to this reaction switch is the enhanced gas–liquid mass transfer in the pressurized regime.

O N H N

Br +

+

CO

Pd(OAc)2/Xantphos

O

NC

+ O

DBU

NC

O

O N O

NC

Scheme 6.17 Aminocarbonylation of 4-bromobenzonitrile with morpholine in microflow [38].

2.5 O

O N

2.0

NC

O 39-B

O N

NC

O 39-A

1.5 Ratio 39-B/39-A 1.0

14.8 bar 7.9 bar 4.5 bar

0.5

0.0 95

105

115

125

135

145

155

165

Temperature (°C)

Figure 6.16 Influence of temperature and pressure on the product outcome for the aminocarbonylation in flow. (Murphy et al. [38]; reproduced with permission from Wiley-VCH).

6.5

New Temperature Field for Product and Material Control

6.5.2 Aminolysis of Epoxides

For the epoxide aminolysis, a kinetic study of the model β-amino alcohol formation was made using styrene oxide and 2-aminoindane as the reagents [18]. Parameter screening was performed in a short time with very low reagent consumption, enabling manifold experiments. Due to the asymmetry of styrene oxide and the presence of a primary amine (i.e., 2-aminoindane) three different regioisomers can be formed. The desired product was the monosubstituted isomer. In total, morpholine (Scheme 7.1 and Figure 7.5). The rate enhancements follow this sequence, being 2.7, 1.7, and 1.5, respectively, at 600 bar. The dependence of the second-order reaction rate constants versus pressure yields the activation volumes ΔV≠ for the different reactions. The reactivity (kinetic constants) order of the leaving group (F > Cl > Br) mirrors in the ΔV≠ -values of −58.0, −41.7, and −32.7 cm3 mol−1 . H N

F

X

+ O2N

N

THF X

n

n=0 n = 1, X = CH2 n = 1, X = O

n

O2N : pyrrolidine : piperidine : morpholine

Scheme 7.1 Aromatic nucleophilic substitution of 1-fluoro-4-nitrobenzene with cyclic amines [12].

7.3.3 Diels–Alder Reaction with Furylmethanols and Cyclopentadiene

In the same capillary microreactor, the influence of pressure on the Diels–Alder reaction of 2-furylmethanol and 3-furylmethanol with three maleimides – methyl-, benzyl-, and phenylmaleimide – was determined [10]. A 1.6-fold higher reactivity of 3-furylmethanol than 2-furylmethanol at 1 bar was found, which

7.3

k = kobs/(amine)(M–1.s–1)

0.04

Pressure as Direct Means – Activation Volume Effects and More

Pyrrolidine R2 = 0.994

Piperidine 0.03 Morpholine 0.02

R2 = 0.990

0.01

R2 = 0.997

0 100

300

500

700

Pressure (bar) Figure 7.5 Aromatic nucleophilic substitution of 1-fluoro-4-nitrobenzene with cyclic amines: rate constant at four different pressures. (Benito-Lopez et al. [12]; reproduced with permission from Royal Society of Chemistry).

was explained by the change of atomic coefficient of the Highest occupied molecular orbital (HOMO) orbital on to C-2 carbon. At 600 bar, a 1.7-fold increase in reaction rate for 2-furylmethanol and the three maleimides was determined. The reaction of 3-furylmethanol was not influenced by pressure. For another Diels–Alder reaction, larger increases in reaction rates at even lower pressures were found. The reaction of cyclopentadiene and three maleimides yields exclusively the endo-Diels–Alder adduct (Scheme 7.2). The order of reactivity was phenylmaleimide > methylmaleimide > benzylmaleimide in batch experiments at 1 bar. Flow experiments were done in a specially designed high-pressure microreactor chip with a tubular channel and 690 bar resistant fibers for feed/product flows were inserted into powder-blasted inlets/outlets. Determined by the latter interconnection, such assembled microreactor could finally work up to pressures of 300 bar. Diels–Alder experiments at pressures of 150 bar changed the above-mentioned order of reactivity to benzylmaleimide > methylmaleimide > phenylmaleimide, since the respective rate constants were increased 14-, 2.8-, and 2.0-fold compared to batch conditions at 1 bar. This shows the large impact of substituents on the pressure-boost of a reaction, most likely via the activation volume effect. O +

N R O

O RT, CDCl3 N R O R = Me, Ph, Bn

Scheme 7.2 Diels–Alder reaction of cyclopentadiene [10].

135

136

7 Press – High-p Processing

7.3.4 Aza Diels–Alder Reaction

The aza Diels–Alder reaction of Danishefsky’s diene with an imine to 1,2diphenyl-2,3-dihydropyridin-4(1H)-one showed almost a doubling of the reaction rates from 1 to 300 bar (Scheme 7.3) [10]. An activation value ΔV≠ of −33 cm3 mol−1 was determined which is similar to comparable Diels–Alder reactions. OMe N

+

TMSO

Ph

N

MeOH, H+

Ph

O

Ph Ph

Danishefsky’s diene Scheme 7.3 Aza Diels–Alder reaction of Danishefsky’s diene with an imine to yield 1,2diphenyl-2,3-dihydropyridin-4(1H)-one [10].

7.3.5 Esterification of Phthalic Anhydride

The esterification of phthalic anhydride with methanol was performed at various temperatures in a continuous-flow glass microreactor at 110 bar using supercritical CO2 as cosolvent (Figure 7.6) [13] Even by high-pressure operation alone, a substantial rate enhancement of a 53-fold increase was achieved at 110 bar and 60 ∘ C, when compared to batch experiments at 1 bar at the same temperature. Using supercritical CO2 as a co-solvent, this can be further increased to a 5400-fold increase. The effects are related to changes in activation energies, possibly being related to pressure-induced changes in reaction mechanisms, negative molar activation volume, and surface (catalytic) effects.

Heated zone 4 5

1 2

6 3

Cooled zone

Figure 7.6 Microreactor for the esterification of phthalic anhydride with methanol. (BenitoLopez et al. [13]; reproduced with permission from the Royal Society of Chemistry).

7.4

Pressure for Advanced Fluidic Studies – to be Used for Shaping Materials and More

7.4 Pressure for Advanced Fluidic Studies – to be Used for Shaping Materials and More

The operation in new high-p windows may give rise to new dispersion phenomena for a given set of materials; basically because the determining forces can be arranged in a new way. The example of novel supercritical dispersing was taken here as motivation. When dispersing immiscible fluids into each other, pressure can change the volume of the dispersed phase (as given for gas/liquid contacting) or the dispersing-determining properties of both phases such as the interfacial tension, viscosity, and density (as given for gas/liquid and liquid/liquid contacting). In this way, different flow patterns and dispersion features (e.g., droplets of new diameter or shape) may be generated [14]. The new hydrodynamics may even open opportunities for material synthesis. Dispersed droplets can be converted to solids by chemical reaction, in particular by polymerization. This has not been explored so far, but is proposed here.

(a)

200 μm

(b)

(c) Figure 7.7 Microsystems for (a) droplets/emulsions and jet formation, (b) formation of emulsion at low pressure, and (c) formation of emulsions and jets in high-pressure supercritical fluids. (Lorber et al. [15]; reproduced with permission from Royal Society of Chemistry).

137

138

7 Press – High-p Processing

7.4.1 scCO2 Droplets or Jets in Liquid Water

High-pressure devices based on co-flow arrangements of small silica tubings with millimeter internals (“millifluidics”) were fabricated and used to generate stable droplets from pressurized fluids such as supercritical fluid–liquid systems [15]. scCO2 droplets or jets were formed in liquid water without need for stabilization by surfactants (Figure 7.7). The scCO2 jets had various shapes, including large straight jets, thin jets breaking into droplets, wavy jets, and more permutations of such shaping. By surface modification with perfluoroalkyl trichlorosilane, scCO2 wets the tubing and the phases are reversed to liquid water in scCO2 . Variation of the density and viscosity of scCO2 with pressure or temperature allows a transition from the dripping to jetting regime. In some cases, such transition induction is exclusive, that is, at fixed constant velocity, only pressure allows to switch from the dripping to the jetting regime. The dynamic phase diagram can be switched reversibly by simply changing the pressure or the temperature. This gives an alternative to the common switching of flow rates or fluids for the same motif.

References 1. For a review on high pressure

2.

3.

4.

5.

6.

7.

applications in organic chemistry: Benito-Lopez, F., Egberink, R.J.M., Reinhoudt, D.N., and Verboom, W. (2008) Tetrahedron, 64, 10023–10040. Murphy, E.R., Inoue, T., Sahoo, H.R., Zaborenko, N., and Jensen, K.F. (2007) Lab Chip, 7, 1309–1314. Tiggelaar, R.M., Benito-Lopez, F., Hermes, D.C., Rathgen, H., Egberink, R.J.M., Mugele, F.G., Reinhoudt, D.N., van den Berg, A., Verboom, W., and Garderniers, H.J.G.E. (2007) Chem. Eng. J., 131, 163–170. Trachsel, F., Hutter, C., and von Rohr, P.R. (2008) Chem. Eng. J., 131, S309–S316. Choe, J., Lee, H., Kim, Y., Lee, S.M., and Song, K.H. (2008) J. Ind. Eng. Chem., 14, 66–70. Pimparkar, K. (2009) Studies in the hydrogenation of biorenewable feedstocks. PhD dissertation. Michigan State University. Lee, H.-J., Shi, T.-P., Busch, D.H., and Subramanian, B. (2007) Chem. Eng. Sci., 62, 7282–7289.

8. Abdallah, R., Meille, V., Fumey, B.,

9.

10. 11.

12.

