Microporous Materials for Separation Membranes 3527343970, 978-3527343973

Citation preview

Microporous Materials for Separation Membranes

Microporous Materials for Separation Membranes Xiaoqin Zou Guangshan Zhu

Authors Xiaoqin Zou

Northeast Normal University Faculty of Chemistry No. 5268 Renmin Street 130024 Changchun China

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.

Guangshan Zhu

Northeast Normal University Faculty of Chemistry No. 5268 Renmin Street 130024 Changchun China

Library of Congress Card No.:

applied for British Library Cataloguing-in-Publication Data

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

The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at . © 2019 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: 978-3-527-34397-3 ePDF ISBN: 978-3-527-34406-2 ePub ISBN: 978-3-527-34409-3 oBook ISBN: 978-3-527-34399-7 Typesetting SPi Global, Chennai, India Printing and Binding

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

v

Contents Preface xi 1

Introduction of Microporous Membranes 1

1.1 1.2 1.3 1.3.1 1.3.1.1 1.3.1.2 1.3.1.3 1.3.2 1.3.3 1.3.4 1.3.5 1.3.5.1 1.3.5.2 1.4 1.4.1 1.4.2 1.4.2.1 1.4.2.2 1.4.2.3 1.5 1.5.1 1.5.2 1.5.2.1 1.5.2.2 1.5.2.3 1.5.2.4 1.5.2.5 1.5.2.6 1.5.2.7 1.5.3 1.5.4 1.5.4.1

Introduction 1 Historical Development of Membranes 1 Microporous Materials 4 Carbonaceous Materials 4 Activated Carbon 5 Carbon Nanotubes 8 Graphene 11 Microporous Silica 14 Zeolites 16 Metal–Organic Frameworks 19 Highly Porous Polymers 21 High Free Volume Polymers 21 Porous Organic Frameworks 23 Fundamentals of Membrane Separation 27 Membrane Definition 27 Transport Theory 27 Membrane Transport for Gas Systems 27 Membrane Transport for Liquid Systems 29 Transport Mechanism in ED Membrane 32 Membrane Configurations 35 Membrane Structures 35 Preparation Techniques 36 Solution Casting 36 Melt Extrusion 36 Spinning 37 Spin Coating 37 Slip Coating Sintering 37 Sol–Gel Technique 37 Carbonization 38 Membrane Technology 38 Membrane Modules 40 Plate-and-Frame Module 41

vi

Contents

1.5.4.2 1.5.4.3 1.5.4.4 1.6 1.7

Tubular Membrane Module 41 Spiral Wound Module 41 Hollow Fiber Module 42 Features of Microporous Membranes 43 Conclusions 45 References 45

2

Microporous Silica Membranes 53

2.1 2.2 2.2.1 2.2.2 2.2.3 2.3 2.4 2.5 2.6 2.7 2.8 2.9 2.10

Introduction 53 Membrane Synthesis 53 Sol–Gel Synthesis Method 54 Templating Approach 55 Chemical Vapor Deposition 56 Intermediate Layers 57 Support 58 Modification of Silica Membranes 58 Microporous Silica Membranes for Hydrogen Separation 60 Microporous Silica Membranes for Carbon Dioxide Separation 63 Microporous Silica-Based Membranes for Pervaporation Process 66 Microporous Silica-Based Membranes for Desalination 69 Conclusions and Future Trends 72 References 73

3

Carbon-Based Membranes 77

3.1 3.2 3.2.1 3.2.1.1 3.2.1.2 3.2.1.3 3.2.1.4 3.2.2 3.3 3.4 3.4.1

Introduction 77 Carbon Membrane Preparations 78 Carbon Molecular Sieve Membranes 78 Precursors of Carbon Molecular Sieves 79 Pyrolysis Environment for CMS Membranes 80 Pretreatments of CMS Membranes 81 Posttreatment of CMS Membranes 81 Module Construction of Carbon Membranes 82 Selective Surface Flow Membranes 84 Advantages and Disadvantages of Carbon Membranes 86 Advantages of Carbon Membrane Versus Conventional Polymer Membrane 86 Disadvantages of Carbon Membranes 87 Carbon Nanotubes 87 Types of CNT Membranes 87 CNT Functionalization 90 Porous Graphene 92 Carbon-Based Mixed Matrix Membranes (MMMs) 99 New Advances and Challenging Aspects 102 Concentration Polarization 103 Fouling 104 Mechanical Stability 105 Scalability 106

3.4.2 3.5 3.5.1 3.5.2 3.6 3.7 3.8 3.8.1 3.8.2 3.8.3 3.8.4

Contents

3.8.5 3.9

Cost 106 Conclusions 107 References 107

4

Microporous Membranes for Water Purification 115

4.1 4.2

Introduction 115 Types and State-of-the-Art Microporous Membranes in Water Purification 116 Removal of Water Contaminants (Inorganics, Organics, Biological) 118 Inorganic Pollutants 118 Organic Pollutants 124 Biological Pollutants 131 Membrane Desalination 134 Membrane Surface Engineering 144 Membrane Surface Engineering Using Nanoparticles 147 Using Hydrophilic Components to Hydrophobic Membranes 151 Conclusions 153 References 153

4.3 4.3.1 4.3.2 4.3.3 4.4 4.5 4.5.1 4.5.2 4.6

5

Mixed Matrix Membranes 161

5.1 5.2 5.2.1 5.2.2 5.2.3 5.2.4 5.3 5.3.1 5.3.2 5.3.3 5.3.4 5.4 5.4.1 5.4.2 5.4.3 5.4.4 5.4.5 5.5 5.6 5.7

Introduction 161 Principles of Mixed Matrix Membranes 161 Polymer Phase 162 Solvents 163 Fabrication and Drying Techniques for MMMs 163 Mass Transport Theory and Models in MMMs 166 MMMs Made of Zeolites 169 Interfacial Modification with Silane Agents 172 Addition of Low Molecular Weight Materials 172 Annealing 173 Priming Method 173 MOF-Based MMMs 173 UiO-66 Series 174 Zeolitic Imidazolate Frameworks 176 MIL Series 177 Cu-MOFs 178 MOF-74 Series 179 POF-Derived MMMs 181 MMMs Containing Other Porous Fillers 184 Conclusions 189 References 189

6

Zeolite Membranes 195

6.1 6.2 6.3

Introduction 195 Synthesis Techniques for Zeolite Membranes 196 Crystal Growth in Zeolite Layers 198

vii

viii

Contents

6.3.1 6.3.2 6.3.3 6.3.4 6.3.5 6.3.6 6.3.7 6.3.8 6.4 6.5 6.6

Conventional Hydrothermal Synthesis 198 Two-Step Crystallization 199 Crystallization by Microwave Heating 200 Use of Intergrowth Supporting Substances 201 Growth of Oriented Zeolite Layers on Supports 205 Bilayer Membranes 207 Functional Zeolite Films 208 Mixed Matrix Membranes 210 Microstructures of Zeolite Films 210 Membrane Characterizations 211 Conclusions and Outlook 215 References 216

7

Gas Separations with Zeolite Membranes 225

7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9

Introduction 225 H2 Recovery 226 Air Separation 228 CO2 Capture 230 N2 /CH4 Separation 235 H2 S Capture 238 Kr/Xe Separation 238 Post-modification of Zeolite Membranes 245 Conclusions and Outlook 247 References 248

8

Pervaporation with Zeolite Membranes 255

8.1 8.2 8.3 8.4 8.5 8.6 8.6.1 8.6.2 8.6.3 8.6.4 8.6.5 8.6.6 8.6.7 8.6.8 8.7

Introduction 255 Pervaporation Process 255 Alcohol Dehydration 260 Organic–Organic Separation 264 Acid Separation 266 Membrane Reactors for Various Separations 268 Water Separation 268 Hydrogen Separation in Dehydrogenation Reactions 271 Hydrogen Separation in Water–Gas Shift Reaction 273 Hydrogen Separation in Syngas Production 273 Methanol Separation in Hydrogenation Reaction 274 Metathesis of Propene 274 Separation of Isomers in Isomerization Reaction 275 Other Separations 277 Conclusions and Outlook 277 References 278

9

MOF Membranes and Their H2 Separation Properties

9.1 9.2 9.2.1 9.2.1.1

287 Introduction 287 Fabrication of MOF Membranes 288 Fabrication Methods for Polycrystalline Membranes 288 Direct Synthesis 288

Contents

9.2.1.2 9.2.1.3 9.2.1.4 9.2.1.5 9.2.1.6 9.2.2 9.2.2.1 9.2.2.2 9.3 9.3.1 9.3.2 9.3.3 9.3.4 9.3.5 9.3.6 9.3.7 9.3.8 9.4 9.5

Seeded Growth 292 Electrochemical Deposition 294 Stepwise Dosing of Reagents 296 Assembly of MOF Nanocrystals 296 Chemical Vapor Deposition 296 Fabrication Methods for SURMOF Membranes 296 Liquid-Phase Epitaxy 296 Interfacial Synthesis 298 Controlling and Characterizing MOF Membranes 300 Powder X-Ray Diffraction 302 Nitrogen Adsorption and Desorption 303 Thermal Gravimetric Analysis 304 Scanning Electron Microscopy 305 Inductively Coupled Plasma Optical Emission Spectroscopy 306 NMR Spectroscopy 306 SS-NMR Spectroscopy 307 Diffuse Reflectance Infrared Fourier Transform Spectroscopy 307 Pore Chemistry for H2 Separation 308 Conclusions 316 References 316

10

CO2 Capture with MOF Membranes 323

10.1 10.2 10.2.1 10.2.2 10.2.3 10.2.4 10.2.5 10.2.6 10.2.7 10.2.8 10.2.8.1 10.2.8.2 10.2.8.3 10.2.8.4 10.2.8.5 10.2.9 10.3 10.3.1 10.3.2 10.3.3 10.3.4 10.3.5 10.4 10.5

Introduction 323 CO2 Capture and Separation Strategies 324 Post-combustion CO2 Capture 325 Pre-combustion CO2 Capture 326 Oxy-fuel Combustion 326 Chemical Absorption 327 Physical Absorption 328 Physical Adsorption 330 Cryogenic 331 Membrane Technology 331 Polymeric Membranes 332 Inorganic Membrane 333 Zeolite Membrane 333 Silica Membrane 334 MOF Membrane 334 Chemical-Looping Combustion 334 Chemistry of MOFs for CO2 Recognition 335 Unsaturated Metal Sites 335 Polar Functional Groups 336 Pore Size and Function Control 339 Core–Shell MOF Structure 341 Alkylamine Incorporation 341 Membrane Design for CO2 Separation 344 Conclusions 353 References 354

ix

x

Contents

11

MOF Membranes for Vapor or Liquid Separation 361

11.1 11.2 11.2.1 11.2.2 11.3 11.4 11.5 11.6

Introduction 361 Selective Separation of Chemicals Via Pervaporation 364 Polycrystalline MOF Membranes for Pervaporation 365 MOF-Based MMMs for PV Process 373 Organic Solvent Nanofiltration 381 Chiral Resolution 394 Stability of MOF Membranes 401 Conclusions and Outlook 405 References 406

12

Microporous Organic Framework Materials for Membrane Separations 413

12.1 12.2 12.3 12.4 12.5 12.6 12.7 12.8

Introduction 413 Porous Structures and Free Volumes 413 Hydrogen Recovery 416 Carbon Dioxide Capture 419 Air Separation 423 Other Gas Separations 427 Emerging Liquid Separations 428 Conclusions 431 References 432 Index 437

xi

Preface Separation represents a very important industrial activity because a lot of labors and investments are devoted to separate the components of large quantities of mixtures into pure or purer forms. These processes involved account for 10–15% of the world’s energy consumption. Membrane technology is widely accepted as a more efficient and less energy-consuming separation and purification process compared with conventional ones (e.g. distillation, sublimation, crystallization, etc.). Its efficiency reflects not only on the large quantity processed per day but also on its unique powerfulness in some specific systems such as azeotropic liquids or solutes. Membrane-based separation would use 90% less energy than distillation. Thus, membrane separation has attracted much attention from both academic and engineering fields, and this interest will definitely keep on increasing. Membrane separation works on selective transport of a particular species through a semipermeable membrane layer under a certain driving force. Pressure driven is the simplest in terms of the ability to separate species in gas or liquid feed streams with high flux. According to the membrane definition, materials play a central role because they govern the mass transport behaviors; thus consequently the selectivity and the permeability are two main parameters for membrane separation. Although some other books are available for membrane separation, this book is solely focused on the development of advanced microporous materials for membrane application from materials perspective. Microporous materials possess one unique and important feature of pores with diameters at the molecular level ( 10, which results in the lighter molecules permeating through the pores. In this case, selectivity is limited and can be calculated with the square root of the ratio of the molar masses of the gases involved. The smaller the Knudsen number, the larger the pores become (relative to the mean free path of the gas molecules), and the lower selectivity becomes. For K n < 1.0, the dominant transport mechanism is viscous flow, which is nonselective. Surface diffusion may occur in parallel with Knudsen diffusion. Gas molecules are adsorbed on the pore walls of the membrane and migrate along the surface. Surface diffusion increases the permeability of the components adsorbing strongly to the membrane pores. At the same time, the effective pore diameter is

1.4 Fundamentals of Membrane Separation

reduced. Consequently, transport of non-adsorbing components is reduced and selectivity is increased. This positive contribution of surface diffusion works for certain temperature ranges and pore diameters. If pore sizes become sufficiently small (3–5 Å), the mechanism of molecular sieving is applicable in separating molecules that differ in kinetic diameter. The pore size becomes so small that only the smaller gas molecules can permeate through the membrane. 1.4.2.2

Membrane Transport for Liquid Systems

In membrane filtration, the separation process is accomplished by using a differential driving potential across a membrane having selective permeability. For example, the differential driving potential used to transport solvent across UF membrane is the hydrostatic pressure. UF is commonly used to separate suspended solids, colloids, and macromolecules from water. Whenever the solvent of a mixture flows through the membrane, retained species are locally concentrated at the membrane surface, thereby resisting the flow. This localized concentration of solute normally results in precipitation of a solute gel over the membrane. Hence, UF throughput depends on physical properties of the membrane, such as permeability, thickness, and process/system variables such as feed concentration, system pressure, velocity, and temperature. Two models of gel polarization and resistance with different approaches are described below. Gel Polarization Model The basic assumptions of this model are as follows:

1. UF membranes have skin that offers minimum resistance to flow, and the asymmetry of the pore virtually eliminates internal pore fouling. 2. Concentration buildup at the membrane surface rises up to the point of incipient gel precipitation, forming a dynamic secondary membrane on top of the primary structure. 3. The secondary membrane offers the major resistance to flow. 4. The gel layer grows in thickness until the pressure-activated convective transport of solute with solvent toward the membrane surface just equals the concentration gradient-activated diffusive transport away from the membrane surface. 5. Beyond a certain threshold pressure, increase in pressure does not improve the flux because the gel layer grows thicker to offer more resistance to the increased driving force. This is called critical flux: Jw =

P Rc + Rm

where J w is the water flux P is the transmembrane pressure Rc is the resistance of the deposited cake Rm is the hydraulic resistance of the membrane 6. Eventually, the concentration at the membrane surface will be high enough to form a gel.

29

30

1 Introduction of Microporous Membranes

7. In the steady state, the convective transport to the membrane must equal the back-diffusive transport away from the membrane: ( ) dc J = −D dx where J is the solvent flux through the membrane C is the concentration of solutes or colloids retained in membrane D is the solute diffusivity x is the distance from the membrane surface By integration, it gives ( ) ( ) ( ) Cg Cg D J= ⋅ ln = k ⋅ ln 𝛿 Cb Cb where k is the mass transfer coefficient 𝛿 is the boundary layer thickness considering D as constant C g and C b represent the maximum solute concentration in the gel layer and the concentration of solutes in the bulk of the feed, respectively 8. Lower solute concentration (C b ) will have higher threshold pressure because much higher flux is required to transport enough solute to the membrane to begin to form a gel. Resistance Model The mechanism of separation in UF involves not only the size

exclusion but also the adsorption and surface charge characteristics of membranes. In the absence of a solute, the water flux through a microporous membrane is defined by Darcy’s law, which states that pure solvent flux is directly proportional to the applied pressure differential (ΔP) and inversely proportional to pure solvent viscosity (𝜇w ): Jw =

ΔP Rm 𝜇w

where Rm is the membrane hydraulic resistance, which is a function of pore size, tortuosity, membrane thickness, and porosity. If the feed solution contains solutes that are retained at the membrane interface, the water flux in UF is generally lower than pure water flux. A number of phenomena have been suggested to account for this flux reduction, such as resistance due to gel layer formation, resistance due to concentration polarization, and resistance due to an absorption layer and pore plugging. For a macromolecular solute of high molecular weight at low concentration, the osmotic pressure effect can be neglected. The effect of the gel layer can be represented as Juf =

ΔPappl 𝜇w (Rm + Rp )

where Rp is the resistance due to gel polarization.

1.4 Fundamentals of Membrane Separation

The time-dependent case of above equation can be represented as Juf (t) =

ΔPappl 𝜇w [Rm + Rp (t)]

After testing, if the membrane is thoroughly washed with appropriate washing solution and the pure water flux (J w ) is determined at the same ΔP, it may be found to be less than J w but still greater than J uf . The difference between J w and J uf may be accounted for by the irreversible fouling due to adsorption of solute on the membrane, and this loss in flux can be visualized as additional resistance to the flux (Ra ): Jw =

ΔPappl 𝜇w (Rm + Ra )

Hence, incorporating Ra , above equation can be written as Juf (t) =

ΔPappl 𝜇w [Rm + Ra + Rp (t)]

It is noted that J uf reaches an almost constant final flux J uf (F) and the time corresponding to this J uf (F) is t(F). At this stage, Rp (t) becomes constant Rp (F): Juf (F) =

ΔPappl 𝜇w [Rm + Ra + Rp (F)]

There also exists a concentration polarization resulting from the relative rate of solute transport to the membrane surface by convection and the back-diffusive solute flux. Although both concentration polarization and fouling reduce the membrane flux, they have opposing effects on the observed percent rejection. Another way to distinguish the two phenomena is through their time dependence. Concentration polarization is dependent on operating parameters such as pressure, temperature, feed concentration, and velocity but is not a function of time. Fouling is partially dependent on these variables, particularly feed concentration, but is also a function of time. The change of flux with time due to different kinds of resistances is given in Figure 1.10 for a typical UF membrane [155]. It shows asymptotic behavior after a particular duration of time. The mass transfer coefficient, k, can be calculated from the following equation: [

k= 3600MB S

1+m(1−f )MA 1000

PR ] )] [ ( X −X c(1 − X3 ) ln X2 −X3 1

3

where PR refers to the product rate f refers to the solute separation with reference to the chosen reference solute MA and MB refer to the molecular weights of solute and water, respectively S refers to the membrane area m is the solute molarity c is the molar density of feed

31

32

1 Introduction of Microporous Membranes

Jw Drop due to adsorption (Ra) Jw′ Drop due to ΔІІ

Drop due to RF

Juf

Juf Juf (F) t (F)

Time

Figure 1.10 Change of flux with time for UF membrane showing asymptotic behavior after a particular duration of time.

X 1 , X 2 , and X 3 refer to the mole fractions of solute at bulk, membrane–solution interface, and membrane-permeated product, respectively 1.4.2.3

Transport Mechanism in ED Membrane

The transport mechanism in the ED membrane is shown in Figure 1.11, showing concentration and potential gradients in a well-stirred ED cell. In this example, chloride ions easily permeate the anionic membranes containing fixed positive groups and are stopped by the cationic membranes containing fixed negative groups. Similarly, sodium ions permeate the cationic membranes but are stopped by the anionic membranes. The overall result is increased salt concentration in A

C Cl–

Anode (+)

A Cl–

C Cl–

Cl–

Na+

Na+

Na+

Na+

Conc.

Dilute

Conc.

Dilute

Cathode (–)

Electrical potential Concentration

Figure 1.11 Schematic of concentration and potential gradients in a well-stirred ED cell.

1.4 Fundamentals of Membrane Separation

alternating compartments, whereas the other compartments are simultaneously depleted of salt. The voltage potential drop caused by the electrical resistance takes place entirely across the ion-exchange membrane. In a well-stirred cell, the flux of ions across the membranes and hence the productivity of the ED system can be increased without limit by increasing the current across the stack. In practice, however, the resistance of the membrane is often small in proportion to the resistance of the water-filled compartments, particularly in the dilute compartment where the concentration of ions carrying the current is low. In this compartment, the formation of ion-depleted regions next to the membrane places an additional limit on the current and hence the flux of ions through the membranes. Ion transport through this ion-depleted aqueous boundary layer generally controls ED system performance. The formation of concentration gradients caused by the flow of ions through a cationic membrane is shown in Figure 1.12. It shows the concentration gradient of univalent sodium ion next to a cationic membrane. Exactly equivalent gradient of anion, such as chloride ion, forms adjacent to the anionic membranes in the stack. The ion gradient formed on the left dilute side of the membrane can be described by Fick’s law. Thus, the rate of diffusion of cations to the surface is given by J+ =

D+ (c+ − c+(0) ) 𝛿

where D+ is the diffusion coefficient of the cation in water c+ is the bulk concentration of the cation in the solution c+(0) is the concentration of the cation in the solution adjacent to the membrane surface (0) 𝛿 is film thickness Figure 1.12 Schematic of concentration gradients adjacent to a single cationic membrane in an ED stack.

Cation membrane Na+ Diluate chamber

Concentrate chamber

l δ

c+

l

c+

c(o) o

l

+

c(l)

33

34

1 Introduction of Microporous Membranes

The rate at which the cations approach the membrane by electrolyte transport is (t + I/F). The total flux of sodium ions to the membrane surface (J + ) is the sum of these two terms: D+ (c+ − c+(0) ) ( t + I ) + J+ = 𝛿 F Transport through the membrane is also the sum of two terms (i) due to the voltage difference and (ii) due to the diffusion caused by the difference in ion concentrations on each side of the membrane. Thus, the ion flux through the membrane can be written as ( + ) P+ (c+(0) − c+(1) ) t(m) I + + J = F l where P+ is the permeability of the sodium ions in a membrane of thickness l. The quantity P+ (c+(0) − c+(1) )∕l is much smaller than transport due to the voltage gradient, so above two equations can be combined and simplified to ( + ) D+ (c+ − c+ ) ( t + I ) (0) (m) t I = + F 𝛿 F + = 1, it can be furFor a selective cationic ion-exchange membrane for which t(m) ther simplified to ( ) (c+ − c+ ) (0) FD+ I= 1 − t+ 𝛿

This important equation has a limiting value when the concentration of the ion at the membrane surface is zero (c+(0) = 0). At this point, the current reaches its maximum value. The limiting current is given by the equation Ilim =

FD+ c+ 𝛿(1 − t + )

The limiting current, I lim , is the maximum current that can be employed in an ED process. If the potential required to produce this current is exceeded, the extra current will be carried by other processes, first by transport of anions through the cationic membrane and, at higher potentials, by hydrogen and hydroxyl ions formed by dissociation of water. Both of these undesirable processes consume power without producing any separation. This decreases the current efficiency of the process. The limiting current density for an ED system operated at the same feed solution flow rate is a function of the feed solution salt concentration. As the salt concentration in the solution increases, more ions are available to transport current in the boundary layer, so the limiting current density also increases. For this reason, large ED systems with several ED stacks in series will operate with different current densities in each stack, reflecting the change in the feedwater concentration as salt is removed.

1.5 Membrane Configurations

1.5 Membrane Configurations As discussed above, membrane materials with the appropriate chemical, mechanical, and permeation properties are crucial for high-performance membranes. In addition, the technology contributes greatly in successfully fabricating this material into a robust, thin, defect-free membrane and then to package the membrane into an efficient, economical, high-surface-area module. This chapter covers the membrane structures, preparation techniques, technology, and modules. 1.5.1

Membrane Structures

As shown in Figure 1.13, membrane structures are organized in a sequence of microporous membranes, homogeneous membranes, asymmetric membranes, electrically charged membranes, and liquid membranes. 1. Microporous membranes: The membrane behaves almost like a fiber filter and separates by a sieving mechanism determined by the pore diameter and particle size. Materials such as inorganics or organics are used in making such membranes. The pores in the membrane may vary between 0.3 nm and 100 μm. 2. Homogeneous membranes: This is a dense film through which a mixture of molecules is transported by pressure, concentration, or electrical potential gradient. Using these membranes, chemical species of similar size and diffusivity can be separated efficiently when their concentrations differ significantly. 3. Asymmetric membranes: An asymmetric membrane comprises a very thin (0.1–1.0 μm) skin layer on a highly porous (50–500 μm) thick substructure. Symmetrical membranes Isotropic microporous membrane

Nonporous dense membrane

Anisotropic membranes

Electrically charged membrane

Supported liquid membrane

Liquid-filled pores Loeb–Sourirajan anisotropic membrane

Thin-film composite anisotropic membrane

Polymer matrix

Figure 1.13 Schematic diagrams of the commonly used membranes.

35

36

1 Introduction of Microporous Membranes

The thin skin acts as the selective membrane. Its separation characteristics are determined by the nature of membrane material or pores size, and the mass transport rate is determined mainly by the skin thickness. Porous sublayer acts as a support for the thin, fragile skin and has little effect on the separation characteristics. 4. Electrically charged membranes: These are necessarily ion-exchange membranes consisting of highly swollen gels carrying fixed positive or negative charges. These are mainly used in the ED. 5. Liquid membranes: A liquid membrane utilizes a carrier to selectively transport components such as metal ions at relatively high rate across the membrane interface.

1.5.2

Preparation Techniques

Large-scale and industrial processes for preparing membranes can be mainly catalogued as (i) solution casting, (ii) melt extrusion, (iii) spinning, (iv) spin coating, (v) slip coating sintering, (vi) sol–gel technique, and (vii) carbonization. 1.5.2.1

Solution Casting

Solution casting is often used to prepare small samples of membrane for laboratory characterization experiments. An even film of an appropriate solution suspension is spread across a flat plate with a casting knife. The casting knife consists of a steel blade, resting on two runners, arranged to form a precise gap between the blade and the plate onto which the film is cast. After casting, the solution is left to stand, and the solvent evaporates to leave a thin and uniform film [156]. Appropriate solution viscosity and solvent type are required for preparing high-quality membranes or films. The solution used for solution casting should be sufficiently viscous to prevent it from running over the casting plate, so typical concentrations are in the range of 15–20 wt%. Preferred solvents are moderately volatile liquids such as acetone, ethyl acetate, and cyclohexane. Films cast from these solutions are dry within a few hours. When the solvent has completely evaporated, the dry film can be lifted from the glass plate. If the film adheres to the plate, soaking in a swelling non-solvent such as water or alcohol will usually loosen the film. 1.5.2.2

Melt Extrusion

Many polymers or inorganic polymer hybrids, including polyethylene, polypropylene, nylons, and silica gel, do not dissolve in appropriate solvents at room temperature, so membranes cannot be made by solution casting. The polymer is compressed between two heated plates. Typically, a pressure of 2000–5000 psi is applied for one to five minutes, at a plate temperature just below the melting point of the polymer or the matrix in the hybrid. Melt extrusion is also used on a very large scale to make dense films for packaging applications, either by extrusion as a sheet from a die or as blown film [157].

1.5 Membrane Configurations

1.5.2.3

Spinning

Spinning is a manufacturing process for creating fiber membranes [158]. It is a specialized form of extrusion that uses a spinneret to form multiple continuous filaments. There are many types of spinning: wet, dry, dry jet-wet, melt, gel, and electrospinning. First, the precursor usually a polymer being spun must be converted into a fluid state. The fluid can be achieved from melting or dissolution in a solvent. For melted fluid, it is forced through the spinneret, and then it cools to a rubbery state and then a solidified state. For hot solution, the solvent is removed after being forced through the spinneret. 1.5.2.4

Spin Coating

Spin coating is widely used in the electronics industry to coat photoresists and photolithographic films onto silicon wafers [159]. The technique is also used in the laboratory to make composite membranes of 0.5–10 μm thick. An excess of dilute solution is placed on the substrate, which is then rotated at high speed. Fluid spins off the edge of the rotating substrate until the desired film thickness is achieved. The coating layer thickness can be decreased by increasing the rotation speed or decreasing the concentration in the applied solution. 1.5.2.5

Slip Coating Sintering

Slip-coating-sintering procedure was widely adopted for producing inorganic membranes, particularly for ceramic membranes with pore diameters in MF and UF range from 0.01 to 10 μm. In the slip-coating-sintering process, a porous ceramic support tube is made by pouring a dispersion of a fine-grain ceramic material and a binder into a mold and sintering at high temperature. The pores between the particles that make up this support tube are large. One surface of the tube is then coated with a suspension of finer particles in a solution of a cellulosic polymer or poly(vinyl alcohol), which acts as a binder and viscosity enhancer to hold the particles in suspension. This mixture is called a slip suspension; when dried and sintered at high temperatures, a finely microporous surface layer remains. Usually several slip-coated layers are applied in series, each layer being formed from a suspension of progressively finer particles and resulting in an anisotropic structure. Most commercial ceramic UF membranes are made this way, generally in the form of tubes or perforated blocks [160]. 1.5.2.6

Sol–Gel Technique

Sol–gel method is used to produce membranes with pores from 10 to 100 Å. In this process, slip coating is taken to the colloidal level. Generally, the substrate to be coated with the sol–gel is a microporous ceramic tube formed by the slip-coating-sintering technique. This support is then solution coated with a colloidal or polymeric gel of an inorganic hydroxide. These solutions are prepared by controlled hydrolysis of metal salts or metal alkoxides to hydroxides. Sol–gel methods fall into two categories, depending on how the colloidal coating solution is formed [161]. In the particulate-sol method, a metal alkoxide dissolved in alcohol is hydrolyzed by addition of excess water or acid. The precipitate that results is maintained as a hot solution for an extended period during which the precipitate forms a stable colloidal solution. This process is called peptization. The

37

38

1 Introduction of Microporous Membranes

colloidal solution is then cooled and coated onto the microporous support membrane. The layer formed must be dried carefully to avoid cracking the coating. In the final step the film is sintered at 500–800 ∘ C. In the polymeric sol–gel process, partial hydrolysis of a metal alkoxide dissolved in alcohol is accomplished by adding the minimum of water to the solution. The active hydroxyl groups on the alkoxides then react to form an inorganic polymer molecule that can then be coated onto the ceramic support. On drying and sintering, the inorganic film forms. Depending on the starting material and the coating procedure, a wide range of membranes can be made by the sol–gel process. 1.5.2.7

Carbonization

Carbonization is a particular method for producing carbon-based membranes. A wide variety of precursor polymer membranes can be used. Polyacrylonitrile, poly(vinylidene chloride), poly(furfuryl alcohol), and PIs easily carbonize and have been widely used. As the precursor membrane is heated, there is a gradual loss of weight. The amount and composition of the material lost depends on the polymer. Most polymers lose 10–20 wt% of their weights when the polymer has been heated to 300–500 ∘ C. At this point, the polymer starts the pyrolysis by releasing backbone hydrogens and becomes yellow to brown. Heating at higher temperatures produces more weight loss, and most polymers lose their heteroatoms by the time the polymer reaches 800–1000 ∘ C. During this carbonization process, the membrane usually becomes compact and pores are generated in most cases [162]. 1.5.3

Membrane Technology

An overview of membrane technologies is given in Table 1.2. MF, UF, nanofiltration (NF), RO, and ED are well-established technologies. The pressure-driven membrane separation processes of RO, NF, UF, and MF are illustrated in Figure 1.14. The relative size of different solutes removed by each class of membrane is given. RO, NF, UF, MF, and conventional filtration are similar processes differing mainly in the average pore diameter of the membrane. The Table 1.2 Industrial membrane technologies. No.

Category

Process

1

Membrane separation technologies well established in the industries

MF, UF, RO, ED

2

Upcoming membrane separation technologies for the industries

Gas separation, pervaporation

3

Membrane separation technologies of interest for the industries

Carrier-facilitated transport membrane contactors, piezodialysis, etc.

1.5 Membrane Configurations

Ultrafiltration

Microfiltration

Conventional filtration

RO

1Å

10 Å

100 Å

1000 Å

1 μm

10 μm

100 μm

Pore diameter

Figure 1.14 Pressure-driven membrane separation processes.

mode of separation in case of NF, UF, and MF is molecular sieving through increasingly fine pores. MF membranes filter colloidal particles and bacteria from 0.1 to 10 μm in diameter. UF membranes can be used to filter dissolved macromolecules, such as proteins from solutions. NF membranes have pore sizes from 1.0 to 10 nm, which is used most often with low total dissolved solids water such as surface water and fresh groundwater, with the purpose of softening and removal of disinfection by-product precursors. MF, UF, and NF membranes are supposed to contain a series of cylindrical capillary pores of specific diameters. RO membranes are so dense that discrete pores do not exist. Transport occurs through the statistically distributed free volume areas. The pores of the membrane range from 3.0 to 5.0 Å in diameter, which is within molecular level. The mechanism of transport through the RO membrane is governed by the solution–diffusion model. According to this model, solutes permeate the membrane by dissolving in the membrane material and diffusing down a concentration gradient. Separation occurs because of the difference in solubilities and mobilities of different solutes in the membrane. Table 1.2 shows gas separation and PV as upcoming membrane technologies. Gas separation with membranes has higher potential of application. Several companies worldwide use membrane-based gas separation systems for a variety of applications. In gas separation, a gas mixture at an elevated pressure is transported across the surface of a membrane that is selectively permeable to one component of the feed mixture. Major current applications of gas separation membranes are the separation of hydrogen from nitrogen, argon, and methane in ammonia plants, the production of nitrogen from air, and the separation of carbon dioxide from methane in natural gas operations. In PV, a liquid mixture is fed in the membrane system, and permeate in the form of vapor is removed. The driving force for the process is pressure drop across the membrane. The separation obtained is proportional to the rate of permeation of the particular component through the selective membrane. PV offers the possibility of separating closely boiling mixtures or azeotropes that are difficult to separate by distillation or other means. Currently, the main industrial application of PV technology is for the dehydration of organic solvents, such as the dehydration of 90–95% ethanol

39

40

1 Introduction of Microporous Membranes

solutions, which is a difficult separation problem because of the ethanol/water azeotrope at 95% ethanol. PV membranes that selectively permeate water can produce more than 99.9% ethanol from these solutions. PV processes are also being developed for the removal of dissolved organics from water and for the separation of organic mixtures. A number of other industrial membrane processes such as carrier-facilitated transport are under development, which often employs liquid membranes containing a complexing or carrier agent. The carrier agent reacts with one component of a mixture on the feed side of the membrane and then diffuses across the membrane to release permeate on the product side. The reformed carrier agent then diffuses back to the feed side of the membrane. Thus, the carrier agent acts as a shuttle to selectively transport one component from the feed to the product side of the membrane. Facilitated transport membranes can be used to separate gases. In this case, membrane transport is driven by a difference in the partial pressure across the membrane. Metal ions can also be selectively transported across a membrane, driven by a flow of H+ or OH− in the other direction. This process is sometimes called coupled transport. As the carrier-facilitated transport process employs a reactive carrier species, very high membrane selectivity can be achieved. The selectivity is often far larger than the one achieved by other membrane processes. This has maintained interest in facilitated transport for so many years. However, commercial deployment is not yet to be deployed due to challenges faced with respect to (i) the physical instability of the liquid membrane and (ii) the chemical instability of the carrier agent. In recent years, a number of potential solutions to this problem have been developed, which may make carrier-facilitated transport a viable process. The membrane separation processes described earlier represent the bulk of the industrial membrane separation industry. Another process, dialysis, is used on a large scale in the field of medical application to remove toxic metabolites from the blood in patients suffering from kidney failure. The first successful artificial kidney was based on cellophane (regenerated cellulose) dialysis membranes and was developed in 1945. Over the past 50 years, much advancement has been made. Currently, most artificial kidneys are based on hollow fiber membranes formed into modules having a membrane area of about a square meter (1.0 m2 ) to remove urea and other toxic elements. Following the success of the artificial kidney, similar devices were developed to remove carbon dioxide and deliver oxygen to the blood. These so-called artificial lungs are used in surgical procedures during which the patient’s lungs cannot function. Another major medical use of membranes is in controlled drug delivery. Controlled drug delivery can be achieved by a wide range of techniques, most of which involve membranes. 1.5.4

Membrane Modules

Membrane plants often require large membrane surface areas to perform the separation required on an industrial scale. Prior to separation, modules are needed for economically and efficiently packaging large areas of membranes. The membranes are cast as flat sheets, tubes, and fine hollow fibers. For accommodating such shapes and structures, different types of membrane modules

1.5 Membrane Configurations Metal spacer

Pressure shell

Wire screen Product manifold

Paper

Permeable membrane

Product

Heat seal

Permeate outlet Product

Feed gas passed through shell Septa stacked and spaced by internal metal strip Stack joined by staybolts

Figure 1.15 An early plate-and-frame design for the separation of helium from natural gas. Source: Reproduced with permission from Reference [163], Copyright John Wiley & Sons.

including plate-and-frame, tubular, spiral wound, and hollow fiber modules have been widely developed for industrial purposes [163]. 1.5.4.1

Plate-and-Frame Module

Plate-and-frame modules (Figure 1.15) are one of the earliest types of membrane configuration, in which membrane, feed spacers, and product spacers are layered between two end plates [163, 164]. The feed is sent across the surface of the membrane. A portion of it passes through the membrane, enters the permeate channel, and makes its way to a central permeate collection manifold. Plate-and-frame modules are used in ED and PV systems. A modified version of plate-and-frame module known as disk and tube module configuration has become a popular approach for treating wastewater with highly fouling feed streams. 1.5.4.2

Tubular Membrane Module

Tubular modules are generally applied in MF and UF. The biggest benefit of this module is the high resistance to membrane fouling due to good fluid hydrodynamics. However, the high cost is the main drawback. Typically, the tubes consist of a porous paper or fiberglass support with the membrane on the inside of the tubes. In a typical tubular membrane system, a large number of tubes are arranged in series. The permeate stream from each tube is collected in the permeate collection header. A tubular system is shown in Figure 1.16 [163]. 1.5.4.3

Spiral Wound Module

In the spiral wound module, the support fabric, feed spacer, and permeate carrier encase the membrane, providing structural integrity, as shown in Figure 1.17 [163]. Feed solution passes across the membrane surface. A portion passes through the membrane and enters the membrane envelope where it spirals inward to the central perforated collection pipe. The feed enters the module. The

41

42

1 Introduction of Microporous Membranes

Tube sheet

Manifold outlet U-bend Outlet Support tube

Permeate manifold

Inlet

Figure 1.16 A tubular UF membrane system in which tubes are connected in series. Source: Reproduced with permission from Reference [163], Copyright John Wiley & Sons. Perforated permeate collection pipe Feed spacer Membrane

Membrane envelope

Figure 1.17 An unfolded view of a spiral wound module. Source: Reproduced with permission from Reference [163], Copyright John Wiley & Sons.

permeate stream and concentrate (reject) stream come out of the module. Spiral wound modules are commonly used by desalination industries for brackish and seawater desalination. 1.5.4.4

Hollow Fiber Module

Hollow fiber membrane modules can be classified into two categories based on feed arrangement. The first is the shell-side feed design illustrated in Figure 1.18a [163]. In such a module, a loop or a closed bundle of fibers is contained in a pressure vessel. The system is pressurized from the shell side and the permeate passes

1.6 Features of Microporous Membranes

Residue

Permeate

Hollow fibers

Feed

Residue

Feed

(a)

Permeate

(b)

Hollow fibers

Figure 1.18 Two types of hollow fiber modules: (a) shell-side feed and (b) bore-side feed. Source: Reproduced with permission from Reference [163], Copyright John Wiley & Sons.

through the hollow fiber. Because the fiber wall must support considerable hydrostatic pressure, the fibers usually have small diameters and thick walls, typically 50 and 100–200 μm for respective inner and outer diameters. The second type of hollow fiber module is the bore-side feed type illustrated in Figure 1.18b. The fibers in this type of unit are open at both ends, and the feed fluid is circulated through the bore of the fibers. To minimize the pressure drop inside the fibers, the diameters are usually larger than those of the fine fibers used in the shell-side feed system. The hollow fibers are generally made by solution spinning. The modules are popular for UF and PV operations. They are used for low- to medium-pressure gas applications. Feed pressures are usually limited to below 10 bars in this type of module. Capillary fiber, which is a modified version of hollow fiber, appears promising for several applications where concentration polarization and fouling are faced in hollow fiber modules.

1.6 Features of Microporous Membranes In contrast to dense membranes, microporous membranes endow the porous nature. That means membrane materials possess large free volumes and open pores. Figure 1.19 illustrates the most important parameters of microporous materials for membrane application. The features of microporous membrane are elucidated from the perspective of pore chemistry because this chemistry plays a central role in membrane separation. The pore properties can be determined using the terms of pore size, configuration, dimensionality, and functionality. Pore configuration describes the shapes and connectivity modes of pores. Conforming to the IUPAC recommendation [165], pores that have continuous connection pathways with the

43

44

1 Introduction of Microporous Membranes

Figure 1.19 Four important parameters – pore size, interconnected pores, pore dimensionality, and surface functionality – of microporous materials for membrane application.

d1 d2 Pore size

Pore dimensionality

Interconnected pores

Surface functionality

outer surfaces of the porous structure are called open pores. On the other hand, pores that are detached from other pores are called closed pores. The open pores play a pivotal role in fluid dynamics and gas adsorption. From application-based perspectives, materials chemists are interested in open pores. The existence of open pores is prerequisite for separations because interconnected pores form a free pathway for gases or liquids to pass through. The advantage of highly connected pore system is the significant contribution to high permeability or flux. Many microporous materials exhibit interconnected pore configuration because of their uniform pore structures or crystalline phases. For example, zeolites discussed in this book possess well-ordered pores, and the pores are cylinder straight or crossing connected. The occurrence of open pores in zeolites is ascribed to their crystalline phases with all atoms placed regularly in the lattice. For microporous membranes, pore sizes are usually in the range of 3–20 Å. Pore size poses different constraints on molecular diffusions through the pores. According to Lennard-Jones plot [166], a huge potential is measured in micropores. As consequence, smaller molecules diffuse faster than larger ones since there is less hindrance for small molecules. Thus, this effect can be utilized to sieve a particular molecule of interest with good selectivity. After a survey of microporous materials, it can be found that their pores are in the range of 3–20 Å, for instance, zeolites have pores in sizes of 3–14 Å, MOFs own pores in 3–20 Å, CNTs have defined channels of 5–10 Å, etc. Pore dimensionality sometimes also influences the behavior of mass transport. For instance, in a porous membrane with 3D pore systems, a molecule is able to diffuse into the pores easily and subsequently to pass through the membrane in each direction. In contrast, elaborative control on the pore orientation should be made in order to render molecules pass along the 1D channel. The effect of pore dimensionality was in detail investigated by random- and b-oriented MFI zeolite membranes [167]. Most of pores are blocked or unused for mass transport in random-oriented membranes. However, great enhancement in xylene flux was achieved by b-oriented MFI zeolite membranes. The functionality on the

References

pore surface is another important parameter that determines the membrane separation performance. The functional groups on the bodies can impose either attractive or repulsive forces toward a particular molecule in the mixture. For example, the adsorption of CO2 molecules is favored on pore surfaces with high polarities because of the strong van der Waals interactions between CO2 molecules and polar groups. The result of preferred adsorption is an increase in the adsorption selectivity and consequently an enhancement to the overall separation selectivity. The strong interaction also facilitates the close packing of CO2 molecules along the pore wall, which eventually increases CO2 permeability in the process of membrane separation. The influence of functionality in separation performance was symmetrically studied by MOF membranes in gas separations [122]. A variety of functionalities such as unsaturated metal centers in clusters, amino groups on ligands, and polar OH entities were introduced in order to improve CO2 selectivity in MOF material-based membranes. From the above basics, it can be concluded that pore chemistry is of vital importance in membrane separation.

1.7 Conclusions This chapter has made short introductions in a wide span of subjects. In the beginning, we have introduced the basics of membranes, and then we reviewed the original membranes to the most recent modern membranes. A particular attention has been paid to microporous materials including nomenclature, types, and structures of carbon, silica, zeolite, MOF, and porous organic polymer. This overview can help us to understand microporous materials comprehensively from materials perspective. In parallel, fundamentals of membrane separation have been discussed in depth, including separation theory, membrane configuration, membrane fabrication, and membrane module. This discussion can help us to comprehend the separation process from the point of view of membrane engineering. In the end, we were closely looking at the features of microporous membranes, which would play a central role in membrane-based separations. We hope all the basics in this chapter provide a solid platform for our dear readers to digest other chapters more easily and offer a solution to equip us with rich knowledge.

References 1 Baker, R.W. (2004). Membrane Technology and Applications, 2e. Menlo Park, 2 3 4 5 6

CA: Membrane Technology and Research Inc. Bechhold, H. (1907). Z. Phys. Chem. 60: 257. Elford, W.J. (1937). Trans. Faraday Soc. 33: 1094. Zsigmondy, R. and Bachmann, R. (1918). Z. Anorg. Chem. 103: 119. Ferry, J.D. (1936). Chem. Rev. 18: 373. Loeb, S. and Sourirajan, S. Sea water demineralization by means of an osmotic membrane. In: Saline Water Conversion-II, Advances in Chemistry

45

46

1 Introduction of Microporous Membranes

7 8 9 10 11 12 13 14 15 16 17

18

19 20

21

22

23

Series Number, 1963, vol. 38 (ed. R.F. Gould), 117–132. Washington, DC: American Chemical Society. Kolf, W.J. and Berk, H.T. (1944). Acta Med. Scand. 117: 121–134. Henis, J.M.S. and Tripodi, M.K. (1980). Sep. Sci. Technol. 15: 1059–1068. de Vos, R.M. and Verweij, H. (1998). Science 279: 1710–1711. Freni, A., Dawoud, B., Bonaccorsi, L. et al. (2015). Characterization of Zeolite-Based Coatings for Adsorption Heat Pumps. Springer. Wright, P.A. and Connor, J.A. (eds.) (2007). Microporous Framework Solids, 1e. Royal Society of Chemistry. IUPAC (1972). Pure Appl. Chem. 31: 577–638. Romanos, J., Beckner, M., Rash, T. et al. (2012). Nanotechnology 23: 015401. Yang, R.T. (2003). Adsorbents: Fundamentals and Applications, 208–213. Wiley. Mendez-Linan, L., Lopez-Garzon, F.J., Domingo-Garcia, M., and Perez-Mendoza, M. (2010). Energy Fuel 24: 3394–3400. Serp, P. and Figueiredo, J.L. (2009). Carbon Materials for Catalysis, 550. Wiley. (a) Gómez-Serrano, V., Piriz-Almeida, F., Durán-Valle, C.J., and Pastor-Villegas, J. (1999). Carbon 37: 1517–1528. (b) Machnikowski, J., Kaczmarska, H., Gerus-Piasecka, I. et al. (2002). Carbon 40: 1937–1947. (c) Petrov, N., Budinova, T., Razvigorova, M. et al. (2000). Carbon 38: 2069–2075. (d) Garcia, A.B., Martinez-Alonso, A., Leon, C.A., and Tascon, J.M.D. (1998). Fuel 77: 613–624. (e) Polovina, M., Babic, B., Kaluderovic, B., and Dekanski, A. (1997). Carbon 35: 1047–1052. (f ) Fanning, P.E. and Vannice, M.A. (1993). Carbon 31: 721–730. (g) Youssef, A.M., Abdelbary, E.M., Samra, S.E., and Dowidar, A.M. (1991). Indian J. Chem. 30: 839–843. (a) Arriagada, R., Garcia, R., Molina-Sabio, M., and Rodriguez-Reinoso, F. (1997). Microporous Mater. 8: 123–130. (b) Molina-Sabio, M., Gonzalez, M.T., Rodriguez-Reinoso, F., and Sepulveda-Escribano, A. (1996). Carbon 34: 505–509. (c) Bradley, R.H., Sutherland, I., and Sheng, E. (1996). J. Colloid Interface Sci. 179: 561–569. Saha, B., Tai, M.H., and Streat, M. (2001). Process Saf. Environ. 79: 211–217. (a) Sutherland, I., Sheng, E., Braley, R.H., and Freakley, P.K. (1996). J. Mater. Sci. 31: 5651–5655. (b) Rivera-Utrilla, J. and Sanchez-Polo, M. (2002). Carbon 40: 2685–2691. (c) Valdés, H., Sánchez-Polo, M., Rivera-Utrilla, J., and Zaror, C.A. (2002). Langmuir 18: 2111–2116. (a) Pradhan, B.K. and Sandle, N.K. (1999). Carbon 37: 1323–1332. (b) Acedo-Ramos, M., Gomez-Serrano, V., Valenzuella-Calahorro, C., and Lopez-Peinado, A.J. (1993). Spectrosc. Lett. 26: 1117–1137. (c) Gomez-Serrano, V., Acedo-Ramos, M., Lopez-Peinado, A.J., and Valenzuela-Calahorro, C. (1991). Thermochim. Acta 176: 129–140. (a) St˝ohr, B., Boehm, H.P., and Schl˝ogl, R. (1991). Carbon 29: 707–720. (b) ´ Biniak, S., Szyma´nski, G., Siedlewski, J., and Swiatkowski, A. (1997). Carbon 35: 1799–1810. (a) Boudou, J.P., Chehimi, M., Broniek, E. et al. (2003). Carbon 41: 1999–2007. (b) Radkevich, V.Z., Senko, T.L., Wilson, K. et al. (2008). Appl. Catal. A Gen. 335: 241–251.

References

24 (a) Evans, M.J.B., Halliop, E., Liang, S., and MacDonald, J.A.F. (1998). Car-

25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53

bon 36: 1677–1682. (b) Papirer, E.N., Lacroix, R., Donnet, J.B. et al. (1995). Carbon 33: 63–72. Papirer, E., Lacroix, R., Donnet, J.B. et al. (1994). Carbon 32: 1341–1358. Nansé, G., Papirer, E., Fioux, P. et al. (1997). Carbon 35: 515–528. Aldana-Pérez, A., Lartundo-Rojas, L., Gómez, R., and Niño-Gómez, M.E. (2012). Fuel 100: 128–138. Diyuk, V.E., Zaderko, A.N., Grishchenko, L.M. et al. (2012). Catal. Commun. 27: 33–37. Budarin, V.L., Clark, J.H., Tavener, S.J., and Wilson, K. (2004). Chem. Commun. 2736–2737. EPA Alumni Association. Senior EPA officials discuss early implementation of the Safe Drinking Water Act of 1974. Iijima, S. (1991). Nature 354: 56–58. Lalwani, G., Kwaczala, A.T., Kanakia, S. et al. (2013). Carbon 53: 90–100. Lalwani, G., Gopalan, A., D’Agati, M. et al. (2015). J. Biomed. Mater. Res. A 103: 3212–3225. Takeuchi, K., Hayashi, T., Kim, Y.A. et al. (2014). Nanosystems: Phys. Chem. Math. 5: 15–24. Karousis, N., Tagmatarchis, N., and Tasis, D. (2010). Chem. Rev. 110: 5366–5397. Yu, M.F., Lourie, O., Dyer, M.J. et al. (2000). Science 287: 637–640. Peng, B., Locascio, M., Zapol, P. et al. (2008). Nat. Nanotechnol. 3: 626–631. Filleter, T., Bernal, R., Li, S., and Espinosa, H.D. (2011). Adv. Mater. 23: 2855–2860. Wang, Z.K., Ci, L.J., Chen, L. et al. (2007). Nano Lett. 7: 697–702. Lu, X. and Chen, Z. (2005). Chem. Rev. 105: 3643–3696. Hong, S. and Myung, S. (2007). Nat. Nanotechnol. 2: 207–208. Charlier, J.C. and Roche, S. (2007). Rev. Mod. Phys. 79: 677–732. Haddon, R.C., Laura, P.Z., Bin, Z., and Hui, H. (2006). Nano Lett. 6: 562–567. Reibold, M., Paufler, P., Levin, A.A. et al. (2006). Nature 444: 286. Liu, F., Wagterveld, R.M., Gebben, B. et al. (2014). Colloid. Interface Sci. Comm. 3: 9–12. Pyrhönen, J., Montonen, J., Lindh, P. et al. (2015). Int. Rev. Electr. Eng. 10: 12. Cooper, D.R., D’Anjou, B., Ghattamaneni, N. et al. (2012). ISRN Condens. Matter Phys. 501686. Geim, A. (2009). Science 324: 1530–1534. Kusmartsev, F.V., Wu, W.M., Pierpoint, M.P., and Yung, K.C. (2014). Condens. Matter. Mater. Sci. 14060809. Geim, A.K. and MacDonald, A.H. (2007). Phys. Today 60: 35–41. Jayasena, B. and Subbiah, S. (2011). Nanoscale Res. Lett. 6: 95. Eigler, S., Enzelberger-Heim, M., Grimm, S. et al. (2013). Adv. Mater. 25: 3583–3587. El-Kady, M.F., Strong, V., Dubin, S., and Kaner, R.B. (2012). Science 335: 1326–1330.

47

48

1 Introduction of Microporous Membranes

54 Hernandez, Y., Nicolosi, V., Lotya, M. et al. (2008). Nat. Nanotechnol. 3:

563–568. 55 Alzari, V., Nuvoli, D., Scognamillo, S. et al. (2011). J. Mater. Chem. 21:

8727–8733. 56 Nuvoli, D., Valentini, L., Alzari, V. et al. (2011). J. Mater. Chem. 21:

3428–3431. 57 Woltornist, S.J., Oyer, A.J., Carrillo, J.M.Y. et al. (2013). ACS Nano 7:

7062–7066. 58 Bointon, T.H., Barnes, M.D., Russo, S., and Craciun, M.F. (2015). Adv. Mater.

27: 4200–4206. 59 Das, S. and Drucker, J. (2017). Nanotechnology 28: 10. 60 Chiu, P.L., Mastrogiovanni, D.D.T., Wei, D. et al. (2012). J. Am. Chem. Soc. 61 62 63 64 65

66 67 68 69 70 71 72 73 74 75 76 77

78 79 80 81 82 83

134: 5850–5856. Patel, M., Feng, W., Savaram, K. et al. (2015). Small 11: 3358–3368. Li, Z., Chen, L., Meng, S. et al. (2015). Phys. Rev. B 91: 094429. Bonaccorso, F., Colombo, L., Yu, G. et al. (2015). Science 347: 41–52. Avouris, P., Chen, Z., and Perebeinos, V. (2007). Nat. Nanotechnol. 2: 605–615. (a) Novoselov, K.S., Geim, A.K., Morozov, S.V. et al. (2005). Nature 438: 197–200. (b) Morozov, S.V., Novoselov, K., Katsnelson, M. et al. (2008). Phys. Rev. Lett. 100: 016602. (c) Chen, J.H., Jang, C., Xiao, S. et al. (2008). Nat. Nanotechnol. 3: 206–209. Akturk, A. and Goldsman, N. (2008). J. Appl. Phys. 103: 053702. Bernardo, A.D., Millo, O., Barbone, M. et al. (2017). Nat. Commun. 8: 14024. Baringhaus, J., Ruan, M., Edler, F. et al. (2014). Nature 506: 349–354. Neto, A.C., Peres, N.M.R., Novoselov, K.S., and Geim, A.K. (2009). Rev. Mod. Phys. 81: 109–162. Chen, J.H., Jang, C., Adam, S. et al. (2008). Nat. Phys. 4: 377–381. Schedin, F., Geim, A.K., Morozov, S.V. et al. (2007). Nat. Mater. 6: 652–655. Adam, S., Hwang, E.H., Galitski, V.M., and Das, S.S. (2007). Proc. Natl. Acad. Sci. 104: 18392–18397. Steinberg, H., Barak, G., Yacoby, A. et al. (2008). Nat. Phys. 4: 116–119. Trisetyarso, A. (2012). Quantum Inf. Comput. 12: 989. Pachos, J.K. (2009). Contemp. Phys. 50: 375–389. Balandin, A.A., Ghosh, S., Bao, W. et al. (2008). Nano Lett. 8: 902–907. (a) Cai, W., Moore, A.L., Zhu, Y. et al. (2010). Nano Lett. 10: 1645–1651. (b) Faugeras, C., Faugeras, B., Orlita, M. et al. (2010). ACS Nano 4: 1889–1892. (c) Xu, X., Pereira, L.F.C., Wang, Y. et al. (2014). Nat. Commun. 5: 3689. (d) Lee, J.U., Yoon, D., Kim, H. et al. (2011). Phys. Rev. B 83: 081419. Denis, P.A. and Iribarne, F. (2013). J. Phys. Chem. C 117: 19048–19055. Yamada, Y., Murota, K., Fujita, R. et al. (2014). J. Am. Chem. Soc. 136: 2232–2235. Eftekhari, A. and Jafarkhani, P. (2013). J. Phys. Chem. C 117: 25845–25851. Yamada, Y., Yasuda, H., Murota, K. et al. (2013). J. Mater. Sci. 48: 8171–8198. Yamada, Y., Kim, J., Murota, K. et al. (2014). Carbon 70: 59–74. Dissanayake, D.M.N.M., Ashraf, A., Dwyer, D. et al. (2016). Sci. Rep. 6: 21070.

References

84 Zhong, M., Xu, D., Yu, X. et al. (2016). Nano Energy 28: 12–18. 85 Akinwande, D., Tao, L., Yu, Q. et al. (2015). IEEE Nanotechnol. Mag. 9: 6–14. 86 (a) Lalwani, G., Henslee, A.M., Farshid, B. et al. (2013). Biomacromolecules

87 88 89

90 91 92 93 94

95 96 97 98 99 100 101 102

103 104 105

14: 900–909. (b) Rafiee, M.A., Rafiee, J., Wang, Z. et al. (2009). ACS Nano 3: 3884–3890. Shehzad, K., Xu, Y., Gao, C., and Duan, X. (2016). Chem. Soc. Rev. 45: 5541–5588. Barboiu, C., Sala, B., Bec, S. et al. (2009). J. Membr. Sci. 326: 514–525. Ayral, A., Julbe, A., Rouessac, V. et al. (2008). Microporous silica membrane: basic principles and recent advances. In: Membrane Science and Technology (ed. M. Reyes and M. Miguel), 33–79. Elsevier: Amsterdam. Tsuru, T. (2003). Porous Ceramics for Filtration, Handbook of Advanced Ceramics, 291–312. Oxford: Academic Press. Nair, B.N., Elferink, W.J., Keizer, K., and Verweij, H. (1996). J. Colloid Interface Sci. 178: 565–570. Witoon, T., Tatan, N., Rattanavichian, P., and Chareonpanich, M. (2011). Ceram. Int. 37: 2297–2303. Naito, M., Nakahira, K., Fukuda, Y. et al. (1997). J. Membr. Sci. 129: 263–269. Kanezashi, M. and Tsuru, T. (2011). Gas permeation properties of helium, hydrogen, and polar molecules through microporous silica membranes at high temperatures: correlation with silica network structure. In: Membrane Science and Technology (ed. S.T. Oyama and M.S.W. Susan), 117–136. Elsevier: Amsterdam. Lin, Y.S., Kumakiri, I., Nair, B.N., and Alsyouri, H. (2002). Sep. Purif. Rev. 31: 239–248. Wang, X., Miao, X.R., Li, Z.M., and Deng, W.L. (2011). Appl. Surf. Sci. 257: 2481–2488. Eswaramoorthy, M., Neeraj, S., and Rao, C.N.R. (1999). Microporous Mesoporous Mater. 28: 205–210. Guo, L., Dong, X., Cui, X. et al. (2009). Mater. Lett. 63: 13–14. Raman, N.K. and Brinker, C.J. (1995). J. Membr. Sci. 105: 273–279. Labropoulos, A.I., Romanos, G.E., Karanikolos, G.N. et al. (2009). Microporous Mesoporous Mater. 120: 177–185. Breck, D.W. (1974). Zeolite Molecular Sieves: Structure, Chemistry, and Use. New York: Wiley. Cronstedt, A.F. (1993). Observations and descriptions on an unknown mineral species called zeoloites. Proceedings of 9th International Zeolite Conference (ed. R. van Ballmoos, J.B. Higgins, and H.M.J. Treacy), 3–9. Boston: Butterworth-Heinemann. Deville, H.d.S.C. (1862). C. R. Seances Acad. Sci. 54: 324. Wadlinger, R.L., Kerr, G.T., and Rosinski, E.J. (1967). Catalytic composition of a crystalline zeolite. US Patent 3308069. (a) Barrer, R.M. (1978). Zeolites and Clay Minerals as Sorbents and Molecular Sieves. London: Academic Press. (b) Barrer, R.M. (1982). Hydrothermal Chemistry of Zeolites. London: Academic Press.

49

50

1 Introduction of Microporous Membranes

106 (a) Plank, C.J. and Rosinski, E.J. (1964). Process for carbonizing coal. US

107

108

109 110

111 112 113 114

115 116 117 118 119 120 121 122 123 124 125 126 127

128

Patent 3140240; (b) Breck, D.W. (1964). Crystalline zeolite y. US Patent 31300007; (c) Biswas, J. and Maxwell, I.E. (1990). Appl. Catal. 63: 197–258. (a) Argauer, R.J. and Landolt, G.R. (1972). Crystalline zeolite ZSM-5 and method of preparing the same. US Patent 3702886; (b) Flanigen, E.M., Bennett, J.M., Grose, R.W. et al. (1978). Nature 271: 512–516. (c) Szostak, R. (1992). Handbook of Molecular Sieves. New York: Springer. (a) Wilson, S.T., Lok, B.M., Messina, C.A. et al. (1982). J. Am. Chem. Soc. 104: 1146–1147. (b) Flanigen, E.M., Patton, R.L., and Wilson, S.T. (1988). Stud. Surf. Sci. Catal. 37: 13–27. International Zeolite Association (2007). Atlas of Zeolite Framework Types, 6ee. http://www.iza-online.org. Rollmann, L.D. and Valyocsik, E.W. (1995). Zeolite molecular sieves. In: Inorganic Syntheses: Nonmolecular Solids, vol. 30 (ed. D.W. Murphy and L.V. Interrante), 227–234. Wiley. Earl, D.J. and Deem, M.W. (2006). Ind. Eng. Chem. Res. 45: 5449–5454. Auerbach, S.M., Carrado, K.A., and Dutta, P.K. (2003). Handbook of Zeolite Science and Technology. New York: Marcel Dekker Inc. Cheetham, A.K., Férey, G., and Loiseau, T. (1999). Angew. Chem. Int. Ed. 38: 3268–3292. (a) Bucar, D.K., Papaefstathiou, G.S., Hamilton, T.D. et al. (2007). Eur. J. Inorg. Chem. (29): 4559–4568. (b) Parnham, E.R. and Morris, R.E. (2007). Acc. Chem. Res. 40: 1005–1013. Wang, Z. and Cohen, S.M. (2009). Chem. Soc. Rev. 38: 1315–1329. (a) Dinc˘a, M. and Long, J.R. (2008). Angew. Chem. Int. Ed. 47: 6766–6779. (b) Das, S., Kim, H., and Kim, K. (2009). J. Am. Chem. Soc. 131: 3814–3815. Pichon, A. and James, S.L. (2008). CrystEngComm 10: 1839–1847. Murray, L.J., Dinc˘a, M., and Long, J.R. (2009). Chem. Soc. Rev. 38: 1294–1314. Czaja, A.U., Trukhan, N., and Müller, U. (2009). Chem. Soc. Rev. 38: 1284–1293. Mendoza-Cortés, J.L., Han, S.S., and GoddardIII, W.A. (2012). J. Phys. Chem. A 116: 1621–1631. Choi, S., Drese, J.H., and Jones, C.W. (2009). ChemSusChem 2: 796–854. Sumida, K., Rogow, D.L., Mason, J.A. et al. (2012). Chem. Rev. 112: 724–781. Millward, A.R. and Yaghi, O.M. (2005). J. Am. Chem. Soc. 127: 17998–17999. Berger, A.H. and Bhown, A.S. (2011). Energy Procedia 4: 562–567. Masuda, T. (2007). J. Polym. Sci., Part A: Polym. Chem. 45: 165–180. Grubbs, R.H. (ed.) (2003). Handbook of Metathesis. Weinheim: Wiley-VCH. (a) Nagai, K., Masuda, T., Nakagawa, T. et al. (2001). Prog. Polym. Sci. 26: 721–797. (b) Aoki, T. (1999). Prog. Polym. Sci. 24: 951–993. (c) Ulbricht, M. (2006). Polymer 47: 2217–2262. (d) Freeman, B.D., Yampolskii, Y., and Pinnau, I. (eds.) (2006). Materials Science of Membranes for Gas and Vapor Separation. Chichester: Wiley. (a) Finkelshtein, E.S., Makovetskii, K.L., Gringolts, M.L. et al. (2006). Macromolecules 39: 7022–7029. (b) Starannikova, L., Pilipenko, M., Belov, N. et al. (2008). J. Membr. Sci. 323: 134–143.

References

129 Kim, J.F. and Drioli, E. (2016). Perfluoropolymers in membranes. In: Encyclo-

pedia of Membranes (ed. E. Drioli and L. Giorno), 1479–1480. Springer. 130 Anolick, C., Hrivnak, J.A., and Wheland, R.C. (1998). Adv. Mater. 10:

1211–1214. 131 Liaw, D.J., Wang, K.L., Huang, Y.C. et al. (2012). Prog. Polym. Sci. 37:

907–974. 132 (a) Tanaka, K., Okano, M., Toshino, H. et al. (1992). J. Polym. Sci. B Polym.

133 134

135 136 137 138 139

140 141 142 143

144

145 146 147 148 149 150 151 152

Phys. 30: 907–914. (b) Liu, Y., Pan, C., Ding, M., and Xu, J. (1999). Polym. Int. 48: 832–836. (c) Lin, W.H., Vora, R.H., and Chung, T.S. (2000). J. Polym. Sci. B Polym. Phys. 38: 2703–2713. Park, H.B., Jung, C.H., Lee, Y.M. et al. (2007). Science 318: 254–258. (a) Feng, X., Ding, X.S., and Jiang, D.L. (2012). Chem. Soc. Rev. 41: 6010–6022. (b) Ding, S.Y. and Wang, W. (2013). Chem. Soc. Rev. 42: 548–568. (c) Cooper, A.I. (2009). Adv. Mater. 21: 1291–1295. (d) Thomas, A. (2010). Angew. Chem. Int. Ed. 49: 8328–8344. (e) Zou, X.Q., Ren, H., and Zhu, G.S. (2013). Chem. Commun. 49: 3925–3936. Côté, A.P., Benin, A.I., Ockwig, N.W. et al. (2005). Science 310: 1166–1170. El-Kaderi, H.M., Hunt, J.R., Mendoza-Cortés, J. et al. (2007). Science 316: 268–272. Uribe-Romo, F.J., Hunt, J.R., Furukawa, H. et al. (2009). J. Am. Chem. Soc. 131: 4570–4571. Shun, W., Jia, G., Jangbae, K. et al. (2008). Angew. Chem. Int. Ed. 47: 8826–8830. (a) Kuhn, P., Forget, A., Su, D. et al. (2008). J. Am. Chem. Soc. 130: 13333. (b) Kuhn, P., Antonietti, M., and Thomas, A. (2008). Angew. Chem. Int. Ed. 47: 3450. Palkovits, R., Antonietti, M., Kuhn, P. et al. (2009). Angew. Chem. Int. Ed. 48: 6909. Zhang, Y. and Riduan, S.N. (2012). Chem. Soc. Rev. 41: 2083–2094. McKeown, N.B., Makhseed, S., and Budd, P.M. (2002). Chem. Commun. 2780–2781. (a) Eastmond, G.C. and Paprotny, J. (1999). Chem. Lett. 479–480. (b) Eastmond, G.C., Paprotny, J., Steiner, A., and Swanson, L. (2001). New J. Chem. 25: 379–384. (a) Budd, P.M., Ghanem, B.S., Makhseed, S. et al. (2004). Chem. Commun. 230–231. (b) Budd, P.M., Elabas, E.S., Ghanem, B.S. et al. (2004). Adv. Mater. 16: 456–459. Budd, P.M. and McKeown, N.B. (2010). Polym. Chem. 1: 63–68. Ben, T., Ren, H., Ma, S. et al. (2009). Angew. Chem. Int. Ed. 48: 9457–9460. Ben, T., Pei, C., Zhang, D. et al. (2011). Energy Environ. Sci. 4: 3991–3999. Ben, T. and Qiu, S. (2013). CrystEngComm 15: 17–26. Ren, H., Ben, T., Wang, E. et al. (2010). Chem. Commun. 46: 291–293. (a) Ben, T., Shi, K., Cui, Y. et al. (2011). J. Mater. Chem. 21: 18208–18214. (b) Meng, S., Zou, X.Q., Liu, C.F. et al. (2016). ChemCatChem 8: 2393–2400. Inzelt, G. (2008). Conducting Polymers: A New Era in Electrochemistry. Berlin, Heidelberg: Springer. Liu, Q., Tang, Z., Wu, M., and Zhou, Z. (2014). Polym. Int. 63: 381–392.

51

52

1 Introduction of Microporous Membranes

153 Boucle, J., Ravirajan, P., and Nelson, J. (2007). J. Mater. Chem. 17:

3141–3153. 154 Kluiters, S.C.A. (2004). Status review on membrane system for hydrogen

separation. Intermediate Report EU Project MIGREYD NNE5-2001-670. 155 Ghosh, A.K. (2003). Ultrafiltration membrane. PhD thesis, Bombay Univer-

sity, Mumbai, India. 156 Siemann, U. (2005). Solvent cast technology – a versatile tool for thin film

157

158 159 160

161

162

163 164 165 166 167

production. In: Progress in Colloid and Polymer Science, vol. 130 (ed. C. Papadakis, A.M. Schmidt and F. Kremer), 1–14. Springer. Mackenzie, K.J. (1992). Film and sheeting material. In: Encyclopedia of Chemical Technology, 4e, vol. 10 (ed. Kirk-Othmer), 761. Hoboken, NJ: Wiley. Ziabicki, A. (1976). Fundamentals of Fiber Formation. London: Wiley. Schubert, D.W. and Dunkel, T. (2003). Mater. Res. Innov. 7: 314–321. (a) Li, K. (2007). Ceramic Membranes for Separation and Reaction. Chichester: Wiley. (b) Keizer, K., Uhlhorn, R.J.R., and Burggraaf, T.J. (1995). Gas separation using inorganic membranes. In: Membrane Separation Technology, Principles and Applications (ed. R.D. Noble and S.A. Stern), 553–584. Amsterdam: Elsevier. (c) Burggraaf, T.J. and Keizer, K. (1991). Synthesis of inorganic membranes. In: Inorganic Membranes Synthesis, Characteristics, and Applications (ed. R.R. Bhave), 10–63. New York: Chapman & Hall. (a) Larbot, A., Fabre, J.P., Guizard, C., and Cot, L. (1988). J. Membr. Sci. 39: 203–212. (b) Anderson, M.A., Gieselmann, M.J., and Xu, Q. (1988). J. Membr. Sci. 39: 243–258. Ismail, A.F. and Salleh, W.N.W. (2013). Carbon membranes. In: Encyclopedia of Membrane Science and Technology (ed. E.M.V. Hoek and V.V. Tarabara), 1–29. Inc.: Wiley. Baker, R.W. (2004). Membranes and modules. In: Membrane Technology and Applications, 89–160. Wiley. Stern, S.A., Sinclaire, T.F., Gareis, P.J. et al. (1965). Ind. Eng. Chem. 57: 49–60. Rouquerol, J., Avnir, D., Fairbridge, C.W. et al. (1994). Pure Appl. Chem. 66: 1739–1758. Lennard-Jones, J.E. (1924). Proc. R. Soc. Lond. A 106: 463–477. Snyder, M.A. and Tsapatsis, M. (2007). Angew. Chem. Int. Ed. 46: 7560–7573.

53

2 Microporous Silica Membranes 2.1 Introduction Microporous silica is an inorganic material built by connected SiO2 tetrahedra and mostly contains exceptionally small pores with mean sizes of 3–5 Å. As clearly shown in Figure 2.1, silica membranes are generally composed of three layers: (i) a membrane layer, (ii) an intermediate layer, and (iii) a support. A membrane layer has an asymmetric structure with an actual selective microporous silica positioned on a support structure comprising several α- and γ-alumina layers [1]. The main purpose of the support is to provide strength to the silica top layer, while the silica top layer is responsible for the gas permeation and separation. An intermediate layer often exists between membrane layer and support, which provides a good connection between these two layers in order to smooth the transition from support to membrane for minimizing support-derived defects and to strengthen the adhesion between the membrane layer and the support as well. Much research has focused on each component to determine the structure–property relationship between material and separation ability. Each of the three layers that comprise a silica membrane will be more fully explained in this chapter. Silica has been chosen as one of the preferred materials to be prepared as top layers on mesoporous supports because of its high porosity due to its amorphous structure in a wide temperature range, and it is able to be sintered in dry atmosphere at temperatures around 400–700 ∘ C. Microporous silica membranes have shown promising molecular sieving characteristics and can accommodate the separations of various gas or liquid species such as H2 , CO2 , and H2 O [2]. Silica membranes are one of the candidates for separations due to their ease of fabrication, low cost of production, and stability. Thus, the potential of microporous silica membranes in separations will be discussed as well.

2.2 Membrane Synthesis Several techniques have been widely developed to effectively control the pore size of silica-derived membranes, including sol–gel, templating approach, and chemical vapor deposition (CVD). Microporous Materials for Separation Membranes, First Edition. Xiaoqin Zou and Guangshan Zhu. © 2019 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2019 by Wiley-VCH Verlag GmbH & Co. KGaA.

54

2 Microporous Silica Membranes

Microporous silica layer

Intermediate layers

Support layer

Figure 2.1 Schematic diagram of a microporous silica membrane.

2.2.1

Sol–Gel Synthesis Method

Among the abovementioned methods, sol–gel processing attracts most attention due to its excellent processability, its potential to precisely control pore size and pore structure, and also its ease to be scaled up in industry [3]. Besides having great advantage of pore size control based on starting precursors and size of silica sols, sol–gel processing also allows for low temperature synthesis of mixed oxides by mixing metal ions into the silica matrix [4]. Membranes prepared by sol–gel process have relatively high permeation rates because of very thin top layers of 50–200 nm [5]. The characteristics of sol–gel-derived membranes are determined by a large number of sol–gel process parameters. Variation of one or more parameters such as pH of the solution, reaction temperature, and water–silica ratios can dramatically affect the microstructure of the resulting silica xerogels [6]. Therefore, it is important that the effects of these parameters are well understood to make sol–gel process a reliable and practical technology for membrane fabrication [7]. Sol–gel synthesis processes are divided into three main procedures: (i) polymeric route and slip casting, (ii) the route of colloidal sol and hot coating, and (iii) the surfactant-templated method. Two main types of sol–gel routes that can be distinguished are the polymeric route and the colloidal route. In the polymeric route, the elemental unit that is usually a silicon alkoxide, Si(OR)4 , is initially used as a molecular precursor. The most common precursors include tetramethoxysilane Si(OCH3 )4 (TMOS) and tetraethoxysilane Si(OC2 H5 )4 (TEOS) in solution in their associated alcohols. In the colloidal route, the elemental unit is a solid nanoparticle dispersed as stable sol in a liquid. The sol stability is related to electrostatic and steric repulsive interactions [8]. In comparison with polymeric route, hydrolysis and condensation reaction is fast in the colloidal route, which is achieved by adjusting the reaction conditions (type of alkoxide, solvent, catalyst, composition of reactants, molar ratio of water, temperature, etc.), resulting in a fully hydrolyzed alkoxide (highly branched polymer). This rapid condensation process causes particulate growth and/or the formation of precipitates [9]. Preparation of mesoporous silica layers can be considered by using colloidal dispersion; however, for preparation of microporous silica layers, the most convenient way is the polymeric route from solutions of molecular precursors like alkoxides. The polymeric gel synthesis consists of hydrolysis and condensation reactions, which can be either acid

2.2 Membrane Synthesis

Sol–gel route

Colloidal route

Polymeric route Elemental unit: alkoxide molecules Si(OR)4

Elemental unit: oxide nanoparticles

Accessible porosity: micropores or mesopores

Accessible porosity: mesopores

Acid catalysis results in microporous materials

Base catalysis does not result in microporous materials due to large and highly cross-linked particles

Figure 2.2 Sol–gel process of microporous silica by polymeric and colloidal routes.

or base catalyzed. However, base-catalyzed synthesis does not result in microporous materials due to large and highly cross-linked particles, whereby only acid-catalyzed reactions will result in microporous materials. Figure 2.2 shows the graphical summary of sol–gel process of microporous silica from colloidal dispersions or molecular solutions [10]. The key mechanism is that in the hydrolysis step, a silanol is formed, which can eventually be developed into a polysilicate in the condensation step. In this way, inorganic polymers are formed, and the degree of polymerization will depend on the ratio of hydrolysis rate over the condensation rate [11]. In the acid catalytic system of polymeric sol route, the hydrolysis reaction is kept slower and typically achieved by adding a small amount of water, resulting in a partially hydrolyzed alkoxide and the formation of a linear inorganic polymer. Through the subsequent gelation process, polymeric sols form a gel network. By controlling the reactant concentrations and the synthesis conditions, the sol–gel morphology can be changed. This is achieved by the control of rates of hydrolysis and condensation by changing amount of water or catalyst used. The structure of the resulting polymers could change from linear to weakly branched polymers; however, for formation of microporous silica membranes, short branched linear polymers are the most preferred in this case [10, 11]. 2.2.2

Templating Approach

Templating approach can also be applied to control the microporosity of the silica layers (Figure 2.3). The first strategy consists of using templating molecules (incorporated in the gelation medium) that are inert toward the chemical process leading to the inorganic network. A second approach consists of using the modified alkoxides where a molecular group acting as a template is covalently bonded to the Si atom.

55

56

2 Microporous Silica Membranes Silica precursor Assembling

Removing templates Extraction/calcination

Templates (surfactants or silanes)

Microporous silica with templates

Microporous silica without templates

Figure 2.3 Schematic of the template approach for the synthesis of microporous silica.

The synthesis conditions are selected in order to directly associate the porosity of the final materials with the thermal elimination of the templating species. There are mainly two types of templates that have been used for tailoring porous structures, i.e. surfactant molecules and organic ligands/polymers. Surfactant molecules incorporated in the matrix could arrange matrix molecules around them by means of non-covalent interactions. The thermal elimination of the template then leaves a residual porosity in the membrane. Thus, the control of shape and pore size could be easily achieved by this method. The same case goes for organic ligands/polymers; however, they are different in the sense that they are bonded covalently to the siloxane matrix [12]. Wang et al. adopted the templated approach to fabricate microporous hollow silica microspheres using nonionic surfactant nonylphenol ethoxylated decyl ether (NP-10) micelles as the template, n-octadecane as the core, and sodium silicate as the silica precursor [13]. The core materials were removed by ethanol during the reaction; thus, the hollow structure can be formed without calcination or chemical etching. Eswaramoorthy et al. also synthesized hexagonal microporous silica employing a short-chain amine as the template molecule and succeeded in obtaining these materials with well-defined pore distributions between 1.0 and 2.0 nm. The surface area of SiO2 after removal of the template by calcination is 800 m2 g−1 [14]. Guo et al. synthesized uniform hollow microporous silica spheres using polystyrene spheres and cetyltrimethylammonium bromide as co-templates at room temperature in concentrated aqueous ammonia [15]. In order to have template approach successfully implemented, the following criteria must be satisfied: (i) the templating molecules must be uniformly incorporated in the inorganic matrix without aggregation or phase separation to avoid creating pores larger than the size of individual templating molecules, (ii) the synthesis and processing conditions should result in dense embedding matrix so that pores are created only by the template removal, and (iii) template removal should be achieved without collapse of the matrix so that the resultant pores preserve the original size and shape of the template [16]. 2.2.3

Chemical Vapor Deposition

CVD is another way of producing microporous silica membranes. The setup is schematically displayed in Figure 2.4. It could be effectively applied for the deposition of silicon oxide or metal oxide layers on porous substrates in order to modify their porous structure and permeation properties.

2.3 Intermediate Layers

Pressure gauge Furnace To disposal of exhaust gases

Load door

Substrate

Deposited SiO2 layer

Gas inlet

Figure 2.4 Illustration of CVD setup for depositing microporous silica layers.

Through CVD technique, a variety of microporous or nearly dense ceramic membranes can be produced. In order to obtain a silica film through CVD reaction, the process consists of the thermal decomposition of a silicon-based precursor and the chemical reaction with an oxidant gas in contact with the hot substrate where the growth of the film takes place. The gaseous precursors are mainly TEOS and TMOS fed to the hot surface by argon or nitrogen as carrier gas. Other possible precursors include silicon tetrachloride (SiCl4 ) or silane (SiH4 ), whereas oxidant gases are air, pure O2 , O3 , N2 O, or water vapor [8, 17]. Two kinds of supports that have been widely used for silica membranes prepared by CVD are Vycor glass and alumina. Vycor glass is compatible with SiO2 in terms of thermal expansion coefficients; thus, they suffer less from cracking due to thermal cycling. However, the low inherent permeance through Vycor glass led to use of other supports such as α-alumina, γ-alumina, and coated γ-alumina tubes [4].

2.3 Intermediate Layers Intermediate layers are the layers between membrane layer and the support (Figure 2.1), which are prepared by dip coating of nanoparticle dispersions with subsequent drying and calcination. Typical compositions include transition aluminas, silica, and zirconia. Precursor particles are made by precipitation from simple salt solutions or hydrolysis of organometal reagents. To obtain homogeneously packed layers with little shrinkage, it is important to control particle agglomeration during synthesis and to remove any agglomeration after synthesis. A well-known synthesis of γ-alumina layers starts with hydrolysis of ATSB at 90 ∘ C, followed by HNO3 addition, resulting in the partial dissolution of the boehmite precipitate and redispersion of agglomerates [18]. The HNO3 addition also ensures colloidal charge stabilization by preferential proton sorption. The hydrolysis/peptization method has a favorable yield but suffers from the presence of residual agglomerates. The agglomerates can be removed by high-speed centrifugation [19], which leads to stable and homogeneous layers. Sonochemical and modified emulsion precipitation methods help avoid formation of the agglomerates. More recently, methods were developed in which

57

58

2 Microporous Silica Membranes

agglomeration is completely avoided up front, in particular for sonochemical and modified emulsion precipitation [20, 21]. Intermediate layer formation from a nanoparticle dispersion is often assisted by additions of linear chain polymers such as PVA to the dispersion or by pretreatment of the support with polymers to minimize penetration. An example includes the repair of a commercial supported γ-alumina membrane by dip coating with a boehmite precursor nanoparticle dispersion [22].

2.4 Support The support layer is to provide additional strength for supporting membrane layer without easily breaking (Figure 2.1). The suitable materials for making supports can be varied from ceramics to metals, and the configurations can be different including disks, tubes, fibers, mesh, etc. (Figure 2.5). A variety of methods for fabricating support layers have been employed. One common method is the dip coating of agglomerate-free submicron particle dispersions, made from commercially available α-Al2 O3 powers. Colloidal stabilization can be adjusted such that coherent dense-packed layers are formed with 25 nm surface roughness after slight sintering at temperatures around 1000 ∘ C [23]. Other methods result in thicker support layers and include colloidal filtration and centrifugal casting [24], thus resulting in high strength, excellent surface properties, and roundness of the tubes. Carrier structures are generally made with conventional ceramic forming methods using commercially available and coarse α-Al2 O3 powders. Dry pressing is used to make small disks for research purposes. Extrusion is used for tubes and multichannel honeycomb structures [25, 26]. These forming methods are very suitable for fairly cheap mass-scale production but have limited near-net-shape capabilities. Non-roundness and other dimensional limitations may result in sealing and construction problems in high temperature membrane reactors. It is for this reason that gel-casting methods are also considered [27]. Such methods allow for a higher initial solid load, better control of overall homogeneity during forming, and, hence, better dimensional specifications. The large pore diameter requires very high sintering temperatures and may result in poor mechanical strength and reliability. This problem might be addressed by application of wet-chemical techniques such as phosphate bonding [28], which provides thermochemical stability up to 900 ∘ C.

2.5 Modification of Silica Membranes The major problem for the application of amorphous microporous silica membranes in separation processes particularly for hot streams is their poor thermal stability in moist atmospheres. The possible reason is that silica specimens on the pore walls are quite vulnerable to guest (e.g. moisture). These specimens are prone to hydrolyze upon contacting water at high temperature. The hydrolyzed

2.5 Modification of Silica Membranes

Al2O3 disk

Glass fiber

Ceramic tube

Metal mesh

Figure 2.5 Various shapes of supports made of different materials. HO

OH OH

HO

OH OH

HO

OH

OH

Initial pore structure

OH

HO

Water HO

OH

HO

HO

Temperature

HO

OH HO

OH

Water

OH

HO

OH

OH

Temperature

OH OH HO

Hydrolysis of siloxanes and formation of mobile silica

Mobile silica fills small pore and then condenses to close and large pore widened

Figure 2.6 Proposed mechanism for the instability of the pore structure in microporous silica membrane under humidity and elevated temperature.

silica is quite mobile with consequences that mobile silica condenses the small pores and the large pore is widened (Figure 2.6). The hydrothermal stability of amorphous silica membranes can be improved by doping a second metal. The metal-doped silica membranes, prepared by either the sol–gel [29–31] or the CVD method [32, 33], exhibited separation properties similar to the pure silica membranes, but the hydrothermal stability of the metal-doped silica membranes was greatly improved. Korean researchers found that for the aluminum-doped silica membrane, significantly less loss of hydrogen permeance was observed after exposure to stream at high temperature as compared with the pure silica membrane [33]. Besides, the metal-doped silica membranes could also improve the performance in pervaporation separation involving water [34, 35]. Another strategy to improve the hydrothermal stability of silica membranes is to eliminate surface silanol groups by surface modification (i.e. the preparation of organic–inorganic hybrid membrane). De Vos et al. first successfully produced hydrophobic silica films using a hydrophobic methyl template [36]. The incorporation of methyl groups in the microporous silica membranes was proven to enhance the service time in the dehydration of butanol/water mixture from a few weeks to more than 18 months (water flux of 4 kg m−2 h−1 with selectivity of 500–20 000) [37]. Recently, Castricum et al. developed surface-functionalized hybrid ceramic membranes using bridged bis-silyl precursors of (EtO)3 Si—CH2 CH2 —Si(OEt)3 and methyltriethoxysilane [38]. Wei et al. also successfully synthesized hydrophobic microporous silica

59

60

2 Microporous Silica Membranes

membranes using TEOS and ethylenetriethoxysilane [39]. The hydrothermal stability of silica membrane was also improved by modifying the surface with carbon. The modified membranes were referred to carbonized template molecular sieving silica (CTMSS) membranes [40]. In Duke’s study, CTMSS and MSS were structurally similar, and both contained silanol contents on their surfaces [41]. However, the CTMSS bulk samples and permselective membranes showed improved integrity and resilience after hydrothermal treatment. The stability improvement for the CTMSS was attributed to the presence of carbon templates in the micropores, which prevented the formation of mobile silica specimens. The carbon templates acted as molecular barriers in the micropore spaces to inhibit the silanol migration and condensation. Thus, the micropore closure caused by the condensation reaction was effectively mitigated in the CTMSS. Such a surface functionalization was able to improve the hydrothermal stability of the porous structure of silica without significant loss of hydrophilic properties. In many cases, microporous silica membranes fabricated by CVD contained a large amount of silanol groups, leading to hydrothermal instability. Gu and Oyama and coworkers reported high temperature (600 ∘ C) CVD-synthesized ultrathin silica membranes on graded mesoporous supports. This membrane exhibited a high hydrogen permeance on an order of 10−7 mol m−2 s−1 Pa−1 with H2 /CH4 selectivity above 1000 at 600 ∘ C [22, 42].

2.6 Microporous Silica Membranes for Hydrogen Separation Microporous silica membranes were considered as an alternative class of membrane materials for hydrogen separation owing to their small pores and high stabilities. In Table 2.1, the most representative results for microporous silica membranes in the application of hydrogen separation are summarized. As shown in Table 2.1, we can see that most works have been done in past two decades (1996–2014). In 1996, Sea et al. employed different silica sources of tetraethyl orthosilicate (TEOS), phenyltriethoxysilane (PTES), and diphenyldiethoxysilane (DPDES) to synthesize microporous silica membranes. Amorphous silica layer was deposited in the mesoporous structure of a γ-alumina film coated on a porous α-alumina tube and then by evacuating the reactant through the porous wall. Their studies indicated that hydrogen permeance reached 10−7 mol m−2 s−1 Pa−1 at the temperature of 600 ∘ C and did not depend on silica sources. One membrane formed using TEOS as the silica source showed H2 /N2 selectivity of 100 at permeation temperatures of 500–600 ∘ C. When using DPDES as the silica source, the selectivity of the derived membranes was higher. Substituting DPDES by PTES, the prepared membrane showed separation performances between TEOS- and DPDES-derived membranes. According to their results, the use of PTES and DPDES was useful for controlling the micropore size [44]. In 2001, Kim et al. prepared microporous silica membranes on alumina supports by the sol–gel method from the TEOS and MOTMS precursors. The composition of the final

Table 2.1 Summary of microporous silica membranes for hydrogen separations. Membrane type

Configuration

Synthesis method

T (∘ C)

dP (bar)

Permeance (mol m−2 s−1 Pa−1 )

H2 /N2

SiO2 /Al2 O3

Disk

Sol–gel

200

1

4.02 × 10−7

SiO2 /Al2 O3

Tube

CVD

200

—

5.0 × 10−7

SiO2

Tube

CVD

600

—

SiO2 /Vycor

Tube

CVD

600

SiO2

Tube

Sol–gel

SiO2 /Al2 O3

Disk

SiO2 /Al2 O3

H2 /CO2

Refs.

64

7.5

[43]

64

—

[44]

6.0 × 10−7

160

—

[45]

0.2

1.8 × 10−8

—

8200

[46]

300

—

2.0 × 10−7

6200

295

[47]

Sol–gel

150

0.7

7.0 × 10−9

112

101

[48]

Tube

Sol–gel

100

—

2.1 × 10−7

12

—

[49]

SiO2 /Al2 O3

Tube

CVD

600

1

1.8 × 10−7

—

2222

[50]

SiO2 /Al2 O3

Tube

Sol–gel

600

—

4.0 × 10−6

30

—

[51]

SiO2 /Al2 O3

Tube

Sol–gel

500

0.9

4.2 × 10−6

400

—

[52]

SiO2 /Al2 O3

Tube

Sol–gel

600

—

5.0 × 10−7

6000

—

[53]

SiO2 /steel

Disk

Sol–gel

350

0.4

5.0 × 10−7

115

—

[54]

SiO2 /Al2 O3

Tube

Sol–gel

250

—

1.0 × 10−8

—

1000

[55]

SiO2 /Al2 O3

Disk

CVD

250

—

1.5 × 10−8

19.4

81

[56]

SiO2 /Al2 O3

Disk

Sol–gel

300

—

2.2 × 10−7

—

10.2

[57]

SiO2 /Al2 O3

Tube

Sol–gel

600

5

1.9 × 10−7

—

1500

[58]

SiO2 /Al2 O3

Tube

CVD

600

—

4.0 × 10−8

1265

421

[59]

SiO2 /Al2 O3

Tube

Sol–gel

500

1

6.0 × 10−7

70

40

[60]

62

2 Microporous Silica Membranes

solution was xMOTMS: (1 − x)TEOS: 9EtOH: 1.0H2 O: 0.2HCl on a molar basis where the mole fraction of MOTMS (x) was varied from 0 to 0.3. During the synthesis, silane coupling agent of MOTMS was added as a template in order to control the pore structure to TEOS. Also, the effects of thermal treatment on the micropore structure of the resultant silica membranes were investigated. Gas permeation tests revealed a H2 permeance around 7 × 10−7 mol m−2 s−1 Pa−1 at 300 ∘ C, and the permeation selectivities for H2 over CO2 and over CH4 in equimolar gas mixtures were 36 and 132, respectively [47]. In 2000s, there were some preliminary studies on water–gas shift (WGS) reaction and catalytic reforming of methane and methanol in H2 -selective microporous silica membranes. The WGS experiments were conducted for clean feed streams only containing steam and CO at low temperatures of less than 290 ∘ C. Giessler et al. tested WGS in a hydrophobic MSS packed with a Cu/ZnO/Al2 O3 catalyst at low temperature. The MSS membrane exhibited a high H2 permeance of 1.5 × 10−6 mol m−2 s−1 Pa−1 and relatively low H2 /CO2 and H2 /N2 selectivities of 6 and 8, respectively [61]. The same research group also reported WGS results in a Co-doped MSS with H2 permeance of 5 × 10−8 mol m−2 s−1 Pa−1 and H2 /CO selectivity of 13 and H2 /CO2 selectivity of 4 [62]. In 2006, a Ni-doped porous silica membranes containing various Ni contents (Si/Ni = 4/1–1/1) were prepared using the sol–gel technique with a purpose of H2 separation at high temperature (around 500 ∘ C) [52]. In this study, H2 -selective permeation characteristics and hydrothermal stability of the membranes were tested in steam at 500 ∘ C to demonstrate that hydrothermal treatments of the membranes before exposure to H2 were quite effective to prevent the further densification of Ni-doped amorphous silica networks due to reduction in H2 and sintering in steam (500 ∘ C, 70 kPa). The results showed that an asymptotic steady H2 permeance of 4.6 × 10−9 m3 m−2 s−1 Pa−1 with a H2 /N2 selectivity of 400 (Figure 2.7). For more practical separation exercise, Brands et al. in 2010 investigated the long-term effects of exposing a silica membrane for 1100 hours in a flue gas stream of a coal power plant. Single gas permeation testing following flue gas exposure revealed a maximum permeation of 1.85 × 10−8 and 2.13 × 10−8 mol m−2 s−1 Pa−1 for helium and hydrogen, respectively, and respective selectivity of 5.1 and 5.2 for He/N2 and H2 /CO2 was achieved at a pressure difference of 2 bar at 200 ∘ C. The permeation behavior of the membrane appeared to be altered as a result of flue gas exposure with the membrane, displaying a reduced H2 flux in contrast to an unexposed but otherwise identical membrane, which displayed fluxes an order of magnitude higher than the membrane used in the power plant. This change in permeation behavior resulted from the densification of silica matrix after long-term exposure to flue gas containing water vapor [63]. More recently in 2014, Jabbari et al. carried out an experimental study on the synthesis of silica membrane for hydrogen purification, in which synthesis of γ-alumina intermediate layer using cheaper and safer source was investigated. For this purpose, aluminum hydroxide was selected, and the boehmite sols were prepared by acid- or base-catalyzed hydrolysis of different salts for comparing with alkoxide source. After γ-alumina layer formation, the gas permeance mechanism

2.7 Microporous Silica Membranes for Carbon Dioxide Separation

10–4 Permeance (m3(STP)· m–2· s–1· kPa–1)

He H2

10–5

10–6

10–7 N2 10–8

10–9

0

20

40

60

80

Time (h)

Figure 2.7 The changes of gas permeances in function of time for Ni-doped silica membranes (Si/Ni = 2/1) after exposed to steam (500 ∘ C, 70 kPa) and fired at 550 ∘ C in steamed atmosphere [52]. Source: Copyright 2006, reproduced with permission from Elsevier.

was approximately changed. These results were similar to SEM results and N2 permeance experiments of sample 3 in which the substrate was coated with alkoxide sol. However, the γ-alumina layer of sample 2 had no good adhesion to the substrate. Nevertheless, the use of aluminum hydroxide can be promised to the synthesis of γ-alumina layer; the membrane was synthesized on the modified support with aluminum tri-sec-butylate sol. In particular, in the synthesized silica membrane as the temperature increased, permselectivity of H2 /CO2 and H2 /N2 increased from 4.7 and 7.3 at room temperature to 9.4 and 11.6 at 100 ∘ C and to 23.4 and 31.3 at 200 ∘ C, respectively [64]. After year-by-year attempts, the performances in hydrogen separations using microporous silica membranes were significantly improved assisted by more mature technology in membrane synthesis.

2.7 Microporous Silica Membranes for Carbon Dioxide Separation Another branch of gas separation using microporous silica membranes is CO2 capture. Table 2.2 summarizes the data in previous works related to microporous silica membranes for CO2 separation. Kuraoka et al. reported the preparation of asymmetric microporous silica xerogel membrane and its application to CO2 separation. The silica xerogel layer was prepared by the sol–gel method with molar ratio in the silicate sol composition of 1.0TEOS: 20C2 H5 OH: 2H2 O: 0.01HCl [65]. Dip-coating procedure was performed at 22 ∘ C and 50% relative humidity. The membranes were subjected to

63

64

2 Microporous Silica Membranes

Table 2.2 Microporous silica membranes for CO2 separation. Membrane material

Synthesis method

Selectivity

CO2 permeance (mol m−2 s−1 Pa−1 )

Refs.

Asymmetric silica xerogel

Sol–gel

66 (CO2 /N2 )

1.7 × 10−7

[65]

Aminosilicate

Sol–gel

100 (CO2 /N2 )

1.0 × 10−7

[66]

Silica

Sol–gel

18 (CO2 /CH4 ) 13 (CO2 /N2 )

1.6 × 10−8 2.0 × 10−8

[47]

Silica

Sol–gel

36 (CO2 /N2 )

7.0 × 10−7

[67]

Nanoporous silica composite

Sol–gel

19 (CO2 /N2 )

10−9 –10−8

[68]

heating at 150 ∘ C for two hours in air. The membranes were then obtained after two times repeating of dip-coating and heating procedures. The selectivity of CO2 /N2 for this membrane attained in the permeation test was more than 60 at 25 ∘ C, which was higher than the Knudsen coefficient of 0.8. Besides, CO2 in the air was concentrated to more than 1000 ppm by using this membrane at 25 ∘ C. From their study where the membrane could concentrate CO2 from such a low CO2 gas concentration (about 300 ppm) to high concentration (about 1000 ppm), it is concluded that this asymmetric microporous silica xerogel membrane can be applied to the recovery of CO2 from the gases with low CO2 concentration. Xomeritaakis et al. introduced the concept of a novel microporous aminosilicate membrane with fixed amine ligands incorporated in the silica matrix in order to enhance membrane affinity for CO2 separation [66]. Incorporation of amine groups were carried out by two different procedures. In the first procedures based on amine salt aqueous solution, they employed sodium glycinate salt CH2 NH2 COONa (GlyNa) as the amine source. In this case, the silica sol was prepared by adding suitable amounts of GlyNa, H2 O, and HCl in the stock sol. The molar ratios of NH2 /Si were adjusted at 0.1, 0.17, 0.25, and 0.33. The second procedure is based on co-condensation of TEOS with an amine-containing alkoxysilane, e.g. 3-aminopropyltriethoxysilane (APTES). The molar ratio of NH2 /Si was 0.2. With a feed of 1–20 vol% CO2 and 0–40% relative humidity at 22 ∘ C, the highest CO2 /N2 separation factor achieved was in the range of 100–200, and the CO2 permeance was in the range of 0.067–1.0 × 10−7 mol m−2 s−1 Pa−1 . Yoshioka et al. prepared sol–gel-derived microporous silica membranes that showed high performance in CO2 gas permeance and permeation selectivity of CO2 /N2 [67]. The microporous silica membranes used in the study were prepared through a two-step sol–gel procedure: (i) preparation of the colloidal sols and (ii) sol coating on porous substrate and firing. The silica colloidal sols were prepared through hydrolysis, polymerization, and condensation of TEOS aqueous solution with a small amount of HNO3 as the catalyst. The coating was done by making a cloth wetted with the colloidal sols, gently contacting the hot substrate. These hot-coating procedures allowed gelation to occur instantly on the substrate and prevented the sol from penetrating deep into the substrate pores,

2.7 Microporous Silica Membranes for Carbon Dioxide Separation

Figure 2.8 Permeance vs. kinetic diameter for several gas species tested on microporous silica membranes [67]. Source: Copyright 2001, reproduced from permission, John Wiley & Sons.

O2

CH4

Ar CO2 He H2

10–5

C3H8

C2H4

l–C4H8

N2

iso–C4H10

Pobs (mol m−2 s−1 Pa−1)

Mem.-A (150 °C) Mem.-B (200 °C)

10–6

10–7

10–8

2

3

4

5

6

Kinetic diameter (Å)

which resulted in the formation of a very thin separation layer. After colloids coating, the membrane was calcined at 360–570 ∘ C for about 15 minutes in air. After repeating the coating and calcination procedures several times with each colloidal sol on the order of its particle size, membranes with sub-nanopores were obtained, which exhibited molecular sieving property in gas transport even at 150–200 ∘ C (Figure 2.8). Additionally, this membrane gave a measured CO2 permeance of 7.0 × 10−7 mol m−2 s−1 Pa−1 , and the permeation selectivity for CO2 over N2 was 36 at 50 ∘ C. Lee et al. studied nanoporous silica composite membranes that were synthesized by sol–gel coating silica sols templated with covalently bonded MTES as well as non-reacted TPABr on SiO2 /α-Al2 O3 composite tubular supports to enhance the CO2 /N2 separation factor [68]. The organic-templating silica sols were prepared from the mixture of a silica sol and an organic template MTES or TPABr at room temperature. Silica sols were templated by TPABr via adding 6 wt% of TPABr to the silica sol with a final molar ratio of TEOS: 3.8EtOH: 5.1H2 O: 0.056HCl (1.0 M solution), followed by aging for 6–18 hours at 500 ∘ C after mixing vigorously at room temperature for 6 hours. The final molar ratio of the MTES-templating silica sol was 0.1MTES: 0.9TEOS: 3.8EtOH: 5.1H2 O: 0.056HCl (1.0 M). CO2 permeance and CO2 /N2 separation factor for the MTES-templated silica composite membranes were enhanced to about 10−9 –10−8 mol m−2 s−1 Pa−1 and 19 at room temperature, respectively. Boffa et al. synthesized microporous niobia-silica membrane where its separation was based on a combination of size-based sieving and variations in molecule wall interactions between gas molecules [69]. Sols were prepared from niobium(V) penta(n-butyloxide) and TEOS in an acidic alcoholic solution. TEOS was

65

66

2 Microporous Silica Membranes

NS top layer γ-Alumina NS top layer

α-Alumina

1 μm (a)

100 nm (b)

Figure 2.9 SEM images of the cross section of an NS membrane: (a) asymmetric three-layer structure and (b) magnified view of the 150 nm thick NS top layer [69]. Source: Copyright 2008, reproduced from permission, John Wiley & Sons.

pre-hydrolyzed before reaction with niobium(V) penta(n-butyloxide) due to that metal alkoxides derived from silicon were less sensitive to hydrolysis than those derived from niobium. Microporous films were formed by coating asymmetric γ-alumina disks with the polymeric sol (Si/Nb = 3 : 1), followed by calcination at 500 ∘ C (Figure 2.9). The ideal selectivity of CH4 /CO2 for niobia-silica membrane was increased from 3.3 at 80 ∘ C to 7.0 at 200 ∘ C. Molecular sieving behaviors are expected to hold in defect-free microporous silica membranes. However, defect-free microporous silica membranes are difficult to fabricate. Although microporous silica membranes showed promising molecular sieving characteristics, there are still room of improvement on flux and selectivity of microporous silica membranes.

2.8 Microporous Silica-Based Membranes for Pervaporation Process Pervaporation is a process that can achieve separations of liquid mixtures by selective permeation and vaporization through a membrane. As shown in Figure 2.10, in pervaporation, a liquid mixture is fed to one side of the membrane, and permeate driven by the pressure gradient comes out from the opposite side in a vapor phase, and the permeate vapor is condensed and collected. Pervaporation has been applied in diverse separations such as the dehydration of

2.8 Microporous Silica-Based Membranes for Pervaporation Process

Figure 2.10 A schematic diagram for the pervaporation process.

Feed

Membrane

Retentate

Condenser Vacuum pump

Permeate

organics, the removal of organics from an aqueous phase, and organic–organic separation. As for the membranes, microporous silica and silica-based materials are also excellent candidates to make membranes for pervaporation. The first and the most important application of pervaporation is the removal of water from azeotropic alcohol–water mixtures. Early in 1995, van Gemert et al. proposed that amorphous silica membranes should have high permselectivity in alcohol dehydration and evaluated more dense silica membranes in dehydration of methanol, ethanol, and isopropyl alcohol (IPA) [70]. Since early 2000s, efforts were sprout out in developing silica-based membranes for dehydration of alcohols. Asaeda et al. in 2002 removed water from a 90 wt% IPA–water solution with pervaporation at a normal boiling point through a silica–zirconia composite membrane and reported a stable water flux that reached as high as 9 kg m−2 h−1 with a separation factor of approximately 2000 [71]. Wang et al. evaluated the pervaporation performances of BTESE-derived hybrid silica membranes for dehydration of ethanol and IPA. In this study, BTESE-derived membranes prepared at higher calcination temperature tended to show a higher separation factor. The membrane calcined at 300 ∘ C showed a water flux of 3.4 kg m−2 h−1 and a separation factor of 4370 in the dehydration of 90 wt% IPA at 75 ∘ C [72]. More recently, Tsuru et al. evaluated the stability of BTESE-derived hybrid silica membranes during long-term pervaporation with a water-rich stream (up to 50 wt% water content in IPA) and at high temperature in a vapor permeation mode (up to 110 ∘ C using a 90 wt% IPA mixture) [73]. Figure 2.11 shows the time course of pervaporation performance at 75 ∘ C for dehydrating IPA mixtures with various IPA concentrations. The operation was initiated using a 90 wt% IPA solution. The water content in the feed was increased in a stepwise manner and finally returned to a 90 wt% IPA solution. The water flux increased with a decrease in the feed IPA concentration. This can be ascribed mainly to an increase in the partial pressure of the water in the feed, but, interestingly, the degree of increase in water flux was greater than that in the water partial pressure. The water flux was increased from 2.3 to 6.4 kg m−2 h−1 (a 2.5-fold increase) with a decrease in the feed IPA concentration from 90 to 50 wt%, while water partial pressure was increased only 1.5-fold. This suggested that IPA molecules inhibited the diffusion of water molecules in the

67

2 Microporous Silica Membranes

Separation factor

Figure 2.11 Time course of the pervaporation performance on a BTESE-derived microporous silica membrane for dehydrating IPA mixtures with various IPA concentrations at 75 ∘ C [73]. Source: Copyright 2016, reproduced with permission from Elsevier.

50/50

104 H2O/IPA = 10/90

103

20/80

30/70

40/60

10/90

102

101 15 Water flux (kg m–2 h–1)

68

10

5

0

0

5

10

15

20

25

30

35

Time (h)

micropores. The adverse impact of IPA on water permeation can be explained by the fact that the adsorbed IPA on pore surface blocked water flow into the micropore. The membrane possessed excellent stability for pervaporation under high water content mixtures, shown by stable water flux and separation factor for a 50 wt% IPA aqueous solution over several hours. Dehydration of ethanol and butanol was also exemplified on microporous silica membranes, and the reported data is summarized in Table 2.3 [37, 38, 72, 75, 77, 84]. Dehydration of acid solutions such as highly concentrated acetic acid is also an interesting industrial application for pervaporation. In 1990, Asaeda and coworkers reported the first dehydration of an aqueous acetic acid solution using silica membranes [78]. After modifying this membrane, they reported an improved performance in water flux of 3.06 kg m−2 h−1 with a high separation factor of 800 for a 90 wt% acetic acid/water mixture at 75 ∘ C. Silica and silica–titania composite membranes were stable in highly concentrated acetic acid solutions, but degradation was observed for a water-rich system [79]. Recently, Tsuru et al. reported that organic–inorganic hybrid silica membranes derived from BTESE showed excellent stability in the dehydration of acetic acid solutions with a water flux of 2.0–4.0 kg m−2 h−1 and a separation factor of 200–500 at 90 wt% of acetic acid at 75 ∘ C [81]. Stable pervaporation performance was confirmed after long-term exposure (70 days) to a 90 wt% acetic acid solution at 75 ∘ C and for acetic acid concentrations that ranged from 30 to 90 mol%. Separation of organic–organic mixtures is important in the chemical and petrochemical industries. Regarding the application of silica membranes for organic–organic mixtures, successful separation was reported for benzene/cyclohexane and methanol/dimethyl carbonate mixtures. Matsuyama

2.9 Microporous Silica-Based Membranes for Desalination

Table 2.3 Membrane pervaporation performances for organic–water and organic–organic mixtures.

Membrane

Feed

Total flux (kg m−2 h−1 )

Separation factor

Silica

IPA/H2 O

0.25

500

[70]

SiO2 —ZrO2

IPA/H2 O

9.0

1500

[34]

Refs.

SiO2 —ZrO2

IPA/H2 O

2.16

>10 000

[74]

ECN silica

IPA/H2 O

1.9

1150

[75]

BTESE-SiO2

IPA/H2 O

3.4

4370

[72]

BTESE-SiO2

IPA/H2 O

2.3

2500

[76]

Co—SiO2

EtOH/H2 O

1.1

2530

[77]

ECN silica

EtOH/H2 O

1.6

350

[75]

BTESE-SiO2

EtOH/H2 O

3.15

100

[72]

Silica

AcOH/H2 O

2.25

280

[78]

Silica

AcOH/H2 O

5.9

525

[79]

SiO2 —TiO2

AcOH/H2 O

2.2

2100

[79]

ECN silica

AcOH/H2 O

1.9

60

[80]

BTESE-SiO2

AcOH/H2 O

4.0

500

[81]

CVD silica

C6 H6 /C6 H12

2.2 × 10−4

113

[82]

Silica

MeOH/DMC

5.6

140

[83]

et al. evaluated silica membranes prepared by a counter-diffusion CVD for the pervaporation of 50 : 50 wt% benzene/cyclohexane at room temperature and reported a separation factor of 113 for benzene over cyclohexane [82]. Tsuru et al. demonstrated the pervaporation of methanol/dimethyl carbonate mixtures at 50 ∘ C using SiO2 , SiO2 –ZrO2 , and SiO2 –TiO2 membranes [83]. Although SiO2 –ZrO2 membranes showed a separation factor of less than 10, porous SiO2 membranes had an increased separation factor of from 10 to 160. Silica membranes with an average pore size of 0.3 nm showed the highest permselectivity of methanol with a separation factor of 140 and a total flux of 5.8 kg m−2 h−1 for a 50 : 50 mol% methanol/dimethyl carbonate mixture.

2.9 Microporous Silica-Based Membranes for Desalination Water is essential for each life and every industrial sector. One of the major challenges of the twenty-first century is water scarcity, which has a resounding impact on all levels of society from the general public to health and politics. Desalination is becoming an increasingly important tool in the fight to the global demand for clean water. Membrane technology has been successfully applied to the desalination industry. Inorganic membranes in particular for microporous silica are attractive candidates for water desalination due to the

69

70

2 Microporous Silica Membranes

advantages of their tunable pore sizes and morphology, thereby offering higher selectivity. Furthermore, interest in amorphous silica-based membranes is gaining momentum because of their simple fabrication techniques, relatively low cost, and excellent molecular sieving properties. Microporous silica membranes feature in molecular sieving structures with pore sizes on the order of the kinetic diameter of the species to be separated (dp = 3–5 Å), and therefore the membrane acts as a selective barrier between the water molecule (dk = 2.6 Å) and the hydrated salt ions (e.g. Na+ : dk = 7.2 Å and Cl− : dk = 6.6 Å), thus allowing the separation of water and salt [85, 86]. Professor J. C. Diniz da Costa from the University of Queensland (Australia) did a lot of work in developing microporous silica membranes for water desalination. His first try was initiated in 2007 during which he prepared a range of inorganic membranes (alumina and molecular sieve silica) with pore size ranging from 0.3 to 500 nm and evaluated their performances in membrane desalination. Best separation results were achieved for the silica membrane pressurized to only seven bars, exhibiting a flux of around 1.8 kg m−2 h−1 and NaCl rejection of 98% with 3.5 wt% (seawater-like) feed. The higher rejection degree for silica rather than alumina is rationalized by the fact that rejection of salt ions occurs when the feed solution pH = 7.0 is far from the isoelectric point (IEP) of the silica membrane surface (pH = 2.0) while the alumina membrane surface has a similar IEP of 8.3 to the feed solution [87]. Two years later, he prepared carbonized template silica membranes by employing a two-step sol–gel synthesis. TEOS as the silica precursor was first hydrolyzed, and then certain amount of surfactants (hexyltriethylammonium bromide, C6 ; dodecyltrimethylammonium bromide, C12 ; hexadecyltrimethylammonium bromide, C16 ) were added into the precursor. The dip-coated silica layers were calcined at vacuum in order to maintain the carbonized surfactant in the silica layer. Carbon template silica (CTS) membranes were hydrostable in contrast to normal silica membranes because water interacted with the silanol groups and enlarged the silica pore sizes. Desalination tests revealed that CTS membranes exhibited higher surface area and pore volume as well as improved salt rejection as compared with conventional silica membranes due to typical microporous structural characteristics. On the other hand, the longest 16-carbon chain (C16 ) surfactant-derived CTS membrane showed highest NaCl rejection up to 97% with flux on the order of 3 kg m−2 h−1 at 1.0 bar pressure difference across the membrane and seawater concentrations. This translated to 1200 ppm of NaCl in permeate, being close to the standard for drinking water [88]. He extended this method in 2011 by using glycol–polypropylene glycol–polyethylene glycol tri-block copolymer as a template for a hybrid carbon/silica membrane for application in seawater desalination [89]. In a recent year, he modified silica membranes by cobalt oxides (Figure 2.12). CoOx Si xerogels were synthesized via a sol–gel method including TEOS, cobalt nitrate hydrate, and peroxide. The introduction of cobalt oxide species in the sol–gel precursor minimized the nonselective pores, the consequence of which was narrowing the pore size, thus rendering the membrane more molecular selective. Further, he investigated the potential of cobalt oxide silica (CoOx Si) membranes for desalination of brackish (1.0 wt% NaCl), seawater (3.5 wt% NaCl), and brine (7.5–15 wt% NaCl)

2.9 Microporous Silica-Based Membranes for Desalination

Feed

Feed

CoOx Modification

Permeate Conventional silica membrane

CoOx

SiO2

Permeate Co-modified silica membrane

Cl–

H2O

Na+

Figure 2.12 Schematic of water and hydrated ion diffusion through the pores of conventional and co-modified silica membranes.

H O

H O

H

O

–

Cl

H

H

O

H

H

H

H

H

Na +

H

H

H O

O

H

H

H

O

O

H

H O

Figure 2.13 Schematic of the structure of the microporous organosilica membrane and water diffusion through this membrane.

O H

concentrations at feed temperatures between 25 and 75 ∘ C. The maximum flux observed was 1.8 kg m−2 h−1 at 75 ∘ C for a 1.0 wt% NaCl feed concentration. The salt rejection was consistently in excess of 99%, independent of either the testing temperature or salt feed concentration [90]. Professor Toshinori Tsuru in Hiroshima University (Japan) also did a representative work in water desalination with silica membrane, which demonstrated a microporous organosilica membrane functioning in reverse osmosis mode (Figure 2.13). The organosilica membranes derived from BTESE exhibited superior molecular sieving ability for a variety of solutes with low molecular weights, for example, 95–99% rejection of NaCl.

H

Si

C

O

O HO Si

C

Si Si

O

C Si

O

Si

Si

C

C

Si O

Si

Si

O C

O

O

C

O

O O

C

O

C

Si

Si

C

Si

Si

O

OH

Si

O

C OH

Si

C

71

72

2 Microporous Silica Membranes

Table 2.4 Performance comparison of silica-based membranes for desalination.

Membrane

C6 CTMSS MTES-SiO2 C6 CTMSS C12 CTMSS C16 CTMSS PPG-CTMSS CoOx —SiO2

Test condition

20 ∘ C, P = 7 bara) 20 ∘ C, P = 7 bara) 20 ∘ C, ΔP < 1 barc) 20 ∘ C, ΔP < 1 barc) 20 ∘ C, ΔP < 1 barc) 20 ∘ C, ΔP < 1 barc) 20 ∘ C, ΔP < 1 bar 50 ∘ C, ΔP < 1 bar 75 ∘ C, ΔP < 1 bar

Feed concentration (wt%)

Water flux (kg m−2 h−1 )

Rejection (%)

Refs.

0.3/3.5b) 0.3/3.5b)

2.1/1.9 4.7/2.5

99.9/98 93.7/83

[87]

0.3/3.5 0.3/3.5 0.3/3.5

3.2/1.4 2.8/1.6 3/2

86/92 84/94 91/97

[88]

0.3/3.5

1.5/6.3

87/99.8

[89]

0.3/15 0.3/15 0.3/15

0.4/0.3 0.9/0.35 1.8/0.55

99.7/99.9 99.5/99.9 99.5/99.9

[90]

a) Feed pressurizing up to 7 bar and permeate vacuum pumping. b) Seawater. c) Permeate vacuum pumping, resulting in a pressure difference ΔP across the membrane less than 1.0 bar.

Exceptional hydrothermal stability has been obtained due to the introduction of an inherently stable, organically bridged silica network structure, significantly broadening the application fields of the organosilica membranes [91]. For an overview, the detailed information is provided in Table 2.4.

2.10 Conclusions and Future Trends Microporous silica membranes have attracted wide attention in both academy and industry owing to their structural advantages of silica components and exceptionally small pores in size of 3–5 Å, as well as their economic merits of cheapness and facile processing procedures. In this chapter, we first reviewed the currently used membrane synthesis strategies of sol–gel, templating, and CVD. Many examples showed the powerfulness of abovementioned strategies to prepare continuous membranes. For precisely controlling the microstructures of silica-based membranes, new and cheap strategies are needed to be innovated since new synthesis strategy offers the membrane with new structure and then new functions. Therefore, extensive researches are underway to overcome the synthesis barriers and obstacles to fabricate defect-free membranes. According to the typical layout of a silica membrane (membrane layer, intermediate layer, support), we have comprehensively discussed each sector including individual plays and corresponding functions. Meanwhile, modification methods have been introduced to functionalize membranes in order to tune membrane properties and eventually to tailor the functions. Microporous silica membranes have been shown to provide excellent molecular sieving properties for gas and liquid separations. In the end, we have summarized all the relevant membrane

References

separations, represented by hydrogen separation, carbon dioxide capture, pervaporation, and desalination. The incomparable characteristics of silica membranes make them one of the promising candidates for gas separations. This chapter has highlighted the importance and broadness of microporous silica membranes for producing hydrogen and for capturing carbon dioxide from diverse sources. In parallel, various types of membranes consisting of amorphous silica, composites of silica and other oxides, and organic–inorganic hybrid silica have been developed and used in pervaporation and desalination. In contrast to gas separations, the reported uses in liquid separations are limited to pervaporation and desalination, primarily due to the sensitivity toward hydrothermal stability. For the development of new separation membranes, the flux or permeability must be increased, while the selectivity either remains unchanged or is increased. The final challenge for this research community is to establish the conditions under which separation using silica-based membranes is most technically and economically viable.

References 1 Peters, T.A., Fontalvo, J., Vorstman, M.A.G. et al. (2005). J. Membr. Sci. 248:

1–2. 2 Tsai, C.Y., Tam, S.Y., Lu, Y., and Brinker, C.J. (2000). J. Membr. Sci. 169: 3 4

5 6 7 8

9 10 11 12 13

255–268. Barboiu, C., Sala, B., Bec, S. et al. (2009). J. Membr. Sci. 326: 514–525. Kanezashi, M. and Tsuru, T. (2011). Gas permeation properties of helium, hydrogen, and polar molecules through microporous silica membranes at high temperatures: correlation with silica network structure. In: Membrane Science and Technology (ed. S.T. Oyama and M.S.W. Susan), 117–136. Amsterdam: Elsevier. Hench, L.L. and West, J.K. (1990). Chem. Rev. 90: 33–72. Witoon, T., Tatan, N., Rattanavichian, P., and Chareonpanich, M. (2011). Ceram. Int. 37: 2297–2303. Naito, M., Nakahira, K., Fukuda, Y. et al. (1997). J. Membr. Sci. 129: 263–269. Ayral, A., Julbe, A., Rouessac, V. et al. (2008). Microporous silica membrane: basic principles and recent advances. In: Membrane Science and Technology (ed. M. Reyes and M. Miguel), 33–79. Amsterdam: Elsevier. Tsuru, T. (2003). Porous ceramics for filtration. In: Handbook of Advanced Ceramics, 291–312. Oxford: Academic Press. Nair, B.N., Elferink, W.J., Keizer, K., and Verweij, H. (1996). J. Colloid Interface Sci. 178: 565–570. Uhlhorn, R.J.R., Keizer, K., and Burggraaf, A.J. (1992). J. Membr. Sci. 66: 271–287. Lin, Y.S., Kumakiri, I., Nair, B.N., and Alsyouri, H. (2002). Sep. Purif. Rev. 31: 229–379. Wang, X., Miao, X.R., Li, Z.M., and Deng, W.L. (2011). Appl. Surf. Sci. 257: 2481–2488.

73

74

2 Microporous Silica Membranes

14 Eswaramoorthy, M., Neeraj, S., and Rao, C.N.R. (1999). Microporous Meso-

porous Mater. 28: 205–210. 15 Guo, L., Dong, X., Cui, X. et al. (2009). Mater. Lett. 63: 1141–1143. 16 Raman, N.K. and Brinker, C.J. (1995). J. Membr. Sci. 105: 273–279. 17 Labropoulos, A.I., Romanos, G.E., Karanikolos, G.N. et al. (2009). Microp-

orous Mesoporous Mater. 120: 177–185. 18 Kusakabe, K., Shibao, F., Zhao, G. et al. (2003). J. Membr. Sci. 215: 321. 19 Mottern, M.L., Quickel, G.T., Shi, J.Y., et al. (2004). Processing and proper-

20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

ties of homogeneous supported γ-alumina membranes. Proceedings of the 8th International Conference on Inorganic Membranes (ed. Y.S. (Jerry) Lin), 26–29, Cincinnati, OH (18–22 July). Chicago, IL: Adams Press. Liang, J., Jiang, X., Liu, G. et al. (2003). Mater. Res. Bull. 38: 161. Shi, J.Y. and Verweij, H. (2005). Langmuir 21: 5570. Lee, D., Zhang, L., Oyama, S.T. et al. (2004). J. Membr. Sci. 231: 117. Brinker, C.J., Ward, T.L., Sehgal, R. et al. (1993). J. Membr. Sci. 77: 165. Nijmeijer, A., Huiskes, C., Sibelt, N.G.M. et al. (1998). Am. Ceram. Soc. Bull. 77: 95. Yoshino, Y., Suzuki, T., Nair, B.N. et al. (2005). J. Membr. Sci. 267: 8. Brinkman, H.W., van Eijk, J.P.G.M., Meinema, H.A., and Terpstra, R.A. (1999). Am. Ceram. Soc. Bull. 78: 51. Millan, A.J., Nieto, M.I., Moreno, R., and Baudin, C. (2002). J. Eur. Ceram. Soc. 22: 2223. Toy, C. and Whittemore, O.J. (1989). Ceram. Int. 15: 167. Yoshida, K., Hirano, Y., Fujii, H. et al. (2001). J. Chem. Eng. Jpn 34: 523. Kanezashi, M. and Asaeda, M. (2006). J. Membr. Sci. 271: 86. Boffa, V., Blank, D.H.A., and ten Elshof, J.E. (2008). J. Membr. Sci. 323: 347. Gu, Y., Hacarlioglu, P., and Oyama, S.T. (2008). J. Membr. Sci. 310: 28. Nam, S.W., Ha, H.Y., Yoon, S.P. et al. (2001). Korean Mem. J. 3: 69. Asaeda, M., Sakou, Y., Yang, J., and Shimasaki, K. (2002). J. Membr. Sci. 209: 163–175. Sekuli, J., Luiten, M.W.J., ten Elshof, J.E. et al. (2002). Desalination 148: 19. de Vos, R.M., Maier, W.F., and Verweij, H. (1999). J. Membr. Sci. 158: 277. Campaniello, J., Engelen, C.W.R., Haije, W.G. et al. (2004). Chem. Commun. 834–835. Castricum, H.L., Sah, A., Kreiter, R. et al. (2008). Chem. Commun. 1103–1105. Wei, Q., Wang, Y.L., Nie, Z.R. et al. (2008). Microporous Mesoporous Mater. 111: 97. Duke, M.C., da Costa, J.C.D., Lu, G.Q. et al. (2004). J. Membr. Sci. 241: 325. Duke, M.C., da Costa, J.C.D., Do, D.D. et al. (2006). Adv. Funct. Mater. 16: 1215. Gu, Y. and Oyama, S.T. (2007). Adv. Mater. 19: 1636. De Vos, R.M. and Verweij, H. (1998). J. Membr. Sci. 143: 37–51. Sea, B.K., Watanabe, M., Kusakabe, K. et al. (1996). Gas Sep. Purif. 10: 187–195. Hwang, G.J., Onuki, K., Shimizu, S., and Ohya, H. (1999). J. Membr. Sci. 162: 83–90.

References

46 Prabhu, A.K., Liu, A., Lovell, L.G., and Oyama, S.T. (2000). J. Membr. Sci.

177: 83–95. 47 Kim, Y.S., Kusakabe, K., Morooka, S., and Yang, S.M. (2001). Korean J. Chem.

Eng. 18: 106–112. 48 Lee, D.W., Sea, B., Lee, K.Y., and Lee, K.H. (2002). Ind. Eng. Chem. Res. 41:

3594–3600. 49 Kusakabe, K., Shibao, F., Zhao, G. et al. (2003). J. Membr. Sci. 215: 321–326. 50 Oyama, S.T., Lee, D., Hacarlioglu, P., and Saraf, R.F. (2004). J. Membr. Sci. 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73

244: 45–53. Yoshino, Y., Suzuki, T., Nair, B.N. et al. (2005). J. Membr. Sci. 267: 8–17. Kanezashi, M. and Asaeda, M. (2006). J. Membr. Sci. 271: 86–93. Gu, Y. and Oyama, S.T. (2007). J. Membr. Sci. 306: 216–227. Lee, D.W., Yu, C.Y., and Lee, K.H. (2008). J. Colloid Interface Sci. 325: 447–452. Battersby, S., Smart, S., Ladewig, B. et al. (2009). Sep. Purif. Technol. 66: 299–305. Koutsonikolas, D.E., Kaldis, S.P., Sklari, S.D. et al. (2010). Microporous Mesoporous Mater. 132: 276–281. Qi, H., Han, J., and Xu, N. (2011). J. Membr. Sci. 382: 231–237. Smart, S., Vente, J.F., and Da Costa, J.C.D. (2012). Int. J. Hydrog. Energy 37: 12700–12707. Han, H.H., Ryu, S.H., Nakao, S., and Lee, Y.T. (2013). J. Membr. Sci. 431: 72–78. Ballinger, B., Motuzas, J., Smart, S., and da Costa, J.C.D. (2014). J. Membr. Sci. 451: 185–191. Giessler, S., Jordan, L., da Costa, J.C.D., and Lu, G.Q. (2003). Sep. Purif. Technol. 32: 255. Battersby, S., Duke, M.C., Liu, S. et al. (2008). J. Membr. Sci. 316: 46–52. Brands, K., Uhlmann, D., Smart, S. et al. (2010). J. Membr. Sci. 359: 110–114. Jabbari, A., Ghasemzadeh, K., Khajavi, P. et al. (2014). Int. J. Hydrog. Energy 39: 18585–18591. Kuraoka, K., Kubo, N., and Yazawa, T. (2000). J. Sol-Gel Sci. Technol. 19: 1–3. Xomeritakis, G., Tsai, C.Y., and Brinker, C.J. (2005). Sep. Purif. Technol. 42: 249–257. Yoshioka, T., Nakanishi, E., Tsuru, T., and Asaeda, M. (2001). AIChE J. 9: 2052–2063. Lee, Y.E., Kang, B.S., Hyun, S.H., and Lee, C.H. (2005). Sep. Sci. Technol. 39: 3541–3557. Boffa, V., ten Elshof, J.E., Petukhov, A.V., and Blank, D.H.A. (2008). ChemSusChem 1: 437–443. van Gemert, R.W. and Cuperus, F.P. (1995). J. Membr. Sci. 105: 287–291. Asaeda, M., Yang, J., and Sakou, Y. (2002). J. Chem. Eng. Jpn 35: 365–371. Wang, J., Kanezashi, M., Yoshioka, T., and Tsuru, T. (2012). J. Membr. Sci. 415–416: 816–820. Nagasawa, H., Matsuda, N., Kanezashi, M. et al. (2016). J. Membr. Sci. 498: 336–344.

75

76

2 Microporous Silica Membranes

74 Yang, J., Yoshioka, T., Tsuru, T., and Asaeda, M. (2006). J. Membr. Sci. 284:

205–213. 75 van Veenm, H.M., van Delft, Y.C., Engelen, C.W.R., and Pex, P. (2001). Sep. 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91

Purif. Technol. 22–23: 361–366. Gong, G., Wang, J., Nagasawa, H. et al. (2014). J. Membr. Sci. 472: 19–28. Wang, J. and Tsuru, T. (2011). J. Membr. Sci. 369: 13–19. Kitao, S. and Asaeda, M. (1990). J. Chem. Eng. Jpn 23: 367–370. Asaeda, M., Ishida, M., and Waki, T. (2005). J. Chem. Eng. Jpn 38: 336–343. Sommer, S. and Melin, T. (2005). Chem. Eng. Process. 44: 1138–1159. Tsuru, T., Shibata, T., Wang, J. et al. (2012). J. Membr. Sci. 421–422: 25–31. Matsuyama, E., Kimura, S., Monma, K. et al. (2013). Kagaku Kogaku Ronbunshu 39: 98–103. Tsuru, T., Sasaki, A., Kanezashi, M., and Yoshioka, T. (2011). AIChE J. 57: 2079–2089. Castricum, H.L., Sah, A., Kreiter, R. et al. (2008). J. Mater. Chem. 18: 2150–2158. Li, L.X., Dong, J.H., Nenoff, T.M., and Lee, R. (2004). Desalination 170: 309–316. Lin, J. and Murad, S. (2001). Mol. Phys. 99: 1175–1181. Duke, M.C., Mee, S., and da Costa, J.C.D. (2007). Water Res. 41: 3998–4004. Wijaya, S., Duke, M.C., and da Costa, J.C.D. (2009). Desalination 236: 291–298. Ladewig, B.P., Tan, Y.H., Lin, C.X.C. et al. (2011). Materials 4: 845–856. Lin, C.X.C., Ding, L.P., Smart, S., and da Costa, J.C.D. (2012). J. Colloid Interface Sci. 368: 70–76. Xu, R., Wang, J.H., Kanezashi, M. et al. (2011). Langmuir 27: 13996–13999.

77

3 Carbon-Based Membranes 3.1 Introduction Carbons are promising building blocks with unique structures and tunable physicochemical properties suitable for membrane-based separations [1]. The carbon materials suitable for membrane separations are microporous carbons, which can be categorized into carbon molecular sieve (CMS), carbon nanotube (CNT), and graphene. CMS are amorphous materials with average pore dimensions similar to the critical dimensions of small molecules (Figure 3.1a). They are prepared by the pyrolysis of natural and synthetic precursors, including coal, coconut shells, pitch, phenol-formaldehyde resin, polyfurfuryl alcohol (PFA), polyacrylonitrile (PAN), and polyvinylidene chloride. Properties of the CMS are quite variable, spanning pore sizes from 3.0 to 12.0 Å, pore volumes from 0.2 to 0.5 cm3 g−1 , and surface areas from 300 to 1500 m2 g−1 . Physicochemical properties of a particular carbon are controlled by the choice of precursor, additives or pore formers, and pyrolysis conditions [1]. The pore structure in molecular sieving carbon is typically ascribed to arise from at least three different mechanisms of precursor in shrunken form, fissures and cracks in carbon matrix, and volatilization of small molecules [1]. Graphene is an allotrope of carbon in the form of a two-dimensional (2D) atomic-scale hexagonal lattice in which one atom forms each vertex (Figure 3.1b) [2]. It can be considered as an indefinitely large aromatic molecule, the ultimate case of the family of flat polycyclic aromatic hydrocarbons. Graphene has many unusual properties. It is about 200 times stronger than the strongest steel. It efficiently conducts heat and electricity and is nearly transparent [3, 4]. Graphene shows a large and nonlinear diamagnetism [4], greater than graphite, and can be levitated by neodymium magnets. Graphene has a theoretical specific surface area (SSA) of approximately 2600 m2 g−1 . This is much larger than that reported to date for carbon black (typically smaller than 900 m2 g−1 ) and is similar to activated carbon [5]. The basic structural element of allotropes includes pristine graphene, graphene oxide (GO), and reduced graphene oxide (rGO). CNTs are allotropes of carbon with a cylindrical nanostructure (Figure 3.1c). Nanotubes are members of the fullerene structural family. Their long and hollow structures with the walls are formed by one-atom-thick sheets of carbon at specific and discrete angles. The combination of the rolling angle and radius decides the nanotube properties. Nanotubes are Microporous Materials for Separation Membranes, First Edition. Xiaoqin Zou and Guangshan Zhu. © 2019 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2019 by Wiley-VCH Verlag GmbH & Co. KGaA.

78

3 Carbon-Based Membranes

Pore

Carbon molecular sieve

Graphene

Carbon nanotube

Figure 3.1 The basic structures of (a) carbon molecular sieve, (b) graphene, and (c) carbon nanotube.

categorized as single-walled nanotubes (SWNTs) and multiwalled nanotubes (MWNTs). Individual nanotubes naturally align themselves into “ropes” held together by van der Waals forces, more specifically π-stacking. These cylindrical carbon molecules have unusual properties that are valuable for nanotechnology, electronics, optics, and other fields of materials science and technology [6].

3.2 Carbon Membrane Preparations 3.2.1

Carbon Molecular Sieve Membranes

Carbon membranes can be produced through the carbonization or pyrolysis process of suitable precursors [7]. Typically, a precursor membrane is pyrolyzed under inert atmosphere (usually N2 ) or vacuum by slowly increasing the temperature, in most cases up to 773–1073 K, holding at that temperature for a certain thermal soak time, and then cooling down (Figure 3.2). The pyrolysis gases evolved leave behind pores, and at the same time the precursor forms the rigid turbostratic structures, along with a shrinking of the precursor membrane to some extent. For an example, Matrimid (a typical polyimide [PI] precursor) is given here for helping us to understand the basics of the carbonization process. Under pyrolysis in vacuum, the only species released up to 673 K is water. Large species or molecules (50–240 Da) are only released from 673 to 873 K, and this gives rise to the formation of micropores. From 823 K, significant amounts of methane and carbon oxides are given off due to the degradation of imide and benzene rings. The decomposition rigidifies the structure of the precursor, and the release of small molecules reasonably gives rise to the formation of the ultramicropores. From 973 K upward, H2 release becomes significant and leads to the formation of an amorphous carbon material [8] and eventually to graphitic structures at even higher temperatures. The micropores formed during the low temperature pyrolysis of Matrimid (673–873 K) shrink at higher temperatures [9]. In the membrane preparation process, several parameters are important in terms of finely tuning the structure and the performance of carbon membranes: the nature of the precursor polymer, the pyrolysis temperature and atmosphere, and the pretreatment or posttreatment. In the following sections, the effects of these parameters on the performances of carbon membranes are described consecutively.

3.2 Carbon Membrane Preparations

Pyrolysis Carbonization

Polymer precursor

Carbon molecular sieve

(a) Pressure transmitter End caps

Diaphragm valve Alumina tube

Furnace

Cold trap

LN2

Temperature controller

Particle filter Vacuum pump

(b)

Figure 3.2 (a) The production of carbon molecular sieve membranes through carbonization or pyrolysis using polymer precursors and (b) process diagram of the carbonization setup.

3.2.1.1

Precursors of Carbon Molecular Sieves

The precursors can be classified into two catalogues of natural and synthetic ones. Naturally carbon-rich materials such as resins, coal, pitch, and plants are employed for the preparation of carbon membranes at the early stage. Recently, numerous synthetic precursors have been used to form carbon membranes. The most common thermosetting polymers used to date for the preparation of CMS membranes are PAN, PIs, phenolic resins (PRs), and PFA [10, 11]. The first CMS hollow fibers, however, were produced by Koresh and Soffer from cellulose [12], which is particularly convenient for its low cost when compared with PIs. PIs are considered the best precursors because of the separation performance and mechanical properties of the resulting CMS membranes [13]. PFA and PRs are other cheap alternatives to PIs, but since they are liquids, they need to be coated on supports before the pyrolysis step. The other polymers are preformed as self-standing membranes, in flat configurations, or as hollow fibers and pyrolyzed as such, maintaining their shape and morphology.

79

80

3 Carbon-Based Membranes

One strategy for the improvement of the productivity of CMS membranes is the introduction in the precursor membrane of porogens, i.e. species that serve as templates for the formation of microporosity in carbon. This can be accomplished by blending with labile polymers (e.g. polyvinylpyrrolidinone [PVP]), which completely gasify during the pyrolysis step, by copolymerization with labile monomers, or by the introduction of thermally labile groups (e.g. —SO3 H, —SO3 − , alkyl) or ions (NO3 − ), which can be decomposed. During the pyrolysis of sulfonated PIs [14], the sulfur-containing groups have already decomposed at 723 K; when instead the loss of a sulfone moiety requires the breakage of two covalent bonds, the maximum peaks for the release of sulfur oxides are observed at or below 773 K [15]. In sulfonated poly(2,6-dimethylphenylene oxide) (PPO), the loss of the sulfonic acid group takes place at around 473 K [13]. PVP has been used as a porogen in blends with PIs [16], PPO [17, 18], and cellulose [19]. Silica-containing CMS membranes prepared by pyrolysis of poly(imide-co-dimethylsiloxane) on alumina supports demonstrated good separation factors for He/N2 , H2 /N2 , CO2 /N2 , and O2 /N2 pairs [20, 21]. Larger polydimethylsiloxane blocks in the copolymer caused an increase in permeability at the expense of selectivity [22]. Another strategy for the improvement of the productivity of CMS membranes is the dispersion of cations in the precursor that catalyze the thermal decomposition of the polymer. Trivalent cations (mainly Fe3+ ) introduced as nitrates proved the most effective species for the improvement of CO2 /CH4 and O2 /N2 separation factors of cellulose-derived carbon membranes, and MgO increased the H2 /CO2 selectivity by reducing CO2 permeability [23]. Chung and coworkers produced flat CMS membranes by cross-linking a polyetherketone with 2,6-bis(4-azidobenzylidene)-4-methylcyclohexanone and pyrolyzing at relatively low temperatures (723–923 K) [24, 25]. The decomposition of the bis-azide blended with the polymer produced extremely reactive nitrene intermediates, which were able to bind to the surrounding chains and to stabilize the thermally labile polyetherketone. The membranes prepared at 823 K exhibited excellent propene/propane ideal selectivity of 44 with a propane permeability of 48 Barrer, which decreased in real mixtures to 32 and 3.6 Barrer, respectively [24]. Similar membranes prepared at 1073 K displayed a CO2 permeability of 280 Barrer and ideal CO2 /CH4 selectivity of 164 [25]. 3.2.1.2

Pyrolysis Environment for CMS Membranes

The role of the pyrolysis atmosphere has been studied by several researchers. Vacuum pyrolysis of PIs yields more selective but less permeable membranes than inert gas purge pyrolysis [8, 26, 27], and this effect has been explained with an increased heat and mass transfer from the gas phase, which accelerates the carbonization reaction and produces a more open porous matrix. At low flow rates of inert sweep gas, the nonvolatile by-products are not removed quickly enough, and they can presumably degrade further, to leave carbon deposits on the surface, which may explain the sharp reduction of flux observed in these cases [26]. The role of small amounts of oxygen in the inert sweep gas has been investigated by Koros and coworkers [28, 29]. During the pyrolysis of Matrimid membranes, oxygen impurities (3–100 ppm in Ar) reduce both the permeance

3.2 Carbon Membrane Preparations

and the CO2 /CH4 selectivity of the resulting CMS membranes. Instead for a 6FDA/BPDA-DAM PI, the increase of O2 concentration in the Ar sweep from 4 to 30 ppm brings about a reduction in flux but with an increase in CO2 /CH4 selectivity, whereas 50 ppm O2 provokes a reduction of flux and selectivity. According to the authors, the oxidation at ppm levels of O2 takes place preferentially at the more reactive sp2 carbon edges of the selective pores, and as a consequence, the size of the ultramicropores is reduced. The oxidation process is controlled by the oxygen concentration in the inert atmosphere [15]. This oxidative process represents a powerful tool for finely tuning the pore size and the selective mass transport of CMS membranes. 3.2.1.3

Pretreatments of CMS Membranes

The requirement of a precursor polymer for making self-supported CMS membranes is that it does not melt before the pyrolysis. The plasticization of the polymer at temperatures higher than the glass transition temperature provokes the thickening of the membrane skin, with the closure of the finer pores present in the porous sublayer. This phenomenon can be exploited to prepare defect-free CMS membranes from defective or even porous precursor membranes [30, 31], but in most cases the densification provokes the formation of a very thick skin layer, which, in turn, causes unacceptable reductions in flux [32]. In order to avoid or minimize this effect, the most common pretreatments are the pre-oxidation in air at 373–673 K and the chemical cross-linking of the precursor membranes. Okamoto and coworkers stabilized asymmetric PI hollow fibers at 673 K in air for 30 minutes before pyrolysis in nitrogen at 873–903 K, obtaining thin skin layers (200 nm) in the resulting CMS membranes that had higher permeances and sustained selectivities for C3 H6 /C3 H8 and 1,3-butadiene/n-butane pairs [33]. The pre-oxidation stage at 673 K was accompanied by the decomposition of the sulfone group contained on the aromatic diamine monomer, giving rise to sulfur oxides in the gas phase. Yoshimune and Haraya pre-oxidized sulfonated poly(phenylene oxide) hollow fibers in air at 533 K for two hours before pyrolysis under vacuum, obtaining very thin (280 nm) and flexible CMS hollow fibers with high CO2 /CH4 and H2 /CH4 separation factors (118 and 392) and CO2 permeance in excess of 10−8 mol m−2 s−1 Pa−1 [34]. Also in this case, the sulfur-containing groups decomposed almost completely during the pre-oxidation stage. Another interesting pretreatment that can improve the selectivity of PI P84-derived CMS membranes is soaking the precursor membrane in a linear alcohol (the best one was given by ethanol) prior to pyrolysis [35]. The experimental evidence indicates that weakening intramolecular cohesion forces caused by the sorption of ethanol allows a more effective structural reorganization of the polymer chains leading to narrower pores. 3.2.1.4

Posttreatment of CMS Membranes

It is not clear whether it is more appropriate to define the synthetic method used by carbon membranes as a posttreatment of CMS membranes or as a real synthesis of a CMS membrane on a carbon porous support. The method introduced by Soffer and coworkers consisted of the pyrolytic chemical vapor deposition (CVD) of carbon in the lumen side of the porous carbon hollow fibers

81

82

3 Carbon-Based Membranes

derived from cellulose by using propene as the carbon source [36]. The CVD plugged the pores of the support but could also create a carbon layer on the surface of the membrane. The CVD was followed by a so-called activation step that consisted of a high temperature oxidation, which created a tailored pore structure within the CVD layer [37]. A similar method (propene or methane CVD and air oxidation at 523–623 K) was adopted by Blue Membranes GmbH for their honeycomb carbon membranes [38]. Pyrolytic carbon deposition of benzene had been also used for the enhancement of the sorption selectivity of activated carbon fibers (ACFs) for pressure swing adsorption (PSA) [39–41]. At a deposition temperature of 1173 K, benzene already pyrolyzed in the gas phase to form larger molecules that could not enter the micropores so that carbon deposited mainly on the external surface of the fibers and produced fibers with limited sorption capacity and poor CO2 /CH4 sorption selectivity. At 1000 K instead, the experimental evidence indicated that benzene could be adsorbed as such into the micropores before pyrolysis, restricting the size of the micropores: as a consequence, the sorption amount of CO2 reduced slightly with respect to the pristine ACFs, whereas a very sharp reduction of CH4 was observed, with high CO2 /CH4 sorption selectivity [40]. Morooka and coworkers used propene CVD at 923 K on their supported CMS membranes derived from PI to reduce the pore size and to increase O2 /N2 and CO2 /N2 selectivities. It turned out that an optimum CVD time was two minutes, because at longer times both permeances and selectivities were reduced [42]. Another popular posttreatment method is oxidation. Morooka et al. treated their carbon membranes with 10 and 20% O2 for three hours at 573 K, obtaining a generalized increase of permeance and almost no change in selectivity. The treatment in air at 373 K for 30 days led to a decrease in permeability with an increase in selectivity; since the former performance could be recovered after heating in N2 at 873 K, the authors raised the hypothesis that oxidation at 373 K produced surface oxides that reduced the size of the ultramicropores [43]. In both cases, the change in elemental composition of the membranes caused by the oxidation at 373–573 K was modest, suggesting an increase in the micropore volume of the membrane, with almost no change in the pore size distribution [44]. The post-oxidation of PPO-derived CMS membranes supported on alumina (air at 373–673 K) instead produced a broadening pore size distribution and a drop in all selectivity values [45–47]. BPDA-ODA PI- and PEI-derived CMS membranes were coated with PPO and pyrolyzed again at 873 K [48]. The second pyrolysis caused the expansion of the pores of the pristine CMS membranes, and at the same time the decomposition of the thermally labile PPO produced the deposition of carbon at the mouth of the pores. As a result, more permeable and more selective membranes were formed. In particular, the PI-derived membrane showed outstanding CO2 and H2 permeabilities (1320 and 1450 Barrer, respectively) coupled with very high selectivities for CO2 /N2 (156), CO2 /CH4 (157), and H2 /CH4 (172). 3.2.2

Module Construction of Carbon Membranes

The geometry and installation of a membrane in a suitable device (i.e. a module) are important to its separation ability [49]. For instance, the brittleness of CMS

3.2 Carbon Membrane Preparations

(a)

(b) Feed

Retentate

(c)

Permeate

Figure 3.3 SEM pictures of a hollow fiber manufactured by Carbon Membranes Ltd. (left) [37]. Source: Copyright 2004, reproduced with permission from Elsevier B.V. Scheme for the preparation of the Blue Membranes GmbH honeycomb CMSM module (right) [38]. Source: Copyright 2005, reproduced with permission from Elsevier B.V.

membranes is a major problem for the preparation and durability of a module. The selection of a membrane module is mainly determined by economic considerations, including all the cost factors plus the cost of the module. For commercial applications of membranes, it is preferable to fabricate a module with an asymmetric structure and capillary or hollow fiber configurations in order to increase the rate of permeation of the products [50]. Carbon Membranes Ltd. (Israel) was the first company to produce high-quality hollow fiber membrane modules on an industrial scale [36]. The hollow fibers had a length of 1.0 m, a diameter of 170 μm, and a thickness of 9 μm (Figure 3.3) [37]. The packing density was about 2000 m2 m−3 and each module contained 4.0 m2 . Carbon Membranes Ltd. started the commercialization of its membranes and modules at the end of the 1990s, but ceased its activity in 2001. Blue Membranes GmbH, based in Wiesbaden, Germany, was the second company to develop and construct a CMS membrane module from a quite different concept, in which the membrane and the module are produced at the same time [38]. Bulky ceramic supports severely limit the maximum membrane packing density of the obtained CMS membrane, interfere with the shrinking of the pyrolyzing material, and introduce mechanical stresses that reduce the resistance of the membrane. In this regard, the choice of a sheet of paper

83

84

3 Carbon-Based Membranes

reinforced with ceramic fibers is able to reduce these risks greatly. In more detail, the paper was first impregnated with a mixture of PR and epoxy glue dissolved in 2-butanone; then a diagonal pattern was stamped on the dry composite, so the direction of the wavy motif had a diagonal orientation with respect to the sheet side (Figure 3.3a, right). The rippled sheet was then pleated (Figure 3.3b, right), creating flow channels for both feed and permeate. Next, alternate pairs of folds were sealed at their edges, as shown in Figure 3.3c (right), isolating in this way the feed from the permeate side. The pockets resulted at this stage of the preparation process resemble the pockets to be glued on the central permeate collector tube in a spiral wound module; in this case, however, no spacer was needed to separate the facing membranes within the pockets. After curing at moderate temperature to bond the points of contact of the sheets, the nascent membranes were pyrolyzed under nitrogen up to 1053 K and then underwent a CVD treatment followed by air oxidation to finely tune the gas separation properties. The structure of the membranes, 50–100 μm thick, comprised a macroporous support with a separating layer of 10–30 μm. The resulting membranes had a high packing density up to 2500 m2 m−3 and were housed in modules containing up to 10.0 m2 of membrane area.

3.3 Selective Surface Flow Membranes Rao and Sircar introduced new membranes in 1993 for the separation of gas mixtures that they called selective surface flow (SSF) membranes [51, 52]. These membranes were synthesized by coating a macroporous graphite disk with layers of polyvinylidene chloride–acrylate terpolymer latex, which contained 0.1 mm polymer beads in aqueous solution. After deposition of each layer, the disk is dried and heated to 1000 ∘ C, which produces a porous carbon membrane through sequential cross-linking and carbonization of the underlying polymer. The resulting thickness of each layer was approximately 0.5 mm, and two to five layers were applied in order to yield a complete module between 1.0 and 2.5 mm. The permeability for H2 in a mixture with hydrocarbons was reduced by several orders of magnitude over that of pure hydrogen. These membranes became promising for H2 separation because the hydrocarbon selective adsorption hindered pore diffusion by hydrogen. Further fine-tuning of the pore structure can be facilitated through various synthetic methods. They include an increase in the oxidation time and temperature, allowing for controlled increase of the pore size and permeabilities of all components, and a variation in the kinetic selectivity [53]. Excessive oxidation, however, may render the pores too large to be selective. Optimization of these membranes has led to even further gas separation advantages, such as hydrogen gas purification. However, this process works oppositely to other inorganic membranes. Since adsorption occurs on the high-pressure side, the partial pressure of the component to be adsorbed can be low. The partial

3.3 Selective Surface Flow Membranes

pressure gradient across the membrane does not need to be high to attain separation, since the driving force for mass transfer across the membrane is the difference in the concentration of the adsorbed species (i.e. concentration gradient). It should also be noted that the activation energy for surface diffusion is typically lower than that for transport across the membrane. Furthermore, adsorption decreases the effective pore volume, hindering the Knudsen diffusion mechanism of non-adsorbed molecules, which would ultimately diminish the degree of separation. Separation processes utilizing SSF membranes are based on the adsorption properties of the components, and larger or more polar species can be separated from the mixture. For example, in the methane reforming process, hydrogen remains on the high-pressure side of the membrane while unwanted species are passed through, eliminating subsequent compression of the H2 gas for many applications. A further advantage of the technique lies in the fact that adsorption capacity and selectivity increase with decreasing temperature, reducing operational cost. This is the reverse of molecular sieving [52]. Furthermore, the PSA process has been integrated with SSF membranes for enhanced performance in the extraction of hydrogen from steam-methane reformer gas [54]. In the commercial production of hydrogen, this reformer gas is subjected to water–gas shift (WGS) reactions followed by hydrogen purification by PSA. Typical PSA cycles consist of alternating pressurization and depressurization of feed gas and hydrogen enriched gas to augment hydrogen recovery. If an SSF membrane is used for purification of the waste gas of the PSA process, hydrogen recovery can be increased from 78 to 85% in the integrated process. Due to a reduction in the compression duty and the membrane area, the process is particularly economic when the PSA waste gas is first fractionated, and only the hydrogen-rich portion is used as feed for the SSF membrane. Viera-Linhares and Seaton used molecular dynamics (MD) calculations and critical path analysis to model the separation process in SSF membranes for methane/hydrogen mixtures [53, 55]. They showed that pore width is critical for the separation process, since it controls adsorption capacity and transport properties through the material. They defined three distinct regions, characterized by pore size. If the size is smaller than 6 Å, a sieving effect results and separation occurs solely based on molecular size. Between 6 and 10 Å, selective adsorption of methane occurs with very little dependence on pressure, due to the pores being filled almost to capacity. Maximum selectivity is achieved at 7–8 Å, while permeability is optimal at 9 Å. Larger pore sizes give rise to a regime in which methane is preferentially adsorbed on the pore walls and hydrogen occupies the center of the pore. It is less effective for H2 purification, as the hydrogen can diffuse through the low-density region of the porous network. Further research showed that for the separation of methane from hydrogen [55], both species passed through distinct pore subnetworks of the membranes, with methane populating the larger pores. The importance of this work lies in the fact that a selective blockage of smaller pores would reduce the permeability to hydrogen and enhance the effectiveness of the separation.

85

86

3 Carbon-Based Membranes

3.4 Advantages and Disadvantages of Carbon Membranes 3.4.1 Advantages of Carbon Membrane Versus Conventional Polymer Membrane 1. Carbon membranes display superior permeability–selectivity combination than polymer membranes [7, 13, 56–58]. 2. Carbon membranes are mechanically much stronger and can withstand higher pressure differences for a given wall thickness [56]. Carbon membranes have higher elastic modulus and lower breaking elongation than the polymer membranes [58]. 3. The permeation properties of carbon membranes are hardly affected by the feed pressure, because carbon membranes do not possess compaction and swelling problems [59]. 4. The permeation properties of carbon membranes will not be time dependent [58]. 5. The activation energies required for the diffusion in the carbon membranes are smaller than those in the polymer membranes. It means that the diffusing gas (especially with large molecule size) is much influenced by the activation energies when it diffuses in the polymer membranes compared with carbon membranes. Therefore, the selectivity of polymer membranes decreases remarkably if the measurement temperature increases [60]. This situation will not happen in the carbon membranes. 6. Carbon membranes offer the advantage of operation in environments prohibitive to polymeric materials and have superior stability in the presence of organic vapor or solvent and non-oxidizing acid or base environments. They can perform very well with high purity and dry feeds. They are ideal for corrosive applications [61]. 7. Carbon membranes are far more stable thermally than polymer membranes. They are suitable to be used in the separation processes with high temperature in the range of 500–900 ∘ C. On the other hand, organic polymer membranes cannot resist very high temperature and begin to decompose or react with certain components. 8. The same starting material can be used to develop membranes of different permeation properties for different gas mixtures. 9. The pore dimension and distribution of the carbon membranes can be finely adjusted by simple thermochemical treatment to meet different separation needs and objectives. 10. Carbon membranes have superior adsorption ability toward some specific gases, which can enhance its gas separation capacity. 11. Carbon membranes have the ability to be back flushed, steam sterilized, or autoclaved.

3.5 Carbon Nanotubes

3.4.2

Disadvantages of Carbon Membranes

1. Carbon membrane is very brittle and fragile. Therefore, it requires more careful handling [7, 56, 62]. This may be avoided to a certain degree by optimizing precursors and preparation methods. Therefore, it is difficult to process and expensive to fabricate carbon membranes. 2. Carbon membranes require a pre-purifier for removing traces of strongly adsorbing vapors, which can clog up the pores due to the transport through a pore system rather than through the bulk system. This is typical of many industrial adsorption separators. This problem may be avoided by operating at sufficiently high temperatures [56]. 3. Carbon membranes only demonstrate high selectivities for certain gas mixture, which limit to gases with molecular sizes smaller than 4.0–4.5 Å. Carbon membranes are not suitable to separate gas mixtures, such as iso-butane/n-butane and gas–vapor mixtures, for instance, air/hydrocarbons and H2 /hydrocarbons [63].

3.5 Carbon Nanotubes 3.5.1

Types of CNT Membranes

CNT membranes can be classified into two major categories according to the fabrication methods: (i) vertically aligned CNT (VA-CNT) membranes and (ii) mixed or composite CNT membranes. Their specific features are summarized in Table 3.1. For VA-CNT membranes, nanotubes are arranged straight up and perpendicular to the membrane surface. In this configuration, nanotubes are bound to each other by organic or inorganic filler materials. On the other hand, mixed CNT membranes have a structure of the top layer mixed with nanotubes and a polymer Table 3.1 Features of vertically aligned (VA)-CNT membranes and mixed (composite) CNT membranes. VA-CNT membranes

Mixed CNT membranes

• CNTs are aligned vertically

• CNTs are mixed with polymers

• CNTs’ forest is compacted densely

• Membranes are supported by transition layers and non-woven substrates

• Water flux is supposed to be very fast

• Water flux is moderately fast

• Functional group can be conveniently attached at the tip of CNTs or on the membrane surface

• Low or anti-membrane fouling

• Fabrication procedures are complicated

• Fabrication processes are quite simple

• Specially adjusted operating system may be needed

• Operationally feasible to conventional membrane processes

87

88

3 Carbon-Based Membranes

(a)

CNT

Top layer (polyamide ∼0.2 μm) Transition layer (polysulfone ∼45 μm)

Support layer (non-woven fabric ∼100 μm) (b)

Figure 3.4 Conceptual illustration for the structures of CNT membranes: (a) vertically aligned (VA) CNT membranes and (b) mixed CNT membranes.

such as polyamide (PA). Conceptual images of both CNT membranes are shown in Figure 3.4. The VA-CNT membranes are synthesized by arranging perpendicular CNTs with supportive filler contents between the tubes. These membranes are high-density molecular sieves with intercalated filler matrix such as polymer between them (Figure 3.4a). The fillers may be epoxy, silicon nitride, and others with no water permeability. The VA-CNT membrane was first introduced by Hinds et al. with polystyrene as filler material between CNTs [64]. The fabrication procedure was simple, but the pore sizes were irregular. The membrane could not retain Ru(NH3 )6 3+ ions initially with plasma and chemical treatments. However, functionalization of CNT core with negatively charged carboxylate groups trapped the positively charged Ru(NH3 )6 3+ ions. Biotin and streptavidin attachment onto the functionalized CNT membranes reduced ion transport by 5.5–15 times. Such functionalized membranes worked as gatekeeper-controlled chemical separators or an ion-channel mimetic sensor. Holt et al. introduced microelectromechanical method for synthesizing void-free VA-CNT membranes with sub-2 nm nanotube pores [65]. Silicon nitride (Si3 N4 ) fillers were incorporated between the nanotube spaces to inhibit any mass flow between nanotube gaps. The flux measured for air flow through the membranes exceeded

3.5 Carbon Nanotubes

the flux predicted by the Knudsen model by at least one to two orders of magnitude. They interpreted the flow enhancement as the consequence of the intrinsic smoothness of the nanotube surface. Further, they investigated the nanofluidic function by water transport though the membrane. The water flux was increased by greater than threefold over other no-slip and hydrodynamics flow. Compared with MWCNT membranes, enhanced ion selectivity was observed on the VA-CNT membrane. The membrane transported Ru2+ (bipyr)3 species with sizes up to 1.3 nm but blocked 2.0 nm Au particles, suggesting that their pore sizes were between 1.3 and 2.0 nm. High selectivity to multiple variants, high water fluxing, and low energy consumption were obtained using the membranes. The synthesis of homogeneous CNT membranes is challenging [66]. The CVD is probably the best method to synthesize VA-CNT membranes [67]. The use of catalysts in the CVD method makes uniform CNT membranes of 20–50 nm in diameter and 5–10 μm in length [68]. The details for their process to fabricate VA-CNT membranes using CVD method are illustrated in Figure 3.5. On the other hand, a composite CNT membrane (also called mixed matrix membrane [MMM]) consists of several layers of polymers or other composite materials (Figure 3.4b). Such membrane can be easily fabricated with reduced cost and was introduced by Zimmerman et al. to overcome the disadvantages of polymer membranes for gas purification [69]. However, this membrane

Si substrate

Fe Catalyst@Si

Electron beam

O2 oxidation

FeOx@Si

Fe surface

Evaporation

Depositing SiNx

Fe@CNT@Si

C2H4

CVD

H2

Fe@CNT@SiNx@Si

Fe@Si

850 °C Reactive ion etching (RIE)

Open backside of Si substrate

Etching

RIE removing

Exposed Si

Excess SiNx

Fe@CNT@SiNx@Si

Fe@CNT@SiNx@Si

Fe@CNT@SiNx@Si Exposing

Bottom Fe

Oxidation

Fe removal

CNTs removal

SiNx membrane

Pure CNT membrane

Fe@CNT@SiNx@Si

Figure 3.5 Process flow for the fabrication of VA-CNT membranes using the CVD method.

89

90

3 Carbon-Based Membranes

has significantly strengthened the water purification ability of the existing membranes. Membrane fouling and pollutant precipitation are major problems in separation technology. Irregular pore size, deleterious micropollutants, influent water quality, and pH variations decrease membrane capacities. Precipitation takes place in membranes when water pollutants exceed their solubility limit in water. Additionally, fouling creates defects in membrane pores, causing pore blocking and complicating membrane regeneration. Composite CNT membranes could be a great solution to overcome the bottlenecks of current separation technologies. Choi et al. have used composite CNT membranes to improve filtration capacity of ultrafiltration (UF) membranes [70]. It increases water permeability, solute retention, and mechanical robustness of the membrane [71]. The details about this type of membranes will be discussed in Section 3.7. 3.5.2

CNT Functionalization

CNT functionalization is often a precondition for CNT-based membrane separation in particular for water treatment or purification. CNT membranes can be functionalized at tube mouth (called tip functionalization) and on the body (called core functionalization). As shown in Figure 3.6, tip-functionalized CNT membranes have selective functional groups on the nanotube mouth, and the core-functionalized CNT membranes have functionalities at the sidewall or interior core. The end caps of nanotubes tend to be composed of highly curved fullerene-like hemispheres, which are therefore highly reactive, as compared with the sidewalls. The sidewalls themselves contain defect sites such Figure 3.6 Functionalization fashions of CNT membranes.

Pristine CNT Tube mouth Tip functionalized CNT Tube body

Core functionalized CNT Functional groups (positive, negative, hydrophilic, ...)

3.5 Carbon Nanotubes

as pentagon–heptagon pairs called Stone–Wales defects, sp3 -hybridized defects, and vacancies in the nanotube lattice. There are different approaches for functionalizing CNT membranes including covalent functionalization and non-covalent functionalization. Covalent functionalization is quite popular for initial investigation of CNT functionalization because the CNT sidewalls are expected to be inert. Mickelson et al. and Bianco et al. developed a fluorination method [72, 73]. The fluorinated CNTs have C—F bonds that are weaker than those in alkyl fluorides, thus providing substitution sites for additional functionalization [74]. Successful replacements of the fluorine atoms by amino, alkyl, and hydroxyl groups have been achieved [75]. Other methods including cycloaddition such as Diels–Alder reaction, carbene and nitrene addition, chlorination, bromination, hydrogenation, and azomethine ylides have also been successfully employed [76–78]. Baek et al. reported direct Friedel–Crafts acylation technique to covalently functionalize CNTs under a mild condition [79]. The CNTs functionalized by this strategy were highly dispersed in solvents, which allowed improved chemical affinity with matrices. Treatment of CNTs with strong acid such as HNO3 , H2 SO4 , or a mixture of them [80] or with strong oxidants such as KMnO4 [81], ozone [82], or reactive plasma [83] tends to open these tubes and to subsequently generate oxygenated functional groups such as carboxylic acid, ketone, alcohol, and ester groups that serve to tether many different types of chemical moieties onto the ends and defect sites of these tubes. These functional groups have rich chemistry, and the CNTs can be used as precursors for further chemical reactions, such as silanation, polymer grafting, esterification, and thiolation, and even some biomolecules [84–87]. The CNTs functionalized by the covalent method have a good advantage of solubility in various organic solvents because the CNTs possess many functional groups such as polar or nonpolar groups. Non-covalent functionalization is another feasible approach because it does not destroy the conjugated system of the CNT sidewalls. The CNTs are functionalized non-covalently by aromatic compounds, surfactants, and polymers, employing π–π stacking or hydrophobic interactions for the most part. The non-covalent modifications of CNTs can do much to preserve their desired properties while improving their solubility quite remarkably. Aromatic molecules, such as pyrene, porphyrin, and their derivatives, can interact with the sidewalls of CNTs by means of π–π stacking interactions, thus opening up the way for the non-covalent functionalization of CNTs. Dai and coworkers reported a general and attractive approach to the non-covalent functionalization of CNT sidewalls and the subsequent immobilization of biological molecules onto CNTs with a high degree of control and specificity [88]. Hecht et al. fabricated CNTs/FET devices functionalized non-covalently with a zinc porphyrin derivative, which was used to detect directly a photoinduced electron transferring within the zinc porphyrin derivative-CNT system [89]. Hu et al. prepared CdSe–CNT hybrids by self-assembling the pyrene-functionalized CdSe (pyrene/CdSe) nanoparticles onto the surfaces of the CNTs [90]. Polymers, especially conjugated polymers, have been shown to serve as excellent wrapping materials for the non-covalent

91

92

3 Carbon-Based Membranes

functionalization of CNTs as a result of π–π stacking and van der Waals interactions between the conjugated polymer chains containing aromatic rings and the surfaces of CNTs [91–93]. Those reported some organic-soluble conjugated poly(m-phenylenevinylene)-co-(2,5-dioctoxy-p-phenylene) vinylene (PmPV), poly(2,6-pyridinlenevinylene)-co-(2,5-dioctoxy-p-phenylene)vinylene (PPyPV), poly-(5-alkoxy-m-phenylenevinylene)-co-(2,5-dioctoxy-p-phenylene)-vinylene (PAmPV), and stilbene-like dendrimers to investigate their non-covalent functionalization for CNTs [91, 94–96]. Surfactants were also employed to functionalize CNTs: (i) nonionic surfactants, such as polyoxyethylene 8 lauryl or C12 EO8 [97] and polyoxyethylene octyl phenyl ether (Triton X-100) [98]; (ii) anionic surfactants, such as sodium dodecyl sulfate (SDS), sodium dodecylbenzenesulfonate (NaDDBS), and poly(styrene sulfonate) (PSS) [99, 100]; and (iii) cationic surfactants, such as dodecyltrimethylammonium bromide (DTAB) [101] and cetyltrimethylammounium 4-vinylbenzoate [102]. The physical adsorption of surfactant on the CNT surface lowered the surface tension of CNTs that effectively prevented the aggregation of CNTs.

3.6 Porous Graphene Graphene composes of a single layer of sp2 -bonded carbon atoms arranged in an aromatic structure [103], exhibiting good chemical stability [104, 105], excellent thermal conductance [106–108], strong mechanical strength [109, 110], and remarkable electronic properties [110–113]. The unique property of graphene (i.e. one atom thickness) results in many promising applications in a wide range of fields [114], in particular for molecular separations. The pristine graphene is demonstrated to be permeable to proton [115, 116], but impermeable to any molecules, even the smallest helium atoms [117–119]. GO is an analogue of graphene by asymmetrically modified with oxygen-containing functional groups (hydroxyl, epoxy, carboxyl, carbonyl, phenol, etc.) on the edges and planes. Therefore, for molecular separation, a graphene-based membrane has to be functionalized with nanopores or nanochannels through chemical or physical approaches. As shown in Figure 3.7, the typical graphene-based materials for this purpose are nanoporous graphene (NPG) and graphene oxide membranes (GOMs) [120]. MD simulations have predicted that monolayer graphene with sub-nanometer pores could act as a highly selective and permeable separation membrane with much higher efficiencies than those of state-of-the-art polymeric filtration membranes [121–125]. For separations, nanopores with controlled sizes and predetermined shapes are highly desirable. Extensive efforts have been devoted to reaching this aim [126–131]. A variety of methods are summarized in Table 3.2 with pore sizes, their advantages, and their limitations. Focused electron beam irradiation of single-layer graphene can create nanopores with controlled sizes [128]. However, this approach is inefficient and not scalable for fabricating large-area porous graphene with high pore density. Bunch et al. created pores in graphene by ultraviolet-induced oxidative etching

3.6 Porous Graphene

Nanopore

Nanoporous graphene

Stacked Interstice

GO membrane Figure 3.7 Schematic pictures of the two types of graphene-based membranes: nanoporous membrane consists of a single graphene layer with defined pores, and GO membrane consists of stacked GO sheets with interstices between the sheets [120]. Source: ACS.

Table 3.2 Advantages and limitations of different perforation methods. Method

Pore size (nm)

Advantage

Limitation

Refs.

Focused electron beam ablation

∼3.5

Tunable and well-defined pore sizes

Limited to small area

[125]

Ability to treat large-area samples

Low pore density and wide size distribution

[126]

Ultraviolet-induced 103 • Robust (good pH range, oxidants) • Good mechanical strength • Not hydrophobic • Readily fabricated (FS and/or HF)b) • Modest cost

• • • •

Ultrafiltration (UF)

• Mean pore sizes: 20–50 nm • Water permeabilitya) >500 • As for MF

• Isoporous (very narrow distribution) • Water permeability >103 • As for MF

Nanofiltration (NF)

• Mean pore size: ≤ 2 nm • Water permeabilitya) ≤ 10 • Retention: M+ (partial), M++ (good) • Partial fractionation of organics • Relatively robust (good pH range) • Good mechanical strength • Not hydrophobic • Readily fabricated (FS)c), d) • Modest cost

• Isoporous (very narrow distribution) • Water permeability > 15 • Retention: M+ (zero), M++ (complete) • Fractionation of organics • Suit wide pH and strong oxidants • Mechanically very strong and durable • Very hydrophilic • Easily fabricated (FS and HF) • Low cost • Intrinsically antifouling • Tunable (surface charge)

Reverse osmosis (RO)

• • • • • • • • •

Forward osmosis (FO) a) b) c) d) e)

Mean pore size: ≤ 1 nm Water permeabilitya) ≤5 Retention: NaCl (> 99%) Retention: boron (> 90%) Relatively robust (good pH range) Good mechanical strength Not hydrophobic Readily fabricated (FS)c), d) Modest cost

• As above, and with support layer S factore) < 500 μm

Water permeability of l m−2 h−1 bar−1 . FS: flat sheet (flexible). FS as in spiral wound modules. HF is available, but not as thin-film composite. S = thickness × tortuosity/porosity.

Desired properties

• • • • • •

• • • • • • • • • • •

Isoporous (very narrow distribution) Water permeability >2 × 103 Suit wide pH and strong oxidants Mechanically very strong and durable Very hydrophilic Easily fabricated (FS and HF) Low cost Intrinsically antifouling Tunable (surface charge, pore size) Self-healing (integrity)

Isoporous (very narrow distribution) Water permeability > 5 Retention: NaCl (> 99.9%) Retention: boron (> 95%) Suit wide pH and strong oxidants Mechanically very strong and durable Very hydrophilic Easily fabricated (FS and HF) Low cost Intrinsically antifouling Tunable (surface charge)

• As above, and with S factor < 200 μm

121

122

4 Microporous Membranes for Water Purification

remove inorganic pollutants in recent years; among these, polymer materials have been widely applied for developing water purification membranes or as supporting substrates for fabricating hybrid membranes due to their excellent permeability and mechanical and chemical stabilities. In this perspective, Feng and coworkers fabricated a PVDF nanofibrous membrane using electrospinning approach for water desalination process [20]. The prepared porous membrane produced pure drinking water successfully (concentration of NaCl lower than ∼280 ppm, 6 wt% NaCl) by desalination, which is comparable with the performance of commercial membranes. The observed results depicted that the membrane flux reached 5–28 kg m−2 h−1 at the temperature ranging from 288 to 356 K. Furthermore, the recyclability of this membrane was outstanding after several uses of water purification. Sun et al. fabricated an NF membrane by immobilizing Aquaporin Z-reconstituted liposomes onto a polydopamine (PDA)-coated microporous membrane support [21]. The obtained membrane retained 66.2% of sodium chloride (NaCl) and 88.1% magnesium chloride (MgCl2 ), respectively. This was attributed to the fact that Aquaporin Z is a hydro-channel protein and exhibit extraordinary water selectivity and permeability. The membranes modified with an Aquaporin Z-to-lipid weight ratio of 1 : 100 improved the water flux by 65% when compared with the membranes without Aquaporin Z functionalization. To prove the positive effects of surface modification on the membrane network, another study was performed by Tijing et al. They developed a superhydrophobic PVDF-co-hexafluoropropylene nanofiber membrane containing distinct weight percent of carbon nanotubes (CNTs) by one-step electrospinning approach [22]. The hydrophobic and mechanical characteristics of prepared membrane could be organized by regulating the concentration of the CNT filler. The observed results showed that the 5 wt% CNT membrane led to highest water flux (24–29.5 l m−2 ) and 99.99% salt retention under an external pressure (≥ liquid entry pressure) of 99 kPa. Subsequently, two-dimensional (2D) graphene-based materials are superb building blocks for fabricating nanoporous membranes for removal of inorganic contaminants in water purification. For example, Chen and coworkers developed graphene oxide (GO) membranes (GOMs)) precisely controlled by added cations through the strong cation–π interactions in the system [23]. The membrane spacing selectively controls one type of cation efficiently and eliminates other larger cations [24]. Using classical molecular dynamics (MD) simulations, Cohen-Tanugi et al. reported that nano-sized pores in single-layer freestanding graphene membrane can efficiently remove NaCl from water [24]. The simulation results clearly showed that the desalination performance is mainly dependent on the pore diameter of a membrane as well as critical role of chemical functional groups (—OH) bonded to the edges of graphene pores. Both significantly improved the water flux, prevented salt, and permitted water flow selectively (Figure 4.2). Similarly, Sumedh et al. discovered tailor-made nanoscale pores on a monolayer graphene membrane through an oxygen plasma etching approach. The obtained membrane exhibited a 100% retention of salt ions (K+ , Na+ , Li+ , Cl− ) with a water flux of 106 g m−2 s−1 at 313 K [25]. Even so, controlling the pore size of monolayer membranes remains a significant challenge, but the nanoporous

4.3 Removal of Water Contaminants (Inorganics, Organics, Biological)

(a)

(b)

Hydrophobic pore

Hydrophilic pore

4.5 nm

(c)

ΔP

H2O

Cl–1 Na+1

Figure 4.2 Graphene pores of (a) hydrogenated and (b) hydroxylated functional groups and (c) lateral view of the computational desalination [24]. Source: Copyright 2012, ACS.

graphene membranes are potential candidate for desalination applications. Moreover, 2D graphene-like materials show similar separation aptitude as graphene-based membranes. For example, Ren and coworkers assembled a 2D nm thin membrane of Ti3 C2 Tx (MXene) for charge- and size-selective removal of ionic and molecular species [26]. The prepared MXene membrane (interlayer spacing of 6 Å) exhibited ultrafast water flux of 37.4 l m−2 h−1 bar−1 and high selectivity toward different-sized metallic and dye cations (MB+ , Ca2+ ) and slower permeation compared with single-charged cations (Na+ ). For the removal of anionic species, Henmi et al. fabricated a bicontinuous cubic (Cubbi) membrane through the self-assembly of thermotropic liquid-crystal (LC) molecules [27]. The obtained Cubbi membrane had unique porosity with average size of 0.6 nm and exhibited exceptional anion selectivity and rejection properties. The observations revealed that the Cubbi membrane removed 83% of Br− , 59% of Cl− , 33% of SO4 2− , and 81% of NO3 − , respectively. This self-organized nanostructured membranes have great potential for removing inorganic contaminants and producing high-quality water. Similarly, Mezzenga

123

124

4 Microporous Membranes for Water Purification

and coworkers [28] prepared a hybrid composite membrane by incorporating β-lactoglobulin, amyloid fibrils, and activated carbon through vacuum filtration technique. The obtained hybrid membrane could eradicate more than 99% of AsO3 4− and AsO3 3− from arsenic-contaminated water via strong supramolecular metal–ligand interactions and had excellent reusability. When compared with the cation-rejection membrane, it can be concluded that the anion-rejection membrane repels anions due to electrostatic interaction or channel surface coordination groups, while the cation-rejection membrane prevents cations due to the pore size of the membrane network. In the case of NP contaminants, NPs can easily enter the human body through the food chain (such as drinking water), causing cellular oxidative stress and inflammation. Thus, the removal of NPs from water is of great importance. Previously, DesOrmeaux et al. produced a nanoporous silicon nitride membrane using porous nanocrystalline silicon support for removing gold (Au) NPs from water [29]. The prepared membrane exhibited pore sizes from 40 to 80 nm, which was controlled by ion etching conditions and mask layer modification. The observation showed that the resultant membrane removed more than 80% of Au NPs (diameter ≥ 60 nm) from water. Similarly, Zhang et al. reported a unique methodology to fabricate ultrathin nanoporous membranes with cellulose nanofibers (∼7.5 nm in diameter) using a freeze-extraction approach [30]. The thickness of obtained membrane could be controlled to as thin as 23 nm, with pore sizes ranging from 2.3 to 12 nm. The experimental results showed that the 30 nm thick membrane had 1.14 × 104 of water flux and 3.96 × 104 l m−2 h−1 bar−1 of acetone flux, respectively. The as-prepared cellulose membranes with sub-10 nm pores exhibited wide application in the fast removal of NPs and substrates with diameters larger than 10 nm from water. Moreover, Li et al. fabricated 3D CNT sponge network into a thin porous membrane for sieving NP pollutants and dye molecules from contaminated water. The prepared CNT membrane could eliminate 80% of CdS NPs (2–4 nm), 100% of Au NPs (8 nm), 100% of TiO2 NPs (12 nm), and approximately 100% of methyl orange and rhodamine B (RhB) molecules from contaminated water, respectively [31]. The CNT membrane is not only thermochemically stable but also has filter capacities up to 45 l g−1 . 4.3.2

Organic Pollutants

The organic compounds that are resistant to environmental degradation through chemical, biological, and photolytic processes including pesticides, hydrocarbons, phenols, oils, and pharmaceuticals are called organic pollutants/contaminants. Because of their persistence in ambient conditions, they have potential adverse impact on human health, endangering the aquatic organisms and damaging the ecosystem, and can reduce the amount of DO in water in their oxidative decomposition process. However, some organic pollutants are produced naturally such as volcanoes and numerous biosynthetic pathways, but most are man-made via industries, vehicle fuel consumption, and total synthesis. Many useful porous materials and technologies have been reported to remove organic pollutants from the natural and wastewater system.

4.3 Removal of Water Contaminants (Inorganics, Organics, Biological)

Among these, membrane technology is getting remarkable attention due to its captivating structural features. In this perspective, Karim and coworkers prepared a nanoporous membrane with cellulose nanocrystals using a freeze-drying approach [32]. Even though the obtained membrane exhibited low water flux (6.4 l m−2 h−1 bar−1 ), it could remove 98% of organic dyes (Victoria Blue 2B, 84% of methyl violet 2B, and 70% of rhodamine 6G) from industrial wastewater. The mechanism suggested that the dye adsorption is mainly through host–guest hydrogen bonds and electrostatic attraction. Later on, scientists found new membranes such as CNT membrane for water purification technologies. Well-aligned CNT can serve as robust pores in membrane framework for water desalination and decontamination technologies [33]. The hollow CNT structure offers frictionless passage of water molecules, and this makes them applicable for the development of high-flux separation. Appropriate pore diameters can constitute energy barriers at the channel entries, rejecting pollutant species and allowing water flow through the nanotube voids [34]. It is also possible to functionalize CNT channel walls to selectively remove contaminants [35]. Thus, CNT membrane plays a role as a “gatekeeper” for size-controlled separation of multiple pollutants. Efforts have been performed to prove gold rush possibility of potential CNT membranes to resolve both the seawater and brackish water desalination/purification. In this domain, Lee and coworkers fabricated high-performance superhydrophobic/superoleophilic membrane using vertically aligned multiwalled carbon nanotubes (VAMWNTs) on a stainless steel mesh substrate [36]. The obtained membrane has contact angles for water and diesel of 163∘ and 0∘ , respectively. It could efficiently separate diesel and high-viscosity lubricating oil from water system. These observations make the VAMWNT-based membrane very promising for the oil/water separation application. Afterward, Liang and coworkers prepared a uniform freestanding carbonaceous nanofiber (CNF) membrane using simple casting approach for removal of dyes and toxic ionic pollutants (methylene blue, Pb, Cr) from water [37]. This approach combines the excellent adsorption behavior of CNFs and the advantages of membrane filtration over traditional adsorption techniques, which include simple scale-up, reduced time, and lower energy consumption. The obtained CNF membrane exhibited large surface, several active sites, unique porosity, and high uniformity (Figure 4.3). Therefore, membrane filtration experiments proved that the obtained membrane could efficiently remove methylene blue at a very high flux of 1580 l m−2 h−1 , 10–100-fold greater than that of commercial NF or UF membranes. In addition, CNF membranes are easily regenerated in adsorption–desorption performance over several cycles and simple washing with HCl. To promote the surface modification and ease of incorporation of CNTs with polymeric system, Gu and his coworkers fabricated a novel Ag/polyacrylic acid (PAA)–CNT hybrid microporous membrane for treating oil/water/solid three-phase separation [38]. From Figure 4.4ac, it can be seen that the obtained microporous membranes cannot be wetted by oil. The resultant membrane provides excellent superoleophobic behavior for numerous organic solvents including dichloromethane (DCM) and carbon tetrachloride (CCl4 ) while compared with water. The wetting behavior under toluene confirms the

125

126

4 Microporous Membranes for Water Purification

(a)

(b)

~50 μm

5 μm

20 μm

Figure 4.3 SEM images of the CNF membrane: (a) upper view with an inset photograph and (b) cross-sectional view [37]. Source: John Wiley & Sons. (a)

(b)

(c)

Oil Water

Before (e)

(d)

After (f)

Oil Water

Before

After

Figure 4.4 Contact angle measurements of Ag/PAA–CNTs membrane under (a–c) oil and (d–f ) under water [38]. Source: Copyright 2016, RSC.

superhydrophilicity of the resultant membrane (Figure 4.4d–f ). These findings reveal that the newly developed membrane can successfully separate an extensive range of surfactant-stabilized oil-in-water emulsions with a significant flux of 3000 l m−2 h−1 bar−1 . In another study, Cao et al. reported a highly water-selective hybrid membrane for water/ethanol separation by incorporating g-C3 N4 nanosheets into a matrix of sodium alginate. The obtained hybrid membrane containing 3% CNs exhibited an optimal pervaporation performance with a permeation flux of 2469 g m−2 h−1 and a separation factor of 1653 for the dehydration of 10 wt% water/ethanol mixture at 350 K. Besides, the hybrid membranes revealed good long-term stability [39]. Zhang and coworkers fabricated a superoleophobic poly(acrylic acid)-grafted PVDF (PAA-g-PVDF) membrane using a salt-induced phase inversion approach [40]. The as-prepared membrane could separate both surfactant-free and surfactant-stabilized oil-in-water emulsions, either under the pressure of 0.1 bar or under gravity, with high separation efficiency (> 99.99 wt% pure water) and high flux (23 200, 16 800, and 15 500 l m−2 h−1 bar−1 of flux for hexadecane/H2 O, toluene/H2 O, and diesel/H2 O, respectively). Besides, the obtained membrane

4.3 Removal of Water Contaminants (Inorganics, Organics, Biological)

had good recyclability after a simple washing with water. Subsequently, Obaid and coworkers reported an oil/water separation membrane by incorporation of NaOH NPs inside the polysulfone (PSF) nanofibers [41]. The NaOH-modified PSF nanofiber membrane could remove almost 100% oil from water due to hydrophobic and hydrophilic sites in the membrane network. Although polymeric membranes incorporating carbon nanomaterials exhibit promising oil/water separation efficiency, they are unstable under severe conditions due to low thermochemical stability of polymeric membrane support [42, 43]. Moreover, the intricate issue about the valuable oil recovery and recycling of carbon nanomaterials is to be considered from the point of economic view. In this perspective, carbon-based membranes have drawn great attentions among membrane technologies due to their low cost, superior chemical and mechanical stability, and highly integrated operation. They are considered as an appropriate membrane material for treatment of oily wastewater owing to their stability in harsh environments. Over the past few years, several reports have been published on the implementation of carbon-based membrane in oily wastewater purification. Hsieh and coworkers [44] studied the surface properties and oil separation efficiency of carbon fiber (CF) membrane modified with fluorinated CNTs. As compared with the unmodified CF membrane, the modified CF membrane provided better separation performances. The separation efficiency of CNT–CF membrane reached as high as 99.7%. Additionally, they also found that the thinner the CF membrane, the faster the permeability was achieved, and the presence of CNTs was favorable for the permeability at the same thickness of CF membrane. Furthermore, modified CF membrane could reach an oil permeability of 180 l m−2 h−1 bar−1 , better than CF membrane without the CNTs (95 l m−2 h−1 bar−1 ). The result was consistent with the contact angle measurements using different nature of liquids. The superoleophobicity of membrane surface was probed using three different types of liquid droplets including distilled water, bead sauce, and olive oil as shown in Figure 4.5. Besides water droplet, the other two types of oil droplets completely wetted the CF surface (Figure 4.5a–c), indicating the superoleophilic behavior of CF membrane. On the other side, the fluorinated CF surface membrane exhibited a significant water and oil repellence (Figure 4.5d–f ). This outcome reflected that the fluorination membrane had efficiently reduced the surface energy of the CF substrate. In contrast, three unlike liquid droplets displayed the same quasi-spherical form on the CNT–CF membrane surface (Figure 4.5g–i). Consequently, the fluorinated CF membranes incorporated with CNTs unveiled the highest oil and water repellence than unmodified carbon-based membrane. Although carbon-based membranes are capable to attain very high performance, they are not able to be mass-produced due to their relatively high cost in the performance regeneration process induced by the tiny emulsified oil droplets within the surface channels. Furthermore, the intricate matter and the environmental impact face us when using surplus chemical reagents on membrane cleaning process. Moreover, a complete economic assessment in the aspect of energy consumption using carbon-based membrane must be carefully evaluated in oil and gas industries.

127

128

4 Microporous Membranes for Water Purification

(a)

(d)

(g)

(b)

(e)

(h)

(c)

(f)

(i)

Figure 4.5 Representative photographs of contact angle measurements: (a–c) unmodified CF membrane, (d–f ) fluorinated CFs membrane, and (g–i) fluorinated CNT–CF membrane (top, middle, and bottom rows present water, soya sauce, and olive oil, respectively) [44]. Source: Copyright 2015, Elsevier.

GO, an atomic sheet produced by oxidative exfoliation of graphite materials, has emerged as one of the most studied nanomaterials in recent years [1, 45]. Its polar oxygenated functionalities such as epoxide, hydroxyl, and carboxyl groups give the material with potential uses in various fields, such as optics, gas sensing, composite materials, gas barriers, and nanobiotechnology [2, 45]. Recently, GO-based nanoporous membranes have been reported as promising medium for gas and liquid separation as well as for removing organic pollutants. In this perspective, Han and coworkers fabricated ultrathin (∼22–53 nm thick) 2D graphene membrane on microporous substrate using filtration-assisted assembly approach for water purification [46]. The well-packed graphene layer membrane (Figure 4.6) could sieve 99.8% of methyl blue and 99.9% of direct red 81 and showed moderate retention (≈20–60%) for ion salts. In addition, the water flux of the membrane reached 21.8 l m−2 h−1 bar−1 . The water purification

4.3 Removal of Water Contaminants (Inorganics, Organics, Biological)

(b)

(a)

HO2C

O

HO2C

CO2H

HO2C

HO2C

HO2C

OH

CO2H

O O O HO2C

1 cm

O O

CO2H

CO2H

CO2H

CO2H CO2H

O HO2C

Figure 4.6 (a) Photographs of ultrathin graphene nanofiltration membranes (uGNM) coated on anodic aluminum oxide disk and a twisted form coated on a PVDF membrane and (b) structure of base-washed graphene oxide [46]. Source: John Wiley & Sons.

mechanism analysis revealed that this type of negatively charged membrane rejected dyes mainly through both physical sieving and electrostatic interaction. Huang and coworkers fabricated a continuous GOM on ceramic hollow fiber support using vacuum suction approach [45]. The obtained GOM could successfully remove dimethyl carbonate from water due to its favored water sorption and fast water diffusivity through the GO pores and layers. At 298 K and 2.6 wt% feedwater content, the permeate water content reached 95.3 wt% with a permeation flux of 1702 g m−2 h−1 . Similarly, Tang and coworkers reported a freestanding GOM using a pressurized UF strategy for separation of ethanol/water mixtures (Figure 4.7) [47]. The obtained GOM exhibited a uniform layered microstructure with high hydrophilicity and structural stability. Due to numerous active sites, unique porosity, intermolecular hydrogen bonding between host and guest, and high surface area of the layered GOM, the resultant membrane showed high retention for water (water/ethanol selectivity is 227) and could dehydrate 85 wt% ethanol aqueous solution at 297 K. Besides GO, TiO2 NPs have also been used for organic wastewater purification due to their antifouling and exceptional photocatalytic oxidation properties. For instance, Gao and coworkers fabricated a high-performance GO-TiO2 microsphere hierarchical membrane using assembling the GO-TiO2 microsphere composites on the surface of a polymer filtration membrane for concurrent water filtration and photodegradation [48]. The rejection experiments revealed that the permeate flux of the obtained membrane reached 60 l m−2 h−1 , ninefold more as compared with commercial membranes. Additionally, under ultraviolet (UV) irradiation, TiO2 was excited to generate highly oxidative species, electron holes, and hydroxyl radicals (OH⋅ ); these reactive species caused degradation of organic contaminants. The GO-TiO2 membrane could eradicate more than 90% of dyes (rhodamine B, acid orange 7) and humic acid (HA) from water under UV irradiation. This type of microporous membrane has a bright future in the arena of clean water production. Subsequently, some other types of 2D nanosheets were applied to fabricate porous membranes for water purification. In order to achieve high water flux, Sun et al. fabricated laminar MoS2 and WS2 membranes for the separation of Evans blue molecules in succession [49, 50]. The obtained laminar MoS2

129

20.0

500

16.0

400 300

12.0 P (Water)

8.0

P (Ethanol)

100

4.0 0.0

× 5000

200

Selectivity

500 400 300 200

Selectivity

(c)

Ethanol permeability/Barrer

(b)

Water permeability/103 barrer

(a)

100 0

2

5

10

20

Pressure/bar

Figure 4.7 Morphology of the GO membrane: (a) digital photo, (b) SEM cross-sectional view, and (c) the pervaporation performance [47]. Source: Copyright 2014, Elsevier.

4.3 Removal of Water Contaminants (Inorganics, Organics, Biological)

membrane prevented Evans blue molecules about 89% with a water flux of 245 l m−2 h−1 bar−1 . Likewise, it was found that the layered WS2 nanosheet membrane could avert Evans blue molecules over 90% with a water flux of 730 l m−2 h−1 bar−1 . In summary, the above-discussed and more examples of nanoporous and microporous membranes for organic pollutants capture are listed in Table 4.3. 4.3.3

Biological Pollutants

Biological pollution is the impact of humanity’s actions on the quality of aquatic and terrestrial environment. Particular examples of some biological pollutants/pathogenic microorganisms such as algae, protozoa, planktons, bacteria, and virus are seriously responsible for causing several waterborne diseases. Several efforts have been performed for the purification of water from biological contaminants in recent years. For example, Yang et al. fabricated a double-layer nanoporous membrane for virus filtration with good stability, unique porosity (Figure 4.8a and b), and selectivity under high pressures of 2 bar [54]. The upper layer is a nanoporous film with a pore size of approximately 17 nm and a thickness of approximately 160 nm as shown in the Figure 4.8c, which was produced by a polystyrene-block-poly(methyl methacrylate) (PS-b-PMMA) copolymer. The lower layer is a conventional MF membrane acted as a substrate, which provides strength to mechanical feature. The rejection experiments depicted that the obtained double-layer membrane had high selectivity toward human rhinovirus type 14 and could sieve nearly 100% of human rhinovirus type 14 from phosphate buffer solution. Similarly, for the removal of biological contaminants, another highperformance porous membrane was reported by Sato and coworkers. They fabricated a nanofibrous composite membrane using infusing surface-functionalized ultrafine cellulose nanofibers (diameter of ∼5–10 nm) into an electrospun nanofibrous scaffold [55]. The experimental observations revealed that the water flux of the created membrane could reach 85 l m−2 h−1 bar−1 , as well as removed 99.99% of bacteriophage virus and 99.9999% of Escherichia coli simultaneously. The mechanism analysis showed that the nanofibrous composite membrane significantly sieved bacteria from the water because of its relevant pore size and it adsorbed viruses (negatively charged) through electrostatic interactions. Zhang and coworkers fabricated high-porosity (21.3%) TiO2 nanowire UF membranes with layered hierarchical structure through alkaline hydrothermal synthesis and hot-press approaches [56]. Figure 4.9 shows two distinct sizes of nanowires with different roles, for example, the TiO2 nanowires (diameter ∼10 nm named as TNW10 ) act as the functional layer, while TiO2 nanowires with a diameter of 20 nm (TNW20 ) are laid as the supportive layer. The obtained TiO2 UF membrane could successfully separate polyethylene glycol (PEG), polyethylene oxide (PEO), HA, and E. coli from water. It could also destroy organic pollutants (such as PEG and PEO) and inactivate biological pollutants (such as E. coli) under UV irradiation. The hierarchical TiO2 UF membrane exhibits excellent multifunctional capabilities under UV irradiation, such as antifouling, antibacterial property, concurrent separation, photodegradation,

131

Table 4.3 Microporous membranes used for removing targeted organic contaminants, the fabrication methods, and their ultimate water fluxes. Membrane

Fabrication method

Target and retention (%)

Permeance/water flux

Reference

GO membrane

Shear-induced alignment

Organic probe molecules (>90%)

71 ± 5 l m−2 h−1 bar−1

[51]

CNCs

Freeze-drying process

Victoria Blue 2B (98%), methyl violet 2B (84%), rhodamine 6G (70%)

6.4 l m−2 h−1 bar−1

[32]

VAMWNTs

Chemical vapor deposition (CVD)

Lubricating oil

1580 l m−2 h−1

[36]

Ag-APAN

Electroless plating, surface modification

1,2-Dibromoethane

—

[52]

CNs-SA

Thermal oxidation etching

Ethanol

2469 g m−2 h−1

[39]

PAA-g-PVDF

Phase inversion

Hexadecane, toluene, diesel (>99.99%)

15 500–23 200 l m−2 h−1 bar−1

[40]

PSF nanofibers

Electrospinning, interfacial polymerization

Soybean oil (∼100%)

5.5 m3 m−2 day−1

[41]

uGNM

Filtration-assisted assembly

99.8% of methyl blue and 99.9% of direct red 81

21.8 l m−2 h−1 bar−1

[46]

GO

Vacuum suction

Dimethyl carbonate (∼95.2%)

1702 g m−2 h−1

[45]

GO

Pressurized ultrafiltration

Ethanol (∼100%)

—

[47]

GO-TiO2

Self-assembly

Rhodamine B, acid orange 7, humic acid (HA) (>90%)

60 l m−2 h−1

[48]

MoS2

Vacuum filtration

Evans blue (89%)

245 l m−2 h−1 bar−1

[49]

WS2

Vacuum filtration

Evans blue (90%)

730 l m−2 h−1 bar−1

[50]

rGO-CNT

Vacuum-assisted filtration

Nanoparticles, dyes, BSA, sugars, and HA (>99%)

20–30 l m−2 h−1 bar−1

[53]

TiO2 nanowire

Hydrothermal synthesis, hot-press process

Polyethylene glycol, polyethylene oxide, HA, Escherichia coli

—

[30]

PMMA

Ultraviolet irradiation, acid rinsing

Human rhinovirus type 14 (∼100%)

—

[54]

MCCNs-PEI

Electrospinning

MS2 bacteriophage virus (99.99%), E. coli (99.99%)

85 l m−2 h−1 bar−1

[55]

APAN, polyacrylonitrile; CNCs, cellulose nanocrystals; g-C3 N4 nanosheets incorporated into sodium alginate matrix; GO, graphene oxide; MCCNs, microcrystalline cellulose nanofibers; PAA-g-PVDF, poly(acrylic acid)-grafted polyvinylidene fluoride; PEI, polyethylenimine; PMMA, polystyrene-block-poly(methyl methacrylate); PSF, polysulfone; rGO-CNT, reduced graphene oxide carbon nanotube; uGNM, ultrathin graphene nanofiltration membrane; and VAMWNTs, vertically aligned multiwalled carbon nanotubes.

4.3 Removal of Water Contaminants (Inorganics, Organics, Biological)

(a)

(c)

(b)

Double layer

200 nm

200 nm

200 nm

Figure 4.8 SEM images for the double-layer nanoporous membrane fabricated by PS-b-PMMA: (a) upper surface, (b) bottom surface, and (c) cross-sectional view [54]. Source: John Wiley & Sons. (a)

TNW10

(c)

(d)

20 nm

(b)

TNW20

100 nm

Figure 4.9 The hierarchical layer of TiO2 nanowires UF membrane: (a and b) TEM images of titania nanowires, (c) schematic view of TiO2 nanowires UF membrane, and (d) photograph of TiO2 nanowires UF membrane [56]. Source: John Wiley & Sons.

and photocatalytic oxidation. Similarly, Chen et al. developed a reduced graphene oxide nanofiltration (rGO-CNT-NF) hybrid membrane intercalated with CNTs on porous ceramic MF membranes using a facile vacuum-assisted filtration approach [53]. The obtained hybrid membrane exhibited 20–30 l m−2 h−1 bar−1 in permeability and more than 99% retention of NPs, dyes, bovine, organophosphate, sugar, and HA. Based on high permeability and separation performance of the rGO-CNT hybrid NF membranes, it is attributed to the fact that the use of block copolymers (BCPs) as a surfactant can enhance steric repulsion and thus disperse CNTs effectively, placing well-dispersed 1D CNTs within 2D graphene sheets to permit a uniform network, which can offer numerous mass transfer pore channels through the continuous 3D nanoporous framework. The aforementioned nano-/microporous membranes are discussed for the separation of three types of contaminants including inorganic, organic, and biological. Commonly, phase inversion and electrospinning technologies are used for fabricating inorganic, organic, and hybrid membranes. Furthermore,

133

134

4 Microporous Membranes for Water Purification

organic membranes (polymeric membrane) exhibit a higher water flux, but a lower separation rate compared with inorganic membrane materials. In contrast, inorganic membranes have a higher separation rate and selectivity toward target contaminants, but a lower water flux when compared with organic membrane materials. Therefore, inorganic–organic hybrid membranes get the advantages of their individuals, not only sustaining high water flux but also exhibiting high selectivity and removal rate for target species. In the next section, we will explain and discuss microporous (CNTs and graphene-based membranes and other) membranes for desalination process. We also highlight the current hurdles and future challenges of membrane materials.

4.4 Membrane Desalination The development of nanotechnology has given enormous opportunities to purify water even at ionic state called desalination. The numerous nanostructured materials have been fabricated with captivating features including high aspect ratio, reactivity and tunable pore volume, pore size, and electrostatic, hydrophilic, and hydrophobic interactions that are beneficial in adsorption, catalysis, sensing, and optoelectronics. Nanoscale metals (silver, titanium, gold, and iron) and their oxides have been extensively applied in environmental mitigation. For example, silver (Ag) NPs are effective in disinfecting biological pollutants such as bacteria, viruses, and fungi [57]. Titanium (Ti) NPs have been reported in micropollutants, transforming redox reactions [58]. Photocatalyst nano-TiO2 can degrade phenolic recalcitrant compounds and microbial and odorous chemicals into harmless species [59, 60]. Most of the nanomaterials have enhanced surface porosity, which increases salt rejection and prevents formation of macropores [61]. Nanomaterials have played main roles in degrading various recalcitrant dyes and halogenated compounds and removal of heavy metals with disinfecting microbes for water purification. Moreover, gold (Au) and iron (Fe) NPs are quite appropriate for removing inorganic heavy metallic ions from surface water and wastewater [57]. Nanomaterials are not only effective in disintegrating various contaminants, but also they are active in bimetallic coupling with other metals and metal oxides that synergistically improve pollution catalysis. Although nanomaterials are appreciated in the field of water desalination and purification, they are also not free from limitations. Limitations include thermal instability, necessity of high pressure, fouling, contaminants precipitation, pore blocking, low flux, slow reaction, formation of toxic intermediates, formation of ion particles, and high rate of aggregation on storage. Moreover, low chances of reusability and unknown risks to ecosystems are also other concerns [57, 61]. Therefore, numerous nanostructured materials can be applied in the form of “composite membranes” for better performances and lack of certain limitations. The membrane technology improves salt retention ability and curtails costs, land area, and energy for desalination. For example, CNTs and graphene-based membranes were extensively employed for adsorption-based desalination because of their extraordinary adsorption capacities [62, 63]. Besides, zeolite NPs are mixed with polymer matrix to form

4.4 Membrane Desalination

thin-film RO membranes [64, 65]. They improve water transport with more than 99.7% salt retention [65]. Likewise, silica NPs were doped with RO polymer matrices for water desalination [66]. They developed polymeric networks, pore sizes, and transport characteristics. CNTs are composed of cylindrical graphite sheets (allotropic form of carbon) rolled up in a tubelike structure with the appearance of latticework fence [67]. Categorically, single-walled carbon nanotubes (SWCNTs) have cylindrical shape composed of a single shell of graphene sheet. On the other side, multiwalled carbon nanotubes (MWCNTs) are composed of multiple layers of graphene sheets. Both types have been significantly applied for direct water desalination process [68–70]. CNTs are getting enormous attentions in advanced membrane technologies for water desalination because CNT membranes provide near frictionless water flow through them with the retention of a broad spectrum of water pollutants and the inner hollow cavity of CNTs offers a huge possibility for separating ions in desalination. Additionally, the high aspect ratios, smooth hydrophobic walls, and inner cavities of CNTs permit ultraefficient transport of water molecules. Mainly, CNT membrane performances often rely on its fabrication methods. Currently, two types of nanotube membranes are available: (i) vertically aligned (VA) and (ii) mixed matrix (MM) CNT membranes. The VA-CNTs are of special interest for the fabrication of membranes (Figure 4.10a). The pore size has special effects on the water passages through the membranes Water molecule

Salts Saline water

Comparison between VA and MM CNT membranes ∗ Vertical CNT arrangement

Filler

(a)

∗ Compact CNT network

Functional groups

∗ Water flux rate is high ∗ Easier to add functional groups at CNT tips/surface ∗ Complicated fabrication procedure Pure drinking water

CNT membrane

Doped CNTs Polymer phase 1 Polymer phase 2

(b)

CNT-Polymer composite

Saline water

∗ Operating system adjustable

∗ Mixed polymeric materials ∗ CNT networks loosely fit ∗ Water flux is moderately fast ∗ Antifouling membrane ∗ Simple fabrication process ∗ Operating system feasible

se

Ba

Pure drinking water

Figure 4.10 Trapping of salts and movement of water molecules from saline water through (a) VA-CNT and (b) mixed matrix CNT membranes [71]. Source: Copyright 2013, Elsevier.

135

136

4 Microporous Membranes for Water Purification

consisting of aligned CNTs. To prove this feature, Majumder and coworkers [72] developed vertically CNT membrane, and experiments observed the frictionless movement of water molecules with high velocities from 9.5 to 43.0 cm s−1 bar−1 through a membrane of pore channel (diameter ∼7 nm), higher than other organic solvents (ethanol, isopropanol, hexane, and decane). The experimental observations revealed that the water flow rates were four- to fivefold faster than those of traditional flow of between 0.00015 and 0.00057 cm s−1 bar−1 . Furthermore, the molecular simulation data exhibited that the water conductance of the (7,7) and (8,8) tubes are roughly double and quadruple than that of the (6,6) tube, respectively [34]. Though a single water chain forms in both the (5,5) and (6,6) tubes, the permeability of the former is a little under half the latter due to the fact that water chains only form across the narrower pore half of the time [34]. On the other hand, a MM-CNT membrane entails numerous polymeric layers or other composite materials (Figure 4.10b). In this perspective, Zimmerman and coworkers fabricated polymeric membranes [73] for gas purification. Simultaneously, the obtained polymeric membrane has significantly strengthened for the water purification ability compared with the existing membrane materials. Mostly, membrane fouling and pollutant precipitation are major issues in separation technology. Irregular pore size, deleterious micropollutants, influent water quality, and variable pH reduce membrane separation ability, pollutant selectivity, and adsorption capacities. Precipitation takes place in membranes when water pollutants exceed their solubility limit in water. Additionally, fouling creates defects in membrane pores causing pore blocking and complicating membrane regeneration. To overcome these problems, MM-CNT membranes could be a great solution to resolve the bottlenecks of current separation technologies. Choi and coworkers [74] developed MM-CNT membranes to enhance filtration capacity of UF membranes. They dramatically increase the water permeability, solute retention, and mechanical robustness of the membrane [75]. Therefore, CNT types and conformation play crucial roles in water passage and permeability. Afterward, functionalization of CNT membranes is often a recognized strategy for getting improved performance in water purification and desalination. CNTs in pristine CNT membrane often aggregate, which significantly reduce water flux and pollutant rejection. CNTs are generally contaminated with metal catalysts, impurities, and physical heterogeneities [76]. Also, CNTs are capped into hemisphere-like fullerene-type curvature during synthesis and purification process [77]. These caps are expanded into open tips that could be oxidized into specific functional groups to capture selective contaminants. Chemical functionality can be positive (—NH3 + ), negative (—COO− , SO3 − ), hydrophobic (aromatic rings), and hydrophilic (OH, NH2 , —COOH) on CNT surfaces [61, 78]. This creates CNT membranes more selective for specific pollutant retention and also improves water influx through the nanotube hole. Modified CNT membranes exhibit high water permeability, thermomechanical stability, fouling resistance, pollutant degradation, and self-cleaning functions [79]. Tip-modified CNT membranes have selective functionalities on the nanotube mouth, and the core-modified CNT have functionalities at interior parts [80]. Consequently, strategies of functionalization, controlling the pore sizes,

4.4 Membrane Desalination

regulating the desired host–guest interaction are effective to get high selectivity and to improve permeability, making CNT membranes suitable for the water desalination process [61, 81]. The major findings in this field are summarized in Table 4.4. Graphene, a single-atom-thick sheet of sp2 -hybridized carbon atoms in 2D and hydrophobic layer forming a honeycomb pattern, has opened the window for assembling separation membranes with enhanced separation abilities [95]. The perceptions of graphene-based materials have been explored for 10 decades. As early as in 1859, Brodie prepared graphitic oxide with a strong oxidizing mixture [96]. In 1958, Hummers developed a redox strategy to prepare graphitic oxide [97]. Since its first isolation in 2004, graphene has drawn increasing attention in various branches of science and technology due to its remarkable physicochemical properties, such as monoatomic thickness, high mechanical strength, and substantial chemical inertness [98]. Therefore, graphene has promising potential for atomic permeation, water transportation, water purification, and desalination [99–101]. For mass production, controllable pore distribution, and application under high pressure conditions, graphene’s oxygen-containing analogue, GO, is being considered for filtration/membrane fabrication and utilization. The possibility of facile and large-scale production of GO and its unique properties have opened up new opportunities for the production of advanced membrane separation technology. Therefore, GO-based membranes have been employed in water treatment and liquid organic molecular separation [102–104]. It should be noted that the physicochemical status of the oxygen-containing groups in GO sheets plays a crucial role in determining their separation efficiencies. In water purification or desalination process, two key features govern the productivity and separation performance of graphene and GO-based membranes, that is, water flux and rejection of dissolved impurities. When the pore diameter is greater than 0.8 nm, the graphene membranes exhibit higher water flux than CNT-based membranes owing to higher velocity in the center area [105]. An exciting study designated that graphene nanopores are capable to block salt ions with twp to three orders of magnitude greater water permeability than commercial RO membranes [106]. However, molecular behavior and mechanisms of water permeation through the layer-by-layer and porous microstructure of the GOM are anticipated using theoretical studies or MD simulations. In this perspective, Wei and coworkers [107] fabricated a high-performance microporous graphene membrane and observed the improved water flow and pointed out that the side-pinning effect caused by H-bonds between water molecules and oxidized regions led to fast water transport in graphene channel walls. Also the enhanced water permeability, a small sheet size, high density of inter-edge spaces, and extensive nanochannels are suggested in GOM fabrication [107]. Similarly, interlayer distance can be tuned using different-sized cross-linkers [108]. Moreover, Cohen-Tanugi et al. proposed that water flux through monolayer graphene has a linear relationship with pore area using the computational approach [24]. They demonstrated that the hydrophilic pores improved water permeation due to H-bonding between host and guest, and quantitatively the hydrophilic edges of the pores contributed more to high water flux compared

137

138

4 Microporous Membranes for Water Purification

Table 4.4 Relevant findings of vertically aligned (VA) and mixed matrix (MM) CNT membranes for water desalination. Membrane type

Substrate/filler

Major observations

Reference

VA

Polystyrene

[82]

VA

Polystyrene

VA

SiNx

VA

Si3 N4

VA

Epoxy resin

• Used for chemical separations and sensing • Modest interaction of the ion with the CNT tips and the core is secured • Removed heavy metals and hydrocarbons • Filtered E. coli and poliovirus (∼25 nm) • Bamboo-shaped CNTs can be used in separation and chemical sensing • Water flux rate was 2.4 × 10−8 mol m−2 s−1 • Water flow reached > 3 orders • Pores (> 2 nm) improved ion selectivity and water permeability • Used for water purification • Improved water permeability

VA

Polyethyleneimine

[87]

MM

Polysulfonate

MM

Polysulfonate

MM

Poly(methyl methacrylate)

MM

Polysulfone

• Degraded recalcitrant pollutants after modification • Increased hydrophilicity • Increased water permeability up to 2 wt% • Decreased solute rejection up to 2 wt% • Increased surface roughness • Reformed surface hydrophilicity • Enhanced mechanical stability • Increased water flux (62%) with improved selectivity and sensitivity • Retained Na2 SO4 (99%) for desalination • Increased fouling resistance

MM

Polysulfone

• Enhanced water flux (60–100%)

[91]

MM

Polysulfone

[92]

MM

Polysulfone

• Enhanced thermal stability • Increased heavy metal rejection • Increased water flux up to 160% • Reduced fouling tendency • Improved self-cleaning ability • Retained salts (R(Na2 SO4 ) > R(MgSO4 ) > R(NaCl)) in desalination

[94]

MM

Polyvinylidene fluoride

[83]

[84]

[85]

[86]

[74]

[88] [89]

[90]

[93]

4.4 Membrane Desalination

to the pore area and size of graphene membranes. On the other side, the flux of conventional polymeric membranes is inversely proportional to the membrane thickness. The similar argument seems invalid for the GOMs. Afterward, Hu and coworkers [109] fabricated GOMs with numerous layers and examined the water purification performance and found no obvious correlation between water flux and the number of layers in the membrane. After that, Han and coworkers [46] investigated a base-reflux-induced reduced GO and suggested that water flux enhanced linearly with pressure and reduced with GO loading contents. Therefore, the aforementioned two features showed contradictory conclusions on the relationship between GO thickness and water flux, and there is a requirement to comprehensively investigate this aspect. Graphene and GOMs are also intensively studied for the rejection of ion from the water. Here, the selectivity is primarily ascribed to size exclusion and interactions (including chemical and electrostatic ones) with functional groups of membranes. To prove this feature, Grossman and coworkers proposed the high separation performance of NaCl from water using monolayer graphene membrane by computational studies [24]. The computational investigations revealed that the small pores, low pressure, and hydrophobic pores rejected salts more efficiently owing to size exclusion, while the larger effective volume of ions and the lack of H-bonding led to a higher energy barrier to ionic passage. Hu and coworkers reported that GOMs showed noteworthy rejection of monovalent and divalent salts due to robust electrostatic interactions and size exclusion [109]. Similarly, Xu and coworkers [110] demonstrated GO’s selective permeation of different metal salts and reported quick permeation of sodium salts, slow permeation of heavy metal salts, and no permeation of copper ions and organic impurities as shown in Figure 4.11. A large number of oxygen-containing functionalities were grafted on the surfaces and edges of the GO sheets, and the GO sheets were regarded as large organic molecules. The electrostatic attractions and chemical interactions between these functionalities and hydrated ions are responsible for the selective penetration of GOMs indicated by the rejection experimental results. Sun and coworkers fabricated nano- and micron-sized porous GOMs as cation exchanger for acid recovery from FeCl3 solution [111]. The nano-GOM exhibited improved penetration of both cations including H+ and Fe3+ . H+ transported through the nano-GOM about two times higher than Fe3+ , and no Fe3+ transportation was observed at very low concentrations (0.01 mol l−1 ). Huang and coworkers [45] employed a GOM to separate a dimethyl carbonate–water mixture through a pervaporation process. The separation technology was believed to obey a sorption–diffusion mechanism. Owing to their exceptional intrinsic mechanical strength, high chemical stability, high antibacterial activity, and exquisite antifouling characteristics, GOMs proved substantial promise for applications in water desalination [112–115]. By regulating the pore size and nanochannels via induction and chemical functionalization of surface oxygen-containing groups, GOMs can be directed to a narrow pore size distribution, which is beneficial for specific sieving based on the size exclusion mechanism.

139

140

4 Microporous Membranes for Water Purification

Na+

(a)

Mn2+ Cd2+ Cu2+ Na+ SO42–

(b)

(d)

(e)

10 μm

(c)

(f)

(g)

~1.1 nm 500 nm

20 μm

1 μm

Figure 4.11 (a) Schematic representation of the penetration processes of different ions through GO membranes, (b) GO colloidal suspension (2 mg ml−1 ) used in the penetration experiments, (c) AFM image of the GO flakes (the inset shows the height profile of a layered GO sheet), (d) photograph of a freestanding GO membrane prepared by drop casting of a 2 mg ml−1 GO suspension, SEM images of GO membrane (e) surface and (f ) a cross-sectional view, and (g) the enlarged view of panel f [110]. Source: Copyright 2013, ACS.

The development of “thin-film composite (TFC) membrane” was a major breakthrough in the field of RO and water desalination. This membrane contains two layers: the upper layer, an active PA film fabricated by the reaction between m-phenylenediamine (MPD) and TMC on a microporous polysulfone substrate (PSU), and the bottom layer [116, 117], as shown in Figure 4.12. Experimental observations concluded that the permeance or water flux through the composite membranes mainly depends upon the hydrophilicity of the membrane surface and the physical features of the porous substrate, while salt rejection performance significantly relies on the surface charges and PA structure [118]. Because membrane capability is extensively dependent on a thin-film structure and its chemical properties as mentioned before, different alternative monomers have been used to fabricate the active layer PA and gives different results [116, 119–127]. For example, Li and coworkers [128] used three distinct isomeric biphenyl acid chlorides (mm-BTCE, om-BTCE, op-PTCE) to react, separately, with MPD on a porous substrate. Consequences indicated that the membrane fabricated from op-PTCE showed higher water permeance and lower salt rejection efficiency, while those produced from mm-BTCE and om-BTCE revealed lower water permeance and higher salt rejection. The intention behind improvement of permeability could be owing to the high

4.4 Membrane Desalination

PA active layer

PA PA

1 μm PES

PES microporous support 500 nm

PES

1 μm

Figure 4.12 Structure and morphology of a polyamide (PA) thin-film composite (TFC) membrane [117]. Source: Copyright 2016, Springer Nature.

density of the carboxyl functionalities on the membrane’s surface prepared from op-PTCE, which directed to better affinity with water. Instead, the higher salt rejection might be because of the thicker PA layer formed by mm-BTCE and om-BTCE reaction. Another investigation was reported by Wang and coworkers, who fabricated more hydrophilic, thinner, and continuous membrane using 3,5-diamino-N-(4-2-aminophenyl)benzamide (DABA) and TMC through interfacial polymerization (IP) approach [129]. On the other hand, one challenge facing the TFC membrane is the degradation of the PA layer by chlorination [130]. In this perspective, Liu and coworkers [131] used three different types of polyacyl chloride monomers to fabricate the TFC membrane with high stability under chlorine environment. Experimental observations revealed that the membrane prepared from MPD-CFIC and MPD-TMC possessed improved chlorine stability as compared with MPD-ICIC membrane. It has been pointed out that the urea bond (NHCONH-) in MPD-ICIC could be easily attacked by chlorine moieties. Recently, a TFC membrane with high chlorine resistance has produced IP of hexafluoroalcohol (HFA)-aromatic diamine and TMC [132]. The steric and electron-withdrawing properties of HFA functionalities lessened the chance of chlorine attack on the stable benzene rings or amide groups in the PA layer, which makes it more durable with high chlorine tolerance. Likewise, several treatments have altered/improved the structural features and characteristics of TFC membrane technology to make more efficient material for water purification and desalination. These treatments could be (i) appropriate selection of monomers for designing PA layer, (ii) functionalization of membrane surface, (iii) optimization of polymerization reaction, and (iv) incorporation of different sizes and types of NPs into PA layer. Jeong and coworkers [65] reported that doping NaA zeolite NPs as filler into the PA matrix could increase water permeability without decreasing salt rejection. Superior hydrophilicity, high negative surface charge, and internal pores of zeolite nanomaterials assisted water diffusion across the membrane while maintaining high salt rejection by Donnan exclusion. In another study by Fathizadeh and coworkers [64], they fabricated thin-film

141

142

4 Microporous Membranes for Water Purification

nanocomposite (TFN) membrane by increasing MPD and TMC concentrations (3 and 0.15% w/v) and zeolite NPs doped in membrane with higher water flux or permeance but lower NaCl rejection. The low solute rejection was ascribed to the poor dispersion of NPs and aggregation in the active PA layer, generating micro-holes, which caused the brackish water to pass through. The observation suggested that the relation between NPs and IP condition is also another significant factor that needs to be addressed. Baoling and coworkers fabricated TFN membrane using porous MCM-41 NPs as filler and MPD–TMC as matrix solution by in situ IP process. Experimental observations indicated that the MCM-41 NPs uniformly dispersed in PA layer and enhanced membrane performances under optimal concentrations. By increasing concentration of MCM-41 NPs (0–0.1 wt%), hydrophilicity, roughness, and zeta potential of TFN membranes all were increased. Particularly, the permeate water flux improved from 28.571 to 46.671 l m−2 h−1 with the incorporation of filler while maintaining high rejections of NaCl and Na2 SO4 (97.9% and 98.6%, respectively) [133]. Afterward, the same group prepared TFN membrane using GO nanosheets as filler in the same matrix solution (MPD–TMC) by in situ IP approach [134]. Experimental results revealed that the GO nanosheets were uniformly dispersed in the PA thin film and significantly improved membrane’s desalination performances. As the concentration of GO from 0 to 0.015 wt% increased during the fabrication, the permeate flux under 300 psi improved from 39.0 ± 1.6 to 59.4 ± 0.4 l m−2 h−1 , while rejections of NaCl and Na2 SO4 slightly decreased from 95.7% to 93.8% and 98.1% to 97.3%, respectively. This was because the interlayer distance of GO nanosheets had crucial role as water channels and hence contributed to the water permeability improvement. Islam Aljundi fabricated TFC membrane using ZIF-8 with MPS-TMC organic solution. The addition of ZIF-8 in the membrane active layer is to significantly improve the fouling resistance of membrane because fouling resistance improvement of RO membranes is still required in the desalination industry. The contact angle of 33.2∘ for obtained membrane was much lower than that of the control membrane (62.8∘ ), suggesting higher surface hydrophilicity. Therefore, the permeate water flux was increased to 53% while maintaining high salt rejection of 99.4%. This investigation revealed that ZIF-8 NPs can significantly improve the PA fouling resistance of membrane [135]. Similarly, Ma and coworkers fabricated TFN membrane using UiO-66 (MOF) NPs in the PA selective layer, which significantly alters the morphology and chemistry of the membrane, leading to an enhancement of intrinsic separation characteristics owing to the molecular sieving and superhydrophilic nature of MOF. Incorporation of 0.1 wt% UiO-66 nanofiller produced an ultimate water flux of 40 and 25% over the TFC control under FO modes (52% increase of water permeability) and also maintained salt rejection levels (∼95%) [136]. However, researchers have failed to find an alternative material to the PA barrier layer or to suggest a new support layer for nucleation. We do agree that PA above PSU/PES shows robust efficiency in RO and other water treatment applications, but these membranes have been used few decades before, and scholars have successfully addressed almost all the challenges facing the progress of such membranes.

4.4 Membrane Desalination

Zeolites are aluminosilicate minerals with a microstructure composed of 3–8 Å pore diameters. Zeolite crystals occur naturally or can be synthesized in a laboratory via a high temperature furnace [137]. Crystal sizes can be controlled from a few nanometers to centimeters by changing synthesis conditions (temperature and time) [137]. Properties such as adsorption capabilities, geometry, ion-exchange characteristics, and catalytic behavior can be tailored to a precise application using the optimal composition. Similarly, the porosity of zeolites varies between 30 and 40% [137]. Most zeolites have a tight pore distribution less than the diameter of a hydrated salt ion; a membrane produced from zeolites has the potential to fully reject salt ions while approving water molecules to permeate through. In this perspective, MD simulations have delivered mechanistic awareness into these processes. Murad and coworkers investigated NaCl/water separation using a single ZK-4 zeolite material with 4.4 Å pore sizes [138, 139]. The solvated ions are too large to pass through the pores and trapped with zeolite network, and only water molecules permeate through the zeolite. These simulation studies have motivated researchers to fabricate a variety of zeolite-based membranes for RO, achieving high mechanical stability, suitable fouling resistance, high flux, and exceptional ion rejection. Thus, Li and coworkers developed 0.5–3 μm thick membranes consisting of hydrophobic zeolites with an average pore size of 5.6 Å on a porous α-alumina substrate using hydrothermal synthesis approach (Figure 4.13a) [142]. Experimental observations revealed that the obtained membranes rejected 76% of Na+ ions while permitting a water flux of 0.112 kg m−2 h−1 under 2.07 MPa (20.7 bar) and with 0.1 M NaCl feedwater. Subsequently, the same group reduced the silicon/aluminum concentration ratio of the zeolite, which lowered the hydrophobicity and increased the flux to 10.21 kg m−2 h−1 under 3.5 MPa (35 bar) and 0.1 M NaCl feedwater [140]. Ion rejection was also dramatically enhanced from 76 to 98.6% (Figure 4.13b). Conversely, as the salt concentration increased, the salt rejection extensively decreased to 90% for a 0.3 M NaCl tested solution [141]. These consequences suggest that the salt rejection ability relies on the formation and size of intercrystalline defects (Figure 4.13c and d). Furthermore, nano-sized intercrystalline defects control the majority of ion transport and offer a challenge to the zeolite membrane fabrication technology. Jeong and coworkers prepared TFN membranes using cubic Linde type A (LTA) zeolite crystals (100 nm) with pore size of 4.1 Å by IP process [65]. This hydrophilic zeolite created preferential flow paths for water to permeate through and also improved ion transport rejection. However, even with the optimal zeolite loading of approximately 40%, the water flux was found to be lower than some commercial state-of-the-art RO membranes. Even though further work is needed to optimize the zeolite-loaded membrane and synthesis strategies, the permeability and salt rejection results show the potential of zeolite-based membranes for RO desalination technology. However, a challenge with the zeolite membranes mentioned above is that water molecules and solvated ions can permeate around the zeolite microstructure. In this configuration, it is difficult to determine the role that zeolites have in water permeability and salt rejection. Additionally, if salt can transport

143

4 Microporous Membranes for Water Purification

α-Alumina support 1 μm

(c)

Zeolite pore

o

Intercrystal defect

(b) 100 95 90 85 80 75 70 65 60 0.001

25 20 15 10 Ion rejection

5

Water flux

Water flux (mol m–2 h–1)

Active layer

0 0.01 0.1 Ion concentration (M)

1

(d)

1

Potential

(a)

Rejection (%)

144

2

Intercrystal defect

Figure 4.13 (a) Cross-sectional view of a zeolite membrane grown directly on porous α-alumina support [140], (b) experimental results of ion rejection and water flux for directly grown zeolite membranes as a function of ion concentration [141], (c) schematic of intercrystalline pore structure of directly grown zeolite membranes [142], and (d) schematic of overlapping of the EDL between zeolite crystals at low salt concentration (1) and high salt concentration (2) [142]. Source: Copyright 2009, 2004, 2008, Taylor & Francis and Elsevier.

around the zeolite crystal, the desired selectivity cannot be achieved. Further research is needed to determine the specific transport mechanisms within the sub-nanometer zeolite pores and in fabricating membranes that limit the transport to the intrinsic zeolite pores. These types of studies will help clarify the benefit of zeolites for water desalination. The major findings related in this section are summarized in Table 4.5.

4.5 Membrane Surface Engineering In current water purification technologies, UF, MF, NF, and RO membranes required subsequent cleaning to remove or reduce the effect of foulant after the purification treatment. Generally, there are two major techniques adapted to lower the effect of membrane fouling: (i) membrane surface decoration using NPs and (ii) modification of hydrophobic polymer with hydrophilic components. The hydrophilic surfaces exhibit high surface energy than hydrophobic ones. Therefore, hydrophobic surfaces can be the growth sites for foulants during cross-filtration experiments.

Table 4.5 Efficiency of various porous membranes in water desalination with respect to their pore diameters. Membrane types

Pore size (nm)

Types of ions +

−

Efficiency (%)

Permeance/water flux −2

−1

CNT-PcH

0.4–1.0

Na and Cl

99.99

24–29 l m

CNTs

0.32, 1.5

CdS, Au, and TiO2 nanoparticles

80, 100, and 100, respectively

—

[31]

CNTs

0.32, 0.49, 0.59, 0.69, and 0.75

Na+ and Cl−

100, 100, 95, and 58, respectively

—

[34]

Graphene-CNTs

0.69, 0.96, 1.2, 1.5, and 1.8

Na+ and Cl−

84, 84, 83, 83, and 82, respectively

1593, 1638, 1689, 1765, and 1846 l m−2 h−1 , respectively

[143]

CNTs-PA

1.5

Na+ and Cl−

98.4–98.7

25.8–31.4 l m−2 h−1

[144]

SWCNT

0.34–0.61

Na+ , K+ , and Cl−

100–90

—

[145]

Functionalized CNTs

1.1

Na+ and Cl−

28 and 86

23.1–107.8 l m−2 h−1

[146]

CNTs

0.4, 0.6, 0.7, 0.8, 1.0, and 1.4

Na+

100,100,100, 100, 35, and 5, respectively

574,744, 936, 1143, 1380, and 1643 l m−2 h−1 , respectively

[147]

Monolayer graphene

1.3

NaCl, MgSO4

20 and 70

3.1 × 10−4 and 4.13 × 10−5 l m−2 h−1

[148]

Nanoporous graphene

0.15–0.62

NaCl

h

Reference

−2

−1

[22]

60–100 at different applied pressures

10–100 l m h at different applied pressures

[24]

PVDF

—

NaCl

92

5–28 kg m−2 h−1

[20]

Aquaporin reconstituted

—

NaCl and MgCl2

66.2 and 88.1

6–30.1 kg m−2 h−1

[21] [109]

GO membrane

10–30

Salt cations

6–46

27.6 l m−2 h−1 bar−1

GO-PAN

123 nm to 0.8 μm

Na2 SO4

56.7

8.2 l m−2 h−1 bar−1

[149] (Continued)

Table 4.5 (Continued) Membrane types

Pore size (nm)

Types of ions

Efficiency (%)

Permeance/water flux

Reference

rGO

0.86

Cu2+ , Na+

Cubbi

—

Br− , Cl− , SO4 2− , and NO3 −

100

12.0 l m−2 h−1 bar−1

[150]

83, 59, 33, and 81 respectively

2.8–5.7 l m−2 h−1 bar−1

[27]

Zr-MOF

0.6–1.0

Al3+ , Mg2+ , and Ca2+

99.3, 98, and 86.3, respectively

0.28 l m−2 h−1 bar−1

[151]

MCM-41-PA-TFN GO-PA-TFN

3.03

NaCl, Na2 SO4

(97.9 ± 0.3%), (98.5 ± 0.2%)

46.6 ± 1.1 l m−2 h−1

[133]

0.83

NaCl, Na2 SO4

(93.8 ± 0.6%), (97.3 ± 0.3%)

59.4 ± 0.4 l m−2 h−1

[152]

K+ -controlled GO

1.14

Mg2+ , Ca2+ , Na+

∼100

0.36 l m−2 h−1

[24]

Monolayer graphene

—

K+ , Na+ , Li+ , Cl−

∼100

106 g m−2 s−1

[25]

CNTs, carbon nanotubes; CNT-PcH, carbon nanotube incorporated polyvinylidene fluoride-co-hexafluoropropylene nanofiber membrane; PA, polyamide; SWCNT, single-walled carbon nanotube; PVDF, polyvinylidene fluoride; GO, graphene oxide; PAN, polyacrylonitrile; rGO, reduced graphene oxide; MCM41-PA-TFN, MCM-41 silica nanoparticle enhanced polyamide thin-film nanocomposite membrane; and Cubbi , thermotropic bicontinuous cubic membrane.