13.

and de Bellefon, C. (2006) IMRET 9, International Conference on Microreaction Technology, Book of Abstracts, Potsdam, Germany, 2006, pp. 58–59. Nimmanwudipong, T., Runnebaum, R.C., Brodwater, K., Heelan, J., Block, D.E., and Gates, B.C. (2014) Energy Fuels, 28 (2), 1090–1096. Verboom, W. (2009) Chem. Eng. Technol., 32, 1695–1701. Kobayashi, H., Driessen, B., van Osch, D.J.G.P., Talla, A., Ookawara, S., Noël, T., and Hessel, V. (2013) Tetrahedron, 69 (14), 2885–2890. Benito-Lopez, F., Verboom, W., Kakuta, M., Garderniers, H.J.G.E., Egberink, R.J.M., Oosterbroek, E.R., van den Berg, A., and Reinhoudt, D.N. (2005) Chem. Commun., 2857–2859. Benito-Lopez, F., Tiggelaar, R.M., Salbut, K., Huskens, J., Egberink, R.J.M., Reinhoudt, D.N., Garderniers, H.J.G.E., and Verboom, W. (2007) Lab Chip, 7, 1345–1351.

References 14. Marre, S., Roig, Y., and Aymonier, C.

(2012) J. Supercrit. Fluids, 66, 251–264. doi: 10.1016/j.supflu.2011.11.029 15. Lorber, N., Sarrazin, F., Guillot, P., Panizza, P., Colin, A., Pavageau, B.,

Hany, C., Maestro, P., Marre, S., Delclos, T., Aymonier, C., Subra, P., Prat, L., Gourdon, C., and Mignard, E. (2011) Lab Chip, 11, 779–787.

139

141

8 Collide and Slide – High-c and Tailored-Solvent Processing It is a general aim of the chemist and process chemist to work at high concentrations, to have high amounts of products (for analysis), and to have high productivity at lower solvent load. The use of concentration to direct a reaction has in the recent years increasingly put into common focus in flow chemistry- research. Thus, high-c is on the way to become as powerful as high-T and probably more than high-p. Thus, before referring to the small state-of-the-art in high-c flow processing, some stimulation is provided for more dedicated high-c opportunities in batch. The high-c trend also has an industrial dimension. Eighty percent and 90% of mass utilization in a typical pharmaceutical/fine chemicals batch chemical operation is devoted to the solvent. The solvent also has a high fingerprint on the total toxicity profile. Chemists typically hardly give attention to solvent and solvent–reactant interactions, separability, heat/mass transfer, and so on. Pharmaceutical industry recently has given greater awareness to solvent issues in batch chemical operations [1–3].

8.1 Batch Process-Based Inspirations for High-c Flow Processes

Solvent-free reactions are one of the final green visions in chemical processing. Several papers from industry highlight the importance of reducing the solvent load [1]. Even more advanced ambitions can be followed, as are also given below. 8.1.1 Polypropylene and Polycarbonate Polymerizations

Radical polymerizations toward polypropylene and polycarbonate initiated by transition metal complexes require no solvents [4]. Several enzyme-catalyzed and ionic reactions can be also processed without solvent or with green solvents. The higher yield and/or higher selectivity are attributed to the higher concentration of the reactants.

Novel Process Windows: Innovative Gates to Intensified and Sustainable Chemical Processes, First Edition. Volker Hessel, Dana Kralisch and Norbert Kockmann. © 2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

142

8 Collide and Slide – High-c and Tailored-Solvent Processing

8.1.2 Enantioselective Thermal and Photochemical Solid-State Reaction

Solvent-free reactions have not only been performed with pure liquid reactants but also with solid-state reactants [5]. An instructive example is the enantioselective thermal and photochemical solid-state reaction of a chiral host compound with a prochiral guest. Reactions that proceed otherwise sluggishly, can be sped up this way. 8.2 Solvent-Free or Solvent-Less Operation – “Highest-c” 8.2.1 Bromination of 3-Bromo-imidazo[1,2-a]Pyridine

The synthesis of several hundreds of grams of 3-bromo-imidazo[1,2-a]pyridine by bromination was accomplished in a microreactor [6]. Although the yield was not improved, clear advantages were stated in terms of process safety, greener chemistry with less waste, and higher rates of reaction outputs in a robust process. The green approach was reasoned by higher concentrations and the absence of CCl4 generation as compared to the reference batch process. 8.2.2 Thiophene Bromination

The fast (milliseconds) bromination of thiophene was carried out solvent-free with pure bromine and thiophene [7]. These streams were fed into a micromixer at 50 ∘ C. Complete thiophene conversion was achieved. By varying the molar reactant ratio via the feed streams from 1.0 to 5.0, control over the degree of substitution from mono- up to fourfold-substituted thiophenes was possible (Figure 8.1). 8.2.3 Claisen Rearrangement of Substituted Phenyl Phenols

A green flow processing – without solvent and work-up – to synthesize allyl parasubstituted phenyl phenols (R = Me, t-Bu, Ph, MeO, CN) via Claisen rearrangement was developed in a capillary microreactor (Scheme 8.1). [8] Compared with the conventional batch processing under reflux, much higher yields and purities could be achieved at much shorter reaction time. A quantitative yield of 2-allyl4-methoxyphenol was obtained in 20 min reaction time in the microreactor. In contrast, batch processing resulted in a yield of only 80% for 3 h and 32% for 20 min reaction time, respectively. The microreactor crude product is transparent, whereas the batch crude product is black and viscous. This indicates a higher purity, as demonstrated then by finding several impurities formed in the HPLC (high-pressure liquid chromatography) traces of the conventional reaction.

8.2

Solvent-Free or Solvent-Less Operation – “Highest-c”

Selectivity (%)

100 90

2-BrT

80

2,5-DiBrT 2,3,5-TriBrT

70

2,3,4,5-TetraBrT

60 50 40 30 20 10 0 0

1

2 3 4 5 Molar ratio bromine : thiophene

6

Figure 8.1 Influence of the molar ratio bromine to thiophene on the selectivity of the bromination reaction; BrT denotes bromothiophene. (Loeb et al. [7]; reproduced by permission of Bentham Science Publishers).

O R

OH

Neat R

- 73–99% yield - T = 200–245 °C - t = 24–36 min - Tubular microreactor Scheme 8.1 Solvent-free Claisen rearrangement of allyl para-substituted phenyl ethers in a tubular microreactor [8].

This high reaction efficiency is enabled by the intensified heat exchange in the microreactor: the activation energy is directly provided when entering the channel, whereas in a batch reactor a longer heating-up period is required. It is only provided for the intrinsic kinetic needs; when reacted, the product leaves immediately and is not exposed to additional energy. Hence, the chance of overheating and decomposition/carbonization is reduced. Another effect is the sealed operation in a closed system which does not allow air to enter and cause oxidation. 8.2.4 Michael Addition

Michael addition reactions of several α,β-unsaturated carbonyl compounds were performed under solvent-free conditions by Löewe et al. (Scheme 8.2) [9]. Due to the exothermicity of these reactions, the addition of the olefin has to be performed slowly into a diluted amine solution to avoid detrimental accidents. The total reaction time in these cases ranges from 17 to 25 h. The reaction time was

143

144

8 Collide and Slide – High-c and Tailored-Solvent Processing

H N

+

N

CN

CN

- 99% yield - T = 80 °C - p = 3–4 bar -t =6s - Tubular microreactor + micromixer Scheme 8.2 Michael addition of dimethylamine to acrylonitrile in a tubular microreactor [9].

brought down to seconds when the batch procedure was transferred to continuous and performed in a microstructured reactor. The space-time yields were up to 650 times increased. 8.2.5 Peroxidation of Methyl Ethyl Ketone

The methyl ethyl ketone peroxidation cannot be accelerated through high temperatures, since significant decomposition of the peroxides would occur [10]. However, higher H2 O2 and H+ concentration increased the active oxygen content and thereby the yield. 8.2.6 Beckmann Rearrangement (High-c)

The sulfuric-acid-catalyzed Beckmann rearrangement of cyclohexanone oxime to ε-caprolactam dissolved in cyclooctane was investigated in a microreactor at high concentration of ε-caprolactam [11] (Scheme 8.3). HO

H

N H2SO4/SO3

Cyclohexanone oxime

N

O

ε-Caprolactam

Scheme 8.3 Beckmann rearrangement of cyclohexanone oxime to ε-caprolactam [11].

The microreactor used comprises a low-temperature split-and-recombine mixing zone followed by a high-temperature microchannel tube reactor (internal diameter: 250 μm; length: 0.50 m) operated at 100–127 ∘ C. The residence time of the reactants in the microreactor setup was about 10 s. Complete conversion and a selectivity of 99% were achieved (Figure 8.2). In literature, a selectivity of about 95% for the same reaction in a similar setup is reported, but at a uniform temperature of 120–130 ∘ C for mixing and reaction. A

8.2

Solvent-Free or Solvent-Less Operation – “Highest-c”

100

Selectivity (%)

98

96

94

Single temperature

92

Two temperatures 90

70

90

110

130

Temperature (°C) Figure 8.2 Selectivity of the conversion of cyclohexanone oxime into ε-caprolactam in a microreactor setup with one or two reaction temperatures; in the latter case, the

temperature of the mixing zone is kept at 65 ∘ C. (Zuidhof et al. [11]; reproduced by permission of Wiley-VCH).

key to enhance the selectivity to ε-caprolactam is to suppress the reaction during mixing. Thereafter, continuous processing (at 100 ∘ C) with internal recirculation was developed to allow constant recycling of the oleum/ε-caprolactam mixture, feeding of pure fuming sulfuric acid, addition of the cyclohexanone oxime/cyclooctane solution, and continuous separation of the cyclooctane-oleum/ε-caprolactam emulsion (see Figure 8.3). However, the purity of the produced ε-caprolactam is lower than given for commercial products without co-feeding of a solvent to the cyclohexanone oxime. The found purity of 98.7 wt% for the recycled microreactor operation (commercial purity: 99.5 wt%) means to demand for a removal of more than twice the amount of impurities. This results in higher investments and thus the current microreactor process provides no economically viable solution for an industrial processing. 8.2.7 [2+2] Photocycloaddition of a Chiral Cyclohexenone (High-c)

Asymmetric [2+2] photocyclization can generate enantiomeric pure cyclobutane scaffolds which are synthetically attractive. A diastereo-differentiating [2+2] photocycloaddition of a chiral cyclohexenone with cyclopentene was done using

145

146

8 Collide and Slide – High-c and Tailored-Solvent Processing

Cyclohexanone oxime

C1

M1

C2

First stage P1

M2 Second stage

S1

P2

S2

Oleum ε-Caprolactam-oleum mixture to 2nd stage Figure 8.3 Schematic process scheme for continuous processing of cyclohexanone oxime to ε-caprolactam (at 100 ∘ C) with internal recirculation of the oleum/ ε-caprolactam mixture and continuous

ε-Caprolactam-oleum mixture to neutralization separation of the cyclooctane-oleum/εcaprolactam emulsion. C: cooling, P: pump, M: mixer, s: splitter. (Zuidhof et al. [11]; reproduced by permission of Wiley-VCH).

a special microreactor equipped with UV-LED (light-emitting diode) lamps for photoreaction (see Scheme 8.4 and Figure 8.4) [12].