4.5 Membrane Surface Engineering

4.5.1

Membrane Surface Engineering Using Nanoparticles

Numerous nanostructures have been determined to exhibit promising antifouling agents, in particular against microorganisms. For example, NPs including silver, titanium dioxide (TiO2 ), zinc oxide (ZnO), and copper (Cu) and even conducting NPs like GO and CNTs are known to have antibacterial activity [153]. Surface modification of membrane using NPs can be achieved through two ways as shown in Figure 4.14. NPs can be physically attached to the membrane surface termed as “coating” while as through the chemical bond is called “grafting.” Grafting is more persistent compared with coating due to less time leaching of NPs. The compiled recent literature stating the antimicrobial activity of NPs for modifying membranes is summarized in Table 4.6. The biocidal activity of the silver is renowned to the scientific community from ages old [165]. First evidence for the usage of silver is from Greek times; ancient people were aware that silver could disinfect water and prevent milk from spoilage [166]. In this perspective, Gibbard [167] first reported the biocidal activity of silver using an agar medium containing Balantidium coli. Silver is one of the extensively studied antimicrobial agents in the membrane research [154, 168]. Among all types of NPs, silver has the broad antimicrobial spectrum, making it one of most promising elements. This exhibits resistance to E. coli and Staphylococcus and to an extent to some virus strains dramatically [169]. Silver NPs can be grafted or coated over the membrane surfaces to disclose antimicrobial effects. In recent times, several researches explored the effect of silver NPs against microorganism as antifoulant on the membrane surfaces, and synergistic effect of silver NP together with other NPs was also well examined [155, 170, 171]. TiO2 , an economical antimicrobial agent, has full biocidal activity against gram-negative and gram-positive bacteria. In recent times, Kubacka and coworkers reported various investigations using titania NPs and also described the mechanism of titania microbial action using Pseudomonas aeruginosa in Surface modification

Coating

Grafting

Nanoparticle coating

Nanoparticle

Membrane

Chemical bonding Modification by blending

Modification by chemical treatment

Figure 4.14 Schematic representation of membrane modification using nanoparticles.

147

Table 4.6 Recent studies on the surface modification for membrane antifouling.

Membrane

Modification agent

Fouling agent

Water contact angle (∘ ) before and after

Flux before fouling −2

−1

−2

∼0.4 l m

Reference

−1

[154]

PVDF

Silver

Escherichia coli

89

61

1.1 l m

PS

Silver

E. coli

76.8 ± 4

68.6 ± 6

408 ± 180 l m−2 s−1 bar−1

532 ± 117 l m−2 s−1 bar−1

[155]

PVDF

TiO2

E. coli

∼92

∼64

1068 kg m−2

616 kg m−2

[156]

PE

Chitosan

E. coli

92

64

4062 ± 266 l m−2 s−1

4749 ± 252 l m−2 s−1

[157]

PS

Silver

E. coli

∼77

∼30

649 LMH

—

[158]

PES

ZnO

E. coli

80

∼50

0.7 l m−2 s−1

50 l m−2 s−1

[159]

PVDF/PEG

TiO2

Soluble microbial products

74.4

70

72.8 l m−2 s−1

58.8 l m−2 s−1

[160]

PVDF

Polystyrene-b-poly (ethylene glycol) methacrylate

E. coli and Staphylococcus epidermidis

85 ± 1

59 ± 2

—

—

[161]

PES

Glycosyl

BSA

—

—

—

—

[162]

PES

Poly(2-hydroxyethyl methacrylate) (PHEMA)-grafted silica (SiO2 )

BSA

85

66

90 l m−2 s−1

180 l m−2 s−1

[163]

PVDF

Polyacrylic acid

BSA

91

52

90 l m−2 s−1

145 l m−2 s−1

[164]

a)

s

Flux after modification

s

BSA, bovine serum albumin; PE, polyethylene; PEG, polyethylene glycol; PES, polyethersulfone; PS, polysulfone; PVDF, polyvinylidene fluoride.

4.5 Membrane Surface Engineering

the titania–ethylene vinyl alcohol (EVOH) composite-based materials [172]. These activities coupled with cell wall engineering made the titania NPs more promising biocidal agents. In this perspective, Qin and coworkers fabricated the PVDF membrane incorporated with titania NPs and polyvinyl alcohol (PVA) as cross-linker. TiO2 NPs were uniformly distributed through chemical linkage with PVDF and PVA matrix solution. This significantly enhanced the hydrophilic behavior, and water contact angle reduced from 84∘ for control to 24∘ for obtained membrane. Moreover, there was a gradual decrease in water contact angle with the increased titania that then reduced to 24∘ when PVA cross-linker was introduced [173]. They also studied the antifouling efficiency and pure water permeability of the prepared hydrophilic membranes. It has been determined that a substantial increase in flux recovery ratio (i.e. 80.5%) was achieved for modified membrane than that of control PVDF membrane. Li and coworkers investigated the impact of TiO2 NP over the PES membrane morphology and rejection performance. They observed that the TiO2 NPs increased the pure water flux/permeance, elongation at break, and average pore diameter and reduced the water contact angle and porosity because of the scramble in total solids content. The titania contents greatly affected the morphology of the PES membrane. Furthermore, the better antifouling characteristic correlated with the surface roughness of the membranes. Results showed that the higher the surface roughness, the lower was the fouling (more hydrophilic). Therefore, the hydrophilicity is the reason for gaining less flux reduction coefficient and high roughness in TiO2 -modified membranes as compared with control membranes [174]. Recently, several groups reported the proper understanding and correlation among antifouling property and mechanical and chemical behavior of titania NP-doped membranes [175]. Copper is one of the potential elements present in almost all living organisms in the form of enzymes, proteins (e.g. cytochrome C), and certain superoxide dismutase [176]. Though, it was later delimited because of the toxic effect of the same in harbor due to enhanced leaching [177]. In this perspective, Cioffi and coworkers [178] investigated the antibacterial and antifungal properties of copper nanoparticles (Cu NPs) incorporated in three distinct polymer matrix membranes using sacrificial anode electrolysis. The obtained polymer–copper nanocomposite displayed the antibacterial and antifungal activities. Afterward, numerous reports have been exploited on the antimicrobial and antifungal activities of the copper [179]. On the other hand, few reports are published on the bactericidal property of Cu NPs in distinct polymer membranes [180]. The antibacterial activity of ceramic powders such as ZnO, MgO, and CaO has recently fascinated the attention of several researchers because of their remarkable antibacterial property. Inorganic nano-dimensional metal oxides are typically important because along with antimicrobial characteristics, they provide very high surface area and biocompatibility. Therefore, these biocidal agents show advantages including low toxicity to human being, durability, and greater selectivity toward microorganisms. The antibacterial activity of ZnO, MgO, and CaO are owed to the formation of reactive oxygen species on the surface of the particle. In this domain, Sawai and coworkers reported the antibacterial efficiency of magnesium oxide nanostructure. They found that the active oxygen species (O2 − )

149

150

4 Microporous Membranes for Water Purification

in the slurry could eventually inactivate the bacteria. Furthermore, more shaking can also kill microbes by the mutual contact with MgO and microbes’ cell membranes [181]. Another report by Nagarajan and coworkers examined the effect of particle size of ZnO NP on its biocidal activity. They observed that NP-sized ZnO displayed much improved bactericidal activity as compared with micron-sized ZnO [182]. Recently, Liang and coworkers fabricated a ZnO-incorporated PVDF membrane and examined the effect of ZnO NP loading on the antifouling effectiveness. Experimental observations revealed that addition of ZnO varied the membrane morphology analogous to that of TiO2 and an addition of 6.7% raised the membrane flux and also altered the fingerlike cross-sectional morphology into macro-cavities [183]. Moreover, water contact angle of the membrane reduced with the rise in ZnO contents. Moreover, Hong and coworkers [184] also made the similar observation, wherein the incorporation of a higher amount of the ZnO increased the PVDF dope viscosity, which dramatically reduced the precipitation rate of membrane during phase inversion process. They also found that the antifouling resistance and surface roughness increased with the amount of ZnO. Therefore, surface roughness influenced two vital effects on membrane surface: (i) high surface roughness raised the effective area of filtration and increased the permeability, and (ii) high surface roughness was hydrophilic in nature, lessening the surface fouling. Likewise, Sawai observed that the behavior of these metal oxides was different for different microorganisms. For example, CaO was more effective against E. coli (followed by MgO and then ZnO), whereas ZnO was more effective against Staphylococcus aureus microorganism [185]. Graphene invention was a big breakthrough by Andre Geim and Konstantin Novoselov who won Nobel Prize in 2010 for this achievement. GO is a one-atom-thick graphene layer comprising carboxyl and epoxide functionalities and shows better tensile strength and conducting and amphiphilic properties [186]. Recently graphene-based materials confirmed to be efficient antibacterial, antifouling, and less cytotoxic, making them a promising candidate for biological and medical applications. Thus, numerous studies have been reported on the incorporation of GO in the polymeric membranes for biological applications [187, 188]. For example, Wang and coworkers [189] fabricated GO-PVDF membranes to study antifouling property. Experimental observations revealed that the addition of GO raised the hydrophilicity, water permeability, flux recovery ratio, and tensile strength of the prepared membranes [190]. Similarly, Zhao and coworkers [191] fabricated UF PVDF membrane and studied the effect of GO concentration on the various characteristics including porosity, average pore size, flux recovery ratio, permeability, and antifouling. Analogous observations were reported by other research groups [192] stating the synergistic effect of modified GO in the PVDF membranes. Experimental observations revealed that the GO immobilization on the membrane’s surface led to the long-term antifouling and antibacterial efficacy and removal of the secondary pollution caused by the uncontrolled release of bactericide from the membrane [193]. In summary, this exceptional 2D material is a potential candidate for water filtration applications and proven to be effective separation of biological contaminants such as E. coli and S. aureus [194].

4.5 Membrane Surface Engineering

Like graphene, CNTs have also acquired a lot of attention owing to their striking antibacterial and antifouling activity. Similar to graphene, CNTs can also break the bacterial cell wall through direct contact or endocytosis. In this domain, Wallace and coworkers [195] examined the interactions between lipid bilayer and CNTs and suggested the usage of CNTs as nano-injector to transport therapeutic agents into the cells. Similar observations were reported for the molecular transport by CNTs [196]. On the other hand, like GO, CNT membranes can also be applied for water filtration and desalination. In this perspective, Lee and coworkers [197] fabricated the millimeter-thick CNT membrane representing high water permeability, less biofilm formation, and enhanced rate of rejection. The CNT membranes can be excellent filters due to their lightweight, optimal pore size, high surface area, ease of cleaning, and high thermal stability. Furthermore, the membrane fabricated by Lee and coworkers shows an anti-adhesion activity to E. coli and also inhibits the biofilm creation on the surface of the membrane [197]. Brady-Estévez and coworkers employed the CNT membranes for size-selective filtration [198]. Currently, incorporation of CNT in the polymer matrix is found to impart excellent mechanical and antibacterial behavior. Sengur-Tasdemir and coworkers [199] examined the effect of concentration and functionalization of COOH-functionalized MWNTs in three distinct polymer membranes and observed enhanced fouling, bacterial resistance, improved hydrophilicity, and flux in all the prepared composites. 4.5.2

Using Hydrophilic Components to Hydrophobic Membranes

Hydrophobic membrane surfaces tend to foul more and therefore reduce the water permeate flux. It is one of the serious problems of hydrophobic membranes with the mutual attraction between the organic, inorganic, and biological contaminants. These pollutants will further clog on the membrane surfaces and also deposit in the pore cavities. Various approaches have been reported to achieve the hydrophilicity in the hydrophobic membranes. The approaches can be widely classified as (i) incorporation of NPs (as discussed above) and (ii) blending or copolymerization with hydrophilic polymers. Polymers including PEG, PVA, polyvinylpyrrolidone (PVP), and polyvinyl acetate (PVAc) are particularly applied to endorse the hydrophilicity in the hydrophobic membrane (e.g. PVDF and PES). In this perspective, Kanagaraj and coworkers [200] fabricated the polyetherimide (PEI) UF membrane with improved hydrophilicity using the incorporation of PVP as a modification agent. The hydrophilicity is determined from water contact angle measurements as shown in Figure 4.15. The presence of PVP in the membrane results in instantaneous demixing in the coagulation bath, decreases the thermodynamic stability of the active layer, and causes the creation of macropore in the membranes. Numerous reports observed similar consequences and can generalize that blending or copolymerization with the hydrophilic polymer will significantly enhance the surface wettability of hydrophobic membrane and the resistance against interaction with hydrophobic substances and other microbes [201]. Recently, some novel efforts to enhance the antifouling activity of the polymeric surfaces are gaining utmost attention.

151

152

4 Microporous Membranes for Water Purification

PVP 0

θ = 96.4°

PVP 6

PVP 2

θ = 60.0°

θ = 71.1°

PVP 8

PVP 4

θ = 62.2°

θ = 56.1°

Figure 4.15 Variation of hydrophilicity with a change in concentration of PVP [200]. Source: Copyright 2015, ACS.

For example, Messersmith and coworkers [202] explored the adhesive protein found in the mussel, which can be used for attachment to the different substrates such as dopamine. The highly hydrophilic dopamine has the advantage of averting the hydrophobic–hydrophobic interaction and can act as a protective coating on the substrate [203]. This protective characteristic is beneficial for membrane surface modification to reduce the fouling activity caused by the microorganisms after water purification. Afterward, the authors prepared the polydopamine coating on three different polymeric membranes and examined the antifouling activities after surface engineering [204]. Even though the membrane permeability slightly decreased after chemical modification, the hydrophilicity of the membranes improved with the coating time, which was very operative for antifouling activities [205]. In this domain, Xi and coworkers fabricated dopamine-coated polymeric membrane and examined improved durability and hydrophilicity by increasing dopamine concentration [206]. Subsequently, the decrease in permeability was resolved by the addition of pore-forming agents such as PEG and PVP, which can enlarge the pore size and facilitate the permeability even after polydopamine coating. Similarly, Shao and coworkers fabricated the PVDF/PD-TiO2 UF membrane in which polydopamine acted as coupling agent between TiO2 and PVDF and resulted in the increase in hydrophilicity of membranes [207]. Gu and coworkers prepared the antifoulant, antimicrobial, and less cytotoxic lysozyme protein extracted from the hen egg white. The lysozyme has been verified to be an excellent adhesive to several substrates and exhibited a broad spectrum of antibacterial efficiency toward both gram-negative and gram-positive bacteria as well as toward fungi [208]. This protein found antibacterial efficiency of 95, 92, and 94% toward E. coli, S. aureus, and Candida albicans, respectively. The improved antimicrobial effect is the characteristic of surface hydration as well as the positive charge of lysozyme protein nanofilm. Moreover, other attempts to design antifouling surfaces using

References

bioactive and bioinert polymer substrates are recently gaining attention [209]. Mechanistically, photodissociation of amine group in carbonyl compounds can abstract a proton from the polymer, resulting in the transfer of the free radical site. This free radical site can again recombine with amine group, thus forming primary amine-polymer bioactive surfaces. These bioactive surfaces can directly undergo bioconjugation reactions with target molecules and can be applied for biomedical applications. On the other hand, the bioinert surfaces produced by the same free radical-initiated polymer substrate can be taken for graft polymerization with antifouling monomers and can be used for membrane applications.

4.6 Conclusions In this chapter, various types of microporous membranes have been overviewed for the removal of water contaminants (organic, inorganic, and biological). This chapter also dealt with the type and state-of-the-art microporous membranes for water purification. Moreover, we mainly discussed the various porous membranes for desalination (CNTs, graphene, GO, TFC, and zeolite) and surface engineering approaches (NP incorporation and hydrophilic end groups) used in the recent years. With the increasing exploitation of water resources for domestic purpose and industry and in many parts of the world, there is growing need for improving efficient materials and purification/desalination technologies. In this perspective, several studies have been performed to design advanced functional materials and membranes for contaminant/water separation. Among these, carbon-based nanomaterials are capable to provide leap-forward opportunities to wastewater treatment. On the other side, the astonishing properties of membrane technology such as CNT-based membrane, graphene or GO-based membrane, and carbon nanofiber have proven their potential benefits to the practical implementation due to excellent physical and chemical characteristics as compared with polymeric and zeolite membranes. On the contrary, it faces some challenges regarding uniform pore size, uniform distribution, interfacial defects, and stability, thus inspiring surface engineering approaches to be carried out to overcome these problems. In our opinion, the development direction of the next generation of microporous membranes for water purification may comprise the following features such as high membrane selectivity, improved fouling resistance of porous membrane, stability under chlorine and ozone, and high separation efficiency. Although there are still many challenges to be overcome for the industrial production of low-cost and efficient membranes for water purification/desalination, we believe that it will be realized in the near future under the joint efforts of scientists, industrial engineers, and investors.

References 1 Camargo, J.A. and Alonso, A. (2006). Environ. Int. 32: 831–849. 2 Schaider, L.A., Rudel, R.A., Ackerman, J.M. et al. (2014). Sci. Total Environ.

468: 384–393.

153

154

4 Microporous Membranes for Water Purification

3 Tchounwou, P.B., Ayensu, W.K., Ninashvili, N., and Sutton, D. (2003).

Environ. Toxicol. 18: 149–175. 4 Daniel, S., Limson, J.L., Dairam, A. et al. (2004). J. Inorg. Biochem. 98:

266–275. 5 Miyahara, T., Tsukada, M., Mori, M.A., and Kozuka, H. (1984). Toxicol. Lett.

22: 89–92. 6 Goktas, R.K. and MacLeod, M. (2016). Environ. Pollut. 217: 33–41. 7 Lonnen, J., Kilvington, S., Kehoe, S.C. et al. (2005). Water Res. 39: 877–883. 8 Matlock, M.M., Howerton, B.S., and Atwood, D.A. (2002). Water Res. 36:

4757–4764. 9 Ali, I. and Gupta, V.K. (2006). Nat. Protoc. 1: 2661–2667. 10 Rengaraj, S., Yeon, K.H., and Moon, S.H. (2001). J. Hazard. Mater. 87:

273–287. 11 Zularisam, A.W., Ismail, A.F., and Salim, R. (2006). Desalination 194:

211–231. 12 Vasudevan, S. and Oturan, M.A. (2014). Environ. Chem. Lett. 12: 97–108. 13 Rezania, S., Ponraj, M., Talaiekhozani, A. et al. (2015). J. Environ. Manag. 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

163: 125–133. Upadhyay, R.K., Soin, N., and Roy, S.S. (2014). RSC Adv. 4: 3823–3851. Baker, R.W. (2012). Membrane Technology and Applications. Wiley. Gin, D.L. and Noble, R.D. (2011). Science 332: 674–676. Shaffer, D.L., Werber, J.R., Jaramillo, H. et al. (2015). Desalination 356: 271–284. Anthony, G.F., Wang, R., and Matthew, X.H. (2015). Angew. Chem. Int. Ed. 54: 3368–3386. Kim, S.H., Kwak, S.Y., and Suzuki, T. (2005). Environ. Sci. Technol. 39: 1764–1770. Feng, C., Khulbe, K.C., Matsuura, I. et al. (2008). J. Membr. Sci. 311: 1–6. Sun, G.F., Chung, T.S., Jeyaseelan, K., and Armugam, A. (2013). Colloids Surf. B 102: 466–471. Tijing, L.D., Woo, Y.C., Shim, W.G. et al. (2016). J. Membr. Sci. 502: 158–170. Chen, L., Shi, G.S., Shen, J. et al. (2017). Nature 550: 380–383. Cohen-Tanugi, D. and Grossman, J.C. (2012). Nano Lett. 12: 3602–3608. Surwade, S.P., Smirnov, S.N., Vlassiouk, I.V. et al. (2015). Nat. Nanotechnol. 10: 459–464. Ren, C.E., Hatzell, K.B., Alhabeb, M. et al. (2015). J. Phys. Chem. Lett. 6: 4026–4031. Henmi, M., Nakatsuji, K., Ichikawa, T. et al. (2012). Adv. Mater. 24: 2238–2241. Bolisetty, S., Reinhold, N., Zeder, C. et al. (2017). Chem. Commun. 53: 5714–5717. DesOrmeaux, J.P.S., Winans, J.D., Wayson, S.E. et al. (2014). Nanoscale 6: 10798–10805. Zhang, Q.G., Deng, C., Soyekwo, F. et al. (2016). Adv. Funct. Mater. 26: 792–800.

References

31 Li, H.B., Gui, X.C., Zhang, L.H. et al. (2010). Chem. Commun. 46:

7966–7968. 32 Karim, Z., Mathew, A.P., Grahn, M. et al. (2014). Carbohydr. Polym. 112:

668–676. 33 Elimelech, M. and Phillip, W.A. (2011). Science 333: 712–717. 34 Corry, B. (2008). J. Phys. Chem. 5: 1427–1434. 35 Bakajin, O., Noy, A., Fornasiero, F. et al. (2009). Nanofluidic carbon nan-

36 37 38 39 40 41 42 43

44 45 46 47 48 49 50 51 52 53 54 55 56 57 58

otube membranes: applications for water purification and desalination. In: Nanotechnology Applications for Clean Water (ed. A. Street, R. Sustich, J. Duncan and N. Savage), 77–93. New York: Elsevier. Lee, C. and Baik, S. (2010). Carbon 48: 2192–2197. Liang, H.W., Cao, X., Zhang, W.J. et al. (2011). Adv. Funct. Mater. 21: 3851–3858. Gu, J., Xiao, P., Zhang, L. et al. (2016). RSC Adv. 6: 73399–73403. Cao, K.T., Jiang, Z.Y., Zhang, X.S. et al. (2015). J. Membr. Sci. 490: 72–83. Zhang, W.B., Zhu, Y.Z., Liu, X. et al. (2014). Angew. Chem. Int. Ed. 53: 856–860. Obaid, M., Barakat, N.A.M., Fadali, O.A. et al. (2015). Chem. Eng. J. 259: 449–456. Ng, L.Y., Mohammad, A.W., Leo, C.P., and Hilal, N. (2013). Desalination 308: 15–33. Le, N.L., Duong, P.H.H., and Nunes, S.P. (2017). Advanced polymeric and organic-inorganic membranes for pressure-driven processes. In: Comprehensive Membrane Science and Engineering (ed. E. Drioli, L. Giorno and E. Fontananova), 120–136. Elsevier. Hsieh, C.T., Hsu, J.P., Hsu, H.H. et al. (2016). Surf. Coat. Technol. 286: 148–154. Huang, K., Liu, G.P., Lou, Y.Y. et al. (2014). Angew. Chem. Int. Ed. 53: 6929–6932. Han, Y., Xu, Z., and Gao, C. (2013). Adv. Funct. Mater. 23: 3693–3700. Tang, Y.P., Paul, D.R., and Chung, T.S. (2014). J. Membr. Sci. 458: 199–208. Gao, P., Liu, Z.Y., Tai, M.H. et al. (2013). Appl. Catal. B Environ. 138: 17–25. Sun, L.W., Huang, H.B., and Peng, X.S. (2013). Chem. Commun. 49: 10718–10720. Sun, L.W., Ying, Y.L., Huang, H.B. et al. (2014). ACS Nano 8: 6304–6311. Akbari, A., Sheath, P., Martin, S.T. et al. (2016). Nat. Commun. 7: 10891. Li, X., Wang, M., Wang, C. et al. (2014). ACS Appl. Mater. Interfaces 6: 15272–15282. Chen, X.F., Qiu, M.H., Ding, H. et al. (2016). Nanoscale 8: 5696–5705. Yang, S.Y., Park, J., Yoon, J. et al. (2008). Adv. Funct. Mater. 18: 1371–1377. Sato, A., Wang, R., Ma, H.Y. et al. (2011). J. Electron Microsc. 60: 201–209. Zhang, X.W., Zhang, T., Ng, J., and Sun, D.D. (2009). Adv. Funct. Mater. 19: 3731–3736. Khin, M.M., Nair, A.S., Babu, V.J. et al. (2012). Energy Environ. Sci. 5: 8075–8109. Kamat, P.V. and Meisel, D. (2003). C. R. Chim. 6: 999–1007.

155

156

4 Microporous Membranes for Water Purification

59 Canela, M.C. and Jardim, W.F. (2008). Environ. Technol. 29: 673–679. 60 Nevim, S., Arzu, H., Gulin, K., and Cinar, Z. (2001). J. Photochem. Photobiol. 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87

88 89 90 91 92

A Chem. 139: 225–232. Goh, P.S., Ismail, A.F., and Ng, B.C. (2013). Desalination 308: 2–14. Mishra, A.K. and Ramaprabhu, S. (2011). Desalination 282: 39–45. Tofighy, M.A. and Mohammadi, T. (2010). Desalination 258: 182–186. Fathizadeh, M., Aroujalian, A., and Raisi, A. (2011). J. Membr. Sci. 375: 88–95. Jeong, B.H., Hoek, E., Yan, Y. et al. (2007). J. Membr. Sci. 294: 1–7. Jadav, G.L. and Singh, P.S. (2009). J. Membr. Sci. 328: 257–267. Thostenson, E.T., Ren, Z., and Chou, T.W. (2001). Sci. Technol. 61: 1899–1912. Dai, K., Shi, L., Zhang, D., and Fang, J. (2006). Chem. Eng. Sci. 61: 428–433. Li, H. and Zou, L. (2011). Desalination 275: 62–66. Nasrabadi, A.T. and Foroutan, M. (2011). Desalination 277: 236–243. Das, R., Ali, M.E., Hamid, S.B.A. et al. (2014). Desalination 336: 97–109. Majumder, M., Chopra, N., Andrews, R., and Hinds, B.J. (2005). Nature 438: 44. Zimmerman, C.M., Singh, A., and Koros, W.J. (1997). J. Membr. Sci. 137: 145–154. Choi, J.H., Jegal, J., and Kim, W.N. (2006). J. Membr. Sci. 284: 406–415. Yang, H.Y., Han, Z.J., Yu, S.F. et al. (2013). Nat. Commun. 4. Mauter, M.S. and Elimelech, M. (2008). Environ. Sci. Technol. 42: 5843–5859. Li, J., Ng, H.T., Cassell, A. et al. (2003). Nano Lett. 3: 597–602. Kar, S., Bindal, R.C., and Tewari, P.K. (2012). Nano Today 7: 385–389. Qu, X., Alvarez, P.J., and Li, Q. (2013). Water Res. 47: 3931–3946. Majumder, M. and Corry, B. (2011). Chem. Commun. 47: 7683–7685. Van Hooijdonk, E., Bittencourt, C., Snyders, R., and Colomer, J.F. Beilstein Int. J. Nanotechnol. 3013 (4): 129–152. Hinds, B.J., Chopra, N., Rantell, T. et al. (2004). Science 303: 62–65. Srivastava, A., Srivastava, O.N., Talapatra, S. et al. (2004). Nat. Mater. 3: 610–614. Holt, J.K., Noy, A., Huser, T. et al. (2004). Nano Lett. 4: 2245–2250. Holt, J.K., Park, H.G., Wang, Y. et al. (2006). Science 312: 1034–1037. Du, F., Qu, L., Xia, Z. et al. (2011). Langmuir 27: 8437–8443. Wu, P. and Ima, M. (2012). Novel biopolymer composite membrane involved with selective mass transfer and excellent water permeability. In: Advancing Desalination (ed. R.Y. Ning), IntechOpen, 57–82. Brunet, L., Lyon, D.Y., Zodrow, K. et al. (2008). Environ. Eng. Sci. 2: 565–576. Shen, J.N., Yu, C.C., Ruan, H.M. et al. (2013). J. Membr. Sci. 442: 18–26. Kar, S., Subramanian, M., Pal, A. et al. (2013). AIP Conf. Proc. 1538: 181–185. Yin, J., Zhu, G., and Deng, B. (2013). J. Membr. Sci. 437: 237–248. Shah, P. and Murthy, C.N. (2013). J. Membr. Sci. 437: 90–98.

References

93 Amini, M., Jahanshahi, M., and Rahimpour, A. (2013). J. Membr. Sci. 435:

233–241. 94 Madaeni, S.S., Zinadini, S., and Vatanpour, V. (2013). Sep. Purif. Technol. 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127

111: 98–107. Huang, H., Ying, Y., and Peng, X. (2014). J. Mater. Chem. A 2: 13772–13782. Brodie, B.C. (1859). Philos. Trans. R. Soc. Lond. 149: 249–259. Hummers, W.S. and Offeman, R.E. (1958). J. Am. Chem. Soc. 80: 1339–1339. Novoselov, K.S., Geim, A.K., Morozov, S.V. et al. (2004). Science 306: 666–669. Tsetseris, L. and Pantelides, S.T. (2014). Carbon 67: 58–63. Gai, J.G., Gong, X.L., Wang, W.W. et al. (2014). J. Mater. Chem. A 2: 4023–4028. Liu, H., Dai, S., and Jiang, D. (2013). Solid State Commun. 175: 101–105. Goh, K., Setiawan, L., Wei, L. et al. (2015). J. Membr. Sci. 474: 244–253. Liu, R., Arabale, G., Kim, J. et al. (2014). Carbon 77: 933–938. Tao, Y., Xue, Q., Liu, Z. et al. (2014). ACS Appl. Mater. Interfaces 6: 8048–8058. Suk, M.E. and Aluru, N. (2010). J. Phys. Chem. Lett. 1: 1590–1594. Pendergast, T.M., Mary, T., and Hoek, M. (2011). Energy Environ. Sci. 4: 1946–1971. Wei, N., Peng, X., and Xu, Z. (2014). ACS Appl. Mater. Interfaces 6: 5877–5883. Mi, B. (2014). Science 343: 740–742. Hu, M. and Mi, B. (2013). Environ. Sci. Technol. 47: 3715–3723. Sun, P., Zhu, M., Wang, K. et al. (2013). ACS Nano 7: 428–437. Sun, P., Wang, K., Wei, J. et al. (2014). J. Mater. Chem. A 2: 7734–7737. Liu, Y., Xie, B., Zhang, Z. et al. (2012). J. Mech. Phys. Solids 60: 591–605. Dreyer, D.R., Park, S., Bielawski, C.W., and Ruoff, R.S. (2010). Chem. Soc. Rev. 39: 228–240. Hu, W., Peng, C., Luo, W. et al. (2010). ACS Nano 4: 4317–4323. Dikin, D.A., Stankovich, S., Zimney, E.J. et al. (2007). Nature 448: 457–460. Cadotte, J.E. (1979). Interfacially synthesized reverse osmosis membrane. US Patent 4277344. Khorshidi, B., Thundat, T., Fleck, B.A., and Sadrzadeh, M.A. (2016). Sci. Rep. 6: 22069. Porter, M.C. (1977). AIChE Symp. Ser. 73: 83–103. Nathaniel, G., Lim, J., and Jung, B. (2012). Desalination 291: 69–77. Xie, W., Geise, G., Freeman, B. et al. (2012). J. Membr. Sci. 404: 152–161. Zhang, Z., Wang, S., Chen, H., and Wang, T. (2013). Desalination 331: 16–25. Sum, J.Y., Ahmad, A., and Ooi, B.S. (2014). J. Membr. Sci. 466: 183–191. Li, Y., Su, Y., Dong, Y. et al. (2014). Desalination 333: 59–65. Zhao, J., Su, Y., He, X. et al. (2014). J. Membr. Sci. 465: 41–48. Jewrajka, S., Reddy, A.V.R., Rana, H. et al. (2013). J. Membr. Sci. 439: 87–95. Wang, T., Dai, L., Zhang, Q. et al. (2013). J. Membr. Sci. 440: 48–57. Yong, Z., Sanchuan, Y., Meihong, L., and Congjie, G. (2006). J. Membr. Sci. 270: 162–168.

157

158

4 Microporous Membranes for Water Purification

128 Li, L., Zhang, S., Zhang, X., and Zheng, G. (2008). J. Membr. Sci. 315: 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161

20–27. Wang, H., Li, L., Zhang, X., and Zhang, S. (2010). J. Membr. Sci. 353: 78–84. Koo, J., Petersen, R., and Cadotte, J. (1986). ACS Polym. Prepr. 27: 391–392. Liu, M., Wu, D., Yu, S., and Gao, C. (2009). J. Membr. Sci. 326: 205–214. La, Y., Sooriyakum, R., Miller, D. et al. (2010). J. Mater. Chem. 20: 4615–4620. Yin, J., Kim, E.S., Yang, J., and Deng, B. (2012). J. Membr. Sci. 423: 238–246. Yin, J., Zhu, G.C., and Deng, B.L. (2016). Desalination 379: 93–101. Aljundi, I.H. (2017). Desalination 430: 12–20. Ma, D., Peh, S.P., Han, G., and Chen, S.B. (2017). ACS Appl. Mater. Interfaces 9: 7523–7534. Cejka, J. (2007). Introduction to Zeolite Science and Practice, 3e, 1058. Elsevier. Murad, S.M. and Lin, J.C. (2002). Ind. Eng. Chem. Res. 41: 1076–1083. Murad, S., Oder, K., and Lin, J. (1998). Mol. Phys. 95: 401–408. Li, L.X. and Lee, R. (2009). Sep. Sci. Technol. 44: 3455–3484. Li, L., Liu, N., Pherson, B.M., and Lee, R. (2008). Desalination 228: 217–225. Li, L., Dong, J., Nenoff, T.M., and Lee, R. (2004). J. Membr. Sci. 243: 401–404. Goldsmith, J. and Martens, C.C. (2009). J. Phys. Chem. Lett. 1: 528–535. Chan, W.F., Chen, H.Y., Surapathi, A. et al. (2013). ACS Nano 7: 5308–5319. Song, C. and Corry, B. (2009). J. Phys. Chem. B 113: 7642–7649. Corry, B. (2011). Energy Environ. Sci. 4: 751–759. Jia, Y.X., Li, H.L., Wang, M. et al. (2010). Sep. Purif. Technol. 75: 55–60. O, Hern, S.C., Jang, D., Bose, S. et al. (2015). Nano Lett. 15: 2354–2360. Wang, J.Q., Zhang, P., Liang, B. et al. (2016). ACS Appl. Mater. Interfaces 8: 6211–6218. Liu, H.Y., Wang, H.T., and Zhang, X.W. (2015). Adv. Mater. 27: 249–254. Liu, X.L., Demir, N.K., Wu, Z.T., and Li, K. (2015). J. Am. Chem. Soc. 137: 6999–7002. Yin, J., Zhu, G.C., and Deng, B.L. (2016). Desalination 379: 93–101. Kang, S., Herzberg, M., Rodrigues, D.F., and Elimelech, M. (2008). Langmuir 24: 6409–6413. Sharma, M., Padmavathy, N., Remanan, S. et al. (2016). RSC Adv. 6: 38972–38983. Zodrow, K., Brunet, L., Mahendra, S. et al. (2009). Water Res. 43: 715–723. Rahimpour, A., Jahanshahi, M., Rajaeian, B., and Rahimnejad, M. (2011). Desalination 278: 343–353. Mural, P.K.S., Kumar, B., Madras, G., and Bose, S. (2016). ACS Sustain. Chem. Eng. 4: 862–870. Nisola, G.M., Park, J.S., Beltran, A.B., and Chung, W.J. (2012). RSC Adv. 2: 2439–2448. Jo, Y.J., Choi, E.Y., Choi, N.W., and Kim, C.K. (2016). Ind. Eng. Chem. Res. 55: 7801–7809. Zhang, J., Wang, Z., Zhang, X. et al. (2015). Appl. Surf. Sci. 345: 418–427. Venault, A., Liu, Y.H., Wu, J.R. et al. (2014). J. Membr. Sci. 450: 340–350.

References

162 Angione, M.D., Duff, T., Bell, A.P. et al. (2015). ACS Appl. Mater. Interfaces

7: 17238–17246. 163 Zhu, L.J., Zhu, L.P., Jiang, J.H. et al. (2014). J. Membr. Sci. 451: 157–168. 164 Zhao, X., Xuan, H., Qin, A. et al. (2015). RSC Adv. 5: 64526–64533. 165 Jose Ruben, M., Jose Luis, E., Alejandra, C. et al. (2005). Nanotechnology 16:

2346. 166 Smetana, A.B., Klabunde, K.J., Marchin, G.R., and Sorensen, C.M. (2008).

Langmuir 24: 7457–7464. 167 Gibbard, J. (1937). Am. J. Public Health Nations Health 27: 112–119. 168 Jain, P. and Pradeep, T. (2005). Biotechnol. Bioeng. 90: 59–63. 169 Banerjee, S.L., Khamrai, M., Sarkar, K. et al. (2016). Int. J. Biol. Macromol.

85: 157–167. 170 Mural, P.K.S., Sharma, M., Shukla, A. et al. (2015). RSC Adv. 5:

32441–32451. 171 Zhang, D.Y., Liu, J., Shi, Y.S. et al. (2016). J. Membr. Sci. 516: 83–93. 172 Kubacka, A., Serrano, C., Ferrer, M. et al. (2007). Nano Lett. 7: 2529–2534. 173 Qin, A., Li, X., Zhao, X. et al. (2015). ACS Appl. Mater. Interfaces 7:

8427–8436. 174 Li, J.F., Xu, Z.L., Yang, H. et al. (2009). Appl. Surf. Sci. 255: 4725–4732. 175 Razmjou, A., Mansouri, J., and Chen, V. (2011). J. Membr. Sci. 378: 73–84. 176 Alloway, B.J. (2013). Introduction. In: Heavy Metals in Soils: Trace Metals

177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192

and Metalloids in Soils and their Bioavailability (ed. B.J. Alloway), 3–9. Dordrecht: Springer Netherlands. Schiff, K., Diehl, D., and Valkirs, A. (2004). Mar. Pollut. Bull. 48: 371–377. Cioffi, N., Torsi, L., Ditaranto, N. et al. (2005). Chem. Mater. 17: 5255–5262. Avery, S.V., Howlett, N.G., and Radice, S. (1996). Appl. Environ. Microbiol. 62: 3960–3966. Ben-Sasson, M., Lu, X., Nejati, S. et al. (2016). Desalination 388: 1–8. Sawai, J., Kojima, H., Igarashi, H. et al. (2000). World J. Microbiol. Biotechnol. 16: 187–194. Nagarajan, P. and Rajagopalan, V. (2008). Sci. Technol. Adv. Mater. 9: 035004. Liang, S., Xiao, K., Mo, Y., and Huang, X. (2012). J. Membr. Sci. 394–395: 184–192. Hong, J. and He, Y. (2012). Desalination 302: 71–79. Sawai, J. (2003). J. Microbiol. Methods 54: 177–182. Weiss, N.O., Zhou, H., Liao, L. et al. (2012). Adv. Mater. 24: 5782–5825. Wang, J., Zhang, P., Liang, B. et al. (2016). ACS Appl. Mater. Interfaces 8: 6211–6218. Choi, W., Choi, J., Bang, J., and Lee, J.H. (2013). ACS Appl. Mater. Interfaces 5: 12510–12519. Wang, Z., Yu, H., Xia, J. et al. (2012). Desalination 299: 50–54. Zinadini, S., Zinatizadeh, A.A., Rahimi, M. et al. (2014). J. Membr. Sci. 453: 292–301. Zhao, C., Xu, X., Chen, J., and Yang, F. (2013). J. Environ. Chem. Eng. 1: 349–354. Chang, X., Wang, Z., Quan, S. et al. (2014). Appl. Surf. Sci. 316: 537–548.

159

160

4 Microporous Membranes for Water Purification

193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209

Zhu, J., Wang, J., Hou, J. et al. (2017). J. Mater. Chem. A 5: 6776–6793. Chen, J., Peng, H., Wang, X. et al. (2014). Nanoscale 6: 1879–1889. Wallace, E.J. and Sansom, M.S.P. (2008). Nano Lett. 9: 2751–2756. Kam, N.W.S., Jessop, T.C., Wender, P.A., and Dai, H. (2004). J. Am. Chem. Soc. 22: 6850–6851. Lee, B., Baek, Y., Lee, M. et al. (2015). Nat. Commun. 6: 7109. Brady-Estévez, A.S., Kang, S., and Elimelech, M. (2008). Small 4: 481–484. Sengur-Tasdemir, R., Mokkapati, V.R.S.S., Koseoglu-Imer, D.Y., and Koyuncu, I. (2017). Environ. Technol. 1–12. Kanagaraj, P., Nagendran, A., Rana, D. et al. (2015). Ind. Eng. Chem. Res. 54: 4832–4838. Du, J.R., Peldszus, S., Huck, P.M., and Feng, X. (2009). Water Res. 43: 4559–4568. Lee, H., Dellatore, S.M., Miller, W.M., and Messersmith, P.B. (2007). Science 318: 426–430. Liu, Y., Ai, K., and Lu, L. (2014). Chem. Rev. 114: 5057–5115. McCloskey, B.D., Park, H.B., Ju, H. et al. (2010). Polymer 51: 3472–3485. Cheng, C., Li, S., Zhao, W. et al. (2012). J. Membr. Sci. 417–418: 228–236. Xi, Z.Y., Xu, Y.Y., Zhu, L.P. et al. (2009). J. Membr. Sci. 327: 244–253. Shao, L., Wang, Z.X., Zhang, Y.L. et al. (2014). J. Membr. Sci. 461: 10–21. Gu, J., Su, Y., Liu, P. et al. (2017). ACS Appl. Mater. Interfaces 9: 198–210. Wu, Z., Wang, D., and Yang, P. (2014). Ind. Eng. Chem. Res. 53: 9401–9410.

161

5 Mixed Matrix Membranes 5.1 Introduction Membrane separation is a rapidly growing field for gas separations [1–3]. Separations of components by means of the membrane process are performed by differential permeation through a membrane. Membrane separation has some advantages compared with other methods such as absorption, adsorption, and cryogenic distillation for separation of gases or liquids. The main advantages include (i) low operating and capital cost, (ii) low energy consumption, and (iii) relative ease of operation [4–7]. These benefits make the membrane process economically favorable for industrial applications [8, 9]. Usually, three types of membrane are applied for separations especially for the widely studied gas separations: polymeric membranes, inorganic membrane, and mixed matrix membrane (MMM) [10, 11]. MMM is a new membrane category for separation and plays a vital role for the advancement of current membrane-based separation technology. A survey in Figure 5.1 from Web of Science clearly demonstrates that there are a sheer number of publications reported in the latest decade, and it shows an increasing trend, encouraging us to review the state-of-art MMMs in a separate chapter.

5.2 Principles of Mixed Matrix Membranes As shown in Figure 5.2, MMMs are composite or hybrid membranes made by blending an inorganic or inorganic–organic material as a filler in the form of micro- or nanoparticles (dispersed phase) and a polymer matrix (continuous phase). By using two materials with different transport properties, such membranes have the potential to combine synergistically the easy processability of polymers and the superior separation performance of porous filler materials. MMMs may provide separation properties that can be above the Robeson upper bound by overcoming the trade-off between the selectivity and permeability that is typical for pure polymer membranes [12].

Microporous Materials for Separation Membranes, First Edition. Xiaoqin Zou and Guangshan Zhu. © 2019 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2019 by Wiley-VCH Verlag GmbH & Co. KGaA.

162

5 Mixed Matrix Membranes

7000 6000 5000 4000 3000 2000 1000 0 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018

Figure 5.1 Historical growth of academic publications related with the topics of mixed matrix membranes or composite membranes or hybrid membranes. Source: data collected from Web of Science. Figure 5.2 Systematic diagram of a mixed matrix membrane consisting of filler and polymer components.

Filler (dispersed phase)

5.2.1

Polymer (continuous phase)

Polymer Phase

In MMMs, the polymer acts as the continuous phase and the inorganic filler as the dispersed phase. Rubbery polymers are not commonly applied due to their limitations. Transition temperature (T g ) and polarity are the two main factors by which polymers can be differentiated. The selection of the fillers is usually governed by their pore size, pore structure, and surface polarity [10, 13]. The selection of materials for membrane fabrication is a complex task especially when glassy polymers like cellulose acetate (CA), polysulfone (PSU), polyimide (PI), polyamide (PA), polyethersulfone (PES), polypropylene (PP), polyvinylidene fluoride (PVDF), etc. are proposed for the fabrication of MMMs [14]. Block copolymers are considered as another type of polymers suitable for preparing defined-nanostructure membranes. This polymer can offer an advanced function and nanostructure if it is processed into a membrane [15]. Block copolymers such as poly(styrene-butadiene-styrene) (SBS), block copoly(ether-urethane-urea), copoly(ether urethane), and poly(ethylene oxide-bamide) (Pebax) were applied for membrane gas separation [16–18]. Pebax was also used with multiwall carbon nanotubes for CO2 removal [19]. As illustrated by Robeson et al. in 1999, the separation capability of the rubbery polymers lies below the upper bound

5.2 Principles of Mixed Matrix Membranes

Table 5.1 Glass transition temperatures (T g ) of common polymers. Polymer

T g (∘ C)

Polyvinylidene fluoride (PVDF)

−35

Polypropylene (PP)

−20/0

Polyvinyl fluoride (PVF)

−20

Poly-3-hydroxybutyrate (PHB)

15

Poly(vinyl acetate) (PVAc)

30

Polyamide (PA)

47–60

Polyethylene terephthalate (PET)

70

Poly(vinyl chloride) (PVC)

80

Poly(vinyl alcohol) (PVA)

85

Polystyrene (PS)

95

Poly(methyl methacrylate) (PMMA)

105

Polytetrafluoroethylene (PTFE)

115

Polysulfone (PSU)

185

Polynorbornene (PNB)

215

limit, while the glassy ones are very close to the boundary [20]. Since rubbery polymers did not provide remarkable enhancement compared with glassy polymers [3], some researchers transferred their work from rubbery polymers to glassy polymers. Table 5.1 lists common polymers and their corresponding glass transition temperature (T g ). 5.2.2

Solvents

Solvent is the second important factor that affecting the membrane property. Table 5.2 provides the properties of different solvents that are used in the fabrication of MMMs. 5.2.3

Fabrication and Drying Techniques for MMMs

Phase inversion and solution casting are mainly two methods for producing MMMs. Membranes for phase inversion can be synthesized from any polymer–solvent mixture that forms a homogeneous solution under certain conditions of temperature and composition but separates into two phases when these conditions are changed. For example, phase inversion can be induced by evaporation of a volatile solvent from a homogeneous polymer solution or by cooling a casting solution that is homogeneous only at elevated temperatures [21]. There are numerous processes for the fabrication of integrally skinned asymmetric membranes from a homogeneous dope. The three main processes in Figure 5.3 are (i) dry process, (ii) wet process, and (iii) dry/wet process. The difference between dry and wet process is whether the outlet of the spinneret or casting knife is submerged directly in a non-solvent coagulant.

163

Table 5.2 A summary of different solvents used for fabricating MMMs. MW (g mol−1 )

Density (g cm−3 )

B. P. (∘ C)

V. P. (kPa)

Solubility in water

Solubility parameter (J1/2 cm−3/2 )

Dichloromethane (DCM) (CH2 Cl2 )

84.93

1.327

39.6

57.3

Not soluble

19.9

Low boiling point; thus filler sedimentation in MMMs is avoided. High volatility can make the membrane with a wavy structure; thus rapid evaporation is avoided

N,N-Dimethylacetamide (DMAc) (C4 H9 NO)

87.12

0.937

165.1

0.3

Soluble

22.1-22.8

Suitable solvent for polyimide synthesis due to polar and aprotic nature. Excellent solvent for high dissolving power for high molecular weight polymers and synthetic resins

N,N-Dimethylformamide (DMF) (C3 H7 NO)

73.10

0.948

152

0.516

Soluble

24.9

Good solvent for salts or compounds with high molecular weights due to high dielectric constant, electron donor ability, and complex-forming ability. Major solvent for producing PUR elastomers, spandex fibers, and synthetic leathers

1-Methyl-2-pyrrolidone (NMP) (C5 H9 NO)

99.13

1.028

202

0.4

Soluble

22.9

Powerful polar solvent with chemical stability and low toxicity

Chloroform (CHCl3 )

119.37

1.489

61.15

25.9

Not soluble

18.9

Common and unreactive solvent in laboratory and factory

Tetrahydrofuran (THF) (C4 H8 O)

72.11

0.889

66

17.4

Soluble

18.6

Suitable for skin layer formation in the membrane fabrication.

Solvent

B.P., boiling point; V.P., vapor pressure at 25 ∘ C; MW, molecular weight.

Remarks

5.2 Principles of Mixed Matrix Membranes

1. Dry phase inversion 2. Wet phase inversion 3. Dry/wet inversion

1

Casting

Evaporation

2 Casting

Quenching Evaporation

3 Casting

Evaporation

Quenching

Figure 5.3 Three types of phase inversion processes.

Solution-casting method is appropriate for small-scale preparation of membrane samples. The polymer solution is spread on a flat glass plate with a casting knife. This knife consists of a steel blade placed on two runners. The thin film forms the membrane that has been left after the solvent evaporation [22]. This method is simple in operation and cheap and the final membrane is crystalline and isotropic. Mahajan et al. tried to solve the problem of void creation between polymer and filler in conventional solution-casting methods for the formation of MMMs and then developed the membrane in two steps under high temperature [23]. Other studies were also carried out for making casting solutions for MMM fabrication [23, 24]. These solutions were prepared by first adding a filler to the solvent to obtain slurry, and then a polymer powder was mixed into this slurry. Drying of MMMs is the subsequent step of producing usable membranes. A group of researchers synthesized MMMs by using PES as the polymer and inorganic zeolite 13X and 4A as the fillers. They adopted three methods for drying the cast film: (a) at room temperature for one day, (b) under atmospheric pressure in nitrogen at 60–80 ∘ C for almost 10 hours, and (c) under partial vacuum with nitrogen circulation at 60–80 ∘ C for 8–10 hours. Methods (a) and (b) were also carried out in quenching and annealing processes [25]. Itta et al. preferred to dry PPO/polyvinylpyrrolidone (PVP) and carbon molecular sieve (CMS) membranes at room temperature for one day [26], as well as the same did in drying PI or polyethylenimine (PEI)/CMS MMMs by Tseng et al. [27]. Another team of researchers used room temperature for drying the Matrimid /CMS MMM followed by vacuum oven drying at 100 ∘ C overnight for completely removing the solvent [28]. Ismail et al. produced a PES/MWCNT MMM through air drying under atmospheric conditions for 24 hours [29]. Different types of inorganic fillers (impermeable and permeable ones) have been used as filler materials [30]. In particular, improvements in separation performance would be expected if nanostructured, highly porous fillers such as CMS, zeolites, metal–organic frameworks (MOFs), or porous organic

®

165

166

5 Mixed Matrix Membranes

Table 5.3 Different microporous materials and their properties. Microporous materials

Properties

Zeolites

Crystalline inorganic solids; high diffusivity and selectivity compared with polymers; quite cheap; difficult fabrication of defect-free membranes

Metal–organic frameworks (MOFs)

Crystalline inorganic–organic hybrids; high surface area; controlled porosity and pore size; good affinity toward specific gases; rigid and flexible frameworks; pore walls can be facilely functionalized; high permeability, but low selectivity; very sensitive to water and heat

Porous organic frameworks (POFs)

Pure organics with open and ordered pores; different pore diameters (2–20 nm) and shapes; excellent thermal and chemical stability; hydrophobic surface; very expensive for sample preparation and difficult for making particles uniformly dispersed

frameworks (POFs) were used. Table 5.3 summarizes the types of microporous fillers and their corresponding properties. Since an MMM is the dispersion of filler particles in a polymeric matrix, the properties of both polymer and filler affect the separation performance [30]. The chemical structure, surface chemistry, particle size distribution, and aspect ratio are other important factors of a filler material. Thus, the selection of the filler depends on the following characteristics: (i) it should have the ability to achieve high permselectivity for the separation of the desired gas [31], (ii) the permeability of all species increases with higher filler particle size [32], and (iii) filler loading percentage [25]. For meeting these requirements, microporous materials are the right choice. After a careful survey, we have found that most of the studies are focused on gas separation with MMMs. Thus, we will detailed discuss the advanced microporous materials (zeolites, MOFs, POFs) as the fillers for making MMMs and their representative gas separations in the following sections. 5.2.4

Mass Transport Theory and Models in MMMs

Ideal MMM morphology is a two-phase system consisting of inorganic fillers and polymer matrix without any defects or distortion at the filler–polymer interface. For this kind of morphology, theoretical models such as Maxwell, Bruggeman, Böttcher and Higuchi, Lewis–Nielsen, Pal, Gonzo–Parentis–Gottifredi (GPG), and Kang–Jones–Nair (KJN) have been adapted for estimating the permeation performance of MMMs. As most MMMs are applied for gas separation, transport models for MMMs have been developed to predict gas effective permeability through MMMs as a function of continuous phase (polymer) and dispersed phase (filler) permeabilities, as well as volume fraction of the dispersed phase. For a given penetrant (A), the permeability coefficient (PA ) can be estimated as the

5.2 Principles of Mixed Matrix Membranes

product of diffusivity coefficient (DA ) and solubility coefficient (SA ): PA = DA × SA Here, DA and SA are usually functions of solubility and pressure. The selectivity, 𝛼 A/B , is the ratio of permeability coefficients of A and B components: 𝛼A∕B =

PA D S = A A PB DB SB

At steady state, Peff describes the effective gas permeability, and a minimum value of Peff is calculated when a series two-layer model is applied: Peff =

Pc Pd ∅c Pd + ∅d Pc

Here, Pc is the permeability in continuous phase, Pd is the permeability in dispersed phase, and 𝜙c and 𝜙d are volume fractions of continuous and dispersed phase, respectively. A maximum value of Peff is reached when both phases are assumed to diffuse through a parallel two-layer membrane (parallel model): Peff = Pc ∅c + Pd ∅d where 𝜙c = (1 − 𝜙d ). Applying the Maxwell–Wagner–Sillars model, the Peff of an MMM with a dilute dispersion of ellipsoids is given by the following expression [33]: Peff = Pc

nPd + (1 − n)Pc − (1 − n)∅d (Pc − Pd ) nPd + (1 − n)Pc + n∅d (Pc − Pd )

Here, n is the particle shape factor. For prolate ellipsoids where the longest axis of the ellipsoid is directed along the applied partial pressure gradient, n value is 0 < n < 1/3. For oblate ellipsoids where the shortest axis of the ellipsoid is directed along the applied partial pressure gradient, n value is 1/3 < n < 1. The limit of n = 0 leads to a parallel two-layer model, and the limit of n = 1 corresponds to a series two-layer model. At the limit of n = 1/3, the original Maxwell equation can be expressed: Peff = Pc

Pd + 2Pc − 2∅d (Pc − Pd ) Pd + 2Pc + ∅d (Pc − Pd )

The Bruggeman model, originally developed for the electric constant of particulate composites, can be also adapted to estimate gas permeability in MMMs as shown [34]: ( ) 𝛼−1 = (1 − ∅d )−1 (Pr )1∕3 𝛼 − Pr Here, Pr = Peff /Pc , where 𝛼 is the permeability ratio of Pd /Pc . The Böttcher and Higuchi models, originally applied to a random dispersion of spherical particles, are respectively expressed in the following [35, 36]: )( ) ( Peff P 𝛼+2 = 3∅d (𝛼 − 1) 1− c Peff Pc

167

168

5 Mixed Matrix Membranes

Pr =

Peff Pc

=1+

3∅d 𝛽 1 − ∅d 𝛽 − KH (1 − ∅d )𝛽 2

Here, K H is treated as an empirical constant and assigned a value of 0.78. 𝛽 is useful to define the “reduced permeation polarizability” or a convenient measure of permeability difference between dispersed spheres and polymer matrix: 𝛽=

P − Pc 𝛼−1 = d 𝛼 + 2 Pd + 2Pc

Here, 𝛽 is bounded by −0.5 ≤ 𝛽 ≤ 1, where lower and upper limits correspond to totally non-permeable and to perfectly permeable fillers. The Lewis–Nielsen model, originally proposed for an elastic modulus of particulate composites, is also possible to apply for predicting the effective permeability in MMMs [37]: Pr =

Peff Pc

=

1 + 2∅d (𝛼 − 1)∕(𝛼 + 2) 1 − 𝜓∅d (𝛼 − 1)∕(𝛼 + 2)

𝜓 here is defined as ( ) 1 − ∅m 𝜓 =1+ ∅d ∅2m where 𝜙m is the maximum packing volume fraction of the filler and is usually considered to be 0.64 for a random close packing of uniform spheres. By considering the effect of particle morphology on permeability, this model might represent a correct behavior of permeability over the range of 0 < 𝜙d < 𝜙m . The relative permeability Peff at 𝜙d = 𝜙m is found to be, however, diverging when the permeability ratio 𝛼 close to ∞. The Pal model, originally applied for thermal conductivity of particulate composites, was also adapted for prediction of permeability [38]: ( ) ( ) ∅ −∅m 𝛼−1 (Pr )1∕3 = 1− d 𝛼 − Pr ∅m Gonzo and coauthors proposed an extension of the original Maxwell model (called the GPG model) in terms of 𝜙d [39]: Pr =

Peff Pc

= 1 + 3𝛽∅d + K∅2d + O(∅d )3

Here, K and O are the needed corrections of Maxwell expression. The coefficient K is a function both of 𝛽 and 𝜙: K = a + b∅1.5 d The parameter a, b is a function of 𝛽: a = −0.002254 − 0.123112𝛽 + 2.93656𝛽 2 + 1.690𝛽 3 b = 0.0039298 − 0.803494𝛽 − 2.16207𝛽 2 + 6.48296𝛽 3 + 5.27196𝛽 4

5.3 MMMs Made of Zeolites

Filler

Interfacial voids (a)

Matrix

Rigidified polymer chain layer (b)

Pore blockage region (c)

Figure 5.4 Schematic diagram of three nonideal morphologies in MMMs: interfacial voids (a), rigidified polymer chain layer (b), and pore blockage (c).

The KJN model was specifically derived for an ideal composite membrane with tubular fillers having a fixed orientation that possesses perfectly anisotropic 1D transport properties. It can be expressed as follows [40]: ) }−1 ) ( {( Peff Pc cos 𝜃 1 ∅d = 1− ∅d + Pc Pd cos 𝜃 + 1 sin 𝜃 cos 𝜃 + 𝛼1 sin 𝜃 𝛼 Here, 𝛼 = l/d is the aspect ratio of tubular fillers, and 𝜃 is the filler orientation angle with respect to the membrane transport direction, varying from 0 to 𝜋/2 radians. In reality, an MMM has a nonideal morphology or three-phase system containing inorganic–organic interface defects. As shown in Figure 5.4, interface defects can be classified into the following three major categories: (i) interface voids or sieves in a cage, (ii) rigidified polymer layer around the inorganic fillers, and (iii) particle pore blockage. For a nonideal MMM, interfacial defects affect the membrane performance and should be considered in prediction models. Thus, several modified models have been proposed, and all of them are derived from their corresponding ideal ones [41].

5.3 MMMs Made of Zeolites Zeolites as a traditional class of molecular sieves are a promising candidate as a filler for producing MMMs. Zeolites possess uniform molecule-sized pores, tetrahedral and arranged by shared oxygen at the corners of aluminosilicate geometric patterns with high thermal and chemical stability. Considering the abovementioned properties, zeolites overcome the limitations of polymer membranes, such as thermal decomposition or deformation at high temperature, high pressure, and oxidative atmospheres. There have been numerous attempts to incorporate zeolite particles into polymer matrices to form zeolite-based MMMs. Polymer segments are interacted with zeolite surfaces to form continuous membranes with potentiality in gas separation [42]. The different forms of zeolite and zeo-type fillers are involved in gas separation such as A, X, Y,

169

170

5 Mixed Matrix Membranes

Table 5.4 MMMs made of zeolites and polymers for gas separation. Zeolite

Polymer

Application

References

LTA

PDMS

H2 /CO2 , H2 /CH4 , CO2 /CH4

[43]

LTA, CHA, RHO

PTMSP

O2 /N2

[44]

MFI

PEBAX, Matrimid

CO2 /CH4 , CO2 /N2

[45–47]

FAU

PEBAX

CO2 /N2 , O2 /N2

[48]

TS-1, ETS-10

Matrimid

CO2 /CH4

[49]

CHA

PEBAX

CO2 /CH4 , CO2 /N2

[50]

LTA

PES

He/H2 , H2 /N2 , O2 /N2 , CO2 /CH4

[51]

Nu-6(2)

PSF

H2 /CH4

[52]

LTA

PVAc

CO2 /CH4

[53]

MFI

PSF, Matrimid

O2 /N2 , CO2 /CH4

[54]

FAU, LTA

PI

CO2 /N2 , CO2 /CH4

[55]

MFI

PDMS

CO2 /N2

[56]

FAU, EMT

PI

CO2 /CH4

[57]

ZSM-5, TS-1, ETS-10, and SAPO-34, and the reported results are summarized in Table 5.4. Rezakazemi et al. studied the gas transport properties of zeolite-reinforced polydimethylsiloxane (PDMS) MMMs [43]. MMMs were prepared by incorporating zeolite 4A nanoparticles into a PDMS matrix using via an improved solution-casting technique. The zeolite filler was dispersed homogeneously in the matrix without any voids at the zeolite–polymer interface. The permeation rates of H2 , CO2 , CH4 , and C3 H8 were evaluated through a pure PDMS membrane and PDMS/4A MMMs with different zeolite loadings to assess the viability of these membranes for hydrogen purification. The MMMs exhibited both higher selectivity of H2 /CH4 and H2 permeability as compared with the neat PDMS membrane, suggesting the potentiality of these membranes in H2 /CH4 separation. Fernández-Barquín and coworkers prepared several MMMs, which were composed of small-pore zeolites with various topologies (CHA [Si/Al = 5], LTA [Si/Al = 1 and 5], and RHO [Si/Al = 5]) as the dispersed phase and highly permeable poly(1-trimethylsilyl-1-propyne) (PTMSP) as the continuous phase via solution casting [44]. The O2 /N2 gas separation performance of the MMMs was analyzed in terms of permeability, diffusivity, and solubility in the temperature range of 298–333 K. The O2 /N2 selectivity and O2 permeability reached up to 8.43 and 4800 Barrer at 333 K. They found that the O2 /N2 permselectivity of the MMMs increased with temperature, the O2 /N2 selectivities being considerably higher than those of the pure PTMSP. Moreover, most of the prepared MMMs exceeded the Robeson upper bound for the O2 /N2 gas pair, showing the membranes attractive for oxygen-enriched air production. Beiragh et al. fabricated MMMs by incorporating ZSM-5 (MFI-type zeolite) in poly(ether-block-amide) (PEBAX) matrix [45]. PEBAX provided a good interaction with ZSM-5 due to high chain mobility of polyethylene oxide (PEO)

5.3 MMMs Made of Zeolites

soft segments and varied the polymer–zeolite interface by changing ZSM-5 loadings from 5 to 15 wt%. Thus, the superior permeability of CO2 is achieved, about 78% at 5.0 wt% ZSM-5 loading, owing to the availability of more free volume. The permeability of CH4 was reduced compared with CO2 , because the porous ZSM-5 restricts the mobility of polymer chain and leads to the formation of tortuous pathways in polymer. However, ZSM-5 particles were dispersed into Matrimid matrix, up to 30 wt% loading, via the phase inversion approach. The higher loading of ZSM-5 particles led to the void formation at polymer–filler interfaces due to incompatibility with Matrimid chains. Thus, the higher permeability with good selectivity was obtained at 6 wt% ZSM-5 loading [46]. Other group reported better size and shape selectivity of ZSM-5/Matrimid MMMs by reducing nonselective voids between them through the penetration of polymer segments into ZSM-5. The enhanced selectivity of ZSM-5/Matrimid for CO2 /CH4 and CO2 /N2 were 34.7 and 17.8, respectively [47]. Zarshenas et al. employed the combined technique of phase inversion and solvent evaporation to fabricate MMMs by incorporating NaX nano-zeolite into poly(ether-block-amide) (PEBAX-1657) as an active layer on the PES membrane as a support layer [48]. The influence of filler content, zeolite particle size, and operating conditions on gas separation performance of prepared membranes were studied as well. Gas permeation results revealed that incorporation of NaX nano-zeolite into the PEBAX-1657 membrane decreased the gas permeability for all tested gases, while the ideal selectivity for CO2 /N2 and O2 /N2 significantly enhanced. Also, the MMMs containing nano-sized zeolite had superior gas separation performance in comparison with the micron-sized zeolite-loaded membrane. It was found that the CO2 /N2 selectivity enhanced from 98.0 to 121.5 with increasing the feed pressure, while the O2 /N2 selectivity decreased from 7.8 to 4.2. For CO2 /N2 separation, the membrane containing 2 wt% nano-zeolite at the operating pressure of 7 bar and temperature of 25 ∘ C showed fairly good gas separation performance. Martin-Gil et al. synthesized three types of titanosilicate zeolites and used them as fillers for fabricating MMMs with Matrimid [49]. A series of flat dense MMMs were fabricated by varying filler loadings up to 30 wt% and studied for gas separations. TS-1-100 (Si/Ti = 100) and TS-1-25 (Si/Ti = 25) possess tetrahedral and ETS-10 (Engelhard Corporation Titanosilicate) has octahedral structured titanium silicate zeolites. TS-1-100 MMMs revealed the improved CO2 permeability about 90% than neat polymer membrane, because TS-1 particles promoted a better adhesion between fillers and polymeric matrix, which increased CO2 adsorption. However, TS-1-25 showed a better CO2 /CH4 selectivity than TS-1-100 due to some “sieve-in-a-cage” morphology defects. The improved permeability and selectivity of ETS-10 MMMs were 24% and 33%, respectively, than unfilled PI membrane. Zhao et al. reported sodium-alumino-phosphate (SAPO)-based MMMs fabricated by dispersing SAPO-34 homogeneously in the PEBAX [50]. Their results exhibited an increased CO2 permeability of about 250% at 50 wt% SAPO-34 than the neat PEBAX membrane, without altering the selectivity of CO2 /CH4 and CO2 /N2 . As discussed in Figure 5.4, the interfacial morphology of MMMs affects the membrane performance, in particular between inorganic zeolite and organic

171

172

5 Mixed Matrix Membranes

Polymer chain Silane

Si

Si OH OH OH OH

Zeolite surface

OH O

OH OH

Silanated zeolite surface

OH O

OH OH

Polymer attached to zeolite

Figure 5.5 Schematic of the silane coupling process between zeolite and polymer.

polymer. To obtain the optimum interfacial morphology between zeolite and polymer, several key factors should be considered. The first one is to promote the adhesion between zeolite filler and polymer matrix by modifying zeolite surface with silane coupling agents [51]. The second one is to introduce low molecular weight materials (LMWMs) to fill the voids between zeolite and polymer [58]. The third one is to apply high processing temperatures close to glass transition temperature (T g ) of polymer to render chain flexibility during membrane formation [23]. The fourth one is to prime zeolite surface by polymer [59]. In order to improve zeolite–polymer interaction, several strategies have been developed to modify the interface, including interfacial modification with silane agents, addition of LMWMs, annealing, and priming method. 5.3.1

Interfacial Modification with Silane Agents

Due to abundant hydroxyl groups on zeolite surface, these groups can be reacted with silane via covalent or hydrogen bonds. Thus, improving adhesion between zeolite and polymer can be achieved in the membrane [60]. Figure 5.5 shows a schematic silanation of zeolite surface with organic silane coupling agent. Leo et al. investigated the silane effect on the separation performance of MMM for gas permeation [61]. The 3-aminopropyltrimethoxysilane (APMS) was added to modify SAPO-34 zeolite before its impregnation into asymmetric PSF MMMs through dry/wet phase inversion method. PSF/modified SAPO-34 membranes showed great enhancement in terms of selectivity and permeability, compared with the original PSF membrane. The increment of CO2 selectivity and permeability was correlated to diminishing interfacial voids when SAPO-34 zeolite was modified using APMS in ethanol. 5.3.2

Addition of Low Molecular Weight Materials

LMWMs are usually solids at ambient temperature in order to prevent their evaporation during membrane fabrication. LMWMs can induce a hydrogen bond with hydroxyl and carbonyl moiety. Hydrogen bonding brings two consequences:

5.4 MOF-Based MMMs

one is improving the compatibility between zeolite and polymer, and the other is decreasing the free volume of the polymer. In general, adding LMWMs into the membrane formulation results can increase gas permselectivity with some sacrifice in gas permeability. One typical LMWM is 2,4,6-triaminopyrimidine (TAP) containing three primary amine groups, which are able to form hydrogen bonds with both hydroxyl and carbonyl groups [58]. Park et al. used TAP to obtain interfacial void-free PI membranes filled with zeolites [58]. TAP enhanced the contact of zeolite particles with PI chains presumably by forming the hydrogen bonding. As a consequence, the void-free PI/zeolite 13X/TAP membrane showed higher gas permeability for He, N2 , O2 , CO2 , and CH4 with little expense of selectivity compared with the PI/TAP membrane having the same PI/TAP ratio, while the PI/zeolite 4A/TAP membrane showed the lower permeability but higher permselectivity. The difference between both membranes was influenced by the pore size of zeolites. In addition, the molecular sieving effect of zeolites seemed to take place when the kinetic diameter of gas penetrants approached the pore size of zeolites. 5.3.3

Annealing

Annealing zeolite-based MMMs are usually performed above the glass transition temperature (T g ) [23]. Annealing process at temperature above T g results in forming stronger bonds between zeolite and polymer, which can relax the stress imposed to the hollow fiber membrane and also can form higher packing density of polymer chains. Annealing also has some drawbacks: it cannot lead to significant improvement in the membrane morphology; annealing at high T g could form sieve-in-a-cage morphology [23]. 5.3.4

Priming Method

Priming is to coat the surface of filler particles with a dilute polymer, which is the same as the bulk polymer for the fabrication of MMMs [24]. More agglomeration occurs in the polymer matrix when smaller particles are used, especially at high particle loadings, and larger zeolite particles are difficult to be dispersed in solvent. This agglomeration is responsible for defects between zeolite and polymer. Polymer effectively coats zeolite particles by mixing with the priming polymer before adding remaining bulk polymer [62]. The purpose of priming is to reduce stress at the polymer–particle interface, to increase the compatibility between zeolite and polymer in MMMs, and to minimize agglomeration of zeolite particles.

5.4 MOF-Based MMMs MOFs offer various advantages over zeolites and therefore are promising fillers for MMMs [63]. MOFs are chemically mutable, highly porous materials prepared from the combination of metal ions or clusters and organic ligands. In comparison with purely inorganic zeolites, the chemical diversity of MOF structures

173

174

5 Mixed Matrix Membranes

Figure 5.6 Schematic representation of major functionalities in MOF-based MMMs.

Functional groups/ modified surface

UiO-66

Particle size

ZIF-8

Particle morphology

MOF-74

HKUST-1

MIL-53

Figure 5.7 Structures of some typical MOFs as fillers in MMMs.

can be used to facilitate strong interactions with the polymer material. This is achieved through judicious choice of ligands with appropriate chemical functionalities (Figure 5.6), which can reduce the probability in micro-gap formation between inorganic and organic phases existed in the case of zeolite-based MMMs [30]. Furthermore, the chemical mutability of the MOF scaffold can be utilized to provide enhanced adsorption of a particular chemical species and to facilitate improved separation performance. In addition to the incorporation of chemical functionality through judicious choice of starting material, MOFs can also be modified post-synthesis to enhance the separation performance or facilitate stronger filler–polymer interactions. MOFs are not only tunable in terms of their chemical groups but also their pore volumes, pore shapes, and surface areas, thus enabling them to contribute to either the permeability or selectivity of the MMM. Many MOFs display considerably high surface areas and larger pore volumes and can often display performance characteristics commensurate with the polymer matrix. In this section, we will introduce several examples compiling UiO-66s, ZIFs, Cu-MOFs, MILs, and MOF-74s to elucidate their functions as fillers in MMMs (Figure 5.7). 5.4.1

UiO-66 Series

Zirconium-based MOFs (UiO-66; UiO = University of Oslo) is a subfamily of MOFs. UiO-66 was first discovered in 2008, possessing exceptional stability due to the fact that each Zr metal center is connected with benzene-1,4-dicarboxylate (BDC) linkers to form the face-centered cubic (FCC) crystal structure framework, with the formula Zr6 O4 (OH)4 (BDC)6 and aperture size approximately 6 Å [64]. The surface-modified UiO-66 has polar and basic functionalities such as hydroxylated (—OH), amino (—NH2 ), nitro (—NO2 ), and methoxy (—OCH3 ) groups [65].

5.4 MOF-Based MMMs

Shen et al. reported that UiO-66/PEBA MMMs possessed good affinity with CO2 molecules and the enhanced CO2 permeability and CO2 /N2 selectivity were about 80–90% and 40–65%, respectively, than pure PEBA membrane [66]. However, Ti-exchanged UiO-66 in PIM-1 showed improved permeability of CO2 about 8200 Barrer without a loss of selectivity compared with pure PIM-1. The enhanced permeability of these materials was reported due to a strong interaction between MOF and polymer, which led to the formation of interfacial free volume and strong affinity for CO2 [67]. Khdhayyer et al. studied the effect of amino and carboxylic group containing Zr-based UiO-66 filler with PIM-1 matrix. UiO-66 and NH2 -UiO-66 fillers showed that the higher permeability was 1100 and 1600 Barrer, respectively, without disturbing the selectivity of CO2 /CH4 . But (COOH)2 -UiO-66 revealed a lower permeability of about 300 Barrer and a decreasing selectivity of about 25% than PIM-1, due to the presence of free carboxylic groups in (COOH)2 -UiO-66/PIM-1, which reduced the intrinsic microporosity and the free volume with interconnected voids of PIM-1 [68]. Venna et al. modified UiO-66-NH2 nanoparticles with functional groups of phenyl acetyl (PA), decanoyl acetyl (DA), and succinic acid (SA) and added them into Matrimid to fabricate MMMs for gas separations [69]. The surface functionalization of UiO-66-NH2 fillers showed a good interaction with polymer. The effective interactions included π–π interaction and hydrogen bonding between MOF fillers and the polymer matrix with the following order: PA > DA > SA (Figure 5.8). This work provided a powerful toolbox to enhance the polymer–filler compatibility in MMMs. This effect was further proved by CO2 permeation. The Phenyl-acetylfunctionalized UiO-66-NH2 O N H Hydrogen bonding

H3C

CH3

1

TT–TT stacking 2

O

O

O

3

4 3

N

2

1

N

H3C 5O

O4 n

Matrimid

Figure 5.8 Scheme demonstrating favorable interactions between fragments in Matrimid polymer and PA groups in PA-modified UiO-66-NH2 [69]. Source: Copyright 2015, adopted with permission from RSC.

175

176

5 Mixed Matrix Membranes

PA-modified UiO-66-NH2 /Matrimid illustrated that increased permeability of CO2 and CO2 /N2 selectivity were 200% and 32%, respectively, than control experiment. However, DA- and SA-functionalized UiO-66-NH2 showed low transport performance of CO2 compared with PA-containing UiO-66-NH2 , due to poor interaction between the polymer and the fillers (polar of SA, nonpolar alkyl group of DA) and the formation of defective interface that altered the diffusion path for nonselective gases. 5.4.2

Zeolitic Imidazolate Frameworks

Zeolitic imidazolate frameworks (ZIFs) as a subfamily of MOFs possess zeolite topology with tunable pore structures. Zeolites are replaced by these ZIFs due to their high thermal/chemical stabilities [70]. ZIFs are consisted of metal nodes connected to imidazole linkers by replacing the bridging oxygen in the Si/Al-O-Al/Si (zeolite) repeating unit with imidazolates and Si/Al atoms with transition metals. Bond angle of metal-linker-metal (∼145∘ ) in ZIFs resembles the Si/Al-O-Al/Si repeating unit in zeolites [71]. A series of ZIF materials can be produced by varying the organic linkers of imidazolate ligands and metal ions (zinc or cobalt) such as ZIF-7, ZIF-8, ZIF-11, ZIF-22, ZIF-71, ZIF-78, ZIF-90, ZIF-108, and ZIF-302. Generally, ZIFs have large cavities with narrow pore apertures (e.g. 3.4 Å for ZIF-8), which is close to kinetic diameters of gases (e.g. 2.9 Å for H2 and 3.3 Å for CO2 ) and allows a suitable filler material in MMMs for gas separation. Theoretically, ZIF-8 should exhibit a stringent molecular sieving capability for CO2 /N2 separation by allowing the transport of CO2 while blocking the passage of N2 . However, recent studies indicated that MMMs with ZIF-8 only exhibited a moderate CO2 /N2 selectivity originating from the flexible pore structure of ZIF-8 nanocrystals [72]. To achieve the molecular sieving property of ZIF-8 for CO2 /N2 and CO2 /CH4 separations, Ban et al. finely tuned the effective cage size of ZIF-8 by incorporating a room temperature ionic liquid (RTIL) as the cavity occupant in an in situ ionothermal synthesis (Figure 5.9) [73]. The RTIL not only functioned as the cavity occupant but also possessed high affinity to CO2 , significantly improving the CO2 adsorption capacity of original ZIF-8 fillers. Compared with conventional inorganic fillers such as zeolites, ZIF-8 nanocrystals show better compatibility with the polymer matrix. No “sieve-in-a-cage” morphology was observed in ZIF-8-based MMMs in the literature so far [72]. However, several studies revealed that ZIF-8 nanocrystals tended to form agglomerates in MMMs at relatively high loadings (>20 wt%), which might result in the formation of nonselective voids that could deteriorate the membrane selectivity [74]. Moreover, strong interactions between ZIF-8 fillers and polymer matrix, important for improved interfacial compatibility, were not been widely observed in MMMs yet. Diverse strategies, including ammonia modification [75], polymer coating [76], and polymer grafting [77], were demonstrated to be effective in enhancing polymer–ZIF interfacial compatibility, which improved the comprehensive performance of MMMs for separations. Recently, Kong et al. dispersed various MOF nanoparticles (including ZIF-8) into organosilica matrix to fabricate ultrathin MOF/organosilica MMMs on porous

5.4 MOF-Based MMMs

Confinement of ILs

(la

rg

es

e

tp

or

C

ag

di

e

si

ze

am et

er

)

into nanocage

Aperture size (cutoff size)

Cavity occupants Effective cage size (cutoff size)

Cavity Aperture size occupants (largest pore diameter)

Figure 5.9 Schematic illustration of the cavity-occupying concept for tailoring the molecular sieving properties of ZIF-8 by the incorporation of RTILs. Source: Reproduced with permission [73]. Copyright 2015, John Wiley & Sons.

ceramic substrates with simultaneously improved gas permeance and selectivity [78]. The resultant MMMs showed no visible defects or filler agglomeration, attributed to the good adhesion of the two phases. Other ZIFs such as ZIF-7, ZIF-11, ZIF-71, ZIF-90, and ZIF-108 were employed as candidate fillers to fabricate MMMs [79–83]. Besides the progresses on materials design, extensive efforts have been devoted to enhance the performance of ZIF-based MMMs from an engineering perspective. Deng et al. adopted a dry-free, water-based process for fabricating MMMs with ZIF-8 loadings of up to 39 wt% [84]. Conventional MMM fabrication process involved the drying of fillers, resulting in unavoidable filler agglomeration and poor filler dispersion. In this dry-free process, nano-sized ZIF-8 particles were directly added into the aqueous solution of poly(vinyl alcohol) (PVA), allowing the PVA chains to adapt and interact with the surface of ZIF-8 fillers. Zhang et al. pioneered in the preparation of hollow fiber MMMs containing up to 30 wt% of ZIF-8 fillers in a dry jet–wet quench fiber spinning process [85]. The as-prepared MMMs exhibited significantly enhanced C3 H6 /C3 H8 selectivity compared with pristine polymeric membranes, paving the way for scaling up MMMs for industrial applications. 5.4.3

MIL Series

MILs (Materials Institute Lavoisier) are a series of MOFs that are constructed through the organic linker connected with metal ions (Al, Cr, Ti) by corner sharing, forming three-dimensional (3D) structures [86]. MILs exhibit their attractive characteristics such as large pore volume, high surface area, and superior gas adsorption capacity and possess some unique structure-breathing

177

178

5 Mixed Matrix Membranes

properties due to the framework consisting of unidirectional diamond-shaped pore channels. Based on recent literature, MILs act as good filler in the development of membranes for separation because of their uniform dispersion and good compatibility with polymers. MILs are reported in different forms and used as fillers in MMMs, such as MIL-53, MIL-68, MIL-101, and MIL-125. For instance, Zhu et al. reported that surface-modified MMM made of NH2 -MIL-53(Al) and Ultem showed higher CO2 permeability than unmodified MIL-53(Al)-containing MMMs and selectivity also increased by 10% than pure Ultem. The gas transport performance was higher due to the interaction of bridging hydroxyl group of MIL-53(Al) with carbonyl groups of polymers. Moreover, the organic terephthalate ligands and surface -NH2 groups also improved the compatibility with the matrix [87]. Abedini et al. fabricated MMMs by dispersing NH2 -MIL-53 in poly(4-methyl-1-pentyne) (PMP) matrix and found that CO2 permeability was 259 Barrer and selectivity of CO2 /CH4 was 15 at 30 wt% filler loaded MMMs than pure PMP membrane. The separation performance of NH2 -MIL-53/PMP surpassed the Robeson upper bound at 30 ∘ C [88]. Sabetghadam et al. explored the effect of 6FDA-DAM matrix in MMMs. NH2 -MIL-53(Al)/6FDA-DAM revealed the enhanced permeability of CO2 , up to 300 Barrer than pure matrix [89]. Further, the surface-modified 6FDA-DAM-HAB matrix with NH2 -MIL-53(Al) MMMs exhibited higher CO2 separation and approached the 2008 Robeson upper bound [90]. Dong et al. investigated the influence of MIL-68(Al)/PI MMMs for CO2 separation. The results illustrated higher permeability and selectivity of CO2 about 2.25-fold than pure polymer, which was because of the strong interaction of the surface hydroxyl groups of MIL-68(Al) with polymer and the increased adsorption affinity of CO2 [91]. Xin et al. reported about the fabrication of MMMs using the bare and PEI-modified MIL-101(Cr) fillers incorporated into SPEEK. The PEI-modified filler interacted through the surface -NH2 groups of PEI with sulfonic acid group of SPEEK through electrostatic interaction and hydrogen bond, which resulted in an enhanced permeability of 1945 Barrer, and the selectivities of CO2 /CH4 and CO2 /N2 were 47 and 36, respectively [92]. Anjum et al. investigated the effect of NH2 group for gas transportation, exemplified by embedding MIL-125(Al) and NH2 -MIL-125(Al) fillers into the PI matrix. The MMMs showed enhanced permeability of 21 and 44 Barrer and increased selectivity of CO2 /CH4 up to 23% [93]. 5.4.4

Cu-MOFs

HKUST-1 with a formula of Cu3 (BTC)2 is the most typical Cu-MOF as the filler in MMMs. Cu3 (BTC)2 is assembled from di-copper paddlewheel secondary building units (SBUs) and benzene-1,3,5-tricarboxylate (BTC) linkers, forming a 3D framework with an intersecting 3D channel system [94]. Cu3 (BTC)2 possesses channels with a size of 9 Å allowing the fast permeation of guests and triangular windows with a size of 3.5 Å suitable for molecular sieving separation. Upon removing guest molecules via activation, unsaturated coordination copper sites can be exposed in Cu3 (BTC)2 , leading to stronger

5.4 MOF-Based MMMs

interactions between Cu-MOFs and guests. Up to date, Cu3 (BTC)2 crystals have been explored as promising fillers in various kinds of polymers, such as PI, PSF, PDMS, poly(2,6-dimethyl-1,4-phenylene oxide), etc. [95–98] For example, Duan et al. reported an incorporation of Cu3 (BTC)2 into Ultem to form MMMs [99]. Cu3 (BTC)2 particles were uniformly dispersed, and there were no interfacial defects in the prepared MMMs when Cu3 (BTC)2 loading was less than 35 wt%. Gas permeability increased obviously with increasing Cu3 (BTC)2 loading, and CO2 /N2 and CO2 /CH4 selectivities were kept unchanged. For MMM with the best separation property, CO2 permeability increased about 2.6 times, and CO2 /N2 selectivity remained almost unchanged. Kim et al. prepared a Cu-based MOF as the filler and further explored the influence of matrix (such as amorphous poly(2-ethyl-2-oxazoline) [POZ] and semicrystalline PEBAX) on membrane separation [100]. They found that the amide segment in PEBAX would contribute more to the CO2 solubility than ether segment. The selectivity of CO2 /N2 was enhanced significantly with the addition of a Cu-MOF, and the values were higher in the Cu-MOF/PEBAX MMM compared with that in a POZ-based asymmetric MMM. Improvement in the CO2 /N2 selectivity of a Cu-MOF/PEBAX MMM was achieved via facilitated transport by the CO2 -selective Cu-MOFs due to both their high adsorption selectivity of CO2 over N2 and the decreased crystallinity of PEBAX due to the presence of the Cu-MOFs, which would provide a synergic effect on the CO2 separation. 5.4.5

MOF-74 Series

MOF-74 series is composed of M2+ (M = Mg, Mn, Fe, Co, Ni, Cu, or Zn) and 2,5-dioxido-1,4-benzenedicarboxylate (dobdc4− ), possessing 1D hexagonal channels rich of unsaturated metal sites ideal for the selective adsorption of polar gases (e.g. CO2 , H2 S) over other nonpolar gases (e.g. N2 , CH4 ) [101]. One typical example is Mg-MOF-74 with 1D hexagonal channels of 11 Å in diameter, which can provide fast guest transport pathways when embedded in the polymer matrix. Bae and Long explored the effect of Mg-MOF-74 nanoparticles (around 100 nm in size) on membrane separation performance in both glassy and rubbery polymer matrices [102]. They observed simultaneously increased gas permeability and selectivity in the glassy polymer matrix differing from the decreased permeability in the rubbery polymer matrix after adding Mg-MOF-74 fillers. This discrepancy was ascribed to the penetration of flexible polymer chains of the rubbery matrix, therefore blocking MOF pores in MMMs. On the contrary, rigid chains of the glassy matrix hardly penetrated into MOF pores, allowing them to be accessible to gases. Ten-Binh et al. dispersed Mg-MOF-74 into PIM-1 matrix to form defect-free MMMs. In situ cross-linking was performed between polymer and MOF, creating strong interaction between PIM-1 chains and Mg-MOF-74 crystal surfaces [103]. Maserati et al. prepared diamine-appended Mg2 (dobpdc) nanorods and explored them as promising fillers in MMMs [104]. The dynamic uptake and release of CO2 along the diamine groups in MOFs largely enhance the membrane CO2 permeability and CO2 /N2 selectivity. Figure 5.10 summarizes all types of MOF-based MMMs in a representative CO2

179

5 Mixed Matrix Membranes

104 αCO2/CH4 (MMMs)

Selectivity (CO2/CH4 and CO2/N2)

2008 Upper bound for CO2/CH4

αCO2/CH4 (Polymer) αCO2/N2 (MMMs)

103

αCO2/N2 (Polymer)

102

101

100 10–4 10–3 10–2 10–1

(b)

100

101

102

103

104

Permeability of CO2

(a) Selectivity factor (SF) of CO2/CH4 and CO2/N2

180

- UiO-66 - ZIFs - MILs - Cu, Mg, others

Trade-off region

101

B

A 100

100

D

Enhanced region

C

101

Trade-off region

Permeability factor (PF) of CO2

Figure 5.10 Profile of metal–organic framework fillers in mixed matrix membrane for CO2 separation: (a) actual permeability and selectivity and (b) permeability and selectivity factors (filled and unfilled symbols represent the selectivity factor of CO2 /CH4 and CO2 /N2 , respectively) [105]. Source: Copyright 2017, reproduced with permission from Elsevier.

separation. In general, MMMs with MOF fillers shows increased permeability of about 98% than pure polymer membranes (Figure 5.10a). However, about 10–15% MOF filler fall in trade-off region due to low selectivity (Figure 5.10b). From the results, it can be inferred that about 70% MOF fillers in MMMs show 100% enhancement in CO2 permeability and selectivity than pure polymer membrane.

5.5 POF-Derived MMMs

5.5 POF-Derived MMMs Although MOF fillers have attracted considerable attention, alternative crystalline and noncrystalline organic porous materials, such as covalent organic frameworks (COFs), and a broad range of porous organic polymers have also been explored very recently [106]. In contrast to MOFs, these fillers, which can be broadly referred to as POFs, have entirely organic extended frameworks with either crystalline (e.g. COFs) [107] or amorphous structures (e.g. porous aromatic frameworks [PAFs]) [108]. The distinct advantages of POFs are their organic structures, which display excellent chemical compatibility with the organic polymer phase, and the chemical stability conferred by irreversible covalent bonding, particularly for materials such as PAFs and polymers of intrinsic microporosity (PIMs). Selected COFs also have chemically robust structures as a consequence of chemical modifications post-synthesis. In this section, the fabrication of MMMs from POFs, which has only recently been explored, will be outlined. PAFs, one of the classes of porous polymers, are 3D extended materials constructed by linking tetrahedral building units together through irreversible C—C coupling reactions [108]. PAFs can exhibit unusually high porosity and narrow pore size distributions for amorphous solids and, by virtue of their covalent backbone, are chemically robust. MMMs composed of PAFs and PTMSP were first explored by Lau et al. [109]. PTMSP is a super-glassy polymer with excellent potential for gas separations; however, its practical use is limited by rapid aging, which leads to a significant decrease in its free volume (CO2 permeability of PTMSP decreases from 29 796 to 8045 Barrer over one year). An unforeseen benefit of adding PAF particles to PTMSP was that physical aging was dramatically halted in the host polymer. Indeed, CO2 permeability decreased by only approximately 7% after 240 days, nearly sixfold less than pure PTMSP membranes. Solid-state NMR experiments indicated that the pore network of the PAF incorporated the pendant methyl groups of the PTMSP polymer and as a result froze the polymer backbone in place (Figure 5.11). An important observation was that this mechanism was not general for nanoporous fillers (such as MOFs), as the chemical compatibility of the filler and polymer appeared to be essential. To this end, PAFs can also modify the aging properties of PIMs, another class of glassy polymer that show promising separation properties. In this case, the PAFs gave rise to selective aging in the MMM, where the H2 /N2 selectivity increased approximately threefold after 400 days of aging. A feature of PAFs is that their pores can be routinely functionalized to tune adsorbent/adsorbate interactions. This chemistry was exploited by Lau et al., who showed that MMMs composed of PTMSP and a series of functionalized PAFs (—NH2 , —SO3 H, C60 , and Li6 C60 ) showed a significant enhancement in permeability for industrially relevant gases (CH4 , CO2 , H2 , and N2 ) compared with neat PAF MMMs [110]. The best performing composite was found to be PTMSP/PAF-Li6 C60 , which showed both outstanding gas permeability and antiaging properties. For example, PTMSP/PAF-Li6 C60 membranes showed an 85% increase in CO2 permeability compared with their PTMSP/PAF counterparts.

181

182

5 Mixed Matrix Membranes

(a) Time

Figure 5.11 PAF/polymer MMM by intermixing glassy polymer and PAF-1: (a) PTMSP PMP and PIM-1 densify to give a non-permeable conformation, (b) but with the addition of PAF-1, and (c) the original permeable structure is maintained [109]. Source: Copyright 2014, reproduced with permission from John Wiley & Sons.

(b) PAF-1

(c) Time

In addition to PAF-1, our group explored another PAF material (PAF-56P) as the filler and PSF as the polymer matrix to fabricate MMMs [111]. PAF-56P prepared by the cross-coupling reaction of triangle-shaped cyanuric chloride and linear p-terphenyl monomers (Figure 5.12a) exhibited an extended conjugated network with a pore size of 1.2 nm. PAF-56P was subsequently integrated with PSF matrix into PAF-56P/PSF asymmetric hollow fiber membranes via the dry jet–wet quench method employing PAF-56P/PSF suspensions. Fabricated PAF-56P/PSF membranes were further exploited for CO2 capture, which was exemplified by gas separations of CO2 /N2 mixtures (Figure 5.12b). The PAF-56P/PSF membranes showed a high selectivity of CO2 over N2 with a separation factor of 38.9 due to the abundant nitrogen groups in the PAF-56P framework. A preferred permeance for CO2 in the binary CO2 /N2 gas mixture was obtained in the range of 93–141 GPU due to the large CO2 adsorption capacity and a large pore size of PAF-56P. Besides the good separation property, PAF-56P/PSF hollow fiber membranes could be easily scaled up. Hypercrosslinked polymers (HCPs) were also explored as MMM fillers [112]. A clear advantage that this class of materials offer is their facile synthesis relative to PAFs [113]. To this end, the aromatic-rich HCP 𝛼,𝛼 ′ -dichloro-p-xylene (p-DCX) was added to the glassy polymer PTMSP, and the performance characteristics of the material were examined. Indeed, p-DCX was found to be effective filler for decreasing physical aging, especially with respect to CO2 permeability. For example, CO2 permeability of PTMSP/p-DCX MMMs was reduced by only 2% after 60 days of aging, while a 40% reduction in N2 permeability was observed for the as-cast membrane. Such relative aging gave rise to outstanding CO2 /N2 selectivity. NMR studies inferred that the fundamental antiaging mechanisms of the HCP fillers were similar to PAFs, namely, that interactions between the p-DCX

5.5 POF-Derived MMMs

N N N

Cl + Cl

(a)

N

N N

Cl

N N N

N N N

N N N

N N N

AlCl3 CHCl3

N N N

(b)

Figure 5.12 (a) The synthesis protocol for preparing PAF-56P and (b) PAF-56P/PSF asymmetric hollow fiber membranes for CO2 /N2 separation.

and the trimethylsilyl groups of PTMSP froze the main chains of the bulk polymer. HCPs (polystyrene [PS]) have also shown positive antiaging effects when used as fillers in PIM-based membranes [114]. Analogously, aging experiments indicated that the loss of permeability of N2 was more rapid than that of CO2 over time, thus leading to an enhanced CO2 /N2 selectivity from 7 to 12. COFs are a class of extended porous materials that are defined by their crystalline structures [107]. Given their high porosity, organic composition, and wide range of pore sizes, it is surprising that their application as porous MMM additives has not been widely explored. Only a few studies of 2D COFs as additives to polymer matrices have been reported; this may be in part due to challenging synthetic procedures and, for some COFs, limited stability under humid conditions. Recently, water-stable, exfoliated imine-based COFs were added to poly(ether imide) or polybenzimidazole (PBI) to yield MMMs [115]. The as-synthesized COFs were exfoliated by sonication, thereby giving rise to sheets down to monolayer thickness with high aspect ratios. MMMs with up to 30% COF filler were cast, and, according to microscopy studies, the membranes were defect-free. This was attributed to good compatibility between the COF and the polymer matrix. In general, the COF additives lead to increased gas permeability and in some cases improvement in gas selectivity. A clear increase

183

184

5 Mixed Matrix Membranes

in the H2 /CO2 selectivity from 9 up to 31 upon 20 wt% COF loading (surpassing the 2008 Robeson upper bound) was observed. Similarly, robust COFs (TPa-1 and TPBD) were employed to synthesize MMMs with PBI [116]. In this case, membranes composed of up to 50% COF could be cast before defects in the material were observed (Figure 5.13), once again indicating the excellent chemical compatibility of COFs and polymer hosts. These high COF loadings showed significant increases (almost sevenfold) in gas permeability for H2 , N2 , CH4 , and CO2 compared with tert-butylpolybenzimidazole (PBI-BuI). Furthermore, as expected, the gas permeability increased as the pore size of the COF increased. MMMs composed of azine-linked COFs and the commercial polymer Matrimid were also been fabricated [117]. An addition of the COF material led to a notable increase in gas permeability (130% for CO2 in the case of the 16 wt% loaded polymer) and an increase in selectivity for CO2 /CH4 from approximately 19–27 compared with the neat polymer.

5.6 MMMs Containing Other Porous Fillers The recent use of porous molecular compounds rather than extended porous materials as fillers is outlined in this section, firstly summarizing the role of MOPs as fillers before discussing applications of the purely organic equivalents (POCs) [118]. In contrast to extended framework fillers, the porous molecular filler may be fully dispersed within the membrane support or be present as a suspension of crystallites [119]. Thus, the potential porosity of the molecular filler may be intrinsic (utilizing only the shape-persistent cavity of the species), extrinsic (relying on the voids formed from crystalline or frustrated amorphous packing of the cage units), or a combination [120]. The alkyl chain-decorated metal–organic polyhedra (MOP-18 [Cu24 (ddbdc)24 (S)24 ], ddbdc = 5-dodecyloxy-1,3-benzenedicarboxylate) was first reported as a component of MMMs [121]. Inspired by the work of Kim and coworkers who used it as a membrane channel builder for cation transport [122], the authors sought to prepare MMMs by dispersing MOP-18 in Matrimid. MOP-18 was exohedrally decorated with alkyl chain substituents in the 5-position of the ligand, which enhanced its solubility in organic solvents. Homogeneous and phase-integrated MMMs with loadings as high as 80 wt% were achieved. No rigidification occurred at the MOP-18/polymer interface, with up to 44 wt% MOP-18 in MOP-18/Matrimid MMMs. This was presumably due to the strong affinity of the polymer chains for the alkyl chains of MOP-18. The MOP-18/Matrimid membranes became more permeable as the temperature was increased to 70 ∘ C, and CO2 plasticization was minimized. Permeability and solubility data showed that the pore, cage void, and alkyl chains of embedded MOP-18 introduced new sorption sites that significantly affected the gas transport properties of the membranes. A similar approach was implemented by Ma et al. who incorporated a soluble anionic MOP (Na6 H18 -[Cu24 (5-SO3 -1,3-BDC)24 (S)24 ]⋅xS, S = methanol and N,N ′ -dimethylacetamide) as a filler in PSF [123]. The design strategy was to use the aromatic rings of both MOP and PSF to enhance the chemical

3.4 Å

3.5 Å

H 2N

NH 2

+ CHO HO

18 Å

OH

OHC

24 Å

CHO OH

+ H2N

TpBD

TpPa-1 +

(a)

NH2

TpPa-1@PBI-Bul H N

N N H

N

n

PBI-Bul TpBD@PBI-Bul

85 °C,16 h

(b)

Evaporation

(c)

Figure 5.13 (a) Schematic representations of the synthesis of COFs and their packing models showing pore aperture and stacking distance, (b) the solution-casting method for fabricating COF@PBI-BuI membranes, and (c) photographs of TpPa-1 and TPBD(50)@PBI-BuI membranes [116]. Source: Copyright 2016, reproduced with permission from John Wiley & Sons.

186

5 Mixed Matrix Membranes

Aging

Aging

Figure 5.14 Schematic representation showing the postulated role that MOPs have in preventing aging and enhancing selectivity over time [124]. Source: Copyright 2015, reproduced with permission from RSC.

compatibility as well as the sulfonate groups to augment CO2 binding and deliver effective separation of CO2 /CH4 mixtures. Solution casting gave an MMM with homogeneous dispersion of MOP cages, which retained their chemical connectivity. Analysis of the separation performance for 8, 12, and 18 wt% MOP loadings revealed that the permeabilities of CO2 and CH4 both increased with an increase in MOP loading, with permeabilities up to 113% and 76% higher, respectively, than those in the pure PSF membrane. The separation factor for CO2 in the mixture also showed significant improvement compared with a pure PSF membrane (from 28 up to 45, a 60% increase). Although the higher permeability was largely dependent on new diffusion pathways, the improved separation factor was attributed to the polar —SO3 Na groups, which interacted more strongly with quadrupolar CO2 . Kitchin et al. set about carefully controlling the interplay between components in an MMM at the molecular level to achieve control over the process of physical aging [124]. The authors utilized PTMSP, which is highly prone to physical aging, and a series of soluble MOPs with various external chain lengths and polarities, namely, a nonpolar tertiary butyl (t Bu-MOP), polar diethylene glycol (DEG-MOP), polar triethylene glycol (TEG-MOP), and nonpolar dodecane (MOP-18), were used to prepare MOP-MMMs. Although all the MOP-PTMSP combinations had slightly lower initial CO2 permeability, the MOP fillers with shorter chains (i.e. t Bu-MOP, DEG-MOP) had a reduced loss in permeability over a one-year-aging cycle (Figure 5.14): the 20% loaded t Bu-MOP-MMM only lost 20% of its CO2 permeability compared with the 73% loss exhibited by the pure PTMSP membrane. As a predictive tool, the antiaging performance was found to correlate with the viscosity of the casting solutions as well as the level of interaction between polymer and MOP. Thus, MOPs can be incorporated into MMMs without phase segregation, and such materials with the right composition can deliver superior separation performance as well as can more resist physical aging compared with a neat polymer membrane. Porous organic cages (POCs) have also recently attracted significant attention in the membrane application owing to chemical stability, readily solubility, and solely organic skeleton [120]. Jiang et al. investigated Matrimid/β-cyclodextrin membranes for O2 /N2 separation and found changes in permeability and selectivity that were dependent on the loading [125]. Along these lines, Chapala et al. systematically studied the effect of incorporating α-, β-, and 𝛾-cyclodextrins

5.6 MMMs Containing Other Porous Fillers

with Me or Me3 Si substituents into a poly(3-trimethylsilyltricyclononene-7) (PTCNSi1) matrix [126]. They observed that bulky Me3 Si groups led to minor reductions in the permeability coefficients for He and H2 , while a marked decrease was found for other gases. Methyl-substituted cyclodextrins led to reductions in permeability for all gases, although this effect was more significant for larger gas molecules. Thus, an MMM composed of Me-substituted α-cyclodextrin resulted in an increase in selectivity for H2 /N2 from 5.2 for the pure polymer to 9.1. Bushell et al. reported the first synthesis of a POC-based MMM, whereby the cage molecules were induced to crystallize within the matrix [119]. Starting from a solution of PIM-1 and POC, a dispersed but phase-separated composite was generated by in situ crystallization (Figure 5.15). The cage was an imine POC O

HO

NH2

CN NH2

O

+

OH

HO

F

F

+ OH

O

F

F CN

O

CN O

O

O CN (a)

n

(b) 100.0 μm

(c)

Figure 5.15 (a) Porous imine cage CC3 synthesized from 1,3,5-triformylbenzene and (R,R)-1,2-diaminocyclohexane by a condensation reaction, (b) PIM-1 synthesized from 5,5′ ,6,6′ -tetrahydroxy-3,3,3′ ,3′ -tetramethyl-1,1′ -spirobisindane and 1,4-dicyanotetrafluorobenzene, and (c) a cross-sectional SEM image of a CC3/PIM-1 membrane [119]. Source: Copyright 2013, reproduced with permission from John Wiley & Sons.

187

188

5 Mixed Matrix Membranes

formed from 1,3,5-triformylbenzene and (R,R)-1,2-diaminocyclohexane (CC3). This study showed that the incorporation of POCs significantly enhanced the permeability, whereas nonporous cage molecules (in this case generated by chemical reduction of the imine POC CC3) had the opposite effect; in the former case permeability increased with increasing weight percent of the porous cage, while in the latter permeability decreased as the concentration of the nonporous cages increased. The effect of using preformed nanocrystals (PIM-1/nanocage) was also investigated. The authors concluded that the PIM-1/nanocage membranes extended the upper bound of performance for various relevant gas pairs, while the in situ crystallized systems provided better resistance toward physical aging. In a computational study, Evans et al. assessed how POCs acted as soluble fillers that could ameliorate nonselective gas transport pathways in MMMs [127]. Five POCs were investigated, comprising three different families of materials: the tetrahedral imine cages (CC1, CC2, CC3), an adamantoid cage, and a triangular dipyramidal cage. The PIM-1/cage MMMs were generated in silico and benchmarked against experimental data reported [119]. The power of this approach was to allow analysis of numerous polymer/POC compositions, thereby generating the permeabilities and selectivities for 40% volume compositions of MMMs composed of POCs and the polymer hosts (Matrimid, Ultem, PIM-1, PIM-7). This revealed that larger cage structures in the MMMs significantly improved the permeability for H2 /N2 and H2 /CO2 separations and concomitant with a minor increase in the selectivity for H2 . Mao and Zhang investigated MMMs incorporating the waterwheel-shaped POC Noria [128]. Noria was first synthesized from resorcinol and pentanedial and shown to be porous by Atwood and coworkers [129, 130]. The hydroxy groups render Noria soluble in polar solvents and allow its physical properties to be tailored by post-synthesis modification. MMMs were prepared by incorporating Noria as well as its derivatives Noria-Boc and Noria-COt Bu as the fillers in the PI 6FDA-DAM. The substituted derivatives achieved better integration, giving a homogeneous dispersion of the nanoaggregates and close interfacial mixing of the phases, particularly in the cases where hydrophobic substituents were used. The separation performance of the resultant membranes was strongly related to the chemical structures of the fillers. Noria/6FDA-DAM membranes gave a minor improvement in CO2 /CH4 selectivity (increase of 15%) concomitant with a 53% decrease in permeability for CH4 . In contrast, the introduction of Noria-COt Bu tended to increase the free volume and gas permeability of the MMMs (e.g. CH4 permeability increased by nearly 40%). Cooper and coworkers recently also utilized their tetrahedral imine cages to render materials porous by solution co-processing [131]. Scrambled POC cage mixtures were prepared by a co-reaction approach that yielded an amorphous material with a porosity almost twice that of the corresponding phase-pure POCs. A combinatorial approach was then used to explore the effect of doping with a series of nonporous polymers, including PEI, PVP, PMMA, and PS. A notable observation was that the cage-polymer samples maintained their porosity to N2 up to 40 wt% polymer loading, which confirmed the ability to convert nonporous commodity polymers into porous ones simply by combination with POCs. Desirable CO2 uptake was achieved for PEI/POC (17 wt% PEI), with the

References

material delivering a higher CO2 uptake than either of the two isolated organic components as well as an ideal CO2 /N2 gas selectivity of 8 (295 K, 1.0 bar).

5.7 Conclusions In this chapter, various types of advanced microporous fillers such as zeolites, MOFs, and POFs have been overviewed for the fabrications of MMMs. The properties of fillers such as size, shape, porous nature, and surface moieties have been reviewed to improve the interfacial interaction between polymers and fillers and thus effectively reduce the formation of void defects in the MMMs. Matured methods have been introduced on how to produce MMMs. This chapter also deals with the state-of-the-art membrane systems for gas separations. It is also acknowledged that porous fillers can enhance gas separation in MMMs by virtue of allowing gas molecules through fillers and facilitating gas molecules diffusion through matrices via changing the polymer chain orientation. Detailed examples in each catalogue of porous fillers have been included to elucidate the structure–function relationship. MMMs possess remarkable performance and overcome the drawbacks like low permeability and/or selectivity. In conclusion, the MMM has a promising future for gas separation due to its excellent physical and chemical properties as compared with pure organic or inorganic membranes. On the contrary, it faces some challenges regarding interfacial defects or adhesion between the organic and inorganic phases, thus inspiring many modifications that have to be carried out to overcome these problems. Current researches focus on overcoming these challenges and developing new materials that could eliminate such drawbacks. From the available reports, it is understood that among various fillers employed in MMMs, MOFs are recognized as potential fillers to overcome the trade-off limitation of membranes. Surface modifications of the inorganic zeolite phase and changes in the membrane casting conditions have been suggested to eliminate the formation of defects in large-scale MMMs. Much effort is dedicated to MMM formation for gas separation, but more research and development approaches are still needed to exploit the potential of this technology for synthesis of high-performance MMMs for industrial-scale gas separation.

References 1 Maier, G. (1998). Angew. Chem. Int. Ed. 37: 2960–2974. 2 Rousseau, R.W. (1987). Handbook of Separation Process Technology. New

York: Wiley-Interscience. 3 Nunes, S.P. and Peinemann, K.V. (2001). Membrane Technology. Wiley

Online Library. 4 Spillman, R. and Sherwin, M. (1990). Chem. Tech. 20: 378–384. 5 Paul, D.R. and Yampol’skii, Y.P. (1993). Polymeric Gas Separation Mem-

branes. Boca Raton, FL: CRC Press.

189

190

5 Mixed Matrix Membranes

6 Stern, S.A. (1994). J. Membr. Sci. 94: 1–65. 7 Mulder, M. (1996). Basic Principles of Membrane Technology. New York:

Springer. 8 Ho, W. and Sirkar, K.K. (1992). Membrane Handbook. New York: Van

Nostrand Reinhold. 9 Kesting, R.E. and Fritzsche, A. (1993). Polymeric Gas Separation Mem-

branes. New York: Wiley. 10 Chung, T.S., Jiang, L.Y., Li, Y., and Kulprathipanja, S. (2007). Prog. Polym.

Sci. 32: 483–507. 11 Aroon, M., Ismail, A., Matsuura, T., and Montazer-Rahmati, M. (2010). Sep.

Purif. Technol. 75: 229–242. 12 Robeson, L.M. (2008). J. Membr. Sci. 320: 390–400. 13 Zimmerman, C.M., Singh, A., and Koros, W.J. (1997). J. Membr. Sci. 137:

145–154. 14 Idris, A. and Ahmed, I. (2008). J. Appl. Polym. Sci. 108: 302–307. 15 Bohbot-Raviv, Y. and Wang, Z.G. (2000). Phys. Rev. Lett. 85: 3428–3431. 16 Odani, H., Taira, K., Tamaguchi, T. et al. (1975). Bull. Inst. Chem. Res. Kyoto

Univ. 53: 409–423. 17 Knight, P. and Lyman, D. (1984). J. Membr. Sci. 17: 245–254. 18 Blume, I. and Pinnau, I. (1990). Composite membrane, method of prepara-

tion and use. US Patent 4963165. 19 Murali, R.S., Sridhar, S., Sankarshana, T., and Ravikumar, Y. (2010). Ind.

Eng. Chem. Res. 49: 6530–6538. 20 Robeson, L.M. (1999). Curr. Opin. Solid State Mater. Sci. 4: 549–552. 21 Pinnau, I. and Freeman, B.D. (2000). Membrane Formation and Modifica-

tion. Washington, DC: American Chemical Society. 22 Baker, R.W. (2004). Membrane Technology and Applications. New York:

Wiley. 23 Mahajan, R., Burns, R., Schaeffer, M., and Koros, W.J. (2002). J. Appl. Polym. 24 25 26 27 28 29 30 31 32 33 34 35 36

Sci. 86: 881–890. Vu, D.Q., Koros, W.J., and Miller, S.J. (2003). J. Membr. Sci. 211: 311–334. Süer, M.G., Baç, N., and Yilmaz, L. (1994). J. Membr. Sci. 91: 77–86. Itta, A.K., Tseng, H.H., and Wey, M.Y. (2011). J. Membr. Sci. 372: 387–395. Tseng, H.H. and Itta, A.K. (2012). J. Membr. Sci. 389: 223–233. Vu, D.Q., Koros, W.J., and Miller, S.J. (2003). J. Membr. Sci. 221: 233–239. Ismail, A.F., Rahim, N.H., Mustafa, A. et al. (2011). Sep. Purif. Technol. 80: 20–31. Dechnik, J., Gascon, J., Doonan, C.J. et al. (2017). Angew. Chem. Int. Ed. 56: 9292–9310. Miller, S.J., Kuperman, A., and Vu, D.Q. (2007). US Patent 7166146B2. Tantekin-Ersolmaz, S¸ .B., Atalay-Oral, Ç., Tatl𝚤er, M. et al. (2000). J. Membr. Sci. 175: 285–288. Maxwell, J.C. (1873). Treatise on Electricity and Magnetism. London: Oxford University Press. Bruggeman, D.A.G. (1935). Ann. Phys. 416: 636. Böttcher, C.J.F. (1945). Recl. Trav. Chim. Pays-Bas 64: 47. Higuchi, W.I. and Higuchi, T. (1960). J. Am. Pharm. Assoc. (Sci. Ed.) 49: 598.

References

37 Lewis, T.B. and Nielsen, L.E. (1970). J. Appl. Polym. Sci. 14: 1449. 38 Pal, R. (2007). J. Reinf. Plast. Compos. 26: 643. 39 Gonzo, E.E., Parentis, M.L., and Gottifredi, J.C. (2006). J. Membr. Sci. 277:

46. 40 Kang, D.Y., Jones, C.W., and Nair, S. (2011). J. Membr. Sci. 381: 50–63. 41 Vinh-Thang, H. and Kaliaguine, S. (2013). Chem. Rev. 113: 4980–5028. 42 Bastani, D., Esmaeili, N., and Asadollahi, M. (2013). J. Ind. Eng. Chem. 19:

375–393. 43 Rezakazemi, M., Shahidi, K., and Mohammadi, T. (2012). Int. J. Hydrog.

Energy 37: 14576–14589. 44 Fernández-Barquín, A., Casado-Coterillo, C., Valencia, S., and Irabien, A.

(2016). Membranes 6: 28. 45 Beiragh, H.H., Omidkhah, M., Abedini, R. et al. (2016). Asia Pac. J. Chem.

Eng. 11: 522–532. 46 Dorosti, F., Omidkhah, M., and Abedini, R. (2015). J. Nat. Gas Sci. Eng. 25:

88–102. 47 Zhang, Y., Balkus, K.J. Jr., Musselman, I.H., and Ferraris, J.P. (2008). J.

Membr. Sci. 325: 28–39. 48 Zarshenas, K., Raisi, A., and Aroujalian, A. (2016). J. Membr. Sci. 510:

270–283. 49 Martin-Gil, V., López, A., Hrabanek, P. et al. (2017). J. Membr. Sci. 523:

24–35. 50 Zhao, D., Ren, J., Li, H. et al. (2014). J. Energy Chem. 23: 227–234. 51 Li, Y., Guan, H.M., Chung, T.S., and Kulprathipanja, S. (2006). J. Membr. Sci.

275: 17–28. 52 Gorgojo, P., Uriel, S., Téllez, C., and Coronas, J. (2008). Microporous Meso-

porous Mater. 115: 85–92. 53 Adams, R.T., Lee, J.S., Bae, T.H. et al. (2011). J. Membr. Sci. 367: 197–203. 54 Dorosti, F., Omidkhah, M., Pedram, M., and Moghadam, F. (2011). Chem.

Eng. J. 171: 1469–1476. 55 Karkhanechi, H., Kazemian, H., Nazockdast, H. et al. (2012). Chem. Eng.

Technol. 35: 885–892. 56 Hussain, M. and König, A. (2012). Chem. Eng. Technol. 35: 561–569. 57 Chen, X., Nik, O., Rodrigue, D., and Kaliaguine, S. (2012). Polymer 53: 58 59 60 61 62 63 64 65 66

3269–3280. Yong, H.H., Park, H.C., Kang, Y.S. et al. (2001). J. Membr. Sci. 188: 151–163. S¸ en, D., Kal𝚤pç𝚤lar, H., and Yilmaz, L. (2007). J. Membr. Sci. 303: 194–203. Husain, S. and Koros, W.J. (2007). J. Membr. Sci. 288: 195–207. Junaidi, M.U.M., Khoo, C.P., Leo, C.P., and Ahmad, A. (2014). Microporous Mesoporous Mater. 192: 52–59. Li, Y., Krantz, W.B., and Chung, T.S. (2007). AIChE J 53: 2470–2475. Seoane, B., Coronas, J., Gascon, I. et al. (2015). Chem. Soc. Rev. 44: 2421–2454. Cavka, J.H., Jakobsen, S., Olsbye, U. et al. (2008). J. Am. Chem. Soc. 130: 13850–13851. Hu, Z., Peng, Y., Kang, Z. et al. (2015). Inorg. Chem. 54: 4862–4868. Shen, J., Liu, G., Huang, K. et al. (2016). J. Membr. Sci. 513: 155–165.

191

192

5 Mixed Matrix Membranes

67 Smith, S.J.D., Ladewig, B.P., Hill, A.J. et al. (2015). Sci. Rep. 5: 7823. 68 Khdhayyer, M.R., Esposito, E., Fuoco, A. et al. (2017). Sep. Purif. Technol.