RO

O OR

O H

RO

O H

+ hv (365 nm)

+

Toluene at 25 °C

O

15 equiv.

1

O H H 2

O H H 3 cis-syn-cis

R= RO

O H

RO

O H

+ O H H 4

O H H 5

cis-anti-cis OMe a

b Scheme 8.4 Diastereo-differentiating [2+2] photocycloadditions of two chiral cyclohexenones with cyclopentene in toluene at 25 ∘ C. (Terao et al. [12]; reproduced by permission of Wiley-VCH).

Even at very high sample concentration (0.77 M), high conversion was accomplished in the microreactor (up to 80%), whereas the reaction under the same

8.2

R′OOC H

Solvent-Free or Solvent-Less Operation – “Highest-c”

R′ =

OH H OMe Figure 8.4 cis-syn-cis [2+2] Photocycloadduct of (−)-8-(methoxyphenyl)menthyl and 3-D model of the distorted skeleton. (Terao et al. [12]; reproduced by permission of Wiley-VCH).

100

Transmission (%)

80

60

40

20

0 0

2

4

6

8

10

12

Path length (mm) Figure 8.5 Transmission spectrum of (−)-8-(methoxyphenyl)menthyl in toluene solution at 365 nm (black: 47 mM; blue: 88 mM; green: 159 mM; red: 263 mM; gray: 490 mM; light green: 769 mM). (Terao et al. [12]; reproduced by permission of Wiley-VCH).

conditions did not proceed to completion in the batch reactor. Moreover, the highest diastereomeric excess (de) of photoproducts of the flow synthesis is obtained when using aromatic compounds such as naphthalene. The addition of such aromatic compounds can enhance diastereoselectivity, because of a shield effect by the π–π interaction between the aromatic compounds and the aromatic ring in the menthol derivatives. However, naphthalene strongly absorbs the light entered, which can be managed in the microreactor due to its increased quantum efficiency as an effect of much deeper light penetration (see Figure 8.5).

147

148

8 Collide and Slide – High-c and Tailored-Solvent Processing

The space-time yield of the microreactor was found to be much higher than for the batch for all concentrations. Particularly at a very high concentration, the microreactor was about 20 times higher than that in the batch. For the photoreaction with (−)-8-(methoxyphenyl)menthyl, a space-time yield of 55 mmol min−1 l−1 (72% conversion, 770 mM) was determined for the microreactor, whereas 3 mmol min−1 l−1 (63% conversion, 47 mM) was obtained in the batch. Such high spacetime yield values of continuous processing can help to overcome productivity limitations, typically known for photoreactions. Yet, due to the need of six lamps for irradiation for the flow synthesis and only one for the batch, the energy efficiency for the flow was not better. 8.2.8 Bromination of Toluene (Solvent-Free)

The efficiency of the bromination of toluene derivatives was significantly improved by utilizing a combination of hydrogen peroxide (30%) and hydrogen bromide (40%) in a microreactor [13]. The solvent-free reaction was activated visible-light photocatalysis. 8.2.9 Sulfonation of Nitrobenzene (Solvent-Free)

A simple and efficient synthesis protocol for the sulfonation of nitrobenzene under solvent-free conditions was developed in a microreactor under solvent-free conditions [14]. Under optimized reaction conditions, 94% conversion of nitrobenzene and 88% yield of m-nitrobenzene sulfonic acid was obtained at a short residence time of less than 2 s. In addition, the process safety has been significantly improved and reaction time has been reduced compared to the batch method. A combined microreactor-batch operation results in high selectivity of 92% for an overall reaction time of 4 h, which is much shorter than the best batch operation – 65% Oleum, 100–110 ∘ C, 10 h, 100% conversion, 91% selectivity (see Figure 8.6). Such operation has the potential for commercial production of kilograms of sulfonated products per hour. 8.2.10 Synthesis of Nitro Herbicides (High-c, Solvent-Free)

Chen et al. presented a safe, efficient, and selective synthesis of dinitro herbicides via a multifunctional continuous-flow microreactor applying a one-step dinitration with nitric acid as agent [15]. Compared to a conventional two-step batch process, reaction can be conducted without the need for a separation of intermediates, and much less solvent as well as solvent-free conditions have been employed. Risks of the heat release due to exothermic dinitration are eliminated. A batch

8.2

Solvent-Free or Solvent-Less Operation – “Highest-c”

149

Microreactor mode

Temperature control system Pump

Check valve

Coolant in T1 T3

Nitrobenzene

Pump

NaOH solution

T2

Check valve

Microreactor

Coolant out Sample valve

Batch reactor SO3 Figure 8.6 Combined microreactor and batch operation for the sulfonation of benzene. (Chen et al. [14]; reproduced by permission of Royal Society of Chemistry).

NO2 H N

HNO3 30–60 °C

1

1c

+ NO2 H N

NO2 NO N

NO2 H N

H N

NO2 HNO3 30–60 °C

1a

+ + NO2 NO2 N NO2

1d

1e

Scheme 8.5 Conventional two-step dinitration of N-(1-ethylpropyl)-3,4-xylidine (1). (Chen et al. [15]; reproduced by permission of the Royal Society of Chemistry).

operation is impossible or very hazardous under the microreactor conditions performing one-step dinitration under identical conditions because of the heat and mass transfer limitations (Scheme 8.5 and Figure 8.7). Pendimethalin and two other dinitro herbicides were generated at high yields within less than 2 s. For the first, a productivity of 4.32 tons per year is given. Due to the short residence time, undesired by-product is not formed anymore. By changing the HNO3 concentration from 98 to 65%, the molar ratio of products and by-products can be changed in a way so that downstream processing is facilitated.

NO2 1b

150

8 Collide and Slide – High-c and Tailored-Solvent Processing

H N

Mixing micropore

Microchannel

NO2 H N NO2

1

1a

NO2 NO N

HNO3

NO2 1b Figure 8.7 One-step dinitration of N-(1-ethylpropyl)-3,4-xylidine (1) in a microreactor system. (Chen et al. [15]; reproduced by permission of RSC).

8.2.11 Suzuki–Miyaura Reaction over Sol–Gel Entrapped Catalyst SiliaCat DPP-Pd

The Suzuki–Miyaura reaction was realized in flow using sol–gel entrapped SiliaCat DPP-Pd catalyst at notably high concentrated solutions (0.5–1.0 M) (Scheme 8.6) [16]. In addition, the flow rate is increased up to 1.0 ml min−1 which is considered to be a practically relevant threshold. Complete conversion in the coupling of several aryl halides has been achieved.

SiliaCat DPP-Pd K2CO3 X + R

B(OH)2 R

R

R

- 92–99% yield - T = 70 °C - p = 1 bar - t = 2.85 min - Column reactor

Scheme 8.6 Suzuki–Miyaura cross-coupling of diverse substrates using sol–gel entrapped SiliaCat DPP-Pd catalyst. (Pandarus et al. [16]; reproduced by permission of the American Chemical Society).

8.2.12 Enzyme and Coenzyme (High-c in Bioprocessing)

Concentration can be a chemical intensification factor as well in flow bioprocessing which is an emerging new technology [17]. Micro-flow enzymatic reactors are a major tool over here. Reactant concentration cannot always easily be enhanced as this may result in substrate inhibition. A more common way to boost the productivity, especially in the microspaces, is to increase the concentration of the enzyme and coenzyme, since, for example, much higher specific loads can be achieved due to enzyme immobilization, which are still favorably mixed by the passive flow mixing.

8.2

Solvent-Free or Solvent-Less Operation – “Highest-c”

151

The impact of enzyme and coenzyme inlet concentrations and flow ratios on the conversion of hexanol to hexanal and the volumetric productivity was analyzed [18]. The enzymatic oxidation of hexanol to hexanal (as green note fragrance) using nicotinamide adenine dinucleotide (NAD+)-dependent commercial alcohol dehydrogenase (from Saccharomyces cerevisiae) was performed in micro-flow tubular microreactors with internal volumes of 6 and 13 μl. Another micro-flow tubular microreactor with a volume of 2 μl internally comprised micromixers, as one had to deal with mass transfer in immiscible phases. Then, the enzyme and coenzyme inlet concentrations were varied. Best result was achieved for c (hexanol) = 5.5 mmol l−1 , c (NAD+) = 0.55 mmol l−1 , and c (NADH) = 0.092 g l−1 , which outperformed the batch processing. Twelve percent conversion of hexanol was given for the 6 μl microreactor after 72 s, whereas the macroscale reactor had only 5% conversion after 180 s. 8.2.13 Enzyme and Coenzyme (High-c in Bioprocessing)

Besides the enzyme concentration, the concentration ratio between enzyme and substrate is a relevant figure and is key to intensification. A preconditioning study utilized mixing enzyme and substrate in predefined ratio before reaction and so enabled a millisecond temporal analysis of enzymatic reactions [19]. Crucial element was a micromixer generating the flow of four lamination layers through a narrow channel, thereby reducing the diffusion lengths to a few micrometers and providing mixing times in the millisecond range. In this way, the enzymatic hydrolysis was performed using the enzyme/substrate ratio from 1 : 1 up to 3 : 1. Kinetic data collection was achieved in a very short time in the order of minutes. The fast and easy handling of the mixing device makes it a very powerful and convenient instrument for millisecond temporal analysis of bioreactions. The enzymatic activity determination of β-galactosidase was done by optical measurements using fluorescein di-(b-D-galactopyranoside) (FDG) as a substrate which gets converted to fluorescein and galactose by sequential hydrolysis (Scheme 8.7).

CH2OH OH O O OH

O

OO

CH2OH OH OH

O OH

O

OH

CH2OH OH O OH OH β-galactosidase

O

O OH

+ COOH

OH O

CH2OH OH OH

FDG

Fluorescein

OH 2X Galactose

Scheme 8.7 Hydrolysis of FDG using β-galactosidase to produce fluorescein and two galactose molecules. (Buchegger et al. [19]; reproduced by permission of AIP Publishing).

152

8 Collide and Slide – High-c and Tailored-Solvent Processing

8.3 Supercritical Fluids to Combine the Former Separated – Mass Transfer Boost

Supercritical fluids are predominantly processed in high-pressure/hightemperature stainless steel batch or continuous reactors, from milliliter up to liter scale. Disadvantages of the conventional techniques, however, are limitations to implement characterization techniques or to carry out fast screening and optimization of the synthesis conditions. In this context, microreactors can add the following performances which can specifically support supercritical processing [20]:

• advanced feedback control, for example, for temperature and feed streams, and overall better reproducibility in operation; in-situ reaction monitoring using sensor integration; fast screening of process conditions; fast mass and heat transfer for a process medium which favors such operation; low reagent consumption during optimization.

• • • •

These and more advantages of “Supercritical Microfluidics” are summarized in Figure 8.8. Vice versa, supercritical fluids have beneficial properties for microscale flows, such as

• high diffusivity = commonly one to two orders of magnitude higher than for liquid phases;

• gas-like viscosity = a few tens of μPa⋅s; • ability to dissolve a wide range of reagents thanks to the liquid-like density (typically 200–1000 kg m−3 ). Numerous existing processes Green Chemistry/Ecotechnologies Precise control of the operating conditions (reproducibility)

Supercritical fluids

Continuous variation of fluid properties (through P/T variations)

Microfluidics

Fast screening Fast heat and mass transfer

Supercritical fluids

P Liquid Pe

In situ and online characterization

Critical point

Solid Gas

Te

Low volume/reagent consumption (safety)

T

Wide range of solvents available Microfluidics

Figure 8.8 Main advantages when combining supercritical fluids and microfluidic systems. (Marre et al. [20]; reproduced by permission of Elsevier).

8.3

Supercritical Fluids to Combine the Former Separated – Mass Transfer Boost

153

Trachsel et al. - 400 μm wide - 340 μm deep - Si-Py + soldering Murphy et al. - 600 μm wide - 300 μm deep - Si-Py + soldering Tiggelaar et al. - 70 μm wide - 30 μm deep - Si-Py + epoxy glueing Marre et al. - 200 μm wide - 150 μm deep - Si-Py + compression part Goodwin et al. - 1000 μm wide - 127 μm deep - Metal + commercial fitt

50 45

Sustainable (CO2, ...)

40

Hydrothermal (H2O, ...)

Pressure (MPa)

35 30 25 20 15 10 Engineering

5

(Alkanes, NH3, ...) 0

0

100

200

Figure 8.9 Operating windows for several high-pressure, high-temperature supercritical fluids – sustainable engineering and hydrothermal fluids – and data sets for microsystems reported in literature:

300 400 Temperature (°C)

500

600

Trachsel et al. [21], Murphy et al. [22], Tiggelaar et al. [23], Marre et al. [24] and Goodwin and Rorrer [25]. (Marre et al. [20]; reproduced by permission of Elsevier).

The various supercritical microreactors actually differ quite considerably in their operating conditions, opening several high-pressure, high-temperature windows. Figure 8.9 provides here a compilation of the processing windows used by Trachsel et al. [21], Murphy et al. [22], Tiggelaar et al. [23], Marre et al. [24], and Goodwin and Rorrer [25]. The T, p conditions can be classified in three regions termed sustainable, engineering, and hydrothermal fluids. The differences in density, viscosity, fluid velocity, and fluid regime (as characterized by the Ee number) are given when comparing supercritical microand millifluidics (at various dimensions) and when comparing liquid-phase with supercritical microfluidics (see Table 8.1). In the latter cases, the enormous differences in the operating regimes are evident. Yet, this holds similarly when going from milli- to microdimensions. Supercritical operation can become laminar which, for example, may be desirable in case of reproducible droplet formation of immiscible supercritical phases. Yet, turbulent conditions can be realized as well in supercritical microfluidics.

700

154

8 Collide and Slide – High-c and Tailored-Solvent Processing

Table 8.1 Comparison of fluid and hydrodynamic parameters for water flows under supercritical or liquid conditions at different scales. Density Viscosity Typical fluid dh (mm) Re (kg m− 3 ) (𝛍Pa s) velocity (ms− 1 )

Sc-millifluidics (1/4 in. tubing) (sc-water, 400 ∘ C − 25 MPa)

166

29

10− 1 to 1

3.8

2000–20 000

Sc-millifluidics (1/8 in. tubing) (sc-water, 400 ∘ C − 25 MPa)

166

29

10− 2 to 10− 1

1.6

100–1000

Liquid microfluidics (liquid water, 25 ∘ C − 0.1 MPa) Sc-microfluidics (sc-water, 400 ∘ C − 25 MPa)

997

890

10− 3 to 10− 2

0.1

0.1–1

166

29

10− 3 to 10− 1

0.1

0.5–500

Marre et al. [17]; redrawn after [20].

The beneficial mass transfer effect is evident when coaxially injecting a Rhodamine-dyed ethanol flow into pure ethanol, which can easily be realized by inserting two capillaries in each other. An operation at 25 ∘ C, with EtOH being liquid, is compared to an operation at 250 ∘ C, with EtOH being supercritical (T c = 241 ∘ C and Pc = 6.1 MPa). In the liquid operation case, a flow focusing effect is observed with the inner flow forming a thinning jet (see Figure 8.10). In the supercritical case, the width of the initial jet appears larger than for the liquid case. Fast diffusion of Rhodamine molecules to the outer pure ethanol flow is given. This is followed by a further blurring and widening of the jet-diffusion zone, finally resulting in a fast pink-reddish coloration of the outer flow. Residence time distribution is much impacted by the facilitated Taylor–Aris dispersion in microchannels under supercritical conditions (Figure 8.11). This is evidenced by the effect of miniaturization of the microchannel width (respectively, diameter) and the fluid viscosity which stands for the change to supercritical. This Liquid EtOH, 25 MPa – 25 °C EtOH Rhodamine in EtOH

Flow focusing effect

Slow diffusion process 100 μm

EtOH

scEtOH, 25 MPa – 250 °C EtOH Rhodamine in EtOH EtOH

Fast diffusion process 100 μm

Figure 8.10 (a,b) Optical imaging of the jet (Rhodamine in EtOH) shaping after coaxial injection into an EtOH solution at 25 MPa, for two different temperatures (sub- and supercritical conditions). (Marre et al. [20]; reproduced by permission of Elsevier).

Supercritical Fluids to Combine the Former Separated – Mass Transfer Boost

100 μm

100 μPa.s

250 μm

500 μPa.s

500 μm

1000 μPa.s

0.7

(a)

0.9

1.1

155

30 μPa.s

E(θ)

E(θ)

8.3

1.3

θ = t / taverage

Figure 8.11 E curves (RTD, residence time distribution) calculated through the Taylor dispersion model when varying (a) the fluid viscosity (in a 100 μm wide capillary) and (b) the cylinder diameter (for a fluid viscosity of

1.5

1.6 mm

0.7

(b)

0.9

1.1

1.3

θ = t / taverage

10−4 Pa s). This is based on the diffusion of a 3 nm particle or large molecule in fluids flowing at v = 10−2 m s−1 and T = 50 ∘ C in a 1 m long flow reactor. (Marre et al. [20]; reproduced by permission of Elsevier).

can be decisive in the (nano-)particle synthesis and results in narrow particle size distributions, as demonstrated for the CdSe quantum dot synthesis in a microfluidic device. 8.3.1 Supercritical Hydrogenation of Cyclohexene

Supercritical CO2 (scCO2 ) is a promising solvent for environmentally friendly chemical synthesis and can replace more harmful solvents conventionally used. scCO2 has high diffusivity and can dissolve gases in large proportions. This can be of great help if mass diffusion limited processes are to be considered. In this way, a three-phase reaction (solid/liquid/gas) can be converted to a two-phase reaction (solid/supercritical) of much higher productivity. scCO2 was used as reaction solvent for the hydrogenation of cyclohexene in a packed bed silicon/glass microreactor (Figure 8.12) [21]. The mass-transfer limitations of conventional multiphase processing (gas-liquid-solid; 1 bar, 25 ∘ C) are overcome through single-phase flow processing (136 bar; 25 ∘ C; CO2 :C6H10:H2 = 90 : 5 : 5). The space-time yield is increased by one order of magnitude. The absence of any mass-transfer limitation is quite striking by the fact that further pressure increases have no notable impact. Only increasing temperature has a positive effect on the product yield. 8.3.2 Supercritical Hydrogenations of Double and Triple Bounds

A Pd-immobilized microchannel reactor (width: 200 μm, depth: 100 μm, length: 40 cm) was used for hydrogenation reactions in scCO2 [26]. The Pd immobilization was achieved via amino groups which were introduced onto the surface of

1.5

156

8 Collide and Slide – High-c and Tailored-Solvent Processing H2

Two phase flow

CO2

Mixing action Packed bed

Outlet

400 μm C6H4

Cat. Flow direction

50 μm scCO2

Figure 8.12 Schematic representation of the used Si/glass microreactor using supercritical CO2 as the reaction solvent. (Trachsel et al. [21]; reproduced by permission of Elsevier).

the glass channel. A solution of microencapsulated palladium was fed through the microchannel reactor and cross-linking of the polymer was achieved via heating. The loading of palladium catalyst was about 10 μg of palladium (inductively coupled plasma (ICP) analysis). A variety of substrates were converted to the desired products in nearly quantitative yields. Terminal and trans olefins and triple bonds were reduced fully to the respective alkanes. A strong activity boost was observed, since for all substrates investigated reaction times were less than 1 s. Selectivity issues could be resolved such as the specific reduction of a triple bond in the presence of a benzyloxy group. The activity boost was attributed, besides to the high diffusivity and the high solubility of hydrogen in scCO2 under the supercritical conditions, to the large surface area of the catalyst. In such a way, the productivity was increased by one order of magnitude as compared to an earlier developed triphase microfluidic system (from 0.01 mmol h−1 per channel to 0.1 mmol h−1 per channel). Assuming a molar mass of 200 gmol−1 and 100 parallel microchannels in operation, this would give 2 g h−1 , which is fine for precious substances. The microchannel reactor could be reused several times without loss of activity.