173: 304–313. 69 Venna, S.R., Lartey, M., Li, T. et al. (2015). J. Mater. Chem. A 3: 5014–5022. 70 Phan, A., Doonan, C.J., Uribe-Romo, F.J. et al. (2010). Acc. Chem. Res. 43:

58–67. 71 Park, K.S., Ni, Z., Côté, A.P. et al. (2006). Proc. Natl. Acad. Sci. U. S. A. 103:

10186–10191. 72 Song, Q., Nataraj, S.K., Roussenova, M.V. et al. (2012). Energy Environ. Sci.

5: 8359–8369. 73 Ban, Y., Li, Z., Li, Y. et al. (2015). Angew. Chem. Int. Ed. 127: 15703–15707. 74 Ordonez, M.J.C., Balkus, K.J., Ferraris, J.P., and Musselman, I.H. (2010). J.

Membr. Sci. 361: 28–37. 75 Md Nordin, N.A.H., Racha, S.M., Matsuura, T. et al. (2015). RSC Adv. 5:

43110–43120. 76 Wang, Z., Wang, D., Zhang, S. et al. (2016). Adv. Mater. 28: 3399–3405. 77 Gao, Y., Qiao, Z., Zhao, S. et al. (2018). J. Mater. Chem. A 6: 3151–3161. 78 Kong, C., Du, H., Chen, L., and Chen, B. (2017). Energy Environ. Sci. 10:

1812–1819. 79 Li, T., Pan, Y., Peinemann, K.V., and Lai, Z. (2013). J. Membr. Sci. 425–426:

235–242. 80 Boroglu, M.S. and Yumru, A.B. (2017). Sep. Purif. Technol. 173: 269–279. 81 Hao, L., Liao, K.S., and Chung, T.S. (2015). J. Mater. Chem. A 3:

17273–17281. 82 Bae, T.H., Lee, J.S., Qiu, W. et al. (2010). Angew. Chem. Int. Ed. 49:

9863–9866. 83 Ban, Y., Li, Y., Peng, Y. et al. (2014). Chem. Eur. J. 20: 11402–11409. 84 Deng, Y.H., Chen, J.T., Chang, C.H. et al. (2016). Angew. Chem. Int. Ed. 55:

12793–12796. 85 Zhang, C., Zhang, K., Xu, L. et al. (2014). AIChE J 60: 2625–2635. 86 Loiseau, T., Serre, C., Huguenard, C. et al. (2004). Chem. Eur. J. 10:

1373–1382. 87 Zhu, H., Wang, L., Jie, X. et al. (2016). ACS Appl. Mater. Interfaces 8:

22696–22704. 88 Abedini, R., Omidkhah, M., and Dorosti, F. (2014). RSC Adv. 4:

36522–36537. 89 Sabetghadam, A., Seoane, B., Keskin, D. et al. (2016). Adv. Funct. Mater. 26:

3154–3163. 90 Tien-Binh, N., Vinh-Thang, H., Chen, X.Y. et al. (2015). J. Mater. Chem. A 3:

15202–15213. 91 Dong, X., Liu, Q., and Huang, A. (2016). J. Appl. Polym. Sci. 133: 43485. 92 Xin, Q., Ouyang, J., Liu, T. et al. (2015). ACS Appl. Mater. Interfaces 7:

1065–1077. 93 Anjum, M.W., Bueken, B., De Vos, D., and Vankelecom, I.F.J. (2016). J.

Membr. Sci. 502: 21–28. 94 Chui, S.S.Y., Lo, S.M.F., Charmant, J.P.H. et al. (1999). Science 283:

1148–1150.

References

95 Yu, S., Jiang, Z., Ding, H. et al. (2015). J. Membr. Sci. 481: 73–81. 96 Basu, S., Cano-Odena, A., and Vankelecom, I.F.J. (2010). J. Membr. Sci. 362:

478–487. 97 Zornoza, B., Seoane, B., Zamaro, J.M. et al. (2011). ChemPhysChem 12:

2781–2785. 98 Ge, L., Zhou, W., Rudolph, V., and Zhu, Z. (2013). J. Mater. Chem. A 1:

6350–6358. 99 Duan, C., Kang, G., Liu, D. et al. (2014). J. Appl. Polym. Sci. 131: 8828–8837. 100 Kim, J., Choi, J., Kang, Y.S., and Won, J. (2016). J. Appl. Polym. Sci. 133:

42853. 101 Rosi, N.L., Kim, J., Eddaoudi, M. et al. (2005). J. Am. Chem. Soc. 127:

1504–1518. 102 Bae, T.H. and Long, J.R. (2013). Energy Environ. Sci. 6: 3565–3569. 103 Tien-Binh, N., Vinh-Thang, H., Chen, X.Y. et al. (2016). J. Membr. Sci. 520:

941–950. 104 Maserati, L., Meckler, S.M., Bachman, J.E. et al. (2017). Nano Lett. 17:

6828–6832. 105 Vinoba, M., Bhagiyalakshmi, M., Alqaheem, Y. et al. (2017). Sep. Purif. Tech-

nol. 188: 431–450. 106 Zou, X.Q., Ren, H., and Zhu, G.S. (2013). Chem. Commun. 49: 3925–3936. 107 Waller, P.J., Gándara, F., and Yaghi, O.M. (2015). Acc. Chem. Res. 48:

3053–3063. 108 Ben, T., Ren, H., Ma, S. et al. (2009). Angew. Chem. Int. Ed. 48: 9457–9460. 109 Lau, C.H., Nguyen, P.T., Hill, M.R. et al. (2014). Angew. Chem. Int. Ed. 53:

5322–5326. 110 Lau, C.H., Konstas, K., Doherty, C.M. et al. (2015). Chem. Mater. 27:

4756–4762. 111 Meng, L.B., Zou, X.Q., Guo, S.K. et al. (2015). ACS Appl. Mater. Interfaces

7: 15561–15569. 112 Lau, C.H., Mulet, X., Konstas, K. et al. (2016). Angew. Chem. Int. Ed. 55:

1998–2001. 113 Dawson, R., Cooper, A.I., and Adams, D.J. (2012). Prog. Polym. Sci. 37:

530–563. 114 Mitra, T., Bhavsar, R.S., Adams, D.J. et al. (2016). Chem. Commun. 52:

5581–5584. 115 Kang, Z., Peng, Y., Qian, Y. et al. (2016). Chem. Mater. 28: 1277–1285. 116 Biswal, B., Chaudhari, H.D., Banerjee, R., and Kharul, U.K. (2016). Chem.

Eur. J. 22: 4695–4699. 117 Shan, M., Seoane, B., Rozhko, E. et al. (2016). Chem. Eur. J. 22:

14467–14470. 118 Tranchemontagne, D.J., Ni, Z., O’Keeffe, M., and Yaghi, O.M. (2008). Angew.

Chem. Int. Ed. 47: 5136–5147. 119 Bushell, A.F., Budd, P.M., Attfield, M.P. et al. (2013). Angew. Chem. Int. Ed.

52: 1253–1256. 120 Evans, J.D., Sumby, C.J., and Doonan, C.J. (2015). Chem. Lett. 44: 582–588. 121 Perez, E.V., Balkus, K.J., Ferraris, J.P., and Musselman, I.H. (2014). J. Membr.

Sci. 463: 82–93.

193

194

5 Mixed Matrix Membranes

122 Jung, M., Kim, H., Baek, K., and Kim, K. (2008). Angew. Chem. Int. Ed. 47:

5755–5757. 123 Ma, J., Ying, Y., Yang, Q. et al. (2015). Chem. Commun. 51: 4249–4251. 124 Kitchin, M., Teo, J., Konstas, K. et al. (2015). J. Mater. Chem. A 3:

15241–15247. 125 Jiang, L.Y. and Chung, T.S. (2009). J. Membr. Sci. 327: 216–225. 126 Chapala, P.P., Bermeshev, M.V., Starannikova, L.E. et al. (2015). Polym. Com-

pos. 36: 1029–1038. 127 Evans, J.D., Huang, D.M., Hill, M.R. et al. (2014). J. Phys. Chem. C 118:

1523–1529. 128 Mao, H. and Zhang, S. (2017). J. Colloid Interface Sci. 490: 29–36. 129 Kudo, H., Hayashi, R., Mitani, K. et al. (2006). Angew. Chem. Int. Ed. 45:

7948–7952. 130 Tian, J., Thallapally, P.K., Dalgarno, S.J. et al. (2009). Angew. Chem. Int. Ed.

48: 5492–5495. 131 Jiang, S., Chen, L., Briggs, M.E. et al. (2016). Chem. Commun. 52:

6895–6898.

195

6 Zeolite Membranes 6.1 Introduction The introduction of industrial membrane-based separation technologies can dramatically reduce the separation costs in comparison with thermally based separation technologies. In addition, membrane technologies allow the energy effective use and recovery of both valuable raw materials and the separation of wastes. Organic polymer membranes are increasingly used, but they suffer from stability at elevated temperatures and toward attack of organic solvents. Therefore, much effort is put into the development of temperature-stable and solvent-resistant inorganic membranes. In recent years there has been a strong focus on developing continuous membrane processes to replace traditional separation methods for applications that require significant cost and energy. Zeolites are potential candidates for use in such processes due to their microstructure, which consists of uniform molecule-sized pores connected by continuous diffusion pathways. The pore sizes of these materials are on the order of the molecule size of many industrially important compounds, which provide molecular sieving capability and unique adsorption properties. Additionally, the operating conditions able to be sustained by inorganic membranes far exceed that of polymeric membranes, which are commonly used in commercial separation processes. Development of new separation technologies such as high temperature gas separations and catalytic membrane reactors requires materials that are able to withstand extreme operating environments and zeolites present themselves as capable candidates. Suzuki reported the first zeolite membranes in 1987 [1]. Since then, significant progress has been made to improve the quality of zeolite membranes and widen their range of applications. Although there are a total of 239 zeolite structures available [2], only a few have been studied extensively for membrane applications. About 14 zeolite structures, including AFI [3], ANA [4], ATN [5], BEA [6], CHA [7, 8], DON [9], FAU [10–12], FER [13], GIS [14], LTA [15–17], MEL [18], MFI [19–28], MOR [29–31], and OFF [32], have been prepared as membranes. One of the most critical selection criteria when choosing a zeolite for a particular application is the pore size exhibited by the material. By far the most common membrane material studied is MFI-type zeolite (ZSM-5, silicalite-1) due to ease of preparation, pore size, control of microstructure, and versatility Microporous Materials for Separation Membranes, First Edition. Xiaoqin Zou and Guangshan Zhu. © 2019 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2019 by Wiley-VCH Verlag GmbH & Co. KGaA.

196

6 Zeolite Membranes

of applications. Silicalite-1 is made up of pure silica, whereas ZSM-5 has Al substituted for some of the Si atoms. Zeolite structures are made up of TO2 , where T represents tetrahedral framework atoms, such as Si, Al, B, Ge, Fe, and P. Most often, zeolites contain Si with other metals substituted into the framework. The zeolite pores are made from rings in the framework and are designated by the number of oxygen atoms making up the ring. Small-pore zeolites include those structures made up of 8-membered rings (MR), medium-pore zeolites have 10 MR, and large-pore zeolites have 12 MR. Some zeolites have a three-dimensional pore system with pores along all crystal axes. Other zeolites, however, have only one- or two-dimensional pore systems. Some of these structures have windows that allow limited movement along the axes with no pores, but others do not. The orientation of these types of zeolite membranes is therefore important. This chapter will cover the synthesis techniques of zeolite membranes, crystal growth in zeolite layers, microstructures in zeolite films, and various methods used to characterize zeolite membranes.

6.2 Synthesis Techniques for Zeolite Membranes Zeolite membranes are most often prepared by hydrothermal synthesis, although dry gel methods have also been used [30, 33, 34]. Hydrothermal synthesis involves crystallization of a zeolite layer onto a porous support from a gel that is usually composed of water, amorphous silica, a source for tetrahedral framework atoms other than Si, a structure-directing organic template, and sometimes a mineralizing agent, such as NaOH. This gel is placed in contact with the support in an autoclave. Zeolite membranes are generally synthesized as a thin continuous film about 2–20 μm thick on either mechanically stable metallic or ceramic porous supports (e.g. alumina, zirconia, quartz, silicon, stainless steel) to enhance their mechanical strength [21, 27]. Typical supported membrane synthesis follows one of two common growth methods: (i) in situ crystallization or (ii) secondary growth. The direct in situ crystallization under hydrothermal conditions is the most widely used method to prepare zeolite membranes on a porous ceramic or stainless-steel support [35]. During in situ crystallization, zeolite crystals nucleate and grow on the support surface. In general, dilute gels are preferred to avoid nucleation in the bulk solution [36]. Techniques have been developed to prepare membranes without organic template molecules [12, 37], and A-type membranes are usually prepared without a template, but if a template is used, the zeolite structure crystallizes around the organic template molecules, which form the pores upon template removal by calcination. A polycrystalline zeolite layer forms on the support, and this acts as the separating layer. In this route, the support is immersed in the zeolite synthesis solution, which can be a clear solution or an aqueous gel contained in polymer bottles for hydrothermal syntheses below 100 ∘ C or in an autoclave for syntheses at much higher temperature. For example, FAU or LTA membranes were prepared in a polypropylene

6.2 Synthesis Techniques for Zeolite Membranes

bottle between 80 and 100 ∘ C, while Teflon-lined autoclave was used for the preparation of MFI-type membranes at 180 ∘ C. Seed crystals are sometimes added on the support prior to the crystallization step to provide sites for zeolite growth and improve control of crystal growth. Using seed crystals is referred to as two-step crystallization. A two-step growth of MFI membranes was proposed by Vroon et al. [38]. In a first step, seed crystallites of 275–700 nm at relatively low temperature and high concentration of the crystallization batch are directly deposited on the support surface. Upon repeating the crystallization with fresh sol at elevated temperatures, a continuous zeolite layer with a thickness of 2–7 mm forms, which shows a separation factor for n-/i-butane of 50 at 25 ∘ C. Further repetitions of the crystallization step did not give any improvements. On the contrary, the oxidative decomposition of the template resulted in a crack formation of the thick zeolite membrane layer. Although the direct in situ crystallization can provide membranes of proven quality in gas separation, it has limitations. There is little scope for the control of the microstructure of the final films since synthesis conditions have to be optimized for nucleation and growth [39]. Two principal methods are used to suppress the effect of a homogeneous nucleation. One route is the so-called dry gel conversion [40, 41], either as vapor-phase transport method when the SDA is in the vapor but not in the dry parent gel [42] or a steam-assisted crystallization with a dry gel containing the SDA [34]. The use of zeolite nanoblocks is believed to trigger a new generation of extremely thin high-flux zeolite membranes [43]. The concept is based on coating a porous support with recently developed silica nanoblocks [44]. Silica polymerization in the presence of organic template molecules can yield identical rectangular silicalite-1 nanoblocks with the size of a few nanometers, for example, 4 × 4 × 1.4 nm3 . These nanoblocks can be isolated and applied for membrane preparation using the self-assembly properties of these nanoblocks supported by surfactants. The silica nanoblocks are negatively charged (zeta potential) like the silicalite-1 seed crystals. Therefore, the nanoblocks can be organized by cationic surfactants. Surfactants and organic templates can be removed by calcinations, and it should be noted that the surfactant molecule should not be too large since it can cause the formation of mesopores upon thermal decomposition. Coatings were made on porous alumina supports with pore sizes of approximately 100 and 50–60 nm. To prevent intrusion of nanoblocks into the support pores, the supports were first coated with one or two intermediate colloidal titania sol–gel layers, decreasing the pore size to approximately 2–3 nm [45]. Coatings were made using a mixture of silicalite-1 nanoblocks and surfactants by dip-coating flat or tubular supports. After drying and calcinations, an extremely thin supported silicalite-1 membrane is obtained. However, the separation behavior of these new membranes is still poor; the best membranes have cutoffs of 250 Da. Nevertheless, the use of small-scale nanoblocks opens new perspective for the preparation of ultrathin defect-free membranes. The challenges consist in achieving a perfect stacking with sufficient adhesion onto the porous support after removal of the surfactant molecule and in the intergrowth of the nanoblocks.

197

198

6 Zeolite Membranes

6.3 Crystal Growth in Zeolite Layers 6.3.1

Conventional Hydrothermal Synthesis

In conventional in situ synthesis of zeolite membrane under hydrothermal conditions, a gel layer on the surface of the support is formed by precipitation of the silica sol particles under certain concentration ranges and at given temperatures [46–50]. In the case of the MFI synthesis, the tetrapropylammonium ions (TPA+ ) are found only in the solution but not in the precipitated gel layer. It is concluded, therefore, that MFI crystallization starts at the phase boundary between the liquid phase (as TPA+ source) and the gel layer (as Si source). The crystals grow into the gel layer consuming the Si gel until the growing MFI crystals have reached the support. Using this synthesis route, different orientations of the MFI channel system relative to the support surface can be found. Often a b-orientation with straight channels perpendicular to the support can be obtained, which is a favorable orientation from the point of view of the anisotropy of mass transport in the MFI structure [51]. This b-orientation can be explained as follows: those nuclei that are oriented in parallel with the interface of the two nutrient pools in their fastest growth directions c and a show the largest growth rate and dominate the crystal orientation in the layer. To get pinhole-free zeolite membranes, the crystal size in the zeolite layer should be less than 1 mm [52] since large well-faceted polyhedral crystals do not grow together to a defect-free layer. Another important issue for obtaining high-quality membranes by in situ crystallization on supports is (i) a smooth support surface and (ii) a hydrophilic support surface with good wettability [53]. Rougher support surfaces provide larger contact angles (like the lotus effect) and the mass transfer in the in situ crystallization is reduced. Using porous titania supports (rutile), their wettability can be improved by UV radiation, which results in a better membrane quality. The deposited gel can either form a surface layer or can be soaked into the pore system of the support, forming zeolite plugs [54]. In the latter case, the plugged zeolite membranes exhibit an improved mechanical stability but relatively low fluxes. Whereas in the literatures methods are described to grow the zeolite membrane layer for stability reasons within the pores of the support rather than on its surface [36, 55, 56], other authors propose to seal the support pores by adsorbed species to prohibit the penetration of the gel into the pores of the support, thus having a high-flux membrane as a supported thin film. Using the latter technique to get higher fluxes, MFI membranes were prepared by in situ crystallization on porous α-Al2 O3 disks that contained a diffusion barrier to limit the excessive penetration of siliceous species into the support pores [57]. The barrier was introduced into the ceramic pores by polymerizing a previously adsorbed mixture of furfuryl alcohol and tetraethyl orthosilicate followed by carbonization and a partial carbon burn-off to generate a carbon-free region for chemical bonding of the MFI layer to the support. The resulting MFI membrane had a smaller thickness and showed increased flows.

6.3 Crystal Growth in Zeolite Layers

6.3.2

Two-Step Crystallization

The secondary growth of a supported seed layer is an effective approach for the formation of consolidated supported membranes and films with good quality and reproducibility [58, 59]. By decoupling the nucleation step (at high supersaturation) from crystal growth (at low supersaturation), the seeds can grow in low concentrated solutions under suppression of the secondary nucleation. Unlike the X-ray amorphous metal oxide membranes, a polycrystalline zeolite layer prepared by hydrothermal synthesis from seed crystals deposited before on the support surface brings about the necessity to control the crystal intergrowth so that pores between the individual zeolite crystals are avoided. This technique also requires a certain minimum membrane thickness or special techniques to achieve an orientation of the MFI zeolite crystals. Like in the direct in situ crystallization of zeolite membranes, masking techniques can be used to avoid the penetration of seeds and synthesis gel into the support pores. By a sophisticated polymethyl methacrylate (PMMA)–polyethylene wax treatment in a laminar flow bench at high temperatures under vacuum, the support pores were filled by the wax, which has a melting point above the synthesis temperature of 100 ∘ C. By this procedure, the pores of the support were protected from the synthesis solution. Using colloidal nucleation seeds followed by hydrothermal growth at 100 ∘ C, high-flux membranes with a thickness of approximately 0.5 mm could be prepared [20]. The power of the seeding technique was demonstrated by Matsukata et al. [60], showing that mordenite- or ZSM-5-type membranes could be prepared from identical organic-free aluminosilicate solutions under the same hydrothermal conditions by using either mordenite or MFI seeds. The seeding approach on zeolite membrane preparation was implemented since 1994 [61], and the use of seed crystals facilitates the formation of zeolite membranes since a seeded support grows to a pure-phase zeolite membrane more easily even when the crystallization conditions and the chemical batch composition are not optimum. There are three main ways to attach the seeds to the support: (i) variation of the pH to achieve that seeds and support have opposite surface charges (zeta potentials) for an electrostatic attachment, (ii) adsorption of positively charged polymers to recharge the surface as condition for the following electrostatic attachment of negatively charged seeds, and (iii) immersion of the dried support into a seed solution. Tsapatsis et al. [23, 62, 63] change the pH of the solution to adjust different zeta potentials between the ceramic support (e.g. α-Al2 O3 ) and the zeolite nanocrystals to be attached (e.g. silicalite-1 as pure SiO2 ). In Sterte’s reports [64–66] cationic polymers were adsorbed on the support to create a positive surface charge and then the negatively charged zeolite seeds such as silicalite-1 were attached. Later, this method, which was first developed for coating Si wafers, was successfully transformed for seeding porous ceramic supports for membrane preparation [67]. The use of seeded supports usually results in a c-orientation of the MFI layer, but under certain conditions also for secondary growth, the desired b-orientation can be obtained [58].

199

200

6 Zeolite Membranes

6.3.3

Crystallization by Microwave Heating

Since 1990s, good experience in the crystallization of large and phase-pure zeolite crystals was made by using microwaves for heating the autoclaves. Pioneering papers reported the successful synthesis of MFI (ZSM-5) and FAU(Y) [68, 69], LTA and FAU(X) [70], and AFI (CoAPO-5 [71] and AlPO4 -5 [72]). Surprisingly, large single crystals were obtained in relative short synthesis times, influenced by the quick energy input and the fast heating rate, which brings the zeolite batch quickly to the crystallization temperature and suppresses kinetically the formation of nuclei. Hence, microwave heating can shorten the nucleation period. Furthermore, due to the accelerated heating of the synthesis mixture, the silicate species, due to a kinetic effect, are not in their thermal equilibrium. The transport and the reactivity of alumina and silica species to precursor building units are influenced by microwave heating. Increasingly, microwave heating is understood and can be applied in zeolite membrane synthesis [73]. Progress was achieved by Julbe and coworkers in the microwave-assisted hydrothermal synthesis of silicalite-1 seeds for membrane preparation [74] and the silicalite-1 membrane crystallization itself [75]. Silicalite-1 membranes with a controllable thickness and high crystallinity can be derived within a few hours when seeded supports are microwave heated. By different synthesis temperatures and different methods for seeding the support, oriented silicalite-1 layers with a (101) channel orientation are obtained [76]. A remarkable progress in the utilization of microwave heating was achieved by Yang et al. in the past few years [77–79] who developed the “in situ aging–microwave synthesis” method (AM method) [80, 81]. In a “first-stage synthesis,” the polished, ultrasonicated, and calcined support is contacted with a clear solution synthesis mixture. The gel layer formed is aged in situ in an air-conditioned oven. Then the membranes are microwave treated for crystallization. This process is repeated in a so-called second-stage synthesis. This way, LTA membranes could be synthesized with high reproducibility. The “in-situ aging” step was found to be necessary for the subsequent successful microwave synthesis. The LTA membrane consists of spherical grains without well-developed crystal faces. This procedure considers that the support does not absorb microwaves and remains unheated, but the microwaves selectively couple with the gel layer because of its higher dielectric loss factor. The gel layer first formed on the support after in situ aging contains plenty of pre-nuclei. During the following microwave heating, these pre-nuclei rapidly and simultaneously develop into crystal nuclei. Then, crystal growth proceeds by propagation through the amorphous primary particles (size of ∼50 nm), and finally these particles transform into LTA crystals of about the same size. In this way, compact LTA zeolite membranes of spherical grains with undefined crystal facets are formed. A recent review reflects the state of the art of zeolite membranes using microwave synthesis [82]. Whereas most zeolite membrane preparations deal with the MFI (silicalite-1, ZSM-5) structure, the majority of the microwave heating membrane preparations are focused on LTA membranes. The general concept is to shorten the zeolite crystallization time so as to reduce the

6.3 Crystal Growth in Zeolite Layers

Short synthesis time Microwave heating

Sol adhesion

Conventional heating Long synthesis time Porous support

Gel layer Zeolite crystal

Zeolite membrane

Figure 6.1 Comparative synthsis models of zeolite membrane preparation by microwave and conventional heating [82]. Source: Copyright 2007, Elsevier.

membrane thickness and to improve the flux (Figure 6.1) [82]. Pre-seeding the support with nano-LTA was needed to overcome the nucleation-related bottleneck [83, 84]. Compared with conventional heating, the synthesis time was shortened by 8–12 times by using microwaves, and the permeance was increased by 4 times while keeping comparable permselectivity for H2 /n-C4 H10 [85, 86]. To further improve the permeance of LTA membranes, the macroporous alumina support was covered with a thin mesoporous top layer to prevent the penetration of the reagent into the support [77]. However, despite the remarkable progress in the LTA membrane synthesis, the permselectivities of the membranes are so far only slightly superior over the Knudsen separation factor. 6.3.4

Use of Intergrowth Supporting Substances

The International Zeolite Association (IZA) database contains 239 different zeolite structures [2]. It was found experimentally that only the high-silica-type zeolites show a real shape-selective separation behavior, especially silicalite-1 (as the Al-free MFI structure) and the DDR type. Most progress in the development of molecular sieve membranes was achieved, therefore, for silicalite-1 membranes since their preparation is relatively easy, and these highly siliceous zeolite membranes provide chemical stability and allow oxidative regeneration. On the contrary, when the high Al-containing zeolite membranes such as LTA and FAU are tested in shape-selective gas or steam permeation, usually Knudsen separation pattern is found, which indicates a high contribution of defect mesoand macropores to the mass transport. For ZSM-5 membrane series with

201

6 Zeolite Membranes

systematically increasing Al content, it was found that the intercrystalline defect transport is enhanced (both mixture separation factors and permselectivities decrease with increasing Al content) and high residual nitrogen permeances are found in permporosimetry. It seems to be a general problem, therefore, to crystallize thin defect-free Al-containing zeolite membrane layers as shown by recent papers [87, 88]. Searching for the reasons for this behavior, an increase of the negative surface charge (zeta potential) of zeolite crystals with enhanced Al content was found, which is independent on the structure type. Since the zeolite precursors in the synthesis solution are negatively charged like the growing zeolite layer, it is assumed that a hindered diffusive transport and attachment of the precursors into narrow slits between growing crystals are hindered. In the case of narrow distances between the crystals, the negative surface charges overlap and block the diffusive transport of the negatively charged silicate species. This mechanism seems to cause the poor intergrowth of the Al-containing crystals to a continuous tight membrane layer. By use of ISS, the crystal surface can be recharged and the crystal intergrowth is improved. The strong negative surface charge can be indeed compensated by adsorption of an ISS (Figure 6.2) [89]. Suitable ISS are small positively charged molecules and stable under the alkaline conditions during the membrane synthesis (e.g. at 180 ∘ C in the case of MFI membrane crystallization) and can be decomposed by calcination. Several ISS have been evaluated in MFI membrane preparation (Table 6.1) [89].

10

In 0.01 m KCI, 25 °C

Membrane synthesis range

0 Zeta potential (mV)

202

–10

ISS

–20 –30 –40 –50 3

4

5

6

7

8

9

10

11

12

pH Si/Al 1000 Si/Al 286 Si/Al 96 Si/Al 57

With HMEDA-l2

Figure 6.2 Zeta potentials of suspended MFI crystals with different Si/Al ratios at room temperature: after addition of hexamethylethylenediammonium diiodide (HMEDA-I2 ) (0.01 M in the electrolyte) as an ISS, the zeta potentials became less negative [89]. Source: Copyright 2006, Elsevier.

Table 6.1 Chemical structures and abbreviations of possible ISS types [89]. Cationic molecular structure

H3C

Name

Acronym

N,N,N,N ′ ,N ′ ,N ′ -Hexamethylethylenediammonium diiodide

HMEDA-I2

N,N,N,N ′ ,N ′ ,N ′ -Hexamethylhexylenediammonium diiodide

HMHDA-I2

N,N,N,N ′ ,N ′ ,N ′ -Hexapropylhexylenediammonium diiodide

HPHDA-I2

N,N,N ′ ,N ′ -Tatramethyldiethylenediammonium diiodide

TMDEDA-I2

N,N ′ -Dimethyltriethylenediammonium diiodide

DMTEDA-I2

CH3 +

H3C

CH2

N

+

CH2

N

CH3

H3C

CH3 CH3

H3C +

H3C

+

CH2 CH3

N

N

CH3

3

H3C

CH3 C3H7

C3H7 +

C3H7

+

C3H7

N

CH2 CH2

N

3

C3H7

C3H7

H2C

H3C

CH2 N

N H3C

H2C +

H3C

N

CH3 +

+

CH2

CH2

CH2

CH2

CH2

CH2

CH2

Source: Elsevier.

CH3 +

N

CH3

204

6 Zeolite Membranes

Table 6.2 Increase of the permselectivities (PS) derived for different gas mixtures from the corresponding single gas permeances at 105 ∘ C for two MFI membranes when HMEDA-I2 is used as ISS [89]. Permselectivity (PS) H2 /n-butane Si/Al

Without ISS

57

2.1

96

2.9

H2 /i-butane

With ISS

2.3 562

H2 /SF6

Without ISS

With ISS

Without ISS

49

98.5

6.4

12.1

68.2

5.0

128.6

2.6

With ISS

Source: Elsevier.

After evaluating different ISS (Table 6.1) and determining their optimum concentration range, the effect of using ISS on the membrane quality was studied for HMEDA-I2 in the preparation of MFI membranes with different Si/Al ratios (Table 6.2). The concept of enhanced crystal intergrowth by using an ISS is an effective tool for decreasing the intercrystalline defect transport, thus increasing the selectivity of Al-containing MFI membranes. The effect of the ISS molecules to improve crystal intergrowth decreases in the order of TMDEDA2+ > HMEDA2+ > DMTEDA2+ > HMHDA2+ > HPHDA2+ . By the use of 0.1 M ISS solution, the zeta potential is changed to values close to the isoelectric point (IEP), but at this high ISS concentration, ISS molecules became incorporated during the membrane synthesis and created mesopores upon thermal decomposition of the ISS during calcination. The optimal ISS concentration was found to be 0.01 M for HMEDA2+ . So the improvement of the permeation properties when using an ISS varied and depended on the Al content: nearly no ISS effect and already good permeation properties without ISS for Si/Al > 200; strong ISS effect and improved selectivities for 200 > Si/Al > 96; and nearly no ISS effect for 96 > Si/Al > 57. This ISS concept was first developed for Al-containing MFI membranes (ZSM-5) and later successfully transformed to the synthesis of LTA and FAU membranes. LTA and FAU membranes can separate water/organic mixtures in an excellent way, but they fail in shape-selective gas separations. Therefore, many attempts were made to improve the separation properties of LTA and FAU membranes for gases. Zeta potential measurements on the Al-rich crystals of zeolites LTA and FAU also showed a strong negative surface charge like it was found for Al-rich MFI crystals (ZSM-5). By adsorption of an ISS, this negative zeta potential could be shifted close to the IEP, which improved the intergrowth of the seed crystals on the support to a continuous membrane layer. This improvement of the LTA and FAU membrane quality could be concluded from permporosimetry measurements (Figure 6.3). Nevertheless, the LTA and FAU membranes prepared with ISS were still far from being defect-free, and their permselectivities were in the range of the Knudsen factor. Improvement of the LTA and FAU membrane quality by using an ISS in the membrane synthesis was shown in Figure 6.3 by permporosimetry studies [90].

6.3 Crystal Growth in Zeolite Layers

120

FAU-3 without ISS

FAU-3 with ISS

LTA-3 with ISS

LTA-3 without ISS M 1000-1 without ISS

Relative N2 permeance (%)

100

80 ISS 60

40

ISS

20

0 0.0

0.2

0.6 0.4 n-Hexane p/ps

0.8

1.0

Figure 6.3 Improvement of the LTA and FAU membrane qualities by using an ISS as measured by permporosimetry [90]. Source: Copyright 2007, Elsevier. The arrows indicate the shift of the residual N2 flux as a measure for reduced defect formation when an ISS is used. FAU-3 and LTA-3 denote three-layer FAU- and LTA-type membranes obtained by repeating three times the membrane synthesis. M 1000-1 denotes a one-layer MFI-type membrane with Si/Al ratio of approximately 1000.

6.3.5

Growth of Oriented Zeolite Layers on Supports

By seeding the support, MFI membrane layers of different crystallographic orientation can in principle be obtained, but most often the MFI-type zeolite membranes show a crystallographic orientation of the c-axis of the zeolite layer perpendicular to the plane of the support surface [58]. The c-orientation can be explained by the competitive growth model. Usually, the nanocrystallites used as seeds do not show developed crystal faces, and, therefore, these crystallites are randomly oriented. If crystal growth is anisotropic, those crystallites with their fastest growth direction pointing away from the seeded surface will grow faster than crystallites in other orientations. Finally, the crystals with the fastest growth direction perpendicular to the plane of the membrane will dominate. For MFI crystals, usually the c-axis is the longest dimension, and, consequently, the c-axis is the fastest growth direction. Under certain growth conditions, other crystallographic orientations were observed like a-orientation [91, 92], b-orientation [93, 94], or intermediate orientations [95]. From studies on the diffusion anisotropy of the MFI structure, it can be expected that permeation through c-oriented MFI membranes perpendicular to the support is less favorable [51]. A b-oriented MFI layer is expected to exhibit

205

6 Zeolite Membranes

higher fluxes. Recently, Tsapatsis et al. have prepared a b-oriented MFI silicalite-1 membrane. They used relatively large seeds (0.5 × 0.2 × 0.1 mm3 ) with developed crystal faces and attached the seeds as a b-oriented mono-seed layer to the support surface. By using di- and trimers of tetrapropylammonium hydroxide (TPAOH), the growth of the b-oriented seeds in b-direction could be enhanced. The resulting polycrystalline silicalite-1 films are approximately 1 mm thick and consist of large b-oriented single crystals with straight channels in the direction of the thickness of the membrane. This very careful membrane preparation resulted in a superior separation performance, which was demonstrated in the separation of xylene isomers (Figure 6.4). The development of high-flux and high-selectivity MFI membranes for xylene separation by the Tsapatsis group [58] showed the importance of channel orientation and the significant influence of a seeded growth of an oriented particle monolayer whereas a c-orientation had no separation at all. It is interesting to note that in the latter case the separation factor increases with increasing temperature. This experimental finding is characteristic for the interplay of adsorption and diffusion effects. At low temperatures, the zeolite pores were filled to a certain degree, and single file-like behavior was observed. That is to say, the more mobile p-xylene isomer b-axis

c-axis

1 μm

10 000

10 000

p-Xylene

1000

1000

p-Xylene SP

100

100

o-Xylene 10

10 o-Xylene 1 0.1 100

1

SP

120

140 160 Temperature (°C)

180

200

100

120

140 160 Temperature (°C)

180

Separation factor (SP)

5 μm

Permeance 10–10 mol m–2 s–1 Pa–1

206

0.1 200

Figure 6.4 Supported silicalite-1 membrane in the separation of p/o-xylene mixtures: influence of the preparation mode and channel orientation on the flux and the selectivity [58]. Source: Copyright 2003 AAAS.

6.3 Crystal Growth in Zeolite Layers

cannot move faster through the pore network than the less mobile o-xylene. This situation changed dramatically at lower pore filling, which was found at higher temperatures and/or lower partial pressures. Now, the mobile p-xylene can move more or less independently from the presence of o-xylene. The permeation experiments were carried out at very low loadings corresponding to a low total pressure of xylene due to a high content of inert gas in the feed stream and the high temperature. The concept of Tsapatsis [96] was further developed, and highly b-oriented and intergrown MFI films could be produced by carrying out secondary growth of b-oriented seed layers under hydrothermal conditions using trimeric tetrapropylammonium iodide as SDA. To deposit the seeds in b-orientation, the stainless-steel support had to be smoothed by using an intermediate layer of mesoporous silica. The MFI seed monolayer was covalently attached to the intermediate silica layer by using 3-chloropropyltrimethoxysilane. X-ray diffraction (XRD) measurements showed the strong b-orientation of the seeds on the silica smoothed stainless-steel support. This b-orientation was preserved during secondary growth using trimeric tetrapropylammonium iodide as the SDA. 6.3.6

Bilayer Membranes

Different aims are followed when synthesizing multilayer zeolite membranes: (i) improved separation selectivity by repeated crystallization of one and the same zeolite type, (ii) novel properties by combination of layers of different zeolite types, and (iii) new fields of application by combination of zeolite layers with other inorganic membrane layers. First, to improve the quality of MFI membranes, Vroon [38] proposed to repeat the crystallization step. Whereas a two-step growth was found to be beneficial for the quality of MFI membranes, further repetitions of the crystallization step did not improve the membrane quality since, as a result of the oxidative template removal, crack formation was observed for increased membrane thickness. Second, different zeolite structure types have been investigated in zeolite membrane applications. In particular, not only MFI-, LTA-, and FAU-type membranes but also BEA-, MOR-, FER-, OFF-, ANA-, CHA-, and ERI-type membranes have been studied [97]. On the contrary, chemical modifications have been made primarily for MFI-type membranes, for example, isomorphous substitution to give Al-, Fe-, B-, and Ge-ZSM-5 membranes [98, 99], variation in Si/Al ratios [100, 101], and ion exchange [102]. These numerous possibilities allow a fine-tuning of the membrane characteristics to tackle many different liquid and gas separation problems. Multilayered zeolite membranes with gradients of chemical composition or structure in the zeolite layers have the potential to expand the applications of zeolite membranes even further. Such membranes allow the intimate combination of different functions or characteristics in a single membrane, for example, catalytic activity/inertness, hydrophobic/hydrophilic character, and different pore sizes. Membranes that combine a catalytically active zeolite layer with an inert one are interesting for membrane reactors because they possess reactive and inert environments

207

208

6 Zeolite Membranes

adjacent to each other, which is a prerequisite for staged reaction/nonreactive separation and/or for the passivation of non-shape-selective catalytic sites at the external surface [103–105]. Lai and Corcoran [104] patented the fabrication of multilayered zeolite membranes and demonstrated both seeded and epitaxial growth of ZSM-5 on silicalite-1 layers supported on porous alumina and stainless-steel supports. They reported that the growth of ZSM-5 layers on calcined silicalite-1 layers led to partial erosion of the underlying silicalite-1 layer and that this erosion could be prevented when the silicalite-1 layer was not calcined prior to the synthesis of the second layer. Gora et al. [106] reported the seeded synthesis of silicalite-1 layers on top of zeolite LTA layers on porous Trumen supports by identifying conditions that allowed for the growth of the second layer without dissolution of the first one. The preparation of the opposite layer sequence was not successful. The crystallization of LTA and FAU layers on a silicalite-1 layer turned out to be more complicated because the high alkalinity of the LTA and FAU synthesis batches caused the dissolution of the silicalite-1 layer already formed. Bilayered ZSM-5/silicalite-1 films were also prepared on nonporous quartz and silicon substrates by Li et al. [107]. Recently, bilayered silicalite-1/ZSM-5 membranes were synthesized, and their permeation and separation properties were examined [108]. If the first layer on the support is silicalite-1 (shape selective, hydrophobic), a second ZSM-5 layer (Brønsted acid sites, hydrophilic) can be crystallized onto the first one. Further synthesis work seems to be necessary for successful preparation of bilayered membranes. Third, there are ambitious attempts to combine zeolite layers with other inorganic membrane layers. As an example, for shape-selective oxidations, a thin silicalite-1 layer was crystallized on an oxygen-transporting perovskite membrane [109]. Assuming that a mixture of the xylene isomers would be in contact with this bilayer membrane facing the Ti-modified silicalite-1 layer, mainly the p-xylene isomer would enter the silicalite-1 layer and could be oxidized to terephthalic acid with the oxygen released from the perovskite membrane. 6.3.7

Functional Zeolite Films

In addition to its use as separation membrane and catalytic membrane reactor, zeolite layers can act as functional film in chemical sensors, as electrode, as optoelectronic device or low dielectric constant material, as protection or insulation layer, as corrosion-resistant coatings [110], as hydrophilic antimicrobial coatings [111], or as sulfonated zeolite BEA for proton-exchange membranes [112]. As an alternative method for crystallizing a zeolite layer in the liquid phase, by the so-called dry gel conversion, a dry (alumino)silicate gel can be converted into a zeolite layer in the presence of vapors [40]. The vapor phase can be only steam or a mixture of steam and a SDA such as TPAOH. In contrast to the conventional steam-assisted crystallization in which the substrate is coated with all the nutrients and then steam-treated by a novel steam-assisted method, the oxidized surface layer of a silicon wafer can be transformed into a silicalite-1 zeolite film [113]. A silicon wafer is coated with 1 M TPAOH, and then zeolite films were crystallized under steam by adding some water to the autoclave. In this method, the silicon wafer serves both as support of the grown zeolite film and as the Si source

6.3 Crystal Growth in Zeolite Layers

(a) 2 h

(b) 4 h

10 μm (d) 8 h

(c) 6 h

10 μm (e) 48 h

10 μm (f) 48 h cross section

Zeolite

Silicon wafer

10 μm

10 μm

10 μm

Figure 6.5 FESEM images of a- and b-oriented silicalite-1 films obtained by steaming (at different time intervals) a SDA-oxidized silicon wafer [113]. Source: Copyright 2007, ACS.

for the formation of the silicalite-1 film. Pure silcalite-1 films with preferential a- and b-out-of-plane orientation were obtained in a temperature window from 100 to 200 ∘ C (Figure 6.5). This method seems to be very useful for zeolite film applications as chemical sensor or for optoelectronics. Porous pure silica materials are attractive as the insulator material in on-chip interconnects due to their high porosity, hydrophobicity, acceptable heat conductivity, and low dielectric constant. A remarkable improvement could be achieved by the UV treatment of spin-on silicalite-1 films to induce hydrophobization [114]. In this posttreatment method during the removal of the organic template in combination with a thermal treatment, UV radiation decreases drastically the quantity of silanols. In parallel, methylation of the silica surface is obtained by decomposition and reaction of the TPA ions as SDA. By this method, the formation of cracks during the removal of the organic template is minimized. Pure silica zeolites have a remarkably higher mechanical strength and hydrophobicity than amorphous porous silica due to their crystalline structure, making them a likely dielectric material for enabling smaller feature sizes in future generation of microprocessors [115]. Zeolite nanocrystals can be coated on substrates to form a transparent film with approximately 20 nm surface roughness. The resulting coating showed a broadband anti-reflection effect with less than 1% average reflection over the visible range. With proper control of the film thickness, one can shift the reflection minimum to achieve a neutral color [116]. This broadband neutral color anti-reflection coating was achieved in a single-step sol–gel process and can find application in the display industry.

209

210

6 Zeolite Membranes

Last but not least, high-silica zeolite coating on metals and metal alloys can be a promising technology for corrosion protection of metals [117]. The as-synthesized SDA-containing MFI films are nonporous and allow the coating of complex shapes and in confined spaces by in situ crystallization [118]. 6.3.8

Mixed Matrix Membranes

Soon after the first preparation of synthetic zeolites, the idea was born to incorporate zeolite crystals as the modifier into a polymer matrix, thus using the easy processing of polymers [119–121]. Mixed matrix membranes are interesting systems to enhance the properties of the host matrix taking advantage of the peculiar properties of specific inorganic fillers [122, 123]. Today a renaissance of this concept can be observed in a series of reports [124–129]. Current polymeric membranes seem to have reached a limit in the trade-off between permeability and selectivity. Therefore, research efforts have been focused on mixed matrix membranes that contain porous (nano)particles [130] in a polymeric matrix and that can be processed by the usual spinning technology [131–133]. By surface treatment of the inorganic modifiers [134] and by polymer chain rigidification [135], mixed matrix membranes with improved separation patterns could be obtained. Various especially small-pore zeolites have been used for their application in mixed matrix membranes. Although early studies were focused on LTA zeolite, this hydrophilic zeolite turned out to be less attractive for gas separation from humid feeds because of the pore blocking by water [127]. Recent studies have focused, therefore, on more hydrophobic zeolites with a high molar SiO2 /Al2 O3 ratio. As an example, zeolite CHA with a SiO2 /Al2 O3 ratio 30 with a particle size of 1 mm can be processed to mixed matrix membrane layers of approximately 1 mm thickness [136]. However, only low selectivities were found for zeolite/polymer mixed matrix membranes, mainly because of the poor wetting [137, 138]. This interfacial problem does not occur for mesoporous MCM-41 in PSF [139] and mesoporous ZSM-5 in Matrimid [140], which suggests that the polymer chains can penetrate into the mesopores. As a result of this penetration, both the glass temperature T g and Young’s modulus of the mixed matrix are higher than those of the pure polymers. In comparison with the pure polymers, the mixed matrix membranes show substantially increased permeability and permselectivity.

6.4 Microstructures of Zeolite Films Whereas the hydrophilic membranes of type LTA and FAU do an excellent job in water separation, they fail in shape-selective separations. From in situ XRD studies, there were indications from literature that during the drying of zeolite powders of the types LTA [141], FAU [142], and MOR [143, 144], there occur extreme irregular expansions/shrinkages of the unit cell. Recently, the change of the unit cell of LTA and FAU crystals was studied by heating wet zeolite samples and measuring the change of the unit cell by in situ XRD [145]. For LTA an

6.5 Membrane Characterizations

extreme shrinking of the unit cell and for FAU an extreme expansion of the unit cell as a result of the dewatering were observed for a temperature range between room temperature and 100 ∘ C. So, hydrated LTA shrinks from 2.460 nm at 50 ∘ C to 2.445 nm at 100 ∘ C, whereas hydrated FAU expands due to the dewatering when heated in air from 2.478 nm at 50 ∘ C to 2.490 nm at 100 ∘ C. In contrast to the drastic changes due to the dewatering, only slight changes of the unit cell were found when the dried LTA and FAU crystals were heated and cooled, respectively. This extreme shrinkage/expansion behavior was only found for the hydrophilic zeolites like LTA and FAU. More hydrophobic ones such as MFI or MOR did not show this behavior. The amount of expansion and shrinkage, respectively, can be controlled by the heating rate.

6.5 Membrane Characterizations Zeolite membranes have been characterized using some of the methods described below. Zeolite crystal size and shape and membrane thickness on top of the support are measured with scanning electron microscopy (SEM). Membranes ranging from 0.5 μm to approximately 500 μm thickness have been reported. In addition, SEM gives a qualitative idea of layer uniformity and continuity. Zeolite framework structure and crystallinity are typically determined using XRD [146]. Zeolite composition is measured with inductively coupled plasma (ICP) or electron probe microanalysis (EPMA), which is also known as energy-dispersive X-ray spectroscopy (EDX). Several studies have also used EPMA to determine the distance where the zeolite layer penetrates into the pores of the support [147–149]. The surfaces of flat membranes are sometimes analyzed without damaging the membranes, but tubular and monolithic membranes must be broken to carry out these measurements. Zeolite powders that form during membrane synthesis can be analyzed with XRD and ICP, and though these crystals have been shown to be similar to those in the membranes, this is not a direct measure of the membrane. Gas permeation is also a common method of membrane characterization. Ideal selectivity is defined as the ratio of single gas permeances and is often used as an indication of membrane quality. Permeation depends on both adsorption and diffusion, but molecules close to or larger than the zeolite pore size have difficulty entering zeolite channels. Molecules larger than the zeolite XRD pore diameters sometimes fit into the zeolite pores, and this is possibly because some zeolite frameworks are quite flexible [150, 151]. Moreover, XRD pore diameters may not be accurate representations of zeolite pore size [152], and kinetic diameters may not accurately represent molecular sizes because the diameters assume molecules are hard spheres. Appreciable permeation of molecules significantly larger than zeolite pores, however, indicates flow through non-zeolite pores. Gas mixture permeation is an additional characterization method. Some mixtures permeate differently from single components because of competitive adsorption. Also, some molecules are inhibited, and some are sped up by the presence of other molecules [153–155]. These effects are generally larger

211

212

6 Zeolite Membranes

when adsorption coverages are high. Several studies have characterized 10 MR zeolite membranes by measuring n-C4 H10 /i-C4 H10 separation selectivities [33, 156–159]. Because these molecules fit into the zeolite pores, mixtures that are separated by molecular sieving are probably better for characterizations. Lai et al. observed that 1 μm thick silicalite-1 membranes with significant grain boundary defects separated n-C4 H10 /i-C4 H10 and N2 /SF6 , but did not separate vapor-phase p-/o-xylene mixtures [58]. Only membranes with few grain boundary defects separated p-xylene (0.58 nm in kinetic diameter) and o-xylene (0.68 nm in kinetic diameter) mixtures. In another study, Lai and Tsapatsis showed that these high-quality membranes exhibited separation factors of 15 and 25 for benzene (0.59 nm)/cyclohexane (0.60 nm) and benzene/2,2-dimethylbutane (DMB) (0.62 nm) vapor mixtures, respectively, at 473 K, even though their n-C4 H10 /i-C4 H10 and N2 (0.36 nm)/SF6 (0.55 nm) separation factors were only 5 and 2.5, respectively, at this temperature [160]. The feeds for these separations were approximately 50/50 mixtures except for benzene/cyclohexane, which was approximately 25/75. For the benzene/cyclohexane and benzene/2,2-DMB binary separations, the benzene, cyclohexane, and 2,2-DMB feed partial pressures were 6.25, 21.2, and 6.45 kPa, respectively. For the n-C4 H10 /i-C4 H10 and N2 /SF6 binary separations, the feed partial pressure of each component was 50 kPa. Mixtures of n-hexane (0.43 nm) and 2,2-DMB have also been used to characterize MFI membranes [56]. Furthermore, separations at high temperatures (∼470 K and above) are better for characterizations because adsorption is low, and high temperatures are a harsher test of the membranes [161]. Gas permeation characterizations, however, are not always a good indication of membrane performance for pervaporation. For example, A-type membranes have effectively dehydrated organic compounds even though their gas selectivities are near the Knudsen values [162]. For characterization of smaller-pore 8 MR zeolite membranes, light gas separations such as H2 /CH4 [7], CO2 /CH4 [163], and H2 /N2 [15] have been used in several studies. Similarly, larger-pore zeolite membranes require larger test molecules. Permporosimetry is one of the important methods used to characterize zeolite membrane. The basic concept of permporosimetry is that an inert non-condensable and less adsorbing gas (He, N2 ) and a vapor that prefers to fill the regular micropores (e.g. n-hexane, water) are sent as co-feed through the membrane. The highly adsorbing vapor such as n-hexane for hydrophobic membranes such as silicalite-1 or water for hydrophilic membranes such as FAU or LTA is mixed to the inert gas with increasing p/ps ratios of the strongly adsorbing component with p and ps denoting the real and the maximum, that is, saturation vapor pressure at the given temperature, respectively. The vapor fills the regular micropore system of the membrane and blocks them for the passage of the less adsorbing He or N2 . A remaining He or N2 flux indicates the presence of defect pores in the mesopore region. Usually, for real zeolite membranes, a superimposition of two fluxes is observed: the intracrystalline shape-selective flux through the regular zeolite pores and an additional nonselective flux through defect mesopores that are larger than the zeolite pore sizes. The flux through these defect pores in the

6.5 Membrane Characterizations

mesopore range can spoil completely any shape selectivity and result in very low separation factors. For a quantitative evaluation of the flux through the defect pores, the permporosimetry can be used. Permporosimetry is similar to the terms permporometry [164–169], dynamic capillary condensation porometry [170, 171], and dynamic flow-weighted pore size distribution technique [172]. To the authors’ knowledge, permporosimetry was originally developed for the characterization of pores where the Kelvin equation is valid. Later, permporosimetry was extended to microporous membranes [173, 174], first applied to zeolite membranes by Deckman [175] and further developed by different groups [176, 177]. According to the adsorption isotherm, at a certain p/ps ratio, the zeolite pores are filled, and the remaining flux of the inert gas can be assigned exclusively to non-regular zeolite mesopores. One has to be aware of the effect that when the p/ps ratio of the strongly adsorbing component is continuously increased, also narrow mesopores of increasing diameter become filled according to the Kelvin equation and therefore blocked for the flux of the inert gas. A typical permporosimetry experiment at constant temperature and pressure difference Dp across the membrane consists of the following steps: (i) Measure the flux of an inert non-condensable gas such as He or N2 through an outgassed porous membrane and set this to 100%. (ii) Select a suitable vapor species that is well adsorbed by the zeolite, for example, n-hexane for the hydrophobic silicalite-1 membrane and water for the hydrophilic LTA membrane. (iii) Send a part of the non-condensable gas through a saturator filled with the well-adsorbing liquid, mix this gas stream with the pure non-adsorbing gas to fine-tune the p/ps, and send the blended gas through the membrane. (iv) Measure the relative decrease of the flux of the inert gas for increasing p/ps of the well-adsorbing species. (v) Calculate the relative decrease of the inert gas flux. As mentioned earlier, the well-adsorbing vapor is expected to fill completely the regular micropore structure of the zeolite membrane under study so that the remaining flux can be attributed completely to defect pores. As an example, permporosimetry on two SiO2 membranes will be compared: silicalite-1 and a SiO2 sol–gel membrane. It is found that n-hexane even at p/ps ≈ 0.05 completely fills the micropore volume of this rather perfect silicalite-1 membrane. However, there is a remaining nitrogen flux if water is used since water is not well adsorbed by the hydrophobic silicalite-1 membrane. On the contrary, in the case of the hydrophilic SiO2 sol–gel membrane, water is better adsorbed than n-hexane and causes a steeper decrease of the nitrogen flux as compared with n-hexane. However, it was found that n-hexane adsorption causes the size of the defects to decrease in MFI membranes. Therefore, the use of n-hexane in permporosimetry is questionable, and more inert molecules such as benzene should be used. The capillary condensation of dimethylbutane (DMB) can be used to estimate relative sizes of non-zeolite pores [178, 179]. Figure 6.6 showed a two-dimensional representation of flow through a section of the polycrystalline zeolite layer. The n-hexane (n-C6 ) flux depends on which C6 isomer was added to the feed first because larger pores exist in series with the zeolite crystals. Thus, when n-hexane diffuses through the zeolite crystals, it desorbs from one zeolite crystal and diffuses into intercrystalline pores and then adsorbs and

213

214

6 Zeolite Membranes

DMB DMB

DMB

DMB

DMB

n-C6 < 0.6 nm

2–4 nm Zeolite crystal

n-C6

Zeolite crystal

Zeolite crystal

(a)

Zeolite crystal

(b) n-C6

DMB < 0.6 nm Zeolite crystal

n-C6

Zeolite crystal

(c)

Figure 6.6 Schematic representation of a MFI zeolite membrane with defects [178]: (a) DMB transport through the defects in an empty membrane. (b) n-Hexane adsorbed in and transported through zeolite crystals for exposure to n-hexane and then DMB. The defects shrink in the presence of n-hexane so that DMB permeation is blocked. (c) Both n-hexane and DMB permeation are blocked when the membrane is exposed to DMB first. The pores shrink and trap DMB so that almost all transport is blocked.

diffuses through another crystal. When the zeolite crystals expanded due to n-hexane adsorption, they apparently trapped DMB in the larger pores as they closed. The DMB flux was essentially the same, independent of the order DMB was added, because in the presence of n-hexane, DMB only diffused through pores that are larger than 0.6 nm and were not closed by the expanding zeolite crystals. The hexane flux was much lower when DMB was added first, because n-hexane essentially cannot transport through intercrystalline pores that have been filled with DMB. The DMB cannot be readily removed from the pores because it is too large to fit into the zeolite pores at these conditions. Thus, the only transport pathways remaining are pores that are too large to be sealed by hexane adsorption. Both hexane and DMB transported through these pores, which were not selective; when DMB was added first, the selectivity was one. After filling the regular micropores of a zeolite membrane, with increasing p/ps, also narrow mesopores can be filled. Assuming perfect wettability, it can be estimated by the Kelvin equation, at which p/ps a pore is “closed” by capillary condensation. In calculating the pore size due to capillary condensation, it is important to consider the thickness of the already adsorbed molecules (Figure 6.7) [180]. It can be advantageous in permporosimetry to use molecules of different size. Small probe molecules such as water or n-hexane can completely fill at low p/ps the regular micropores of a zeolite membrane, and the remaining

6.6 Conclusions and Outlook

Halsey equation dmicro

d1

Kelvin equation

d2

d3

d4

d5

d6 d-2t

t

0.005 0.01 0.05 0.1

0.2

0.3

0.4

Micropore

0.5

0.6

0.7

0.8

Mesopore

0.95 1.0 50

monolayer covering

Capillary condensation t = tM · 3 –

5 In p/ps

p/ps

Macropore

2.0 Total pore filling by overlap of surface potentials

0.9

rpore = –2Vm γ·cos θ RT In p/ps

tM thickness of monolayer

Vm γ θ R T

Pore coverage (nm)

at p/ps near 1 only monolayer

molar volume surface tension contact angle gas constant temperature

Figure 6.7 Schematic presentation of the pore filling process (a redrawn picture according to the concept) [180]. Source: DeGruyter.

N2 stream can be ascribed to the defect pores. In contrast, large probe molecules such as perfluorotributylamine (1.03 nm), triethylamine (0.74 nm), trimethylbenzene (0.62 nm), triethylbenzene (0.84 nm), and tri-i-propylbenzene (0.85 nm) can fill at high p/ps the mesopores of the membrane by capillary condensation (because of their bulkiness, these molecules cannot enter micropores), and the remaining N2 stream is due to the transport through the regular pore system of the zeolite membrane. Permporosimetry measurements give a quick insight on the existence of defect pores and their pore size distribution and can forecast the separation behavior of a membrane. As an example, the correlation of permporosimetry data and the mixture separation factor was done for ZSM-5 membranes with different Al contents. It is found by permporosimetry that an increasing Al incorporation into the MFI structure gives membranes with high concentrations of defects. Consequently, the membrane with the highest Al content shows the highest residual nitrogen flux and has the lowest separation factor. On the contrary, the silicalite-1 membrane (Si/Al = ∞) has no measurable residual nitrogen flux at p/ps > 0.05 and shows the highest separation factor.

6.6 Conclusions and Outlook The industrial separation processes are highly competitive both within the membrane field itself and with other gas separation technologies. The current rise in energy costs makes membrane separations more attractive. Significant progress has been made during the last decade in the understanding of separation mechanisms of zeolite membranes and the synthesis of thin, high-flux, defect-free zeolite membranes applying new techniques of preparation [181], modification, [182] and new materials [183]. The zeolite’s unique properties combined with the membranes’ continuous separation properties make zeolite membranes

215

216

6 Zeolite Membranes

attractive for a wide variety of industrial applications. The technologies are now available to prepare zeolite membranes of sufficient quality and reliability. In the near future, while having hard competition with other separation techniques, the exploitation of new and more advanced zeolite membranes for industrial separations will continue. Zeolite membranes demonstrated high efficiency in separations that are still not attainable by conventional methods. The main field of application for zeolite membranes in the future is expected to be the shape-selective separation of C4–C8 hydrocarbon isomers because no other separation technique for such separation is available. On a medium time scale, therefore, new zeolite membranes will be developed, which can do the unique job that no other membrane can do: molecular sieving of molecules of almost identical or similar mass but different size and shape. However, to make zeolite membranes a viable option, further improvements, especially cost reduction, membrane reliability, reproducibility, and operational stability, should be endeavored to facilitate the wide introduction of zeolite-based membranes into industrial practice.

References 1 Suzuki, H. (1987). Composite membrane having a surface layer of an ultra-

2 3 4 5 6 7 8 9 10 11 12 13 14 15

thin film of cage-shaped zeolite and processes for production thereof. US Patent 4,699,892 International Zeolite Association (2007). Atlas of Zeolite Framework Types, 6ee. http://www.iza-online.org. Chiou, Y.H., Tsai, T.G., Sung, S.L. et al. (1996). J. Chem. Soc. Faraday Trans. 92: 1061–1066. Nishiyama, N., Ueyama, K., and Matsukata, M. (1996). Microporous Mater. 7: 299–308. Washmon-Kriel, L. and Balkus, K.J. (2001). Microporous Mesoporous Mater. 38: 107–121. Tuan, V.A., Li, S.G., Falconer, J.L., and Noble, R.D. (2002). Chem. Mater. 14: 489–492. Kalipcilar, H., Bowen, T.C., Noble, R.D., and Falconer, J.L. (2002). Chem. Mater. 14: 3458–3464. Poshusta, J.C., Tuan, V.A., Falconer, J.L., and Noble, R.D. (1998). Ind. Eng. Chem. Res. 37: 3924–3929. Munoz, T. and Balkus, K.J. (1999). J. Am. Chem. Soc. 121: 139–146. Kita, H., Asamura, H., Tanaka, K. et al. (1997). Abstr. Pap. Am. Chem. Soc. 214: 269. Kusakabe, K., Kuroda, T., Uchino, K. et al. (1999). AIChE J. 45: 1220–1226. Li, S., Tuan, V.A., Falconer, J.L., and Noble, R.D. (2002). Microporous Mesoporous Mater. 53: 59–70. Nishiyama, N., Ueyama, K., and Matsukata, M. (1997). Stud. Surf. Sci. Catal. 105: 2195–2202. Dong, J.H. and Lin, Y.S. (1998). Ind. Eng. Chem. Res. 37: 2404–2409. Aoki, K., Kusakabe, K., and Morooka, S. (1998). J. Membr. Sci. 141: 197–205.

References

16 Kita, H., Horii, K., Ohtoshi, Y. et al. (1995). J. Mater. Sci. Lett. 14: 206–208. 17 Hedlund, J., Schoeman, B., and Sterte, J. (1997). Chem. Commun. 13:

1193–1194. 18 Tuan, V.A., Li, S.G., Noble, R.D., and Falconer, J.L. (2001). Chem. Commun.

0: 583–584. 19 Bakker, W.J.W., van den Broeke, L.J.P., Kapteijn, F., and Moulijn, J.A. (1997).

AIChE J. 43: 2203–2214. 20 Hedlund, J., Sterte, J., Anthonis, M. et al. (2002). Microporous Mesoporous

Mater. 52: 179–189. 21 Bowen, T.C., Kalipcilar, H., Falconer, J.L., and Noble, R.D. (2003). J. Membr.

Sci. 215: 235–247. 22 Sano, T., Yanagishita, H., Kiyozumi, Y. et al. (1994). J. Membr. Sci. 95:

221–228. 23 Lovallo, M.C., Gouzinis, A., and Tsapatsis, M. (1998). AIChE J. 44:

1903–1913. 24 Burggraaf, A.J., Vroon, Z.A.E.P., Keizer, K., and Verweij, H. (1998).

J. Membr. Sci. 144: 77–86. 25 Noack, M., Kolsch, P., Caro, J. et al. (2000). Microporous Mesoporous Mater.

35: 253–265. 26 Wegner, K., Dong, J.H., and Lin, Y.S. (1999). J. Membr. Sci. 158: 17–27. 27 Lin, X., Kita, H., and Okamoto, K. (2000). Chem. Commun. 1889–1890. 28 Gardner, T.Q., Flores, A.I., Noble, R.D., and Falconer, J.L. (2002). AIChE J.

48: 1155–1167. 29 Navajas, A., Mallada, R., Téllez, C. et al. (2002). Desalination 148: 25–29. 30 Nishiyama, N., Korekazu, U., and Masahiko, M. (1995). J. Chem. Soc. Chem.

Commun. 1967–1968. 31 Yamazaki, S. and Tsutsumi, K. (1997). Adsorption 3: 165–171. 32 Tanaka, K., Yoshikawa, R., Ying, C. et al. (2001). Catal. Today 67: 121–125. 33 Matsufuji, T., Nishiyama, N., Matsukata, M., and Uyama, K. (2000). J.

Membr. Sci. 178: 25–34. 34 Alfaro, S., Arruebo, M., Coronas, J. et al. (2001). Microporous Mesoporous

Mater. 50: 195–200. 35 Ramsay, J.D.F. and Kallus, S. (2000). Recent Advances in Gas Separation

36 37 38 39

40 41 42

by Microporous Ceramic Membranes, Membr. Sci. and Techn. Ser. 6 (ed. N.K. Kanellopoulos), 373. Amsterdam: Elsevier. Coronas, J. and Santamaria, J. (1999). Sep. Purif. Meth. 28: 127–177. Pan, M. and Lin, Y.S. (2001). Microporous Mesoporous Mater. 43: 319–327. Vroon, Z.A.E.P., Keizer, K., Gilde, M.J. et al. (1996). J. Membr. Sci. 113: 293–300. Tsapatsis, M. and Gavalas, G.R. (1999). Synthesis of porous inorganic membranes. In: Membranes and Membrane Processes (ed. V.N. Burganos) MRS Bulletin, 24, 30. Materials Research Society and Cambridge University Press. Matsukata, M., Ogura, M., Osaki, T. et al. (1999). Top. Catal. 9: 77–92. Dong, J., Dou, T., Zhao, X., and Gao, L. (1992). J. Chem. Soc. Chem. Commun. 1056–1058. Matsufuji, T., Nishiyama, N., Ueyama, K., and Matsukata, M. (2000). Catal. Today 56: 265–273.

217

218

6 Zeolite Membranes

43 Buekenhoudt, A., Servaes, F., Wyns, K., et al. (2006). A new generation

44 45 46

47 48 49 50 51 52 53

54 55 56 57 58 59 60 61

62 63 64

of zeolite membranes. Proceedings of the 9th International Conference on Inorganic Membranes, Lillehammer, Norway (25–29 June 2006), p. 286. Kirschock, C., Buschmann, V., Kremer, S. et al. (2001). Angew. Chem. Int. Ed. 40: 2637–2640. Van Gestel, T., Vandecasteele, C., Buekenhoudt, A. et al. (2002). J. Membr. Sci. 207: 73–89. Jansen, J.C., Nugroho, W., and van Bekkum, H. (1993). Controlled growth of thin films of molecular sieves on various supports. Proceedings of the 9th International Conference on Zeolite Conference, Butterworth-Heinemann, Montreal (1992), p. 247. Koegler, J.H., Arafat, A., van Bekkum, H., and Jansen, J.C. (1997). Stud. Surf. Sci. Catal. 105: 2163–2170. Geus, E.R., van Bekkum, H., Bakker, W.J.W., and Moulijn, J.A. (1993). Microporous Mater. 1: 131–147. Jansen, J.C. and Rosmalen, G.M. (1993). J. Cryst. Growth 128: 1150–1156. Kapteijn, F., Bakker, W.J.W., van de Graaf, J. et al. (1995). Catal. Today 25: 213–218. Caro, J., Noack, M., Marlow, F. et al. (1993). J. Phys. Chem. 97: 13685–13690. Jansen, J.C., Kashchiev, D., and Erdem-Senetalar, A. (1994). Stud. Surf. Sci. Catal. 85: 215–250. Kuhn, J., Gross, J., Jansen, J.C., and Jansens, P.J. (2006). Mass transport through zeolite membranes: investigating framework polarity. Proceedings of the 9th International Conference on Inorganic Membranes, Lillehammer, Norway (25–29 June 2006), p. 256. Piera, E., Giroir-Fendler, A., Dalmon, J.A. et al. (1998). J. Membr. Sci. 142: 97–109. Uzio, D., Peureux, J., Giroir-Fendler, A. et al. (1994). Stud. Surf. Sci. Catal. 87: 411–418. Coronas, J., Falconer, J.L., and Noble, R.D. (1998). Ind. Eng. Chem. Res. 37: 166–176. Yan, Y., Davis, M.E., and Gavalas, G.R. (1997). J. Membr. Sci. 126: 53–65. Lai, Z., Bonilla, G., Diaz, I. et al. (2003). Science 300: 456–460. Hedlund, J., Jareman, F., Bons, A.J., and Anthonis, M. (2003). J. Membr. Sci. 222: 163–179. Li, G., Kikuchi, E., and Matsukata, M. (2003). Microporous Mesoporous Mater. 62: 211–220. Lai, W.F., Deckman, H.W., McHenry, J.A., and Verduijn, J.P. (1994). Supported zeolite membranes with controlled crystal width and preferred orientation grown on a growth enhancing layer. US Patent 5,871,650. Gouzinis, A. and Tsapatsis, M. (1998). Chem. Mater. 10: 2497–2504. Xomeritakis, G., Gouzinis, A., Nair, S. et al. (1999). Chem. Eng. Sci. 54: 3521–3531. Hedlund, J., Schoeman, B.J., and Sterte, J. (1997). Stud. Surf. Sci. Catal. 105: 2203–2210.

References

65 Mintova, S., Hedlund, J., Valtchev, V. et al. (1997). Chem. Commun. 2:

15–16. 66 Mintova, S., Hedlund, J., Valtchev, V. et al. (1998). J. Mater. Chem. 8:

2217–2221. 67 Hedlund, J., Noack, M., Kölsch, P. et al. (1999). J. Membr. Sci. 159: 263–273. 68 Chu, P., Dwyer, F.G., and Clarke, V.J. (1990). EP 358827. 69 Arafat, A., Jansen, J.C., Ebaid, A.R., and van Bekkum, H. (1993). Zeolites 13:

162–165. 70 Jansen, J.C., Arafat, A., Barakat, A.K., and Van Bekkum, H. (1992). Synthesis

of Microporous Materials, vol. I, 507. New York: Van Nostrand Reinhold. 71 Girnus, I., Hoffmann, K., Marlow, F. et al. (1994). Microporous Mater. 2:

537–541. 72 Girnus, I., Jancke, K., Vetter, R. et al. (1995). Zeolites 15: 33–39. 73 Weh, K., Noack, M., Sieber, I., and Caro, J. (2002). Microporous Mesoporous

Mater. 54: 27–36. 74 Motuzas, J., Julbe, A., Noble, R.D. et al. (2005). Microporous Mesoporous

Mater. 80: 73–83. 75 Motuzas, J., Julbe, A., Noble, R.D. et al. (2006). Microporous Mesoporous

Mater. 92: 259–269. 76 Motuzas, J., Mikutaviciute, R., Noble, R.D., et al. (2006).

77 78 79 80 81

82 83 84 85 86 87 88 89

Microwave-synthesized MFI membranes reproducibility and permeation behaviour at high temperature. Proceedings of the 9th International Conference on Inorganic Membranes, Lillehammer, Norway (25–29 June 2006), p. 532. Chen, X., Yang, W., Liu, J., and Lin, L. (2005). J. Membr. Sci. 255: 201–211. Chen, X., Yang, W., Liu, J., and Lin, L. (2004). J. Mater. Sci. 39: 671–673. Xu, X., Bao, Y., Song, C. et al. (2004). Microporous Mesoporous Mater. 75: 173–181. Li, Y.S., Liu, J., and Yang, W.S. (2006). J. Membr. Sci. 281: 1–6. Yang, W. and Li, Y. (2006). Formation Mechanism of Microwave Synthesized LTA Zeolite Membranes and the Thermal/Non-thermal Effects, Proceedings of the 9th International Conference on Inorganic Membranes, Lillehammer, Norway (25–29 June 2006), p. 273. Li, Y. and Yang, W. (2008). J. Membr. Sci. 316: 3–17. Slangen, P.M., Jansen, J.C., and van Bekkum, H. (1997). Microporous Mater. 9: 259–265. Cundy, C.S., Plaisted, R.J., and Zhao, J.P. (1998). Chem. Commun. 1465–1466. Xu, X.C., Yang, W.H., Liu, J., and Lin, L.W. (2001). Sep. Purif. Technol. 25: 241–249. Xu, X.C., Yang, W.S., Liu, J., and Lin, L.W. (2000). Adv. Mater. 12: 195–198. Noack, M., Kölsch, P., Seefeld, V. et al. (2005). Microporous Mesoporous Mater. 79: 329–337. Noack, M., Mabande, G.T.P., Caro, J. et al. (2005). Microporous Mesoporous Mater. 82: 147–157. Noack, M., Kölsch, P., Dittmar, A. et al. (2006). Microporous Mesoporous Mater. 97: 88–96.

219

220

6 Zeolite Membranes

90 Noack, M., Kölsch, P., Dittmar, A. et al. (2007). Microporous Mesoporous

Mater. 102: 1–20. 91 Hedlund, J., Mintova, S., and Sterte, J. (1999). Microporous Mesoporous

Mater. 28: 185–194. 92 Wang, Z. and Yan, Y. (2001). Chem. Mater. 13: 1101–1107. 93 Koegler, J.H., van Bekkum, H., and Jansen, J.C. (1997). Zeolites 19: 262–269. 94 Lai, S.M., Au, L.T.Y., and Yeung, K.L. (2002). Microporous Mesoporous

Mater. 54: 257–268. 95 Bons, A.J. and Bons, P.D. (2003). Microporous Mesoporous Mater. 62: 9–16. 96 Avhale, A., Mabande, G.T.P., Schwieger, W. et al. (2006). Preparation of

97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114

b-oriented MFI films on the stainless-steel supports. Proceedings of the 9th International Conference on Inorganic Membranes, Lillehammer, Norway (25–29 June 2006), p. 279. Noack, M. and Caro, J. (2002). Handbook of Porous Solids, 2433–2507. Weinheim, Germany: Wiley-VCH. Tuan, V.A., Li, S., Falconer, J.L., and Noble, R.D. (2002). J. Membr. Sci. 196: 111–123. Tuan, V.A., Falconer, J.L., and Noble, R.D. (2000). Microporous Mesoporous Mater. 41: 269–280. Jareman, F., Hedlund, J., and Sterte, J. (2003). Sep. Purif. Technol. 32: 159–163. Mabande, G.T.P., Pradhan, G., Schwieger, W. et al. (2004). Microporous Mesoporous Mater. 75: 209–220. Tatlier, M., Tantekin-Ersolmaz, S., Atalay-Oral, C., and Erdem-Senatalar, A. (2001). J. Membr. Sci. 182: 183–193. Sterte, J., Hedlund, J., Creaser, D. et al. (2001). Catal. Today 69: 323–329. Lai, W. and Corcoran, E. (2000). Compositions having two or more zeolite layers. US Patent 6037292. Li, Q., Hedlund, J., Creaser, D., and Sterte, J. (2001). Chem. Commun. 527–528. Gora, L., Clet, G., Jansen, J., and Maschmeyer, T. (2001). Stud. Surf. Sci. Catal. 135: 162. Li, Q., Hedlund, J., Sterte, J. et al. (2002). Microporous Mesoporous Mater. 56: 291–302. Mabande, G.T.P., Noack, M., Avhale, A. et al. (2007). Microporous Mesoporous Mater. 98: 55–61. Caro, J., Wang, H., Noack, M., et al. (2006). EP 1127643. Mitra, A., Wang, Z.B., Cao, T.G. et al. (2002). J. Electrochem. Soc. 149: B472–B478. McDonnell, A.M.P., Beving, D., Wang, A.J. et al. (2005). Adv. Funct. Mater. 15: 336–340. Holmberg, B.A., Hwang, S.J., Davis, M.E., and Yan, Y.S. (2005). Microporous Mesoporous Mater. 80: 347–356. Chaikittisilp, W., Davis, M.E., and Okubo, T. (2007). Chem. Mater. 19: 4120–4122. Eslava, S., Iacopi, F., Baklanov, M.R. et al. (2007). Stud. Surf. Sci. Catal. 170: 594–599.

References

115 Lew, C.M. and Yan, Y.S. (2007). Stud. Surf. Sci. Catal. 170: 1502–1507. 116 Chiang, A.S.T., Wong, L.J., Li, S.Y. et al. (2007). Stud. Surf. Sci. Catal. 170:

1583–1589. 117 Beving, D., O’Neill, C., and Yan, Y.S. (2007). Stud. Surf. Sci. Catal. 170:

1629–1634. 118 Cheng, X.L., Wang, Z.B., and Yan, Y.S. (2001). Electrochem. Solid-State Lett. 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144

4: B23–B26. Wyllie, M.R.J. and Patnode, H.W. (1950). J. Phys. Chem. 54: 204–227. Barrer, R.M. and James, S.D. (1960). J. Phys. Chem. 64: 417–421. Robb, W.L. (1968). Ann. N. Y. Acad. Sci. 146: 119–137. Clarizia, G., Algieri, C., and Orioli, E. (2004). Polymer 45: 5671–5681. Moore, T. and Koros, W.J. (2005). J. Mol. Struct. 739: 87–98. Tantekin-Ersolmaz, S.B., Atalay-Oral, C., Tatler, M. et al. (2000). J. Membr. Sci. 175: 285–288. Koros, W.J. and Mahajan, R. (2001). J. Membr. Sci. 181: 141. Mahajan, R. and Koros, W.J. (2000). Ind. Eng. Chem. Res. 39: 2692–2696. Moore, T.T., Vo, T., Mahajan, R. et al. (2003). J. Appl. Polym. Sci. 90: 1574–1580. Moore, T.T., Mahajan, R., Vu, D.Q., and Koros, W.J. (2004). AIChE J. 50: 311–321. Chung, T.S., Jiang, L.Y., Li, Y., and Kulpathipanja, S. (2007). Prog. Polym. Sci. 32: 483–507. Jiang, L.Y., Chung, T.S., Cao, C. et al. (2005). J. Membr. Sci. 252: 89–100. Jiang, L.Y., Chung, T.S., and Kulprathipanja, S. (2006). AIChE J. 52: 2898–2908. Jiang, L.Y., Chung, T.S., and Kulprathipanja, S. (2006). J. Membr. Sci. 276: 113–125. Li, Y., Chung, T.S., Huang, Z., and Kulprathipanja, S. (2006). J. Membr. Sci. 277: 28–37. Li, Y., Guan, H.M., Chung, T.S., and Kulprathipanja, S. (2006). J. Membr. Sci. 275: 17–28. Li, Y., Chung, T.S., Cao, C., and Kulprathipanja, S. (2005). J. Membr. Sci. 260: 45–55. Zones, S.I., Yuen, L., and Miller, S.J. (2004). Small crystallite zeolite CHA. US Patent 6709644. Stern, S.A. (1994). J. Membr. Sci. 94: 1–65. Vu, D.Q., Koros, W.J., and Miller, S.J. (2003). J. Membr. Sci. 211: 311–334. Reid, B.D., Ruiz-Trevino, F.A., Musselmann, I.H. et al. (2001). Chem. Mater. 13: 2366–2373. Zhang, Y., Balkus, K.J., Musselmann, I.H., and Ferraris, J.P. (2008). J. Membr. Sci. 325: 28–39. Readman, J.E., Gameson, I., Hriljac, J.A., and Anderson, P.A. (2005). Microporous Mesoporous Mater. 86: 96–105. Bhange, D.S. and Ramaswamy, V. (2006). Mater. Res. Bull. 41: 1392–1402. Schlenker, J.L., Pluth, J.J., and Smith, J.V. (1979). Mater. Res. Bull. 14: 751–758. Rudolf, P.R. and Garces, J.M. (1994). Zeolites 14: 137–146.

221

222

6 Zeolite Membranes

145 Noack, M., Schneider, M., Dittmar, A. et al. (2009). Microporous Meso-

porous Mater. 117: 10–21. 146 Treacy, M.M.J. and Higgins, J.B. (2001). Collection of Simulated XRD Pow-

der Patterns for Zeolites, 4e. Amsterdam: Elsevier. 147 Poshusta, J.C., Tuan, V.A., Pape, E.A. et al. (2000). AIChE J. 46: 779–789. 148 Coronas, J., Falconer, J.L., and Noble, R.D. (1997). AIChE J. 43: 1797–1812. 149 Kondo, M., Komori, M., Kita, H., and Okamoto, K. (1997). J. Membr. Sci. 150 151 152

153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171

133: 133–141. Boulicaut, L., Brandani, S., and Ruthven, D.M. (1998). Microporous Mesoporous Mater. 25: 81–93. van Koningsveld, H. and Jansen, J.C. (1996). Microporous Mater. 6: 159–167. Cook, M. and Conner, W.M.C. (1999). How big are the pores of zeolites? In: Proceedings of the 12th International Zeolite Conference, 409. Baltimore, MD: Materials Research Society. Snurr, R.Q. and Karger, J. (1997). J. Phys. Chem. B 101: 6469–6473. Paschek, D. and Krishna, R. (2001). Langmuir 17: 247–254. Bowen, T.C., Wyss, J.C., Noble, R.D., and Falconer, J.L. (2004). Ind. Eng. Chem. Res. 43: 2598–2601. Bakker, W.J.W., Kapteijn, F., Poppe, J., and Moulijn, J.A. (1996). J. Membr. Sci. 117: 57–78. Vroon, Z., Keizer, K., Burggraaf, A.J., and Verweij, H. (1998). J. Membr. Sci. 144: 65–76. Li, S.G., Tuan, V.A., Noble, R.D., and Falconer, J.L. (2001). Ind. Eng. Chem. Res. 40: 4577–4585. Keizer, K., Burggraaf, A.J., Vroon, Z., and Verweij, H. (1998). J. Membr. Sci. 147: 159–172. Lai, Z. and Tsapatsis, M. (2004). Ind. Eng. Chem. Res. 43: 3000–3007. Tuan, V.A., Falconer, J.L., and Noble, R.D. (1999). Ind. Eng. Chem. Res. 38: 3635–3646. Okamoto, K., Kita, H., Horii, K. et al. (2001). Ind. Eng. Chem. Res. 40: 163–175. Poshusta, J.C., Noble, R.D., and Falconer, J.L. (2001). J. Membr. Sci. 186: 25–40. Ash, R., Barrer, R.M., and Pope, C.G. (1963). Proc. Roy. Soc. Lond. A. 271: 1–18. Ash, R., Barrer, R.M., and Lowton, R.T. (1973). J. Chem. Soc. Faraday Trans. 1: 2166–2178. Katz, M.G. and Baruch, G. (1986). Desalination 58: 199–211. May-Marom, A. and Katz, M.G. (1986). J. Membr. Sci. 27: 119–130. Cuperus, F.P., Bargeman, D., and Smolders, C.A. (1992). J. Membr. Sci. 71: 57–67. Cao, G.Z., Meijerink, J., Brinkman, H.W., and Burggraaf, A.J. (1993). J. Membr. Sci. 83: 221–235. Gallaher, G.R. and Liu, P.K.T. (1994). J. Mater. Sci. 92: 29–44. Lin, C.L., Flowers, D.L., and Liu, P.K.T. (1994). J. Membr. Sci. 92: 45–58.

References

172 Fain, D.E. (1989). A dynamic flow-weighted pore size distribution.

173 174

175 176 177 178 179 180 181 182 183

Proceedings of the 1st International Conference on Inorganic Membranes, Montpellier, France (3–6 July 1989), p. 199. Julbe, A. and Ramsay, J.D.F. (2000). Fundamentals of Inorganic Membrane Science and Technology. Elsevier. Steriotis, T.A., Stefanopoulos, K.L., Mitropoulos, A.C., and Kanellopoulos, N.K. (2000). Recent Advances in Gas Separation by Microporous Ceramic Membranes. Elsevier. Clark, T.E., Deckman, H.W., Cox, D.M., and Chance, R.R. (2004). J. Membr. Sci. 230: 91–98. Caro, J. (2002). Handbook of Porous Solids, 363. Wiley-VCH. Beltsios, K.G., Steriotis, T.A., Stefanopoulos, K.L., and Kanellopoulos, N.K. (2002). Handbook of Porous Solids, 2382. Wiley-VCH. Yu, M., Falconer, J.L., Amundsen, T.J. et al. (2007). Adv. Mater. 19: 3032–3036. Yu, M., Amundsen, T.J., Hong, M. et al. (2007). J. Membr. Sci. 298: 182–189. Sing, K.S.W. (1982). Pure Appl. Chem. 54: 2201–2218. McLeary, E.E. and Jansen, J.C. (2004). Top. Catal. 29: 85–92. Hong, M., Falconer, J.L., and Noble, R.D. (2005). Ind. Eng. Chem. Res. 44: 4035–4041. Li, S.G., Falconer, J.L., and Noble, R.D. (2004). J. Membr. Sci. 241: 121–135.

223

225

7 Gas Separations with Zeolite Membranes 7.1 Introduction The development of membranes for gas separation dates back to 1961 with the production of high-flux asymmetric membranes by Loeb and Sourirajan [1]. Since then, different polymer families have been widely investigated for gas separation, such as polycarbonate, polyester, polysulfone, polyethersulfone (PES), polyetherimide, polyimide, and others [2–4]. Nowadays, gas separation is widely used in industrial processes such as hydrogen recovery [5–7], carbon capture and storage (CCS) [8–12], natural gas upgrading [13–15], recovery of alkenes from alkanes [16], benzene derivative separation [17, 18], and air purification [19–21]. The commonly used technologies in these processes include cryogenic distillation [22] and absorption and adsorption [23, 24], as well as the use of membrane technology [25–27]. Processes that rely on heat, such as distillation and absorption, account for more than 10% of the energy consumption in the world, increasing global emissions and pollution [28, 29]. Membrane-based separation processes do not require heat and are therefore a competitive approach for gas separation. Approximately 90% of the cost associated with heat generation would be saved by replacing distillation processes with membrane-based separation technology [30]. In addition, membrane separation offers other intrinsic merits, such as a smaller environmental footprint (no emissions), continuous operation, and great simplicity [31–34]. The ideal gas separation membrane is highly permeable and selective to target gases of interest. Unfortunately, all commercially available membranes are subjected to the well-known trade-off between permeability (or permeance) and selectivity, in spite of the progress in membrane technology over the last four decades [35–38]. The goal of simultaneously achieving high separation capacity (gas permeation or flux rate) together with excellent selectivity still remains as difficult as ever and continues to attract interest from both academia and industry. A number of highly permeable and selective membranes have been prepared based on materials with porous architectures for gas and liquid separations. An overview of these membranes based on their compositions is shown in Figure 7.1. Among them, inorganic materials with high thermal and chemical stabilities are the best candidates for separating gases under harsh conditions. Zeolites are the most studied inorganic materials in shape-selective catalysis and gas Microporous Materials for Separation Membranes, First Edition. Xiaoqin Zou and Guangshan Zhu. © 2019 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2019 by Wiley-VCH Verlag GmbH & Co. KGaA.

eo

lit e

s

ks or

Microporous membranes

Inorg

z a nic

Metal–o rgan ic f ram ew

7 Gas Separations with Zeolite Membranes

Figure 7.1 Examples of the state-of-the-art microporous materials as building blocks for the preparation of gas separation membranes: zeolites, metal organic frameworks, and porous organic frameworks are listed considering the enormous number of publications in this area.

ganic fram ous or ewo Por rks

226

separation. Zeolite membranes are known to have high permeabilities in gas separations. Due to the well-defined pore structures, zeolite membranes can also offer high selectivities. In addition, zeolite-based membranes have high chemical, mechanical, and thermal stability, i.e. can potentially be used at both very high and very low temperatures, offering a great advantage over polymeric membranes. The present chapter will focus on gas separation applications of zeolite membranes, i.e. H2 recovery, air separation, CO2 capture, H2 S capture, and Kr/Xe and N2 /CH4 separation. Post-synthesis modifications of zeolite membranes to enhance separation performance are highlighted as well. Finally, general conclusions about the state of the art and an outlook on important development directions are discussed.

7.2 H2 Recovery Gas separation membranes are widely used for H2 recovery [39]. Among the microporous membranes, the X-ray amorphous metal oxide membranes, mainly silica [40–43], and zeolite membranes, especially the MFI-type zeolites [44–46], are the most common ones. Carbon molecular sieves (CMS) are another commonly used membranes in H2 recovery process [47–53]. Since the kinetic diameter of H2 is around 0.29 nm, the pore diameter of the membrane should be larger than this but smaller than the kinetic diameters of the molecules from which H2 is to be separated. The microporous membranes based on amorphous metal oxides (SiO2 , TiO2 , ZrO2 ) are most often prepared by a sol–gel technique using spin coating or dip coating. On the one hand, SiO2 sol–gel layers show excellent separation factors and fluxes but have a very limited hydrothermal stability, which excludes their use for H2 removal from atmospheres containing steam at high

7.2 H2 Recovery

temperature. On the other hand, TiO2 and ZrO2 sol–gel layers are much more stable, but it does not succeed to prepare highly selective narrow-pore membrane layers for gas separation. In contrast to these X-ray amorphous metal oxide membranes, crystalline zeolite membranes offer much better thermal and hydrothermal stability. Gas permeation results for silicalite-1 membranes have been reported in many studies in the literature [54–69]. The mixture separation factor (𝛼) of H2 from i-butane increases from approximately 𝛼 ≈ 1.5 at room temperature to 𝛼 ≈ 70 at 500 ∘ C (Figure 7.2). It must be considered here that both H2 and i-butane can pass the 0.55 nm pores of the silicalite-1 membrane due to their smaller kinetic diameter (0.29 and 0.50 nm, respectively). At low temperature, mainly i-butane is adsorbed inside the silicalite-1 pores, and the slowly moving i-butane molecules block the diffusion paths of the rarely adsorbed highly mobile H2 . With increasing temperature, less i-butane is adsorbed, and H2 with its higher diffusivity can now move fast in the resulting free volume. It is also remarkable that no degradation during one week of operation at 500 ∘ C including five oxidative regenerations to burn off carbonaceous residues was observed [70]. The reasonable H2 separation factor of 𝛼 ≈ 70 and the H2 permeance of ΠH2 ≈ 1.0 m3 m−2 h−1 bar−1 at 500 ∘ C suggest that this MFI membrane is a candidate for an extractor membrane reactor with selective H2 removal. Summarizing, one can state that silicalite-1 as the only zeolite membrane with a neglecting intercrystalline defect flux developed so far can hardly compete with other organic and inorganic membranes. However, zeolite membranes can become important tools for H2 separation if it succeeds to develop narrow-pore all-silica zeolite membranes with pore sizes near 0.3 nm as a thin (μm or nm 100 Fresh MFI membrane

90

MFI membrane after one week at 500 °C including five oxidative regenerations

80 Separation factor (H2/i-C4)

Figure 7.2 Mixture separation factor 𝛼 for H2 /i-butane (a feed composition of 1 : 3, which is representative for the equilibrium composition of i-butane dehydrogenation at 500 ∘ C) at different temperatures [70]. Source: Copyright 2001, Elsevier.

70 60 50 40 30 20 10 0 0

200

400

Temperature (°C)

600

227

228

7 Gas Separations with Zeolite Membranes

Table 7.1 Separation pattern of the AIPO4 -5 membrane tested at 91 ∘ C and at a pressure difference of 1.0 bar over the membrane [71].

n-Heptane (single gas)

n-Heptane/ toluene

n-Heptane/ mesitylene

n-Heptane/ triethylbenzene

n-Heptane/ triisopropyl benzene

Permeation flux (density × 106 / mol s−1 cm−2 )

3.9

0.85

0.43

1.82

0.94

Flux densities relative to pure n-heptane

100%

22%

11%

47%

24%

Separation factor 𝛼 = [n-heptane]/ [aromatic]/in the permeate

—

0.8

1.7

105

1220

Feed binary mixtures 1 : 1 (in mole) of n-heptane and an aromatic compound, measurements after one hour of permeation in steady state. The membrane contained 595 oriented AlPO4 -5 crystals of a size of 47 μm × 10 μm, giving a net molecular sieve area for permeation of 4.8 × 10−4 cm2 . Source: John Wiley & Sons.

thick) separation layer. These requirements are based on experiences obtained from a model membrane of aligned AlPO4 -5 crystals with a 1D pore system (Table 7.1) [71, 72]. For the case of real molecular sieving, permeation of the component that is to be separated is not influenced by the presence of other mixture components, and the flux of this component can be quite high. In the usual case for silicalite-1 membranes, all mixture components can enter the pores of the zeolite membrane due to its size, and the observed separation effect is the result of a complicated interplay of mixture diffusion and mixture adsorption. In this transport mechanism, strongly adsorbed or bulky components can drastically reduce the permeation of more mobile components. On the contrary, the narrow pore size and the rather compactness (low density of pores per unit area) of the suitable zeolite structures for H2 sieving require thin membrane layers for reasonable fluxes.

7.3 Air Separation Zeolite membranes have been widely studied for potentially highly selective molecular sieving gas separations. Among various gas pairs, separation of air into oxygen and nitrogen is an important one and is currently achieved primarily by energy-intensive cryogenic distillation [73]. Piera et al. separated CO/air at an extremely low CO concentration of 160 ppbv using an MFI-type membrane in the temperature range of 243–303 K [74]. Van den Bergh et al. investigated permeation and separation properties of various gases (CO2 , CH4 , N2 , and O2 ) and their mixtures using DDR membranes in the temperature range of 220–373 K [75]. The mixture selectivity increased as temperature was decreased, especially for CO2 /CH4 , CO2 /air and N2 O/air mixtures. The CO2 /CH4 separation factor

7.3 Air Separation

was higher than 1000 at temperatures below 250 K. At the same time, the O2 /N2 selectivity was low, amounting to only cal. 2.0, and the flux was very low. Hong et al. studied CO2 /H2 separation using SAPO-34 membranes in the temperature range of 253–308 K [76]. Hedlund and coworkers developed ultrathin (∼0.5 μm in thickness) MFI membranes on open-graded alumina supports [77]. Due to the low thickness and well-defined pore structure, the membranes can display both high flux and high selectivity. These membranes displayed excellent performance for CO2 separation from synthesis gas at high pressure and low temperature [78]. The greater separation performance of the membranes at lower temperature was attributed to stronger adsorption of CO2 in the zeolite pores, effectively blocking the transport of H2 and O2 . A stainless-steel-net-supported zeolite NaA membrane with high permeance and high permselectivity for O2 over N2 was studied by our group [79]. Zeolite A possesses uniform pore size of 4 Å, which falls between the kinetic diameters of N2 (4.2 × 3.2 Å) and O2 (3.8 × 2.8 Å); therefore zeolite A is well suited for air separation. The membranes demonstrated high oxygen permeance (∼2.6 × 10−7 mol m−2 s−1 Pa−1 ) and high permselectivity for O2 over N2 (∼7) as shown in Figure 7.3. Ultrathin MFI membranes were for the first time evaluated for air separation at low feed pressures ranging from 100 to 1000 mbar at cryogenic temperature [80]. The O2 /N2 separation factor at optimum temperature increased as the feed pressure was decreased and reached 5.0 at 100 mbar feed pressure and a membrane temperature of 67 K. The corresponding membrane selectivity was 6.3, and the O2 permeance was as high as 8.6 × 10−7 mol m−2 s−1 Pa−1 . This permeance was about 100 times higher than that reported for promising polymeric membranes. The membrane selectivity and high O2 permeance were most likely a result of O2 /N2 10 Separation factor Permeance

8

Separation factor

8 6 6 4 4

2 0

2

Permeance ×10–7(mol m–2s–1Pa–1)

10

–2 0

20

40

60

80 100 120 140 160 180 200 Time (min)

Figure 7.3 The O2 /N2 separation factor and the oxygen permeance of the stainless-steel-net-supported zeolite NaA membrane at different time intervals [79]. Source: John Wiley & Sons.

229

230

7 Gas Separations with Zeolite Membranes

adsorption selectivity. The increase in O2 /N2 separation factor with decreasing pressure and temperature was probably due to increased adsorption selectivity at reduced temperature. This work has demonstrated the potential of MFI zeolite membranes for O2 /N2 separations at cryogenic temperature. Ye et al. evaluated membranes for air separation at cryogenic temperatures (i.e. below 123 K) [81]. The membranes were oxygen selective, and when the feed pressure was decreased, the maximum O2 /N2 separation factor increased, whereas the optimum separation temperatures decreased. In the case of the separation of CO2 /N2 mixtures using zeolite membranes, Kusakabe et al. have reported the separation of CO2 /N2 using Faujasite-type zeolite (NaY) membranes with separation factor of 100 and CO2 permeance of 1.5 × 10−7 mol m−2 s−1 Pa−1 [82–84]. MFI membranes have also been used in the separation of CO2 /N2 mixtures, due to their low Al content, which gives these membranes a good reproducibility and chemical stability [54, 85–89]. Ando et al. found a separation factor of 25.5 at a CO2 permeance of 6.6 × 10−7 mol m−2 s−1 Pa−1 [85]. Dunne et al. reported that at 300 K, the CO2 and N2 isosteric heats of adsorption at the limit of zero coverage increased in an order of silicalite-1 < H-ZSM-5 < Na-ZSM-5 and that for Na-ZSM-5 (Si/Al = 30) was as high as for NaX (Si/Al = 1.23) [90, 91]. The fact was that Na+ ions provided the strongest electric field within the pore. However, at high loadings, the effects were less pronounced, and the adsorption heat was similar for silicalite-1, H-ZSM-5, and Na-ZSM-5. Bernal and coworkers tested the separation of CO2 /N2 mixtures on MFI-type zeolite (ZSM-5) membranes supported on alumina and stainless-steel tubular supports [92]. The selective separation of CO2 from N2 took place because of the preferential adsorption of CO2 in zeolite pores, which hindered the permeation of N2 through the zeolite pore network. CO2 has a stronger electrostatic quadrupole than N2 , leading to a more intense interaction of CO2 with the zeolite material, which translates into a preferential adsorption of CO2 in CO2 /N2 mixtures. Thus, it can be expected that surface diffusion of CO2 would make a significant contribution to its permeation, while at the same time adsorbed CO2 would reduce the N2 permeation flux through the membrane; both factors would lead to selective CO2 /N2 separations [54, 82, 88, 89, 93].

7.4 CO2 Capture The development of proper separation technologies for the removal of CO2 from exhaust gases and from natural gas is still a challenging problem. In applications for natural gas treatment, the feed gas usually stems directly from gas wells in a wide pressure range from 20 to 70 bar with 5–50% CO2 . The product gas must contain less than 2% CO2 . At the moment, it is not economic to produce high-purity gas from gas fields with CO2 content higher 10%. Glassy polymer membranes are used for natural gas purification (removal of CO2 , H2 O, and H2 S), but they suffer from swelling-induced plastification by incorporation

7.4 CO2 Capture

of CO2 and hydrocarbons [93], which reduces their selectivity. This kind of membrane failure would not happen with zeolite membranes since they are chemically stable toward organic solvents and against plastification due to gas absorption. Although polymer membranes with a high performance for CO2 /CH4 separation exist, these membranes have only a rather low separation performance in the CO2 /N2 separation because of low diffusivity and solubility selectivity due to the similar size of CO2 and N2 [94–96]. CO2 (0.33 nm in kinetic diameter), N2 (0.364 nm in kinetic diameter), and CH4 (0.38 nm in kinetic diameter) are relative small molecules, that is to say, much smaller than the pores of large- and medium-pore zeolites. Therefore, the separation of CO2 from N2 or CH4 using zeolite membranes will be based on competitive adsorption, and the selectivity was found to be rather low. Nevertheless, most often the MFI-type membrane was studied [54, 55, 57, 61, 68, 88, 97–104]. As an example, Lovallo et al. [54] obtained a selectivity of approximately 10 for a silicalite-1 membrane at 120 ∘ C. CO2 has a stronger electrostatic quadrupole moment than N2 , leading to a preferential adsorption of CO2 from N2 /CO2 mixtures [105]. Thus, it can be expected that surface diffusion of CO2 contributes significantly to its permeation and simultaneously reduces the N2 permeation flux. The best results for the separation of CO2 /N2 mixtures on large-pore zeolite membranes were reported by Kusakabe et al. using FAU-type membranes [83, 106]. In contrast, small-pore zeolites such as zeolite T (0.41 nm pore size), DDR (0.36 nm × 0.44 nm), and SAPO-34 (0.38 nm) have pores that are similar in size to CH4 but larger than CO2 . It can be expected, therefore, that these membranes show high CO2 /CH4 selectivity due to a combination of differences in diffusion and adsorption. For T-type zeolite membranes, Cui et al. [107] found a mixture separation factor 𝛼 = 400 with a CO2 permeance of Π = 4.6 × 10−8 mol m−2 s−1 Pa−1 at 35 ∘ C. Tomita et al. obtained a CO2 /CH4 separation factor of 𝛼 = 220 with a CO2 permeance of Π = 7 × 10−8 mol m−2 s−1 Pa−1 at 28 ∘ C using a DDR membrane [108]. Very powerful SAPO-34 membranes were recently synthesized by in situ crystallization on a porous tubular stainless-steel support by Noble and Falconer [109]. For a SAPO-34 membrane synthesized from a Si/Al gel ratio of 0.1, a CO2 /CH4 selectivity of 𝛼 = 170 with a CO2 permeance of Π = 1.2 × 10−7 mol m−2 s−1 Pa−1 was found at 22 ∘ C. With decreasing temperature, the selectivity increased, and at −21 ∘ C a CO2 /CH4 separation factor 𝛼 = 560 was found. A SAPO-34 membrane prepared from a gel with (the higher) Si/Al of 0.15 had a slightly lower selectivity (𝛼 = 115) but a higher CO2 permeance (Π = 4 × 10−7 mol m−2 s−1 Pa−1 ) at 35 ∘ C. At 7 MPa, the SAPO-34 membrane showed a CO2 /CH4 selectivity 𝛼 = 100 for a 50%/50% feed at room temperature over about a week [109]. In a previous paper, the same authors found that SAPO-34 membranes could separate CO2 from CH4 best at low temperatures with a selectivity of 𝛼 = 270 at −20 ∘ C [110]. The SAPO-34 membranes effectively separate CO2 from CH4 for conditions at or near industrial requirements (Figure 7.4). However, CO2 flux and selectivity decreased in the presence of water since water has a strong affinity to the hydrophilic SAPO-34 membrane [111].

231

7 Gas Separations with Zeolite Membranes

Upper bound

CO2/CH4 selectivity

100

M3 S1

10 Polymers

0.1

0.1

1

10

100

Figure 7.4 Comparison of the CO2 /CH4 separation selectivity versus the CO2 permeability for polymeric membranes and two SAPO-34 membranes (M3 and S1) at room temperature (feed and permeate pressures of 222 and 84 kPa, respectively) [109]. Source: John Wiley & Sons.

1000 10000

CO2 permeability ×1010 (cm2(STP)/(s.cmHg))

Therefore, hydrophobic small-pore zeolite membranes are more appropriate to separate CO2 from humid gases. Consequently, DD3R membranes show high CO2 flux and selectivity and a negligible water influence on the performance in the CO2 separation from natural gas [112]. Studies of single and binary mixture permeation of CH4 and CO2 through silicalite-1 membranes have shown that the CO2 selectivity in the permeation is due to the favorable CO2 adsorption [88]. The GMS equations, in combination with the ideal adsorbed solution (IAS) theory, were used to model their binary permeation. It was found that the use of accurate adsorption data is of utmost importance for extracting transport properties from the single-component permeation as well as for modeling multicomponent permeation. In detail, both CH4 and CO2 fluxes in the mixture increase with increasing total pressure at 30 ∘ C (Figure 7.5a) [98]. For a fixed total pressure of 101.3 kPa, the CO2 flux in the binary permeate decreases monotonically with temperature, whereas the CH4 flux remains almost constant (Figure 7.5b). Owing to these component fluxes in the binary mixture permeation, the mixture selectivity is almost constant around a value of 4 at 30 ∘ C with increasing gas pressure (Figure 7.6a), but it decreases at 101.3 kPa with increasing temperature 10–1 Flux (mol m–2 s–1)

10–1 Flux (mol m–2 s–1)

232

10–2

10–3 (a)

10–2

10–3 80 120 160 200 240 280 320 pf/tot (kpa)

(b)

300 320 340 360 380 400 420 Temperature (K)

Figure 7.5 Component fluxes of the binary (50 : 50) mixture of CH4 [7] and CO2 [8] through a silicalite-1 membrane: (a) as a function of the total feed pressure at 30 ∘ C and (b) as a function of temperature at a total feed pressure of 101.3 kPa [98]. Source: Copyright 2006, ACS.

6

Selectivity for CO2

Selectivity for CO2

7.4 CO2 Capture

4

2 80

(a)

120

160

200

pf,tot (kPa)

240

6

4

2 300 320 340 360 380 400 420

(b)

Temperature (K)

Figure 7.6 Mixture permeation selectivity for CO2 : (a) as a function of the total feed pressure at 30 ∘ C and (b) as a function of temperature at a total feed pressure of 101.3 kPa; open symbols (J) indicate the ideal selectivity, and the filled ones (K) the real mixture selectivity; the lines are the GMS model predictions [98]. Source: Copyright 2006, ACS.

(Figure 7.6b) [98]. Summarizing, the GMS equations in combination with the IAS theory enables one to predict the binary gas permeation through zeolite membranes. Another separation problem with relevance to practical applications is the CO2 removal from N2 in exhaust gases for CO2 sequestration. Because a CO2 separation will take place at elevated pressures, the CO2 permeation from pressurized feeds on a silicalite-1 membrane on different supports has been studied [113]. A maximum value of 12–13 for the mixture separation factor (CO2 /N2 ) was found between 6 and 16 bar total retentate pressure. The CO2 /N2 selectivity was found to depend on (i) the kind of support and (ii) the modification of the MFI structure. Boron-ZSM-5 was found to have a higher selectivity toward CO2 than Na-ZSM-5, indicating that the adsorption mechanism includes electrostatic components. Furthermore, MFI membranes prepared on stainless-steel supports showed higher CO2 /N2 selectivities than those deposited on alumina since aluminum is believed to leach from the support and to become incorporated into the MFI layer. There are contradicting statements on the separability of CO2 from gas mixtures by zeolite membranes. As an example, for equimolar mixtures of CO2 /N2 on FAU membranes [114], the separation factor was determined to be 𝛼(CO2 /N2 ) ≈ 2–5 at 30 ∘ C, whereas another selectivity with 𝛼(N2 /CO2 ) ≈ 5–8 was reported [115]. These different experimental findings can be explained by the role of moisture. ZSM-5-type zeolite membranes showed a permeance of approximately 3.6 × 10−8 mol m−2 s−1 Pa−1 and a separation factor 𝛼(CO2 /N2 ) ≈ 54.3 at 25 ∘ C and 𝛼(CO2 /N2 ) ≈ 14.9 at 100 ∘ C [116]. However, the separation factor of the ZSM-5 membrane increased as the permeation time increased (Figure 7.7) [116]. This experimental finding is explained by the mechanism that moisture occupies large pores through which mainly the N2 flows, and as a result, the separation factor a (CO2 /N2 ) is higher for moisture-saturated feed gases than for dry feed gases. The same finding was observed independently by Gu et al., namely, that the presence of water vapor significantly enhances the CO2 selectivity of a FAU membrane in the CO2 /N2 mixture separation at

233

7 Gas Separations with Zeolite Membranes

60 Separation factor (CO2/N2)

234

Feed pressure = 400 kPa Permeation temperature = 25 °C Feed flow rate = 350 ml min–1 He sweeping rate = 100 ml min–1

50 40 30 20 10 0 0

10

20

30

40

50

60

Test duration (min)

Figure 7.7 CO2 /N2 separation factor versus gas permeation test duration for a ZSM-5-type zeolite membrane with moisture-saturated feed gases [116]. Source: Copyright 2005, Elsevier.

110–200 ∘ C [117]. Alternatively, a simpler explanation of this experimental finding would be that CO2 is dissolved by a water film acting like a supported liquid film membrane. Another zeolite membrane for CO2 separation is SAPO-34 as a silicon-substituted eight-membered ring (8MR) aluminum phosphate. From CO2 /CH4 mixtures, the smaller CO2 molecules preferentially permeated, and high CO2 selectivities under praxis-relevant test conditions of 30 bar at 50 ∘ C were found [118]. Maghsoudi and Soltanieh employed a high-silica (Si/Al = 5) CHA-type membrane for simultaneous separation of both acid gases (H2 S, CO2 ) from CH4 [119]. Results showed a permeance of 3.39 × 10−8 mol m−2 s−1 Pa−1 for pure CO2 with CO2 /CH4 ideal selectivity of 21.6 at 303 K and 100 kPa pressure difference across the membrane. CO2 /CH4 separation results for pure, binary (CO2 —CH4 ), and ternary (H2 S—CO2 —CH4 ) gas mixtures, with compositions near the real sour natural gas (CO2 : 2.13 mol%, H2 S: 0.3 mol%, and CH4 : 97.57 mol%), at 298 K and 400 kPa pressure difference were showed in Figure 7.8. According to this figure, CO2 permeate flux was reduced from 4.58 × 10−8 mol m−2 s−1 Pa−1 at 303 K for pure gas to 3.56 × 10−8 mol m−2 s−1 Pa−1 for equimolar CO2 —CH4 binary gas mixture at 298 K and then to 2.77 × 10−8 mol m−2 s−1 Pa−1 for ternary H2 S—CO2 —CH4 at 298 K at gas mixture with CO2 concentration of 2.13 mol%. Also, CO2 /CH4 selectivity continuously reduced from ideal selectivity of 13.1–8.0 for equimolar CO2 —CH4 mixture and then to 5.3 for ternary H2 S—CO2 —CH4 gas mixture. The reduction of CO2 permeance and CO2 /CH4 selectivity with decreasing CO2 concentration is due to diffusion selectivity rather than desorption selectivity. In other words, it can be stated that slowly diffusing CH4 molecule may hinder the fast-diffusing CO2 molecule.

7.5 N2 /CH4 Separation

Pure gas at 303 K 20 Binary gas mixture (CO2/CH4)

4

15 3 Ternary gas mixture (0.3 mol% H2S)

10

2 e ixtur as m ) g y r Bina O 2/CH 4 (C

1

CO2/CH4 selectivity

CO2 permeance (10–8 × mol m–2 s–1 Pa–1)

5

5 Permeance Selectivity

0 0

10

20

30

40

50

60

70

80

0 90 100

CO2 concentration in the feed (mol%)

Figure 7.8 CO2 permeance and CO2 /CH4 selectivity at 298 K and 400 kPa pressure difference for different feed compositions [119]. Source: Copyright 2014, Elsevier.

7.5 N2 /CH4 Separation Natural gas consists primarily of methane, higher alkanes, carbon dioxide, nitrogen, and hydrogen sulfide. In particular, CO2 and N2 decrease the heat value of the natural gas. Therefore, it is highly desirable to remove CO2 and N2 from natural gas to improve its heat content and to avoid the erosion of pipelines from CO2 in the presence of moisture. Zeolites membranes like AlPO-18, SAPO-34, and SSZ-13 have been reported to separate N2 /CH4 mixtures. AlPO-18, a microporous aluminophosphate with AEI topology, possesses a three-dimensional framework with a micropore size of 0.38 nm [120]. Based on the micropore size of AlPO-18 (0.38 nm), this aluminophosphate could potentially molecular sieve N2 (0.36 nm) over CH4 (0.38 nm) or at least can make that N2 diffuse through the pores more rapidly than CH4 . The first example of continuous AlPO-18 membranes exhibiting molecular sieving properties for CO2 /CH4 gas mixtures was demonstrated by Carreon et al. [121]. It was then confirmed by the group of Zhou for the molecular sieving properties of AlPO-18 membranes for the separation of CO2 /CH4 , CO2 /N2 , and H2 /CH4 binary gas mixtures [122, 123]. Carreon et al. continued their investigation on AlPO-18 membranes for the separation of equimolar of N2 /CH4 mixtures [124]. AlPO-18 membranes displayed N2 permeances as high as 3076 GPU (1.0 GPU = 10−6 cm3 (STP) cm−2 s−1 cmHg−1 = 3.35 × 10−10 mol m−2 s−1 Pa−1 ), with N2 /CH4 separation selectivity as high as 4.7. They have conducted an economic evaluation for N2 /CH4 separation [125] and suggested that high N2 permeances are essential to reduce the N2 rejection cost employing membrane technology.

235

236

7 Gas Separations with Zeolite Membranes

SAPO-34, a CHA-type silicoaluminophosphate zeolite, with average pore size of 0.38 nm represents an ideal candidate in membrane form to molecular sieve N2 from CH4 . The first example of a continuous SAPO-34 membrane was reported by Zhang et al. dating back to the year 1997 [126]. The Noble and Falconer groups have demonstrated the high-performance SAPO-34 membranes that efficiently separated diverse gas mixtures including CO2 , H2 , CH4 , and other light hydrocarbons [127, 128]. The authors reported a N2 /CH4 selectivity of 5–7 for a feed pressure of 350 kPa and N2 permeance of approximately 300 GPU. Although the separation selectivity was half of that observed for SSZ-13 membranes, the observed N2 permeance was approximately five times higher. The higher N2 permeance of SAPO-34 membranes as compared with SSZ-13 membranes was likely related to the membrane thicknesses. By optimizing the synthesis of SAPO-34 membranes, some optimized SAPO-34 membranes were prepared by Li et al. [125]. Separation of N2 from CH4 revealed N2 permeance as high as 500 GPU and N2 /CH4 separation selectivity as high as 8 at 24 ∘ C for a 50/50 N2 /CH4 mixture. They also found that the heat of adsorption was higher for CH4 (15 kJ mol−1 ) than N2 (11 kJ mol−1 ), leading to a preferential adsorption of CH4 over N2 in the N2 /CH4 mixture. On the other hand, the N2 molecule diffused much faster than the CH4 molecule, suggesting that differences in diffusivity played a more critical role than the competitive adsorption. As such, the SAPO-34 membranes were selective for N2 over CH4 in the mixture. A high-aspect-ratio SAPO-34 seeds were employed by the groups of Yu and Li to synthesize high-quality thin SAPO-34 membranes (i.e. 10 unit cells in thickness) for N2 /CH4 separation [129]. The resulting thin SAPO-34 membranes were tested for separation of equimolar N2 /CH4 mixtures at 22 ∘ C and at a fixed feed and permeate pressure of 275 kPa and 101 kPa, respectively. These SAPO-34 membranes displayed N2 /CH4 selectivity as high as 11.2 and N2 permeance of approximately 860 GPU. Zong et al. optimized further synthesis conditions for SAPO-34 membranes toward N2 /CH4 separation by diluting the membrane synthesis gels [130]. Cross-sectional SEM images of the membranes (Figure 7.9b) showed that the thickness of the membranes decreased from approximately 6.8 to approximately 2.7 μm as the water content increased from 150 to 400, indicating that diluted gel formed a thinner membrane layer. These SAPO-34 membranes offered N2 permeance as high as 1300 GPU and selectivity of 7.4 at 23 ∘ C for a 50/50 N2 /CH4 mixture. In addition, these membranes were reproducible and effectively separated gas mixtures having low N2 content, which are relevant compositions of natural gas. In another study, SAPO-34 membranes were further optimized by Zong et al. through hydrothermal synthesis in the water bath, ice/water bath, and flowing water cooling [131]. The resulting thin SAPO-34 membranes led to a higher permeance for N2 (∼2600 GPU) and a slightly improved N2 /CH4 selectivity (7.4). They found that the heat flux rate of the stainless-steel autoclaves during membrane preparation was two orders of magnitude higher than that of the membranes prepared in conventional Teflon liners. The higher heat flux rate slowed

7.5 N2 /CH4 Separation

(A)

(B)

Zeolite layer 6.8 μm

(a)

Zeolite layer 6.3 μm

(b)

Zeolite layer 4.2 μm

(c)

Zeolite layer 3.0 μm

(d)

Zeolite layer 2.7 μm

(e)

Figure 7.9 Top (A) and cross-sectional view SEM (B) images of SAPO-34 membranes prepared with synthesis gel compositions of 1.0Al2 O3 : 1.0P2 O5 : 0.3SiO2 : 1.0TEAOH : 1.6DPA : xH2 O: (a) x = 150, (b) x = 200, (c) x = 250, (d) x = 300, and (e) x = 400 [130]. Source: Copyright 2015, Elsevier.