8.3

Supercritical Fluids to Combine the Former Separated – Mass Transfer Boost

8.3.3 Ascaridole Synthesis under Photo-Supercritical Conditions

High-power LED technology was combined with high-pressure scCO2 operation in a photocatalytic milliliter-scale reactor [27, 28]. The simple tubular sapphire reactor provides a long irradiation zone which is illuminated by high-power visible LED, such as those used for automobile headlights. The efficiency of current LEDs is much higher in the visible region than in the UV. Accordingly, high concentrations of singlet 1 O2 were created, which allowed reaching high space-time yields. This was demonstrated in the synthesis of ascaridole from α-terpinene using 5,10,15,20-tetrakis(pentafluorophenyl)porphyrin (TPFPP) as a photosensitizer (Scheme 8.8). scCO2 operation led to at least two-times faster reaction times as compared to the conventionally used CCl4 . Quantitative conversion and stable operation over 8 h without fouling was achieved for flow rates of up to 12 ml h−1 . The ascaridole synthesis was scaled up by a factor of 3000, facilitated by the ease in scaling up the photolysis cell. 8.3.4 Citronellol Oxidation under Photo-Supercritical Conditions

The new photoreactor was also applied to the photo-oxidation of citronellol which is part of the synthesis of Rose Oxide, a fragrance of commercial value [27]. Supercritical operation at 180 bar led to 100% conversion to yield two compounds (52 : 42 selectivity). These products, as well as their hydroperoxide intermediates, are potentially explosive. The new flow photoreactor had a space-time yield of 70 mmol l−1 min−1 compared to 0.1 mmol l−1 min−1 of a conventional Schlenk reactor with an immersed LED array. 8.3.5 Near-scCO2 Enzymatic Biodiesel Synthesis

Enzymatic microreactors are gaining increasing importance. So far, no supercritical operation was taken here to intensify the biocatalytic processing. The following

O2, sCO2 TPFPP, hυ

O O

- 99% yield - p = 48 bar - Sapphire tube reactor Scheme 8.8 Continuous photo-oxidation of α-terpinene in supercritical CO2 [27].

157

8 Collide and Slide – High-c and Tailored-Solvent Processing

high-pressure enzymatic processing gives evidence for impact for a process of huge capacity, and can exemplify what might be possible for flow processing. An enzymatic biodiesel batch synthesis was improved by high-pressure, nearcritical carbon dioxide processing [29]. The biodiesel synthesis was evaluated both in batch and semi-continuous systems. Three immobilized lipases – Novozyme 435, Lipozyme RM IM, and Lipozyme TL IM from Novozymes – were tested. Lipozyme TL IM is the most effective lipase, showing the highest conversion. Biodiesel conversion from several edible and non-edible oil feedstocks was tested. Almost quantitative conversion (99.0%) was obtained using canola oil by employing repeated batch processes at 100 bar (Figure 8.13). The enzyme maintained 80% of its initial stability after being reused eight times. The authors argued that this method produces biodiesel energy-efficiently and environment-friendly. 8.3.6 Supercritical Water, Non-Catalytic Beckmann Rearrangement

There are only a few examples so far about the use of supercritical water (scH2 O) as solvent/reactant in micro-flow. Naturally, the harsh process conditions and the ε-caprolactam were generated via the non-catalytic Beckmann rearrangement of cyclohexanone oxime in scH2 O (up to 400 ∘ C and 25 MPa) in a micro-flow reactor [30]. The reaction was monitored with an in-line high-pressure and high-temperature Fourier transform infrared spectroscopy (FTIR). A yield of 100

80 FAME conversion (%)

158

60

40

20

One-pot reaction Stepwise addition of methanol Repeated batch reaction

0 1

2

3

4

5

6

7

8

9

10

Number of use Figure 8.13 Changes in stability of Lipozyme TL IM by fatty-acid methyl ester (FAME) conversion during different biodiesel synthesis processes. (Lee et al. [29]; reproduced by permission of Springer).

8.3

Supercritical Fluids to Combine the Former Separated – Mass Transfer Boost

almost 100% at 100% selectivity is achieved for a residence time as short as 1 s. This intensification was attributed to the much superior heat transfer provided by the microscale conditions. 8.3.7 Supercritical Water, Non-Catalytic Pinacol Rearrangement

Significant acceleration of the pinacol rearrangements was achieved by using scH2 O, especially near the critical point, even in the absence of any acid catalysts [30]. In turn, scH2 O replaces the conventional acid catalysts for both rearrangements. The rate of the pinacol rearrangement with scH2 O is massively increased by the factor of 28 200 (0.871 M HCl, 46.7 MPa, distillation conditions) (Table 8.2). The activation energy under supercritical conditions is substantially reduced to about one-third of the mentioned conventional process. This was attributed to a much higher local proton concentration encompassing the organic reactants (Table 8.3). Selectivity gains can be achieved by smart tuning of the acidity of scH2 O which opens a new reaction pathway (Figure 8.14). Reaction at weak acidity, as given in the near-critical region, prevents the path through the pinacol rearrangement, but rather proceeds via Diels–Alder addition between the dehydrated pinacol products. 8.3.8 Supercritical Water Oxidation

Supercritical water oxidation (SCWO) is used for the degradation of organic molecules [31, 32]. SCWO is typically run at slightly lower temperature than gasification (380–500 ∘ C, 25–30 MPa). Heat transfer is the key issue in such biomass conversion processes. For such issue, microreactors are advantageous. A quartz capillary (580 ∘ C, 28 MPa) was used to study the SCWO of acetic acid [33]. A more advanced design was then developed for a chip silicon–Pyrex microsystem (400 ∘ C, 25 MPa) (Figure 8.15) [24]. Higher performance is intended Table 8.2 First-order rate constants determined for the pinacol rearrangement. Temperature (K)

103 k1 (s− 1 )

623.0 637.5 647.5 652.5 667.5

Not detectable 0.35 × 0.02a) 2.46 ± 0.36a) 8.16 ± 0.75a) 1.40 ± 0.17a)

a) 95 % confidence limit. Ikushima et al. [30]; reproduced by permission of the American Chemical Society.

159

160

8 Collide and Slide – High-c and Tailored-Solvent Processing

Table 8.3 Activation energies and entropies for the pinacol rearrangement. P (MPa)

E a (kJ mol− 1 )

𝜟S(J k− 1 mol− 1 )

22.5 25.0 30.0 35.0 0.1b) 46.7b) 95.5b) 144.7b)

28.3 ± 2.1a) 55.7 ± 4.1a) 80.9 ± 5.5a) 93.4 ± 2.0a) 151.3 155.0 158.8 162.2

− 246.7 ± 13.3a) − 202.0 ± 11.3a) − 161.0 ± 9.0a) − 142.0 ± 6.7a) 93.4 103.1 113.1 121.9

a) 95% confidence limit. b) In 47.5 wt% H 2 SO4 aq solution at 343 K. Ikushima et al. [30]; reproduced by permission of the American Chemical Society.

H3C H3C

C

C

OH

OH

CH3 CH3

I No catalyst

No catalyst

scH2O and superheated H2O

scH2O

(573 ~ 723 K, 20 ~ 35 MPa)

(648 ~ 653 K, 22.5 MPa) H2C

C

CH2

C

C H3

CH3

H3 C III No catalyst

H3C

H3C

C C

O

CH3

C

C

II

CH3

CH3

Diels–Alder reaction H2 C

C H2

CH2 C

C CH3

CH2 CH3

IV Figure 8.14 Non-catalytic, scH2 O-mediated reaction paths yielding pinacol through rearrangement and 1,25-trimethyl-5-isopropenyl-1-cyclohexene through Diels–Alder reaction. (Ikushima et al. [30]; reproduced by permission of the American Chemical Society).

8.3

Supercritical Fluids to Combine the Former Separated – Mass Transfer Boost

T = 25 °C P = 25 MPa

300 °C < T < 400 °C

Product Water + H2O2

Temperature gradient 25 °C mm−1

Water + R–OH

5 mm

(a)

Gas–liquid segmented flow

Homogeneous flow

5 mm (b)

T = 350 °C/subcritical

Figure 8.15 (a) Scheme of the microsystem for SCWO of MeOH and PhOH with lowand high-temperature regions separated by confined zone acting as insulation (45 μl microreactor) and (b) gas–liquid segmented

T = 380 °C/supercritical methanol/water flow under subcritical (left) and fully mixed, single-phase under supercritical water conditions (right). (Marre et al. [20]; reproduced by permission of the American Chemical Society).

due to integrating special injection schemes to mix reagent in the early stage of the reaction. The sub- and supercritical water oxidation of methanol and phenol was studied therein. The corresponding aqueous solutions were mixed with a hydrogen peroxide solution at room temperature and the mixtures were passed to a highT section. The oxidation of phenol and methanol is within seconds completed at even higher conversion values as reported in literature. For methanol, up to 37% conversion is achieved in 2 s at 300 ∘ C, that is, under subcritical conditions. A massive conversion boost up to 80% is possible under supercritical conditions (380 ∘ C). This resembles reported performance in larger stainless steel equipment which, however, has to use much higher temperature (460 ∘ C).

161

162

8 Collide and Slide – High-c and Tailored-Solvent Processing

8.3.9 Supercritical-Acetonitrile, Nitriles from Carboxylic Acids

Organic nitriles are synthesized from the corresponding carboxylic acids in flow manner. A process-simplified method is developed based on the acid–nitrile exchange reaction with acetonitrile, which also constitutes the solvent (Scheme 8.9 and Figure 8.16) [34].