237

238

7 Gas Separations with Zeolite Membranes

down the kinetics of crystal growth, translating into thinner and more N2 permeable membranes. Wu et al. prepared SSZ-13 zeolite membranes via secondary seeded growth on the outer surface of mullite porous tubes [127]. The resultant membranes displayed remarkably high N2 /CH4 separation selectivity of 13 but relatively low N2 permeance (66 GPU) at 20 ∘ C for a feed pressure of 270 kPa.

7.6 H2 S Capture As natural gas demand is expected to nearly double in the coming 25 years and the raw natural gas has a varying composition depending on its origin [132], there is an opportunity for membrane technology to remove impurities including water, carbon dioxide, nitrogen, hydrogen sulfide, and other hydrocarbons [133]. The presence of H2 S in the feed stream resulted in an increased separation cost for meeting the specifications of both H2 S and CO2 (4 ppm and 2%, respectively). Maghsoudi and Soltanieh conducted the simultaneous separation of H2 S from CH4 using a high-silica CHA-type zeolite membrane [119]. Ternary mixture (2.13% CO2 , 0.30% H2 S in CH4 ) permeance was measured at room temperature, and the H2 S permeance was 1.70 × 10−8 mol m−2 s−1 Pa−1 with H2 S/CH4 selectivity of 3.24.

7.7 Kr/Xe Separation The separation of krypton from xenon is an industrially relevant problem. Kr and Xe are widely used in fluorescent light bulbs. High-purity Xe has been used in commercial lighting, medical imaging, anesthesia, and neuroprotection [134]. In addition, separating Kr from Xe is an important issue for nuclear industries. Specifically, separating Kr from Xe is a critical step in removing radioactive 85 Kr during treatment of spent nuclear fuel [135]. Effective separation of Kr from Xe in nuclear reprocessing plants would lead to a considerable reduction in storage costs. The current conventional technology produces these gases from the cryogenic distillation of air in which these noble gases are present in very small concentrations [136], which is again an energy-intensive process. Membrane technology could play a key role in making this separation less energy intensive and economically feasible since it does not involve any phase transformation. In addition, membrane process is simple and easy to operate, control, and scale up than the conventional cryogenic distillation. Based on the kinetic diameter of Kr (∼3.69 Å) and Xe (∼4.10 Å), in principle, SAPO-34, a CHA-type small-pore silicoaluminophosphate zeolite displaying average pore size of 3.8 Å in between these two gases, is considered a suitable candidate to separate Kr from Xe by molecular sieving if prepared in membrane form. In the past, only several membranes have been proposed to separate Kr and Xe (Table 7.2). However, all these membranes have been evaluated only for single gas permeation experiments and not for gas mixtures. Furthermore,

7.7 Kr/Xe Separation

239

Table 7.2 Membranes for single gas permeation of Kr and Xe.

Membrane

Kr permeance (mol m−2 s−1 Pa−1 )

Xe permeance (mol m−2 s−1 Pa−1 )

Kr/Xe ideal selectivity

Reference

4-Fluoroethylene

3.7 × 10−7

2.9 × 10−7

1.3

[137]

Cellulose acetate

7.3 × 10−12

2.2 × 10−12

3.7

[138]

−11

−11

[139]

Carbon

6.3 × 10

2.1 × 10

3.0

DDR

2.9 × 10−10

5.4 × 10−11

5.4

SAPO-34

2.4 × 10−9

2.4 × 10−10

9.9

Table 7.3 Performance comparison of various membranes in Kr/Xe separations.

Feed composition molar ratio

Separation factor (𝜶)

Kr permeance (mol m−2 s−1 Pa−1 )

Reference

35

10 × 10−8

[140]

Zeolite

Support

SAPO-34

α-Al2 O3

Kr/Xe 9 : 1 (138 kPa, 298 K)

AlPO-18

α-Al2 O3

Kr/Xe 92 : 8 (107 kPa, 298 K)

6.4

15.5 × 10−8

[141]

Silicalite-1

Polyethersulfone (PES)

Kr or Xe (single gas)

0.22

1.82 × 10−9

[142]

ZIF-8

α-Al2 O3

Kr/Xe 92 : 8 (140 kPa, 298 K)

10.6

0.8 × 10−8

[143]

SAPO-34

Ceramic

Kr/Xe 10 : 90 (1.0 atm, 255 K)

30

0.36 × 10−8

[144]

K-SAPO-34

α-Al2 O3

Kr/Xe 10 : 90 (140 kPa, 295 K)

37

6.3 × 10−9

[145]

Hollow fiber K-SAPO-34

α-Al2 O3

Kr/Xe 10 : 90 (140 kPa, 295 K)

48

5.0 × 10−9

[146]

MFI

γ-Al2 O3

CO2 /Xe 50 : 50 (1.5 bar, 448 K)

1213 GPU for CO2 , 200 GPU for Xe

[146]

5.6

these membranes show low ideal separation selectivity and low Kr permeance. However, significant progress has been made recently regarding the separation of binary mixture of Kr and Xe, and the representative separation data are summarized in Table 7.3. The first separation of Kr/Xe mixture was studied by the group of Carreon over a SAPO-34 (CHA-type) zeolite membrane [140]. SAPO-34 membranes supported on α-Al2 O3 substrates separated Kr/Xe mixtures with Kr permeance as high as 1.2 × 10−7 mol m−2 s−1 Pa−1 and separation selectivity of 35 for molar compositions close to typical concentrations of these two gases in air. Molecular gas mixtures can be separated over zeolite membranes by at least one of the following separation mechanisms: molecular sieving, differences in diffusivity, and competitive adsorption. Figure 7.10 showed that CHA membrane

7 Gas Separations with Zeolite Membranes

(a)

2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0

50 Kr/Xe selectivity

Kr permeance ×10–7 (mol m–2 s–1 Pa–1)

240

40 30 20 10

More selective Less selective

0 0 1 2 3 4 5 6 7 8 9 10 Membrane thickness (μm) (b)

1 2 3 4 5 6 7 0 Membrane average crystal size (μm)

Figure 7.10 (a) Kr permeance as a function of membrane thickness and (b) separation selectivity as a function of membrane average crystal size: membrane average crystal size for 150, 200, 250, 300, and 350 water mol compositions were 5.9 ± 0.9 μm, 4.8 ± 1.0 μm, 3.0 ± 0.7 μm, 2.7 ± 0.7 μm, and 2.4 ± 0.6 μm, respectively [140]. Source: Copyright 2016, ACS.

thickness and membrane average crystal size have a profound effect on Kr permeance and separation selectivity. Adsorption experiments showed that Xe adsorbed more strongly than Kr on SAPO-34 crystals due to its higher dipole polarizability (26.85–28.7 atomic units) as compared with Kr (16.44–18.0 atomic units), favoring the separation of Xe over Kr in the gas mixture. Stronger electrostatic interactions between the negatively charged SAPO-34 surface and the higher dipole moment of Xe promote the preferential adsorption of Xe over Kr. However, the smaller Kr molecule should diffuse much faster than the larger Xe molecule. Although competitive adsorption and diffusivity/molecular sieving compete in separation mechanisms, separation data suggest that diffusivity differences and molecular sieving are the two dominant mechanisms leading to highly permeable Kr selective membranes. Carreon and coworkers have done a comparison between different molecular sieves, i.e. AlPO-18, SAPO-34, and ZIF-8 on Kr/Xe separation [141]. The results are shown in Figure 7.11. AlPO-18 membranes separated Kr/Xe mixtures with average Kr/Xe separation selectivity of 6.4 and unprecedented high average Kr permeance of 1.6 × 10−7 mol m−2 s−1 Pa−1 . SAPO-34 membranes displayed the best overall separation performance, while AlPO-18 membranes displayed the highest Kr permeances. The key factors affecting both the separation selectivity and permeance were the presence of rigid micropores with size lying between Kr and Xe atomic sizes, lower Xe/Kr uptakes (adsorption selectivity), and lower concentration of nonselective pores. The Kr permeances for these three microporous crystalline membranes decreased exponentially as membrane thickness increased (Figure 7.12). Kwon et al. also reported a detail study on the Kr and Xe adsorption on SAPO-34 [144]. Membrane transport measurements reveal that SAPO-34 membranes can separate Kr from Xe by molecular sieving, with Kr permeabilities around 50 Barrer (1.0 Barrer = 10−10 cm3 (STP) cm cm−2 s−1 cmHg−1 = 3.35 × 10−16 mol m m−2 s−1 Pa−1 ) and mixture selectivities of 25–30 for Kr/Xe at ambient or slight sub-ambient conditions. The membrane transport

7.7 Kr/Xe Separation

Kr/Xe selectivity

100

SAPO-34 ZIF-8

10

AIPO-18 1

10

1

100

1000

Kr permeance (GPU)

Figure 7.11 Kr/Xe separation selectivity vs. Kr permeance over AlPO-18, SAPO-34, and ZIF-8 membranes for a Kr-rich feed gas mixture (molar gas mixture composed of 92:8 Kr/Xe was employed for AlPO-18 (orange ▴), SAPO-34 (green ⧫), and ZIF-8 (black ⬧) and 9:1 Kr/Xe molar gas mixture for SAPO-34 (green ×) [141]. Source: Copyright 2017, ACS.

Kr permeance (mol m–2 s–1 Pa–1)

4E–7

3E–7 AIPO-18 SAPO-34

2E–7

ZIF-8

1E–7 Trendline: P = 4 × 10–7 d–1.297 0 0

10

20

30

Membrane thickness (μm)

Figure 7.12 Kr permeance for SAPO-34, ZIF-8, and AlPO-18 membranes as a function of membrane thickness over Kr-rich feed gas mixture. The dashed line indicates the fitting equation [141]. Source: Copyright 2017, ACS.

characteristics are modeled by the Maxwell–Stefan equations, whose predictions are in very good agreement with experiment and confirm the minimal competing effects of adsorption and diffusion. In another study, Kwon et al. extended their study on Kr/Xe separation by using ion-exchanged SAPO-34 membranes [145]. SAPO-34 that has been ion exchanged to H, Li, Na, and K form were analyzed for their performance in Kr/Xe separation. For all the SAPO-34 membranes considered here, the diffusion selectivity (molecular sieving effect) for Kr exceeds the preferential adsorption of Xe, leading to high Kr/Xe permeation selectivity. It is seen that the M−S diffusivities of Kr and Xe decrease monotonically as

241

Permeability or selectivity change (%)

7 Gas Separations with Zeolite Membranes

Figure 7.13 Percent changes in Kr/Xe permeabilities and their selectivities as a function of cation radius [145]. Source: Copyright 2018, ACS.

80

40

0

–40

–80

Kr Xe Selectivity

Li+ (0.076)

Na+ (0.102)

K+ (0.138)

Ion type (ionic radius in nm)

the size of the cation increases. This is attributed to the reduction in effective pore size, which hinders the passage of molecules through the 8MR windows. Figure 7.13 show that the Kr/Xe diffusion selectivity and permeation selectivity of K-SAPO-34 are higher than that of Na-SAPO-34. Among the ion-exchanged membranes, K-SAPO-34 appears to be the best candidate due to its high ideal selectivity (∼37) and only modestly reduced Kr permeance (6.3 × 10−9 mol m−2 s−1 Pa−1 ). Kr permeances in binary mixture increase with decreasing temperature, whereas Xe permeances decrease. As a result, the binary Kr/Xe separation factor increases with decreasing temperature. The Kr/Xe separation factor in K-SAPO-34 membranes is 48 at 253 K and 30 at 298 K. In addition, they also studied the separation on the hollow fiber SAPO-34 membranes before and after ion exchange with potassium cations. Figure 7.14 100 50 Separation factor (α)

Permeance (×10–10 mol m–2 s–1 Pa–1)

242

10

40

30

20

1 HF-M2 K-HF-M2

260

Kr Kr

280

HF-M2 K-HF-M2

Xe Xe

300

Temperature (K)

260

280

300

Temperature (K)

Figure 7.14 Temperature dependence of binary Kr and Xe permeances and Kr/Xe separation factors in hollow fiber H-SAPO-34 and ion-exchanged K-SAPO-34 hollow fiber membranes [145]. Source: Copyright 2018, ACS.

7.7 Kr/Xe Separation

shows the temperature dependence of Kr/Xe binary mixture permeation. The binary mixture separation shows improved selectivities (∼48) after ion exchange at all temperatures with clear molecular sieving effects. Amiri et al. investigated a PES-supported silicalite-1/polydimethylsiloxane (PDMS) mixed matrix membrane in the single gas separation of xenon and krypton [142]. The average permeability of krypton and xenon gases through the PDMS polymeric membrane is calculated as 1.25 × 10−9 and 1.78 × 10−9 cm mol cm−2 s−1 kPa−1 , respectively, while by adding only 5 wt% of nanosilicalite-1 to the polymeric matrix of the membrane, their amounts are increased to 1.82 × 10−9 and 8.07 × 10−9 cm mol cm−2 s−1 kPa−1 , respectively. In addition, the presence of nanosilicalite-1 as the filler leads to an increase in the selectivity of xenon to krypton up to 4.38. Anderson et al. conducted a grand canonical Monte Carlo and biased molecular dynamics simulations (using adaptive biasing force) to elucidate the nature of adsorption- and diffusion-based Kr/Xe separation mechanisms in a set of nanoporous materials: SAPO-34, ZIF-8, UiO-66, and IRMOF-1 [147]. Figure 7.15 shows the crystallographic structures of ZIF-8, IRMOF-1, UiO-66, and SAPO-34 respectively [147]. Xenon is found to preferentially adsorb on all

(a)

(b)

(c)

(d)

Figure 7.15 Crystallographic structures of (a) ZIF-8, (b) IRMOF-1, (c) UiO-66, and (d) SAPO-34 microporous materials [147]. Source: Copyright 2017, ACS.

243

7 Gas Separations with Zeolite Membranes

materials, but diffusion selectivity for krypton is found to dominate the overall membrane separation selectivity. They find that large pore cages are desirable to increase adsorption selectivity for krypton, while stiff pore windows with a diameter smaller than xenon are desirable to increase diffusion selectivity for krypton. No perfect molecular sieving is found, but the relatively rigid SAPO-34 is more effective at excluding xenon than the more flexible ZIF-8. During xenon “window crossing,” the SAPO-34 window opens to only 3.8 Å, while the ZIF-8 window opens to 4.1 Å, resulting in a lower free energy “diffusion” barrier for xenon in ZIF-8. Therefore, an ideal membrane material for Kr/Xe separation should be rigid and should have large pore cages and small pore windows. Despite great achievements in separation of Xe from Kr or air by chabazite SAPO-34 [140, 144, 145], AlPO-18 [141], silicalite-1 [142], and ZIF-8 [143] membranes, CO2 /Xe mixture separation using microporous membranes has been overlooked in the past decades. Recently, Wang et al. employed b-oriented MFI zeolite membranes supported on γ-Al2 O3 support for Xe recovery from CO2 /Xe mixture [146]. They proposed a novel Xe recovery approach from exhaled anesthetic gas based on b-oriented MFI zeolite membranes. This can be used to continuously remove the major impurity of CO2 from the closed-circuit anesthesia system. As shown in Figure 7.16, the separation factor of CO2 to Xe was 5.6, higher than the ideal selectivity and Knudsen constant. The permeance of pure CO2 exhibited a maximum as a function of temperature, proving that surface diffusion dominated CO2 transport in MFI channels. The improved in-plane growth by using a zeolite growth modifier led to fusion of independent crystals and eliminates boundary gaps, giving good selectivity in the separation of CO2 /Xe mixtures. The fast diffusion of CO2 dominated the overall membrane selectivity toward the CO2 /Xe mixture. The good reproducibility and long-term hydrothermal stability (>260 hours) endow b-oriented MFI zeolite membranes with a great potential for practical application of Xe recovery from exhaled anesthetic gas.

CO2

1.5

6

4

1.0

2

0.5 Xe –

+

+

–

–

–

–

+

0.0

0 0

25

50

75 100 125 150 175 200 Temperature (°C)

CO2/Xe separation factor

Separation factor

Permeance/103GPU

244

Figure 7.16 Temperaturedependent separation performance of b-oriented MFI zeolite membrane (M11) for 50/50 CO2 /Xe mixture at an absolute feed pressure of 1.5 bar; the cyan color indicates Knudsen constant of CO2 /Xe (1.7) [146]. Source: Copyright 2018, ACS.

7.8 Post-modification of Zeolite Membranes

7.8 Post-modification of Zeolite Membranes Post-synthesis modification is a convenient way to alter properties or reduce the number of defects within the zeolite film. Such modification alters the zeolite adsorption properties and sometimes allows more selective separation [148]. Among the numerous strategies, which have been adopted to produce zeolite membranes with either a reduced number of defects or with specific properties, a series of modification techniques has been tested, such as isomorphous substitutions of Al, Fe, B, or Ge in the structure, ion exchange, silylation to decrease pore size [149] and to increase hydrophobicity [150, 151], atomic layer deposition (ALD) [152], chemical vapor deposition (CVD) [153], selective coking [154] to fill non-zeolite pores, catalytic cracking, and reactions with adsorption sites in the zeolite structure. Impregnation with triisopropylbenzene (TIPB) (molecular size 8.4 Å) and high temperature pyrolyzing of TIPB for coking were used to increase the selectivity of ZSM-5 membranes by reducing macro-voids and defects between zeolite crystals without affecting the intracrystalline ones [155]. Noble et al. applied soaking in cyclodextrin solution to improve the separation performance of SAPO-34 membranes [156]. Since cyclodextrin molecules are too large to penetrate into zeolite pores, they selectively block the intercrystalline defects. As a result, the CO2 /CH4 separation tests showed significant increase of CO2 /CH4 selectivity without affecting much the CO2 permeance through the modified membranes. Similar methods to seal the defects including dip coating of MFI membranes with amorphous silica layer [157] and silicone rubber [158] have been proposed. Another relevant approach is fabrication of poly-layer zeolite membranes in order to plug the defects of the bottom layer by growing of another layer of the same material on top [159] or to combine desired properties of two different zeolite types in one structure [160]. Permeable polymer like PDMS has been used to seal defects of microporous silica and zeolite Y membranes (Figure 7.17) [161]. The only limitation of using PDMS is the thermal stability. Fluoropolymers or other temperature-resistant polymers may be able to overcome this limitation. CVD modification is another widely used post-modification technique. Yan et al. tuned the zeolite pore size by CVD of silica or carbon layers [154]. Lin and coworkers have performed CVD treatment on high-quality MFI membranes, and they concluded that the transport was governed by Knudsen diffusion for small non-adsorbing molecules like H2 and CO2 . Thus, plugging of large intercrystalline defects by CVD using a molecule like methyldiethoxysilane (MDES) or TEOS increased the H2 /CO2 separation factor. However, this was true only for membranes with good initial quality, i.e. those lacking large intercrystalline voids (Figure 7.18) [162]. Importantly, membranes silylated by MDES exhibited improved thermal and hydrothermal stability [163]. Thus, pore size tuning by deposition of molecular silica inside the zeolite pores offers a possibility to turn initially adsorption selective into stable and truly molecular sieving membranes with high selectivity toward small molecules (mainly H2 ). Nomura et al. applied the counter-diffusion CVD technique in which TEOS and

245

246

7 Gas Separations with Zeolite Membranes

PDMS

SiO2 200 nm

γ-Alumina

α-Alumina

Acc.V Spot Magn Det WD 5.00 kV 3.0 10000x TLD 5.8

2 μm

Figure 7.17 SEM cross section of a PDMS-coated silica membrane supported on γ-alumina/α-alumina support [161]. Source: Copyright 2011, Elsevier. H2

(a)

(c)

CO2

(b)

(d)

Figure 7.18 Schematic illustrations of H2 and CO2 permeation through the unmodified and CVD modified MFI membranes with different initial qualities: (a) unmodified membrane with relatively bad quality, (b) unmodified membrane with initial relatively good quality, (c) CVD modified membrane with bad initial quality, and (d) CVD modified membrane with good initial quality [162]. Source: Copyright 2010, ACS.

7.9 Conclusions and Outlook

ozone cause amorphous silica to deposit on the membrane [164]. This plugged the intercrystalline defects without blocking the zeolite pores since TEOS is too large to enter them. A similar TEOS–O3 system was used recently for plugging of defects in porous silica membranes of pore size approximately 1.0 nm [165]. NaA membranes were modified by CVD of a Pd compound or impregnation with a Pd salt solution in order to minimize the nonselective defects and increase the H2 selectivity [166]. However, in most cases, the increased membrane selectivity was coupled with a substantial decrease of the membrane flux. Catalytic cracking deposition (CCD) was used to reduce the pore size and better separation of H2 and CO2 . Hong et al. tuned pores of B-ZSM-5 and SAPO-34 membranes by CCD of MDES and observed a significant improvement in H2 /CO2 selectivity without proportional decrease of H2 permeance [166]. Tang et al. observed an even more pronounced increase of the H2 /CO2 selectivity after several consecutive CCD steps [167]. Dong et al. reported that a catalytic membrane reactor based on high-quality MFI membrane modified by CCD was capable of surpassing the equilibrium limits in the WGS reaction. The conversion of CO in such zeolite membrane reactor was 81.7%, well above the equilibrium value of 65% at 550 ∘ C and 1.5 bar. However, the limitations of CVD and CCD post-modifications are the expensive equipment and difficulty to scale up. Moreover, amorphous silica deposits tend to undergo densification upon prolonged heating, especially in the presence of water vapor [168]. Therefore, the long-term stability of these membranes is not guaranteed [169]. Apart from blocking membrane defects and decreasing the pore size, post-synthesis modification can be used to tune zeolite adsorption properties. For example, grafting with methylamine enhances the performance of silicalite-1 membranes in CO2 /H2 and CO2 /CH4 separations [170]. The bridge of Si—N—Si strengthens the overall framework basicity, leading to increased CO2 adsorption [171]. Another way to improve CO2 affinity and selectivity is by impregnation with calcium nitrate followed by its thermal decomposition, resulting in CO2 specific adsorption sites [172]. Silylation is applied to enhance the hydrophobic properties of the MFI membranes. Grafting with triethoxyfluorosilane (TEFS) yields materials with drastically increased hydrophobicity by partial replacement of polar Si—OH groups with Si—F ones within the zeolite layer [173]. In conclusion, post-synthesis modification provides a convenient and versatile way to tune the specific separation, adsorption properties, and the overall quality of zeolite membranes.

7.9 Conclusions and Outlook Membrane technology constitutes an increasingly important, convenient, and versatile way of separating gas mixtures. Compared with other approaches, membranes reduce energy and other operational cost for gas separation. Moreover, membrane operations are more scalable than conventional separation unit operations in the chemical industry. Industrial processes would benefit from the use of membranes, such as air separation, recovery of hydrogen from mixtures,

247

248

7 Gas Separations with Zeolite Membranes

hydrocarbon separations, and CO2 capture from natural gas, flue gas, biogas, and syngas. Within the next few years, despite much progress in the development of supported thin-layer zeolite membranes, there is still no application of zeolite membranes in industrial gas separation. Although usually separation performance and thermal and chemical stability of zeolite membranes are unsurpassed by other materials, their commercialization for gas separation requires significant reduction of fabrication costs and improvement of preparation reproducibility. The relatively high costs of the supported zeolite membranes are due to the expensive asymmetric graded ceramic multilayer support. Therefore, implementation of cheaper extruded porous supports for fabrication of membranes is crucial. It is now becoming possible because of remarkable advances recently achieved in coating very rough support surfaces with zeolite seed layers and blocking the membrane defects by post-synthesis modification methods. Emphasis on reproducibility and stability of performance under multicomponent mixtures should remain the focus of fundamental studies. In most cases, reproducible production of novel zeolite films requires parallel development of zeolite membrane microstructures (e.g. preferred orientation, designed interfaces, grain boundary control) and support layers in terms of permeability, chemical properties, and morphology. The importance of the support would only increase with inevitable and ever-ongoing decrease of membrane thickness. Long-term stability studies (e.g. high temperature, pressure, presence of aggressive impurities) are inevitable to evaluate the robustness of zeolite membranes.

References 1 Loeb, S. and Sourirajan, S. (1962). Adv. Chem. Ser. 38: 117–132. 2 Koros, W.J., Coleman, M.R., and Walker, D.R.B. (1992). Annu. Rev. Mater.

Sci. 22: 47. 3 Baker, R.W. (2002). Ind. Eng. Chem. Res. 41 (6): 1393–1411. 4 Spillman, R.W. (1989). Chem. Eng. Prog. 85 (1): 41–62. 5 Huang, A., Bux, H., Steinbach, F., and Caro, J. (2010). Angew. Chem. Int. Ed.

122: 5078–5081. 6 Rubio, C., Zornoza, B., Gorgojo, P. et al. (2014). Curr. Org. Chem. 18:

2351–2363. 7 Li, B., He, G., Jiang, X. et al. (2016). Front. Chem. Sci. Eng. 10: 255–264. 8 Merkel, T.C., Lin, H., Wei, X., and Baker, R. (2010). J. Membr. Sci. 359:

126–139. 9 Rezakazemi, M., Amooghin, A.E., Montazer-Rahmati, M.M. et al. (2014).

Prog. Polym. Sci. 39: 817–861. 10 Boot-Handford, M.E., Abanades, J.C., Anthony, E.J. et al. (2014). Energy

Environ. Sci. 7: 130–189. 11 Mason, J.A., McDonald, T.M., Bae, T.H. et al. (2015). J. Am. Chem. Soc. 137:

4787–4803. 12 Dong, G., Hou, J., Wang, J. et al. (2016). J. Membr. Sci. 520: 860–868.

References

13 Guo, X., Huang, H., Ban, Y. et al. (2015). J. Membr. Sci. 478: 130–139. 14 Bachman, J.E. and Long, J.R. (2016). Energy Environ. Sci. 9: 2031–2036. 15 Khakpay, A., Rahmani, F., Nouranian, S., and Scovazzo, P. (2017). J. Phys.

Chem. C 121: 12308–12320. 16 Xu, L., Rungta, M., Brayden, M.K. et al. (2012). J. Membr. Sci. 423: 314–323. 17 Kim, D., Jeon, M.Y., Stottrup, B.L., and Tsapatsis, M. (2018). Angew. Chem.

Int. Ed. 57: 480–485. 18 Agrawal, K.V., Topuz, B., Pham, T.C. et al. (2015). Adv. Mater. 27:

3243–3249. 19 Ackley, M.W., Rege, S.U., and Saxena, H. (2003). Microporous Mesoporous

Mater. 61: 25–42. 20 DeCoste, J.B. and Peterson, G.W. (2014). Chem. Rev. 114: 5695–5727. 21 Jasuja, H., Peterson, G.W., Decoste, J.B. et al. (2015). Chem. Eng. Sci. 124:

118–124. 22 Tuinier, M., van Sint Annaland, M., Kramer, G.J., and Kuipers, J. (2010). 23 24 25 26 27 28

29 30 31 32 33 34 35 36 37 38 39 40 41 42 43

Chem. Eng. Sci. 65: 114–119. Rochelle, G.T. (2009). Science 325: 1652–1654. Li, H., Wang, K., Sun, Y. et al. (2018). Mater. Today 21: 108–121. Bae, T.H. and Long, J.R. (2013). Energy Environ. Sci. 6: 3565–3569. Sjöberg, E., Barnes, S., Korelskiy, D., and Hedlund, J. (2015). J. Membr. Sci. 486: 132–137. Rodenas, T., Luz, I., Prieto, G. et al. (2015). Nat. Mater. 14: 48–55. Angelini, P., Armstrong, T., Counce, R. et al. (2005). Materials for Separation Technologies: Energy and Emission Reduction Opportunities. United States. Department of Energy. Office of Energy Efficiency and Renewable Energy. Humphrey, J.L. (1997). Separation Process Technology. Canada: McGraw-Hill. Sholl, D.S. and Lively, R.P. (2016). Nature 532: 435–438. Wang, S., Li, X., Wu, H. et al. (2016). Energy Environ. Sci. 9: 1863–1890. Zhu, X., Tian, C., Do-Thanh, C.L., and Dai, S. (2017). ChemSusChem 10: 3304–3316. Ghalei, B., Sakurai, K., Kinoshita, Y. et al. (2017). Nat. Energy 2: 17086. Li, C., Meckler, S.M., Smith, Z.P. et al. (2018). Adv. Mater. 30: 1704953. Baker, R.W. and Low, B.T. (2014). Macromolecules 47: 6999–7013. Park, H.B., Kamcev, J., Robeson, L.M. et al. (2017). Science 356: 1137. Dechnik, J., Gascon, J., Doonan, C.J. et al. (2017). Angew. Chem. Int. Ed. 56: 9292–9310. Dai, Z., Ansaloni, L., and Deng, L. (2016). Green Energy Environ. 1: 102–128. Qi, R. and Henson, M.A. (1998). Sep. Purif. Technol. 13: 209. de Vos, R.M. and Verweij, H. (1998). J. Membr. Sci. 143: 3781. Kim, Y.S., Kusakabe, K., Morooka, S., and Yang, S.M. (2001). Korean J. Chem. Eng. 18: 106. Schafer, R., Noack, M., Kolsch, P. et al. (2001). Sep. Purif. Technol. 25: 3. Peters, T.A., Fontalvo, J., Vorstman, M.A.G. et al. (2005). J. Membr. Sci. 248: 73.

249

250

7 Gas Separations with Zeolite Membranes

44 Caro, J., Noack, M., Kolsch, P., and Schafer, R. (2000). Microporous

Mesoporous Mater. 38: 3. 45 McLeary, E.E., Jansen, J.C., and Kapteijn, F. (2006). Microporous Mesoporous

Mater. 90: 198. 46 Bakker, W.J.W., Kapteijn, F., Poppe, J., and Moulijn, J.A. (1996). J. Membr. 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72

Sci. 117: 57. Kusuki, Y., Shimazaki, H., Tanihara, N. et al. (1997). J. Membr. Sci. 134: 245. Centeno, T.A. and Fuertes, A.B. (1999). J. Membr. Sci. 160: 201. Shiftlett, M.B. and Foley, H.C. (1999). Science 285: 1902. Shiftlett, M.B. and Foley, H.C. (2000). J. Membr. Sci. 179: 275. Centeno, T.A. and Fuertes, A.B. (2000). Carbon 38: 1067. Wang, H., Zhang, L., and Gavalas, G.R. (2000). J. Membr. Sci. 177: 25. Kita, H., Nanbu, K., Yoshino, M. et al. (2001). Polym. Mater. Sci. Eng. 85: 296. Lovallo, M.C., Gouzinis, A., and Tsapatsis, M. (1998). AIChE J. 44: 1903. Hedlund, J., Noack, M., Kolsch, P. et al. (1999). J. Membr. Sci. 159: 263. Coronas, J., Falconer, J.L., and Noble, R.D. (1997). AIChE J. 43: 1797. Burggraaf, A.J., Vroon, Z.A.E.P., Keizer, K., and Verweij, H. (1998). J. Membr. Sci. 144: 77. Casanave, D., Ciavarella, P., Fiaty, K., and Dalmon, J.A. (1999). Chem. Eng. Sci. 54: 2807. Bai, C., Jia, M., Falconer, J.L., and Noble, R.D. (1995). J. Membr. Sci. 105: 79. Giroir-Fendler, A., Peureux, J., Mozzanega, H., and Dalmon, J.A. (1996), in C.H. Bartholomew and G.A. Fuentes (eds.), Stud. Surf. Sci. Catal. 111: 127. Kusakabe, K., Yoneshige, S., Murata, S., and Morooka, S. (1996). J. Membr. Sci. 116: 39. van de Graaf, J.M., van der Bijl, E., Stol, A. et al. (1998). Ind. Eng. Chem. Res. 37: 4071. Aoki, K., Tuan, V.A., Falconer, J.L., and Noble, R.D. (2000). Microporous Mesoporous Mater. 39: 485. Matsufuji, T., Nishiyama, N., Matsukata, M., and Ueyama, K. (2000). J. Membr. Sci. 178: 25. Ciavarella, P., Moueddeb, H., Miachon, S. et al. (2000). Catal. Today 56: 253–264. Bernal, M.P., Coronas, J., Menendez, M., and Santamaria, J. (2002). J. Membr. Sci. 195: 125. Bernal, M.P., Coronas, J., Menendez, M., and Santamaria, J. (2002). Ind. Eng. Chem. Res. 41: 5071. Takata, Y., Tsuru, T., Yoshioka, T., and Asaeda, M. (2002). Microporous Mesoporous Mater. 54: 257. Hanebuth, M., Dittmeyer, R., Mabande, G., and Schwieger, W. (2003). Chem. Ing. Tech. 75: 221. Illgen, U., Schafer, R., Noack, M. et al. (2001). Catal. Commun. 2: 339–345. Girnus, I., Pohl, M.M., Richter-Mendau, J. et al. (1995). Adv. Mater. 7: 711–714. Noack, M., Kolsch, P., Venzke, D. et al. (1994). Microporous Mater. 3: 201.

References

73 Yang, R.T. (1987). Gas separation by Adsorption Processes. Stoneham, MA:

Butterworth Publications. 74 Piera, E., Brenninkmeijer, C.A.M., Santamar𝚤´ a, J., and Coronas, J. (2002).

J. Membr. Sci. 201: 229. 75 van den Bergh, J., Zhu, W., Gascon, J. et al. (2008). J. Membr. Sci. 316: 35. 76 Hong, M., Li, S., Falconer, J.L., and Noble, R.D. (2008). J. Membr. Sci.

307: 277. 77 Hedlund, J., Sterte, J., Anthonis, M. et al. (2002). Microporous Mesoporous

Mater. 52: 179–189. 78 Sandström, L., Sjöberg, E., and Hedlund, J. (2011). J. Membr. Sci. 380: 232. 79 Yin, X., Zhu, G.S., Yang, W. et al. (2005). Adv. Mater. 17: 2006–2010. 80 Ye, P., Korelskiy, D., Grahn, M., and Hedlund, J. (2015). J. Membr. Sci. 487:

135–140. 81 Ye, P., Sjöberg, E., and Hedlund, J. (2014). Microporous Mesoporous Mater.

192: 14. 82 Makertihartha, I.G.B.N., Dharmawijaya, P.T., Zunita, M., and Wenten, I.G.

(2017). AIP Conf. Proc. 1818: 020074. 83 Kusakabe, K., Kuroda, T., Murata, A., and Morooka, S. (1998). J. Membr. Sci.

148 (1): 13. 84 Kusakabe, K., Kuroda, T., Uchino, K. et al. (1999). AIChE J. 45: 1220. 85 Ando, Y., Hirano, Y., Mase, S., and Taguchi, H. (1998). Preparation and

86

87

88 89 90 91 92 93 94 95 96 97

characterization of monolithic and planar elements of zeolite MEMBRANES. Proceedings of the 5th International Conference on Inorganic Membranes, Nagoya, Japan (1998), p. 124. Masuda, T., Hashimoto, K., Kapteijn, F., and Moulijn, J.A. (1998). Selective permeation of CO2 from mixture gas of CO2 and N2 through ZSM-5 zeolite MEMBRANE. Proceedings of the 5th International Conference on Inorganic Membranes, Nagoya, Japan (1998), p. 128. Mase, S., Ando, Y., Hirano, Y., and Taguchi, H. (1998). Planar type module with zeolite (MFI) membranes. Proceedings of the 5th International Conference on Inorganic Membranes, Nagoya, Japan (1998), p. 652. Van den Broeke, L.J.P., Kapteijn, F., and Moulijn, J.A. (1999). Chem. Eng. Sci. 54: 259. Bernal, M.P., Coronas, J., Menendez, M., and Santamaria, J. (2002). J. Membr. Sci. 195 (1): 135. Dunne, J.A., Mariwala, R., Rao, M. et al. (1996). Langmuir 12: 5888. Dunne, J.A., Rao, M., Sircar, S. et al. (1996). Langmuir 12: 5896. Bernal, M.P., Coronas, J., Menendez, M., and Santamaria, J. (2004). AIChE J. 50: 127. Wind, J.D., Paul, D.R., and Koros, W.J. (2004). J. Membr. Sci. 228: 227. Sebastian, V., Kumakiri, I., Bredesen, R., and Menendez, M. (2006). Desalination 199: 464. Tanaka, K., Osada, Y., Kita, H., and Okamoto, K. (1995). J. Polym. Sci. 33: 1907. Suzuki, H., Tanaka, K., Kita, H. et al. (1998). J. Membr. Sci. 146: 31. Bakker, W.J.W., van de Broeke, L.J.P., Kapteijn, F., and Moulijn, J.A. (1997). AIChE J. 43: 2203.

251

252

7 Gas Separations with Zeolite Membranes

98 Zhu, W., Hrabanek, P., Gora, L. et al. (2006). Ind. Eng. Chem. Res. 45:

767–776. 99 Yan, Y., Davis, M.E., and Gavalas, G.R. (1995). Ind. Eng. Chem. Res. 34:

1652–1661. 100 Poshusta, J.C., Noble, R.D., and Falconer, J.L. (1999). J. Membr. Sci. 160: 115. 101 van den Broeke, L.J.P., Bakker, W.J.W., Kapteijn, F., and Moulijn, J.A. (1999).

Chem. Eng. Sci. 54: 245. 102 Gardner, T.Q., Flores, A.I., Noble, R.D., and Falconer, J.L. (2002). AIChE J.

48: 1155. 103 Algieri, C., Bernardo, P., Golemme, G. et al. (2003). J. Membr. Sci. 222: 181. 104 Bonhomme, F., Welk, M.E., and Nenoff, T.M. (2003). Microporous Meso-

porous Mater. 66: 181. 105 Vlugt, T.J.H. (2000). Adsorption and diffusion in zeolites: a computational

study. PhD Thesis, University of Amsterdam. 106 Kusakabe, K., Kuroda, T., Murata, A., and Mooroka, S. (1997). Ind. Eng.

Chem. Res. 36: 649–655. 107 Cui, Y., Kita, H., and Okamoto, K.I. (2004). J. Mater. Chem. 14: 924. 108 Tomita, T., Nakayama, K., and Sakai, H. (2004). Microporous Mesoporous

Mater. 68: 71. 109 Li, S.G., Martinek, J.G.J.G., Falconer, J.L. et al. (2005). Ind. Eng. Chem. Res.

44: 3220–3228. 110 Li, S., Falconer, J.L., and Noble, R.D. (2006). Adv. Mater. 18: 2601–2603. 111 Poshusta, J.C., Noble, R.D., and Falconer, J.L. (2001). J. Membr. Sci. 186:

25–40. 112 Tomita, T., Nakayama, K., and Sakai, H. (2004). Microporous Mesoporous

Mater. 68: 71–75. 113 Sebastian, V., Kumakiri, I., Bredesen, R., and Menendez, M. (2006). Zeolite

114 115 116 117 118 119 120 121 122 123 124 125 126

membrane for CO2 removal: Operating at high pressure. Proceedings of the 9th International Conference on Inorganic Membranes, Lillehammer, Norway (25–29 June 2006), p. 290. Clet, G., Gora, L., Nishiyama, N. et al. (2001). Chem. Commun. 41–42. Weh, K., Noack, M., Sieber, I., and Caro, J. (2002). Microporous Mesoporous Mater. 54: 27. Shin, D.W., Hyun, S.H., Cho, C.H., and Han, M.H. (2005). Microporous Mesoporous Mater. 85: 313–323. Gu, X., Dong, J., and Nenoff, T.M. (2005). Ind. Eng. Chem. Res. 44: 937. Noble, R.D. and Falconer, J.L. (2005). Ind. Eng. Chem. Res. 44: 3220. Maghsoudi, H. and Soltanieh, M. (2014). J. Membr. Sci. 470: 159–165. Lok, B.M., Messina, C.A., Patton, R.L. et al. (1984). J. Am. Chem. Soc. 106: 6092. Carreon, M.L., Li, S., and Carreon, M.A. (2012). Chem. Commun. 48: 2310. Wu, T., Wang, B., Lu, Z.H. et al. (2014). J. Membr. Sci. 471: 338. Wang, B., Hu, N., Wang, H.M. et al. (2015). J. Mater. Chem. A 3: 12205. Zong, Z., Elsaidi, S.K., Thallapally, P.K., and Carreon, M.A. (2017). Ind. Eng. Chem. Res. 56: 4113. Li, S., Zong, Z., Zhou, S.J. et al. (2015). J. Membr. Sci. 487: 141. Zhang, L., Jia, M., and Enze, M. (1997). Stud. Surf. Sci. Catal. 105: 2211.

References

127 Wu, T., Diaz, M.C., Zheng, Y. et al. (2015). J. Membr. Sci. 473: 201. 128 Poshusta, J.C., Tuan, V.A., Falconer, J.L., and Noble, R.D. (1998). Ind. Eng.

Chem. Res. 37: 3924. 129 Huang, Y., Wang, L., Song, Z. et al. (2015). Angew. Chem. Int. Ed. 54: 10843. 130 Zong, Z., Feng, X., Huang, Y. et al. (2016). Microporous Mesoporous Mater.

224: 36–42. 131 Zong, Z. and Carreon, M.A. (2017). J. Membr. Sci. 524: 117. 132 Shah, M.S., Tsapatsis, M., and Siepmann, J.I. (2017). Chem. Rev. 117:

9755–9803. 133 Pullumbi, P. (2016). Gas Separation by Polymer Membranes: Research

134 135

136 137 138 139

140 141 142 143 144 145 146 147 148

149 150

Activity and Industrial Applications. In: Membrane Reactor Engineering: Applications for a Greener Process Industry (ed. A. Basile, M. De Falco, G. Centi and G. Iaquaniello), 259. New York: Wiley. Chen, L., Reiss, P.S., Chong, S.Y. et al. (2014). Nat. Mater. 13: 954–960. Izumi, J. (2003). Waste gas treatment using zeolites in nuclear-related industries. In: Handbook of Zeolite Science and Technology (ed. S.M. Auerbach, K.A. Carrado and P.K. Dutta), 1–17. New York: Marcel Dekker. Kerry, F.G. (2007). Industrial Gas Handbook: Gas Separation and Purification. Boca Raton, FL: CRC Press/Taylor & Francis Group. Ohno, M., Ozaki, O., Sato, H. et al. (1977). J. Nucl. Sci. Technol. 14: 582–589. Nakai, Y., Yoshimizu, H., and Tsujita, Y. (2005). J. Membr. Sci. 256: 72–77. Crawford, P.G. (2013). Zeolite membranes for the separation of krypton and xenon from spent nuclear fuel reprocessing off-gas. MS Thesis, Georgia Institute of Technology, Atlanta, GA. Feng, X., Zong, Z., Elsaidi, S.K. et al. (2016). J. Am. Chem. Soc. 138: 9791–9794. Wu, T., Lucero, J., Zong, Z. et al. (2018). ACS Appl. Nano Mater. 1: 463–470. Amiri, H., Charkhi, A., Moosavian, M.A. et al. (2017). Chem. Pap. 71: 1587–1596. Wu, T., Feng, X., Elsaidi, S.K. et al. (2017). Ind. Eng. Chem. Res. 56: 1682–1686. Kwon, Y.H., Kiang, C., Benjamin, E. et al. (2017). AIChE J. 63: 761–769. Kwon, Y.H., Min, B., Yang, S. et al. (2018). ACS Appl. Mater. Interfaces 10: 6361–6368. Wang, X., Karakiliç, P., Liu, X. et al. (2018). ACS Appl. Mater. Interfaces 10: 33574–33580. Anderson, R., Schweitzer, B., Wu, T. et al. (2018). ACS Appl. Mater. Interfaces 10: 582–592. Kassaee, M.H. (2012). Internal surface modification of zeolite MFI particles and membranes for gas separation. PhD thesis, Georgia Institute of Technology, USA. Masuda, T., Fukumoto, N., Kitamura, M. et al. (2001). Microporous Mesoporous Mater. 48: 239. Park, D.H., Nishiyama, N., Egashira, Y., and Ueyama, K. (2001). Ind. Eng. Chem. Res. 40: 6105.

253

254

7 Gas Separations with Zeolite Membranes

151 Sano, T., Hasegawa, M., Ejiri, S. et al. (1995). Microporous Mater. 5: 179. 152 Falconer, J.L., George, S.M., Ott, A.W. et al. (2000). Modification of zeolite

153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173

or molecular sieve membranes using atomic layer controlled chemical vapor deposition. US Patent 6,043,177. Nomura, M., Yamaguchi, T., and Nakao, S. (1997). Ind. Eng. Chem. Res. 36: 4217. Yan, Y.S., Davis, M.E., and Gavalas, G.R. (1997). J. Membr. Sci. 123: 95. Navajas, A., Mallada, R., Tellez, C. et al. (2006). J. Membr. Sci. 270: 32. Zhang, Y., Avila, A.M., Tokay, B. et al. (2010). J. Membr. Sci. 358: 7. Nair, S., Lai, Z., Nikolakis, V. et al. (2001). Microporous Mesoporous Mater. 48: 219. Matsuda, H., Yanagishita, H., Negishi, H. et al. (2002). J. Membr. Sci. 210: 433. Huang, A., Wang, N., and Caro, J. (2012). Microporous Mesoporous Mater. 164: 294. Kiyozumi, Y., Nagase, T., Hasegawa, Y., and Mizukami, F. (2008). Mater. Lett. 62: 436. Chiu, W.V., Park, I.S., Shqau, K. et al. (2011). J. Membr. Sci. 377: 182–190. Zhu, X., Wang, H., and Lin, Y.S. (2010). Ind. Eng. Chem. Res. 49: 10026–10033. Gu, X., Tang, Z., and Dong, J. (2008). Microporous Mesoporous Mater. 111: 441–448. Nomura, M., Yamaguchi, T., and Nakao, S.I. (2001). J. Membr. Sci. 187: 203–212. Labropoulos, A.I., Athanasekou, C.P., Kakizis, N.K. et al. (2014). Chem. Eng. J. 255: 377–393. Hong, M., Falconer, J.L., and Noble, R.D. (2005). Ind. Eng. Chem. Res. 44: 4035. Tang, Z., Dong, J., and Nenoff, T.M. (2009). Langmuir 25: 4848. Tsapatsis, M. and Gavalas, G. (1994). J. Membr. Sci. 87: 281–296. Rangnekar, N., Mittal, N., Elyassi, B. et al. (2015). Chem. Soc. Rev. 44: 7128. Lindmark, J. and Hedlund, J. (2010). J. Mater. Chem. 20: 2219. Han, A., He, H., Gou, J. et al. (2005). Microporous Mesoporous Mater. 79: 177. Lindmark, J., Hedlund, J., Wirawan, S.K. et al. (2010). J. Membr. Sci. 365: 188. Kosinov, N., Sripathi, V.G.P., and Hensen, E.J.M. (2014). Microporous Mesoporous Mater. 194: 24.

255

8 Pervaporation with Zeolite Membranes 8.1 Introduction Separation of organic–organic mixtures using membrane separation techniques is being investigated extensively owing to its great importance in chemical and petrochemical industries. Pervaporation (PV) allows separations of some mixtures that are difficult to separate by distillation, extraction, and sorption. PV has advantages in separating azeotropes, close-boiling mixtures, and thermally sensitive compounds and removing species present in low concentrations. These advantages make membrane processes or hybrid processes involving membranes economically attractive in many industrial applications. Zeolite membranes have uniform, molecular-sized pores, and they separate molecules based on differences in the molecules’ adsorption and diffusion properties. Zeolite membranes are thus well suited for separating liquid-phase mixtures by PV, and the first commercial application of zeolite membranes has been implemented in dehydrating of organic compounds. Besides that, zeolite membranes have also been used to remove organic compounds from water, separate organic mixtures, and remove water from acid solutions on the laboratory scale. In this chapter, the fundamental aspects of separations by PV through zeolite membranes are presented. The applications of PV in alcohol dehydration, organic–organic separation, and acid separation are reviewed. Finally, the application of membrane reactors for various separations is presented in the last part of this chapter.

8.2 Pervaporation Process Pervaporation is a contraction of the terms permeation and evaporation because the feed is a liquid [1] and the low-pressure vapor exits the membrane on the permeate side as shown by the diagram in Figure 8.1 [1]. Only a fraction of a mixture is vaporized during PV, and lower temperatures than those required in distillation are usually used [2]. In addition, membranes operate continuously without requiring sorbent regeneration, and they are modular, which allows design flexibility. For laboratory-scale studies, the vapor is enriched in the preferentially permeating component and is condensed in a liquid N2 trap for future processing. Meanwhile, the retentate is enriched in the non-preferentially permeating Microporous Materials for Separation Membranes, First Edition. Xiaoqin Zou and Guangshan Zhu. © 2019 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2019 by Wiley-VCH Verlag GmbH & Co. KGaA.

256

8 Pervaporation with Zeolite Membranes

Figure 8.1 Schematic of the pervaporation process. Feed

Membrane Adsorption

Desorption Diffusion

Liquid retentate Vapor permeate

component and can either be used in another process or recycled for further separation [3–8]. The liquid feed mixture to be separated flows along one side of the membrane, while the various feed components are permeating in and through the membrane at different rates. Therefore, the liquid retentate leaving the unit on the same side of the membrane as the feed enters is depleted in the components permeating preferentially. Consequently, the vapor permeate is enriched in the preferentially permeating component. During PV, membranes are sealed into a membrane module. Disk-shaped membranes are usually sealed between two plates in a manner similar to that shown in Figure 8.2a. The feed is typically on the zeolite side of the support so that concentration polarization is minimized. Tubular membranes have nonporous ends on each side of the porous support to allow O-rings to seal the membrane, as shown in Figure 8.2b. Nonporous tubes are welded to the ends of porous stainless-steel supports, and the ends of ceramic and SiC porous supports are glazed to create the nonporous ends. Tubular membrane modules have been operated with the feed on either the inside or outside of the tube, but as with the flat membranes, the feed is typically on the zeolite side of the membrane. To maximize space efficiency, commercial membrane modules house banks of tubular membranes [9]. A more efficient multichannel monolith support (Figure 8.2c) has been used on the lab scale, and it has the potential to be more efficient in commercial applications because its surface area to volume ratio is higher than that of tubes [10, 11]. Monolith membranes are sealed in a module similar to the tubular membrane module. Zeolite membranes have also been prepared on a ceramic hollow fiber [12], which also has a high surface area to volume ratio, but this configuration has not been used for PV. PV fluxes in lab-scale studies are usually measured by weighing the amount of permeate collected in a liquid N2 trap during a given time period. Fluxes are typically reported in units of kg m−2 h−1 because these are the units used in industrial separations. Molar fluxes in mol m−2 h−1 , however, are more useful for comparing permeation of components with different molecular weights, and fluxes are reported in these units in this review. For comparison, a 1.0 kg m−2 h−1 water flux is the same as 55 mol m−2 h−1 .

8.2 Pervaporation Process

Zeolite Feed

Gasket

(a)

Stir bar

Retentate

Porous support

Permeate

Porous section

Feed

(b)

Retentate

Nonporous section

Permeate

Channels Nonporous section

(c)

O-rings

Porous section

Figure 8.2 Cross-sectional sketches of modules for disk-shaped (a) and tubular (b) zeolite membranes and a sketch of a multichannel monolith membrane (c) [2]. Source: Copyright 2004, Elsevier.

Gas or liquid chromatography is usually used offline to measure permeate concentrations, but online mass spectroscopy [13–16] and titration [17] have also been used. Separation performance in PV is usually reported by a separation factor (SP): 𝛼a∕b =

ya ∕yb xa ∕xb

where y and x are permeate and feed compositions, respectively. This definition is analogous to the definition of relative volatility in vapor–liquid equilibrium. SP are used instead of flux ratios because PV feeds usually contain low concentrations of one component in a mixture. Note that the SP is a ratio of ratios and small composition changes can lead to large changes in SP, especially at low feed concentrations and high permeate concentrations.

257

258

8 Pervaporation with Zeolite Membranes

The impurities adsorbed during PV can significantly affect permeation and separation performance [18]. Even if no impurities adsorb, some components used during PV are strongly adsorbed in the zeolite pores that affect permeation of other components if the feed is changed. Nomura et al. [19] and Li et al. [20] reported that fermentation broth fluxes through zeolite membranes decreased after 50–70 hours of PV because minor components in the broths adsorbed in the membranes. The fluxes in both studies returned to the original values after the membranes were calcined at high temperature. Calcination between 550 and 650 K is usually the easiest way in the laboratory to remove unwanted molecules adsorbed in the zeolite pores. Another method for removing impurities is purging the membrane with a weakly adsorbing gas like N2 or He. Funke et al. [18] recommended that laboratory membranes be stored in a vacuum oven or similar inert environment when not in use to reduce adsorption of impurities. These methods of removing undesired adsorbates are less feasible for large-scale separations. Few studies have investigated the effects of zeolite membrane fouling, how to prevent it, or better ways to regenerate a fouled membrane. The use of PV as an alternative or complement to conventional processes, such as distillation for difficult separations, has been suggested for a number of cases [21–29]. Zeolite membranes have been used for PV both industrially and in laboratory studies. These membranes are polycrystalline zeolite layers deposited on porous inorganic supports, and they offer several advantages over polymeric membranes: (i) zeolite membranes do not swell, whereas polymeric membranes do; (ii) zeolites have uniform, molecular-sized pores that cause significant differences in transport rates for some molecules and allow molecular sieving in some cases; (iii) most zeolite structures are more chemically stable than polymeric membranes, allowing separations of strong solvents or low pH mixtures; and (iv) zeolites are stable at high temperatures (as high as 1270 K for some zeolites). Transport through zeolite pores takes place by surface diffusion at low temperatures and activated gaseous diffusion at high temperatures where adsorption is insignificant. Surface diffusion follows an adsorption–diffusion mechanism. That is, molecules first diffuse from the bulk feed to the zeolite surface. Next, they adsorb to the sites on the zeolite surface and in the zeolite pores. After entering the zeolite pores, the molecules diffuse along the surface of the pores by jumping from site to site, driven by the chemical potential gradient within the pore. At the permeate side of the membrane, molecules desorb from the zeolite and diffuse into the bulk permeate through the support pores. For many mixtures, zeolite membranes separate because of adsorption differences, and this is especially true for water/organic separations because of the hydrophobic/hydrophilic nature of zeolites. Diffusion differences are also important in some separations. Most zeolite membrane separations occur in the adsorption and diffusion steps, but concentration polarization and support resistance also affect selectivity. The performance of a PV membrane is characterized by three main features: (i) membrane productivity, (ii) selectivity, and (iii) stability. The productivity of a membrane is given by the permeate flux. The flux through zeolite pores in a

8.2 Pervaporation Process

zeolite membrane is calculated by the following expression: J = −𝜌Dt (q)

dq dz

where q is the coverage or occupancy, which is the amount adsorbed in the zeolite pores, z the position along the transmembrane direction, Dt (q) the coverage-dependent transport diffusivity, and 𝜌 the zeolite density. The selectivity is quantified by parameter 𝛼, defined as in gas separation [30]. Several authors call this selectivity [31–33] and many others separation factor [34–41]. Regarding PV, the physical meaning of 𝛼 is not so clear as in gas separation, as the components of the feed mixture do not permeate independently, but the flux of one affects the flux of the other; sometimes the coupling of fluxes is very strong, and one component can end up blocking the other components in the mixture [42, 43], and the flux coupling, the concentration-dependent diffusion coefficient, and the equilibrium on the membrane surface should be considered. The stability of the membrane consists of its ability to maintain the same permeability and selectivity for long periods of time. Given the extensive time requirements of stability tests, there are but few reports [44, 45] published related to the characterization of PV membranes in transient conditions. The best-known examples of zeolite membranes for dehydration of organic compounds are NaA-type zeolite membranes [46–49], but ZSM-5, MOR, and X- and Y-type zeolites [41, 48] have also been investigated. On the other hand, zeolite membranes, because of their intrinsic characteristics, such as a narrow pore size distribution, showed higher SP in the dehydration of alcohols [48, 50], and hydrophobic zeolite membranes, such as silicalite-1, have been successfully used to remove organic components from water, despite the smaller size of the water molecule with respect to organic molecules. For zeolite membranes, the larger the Si/Al ratio, the lower the water flux through the membrane, which turns in general more hydrophobic. The SP are therefore one to three orders lower in hydrophobic than in dehydration processes. The transport mechanism in PV is a combination of adsorption and diffusion. This applies to both sol–gel [32] and zeolite membranes [50]. Diffusion favors water permeation because the organic components to be separated are in general larger than water molecules. The hydrophobic zeolites overcome this problem as a result of the preferential adsorption of organic compounds. Some organic mixtures also exhibit differences in adsorption and diffusion [51, 52], thus allowing the separation. To sum up, sol–gel membranes give in general higher fluxes and lower SP than zeolite membranes, which provide higher SP because of their intrinsic characteristics, such as the narrow pore size distribution, allowing discriminating the large molecules in favor of the small ones in the mixture. Recently, several groups have developed methylated silica [49] membranes or silica doped with stable metal oxides as zirconia or titania [53–57], providing high flux and selectivity of the same order of magnitude as zeolite membranes.

259

260

8 Pervaporation with Zeolite Membranes

8.3 Alcohol Dehydration The global demand for dehydrated ethanol has been increasing significantly from the last decade, driven by the petrochemical industry. When ethanol is used as a blend for gasoline, the water content must be lower than 2000 ppm, and for the production of ethyl tertiary butyl ether (ETBE) from i-butene and ethanol, the water content must be below 500 ppm. An azeotrope with ethanol content up to 95.6 wt% can be obtained by distillation of an ethanol/water fermentation broth. However, the conventional dehydration of alcohols by azeotropic, extractive, or two-pressure distillation is energy intensive and requires a complex process layout. Thus for economic reasons, ethanol with 92–93 wt% of yield is produced from the optimized membrane processes. Especially for ethanol/water and other water-containing azeotropes, two alternative processes are available: (i) pressure swing adsorption employing type 3A or type 4A LTA molecular sieves and (ii) steam permeation/PV using hydrophilic organic or inorganic (4A molecular sieve) membranes. The application of membrane processes is especially beneficial for systems of low relative volatility [58]. Type A zeolite membranes are nearly ideally suited for organic dehydration because they are highly hydrophilic and their X-ray diffraction (XRD) pore diameter (0.4 nm) is smaller than almost all organic molecules but larger than water. A hydrophilic LTA zeolite layer is extremely selective in the separation of water from organic solutions by steam permeation and PV to give water-free ethanol. These properties allow preferential permeation of water over organic compounds with SP that are often over 1 000 and sometimes higher than 10 000. These high SP are sensitive to permeate concentration because the water concentrations are often higher than 98%. Hydrophilic LTA membranes have been applied in industrial plants for dehydration of the crude ethanol stream [59, 60]. The water flux measured in PV operation for 90 wt% ethanol solution at 75 ∘ C is approximately 7 kg m−2 h−1 . Ethanol scarcely leaks through the membrane, resulting in an SP of water/ethanol ≈ 10 000. Two different tubular ZeoSepA membranes are produced: large-size elements with 16 mm outer diameter and 1 m length for dewatering of bioethanol and small-size elements of 12 mm outer diameter and 0.8 m length for the recovery of i-propanol. The supports are in both cases porous 𝛼-alumina tubes. Recently, by combined SEM, TEM, FTIR, focused ion beam, and XRD characterization, the molecular structure of the Mitsui-BNRI LTA membranes could be solved [61–64]. From April to September 2003, Mitsui-BNRI tested successfully dewatering of bioethanol in the pilot scale by using LTA membranes for vapor permeation in Piracicaba (São Paulo, Brazil). The capacity was 100 l h−1 working with a feed containing 93 wt% ethanol, and the product was 99.65 wt% ethanol. The demonstration plant in Brazil is driven by electricity applying vapor compressor energy recycling. As the next step, Mitsui-BNRI installed a larger steam permeation capacity at Daurala Sugar Works (Uttar Pradesh, India). The capacity of 30 000 l day−1 can be achieved with a LTA membrane area of 30 m2 . Each of the membranes is inserted into a sheath tube (i.e. 19 mm). The feed is evaporated by heating with steam. The feed contains 93 wt% bioethanol, and the product purity

8.3 Alcohol Dehydration

is 99.8 wt% ethanol suitable for blending with gasoline. The ethanol content in the permeate is below 0.1 wt%. The operating pressure and temperature of the membrane are 600 kPa and 130 ∘ C, respectively. Extremely high ethanol fluxes are reported: 11.9, 14.9, 17.6, and 22.4 kg m−2 h−1 at 100, 110, 120, and 130 ∘ C, respectively [65]. The plant is in permanent operation since January 2004 (Figure 8.3). The smaller ZeoSepA membranes are mainly used for the recovery of i-propanol (IPA process) in the Japanese electronics industry using vapor permeation (azeotrope: 87.9 wt% i-propanol, 12.1 wt% water). For a 90%/10% i-propanol/water mixture, the water flux at 75 ∘ C is 3.5–3.7 kg m−2 h−1 with an SP of water/i-propanol ≈ 10 000. Zeolites as crystalline materials are much more stable toward phase transformation and densification compared with X-ray amorphous metal oxide membranes from sol–gel techniques. Because of the high Al content, LTA membranes should be operated at 6.5 < pH < 7.5. Therefore, Mitsui-BNRI developed with the same supports used for the ZeoSepA element, a zeolite FAU membrane with lower Al content (namely, Si/Al between 1.5 and 1.6). This FAU membrane was tested successfully in vapor permeation for dewatering of a spent IPA solution with a starting content of 12 wt% water to 0.46 wt%. The water flux of this membrane was evaluated at 75 ∘ C in a PV experiment with a model feed containing 90 wt% ethanol and gave a water flux of 7–10 kg m−2 h−1 and an SP 𝛼(water/ethanol) ≈ 300. For the LTA membrane tested under the same conditions, a flux of 13.5 kg m−2 h−1 with an SP 𝛼(water/ethanol) ≈ 4800 was found [66]. The described FAU membrane was also successfully tested in alcohol/ether separations. As an example, ethanol is separated from a 5/95 wt% mixture of ethanol/ETBE with fluxes of 2.2 and 4.1 kg m−2 h−1 and selectivities 𝛼(ethanol/ETBE) of 2800 and 1600 at 90 and 110 ∘ C, respectively, by vapor permeation [67]. For a 10/90 wt% mixture of methanol/methyl tert-butyl ether (MTBE) at 100 ∘ C in vapor permeation, methanol fluxes of 10 kg m−2 h−1 with selectivities 𝛼(methanol/MTBE) of 3000 are found [68]. The stability of the hydrophilic FAU membranes could be increased having a Si/Al ≈ 2.2 by using USY12 seed crystals (S. Ikeda, Bussan Nanotech Research Institute, personal information). A further increase of the stability of hydrophilic membranes was achieved by developing a weakly hydrophilic MFI-type membrane with Si/Al ≈ 120 [69, 70]. In PV tests of a 90/10 wt% i-propanol/water model mixture, water fluxes of 3.1 kg m−2 h−1 with 𝛼(water/i-propanol) ≈ 690 were determined at 75 ∘ C. Despite these developments, in recent installations in Europe, the adsorptive drying of ethanol was implemented instead of the membrane technology. Here, ethanol is purified by distillation, which is coupled with a zeolite adsorption section. By the so-called DELTA-T technology (www.deltatcorp.com and http://www.sasol.com), ethanol is purified to less than 100 ppm water. The preference of the molecular sieve adsorption process with zeolites for dewatering of ethanol may be because the azeotrope ethanol/water contains only a relative low amount of water (4.4 wt%) so that the heat management of the cyclic adsorption processes can be managed. Since the azeotrope i-propanol/water contains more water (12.1 wt%), steam permeation using hydrophilic organic or inorganic membranes has a higher chance for realization in comparison with the ethanol/water mixture.

261

262

8 Pervaporation with Zeolite Membranes

Vapor permeation

Poor produced from ligno cellulose

Pervaporation

Cooler

Reflux

Product (EtOH 99.6 wt%) Heater

Reboiler

Stripping column

Pilot plant process flow

(a)

(b)

Figure 8.3 Plant for the dehydration of bioethanol by steam permeation using LTA membranes at Daurala Sugar Works (Uttar Pradesh, India) with a capacity of 30 000 l day−1 [65] (S. Ikeda, Bussan Nanotech Research Institute, personal information). (a) The flow sheet. (b) A view of the plant.

In Europe, the company inocermic GmbH, which is a 100% subsidiary of HITK, Hermsdorf, Germany, produces NaA membranes for dewatering processes, especially for bioethanol, by PV and vapor permeation. The membrane layer was inside a α-Al2 O3 4-channel support thus protected against mechanical

8.3 Alcohol Dehydration

damage. Organic solutions can be dried to water levels as low as 0.1 wt% by PV/steam permeation. For a feed with 90 wt% ethanol at 100–120 ∘ C, water fluxes of 7–12 kg m−2 h−1 with SP H2 O/ethanol ≈ 1000 are found [71]. The dewatering behavior of these semi-industrially produced NaA membranes was tested by PV with bioethanol feed stocks from real fermentation processes. The impurities in the bioethanol from grain fermentation or wine production lowered the specific permeate flux by only 10–15% as compared with synthetic ethanol/water mixtures [72]. All bioethanol samples were dewatered to greater than 99.5 wt% ethanol [71]. Hydrophilic membranes such as LTA and FAU can be used for water extraction from aqueous ethanol or i-propanol mixtures to obtain concentrated alcohol [73]. An opposite target can be the continuous removal of ethanol from the fermentation broth since the fermentation process stops at ethanol concentrations at 15 wt%. Hydrophobic membranes to solve this problem, such as the MFI type, are under development. The ethanol fluxes are between 0.2 and 1.4 kg m−2 h−1 with SP between 30 and 70. A typical result is a flux of approximately 1.0 kg m−2 h−1 of 85 wt% ethanol from a feed with 8 wt% ethanol (Figure 8.4), which corresponds to an SP of 57 [74]. The relative low ethanol fluxes are due to the test conditions with real fermentation broths. By optimizing the support structure, reducing the membrane thickness, and increasing the Si/Al ratio of the MFI membrane, it should be possible to increase the ethanol fluxes. By changing the top layer of the support from a 5 nm γ-Al2 O3 nanofiltration layer to a 250 nm α-Al2 O3 microfiltration layer, the ethanol-enriched flux (between 70 and 80 wt% ethanol) could be increased from 0.8 to 1.4 kg m−2 h−1 (Voigt, I. and Richter, H., Hermsdorf Institute for Technical Ceramics, HITK, personal communication). In a more recent year, we developed an economic approach for synthesizing an aluminum-rich MFI-type zeolite membrane with an ultimate goal of alcohol dehydration [75]. The economy of this approach reflects on no need of additional organic templates and mild synthesis conditions. In the synthesis, calcined MFI seeds were introduced as the nucleation sites for ZSM-5 membrane on the 100 90 80

4

70 60

3

50 2

40

Permeate flux Ethanol in permeate

30 20

1

Ethanol in permeate in wt%

Permeate flux in (kg m–2 h–1)

5

10 0

0

2

4

6

8

0 10

Ethanol in feed in wt%

Figure 8.4 Pervaporation test for water/ethanol at 40 ∘ C [74]. Source: Copyright 2006, Elsevier.

263

264

8 Pervaporation with Zeolite Membranes

(a)

(b)

100 μm NONE

SEI

5.0 kV

× 15,000

1 μm

WD 6.4 mm

(d)

(c)

1 μm MFI membrane

4 μm

Net support NONE

SEI

3.0 kV

× 13,000

1 μm

WD 6.1 mm

NONE

SEI

5.0 kV

× 8,000

1 μm

WD 6.0 mm

Figure 8.5 SEM images of (a) the seeded support, (b) optical picture, and (c) top and (d) side views of grown ZSM-5 membranes [75]. Source: RSC.

top of seed layer (Figure 8.5). The membrane crystallization proceeded at low temperature (100 ∘ C) and ambient pressure, the consequence of which was forming a continuous and crack-free membrane. Besides, as-prepared membranes exhibited hydrophilic properties with selective sorption toward water rather than ethanol and isopropanol. This hydrophilicity was originated by an insertion of much aluminum into the zeolite framework. Applying these membranes to the dehydration of ethanol and isopropanol, impressive SP of approximately 90 000 for EtOH-H2 O and IPA-H2 O as well as high water flux (3.43–6.88 kg m−2 h−1 ) were achieved.

8.4 Organic–Organic Separation PV membranes are being used in the separation of four major categories of organic–organic mixtures: polar/nonpolar mixtures, aromatic/alicyclic mixtures, aromatic/aliphatic mixtures, and aromatic isomers. These membranes are made of materials that are organic as well as inorganic in nature. Some nonaqueous organic mixtures exhibit adsorptions with zeolite membranes. Methanol (1.7 Debye dipole moment) is more polar than MTBE (1.4 Debye) and benzene, which does not have a permanent dipole. Thus, NaY and NaX zeolite membranes that have high aluminum contents separate these mixtures better than a silicalite-1 membrane because of their localized electrostatic poles.

8.4 Organic–Organic Separation

10 000

0.3 1 000

α

Q (kg m–2 h–1)

0.4

0.2

0.1 0.0

0.2 0.4 0.6 0.8 Mole fraction of MeOH in feed

100 1.0

Figure 8.6 Flux Q (◽) and separation factor 𝛼 (○) as a function of methanol feed concentration for pervaporation of methanol/MTBE mixtures through NaY zeolite membrane at 50 ∘ C [76]. Source: Copyright 1997, RSC.

Kita et al. showed that methanol/MTBE total fluxes through a NaY zeolite membrane at 323 K were nearly independent of feed concentration, whereas the SP decreased from 10 000 to 800 as the methanol feed concentration increased from 5 to 75% (Figure 8.6) [76]. Hexane isomers and benzene/p-xylene mixtures have also been separated by PV through zeolite membranes. Matsufuji et al. [77] observed SP as high as 270 for n-hexane/2,3-DMB separations using a ZSM-5 membrane, but the fluxes were about two orders of magnitude lower than those obtained by Flanders et al. [78] for the n-hexane/2,2-DMB separations. The membrane prepared by Matsufuji et al. was on α-Al2 O3 support, and the continuous zeolite layer was in the support pores instead of on the surface of the support [79]. In addition, the TIPB flux was less than the detection limit (4 × 10−6 mol m−2 h−1 ). Thus, the low n-hexane/2,3-DMB flux was attributed to flow only through zeolite pores and reduced effective permeable area because of the porosity of the support (0.4). The membrane prepared by Flanders et al. was a ZSM-5 zeolite layer on a stainless-steel support. Sommer et al. obtained good fits for the hexane isomer single-component and mixture fluxes and selectivities through this membrane [80]. The data were fitted using adsorption isotherms and diffusivities from the literature, a Maxwell–Stefan diffusion model for flow through zeolite pores, and a Knudsen diffusion model for flow through non-zeolite pores. They determined that 2,2-DMB transported almost exclusively by activated Knudsen diffusion through non-zeolite pores. Nishiyama et al. [81, 82] used a method similar to that of Matsufuji et al. [77] to prepare mordenite and ferrierite zeolite membranes that had SP of higher than 100 for benzene/p-xylene mixtures. The fluxes, however, were less than 1.0 mol m−2 h−1 . These membranes also had their continuous zeolite layers inside the support pores and TIPB fluxes below the detection limit, and this may explain their low fluxes. Furthermore, the FER membrane had a lower flux than the MOR membrane, and this was probably because it has smaller zeolite pores.

265

266

8 Pervaporation with Zeolite Membranes

Although zeolite membranes have been used to separate vapor mixtures of xylene isomers with p-/o-xylene SP as high as 500 in a silicalite-1 membrane, few PV studies of xylene isomer separations have been reported for zeolite membranes [79, 83–85]. Lai et al. showed that the b-oriented MFI films made by the use of trimer TPA show the highest selectivity and fluxes approaching those of the alumina support [79]. The p-xylene permeance is about twice as high as the flux through the composite MFI membranes. The high SP and flux are retained throughout the temperature range tested, in contrast to previous studies that show flux maxima at intermediate temperatures. For p-/m-xylene and p-/o-xylene mixtures, Matsufuji et al. [86] and Nishiyama et al. [87] observed that p-xylene preferentially permeated with PV SP between 2 and 16, but the fluxes were below 0.01 mol m−2 h−1 . Wegner et al. [88] obtained SP that were approximately 1.0. Recently, Yuan et al. used a template-free synthesis method to prepare a silicalite-1 membrane and obtained a p-/o-xylene SP of 40 at 323 K; the total flux was 1.3 mol m−2 h−1 [89]. In recent years, 2D zeolites of atomic thickness are emerging as nano-building blocks to develop high-performance separation membranes that feature unique nanopores and nanochannels to allow ultrafast and highly selective molecular transport [90]. The Tsapatsis group realized high-purity exfoliated MFI nanosheet (1.5 unit cell thick, 300 nm lateral size) suspensions by developing a polymer-melt-compounding exfoliation technique combined with a density gradient centrifugation purification method (Figure 8.7a–c) [91, 92]. The 2D MFI framework contains 10-membered ring pores that run inside and through the layers; these are of extreme importance because they provide shape-selective diffusion inside the layers while reducing the characteristic diffusion length to cal. 1.0 nm. Filtration of suspensions through the porous supports is used to deposit oriented nanosheet multilayers with thickness of sub-100 nm. However, the filtered MFI nanosheet film contains nonselective gaps and thus is unable to exhibit separation selectivity. These gaps were reduced by mild secondary growth of the nanosheet film, and the established orientation is preserved by using appropriate structure-directing agents such as tetrapropylammonium cations. The thickness of the current state-of-the-art 2D zeolite membranes is around 100 nm (Figure 8.7d–e), which is 10 times thinner than typical three-dimensional (3D) zeolite membranes. The intergrown MFI nanosheets were preferentially b-oriented, as confirmed by XRD, which indicates that the 5.5 Å straight pore MFI channels were normal to the membrane surface. As a result, these 2D MFI membranes exhibited a high flux and selectivity for xylene separation. The p-xylene permeance was 1.3–3.6 × 10−7 mol m−2 s−1 Pa−1 , and the SP was 185–1050, with an equimolar feed of p- and o-xylene at 150 ∘ C.

8.5 Acid Separation A few studies have investigated PV with acidic feeds, which are industrially interesting because finding a suitable stable material for these separations is often difficult and some aqueous acids have azeotropes. Hydrophilic zeolites,

8.5 Acid Separation

(b)

F

h (nm)

(a)

(d)

(c)

4 2 0 0

200 x (nm)

400

(e)

100 nm thick

500 nm

100 nm

Figure 8.7 Zeolite nanosheet membrane: (a) TEM images of b-oriented MFI nanosheets, (b) atomic force microscopy (AFM) topographical images of MFI nanosheets, and (c) relaxed surface structures of MFI nanosheets obtained by structure optimization of 1.5 unit cell thick MFI structures with Car–Parrinello molecular dynamics: Si, O, and H atoms are yellow, red, and white, respectively; viewed along the c (upper) and b (bottom) axes; SEM top (d) and cross-sectional (e) images of 100 nm thick MFI nanosheet membrane, made by secondary growth of deposited nanosheets [91, 92]. Source: Copyright 2011, The American Association for the Advancement of Science and 2013, Wiley-VCH.

in general, are not stable in low pH environments because acid leaches Al from the framework. In addition to steaming [93], acid leaching is commonly used to dealuminate zeolite structures [94]. Zeolite membranes used for low pH PV, therefore, need to have relatively high Si/Al ratios so that the framework is not destroyed when Al is removed. Stainless steel supports are usually used for these applications because Al2 O3 supports are susceptible to degradation by acids. A Ge-ZSM-5 membrane removed acetic acid from a 5 wt% acetic acid/H2 O mixture at 363 K with 𝛼 = 14 and a 16.8 mol m−2 h−1 flux [95]. The Ge-ZSM-5 membrane had higher SP than silicalite-1 membranes in the same study. Sano et al. reported acetic acid/H2 O SP as high as 2.7, but the fluxes were approximately five times higher than those through the Ge-ZSM-5 membrane [96]. In addition, they also showed that the permeate concentration during PV of 15 vol% acetic acid through a silicalite-1 membrane remained nearly unchanged after almost one month of operation. Li et al. used a more hydrophilic ZSM-5 zeolite membrane (Si/Al = 50) to dehydrate 50 wt% H2 O/acetic acid mixtures [97]. To further increase hydrophilicity, the membrane was treated with an alkali solution to extract Si from the zeolite

267

268

8 Pervaporation with Zeolite Membranes

framework to create structure defects containing silanol groups. After this treatment, the flux and SP were 44 mol m−2 h−1 and 380, respectively. SSZ-13 zeolite membranes (CHA structure, Si/Al = 14) was employed in dehydration of HNO3 , a significantly stronger acid than acetic acid, and exhibits an azeotrope at approximately 69 wt% HNO3 [17]. The H2 O/HNO3 SP were approximately 3.3 using a 69 wt% HNO3 feed at 298 K, and the fluxes were about 2.8 mol m−2 h−1 . About 5% of the initial Al was removed from SSZ-13 powder during three days of exposure to 69 wt% HNO3 , but a membrane continued to separate with a flux of 8.5 mol m−2 h−1 and 𝛼 = 2.6, even after 13 days of PV with approximately 69 wt% HNO3 feed.

8.6 Membrane Reactors for Various Separations The concept of combining membranes and reactors is being explored in various configurations, which can be classified in three groups, related to the role of the membrane in the process [98]. A (catalytic) membrane reactor combines a chemical reaction with an in situ separation in one unit [99–101]. Catalytic membrane reactors can be classified according to their function into (i) extractor-, (ii) distributor-, and (iii) contactor-type reactors [102]. The extractor mode especially requires a high separation selectivity, which can be provided by zeolitic molecular sieve membranes. In addition to their molecular sieving function, zeolite membranes are relatively stable at the temperatures of most chemical reactions, and they are stable against solvents in comparison with organic polymer membranes. The use of a zeolite membrane as a distributor for a reactant has been attempted for the partial oxidation of alkanes such as propane to propene [103] or n-butane to maleic anhydride [104]. Limited performances were obtained because the back-diffusion of the alkane is hardly controllable with this type of microporous membrane [105]. Zeolite membranes can be used as active contactors if they are catalytically active. Immobilizing transition metal ions in zeolites by ion exchange or by incorporation into the lattice leads to stable isolated and well-defined redox-active catalytic sites [106]. Typical applications of zeolite membranes in reactors include (i) conversion enhancement by equilibrium displacement (product removal), by removal of catalyst poisons/inhibitors, or by improved diffusion of the reactant to the catalytic sites and (ii) selectivity enhancement either by control of residence time or by control of reactant traffic. A large number of examples are reported and discussed in the literatures [107–109]. 8.6.1

Water Separation

There are increasing activities to develop hydrophilic water-selective membranes that are more stable than LTA. Zeolite SOD is hydrophilic like LTA, but with a higher chemical and thermal stability thanks to its higher framework density [110]. The small pore size of the six-membered ring in H-SOD (pore opening, 2.65 Å) allows selective permeation of small molecules like H2 O

8.6 Membrane Reactors for Various Separations

(kinetic diameters, 2.6 Å). Therefore, due to their hydrophilicity and molecular sieving properties, SOD membranes can accomplish the removal of steam under harsh reaction conditions [111–113] and could be used in the synthesis of methanol (MeOH; Equation 8.1), dimethyl ether (DME) (Equation 8.2), and dimethyl carbonate (DMC) (Equation 8.3) in catalytic membrane reactors with carbon dioxide, hydrogen, or methanol as reagents [114]. CO2 + 3H2 ⇆ CH3 OH + H2 O

(8.1)

2CO2 + 6H2 ⇆ CH3 OCH3 + 3H2 O

(8.2)

CO2 + 2MeOH ⇆ CH3 OCOOCH3 + H2 O

(8.3)

The separation performances of SOD membranes for equimolar mixtures of steam with H2 , CO2 , MeOH, DME, or DMC were evaluated in the temperature range from 125 to 200 ∘ C. The mixture SP for steam from DME and DMC through the SOD membrane were found to be higher than 200 and 1000, respectively. To increase the hydrothermal stability of SOD membranes, SOD can be doped with sulfur [115]. MFI membranes modified by catalytic cracking deposition (CCD) have been successfully evaluated in catalytic high-temperature water–gas shift (WGS) reactors [116]. Despite the relatively modest hydrogen selectivity, the CO conversion could be increased over the thermodynamic limit. MFI membranes modified by CCD coking have also been successfully evaluated in a low-temperature WGS reaction to overcome the equilibrium constraints (Figure 8.8) [117]. Since esterification is equilibrium-limited reaction, the removal of water (i.e. side product in esterification) by separation directs equilibrium toward ester production and increases the conversion. Hydrophilic organic polymer membranes were used for water separation in low-temperature esterification. However, in the case of high-temperature esterification, hydrophilic inorganic membranes with high stability against strong acids have to be used. MFI-type zeolite membranes are suitable candidates to fulfil these demands. The benefits of a membrane-assisted esterification were shown for the reaction of n-propanol with propionic acid using a MFI-type ZSM-5 membrane with a Si/Al molar ratio of 96 [59]. In Figure 8.9, the hatched area indicates the optimum working CO + H2O

CO2

LT-WGS reaction

Modified MFI zeolite membrane Sweep gas

Sweep gas + H2

Figure 8.8 Schematic of a membrane-supported water–gas shift reaction: in an extractor-type membrane reactor, hydrogen is selectively removed through an MFI membrane modified by coking [117]. Source: Copyright 2012, Elsevier.

269

8 Pervaporation with Zeolite Membranes

1.0 0.92

0.9 0.8 Relative concentration

270

0.7 0.6 0.52

0.5 0.4 Ester without membrane

0.3

Ester with membrane Water with membrane

0.2 0.10

0.1

0.05

0.0 0

20

40

60

80 100 120 Time (min)

140

160

180

200

Figure 8.9 Conversion enhancement by water removal via a hydrophilic ZSM-5 membrane with Si/Al ratio of 96 in a membrane reactor for the esterification of propionic acid with i-propanol to yield the corresponding ester and water at 70 ∘ C [59]. Source: Copyright 2005, Springer Nature.

range of the membrane reactor. The water content should be reduced to values between 5 and 10%. A reduction of the water content to 5% led to an increase of the ester yield from 52 to 92%. Further reduction of the water content in the esterification mixture was not recommended since this would require very large membrane area but contributed only slight improvement to the ester yield. Tanaka et al. proposed zeolite T membranes for the PV-assisted esterification of acetic acid or lactic acid with ethanol, catalyzed by Amberlyst 15 in a batch reactor at 343 K [118, 119]. Such membrane reactor displayed good selectivity in water removal, and nearly full conversion was observed within eight hours of operation. De la Iglesia et al. employed mordenite and LTA zeolite membranes in a continuous membrane reactor packed with Amberlyst catalyst for esterification of ethanol and acetic acid [120]. Mordenite membrane offered conversion of nearly 90% for five days of experiment attributed to its stability under the conditions applied, while zeolite LTA membrane with relatively low stability caused tremendous conversion loss. The synthesis of multilayer bifunctional layers also has been proposed where a two-layered MOR (acid-resistant separative layer)/H-ZSM-5(catalyst) membrane has been prepared and tested for an esterification process [121]. Besides esterification, the Amberlyst coupled mordenite or NaA zeolite membranes were also implemented in the gas-phase synthesis of MTBE from tert-butanol (TBA) and methanol (MeOH) [122]. The selective permeation of water across the zeolite membranes was attributed to their high polarities resulting in preferential adsorption and pore blocking for the other components. Experiments for the water removal in the membrane

8.6 Membrane Reactors for Various Separations

reactor showed 6.7% absolute increase in the MTBE yield compared with the conventional configuration without water removal. NaA zeolite membranes with γ-Al2 O3 as the catalyst were used as water extractors in the methanol dehydration to DME in a conventional fixed-bed reactor [123]. Separation of water resulted in more than 20% absolute improvement in methanol conversion. Apart from overcoming equilibrium-limited reactions, hydrophilic zeolite membranes were employed in reactions where water removal was crucial to minimize catalyst deactivation. NaA zeolite membranes coupled with sulfonated ion-exchange resin catalysts were employed for water removal in liquid-phase etherification of n-pentanol to di-n-pentyl ether (DNPE) [124]. The excellent dehydration performance of the membrane showed twofold increase in n-pentanol conversion. Zeolite membranes might be used for the direct DME synthesis from a mixture of syngas and carbon dioxide [125]. However, in the multistep reaction via methanol, water is being released [126, 127]. DME yield is highly dependent on the membrane water permselectivity. For membranes with low water permselectivity, the DME yield (7.0%) was lower than that obtained from a conventional reactor (14.8%). Further studies showed that approximately 30% DME yield could be obtained in the zeolite membrane reactor at high recirculation factors of the sweep gas stream due to reduction in the methanol loss across the membrane. The advantage of hydrophilic membrane reactors for selective removal of water has been considered for Fischer–Tropsch synthesis: (i) improvement of the catalyst lifetime, (ii) increasing reactor productivity, and (iii) displacement of the WGS equilibrium in favor of CO. [128–130]. Hydroxy sodalite zeolite membranes with a layer thickness of 2 μm showed great potential for the in situ water removal due to their extraordinary separation and permeation performance [111]. Configuration that combines a fixed-bed water-permselective membrane reactor equipped with H-SOD zeolite membrane coated on α-Al2 O3 support and a fluidized-bed hydrogen-permselective membrane reactor equipped with Pd—Ag membrane has been proposed for high-temperature Fischer–Tropsch process to improve production of gasoline from syngas and reduced CO2 yield [131, 132]. 8.6.2

Hydrogen Separation in Dehydrogenation Reactions

Since dehydrogenation of alkanes to olefins is a strongly endothermic and thermodynamically equilibrium-limited reaction [133], removing products from the reaction medium will shift the reaction equilibrium toward the product side, thus increasing the overall conversion. Selective H2 (product) removal by hydrogen-selective zeolite membranes and its effect on the reaction efficiency have been studied in dehydrogenation of alkanes. Dalmon and coworkers investigated dehydrogenation of i-butane in Pt—In—Ge/MFI [134], Pt—In/silicalite [135], or commercial Pt—Sn/γ-Al2 O3 [136] membrane reactors. MFI zeolite membranes were used in catalytic dehydrogenations of i-butane although both H2 and i-butane can pass the 0.55 nm pores due to their kinetic diameters (0.29 and 0.50 nm, respectively) since the interplay of mixture adsorption and mixture diffusion results in H2 selectivity at high temperatures [135, 137]. The MFI

271

272

8 Pervaporation with Zeolite Membranes

membrane reactor showed up to four times higher dehydrogenation yield than the conventional reactor. The performance of the reactor was limited by the catalyst activity under countercurrent sweep flow conditions because in most cases the catalyst activity could not catch up with the high membrane permeability. In addition, the reactor performance was controlled by the membrane permeation efficiency, and insufficient selectivity will lead to reactants (i.e. i-butane) loss. A study conducted by the Caro group showed that when hydrogen was removed from the shell side of the membrane reactor through the sweep gas, the conversion of i-butane increased by approximately 15% [138]. Removal of the hydrogen leads to hydrogen-depleted conditions as compared with the conventional fixed bed. This has two positive effects: (i) the conversion of i-butane is increased, and (ii) the selectivity to i-butene is increased since hydrogenolysis is suppressed. As a result, at the beginning of the reaction, the yield for i-butene in the membrane reactor is higher by approximately 1/3 than in the conventional fixed bed. However, because of the hydrogen removal, coking is promoted, and after approximately two hours time on stream, the olefin yield of the membrane reactor drops below that of the classical packed bed [139]. After an oxidative regeneration, however, the activity and selectivity of the membrane reactor (membrane and catalyst) are restored completely. Van den Bergh et al. revealed the benefit of using small-pore zeolite DD3R membranes coupled with Cr2 O3 /Al2 O3 catalysts in a membrane reactor for the dehydrogenation of i-butane [140]. The DD3R membrane effectively separated H2 and i-butane by molecular sieving effects in which H2 permeated while i-butane retained due to its bigger size. Thus, the membrane exhibited excellent H2 /i-butane selectivity of over 500 and 50% increase in yield due to the effective removal of H2 from the reaction zone at 773 K. Kusakabe and coworkers evaluated the performance of FAU-type zeolite membranes coupled with a Pt/Al2 O3 catalyst in catalytic dehydrogenation of cyclohexane [141, 142]. According to their findings, such membrane reactor could simultaneously remove hydrogen and benzene from the reaction zone. They concluded that in terms of industrial practice high membrane permeability accompanied with reasonable selectivity was more favorable option than a high selectivity at the expense of low permeability. Defect-free silicalite-1 zeolite membranes were used for the catalytic dehydrogenation of ethylbenzene to styrene in membrane reactors packed with Fe2 O3 and illustrated superior performance (74.8% conversion) over a conventional reactor (67.5% conversion) at 610 ∘ C due to an instant extraction of the produced H2 across the membrane [143]. Increasing sweep gas to reactant feed ratio from 0.5 to 2 contributed approximately 4% increase in the ethylbenzene conversion since higher ratios induced larger driving force for H2 permeation. Alternatively, small-pore zeolite SOD membranes have been employed in selective removal of H2 in catalytic dehydrogenation of ethylbenzene to styrene [144]. The benefit of the membrane reactor for effective H2 separation and removal was once again proven in which the absolute ethylbenzene conversion was 3.45% and the yield was 8.99% higher than the conventional reactor. Natural mordenite membranes packed with Pt/Al2 O3 catalysts has been used in

8.6 Membrane Reactors for Various Separations

dehydrogenation reaction of ethane [145], and the membrane was able to shift the reaction equilibrium due to its hydrogen-selective properties. Increasing the permeation area to reactor volume ratio (A/V ratio) from 0.04 to 0.16 m−1 contributed to an additional separation of the H2 in the reaction zone, which consequently increased the reaction rate and ethane conversion. 8.6.3

Hydrogen Separation in Water–Gas Shift Reaction

WGS reaction represents another equilibrium-limited reaction for which the application of extractor-type membrane reactors has been reported. The adoption of a H2 -selective membrane in WGS reaction can break the equilibrium constrains and facilitate the CO conversion, thus intensifying the process and resulting in economically beneficial application. Tang et al. [146–148] modified the pore diameters of MFI zeolite membranes by in situ CCD of silane precursors to reduce the effective pore size to below 0.36 nm that excluded permeation of CO2 into the pore channels, giving rise to a H2 /CO2 permselectivity of 68.3 and equimolar mixture SP of nearly 38 at 550 ∘ C [138]. Apart from that, Kim and coworkers applied cerium-doped ferrite catalysts on a modified MFI membrane reactor in WGS reaction that resulted more than 99.5% CO conversion at 550 ∘ C, 50 atm, and H2 O/CO ratio around 5.0 [116]. In another study, Lin et al. evaluated the performance of ZSM-5/silicalite bilayer membranes packed with Fe—Cr—Cu catalysts in WGS reaction, and over 95% of CO conversion and more than 90% H2 recovery were reported under the studied optimal conditions [149]. For low-temperature WGS reactions, Zhang et al. illustrated the benefit of a H2 -permselective CuO/ZnO/Al2 O3 /MFI membrane reactor post-modified by CCD of methyldiethoxysilane over the conventional packed-bed reactor [117]. A CO conversion exceeding equilibrium was obtained due to an enhanced permeation of H2 . In a following study, Zhang et al. proposed the idea of using steam as sweep gas instead of inert gas (i.e. N2 ) in order to avoid subsequent separation to obtain pure hydrogen [150]. Figure 8.10 illustrates the scheme of this modified hollow fiber MFI zeolite membrane reactor configuration where the steam is applied in a counter-diffusion toward the reactions side in order to remove H2 . Sweeping with pure steam has the advantage of enhancing the conversion and allowing direct acquisition of the pure H2 . 8.6.4

Hydrogen Separation in Syngas Production

Liu et al. employed a combination of catalytic composite zeolite membranes with either La2 NiO4 /NaA or La2 NiO4 /NaY supported on γ-Al2 O3 /α-Al2 O3 and NiO-La2 O3 /γ-Al2 O3 as catalyst for selective product permeation in CO2 reforming of methane for syngas production [151, 152]. The purpose for such mixed configuration was to improve the separation efficiency of the as-prepared inert zeolite membranes for the H2 /CH4 mixture and improving diffusion of methane during reforming. The permselective permeation of CO and H2 across the membrane offered excellent CH4 and CO2 conversion of 73.6 and 82.4 mol%

273

274

8 Pervaporation with Zeolite Membranes

Pure H2 CO +

H2O

H2 + CO2

Steam

H2O + H2 HT-WGS catalyst

Dead end

Modified hollow fiber MFI zeolite membrane

Figure 8.10 Schematic illustration of a WGS membrane reactor with modified hollow fiber MFI zeolite membrane swept by steam applied in a counter-diffusion toward the reaction side [150]. Source: John Wiley & Sons.

at 700 ∘ C. In addition, coke deposition and catalyst deactivation were remarkably reduced in such membrane reactor than the conventional reactor. 8.6.5

Methanol Separation in Hydrogenation Reaction

Technology that capable of reducing CO2 concentration in the atmosphere and thus of mitigating the greenhouse effect remains as a hot topic now and in the future. The conversion of CO2 into methanol was studied by Gallucci et al. in a zeolite NaA coupled CuO-ZnO/Al2 O3 fixed-bed catalyst membrane reactor [153]. The membrane reactor displayed higher CO2 conversion and selectivity than the traditional reactor mainly attributed to the selective removal of CH3 OH via the zeolite membrane. Moreover, the authors stressed the positive aspect of the reduced energy demand since the membrane reactor was able to reach the CO2 conversions of the traditional reactor at milder conditions (i.e. at 225 ∘ C vs. 265 ∘ C in a conventional reactor). 8.6.6

Metathesis of Propene

The selective product removal in the equilibrium-limited metathesis of propene to ethene and 2-butene and simultaneously occurring geometrical isomerization of cis-2-butene into trans-2-butene was evaluated on a membrane reactor equipped with a silicalite-1 zeolite membrane supported on stainless steel and 16.4 wt% Re2 O7 /γ-Al2 O3 as catalyst [154, 155]. Experimental tests on the membrane reactor showed 13% absolute improvement of the propene conversion and 32% increase in the trans-2-butene/cis-2-butene ratio due to the preferential permeation of trans-2-butene.

8.6 Membrane Reactors for Various Separations

8.6.7

Separation of Isomers in Isomerization Reaction

Xylene isomers with a typical composition of 18% p-xylene, 40% m-xylene, 22% o-xylene, and 20% ethylbenzene are generally produced from petroleum reformate streams [156]. Separation is required to obtain pure isomers. The close boiling points of p- and m-xylene have made their separation by distillation not possible; zeolite membrane reactor with molecular sieve properties evolved as an excellent candidate for such application. Membrane reactors coupling xylene isomerization and selective simultaneous recovery have gained increasing importance due to an increasing demand of xylenes. Isomerization of xylene is a very attractive application for MFI zeolite membranes [157–159]. The commercialization of a new zeolite-based membrane is being planned by NGK Insulators for p-xylene production from other xylene isomers. p-Xylene molecules, which are smaller than those of m- or o-xylene, are sieved from the mixture using the membrane, thus cutting significantly the production costs [160]. H-ZSM-5 membranes can also play the role as a catalyst in the isomerization reactor [158]. The development of silicalite-1-coated ZSM-5 catalysts was proposed to increase the p-xylene selectivity by inhibiting side reactions occurring on the external surface [159]. Separation of xylenes with MFI zeolite membranes demonstrating high permselectivity for p-xylene over other isomers was attributed mainly to the preferential permeation of p-xylene since the zeolite pores exposed sterical hindrance for the permeation of the bulkier m- and o-xylene isomer molecules [85, 89, 161–163]. Lai et al. achieved a dramatic improvement of the p-/o-xylene separation by b-oriented silicalite-1 membranes [79]. Figure 8.11 shows a comparison of permeances and SP as a function of the operating temperature on a (a) c-oriented, (b) [h0l]-oriented, (c) a- and b-oriented, and (d) b-oriented MFI membranes. It is clearly evidenced that MFI membranes with different crystal orientation showed different behavior in separation of xylene isomer, indicating the impact of the crystal orientation on the membrane performance for selective separation of isomers. For xylene isomerization in a membrane reactor, the separation of the produced p-xylene from the reaction zone will shift the equilibrium and enhance the selectivity and yield p-xylene. Zhang et al. tested silicalite-1/α-Al2 O3 zeolite membranes packed with H-ZSM-5 catalysts in the isomerization of m-xylene to p-xylene [164]. Two different catalysts packing methods have been studied: (i) depositing the catalyst on the Al2 O3 support on the opposite side of the membrane and (ii) packing the catalysts in the tube in contact with the membrane layer. They found that the latter packing method was more effective since it gave rise to higher p-xylene yield and selectivity than in conventional reactor thanks to an immediate removal of p-xylene from the reaction zone due to molecular sieving effects. van Dyk et al. [157] used a pore-plugging-type zeolite/alumina nanocomposite membrane (i.e. zeolite crystals grown as film inside the pores of the porous tubular support) equipped with a commercial xylene isomerization Pt catalyst (ISOXYL) for m-xylene isomerization. By combining the retentate and permeate fractions, p-xylene selectivity of 65% and yield of 23% were reported.

275

8 Pervaporation with Zeolite Membranes c-axis

c-axis

1 000 100

1 000 100

o-Xylene

SP 10

10

o-Xylene SP

1 0.1 100

120

1

140

160

180

a/b mixed

(a)

10 000 1 000

200 100

120

160

180

0.1 200

10 000

p-Xylene

p-Xylene

1 000

o-Xylene SP

100 10

10 1

140

b-axis

(b)

100

0.1 100

(c)

p-Xylene

p-Xylene

o-Xylene SP

120

1

140

160

180

200 100

Separation factor (SP)

10 000

120

140

160

180

Separation factor (SP)

Permeance 10–10 mol m–2 s–1 Pa–1

10 000

Permeance 10–10 mol m–2 s–1 Pa–1

276

0.1 200

(d) Temperature (°C)

Figure 8.11 ZSM-5 membrane performance in xylene isomer separation (feed partial pressure of p-xylene and o-xylene are 0.45 kPa and 0.35 kPa, respectively): p-xylene, o-xylene permeance, and mixture separation factor (SP) in dependence on temperature for (A) c-oriented, (B) [h0l]-oriented, (C) a- and b-oriented, and (D) b-oriented membranes [79]. Source: Copyright 2003, AAAS.

Daramola et al. applied a “filmlike” MFI–alumina membrane packed with Pt/H-ZSM-5 catalyst for m-xylene isomerization [165]. Decreasing the operating temperature resulted in linear increase in the p-xylene yield due to an effective extraction of p-xylene from the reaction zone. In a parallel study [166], Daramola and coworkers managed to synthesize membranes of higher quality in terms of selectivity for p-xylene (i.e. p-xylene/o-xylene > 400). The zeolite membrane reactor outperformed the conventional one at the studied reaction temperature of 473 K. In another study, they employed MFI–ceramic hollow fiber nanocomposite membranes for the separation of xylene isomers. In their findings, nearly 30% increase in p-xylene flux with reasonable selectivity was observed [167]. Finally, the utilization of hollow fiber membranes in zeolite membrane reactors for the isomerization of xylene could offer the essential increase in the p-xylene

8.7 Conclusions and Outlook

permeation flux in order to overcome one of the considerable limitations of this reactor configuration in competition with the existing technologies. 8.6.8

Other Separations

The concept of zeolite membrane encapsulated catalyst (ZMEC) is an original way to use zeolite membranes in catalytic reactors while limiting the influence of defects (non-zeolite pores) on large-scale membranes [168, 169]. Coating catalyst particles with a selective zeolite layer can reveal useful for either increasing reactant selectivity [169] (e.g. selective hydrogenation of linear molecules with silicalite-1-coated Pt/TiO2 ) or product selectivity [159] (e.g. alkylation of toluene with methanol or isomerization of m-xylene with silicalite-1-coated ZSM-5). The method can also be useful to protect the catalyst from poisoning. Polymer–zeolite composite membranes are also studied in reactors, either as interphase contactors in liquid-phase oxidation processes or for improving the properties of Nafion in fuel cell applications [170, 171].

8.7 Conclusions and Outlook PV is advantages in separating azeotropes, close-boiling mixtures, and thermally sensitive compounds, but only at low concentration due to heat transfer limitation. Advances in the following areas are required in order to understand and improve the effectiveness of PV through zeolite membranes: (i) investigate the PV performance of thin membranes; (ii) increase hydrophobicity in zeolite membrane for removal of organic compounds from water; (iii) improve membrane production techniques and reproducibility, particularly on the oriented zeolite membranes; (iv) understand PV transport mechanism by investigating mixture isotherms and multicomponent diffusivities during PV; (v) conduct modeling and simulations of transport through zeolites; (vi) study the behavior and transport of polar molecules and other components during PV; and (vii) investigate the effect of fouling on zeolite membranes based on the adsorption strengths of the molecules during PV. Most mixtures separated in industrial applications contain impurities or are multicomponent mixtures, but few studies on PV using these mixtures have been performed. In the future, it is believed that PV/hybrid process will find new opportunities especially in tackling challenging separation problems encountered in petrochemical industry, thereby opening new avenues for its widespread industrialization. After the successful realization of an industrial separation process using zeolite membranes, the development of catalytic membrane reactors has emerged. Most probably, here again the shape-selective separation behavior of zeolite membranes will be exploited, which recommends the application of an extractor-type membrane reactor. More studies on the economic feasibility of zeolite membrane reactor processes, long-term stability of membranes, and upscaling are needed. A closer look at temperature processes and control within the synthesis autoclave is necessary for large-scale membrane production; here

277

278

8 Pervaporation with Zeolite Membranes

microwave heating, feasible at small-scale preparation, might be a solution to the problem although this should be further explored. Further improvements in especially cost reduction and membrane reliability should still be endeavored to facilitate the wide introduction of zeolite-based membrane reactors into industrial practice. Finally, reaction and separation microdevices based on zeolite membranes are now also seriously considered [172]. Lab-on-a-chip (i.e. scaled-down analogue of a chemical processing plant) is becoming a familiar concept to express the miniaturization of chemical and biological analyses, with a drastic reduction of reactant consumption [173]. According to the large number of studies emerging in this field in the literatures [107, 174], micro-separators and micro-reactors with zeolite membranes will certainly be developed very rapidly during the coming years.

References 1 Coterillo, C.C., Mendia, A.M.U., and Uribe, I.O. (2008). Pervaporation and

2 3 4 5 6

7 8 9 10 11 12 13 14 15 16 17

gas separation using microporous membranes, In Inorganic Membranes: Synthesis, Characterization and Applications, Eds. R. Mallada and M. Menéndez. Membr. Sci. Technol. 13: 217–253. Bowen, T.C., Noble, R.D., and Falconer, J.L. (2004). J. Membr. Sci. 245: 1–33. Baker, R.W. (2004). Membrane Technology and Applications. Chichester: Wiley. Drioli, E. and Romano, M. (2001). Ind. Eng. Chem. Res. 40: 1277–1300. Feng, X. and Hang, R.Y.M. (1997). Ind. Eng. Chem. Res. 36: 1048–1066. Fleming, H.L. and Slater, C.S. (1992). Pervaporation. In: Membrane Handbook (ed. W.S.W. Ho and K.K. Sirkar), 105–116. New York: Chapman & Hall. Jonquières, A., Clément, R., Lochon, P. et al. (2002). J. Membr. Sci. 206: 87–117. Karlssson, H.O.E. and Trägardh, G. (1993). J. Membr. Sci. 76: 121–146. Morigami, Y., Kondo, M., Abe, J. et al. (2001). Sep. Purif. Technol. 25: 251–260. Bowen, T.C., Kalipcilar, H., Falconer, J.L., and Noble, R.D. (2003). J. Membr. Sci. 215: 235–247. Kalipcilar, H., Gade, S.K., Noble, R.D., and Falconer, J.L. (2002). J. Membr. Sci. 210: 113–127. Xu, X., Yang, W., Liu, J. et al. (2004). J. Membr. Sci. 229: 81–85. Bowen, T.C., Wyss, J.C., Noble, R.D., and Falconer, J.L. (2004). Ind. Eng. Chem. Res. 43: 2598–2601. Holmes, S.M., Schmitt, M., Markert, C. et al. (2000). Chem. Eng. Res. Des. 78: 1084–1088. Tanaka, K., Kita, H., Okamoto, K.I. et al. (2002). J. Membr. Sci. 197: 173–183. Bowen, T.C., Wyss, J.C., Noble, R.D., and Falconer, J.L. (2004). Microporous Mesoporous Mater. 71: 199–210. Kalipcilar, H., Bowen, T.C., Noble, R.D., and Falconer, J.L. (2002). Chem. Mater. 14: 3458–3464.

References

18 Funke, H.H., Frender, K.R., Green, K.M. et al. (1997). J. Membr. Sci. 129:

77–82. 19 Nomura, M., Bin, T., and Nakao, S. (2002). Sep. Purif. Technol. 27: 59–66. 20 Li, S.G., Tuan, V.A., Falconer, J.L., and Noble, R.D. (2001). Ind. Eng. Chem.

Res. 40: 1952–1959. 21 Aptel, P., Challard, N., Cuny, J., and Néel, J. (1976). J. Membr. Sci. 1:

271–287. 22 Gallego-Lizón, T., Ho, Y.S., and Freitas dos Santos, L. (2002). Desalination

149: 3–8. 23 Gallego-Lizón, T., Edwards, E., Lobiundo, G., and Freitas dos Santos, L.

(2002). J. Membr. Sci. 197: 309–319. 24 Jou, J.D., Yoshida, W., and Cohen, Y. (1999). J. Membr. Sci. 162: 269–284. 25 Lipnizki, F., Field, R.W., and Ten, P.K. (1999). J. Membr. Sci. 153: 183–210. 26 Liu, Q., Noble, R.D., Falconer, J.L., and Funke, H.H. (1996). J. Membr. Sci.

117: 163–174. 27 McCabe, W., Smith, J., and Harriot, P. (1993). Unit Operations of Chemical

Engineering, 5e. New York: McGraw-Hill. 28 Martin, N. (2003). Reprint. Sulzer Technical Review. 29 Tuan, V.A., Li, S., Falconer, J.L., and Noble, R.D. (2002). J. Membr. Sci. 196:

111–123. 30 Burggraaf, A.J. and Cot, L. (1996). Fundamentals of Inorganic Membrane

31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

Science Technology, Membrane Science and Technology Series 4, 1–681. Amsterdam: Elsevier. Van Gemert, R.W. and Cuperus, F.P. (1995). J. Membr. Sci. 105: 287–291. Van Veen, H.M., van Delft, Y.C., Engelen, C.W.R., and Pex, P.P.A.C. (2001). Sep. Purif. Technol. 22: 361–366. Xiao, J. and Wei, J. (1992). Chem. Eng. Sci. 47: 1123–1141. van den Broeke, L.J.P., Kapteijn, F., and Moulijn, J.A. (1999). Chem. Eng. Sci. 54: 259–269. Coronas, J. and Santamaría, J. (1999). Sep. Purif. Methods 28: 127–177. Cui, Y., Kita, H., and Okamoto, K.I. (2004). J. Membr. Sci. 236: 17–27. Elshof, J.E., Rubio Abadal, C., Sekulic, J. et al. (2003). Microporous Mesoporous Mater. 65: 197–208. Gao, Z., Yue, Y., and Li, W. (1996). Zeolites 16: 70–74. He, W., Chan, W.H., and Ng, C.F. (2001). J. Appl. Polym. Sci. 82: 1323–1329. Huang, R.Y.M., Shao, P., Feng, X., and Anderson, W.A. (2002). Ind. Eng. Chem. Res. 41: 2957–2965. Li, S., Tuan, V.A., Noble, R.D., and Falconer, J.L. (2001). Ind. Eng. Chem. Res. 40: 4577–4585. Masuda, T., Otani, S.H., Tsuji, T. et al. (2003). Sep. Purif. Technol. 32: 181–189. Zah, J., Krieg, H.M., and Breytenbach, J.C. (2007). J. Membr. Sci. 287: 300–316. Asaeda, M., Sakou, Y., Yang, J., and Shimasaki, K. (2002). J. Membr. Sci. 209: 163–175. Campaniello, J., Engelen, C.W.R., Haije, W.G. et al. (2004). Chem. Commun. 7: 833–834.

279

280

8 Pervaporation with Zeolite Membranes

46 Casado, L., Mallada, R., Téllez, C. et al. (2003). J. Membr. Sci. 216: 135–147. 47 Kita, H. (2002). Fabrication and application of zeolite membrane. In:

48 49 50 51 52 53

54 55 56 57

58 59 60 61 62 63 64 65

66

67 68

Membranes-Preparation, Properties and Applications, vol. 752 (ed. V.N. Burganos, R.D. Noble, M. Asaeda, et al.), 239. Warrendale, PA: Materials Research Society. Okamoto, K., Kita, M., Horii, K. et al. (2001). Ind. Eng. Chem. Res. 40: 165–175. Urtiaga, A., Gorri, E.D., Casado, C., and Ortiz, I. (2003). Sep. Purif. Technol. 32: 207–213. Shah, D., Kissick, K., Ghorpade, A. et al. (2000). J. Membr. Sci. 179: 185–205. Dafinov, A., Garcia-Valls, R., and Font, J. (2002). J. Membr. Sci. 196: 69–77. Li, Y., Zhu, G., Liu, J., and Yang, W. (2007). J. Membr. Sci. 297: 10–15. Asaeda, M., Kanezashi, M., Yoshioka, T., and Tsuru, T. (2002). Gas permeation characteristics and stability of composite silica-metal oxide membranes. In: Membranes-Preparation, Properties and Applications, vol. 752 (ed. V.N. Burganos, R.D. Noble, M. Asaeda, et al.), 213–218. Warrendale, PA: Materials Research Society. Asaeda, M. and Yamasaki, S. (2001). Sep. Purif. Technol. 25: 151–159. Urtiaga, A., Casado, C., Asaeda, M., and Ortiz, I. (2006). Desalination 193: 97–102. Sekulic, J., ten Elshof, J.E., and Blank, D.H.A. (2006). J. Membr. Sci. 254: 267–274. Yang, J. (2002). A study on pervaporation separation of aqueous solutions of organic solvents with microporous SiO2-ZrO2 membranes, PhD thesis, University of Hiroshima, Japan. Brinkmann, T., Puhst, J., and Pingel, H. (2006). Desalination 200: 262–264. Caro, J., Noack, M., and Kölsch, P. (2005). Adsorption 11: 215–227. Caro, J. (2004). Stud. Surf. Sci. Catal. 154: 80–93. Kyotani, T., Sato, K., Mizino, T. et al. (2006). Anal. Sci. 21: 321–325. Kyotani, T., Inoue, S., Kakui, S., and Saito, J. (2006). Anal. Sci. 22: 317–319. Kyotani, T., Shimotsuma, N., and Kakui, S. (2006). Anal. Sci. 22: 325–327. Kyotani, T., Kajui, S., Mizuno, T. et al. (2006). Anal. Sci. 22: 1031–1034. Inoue, S., Mizuno, T., Saito, J., et al. (2006). Development of 1000 mm Long LTA zeolite membrane for alcohols dehydration with high performance. Proceedings of the 9th International Conference on Inorganic Membranes, Lillehammer, Norway (25–29 June 2006), p. 416. Sato, K., Take, M., Sugimoto, K., and T. Nakane (2006). Super high flux LTA and FAU zeolite membranes; influence of feed-flow rate on the permeate flux in PV and VP dehydration processes. Proceedings of the 9th International Conference on Inorganic Membranes, Lillehammer, Norway (25–29 June 2006), p. 670. Lai, S.M., Martin-Aranda, R., and Yeung, K.L. (2003). Chem. Commun. 2: 218–219. Sato, K., Sugimoto, K., and Nakane, T. (2006). High-flux FAU zeolite membranes for application to organic-organic separations. Proceedings of the 9th International Conference on Inorganic Membranes, Lillehammer, Norway (25–29 June 2006), p. 480.

References

69 Li, G., Kikuchi, E., and Matsukata, M. (2003). Microporous Mesoporous

Mater. 62: 211–220. 70 Sato, K., Sugimoto, K., Nakane, T., and Matsukata, M. (2006). Dehydrations

71

72 73

74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93

behaviour of MFI zeolite membrane for various kinds of solvents. Proceedings of the 9th International Conference on Inorganic Membranes, Lillehammer, Norway (25–29 June 2006), p. 333. Richter, H., Voigt, I., and Kühnert, J.T. (2006). NaA-membranes for bioethanol dewatering by pervaporation and vapour permeation. Proceedings of the 9th International Conference on Inorganic Membranes, Lillehammer, Norway (25–29 June 2006), p. 552. Richter, H., Voigt, I., and Kühnert, J.T. (2006). Desalination 199: 92–93. Richter, H., Weyd, M., Fischer, G., et al. (2006). Separation of bioethanol by pervaporation with hydrophobic ZSM-5 membranes. Proceedings of the 9th International Conference on Inorganic Membranes, Lillehammer, Norway (25–29 June 2006), p. 329. Weyd, M., Richter, H., Voigt, I. et al. (2006). Desalination 199: 308–309. Gao, X., Zou, X.Q., Zhang, F. et al. (2013). Chem. Commun. 49: 8839–8841. Kita, H., Inoue, T., Asamura, H. et al. (1997). Chem. Commun. 45–46. Matsufuji, T., Watanabe, K., Nishiyama, N. et al. (2000). Ind. Eng. Chem. Res. 39: 2434–2438. Flanders, C.L., Tuan, V.A., Noble, R.D., and Falconer, J.L. (2000). J. Membr. Sci. 176: 43–53. Lai, Z.P., Bonilla, G., Diaz, I. et al. (2003). Science 300: 456–460. Sommer, S., Melin, T., Falconer, J.L., and Noble, R.D. (2003). J. Membr. Sci. 224: 51–67. Nishiyama, N., Ueyama, K., and Matsukata, M. (1996). Microporous Mater. 7: 299–308. Nishiyama, N., Matsufuji, T., Ueyama, K., and Matsukata, M. (1997). Microporous Mater. 12: 293–303. Hedlund, J., Sterte, J., Anthonis, M. et al. (2002). Microporous Mesoporous Mater. 52: 179–189. Keizer, K., Burggraaf, A.J., Vroon, Z., and Verweij, H. (1998). J. Membr. Sci. 147: 159–172. Xomeritakis, G., Lai, Z.P., and Tsapatsis, M. (2001). Ind. Eng. Chem. Res. 40: 544–552. Matsufuji, T., Nakagawa, S., Nishiyama, N. et al. (2000). Microporous Mesoporous Mater. 38: 43–50. Nishiyama, N., Ueyama, K., and Matsukata, M. (1997). Stud. Surf. Sci. Catal. 105: 2195–2202. Wegner, K., Dong, J.H., and Lin, Y.S. (1999). J. Membr. Sci. 158: 17–27. Yuan, W., Lin, Y.S., and Yang, W. (2004). J. Am. Chem. Soc. 126: 4776–4777. Liu, G., Jin, W., and Xu, N. (2016). Angew. Chem. Int. Ed. 55: 13384–13397. Varoon, K., Zhang, X., Elyassi, B. et al. (2011). Science 334: 72–75. Tsapatsis, M. (2014). AIChE J. 60: 2374–2381. van Donk, S., Janssen, A.H., Bitter, J.H., and de Jong, K.P. (2003). Catal. Rev. 45: 297–319.

281

282

8 Pervaporation with Zeolite Membranes

94 Kooyman, P.J., van der Waal, P., and van Bekkum, H. (1997). Zeolites 18:

50–53. 95 Li, S.G., Tuan, V.A., Noble, R.D., and Falconer, J.L. (2001). Ind. Eng. Chem.

Res. 40: 6165–6171. 96 Sano, T., Ejiri, S., Yamada, K. et al. (1997). J. Membr. Sci. 123: 225–233. 97 Li, G., Kikuchi, E., and Matsukata, M. (2003). J. Membr. Sci. 218: 185–194. 98 Dalmon, J.A. (1997). Catalytic membrane reactors. In: Handbook of Heteroge-

neous Catalysis (ed. H. Knozinger, G. Ertl and J. Weitkamp). VCH Pub. 99 Drioli, E. and Fontanova, E. (2010). Ullmann’s Encyclopaedia of Industrial

Chemistry. Weinheim: Wiley-VCH. 100 Fontanova, E. and Drioli, E. (2014). Chem. Ing. Tech. 86: 2039–2050. 101 Dittmeyer, R. and Caro, J.(ed. G. Ertl, H. Knözinger, F. Schüth and J.

Weitkamp). Handbook of Heterogeneous Catalysis, 2198–2247. Wiley-VCH. 102 Rangnekar, N., Mittal, N., Elyassi, B. et al. (2015). Chem. Soc. Rev. 44:

7128–7154. 103 Pantazidis, A., Dalman, J.A., and Mirodatos, C. (1995). Catal. Today 25:

403–408. 104 Mota, S., Miachon, S., Volta, J.C., and Dalman, J.A. (2001). Catal. Today 67:

169–176. 105 Farrusseng, D., Julbe, A., and Guizard, C. (2001). Sep. Purif. Technol. 25:

137–149. 106 Julbe, A., Farrusseng, D., Jalibert, J.C. et al. (2000). Catal. Today 56:

199–209. 107 Coronas, J. and Santamaria, J. (2004). Top. Catal. 29: 29–44. 108 Sanchez-Marcano, J.G. and Tsotsis, T.T. (2002). Catalytic Membranes, Mem-

brane Reactors. Weinheim: Wiley VCH. 109 Kapteijn, F., Zhu, W., Moulijn, J.A., and Gardner, T.Q. (2005). Zeolite mem-

110 111 112 113 114 115 116 117 118 119 120

branes – modelling, application. In: Structured Catalysts and Reactors, 2e (ed. J.A. Moulijn and A. Cybulski), 701–747. New York: Marcel Dekker. Khajavi, S., Sartipi, S., Gascon, J. et al. (2010). Microporous Mesoporous Mater. 132: 510–517. Khajavi, S., Kapteijn, F., and Jansen, J.C. (2007). J. Membr. Sci. 299: 63–72. Lee, S.R., Son, Y.H., Julbe, A., and Choy, J.H. (2006). Thin Solid Films 495: 92–96. Julbe, A., Motuzas, J., Cazeville, F. et al. (2003). Sep. Purif. Technol. 32: 139–149. Wang, N., Liu, Y., Huang, A., and Caro, J. (2015). Microporous Mesoporous Mater. 207: 33–38. Günther, C., Richter, H., and Voigt, I. (2013). Chem. Eng. Trans. 32: 1963–1968. Kim, S.J., Yang, S., Reddy, G.K. et al. (2013). Energy Fuel 27: 4471–4480. Zhang, Y., Wu, Z., Hong, Z. et al. (2012). Chem. Eng. J. 197: 314–321. Tanaka, K., Yoshikawa, R., Ying, C. et al. (2001). Catal. Today 67: 121–125. Tanaka, K., Yoshikawa, R., Ying, C. et al. (2002). Chem. Eng. Sci. 57: 1577–1584. de la Iglesia, Ó., Mallada, R., Menéndez, M., and Coronas, J. (2007). Chem. Eng. J. 131: 35–39.

References

121 de la Iglesia, Ó., Irusta, S., Mallada, R. et al. (2006). Microporous Mesoporous

Mater. 93: 318–324. 122 Salomón, M.A., Coronas, J., Menéndez, M., and Santamar𝚤´ a, J. (2000).

Appl. Catal., A 200: 201–210. 123 Fedosov, D.A., Smirnov, A.V., Shkirskiy, V.V. et al. (2015). J. Membr. Sci. 486:

189–194. 124 Pera-Titus, M., Llorens, J., and Cunill, F. (2009). Chem. Eng. Process. Process

Intensif. 48: 1072–1079. 125 Diban, N., Urtiaga, A.M., Ortiz, I. et al. (2013). Chem. Eng. J. 234: 140–148. 126 Olah, G.A., Goeppert, A., and Prakash, G.K.S. (2009). J. Org. Chem. 74:

487–498. 127 Aguayo, A.T., Ereña, J., Sierra, I. et al. (2005). Catal. Today 106: 265–270. 128 Rohde, M.P., Unruh, D., and Schaub, G. (2005). Catal. Today 106: 143–148. 129 Rohde, M.P., Schaub, G., Khajavi, S. et al. (2008). Microporous Mesoporous

Mater. 115: 123–136. 130 Espinoza, R.L., du Toit, E., Santamaria, J. et al. (2000). Stud. Surf. Sci. Catal.

130: 389–394. 131 Rahimpour, M.R., Mirvakili, A., and Paymooni, K. (2011). Energy 36:

1223–1235. 132 Rahimpour, M.R., Mirvakili, A., Paymooni, K., and Moghtaderi, B. (2011).

J. Nat. Gas Sci. Eng. 3: 484–495. 133 Cavani, F. and Trifiro, F. (1997). Catal. Today 34: 269–279. 134 Dyk, L.V., Miachon, S., Lorenzen, L. et al. (2003). Catal. Today 82: 167–177. 135 Ciavarella, P., Casanave, D., Moueddeb, H. et al. (2001). Catal. Today 67:

177–184. 136 Casanave, D., Giroir-Fendler, A., Sanchez, J. et al. (1995). Catal. Today 25:

309–314. 137 Casanave, D., Ciavarella, P., Fiaty, K., and Dalmon, J.A. (1999). Chem. Eng.

Sci. 54: 2807–2815. 138 Illgen, U., Schäfer, R., Noack, M. et al. (2001). Catal. Commun. 2: 339–345. 139 Schäfer, R., Noack, M., Kölsch, P. et al. (2003). Catal. Today 82: 15–23. 140 Bergh, J.V.D., Gücüyener, C., Gascon, J., and Kapteijn, F. (2011). Chem. Eng.

J. 166: 368–377. 141 Jeong, B.H., Sotowa, K.I., and Kusakabe, K. (2003). J. Membr. Sci. 224:

151–158. 142 Jeong, B.H., Sotowa, K.I., and Kusakabe, K. (2004). Chem. Eng. J. 103: 69–75. 143 Kong, C.L., Lu, J.M., Yang, H.H., and Wang, J.Q. (2007). J. Membr. Sci. 306:

29–35. 144 Vaezi, M.J., Babaluo, A.A., and Shafiei, S. (2015). J. Chem. Pet. Eng. 49:

51–62. 145 Avila, A.M., Yu, Z., Fazli, S. et al. (2014). Microporous Mesoporous Mater.

190: 301–308. 146 Tang, Z., Dong, J., and Nenoff, T.M. (2009). Langmuir 25: 4848–4852. 147 Tang, Z., Kim, S.J., Reddy, G.K. et al. (2010). J. Membr. Sci. 354: 114–122. 148 Gu, X., Tang, Z., and Dong, J. (2008). Microporous Mesoporous Mater. 111:

441–448.

283

284

8 Pervaporation with Zeolite Membranes

149 Dong, X.L., Wang, H.B., Rui, Z.B., and Lin, Y.S. (2015). Chem. Eng. J. 268: 150 151 152 153 154 155 156

157 158 159

160 161 162 163 164 165 166 167 168 169 170 171

219–229. Zhang, Y., Sun, Q., and Gu, X. (2015). AIChE J. 61: 3459–3469. Liu, B.S. and Au, C.T. (2001). Catal. Lett. 77: 67–74. Liu, B.S., Gao, L.Z., and Au, C.T. (2002). Appl. Catal., A 235: 193–206. Gallucci, F., Paturzo, L., and Basile, A. (2004). Chem. Eng. Process. 43: 1029–1036. Graaf, J.M.V., Zwiep, M., Kapteijn, F., and Moulijn, J.A. (1999). Appl. Catal., A 178: 225–241. Graaf, J.M.V., Zwiep, M., Kapteijn, F., and Moulijn, J.A. (1999). Chem. Eng. Sci. 54: 1441–1445. Lee, G.S., McCain, J., and Bhasin, M. (2007). Synthetic organic chemicals. In: Kent and Riegel’s Handbook of Industrial Chemistry and Biotechnology (ed. J. Kent), 345–403. New York: Springer US. van Dyk, L., Lorenzen, L., Miachon, S., and Dalman, J.A. (2005). Catal. Today 104: 274–280. Haag, S., Hanebuth, M., Mabande, G.T.P. et al. (2006). Microporous Mesoporous Mater. 96: 168–176. Miyamoto, M., van Dung, V., Nishiyama, N. et al. (2006). A silicalite film coating on ZSM-5 catalysts for selective production of p-xylene. Proceedings of the 9th International Conference on Inorganic Membranes, Lillehammer, Norway (25–29 June 2006), p. 34. Degnan, T. (ed.) (2003). p-Xylene production costs reduced by new zeolite membranes. Focus. Catal. 1: 7. Sakai, H., Tomita, T., and Takahashi, T. (2001). Sep. Purif. Technol. 25: 297–306. Gu, X., Dong, J., Nenoff, T.M., and Ozokwelu, D.E. (2006). J. Membr. Sci. 280: 624–633. Gump, C.J., Tuan, V.A., Noble, R.D., and Falconer, J.L. (2001). Ind. Eng. Chem. Res. 40: 565–577. Zhang, C., Hong, Z., Gu, X.H. et al. (2009). Ind. Eng. Chem. Res. 48: 4293–4299. Daramola, M.O., Deng, Z., Pera-Titus, M. et al. (2010). Catal. Today 156: 261–267. Daramola, M.O., Burger, A.J., Giroir-Fendler, A. et al. (2010). Appl. Catal., A 386: 109–115. Daramola, M.O., Burger, A.J., Pera-Titus, M. et al. (2009). J. Membr. Sci. 337: 106–112. Nishiyama, N., Miyamoto, M., Egashira, Y., and Ueyama, K. (2001). Chem. Commun. 1746–1747. Nishiyama, N., Ichioka, K., Park, D.H. et al. (2004). Ind. Eng. Chem. Res. 43: 1211–1215. Langhendries, G., Baron, G.V., and Jacobs, P.A. (1999). Chem. Eng. Sci. 54: 1467–1472. Tricoli, V. and Nannetti, F. (2003). Electrochim. Acta 48: 2625–2633.

References

172 Wan, Y.S.S., Chau, J.L.H., Gavriilidis, A., and Yeung, K.L. (2002). Chem.

Commun. 878–879. 173 Chow, A.W. (2002). AIChE J. 48: 1590–1595. 174 Yeung, K.L. (2006). Design, fabrication and application of zeolite membrane.

Proceedings of the 9th International Conference on Inorganic Membranes, Lillehammer, Norway (25–29 June 2006), p. 218.

285

287

9 MOF Membranes and Their H2 Separation Properties 9.1 Introduction The metal–organic framework (MOF) field has exhibited tremendous growth over the last 20 years. The existing MOFs, about 20 000 in 2013 [1] and more than 70 000 in early 2017 [2], have been obtained by combining roughly 100 different nodes with roughly 1 000 different ditopic, tritopic, or quadrutopic linkers. The number of MOFs that have an associated theoretical description is roughly 1.0 million. Considering that more than 100 million organic compounds have been synthesized so far [3, 4], which are all, in principle, suitable as MOF linkers after additional modification, there is virtually no upper limit to the number of possible structures. MOFs have exceptionally high porosity, uniform but tunable pore sizes, and well-defined molecular adsorption sites. As a new class of porous solid materials, exploration of their performances in separation and purification applications is attracting extreme interest from researchers in fields such as chemistry, chemical engineering, and materials science [5–9]. Membrane separation, which has been proven to be very promising in addressing energy and environmental challenges, has experienced rapid growth in the past few decades [10–12]. Compared with traditional gas separation technologies such as PSA/TSA operation, membrane-based gas separation provides extensive application in terms of its lower energy consumption, smaller carbon footprint, and easy operation. As emerging inorganic and organic hybrid materials, MOF materials are very appealing to be assembled into MOF membranes for gas separation applications for mainly two reasons: their pores can be rationally controlled by the interplay of both inorganic metal ions and organic linkers, and their pore surfaces can be readily functionalized through a variety of methodologies [13]. Commonly, separation is based on the size and shape of the molecules to be separated or on their interaction with the membrane material. For MOF membranes, there are various factors, such as the limiting pore size and pore size distribution (PSD), surface diffusion, capillary condensation, shape selectivity, and molecular sieving, that contribute to the separation properties. These advantages make the MOF membrane more favorable for separation applications [14]. In this chapter, we first discuss a brief introduction to the current methodologies for the fabrication of MOF membranes including polycrystalline MOFs Microporous Materials for Separation Membranes, First Edition. Xiaoqin Zou and Guangshan Zhu. © 2019 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2019 by Wiley-VCH Verlag GmbH & Co. KGaA.

288

9 MOF Membranes and Their H2 Separation Properties

and surface-supported metal–organic frameworks (SURMOFs). Secondly, we highlight the controlling factors and characterization of MOF membranes. Finally, we discuss the role of pore sizes for H2 separation in various MOF membranes.

9.2 Fabrication of MOF Membranes In this section, two prime classes of MOF membranes will be described. The first conventional class of MOF membrane is “polycrystalline membranes.” In this class, MOF crystals are more or less randomly oriented and intergrown on the surface, which compromise elite features of the membrane network like continuous surface, regular crystalline growth, and controlled thickness (few micrometers). The characteristics of polycrystalline membranes are rather difficult to evaluate. Currently, another most spectacular and captivating class of MOF membrane is termed “SURMOFs.” These MOF membranes are ultrathin (few nanometers) multilayers that are well oriented, with unidirectional growth of specimen and continuous surface compared with polycrystalline films. It is more challenging to fabricate a continuous and defect-free MOF membrane. Despite the precisely controlled thickness and ideally intergrown network of SURMOFs, their enhanced properties and vibrant practical applications are due to close surface proximity. In fact, the obtained SURMOF structure does not exist as bulk powder or single-crystalline material but as a uniform membrane material that is only bound to an appropriate substrate. The choice of an appropriate substrate is crucial for the deposition and surface modification of MOF membranes. Recently, different kinds of substrate have been developed for coating such as planar substrates (Au, Si, ITO, and FTO), nonplanar substrates (metallic foam, metal oxide, polymeric particles), and flexible plastics. Synthetic strategies for polycrystalline and SURMOF membranes will be discussed here. A number of strategies have been used for the fabrication of MOF membranes as shown in Figure 9.1. These strategies are “solution-based” and “vacuum-based” fabrication techniques. 9.2.1 9.2.1.1

Fabrication Methods for Polycrystalline Membranes Direct Synthesis

In direct synthesis, we group such methods where substrates (i.e. self-assembled monolayer [SAM]) modified by organic molecules (—COOH, —NH2 , —OH, and pyridine) [15–18] and unmodified substrates are used for appropriate growth solution. Actually, growth occurs at the substrate surface or in the solution simultaneously. This surface growth leads to crystallization of MOF in a continuous manner. In Situ Crystallization Direct synthesis through in situ crystallization is simply

defined as when the substrates (modified or unmodified) are immersed into a mother solution (vertically lying down is better to avoid sedimentation) and then heating the whole closed system solvothermally. The growth, nucleation,

ye de r-by po -la sit ye ion r

Che mic a dep l vapo u osit ion r

t rec s Di hesi nt sy

ts

See d grow ed th

sem na bly o no cry f MO sta ls F

As

al mic he o-c on ctr cati Ele fabri

n ge rea e e is pw sag Ste do

Linker solution

La

Figure 9.1 Fabrication methods for MOF films or membranes.

e has d-p y i u Liq pitax e

9.2 Fabrication of MOF Membranes

Time

Figure 9.2 Schematic representation of in situ crystallization of MOFs in the gel layer [15]. Source: Copyright 2011 ACS.

and intergrowth of MOF crystals occur on the substrate and all happen during the same fabrication process. In some cases, homogeneous and dense crystalline membranes of few micrometers thick have been obtained by this method. A general description of in situ crystallization is shown in Figure 9.2. Lai and Jeong et al. synthesized an MOF-5 membrane using unmodified porous alumina substrate at 105 ∘ C by in situ solvothermal approach. Here, the ligand (—COOH) formed covalent bonds with (—OH) group of the substrate [19]. Similarly, ZIF-8 and ZIF-69 membranes were fabricated using the same approach [20]. Despite its simplicity, this approach is not that powerful, and it remains difficult to prepare crack-free and continuous membranes due to lack of interfacial interaction and heterogeneous nucleation on the unmodified substrate. So there is a need to increase interfacial surface interaction between the substrate and the MOF network. Then in 2009, the authors reported a metallic (Cu) net (400 mesh) for the fabrication of HKUST-1 membrane at 120 ∘ C for three days, which produced a continuous and defect-free membrane of about 60 μm thick [21]. Later on Z. Kang et al. reported a continuous and thin membrane of Ni(L-asp)2 (bipy) crystals grown using nickel net without addition of a metallic precursor because nickel net played both as an ionic source and as a supporting substrate [22].

289

290

9 MOF Membranes and Their H2 Separation Properties

NH2 NH2 NH2 EtO

Si OEt

OEt

Toluene 110 °C, 2 h

NH2 NH2

NH2

NH2

O

H

N

N

Zn

Zn

DMF 100 °C, 18 h

Covalent linkages

Figure 9.3 Scheme for preparing ZIF-90 membranes using 3-aminopropyltriethoxysilane (APTES) as a covalent linker between a ZIF-90 membrane and an Al2 O3 support via an imine condensation reaction [27]. Source: Copyright 2010 ACS.

A common challenge facing MOF membranes is poor attraction between membrane network and supporting substrate. Another effective direct synthesis method is chemical modification of the desired substrate with appropriate functional groups (e.g. —COOH, NH2 , OH, and pyridine) [15–18] to enhance the heterogeneous nucleation and direct growth of MOF membranes on the molecular level [23, 24]. Numerous inorganic compounds such as lead sulfate, calcium carbonate, metal (Ti, Zn, Fe) oxides, and zeolites have been grown using SAM. McCarthy et al. reported the fabrication of ZIF-based continuous and well-intergrown thin film at 200 ∘ C using modified porous alumina substrate with organic linker [25]. Also Huang et al. reported defect-free, high-quality ZIF-22 membrane with proper thickness (40 μm) using 3-aminipropyltriethoxysilane (APTES) functionalized porous titania substrate [26]. The same covalent functionalized strategy was adopted for fabrication of ZIF-90 as shown in Figure 9.3 [27]. Commonly, MOF synthesis is performed through conventional oven heating, but recently microwave heating is employed for MOF membrane fabrication [28]. Huang et al. synthesized high-quality, crack-free ZIF-8 membrane with a uniform thickness of about 40 μm using Al2 O3 as substrate via microwave heating and compared it with conventional ZIF-8 membrane [29]. As mentioned above, the fabrication process of MOF membranes often needs multiple steps and is not straightforward because it not generally favors the intergrowth and homogeneous nucleation of MOF crystals on porous substrate. A key problem of the in situ fabrication process arises from high temperature (>100 ∘ C) treatment because majority of supporting substrates (Au-SAM) are temperature sensitive. To mitigate this problem, several researchers adopted crystal growth from solution rather than mother solution directly. In this area, the first case study was introduced in MOF-5 fabrication on Au-SAM substrates in 2005. Hermes and coworkers prepared a mother solution of MOF-5 by first heating it at 75 ∘ C for three days, then quickly heating it to 105 ∘ C to initiate crystallization, and finally cooling it to room temperature to obtain a turbid solution, and they then immersed Au-SAM substrates into the crystal solution for 24 hours. A thin film of ultra-refined and uniform thickness of 5 μm was found without damaging gold substrates [30]. Furthermore, Bein et al. extended this useful strategy to many other MOF membranes such as CAU-1, HKSTU-1, and Fe-MIL-88B

9.2 Fabrication of MOF Membranes

Nylon membrane

Zinc nitrate solution (a)

Hmim solution

Zn2+

Hmim

ZIF-8 film

(b)

Figure 9.4 (a) Diffusion cell for ZIF-8 membrane fabrication and (b) schematic of the formation of ZIF-8 films on both sides of the nylon substrate via contra-diffusion of Zn2+ and Hmim through the pores of the nylon support [34]. Source: RSC.

[31–33]. Meanwhile, a few limitations of this method emerged such as (i) long processing duration (up to several days), (ii) multiple heating steps, (iii) noncontinuous membrane, and (iv) poor morphology and intergrowth. Slow Diffusion of Reactants As we know, the crystals are preferably grown in

homogeneous medium rather than at the liquid–solid interfaces of the substrates. To resolve this issue, Bein and coworkers introduced a novel fabrication strategy for MOF membranes. In this strategy, the macroporous substrate was used as “permeable membrane” between two homogeneous solutions, i.e. salt solution and organic linker solution, because the metal and organic linker ions had the ability to diffuse into the macroporous substrate, and crystallization occurred at the interface as shown in Figure 9.4. Recently, Yao and coworkers applied this method for fabrication of ZIF-8 thin film on macroporous nylon membrane [34]. After 72 hours in static medium of zinc nitrate and Hmim (organic linker) solutions, a continuous crystalline layer of about 18 μm thick was observed on each side of nylon substrate. Moreover, the thickness and morphology of the membrane could be controlled by changing synthesis time, repetitions, and concentration of mother solutions. The MOF membrane obtained by slow diffusion process is not of good quality due to uncontrolled deposition of crystals at the substrate surface. Controlled Crystallization on Surfaces Another useful method of controlled depo-

sition of MOF crystals on surfaces was proposed by Ameloot and coworkers [35]. In this method, two homogeneous precursor solutions (clear solution) were mixed together in which no MOF precipitates exist at ambient conditions, but MOF crystals (1–2 μm in size) are obtained after solvent evaporation from the clear solution. The authors observed that MOF crystals showed preferred growth orientation along [36] direction, which was unlike to the nature of substrate orientation. A similar case was established by Carbonell and coworkers [37]. They prepared an array of HKUST-1 single crystals. In their approach, droplets of clear precursor solution were deposited on the functionalized SAM substrates. Then controlled array of HKUST-1 crystals was obtained after solvent evaporation under ambient condition. The growth orientation of crystals was again

291

292

9 MOF Membranes and Their H2 Separation Properties

along [111] direction. In the case of hydrophilic (—COOH and —OH) terminated SAMs, the droplets spread completely, while in the case of hydrophobic (—CH3 and —CF3 ) terminated SAMs, the droplets were confined, and approximately 1.0 μm single crystals were observed. Another interesting feature that controls crystal size is the volume of droplets, which is controlled by plasma treatment. 9.2.1.2

Seeded Growth

Seeded growth is also called secondary growth. This method is usually used for MOF membrane fabrication, particularly zeolite-based membranes on porous ceramic supports [38, 39]. The above-stated direct growth method has some limitations in substrate and manufacturing processes. However, in this method, the substrate surface is modified by a sequence of operations to produce denser, crack-free, continuous, and oriented membranes. The method requires two main steps: preparation of MOF seeds and controlled deposition of seed to grow into thin membrane. The second step is often performed under solvothermal condition. Different types of seeds may be used in MOF membrane fabrication. Despite this, several seeding techniques are employed including rubbing, wiping, dip coating, spin coating, heating, reactive seeding (RS), and electrospinning [40]. Nano-sized MOF Seeds MOF nanocrystals (size of 20–100 nm) are most com-

monly used as seeds for secondary growth in fabrication of MOF membranes. Their typical syntheses were performed by using suitable capping agent with identical functionality such as sodium carboxylates (sodium formate, sodium acetate, and sodium oxalate) to control size and morphology. The synthesis protocols of nano-sized and nano-shaped MOF seed particles can be highly controlled for well-grown MOF membranes, while micron-sized MOF crystals are not favorable. Afterward, they are deposited onto the porous substrate using dropping, spin coating, and other techniques. A typical case in this area was shown by us [41]: the MOF nanocrystals of In(OH)(bdc) were first synthesized and isolated, and then they were redispersed in MeOH, producing a milky colloidal solution. This suspension was dropped onto different substrates including porous α-alumina disk and a silicon wafer. The dry seeded substrates were vertically placed in mother solution and then kept into autoclave for 72 hours. The obtained MOF layer was about 5 μm in thickness. Often, MOF seed crystals are not fully stable in mother solution (as for MOF-5) for secondary growth because they either dissociate or dissolve from the substrate. Sometimes, dissolution occurs due to acidic nature of the reaction mixture, which can be prevented by adding a base (N-ethyldiisopropylamine [DIPEA]) [21]. Another problem encountered in this step is the lack of adhesion between MOF seeds and the porous substrate. To overcome this problem, Ranjan et al. [42] first deposited polyethyleneimine (PEI) layer onto the porous substrate surface and then manually deposited MOF seed crystals on the PEI layered substrate. The PEI layer is responsible for the enhanced interfacial interaction between MOF seeds and substrate via H-bonding. As a result, crack-free, compact, and continuous film of about 20 μm thick was observed, as well as preferred growth orientation along b-direction (perpendicular to the substrate). Similar results were obtained by Li et al. [43, 44] from ZIF-8 thin

9.2 Fabrication of MOF Membranes

films. Again Varela-Guerrero et al. [45], aiming at enhancing adhesion between MOF seeds and substrate without pre-layered organic linker, presented the novel concept of “thermal seeding” for HKUST-1 thin film fabrication on porous α-alumina substrate. They deposited HKUST-1 seed crystals onto a hot substrate (200 ∘ C) rather than room temperature substrate. This helped to anchor seed crystals due to absence of active —OH surface moieties, leading to formation of homogeneous thin films after secondary growth. Another interesting and facile approach called “reactive seeding” was developed by Hu et al. [46], which was used to grow MIL-53(Al) on porous alumina substrate. In this approach, a uniform seed layer was obtained by reaction between inorganic substrate and organic precursor solution in a single stage. Similar to the metal substrate discussed in the in situ crystallization process, the inorganic porous substrate is used as a metal ion source for nucleation or growth of seed crystals via coordination with the organic linker during membrane synthesis. One more versatile and novel technique that is dependent on the choice of supporting substrates was developed by Qiu et al. [47], the so-called electrospinning, which can achieve uniform seed layer with controlled thickness. This seeding growth technique has its own pros and cons although this can be employed to different kinds of supports, particularly tubes as well as economically low-cost ones. A general schematic of this technique is shown in Figure 9.5. Firstly, the authors prepared electrospinning solution of ZIF-8 nanocrystals, polyvinylpropylene (PVP), and methanol and then dropped onto macroporous silica tube with controlled rate and constant speed at high voltage. As a result, silica tubular substrate’s outer surface was uniformly wrapped by electrospun fibers to give smoother, continuous, and crack-free ZIF-8 thin film with uniform thickness. To prove the effectiveness of this novel approach, they prepared different microporous MOF membranes and films on various porous substrates

nt

lve

So

Metal Ligand solution solution S 3

1

olv

2

4

One cycle

Electrospin method ZIF-8/PVP solution en

t

MOF thin film

Syringe pump ZIF-8/PVP composite fiber

High voltage power supply +

Vacuum chunk

Metallic needle

Macroporous SiO2 support

– Collector

(a)

(b)

Figure 9.5 Representation of (a) spin-coating setup [48] and (b) electrospinning process for ZIF-8 on a microporous SiO2 wafer support [47]. Source: Copyright, The Royal Society of Chemistry 2012 and ACS 2016.

293

294

9 MOF Membranes and Their H2 Separation Properties Growth process

IRMOF-1 seed layer

Hybrid IRMOF-3/-1 membrane

Figure 9.6 Schematic illustration of heteroepitaxially growing IRMOF-3 membrane on the surface of an IRMOF-1 seed layer.

such as (i) beta membrane, zeolite NaA membrane, and pure silica zeolite membrane synthesized on porous alumina tube. (ii) Microporous stainless steel nets were used to synthesize zeolite NaY membrane. (iii) MOF (JUC-32) thin film was synthesized on silica disk. All results motivated the potential and versatility of electrospinning technique for practical application. Coordination Polymer as Seed Layer Coordination polymer can be used as seed

layer and catalyze secondary growth in MOF membrane fabrication, which was confirmed by Gascon et al. [49]. The authors chose micron-sized HKUST-1 crystals as case study and one-dimensional Cu–BTC coordination polymer to accelerate growth process and to prevent deprotonation of the H3 BTC (benzene-1,3,5-tricarboxylic acid) organic linker. As a result, they obtained well-defined tagliatelle-shaped microcrystals, which were uniformly distributed on porous alumina substrate with random orientation and thickness of 2–5 μm. To verify this approach, they used other coordination polymers for distinct MOF membrane fabrication. Nan et al. [50] performed stepwise deposition approach using copper carboxylate complexes as coordination polymer on the support surface, and 25 μm thick compact layer of MOF was obtained after the secondary growth step. MOF Thin Film as Seed Layer Despite other seeding materials for secondary growth step, several workers used thin MOF films as a seed layer to obtain thick MOF layer after secondary growth as shown in Figure 9.6. Yusenko and coworkers [51] constructed a [Cu(ndc)(dabco)0.5 ] film using “stepwise dosing of reactants.” The film obtained via secondary growth method was more continuous, oriented, and homogeneous compared with that from direct growth method. Yoo and coworkers [52] studied a similar case related to MOF-5@IRMOF-3 core–shell hybrid thin film, but no satisfactory results were obtained due to poor heterogeneous nucleation rate of IRMOF-3 shell material. Furthermore, to overcome this difficulty, the authors decided to grow MOF-5 seed layer first and then dropped IRMOF-3 mother solution onto the substrate for secondary growth step in one to three hours. The resulting core–shell MOF-5@IRMOF-3 consisted of minor cracks in surface, which may be attributed to the instability of the shell layer (IRMOF-3) toward air. 9.2.1.3

Electrochemical Deposition

Generally, three useful scientific approaches have been developed for MOF membrane fabrication via electrochemical strategy: (i) anodic deposition, (ii) cathodic deposition, and (iii) electrophoretic deposition. The general schematic illustration is shown in Figure 9.7. The first case study was performed

9.2 Fabrication of MOF Membranes

e– e– e–

(b)

Metal ions

e–

Conductive substrate

+

M+

–

RE Working electrode

+

– Electric field

Organic linker

Metal M+

(a)

V

V

+

M+

M+

Counter electrode

V –

(c)

Figure 9.7 Schematic depiction of the electrochemical fabrication of MOF membranes via (a) anodic deposition, (b) electrophoretic deposition, and (c) cathodic deposition [53]. Source: Copyright 2017, RSC.

by Muller and coworkers at BASF [54]. In the anodic deposition process, a metallic electrode (metallic substrate) serves as the anode at high voltage, which oxidizes to produce metal ions into electrolytic solution, which reacts rapidly with organic linker moieties and produces MOF membrane at anode surface. Electrochemical anodic deposition approach is widely applied for fabrication of various MOF membranes on metallic electrodes such as Zn(TPTC) (H3 TPTC=1,3,5-tris[4-(carboxyphenyl)-oxamethyl-2,4,6-trimethylbenzene) on zinc (Zn) anode plate [55], MIL-100(Fe) on Fe anode plate [56], and HKUST-1 (Cu2 (BTC)3 , BTC = benzene-1,3,5-tricarboxylic acid) on Cu plate [57]. Despite this, some nonmetallic substrates coated with a metal or metal oxide (MO) can serve as the anode in MOF membrane fabrication. The literature showed that Cu2 (BTC)3 thin film grew on Cu-coated glass (nonmetallic) substrate [58]. Similarly, more homogeneous Tb-BTC (MOF) thin layer grew on the ZnO substrate [59]. Furthermore, only non-corrosion metal is suitable for anodic electrode growth of MOF membranes [60]. In the cathodic deposition process, MOF membrane fabrication was performed by Dinca et al. [60, 61]. Both working electrode (WE) and counter electrode (CE) are inert in that they cannot participate in the redox reaction, and reactant flow can be directed under an electric field at high voltage, thus leading to the fabrication of MOF membrane. The primary step in the cathodic deposition process is to construct a local alkaline region close to the cathode where organic ligand moieties are deposited after deprotonation. This can only be achieved by reduction of oxoanions to hydroxyl groups. Another useful approach is electrophoretic deposition (EPD). This approach was successfully developed by Farha, Hupp, and coworkers [62]. They use this approach for the fabrication of various MOFs including HKUST-1, Al-MIL-53, NU-1000, and UiO-66 on conductive fluorinated tin oxide (FTO) doped glass substrate. In this process, two conductive electrodes were immersed into MOF-containing solution (electrolyte), and then an electric field was generated by applying certain voltage between conductive electrodes, which deposited MOF charge particles onto the counter charge electrode, thus leading to formation of MOF membrane.

295

296

9 MOF Membranes and Their H2 Separation Properties

9.2.1.4

Stepwise Dosing of Reagents

This approach is closely related to the liquid-phase epitaxy (LPE) technique, but factually, this strategy is concerned to linear layer-by-layer (LBL) growth on the SAM-modified substrates. The stepwise application of reagents on the SAM-modified substrates was investigated using porous alumina [63], inorganic wafers [51], and textiles [64] to examine [Cu2 (ndc)2 (dabco)] membrane. Yusenko and coworkers used this approach for MOF membrane fabrication on various inorganic wafers including silica, alumina, Ta2 O5 , and Si3 N4 . The authors observed that no self-terminated nucleation occurred due to dynamic complex equilibrium between ligand and metal ions at high temperature (>50 ∘ C). After sequential deposition cycles, 100 nm crystalline thin membrane grew to 150 nm with orientation along [001] direction, which was further used as a seed layer for secondary crystallization of thick MOF membrane. 9.2.1.5

Assembly of MOF Nanocrystals

The interesting advantages of this route over other conventional routes are the precise control of MOF crystalline size and the existence of mesoporous voids. These two key features enhance analyte diffusion into the membrane framework, which is useful for many applications. In this method, well-defined MOF nanocrystalline seeding particles were used for the fabrication of MOF membranes after surface dip-coating technique. The case study regarding this approach was published by Sanchez, Serre, and coworkers [65–67]. They prepared three MOF membranes, namely, ZIF-8, MIL-89, and MIL-101[Cr]. Nano-sized (22 nm) MIL-101[Cr] crystals were produced in stable colloidal solution to ethanol in which pristine silicon wafer was dip coated. The resulting thick MOF layer depends upon solution concentration, growth time, and dip-coating cycles. 9.2.1.6

Chemical Vapor Deposition

In 2016, Ameloot and coworkers [68] fabricated uniform microcrystalline network of ZIF-8 membrane using chemical vapor deposition (CVD) as shown in Figure 9.8. A typical designed substrate with numerous silicon pillar arrays was successfully coated around the surface with well-defined and controlled thickness. A uniform MOF coating on inner or outer silicon surface can be produced by precisely controlling vapor-phase deposition. Schematically, CVD consists of mainly two key steps: (i) deposition of MO layer and (ii) vapor–solid reaction at the substrate surface. In details, the authors used atomic layer deposition (ALD) technique to deposit 3–15 nm uniform ZnO thin film onto silicon pillar array substrate and then ZIF-8 growth started when ZnO-coated substrate surface was exposed to 2-methylimidazole (HmIM) at 100 ∘ C for 30 minutes. Consequently, the highly microcrystalline solid network of ZIF-8-membrane was obtained. 9.2.2 9.2.2.1

Fabrication Methods for SURMOF Membranes Liquid-Phase Epitaxy

One of the most frequently and extensively used SURMOF membrane fabrication technique is called “liquid-phase epitaxy,” developed by Wöll and coworkers [69]

9.2 Fabrication of MOF Membranes

Chemical vapor deposition (CVD) (a)

Vaporized precursors

Vapor processing

Step 1 M

Dense metal oxide

Step 2 O

L

Microporous MOF

(b)

(c) ZnO 100 μm ZIF-8

50 μm

5 μm

20 μm

Figure 9.8 (a) Chemical vapor deposition approach for the ZIF-8 membrane fabrication. The procedure consists of a metal oxide vapor deposition (step 1) and a consecutive vapor–solid reaction (step 2). Scanning electron microscopy (SEM) images of (b) ZIF-8-coated silicon pillar arrays and (c) manufactured ZIF-8 patterns [68]. Source: Copyright 2016, Springer Nature.

in 2007. In this method, SAM-functionalized substrate is immersed into separate solutions of organic ligand and metal precursor rather than mixture of reactants used in conventional solvothermal process. Mainly, there are two key steps involved in this method: (i) functionalization of substrate and (ii) deposition process. Substrate Functionalization Generally, organic functional groups (—COOH,

—OH, —NH2 , and pyridine) based on SAM substrates are used because these groups help anchor metal, metal oxide, or organic linkers, enhance adhesion,

297

298

9 MOF Membranes and Their H2 Separation Properties

promote growth of MOF membrane, and lead appropriate orientation with structural perfection. Despite this, several studies support this worthy argument. Proper growth orientation of SURMOF membrane is highly influenced by (i) types of functional groups and (ii) the density of functional groups [70–72]. Separation of Reactants In contrast to the one-pot solution used in conventional

solvothermal approach, separate constituents are adopted in LPE by separately immersing the SAM-modified substrate into solutions. The mysterious concept in this approach is the bulk cluster of metal or metal oxide acting as nodes that anchor organic ligands sequentially in an LBL approach. This methodology offers well-defined, highly oriented, and compact SURMOF at nanometer scale, with high homogeneity and controlled thickness at optimal reaction conditions. Due to high demand of SURMOFs at industrial scale, numerous new LPE strategies have been developed for SURMOF membrane fabrication such as “spin coating,” “dipping robot,” “spray coating,” and “flow-based automation.” Furthermore, typical organic ligands consisting of a N-group at para-position or dicarboxylic group at meta/para-position were adopted for three-dimensional (3D) (pillared SURMOFs) and two-dimensional (2D) SURMOFs, respectively, using LBL approach [73]. The spray setup is quite suitable for HKUST-1 and 2D SURMOFs, [74] while spin coating, dipping robot, and flow-based automation frequently were applied as appropriate tools for 3D SURMOF fabrication [75]. 9.2.2.2

Interfacial Synthesis

The high-quality freestanding MOF membranes were synthesized by new methodology known as “interfacial synthesis.” In this versatile method, 2D MOF membranes are produced via self-coordination of metal ions and organic linkers at two different phases such as liquid–liquid interface and air–liquid interface. Liquid–Liquid Interfacial Synthesis 2D MOF membranes synthesized using

liquid–liquid interfacial approach was first developed by Ameloot and coworkers in 2011 [76]. As apparent from the name, here two immiscible solvents consisting of metal ions and organic ligands are brought into contact at optimal conditions, and then crystallization occurs at the interface level via self-coordination chemistry, which can result in high-quality 2D freestanding MOF membranes. The nucleation rate only depends upon MOF precursor diffusion rate in respective immiscible solvent. A schematic illustration is shown in Figure 9.9. After that, this strategy have been widely used for the various 2D MOFs such as 2D MOF-2 [64], ZIF-8 [76, 78, 79], MIL-53(Al) [80], Cu3 (BTC)2 [76, 81–83], and [Cu2 (ndc)2 (dabco)3 ] [84] (ndc = 1,4-naphthalenedicarboxylate and dabco = 1,4-diazabicyclo[2.2.2]octane). Beyond 2D MOF membranes, Ameloot et al. [76] applied the same methodology for the synthesis of different-shaped 3D hollow capsule MOF membranes. Air–Liquid Interfacial Synthesis Makiura et al. [77] developed a novel strategy

for 2D MOF membrane synthesis called “air–liquid interfacial approach.” This approach provides another excellent platform to construct high-quality 2D MOF system. The authors published the synthesis of NAFS-13 (MOF) nanosheets

(a)

HO

(b)

O

C Metal ions Aqueous solution

∼16.6 Å

N O

N

N

OH

N

O

Pd N

HO

∼16.6 Å

O Pd Cu

MOF crystals

O

O

OH

PdTCPP, 1

X-ray

b a

C Organic solution (c)

H

Cu(NO3)2·3H2O aqueous solution (d)

Liquid 1

Post-injection of concentrated Cu(NO3)2·3H2O aqueous solution

∼3Å

PdTCPP Liquid 2 MOF capsules

Pure water subphase

Single-layer nanosheets

Figure 9.9 Liquid–liquid interfacial synthesis: (a) interfacial preparation of an MOF layer using a biphasic synthesis mixture consisting of an aqueous metal-ion-containing solution (blue) and an organic ligand solution (purple). Crystallite formation takes place primarily at defects remaining in the layer, resulting in self-completing growth. Transparent lighter blue octahedra correspond to new nucleation events resulting from bond formation between metal ions and ligand molecules that meet [76]. (b) Schematic diagram of the assembly process of PdTCPP-Cu MOF nanosheets (NAFS-13) at the air–liquid interface with the crystalline structure of NAFS-13 shown in the inset [77]. (c) Both immiscible liquids are supplied by syringe pumps to a T-junction, where the formation of aqueous solution droplets takes place in the continuous organic phase [76]. (d) Schematic diagram of the post-injection methodology employed in the fabrication of NAFS-13 nanosheets [77]. Source: Copyright 2011 and 2013, Springer Nature.

300

9 MOF Membranes and Their H2 Separation Properties

using this approach. In this particular synthesis, firstly, an ethanolic solution of PdTCPP (5,10,15,20-tetrakis(4-carboxyphenyl)-porphyrinato-palladium(II)) was subsequently spread onto the 1.0 mM aqueous solution of Cu(NO3 )2 ⋅3H2 O. The tetratopic molecular building units (PdTCPP) is oriented on a water subphase and coordinated with Cu metal nodes, thus initiating NAFS-13 crystallization in the form of nanosheets lying flat on the air–liquid interface. Moreover, these 2D NAFS-13 nanosheets further linked to binuclear Cu2 (COO)4 linker and then produced microscale sheet about 300 nm thick, as shown in Figure 9.9. Langmuir–Blodgett (LB) Layer-by-Layer Deposition Ultrathin MOF membranes

fabricated using the “bottom-up approach” were first developed by H. Kitagawa, R. Makiura, and coworkers [85–90]. The nucleation orientation and uniform thickness of these membranes with metalloporphyrin molecular building units and metal ion nodes can be controlled by a combined concept of LBL growth and Langmuir–Blodgett (LB) strategy. A complete demonstration of this methodology with reference to one example is illustrated in Figure 9.10a. R. Makiura and coworkers [86] successfully prepared a NAFS-1-based membrane using this approach. In the first step, a chloroformic/methanolic solution containing molecular building unit organic linkers CoTCPP (5,10,15,20-tetrakis(4-carboxyphenyl)-porphyrinato-cobalt) and pyridine (Py) was spread onto a layer of aqueous solution containing metal ions (CuCl2 ⋅2H2 O). Subsequently, 2D nanosheets of CoTCPP-Py-Cu were crystallized at the interface of chloroform/methanol and aqueous layer by self-coordination assembly. In the second step, as-synthesized CoTCPP-Py-Cu nanosheets (NAFS-1) were transferred onto silicon substrate using Langmuir–Blodgett (LB) approach. The same process was repeated many times, producing stacked NAFS-1 nanofilm comprising a number of nanosheets in parallel fashion. X-ray diffraction (XRD) analysis confirmed the crystallinity and the homogeneity of NAFS-1 membrane and indicated the proper structural orientation along [001] direction and exact 2D stacked nanosheets with an average tilting angle of 0.3∘ . Langmuir–Schäfer (LS) Layer-by-Layer Deposition Unlike the fabrication of 2D

MOF membranes using liquid–liquid interface, air–liquid interface, and LB LBL deposition, Zhang et al. [91, 92] produced an ultrathin membrane using surfactant-assisted approach via Langmuir–Schäfer (LS) method. This new method permits the fabrication of ultrathin bimetallic MOF membrane with a uniform thickness of about 10 nm, such as M-TCPP(Fe) (M = Co, Cu, and Zn and TCPP (Fe) = Fe(III)tetra(4-carboxyphenyl)porphine chloride). During this particular synthesis, polyvinylpropylene (PVP) used as capping agent/surfactant was introduced, which modulated the anisotropic nucleation growth of Co-TCPP(Fe) nanosheets. A schematic illustration is shown in Figure 9.10b and c.

9.3 Controlling and Characterizing MOF Membranes Due to the very small thickness of SURMOF membrane compared with polycrystalline one, SURMOFs raise certain characterization issues, including

9.3 Controlling and Characterizing MOF Membranes

Building unit O

Top view of 2D array

HO

HO

O N N Co N N

O

OH OH

O

CoTCPP

N

Py

2D array formation by LB method and deposition on substrate

Rinse/immerse

CuCl2·2H2O, aqueous solution

(a)

Substrate

Layer-by-layer growth

Co2(COO)4

2D Co-TCPP(Fe) nanosheets

TCPP(Fe) Co

Fe

PVP

(b) Substrate MOF suspension

(c)

Water

MOF thin film on water

SURMOF thin film

Langmuir–Schafer transfer

Figure 9.10 (a) Schematic illustration of the fabrication of NAFS-1 membrane with bottom-up modular assembly by combining layer-by-layer technique with Langmuir–Blodgett method [86]. Source: Copyright 2010, John Wiley & Sons. (b) Schematic illustration of the surfactant-assisted synthesis of 2D Co-TCPP(Fe) nanosheets. The MOF layers are drawn in alternating blue and red to clarify the layered structures. (c) Schematic illustration of the assembly process for the preparation of 2D nanosheet-based MOF membranes [91].

301

302

9 MOF Membranes and Their H2 Separation Properties

crystalline structure, morphology, and porosity. In this perspective, more fundamental MOF characterization techniques are applied to both SURMOF and polycrystalline membranes, including powder X-ray diffraction (PXRD) to verify the crystallinity and phase purity and nitrogen adsorption and desorption to endorse porosity and apparent surface area of the material. Additionally, thermogravimetric analysis (TGA) is to check thermal stability; scanning electron microscopy (SEM) micrographs measure crystal size and morphology; energy-dispersive X-ray spectroscopy (EDS) analyzes the elemental composition and distribution; nuclear magnetic resonance (NMR) spectroscopy determines the bulk purity and quantification of mixed MOFs; solid-state nuclear magnetic resonance (SS-NMR) spectroscopy is a useful tool to probe chemical environment of specific functional in MOF samples; inductively coupled plasma optical emission spectroscopy (ICP-OES) is applied to investigate the elemental ratio and purity in MOFs; diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) is used for the confirmation of infrared (IR)-active functional groups in MOF membrane. It is noted that all these characterization techniques are applied for a specific purpose. In this section, we try to describe brief principle but comprehensive relevant characterizations with controlling parameters using suitable examples of MOF membranes. 9.3.1

Powder X-Ray Diffraction

Generally, PXRD pattern provides information about the crystalline structure, crystal size, chemical composition, and physical properties of a material. Once a sample is confirmed to be crystalline, other information can be obtained from the pattern. PXRD pattern is also used for identifying unknown substance by comparing diffraction data with the data base, and only solid-state (powder) substances are used to evaluate crystallographic studies. In the case of MOFs, PXRD pattern is significantly used to determine bulk crystallinity and phase purity of the sample. High intensity peaks verify the structural crystallinity and phase purity can be confirmed by comparing experimental XRD pattern to the simulated pattern generated from computational modeling. Furthermore, sample preparation is also a key process for PXRD analysis. Commonly, the powder sample is loaded onto a flat glass, plastic, or alumina disk holder or affixed using volatile liquid or oil. Although this preparation method works well for the majority of MOF samples, but not for plate- or needlelike crystals, some of these crystals will deposit with a preferred orientation when samples are prepared in this manner. The simulated powder pattern is based on the assumption that crystallites are randomly oriented; therefore, this can cause significant peak intensity differences between experimental and simulated patterns, as shown in Figure 9.11. To overcome the issue of preferred orientation, the MOF sample should be frequently rotated during data collection, because the orientation of crystallites becomes random relative to the detector by spinning the sample holder or capillary tube [40]. Besides, broad diffraction peaks are associated with poor crystallinity or undersized crystallites. Often, full structural determination can also be accomplished using PXRD, but the process is still challenging [93, 95].

9.3 Controlling and Characterizing MOF Membranes

Figure 9.11 Powder X-ray diffraction patterns of platelike MOF crystals (Zn-RPM) collected using conventional method, spinning capillary method, and simulated modeling [93]. Source: Copyright 2016, ACS.

Capillary Conventional Simulated

5

10

15

20

25

30

35

2θ (°)

9.3.2

Nitrogen Adsorption and Desorption

Adsorption–desorption isotherms for nonreactive gases at cryogenic temperature can be applied to determine apparent surface area, total pore volume and pore accessibility, or pore size for MOFs and other materials. Typically, nitrogen is used due to its inert nature, well-established molecular size, reasonable cost, and availability in high purity. To obtain better information, the MOF sample should be properly activated using solvent exchange method or vacuum drying or supercritical drying method before the analysis, and the amount of the sample is another key point before running operation. As a rule of thumb, the amount of sample (in gram) multiplied by surface area (m2 g−1 ) should be equal to 100 m2 or more to achieve reliable data. Finally, the data analysis protocol is a crucial choice because if the analysis is done incorrectly, then the calculated surface area either is underestimated or overestimated [96]. At present, the best practice for apparent surface area calculation is to use Brunauer–Emmett–Teller (BET) theory, because the pore sizes of most MOFs support multilayer gas adsorption mechanism. The Langmuir theory is valid for monolayer gas adsorption [97, 98], consequently tending to overestimate the apparent surface area, in some cases more than 50% for most MOFs. In 1938, Stephen Brunauer, Paul Hugh Emmett, and Edward Teller published an article about BET theory. So, BET theory named after the initials of the founders’ first names derived the BET equation used to determine the apparent surface area by plotting [(p/p0 )/n(1 − p/p0 )] versus p/p0 , where n indicates the amount of N2 gas adsorbed, p is pressure, and p0 is saturated pressure of nitrogen gas at 77 K. From the obtained plot, the BET surface area was calculated from the isotherms of micro/meso-ZIF-8 crystals prepared by Zhang et al. [99]. N2 adsorption–desorption isotherms and pore distributions for these ZIF-8 materials were measured and are shown in Figure 9.12. The BET surface areas and total pore volumes were 1058.5 m2 g−1 and 0.69 cm3 g−1 for micro-ZIF-8, but 1094.7 m2 g−1 and 1.08 cm3 g−1 for micro/meso-ZIF-8, respectively. The

303

9 MOF Membranes and Their H2 Separation Properties

Figure 9.12 Experimental N2 adsorption–desorption isotherms of micro-ZIF-8 and micro/meso-ZIF-8 [99]. Source: Copyright 2017, RSC.

Volume adsorbed (cm3 g–1)

800 micro-ZIF-8

700

meso-ZIF-8

600 500 400 300 200 100 0 0.0

0.2

0.4

0.6

0.8

1.0

Relative pressure (p/p0)

mesoscopic pores inside micro/meso-ZIF-8 are introduced by a surfactant (CTAB) and can be demonstrated from hysteresis loops at a p/p0 value greater than 0.8. There are several models that can be used to analyze PSD and pore volume of MOF membrane including density functional theory (DFT) model and Barrett–Joyner–Halenda model, but the former is most widely accepted for microporous ( 60 psig) or concentration of feed gas CO2 is high. Here, propylene carbonate is used for the purification of carbon dioxide because of its polar nature and high affinity for CO2 . Some advantages such as no heating for regeneration of solvent, high CO2 solubility, and high loading capacity make it more useful strategy. However, some disadvantages including very expensive cost and high affinity to heavy hydrocarbons of some Fluor solvents lower the absorbent efficiency. 10.2.6

Physical Adsorption

The adsorption process relies on the thermodynamic properties of a porous substance, where the gas diffuses and attaches with the chelating sites of an adsorbent. This adsorption can be either “physisorption” or “chemisorption.” The selective removal of CO2 from the gas stream to the solid porous material is followed by the regeneration/desorption of adsorbed gases, which can be achieved either by decreasing the pressure or increasing the temperature or electric swing process or washing with specific inert liquids. Up to date, several porous materials have been reported regarding CO2 selective capture and storage ability. For example, “molecular sieves” are a range of especially designed sieves that separate gaseous molecules based on their molecular size. This technology is believed to be cost effective and can be adapted to a variety of carbon sequestration processes [15]. Later on, much successful efforts were performed to improve the CO2 adsorption–desorption capacity by chemical post-functionalization of molecular sieves. Generally, adsorbents based on high surface area, unique porosity, and basic functional organic groups (like amines) into the channel walls are of particular interest. This is due to robust interaction between the basic surface of the adsorbent and acidic CO2 gaseous molecules, resulting in high adsorption affinity for the acidic gas. Mesoporous silica substrates such as SBA-1 [29], SBA-15 [17], MCM-41 [29–31], and MCM-48 [32] are attractive because they possess mesoporous channel capillaries that are large enough to be accessed for the chemical functionalization and guest gaseous molecules. Both the unique porosity and surface-functionalized groups mainly facilitate the capture and adsorption of CO2 . Moreover, Chaffee et al. [33] developed new porous solid adsorbents for vacuum swing adsorptive (VSA) based CO2 separation from mixture of flue gas. The solid adsorbents are capable to insensitively operate above ambient temperature range. The authors mainly focused on development of inorganic–organic hybrid materials, where the mesoporous inorganic substances offered both pore volumes and high surface area for improved host–guest interactions [34, 35]. Activated carbons or carbonized carbons have well-developed micro-/ mesoporous materials that are applied in a broad range of industrial processes [36]. Heteroatoms (oxygen, nitrogen, etc.) present on the surface of activated

10.2 CO2 Capture and Separation Strategies

carbons can be acidic, basic, or neutral functional groups [37]. In order to enhance the specific host–guest interaction, the surface of activated carbons can be modified by doping of nitrogen atoms into the matrix, causing an increase in the number of basic groups, charge distribution, and high adsorption affinity of CO2 [38]. Furthermore, Maroto-Valer et al. [39] studied the carbon dioxide capture behavior of steam-activated anthracite. The observed adsorption capacity of the activated anthracite decreased swiftly with increasing temperature. The highest CO2 adsorption capacity was 65.7 mg g−1 at 800 ∘ C for two hours with a surface area of 540 m2 g−1 . The inactivated anthracite with a surface area of 1071 m2 g−1 exhibited CO2 adsorption capacity of about 40 mg CO2 per adsorbent gram. Meanwhile, surface functionalization of the activated anthracite with ammonia (NH3 ) and polyethylenimine (PEI) increased carbon dioxide adsorption affinity at higher temperature, due to the incorporation of nitrogen groups on the surface. Lithium zirconate (Li2 ZrO3 ) and lithium silicate (Li4 SiO4 ) with favorable carbon dioxide adsorption characteristics have been investigated as high temperature CO2 absorbents [40–42]. The large adsorption capacity and wide range of temperature stability make lithium silicate a strong candidate for developing commercially competitive CO2 adsorbent. In the last decade, the progress of developing novel and efficient CO2 adsorbents continued, and most dynamic porous materials were discovered in the arena of gas separation and storage, for example, porous organic polymers (POPs), crystalline organic frameworks (COFs), MOFs, and membranes with outstanding CO2 adsorption ability. 10.2.7

Cryogenic

The cryogenic purification technology involves the separation and liquefaction of desired gases on the basis of fractional condensation and distillation at lower temperature range (−150 ∘ C/123 K/238 ∘ F). This is a commercial process commonly used to purify carbon dioxide from high-purity sources (> 90%). Here, the gases are cooled to a very low temperature so that carbon dioxide can be liquefied and separated from the mixture. The process has the advantage that it permits pure CO2 recovery in the liquid state, which can be transported conveniently to the required site for enhanced oil recovery (EOR). In this perspective, Hart et al. [43] reported cryogenic carbon dioxide capture from natural gas using “CryoCell” technology. This technology eliminates the usage of water consumption, chemicals, and corrosion-related issues. This development indicates improved economic viability of carbon dioxide separation. 10.2.8

Membrane Technology

In this technology, a particular membrane is applied for the separation of specific gases from a gaseous stream. This can be CO2 separation from the mixture of flue gas (post-combustion system), CO2 /H2 separation (pre-combustion system), CO2 /natural gas separation (natural gas system), or N2 /O2 separation (oxy-fuel combustion). The membranes are available in different material types including polymeric membranes (organic system), inorganic membranes (porous

331

332

10 CO2 Capture with MOF Membranes

(a)

(b)

(c)

(d) Retentate

Membrane Selective separation Permeate

Figure 10.5 Schematic representation of separation mechanisms: (a) Knudsen diffusion, (b) molecular sieving, (c) solution–diffusion, and (d) gas separation membrane process [44–47]. Source: OSTI, ACS, Annual Reviews, and WAP.

and nonporous system), silica membranes, zeolite membranes, and MOF membranes, and they are able to separate gaseous components by five possible mechanisms including Knudsen diffusion, molecular sieving, solution–diffusion separation, surface diffusion, and capillary condensation [44–47], of which the first three are schematically shown in Figure 10.5a–c. Membranes are semipermeable barriers that act as filters to separate one or more gases from a feed mixture and generate a specific gas-rich permeate as shown in Figure 10.5d. Two properties hold the membrane performance: permeability, that is, the flux of a specific gas through the membrane, and selectivity, the membrane’s preference to pass one gas molecules over the other. Molecular sieving and solution–diffusion mechanisms are frequently applicable for all gas separation membranes, while the Knudsen separation mechanism is only based on the pore sizes of the membrane. Membranes have several advantages over adsorption and absorption strategies such as easy way of modeling, no waste stream, and no regeneration energy required. However, a limitation of membranes is the instability or sensitivity for sulfur compounds and other trace elements. The most typical examples of some membrane materials are the following. 10.2.8.1

Polymeric Membranes

The diffusion of gas molecules through a polymeric membrane is followed by a solution–diffusion mechanism. The other types of membranes obey the molecular sieve and Knudsen diffusion mechanisms [48]. Commonly, there are two types of polymeric membranes, namely, rubbery membrane and glassy

10.2 CO2 Capture and Separation Strategies

membrane, both operating relative to the glass transition temperature (T g ) of the polymer [49]. The former operates above the glass transition temperature (T g ) of the membrane, and the latter operates below T g . The polymer chains are packed defectively, leading to extra free volume as microscopic voids in the polymeric membrane. Within these cavities, the Langmuir adsorption of gases occurs that raises the solubility. Normally, membrane permeability is inversely proportional to the membrane area required for gas separation. Therefore, high permeability leads to lower cost. For most membranes, there is trade-off between gas selectivity and permeability. Commonly, a highly permeable membrane tends to have low selectivity, and vice versa. To overcome this issue is by improving the performance of the CO2 polymeric membrane that is achievable by two approaches: first is by increasing the solubility of carbon dioxide in the membrane through changes in polymeric composition, and second is by increasing the diffusion of CO2 through varying the polymer packing or cavities within the membrane. The combination of these two approaches has produced an extensive range of efficient polymeric membranes with reasonable permeability and selectivity for carbon dioxide separation [50–53]. 10.2.8.2

Inorganic Membrane

There are two main types of inorganic membranes: porous membranes (alumina, carbon, glass, silicon carbide, titania, zeolite, and zirconia membranes) and nonporous membranes (supported on different substrates such as α-alumina, γ-alumina, zeolite, or porous stainless steel). Porous inorganic membranes are usually inexpensive but less selective. The membrane performance can be improved through surface modification by appropriate functional groups. These functional groups exhibit high chemical affinity for carbon dioxide, and thus the pore channels become more saturated, which increases the permeability. 10.2.8.3

Zeolite Membrane

Zeolites are inorganic crystalline structures with uniform pore sizes and molecular dimensions. Generally, small-, medium-, and large-pore zeolites have been used to fabricate membranes that separate carbon dioxide (CO2 ) from CH4 [54–56]. Since both CO2 (0.33 nm kinetic diameter) and CH4 (0.38 nm) molecules are much smaller than the cavities of large-pore and medium-pore zeolites, separation was mainly based on competitive adsorption. In the case of ZSM-5 membranes, the CO2 /CH4 separation selectivity observed at room temperature was 2.4–5.5 [57]. In the case of Y-type membranes, the CO2 /CH4 separation selectivity was about 10 [58]; for X-type membranes, the CO2 /CH4 separation selectivity as high as 28 was obtained [56]. In contrast, small-pore membranes such as zeolite T (pore size 0.41 nm), DDR (0.36 × 0.44 nm), and SAPO-34 (0.38 nm) have pore sizes similar to CH4 but larger than CO2 . Better CO2 /CH4 selective separation was observed for these membranes due to a combination of differences in diffusivity and competitive adsorption. In this context, another approach was reported by Cui et al., in which they prepared T-type membrane and observed a CO2 /CH4 separation selectivity of 400 and a high CO2 permeance at 308 K [59]. The obtained results revealed that the

333

334

10 CO2 Capture with MOF Membranes

carbon dioxide gas adsorbed more strongly than methane (CH4 ) on SAPO-34 membrane, causing selective separation of both gases. 10.2.8.4

Silica Membrane

Silica membrane with pores smaller than 1.0 nm is appropriate for highly selective membranes. However, diffusion through such narrow pores is usually slow. In order to improve both the performance and the selectivity of silica membranes, several efforts have been reported to control the pore size. Mostly chemical vapor deposition (CVD) and sol–gel method are used to fabricate silica membranes. The latter is often adopted in membrane synthesis or membrane pore modification because of its high controllability and homogeneity [60–62]. In this perspective, Raman and Brinker [63, 64] used silane coupling agent (methyltriethoxysilane [MTES]) for surface modification, and the resulting membrane exhibited both high selectivity and carbon dioxide permeability from mixture of CO2 /CH4 . This is due to the uniform thin top layers with thickness of about 50–200 nm. 10.2.8.5

MOF Membrane

Recently, MOF membranes have been discovered as potential candidate for CO2 capture. The specific structure and chemical properties of MOFs have led to the highest adsorption capacities. An application of MOFs is predicted in so-called mixed matrix membranes (MMMs), which exhibit improved performance in comparison with the pure polymeric membranes. Several types of MOFs have been fabricated in the form of membrane and studied for efficient gas separation than other material’s membrane. A detailed explanation will be discussed in the next section. 10.2.9

Chemical-Looping Combustion

This technology was proposed by Richter and Knoche in 1983 [65]. In this method, reactions are divided into oxidation and reduction reactions. The general operation of CLC system is shown in Figure 10.6. Figure 10.6 Schematic of chemical-looping combustion system.

Combustion products CO2, H2O

Depleted air

Fuel reactor

Air reactor

MeO

Me

Air O2/N2

Fuel

10.3 Chemistry of MOFs for CO2 Recognition

The CLC system has two main reactors, one for air and the other for fuel delivery. The oxygen carriers (metal oxide including CuO, Fe2 O3 , NiO, and Mn2 O3 ) move between the two reactors. In the air reactor, metal oxides are oxidized. In the fuel reactor, the carriers are reduced by the fuel, resulting in CO2 and steam [66]. The amount of energy released or required in the reactors depends on these two reactions. CLC method has many advantages as compared with conventional combustion. The exhaust gas stream from air reactor is harmless, consisting mainly of N2 . In a well-designed system, there should be no thermal formation of NOx , because the regeneration of oxygen carrier occurs without flame and at moderate temperatures. The exhaust gas from the fuel reactor consists of CO2 and H2 O. Separation of CO2 can be performed by a condenser, a prime advantage with CLC that avoids the huge energy penalty necessary in traditional amine-scrubbing process to capture carbon dioxide and thus leads to less operational cost. In this field, Brandvoll and Bolland described that the overall energy penalty could be as low as 400 kJ kg−1 in a natural gas combined-cycle plant, assuming idealized chemical stability of the oxygen carriers (metal oxides) [67].

10.3 Chemistry of MOFs for CO2 Recognition The carbon dioxide (CO2 ) recognition ability of a porous adsorbent is normally evaluated by two key factors: CO2 adsorption capacity and selectivity of the material. Ideal MOF materials with high carbon dioxide capture ability are expected to show both high uptake and high adsorption for CO2 gas over other gases, such as N2 and CH4 . The effectiveness of MOF materials regarding carbon dioxide capture ability is due to the unique structural and chemical features of MOFs. These features include unsaturated metal sites or open metal sites, polar functional groups into the pore channels, alkylamine incorporation, pore size, and function control. The fundamental role of these features in enhancing carbon dioxide capture ability with appropriate examples is explained here. 10.3.1

Unsaturated Metal Sites

Metal atoms in most MOFs are coordinatively saturated through network ligands, but in some MOFs, the metal atoms are partially coordinated by guest solvent molecules. When these solvent molecules are removed, coordinatively unsaturated metal sites or open metal sites are produced within the MOF cavities. These positively charged open metal sites can act as Lewis acid that strongly polarizes carbon dioxide gas molecules in the pores and causes more CO2 capture ability. Several reports on this type of MOFs have been published for the capture and separation of carbon dioxide gas [68–71]. The most exciting member of MOF family is MOF-74, and the series of metal MOF-74 (metal = Mg, Mn, Fe, Co, Ni, Cu, Zn) is a typical example where the unsaturated open metal sites within the pores can play a significant role for high CO2 adsorption capacity and selectivity [72]. The entire MOFs are constructed by similar organic ligand with

335

336

10 CO2 Capture with MOF Membranes

different metal salts, which create the one-dimensional (1D) hexagonal channels with exposed unsaturated metal sites in honeycomb framework. Among these MOFs, Mg-MOF-74 exhibits benchmark results with the highest CO2 adsorption capacity at the low to moderate partial pressures relevant to CO2 capture from flue gas. Even though several MOFs with open metal sites have been studied for CO2 capture, the strategies for the precise control on structure and property at the molecular level are still lacking. In this aspects, Zhou et al. reported the novel idea of “single-molecule trap” (SMT) with confined unsaturated metal sites into the pore for capturing single CO2 molecule [73]. The building units of SMTs are constructed from three-dimensional MOF to ensure the efficient adsorption and desorption of CO2 molecules as shown in Figure 10.7. The distance between the two opposite open metal sites in each SMT is 7.4 Å, which is suitable for the accommodation of one CO2 molecule through the interaction of its two O atoms to the open Cu(II) sites. This strategy dramatically increases the carbon dioxide capture ability and selectivity. Therefore, the method of carefully designing an SMT and therefore its extension into porous frameworks can be applied for the synthesis of other new MOFs and composite materials for CO2 separation through confined cavity or pores with strongly binding active sites. 10.3.2

Polar Functional Groups

One of the most promising characteristics of MOFs is the possibility to tailor the active functionality into the MOF network, giving rise to high adsorption

CO2 trapping

Coordination assembly SMT

MOF framework

Figure 10.7 Schematic representation of the design and construction of an SMT for CO2 adsorption [73]. Source: Copyright 2013, Springer Nature.

10.3 Chemistry of MOFs for CO2 Recognition

selectivities and capacities of the desired gaseous species and reducing the regeneration energy. Introducing polar functional groups that have robust interaction with CO2 in the pore surfaces of MOFs via direct synthesis or post-synthesis approaches offers a great opportunity to increase the adsorption capacity and selectivity of CO2 . Up to date, several MOFs with different polar functionalities have been reported with high CO2 uptake and selectivity. Among these, aromatic amine (—NH2 ) is considered as a promising functional group to increase the CO2 adsorption capacity and selectivity of MOFs due to its strong affinity with CO2 . Numerous MOFs with aromatic amine ligands have been designed such as Zn2 (atz)2 (ox) and bio-MOF-11 as representative examples. The observed CO2 uptake capacity of Zn2 (atz)2 (ox) at 273 K and 1.2 bar is 4.35 mmol g−1 , and the Qst of CO2 was measured to be 40.8 kJ mol−1 [74]. Moreover, the aromatic amine functionalities in Zn2 (atz)2 (ox) are open to the adsorbate within the channel walls, making hydrogen bonds with CO2 , which significantly improve its heats of adsorption and selectivity over other gases. This observation was observed through single-crystal X-ray diffraction studies of the same MOF [75]. It should be noted that integrating functional groups into MOF network always leads to partial occupancy of pore voids and in turn lessens the surface areas and pore volume. A useful approach to relieve this problem is known as “functionality insertion method,” referring to tailoring MOFs with desired functional groups, which are introduced into backbones of organic linkers in MOFs [76]. In this arena, Bai, Zaworotko, and coworkers has performed modification of MOF (NJU-Bai0) channel walls by acylamide groups using the functionality insertion method. The calculated BET surface area is about 3160 m2 g−1 , slightly smaller than that of its analogous MOF, PCN-61 (BET = 3355 m2 g−1 ) [77]. The gas adsorption results revealed that the NJU-Bai0 exhibited a CO2 uptake of 23.53 mmol g−1 at 298 K, 20 bar, higher than that of its analogous MOF (PCN-61). The Qst of NJU-Bai0 was estimated to be 26.3 kJ mol−1 , higher than that of PCN-61 (22 kJ mol−1 ), suggesting a stronger CO2 –framework affinity in NJU-Bai0. Furthermore, the observed ideal adsorption solution theory (IAST) selectivity of the CO2 /N2 gas mixture at 298 K and 1 bar for NJU-Bai0 is 22; this is also higher than that of PCN- 61 (15) at the same conditions as shown in Figure 10.8. The authors believed that amide (—CONH) functional groups have an optimistic effect on the CO2 adsorption process by increasing the initial slopes of CO2 isotherms for NJU-Bai0, resulting in higher Qst , uptake capacity, and selectivity for CO2 adsorption over other gases. To further ameliorate the CO2 capture and adsorption capacity of the acylamide-functionalized MOFs, the research group investigated two expanded isoreticular rht-type MOFs prepared from nano-sized triangular acylamide-bridging hexacarboxylate linkers such as [Cu3 (BTB6− )]n and [Cu3 (TATB6− )]n . The CO2 uptake capacity of both MOFs is of 157 and 111 wt% at 20 bar, at 273 and 298 K, respectively, which is slightly lesser than those of MOF-177 (123.2 wt%) and MOF-205 (114.4 wt%). These outcomes confirm that the insertion of acylamide functional groups within the framework can offer robust interaction sites and play a significant role in highly selective CO2 uptake [78].

337

10 CO2 Capture with MOF Membranes 28

24 20

1

1

PCN-61

PCN-61

(CO2)

26

PCN-66

1

24

PCN-68

Qst (kJ mol−1)

–1

CO2 or N2 uptake (mmol g )

338

16 12 8

PCN-61

18 16

12 10

0

(a)

20

14

(N2)

4

22

0

5

15

10

20

Pressure (bar)

O2C

0

5

10

15

20

–1

(b)

CO2 uptake (mmol g )

CO2

CO2

O2C

NH

O

O O2C

H N

CO2 O2C CO2

(c)

btei6– in PCN-61

CO2

(d)

CO2 HN O

CO2

TPBTM6– in 1

CO2

Figure 10.8 (a) CO2 and N2 sorption isotherms of 1 and the PCN-6X series at 298 K, (b) isosteric heats of CO2 adsorption for 1 and PCN-61, (c) chemical structure of ethyne ligand, and (d) chemical structure of amide-based ligand [77]. Source: Copyright, 2010, ACS.

Recently, a Zr-based MOF named UiO-66 has attracted intense interest for its high stability and good adsorption ability toward carbon dioxide. Many reports have shown that the introduction of polar functional groups to the ligand could significantly improve the CO2 adsorption and separation ability of UiO-66. For example, Serre, Weireld, Maurin, and coworkers reported a —COOH-functionalized hydrostable Zr-MOF (UiO-66(Zr)–(COOH)2 ) for the capture and separation of carbon dioxide from mixture of gases [79]. The gas adsorption experiments revealed that using a 15 : 85 CO2 : N2 gas mixture at 303 K and 1.0 bar exhibited a CO2 /N2 selectivity of 56 for UiO-66(Zr)–(COOH)2 and Qst values of 34.8 and 17.8 kJ mol−1 for CO2 and N2 , respectively, at 303 K. The incorporation of functional groups into the UiO-66 network significantly increases the CO2 capture, and separation ability, however, will significantly reduce the CO2 adsorption capacity due to the relative small pore size and surface area of UiO-66. Therefore, UiO-67 is a more promising platform for functional groups with improved CO2 adsorption capacity and separation abilities due to its larger pore size or cavities and surface area than those of UiO-66 [80]. The gas adsorption results revealed that the maximum CO2 uptake of UiO-67, BUT-10, and BUT-11 at 298 K and 1 bar was 22.9, 50.6, and 53.5 cm3 g−1 , respectively. It is quite exciting that CO2 uptakes in BUT-10 and

O

60

BUT-11 BUT-10 UiO-67

40

O

HOOC

COOH O

HOOC

COOH

HOOC

COOH

20 0 0.0

0.2

0.4 0.6 P (bar)

0.8

CO2 adsorption

BUT-11

Selectivity of CO2 / N2

CO2 uptake at 298K (cm3 g–1)

10.3 Chemistry of MOFs for CO2 Recognition

60 45 30 15 0 0.0

1.0

ZrCl4

analogues

BUT-10

BUT-11 BUT-10 UiO-67

0.2

0.4 0.6 P (bar)

0.8

1.0

CO2 separation

UiO-67

Figure 10.9 CO2 sorption isotherms and separation curves of UiO-67, BUT-10, and BUT-11 at 298 K and 1.0 bar [80]. Source: Copyright, 2014, ACS.

BUT-11 are two times more than that of UiO-67, suggesting that introduced polar functional groups in the ligands are able to significantly increase the CO2 adsorption capacities of the resulting MOFs as shown in Figure 10.9. 10.3.3

Pore Size and Function Control

Molecular sieving can enable ultrahigh selectivity by excluding gaseous molecules with kinetic diameters larger than the pore size (e.g. N2 , 3.64 Å; CH4 , 3.80 Å) while allowing passage of smaller gaseous molecules (e.g. CO2 , 3.30 Å). Unfortunately, due to the difficulty in controlling pore size within the 3–4 Å range, which is most appropriate for gas molecule separations, molecular sieving has rarely been achieved. In this perspective, Eddaoudi, Zaworotko, and coworkers reported the carbon dioxide uptake study on three isoreticular MOFs including SIFSIX-2-Cu, SIFSIX-2-Cu-i, and SIFSIX-3-Zn [81]. The observed pore sizes of SIFSIX-2-Cu, SIFSIX-2-Cu-i, and SIFSIX-3-Zn are 13.05, 5.15, and 3.84 Å, respectively. SIFSIX-2-Cu displays a carbon dioxide uptake of about 1.84 mmol g−1 at 298 K and 1.0 bar; however, the polymorph of SIFSIX-2-Cu, SIFSIX-2-Cu-i, exhibits significantly higher values of 5.41 mmol g−1 under the same conditions. The IAST selectivities of SIFSIX-2-Cu-i for CO2 : CH4 gas mixture (50 : 50) and CO2 : N2 gas mixtures (10 : 90) at room temperature are 33 and 140, respectively, which are dramatically higher than those of SIFSIX-2-Cu (5.3 and 13.7, respectively) under similar conditions. More importantly, the SIFSIX-3-Zn framework with a smaller pore size even shows much higher selectivities (495 and 109 for 10 : 90 CO2 : N2 and 50 : 50 CO2 : CH4 gas mixtures,

339

340

10 CO2 Capture with MOF Membranes

respectively) than those of SIFSIX-2-Cu-i at the same conditions. Moreover, column breakthrough experiments show that the selectivities of CO2 for a 30 : 70 CO2 : H2 gas mixture of SIFSIX-2-Cu-i and SIFSIX-3-Zn are 240 and 1800, respectively, which indicate that these two MOFs are mainly apposite for CO2 separation from syngas. Afterward, Eddaoudi et al. studied another high-performing isoreticular MOF, SIFSIX-3-Cu, for the low concentration of CO2 removal [82]. The resulting MOF is structurally similar to SIFSIX-3-Zn. It is found that SIFSIX-3-Cu framework displays even steeper adsorption isotherms at very low pressure as compared with SIFSIX-3-Zn, because of the relative stronger bonding between the Cu(II) and the ligand pyrazine into the framework, and exhibits smaller pore size (3.5 Å) than the SIFSIX-3-Zn (3.84 Å), indicating relatively stronger CO2 -SIFSIX-3-Cu framework interactions than SIFSIX-3-Zn. On the other side, the kinetic measurements reveal that the carbon dioxide gas adsorbs more swiftly and strongly than other gaseous molecules. Recently, Zhang, Zaworotko, and coworkers reported the well-controlled pore sizes of five MOFs including Qc-5-Cu-sql-α and Qc-5-metal-dia (metal = Co, Ni, Zn, and Cu). Both copper-based frameworks such as Qc-5-Cu-dia and Qc-5-Cu-sql-α exhibit 1D channels with diameters of 4.8 and 3.8 Å, respectively [83], and the pore size of Qc-5-Cu-sql-β framework after the desolvation process is about 3.3 Å as shown in Figure 10.10. Gas adsorption investigations show that the Qc-5-Ni-dia and Qc-5-Cu-dia reveal adsorption capacity of N2 (7.1 and 5.8 cm3 g−1 ) and CH4 (25.6 and 24.6 cm3 g−1 ) at 293 K and 1.0 bar relative pressure, which are much higher than those of Qc-5-Cu-sql-β (N2 = 0.3 cm3 g−1 ; CH4 = 1.3 cm3 g−1 ). Conversely, Qc-5-Cu-sql-β exhibits a higher CO2 uptake (48.4 cm3 g−1 ) than Qc-5-Cu-dia at 293 K and 1.0 bar relative pressure. The IAST

Desolvation

Qc-5-Cu-dia

Pore size: 4.8 Å

Qc-5-Cu-sql-α

Pore size: 3.8 Å

Qc-5-Cu-sql-β

Pore size: 3.3 Å

Figure 10.10 Pore size tuning from supramolecular isomerism in diamondoid (dia) and square lattice (sql) polymorphs of [Cu(Qc)2 ]n . C (gray), Cu (maroon), O (red), N (blue), and H (white) [83]. Source: John Wiley & Sons.

10.3 Chemistry of MOFs for CO2 Recognition

selectivities for a 15 : 85 CO2 : N2 mixture and a 50 : 50 CO2 : CH4 gas mixture of Qc-5-Cu-sql-β are 40 000 and 3 300, respectively, which are much higher than those of Qc-5-Cu-dia (19 and 3) and Qc-5-Ni-dia (36 and 7). The Qst of CO2 at zero coverage of Qc-5-Cu-sql-β is 36 kJ mol−1 , higher than that of Qc-5-Ni-dia (32 kJ mol−1 ) and Qc-5-Cu-dia (34 kJ mol−1 ). 10.3.4

Core–Shell MOF Structure

Another promising strategy to improving the CO2 adsorption capacity and selectivity is the construction of core–shell MOF materials. The core would adsorb carbon dioxide gas, while the shell would perform as a gas sieve and a protection against water. To further insight into the knowledge of this material strategy, Rosi and coworkers described the understanding of such a strategy using bio-MOF-11 as core and bio-MOF-14 as shell [84]. The two MOFs are isoreticular structures with the only difference in the coordinated monocarboxylate to the SBU. The coordinated monocarboxylate in bio-MOF-11 is acetate, while that in bio-MOF-14 is valerate. Gas adsorption results depicts that bio-MOF-11 has a high adsorption capacity for CO2 and low hydrostability, while bio-MOF-14 has a low capacity for CO2 and high CO2 /N2 separation selectivity at 273 and 298 K and is stable in water. Core–shell material for selective CO2 capture would utilize the qualities of bio-MOF-11 (high CO2 adsorption capacity) and bio-MOF-14 (high CO2 /N2 selectivity and water stability), which is represented as bio-MOF-11@bio-MOF-14. Gas adsorption studies reveal that the observed CO2 adsorption capacities at 1.0 bar and 273 K are 92, 58.3, and 44.8 for the core, the core–shell material, and the shell, respectively. The core–shell material adsorbs 30% more CO2 than pure shell material (bio-MOF-14). The reason is because the core–shell materials exhibit more porosity than individual shell. Moreover, the core–shell structure exhibits a much lower N2 uptake at 77 K than the core material and an uptake only slightly higher than that of shell structure. These observations suggest that the bio-MOF-14 shell efficiently averts N2 uptake by the core. In addition, after being soaked in water for 24 hours, the crystallinity of core–shell is retained, indicating that the bio-MOF-14 shell plays a role in protecting the water-sensitive core structure. 10.3.5

Alkylamine Incorporation

Twenty to thirty percent aqueous solutions of alkylamines are usually used in large scale to capture CO2 from industrial streams. Because of the high heat capacity of these solutions, regeneration of solution bears much energy and costs. On the other side, one basic advantage of porous solid materials is the low regeneration energy as compared with that required for aqueous amine solutions. However, this advantage often comes at the expense of low adsorption capacity and poor selectivity. To minimize this problem, several efforts have been performed to develop novel, economical, and efficient adsorbents, for instance, introducing functional sites into the porous framework for high affinity to the gaseous molecules at low pressure and using less regeneration energy as compared with alkylamine solution. In addition, the open metal sites of MOFs

341

342

10 CO2 Capture with MOF Membranes

offer a great opportunity for post-synthetic modification by grafting amine into the pore channel to enhance adsorption and selectivity, being promising for the bridging of the two approaches discussed above. Long and coworkers successfully incorporate N,N ′ -dimethylethylenediamine (mmen) in series of MOFs based on M2 (dobpdc) with extended framework of MOF-74 where M = Mg, Mn, Fe, Co, Ni, or Zn [85]. All prepared MOF materials exhibit exceptionally high CO2 adsorption capacities at low pressures except for the Ni-based MOF as shown in Figure 10.11. For example, Mg-MOF-74 has a carbon dioxide uptake of 2.0 mmol g−1 at 0.39 mbar and 25 ∘ C, conditions relevant to removal of CO2 from air, and 3.14 mmol g−1 at 0.15 bar and 40 ∘ C, conditions relevant to CO2 capture from flue gas. Gas sorption isotherms suggest that the high adsorption capacity of mmen-Mg2 (dobpdc) at relatively low pressure is an extraordinary cooperative process in which CO2 molecules insert into metal–amine bonds. Such type of insertion is known as cooperative insertion of carbon dioxide gas molecules in the diamine-appended MOF. The aforementioned studies are based on the incorporation of “diamine” groups into the unsaturated metal sites of MOFs. Besides these, some “polyamine groups” were also used for the same purpose in the MOF network; the resulting MOFs are more promising than diamine-based MOFs because of more amine groups in the polyamines. In this perspective, Kong, Chen, and coworkers prepared a series of PEI-functionalized MOFs with low molecular weight linear PEI loaded into MIL-101(Cr) for highly efficient CO2 capture [86]. As the PEI concentration (%) increases, the observed BET surface area decreases in PEI-MIL-101(Cr). 125 wt% is the highest amount that can be loaded into the porous material. The CO2 uptake isotherms of unmodified MIL-101 and modified PEI-MIL-101 samples examined at 50 ∘ C and 1.0 bar indicated that all polyamine-based MOF (PEI-MIL-101) samples exhibited significantly enhanced CO2 adsorption capacity as compared with the unmodified MIL-101. Mainly, the observed CO2 adsorption capacity of a modified PEI-MIL-101 with 125 wt% PEI loading was 3.95 and 4.51 mmol g−1 , four times than the unmodified MIL-101 (0.20 and 1.00 mmol g−1 ) at 50 ∘ C, 0.15 and 1 bar, respectively. Moreover, the CO2 : N2 ideal selectivities for a 15 : 75 CO2 : N2 gas mixture of PEI-MIL-101-125 were up to 770 and 1200 at 25 and 50 ∘ C, respectively, which were among the highest values of CO2 : N2 selectivities for zeolite- and MOF-based adsorbents listed in the literature. Later on, Yaghi and coworkers reported that the ligand of IRMOF-74-III can be modified with primary amine (IRMOF-74-III-CH2 NH2 ) through ligand functionalization and used for the selective capture of CO2 . Due to the strong interactions between CO2 and aliphatic amine functionalities into the cavities, the resultant IRMOF-74-III-CH2 NH2 possesses the highest uptake capacity at low CO2 pressure ( N2 > CH4 > CO2 . In the case of binary gaseous mixtures, the separation factor of H2 /CO2 reached up to 72 at 298 K and 1.0 bar. It was believed that carbon dioxide gas can be trapped in the pore structure of ZIF-100, while other gas molecules easily permeate through the membrane because of the high CO2 adsorption affinity and small window aperture (3.35 Å) of ZIF-100 framework. ZIF-78 is a GME topological structure with a cage size of 7.1 Å and a pore window size of 3.8 Å [124]. In this particular MOF membrane, Jin et al. fabricated a ZIF-78 membrane on porous zinc oxide substrate using the reactive seeding strategy [125]. The authors applied a novel “solvent activation-exchange method” to activate the ZIF-78 layer, removing the macroscopic defects and the intercrystalline voids in the membrane. The ideal selectivity factor of H2 /CO2 is 11.0 for the ZIF-78 membrane at room temperature. The low permeance of CO2 relative to that of hydrogen was attributed to the strong interaction between CO2 molecules and ZIF-78 network. The observed binary gas mixture separation factor of H2 /CO2 was found to be 9.5 with a H2 permeance of 0.97 × 10−7 mol m−2 s−1 Pa−1 at 298 K and 1.0 bar. On the other side, Eddaoudi et al. report an exclusive combination of pure inorganic zeolite and MOFs, named as ZMOFs, which exhibit accessible extra-large cavities, chemical stability, and ion-exchange ability [126]. They also fabricated a continuous sod-ZMOF-1 membrane on a porous alumina support. The sod-ZMOF-1 has a sodalite topological anionic framework with extra-framework cations in the pore. The resultant membrane displayed exciting permselectivity for CO2 over other tested gases. The ideal selectivities of CO2 /He, CO2 /H2 , CO2 /N2 , CO2 /O2 , and

347

348

10 CO2 Capture with MOF Membranes

CO2 /CH4 are 3.2, 2.6, 8.7, 5.1, and 3.6, respectively, with a low CO2 permeance of 0.0063 × 10−7 mol m−2 s−1 Pa−1 . Moreover, the gas mixture separation factor of CO2 /CH4 is 4.0 with a CO2 permeance of 0.0049 × 10−7 mol m−2 s−1 Pa−1 at 308 K and 1.0 bar. They believed that the anionic atmospheres of the membrane increased the interactions with the CO2 due to small pore window size (4.1 Å), resulting in high permselectivity of CO2 [127]. Besides ZIFs and ZMOF-based membranes, some metal-carboxylate MOF-based membranes also have been fabricated for gas separations. For example, Peng et al. reported a pressure-assisted room temperature growth approach to fabricate continuous and intergrown HKUST-1 membrane on a PVDF hollow fiber support [128]. In the procedure one end of the PVDF hollow fiber was linked to the pump, and the other side was sealed by Teflon tape. Copper hydroxide nanostrands (CHNs) were filtered onto the PVDF hollow fiber under vacuum pump and operated as a copper source for HKUST-1. The PVDF hollow fiber with a CHN layer was then poured into ligand solution for nucleation under room temperature for about 40 minutes, which resulted in a continuous HKUST-1 layer firmly adhered on the PVDF hollow fiber support. The ideal separation factor of H2 /CO2 for the HKUST-1 membrane is 4.2 from single-component gas permeation tests, which is near to the Knudsen selectivity, because the pore size of HKUST-1 (9.0 Å) is much larger than the kinetic diameters of H2 (2.9 Å) or CO2 (3.3 Å). The separation factor of H2 /CO2 for 50 : 50 binary gases is 8.1 with a H2 permeance of 19.7 × 10−7 mol m−2 s−1 Pa−1 . The authors ascribed this selectivity between CO2 and H2 to sorption and size effect. Later on, the same research group also fabricated a freestanding HKUST-1 membrane and HKUST-1 membranes on other substrates and achieved improved CO2 separation and selectivity [129, 130]. Another acknowledgeable effort was reported by Li et al., who developed a new facile and versatile oriented nanomicrostructure assisted controllable fabrication (ONACF) strategy to fabricate some MOF membranes. In this method, a HKUST-1 membrane was fabricated on oriented Ni3 S2 nanomicrostructure array modified Ni foam by in situ growth. During nucleation, a new phase of copper (Cu) was formed in the initial stage on the Ni3 S2 nanoarrays on Ni foam (Figure 10.13), which was believed to show a crucial role in the fabrication of continuous HKUST-1 membrane. Gas mixture permeation investigations showed that the H2 permeance of the HKUST-1 membrane was 27.2 × 10−7 mol m−2 s−1 Pa−1 with a H2 /CO2 selectivity of 6.8 [131]. Zhang, Meng, and coworkers fabricated a well-intergrown HKUST-1 membrane on rigid polyacrylonitrile (PAN) hollow fiber supports by a chemical modification approach. The support was premodified by hydrolyzing cyano (—CN) functional groups of PAN into carboxyl (—COO) groups, which offered nucleation active sites for growing crystals into a compact layer and enhance the adhesion between the MOF layer and the PAN hollow fiber. The as-synthesized membrane exhibited an ultrahigh H2 permeance of 705 × 10−7 mol m−2 s−1 Pa−1 with a H2 /CO2 separation factor of 7.14 for binary mixtures with high thermal and pressure stability [132]. A continuous and crack-free HKUST-1 membrane was fabricated on polysulfone hollow fiber support, which was premodified with a polydimethylsiloxane (PDMS)/HKUST-1 layer. The authors supposed

10.4 Membrane Design for CO2 Separation

(a)

(b)

300 μm (c)

5 μm (d)

5 μm

20 μm

Figure 10.13 SEM images of (a) Ni foam, (b) Ni3 S2 nanoarrays on Ni foam, (c) Cu/Ni3 S2 microstructure on Ni foam, and (d) HKUST-1 membrane on the modified Ni foam [131]. Source: John Wiley & Sons.

that the mixed PDMS/HKUST-1 layer had four key purposes: (i) serving as seeds for the growth of a continuous and well-intergrown MOF layer, (ii) improving the adhesion of the MOF layer, (iii) improving the capacity of gas separation, and (iv) reducing the mass transfer resistance as compared with entire MMMs or pure polymer layers. The ideal separation factors of H2 /CO2 and N2 /CO2 for the HKUST-1 membrane reached 21.03 and 7.23, respectively, with higher permeances of H2 (4.85 × 10−7 mol m−2 s−1 Pa−1 ) and N2 (1.66 × 10−7 mol m−2 s−1 Pa−1 ). Afterward, they fabricated a continuous and compact NH2 -MIL-53(Al) membrane on macroporous glass frit disks support with pre-covered colloidal seeds [133]. The observed gas permeances were in the order of H2 > CH4 > N2 > CO2 at 288 K. The H2 permeance was 26.71 × 10−7 mol m−2 s−1 Pa−1 , and the H2 /CO2 ideal separation selectivity is 27.3. The gas mixture separation factor of H2 /CO2 is 30.9 with a H2 permeance of 15 × 10−7 mol m−2 s−1 Pa−1 . Similarly, Zhang, Meng, and coworkers also fabricated a continuous NH2 -MIL-53(Al) membrane layer on an ammoniated PVDF hollow fiber support [134]. The obtained NH2 -MIL-53(Al) membrane was well intergrown with a thickness of about 8 μm and showed a H2 permeance of 54.2 × 10−7 mol m−2 s−1 Pa−1 with an ideal separation factor H2 /CO2 of 30.37. The membrane could also efficiently separate the H2 /CO2 gas mixture with a H2 permeance of 42.13 × 10−7 mol m−2 s−1 Pa−1 and a separation factor of 32.35, which signified the highest H2 /CO2 separation factors for reported polymer-supported MOF membranes. Subsequently, CAU-1 made of Al(III) metal centers and 2-amino-1,4-benzene dicarboxylic acid was processed into an amino-functionalized MOF membrane, which seems interesting for carbon dioxide capture due to the acid–base interaction between CO2 and amino functionalities [135]. Yang and coworkers

349

350

10 CO2 Capture with MOF Membranes

(a)

(b)

10 μm (c)

10 μm (d)

10 μm

10 μm

Figure 10.14 SEM images of the CAU-1 membrane: (a) upper view and (b) lateral view. EDS mapping of the CAU-1 membrane; green C; red Al (c and d) [136]. Source: Copyright 2014, RSC.

fabricated a continuous CAU-1 membrane on asymmetric α-alumina tube using the secondary seeded growth approach (Figure 10.14) [136]. For the single gas permeation, the permeance of CO2 for the CAU-1 membrane was 13.2 × 10−7 mol m−2 s−1 Pa−1 , which was higher than other tested gases (H2 , CH4 , N2 , and SF6 ). The ideal selectivities of CO2 /H2 , CO2 /N2 , CO2 /CH4 , and CO2 /SF6 systems were 2.6, 26.2, 14.8, and 79.3 respectively, suggesting that the permeation of CO2 through the CAU-1 membrane was directed by the preferential adsorption. They further estimated the separation performance of the CAU-1 membrane for CO2 /N2 gas mixtures at different concentrations of carbon dioxide. At a low CO2 molar fraction of 0.1–0.2 (flue gas composition), the observed CO2 permeance was 5.0–5.7 × 10−7 mol m−2 s−1 Pa−1 with the separation factor of CO2 /N2 of about 17.4. However, at a high CO2 molar concentration of 0.9, the observed CO2 permeance was 13.4 × 10−7 mol m−2 s−1 Pa−1 and a CO2 /N2 separation factor of 22.8. Similarly, the authors also fabricated a continuous CAU-1 membrane on a α-alumina hollow ceramic fiber support using the secondary seeded growth approach [137]. In contrast, the obtained CAU-1 membrane exhibited preferential permeation of H2 over CO2 , N2 , and CH4 . The mixture separation factor of H2 /CO2 is 12.34 with a H2 permeance of 1.8 × 10−7 mol m−2 s−1 Pa−1 at room temperature. Moreover, Li, Yang, and coworkers lately fabricated a CAU-10-H membrane with a thickness of about 6 μm on α-alumina substrate using in situ solvothermal approach [138]. CAU-10-H is constructed from Al(III) metal centers and the V-shaped linker 1,3-benzenedicarboxylate and exhibits an outstanding

10.4 Membrane Design for CO2 Separation

hydrothermal stability [139, 140]. The obtained CAU-10-H membrane displayed a H2 /CO2 separation factor of 10.5 and a long-term hydrothermal stability. Furthermore, the effect of feed pressure and temperature on the membrane performance was examined for the ternary mixture (H2 /CO2 /H2 O) of 41 : 41 : 18. Overall, the maximum H2 /CO2 and H2 /H2 O mixed gas separation factors were 11.1 and 5.67, respectively, with a H2 permeance of 0.153 × 10−7 mol m−2 s−1 Pa−1 at 423 K and 1.0 bar. It is well known that M-MOF-74 materials exhibit high carbon dioxide adsorption capacity and selectivity due to their uncoordinated metal centers [141]. For the same MOF material, Kim et al. prepared a Ni-MOF-74 membrane on α-alumina substrate using a layer-by-layer seeding approach followed by secondary growth crystallization technique [142]. The observed ideal selective factors of H2 /CO2 , CH4 /CO2 , and N2 /CO2 are 9.1, 3.2, and 3.0, respectively, for Ni-MOF-74 membrane. The consequence was described by arguing that Ni-MOF-74 has higher adsorption affinity than other examined gases, with the permeation of CO2 being dominated by surface diffusion. Similarly, Caro, Huang, and coworkers used magnesium oxide (MgO) as seeds to fabricate a continuous Mg-MOF-74 membrane [143]. Post-functionalization of the Mg-MOF-74 membrane with ethylenediamine offered enhanced separation performance of the Mg-MOF-74 membrane. The H2 /CO2 mixture selectivity improved from 10.5 to 28 at 298 K. It was believed that the functionalization of the Mg-MOF-74 membrane with ethylenediamine significantly increased the CO2 adsorption affinity of the MOF membrane and in turn decreased the permeance of CO2 . Another well-known and stable MOF material, MOF-5-based membrane, was fabricated by Lin and coworkers on porous α-alumina substrate using the secondary seeded growth approach [144]. As-prepared MOF-5 membrane revealed significant permeation for CO2 over H2 or N2 with a separation factor of near to 5 and 60 for CO2 /H2 and CO2 /N2 at room temperature. Similarly, Huang and coworkers fabricated a highly hydrophobic and permselective Zn(BDC)(TED)0.5 membrane using the secondary seeded growth strategy for H2 /CO2 separation [145]. The gas separation results for an equimolar H2 /CO2 gas mixture, the hydrogen permeance of the membrane was 27 × 10−7 mol m−2 s−1 Pa−1 , and the selectivity was about 12.1. It was believed that the Zn(BDC)(TED)0.5 membrane exhibited a H2 /CO2 permselectivity because of the preferential adsorption affinity and capacity to CO2 as well as a highly porous structure with large channels of the network. Tunable pore size is one of the most vital advantages for MOF membranes in gas separation. Kang et al. fabricated an ultramicroporous JUC-150 membrane (0.25 nm × 0.45 nm) on a Ni support using the secondary seeded growth approach [146]. The JUC-150 membrane displayed high thermal and mechanical stability and reusability and preferential permeance of hydrogen relative to other tested gases, which mainly based on pore size of the network. It was believed that only hydrogen could pass through the JUC-150 membrane network with a permeance of 2.96 × 10−7 mol m−2 s−1 Pa−1 from the single-component gas permeation tests. The mixture separation factor of H2 /CO2 was about 38.7. Furthermore, Jin and coworkers fabricated a well-intergrown [Ni2 (mal)2 (bpy)]

351

10 CO2 Capture with MOF Membranes

(Ni-MB) homochiral membrane on α-alumina substrate using the secondary seeded growth approach, in which high-energy ball milling was employed to prepare nano-sized MOF seeds [147]. The single gas permeances of H2 , N2 , CH4 , and CO2 for the Ni-MB membrane were 1.549, 0.481, 0.446, and 0.017 × 10−7 mol m−2 s−1 Pa−1 , respectively, at room temperature and 1.0 bar pressure. The observed ideal selectivities of H2 /CO2 , N2 /CO2 , and CH4 /CO2 were 89, 27, and 25 respectively. Such high permselectivity was suggested to be responsible by strong interactions between carbon dioxide and Ni-MB framework. Because of the quite similar molecular sizes of CH4 and CO2 , it is challenging to attain their separation through shape selectivity in a membrane process as 18

16 CO2

10–6

14

S. F.

12

CH4

10–7

Separation factor

Permeances (mol m–2 s–1 pa–1)

10–5

sweep rate = 50 ml min–1 10

feed flow rate = 50 ml min–1 feed pressure = 1.0 × 10 Pa 5

10

–8

8 0

10

(a)

20 30 40 50 Permeance temperature (°C)

60

16 14 12 Separation factor

352

10 8

sweep rate = 50 ml min–1 feed flow rate = 50 ml min–1

6

feed pressure = 1.0 × 105 Pa permeance temperature = 25 °C

4 2 0 0 (b)

5

10 15 Test time (h)

20

25

Figure 10.15 (a) CO2 /CH4 permeances and separation factor (SF) for the Co3 (HCOO)6 membrane versus the permeation temperature and (b) plot of the CO2 /CH4 separation factor for the same membrane as a function of test time [102]. Source: John Wiley & Sons.

10.5 Conclusions

mentioned above. With their composition of varied metal ions and organic ligands, MOFs have exhibited the selective adsorption of these two gases, providing hope that they might be used to separate CO2 from CH4 through a preferential adsorption. For instance, the authors reported a microporous, continuous, and well-intergrown membrane of Co3 (HCOO)6 on a macroporous glass frit substrate using a secondary growth approach for the separation of CO2 /CH4 mixtures [102]. The obtained membrane exhibited diamondoid connectivity as well as high thermal stability and the overall framework gave rise to 1D zigzag channels with an effective pore size of 5.5 Å. This channel system was appropriate for CO2 separation from CH4 . The gas separation behavior and the possible mechanism were inferred on the basis of sorption isotherms of CO2 and CH4 connected with in situ IR measurements. As the permeance shown in Figure 10.15, the macroporous glass frit-supported Co3 (HCOO)6 membrane exhibited a high permeation flux (2.09 × 10−6 mol m−2 s−1 Pa−1 ) and excellent permeation selectivity for CO2 over CH4 (in the range of 10.37–15.95) at 273–333 K. The CO2 molecules diffused easily through the zigzag channels of the membrane as compared with the bulky CH4 molecules. This result was due to the preferential adsorption mode of CO2 in the micropores and at the external surfaces of the Co3 (HCOO)6 membrane, averting the CH4 sorption from the mixture. The effective pore size of the Co3 (HCOO)6 membrane (5.5 Å) combined with the pore shape did not permit two distinct molecules to pass through simultaneously, such that once CO2 molecules diffused through the pore channels of the framework, the diffusion of CH4 molecules was blocked.

10.5 Conclusions In this chapter, separation and capture strategies, MOF chemistry, and MOF designing have been overviewed for adsorption and separation of CO2 gas. Various striking features of MOFs such as unsaturated active sites, polar end groups, shape, pore size, and flexibility have been outlined to improve the CO2 capture and separation performance. Besides, core–shell structure of MOF and incorporation of different type of alkylamines as ligands into the MOF network could significantly alters the CO2 separation performance. This chapter also dealt with the membrane designing for the CO2 separation, and detailed examples in each catalogue of MOF membranes have been included to elucidate the structure–function relationship. It is revealed that several aspects of the nature of the MOF membranes including structural properties leading to phase transitions during adsorption/desorption, strong electrostatic as well as weaker multipoint adsorbate–active site interactions, and response to stimuli can be properly tuned to achieve specific designs targeting improved selectivity for CO2 separation. Up till now, selectivity and permeability are not straightforward in function of the extent of adsorption. Many other factors affect the molecular or mass transport, thus identifying the synergistic effects of the structural or chemical microenvironments, and the complexity of multicomponent mixtures

353

354

10 CO2 Capture with MOF Membranes

is very useful for fundamental understanding of MOF membrane-based CO2 separation. In addition, membrane stability, material cost, separation reliability, and production scalability are determinant factors for pilot use of MOF membranes in practical CO2 separation or capture.

References 1 Fanchi, J.R. and Fanchi, C.J. (2017). Energy in the 21st Century, 350. World

Scientific Publishing Co. Inc. 2 IPCC (2005). Intergovernmental Panel on Climate Change Special Report

3 4 5 6 7

8 9 10 11 12 13 14 15 16 17 18 19 20 21

on Carbon Dioxide Capture and Storage. Cambridge: Cambridge University Press. Smit, B., Reimer, J.R., Oldenburg, C.M., and Bourg, I.C. (2014). Introduction to Carbon Capture and Sequestration. Imperial College Press: London. Haszeldine, R.S. (2009). Science 325: 1647. Olajire, A.A. (2010). Energy 35: 2610–2628. Jeremy, D. (2000). Economic evaluation of leading technology options for sequestration of carbon dioxide. MS thesis, MIT, Cambridge, MA. Audus, H. (2000). Leading Options for the Capture of CO2 at Power Stations. Fifth International Conference on Greenhouse Gas Control Technologies, Cairns, Australia (13–16 August 2000). Irons, R., Sekkapan, G., Panesar, R., et al. (2007). CO2 capture ready plants. IEA Greenhouse Gas R & D Programme. www.ieagreen.org.uk. IEA special report on carbon dioxide capture and storage (2005). www.ipcc.ch Energy Information Administration (EIA) (1998). Natural gas issues and trends. Adams, D. and Davison, J. (2007). IEA Greenhouse Gas R & D Program, 17. ISBN: 978-1-898373-41-4. Croiset, E. and Thambimuthu, K.V. (2000). Can. J. Chem. Eng. 78: 402–407. White, C.M., Strazisar, B.R., Granite, E.J., and Hoffman, J.S. (2003). J. Air Waste Manage. Assoc. 53: 645–715. Brooks, L.A. and Audrieth, L.F. (1946). Inorg. Synth. 2: 85–92. Stewart, C. and Hessami, M. (2005). Energy Convers. Manag. 46: 403–420. Bai, H. and Yeh, A.C. (1997). Ind. Eng. Chem. Res. 36: 2490–2493. Gray, M.L., Soong, Y., Champagne, K.J. et al. (2005). Fuel Process. Technol. 86: 1449–1455. Idem, R., Wilson, M., Tontiwachwuthikul, P. et al. (2006). Ind. Eng. Chem. Res. 45: 2414–2420. Aronu, U.E., Svendsen, H.F., Hoff, K.A., and Juliussen, O. (2009). Energy Procedia 1: 1051–1057. Fauth, D.J., Frommell, E.A., Hoffman, J.S. et al. (2005). Fuel Process. Technol. 86: 1503–1521. Melien, T. (2005). Capture and Separation of Carbon Dioxide from Combustion Sources, vol. 1, 47–87. Elsevier.

References

22 Duncan, J. (2003). Electro-catalytic oxidation (ECO): results from pilot scale

23

24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46

testing of simultaneous NOx, SO2 , Hg and PM 2.5 removal at first energy’s R.E. Burger plant. Proceedings of 20th Annual International Pittsburgh Coal Conference, Pittsburgh, PA. NPCC (2000). Study on CO2 Sequestration by Spray Concentrated Aqueous Ammonia and Production of Modified NH4 HCO3 Fertilizer e a Proposal for US e China Joint Research. China: State Engineering Technology Research Centre of Combustion of Power Plants (NPCC). Tomita, T., Nakayama, K., and Sakai, H. (2004). Microporous Mesoporous Mater. 68: 71–78. Yeh, A.C. and Bai, H. (1999). Sci. Total Environ. 228: 121–133. Rackley, S.A. (ed.) (2017). Absorption capture systems. In: Carbon Capture and Storage, 2e, 115–149. Elsevier Inc. Weiss, H. (1998). Gas Sep. Purif. 2: 171–176. Knowles, G.P., Graham, J.V., Delaney, S.W., and Chaffee, A.L. (2005). Fuel Process. Technol. 86: 1435–1448. Yoshitake, H., Yokoi, T., and Tatsumi, T. (2003). Chem. Mater. 15: 4536–4538. Song, C. (2006). Catal. Today 115: 20–32. Xu, X., Song, C., Miller, B.G., and Scaroni, A.W. (2005). Fuel Process. Technol. 86: 1457–1472. Huang, H.Y., Yang, R.T., Chinn, D., and Munson, C.L. (2003). Ind. Eng. Chem. Res. 42: 2427–2433. Chaffee, A.L., Knowles, G.P., Liang, Z. et al. (2007). Int. J. Greenhouse Gas Control 1: 11–18. Knowles, G.P., Beyton, V., and Chaffee, A.L. (2006). Am. Chem. Soc., Div. Fuel Chem. 51: 102–103. Liang, Z., Fadhel, B., Schneider, C.J., and Chaffee, A.L. (2006). Am. Chem. Soc., Div. Fuel Chem. 51: 159–161. Bansal, R.C., Donnet, J., and Stoeckli, F. (1998). Active Carbon. Marcel Dekker Inc. Arenillas, A., Rubiera, F., Parra, J.B. et al. (2005). Appl. Surf. Sci. 252: 619–624. Plaza, M.G., Pevida, C., Arias, B. et al. (2008). J. Therm. Anal. Calorim. 92: 601–606. Maroto-Valer, M.M., Tang, Z., and Zhang, Y. (2005). Fuel Process. Technol. 86: 1487–1502. Yeboah, Y.D., Xu, Y., Sheth, A. et al. (2003). Carbon 41: 203–214. Essaki, K., Nakagawa, K., Kato, M., and Uemoto, H. (2004). J. Chem. Eng. Jpn 37: 772–777. Kato, M., Nakagawa, K., Essaki, K. et al. (2005). Int. J. Appl. Ceram. Technol. 2: 467–475. Hart, A. and Gnanendran, N. (2009). Energy Procedia 1: 697–706. Shekhawat, D., Luebke, D.R., and Pennline, H.W. (2003). A review of carbon dioxide selective membrane: a topical report. Luebke, D., Myers, C., and Pennline, H. (2006). Energy Fuel 20: 1906–1913. Lackner, K.S. (2002). Annu. Rev. Energy Environ. 27: 193–232.

355

356

10 CO2 Capture with MOF Membranes

47 Frizsche, A. and Kurz, J. (1990). The Separation of Gases by Membranes,

559–593. William Andrew Publishing. 48 Powell, C.E. and Qiao, G.G. (2006). J. Membr. Sci. 279: 1–49. 49 Paul, D.R. and Yampol, skii, Y.P. (1993). Polymeric Gas Separation

Membranes, 115–208. CRC Press. 50 Stern, S.A. (1994). J. Membr. Sci. 94: 1–65. 51 Illing, G., Hellgardt, K., Wakeman, R.J., and Jungbauer, A. (2001). J. Membr.

Sci. 184: 69–78. 52 Xu, Z.K., Dannenberg, C., Springer, J. et al. (2002). J. Membr. Sci. 205:

23–31. 53 Pixton, M.R. and Paul, D.R. (1995). J. Polym. Sci. B Polym. Phys. 33:

1135–1149. 54 Li, S., Falconer, J.L., and Noble, R.D. (2004). J. Membr. Sci. 241: 121–135. 55 Li, S., Martinek, J.G., Falconer, J.L., and Noble, R.D. (2005). Ind. Eng. Chem.

Res. 44: 3220–3328. 56 Hasegawa, Y., Tanaka, T., Watanabe, K. et al. (2002). Korean J. Chem. Eng.

19: 309–313. 57 Poshusta, J.C., Falconer, J.L., and Noble, R.D. (1999). J. Membr. Sci. 160:

115–120. 58 Kusakabe, K., Kuroda, T., Murata, A., and Morooka, S. (1997). Ind. Eng.

Chem. Res. 36: 649–656. 59 Cui, Y., Kita, H., and Okamoto, K.I. (2004). J. Mater. Chem. 14: 924–931. 60 Kim, H.S., Han, J.W., Chin, K.Y. et al. (1995). Korean J. Chem. Eng. 12:

405–411. 61 Roh, H.S., Chang, J.S., and Park, S.E. (1999). Korean J. Chem. Eng. 16: 62 63 64 65 66 67 68 69 70 71 72 73 74 75

331–339. Lee, K.H., Cho, Y.K., and Han, K.W. (1995). Hwahak Konghak 33: 570–572. Brinker, C.J., Sehgal, R., Hietala, S.L. et al. (1994). J. Membr. Sci. 94: 85–92. Raman, N.K. and Brinker, C.J. (1995). J. Membr. Sci. 105: 273–279. Richter, H.J. and Knoche, K. (1983). ACS Symp. Ser. 235: 71–85. Bao, L. and Trachtenberg, M. (2005). Chem. Eng. Sci. 60: 6868–6875. Brandvoll, O. and Bolland, O. (2004). J. Eng. Gas Turbines Power 126: 316–321. Wang, Q., Bai, J., Lu, Z. et al. (2016). Chem. Commun. 52: 443–452. Gao, W.Y., Palakurty, S., Wojtas, L. et al. (2015). Inorg. Chem. Front. 2: 369–372. Zhang, Z., Yao, Z.Z., Xiang, S., and Chen, B. (2014). Energy Environ. Sci. 7: 2868–2899. Wang, X.S., Chrzanowski, M., Kim, C. et al. (2012). Chem. Commun. 48: 7173–7175. Queen, W.L., Hudson, M.R., Bloch, E.D. et al. (2014). Chem. Sci. 5: 4569–4581. Li, J.R., Yu, J., Lu, W. et al. (2013). Nat. Commun. 4: 1538. Vaidhyanathan, R., Iremonger, S.S., Dawson, K.W., and Shimizu, G.K.H. (2009). Chem. Commun. 45: 5230. Vaidhyanathan, R., Iremonger, S.S., Shimizu, G.K.H. et al. (2010). Science 330: 650–653.

References

76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109

Gao, W.Y., Yan, W., Cai, R. et al. (2012). Chem. Commun. 48: 8898–8900. Zheng, B., Bai, J., Duan, J. et al. (2011). J. Am. Chem. Soc. 133: 748–751. Zheng, B., Yang, Z., Bai, J. et al. (2012). Chem. Commun. 48: 7025–7027. Yang, Q., Vaesen, S., Ragon, F. et al. (2013). Angew. Chem. Int. Ed. 52: 10316–10320. Wang, B., Huang, H., Lv, X.L. et al. (2014). Inorg. Chem. 53: 9254–9259. Nugent, P., Belmabkhout, Y., Burd, S.D. et al. (2013). Nature 495: 80–84. Shekhah, O., Belmabkhout, Y., Chen, Z. et al. (2014). Nat. Commun. 5: 4228. Chen, K.J., Madden, D.G., Pham, T. et al. (2016). Angew. Chem. Int. Ed. 55: 10268–10272. Li, T., Sullivan, J.E., and Rosi, N.L. (2013). J. Am. Chem. Soc. 135: 9984–9987. McDonald, T.M., Mason, J.A., Kong, X. et al. (2015). Nature 519: 303–308. Lin, Y., Yan, Q., Kong, C., and Chen, L. (2013). Sci. Rep. 3: 1859. Knight, H. (2010). New Sci. 206: 44–47. Makogon, Y.F. (2010). J. Nat. Gas Sci. Eng. 2: 49–59. R. F. Service (2004). Science 305: 962–963. Lin, H.Q., Van Wagner, E., Raharjo, R. et al. (2006). Adv. Mater. 18: 39–44. Xu, W., Blazkiewicz, P., and Fleming, S. (2001). Adv. Mater. 13: 1014–1018. Liu, L., Chakma, A., and Feng, X.S. (2005). Ind. Eng. Chem. Res. 44: 6874–6882. Weh, K., Noack, M., Sieber, I., and Caro, J. (2002). Microporous Mesoporous Mater. 54: 27–36. Tomita, T., Nakayama, K., and Sakai, H. (2004). Microporous Mesoporous Mater. 68: 71–75. van den Bergh, J., Zhu, W., Gascon, J. et al. (2008). J. Membr. Sci. 316: 35–45. Lindmark, J. and Hedlund, J. (2010). J. Mater. Chem. 20: 2219–2225. Hermes, S., Schroder, F., Chelmowski, R. et al. (2005). J. Am. Chem. Soc. 127: 13744–13745. Liu, Y. (2009). Microporous Mesoporous Mater. 118: 296–301. Takamizawa, S., Takasaki, Y., and Miyake, R. (2010). J. Am. Chem. Soc. 132: 2862–2863. Xie, Z., Li, T., Rosi, N.L., and Carreon, M.A. (2014). J. Mater. Chem. A 2: 1239. Venna, S.R. and Carreon, M.A. (2009). J. Am. Chem. Soc. 132: 76–78. Zou, X., Zhang, F., Thomas, S. et al. (2011). Chem. – Eur. J. 17: 12076–12083. Liu, Y., Hu, E., Khan, E.A., and Lai, Z. (2010). J. Membr. Sci. 353: 36–40. Liu, Y., Zeng, G., Pan, Y., and Lai, Z. (2011). J. Membr. Sci. 379: 46–51. Aguado, S., Nicolas, C.H., Moizan-Basle, V. et al. (2011). New J. Chem. 35: 41–44. Yoo, Y., Varela-Guerrero, V., and Jeong, H.K. (2011). Langmuir 27: 2652–2657. Bohrman, J.A. and Carreon, M.A. (2012). Chem. Commun. 48: 5130–5132. Bao, Z., Yu, L., Ren, Q. et al. (2011). J. Colloid Interface Sci. 353: 549–556. Qiu, S., Ming, X., and Zhu, G. (2014). Chem. Soc. Rev. 43: 6116–6140.