R

scMeCN

COOH

R

- 72–97% yield - T = 350 °C - p = 65 bar - t = 25 min - Stainless steel tubular microreactor

CN

Scheme 8.9 Formation of nitriles by the acid–nitrile exchange reaction with supercritical acetonitrile [34].

No catalyst or additives are needed by virtue of activation through hightemperature/high-pressure conditions. The latter enable supercritical processing of acetonitrile at 350 ∘ C and 65 bar. Benzoic acid is so converted to benzonitrile requiring only 25 min. Such a procedure can be applied to a variety of aromatic and aliphatic nitriles. 8.3.10 Self-Optimizing Continuous Reactions under Supercritical Conditions

Optimization algorithms such SMSIM using the modified Simplex routine were used for self-optimization of scCO2 flow reactors [35]. The objective function was the maximization of the yield to minimize the E factor, namely the number of kilograms of waste per kilogram of product.

O

NH

O

COOH O

+

O N H B

N C A

H2O C

N

O NH2 + AcOH

D

C

Figure 8.16 Mechanism of the acid–nitrile exchange reaction. (Cantillo and Kappe [34]; reproduced by permission of the American Chemical Society).

References

References 1. Constable, D.J.C., Jimenez-Gonzalez, C.,

2. 3.

4. 5. 6. 7.

8.

9.

10.

11.

12.

13.

14. 15.

16.

17.

and Henderson, R.K. (2007) Org. Process Res. Dev., 11, 133–137. Jessop, P.G. (2011) Green Chem., 13, 1391–1398. Capello, C., Fischer, U., and Hungerbuehler, K. (2007) Green Chem., 9, 927–934. Metzger, J.O. (1998) Angew. Chem., Int. Ed. Engl., 37, 2975–2978. Tanaka, K. and Toda, F. (2000) Chem. Rev., 100, 1025–1074. Pelleter, J. and Renaud, F. (2009) Org. Process Res. Dev., 13, 698–705. Loeb, P., Hessel, V., Klefenz, H., Loewe, H., and Mazanek, K. (2005) Lett. Org. Chem., 2, 767–779. Kong, L., Lin, Q., Lu, X., Yang, Y., Jia, Y., and Zhou, Y. (2009) Green Chem., 11, 1108–1111. Loewe, H., Hessel, V., Hubbard, S., and Loeb, P. (2006) Org. Process Res. Dev., 10, 1144–1152. Wu, W., Qian, G., Zhou, X.-G., and Yuan, W.-K. (2007) Chem. Eng. Sci., 62, 5127–5132. Zuidhof, N.T., de Croon, M.H.J.M., Schouten, J.C., and Tinge, J.T. (2012) Chem. Eng. Technol., 35, 1257–1261. Terao, K., Nishiyama, Y., Aida, S., Tanimoto, H., Morimoto, T., and Kakiuchi, K. (2012) J. Photochem. Photobiol., A, 242, 13–19. Yu, W.-B., Yu, D.-P., Zheng, M.-M., Shan, S.-T., Li, Y.-J., and Gao, J.-R. (2012) J. Chem. Res., 36, 258–260. Chen, Y., Su, Y., Jiao, F., and Chen, G. (2012) RSC Adv., 2, 5637–5644. Chen, Y., Zhao, Y., Han, M., Ye, C., Dang, M., and Chen, G. (2013) Green Chem., 15, 91–94. Pandarus, V., Gingras, G., Béland, F., Ciriminna, R., and Pagliaro, M. (2014, online) Org. Process Res. Dev. doi: 10.1021/op4003449. http: //pubs.acs.org/doi/abs/10.1021/ op500008h?prevSearch=pandarus &searchHistoryKey=. (a) Hessel, V., Tibhe, J., Noël, T., and Wang, Q. (2014) Chem. Biochem. Eng. Q., 28 (2), 167–188; (b) Dencic, I., Noël, T., Meuldijk, J., de Croon, M.H.J.M., and

18.

19.

20. 21.

22.

23.

24.

25. 26. 27.

28.

29.

Hessel, V. (2013) Eng. Life Sci., 13 (4), 326–343; (c) Fernandes, P. (2010) Int. J. Mol. Sci., 11, 858–879; (d) Miyazaki, M. and Maeda, H. (2006) Trends Biotechnol., 24, 463–470; (e) Marques, M.P.C. and Fernandes, P. (2011) Molecules, 16, 8368–8401; (f ) Bolivar, J.M., Wiesbauer, J., and Nidetzky, B. (2011) Trends Biotechnol., 29, 333–342; (g) Krühne, U., Heintz, S., Ringborg, R., Rosinha, I.P., Tufvesson, P., Gernaey, K.V., and Woodley, J.M. (2014) Green Process. Synth., 3, 23–31. Šali, A., Tušek, A., Kurtanjek, Ž., and Zelic, B. (2011) Biotechnol. Bioprocess Eng., 16, 495–504. Buchegger, W., Haller, A., van den Driesche, S., Kraft, M., Lendl, B., and Vellekoop, M. (2012) Biomicrofluidics, 6, 012803. Marre, S., Roig, Y., and Aymonier, C. (2012) J. Supercrit. Fluids, 66, 251–264. Trachsel, F., Tidona, B., Desportes, S., and von Rohr, P.R. (2009) J. Supercrit. Fluids, 48, 146–153. Murphy, E.R., Inoue, T., Sahoo, H.R., Zaborenko, N., and Jensen, K.F. (2007) Lab Chip, 7, 1309–1314. Tiggelaar, R.M., Benito-Lopez, F., Hermes, D.C., Rathgen, H., Egberink, R.J.M., Mugele, F.G., Reinhoudt, D.N., van den Berg, A., Verboom, W., and Gardeniers, H. (2007) Chem. Eng. J., 131, 163–170. Marre, S., Adamo, A., Basak, S., Aymonier, C., and Jensen, K.F. (2010) Ind. Eng. Chem. Res., 49, 11310–11320. Goodwin, A.K. and Rorrer, G.L. (2008) Ind. Eng. Chem. Res., 47, 4106–4114. Kobayashi, J., Mori, Y., and Kobayashi, S. (2005) Chem. Commun., 2567–2568. Bourne, R.A., Han, X., Poliakoff, M., and George, M.W. (2009) Angew. Chem. Int. Ed., 48, 5322–5325. For a review about continuous-flow in supercritical carbon dioxide: Han, X. and Poliakoff, M. (2012) Chem. Soc. Rev., 41, 1428–1436. Lee, M., Lee, D., Cho, J.K., Cho, J., Han, J., Park, C., and Kim, S. (2012) Bioprocess Biosyst. Eng., 35, 105–113.

163

164

8 Collide and Slide – High-c and Tailored-Solvent Processing 30. Ikushima, Y., Hatakeda, K., Sato, O.,

33. Maharrey, S.P. and Miller, D.R. (2001) Yokoyama, T., and Arai, M. (2000) J. Am. AIChE J., 47, 1203–1211. Chem. Soc., 122, 1908–1918. 34. Cantillo, D. and Kappe, C.O. 31. Tester, J.W., Holgate, H.R., Armellini, F.J., (2013) J. Org. Chem., 78 (20), 10567–10571. Webley, P.A., Killilea, W.R., Jong, G.T., and Barner, H.E. (1993) ACS Symp. Ser., 35. Parrott, A.J., Bourne, R.A., Akien, 518, 35–76. G.R., Irvine, D.J., and Poliakoff, M. (2011) Angew. Chem. Int. Ed., 50, 32. Vogel, F., Blanchard, J.I.D., Marrone, 3788–3792. P.A., Rice, S.F., Webley, P.A., Peters, W.A., Smith, K.A., and Tester, J.W. (2005) J. Supercrit. Fluids, 34, 249–286.

165

9 Doing More by Combining – Process Integration Process integration is the concept on large scale to make huge facilities cost and eco-efficient. A superb example is the Verbund strategy, introduced by BASF, which is based on production and process integration at the site, where an interlinked network of value chains secures critical mass for the main components. BASF’s first Verbund site in Asia in Kuantan gives an example of such modern process integration within a chemical park [1]. The Kuantan site has been developed by BASF together with the partner company Petronas, the Malaysian oil and gas company. The main raw material used is propylene as the starting material for the C3 value chain (see Figure 9.1). This creates an acrylic acid value chain and a C4 oxo alcohols value chain. Phthalic anyhydride completes the plasticizer business opened by the oxo alcohols. N-butane is converted finally to butanediol. Here and there, process integration is quoted in the microreactor literature, for example, at the level of multifunctional integration in reaction systems, and also in microreactor patent literature. There is, however, hardly any systematic compilation; also due to the fact that only a few scientific were reported so far. 9.1 Integration of Reaction and Cooling/Heating, Separation, or Other

The challenges of designing integrated microchips with a number of coherently performed microdevice functions and their combination into single microfluidic processes have been analyzed and reviewed [2]. As to be expected, several challenges and difficulties arise from the small scale of operation, such as complexities in the underlying physics and chemistry, and differences in the time constants of the participating units. In the following, concrete examples for reaction/operation integration are given; most often, however, not as chip reactors but as larger microstructured counterparts. 9.1.1 Integrated Micro-Steam Reformer-Catalytic Combustor for Methane Fuel Processing

In micro-fuel processing, system integration means most often reactor–reactor or reactor–process unit integration in one system. Process units are afterburners, Novel Process Windows: Innovative Gates to Intensified and Sustainable Chemical Processes, First Edition. Volker Hessel, Dana Kralisch and Norbert Kockmann. © 2015 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2015 by Wiley-VCH Verlag GmbH & Co. KGaA.

166

9 Doing More by Combining – Process Integration

PETRONAS Propane dehydrogenation

Natural gas

BPC Acrylic acid Propylene

Butylacrylate 2-EH acrylate

Phase I

Syngas/ Hydrogen o-Xylene Phase II

Central utility facility

Glacial AA

N-Butane

Oxo-C4 Alcohols

Oxo-C4 Alcohols

Phthalic anhydride

Plasticisers

Maleic anhydride

Butanediol

Phase III

Figure 9.1 BASF-Petronas production Verbund at Kuantan site. (Reproduced with permission from Springer).

evaporators, micro heat exchangers, or even membranes. With the exception of the latter, thermal coupling to improve efficiency is the motivation and leads to a compact overall system design. Micro-fuel processors consist of several such thermally coupled systems, with typical numbers of 4–6 (see diesel steam reforming, water-gas shift, and preferential oxidation microstructred reactors in Figure 9.2) [3, 4]. In another application, for methane-based hydrogen generation, integration was done for a microstructured steam reformer by coupling to a catalytic combustor [5]. The latter delivers the heat needed for the endothermic steam reforming reaction in co-flow. The reactor system was operated autonomously without external heat supply, under static and dynamic load conditions. Firing of the integrated combustor with a hydrogen/methane mixture gave best performance and was even not disturbed by the presence of carbon dioxide which simulates the anode off-gas firing. Under static load conditions, methane conversion rates were as high as about 70%. Dynamic tests show that start-up needs 10 KWth

Diesel STR/AFB 5 kWel, net PrOX reactor – 5 kWel, net

550 mm

STR/AFB HX-02

Figure 9.2 Diesel steam reforming, water-gas shift, and preferential oxidation microstructred reactors. (O’Connell et al. [5, 6]; Reproduced with permission from FhG-IMM).

9.1.3 Integrated Micro Reactor–Evaporative Cooler

An efficient and fast method for heat removal from a system is claimed in a patent and in open literature via evaporative cooling through a microfluidic Y-junction [7]. By soft-lithography techniques simple fluidic junctions are formed which bridge between channels transporting with refrigerant and N2 gas. This system integration is demonstrated for the vaporization of acetone, isopropyl alcohol, and ethyl ether. A parametric optimization results in refrigeration rates > 40∘ C/s and long lasting subzero cooling in the junction. 9.1.4 Integrated Microwave–Microreactor

An open-ended semi-rigid coaxial cable was utilized as a miniaturized microwave heating source in microwells made of poly(methyl methacrylate) and polydimethylsiloxane for the realization of a microreactor for mobile lab-on-a-chip devices [8]. Microwave fields coupled into the fluid by the open end of the miniature coaxial cable lead to localized dielectric and also, to a minor degree, resistive heating of the fluid in the microreactor. 9.1.5 Integrated Enzyme Microreactor–Extractor

A microreaction system, consisting of an enzyme-immobilized microreactor and a microextractor, was used for production of an optically pure unnatural amino acid from a racemic mixture [9]. An acetyl-D,L-amino acid mixture was

167

168

9 Doing More by Combining – Process Integration

HCI aq.

Enzyme-immobilized microreactor

L-Phe OH

H3N

Micro-extractor

O

O

Racemic Ac-Phe (in PBS) Ethyl acetate

N H

Ac-D-Phe

O

O N H

OH O

OH O

N H

OH O

Organic phase O OH

N O

O OH

N

OH

H N 3

O

Aqueous phase (acidic)

O

Figure 9.3 Continuous flow system for the chiral separation of racemic amino acids. (Honda et al. [9]; reproduced with permission from the Royal Society of Chemistry).

hydrolyzed enantioselectively by the acylase-CEM to give the L-amino acid and the unhydrolyzed acetyl-D-amino acid. A high enantioselectivity (99.2–99.9% ee) for the L-isomer was achieved. By addition of HCl, the acetyl-D-amino acid is neutralized which increases its solubility in ethyl acetate, while the L-amino acid can be separated in the aqueous stream. Separation is done through a membrane on the microchannel surface which enabled the selective extraction of the product (Figure 9.3). 9.1.6 Integrated Membrane Microreactor for Knoevenagel Reaction

Knoevenagel reactions between benzaldehyde and ethyl cyanoacetate, ethyl acetoacetate, and diethyl malonate were carried out in a membrane microreactor. The water generated by the condensation was removed from the reaction mixture by pervaporation through a zeolite micromembrane. The best reaction yield was obtained by locating the CsNaX catalyst adjacent to the membrane. The microreactor performance with membrane was much better than without [10]. The Knoevenagel condensation of benzaldehyde and ethyl acetoacetate was carried out in a microreactor using a Cs-exchanged NaX catalyst and a NaA membrane [11]. For microreactor operation with powder catalyst, a follow-up product is found, which lowered the selectivity to 58%. When a thin CsNaX film was used instead, the selectivity increased to 78%. Lowering diffusion distances by increasing the microchannel aspect ratio from 2 to 5 increased the conversion from 25 to 60%. Through selective removal of water via a membrane, even higher conversions were achieved.

9.1

Integration of Reaction and Cooling/Heating, Separation, or Other

9.1.7 Continuous Multiple Liquid–Liquid Separation: Diazotization of Amino Acids

There is a need for liquid-based separation (not using solid scavenger techniques). With single-stage operation becoming more established and this, as for conventional single-stage separation, approaching its limits in performance, the development of multiple consecutive phase separations is demanded. This has been done and applied to the diazotization of amino acids in flow [12]. For this purpose, a second-generation laboratory-scale, modular liquid–liquid separation device based on computer-controlled high-pressure pumps and a highresolution digital camera was used. The diazotization of amino acids to produce valuable chiral hydroxyacids is demonstrated in flow for the first time. The use of a triple-separator system in conjunction with the developed diazotization process allows the safe and efficient production and automated isolation of multigram quantities of valuable chiral hydroxyacids. The reaction stream from the flow reactor into the extractor mixes the reaction stream with an extraction solvent. The biphasic stream then enters the top of a small (3 mm diameter) glass separating column through a stainless steel tube within a T-piece connector. A small colored polymer float remains at the interface between the organic and aqueous phases due to its intermediate density. The bottom of the phase-separating column is connected to a computer-controlled high-pressure pump. When the flow rate of the high-pressure pump is lower than the mixed incoming streams, an overflow of the upper layer occurs. A digital highresolution camera monitors the colored float in the column. It provides feedback to the high-pressure pump to adjust the flow rate. Flow diazotization and multiple extraction were done for various amino acids at a scale >1 g. Both preparation and extraction turned out to be efficient and remained stereoselective at the elevated temperature mostly used. In some cases, acetone was introduced to prevent crystallization of the less polar hydroxyacid products within the reactor. A 24+ h run demonstrated operational stability at 20+ g scales. No manual intervention was needed and yield did not change over that period of time. The enantiomer of hydroxyvaline synthesized from D-valine was actually used as the starting material for the total synthesis of (−)-enniatin B. 9.1.8 Coupling of the Hydroxylation of Progesterone Using Rhizopus Nigricans with Flow Extraction

The hydroxylation of progesterone in 11α position by Rhizopus nigricans was performed in a microchannel reactor with a Y-mixer and a multiple bended channel [13]. This is one of the key steps in the production of corticosteroid drugs and hormones. The reaction was coupled to a flow extraction of progesterone and 11α-hydroxyprogesterone from water into ethyl acetate [14]. This allows to increase productivity as compared to the batch process both because the reactant

169

170

9 Doing More by Combining – Process Integration

is readily recovered for reuse and the product is easily purified. The improved rheological properties of the medium result in lower operational costs. For increasing the capacity of the system, a numbering-up approach by using parallel microreactors is proposed. It is predicted that a continuous filtration unit is then needed upfront [14]. 9.1.9 Coupling of the Esterification to Isoamyl Acetate Using Lipase B with Flow Extraction

The esterification of isoamyl alcohol and acetic acid to give isoamyl acetate using Lipase B (from Candida Antarctica) was done in a microchannel reactor, for which all details were given above [15]. Isoamyl acetate is one of the most employed flavors in food industries. An increase in temperature from 25 to 45 ∘ C resulted in higher reaction yields. The residence time needed for the flow reaction was even below 1 min. Reaction rates were orders of magnitude higher than for batch esterification were obtained at similar reactant concentrations. A further benefit stems from the direct coupling of the flow reaction to flow extraction. The enzyme is dissolved in aqueous phase and continuously separated from the organic phase (n-hexane) and the product. Complex flow patterns arise due to the amphiphilic nature of the enzyme causing droplet formation aside the induced slugs of the segmented flow. The latter is then surrounded by an increasing corona of droplets (Figure 9.4).

(a)

220 μm

220 μm

(b)

220 μm

slugs. (Andrej Pohar, Igor Plazl and Polona Figure 9.4 (a) Long slug of n-heptane Žnidaršiˇc-Plazl [16]; with permission from with concave tail surrounded with small droplets of n-heptane and (b) fine dispersion Royal Society of Chemistry). between the two neighboring n-heptane

9.2

Integration of Process Control and Sensing

9.2 Integration of Process Control and Sensing

The integration of microreactors with sensors, actuators, and automated fluid handling is useful for gathering chemical information, but yet it is a field largely unexplored (see Outlook). In this way, a fast, continuous automation of reaction screening and optimization using online monitoring can be achieved [17]. 9.2.1 Integrated Process Control for Methanol Steam Reforming

A microreactor system was designed with integrated microfabricated sensors and actuators [18]. Main issue was effective control strategy, especially concerning system start-up and controllability for rapid load changes. This was accomplished via methanol steam reforming combined with proton exchange membrane fuel cells (PEMFCs). Each system component was interfaced with a comprehensive control algorithm based on Proportional-integral-derivative (control) (PID) control. Onchip sensing was made with a MEMS-based thin-film flow sensor to measure the rate of hydrogen production rate. Besides that, temperature sensing and actuation was integrated into the system. In this way, 100% conversion, rapid turn-on, and high responsiveness to varying hydrogen loads was achieved. 9.2.2 Integrated Sensing, Catalyst, and Heating for Ammonia Oxidation

The design of an integrated microreactor system was motivated by fabrication, assembly, and system integration concepts known in microelectronics and computer industry [19–21]. This integrated system incorporates a multilayer laminate gas-phase microreactor and is organized and assembled as computer chassis with modular boards to perform the required process functions (Figure 9.5). The reaction channels contain a 1 μm thin platinum film catalyst which is coated underneath a silicon nitride membrane. Seven platinum heaters and temperature sensors are placed on top of the silicon nitride membrane. Various pins are used for electrical contacting for means of on-board controlling and sensing. Aim was to develop a stand-alone gas-phase microreactor system whose functionality satisfied the objectives of the defense advanced research projects agency (DARPA) Microflumes program. Emphasis was placed on designing and constructing a MEMS-based system from first principles that had the same functionalities that are present in a more conventional laboratory-scale system, such as the DuPont MARS system. The components are designed and fabricated from first principles as prototype electromechanical boards that are configured to perform one or more process specific functions. To demonstrate feasibility, the oxidation of CH4 and NH3 are used as test reactions. Operation