357

358

10 CO2 Capture with MOF Membranes

110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142

Yu, J., Xie, L.H., Li, J.R. et al. (2017). Chem. Rev. 117: 9674–9754. Pan, Y. and Lai, Z. (2011). Chem. Commun. 47: 10275–10277. Pan, Y., Wang, B., and Lai, Z. (2012). J. Membr. Sci. 421: 292–298. Brown, A.J., Brunelli, N.A., Eum, K. et al. (2014). Science 345: 72–75. Guo, H., Zhu, Y., Qiu, S. et al. (2010). Adv. Mater. 22: 4190–4192. Liu, Q., Wang, N., Caro, J., and Huang, A. (2013). J. Am. Chem. Soc. 135: 17679–17682. Huang, A., Liu, Q., Wang, N. et al. (2014). J. Am. Chem. Soc. 136: 14686–14689. Hou, J., Sutrisna, P.D., Zhang, Y., and Chen, V. (2016). Angew. Chem. Int. Ed. 55: 3947–3951. Yao, J. and Wang, H. (2014). Chem. Soc. Rev. 43: 4470–4493. Morris, W., Doonan, C.J., Furukawa, H. et al. (2008). J. Am. Chem. Soc. 130: 12626–12627. Brown, A.J., Johnson, J.R., Lydon, M.E. et al. (2012). Angew. Chem. Int. Ed. 51: 10615–10618. Wang, B., Cote, A.P., Furukawa, H. et al. (2008). Nature 453: 207–211. Huang, A., Chen, Y., Wang, N. et al. (2012). Chem. Commun. 48: 10981–10983. Wang, N., Liu, Y., Qiao, Z. et al. (2015). J. Mater. Chem. A 3: 4722–4728. Banerjee, R., Furukawa, H., Britt, D. et al. (2009). J. Am. Chem. Soc. 131: 3875–3877. Dong, X., Huang, K., Liu, S. et al. (2012). J. Mater. Chem. 22: 19222–19227. Eddaoudi, M., Sava, D.F., Eubank, J.F. et al. (2015). Chem. Soc. Rev. 44: 228–249. Al-Maythalony, B.A., Shekhah, O., Swaidan, R. et al. (2015). J. Am. Chem. Soc. 137: 1754–1757. Mao, Y., Li, J., Cao, W. et al. (2014). ACS Appl. Mater. Interfaces 6: 4473–4479. Guo, Y., Mao, Y., Hu, P. et al. (2016). Chem. Select 1: 108–113. Mao, Y., Cao, W., Li, J. et al. (2013). Chem. Eur. J. 19: 11883–11886. Sun, Y., Yang, F., Wei, Q. et al. (2016). Adv. Mater. 28: 2374–2381. Li, W., Yang, Z., Zhang, G. et al. (2014). J. Mater. Chem. A 2: 2110–2118. Zhang, F., Zou, X., Gao, X. et al. (2012). Adv. Funct. Mater. 22: 3583–3590. Li, W., Su, P., Zhang, G. et al. (2015). J. Membr. Sci. 495: 384–391. Si, X., Jiao, C., Li, F. et al. (2011). Energy Environ. Sci. 4: 4522–4527. Yin, H., Wang, J., Xie, Z. et al. (2014). Chem. Commun. 50: 3699–3701. Zhou, S., Zou, X., Sun, F. et al. (2013). Int. J. Hydrog. Energy 38: 5338–5347. Jin, H., Wollbrink, A., Yao, R. et al. (2016). J. Membr. Sci. 513: 40–46. Reinsch, H., van der Veen, M.A., Gil, B. et al. (2013). Chem. Mater. 25: 17–26. Frohlich, D., Henninger, S.K., and Janiak, C. (2014). Dalton Trans. 43: 15300–15304. Yazaydin, A.O., Snurr, R.Q., Park, T.H. et al. (2009). J. Am. Chem. Soc. 131: 18198–18199. Lee, D.J., Li, Q., Kim, H., and Lee, K. (2012). Microporous Mesoporous Mater. 163: 169–177.

References

143 Wang, N., Mundstock, A., Liu, Y. et al. (2015). Chem. Eng. Sci. 124: 27–36. 144 Zhao, Z., Ma, X., Kasik, A. et al. (2013). Ind. Eng. Chem. Res. 52:

1102–1108. 145 Huang, A., Chen, Y., Liu, Q. et al. (2014). J. Membr. Sci. 454: 126–132. 146 Kang, Z., Xue, M., Fan, L. et al. (2014). Energy Environ. Sci. 7: 4053–4060. 147 Li, Q., Liu, G., Huang, K. et al. (2016). J. Chem. Eng. 11: 60–69.

359

361

11 MOF Membranes for Vapor or Liquid Separation 11.1 Introduction In the chemical industry, the costs associated with separating and purifying products hinder the largest fraction of production [1]. Separation process based on membranes offers greater applicable potential in terms of lower energy consumption, smaller carbon footprint, and ease of operation, compared with other techniques, including adsorption, extraction, and distillation [2]. Among diverse materials used so far for membrane preparation, polymers and zeolites are the types most extensively investigated materials [3]. Polymer membranes have many remarkable advantages, such as ease of processing and low operation cost, which are suitable for industrialized gas separation. However, polymeric membranes are limited by their poor pollution-resistant performances, low chemical and thermal stability, and, in particular, the intrinsic trade-off effect between permeability and selectivity [4]. In comparison, inorganic membranes have distinct superiority, including higher chemical and thermal stability, higher solvent swelling resistance, and more favorable mechanical properties. Quite often, inorganic membranes work well at high temperature under different feed compositions and concentrations. Nevertheless, inorganic membranes are more difficult to be fabricated into large defect-free membranes due to their poor film-forming capabilities and high brittleness [3a, 5]. Metal–organic frameworks (MOFs) or porous coordination polymers (PCPs) are hybrid inorganic–organic materials constructed of metal ions and organic linkers [6]. Their well-defined porosity and tunable chemical functionality enable them great potential to the applications in gas storage, catalysis, and separation. Apart from their use as bulk materials, MOFs are also potential candidates for membrane applications [7]. In the last decade, many efforts have been made to study the applications of MOF materials as membranes for the selective separation of chemicals via pervaporation (PV) and nanofiltration (NF) [8]. Due to the designable structure and functional pore environment, MOF membranes possess desirable separation performances for separation of ions from water (i.e. desalination, heavy metal ion treatment), the separation of organics from organic–aqueous mixtures (i.e. organic removal from groundwater or drinking water, alcohol removal from beer and wine), and separation of organics from organic–organic mixtures (i.e. gasoline desulfurization and Microporous Materials for Separation Membranes, First Edition. Xiaoqin Zou and Guangshan Zhu. © 2019 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2019 by Wiley-VCH Verlag GmbH & Co. KGaA.

362

11 MOF Membranes for Vapor or Liquid Separation

benzene/cyclohexane separation). However, MOF membranes still cannot satisfy the requirements for practical applications at present. On one hand, the poor hydrothermal stability of most MOFs hinders their long-term and practical application in the fields of liquid and vapor environments. Several strategies have been made recently to improve the stability of MOF materials, including surface coating, chemical functionality of ligands, metal and ligand exchanging, etc. [9]. On the other hand, the scale-up of MOF membranes is another challenging issue, as most MOF membranes are prepared on the porous substrates via the solvent thermal method [10]. Therefore, mixed matrix membranes (MMMs), composed of a polymer and inorganic fillers, are proposed to enhance the performance by improving the selective sorption, diffusion, and stability via incorporation of appropriate fillers [11]. The fabrication processes of MMMs are based on well-established polymer membrane technologies [12]. In this chapter, MOF membranes for the selective separation of chemicals by PV, organic solvent nanofiltration (OSN), chiral resolution, and stability improving of MOF membranes will be discussed, and MOF-based MMMs for vapor or liquid separation will be introduced as well (Figure 11.1). Membrane Retentate

Feed

Pump Condenser Permeant Feed Membrane

Retentate

Pervaporation

Condenser Permeant

Polycrystalline MOF membrane Feed

Δp

Organic solvent nanofiltration

Mixed matrix membrane

Permeant Thin-film composite membrane

R/S

R

S R S R S S R R S S S R R R R R R R

Figure 11.1 MOF membranes for vapor or liquid separation.

Chiral resolution

11.1 Introduction

As described and summarized by Professor Koros [13] (“several terminology for membranes and membrane processes”), a basic set of terms applicable to nonliving membranes and membrane processes should be discussed firstly: Penetrant (permeant): Entity from a phase in contact with one of the membrane surface that passes through the membrane. Permeate: Stream containing penetrants that leaves a membrane module. Retentate: Stream that has been depleted of penetrants that leaves the membrane modules without passing through the membrane to the downstream. Permeation flux (J, kg m−2 h−1 ): Number of moles, volume, or mass of a specified component passing per unit time through a unit of membrane surface area normal to the thickness direction: Q J= At where Q is the total weight (kg) of the permeate, A is the effective membrane area (m2 ), and t is the operation time (h). Separation factor (𝛼 i/j ): Ratio of the compositions of components i and j in the permeate relative to the composition ratio of these components in the retentate. Yi∕Y j 𝛼i∕j = X i∕X j where Y i , Y j , X i , X j are mass fractions of i and j in the permeate and feed, respectively. Membrane permeability Pi : Parameter defined as a transport flux, J, per unit transmembrane driving force per unit membrane thickness: Pi =

Ji × l Pi0 − Pil

where J i is the molar flux of component i, l is the thickness of the membranes, and Pio and Pil are the partial pressure of component i on the feed side and the permeate side. Relative recovery, 𝜂 n,B : Amount of substance of a component B collected as a useful product, 𝜂 B,out , divided by the amount of substance of that component entering the process 𝜂 B,in : 𝜂B,out 𝜂n,B = 𝜂B,in Rejection factor, R: Parameter equal to one minus the ratio of the concentration for a component (i) on the downstream and upstream side: ci,downstream R=1− ci,upstream Retention factor, rF : Parameter defined as one minus the ratio of permeate concentration to the retentate concentration of component (i): ci,permeate rF = 1 − ci,retentate

363

364

11 MOF Membranes for Vapor or Liquid Separation

Membrane selectivity, 𝛼: The ratio of the permeability of component i and j (Pi and Pj ): 𝛼=

Pi Pj

11.2 Selective Separation of Chemicals Via Pervaporation Pervaporation, a kind of membrane separation technology, combines membrane permeation and evaporation for selective separation of liquid molecule mixtures [14]. A general process of PV consists of the following steps: 1. Liquid mixture in upstream adsorbed by the membrane. 2. Components in the feed pass through the membrane driven by chemical potential difference due to vacuuming and gas sweeping or thermal difference at the downstream side of the membrane. 3. Permeate vapor desorbs from the downstream side of the membrane [15]. The mass transport process in PV membrane is generally described by solution–diffusion mechanism. Although the phase change from liquid to vapor takes place in PV, only the latent heat of evaporation is required. Therefore, PV is considered as a more energy-saving process compared with broadly utilized distillation operation. Besides, PV process requires no additives (the third components like entrainers or extractants), making it a more environmentally friendly separation process than azeotropic distillation and extractive distillation [16]. PV is applied to four main areas: dehydration of organic solvents, removal of dilute organic compounds from aqueous streams, separation of organic–organic mixtures, and reversible reactions. The solution–diffusion mechanism is the dominant mechanism to predict the performance of dense membranes [17]. It was first proposed to explain the gas permeation through homogeneous membranes. Then, it was accepted by many research fields including PV, NF, reverse osmosis (RO), and dialysis. When the PV membrane is in contact with a liquid feed mixture, there exists solution equilibrium: Cm =K Cfeed where C m and C feed represent the concentrations of a species in the membrane surface and the feed, respectively, and K is for the partition coefficient of a species between the membrane and feed phase. The membrane transport is generally governed by Fick’s first law: dCm d𝛿 where J stands for the permeation flux of a species, D is the diffusion coefficient, and 𝛿 is the position variable. The ideal separation factor 𝛼 of a membrane for J = −D

11.2 Selective Separation of Chemicals Via Pervaporation

species i and j can be defined as 𝛼ji =

Pi D K = i i = (𝛼ji )D (𝛼ji )K Pj Dj Kj

It is concluded from this equation that a maximum separation factor can be achieved by maximizing the differences in P, D, and K. With regard to PV process, the solution and diffusion steps are rate-governed steps since the desorption step usually happens extremely fast [18]. Moreover, it is often assumed that the equilibrium of solution is achieved. Accordingly, Henry’s law (an ideal state of no interaction between permeate and polymer matrix) and Flory–Huggins theory (strong interaction between permeate and polymer matrix) can be applied to calculate the activity coefficient in solution process. Diffusion process in PV normally belongs to anomalous diffusion, which is not exactly obeying Fick’s first law. Therefore, a new practical theory, the free volume theory, was proposed by Fujita and applied to describe the diffusion process of penetrants in polymer matrix [19]. Generally, the diffusion of molecules through polymer networks is attributed to the passage of these molecules through the voids and intermolecular spacing between the polymer chains. Based on this predictive models and solution–diffusion mechanism, a bunch of researchers investigate the mass transport in PV separation. Theoretically, PV, as a similarly sized liquid molecule mixture separation technology, confronts more severe challenges than gas separation technology because of the much stronger interactions between liquid molecules and membrane materials. In general, there are three types of membrane materials for PV: polymeric, inorganic, and hybrid materials. Polymers are firstly and most widely utilized as PV membrane materials owing to their low cost, easy processing, good mechanical stability, and tunable transport properties. However, their poor pollution resistance, low chemical and thermal stability, and in particular the intrinsic trade-off effect between permeability and selectivity extremely limited their applications [16]. Inorganic membranes have higher chemical and thermal stability, higher solvent swelling resistance, and better mechanical properties. Especially, inorganic membranes have better endurance against high temperature under different feed compositions and concentrations [20]. Unfortunately, the large-scale fabrication of defect-free inorganic membranes are challenging, limiting the wide applications [21]. MMMs composed of polymer matrix and inorganic fillers are alternative route for the PV process by combining the selective sorption, diffusion, and stability via incorporation of appropriate fillers. MOFs as a typical branch of inorganics deserve the attention in membrane-based PV in either form of polycrystalline or mixed membranes. 11.2.1

Polycrystalline MOF Membranes for Pervaporation

MOFs are a group of crystalline microporous materials consisting of metal ions linked together by organic ligands. Nowadays, considerable researches have been focused on preparation and application of MOFs, which attract increasing attention due to their diversity and flexibility in structure designing. That offers them great potentials in separations, gas storage, CO2 capture, catalysis,

365

366

11 MOF Membranes for Vapor or Liquid Separation

sensing, etc. A continuous and defect polycrystalline MOF membrane is the first step and important issue for the PV process. As discussed in an early chapter, several methods including in situ and seeding–secondary growth have been successfully applied for the MOF membrane preparation. Besides the preparation method, selecting an MOF structure with suitable pore size, environment, and hydrophilicity/hydrophobicity affects the separation performance as well. PV and membrane distillation processes are attractive in the membrane separation for chemical purification, which require hydrophilic or hydrophobic property of membrane materials. The separation performance of the PV process is mainly determined by the composition, structure, and property of the membrane. Zeolitic imidazolate framework-8 (ZIF-8) [22] was firstly evaluated as molecular sieve membrane in the PV of the two liquid mixtures, n-hexane/benzene and n-hexane/mesitylene [23]. Zeolitic imidazolate frameworks (ZIFs), a subfamily of MOFs with extended three-dimensional structures and zeolite-like topology, are constructed by using tetrahedral transition metal ions (e.g. Zn, Co) and imidazolate-based bridge ligands. Because of their wide topological variety and exceptional thermal and chemical stability, ZIFs are very attractive for gas adsorption and separation [24]. Though it is known from permeation studies of light gases that ZIF-8 membranes show no sharp separation cutoff at the estimated crystallographic pore size of 3.4 Å. The consideration of framework flexibility is crucial to predicting correct separations on MOF membranes. Permeation studies on ZIF-22 [25] and ZIF-90 [26] membranes also showed that no sharp cutoff existed for hydrocarbons with critical diameters larger than the crystallographic pore size. A more appropriate expression for the terminus “breathing” might be “gate opening” [27]. In a recent paper of Kapteijn’s group, it was shown that ethane is adsorbed by ZIF-7 at a lower pressure in comparison with the only slightly lighter and stiffer ethylene [28]. Highly branched or aromatic hydrocarbons greater than C5 could be expected therefore to become rejected by the ZIF-8 pores, thus remaining in the retentate. However, the ZIF-8 membrane showed remarkable fluxes for n-hexane and benzene. Under consideration of the leakage of the apparatus, they stated that n-hexane and benzene could pass the ZIF-8 membrane. Benzene had a lower flux than n-hexane, whereas for mesitylene they could only observe a very small leakage rate through the O-ring gasket. Correspondingly, medium mixture separation factors have been found for the PV separation of a liquid n-hexane/benzene mixture. Additional mixed gas hydrogen/methane separations, adsorption experiments, and leak rate measurements were carried out to evaluate the results. ZIF-71 membrane was then made into a polycrystalline membrane for the PV separation of chemical mixtures [29]. ZIF-71 has a RHO topology with small windows (0.48 nm) and big cages (1.68 nm) and exhibits intrinsic hydrophobic property [30]. A molecular simulation study indicated that alcohols (methanol and ethanol) could be selectively adsorbed from alcohol–water mixtures by ZIF-71 crystals especially under relatively low pressures [31]. In Lin’s work, an integrated ZIF-71 membrane was prepared and used for PV separation of alcohol (methanol and ethanol)–water and dimethyl carbonate (DMC)–methanol mixtures. A robust reactive seeding method developed by Jin was applied in their work to prepare ZIF-71 membrane. To assess the integrity of the ZIF-71

11.2 Selective Separation of Chemicals Via Pervaporation

membranes, the permeances of single gas molecules (He, N2 , SF6 ) through the membrane were measured at different feed pressures. The permeances of all gases were independent of the feed pressure, indicating the absence of macroscopic defects in the membrane. The ZIF-71 membrane exhibited permselectivity for alcohols over water and DMC over methanol for PV separation of alcohol–water and DMC–methanol mixtures. Separation of DMC from water by the ZIF-71 membrane confirmed the “gate opening” effect for ZIF crystals observed by other researchers [27]. This work demonstrates that ZIF-71 is a promising membrane material for the PV separation of not only organics–water but also organics–organics systems. In the same year, Qiu reported a continuous MOF (Zn2 (BDC)2 DABCO) membrane, which was successfully synthesized by the secondary growth approach on the modified porous SiO2 substrate [32]. The seed crystals were prepared using acetic acid as tapped agent. The thickness of the membrane was varied by adjusting the repetitions of growing process. The separation performance tests of xylene isomer mixtures were taken out. The separation of mixed xylene isomers is one of the most challenging issues since xylene isomers are important chemical intermediates. For example, p-xylene is exclusively used as a raw material in the production of terephthalic acid (TPA) and dimethyl terephthalate (DMT), which then react with ethylene glycol to form polyethylene terephthalate (PET). Distillation can be used to remove o-xylene, which fails for the other xylene isomeric compounds because of similar boiling points (p-xylene, 138.37 ∘ C, and m-xylene, 139.12 ∘ C). Nowadays, adsorption is widely used to separate xylene isomers, for example, in Parex or Ebex units in which simulated moving-bed processes are used for the recovery of p-xylene and ethylbenzene [33]. Zeolites X and Y exchanged with cations (Na+ , K+ , and Ba2+ ) discriminate very selectively between different xylene isomers and have been widely used as industrial adsorbents [34]. However, the process control of this method is complex. Zeolite membranes are used to separate xylene isomers as well, which may require high permeance temperature and pressure [35]. On the contrary to the results from zeolite membranes, o-xylene and m-xylene molecules with larger kinetic diameter have greater permeability on this MOF membrane, which was consistent with the selective adsorption results of this MOF. This phenomenon could be explained by the adsorption–diffusion model. The molecules that are easier to be adsorbed by framework materials took precedence inside the channels and hindered other molecules to get through the membrane. In another Qiu’s study, an MOF membrane was applied in the separation of cyclohexanone and cyclohexanol, which are important organic intermediates in the petrochemical industry and difficult to separate [36]. The MOF structure selected in this work for membrane synthesis was ZIF-78, owing to its appropriate pore size, considerable nitro groups in the framework, and relatively high stability. They developed an applicable novel approach to synthesize ZIF-78 membrane by replacing the crystal seeds with nano-sized amorphous precursors, which were obtained rapidly and provided better distributed nucleation sites. During the secondary growth process, the crystal size and intergrowth degree of the membrane were adjusted by adding triethylamine into the reaction solution, and an optimal synthesis condition was finally obtained.

367

368

11 MOF Membranes for Vapor or Liquid Separation

The high-quality membrane synthesized under this method was tested for separation of cyclohexanone/cyclohexanol mixture at room temperature with permselectivity of 1 : 2 and total flux around 8.7 × 10−2 kg m−2 h−1 . Membranes made of large-pore MOFs such as MOF-5, offer potential for separation of xylene isomers in liquid phase. However, MOF-5 exhibits severe degradation upon contact with atmospheric levels of moisture. In the work reported by Lin, he determined the stability of MOF-5 membranes during PV of organic solvents [37]. High-quality MOF-5 membranes with about 10 μm in thickness were prepared by the secondary growth method using ball-milled MOF-5 seed crystals. On-stream p-xylene PV test of the as-synthesized MOF-5 membrane showed that the p-xylene flux declined and leveled off at a steady-state value about 70% of the original value for fresh membrane after 16 hours on the stream. The p-xylene flux was not able to be restored to the original value, suggesting permanent fouling of the MOF-5 membrane upon exposure to p-xylene stream. Subsequent characterization indicated that there was no structural or microstructural degradation that occurred, and evidence of retained xylene isomers in the MOF-5 structure was found using Fourier transform infrared spectroscopy (FITR). Nevertheless, the fouled MOF-5 membranes exhibited stable PV fluxes for organic molecules with sizes smaller than the aperture of the MOF-5 cage and excluded organic molecules with sizes larger than the MOF-5 aperture. The results showed that the MOF-5 membranes, after a permanent reduction in permeation flux by initial exposure to xylene stream, were stable in organic solvent and could be used for separation of liquid organic molecules based on the molecular sieving effects. In a follow-up work, multiple high-quality MOF-5 membranes were reproducibly prepared by the same group for various experimental runs [38]. The PV of pure toluene, o-xylene, and 1,3,5-triisopropylbenzene (TIPB) and the separation of their binary mixtures were studied. The permeation flux and separation factors decreased with PV on-stream time, and steady-state permeation flux could not be reached even after 10 hours of PV. The fouling effects did not change the crystalline structure of the MOF-5 membrane. The PV flux with the mixture feed was lower than the pure component flux, and the reduction in the flux decreased with decreasing affinity of the permeating species with MOF-5. The mixture maximum separation factors for toluene/TIPB and o-xylene/TIPB were about 26.7 and 14.6, respectively, significantly higher than the pure component ideal separation factor. The fluxes and separation factors could not be restored to their original values upon membrane activation at 100 ∘ C in vacuum. Fouling is a critical issue that needs to be addressed for application of MOF membranes for PV separation of organic liquids. Large-pore ZIF-68 membranes offer adsorption-based selectivity for separation of gas mixtures or molecular sieving characteristics for the separation of large liquid molecules [39]. ZIF-68 membranes can be grown on ZnO-modified α-alumina supports by a modified reactive seeding method [40]. The resultant membranes were around 40 μm in thickness and were determined to have limited nonselective defects given their adherence to Knudsen diffusion during single gas permeation measurements. Further PV experiments showed that the ZIF-68 membranes synthesized via the modified reactive seeding method had a p-xylene

11.2 Selective Separation of Chemicals Via Pervaporation

PV flux approximately 5.4 times as large as that reported for similar pore-sized MOF-5 membranes; however, PV flux of a larger molecule, di-tert-butylbenzene, through the same two MOF membranes showed that the flux of the ZIF-68 membrane was 3.4 times smaller than that reported for MOF-5. This reversal in PV flux indicated that the ZIF-68 structure is more readily accessible to molecules smaller than its pore size, but larger molecules were subjected to a staunch cutoff in flux. Water scarcity is increasingly becoming a global challenge to human activities and industrial and agricultural development due to the shortage of freshwater [41]. Therefore, the development of renewable freshwater resources such as seawater desalination plays an important role in resolving the future water supply. Indeed, seawater desalination is expected to offer an unlimited and steady supply of high-quality water without impairing natural freshwater ecosystems [42]. In the past half century, many technologies such as multistage flash distillation (MSFD) and RO) have been developed for seawater desalination [43]. Compared with the thermally based MSFD, the membrane-based RO is considered to be the most promising alternative because of its low energy consumption [44]. However, the present polymeric RO membranes often suffer from biofouling, oxidation, metal oxide fouling, abrasion, and mineral scaling, thus usually resulting in a low rejection and low stability [45]. Inorganic porous membranes are expected to break through these materials-based limitations for desalination due to their uniform pore structure and high thermal and chemical stability [46]. However, it should be noted that the thermal burning of organic templates used in the synthesis of zeolite membranes may result in the formation of cracks, leading to poor reproducibility of membrane preparation. It is highly desired, therefore, to develop novel molecular sieving membranes that can be facilely prepared and activated and thus effectively used for seawater desalination. In Huang’s work, he successfully developed stable ZIF membranes for seawater desalination [47]. Attributing to the small aperture size as well as the high stability in seawater, the developed ZIF membranes displayed good separation performances for seawater desalination as demonstrated by both PV experiments and simulation studies. At 25, 50, 75, and 100 ∘ C, the water fluxes through the ZIF-8 membrane were 5.8, 8.1, 10.8, and 13.5 kg m−2 h−1 , respectively, with ion rejections of over 99.8%, which were superior to conventional zeolite membranes. This high separation performances combined with high stability suggest that the developed ZIF membranes are promising candidates for seawater desalination by ionic sieving. Very recently, zirconium(IV) carboxylate MOFs (Zr-MOFs)) have attracted much attention due to their high chemical and thermal stability. UiO-66 (UiO stands for the University of Oslo), a typical Zr-based MOF, was first reported by Cavka et al. [48] The framework is built up from Zr6 O4 (OH)4 (CO2 )12 clusters linked with 1,4-benzenedicarboxylate (H2 BDC) and has a main pore diameter of 6 Å. This MOF shows excellent thermal stability (up to 500 ∘ C) owing to the presence of the Zr6 O4 (OH)4 inorganic building blocks and remains unaltered toward a number of organic solvents. Accordingly, UiO-66 membranes are expected to demonstrate its prosperous use in liquid separation. UiO-66 membranes are more difficult to be prepared than others, such as Zn-based ZIF membranes; thus there have been few reports on continuous UiO-66

369

370

11 MOF Membranes for Vapor or Liquid Separation

polycrystalline membranes. The continuous UiO-66 membranes were reported by Li [49]. As shown in Figure 11.2, the alumina hollow fiber (HF) used in this study contains two spongelike layers sandwiching a layer of fingerlike voids. The UiO-66 membrane was synthesized on the outer surface of such an HF substrate by an in situ solvothermal synthesis method. It should be noted that anhydrous

400 μm (a)

200 μm (b)

3 μm (c)

500 nm (d)

1 μm (e)

1 μm (f)

Figure 11.2 SEM images (a–c and e, cross section; d, top view) and (f ) EDXS mapping (corresponding to e) of the alumina hollow fiber (HF) supported UiO-66 membranes. Zr signal, red; Al signal, light blue. The membranes were fabricated on the outer surface of the HF [49]. Source: Copyright 2015, ACS.

11.2 Selective Separation of Chemicals Via Pervaporation

chemicals and solvent should be kept fresh and handled with care to avoid deliquescence or moisture sorption. This is because the amount of water in the mother solution for membrane synthesis is critical to the nucleation and intergrowth of UiO-66 crystals. Single gas permeation and ion rejection tests were carried out to confirm membrane integrity and functionality. The membrane exhibited excellent multivalent ion rejection (e.g. 86.3% for Ca2+ , 98.0% for Mg2+ , and 99.3% for Al3+ ) on the basis of size exclusion with moderate permeance (0.14 l m−2 h−1 bar−1 ) and good permeability (0.28 l m−2 h−1 bar−1 μm). Benefiting from the exceptional chemical stability of the UiO-66 material, no degradation of membrane performance was observed for various tests up to 170 hours toward a wide range of saline solutions. The high separation performance combined with its outstanding water stability suggested the developed UiO-66 membrane as a promising candidate for water desalination. In the study reported by the same group, the as-synthesized UiO-66 membranes supported on prestructured YSZ HF provided excellent performance for purifying typical biofuels, biochemicals, and organics under harsh environments [50]. On the basis of the adsorption–diffusion mechanism, the membranes provided a very high flux of up to cal. 6.0 kg m−2 h−1 and excellent separation factor (>45 000) for separating water from i-butanol (next-generation biofuel), furfural (promising biochemical), and tetrahydrofuran (THF) (typical organics). This performance, in terms of separation factor, was one to two orders of magnitude higher than that of commercially available polymeric and silica membranes with equivalent flux. It was comparable to the performance of commercial zeolite NaA membranes. As shown in Figure 11.3, benefiting from the excellent chemical stability of UiO-66 materials and good attachment of membrane layer to substrate, the membrane remains robust during a PV stability test (≈300 hours), including exposure to harsh environments (e.g. boiling benzene, boiling water, and sulfuric acid) where some commercial membranes (e.g. zeolite NaA membranes) cannot survive.

Water / furfural separation

Immersed in boiling benzene and boiling water

1.6

90

H2SO4 added

1.2 0.8 0.4

Water / n-butanol

Water / n-butanol separation

Total flux (kgm–2h–1)

Solvothermal and hydrothermal test

100 2.0

80 70

Water / n-butanol

60 Water / furfural

0.0

Water con. in permeate (wt %)

2.4

On-stream activation

50 0.0

50

100

150

200

250

300

Time (h)

Figure 11.3 Flow chart and PV performance of the UiO-66 membrane for separating water from n-butanol and furfural during on-stream activation and stability test processes at 30 ∘ C. The concentration of water in each feed was kept at 5.0 wt% [50]. Source: John Wiley & Sons.

371

372

11 MOF Membranes for Vapor or Liquid Separation

In Zhang’s wok, stable UiO-66 membranes were successfully prepared on macroporous alumina tubes by secondary growth and were applied in the separation of methanol/methyl tert-butyl ether (MTBE) mixtures by PV [51]. Two key factors affecting the growth of the UiO-66 membranes were investigated. It has been found that acetic acid promoted the crystallization of UiO-66 crystals and water favored the intergrowth of UiO-66 crystals, leading to the formation of high-performance UiO-66 membranes. Under the optimized synthesis conditions, a continuous and dense UiO-66 membrane was obtained, showing a separation factor of nearly 600 and a flux of 1.21 kg m−2 h−1 for 5 wt% methanol/MTBE mixtures at 40 ∘ C. The effects of operation parameters on PV performance of the UiO-66 membranes were also studied in detail. The results revealed that increasing feed temperature or methanol in the feed would enhance permeation flux but reduced separation factor, the permeability of components, and membrane selectivity. It is worth mentioning that the UiO-66 membranes showed excellent operating stability. This work demonstrates that the UiO-66 membranes have great potential for separation of methanol/MTBE mixtures via PV. Zr-MOFs are notoriously known for their intrinsic defects caused by ligand/cluster missing, which may greatly affect the molecular sieving property of Zr-MOF membranes [52]. Zhao presented the mitigation of ligand-missing defects in polycrystalline UiO-66(Zr)-(OH)2 membranes by post-synthetic defect healing (PSDH) [53]. The healed membranes exhibited NaCl rejection of 45% with water permeance of 0.285 kg m−2 h−1 bar-1 and methyl blue (MB) rejection of 99.8% with water permeance of 0.23 kg m−2 h−1 bar-1 . The improved separation performance of membranes after PSDH was confirmed by water treatment tests using aqueous solutions containing Na+ or MB (Figure 11.4). The rejection rate of Na+ was 45% for the membranes after PSDH (74.9% increase compared with that in pristine membranes). Intriguingly, the membranes also exhibited excellent hydrothermal stability in aqueous solutions (>600 hours). As reported by Miyamoto, a UiO-66 membrane was synthesized on an Al2 O3 support by the in situ solvothermal method with a coordination modulation technique [54]. He first confirmed the organic selectivity of the UiO-66 membrane for methanol, ethanol, and acetone in mixtures with water. The membrane displayed highly stable PV performance during the test and exhibited a separation factor of approximately 4.3 and a flux of 1.28 kg m−2 h−1 in ethanol/water. From the water and ethanol vapor adsorption measurement, the water uptake was lower in a binary system than that in a single system, although the ethanol uptake was similar for the binary and single systems. Therefore, the ethanol selectivity of the UiO-66 membrane was attributed to the preferential adsorption of the organic compound rather than water, as confirmed by a binary vapor adsorption measurement. The developed UiO-66-NH2 membranes were evaluated for seawater desalination by PV [55]. It was found that the UiO-66-NH2 membranes showed high desalination performances attributing to the narrow pore size, which is exactly in between the size of water molecules and hydrated ions. With increasing feed temperature from 318 to 363 K, the water fluxes increased from 1.5 to 12.1 kg m−2 h−1 , with ion rejections of above 99.7%. Further, the UiO-66-NH2

11.2 Selective Separation of Chemicals Via Pervaporation

300

150

12.6Å n=6

200

n = 4.68 n=5

Normalized weight (%)

Defective cluster: Zr6O4(OH)4(DOBDC)n

250

100 Pristine crystal PSDH crystal

Defective structure

0.8

Pristine membrane PSDH membrane

80

0.6

60

0.4

40

0.2

20

0.0

(c)

Na+ Methyl blue

Na+ Methyl blue

Rejection (%)

Flux (kg m–2 h–1)

50 (a) 100 200 300 400 500 600 700 800 900 Temperature (°C) 1.0 100

4Å

0

4Å

Healed structure (b)

Figure 11.4 (a) TGA curves of UiO-66(Zr)-(OH)2 crystals before and after PSDH; the weight was normalized with respect to the ZrO2 residue left after heating up to 650 ∘ C in air. (b) Scheme of PSDH by relinking two adjacent Zr6 O4 (OH)4 clusters (blue polyhedron) by one DOBDC ligand. The yellow area indicates the effective aperture size for molecular sieving separation; the cyan area indicates the location of the re-coordinated DOBDC within the framework. (c) Separation performance of the membrane before and after PSDH [53]. Source: Copyright 2017, ACS.

membranes displayed high stability for a long time in seawater desalination, which is very promising for seawater desalination. 11.2.2

MOF-Based MMMs for PV Process

Polymers can be used to seal the gaps between the MOF crystals to form the main part of the MMMs. MOF-based MMMs have the potential to synergistically combine the easy processability of polymers and the high porosity of the MOFs. Unlike polycrystalline MOF membranes that often require supporting substrates, the preparation of MMMs typically results in freestanding membranes, regardless of the original MOF synthetic conditions. Due to the organic linkers in the structure, MOFs interact well with the polymer, avoiding the formation of micro-gaps to some extent. In particular, improvements in the liquid permeation performance of a membrane would be expected if nanostructured and highly porous MOFs were used. Despite their potential, MOF-based MMMs still pose challenges and limitations. MOF-based MMMs have reduced permeability compared with their pure MOF counterparts because of the nonporous nature of many polymer binders. This limitation can be overcome by

373

11 MOF Membranes for Vapor or Liquid Separation

increasing the MOF loading, which shifts the MMMs from polymer-dominant to MOF-dominant composites. Another significant challenge for MOF-based MMMs is the aggregation of the MOF particles and poor dispersion within the MMM. This is a leading source of membrane void defects. In the work reported by Yang [56], as a consequence of the flexible pore apertures and the superhydrophobic pore surface, ZIF-8 nanoparticles (NPs) exhibited exceptional adsorption selectivity and capacity toward isobutanol molecules and showed a reversible gate opening effect upon variation of the isobutanol pressure or temperature. As demonstrated by CB-GCMC simulations, each sodalite cage of ZIF-8 accommodated six isobutanol molecules at 3.5 kPa (Figure 11.5). Encouraged by the experimental and theoretical results, ZIF-8 NPs were introduced into silicone rubber (PMPS) to fabricate organophilic pervaporation (OPV) membranes. The ZIF-8/PMPS membrane showed a very promising performance for recovering bioalcohols from dilute aqueous solution and offered significant potential for the construction of a membrane reactor for

400 Uptake (mg g–1)

374

c

300 200

Isobutanol density

100 0

b 0

1

(a)

2 P (kPa)

3

4 a

(b)

c

b 0.3 nm a (c)

Site I

Site II

Figure 11.5 (a) Experimental and simulated adsorption and desorption isotherms for isobutanol and water with ZIF-8 at 40 ∘ C (filled (empty) circles and red (blue) lines represent experimental and simulated isobutanol adsorption (desorption), filled (empty) triangles represent experimental water adsorption (desorption), and the wine line represents simulated water adsorption). (b) Density contours for isobutanol in one sodalite cage of ZIF-8 at 3.5 kPa and 40 ∘ C. (c) Representations of isobutanol adsorption sites in ZIF-8 (abstracted from (b)) [56]. Source: John Wiley & Sons.

11.2 Selective Separation of Chemicals Via Pervaporation

in situ product recovery (ISPR) applications. Both the membrane selectivity and permeability can be improved by increasing the ZIF-8 loading in the composite membrane. It is, therefore, expected that polycrystalline ZIF-8 membranes will show much better performance. As reported in the later work [57], a homogeneous ZIF-8/silicone rubber nanocomposite membrane with high particle loading was successfully fabricated on a hierarchically ordered stainless steel mesh (HOSSM) employing a novel “plugging–filling” method. The membrane exhibited high PV separation index (separation factor of 53.3 and total flux of 0.90 kg m−2 h−1 ) and excellent stability in a test of more than 120 hours at 80 ∘ C for recovery of furfural (1.0 wt%) from water. This very high performance should be attributed to the exceptional adsorption selectivity and capacity of ZIF-8 toward furfural molecules and the effects of space restriction and physical cross-linking of the HOSSM. This is a good demonstration of the potential of MOF membranes for separating biomass-derived compounds in biorefinery processes. In a study of Jin, MMMs incorporating ZIF-71 particles into polyether-blockamide (PEBA) were prepared for biobutanol recovery from acetone–butanol– ethanol (ABE) fermentation broth by PV [58]. The PV performance of the prepared MMMs with various ZIF-71 loadings for separating n-butanol from its aqueous solution was investigated. As a result, both separation factor and thickness normalized flux of the PEBA membranes were improved by incorporating appropriate amount of ZIF-71 (≤20 wt%). The membrane with 20 wt% ZIF-71 was evaluated in ABE model solution and ABE fermentation broth and exhibited high n-butanol separation performance. ZIF-71 particles were confirmed as promising fillers to enhance the separation performance for butanol recovery because of their excellent compatibility with polymer and organophilicity. This work demonstrated that the ZIF-71/PEBA MMMs could be potential candidates for practical biobutanol production. The ability to obtain a maximum loading of inorganic NPs while maintaining uniform dispersion in the polymer is the key to the fabrication of MMMs with high PV performance in bioalcohol recovery from aqueous solution. Zhang reported the simultaneous spray self-assembly of a ZIF–polymer suspension and a cross-linker/catalyst solution as a method for the fabrication of a well-dispersed ZIF-8/PDMS nanohybrid membrane with an extremely high loading [59]. The process of membrane formation is shown in Figure 11.6; the ZIF-8/PDMS suspension and a solution of the cross-linking agent TEOS and the catalyst dibutyltin dilaurate (DBTDL) were poured separately into two self-stirring pressure barrels and simultaneously sprayed onto a PS substrate. The ZIF-8/PDMS membrane showed excellent biobutanol permselective PV performance. When the ZIF-8 loading was increased to 40 wt%, the total flux and separation factor could reach 4846.2 g m−2 h−1 and 81.6, respectively, in the recovery of n-butanol from 1.0 wt% aqueous solution (80 ∘ C). This new method is expected to have implications for the preparation of defect-free MMMs for many applications. The same group prepared a homogeneous nanodispersed ZIF-8/PDMS membrane by repeated immersion of a polysulfone (PSF) supporting membrane in a dilute ZIF-8/PDMS suspension and subsequent removal of defects using a concentrated PDMS solution [60]. To improve the nanoscale dispersion of ZIF-8,

375

376

11 MOF Membranes for Vapor or Liquid Separation Nozzle1

PDMS

Nozzle2

TEOS / DBTDL ZIF-8

(a)

(b)

(d)

(c)

Figure 11.6 Formation of the ZIF-8/PDMS nanohybrid composite membrane by the simultaneous spraying self-assembly technique [59]. (a) Porous substrate, (b), simultaneous spray self-assembly, (c) crosslinking, and (d) ZIF-8/PDMS nanohybrid membrane. Source: John Wiley & Sons.

the nascent ZIF-8 suspension was directly dispersed in a PDMS solution without drying. This procedure avoids aggregation and redispersion of ZIF-8 NPs after forming a powder. Analyses confirmed that the ZIF-8/PDMS dispersion effectively diminished aggregation between NPs and led to the formation of a well-dispersed ZIF-8/PDMS membrane. A homogeneous and thin ZIF-8/PDMS permselective layer was obtained by adjusting the preparation conditions. The prepared ZIF-8/PDMS membrane exhibited a high separation factor (52.81) and a high flux (2800.5 g m−2 h−1 ) in the separation of 5.0 wt% n-butanol/water solution at 80 ∘ C. By comparing the powder-dispersed ZIF-8/PDMS hybrid membrane with the suspension-dispersed ZIF-8/PDMS membrane, they found that the latter showed much higher performance in butanol separation. [Cu2 (bdc)2 (bpy)]n is found to preferentially adsorb methanol (MeOH) over MTBE. For this reason, [Cu2 (bdc)2 (bpy)]n is incorporated into sulfonated polyarylethersulfone with cardo (SPES-C) to form [Cu2 (bdc)2 (bpy)]n /SPES-C MMMs for the separation of MeOH/MTBE mixtures [61]. Both the sorption selectivity and the diffusion selectivity were found to increase with the addition of [Cu2 (bdc)2 (bpy)]n , leading to an increase in the separation factor. With increasing [Cu2 (bdc)2 (bpy)]n loading from 5 to 20 wt%, the flux increased up to 0.288 kg m−2 h−1 at 20 wt% loading. A threshold of [Cu2 (bdc)2 (bpy)]n loading is approximately 20 wt% where the flux remains almost constant regardless of further increasing the [Cu2 (bdc)2 (bpy)]n content. In Vankelecom’s study, MMMs based on PDMS and ZIF-71 were prepared for separation of alcohols (methanol, ethanol, isopropanol [IPA], or sec-butanol) from aqueous solutions [62]. Experimental results revealed that the PV performance of the PDMS membrane was improved in both flux and separation

11.2 Selective Separation of Chemicals Via Pervaporation

factors upon embedment of ZIF-71. The PV separation factors for alcohols of ZIF-71 filled the PDMS membrane (PDMS : ZIF-71 = 10 : 2) nearly doubled compared with those of unfilled PDMS membranes. Encouraged by this work, the same groups synthesized submicrometer-sized ZIF-71 crystals with different particle sizes (140, 290, or 430 nm) via a facile room temperature synthesis method [63]. These small ZIF-71 particles are ideal fillers for the fabrication of thinner and homogeneous PDMS-based MMMs with excellent filler dispersion and filler–polymer adhesion at high loading up to 40 wt%, as confirmed by scanning electron microscopy (SEM). PV tests using these submicrometer-sized ZIF-71-filled MMMs showed significant improvement for bioethanol recovery. The intrinsic permeability of the membranes was calculated by correcting the driving force as shown in Figure 11.7. The submicrometer-sized ZIF-71-filled PDMS membranes displayed a gradual increase in separation factor and flux with increasing ZIF-71 loading (20–40 wt%). However, the separation factors of the micrometer-sized ZIF-71-filled membranes decreased at higher ZIF-71 loadings. The results suggested that the submicrometer-sized ZIF-71 particles

(a)

Water Total

PDMS

20%

30%

40%

Permeability ×106/(gm h–1 kPa–1)

Ethanol

Water

200 100 0 100

PDMS

20%

30%

40%

200 300 400 Micron-sized

(b) 0.8

Submicron-sized

Submicron-sized

0.7 0.6

8

0.5

6 PDMS

20%

30%

6

40%

0.4 0.3 0.4

PDMS

20%

30%

40%

0.5

8

12

300

600

Micron-sized

0.6 0.7

10

(c)

400

Ethanol

500

12

4

Submicron-sized

500

10 Separation factor

600

Submicron-sized

Selectivity

Normalized flux/ (g m–2 h–1)

1400 1200 1000 800 600 400 200 0 200 400 600 800 1000 1200 1400

0.8

Micron-sized

Micron-sized

(d)

Figure 11.7 Comparison of the pervaporation performance: (a) flux normalized to 5 μm, (b) permeability, (c) separation factor, and (d) selectivity of the micrometer-sized and submicrometer-sized ZIF-71-filled PDMS membranes prepared from 20% PDMS with different ZIF-71 loadings (20, 30, and 40 wt%). The data presented in the reverse order indicates the PV results of the micrometer-sized ZIF-71-filled PDMS membranes. The pervaporation was run at 50 ∘ C using 5% ethanol/water mixtures. The total flux and separation factor values were obtained based on average equilibrium performances of three different membrane coupons. The calculated standard deviations were 2–3% [63]. Source: John Wiley & Sons.

377

378

11 MOF Membranes for Vapor or Liquid Separation

were better suited for the preparation of defect-free thin composite membranes. Interesting phenomena of (i) reversible ethanol–ethanol hydrogen interaction in the ethanol liquid phase and (ii) irreversible hydrogen interaction of ethanol and —Cl functional group in the α-cages and octagonal prismatic cages of ZIF-71 in ethanol vapor phase were discovered for the first time by an FTIR study. In full agreement with molecular simulation results, these explained fundamentally the ZIF-71-filled MMM PV performance. In another study reported by Vankelecom, ZIF-71 and ZIF-8 nanocrystals were grown on the surface of mesoporous silica spheres (MSS) via the seeding and regrowth approach in order to obtain monodispersed MSS-ZIF-71 and MSS-ZIF-8 spheres with a particle size of 2–3 μm [64]. These MSS-ZIF spheres were uniformly dispersed into PDMS matrix to prepare MMMs. These MMMs were evaluated for the separation of ethanol from water via PV. The PV results revealed that the MSS-ZIF-filled MMMs substantially improved the ethanol recovery in both flux and separation factor. These MMMs outperformed the unfilled PDMS membranes and the conventional carbon- and zeolite-filled MMMs. As expected, the mesoporous silica core allowed very fast flow of the permeating compound, while the hydrophobic ZIF coating enhanced the ethanol selectivity through its specific pore structure, hydrophobicity, and surface chemistry. It was seen that ZIF-8 mainly had a positive impact on the selectivity, while ZIF-71 enhanced the flux more significantly. Lind’s group reported the effects of synthesis time, temperature, and reactant ratio on ZIF-71 particle size and the effect of particle size on membrane performance [65]. Temperature had the greatest effect on particle size, as the synthesis temperature varied from −20 to 35 ∘ C. The synthesized ZIF-71 had particle diameters ranging from 150 nm to 1 μm. ZIF-71 particle size was critical in ZIF-71/PDMS composite membrane performance for ethanol and 1-butanol removal from water through PV. The membranes that were made with micrometer-sized ZIF-71 particles exhibited higher alcohol/water selectivity than those with smaller particles. Both alcohol and water permeability increased when larger-sized ZIF-71 particles were incorporated, because the pathways through the membranes were less tortuous. Two kinds of ZIFs (two-dimensional ZIF-L nanosheets and zero-dimensional ZIF-8 NPs) with the same building blocks were synthesized. Both the ZIF-L and ZIF-8 materials were incorporated into sodium alginate (SA) matrix to fabricate hybrid membranes for PV dehydration of ethanol [66]. At the filler content of 4 wt%, the ZIF-L-filled membrane displayed permeation flux of 1218 g m−2 h−1 and separation factor of 1840, while the ZIF-8-filled membrane displayed permeation flux of 879 g m−2 h−1 and separation factor of 678. The superior separation performance of the ZIF-L-filled membrane was due to the following two reasons: the ordered alignment and the regular apertures of ZIF-L rendered ordered water channels for rapid transport of water molecules, and the suitable apertures of ZIF-L rendered the desirable molecular sieving effect. Furthermore, the hybrid membranes exhibited both good swelling resistances and thermal and mechanical stability. This is a step forward in realizing superior performance of hybrid membranes incorporating two-dimensional porous fillers for separation processes.

11.2 Selective Separation of Chemicals Via Pervaporation

In a study reported by Pan [67], bimetallic MOF nanocages, FeIII -HMOF-5, were prepared and incorporated into SA matrix to fabricate water-selective MMMs. Introduction of Fe3+ ions into MOF-5 created more coordinatively unsaturated sites, which led to preferential dissolution of water molecules over ethanol molecules, thus elevating the solubility selectivity. The hollow structure of FeIII -HMOF-5 ensured the free diffusion of water molecules. The synergistic regulation of chemical composition and physical structure of MOFs endowed the hybrid membranes with remarkably elevated separation factor and permeation flux. Taking dehydration of 90 wt% ethanol aqueous solution as a model system, the hybrid membrane containing FeIII -HMOF-5 exhibited the highest separation performance with a separation factor of 3423 and a permeation flux of 1540 g m−2 h−1 , which were much higher than the values obtained for pure SA membrane and the hybrid membranes incorporating hollow MOF-5 (HMOF-5) and MOF-5. Moreover, the hybrid membranes containing Fe-HMOF-5 displayed remarkably superior physicochemical stabilities and long-term operation stability. This study demonstrated a promising prospect of heterometallic hollow MOFs as multifunctional fillers in high performance. Because of the high stability of UIO-66 series structures, they were blended with polymers to form MMMs for PV. UiO-66 NPs were successfully synthesized and incorporated in the polyamide (PA) selective layer to fabricate novel thin-film nanocomposite (TFN) membranes [68]. Compared with unmodified pure PA thin-film composite (TFC) membranes, the incorporation of UiO-66 NPs significantly changed the membrane morphology and chemistry, leading to an improvement of intrinsic separation properties due to the molecular sieving and superhydrophilic nature of UiO-66 particles (Figure 11.8). The top surface TFC

TFN-U1

TFN-U2

TFN-U3

TFN-U4

10 μm

PA

PA

PA

PA

100 nm

UiO-66 UiO-66

UiO-66

UiO-66

1 μm

Figure 11.8 FESEM micrographs of cross section and top surfaces for TFC and TFN membranes with different UiO-66 loadings: TFC (0 wt%), TFN-U1 (0.05 wt%), TFN-U2 (0.1 wt%), TFN-U3 (0.15 wt%), and TFN-U4 (0.2 wt%) [68]. Source: Copyright 2017, ACS.

379

380

11 MOF Membranes for Vapor or Liquid Separation

morphologies showed that although TFN and TFC membranes displayed similar ridge-and-valley structures, they were slightly different for the size of the ridge structure. The difference was larger as the UIO-66 ratio increased, indicating that UiO-66 NPs affected the interfacial polymerization (IP) process. In addition, when the concentration of UiO-66 was high (0.2 wt%), more NPs were exposed at the top surfaces of membranes surrounded by bare PA layers without apparent particle aggregations as labeled in red. Since the average particle size of UiO-66 was larger than the thickness of a typical PA layer (∼200 nm), partial embedment of UiO-66 in the PA selective layer took place with the particle bottom part tucked into the PA matrix and the remainder exposed. The best performing TFN-U2 (0.1 wt% particle loading) membrane not only showed a 52% increase of water permeability but also maintained salt rejection levels (∼95%) similar to the benchmark. The effects of UiO-66 loading on the forward osmosis (FO) performance were also investigated. Incorporation of 0.1 wt% UiO-66 produced a maximum water flux increase of 40 and 25% over the TFC control under PRO and FO modes, when 1 M NaCl was used as the draw solution against deionized water feed. Meanwhile, solute reverse flux was maintained at a relatively low level. In addition, TFN-U2 membrane displayed a relatively linear increase in FO water flux with increasing NaCl concentration up to 2.0 M, suggesting a slightly reduced internal concentration polarization effect. Chung has designed novel MMMs consisting of UiO-66 NPs and 6FDA-HAB/ DABA polyimide for the dehydration of ethanol, IPA, and n-butanol via PV [69]. The UiO-66 NPs have a particle size of around 100 nm. They can be evenly dispersed in the 6FDA-HAB/DABA polyimide matrix without visible agglomeration even at the highest 30 wt% loading. The incorporation of UiO-66 into the 6FDA-HAB/DABA polyimide not only significantly enhances both free volume radius and fractional free volume as confirmed by positron annihilation lifetime spectroscopy but also remarkably improves the normalized flux of MMMs for the dehydration of ethanol/water, IPA/water, and n-butanol/water systems. The MMMs show excellent separation efficiency for the dehydration of IPA and n-butanol. At the highest UiO-66 loading of 30 wt%, the MMMs have the water permeability of 0.329 and 0.292 mg m−1 h−1 kPa−1 and mole-based selectivity of 2209 and 14214, respectively, for IPA/water and n-butanol/water systems, outperforming most literature data. These experimental results strongly suggest that the newly developed UiO-66/polyimide MMMs have great potential for IPA and n-butanol dehydration via PV. An acid-stable Zr-MOF NH2 -UiO-66 was synthesized and incorporated into poly(ethyleneimine) (PEI) to form MMMs for separating acetic acid/water mixtures [70]. The NH2 -UiO-66/PEI MMMs were deposited on the surface of NaA zeolite tubular substrate to form composite membranes using the dip-coating method. The effects of membrane preparation conditions on the separation performance were investigated. The results indicated that the NH2 -UiO-66/PEI composite membranes showed good acetic acid dehydration behavior because of the high porosity and hydrophilicity of the particles. Moreover, the particles had good compatibility with polymer and strong combination with substrate. Therefore, this study may provide a new material and facile strategy for preparing composite membrane in the separation of acetic acid/water mixtures.

11.3 Organic Solvent Nanofiltration

= GO sheets

Pressurization

= MOF

PASA technique

Figure 11.9 Schematic diagram of MOF@GO membranes prepared by PASA filtration technique [72]. Source: Copyright 2016, ACS.

Besides polymer, MOFs can be adopted into other spices such graphene oxide (GO) for PV application. GO membranes assembled by single-atom-thick GO nanosheets have displayed huge potential application in both gas and liquid separation processes due to its facile and large-scale preparation resulting from various functional groups, such as hydroxyl, carboxyl, and epoxide groups [71]. Taking advantage of these characters, GO membranes intercalated by superhydrophilic MOFs as strengthening separation fillers were prepared on modified polyacrylonitrile (PAN) support by a novel pressure-assisted self-assembly (PASA) filtration technique instead of traditional vacuum filtration method for the first time (Figure 11.9) [72]. Compared with GO membrane, these MOF@GO membranes combined the unique properties of MOF and GO and thus had significant enhancements of PV permeation flux and separation factor simultaneously for ethyl acetate/water mixtures (98/2, w/w) through the PV process, which were also superior to the reported other kinds of membranes. Especially, for [email protected] membrane (corresponding MOF loading of 23.08 wt%), the increments were 159% and 244%, respectively, at 303 K, and the permeate water content reached as high as 99.5 wt% (corresponding separation factor of 9751) with a high permeation flux of 2423 g m−2 h−1 . Moreover, the procedures of both the synthesis of MOF and membrane preparation are environmentally friendly that only water is used as the solvent. Such a nano-sized MOF-intercalating approach may be also extended to other laminated membranes, providing valuable insights in designing and developing advanced membranes for effective separation of aqueous organic solution through nanostructure manipulation of the nanomaterials (Tables 11.1 and 11.2).

11.3 Organic Solvent Nanofiltration Organic solvent nanofiltration, also known as solvent-resistant nanofiltration, is an emerging technology for performing membrane separation processes [85], which is significant in expanding the membrane applications from aqueous systems, primarily for water/water-related purification and treatment to filtration and concentration of organic solutions [85a, 86]. Compared with competing technologies, for example, preparative chromatography, distillation, extraction, and crystallization, it is generally more energy efficient. Nevertheless, it does not create extra waste streams and allows for mild operating conditions [87].

381

Table 11.1 Polycrystalline MOF membranes for PV separation process.

Membrane

Mixture

Method

Temperature (∘ C)

Flux or permeance −2

ZIF-8

n-Hexane/benzene/ mesitylene

Secondary growth

RT

3.9 ml m

ZIF-78

Cyclohexanone/ cyclohexanol

Secondary growth

RT

ZIF-8

Aqueous salt solutions

In situ

UiO-66-NH2

Aqueous salt solutions

Zn2 (BDC)2 DABCO

Separation factor or rejection ratio −1

n-Hexane/ benzene(8.7)

[23]

8.7 × 10−2 kg m−2 h−1

Cyclohexanol/ cyclohexanone (2)

[36]

100

13.5 kg m−2 h−1

Ion rejections > 99.8%

[47]

In situ

90

12.1 kg m−2 h−1

Ion rejections > 99.7%

[55]

Xylene

Secondary growth

25–200

2 kg m−2 h−1

m-x/p-x(1.934), o-x/p-x(1.617)

[32]

UiO-66

Aqueous salt solutions

In situ

20

0.14 l m−2 h−1 bar−1

Ca2+ (86.3%), Mg2+ (98.0%), Al3+ (99.3%)

[49]

UiO-66

Water/n-butanol

In situ

30

6.0 kg m−2 h−1

>45 000

[50]

ZIF-71

Alcohol/water

Secondary growth

25

102.89 g m−2 h−1 kPa−1

EtOH/water (6.07), MeOH/water(21.38), DMC/MeOH(5.34)

[29]

MOF-5

Toluene/TIPB

Secondary growth

25

11.7 × 10−4 mol m−2 s−1

27.7

[37]

ZIF-68

p-Xylene/DTBB

Secondary growth

25

1.87 kg m−2 h−1

102a)

[40]

ZIF-8 (surface ligand exchange)

Pure water

Secondary growth

25

0.50 kg m−2 h−1

—

[73]

UiO-66

MeOH/water

In situ

50

1.58 kg m−2 h−1

5

[54]

Ni2 (l-asp)2 bipy

Water/ethanol

Secondary growth

30

27.6 kg m−2 h−1

73.6

[74]

UiO-66

Methanol/MTBE

Secondary growth

40

1.21 kg m−2 h−1

600

[51]

UiO-66-(OH)2

Aqueous salt solutions

Secondary growth

RT

0.285 kg m−2 h−1 bar−1

45%

[53]

a)

Ideal separation factor.

h

Reference

Table 11.2 MMMs based on MOFs for PV separation process.

MOF

Polymer

Mixture

Temperature (∘ C)

Flux or permeance −2

−1

Reference

95%

[68]

UiO-66

PA

Aqueous salt solutions

RT

51.3 l m

UiO-66

6FDAHAB/DABA polyimide

Isopropanol/water

60

0.329 mg m−1 h−1 kPa−1

2209

[69]

Ordered mesoporous silica–(ZIF-8) core–shell spheres (MSS-ZIF8)

Matrimid

Water/ethanol

42

0.19 kg m−2 h−1

137

[75]

®

5218

h

Separation factor or rejection ratio

ZIF-8

PEBA-2533

Phenol/water

70

1310 g m−2 h−1

53

[76]

NH2 -UiO-66

PEI

Water/acetic acid

60

212 g m−2 h−1

356

[70]

ZIF-71

PEBA

n-Butanol/acetone/ water

37

520.2 g m−2 h−1

18.8

[58]

Zn(BDC)(TED)0.5

PEBA

n-Butanol/acetone/ water

37

630.2 g m−2 h−1

17.4

[77]

UiO-66

GO

Dye/water

RT

30.6 l m−2 h−1

95%

[78]

MOP-tBu

Boltorn W3000

Toluene/n-heptane

40

220.5 g m−2 h−1

19

[79]

Cu3 (BTC)2

PVA

Toluene/n-heptane

40

133 g m−2 h−1

17.9

[80]

ZIF-71

PDMS

1-Butanol/water

60

90966 Barrer

63

[65]

ZIF-71

PDMS

1-Butanol/water, ethanol/water

60

1-Butanol (123045 Barrer), ethanol (24809 Barrer)

1-Butanol/water (69.9) ethanol/ water (12.5)

[81]

(Continued)

Table 11.2 (Continued)

MOF

Polymer

Mixture

Temperature (∘ C)

Flux or permeance −2

Separation factor or rejection ratio −1

h

Reference

[Cu2 (bdc)2 (bpy)]n

SPES-C

Methanol/MTBE

40

0.288 kg m

2657

[61]

ZIF-71

PDMS

Acetone/water

60

1236.8 g m−2 h−1

39.1

[82]

ZIF-8

PEK-c

Water/methanol/ MTBE

50

1.48 kg mm−2 h−1

19700

[83]

ZIF-L

Sodium alginate

Ethanol/water

76

1218 g m−2 h−1

1840

[66]

ZIF-8

Sodium alginate

Ethanol/water

76

879 g m−2 h−1

678

[66]

ZIF-71

PDMS

Ethanol/water

50

550 g m−1 h−1 kPa−1

9.9

[62]

ZIF-71

PDMS

Ethanol/water

50

580 g m−1 h−1 kPa−1

8

[63]

MSS-ZIF-71

PDMS

Ethanol/water

50

1 kg m−2 h−1

13

[64]

MSS-ZIF-8

PDMS

Ethanol/water

50

0.75 kg m−2 h−1

15

[64]

UiO-66-PSBMA

PSF

Protein

RT

602 l m−2 h−1

>98%

[84]

ZIF-8

PMPS

Butanol/water

80

6454 GPU

40.1

[56] [57]

ZIF-8

PMPS

Butanol/water

80

0.90 kg m−2 h−1

53.3

UiO-66-(COOH)2

GO

Ethyl acetate/water

60

3632 g m−2 h−1

6076

[72]

ZIF-8

PDMS

n-Butanol/water

80

4846.2 g m−2 h−1

81.6

[59]

ZIF-8

PDMS

n-Butanol/water

80

2800.5 g m−2 h−1

52.81

[60]

11.3 Organic Solvent Nanofiltration

OSN can also well complement these conventional separation techniques into more efficient hybrid processes. In industry, OSN may be applied in many solvent-intensive processes, some with a large economic impact, such as edible oil refining and degumming, catalyst recovery, solvent recycling in the pharmaceutical industry, solvent dewaxing, polymer fractionation, and thermal solvent exchanges. Polymeric membranes are considered to be the most interesting materials for OSN applications. Advantages are the large variety of available polymers, their relatively low price, and the easy fabrication and scaling up. An important limitation of polymeric membranes, however, is their poor thermal and chemical stability. Interactions between organic solvents and the membrane can cause these membranes to swell extensively (or ultimately even dissolve), detrimental to the selectivity. MOFs having versatile structural properties can be filled inside the polymer membranes to prepare MOF-based membranes to further enhance the OSN performance. The organic bridge existence in the MOF structure facilitates a better affinity for polymeric matrices, and it is easier to control MOF/polymer interactions, improving the roughness and mechanical behavior for better solvent separation performances. Moreover, the tunable porosity and pore size of MOFs may create selective cavities and paths to increase the solvent flux and maintain a high rejection. Two different combining types between MOFs and membrane are proposed for OSN applications: physical blending in the membrane matrix to prepare MOF-based MMM or embedding MOFs in the PA layer to design an MOF-based TFN membrane [8]. For the pioneer study, Vankelecom et al. prepared MMMs using a series of MOFs (Cu-BTC, MIL-47, MIL-53(Al), and ZIF-8) as dispersed phases in PDMS membranes [86]. The membranes were applied in the separation of Rose Bengal (RB) from IPA. The membranes showed increased permeance but lower retention compared with unfilled membranes. Membranes were characterized with SEM where poor adhesion of the fillers to PDMS was observed. Chemical modification of the fillers with N-methyl-N-(trimethylsilyl)-trifluoroacetamide (MSTFA) allowed synthesis of defect-free membranes with higher retention of RB, due to good interaction with fillers and decreased membrane swelling. In the study reported by Livingstone, TFN membranes containing a range of 50–150 nm MOF NPs including ZIF-8, MIL-53(Al), NH2 -MIL-53(Al), and MIL-101(Cr) in a PA thin-film layer were synthesized via in situ IP on top of cross-linked polyimide porous supports [88]. MOF NPs were homogeneously dispersed in the organic phase containing trimesoyl chloride (TMC) prior to the interfacial reaction. Membrane performance in OSN was evaluated on the basis of methanol and THF permeances and rejection of styrene oligomers (PS). The effect of different posttreatments and MOF loadings on the membrane performance was also investigated. MeOH and THF permeance increased when MOFs were embedded into the PA layer, whereas the rejection remained higher than 90% (molecular weight [MW] cutoff of less than 232 and 295 g mol−1 for MeOH and THF, respectively) in all membranes. Moreover, the permeance increased with increasing pore size and porosity of the MOF used as filler. The incorporation of nano-sized MIL-101(Cr) (Figure 11.10), with the largest pore size of 3.4 nm, led to an exceptional increase in permeance, from 1.5 to 3.9 and from 1.7 to 11.1 l m−2 h−1 bar−1 for MeOH/PS and THF/PS, respectively. The

385

386

11 MOF Membranes for Vapor or Liquid Separation

1 μm

1 μm

(a)

(b) PI P84 support

Pt coating

PA

Ski

n la

0.2 μm (c)

yer

+M

OF

film

50 nm

P84 support 50 nm

0.1 μm (d)

Figure 11.10 (a) SEM image of TFC membrane surface after DMF dipping, (b) SEM image of TFN-MIL-101(Cr) [0.2% (w/v)] membrane surface after DMF dipping, (c) TEM image of TFN-MIL-101(Cr) [0.2% (w/v)] cross-sectional membrane lamella prepared using the FIB technique, and (d) TEM image of detached PA-MIL-101(Cr) thin-film surface, where the inset is at a higher magnification [88]. Source: Copyright 2013, ACS.

same group used in situ growth (ISG) of MOF within the pores of integrally skinned asymmetric polymer membrane supports for OSN applications [89]. Interfacial synthesis was utilized to produce a thin layer of HKUST-1 on polyimide P84 ultrafiltration supports. Two different fabrication methodologies were employed: methodology (A) resulted in the HKUST-1 layer growing above the polymer membrane surface, and methodology (B) resulted in HKUST-1 growth embedded in the surface of the polymer support membrane. The MOF TFCs produced via methodology (A) were shown to have similar solute retentions as ISG membranes; however the permeance values achieved were over three times higher than that of the ISG membranes. Vankelecom et al. prepared a continuous thin ZIF-8 membrane on porous polymeric supports via a simple interfacial synthesis method [85a]. When a zinc nitrate aqueous solution was brought in contact with a hexane-based 2-methylimidazole solution in the presence of a small amount of ethanol as cosolvent, the ligand and the zinc2+ reacted at the water/organic interface, leading to a white layer at the interface (Figure 11.11). White powder was obtained after removal of the liquid phase by centrifugation, followed by steps of washing and drying. The XRD pattern of this white powder confirmed the ZIF-8 formation. Although these membranes were prepared in just one synthesis cycle, these

11.3 Organic Solvent Nanofiltration

membranes showed excellent RB removal performances from different solvents. Compared with the solvothermal synthesis method, the present method is easy to upscale and can thus be used to prepare MOF membranes with larger surface area on different porous supports. In addition, this method will be extended to other types of MOFs, and the membrane applications will be extended to other membrane separation processes. This group also reported an advanced synthesis protocol for the preparation of TFN membranes. The evaporation-controlled filler positioning (EFP) method carefully pre-positioned the filler (ZIF-8 NPs) exactly at the water/solvent interface before the initiation of the IP. A permeance increase of 220% compared with the counterpart without filler was observed, corresponding to the high permeance increase without rejection loss. Moreover, optimal membrane performance was already obtained at an MOF concentration as low as 0.005 w/v%, which represented an 80 times reduction in the required amount of filler compared with other reported ZIF-8-based TFN membranes. This explicitly proved the importance of pre-positioning the filler right where it was needed, i.e. at the organic/water interface, before the IP. ZIF-8-based TFN membranes were prepared via both the EFP and the conventional synthesis methods, and a comparison between both protocols was made. Additionally, the influence of the size of the filler was investigated. ZIF-8 membranes on porous polyethersulfone (PES) supports were synthesized by using an interfacial synthesis method based on a liquid–liquid interfacial coordination mechanism (Figure 11.11) [90]. All studied parameters dramatically influenced the performance of the resultant ZIF-8 membranes. Higher molar ratios of 2-methylimidazole to zinc nitrate resulted in denser membranes with lower permeance and higher rejection, due to the dense packing and smaller interparticle spaces. By changing the molar ratio of 2-methylimidazole to zinc nitrate from 1.9 to 15.9, ZIF-8/PES membrane permeances varied from 5.6 to 37.5 l m−2 h−1 bar−1 , with RB rejections higher than 98%. The support had a very remarkable effect on the ZIF-8 membrane performance; 2% polymer concentration variation led to more than six times permeance change. The higher mass transfer resistance of a dense support created a concentration distribution of zinc nitrate in the pores during the reaction. The lower zinc nitrate concentration at the interface resulted in a higher ratio of 2-methylimidazole to zinc nitrate and subsequently denser ZIF-8 membranes. Unusually, by increasing the reaction time, higher permeance membranes were achieved, contrasting generally accepted rules from conventional IP. Using octanol as solvent not only created the immiscible interface but also served as a pore collapse preventer. As the higher boiling point octanol diffused into the pores of the supports if the contact time was long enough, the pore collapse was prevented during the later drying step. The present conclusions can serve as a base for other types of continuous MOF membrane synthesis. Wang prepared a tubular ceramic ZIF-8/PSS hybrid membrane through the in situ layer-by-layer (LBL) self-assembly method [91]. ZIF-8 particles in situ grew into PSS during the membrane formation, which resulted in their good compatibility and uniform dispersion within it. It was found that the membrane had excellent NF performances for the dye removal from water. Under optimized conditions, the membrane represented a flux of 210 l m−2 h−1 MPa−1 and rejection of

387

11 MOF Membranes for Vapor or Liquid Separation

(a)

0s

10s

50s Organic phase

3

CH

H N

H

H N

H

N

CH3

N

H N

N

N

CH

CH3

N

2-Methylimidazole

N H

3

N H

N

N

3

CH

H N

N H

H

N

N H

H C

N

3

N H

3

3

CH

C

C

H

40s

N

N

CH

N

N

N

CH3

CH

N

3

N N H

H

N

N

N H

CH

3

30s

N

CH

CH3

N H

3

N

N

CH3

3

CH

N H H

20s 3

N

H N N

(b)

3

CH

N

ZIF-8

Zn2+

Zinc source Supporting membrane

2+

Zn

2+

n

Zn2+

2+

Zn

2+ Z

Zn2+

Zn

2+

(c)

2+

Zn

Zn

2+

2+

Zn 2+

Water phase

2+

Zn 2+ 2+ Zn Zn2+ Zn 2+ Zn 2+

Zn

Zn2+

Zn

2+

Zn 2+

Zn

Zn

2+

2+ Zn Zn Zn 2+

2+

Zn 2+ Zn 2+

Zn

388

2+

Zn

(d)

ZIF-8/PES PES support ZIF-8 REF ZIF-8-i

5 μm

0

10

20 2θ

30

40

Figure 11.11 (a) Images of ZIF-8 synthesis using the interfacial method (ZIF-8-i), (b) schematic graph of interfacial synthesis of PES-supported ZIF-8 (ZIF-8/PES) membrane, (c) SEM images of the interfacially synthesized ZIF-8, and (d) XRD patterns of PES support, ZIF-8/PES membrane, and ZIF-8 prepared in DMF (ZIF-8 REF) and ZIF-8-i [90]. Source: Copyright 2015, RSC.

98.6% toward the MB NF from water. Furthermore, the mechanical stability of the hybrid membrane was enhanced through using the ceramic substrate. As a result, the hybrid membrane showed good pressure resistance ability and running stability. Combined with the advantages of the tubular module, this tubular ceramic ZIF-8/PSS hybrid membrane had a great potential for practical applications such as in the NF of dyes and other large molecules from water. Wang then reported a new concept of the composite membrane, namely, nano-confined composite membranes (NCC) [92], which was prepared via the fine-tuning contra-diffusion method. The barrier layer of the formed NCC membrane was of nearly zero thickness, which decreased the mass transfer resistance of water (Figure 11.12). The formed ZIF-11 NPs were in situ embedded in the nanopores of the PAN substrate. The particle size and distribution was controlled because of the confinement effect of the substrate. As a result, the formed ZIF/PAN membranes showed dramatically increased flux when compared with other NF membranes reported in the literature, without sacrificing rejection. The separation performance for

11.3 Organic Solvent Nanofiltration

Composite layer

Nanoparticle

Substrate

Substrate

Nanopore

Nano-confined layer

Composite layer Nanoporous layer Fingerlike pores Support layer TFC/TFN membrane

NCC membrane

Figure 11.12 Comparison of the microstructures of the TFC/TFN membrane and the NCC membrane [92]. Source: Copyright 2016, ACS.

the ZIF-11/PAN membrane had a flux of 464 l m−2 h−1 MPa−1 and a rejection of 98.4%. This membrane ran stably for 60 hours without performance degradation. The excellent separation performance as well as the good stability of the membrane makes it a great potential in NF. Developing advanced filtration membrane with high flux, good solute rejection, and excellent antifouling performance is highly demanded. Hydrophilic GO nanosheets are attractive fillers for the preparation of composite membranes for water purification [93]. To further improve the separation performance, Wang also grew ZIF-8 NPs onto the surface of GO sheets to form ZIF-8@GO composites, which were co-deposited with PEI matrix on tubular ceramic substrate through a vacuum-assisted assembly method [94]. The dispersion of ZIF-8 NPs in PEI matrix as well as the compactness and uniformity of the composite membranes was readily improved due to the templating effect of lamellar GO sheets and the transmembrane pressure. Membrane performance in OSN was evaluated on the basis of methanol permeance and retention of dye molecules. Methanol permeance increased when ZIF-8@GO laminates were embedded into the PEI layer, whereas the retention remained higher than 99%. The improvement of separation performance might be due to the well dispersion of ZIF-8 NPs in PEI matrix, which offered more well-defined pathways for solvent molecules. With similar idea but different from Wang’s study, Zhong and Liu reported that UiO-66 was specifically anchored to the GO layers as a porous modifier [95]. The incorporated UiO-66 effectively prevented the GO layers from stacking and introduced unique properties into the composite (UiO-66@GO). A series of novel composite membranes were prepared with the obtained UiO-66@GO composite and PES. As a result, the prepared composite membranes (UiO-66@GO/PES) exhibited high hydrophilicity and water purification performance. Especially, the water flux of composite membrane with 3.0 wt% UiO-66@GO loading showed an increase of 351% and 78%, respectively, in comparison with that of the PES and GO/PES membranes. Moreover, the UiO-66@GO/PES membranes exhibited good solute rejection and impressive

389

390

11 MOF Membranes for Vapor or Liquid Separation

antifouling performance, appealing for the application of industrial water purification. Liu and Zhang also reported the fabrication of antimicrobial TFN membranes containing a novel antimicrobial agent of ZIF-8/GO via IP [96]. The resultant hybrid nanosheets not only integrated the merits of both ZIF-8 and GO but also yielded a uniform dispersion of ZIF-8 onto GO nanosheets simultaneously, thus effectively eliminating the agglomeration of ZIF-8 in the active layer of membranes. A ZIF-8/GO thin-film nanocomposite (TFN-ZG) membrane with typical water permeability (40.63 l m−2 h−1 MPa−1 ) allowed for efficient bivalent salt removal (rejections of Na2 SO4 and MgSO4 were 100% and 77%, respectively). Furthermore, the synthesized ZIF-8/GO nanocomposites were verified to have an optimal antimicrobial activity (MIC, 128 μg ml−1 ) in comparison with ZIF-8 and GO separately, which sufficiently endowed the TFN-ZG membrane with excellent antimicrobial activity (84.3% for TFN-ZG3). Besides, the antimicrobial mechanisms of ZIF-8/GO hybrid nanosheets and TFN-ZG membranes were proposed. ZIF-8/GO functionalized membrane with high antimicrobial activity and salt retention denoted its great potential in water desalination, and they suggested that ZIF-8-based crystals might offer a new pathway for the synthesis of a multifunctional bactericide. MOFs are studied for the design of advanced nanocomposite membranes, primarily due to their ultrahigh surface area, regular and highly tunable pore structures, and favorable polymer affinity [97]. However, the development of engineered MOF-based membranes for water treatment lags behind. Here, TFN membranes containing poly(sodium 4-styrenesulfonate) (PSS) modified ZIF-8 (mZIF) in a PA) layer were constructed via a facile IP method [98]. The modified hydrophilic mZIF NPs were evenly dispersed into an aqueous solution comprising piperazine (PIP) monomers, followed by polymerizing with TMC to form a composite PA film. FTIR and XPS analyses confirmed the presence of mZIF NPs on the top layer of the membranes. SEM and AFM images evidenced a retiform surface morphology of the TFN-mZIF membrane, which was intimately linked to the hydrophilicity and adsorption capacity of mZIF NPs. Furthermore, the effect of different ZIF-8 loadings on the overall membrane performance was studied. Introducing the hydrophilic mZIF NPs furnished the PA layer with a better surface hydrophilicity, and more negative charge doubled more the original water permeability while maintaining a high retention of Na2 SO4 . The ultrahigh retentions of reactive dyes (e.g. reactive black 5 and reactive blue 2, >99.0%) for mZIF-functionalized PA membranes ensured their superior NF performance. This facile and cost-effective strategy would provide a useful guideline to integrate with other modified hydrophilic MOFs to design NF for water treatment. Li proposed another strategy, namely, coordination-driven in situ self-assembly, for the fabrication of MOF hybrid membranes [91]. In such a process, metal ions and organic ligands would assemble through the formation of coordination bonds. MOF particles were simultaneously generated in the polymer during the formation of the membrane, thus resulting in their good dispersion. The MOF loading could also be controlled by modifying the concentration of the MOF precursors and other parameters. In particular in some cases, the ability to form additional coordination bonds between MOF particles and the polymer itself not

11.3 Organic Solvent Nanofiltration

only significantly improved their compatibility but also enhanced the membrane stability. Furthermore, the coordination bonds between MOF particles and the polymer also changed the properties of the former, such as their hydrophilicity, adsorption ability, and selectivity. The resultant hybrid membrane should thus show good performance in separation applications. As an experimental proof of concept, a high-quality ZIF-8/PSS membrane was fabricated that showed excellent performance in the NF and separation of dyes from water. The conventional blending fabrication for TFN membranes is to disperse porous fillers in aqueous/organic phases prior to IP, and the aggregation of fillers may lead to the significant decrease in membrane performance [99]. To overcome this limitation, Lei proposed a novel LBL fabrication to prepare a PA/ZIF-8 nanocomposite membrane with a multilayer structure: a porous substrate, a ZIF-8 interlayer, and a PA coating layer [100]. The PA/ZIF-8 (LBL) membrane for NF applications was prepared by growing an interlayer of ZIF-8 NPs on an ultrafiltration membrane through ISG and then coating it with an ultrathin PA layer through IP. The obtained PA/ZIF-8 (LBL) membrane exhibited both better permeance and selectivity than the conventional PA/ZIF-8 TFN membrane because the ZIF-8 in situ grown produced a ZIF-8 interlayer with more ZIF-8 NPs but fewer aggregates. Compared with the pure PA membrane (the flux of 11.2 kg m−2 h−1 and rejection of 99.6%) for dye removal, the obtained PA/ZIF-8 (LBL) membranes achieved a significant improvement in membrane permeance and selectivity (flux up to 27.1 kg m−2 h−1 and the rejection of 99.8%). This LBL fabrication is a promising methodology for other polymer nanocomposite membranes simultaneously having high permeance and good selectivity. Micropollutants such as pharmaceuticals, pesticides, and endocrine-disrupting chemicals are currently the focus of a growing body of scientific research due to their potential risk to both human and aquatic environment [101]. A polytetrafluoroethylene (PTFE) double-layer microfiltration membrane has been modified with ZIF-8 and tested for micropollutant removal from water [102]. The physical properties of the modified membrane and its removal characteristics were tested using progesterone as a model pollutant. Compared with the unmodified membrane, incorporation of ZIF-8 onto the polymer matrix resulted in close to 40% increase in the membrane adsorption capacity and almost doubling its water permeability, which could result in significant reduction in processing specific energy consumption. The membrane maintained near 95% of its original removal efficiency after three regeneration cycles using polyethylene glycol 400. Modeling of the adsorption results agreed satisfactorily with the experimental measurements. The synergy provided by using ZIF-8 particles and the significant regeneration volume reduction ratio demonstrated the effectiveness of the approach for increasing the overall micropollutant removal capacity and allowed for its further consideration for development of viable continuous membrane adsorption techniques for decontamination of water resources. Pharmaceutical compounds that are hydrophilic and biologically persistent are being recognized as new unregulated contaminants in water. The hydrophilic characteristic deters the removal through biological process and adsorption. However, electrostatic repulsion and size sieving in NF can be used for the removal of these compounds. Although both NF and RO are options for removal

391

392

11 MOF Membranes for Vapor or Liquid Separation

of pharmaceutical compounds from aqueous feeds, the high pressure (and therefore high energy consumption) and membrane fouling remain serious concerns. Therefore development of improved membranes continues to be a key research area. MOFs are a class of porous crystalline compounds that offer advantages of large surface area and controlled porosity. MOFs have been used as adsorbents. More recently, MOF membranes and MOF-incorporated polymeric membranes have gained importance. In Basu’s work, MOF-incorporated polymeric membrane was prepared and tested for removal of the common analgesic acetaminophen (or paracetamol, MW 151 g mol−1 ) [103]. Thin-film PA composite membranes were synthesized by IP with ZIF-8. Two different structures were prepared and characterized: (i) PSF support membranes with ZIF-8 and PA separation layer and (ii) LBL PA/ZIF-8 nanocomposite membrane on top of PSF support. The latter synthesis protocol produced defect-free TFC membrane with 55% acetaminophen retention and permeance equivalent to conventional PSF/PA membrane. Shao reported the construction of additional solvent passageways in ultrathin PA NF membranes through the incorporation of hydrothermal stable UiO-66 MOFs [104]. The effects of functionalized ligands (UiO-66-(CH3 )2 , UiO-66-NH2 ) and post-synthetic functionalization (UiO-66(Ti)) on pore size, structure, and defects that consequently impacted membrane molecular transport were also studied. The intrinsic porosity of MOFs enhanced water and solvent permeation of PA membranes. The pure water permeances of PA/UiO-66 membranes were as high as 15.4 l m−2 h−1 bar−1 while rejecting 100% of RB and 97.6% azithromycin dissolved in water. Ascribed to the hydrostability of the UiO-66 MOFs, the high water permeance of the PA/UiO-66 membrane was maintained over 100 hours continuous operation. Functionalized UiO-66 MOFs studied here also imbued excellent separation performances to PA membranes during operation in organic solvents such as alcohols, ketones, and ethyl acetate. Added functionalities and defects in UiO-66 tailored pore sizes, and structures consequently impacted the additional passageways for water and solvent transport and membrane separation efficiency. This indicated that permeance enhancements of these membranes were highly dependent on the MOF surface area. The MOFs of Cr carboxylate MIL-101(Cr) and ZIF-11 have different chemical and textural properties. These properties have been combined to prepare TFN membranes for OSN with the two MOFs simultaneously embedded in the same membrane [105]. The TFN membrane containing ZIF-11 permeated faster than that with MIL-101(Cr) when filtering sunset yellow (SY) (4.9 l m−2 h−1 bar−1 ) and acridine orange (AO) (3.2 l m−2 h−1 bar−1 ), although MIL-101(Cr) alone gave rise in both cases to better rejections (above 90%). The combination of the two MOFs led to a versatile TFN membrane that showed an intermediate performance that improved the rejections given by the TFN membrane synthesized with ZIF-11 and the permeances corresponding to the TFN membrane synthesized with MIL-101(Cr). The effect of the temperature on the TFN membranes designed for this work was also studied, obtaining apparent activation energies of 13.2 ± 2.1 and 8.3 ± 1.1 kJ mol−1 for the OSN of pure methanol and SY methanol, respectively.

11.3 Organic Solvent Nanofiltration

The developments of MOF/polymer hybrid membranes are hampered by the problems of MOF particle agglomeration and harsh synthetic conditions of MOFs. In Wang’s paper, they provided a facile and simple route for fabrication of a uniform MOF/polymer hybrid membrane via the combination of self-assembly and interfacial reaction method [106]. The positively charged PEI molecules and zinc ions were first deposited on the negatively charged hydrolyzed polyacrylonitrile (HPAN) substrate via electrostatic attraction (self-assembly), and then ZIF-8 particles were in situ formed in a PEI layer via interfacial reaction, thereby forming a thin and uniform ZIF-8/PEI hybrid membrane. The prepared membrane possessed a rejection of 99.6% for MB aqueous solutions, with permeance up to 33.0 l m−2 h−1 bar−1 . These results suggested that the ZIF-8/PEI hybrid NF membrane, which was prepared under facile conditions without any complicated posttreatment, can be a promising candidate for dye removal process. Tang reported a novel positively charged NF membrane, which was fabricated by incorporating MOFs into chitosan polymeric matrix for enhanced removal of multivalent cations [107]. The synthesized MOFs, NH2 -MIL-101(Al) and NH2 -MIL-101(Cr), could be homogeneously dispersed in the chitosan polymeric matrix. The morphologies of MOFs had a significant influence on the permeability of NF membranes. NF membrane filled with NH2 -MIL-101(Al) with rodlike structure attained two times higher flux but similar rejection compared with that filled with NH2 -MIL-101(Cr) with grainy structure. These positively charged MOF/chitosan NF membranes were able to reject up to 93.0% of MgCl2 , and the salt rejection followed the order of MgCl2 > CaCl2 > NaCl > Na2 SO4 . Preparation of defect-free and optimized TFN membranes is an effective way to enhance the process of OSN. However, it still remains a great challenge due to poor filler particle dispersibility in organic phase and weak compatibility between fillers and polymers. Aiming at these difficulties, UiO-66-NH2 NPs were surface modified with long alkyl chains and used in the preparation of TFN membranes [108]. As a result, defect-free TFN membranes with ultrathin MOF@polyamide layer were successfully prepared, benefited from the improved particle dispersibility in n-hexane. Significant enhancement was found in methanol permeance after NP incorporation, without comprising the tetracycline rejection evidently. Especially, the novel TFN membrane prepared with organic phase solution containing 0.15% (w/v) modified UiO-66-NH2 NPs showed a superior methanol permeance of 20 l m−2 h−1 bar−1 and a tetracycline rejection of about 99%, appealing to the application in pharmaceutical industry for example. In the study reported by Chen [109], a novel TFN membrane was prepared by incorporating covalent organic frameworks (COFs) into PA layer on a PES substrate through IP method. The porous structure of COFs (SNW-1) provided more passageways for water transport. Meanwhile, the reaction between SNW-1 and TMC during IP formed strong covalent bonds for better interfacial compatibility. The surface hydrophilicity of the hybrid membrane was improved due to the existence of amine-rich SNW-1. The influence of the IP conditions and PIP and the loading of SNW-1 on the membrane performance were investigated. The resultant PA-SNW-1/PES membrane with a SNW-1 loading of 1 g m−2 exhibited an

393

394

11 MOF Membranes for Vapor or Liquid Separation

increased pure water flux from 100 to 192.5 l m−2 h−1 MPa−1 compared with the pristine PA membrane, while the rejection for Na2 SO4 maintained above 80%. Moreover, the membrane also showed a long-term running stability (Table 11.3).

11.4 Chiral Resolution In the last decade, considerable attention has been paid to chiral resolution on account of significant differences in biological and pharmacological properties of the isomers of chiral compounds [116]. One has the desired effect, while the other may be inert or even harmful [117]. The optical resolution of racemic mixtures is challenging due to their similar property. Various separation techniques have been applied to reach this target, including thin-layer chromatography (TLC), gas chromatography (GC), high-performance liquid chromatography (HPLC), etc. [118] However, all of these methods suffer from drawbacks such as small separation amounts per run, batch processing, and high cost. MOFs are typically synthesized by assembling metal ions with organic ligands in appropriate solvents. They have crystalline structures, large internal surface areas, and uniform cavities. It is worth mentioning that the design and modification of MOFs at the molecular level can be generally achieved through tuning or controlling their pore size and shape. MOFs with chiral channels can be built via chiral ligands. Similarly, membranes made from chiral MOFs are expected to offer unique opportunities to be applied in the field of chiral resolution. MOF-based film materials were firstly used to selective adsorption of chiral isomers, and then polycrystalline MOF membranes were applied for chiral resolution via membrane separation process [119]. Enantiomer adsorption and separation of chiral molecules is an important field of chemistry and has many important applications in pharmaceutical, agricultural, and chemical engineering [120]. The chiral resolution was achieved by selective adsorption based on the chiral MOF films. The case of membrane and thin-film fabrication of MOFs on the substrates exhibits striking advantages compared with powder materials [121]. In particular, surface-anchored MOFs (SURMOFs) grown by the liquid-phase epitaxy (LPE) method afford the preparation of hetero-SURMOFs by changing the metal nodes and/or linker molecules in successive deposition cycles. The LPE approach is based on the sequential immersion of a surface-modified substrate into the solutions of metal salts and organic ligands. Between each step, the substrates were washed with solvent to remove the unreacted metal ions or organic linkers. The enantiopure metal–camphorate frameworks Zn2 (d/l-cam)2 dabco (diazabicyclo[2.2.2]octane [dabco]) were grown (20–40 cycles) by dipping the quartz crystal microbalance (QCM) substrate alternately in ethanol solutions of Zn(Ac)2 H2 O and equimolar (d or l)-cam/dabco mixtures, each step followed by immediately rinsing with pure ethanol (Figure 11.13) [123]. The QCM is an ultrasensitive device capable of sensing mass changes in the nanogram range. The deposition of enantiopure directly on the QCM substrate enables the adsorption kinetics of enantiomeric probe molecules to be monitored by QCM and thus allow

Table 11.3 MMMs based on MOFs for organic solvent nanofiltration process.

MOF

Polymer

Mixture

Pressure

Flux or permeance −2

−1

h

−1

bar

Separation factor or rejection ratio

Reference

90%

[86]

ZIF-8

PA

Aqueous salt solutions

4 bar

4lm

MIL-53

PMIA

Brilliant blue G in ethanol

10 bar

0.7 l m−2 h−1 bar−1

94%

[110]

ZIF-8

GO

Aqueous salt solutions

0.4 MPa

40.63 l m−2 h−1 MPa−1

Na2 SO4 (100%), MgSO4 (77%)

[96]

ZIF-8

PA

Dye

8 bar

14.90 l m−2 h−1

>99%

[98]

MIL-101 ZIF-11

PA

Dye

20 bar

4.8 l m−2 h−1 bar−1

87.90%

[105]

MIL-101(Cr)

PA

Aqueous salt solutions

10 bar

3.91 l m−2 h−1 bar−1

>90%

[111]

ZIF-8

PTFE

Micropollutant removal

—

5.48×104 l m−2 h−1 bar−1

95%

[102]

UiO-66-NH2

PA

Tetracycline

5 bar

20 l m−2 h−1 bar−1

99%

[108]

ZIF-8

PA

Dye

1 MPa

27.1 kg m−2 h−1

99.80%

[100] [112]

ZIF-8

PA

Dye

0.4–1.2 MPa

22.6 l m−2 h−1 Pa−1

>98.5%

ZIF-8

PSS

Dye

0.5 MPa

265 l m−2 h−1 MPa−1

98.60%

[91]

NH2 -MIL-53(Al)

PA

MeOH/PS

30 bar

2.3 l m−2 h−1 bar−1

1.8

[88]

MIL-53(Al)

PA

MeOH/PS

30 bar

2.3 l m−2 h−1 bar−1

1.9

[88]

ZIF-8

PA

MeOH/PS

30 bar

2.5 l m−2 h−1 bar−1

2.1

[88]

MIL-101(Cr)

PA

MeOH/PS

30 bar

4.2 l m−2 h−1 bar−1

3.9

[88]

HKUST-1

P84

Polystyrene markers

10 bar

54 l m−2 h−1 bar−1

40%

[89]

UiO-66

PA

Azithromycin

10 bar

15.4 l m−2 h−1 bar−1

97.60%

[104] (Continued)

Table 11.3 (Continued)

MOF

Polymer

Mixture

Pressure

Flux or permeance

Separation factor or rejection ratio

Reference

NH2 -MIL-101

chitosan

Aqueous salt solutions

5–15 bar

4 l m−2 h−1 bar−1

93.00%

[107]

Cu3 (BTC)2

PDMS

Dye

44 bar

0.5 l m−2 h−1 bar−1

>95%

[86]

MIL-47

PDMS

Dye

44 bar

0.5 l m−2 h−1 bar−1

>95%

[86]

MIL-53(Al)

PDMS

Dye

44 bar

0.5 l m−2 h−1 bar−1

>95%

[86]

ZIF-8

PDMS

Dye

44 bar

0.5 l m−2 h−1 bar−1

>90%

[86]

ZIF-8

PES

Dye

—

1.3 l m−2 h−1 bar−1

98.90%

[90]

ZIF-8

PA

Aqueous salt solutions

10 bar

2.5 l m−2 h−1 bar−1

72%

[113]

ZIF-8

PES

Dye

0.2 MPa

37.5 l m

>98%

[114]

ZIF-8

PSS

Dye

0.5 MPa

210 l m−2 h−1 MPa−1

98.6%

[115]

ZIF-11

PAN

Dye

0.5 MPa

464 l m−2 h−1 MPa−1

98.40%

[92]

ZIF-8

PEI

Dye

2 bar

33–51 l m−2 h−1 bar−1

81%-99.6%

[106]

−2

−1

h

−1

MPa

ZIF-8

GO/PEI

Dye

0.5 MPa

3.5 l m−2 h−1 bar−1

99.10%

[94]

UiO-66@GO

PES

Dye

2.5 MPa

15 l m−2 h−1 bar−1

98.30%

[102]

11.4 Chiral Resolution

SAMs Pyridyl or COOH

MII solution

L+P solution

yl

id

r Py

CO

OH

(110)

(001)

Figure 11.13 Top: two principal growth directions and of [Zn2 (cam)2 (dabco)]n . Bottom: schematic illustrations of oriented growth in the (001) orientation on pyridyl-terminated (PPMT) and the (110) orientation on COOH-terminated (MHDA) SAMs on gold substrates [122]. Source: John Wiley & Sons.

assessing the adsorption enantioselectivity of the MOF thin film. The results indicated that Zn2 (d-cam)2 (dabco) displayed a roughly 1.5-fold preference for adsorption of (2R,5R)-2,5-hexanediol (R-HDO) over (2S,5S)-2,5-hexanediol (S-HDO), whereas Zn2 (l-cam)2 (dabco) exhibited the inverse selectivity, a 1.5-fold adsorption of S-HDO over R-HDO. This very significant difference was attributed to the different interaction of R- and S-HDO with MOFs containing d- and l-linkers.

397

398

11 MOF Membranes for Vapor or Liquid Separation

The enantioselectivity of isoreticular chiral MOFs with identical stereogenic centers and different pore sizes was investigated as well. A series of homochiral pillared-layer [Cu2 (d-cam)2 (P)] were prepared with an identical chiral layer [Cu(d-cam)] and different pillar linkers P: dabco, 4,4-bipyridine, (bipy) and 1,4-bis-(4-pyridyl)benzene (bipyb) [124]. The data showed clearly that the pore size had a significant impact on the enantioselectivity. The enantioselectivity did not follow such a simple trend as the adsorption capacity, which increased with increasing pore size. The highest enantiomeric excess (ee) was found for a pore size of 0.8 nm, which was found to be the medium case. It was assumed that the differences of the loadings were caused by the different alignments of the chiral guest molecules adsorbed in the pores, where the stereogenic centers had a different impact on the enantiomer selectivity. If the pore size is “too” small, the guest molecules are “forced” to adsorb in the pores in such a position, where the impact of the stereogenic center in the framework is small. If the pore size is “too” large, the molecules can adsorb all over the large pore and the impact of the stereogenic center is small, too. If the pore size is well adjusted, roughly as large as the guest molecule, the stereogenic center has the highest impact on the guest molecule, resulting in the highest enantiomer separation. This study demonstrates that not only the stereogenic center but also the pore size has to be adjusted for gaining highest enantioselectivities in chiral nanoporous materials and thereupon enabling a significantly more efficient enantiomer separation. Furthermore, the liquid-phase epitaxial growth of MOF [Cu2 (d-cam)2 (dabco)] thin films enables the possibility to investigate the enantiomer separation from a racemic mixture by means of circular dichroism (CD). The ee of (+)-ethyl-d-lactate [(+)EtLt] and (−)-ethyl-l-lactate [(−)EtLt] in [Cu2 (d-cam)2 (dabco)] was investigated by CD [125]. By comparing the CD spectra of the pure EtLt and the enantiomer-pure EtLt adsorbed in the [Cu2 (d-cam)2 (dabco)] with the spectrum measured for the racemic EtLt mixture adsorbed in the [Cu2 (d-cam)2 (dabco)], an ee (+) versus (−) of 28% was found. The enantiomer separation property was studied by GC technology, which is a typical analytical method for separating and analyzing compounds without decomposition. In order to obtain a homogeneous sample and perform the experiment of GC enantiomer separation well, the homochiral MOF [Cu2 (d-cam)2 P] (P = dabco and bipy) thin films were grown in the capillary column directly using LPE in a well-defined LBL fashion [126]. The morphology and cross section of the MOF thin film in the column were checked by the SEM, showing that MOF Cu2 (d-cam)2 P (P = dabco and bipy) was grown in the inner wall of the capillary column with compact and homogeneous film using LPE LBL fashion successfully. By comparing GC retention time of probe molecules in a poly(l-DOPA) (3,4-dihydroxy-l-phenylalanine [l-DOPA]) functionalized capillary column, Cu2 (d-cam)2 P grown in achiral hydroxyl and enantiopure poly(l-DOPA) functionalized capillary columns and [Cu(d-cam)P] grown in a poly(l-DOPA) functionalized column exhibited excellent separation efficiency for methyl lactate, respectively. The explored chiral substrate-mounted homochiral MOF thin film had high efficiency separation, which was effected by the chiral poly(l-DOPA) functionalized substrate. This work not only indicated that the LPE process might be a promising approach for introducing multiple

11.4 Chiral Resolution

chiral layers or films as stationary phases for highly effective separation but also demonstrated that a chiral poly(l-DOPA) substrate had the ability to improve enantiomer separation for homochiral MOF thin films. Gu described the fabrication of hybrid well-ordered porous NP arrays with full three-dimensional periodicity by embedding nanometer-sized metal–organic clusters (MOCs) into MOFs [127]. Although conventional NP@MOF encapsulation procedures failed for the fairly large (1.66 nm diameter) NPs, It achieved maximum loading efficiency (one NP per pore) by using a modified LPE LBL approach to grow and load the MOF. The preformed NPs, homochiral Ti4 (OH)4 (R/S-BINOL)6 clusters (Ti-MOC, BINOL = 1,1′ -bi-2-naphthol), formed a regular lattice inside the pores of an achiral HKUST-1 MOF thin film. Exposure to the different enantiomers of methyl lactate revealed that the NP@MOF metacrystal was quite efficient regarding enantiomer recognition and separation. The approach presented here is also suited for other MOF types and expected to provide a substantial stimulus for the fabrication of metacrystals, crystalline solids made from NPs instead of atoms. The mechanism between chirality and enantio-adsorption/separation is very significant in homochiral porous materials; in particular, the understanding of the relationship between crystalline orientations and chiral behavior is a challenging but important mechanism. In another work reported by Gu [128], homochiral porous crystalline MOF materials were grown on hydroxyl- and carboxyl-functionalized substrates, resulting in homochiral porous thin films with different orientations. The enantioselectivity and adsorption rates in two different oriented homochiral porous thin films were studied by using gas-phase QCM experiment of chiral probe molecules. The different mass uptake and time constant showed that the chiral behavior was obviously influenced by the crystalline orientations on the same homochiral porous thin films. This study not only offers a good model to understand the mechanism of chiral behavior in homochiral porous materials but also provides guidance for developing new homochiral-oriented porous thin films with high enantioselectivity or enantioseparation. A self-polymerized chiral monomer l-DOPA has been introduced into the pores of an achiral SURMOF [129], and then the homochiral poly(l-DOPA) thin film has been successfully formed after UV light irradiation and etching of the SURMOF. Remarkably, such a poly(l-DOPA) thin film exhibited enantioselective adsorption of naproxen. This study opens a SURMOF-templated approach for preparing porous polymer thin films. For the membrane-based separation, the first reported example is the preparation of a homochiral [Zn2 (bdc)(l-lac)(dmf )] membrane by solvothermal reaction [123]. In Jin’s study, reactive seeding was used to ensure strong binding of the membrane onto the porous ZnO support. The membrane was extensively characterized, and no defects or macroporosity was found. The enantiomeric resolution potential was evaluated by a test mixture of R/S-methyl phenyl sulfoxide. A homemade side-by-side diffusion setup was used for the separation experiment. After 48 hours, a maximum ee value of 33% was obtained for a feed concentration of 2%. The performance dropped drastically at higher feed concentrations (5–50 mmol l−1 ). Adsorption experiments using [Zn2 (bdc)(l-lac)(dmf )] powder confirmed the intrinsic chiral selectivity of the material and indicated the origin

399

400

11 MOF Membranes for Vapor or Liquid Separation

of the membrane’s selectivity. Simulations supported this hypothesis as diffusion of the R isomer was lower due to higher interaction energy with the pores, compared with the S isomer. In another study, chiral Ni2 (l-asp)2 (bipy) (Ni-LAB) membrane was prepared on the support of porous aluminum oxide [130]. The membrane was fabricated via a secondary growth process. Submicron-sized seed crystals were obtained by ball-milling Ni-LAB to reduce the crystal size of the as-synthesized sample. In a second step, a ceramic support was seeded using dip coating after which a secondary growth step took place. The separation of 2-methyl-2,4-pentanediol through the membrane was achieved preferentially with the l-asp unit, and ee values up to 35.5 were obtained. For the same structure of MOFs, Kang et al. used a very fine nickel mesh for the preparation of chiral membrane by ISG (Figure 11.14) [131]. The nickel mesh served two roles: it provided the support for the MOF crystals and also served as the single metal source for the MOF synthesis. After the in situ reaction process, a single crystal layer and defect-free membrane was obtained. The separation of chiral molecules (R,S)-2-methyl-2,4-pentanediol was tested in a PV setup and resulted in an ee of 32.5% at 200 ∘ C. A new homochiral MOF membrane [Ni2 (mal)2 (bpy)]⋅2H2 O (Ni-MB) was synthesized by a secondary growth technique, in which high-energy ball milling was employed to prepare nano-sized MOF seeds, and their synthesis conditions were

200 μm

(a)

(c)

(b)

50 μm

(d)

100 μm

50 μm

Figure 11.14 (a) Leica picture of the surface of the Ni2 (L-asp)2 (bipy) membrane; SEM pictures of the surface of (b) the Ni2 (L-asp)2 (bipy) membrane and (c) details of the densely packed crystallites; (d) a cross-sectional SEM picture of the Ni2 (L-asp)2 (bipy) membrane [131]. Source: RSC.

11.5 Stability of MOF Membranes

Table 11.4 MOF membranes or films for chiral separation process.

MOF

Solute

ee value (%) or separation factor

Process

Reference

[Zn2 -(bdc)(l-lac)(dmf )]

Sulfoxides (methyl, phenyl, bromo, nitro)

6.49–33

Membrane

[123]

Ni2 (l-asp)2 (bipy)

2-Methyl-2,4pentanediol

42.4–73.5

Membrane

[130]

Ni2 (l-asp)2 (bipy)

2-Methyl-2,4pentanediol

35.5

Membrane

[131]

(±)Zn2 (cam)2 (dabco)

2,5-Hexanediol

𝛼 = 1.55 (+), 0.66 (−)

Film adsorption

[122]

[Cu2 (d-cam)2 (dabco)]n / [Cu2 (l-cam)2 (dabco)]n

Ethyl-d-lactate

28

Film adsorption

[125]

Cu2 (d-cam)2 (L)

Limonene

34

Film adsorption

[124]

[Cu2 d-cam2 L]n

Methyl lactate

53

Film adsorption

[128]

Ti-MOC@HKUST-1

Methyl d-lactate/methyl l-lactate

—

Film adsorption

[127]

Zn2 (bdc)(l-lac) (dmf ) DMF

1-Phenylethylamine

𝛼 = 2.2

Film adsorption

[133]

optimized [132]. The SEM results indicated that the membrane was continuous and integrated without observable defects; the permeation measurements of single gas showed that the permeate flux of nitrogen was independent of the transmembrane pressure, further confirming the absence of macroporous defects in the prepared chiral MOF membranes. It is expected that the proposed chiral MOF membrane can be potentially applied to enantioselective separation in the near future (Table 11.4).

11.5 Stability of MOF Membranes For most MOFs, one of the major drawbacks is their poor (hydrothermal or chemical) stability, which tells against their practical applications [9a, 134]. Development of novel stable MOF structures and methods to enhance MOF stability are important research topics, which are critical for the separation process on MOF membranes. To date, only a few MOFs are reported to possess satisfactory hydrothermal stability, e.g. the ZIF family [30a], MIL (Materials Institute Lavoisier) analogues [135], and some zirconium- and pyrazolate-based MOFs [48, 136]. Otherwise, the post-enhancement of hydrothermal stability of the already existing MOF materials is also an attractive option and has

401

402

11 MOF Membranes for Vapor or Liquid Separation

become one of the most active research domains nowadays [134, 137]. The post-enhancement methods, however, need to meet the requirements of keeping the porosity of the parent MOFs after modification or reinforcement. The perceived robust stability of ZIFs was regarded as an important feature of ZIFs that have intrigued intensive research efforts. ZIF-8 is currently one of the most stable MOFs. Professor Lin’s work demonstrated that the gas permeance on ZIF-8 membrane for H2 , He, N2 , and SF6 hardly changed during days of storage in lab [138]. The propylene and propane permeance did not change either during the 27 days of the off-stream storage. The ZIF-8 membrane exhibited constant permeance and selectivity for C3 H6 /C3 H8 during the entire period of off-stream and on-stream tests. This work showed that ZIF-8 membranes were extremely stable under both off-stream storage and on-stream C3 H6 /C3 H8 separation conditions. The hydrolysis of ZIF crystals was then identified under hydrothermal conditions. In the study of Lin, the dynamic stability of α-alumina-supported ZIF-8 membranes and static stability of both ZIF-8 membranes and powder in water were studied via PV and batch durability tests at room temperature [139]. The sharp increase in water PV flux from 0.04 to 1.39 kg m−2 h−1 and the disappearance of ZIF-8 structure after PV were consistent with morphology and phase structure data, showing that ZIF-8 layer on the α-alumina support underwent degradation during water PV. The gradual degradation of ZIF-8 layer was also experimentally observed in batch durability tests in pure water. To better understand the mechanism of stability of ZIF-8 structure, control experiments on hydrolysis of ZIF-8 powder in aqueous conditions were designed and performed. The continuous release of Zn2+ from ZIF-8 structure to the aqueous solution was detected along with the change in pH value of the aqueous system. The presence of 2-methylimidazolate in aqueous solution could stabilize the ZIF-8 structure of both membrane and crystals. These evidences confirmed that ZIF-8 crystals degraded in water due to the dissolution of ZIF-8. The rate and extent of the dissolution depended on the ZIF-8 crystal to water mass ratio. To enhance the stability of ZIFs under hydrothermal conditions, a partial ligand exchanging process referred to shell-ligand-exchange-reaction (SLER) was carried out by Li and Yang [134]. 2-MIM ligands in the ZIF-8 crystal out layer were replaced by 5,6-dimethylbenzimidazole (DMBIM), making the surface of crystals hydrophobic. Besides ZIF-8, the SLER methodology was also successfully applied to stabilize other types of ZIFs, such as ZIF-7 and ZIF-93. After the SLER process, the obtained ZIF materials kept the original porosity and gave an improved stability under hydrothermal tests. In the later work, the SLER method was extended to the ZIF-8 membrane by Lin’s group. It showed the improvement of ZIF-8 crystal and membrane stability in water after the replacing methylimidazole ligands on the outer surface of ZIF-8 membranes by the more hydrophobic and bulkier DMBIM [73]. Ligand exchange modification does not change the crystal structure, morphology, or gas permeance as shown in ZIF-8 powder and membrane characterization experiments (Figure 11.15). The mZIF-8 powders and membranes retained identical morphologies and crystallinities after static and dynamic water immersion. The mZIF-8 membrane exhibited stable water PV flux controlled by the ZIF-8 layer, while the unmodified ZIF-8 membrane

11.5 Stability of MOF Membranes

(a)

(b)

70°

10 μm

Intensity (a.u.)

(c)

∗ ZIF-8

∗

∗∗

105°

10 μm

α-Alumina Modified membrane

(d)

∗ ∗∗ Pristine membrane

10

20

30 2 θ (°)

40

50

Figure 11.15 Comparison of ZIF-8 membranes in surface micrographs, (a) pristine and (b) modified ZIF-8 membrane, and corresponding static water contact angle (the inset of (a) and (b)), (c) XRD patterns, and (d) appearance change of a ZIF-8 membrane before and after modification [73]. Source: Copyright 2017, Elsevier.

experienced dissolution of the ZIF-8 layer during water PV. This ligand exchange strategy enabled hydrostable ZIF-8 membranes for practical applications involving aqueous solutions. Another effective method to solve the stability problem is to reverse the surface polarity of MOFs from hydrophilicity to hydrophobicity. In the work reported by Professor Yu’s group, hydrophobic PDMS was deposited facilely on the surface of MOF materials [137b]. This method was applied to several classic MOFs, and the modified materials possessed better moisture or water resistance compared with the pristine MOFs and maintained the original porosity and surface areas. Such a coating process can expand the application of MOF membranes for the gas separation at humid condition. In a recently reported work, this work has been used to switch the surface wettability of MOF-coated mesh for oil/water separation [140]. Similar to Yu’s idea, a facile method was developed to improve the feasibility of water-unstable MOFs in an aqueous environment via imbedding in a polymer monolith [141]. The effect of compartment type during polymerization played a significant role in maintaining the crystalline structure and thermal stability of the MOFs, which was confirmed by PXRD and TGA,

403

404

11 MOF Membranes for Vapor or Liquid Separation

respectively. The MOF/polymer composite prepared in a narrow compartment (column, ID 0.8 mm) had better thermal and chemical stability than that prepared in a broad compartment (vial, ID 7 mm). The developed MOF/polymer composite was applied as an adsorbent in solid-phase micro-extraction of nine nonsteroidal anti-inflammatory drugs (NSAIDs) and could be used for extraction more than 30 times, demonstrating the proposed approach promising for industrial applications. For another kind of famous stable MOF, continuous Zr-MOFs have also been prepared into membranes. The UiO-66 membranes were fabricated on alumina HF using an in situ solvothermal synthesis method [49]. The water amount in the synthesis solution was a key point to obtain high-quality intergrown UiO-66 membrane. Single gas permeation and ion rejection tests were carried out to confirm membrane integrity and functionality. The membranes exhibited chemical stability during the ion removing from a wide range of saline solutions up to 170 hours. The stable MOFs based on Zr with a wide range of pore sizes and various functional groups have been reported in the last few years. Furthermore, post-modification and ligand exchange can also be carried out to adjust the pore size of Zr-MOF. Zhao presented an efficient approach for healing the intrinsic defects in water-stable polycrystalline UiO-66(Zr)-(OH)2 membranes by PSDH [53]. This approach is based on the dynamic feature of coordination bonds in MOFs and is reminiscent of the solvent-assisted linker exchange reactions in bulk MOF crystals. The hydrothermal stability of MOFs is especially dominated by the strength of metal–ligand bonds [142]. Thus, researchers are looking for more robust coordination bonds such as Hf-O (802 kJ mol−1 ) or Zr-O (776 kJ mol−1 ) to construct MOF-based membranes. Despite their superior water stability, Zr-MOFs are increasingly found to have intrinsic defects caused by ligand missing, leading to dramatically enhanced porosity and aperture size that are unfavorable for molecular sieving separation [52]. Theoretically, the non-bridging groups such as water molecules in UiO-66(Zr) [143] can be fully displaced by fitting dicarboxylates, which can seal the framework defects by bridging two adjacent clusters together [144]. Zhao hypothesized that by replacing two mono-coordinated moieties with one bridging dicarboxylate ligand, a process called post-synthetic defect healing, the defective large aperture size can be reduced, affording better separation performance of the resultant membranes. Briefly, the as-synthesized membrane was solvothermally treated in ligand solution at 120 ∘ C for one day. As expected, the coordinative defects were mitigated after PSDH: each Zr6 O4 (OH)4 cluster was coordinated with five ligands on average, meaning 24% of the coordinative defects were mitigated by re-coordinating with ligands. This approach offers a novel platform for further optimization and functionalization of polycrystalline MOF membranes from molecular levels. In the future, intensive research work will be highly desired to optimize the membrane structure as well as to expand the applications in niche markets.