171

172

9 Doing More by Combining – Process Integration

DieMateTM socket with microreactor

Redwood flow manifolds

Redwood flow manifold electronics

Figure 9.5 Reactor board organized and assembled as a computer unit. (Quiram et al. [19]; reproduced with permission from the American Chemical Society.).

of these reactions is also illustrated under commercial relevant conditions using integrated Microelectromechanical Systems (MEMS) components. 9.2.3 Integration of the Esterification to Ethyl Oleate Using Lipase B with Photoionization Mass Spectrometry

Lipid transformation was achieved by the enzyme lipase immobilized on a silica monolith providing a micro-flow network [22]. The monolith was placed into a 320 mm internal diameter fused silica capillary and such micro-flow reactor could be used up to flow rates of 60 ml h−1 (Figure 9.6). Candida antarctica lipase was covalently bound onto the silica monolith using glutaraldehyde as the crosslinking reagent. Triolein was quantitatively transformed into ethyl oleate at room temperature. No loss of activity was found for 15 runs. The micro-flow lipid transformations were coupled online with atmospheric pressure photoionization mass spectrometry. Feasibility could be demonstrated, yet a further reduction of dead times of the prototype system is needed to enable the potential of such innovative analytical monitoring.

9.2

Integration of Process Control and Sensing

173

HPLC, GC analyses

Reactant solutions in syring

Sample collecting (a) N2/ High temp

(b)

Flowrate: μL/min

Mass spec

APPI ion source

LC makeup flow

hν

Mass spectrometer system

Enzymatic microreactor

(c)

×1000

(d)

×1000

Figure 9.6 Enzymatic micro-flow reactor based on monoliths: (a) offline reaction; (b) online coupled with the mass spectrometry system; Scanning electron microscopy (SEM) image of a cross-section of: (c) a silica monolithic capillary (1000×); (d) an

Data output

(e)

×350

(f)

×450

enzymatic monolithic microreactor (1000×); (e) an NaOH-treated silica monolithic capillary (350×); and (f ) a non-treated silica monolithic capillary (450×). (Anuara et al. [22]; with permission from Elsevier).

9.2.4 Integration of the Intramolecular Friedel−Crafts Addition with Ultra-High-Pressure Liquid Chromatography

A microcapillary reactor is coupled to ultra-high-pressure liquid chromatography (UHPLC) for fast online analysis (Figure 9.7) [23]. Such a system can be used for high-throughput screening of homogeneous catalysts. The UHPLC detection at high temperature allows the online determination of substrate, product, and by-product concentrations even at high throughput which is essential for the catalyst discovery. In this way, a series of soluble acid catalysts were tested for an intramolecular Friedel–Crafts addition into an acyliminium ion intermediate (Scheme 9.1). Besides molecular diversity, the catalyst loading, reaction time, and reaction temperature were varied. To detect the side products, however, off-line mass spectrometric detection was needed. All these experiments were done within 1 day and with minimal material investment. High reproducibility in operation is achieved and indeed some promising acid catalysts candidates were identified out of the library tests. The catalyst Er(OTf )3 (trifluoromethanesulfonate or simply triflate) was the hit candidate found in the screening of the cyclization reaction and confirmed that excellence on a preparative scale.

174

9 Doing More by Combining – Process Integration

Autosampler (catalyst 1,2,3...)

Syringe pump 2 (carrier flow)

UHPLC column

UHPLC detector

Waste

Reactor Mixer L2

L1 Syringe pump 1 (reagents)

UV-vis detector

Waste

UHPLC pump

100 μm i.d. teflon tube Zone 4

(a)

Zone 3

Zone 2

Zone 1

0.90 0.80

Absorbance (A)

0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 60 (b)

80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 Time (min.)

Figure 9.7 (a) Schematic of the flow reactor-analysis system. L1 and L2 are loop injectors. (b) absorbance response of 54 reaction zones passing through the UV–vis

optical detector for a residence time of 1 h. (Fang et al. [22]; with permission from the American Chemical Society).

9.2.5 Integration of Pyrane Flow Reaction and Synchrotron-Based IR and X-Ray Beam Analysis

For a deep understanding of complex catalytic organic transformations, the detection of short-lived intermediates is crucial. Therefore, an in situ real-time analysis is needed which is able for (spectral) mapping of the whole flow microreaction

9.2

MeO

MeO

Integration of Process Control and Sensing

O

H N

Acid, CH3CN catalyst

NH

MeO

O

MeO

MeO

O MeO

N

H

MeO

- 63–93% yield - T = 0–40 °C - p = 1 bar - t = 0.5–1 h -Teflon tubular microreactor

Scheme 9.1 Intramolecular Friedel–Crafts cyclization [23].

area, that is, giving temporal–spatial information. Two such advanced techniques are micrometer-sized synchrotron-based IR and X-ray beam analysis [24]. These were applied to monitor pyrane formation as example for a multistep catalytic reaction (Scheme 9.2). O

2 nm Au clusters PhICl2 d-toluene

O

O O

nBu DO

nBu

nBu (1)

(2) - 71% yield (2) - T = RT - p = 1 bar - t = 20 s - Flow microreactor

Scheme 9.2 Pyrane formation in two steps [24].

Using a flow process and a flow reactor, a spatial resolution of 15 μm is hereby achieved. As the catalyst, 2-nm sized Au nanoclusters were supported on mesoporous SiO2 and packed in a flow microreactor. High conversion and tunable product selectivity is achieved. In situ synchrotron-sourced IR microspectroscopy detected the vinyl ether reactant conversion into the primary product allenic aldehyde. The latter is then catalytically transformed into the corresponding acetal as secondary product. A detailed analysis of the reaction kinetics was done by changing process parameters such as residence time. As second analysis tool, in situ micrometer X-ray absorption spectroscopy scanned along the flow reactor axis (Figure 9.8). Thereby, it could be shown that the local catalytic conversion, as detected by IR microspectroscopy, correlates to high concentration of Au(III), that is, which constitutes the catalytically active species.

175

176

9 Doing More by Combining – Process Integration

μm X-ray beam

μm IR beam

Reactants

Products ctor

Flow rea

O

H

Ph

O

n-Bu

n-Bu Reactants and products

DO

O

OR O

Ph

n-Bu

Ph

Catalyst properties

Figure 9.8 Gathering temporal–spatial information by micrometer-sized synchrotron-based IR and X-ray beam analysis to unreveal a multistep catalytic reaction done in flow – pyran formation. (Gross et al. [24]; with permission from the American Chemical Society).

9.2.6 Integration of Multistep Organic Transformations Catalyzed by Au Nanoclusters

Especially valuable in case of the flow synthesis of nanoparticles is the coupling of supercritical microreactor systems to conventional online characterization techniques (Figure 9.9). Suited analytical techniques are microparticle image velocimetry or dynamic light scattering (DLS), to measure flow profiles and

Confocal Raman spectroscopy

Dynamic light scattering

Optical fibers Optical access

Injection zone

Reaction zone

Back pressure regulator

HP syringe pumps

UV / Vis spectroscopy Sample recovery

Figure 9.9 Supercritical microfluidics setup: high-p,T microreactors (Si–Pyrex or borosilicate/borosilicate) combined with in situ and online characterization techniques. (Marre et al. [25a]; reproduced with permission from the American Chemical Society).

9.3

Thermal Integration on a Process Level

particles size distribution (PSD), respectively [25]. Such an experimental setup has been developed for the understanding of the chemistry and nucleation/growth in supercritical fluids and for the development of high-quality nanocrystals [26]. Such a tool is complementary to the time-resolved in situ synchrotron X-ray analysis. In this way, a new laboratory-scale tool is available for process optimization and fast screening of reaction conditions (about 100 wk−1 ). 9.3 Thermal Integration on a Process Level 9.3.1 Thermal Integration of a Methanol Micro-fuel Processor/Fuel Cell

Complete thermal integration of a methanol micro-fuel processor and fuel cell system was aimed to eliminate one of the most severe roadblocks toward reducing full-system size and enabling portability, which are crucial for acceptance of the new technology (Figure 9.10) [27]. Heat losses were experimentally determined for various pathways from the planar microreactor structure for catalytic methanol steam reforming. This allowed to set up an empirical correlation to predict the natural convection heat transfer coefficient of mesoscale to microscale devices. Fundamental heat management issues were addressed such as the transfer of heat between reactor components, control of temperature, insulation, and heat losses. By this, a design (packaging) and scale-up proposal could be given with reduced convective and radiative losses. Complete thermal integration was achieved for a Si-based catalytic MeOH steam reforming reactor [28]. Thermal characterization unrevealed the heat loss mechanisms and effective convective heat coefficients in the planar microreactor structure. This allowed proposing suited packaging schemes with reduced heat losses and to outline scale-up.

(a)

(b)

(c)

Figure 9.10 A steam reformer microreactor equipped with insulation chips: (a) front side; (b) backside; and (c) side view of the integrated microreformer. (Shah, Besser [28]; reproduced with permission from Elsevier).

177

178

9 Doing More by Combining – Process Integration

9.4 Integration of Units on Racks, Backbones, Frames, Interfaces, or Similar Level 9.4.1 Integrated Circuit Socket for Fluidic-Electric Interface

A commercial plastic socket, which is normally used for integrated circuit testing, was adapted for installation of the reactor chip to allow alignment with the electrical contacts (Figure 9.11) [19]. A heated fluidics interface connects the nonmetallic feed and product gas ports on the microreactor chip to metal tubing. This plastic reactor socket can be operated up to 250 ∘ C. The microreactor itself, however, can be heated much higher to about 600 ∘ C, via the heaters placed on the membrane. Step-response tests showed that temperature can be increased by 100 ∘ C in