11.6 Conclusions and Outlook

11.6 Conclusions and Outlook The membranes based on MOF materials have a huge potential for practical separation applications. The successful development of water-stable MOFs and post-synthetic modification of MOFs offer an alternative way to broaden the scope of applications for MOF-based membranes from gas absorption, separation, and catalysis to new fields such as liquid separation. Many new strategies for fabricating MOF-based membranes in liquid separation have emerged; clear improvements have been made in the control of key parameters, such as the dispensability, stability, and hydrophilicity/hydrophobicity. Understanding of the preparation mechanism for MOF-based membranes, as well as for liquid separation applications, is steadily increasing. To date, conventional preparation methods including ISG and secondary growth as well as blending routes have been widely explored to produce continuous MOF membranes or MOF-based MMMs, and IP routes have been developed quickly in recent years to prepare MOF-based TFN/TFC membranes. Moreover, novel membrane materials such as GO have been combined with MOFs to form hybrid membranes for efficient separation. Beyond the classic MOFs that have been made into membrane, an abundance of structures are available to be applied in the field of membrane-based vapor and liquid separation. More potential MOFs with low-cost ligands, scale-up preparation methods, and stable and suitable structures will be found and selected from the database, and computational studies may play the role of a guide for effectively seeking appropriate candidates. Jiang reported an atomistic simulation study for seawater PV through five ZIF membranes including ZIF-8, ZIF-93, ZIF-95, ZIF-97, and ZIF-100 [145]. Salt rejections in these five ZIFs were predicted to be 100%. Taking advantages of the largest pore size, ZIF-100 possessed the highest water permeability of 5 × 10−4 kg mm−2 h−1 bar−1 , which was substantially higher compared with commercial RO membranes, as well as zeolite and GO PV membranes. In ZIF-8, ZIF-93, ZIF-95, and ZIF-97 with similar aperture size, water flux was governed by framework hydrophobicity/hydrophilicity; in hydrophobic ZIF-8 and ZIF-95, water flux was higher than in hydrophilic ZIF-93 and ZIF-97. The lifetime of hydrogen bonds in ZIF-93 was found to be longer than in ZIF-100. This kind of simulation study provides a microscopic insight into water desalination in MOF materials and reveals the key factors (aperture size, polarity of functional group, and framework hydrophobicity/hydrophilicity) governing water flux. For the separation applications, both polycrystalline MOF membranes and MOF-based hybrid membranes have been studied for vapor and liquid separations. Surveying the separation fields of MOF-based membranes over the past decade, several milestones can be defined: from separating single components in liquid environments then to the synchronous separation of multicomponent and multifunctional MOF-based membranes. In addition, the combination of MOFs with other species to form novel multifunctional nanocomposites can further influence wider applications, due to the fact that they have the advantages of both components.

405

406

11 MOF Membranes for Vapor or Liquid Separation

There are still several challenges for the practical application of MOF-based membranes, including the scale-up and long-term stability issues. Simpler, cheaper, and high-yield synthesis methods are expected to lead to the preparation of MOF membranes on the large scale. Additionally, MOF membranes with enhanced stability are required for practical separation conditions.

References 1 Sholl, D.S. and Lively, R.P. (2016). Nature 532: 435–437. 2 (a) Baker, R.W. (2012). Membrane Technology and Applications, 3e. Newark,

CA: Wiley. (b) Baker, R.W. (2002). Ind. Eng. Chem. Res. 41: 1393–1411. 3 (a) Lin, Y.S. and Duke, M.C. (2013). Curr. Opin. Chem. Eng. 2: 209–216.

4 5 6

7 8 9

10 11

12 13 14 15 16 17 18 19 20

(b) Park, H.B., Kamcev, J., Robeson, L.M. et al. (2017). Science 356: eaab0530. Freeman, B.D. (1999). Macromolecules 32: 375–380. Coronas, J. and Santamaría, J. (1999). Sep. Purif. Rev. 28: 127–177. (a) Shah, M., McCarthy, M.C., Sachdeva, S. et al. (2012). Ind. Eng. Chem. Res. 51: 2179–2199. (b) Furukawa, H., Cordova, K.E., O’Keeffe, M., and Yaghi, O.M. (2013). Science 341: 1230444. (a) Kang, Z., Xue, M., Fan, L. et al. (2014). Energ. Environ. Sci. 7: 4053–4060. (b) Kang, Z., Fan, L., and Sun, D. (2017). J. Mater. Chem. A 5: 10073–10091. Li, X., Liu, Y., Wang, J. et al. (2017). Chem. Soc. Rev. 46: 7124–7144. (a) Li, N., Xu, J., Feng, R. et al. (2016). Chem. Commun. (Camb.) 52: 8501–8513. (b) Wang, C., Liu, X., Keser Demir, N. et al. (2016). Chem. Soc. Rev. 45: 5107–5134. Qiu, S., Xue, M., and Zhu, G. (2014). Chem. Soc. Rev. 43: 6116–6140. (a) Dechnik, J., Gascon, J., Doonan, C.J. et al. (2017). Angew. Chem. Int. Ed. 56: 2–21. (b) Erucar, I., Yilmaz, G., and Keskin, S. (2013). Chem. Asian J. 8: 1692–1704. Zhang, C., Zhang, K., Xu, L. et al. (2014). AlChE J. 60: 2625–2635. Koros, W.J., Ma, Y.H., and Shimidzu, T. (1996). Pure Appl. Chem. 68: 1479–1489. (a) Jia, Z. and Wu, G. (2016). Microporous Mesoporous Mater. 235: 151–159. (b) Cheng, X., Pan, F., Wang, M. et al. (2017). J. Membr. Sci. 541: 329–346. Tusel, G. and Brüschke, H. (1985). Desalination 53: 327–338. Shao, P. and Huang, R.Y.M. (2007). J. Membr. Sci. 287: 162–179. Gong, L.L., Zhang, L., Wang, N.X. et al. (2014). Sep. Purif. Technol. 122: 32–40. Mafia, A., Raisi, A., Hatam, M., and Aroujalian, A. (2012). J. Membr. Sci. 423: 175–188. Fujita, H. (1961). Diffusion in polymer-diluent systems. In: Fortschritte Der Hochpolymeren-Forschung, 1–47. Berlin, Heidelberg: Springer. (a) Guo, H., Zhu, G., Li, H. et al. (2006). Angew. Chem. Int. Ed. 45: 7053–7056. (b) Li, Y. and Yang, W. (2008). J. Membr. Sci. 316: 3–17. (c) Zhou, H., Li, Y., Zhu, G. et al. (2009). Sep. Purif. Technol. 65: 164–172.

References

21 Rangnekar, N., Mittal, N., Elyassi, B. et al. (2015). Chem. Soc. Rev. 44:

7128–7154. 22 Park, K.S., Ni, Z., Cote, A.P. et al. (2006). Proc. Natl. Acad. Sci. U. S. A. 103:

10186–10191. 23 Diestel, L., Bux, H., Wachsmuth, D., and Caro, J. (2012). Microporous Meso-

porous Mater. 164: 288–293. 24 (a) Bux, H., Feldhoff, A., Cravillon, J. et al. (2011). Chem. Mater. 23:

25 26 27 28 29 30

31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46

2262–2269. (b) Yao, J., Dong, D., Li, D. et al. (2011). Chem. Commun. (Camb.) 47: 2559–2561. (c) Fan, L., Xue, M., Kang, Z. et al. (2012). J. Mater. Chem. 22: 25272–25276. Huang, A., Bux, H., Steinbach, F., and Caro, J. (2010). Angew. Chem. Int. Ed. 49: 4958–4961. Huang, A., Dou, W., and Caro, J. (2010). J. Am. Chem. Soc. 132: 15562–15564. Fairen-Jimnez, D., Moggach, S.A., Wharmby, M.T. et al. (2011). J. Am. Chem. Soc. 133: 8900–8902. Gücüyener, C., Bergh, J.v.d., Gascon, J., and Kapteijn, F. (2010). J. Am. Chem. Soc. 132: 17704–17706. Dong, X. and Lin, Y.S. (2013). Chem. Commun. (Camb.) 49: 1196–1198. (a) Banerjee, R., Phan, A., Wang, B. et al. (2008). Science 319: 939–943. (b) Morris, W., Leung, B., Furukawa, H. et al. (2010). J. Am. Chem. Soc. 132: 11006–11008. Zhao, X.S., Jiang, J.W., and Nalaparaju, A. (2010). J. Phys. Chem. C 114: 11542. Kang, Z., Ding, J., Fan, L. et al. (2013). Inorg. Chem. Commun. 30: 74–78. Broughton, D.B., Neuzil, R.W., Pharis, J.M., and Brearley, C.S. (1970). Chem. Eng. Prog. 66: 70–75. Cottier, V., Bellat, J.P., Simonot-Grange, M.H., and Methivier, A. (1997). J. Phys. Chem. B 101: 4798–4802. Pham, T.C.T., Kim, H.S., and Yoon, K.B. (2011). Science 334: 1533–1538. Fan, L., Xue, M., Kang, Z. et al. (2014). Microporous Mesoporous Mater. 192: 29–34. Kasik, A. and Lin, Y.S. (2014). Sep. Purif. Technol. 121: 38–45. Ibrahim, A. and Lin, Y.S. (2016). Ind. Eng. Chem. Res. 55: 8652–8658. Banerjee, R., Furukawa, H., Britt, D. et al. (2009). J. Am. Chem. Soc. 131: 3875–3877. Kasik, A., James, J., and Lin, Y.S. (2016). Ind. Eng. Chem. Res. 55: 2831–2839. R. F. Service (2006). Science 311: 1088–1090. Shannon, M.A., Bohn, P.W., Elimelech, M. et al. (2008). Nature 452: 301–310. Marcovecchio, M.G., Mussati, S.F., Aguirre, P.A., and Scenna, N.J. (2005). Desalination 182: 111–122. Khawaji, A.D., Kutubkhanah, I.K., and Wie, J.-M. (2008). Desalination 221: 47–69. El-Saied, H., Basta, A.H., Barsoum, B.N., and Elberry, M.M. (2003). Desalination 159: 171–181. Cho, C.H., Oh, K.Y., Kim, S.K. et al. (2011). J. Membr. Sci. 371: 226–238.

407

408

11 MOF Membranes for Vapor or Liquid Separation

47 Zhu, Y., Gupta, K.M., Liu, Q. et al. (2016). Desalination 385: 75–82. 48 Cavka, J.H., Jakobsen, S., Olsbye, U. et al. (2008). J. Am. Chem. Soc. 130:

13850–13851. 49 Liu, X., Demir, N.K., Wu, Z., and Li, K. (2015). J. Am. Chem. Soc. 137:

6999–7002. 50 Liu, X., Wang, C., Wang, B., and Li, K. (2017). Adv. Funct. Mater. 27:

1604311. 51 Wu, F., Lin, L., Liu, H. et al. (2017). J. Membr. Sci. 544: 342–350. 52 Wu, H., Chua, Y.S., Krungleviciute, V. et al. (2013). J. Am. Chem. Soc. 135:

10525–10532. 53 Wang, X., Zhai, L., Wang, Y. et al. (2017). ACS Appl. Mater. Interfaces 9:

37848–37855. 54 Miyamoto, M., Hori, K., Goshima, T. et al. (2017). Eur. J. Inorg. Chem. 2017:

2094–2099. 55 Wan, L., Zhou, C., Xu, K. et al. (2017). Microporous Mesoporous Mater. 252:

207–213. 56 Liu, X.-L., Li, Y.-S., Zhu, G.-Q. et al. (2011). Angew. Chem. Int. Ed. 50:

10636–10639. 57 Liu, X., Jin, H., Li, Y. et al. (2013). J. Membr. Sci. 428: 498–506. 58 Liu, S., Liu, G., Zhao, X., and Jin, W. (2013). J. Membr. Sci. 446: 181–188. 59 Fan, H., Shi, Q., Yan, H. et al. (2014). Angew. Chem. Int. Ed. Engl. 53:

5578–5582. 60 Fan, H., Wang, N., Ji, S. et al. (2014). J. Mater. Chem. A 2: 20947–20957. 61 Han, G.L., Zhou, K., Lai, A.N. et al. (2014). J. Membr. Sci. 454: 36–43. 62 Li, Y., Wee, L.H., Martens, J.A., and Vankelecom, I.F.J. (2014). J. Mater.

Chem. A 2: 10034–10040. 63 Wee, L.H., Li, Y., Zhang, K. et al. (2015). Adv. Funct. Mater. 25: 516–525. 64 Naik, P.V., Wee, L.H., Meledina, M. et al. (2016). J. Mater. Chem. A 4:

12790–12798. 65 Yin, H., Khosravi, A., O’Connor, L. et al. (2017). Ind. Eng. Chem. Res. 56:

9167–9176. 66 Liu, G., Jiang, Z., Cao, K. et al. (2017). J. Membr. Sci. 523: 185–196. 67 Cheng, X., Jiang, Z., Cheng, X. et al. (2018). J. Membr. Sci. 545: 19–28. 68 Ma, D., Peh, S.B., Han, G., and Chen, S.B. (2017). ACS Appl. Mater. Inter-

faces 9: 7523–7534. 69 Xu, Y.M. and Chung, T.-S. (2017). J. Membr. Sci. 531: 16–26. 70 Wang, N., Zhang, G., Wang, L. et al. (2017). Sep. Purif. Technol. 186: 20–27. 71 (a) Liu, G., Jin, W., and Xu, N. (2016). Angew. Chem. Int. Ed. 55:

72 73 74 75 76

13384–13397. (b) Koros, W.J. and Zhang, C. (2017). Nat. Mater. 16: 289–297. Ying, Y., Liu, D., Zhang, W. et al. (2017). ACS Appl. Mater. Interfaces 9: 1710–1718. Zhang, H., James, J., Zhao, M. et al. (2017). J. Membr. Sci. 532: 1–8. Wang, S., Kang, Z., Xu, B. et al. (2017). Inorg. Chem. Commun. 82: 64–67. Kudasheva, A., Sorribas, S., Zornoza, B. et al. (2015). J. Chem. Technol. Biotechnol. 90: 669–677. Ding, C., Zhang, X., Li, C. et al. (2016). Sep. Purif. Technol. 166: 252–261.

References

77 Liu, S., Liu, G., Shen, J., and Jin, W. (2014). Sep. Purif. Technol. 133: 40–47. 78 Guan, K., Zhao, D., Zhang, M. et al. (2017). J. Membr. Sci. 542: 41–51. 79 Zhao, C., Wang, N., Wang, L. et al. (2014). Chem. Commun. (Camb.) 50: 80 81 82 83 84 85

86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101

102 103 104

13921–13923. Zhang, Y., Wang, N., Ji, S. et al. (2015). J. Membr. Sci. 489: 144–152. Yin, H., Lau, C.Y., Rozowski, M. et al. (2017). J. Membr. Sci. 529: 286–292. Ying, Y., Xiao, Y., Ma, J. et al. (2015). RSC Adv. 5: 28394–28400. Gao, R., Zhang, Q., Lv, R. et al. (2017). Chem. Eng. Res. Des. 117: 688–697. Sun, H., Tang, B., and Wu, P. (2017). ACS Appl. Mater. Interfaces 9: 21473–21484. (a) Hermans, S., Mariën, H., Van Goethem, C., and Vankelecom, I.F.J. (2015). Curr. Opin. Chem. Eng. 8: 45–54. (b) Shi, B., Marchetti, P., Peshev, D. et al. (2017). J. Membr. Sci. 525: 35–47. Basu, S., Maes, M., Cano-Odena, A. et al. (2009). J. Membr. Sci. 344: 190–198. Szekely, G., Jimenez-Solomon, M.F., Marchetti, P. et al. (2014). Green Chem. 16: 4440–4473. Sorribas, S., Gorgojo, P., Tellez, C. et al. (2013). J. Am. Chem. Soc. 135: 15201–15208. Campbell, J., Davies, R.P., Braddock, D.C., and Livingston, A.G. (2015). J. Mater. Chem. A 3: 9668–9674. Li, Y., Wee, L.H., Volodin, A. et al. (2015). Chem. Commun. (Camb.) 51: 918–920. Zhang, R., Ji, S., Wang, N. et al. (2014). Angew. Chem. Int. Ed. Engl. 53: 9775–9779. Wang, N., Li, X., Wang, L. et al. (2016). ACS Appl. Mater. Interfaces 8: 21979–21983. Zhu, J., Tian, M., Hou, J. et al. (2016). J. Mater. Chem. A 4: 1980–1990. Yang, H., Wang, N., Wang, L. et al. (2018). J. Membr. Sci. 545: 158–166. Ma, J., Guo, X., Ying, Y. et al. (2017). Chem. Eng. J. 313: 890–898. Wang, J., Wang, Y., Zhang, Y. et al. (2016). ACS Appl. Mater. Interfaces 8: 25508–25519. (a) Li, J.R., Kuppler, R.J., and Zhou, H.C. (2009). Chem. Soc. Rev. 38: 1477–1504. (b) Gin, D.L. and Noble, R.D. (2011). Science 332: 674–676. Zhu, J., Qin, L., Uliana, A. et al. (2017). ACS Appl. Mater. Interfaces 9: 1975–1986. Duan, J., Pan, Y., Pacheco, F. et al. (2015). J. Membr. Sci. 476: 303–310. Wang, L., Fang, M., Liu, J. et al. (2015). ACS Appl. Mater. Interfaces 7: 24082–24093. (a) Caliman, F.A. and Gavrilescu, M. (2009). Clean: Soil, Air, Water 37: 277–303. (b) Silva, C.P., Otero, M., and Esteves, V. (2012). Environ. Pollut. 165: 38–58. Ragab, D., Gomaa, H.G., Sabouni, R. et al. (2016). Chem. Eng. J. 300: 273–279. Basu, S. and Balakrishnan, M. (2017). Sep. Purif. Technol. 179: 118–125. Cheng, X., Jiang, X., Zhang, Y. et al. (2017). ACS Appl. Mater. Interfaces 9: 38877–38886.

409

410

11 MOF Membranes for Vapor or Liquid Separation

105 Echaide-Gorriz, C., Navarro, M., Tellez, C., and Coronas, J. (2017). Dalton 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120

121 122 123 124 125 126 127 128 129 130 131 132 133 134

Trans. 46: 6244–6252. Yang, L., Wang, Z., and Zhang, J. (2017). J. Membr. Sci. 532: 76–86. Ma, X.-H., Yang, Z., Yao, Z.-K. et al. (2017). J. Membr. Sci. 525: 269–276. Guo, X., Liu, D., Han, T. et al. (2017). AlChE J. 63: 1303–1312. Wang, C., Li, Z., Chen, J. et al. (2017). J. Membr. Sci. 523: 273–281. Zhu, L., Yu, H., Zhang, H. et al. (2015). RSC Adv. 5: 73068–73076. Xu, Y., Gao, X., Wang, Q. et al. (2016). RSC Adv. 6: 82669–82675. Wang, L., Fang, M., Liu, J. et al. (2015). RSC Adv. 5: 50942–50954. Van Goethem, C., Verbeke, R., Hermans, S. et al. (2016). J. Mater. Chem. A 4: 16368–16376. Li, Y., Wee, L.H., Martens, J.A., and Vankelecom, I.F.J. (2017). J. Membr. Sci. 523: 561–566. Wang, N., Liu, T., Shen, H. et al. (2016). AlChE J. 62: 538–546. Collins, A.N., Sheldrake, G.N., and Crosby, J. (1992). Chirality in Industry. New York: Wiley. Aboul-Enein, H.Y. and Wainer, I.W. (1997). The Impact of Stereochemistry on Drug Development and Use. New York: Wiley. Subramanian, G. (2001). Chiral Separation Techniques A Practical Approach. New York: Wiley-VCH. Gu, Z.G., Zhan, C., Zhang, J., and Bu, X. (2016). Chem. Soc. Rev. 45: 3122–3144. (a) Li, G., Yu, W.B., and Cui, Y. (2008). J. Am. Chem. Soc. 130: 4582–4583. (b) Xie, S.M., Zhang, Z.J., Wang, Z.Y., and Yuan, L.M. (2011). J. Am. Chem. Soc. 133: 11892–11895. Falcaro, P., Ricco, R., Doherty, C.M. et al. (2014). Chem. Soc. Rev. 43: 5513–5560. Liu, B., Shekhah, O., Arslan, H.K. et al. (2012). Angew. Chem. Int. Ed. Engl. 51: 807–810. Wang, W., Dong, X., Nan, J. et al. (2012). Chem. Commun. 48: 7022–7024. Gu, Z.G., Grosjean, S., Brase, S. et al. (2015). Chem. Commun. (Camb.) 51: 8998–9001. Gu, Z.G., Burck, J., Bihlmeier, A. et al. (2014). Chemistry 20: 9879–9882. Gu, Z.G., Fu, W.Q., Wu, X., and Zhang, J. (2016). Chem. Commun. (Camb.) 52: 772–775. Gu, Z.G., Fu, H., Neumann, T. et al. (2016). ACS Nano 10: 977–983. Chen, S.M., Liu, M., Gu, Z.G. et al. (2016). ACS Appl. Mater. Interfaces 8: 27332–27338. Gu, Z.G., Fu, W.Q., Liu, M., and Zhang, J. (2017). Chem. Commun. (Camb.) 53: 1470–1473. Huang, K., Dong, X., Ren, R., and Jin, W. (2013). AlChE J. 59: 4364–4372. Kang, Z., Xue, M., Fan, L. et al. (2013). Chem. Commun. 49: 10569–10571. Li, Q., Liu, G., Huang, K. et al. (2016). Asia Pac. J. Chem. Eng. 11: 60–69. Duan, H.-J., Yang, C.-X., and Yan, X.-P. (2015). RSC Adv. 5: 30577–30582. (a) Cychosz, K.A. and Matzger, A.J. (2010). Langmuir 26: 17198–17202. (b) Low, J.J., Benin, A.I., Jakubczak, P. et al. (2009). J. Am. Chem. Soc. 131:

References

135 136

137

138 139 140 141 142 143 144 145

15834–15842. (c) Liu, X., Li, Y., Ban, Y. et al. (2013). Chem. Commun. 49: 9140–9142. Ferey, G., Mellot-Draznieks, C., Serre, C. et al. (2005). Science 309: 2040–2042. (a) Feng, D., Gu, Z.-Y., Li, J.-R. et al. (2012). Angew. Chem. Int. Ed. 51: 10307–10310. (b) Colombo, V., Galli, S., Choi, H.J. et al. (2011). Chem. Sci. 2: 1311–1319. (a) Decoste, J.B., Peterson, G.W., Smith, M.W. et al. (2012). J. Am. Chem. Soc. 134: 1486–1489. (b) Zhang, W., Hu, Y., Ge, J. et al. (2014). J. Am. Chem. Soc. 136: 16978–16981. Liu, D., Ma, X., Xi, H., and Lin, Y.S. (2014). J. Membr. Sci. 451: 85–93. Zhang, H., Liu, D., Yao, Y. et al. (2015). J. Membr. Sci. 485: 103–111. Kang, Z., Wang, S., Fan, L. et al. (2017). Mater. Lett. 189: 82–85. Shih, Y.H., Kuo, Y.C., Lirio, S. et al. (2017). Chemistry 23: 42–46. Low, J.J., Benin, A.I., Jakubczak, P. et al. (2009). J. Am. Chem. Soc. 131: 15834–15842. Trickett, C.A., Gagnon, K.J., Lee, S. et al. (2015). Angew. Chem. Int. Ed. 54: 11162–11167. Deria, P., Mondloch, J.E., Karagiaridi, O. et al. (2014). Chem. Soc. Rev. 43: 5896–5912. Gupta, K.M., Qiao, Z., Zhang, K., and Jiang, J. (2016). ACS Appl. Mater. Interfaces 8: 13392–13399.

411

413

12 Microporous Organic Framework Materials for Membrane Separations 12.1 Introduction Membrane science and technology is recognized as a powerful tool for a variety of separations thanks to the cost effectiveness, energy efficiency, and environmental benefit [1]. Ideal separating membranes should be highly permeable and very selective toward species of interest. Despite the renewal of membrane technology via a physical manner (preparation route, membrane configuration, etc.), the development of new materials with excellent separating abilities has attracted increasing attention from academia. Polymeric membranes have been successfully commercialized since the 1980s. Unfortunately, the performances of polymers are limited by an undesirable trade-off relationship between permeability (P) and selectivity (S); namely, highly selective polymer membranes generally have low permeability and vice versa [2, 3]. These studies suggest that increasing the porosity of the membrane material is an alternative for achieving a better performance. Thus, this chapter will talk about the microporous organic framework materials for membrane separations. Microporous organic framework materials refer to hydrocarbons containing pores or voids at the microporous regime. Their backbones are composed of organic moieties connected by strong covalent bonds, usually resulted in ordered and rigid structures. The family members include polymers of intrinsic microporosity (PIMs), covalent organic frameworks (COFs), conjugated microporous polymers (CMPs), covalent triazine-based frameworks (CTFs), hypercrosslinked polymers (HCPs), porous cages (PCs), porous aromatic frameworks (PAFs), etc. [4–8]. This chapter is intended to report recent research on newly emerged microporous organic framework membranes with defined porous structures. The scope of gas and liquid separations is covered.

12.2 Porous Structures and Free Volumes Porous structures are inherent to natural processes. For instance, the honeycomb, the “magnum opus” of honeybees, is a porous structure. Sponges are aquatic animals that have pores on the surfaces of their bodies. The natural porous structures and their uses provided materials scientists with the idea of mimicking them Microporous Materials for Separation Membranes, First Edition. Xiaoqin Zou and Guangshan Zhu. © 2019 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2019 by Wiley-VCH Verlag GmbH & Co. KGaA.

414

12 Microporous Organic Framework Materials for Membrane Separations

1 2

(a) Pore free volume Interstitial free volume

V

Molecular volume

(b)

Occupied volume (0 K)

Tg

T

Figure 12.1 (a) Cross section of a porous solid illustrating open and closed pores (1 for closed pores and 2 for open pores) and (b) one interpretation of the contributions to the specific volume of a polymer and their temperature dependence.

in artificial structures, like microporous organic materials. Conforming to the recommendations of IUPAC, pores in microporous organic frameworks that have continuous connection pathways with the outer surfaces of the porous structure are called open pores. On the other hand, pores that are detached from other pores are called closed pores. Figure 12.1a illustrates the cross section of a porous medium. As defined, pores 1 are closed pores, while pores 2 are open pores. The open pores play a pivotal role in fluid dynamic and separation, and closed pores contribute to adsorption somehow. The IUPAC defines the attributes of a porous solid as follows: pore volume (V p ) is measured by a given method that must be stated [9]. The porosity is given by the ratio of the total pore volume V p to the apparent volume V of the particle or powder. The surface area of a porous solid is usually represented as the specific surface area that is given by the accessible area of solid surface per unit mass of the material. The pore size is described as the distance between two opposite walls of the pore. The pore size distribution can be represented by the derivatives dAp /drp or dV p /drp in function of rp , where Ap , V p , and rp are the wall area, volume, and radius of the pore. Porous organic frameworks have some attributes abovementioned that may be imperative to describe these materials.

12.2 Porous Structures and Free Volumes

The idea that molecular motion in liquids and solids requires molecule-sized pores was developed by Eyring [10]. These pores, which collectively make up the free volume, are constantly moving in the liquid state. Free volume is an intrinsic property of the organic polymer and is created by the gaps left between entangled polymer chains. Doolittle related the viscosity of a liquid to its relative free volume [11]. In the context of polymers, a pore or free volume element may be considered to be similar in size to a segment of a polymer molecule, and more than one may be required for mobility (i.e. motion in polymers involves cooperative movement of portions of a polymer chain). Fox and Flory [12] suggested that the glass transition, T g , represents an iso-free volume state occurring at a critical fractional free volume for large-scale movement of polymer segments. The fractional free volume, f V , in the glassy state may be calculated as 0.025 (i.e. 2.5% free volume) on the basis of the Williams–Landel–Ferry (WLF) equation [13], which describes the temperature dependence of viscosity and relaxation times for amorphous polymers and which may be derived on the basis of the Doolittle relationship. A substantially larger f V at T g may be inferred from Simha and Boyer’s relationship, (𝛼 R − 𝛼 G ) T g = 0.113, where 𝛼 R and 𝛼 G are the coefficients of thermal expansion in the rubbery and glassy states, respectively [14]. Other attempts to quantify free volume give further different values, depending on the assumptions made about what constitutes free volume. There are a number of contributions to a polymer’s specific volume, V (= 1/𝜌, where 𝜌 is polymer density), as indicated in Figure 12.1b. Firstly, there is the molecular volume, a useful estimate of which is the specific van der Waals volume, V w , which can be calculated by group contribution methods [15, 16]. When molecules are packed in a condensed phase, there is a limit to the packing density that can be achieved, so each molecule actually requires more space than its molecular volume. Typically, V w is multiplied by 1.3, based on the packing density of a molecular crystal at 0 K. Some workers regard this as the occupied volume, in which case the fractional free volume is given by the following equation: V − 1.3Vw V For a variety of glassy polymers, this approach gives values of f V in the range of 0.11–0.23 [17]. The free volume may then be subdivided into two types: interstitial free volume, which is spread among all the molecules, and pore free volume, which is localized but readily redistributed [18]. It is the pore free volume that is of relevance to the WLF equation. While some workers define the occupied volume as the volume of the liquid at equilibrium at 0 K [18], in which case it is independent of temperature, others include the effect of thermal vibrations in the occupied volume. The occupied volume increases approximately linearly with temperature throughout the glassy and rubbery regimes, while the free volume is approximately constant in the glassy state but increases rapidly with temperature in the rubbery state [19, 20]. Despite the lack of consensus as to the definition of free volume, it has proved a remarkably useful semiquantitative concept. The free volume concept has been widely accepted to describe sorption and diffusion in a porous polymer. Under the free volume model, the adsorption and diffusion of molecules in polymers depend greatly on the available free volume. For fV =

415

416

12 Microporous Organic Framework Materials for Membrane Separations

instance, many polymers show a sorption increase as the amount of free volume increases. In general, the fractional free volume of f V is used to express transport correlations within a family of porous polymers [21].

12.3 Hydrogen Recovery Hydrogen is widely accepted as a clean energy carrier and broadly applied in different sectors. H2 is industrially produced either from ammonia or hydrocarbon chemistry. However, H2 usually coexists with other light gases (N2 , CO2 , H2 O, CH4 , etc.) when it is produced from industrial processes. To use hydrogen-rich gas streams as fuels, gas separation is required. The enrichment of H2 from raw gases can be achieved by membrane gas separation. From the principle that connected pores offer highways for gas transports, many microporous organic materials with high surface areas have been created for the membrane application. The typical material is PIM-1, which was first discovered by Budd and McKeown in 2004 [22, 23]. This concept was further widely adopted in the syntheses of PIM-based membranes. Another archetypal PIM-TB was prepared using a greater shape-persistent unit such as ethanoanthracene (EA). Via the efficient formation of contorted bicyclic diamine (Tröger’s base), an extreme rigidity was generated (Figure 12.2a). Nitrogen sorption gave Brunauer–Emmett–Teller (BET) surface areas of 1028 m2 g−1 for PIM-EA-TB NH2 H2N

N

DMM TFA

N

PIM-EA-TB H2N

NH2

N N

DMM TFA

N N

PIM-SBI-TB

(c)

(a)

H2

Macromolecular chain

Inter-chain pore

(d)

CO2

CH4

N2

H2

(b)

Figure 12.2 (a) Synthesis and molecular structures of PIM-EA-TB and PIM-SBI-TB, (b) schematic presentation of generated pores by inefficient packed chains in both PIMs, (c) a solvent-cast film of PIM-EA-TB, and (d) pictorial illustration of H2 separation from gas mixtures using PIM membranes. Source: Reproduced with permission [24]. Copyright 2013, AAAS.

12.3 Hydrogen Recovery

and 745 m2 g−1 for PIM-SBI-TB, which ensured highly interconnected pore systems. The microporosities of both PIMs were arisen from their contorted moieties in the backbones that restricted conformational freedom and inhibited close packing of the macromolecular chains (Figure 12.2b). Due to good solubility in organic solvents, PIM-TB membranes were subsequently prepared by the solution-cast method (Figure 12.2c). Gas separation results revealed that PIM-TB membranes exhibited high gas permeability and selectivity for small gas molecules, such as H2 (Figure 12.2d). For PIM-EA-TB membrane with a thickness of 181 μm, the H2 permeability of 7760 Barrer was obtained. The fast gas transport was benefited from the connected pores. Further, high selectivities of 15.0 and 11.0 were obtained for gas pairs of H2 /N2 and H2 /CH4 , showing molecular sieving characteristics of PIM-EA-TB toward small gas molecules. This performance by virtue of permeability and selectivity for H2 exceeded the latest Robeson upper bound, which promised these membranes in hydrogen recovery from ammonia synthesis and petrochemical refinery streams [24]. Precise control on the pore width is an efficient strategy to realize a better kinetic separation performance. The kinetic separation by microporous membranes depends on the pore diameter. When the pore diameter becomes comparable to the molecular size, gas molecules will continuously feel the interaction with the walls. Due to this strong repulsion, the gas diffusivity is significantly reduced. For even smaller pores in the domain of less than 10 Å, configuration diffusion takes over, and the gas mobility starts to depend on pore dimension and gas size as well (Figure 12.3a) [25]. In this regard, Pinnau et al. engineered the pore sizes of PIM derivatives by using building blocks with specific geometry of the contortion site of approximately 120∘ kink [26]. Two AB-type ladder PIMs (TPIM-1 and TPIM-2) were prepared by polymerizing AB-monomer-containing precursors [27]. Pore size distributions measured by N2 adsorption demonstrated that the dominant pore sizes of TPIM-1 and TPIM-2 were 5.5 Å and 6.0 Å (Figure 12.3b), typical for ultramicroporosity ( O2 > CH4 > N2 , indicating the size-selective diffusion of gases through molecular-level windows in CC3 membrane. H2 separation ability was further assessed on CC3, ASPOC, and CC13 membranes. For a typical 50 nm thick CC3 membrane, the H2 /N2 selectivity was as high as 30, and the H2 permeability was about 226 Barrer. More works have been done to explore the applicability of microporous organic membranes in H2 separations [29–52]. According to Figure 12.5, many membranes (such as TPIM-1, KAUST-PI-1, PIM-Trip-TB, and PIM-EA-TB) exhibit excellent hydrogen separation performances with data points above the trade-off line, which can be explained by the good balance between small pores and high surface areas of these membranes.

12.4 Carbon Dioxide Capture An increase of CO2 level in the atmosphere poses an environmental burden on our modern society. Carbon capture and storage (CCS) is one of the initiatives proposed to mitigate anthropogenic CO2 emissions. Microporous membranes are considered to be a new option for CO2 capture owing to the high capacity of microporous materials in couple with the technological and economic advantages of membrane technology. Membrane materials with high CO2 permeability and easy membrane fabrication are crucial to large-scale applications. To meet this target, microporous organic membranes are evolving as the best candidate for efficient CO2 capture. In 2008, CO2 separation was attempted by using a microporous PIM-1 membrane. The gas permeation results showed CO2 permeability of 4350 Barrer and selectivity of 22.9 over N2 and 13.2 over CH4 for the as-cast membrane. The permeability for CO2 could be improved close to 104 Barrer with the careful manipulation of posttreatment conditions. This result confirmed that high porosity (apparent surface area of ∼860 m2 g−1 for PIM-1) is the prerequisite for high permeability, and the preferential CO2 permeation over other gases stemmed from the existence of a large number of micropores (6–8 Å) in PIM-1 [29]. Similar to the cases of H2 separations, CO2 separation properties of microporous organic membranes can be tailored by controlling membrane microstructures. The feasibility was exemplified by modified PIM membranes using external stimuli (e.g. heat) [28, 53]. Thermal post-synthesis was adopted to produce self-cross-linked PIM-1 membranes with tunable molecular sieving behaviors. Upon the thermal treatment, nitrile groups in the original PIM membrane underwent the cross-linking reaction via forming triazine rings between the chains. The formation of bulky planar triazine rings tended to pull the polymer chains closer to each other, resulting in contracted pores. As the

419

(a)

(b)

N O

H2

H2

N

N

+

O

N

–12 H2O H 2N

NH2

O

N

N

N

N

N CC3

Tortuous pathway

CC13

N N

N

Straight pathway

(c)

(d)

50 nm Single CC3 cage Face-to-face assembled structure

Alumina support

100 nm

Figure 12.4 (a) H2 diffusion in tortuous and straight channels of microporous materials, (b) synthetic routes and chemical structures of CC3 and CC13, (c) molecular simulations of single CC3 cage and assembled 3D crystalline structure, and (d) cross-sectional SEM of 50 nm thick CC3 thin film coated on Al2 O3 support. Source: Reproduced with permission [37]. Copyright 2016, John Wiley & Sons.

12.4 Carbon Dioxide Capture 103

TPIM-1 (d = 5.5, SBET = 862)

102

KAUST-Pl-1 (d = 5.5, SBET = 750) PIM-Trip-TB (SBET = 899) PIM-EA-TB (SBET = 1028)

101

100 101

(a)

102

103 104 105 H2 permeability (barrer)

106

Selectivity (H2/CH4)

Selectivity (H2/N2)

103

TPIM-1 (d = 5.5, SBET = 862)

102

KAUST-Pl-1 (d = 5.5, SBET = 750)

PIM-Trip-TB (SBET = 899) PIM-EA-TB (SBET = 1028)

101

100 101

102

(b)

104 105 103 H2 permeability (barrer)

106

Figure 12.5 H2 permeability and selectivity over N2 (a) and CH4 (b) in selected microporous organic membranes with defined porous structures. The data points in the square areas represent good balances between permeability and selectivity.

cross-linking reaction continued, extensive formations of triazine rings occurred in the polymer chains, leading to the rearrangement of porosity (Figure 12.6) [53]. The highly cross-linked PIM-1 exhibited an interesting CO2 separation property. For instance, PIM-300-2.0d membrane showed CO2 /N2 and CO2 /CH4 selectivities of 41.7 and 54.8, which were higher than the respective values of 20.7 and 14.8 for the original PIM-1. The significant enhancement in CO2 selectivity was contributed by contracted and narrow distributed pores. A small expense in CO2 permeability was observed for the PIM-300-2.0d membrane, thanks to the regained pore volumes after high-degree cross-linking. Another effective approach to enhance CO2 separation performance is to modify microporous organic membranes through chemical functionalization. In 2011, Guiver and his coworkers tried to attach pendant tetrazole groups onto PIM-1 skeleton (TZPIMs) to increase CO2 affinity [54]. In the process of post-modification, nitriles in the chain reacted with azides to yield tetrazoles by Enlarged pores Thermal

High degree of

cross-linking

cross-linking

Excess free volume Original PIM-1

Contracted pores Cross-linked PIM-1

Contracted pores Highly cross-linked PIM-1

Figure 12.6 Two-dimensional representations of the contorted PIM-1 membrane before and after thermal cross-linking reaction via the formation of triazine rings. Source: Reproduced with permission [53]. Copyright 2012, ACS.

421

422

12 Microporous Organic Framework Materials for Membrane Separations O O O

CN

ZnCl2

O

O

NaN3

O CN

PIM-1

N NH

CN O

O

O

O CN

X

N

N

N N H N

N

O

O

O

O CN

y

n

O O z

TZPIM

(a)

N2

N N H N N

TZPIM-1: x = 0, y = 1.0, z = 0 TZPIM-2: x = 0, y = 0.6, z = 0.4 TZPIM-3: x = 0, y = 0, z = 1.0

Diffusion transport

Assisted transport CO2

(b)

Figure 12.7 (a) Reaction scheme for the conversion of PIM-1 to TZPIM (the shadows indicate the conversion degrees) and (b) CO2 facilitated transport in functionalized pore walls of TZPIM membranes during gas permeation process. Source: Reproduced with permission [54]. Copyright 2011, Springer Nature.

the ZnCl2 -catalyzed click reaction (Figure 12.7a). Tetrazoles having both basic character and an acidic hydrogen served as adsorption sites for CO2 molecules. In addition, high-density tetrazoles on the pore walls acted as the carriers for fast CO2 transport through the pores (Figure 12.7b). Both TZPIM-1 and TZPIM-2 exhibited CO2 permeabilities of greater than 2000 Barrer, which was attributed to the facilitated surface diffusion of CO2 molecules along TZPIM channels. Meanwhile, CO2 /N2 selectivities for TZPIM membranes were enhanced in comparison with that of PIM-1. This enhancement was rationalized as the preferential selective adsorption of CO2 in TZPIM. The good performance in terms of CO2 /N2 selectivity and CO2 permeability soundly proved the applicability of TZPIM membranes in extracting CO2 from N2 . In the same manner, Dai et al. later on prepared a triazine-framework-based membrane with intrinsic porosity [55]. With functionalized triazine units, the membrane exhibited an increased selectivity for membrane separation of CO2 over N2 with an ideal CO2 /N2 selectivity of approximately 30, along with a CO2 permeability of approximately 520 Barrer. Both studies indicated that tailoring the pore surfaces with functionalities brought benefits in CO2 capture from post-combustion gases. The practices of introducing CO2 -philic groups

12.5 Air Separation

(e.g. —OH, —COOH, —NH2 ) onto the pore walls of microporous organic membranes were extensively executed by our researchers [56–64]. More recently, our group proposed a new strategy of simultaneously modulating pore surface property and controlling the pore size in a narrow region for improving CO2 separation. In this respect, an N-rich Schiff-based material (SNW-1) was prepared by the condensation between two monomers of melamine and terephthalaldehyde (Figure 12.8a) [65]. The pore system of SNW-1 was identified by the N2 sorption measurement. The sorption data revealed that SNW-1 possessed a 3D network with major pores of 5 Å (Figure 12.8a). Further, the affinity of SNW-1 for CO2 was quantified by gas sorption. A huge uptake for CO2 (50 cm3 g−1 at 25 ∘ C and P/P0 = 1.0) was measured, which is contrast to the small uptakes of 10 and 5 cm3 g−1 for CH4 and N2 . Due to the high dispersion of SNW-1 colloids, SNW-1/PSF membranes were facilely produced by spin-coating SNW-1/PSF suspensions onto porous supports (Figure 12.8b). A representative membrane with 12 wt% SNW-1 afforded CO2 permeability of 22.4 Barrer and selectivities of 34 and 40 for CO2 /CH4 and CO2 /N2 , respectively (Figure 12.8c and d). The selectivities were higher than the corresponding ideal ones (27 and 29 for CO2 /CH4 and CO2 /N2 ), which was interpreted by the synergic effects of preferential adsorption and favorable diffusion for CO2 molecules in functionalized small pores. The robustness and reproducibility rendered this membrane suitable for natural gas sweetening and carbon dioxide capture from post-combustion gases. The method of pore modulation to improve CO2 capture efficiency was extended to other microporous organic membranes including crystalline two-dimensional (2D) and amorphous 3D ones, the work of which is summarized in Figure 12.9 [24, 53–97]. Based on the survey, it can be concluded that introduced surface functionality and small pore contribute significantly to increase the selectivity and fully interconnected pores (equivalently by high surface areas) bring the huge enhancement in gas permeability.

12.5 Air Separation The separation of air is the basic commercial activity for producing industrial gases, such as oxygen and nitrogen, and some rare gases, which are widely used in chemical industry. The majority of the activities are being done in enriching oxygen or nitrogen from atmospheric air, because high-purity oxygen is needed for coal gasification, or enriched nitrogen is required for inerting storage tanks. Air separation by polymeric membranes has already been conducted to generate oxygen or enrich nitrogen. However, the use of polymeric membranes is a small portion of the market of air separation because of low permeability and moderate selectivity. For improving membrane performance, new materials particularly for microporous organic polymers are being explored for the target of air separation. As well known, oxygen and nitrogen have similar physical properties (such as the boiling points of 90.2 K for O2 and 77.3 K for N2 ), which lead to very close

423

N N

N HN

NH

N HN

N HN

N

N N

N HN

NH

N

400 350 300 250 200 150 100 50 0 0.0

5Å

5Å 12 mm 1.0 1.5 0.5 Pore width (nm)

2.0

(b)

(a) 30

20 25 15 20 10

40 Separation factor

Separation factor

25

CO2 permeability (barrer)

35 30

2

4

6 8 10 SNW-1 wt%

12

14

20

30 25

15

20 10

5 0

25

35

15

15

(c)

30

45

5

10

16

0

(d)

CO2 permeability (barrer)

3

–1

Differential pore volume (cm g )

N N

2

4

6 8 10 SNW-1 wt%

12

14

16

Figure 12.8 (a) Schematic representation of chemical structure and molecular model of a fragment of SNW-1 showing its porous network and the pore size distribution derived from N2 adsorption, (b) optical photo of cast SNW-1/PSF membrane, and separation performances of SNW-1/PSF membranes with different SNW-1 contents for (c) CO2 /CH4 and (d) CO2 /N2 gas mixtures. Source: Reproduced with permission [65]. Copyright 2014, John Wiley & Sons.

12.5 Air Separation

102

103

PAF-56P (triazine) SNW-1 (amino) TPDA-DAR (imide) AO-PIM-1 (amidoxime) PIM-1 (SBET = 860) PIM-PI-EA (SBET = 616)

101

TBPIM-25 (SBET = 760) PIM-EA-TB (SBET = 1028)

Selectivity (CO2/CH4)

Selectivity (CO2/N2)

103

102

SNW-1 (amino)

TPDA-DAR (imide) AO-PIM-1 (amidoxime)

6FDA-HSBF (hydroxyl)

PIM-PI-EA (SBET = 616)

101 PIM-EA-TB (SBET = 1028) PAF-1-Li6C60 (SBET = 7360)

100 100

(a)

101 102 103 104 105 CO2 permeability (barrer)

100 100

106

(b)

101 102 103 104 105 CO2 permeability (barrer)

106

Figure 12.9 CO2 permeability and selectivity over N2 (a) and CH4 (b) in selected microporous organic membranes with defined porous structures.

adsorption coefficients of O2 and N2 in most porous polymers. From this aspect, it is very difficult to enrich O2 or N2 from the sorption basics. According to Equation 𝛼 ij = Pi /Pj = (Si /Sj ) × (Di /Dj ) Eq. (4), a high O2 selectivity is possible to be obtained on the diffusivity discrepancies due to different kinetic diameters of the two molecules (3.46 Å and 3.64 Å for O2 and N2 ). For achieving a high O2 selectivity, the group of Guiver tried to modify PIM-1 with trifluoromethylphenylsulfone groups to increase chain stiffness and reduce interchain interaction (Figure 12.10) [98]. The TFMPSPIM membranes yielded O2 /N2 selectivities of 3.4–4.7, along with O2 permeabilities in the range of 156–737 Barrers. The selectivity of TFMPSPIM-1 membrane (4.7) for O2 over N2 was higher than that of the PIM-1 counterpart (3.2), attributed to the shorter interchain distance in TFMPSPIM-1. Molecular modeling showed that the width and the volume of TFMPSPIM-1 pores were decreased because the pendant phenylsulfonyl groups resided between the polymer chains (Figure 12.10). This study along with other reports clearly verified that narrowing the pores could effectively increase the diffusion selectivity of O2 over N2 and consequently increased the overall separation factor [99]. The industrial criteria for membrane process to enrich oxygen from air recommend O2 /N2 selectivity above 2.5 and concurrently O2 permeability greater than 500 Barrer [100]. To satisfy both needs, the groups of McKeown and Pinnau parallel devoted to increase O2 permeability and minimize the sacrifice in O2 /N2 selectivity by virtue of enlarging the porosity and retaining the microporous structure [101–107]. The typical membranes are derived from triptycene-based microporous polymers. McKeown et al. developed a benzotriptycene-based PIM of PIM-BTrip-TB (Figure 12.11), which was prepared from diaminobenzotriptycene monomer [101]. PIM-BTrip-TB demonstrated an apparent BET surface area of 870 m2 g−1 , confirming its highly porous structure. Gas permeation measurement on the cast membrane showed that PIM-BTrip-TB had an O2

425

426

12 Microporous Organic Framework Materials for Membrane Separations O

O

F

CN O

O

F

O

O

O

O CN

PIM-1

F

O S O n

n

TFMPSPIM1

Figure 12.10 Macromolecular structures and fragment models of PIM-1 and TFMPSPIM-1 (phenylsulfonyl groups are shown in shadows). Source: Reproduced with permission [98]. Copyright 2008, ACS.

N N PIM-BTrip-TB

Figure 12.11 Chemical structure and molecular fragment of PIM-BTrip-TB contorted chain showing its porous network. Source: Drawn by Materials Studio.

permeability of 3290 Barrer and an O2 /N2 selectivity of 3.6. The high permeability was ascribed to the large surface area that ensured fully interconnected pores. The rigid triptycene moiety protected PIM-BTrip-TB membrane from physical aging, evidenced by a remarkably high value of 5.4 for O2 /N2 selectivity and an acceptable O2 permeability of 1170 Barrer. To understand why these PIMs displayed unprecedented O2 /N2 separation performances, the microstructures of four PIMs (TPIM-1, KAUST-PI-1,

12.6 Other Gas Separations

Rigid, contorted macromolecules Ethanoanthracene

Triptycene

TPIM-1, PIM-Trip-TB PIM-EA-TB, KAUST-PI-1

Fast molecular sieving ultramicroporosity

Tröger’s base

Figure 12.12 Schematic illustration of three bridged building blocks for forming PIM macromolecules with ultramicroporosity to sieve gas molecules. Source: Reproduced with permission [108]. Copyright 2015, ACS.

PIM-Trip-TB, and PIM-EA-TB) constructed from similar bridged-bicyclic building blocks were examined in more detail (Figure 12.12) [108]. From sorption measurements and permeation tests, they found that double-end open pores in these PIMs originated from inefficient packing of rigid and contorted chains readily permitted fast gas diffusion and thus higher permeability. Moreover, the four PIMs featured a subtle balance between larger pores and smaller ultramicropores, which were responsible for selective O2 permeation. To date, extensive studies have been implemented to create high porosities in microporous polymers in order to attain highly permeable membranes (e.g. polyimides) for separating O2 from N2 (Figure 12.13) [24, 98–118]. The designed syntheses of O2 -selective microporous organic membranes via structure–property optimizations will be the next focus.

12.6 Other Gas Separations The separations described above represent the most studied systems using microporous organic membranes. In addition, there are a number of investigations on membrane separations for other industrially relevant gases, such as the recovery of light hydrocarbons [119, 120]. Olefin/paraffin separation is another important branch of chemical separations. Pinnau et al. exploited PIM-PI-1 and KAUST-PI-1 membranes for C3 H6 /C3 H8 separations [121]. Single gas permeation experiments showed that C3 H6 permeabilities for PIM-PI-1 and KAUST-PI-1 were 393 and 817 Barrer, respectively. The significant increase in permeability from PIM-PI-1 to KAUST-PI-1 was due to the larger microporosity of KAUST-PI-1. Later, the same group modified PIM-6FDA-OH membrane by thermal annealing at 250 ∘ C. The annealed membrane produced a very high

427

12 Microporous Organic Framework Materials for Membrane Separations

Selectivity (O2/N2)

428

101 PIM-EA-TB (SBET = 1028) PIM-Trip-TB (SBET = 899) PIM-BTrip-TB (SBET = 870)

PIM-SBF (SBET = 803)

100

100

101

102

103

104

105

106

O2 permeability (barrer)

Figure 12.13 O2 permeability and O2 /N2 selectivity in selected microporous organic membranes with defined porous structures. Representatives of highly permeable membranes are marked in the square area.

pure gas C3 H6 /C3 H8 selectivity of 30 [122]. The high selectivity resulted from enhanced diffusion selectivity due to the formation of interchain charge transfer complexes. Similarly, Chung et al. modified the PIM-1 membrane with metal ions by first hydrolyzation and subsequent incorporation of Zn2+ , Mg2+ and Ag+ into PIM-1 pores. Due to the π-complexation between metal ions and C3 H6 molecules, improved C3 H6 permeabilities of 568–837 Barrers and enhanced C3 H6 /C3 H8 selectivities of 9.12–15.23 were obtained for metal-functionalized PIM-1 membranes [123]. Moreover, metal treatment helped stabilize the membranes against plasticization. This is probably because metal carboxylation induced the cross-linking in the structure, which inhibited chain mobility even at a relatively high feed pressure (4 bars). Li and Yang et al. predicted that 2D covalent-triazine-framework membranes (CTF-0) with uniform and ultra-small pores would work effectively for helium separation. From first principles calculations, it was found that a monolayer CTF-0 membrane could exhibit unprecedentedly high helium selectivity over other permanent gases [124].

12.7 Emerging Liquid Separations In the chemical or environmental industries, nanofiltration (NF) membranes are widely used for applications such as wastewater treatment and recovery of valuable organic molecules. NF membranes have pore sizes from 1 to 10 nm, and

12.7 Emerging Liquid Separations

they are predominantly created from conventional polymer thin films, which are composed of amorphous polymers. Dense structure, poor porosity, and pore connectivity of these amorphous polymers also hinder the permeation of solvents. COFs are highly porous and crystalline polymers with uniformly arranged ordered pore channels (sizes of 0.5–10 nm). The definition tells us that COFs bear the combined merits of inorganic crystallinity and polymeric network; thus these materials are emerging as candidates for membrane materials. Mainly two approaches are proposed for the fabrication of COF membranes: one is direct synthesis like inorganic membranes, and another is interfacial polymerization like polymer membranes. The group of Caro developed an in situ solvothermal approach for the synthesis of continuous 2D imine-linked COF-LZU1 membrane with a thickness of only 400 nm. As shown in Figure 12.14a, Al2 O3 tube was first modified 3-aminopropyltriethoxysilane (APTES). After this modification, the amino-Al2 O3 tube surface was further functionalized with aldehyde groups by reaction with 1,3,5-triformylbenzene (TFB) at 150 ∘ C for one hour. The successful preparation of a well-intergrown COF-LZU1 layer on the alumina substrate was accomplished after sufficient optimization of the preparation conditions, including the concentrations of aldehyde and amine, the synthesis time and temperature, and the type and amount of solvent (Figure 12.14b). The potentiality of COF-LZU1 in NF membrane was evaluated by rejecting aqueous dyes (chrome black T, methyl blue, acid fuchsin, Congo red, Rose Bengal). The O

NH2

NH2 NH2

O

COF-LZU1 O

O N

NH2

NH2 NH2

NH2 NH2

O

O

O

O N

O

Dioxane, 150 °C, 1 h

Amino-Al2O3 tube

O

O N

O

H2N

NH2

COF layer

Dioxane, 3M acetic acid, 120 °C, 72 h

Aldehyde-Al2O3 tube

Tubular COF-LZU1 membrane

(a)

COF-LZU1 layer

500 nm

1 μm

(b)

Figure 12.14 (a) Diagram for the synthesis of tubular COF-LZU1 membranes and (b) top (left) and side (right) SEM images of the prepared COF-LZU1 membrane supported on Al2 O3 tube. Source: Reproduced with permission [125]. Copyright 2018, John Wiley & Sons.

429

430

12 Microporous Organic Framework Materials for Membrane Separations

+

Br

Br

Br

1. Surface-initiated polymerization 2. PMMA protection Si/SiO2

1. Membrane transfer 2. PMMA removal PMMA layer CMP layer

CMP on PAN substrate

Figure 12.15 Surface-initiated polymerization of the CMP membrane on a bromobenzene-functionalized Si/SiO2 substrate. C, gray; H, white; Br, red. The bottom panel shows a schematic representation of the surface-initiated polymerization; the inset in the red circle (right) shows the all-rigid skeleton of the resulting CMP membrane [126]. Source: Copyright 2018, Springer Nature.

membrane showed water permeance of cal. 760 l m−2 h−1 MPa−1 and favorable rejection rates exceeding 90% for water-soluble dyes larger than 1.2 nm [125]. Shortly after this work, Li and Tang extended this strategy in constructing membranes of CMPs. In their work, the CMP membranes (p-CMP, m-CMP, o-CMP) were first fabricated on bromobenzene-functionalized silica wafers via surface-initiated Sonogashira–Hagihara polymerization of 1,3,5-triethynylbenzene (1,3,5-TEB) with three different dibromobenzenes (1,4-dibromobenzene, 1,3-dibromobenzene, 1,2-dibromobenzene). Afterward, the as-prepared CMP membranes were easily transferred to the top of porous supports to form a composite membrane (Figure 12.15). Due to the nature of the surface-initiated polymerization, the membrane formation occurred only on the support surface, and consequently an ultrathin membrane in 42 nm thickness was yielded. Further, CMP membranes supported on polyacrylonitrile substrates provided excellent retention of solutes and broad-spectrum NF in both nonpolar hexane and polar methanol, the permeance for which reached 32 and 22 l m−2 h−1 bar−1 , respectively [126]. In 2017, Banerjee and coworkers developed a very simple approach for fabricating COF membranes. Their approach includes the formation of an aqueous organic salt containing powdered aromatic diamine and a co-reagent (ptoluenesulfonic acid [PTSA]), followed by adding 1,3,5-triformylphloroglucinol (Tp) into the above mixture to form a viscous dough via thoroughly shaking. The precursor was cast on a clean glass plate, and the reaction between monomers of Tp and diamine was initiated through solid polymerization (Figure 12.16a). The resultant self-standing membranes display high porosity and crystallinity, and they are flexible, continuous, and devoid of any internal defects or cracks (Figure 12.16b). The authors utilized these COF membranes for evaluating challenging separation applications such as wastewater treatment and recovery of valuable active pharmaceutical ingredients from organic solvents as the representative cases. As a significant outcome, these membranes showed high fluxes toward organic solvents such as acetone and acetonitrile. The permeance of acetonitrile (278 l m−2 h−1 bar−1 ) of M-TpTD is 2.5 times magnitude higher than the existing polyamide-nanofilm-based NF membranes (112 l m−2 h−1 bar−1 ) with equivalent solute-rejection performances [127].

12.8 Conclusions

NH2 SO3H

HO

+

CHO OH CHO

OHC OH

PTSA

NH2

BD

tex Veraker sh

(a)

M-TpTD

(b)

Figure 12.16 (a) Schematic representation of COF membrane fabrication and (b) photographs of M-TpTD demonstrating the membrane flexibility [127]. Source: Copyright 2017, Wiley-VCH.

Later on, the group of Lai developed a modified technique called the Langmuir–Blodgett (LB) method to fabricate 2D COF membranes, which was exemplified by a crystalline TpDHF 2D COF membrane constructed from two precursors of Tp and 9,9-dihexylfluorene-2,7-diamine (DHF). A single COF layer was precisely four unit cells thick and can be transferred to different support surfaces layer by layer. The Tp-DHF 2D COF membrane supported on an AAO porous support displayed remarkable permeabilities for both polar and nonpolar organic solvents, which were approximately 100 times higher than that of the amorphous membranes prepared by the same procedure [128]. The membrane exhibited a steep molecular sieving with a molecular weight retention onset (MWRO) of approximately 600 Da and a molecular weight cutoff (MWCO) of approximately 900 Da. Pioneered by these works, increasing attempts have been made to fabricate COF membranes with NF capabilities [129, 130].

12.8 Conclusions In summary, diverse microporous organic frameworks (PIMs, COFs, CMPs, CTFs, HCPs, PCs, and PAFs) were advanced for membrane separation focusing on gas separations of hydrogen, carbon dioxide, air, and other relevant gases,

431

432

12 Microporous Organic Framework Materials for Membrane Separations

as well as water treatment. The following conclusions were made according to pore chemistry perspective: highly interconnected pores provide mass transport channels through the membranes, and suitable pores and surface functionalities contribute to the impressive selectivities for one over another species. Despite great achievements made up to date, many challenges still remain in the next decade: how to create POF materials with precise functions, how to make large-scale POF membranes, how to enhance membrane thermal stability and chemical resistance, and how to globalize a standard procedure for measuring membrane separation capacity.

References 1 Bernardo, P., Drioli, E., and Golemme, G. (2009). Ind. Eng. Chem. Res. 48: 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

4638–4663. Robeson, L.M. (2008). J. Membr. Sci. 320: 390–400. Freeman, B.D. (1999). Macromolecules 32: 375–380. Feng, X., Ding, X.S., and Jiang, D.L. (2012). Chem. Soc. Rev. 41: 6010–6022. Ding, S.Y. and Wang, W. (2013). Chem. Soc. Rev. 42: 548–568. Cooper, A.I. (2009). Adv. Mater. 21: 1291–1295. Thomas, A. (2010). Angew. Chem. Int. Ed. 49: 8328–8344. Zou, X.Q., Ren, H., and Zhu, G.S. (2013). Chem. Commun. 49: 3925–3936. Rouquerol, J., Avnir, D., Fairbridge, C.W. et al. (1994). Pure Appl. Chem. 66: 1739–1758. Eyring, H. (1936). J. Chem. Phys. 4: 283–291. Doolittle, A.K. (1951). J. Appl. Phys. 22: 1471–1475. Fox, T.G. and Flory, P.J. (1950). J. Appl. Phys. 21: 581–591. Williams, M.L., Landel, R.F., and Ferry, J.D. (1955). J. Am. Chem. Soc. 77: 3701–3707. Simha, R. and Boyer, R.F. (1962). J. Chem. Phys. 37: 1003–1007. Bondi, A. (1968). Physical Properties of Molecular Crystals, Liquids and Glasses. New York: Wiley. van Krevelen, D.W. (1997). Properties of Polymers. Amsterdam: Elsevier. Thran, A., Kroll, G., and Faupel, F. (1999). J. Polym. Sci. B Polym. Phys. 37: 3344–3358. Vrentas, J.S. and Duda, J.L. (1986). Encycl. Polym. Sci. Eng. 5: 36–68. Sperling, L.H. (1992). Introduction to Physical Polymer Science. New York: Wiley. Gedde, U.W. (1995). Polymer Physics. London: Chapman & Hall. Budd, P.M., McKeown, N.B., and Fritsch, D. (2005). J. Mater. Chem. 15: 1977–1986. Budd, P.M., Ghanem, B.S., Makhseed, S. et al. (2004). Chem. Commun. 230–231. Budd, P.M., Elabas, E.S., Ghanem, B.S. et al. (2004). Adv. Mater. 16: 456–459. Carta, M., Malpass-Evans, R., Croad, M. et al. (2013). Science 339: 303–307. Post, M.F.M. (1991). Diffusion in zeolite molecular sieves. In: Introduction to Zeolite Science and Practice, Vol. 58 of Studies in Surface Science and

References

26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

Catalysis (ed. H. van Bekkum, E.M. Flanigen and J.C. Jansen), 391–443. Amsterdam: Elsevier. Swaidan, R., Al-Saeedi, M., Ghanem, B. et al. (2014). Macromolecules 47: 5104–5114. Ghanem, B.S., Swaidan, R., Ma, X. et al. (2014). Adv. Mater. 26: 6696–6700. Song, Q.L., Cao, S., Pritchard, R.H. et al. (2014). Nat. Commun. 5: 4813. Budd, P.M., McKeown, N.B., Ghanem, B.S. et al. (2008). J. Membr. Sci. 325: 851–860. Lu, H., Wang, C., Chen, J.J. et al. (2015). Chem. Commun. 51: 15562–15565. Ying, Y.P., Liu, D.H., Ma, J. et al. (2016). J. Mater. Chem. A 4: 13444–13449. Fu, J.R., Das, S., Xing, G.L. et al. (2016). J. Am. Chem. Soc. 138: 7673–7680. Li, F.Y., Xiao, Y.C., Ong, Y.K., and Chung, T.S. (2012). Adv. Energy Mater. 2: 1456–1466. Zhuang, Y.B., Seong, J.G., Do, Y.S. et al. (2014). Macromolecules 47: 3254–3262. Lee, M., Bezzu, C.G., Carta, M. et al. (2016). Macromolecules 49: 4147–4154. Lau, C.H., Konstas, K., Thornton, A.W. et al. (2015). Angew. Chem. Int. Ed. 54: 2669–2673. Song, Q.L., Jiang, S., Hasell, T. et al. (2016). Adv. Mater. 28: 2629–2637. Jimenez-Solomon, M.F., Song, Q.L., Jelfs, K.E. et al. (2016). Nat. Mater. 15: 760–767. Ma, X.H., Mukaddam, M., and Pinnau, I. (2016). Macromol. Rapid Commun. 37: 900–904. Ghanem, B.S., McKeown, N.B., Budd, P.M. et al. (2009). Macromolecules 42: 7881–7888. Ghanem, B.S., McKeown, N.B., Budd, P.M., and Fritsch, D. (2008). Macromolecules 41: 1640–1646. Evans, J.D., Huang, D.M., Hill, M.R. et al. (2014). J. Phys. Chem. C 118: 1523–1529. Lindemann, P., Tsotsalas, M., Shishatskiy, S. et al. (2014). Chem. Mater. 26: 7189–7193. Yong, W.F., Li, F.Y., Chung, T.S., and Tong, Y.W. (2013). J. Mater. Chem. A 1: 13914–13925. Zhou, J.H., Zhu, X., Hu, J. et al. (2014). Phys. Chem. Chem. Phys. 16: 6075–6083. Kim, D., Tzeng, P., Barnett, K.J. et al. (2014). Adv. Mater. 26: 746–751. Tian, Z.Z., Wang, S.F., Wang, Y.T. et al. (2016). J. Membr. Sci. 514: 15–24. Wang, H. and Cao, D. (2015). J. Phys. Chem. C 119: 6324–6330. Keskin, S. (2012). J. Phys. Chem. C 116: 1772–1779. Ma, X.H. and Pinnau, I. (2016). Polym. Chem. 7: 1244–1248. Luo, S.J., Wiegand, J.R., Gao, P.Y. et al. (2016). J. Membr. Sci. 518: 100–109. Smith, S.J.D., Lau, C.H., Mardel, J.I. et al. (2016). J. Mater. Chem. A 4: 10627–10634. Li, F.Y., Xiao, Y.C., Chung, T.S., and Kawi, S. (2012). Macromolecules 45: 1427–1437. Du, N.Y., Park, H.B., Robertson, G.P. et al. (2011). Nat. Mater. 10: 372–375.

433

434

12 Microporous Organic Framework Materials for Membrane Separations

55 Zhu, X., Tian, C.C., Mahurin, S.M. et al. (2012). J. Am. Chem. Soc. 134:

10478–10484. 56 Alaslai, N., Ghanem, B., Alghunaimi, F., and Pinnau, I. (2016). Polymer 91:

128–135. 57 Swaidan, R., Ghanem, B.S., Litwiller, E., and Pinnau, I. (2014). J. Membr. Sci.

457: 95–102. 58 Kang, Z.X., Peng, Y.W., Qian, Y.H. et al. (2016). Chem. Mater. 28: 1277–1285. 59 Meng, L.B., Zou, X.Q., Guo, S.K. et al. (2015). ACS Appl. Mater. Interfaces 7:

15561–15569. 60 Sekizkardes, A.K., Kusuma, V.A., Dahe, G. et al. (2016). Chem. Commun. 52:

11768–11771. 61 Biswal, B.P., Chaudhari, H.D., Banerjee, R., and Kharul, U.K. (2016). Chem.

Eur. J. 22: 4695–4699. 62 Du, N.Y., Robertson, G.P., Dal-Cin, M.M. et al. (2012). Polymer 53:

4367–4372. 63 Ma, X.H., Salinas, O., Litwiller, E., and Pinnau, I. (2014). Polym. Chem. 5:

6914–6922. 64 Ma, X.H., Swaidan, R., Belmabkhout, Y. et al. (2012). Macromolecules 45:

3841–3849. 65 Gao, X., Zou, X.Q., Ma, H.P. et al. (2014). Adv. Mater. 26: 3644–3648. 66 Wang, Z.G., Liu, X., Wang, D., and Jin, J. (2014). Polym. Chem. 5:

2793–2800. 67 Qiao, Z.A., Chai, S.H., Nelson, K. et al. (2014). Nat. Commun. 5: 3705. 68 Regno, A.D., Gonciaruk, A., Leay, L. et al. (2013). Ind. Eng. Chem. Res. 52:

16939–16950. 69 Chernova, E., Petukhov, D., Boytsova, O. et al. (2016). Sci. Rep. 6: 31183. 70 Shamsipur, H., Dawood, B.A., Budd, P.M. et al. (2014). Macromolecules 47:

5595–5606. 71 Ma, X.H., Salinas, O., Litwiller, E., and Pinnau, I. (2013). Macromolecules 46:

9618–9624. 72 Du, N.Y., Dal-Cin, M.M., Robertson, G.P., and Guiver, M.D. (2012). Macro-

molecules 45: 5134–5139. 73 Jeon, E., Moon, S.Y., Bae, J.S., and Park, J.W. (2016). Angew. Chem. Int. Ed.

55: 1318–1323. 74 Angelis, M.G.D., Gaddoni, R., and Sarti, G.C. (2013). Ind. Eng. Chem. Res.

52: 10506–10520. 75 Yong, W.F., Lee, Z.K., Chung, T.S. et al. (2016). ChemSusChem 9: 1953–1962. 76 Mitra, T., Bhavsar, R.S., Adams, D.J. et al. (2016). Chem. Commun. 52:

5581–5584. 77 Lau, C.H., Konstas, K., Doherty, C.M. et al. (2015). Chem. Mater. 27:

4756–4762. 78 Park, H.B., Jung, C.H., Lee, Y.M. et al. (2007). Science 318: 254–258. 79 Wu, X.M., Zhang, Q.G., Lin, P.J. et al. (2015). J. Membr. Sci. 493: 147–155. 80 Yong, W.F., Li, F.Y., Chung, T.S., and Tong, Y.W. (2014). J. Membr. Sci. 462:

119–130. 81 Hao, L., Li, P., and Chung, T.S. (2014). J. Membr. Sci. 453: 614–623.

References

82 Tian, Z.Q., Dai, S., and Jiang, D.E. (2015). ACS Appl. Mater. Interfaces 7:

13073–13079. 83 Volkov, A.V., Bakhtin, D.S., Kulikov, L.A. et al. (2016). J. Membr. Sci. 517:

80–90. 84 Zhao, H.Y., Xie, Q., Ding, X.L. et al. (2016). J. Membr. Sci. 514: 305–312. 85 Song, Q.L., Cao, S., Pritchard, R.H. et al. (2016). J. Mater. Chem. A 4:

270–279. 86 Fritsch, D., Bengtson, G., Carta, M., and McKeown, N.B. (2011). Macromol.

Chem. Phys. 212: 1137–1146. 87 Cao, X.C., Qiao, Z.H., Wang, Z. et al. (2016). Int. J. Hydrog. Energy 41:

9167–9174. 88 Shan, M.X., Seoane, B., Rozhko, E. et al. (2016). Chem. Eur. J. 22:

14467–14470. 89 Tong, M.M., Yang, Q.Y., Ma, Q.T. et al. (2016). J. Mater. Chem. A 4: 124–131. 90 Lasseuguette, E., Carta, M., Brandani, S., and Ferrari, M.C. (2016). Int. J.

Greenhouse Gas Control 50: 93–99. 91 Lau, C.H., Mulet, X., Konstas, K. et al. (2016). Angew. Chem. Int. Ed. 55:

1998–2001. 92 Rogan, Y., Malpass-Evans, R., Carta, M. et al. (2014). J. Mater. Chem. A 2: 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108

4874–4877. Alghunaimi, F., Ghanem, B., Alaslai, N. et al. (2015). J. Membr. Sci. 490: 321–327. Zhuang, Y.B., Seong, J.G., Do, Y.S. et al. (2016). J. Membr. Sci. 504: 55–65. Li, S.H., Jo, H.J., Han, S.H. et al. (2013). J. Membr. Sci. 434: 137–147. Swaidan, R., Ghanem, B., Al-Saeedi, M. et al. (2014). Macromolecules 47: 7453–7462. Zhuang, Y.B., Seong, J.G., Do, Y.S. et al. (2016). Chem. Commun. 52: 3817–3820. Du, N.Y., Robertson, G.P., Song, J.S. et al. (2008). Macromolecules 41: 9656–9662. Wang, Z.G., Wang, D., and Jin, J. (2014). Macromolecules 47: 7477–7483. Bhide, B.O. and Stern, S.A. (1991). J. Membr. Sci. 62: 13–35. Rose, I., Carta, M., Malpass-Evans, R. et al. (2015). ACS Macro Lett. 4: 912–915. Carta, M., Bernardo, P., Clarizia, G. et al. (2014). Macromolecules 47: 8320–8327. Carta, M., Croad, M., Malpass-Evans, R. et al. (2014). Adv. Mater. 26: 3526–3531. Bezzu, C.G., Carta, M., Tonkins, A. et al. (2012). Adv. Mater. 24: 5930–5933. Ghanem, B.S., Swaidan, R., Litwiller, E., and Pinnau, I. (2014). Adv. Mater. 26: 3688–3692. Rogan, Y., Starannikova, L., Ryzhikh, V. et al. (2013). Polym. Chem. 4: 3813–3820. Ghanem, B., Alaslai, N., Miao, X.H., and Pinnau, I. (2016). Polymer 96: 13–19. Swaidan, R., Ghanem, B., and Pinnau, I. (2015). ACS Macro Lett. 4: 947–951.

435

436

12 Microporous Organic Framework Materials for Membrane Separations

109 Thornton, A.W., Doherty, C.M., Falcaro, P. et al. (2013). J. Phys. Chem. C

117: 24654–24661. 110 Hao, L., Zuo, J., and Chung, T.S. (2014). AIChE J. 60: 3848–3858. 111 Ghanem, B.S., McKeown, N.B., Budd, P.M. et al. (2008). Adv. Mater. 20: 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

2766–2771. Yong, W.F., Li, F.Y., Xiao, Y.C. et al. (2012). J. Membr. Sci. 407–408: 47–57. Yong, W.F., Li, F.Y., Xiao, Y.C. et al. (2013). J. Membr. Sci. 443: 156–169. Yong, W.F. and Chung, T.S. (2015). Polymer 59: 290–297. Ghanem, B., Alghunaimi, F., Ma, X.H. et al. (2016). Polymer 101: 225–232. Mao, H.C. and Zhang, S.B. (2017). J. Colloid Interface Sci. 490: 29–36. Chen, Y.R., Chen, L.H., Chang, K.S. et al. (2016). J. Membr. Sci. 514: 114–124. Ma, X.H., Ghanem, B., Salines, O. et al. (2015). ACS Macro Lett. 4: 231–235. Thomas, S., Pinnau, I., Du, N.Y., and Guiver, M.D. (2009). J. Membr. Sci. 333: 125–131. Thomas, S., Pinnau, I., Du, N.Y., and Guiver, M.D. (2009). J. Membr. Sci. 338: 1–4. Swaidan, R.J., Ghanem, B., Swaidan, R. et al. (2015). J. Membr. Sci. 492: 116–122. Swaidan, R.J., Ma, X.H., Litwiller, E., and Pinnau, I. (2015). J. Membr. Sci. 495: 235–241. Liao, K.S., Lai, J.Y., and Chung, T.S. (2016). J. Membr. Sci. 515: 36–44. Wang, Y., Li, J.P., Yang, Q.Y., and Zhong, C.L. (2016). ACS Appl. Mater. Interfaces 8: 8694–8701. Fan, H.W., Gu, J.H., Meng, H. et al. (2018). Angew. Chem. Int. Ed. 57: 4083–4087. Liang, B., Wang, H., Shi, X.H. et al. (2018). Nat. Chem. 10: 961–967. Kandambeth, S., Biswal, B.P., Chaudhari, H.D. et al. (2017). Adv. Mater. 29: 1603945. Shinde, D.B., Sheng, G., Li, X. et al. (2018). J. Am. Chem. Soc. 140: 14342–14349. Zhang, W.X., Zhang, L.M., Zhao, H.F. et al. (2018). J. Mater. Chem. A 6: 13331–13339. Wang, R., Shi, X.S., Xiao, A.K. et al. (2018). J. Membr. Sci. 566: 197–204.

437

Index a AB-type ladder PIMs 417 acid separation 266–268 acid stable Zr-MOF NH2 -UiO-66 380 activated carbon 5–7, 77, 82, 124, 330 activated carbon fibers (ACFs) 6, 82 activated carbon monoliths (ACMs) 6 activated charcoal 5 active polyamide (PA) film 140 addition-type poly(trimethylsilyl norbornene) 22 adsorption-diffusion model 367 Ag/polyacrylic acid (PAA)-CNTs hybrid microporous membranes 125 air-liquid interfacial synthesis 298–300 air separation 226, 228–230, 247, 423–427 alcohol dehydration 260–264 aligned AlPO4 -5 crystals 228 alkylamines 335, 341–343, 353 ALL-CRAFT program 7 allotropes 8, 11, 13, 77 α-Al2 O3 microfiltration layer 263 α-alumina supported ZIF-8 membranes 402 AlPO-18 235, 240 aluminium-rich MFI-type zeolite membrane 263 aluminum tri-sec-butylate sol 63 Amberlyst coupled modernite 270 3-aminopropyl triethoxysilane (APTES) 64, 100, 290, 429 3-aminopropyltrimethoxysilane (APMS) 172, 313

amorphous silica layer 60, 245 amorphous silica membranes 59, 67 amyloid fibrils 124 annealing process 165, 173 anodic aluminum oxide (AAO) 97, 129, 419 archetypal PIM-TB 416 aromatic-rich HCP α,α′ -dichloro-p-xylene (p-DCX) 182 artificial ceramic membranes 3 as-synthesized CNT arrays 9 as-synthesized UiO-66 membranes 371 asymmetric membranes 35, 163, 225 azeotropic alcohol-water mixtures 67

b 1, 4-benzenedicarboxylic acid (H2 BDC) 19 benzene-1,3,5-tricarboxylate (BTC) linkers 178 β-lactoglobulin 124 bicontinuous cubic (Cubbi) membrane 123 bilayered silicalite-1/ZSM-5 membranes 208 bilayer membranes 207–208 bimetallic MOF nanocages 379 bio-diaphragm 1 bioethanol 262, 263, 377 biological pollutants 115, 116, 118, 131–134 bioreactors 3 1,4-bis-(4-pyridyl)benzene (bipyb) 398

Microporous Materials for Separation Membranes, First Edition. Xiaoqin Zou and Guangshan Zhu. © 2019 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2019 by Wiley-VCH Verlag GmbH & Co. KGaA.

438

Index

2,2-bistrifluoromethyl-4,5-difluoro-1, 3-dioxole 22 block copoly(ether-urethane-urea) 162 block copolymers (BCPs) 70, 97, 133, 162 Blue Membranes GmbH 82, 83 Boehmite precursor nanoparticle dispersion 58 b-oriented MFI layer 205 b-oriented MFI zeolite membranes 244 boron-ZSM-5 233 Böttcher and Higuchi models 167 Bruggeman model 167 Brunauer, Paul Hugh Emett and Edward Teller (BET) theory 303 BTESE-derived hybrid silica membranes 67, 68 bulk nanotube materials 10

c carbonaceous materials activated carbon 5–7 carbon nanotubes 8–11 graphene 11–14 carbonaceous nanofiber (CNF) membrane 125 carbon aerogels 6 carbon based membranes advantages of 86–87 CMS post-treatment of 81–82 precursors 79–80 pre-treatments of 81 production of 79 pyrolysis environment for 80–81 structures of 77, 78 CNTs 87–92 disadvantages of 87 graphene electron beam and ultraviolet induced oxidative etching, holes 95 focused electron beam irradiation 92

gas separation and water purification 94 GOMs 92 MD simulations 92 NPG 92 permeability and selectivity of 94, 95 PTMSP surfaces 96 rejection of 98 structure of 77, 78 water flux 97 water treatment membranes 99 MMMs 99–102 SSF 84–85 strategic directions of concentration polarization 103–104 cost 106–107 fouling 104–105 mechanical stability 105–106 scalability 106 carbon capture and storage (CCS) 419 carbon fabric (CF) membrane 127 carbonization 5, 38, 78–80, 84, 198 carbonized-template molecular sieving silica (CTMSS) membranes 60 Carbon Membranes Ltd. (Israel) 83 carbon molecular sieves (CMS) 226 module construction of 82–84 post-treatment of 81–82 precursors 79–80 pre-treatments of 81 production of 79 pyrolysis environment for 80–81 structures of 77, 78 carbon nanotubes (CNTs) 8, 122, 125 functionalization 90–92 mixed membranes 87, 89 MWNTs 78 structure of 77 SWNTs 78 VA-CNT 87 well-aligned 125 well-dispersed 1D 133 carbon template silica (CTS) membranes 70

Index

catalytic cracking deposition (CCD) 247, 269 (catalytic) membrane reactor 268 cellulose acetate (CA) 2, 162 ceramic membranes 3 CHA-type small pore silicoaluminophosphate zeolite 238 chemical absorption 324, 326–328 chemical-looping combustion (CLC) system 324, 334–335 chemical vapor deposition (CVD) 16, 56, 89 graphene 94 ZIF-8 membrane 296 chemisorption 20, 21, 330 China-based 2D Carbon Graphene Material Co. Ltd. 13–14 chiral MOFs 394, 401 chiral Ni2 (l-asp)2 (bipy) (Ni-LAB) membrane 400 chiral resolution 362, 394–401 closed pores 414 CNT-based yarns 11 CO2 capture 230, 419 and separation strategy chemical absorption 327–328 CLC system 334–335 cryogenic purification technology 331 inorganic membranes 333 oxyfuel combustion 326–327 physical absorption 328–330 polymeric membranes 332–333 post-combustion 325–326 pre-combustion 326 silica membrane 334 zeolite membrane 333–334 composite membranes 67, 124, 131, 375, 378, 388, 389, 430 concentration-dependent diffusion coefficient 259 concentration polarization (CP) 30, 31, 43, 103, 256, 258, 380 conjugated microporous polymers (CMP) 24, 413, 430

continuous and defect polycrystalline MOF membrane 366 continuous MOF (Zn2 (BDC)2 DABCO) membrane 367 continuous thin ZIF-8 membrane 386 continuous two-dimensional imine-linked COF-LZU1 membrane 429 continuous UiO-66 polycrystalline membranes 370 CO2 /N2 using Faujasite-type zeolite (NaY) membranes 230 conventional polymeric membrane 86, 102 CoOx Si xerogels 70 copoly(ether urethane) 162 copper 10, 11, 13, 98, 118, 139, 147, 149, 178, 294, 309, 348 CO2 separation, microporous silica membranes 63–66 counter electrode (CE) 295 covalent functionalization 91, 92 covalent organic frameworks (COFs) 24, 181, 413, 429 robust 184 covalent triazine-based frameworks (CTFs) 24, 413 covalent-triazine-framework membranes (CTF-0) 428 critical flux 29 cryogenic purification technology 331 crystalline/semi-crystalline triazines-based materials 24 crystalline TFPDHF 2D COF membrane 431 crystalline zeolite membranes 227 crystallization by microwave heating 200–201 C4 to C8 hydrocarbon isomers 216 Cubbi-membrane 123 cubic LTA (Linde Type A) zeolite crystals 143 Cu-MOFs 178, 179 [Cu2 (bdc)2 (bpy)]n 376 [Cu2 (bdc)2 (bpy)]n /SPES-C MMMs 376 Cussler model 100

439

440

Index

d dabco, 4,4-bipyridine (bipy) 398 3D all-carbon scaffolds 8 DD3R membrane 272 defect-free Al-containing zeolite membrane 202 defect-free silicalite-1 zeolite membranes 272 defect-free thin composite membranes 378 defect-free zeolite membranes 215 defined-nanostructure membranes 162 dehydrated ethanol 260 desalination 134 microporous silica membranes 69–72 3,5-diamino-N-(4-2-aminophenyl)benzamide (DABA) 141 diethanolamine (DEA) 328 diffuse reflectance infrared fourier transform spectroscopy (DRIFTS) 302, 307–308 diffusion-based Kr/Xe separation mechanisms 243 diffusivity coefficient (DA ) 167 9,9-dihexylfluorene-2,7-diamine (DHF) 431 dimethyl terephthalate (DMT) 367 diphenyldiethoxysilane (DPDES) 60 direct dimethyl ether (DME) synthesis 271 disc-shaped membranes 256 dissolved oxygen (DO) 124 dry (alumino) silicate gel 208 dry/wet process 163 dynamic nuclear polarization (DNP) 307

e electrically charged membranes 36 electro-dialysis (ED) 3 electron-deficient metallocenes 10 electron-rich metallocenes 10 electrophoretic deposition (EPD) 294, 295 electro-spinning technique 293

electrospun nanofibrous scaffold 131 enhanced oil recovery (EOR) 331 ethyl tertiary butyl ether (ETBE) 260 evaporation-controlled filler positioning (EFP) method 387

f fabricate defect-free membranes 72 fabricated PAF-56P/PSF membranes 182 FAU-type zeolite membranes 272 6FDA-HAB/DABA polyimide 380 Fick’s first law 364, 365 “film-like” MFI-alumina membrane 276 filtered MFI nanosheet film 266 Fischer–Tropsch synthesis 271 Flory–Huggins theory 365 flue gas desulphurization (FGD) process 325 fluorinated carbon nanotubes 127 fluor process 330 forward osmosis (FO) 116, 120, 380 fouling 29, 31, 41, 43, 87, 103–105, 107, 129, 131, 134, 136, 139, 142, 143, 147, 149, 150, 152, 153, 258, 327 fouling resistance (Rf ) 104 free volume concept 415 Friedel–Crafts acylation technique 91 functionality insertion method 337 functionalized PAFs 181 functional zeolite film 208–210

g γ-alumina layer formation 62 gas/liquid chromatography 257 gas mixture permeation 211, 348 gas permeation 53, 62, 100, 171, 172, 211, 212, 225, 227, 233, 234, 238, 309, 313–315, 346–348, 350, 368, 371, 404, 427 gas separations 39, 427 with zeolite membranes air separation 228–230 CO2 capture 230–235 H2 S capture 238

Index

Kr/Xe separation 238–244 N2 /CH4 separation 235–238 post-synthesis modification 245 g-C3 N4 nanosheets 126 gel polarization model 29–30 Gesellschaft Für Trenntechnik (GFT) 3 Ge-ZSM-5 membrane 207, 267 glassy polymer membranes 230 glassy polymers 162, 163, 179, 182, 230, 415 GO-based nanoporous membranes 128 Gonzo-Parentis-Gottifredi (GPG) model 168 GO/PES 389 grafting 20, 91, 104, 105, 147, 176, 247, 342 graphene 11 electron beam and ultraviolet induced oxidative etching, holes 95 focused electron beam irradiation 92 gas separation and water purification 94 permeability and selectivity of 94, 95 PTMSP surfaces 96 rejection of 98 structure of 77, 78 water flux 97 water treatment membranes 99 graphene oxide (GO) 77, 92, 128, 346, 381 graphene oxide membranes (GOMs) 92, 95, 97, 98, 122 growth of oriented zeolite layers on supports 205

h Henry’s law 328, 365 hetero-SURMOFs 394 hexagonal microporous silica 16, 56 hexane isomers 265 high-aspect-ratio-SAPO-34 236 higher boiling point octanol 387 high free-volume glassy polymers 21 highly cross-linked PIM-1 421

highly permeable and selective membranes 225 highly porous polymers, high free-volume glassy polymers 21–23 high performance GO-TiO2 microsphere hierarchical membrane 129 high-performance superhydrophobic/superoleophilic membrane 125 high-purity exfoliated MFI nanosheet 266 high-quality ZIF-8/PSS membrane 391 HKUST-1 single crystals 291 hollow-fiber membrane modules 42 hollow fibre MFI zeolite membrane reactor configuration 273 hollow fibre SAPO-34 membranes 242 homochiral MOF membrane [Ni2 (mal)2 (bpy)]2 H2 O (Ni-MB) 400 homochiral, pillared-layer [Cu2 (d-cam)2 (P)] 398 homochiral poly(l-DOPA) thin film 399 homochiral [Zn2 (bdc)(l-lac)(dmf )] membrane 399 homogeneous membranes 35, 364 homogeneous nanodispersed ZIF-8/PDMS membrane 375 homogeneous ZIF-8/silicone rubber nanocomposite membrane 375 H2 -permselective CuO/ZnO/Al2 O3 /MFI membrane reactor 273 H2 S capture 226, 238 hybtonite carbon nanoepoxy resins 10 hydrogen recovery 85, 225, 416–419 hydrogen separation in dehydrogenation reactions 271–273 microporous silica membranes 60–63 in syngas production 273–274 in water gas shift reaction 273

441

442

Index

hydrolysis/peptization method 57 hydrophilic graphene oxide (GO) nanosheets 389 hydrophilic LTA membranes 260 hydrophilic membranes 149, 210, 212, 261, 263, 271 hydrophilic organic polymer membranes 269 hydrophilic zeolites 143, 210, 211, 266, 271 hydrophilic ZSM-5 zeolite membrane 267 hydrophobic membranes 151–153, 212, 263 hydrophobic small-pore zeolite membranes 232 hydroxy sodalite zeolite membranes 271 hypercrosslinked polymers (HCPs) 24, 182, 413 H-ZSM-5 membranes 275

i i-butane 227, 271, 272 ideal gas separation membrane 225 immobilizing Aquaporin Z-reconstituted liposomes 122 inductive coupled plasma optical emission spectroscopy (ICP-OES) 306 industrial membrane-based separation technologies 195 inorganic fillers 87, 162, 165, 166, 169, 176, 210, 362, 365 inorganic membranes 333 inorganic-organic hybrid membranes 134 inorganic pollutants 118 in-situ aging 200 in-situ aging–microwave synthesis 200 in-situ crystallization 187, 196–198, 210, 231, 288–291 MOF 289 inter-grown MFI nanosheets 266 intergrowth supporting substances 201 interstitial free volume 415

iso-electric point (IEP) 70 isomer separation 275

k Kang–Jones–Nair (KJN) model 169 KAUST-PI-1 membranes 427 Knudsen diffusion behaviour 309 mechanism 85 Knudsen model 89 Knudsen number 28 Knudsen separation factor 201 Knudsen separation mechanism 332 Kr/Xe separation 238–244 K-SAPO-34 242

l laminar MoS2 membrane 129–131 Langmuir–Blodgett (LB) method 301, 431 Langmuir–Blodgett (LB) strategy 300 Langmuir-Schäfer layer 300 large-pore ZIF-68 membranes 368 layer-by layer (LBL) polyamide/ZIF-8 nanocomposite membrane 392 layer deposition (LB) Langmuir–Blodgett layer 300 Langmuir-Schäfer layer 300 layered WS2 nanosheet membrane 131 Lewis–Nielsen model 168 linear layer by layer (LBL) growth 296 liquid-liquid interfacial synthesis 298 liquid membranes 35, 36, 40 liquid phase epitaxy (LPE) 296–298, 394 liquid separation 1, 27, 72, 101, 225, 361–406, 428 lithium-coated graphene 12 lithium silicate (Li4 SiO4 ) 331 lithium zirconate (Li2 ZrO3 ) 331 low molecular weight materials (LMWMs) 172 LTA zeolite membranes 200, 270

m Materials Institute Lavoisier (MILs) 177–178

Index

Matrimid 78, 80, 165, 170, 171, 175, 176, 184, 188, 210 Matrimid/β-cyclodextrin membranes 186 Maxwell model 99, 168 Maxwell–Wagner–Sillar model 167 medium-pore zeolites 196, 231, 333 melt extrusion 36 membrane-based separation processes 225 membrane characterization 211–215, 402 membrane configurations carbonization 38 melt extrusion 36 membrane modules 40–43 slip coating-sintering procedure 37 sol-gel method 37–38 solution casting 36 spin coating 37 spinning 37 structures 35–36 technologies 38–40 membrane desalination 142, ZIF–8 composite membranes 134 cubic LTA (Linde Type A) zeolite crystals 143 mixed matrix (MM) CNT membranes 135 modified CNT membranes 136 monolayer graphene 137 multi-walled carbon nanotubes 135 nano-and micron-sized porous GO membranes 139 nanoscale metals 134 photocatalyst nano-TiO2 134 porous MCM-41 nanoparticles 142 single-walled carbon nanotubes 135 TFC membrane technology 141 thin film composite membrane 140 tip modified CNT membranes 136 titanium (Ti) nanoparticles 134 vertically aligned (VA) CNT membranes 135 zeolites 143 membrane permeability 152, 272, 333, 363

membrane permeation 1, 27, 272, 364 membrane science and technology 1, 413 membrane selectivity 40, 104, 153, 176, 229, 244, 247, 364, 372, 375 membrane separation 1 definition 27 membrane transport for gas systems 27–29 membrane transport for liquid systems 29 transport mechanism in the ED membrane 32–34 membrane surface engineering hydrophobic membranes 151–153 nanoparticles 147–151 membrane technology 1–3, 38–40, 69, 125, 134, 141, 153, 225, 235, 238, 247, 324, 331–334 membrane transport for gas systems 27–29 for liquid systems 29–32 mesoporous silica spheres (MSS) 378 Me-substituted α-cyclodextrin 187 metal-doped silica membranes 59 metal-organic framework (MOF) membranes characterization techniques DRIFTS 307–308 ICP-OES 306 nitrogen adsorption and desorption 303–304 NMR spectroscopy 306–307 PXRD pattern 302–303 scanning electron microscope (SEM) 305–306 SS-NMR spectroscopy 307 thermal gravimetric analysis (TGA) 304–305 CO2 capture and separation strategy chemical absorption 327–328 CLC system 334–335 cryogenic purification technology 331 inorganic membranes 333 oxyfuel combustion 326–327 physical absorption 328–330

443

444

Index

metal-organic framework (MOF) membranes (contd.) polymeric membranes 332–333 post-combustion 325–326 pre-combustion 326 silica membrane 334 zeolite membrane 333–334 CO2 recognition alkylamine incorporation 341–343 core-shell materials 341 polar functional groups 336–339 pore size and function control 339–341 unsaturated metal sites 335–336 CO2 separation 344–353 fabrications of polycrystalline membranes 288–296 solution based and vacuum-based techniques 288 SURMOFs membranes 296–300 H2 separation 308–315 metal organic frameworks (MOFs) 19, 361 chiral resolution 394–401 organic solvent nanofiltration 381–394 selective separation of chemicals via pervaporation 364–381 stability 401–404 metathesis of propene 274–277 methanol separation in hydrogenation reaction 274 methylated silica membranes 259 methyl diethanolamine (MDEA) 328 MFI-type zeolite (ZSM5) membranes 170, 195, 230, 269 MFI zeolite crystals 199 microfiltration (MF) 2, 3, 116, 263 micrometer-sized ZIF-71 filled membranes 377 micropollutants 90, 115, 134, 136, 391 micropores 6, 44, 60, 62, 68, 78, 82, 212–214, 235, 240, 353, 419, 427 microporous materials carbonaceous materials 4–14

different species of 4 high free-volume glassy polymers 21–23 metal-organic frameworks 19–21 microporous silica 14–16 porous organic frameworks 23–26 zeolites 16–19 microporous membranes 35 features of 43–45 microporous organic framework materials air separation 423–427 CO2 capture 419–423 gas separations 427–428 hydrogen recovery 416–419 liquid separation 428–431 porous structures and free volumes 413–416 microporous polymeric materials 4 microporous silica membranes 14 carbon dioxide separation 63–66 CVD 56–57 desalination 69–72 future aspect 72–73 hydrogen separation 60–63 intermediate layer 53 intermediate layers 57–58 modification of 58–60 pervaporation 66–69 schematic diagram of 54 sol-gel synthesis processes 54 support layer 58 templating approach 55–56 microstructures of zeolite films 210–211 MIL-68(Al)/polyimide MMMs 178 mixed matrix (MM) CNT membranes 135, 138 mixed matrix membranes (MMMs) 99, 210, 362 fabrication and drying techniques 163–166 historical growth of academic publications 161 mass transport theory and models 166–169 MOFs based 173–180, 373

Index

Cu-MOFs 178–179 Materials Institute Lavoisier 177–178 MOF-74 series 179–180 UiO-66 series 174–176 zeolitic imidazolate frameworks 176–177 MOP-18/Matrimid 184 polymer phase 162–163 porous molecular compounds 184 porous organic frameworks 181 solvents 163 zeolites 169–173 mixture separation factor 202, 215, 227, 231, 233, 276, 314, 346–351, 366 modernite membrane 270 modified CNT membranes 136 modified emulsion precipitation methods 57 modified hydrophilic mZIF nanoparticles 390 modified UiO-66-NH2 nanoparticles 175 MOF-based film materials 394 MOF [Cu2 (d-cam)2 (dabco)] thin films 398 MOF metacrystal 399 MOF-74 series 179–180 molecular dynamics (MD) simulations 92, 122 molecular sieves 18, 78–82, 169, 201, 235, 236, 260, 268, 275, 313, 330 monoethanolamine (MEA) 327 monolayer graphene 92, 94, 97, 98, 122, 137, 139 monolith membranes 256, 257 Monsanto Prism membrane 3 8 MR zeolite membranes 212 10 MR zeolite membranes 212 MTES-templating silica sol 65 multi-layer zeolite membranes 207 multistage flash distillation (MSFD) 369 multi-walled carbon nanotubes (MWCNTs) 8, 9, 100, 125, 135

multi-walled nanotubes (MWNTs) 8, 78 MXene-membrane 123

n NaA zeolite membranes 270, 271 nano-and micro-sized porous GO membranes 139 nano-confined composite membranes (NCC) 388 nanofiltration (NF) membranes 122, 428 nanoparticles 91, 102, 118, 147–151, 161, 170, 175, 176, 179, 346 nanoporous graphene (NPG) 92 nanoporous silicon nitride membrane 124 nanoscale metals 134, 323 nanosheets 102, 126, 129, 142, 300, 378, 381, 390 nanosized MIL-101(Cr) 385 narrow pore all-silica zeolite membranes 227 natural rubber 21 natural zeolites 16, 17 NaY and NaX zeolite membranes 264 NaY zeolite membrane 265 N2 /CH4 separation 226, 235–238 n-hexane 101, 212–214, 265, 366, 393 n-hexane/2,2-DMB separations 265 NH2 -UiO-66/PEI composite membranes 380 Ni-doped porous silica membranes 62, 63 nitrocellulose membranes 1 nitrogen adsorption and desorption 303–304 N-methyl-N-(trimethylsilyl)trifluoroacetamide (MSTFA) 385 5-nm γ-Al2 O3 nanofiltration 263 non-covalent functionalization 91, 92 norbornene (NBE) 21 Noria 188 Noria-Boc 188 Noria-COt Bu 188

445

446

Index

Noria/6FDA-DAM membranes 188 N-rich Schiff based material (SNW-1) 423 nuclear magnetic resonance (NMR) spectroscopy 302, 306

o obtained GO-membrane 129 Office of Saline Water (OSW) 2 olefin/paraffin separation 427 one-dimensional conductors 10 O2 /N2 separation factor 64, 65, 80, 229, 230 open pores 43, 44, 414, 427 organic membranes 134 organic–organic mixtures 68, 69, 255, 264, 361, 364 organic pollutants 115, 118, 124–131 organic polymer membranes 86, 195, 268, 269 organic solvent nanofiltration (OSN) continuous thin ZIF-8 membrane 386 evaporation-controlled filler positioning method 387 GO/PES 389 higher boiling point octanol 387 high-quality ZIF-8/PSS membrane 391 hydrophilic graphene oxide (GO) nanosheets 389 micropollutants 391 MMMs 385 nano-confined composite membranes 388 nanosized MIL-101(Cr) 385 N-methyl-N-(trimethylsilyl)trifluoroacetamide (MSTFA) 385 PA-SNW-1/PES membrane 393 PA/UiO-66 membranes 392 polyamide (PA)/ZIF-8 nanocomposite membrane 391 polymeric membranes 385 poly(sodium 4-styrenesulfonate) modified ZIF-8 390

polytetrafluoroethylene double layer microfiltration membrane 391 positively charged MOF/chitosan NF membranes 393 thin-film nanocomposite membranes 385 tubular ceramic ZIF-8/PSS hybrid membrane 387 ZIF-8/GO nanocomposites 390 ZIF-8/GO thin-film nanocomposite membrane 390 ZIF-8/PEI hybrid NF membrane 393 organic-water mixture 69 organic–organic separation 264 organophilic pervaporation (OPV) membranes 374 O2 -selective microporous organic membranes 427 oxyfuel combustion 326 oxygen plasma etching approach 122 o-xylene 207, 212, 367, 368

p PAF-56P/PSF membranes 182 Pal model 168 PA modified UiO-66-NH2 /Matrimid 176 PA-SNW-1/PES membrane 393 pathogenic microorganisms 131 PA/UiO-66 membranes 392 PA/ZIF-8 (LBL) membrane 391 PDA-coated microporous membrane 122 PDMS/HKUST-1 layer 349 peptization 37, 57 perfluoropolymers 21, 22 permporosimetry 212 measurements 204 pervaporation 3, 39, 101 process 66–69 separation index 375 with zeolite membranes acid separation 266–268 alchohol dehydration 260–264 disc-shaped membranes 256 gas/liquid chromatography 257

Index

hydrogen separation in dehydrogenation reactions 271–273 hydrogen separation in syngas production 273–274 hydrogen separation in water gas shift reaction 273 isomer separation 275–277 metathesis of propene 274 methanol separation in hydrogenation reaction 274 monolith membranes 256 multi-channel monolith support 256 organic–organic separation 264–266 pervaporation fluxes 256 schematic 255 surface diffusion 258 tubular membranes 256 vapor permeate 256 water separation 268–271 zeolite membrane encapsulated catalyst 277 phase inversion 101, 126, 150, 163, 165, 171, 172 phenolic resins (PRs) 79 phenyltriethoxysilane (PTES) 60 photocatalyst nano-TiO2 134 physical absorption process 329 physical adsorption 92, 324, 330–331 physisorption 20, 330 PIM-BTrip-TB 425, 426 PIM-1/cage MMMs 188 PIM-300-2.0d membrane 421 PIM-6FDA-OH membrane 427 PIM-1/nanocage membranes 188 PIM-PI-1 membranes 427 PIM-TB membranes 417 pinhole-free zeolite membranes 198 plate-and-frame modules 2, 41 plugging–filling method 375 poly(ethylene oxide-bamide) (Pebax) 162 poly(furfuryl alcohol) (PFA) 79 poly(methyl methacrylate) (PMMA) 188

poly(styrene-butadiene-styrene) (SBS) 162 polyacrylonitrile (PAN) 6, 38, 77, 79, 348, 381, 393, 430 polyacrylonitrile fiber 6 polyamide (PA) 88, 162, 379, 393 polyamide-nanofilm-based NF membranes 430 polyamide (PA)/ZIF-8 nanocomposite membrane 391 poly(zwitterionic) CNT (ZCNT) 104 polycrystalline membranes assembly of MOF nanocrystals 296 CVD 296 direct synthesis controlled deposition 291 in-situ crystallization approach 288–299 slow diffusion of reactants 291 electro-chemical deposition 294–295 seeded growth coordination polymer 294 nanocrystals 292 thin Film 294 stepwise dosage of reagents 296 polycrystalline MOFs membranes 365, 394 polycrystalline UiO-66(Zr)-(OH)2 membranes 372, 404 polycrystalline zeolite layer 196, 199, 213, 258 polydopamine (PDA) 152, 346 polyether-block-amide (PEBA) 375 polyetherimide (PEI) UF-membrane 151 polyethersulfone (PES) supported silicalite-1/polydimethylsiloxane (PDMS) mixed matrix membrane 243 polyethersulfone (PES) 162, 171, 225 polyethylene glycol (PEG) 131, 151, 329, 391 polyethyleneimine (PEI) 188, 292 modified MIL-101(Cr) fillers 178 polyimides (PI) 21, 79, 162

447

448

Index

polymer-copper nanocomposite 149 polymeric membrane 3, 4, 118, 127, 134, 136, 139, 150, 152, 161, 177, 195, 226, 229, 243, 258, 332–334, 344, 346, 361, 385, 392, 413, 423 polymer-melt-compounding exfoliation technique 266 polymers of intrinsic microporosity (PIMs) 24, 181, 413 poly(sodium 4-styrenesulfonate) (PSS) modified ZIF-8 (mZIF) 390 polypropylene (PP) 36, 162, 196 poly(norbornene)s 21, 22 polystyrene (PS) 16, 56, 88, 131, 188 polystyrene-block-poly(methyl methacrylate) copolymer (PS-b-PMMA) 131 polysulfone (PSF) 162, 172 nanofibers 127 support membranes 392 polytetrafluoroethylene (PTFE) double layer microfiltration membrane 391 poly[(1-trimethylsilyl)-1-propyne] [poly(11)] 22 polyvinyl acetate (PVAc) 151 polyvinyl alcohol (PVA) 101, 149, 151 polyvinylidene fluoride (PVDF) 162, 163 polyvinylpyrrolidone (PVP) 151, 165, 188 pore-plugging type zeolite/alumina nanocomposite membrane 275 porous aromatic frameworks (PAFs) 24, 181, 413 porous cages (PCs) 24, 188, 413 porous carbons 77, 99, 107 porous coordination polymers (PCPs) 361 porous MCM-41 nanoparticles 142 porous molecular compounds 184 porous organic cages (POCs) 186, 418 porous organic frameworks (POFs) 4, 23–26, 181, 413–432 porous pure-silica materials 209 porous structures and free volumes 413–416

positively charged MOF/chitosan NF membranes 393 powdered activated carbons (PACs) 6 powder X-ray diffraction (PXRD) 302–303 pressure-assisted self-assembly (PASA) filtration technique 381 pressure swing adsorption (PSA) process 85, 86, 260 priming 172, 173 pristine graphene 77, 92, 95 PSF/modified SAPO-34 membranes 172 PTMSP/PAF-Li6 C60 membranes 181 PTMSP/p-DCX MMMs 182 pure-silica zeolites 209 p-xylene 182, 206–208, 212, 266, 275, 276, 367, 368

r reactive seeding (RS) 292, 293, 347, 366, 368, 399 rectangular silicalite-1 nanoblocks 197 rectisol process 329 reduced graphene oxide (rGO) 77, 133 reduced permeation polarizability 168 rejection factor 363 relative recovery 363 resistance model 30–32 retention factor 363 reverse osmosis (RO) 2, 3, 71, 116, 364, 369 rubbery polymers 25, 162, 163, 179

s Safe Drinking Water Act 7 SAPO-34 236, 243 membranes 231 zeolite 172 scanning electron microscope (SEM) 305 SDA-containing MFI films 210 secondary building units (SBUs) 19, 178 seed crystals 197, 199, 204, 261, 292, 293, 368, 400 segmented polyurethane (SPU) 102

Index

selective separation of chemicals via pervaporation Fick’s first law 364 MOFs based MMMs 373 polycrystalline MOF membranes 365–373 solution-diffusion mechanism 364 selective surface flow membranes (SSF) 84–85 self-cross-linked PIM-1 membranes 419 self-polymerized chiral monomer l-DOPA 399 semipermeable membrane 1, 27 short branched linear polymers 15, 55 Si/Al-O-Al/Si (zeolite) repeating unit 176 silane agents 172 silica-free zeolite-like materials 17 silicalite-1 195–197, 199, 200, 206, 208, 209, 212, 213, 227, 230, 232, 233, 244, 247, 259, 264, 266, 267, 272, 274, 275, 277 silicalite-1/α-Al2 O3 zeolite membranes 275 silicalite-1 coated ZSM-5 catalysts 275 silicalite-1 membrane crystallization 200 silicalite-1 membranes 197, 200, 201, 206, 212, 213, 215, 227, 232, 233, 237, 267, 275 silica membrane 15, 53–73, 334 silica nanoblocks 197 silica-titania composite membranes 68 single-crystal membranes 315 single-layer freestanding graphene membrane 122 single-layer graphene 12, 13, 92, 94, 104–106 single-walled carbon nanotubes 8, 135 single-walled nanotubes (SWNTs) 8, 78 size exclusion mechanism 99, 116, 139 slip coating-sintering procedure 37 small-pore zeolite DD3R membranes 272 small-pore zeolites 170, 196, 210, 231

SNW-1/PSF membranes 423 sodium-alumino-phosphate (SAPO)-based MMMs 171 sol-gel method 37, 60, 70, 334 sol-gel synthesis processes 14, 54 solid state nuclear magnetic resonance (SS-NMR) spectroscopy 302, 307 solubility coefficient 167, 329 soluble anionic MOP (Na6 H18 -[Cu24 (5-SO3 -1,3-BDC)24 (S)24 ]⋅x S 184 solution based fabrication techniques 288 solution-casting method 36, 165, 185 solution-diffusion mechanism 28, 118, 332, 364, 365 solvent resistant nanofiltration 381 solvents 3, 4, 9, 19, 22, 36, 163, 184, 188, 306, 326–328, 364, 368 sonochemicals method 57, 58 specific surface area (SSA) 4, 77, 414 specific van der Waals volume 415 spin coating 36, 37, 226, 292, 293, 298, 419 spinning 36, 37, 43, 122, 133, 177, 210, 302 spin-on silicalite-1 films 209 spiral wound module 41, 42, 84 SSZ-13 zeolite membranes 238, 268 stable UiO-66 membranes 372 1-stage or 2-stage filtration principle 7 stainless-steel-net-supported zeolite NaA membrane 229 starbons 6 submicrometer-sized ZIF-71 crystals 377 submicrometer-sized ZIF-71 filled PDMS membranes 377 substituted polyacetylenes 21, 22 sulfonated PIs 80 superhydrophobic PVDF-co-hexafluoropropylene nanofiber membrane 122 superoleophobic poly(acrylic acid)-grafted PVDF (PAA-g-PVDF) membrane 126

449

450

Index

surface anchored MOFs (SURMOFs) 394 surface diffusion 28, 29, 85, 230, 231, 244, 258, 287, 332, 351, 422 surface supported metal-organic frameworks (SURMOFs) membrane interfacial synthesis air-liquid 298–300 Langmuir–Blodgett (LB) strategy 300 Langmuir–Schäfer layer 300 liquid-liquid 298 liquid phase epitaxy (LPE) 296 syngas production 273–274 synthetic zeolites 18, 210

t Teflon AF1600 22 Teflon AF2400 22 templating molecules 15, 16, 55, 56 terephthalic acid (TPA) 208, 367 tetraethoxysilane Si(OC2 H5 )4 (TEOS) 54 tetraethyl orthosilicate (TEOS) 60, 70 tetraethyl orthosilicate Si(OC2 H5 )4 (TEOS) 14 5,5′ ,6,6′ -tetrahydroxy-3,3,3′ ,3′ tetramethyl-1,1′ -spirobisindane 25 tetramethoxysilane Si(OCH3 )4 (TMOS) 14, 54 tetrapropylammonium ions (TPA+ ) 198 TFC membrane technology 141 TFMPSPIM-1 membrane 425 thermally-rearranged polymers 21 thermogravimetric analysis (TGA) 302, 304 thermotropic liquid-crystal (LC) molecules 123 thin film composite (TFC) membrane 140, 379 thin-film nanocomposite (TFN) membranes 379, 385

three-dimensional (3D) zeolite membranes 266 Ti-exchanged UiO-66 175 TiO2 147 nanoparticles 129 nanowire ultrafiltration (UF) membranes 131 tip functionalization 90 tip modified CNT membranes 136 titanium (Ti) nanoparticles 134 titanosilicate zeolites 171 toluene 125, 126, 277, 368 transmembrane pressure (TMP) 104, 389, 401 2,4,6-triaminopyrimidine (TAP) 173 triazine-framework-based membrane 422 1,3,5-triformylbenzene and (R,R)-1,2-diaminocyclohexane (CC3) 188 1,3,5-triformylphloroglucinol (Tp) 431 1,3,5-triisopropylbenzene (TIPB) 368 triisopropyl orthoformate (TIPO) 101 trimesoyl chloride (TMC) 141, 385 Tsapatsis 207 T-type zeolite membranes 231 tubular ceramic ZIF-8/PSS hybrid membrane 387, 388 tubular membrane modules 41, 256 tubular ZeoSepA membranes 260 tunable pore size, MOF 351 two-dimensional (2D) graphene based materials 122 nanosheets 129 ZIF-L nanosheets 378 2D zeolite membranes 266 two-layered MOR (acid resistant separative layer)/H-ZSM-5(catalyst) membrane 270 two-steps crystallization 199 type-A zeolite membranes 260

u UiO-66 243, 295, 338 UiO-66-NH2 membranes 372 UiO-66/PEBA MMMs 175

Index

UiO-66 series 174–176 ultrafiltration (UF) membranes 3, 90, 116 ultrathin (∼22–53nm thick) 2D-graphene membrane 128 ultra-thin MFI membranes 229 ultra-thin (∼0.5μm in thickness) MFI membranes 229 ultrathin MOF/organosilica MMMs 176 unmodified MIL-53(Al) containing MMMs 178 unsaturated metal sites 19, 179, 323, 335–336

water-stable polycrystalline UiO-66(Zr)-(OH)2 membranes 404 well packed graphene layer membrane 128 wet process 163 Williams–Landel–Ferry (WLF) equation 415 working electrode (WE) 295

x X-ray amorphous metal oxide membranes 199, 226, 227, 261 xylene isomers 206, 266, 275, 276, 367, 368

v vacuum-based fabrication techniques 288 vacuum swing adsorptive (VSA) 330 van’t Hoff equation 1 vertically aligned CNT membranes (VA-CNT) 87, 135 arrays 8 vertically-aligned multi-walled carbon nanotubes (VAMWCNTs) 125 vycor glass 57

w waterborne diseases 115, 131 water desalination 70 water gas shift (WGS) reaction 62, 85, 273 water purification biological pollutants 131–134 inorganic pollutants 118–124 membrane desalination 134–144 membrane surface engineering 147–151 organic pollutants 124–131 types and state-of-the-art microporous membrane 116–118 water separation 97, 125, 127, 153, 268–271 water-stable, exfoliated imine-based COFs 183

z zeolite imidazolate framework (ZIF) based membranes 345 zeolite membrane encapsulated catalyst (ZMEC) 277 zeolite membranes 4, 16, 143, 169, 333 bi-layer membranes 207–208 characterization 211–215 conventional hydrothermal synthesis 198 crystallization by microwave heating 200–201 functional zeolite film 208–210 growth of oriented zeolite layers on supports 205–207 intergrowth supporting substances 201–205 microstructures of zeolite films 210–211 mixed matrix membranes 210 synthesis techniques 196–197 two-steps crystallization 199 zeolitic imidazolate frameworks (ZIFs) 176–177, 312, 366 zeolitic molecular sieve membranes 268 ZeoSepA membranes 260, 261 zero-dimensional ZIF-8 nanoparticles 378

451

452

Index

ZIF-78 347 ZIF-90 290 ZIF-71 filled MMMs pervaporation performance 378 ZIF-8/GO nanocomposites 390 ZIF-8/GO thin-film nanocomposite (TFN-ZG) membrane 390 ZIF-8 membrane 142, 243, 345 MMMs 176 ZIF-71 membrane 366 ZIF-95 membrane 347 ZIF-100 membrane 347 ZIF-8 nanoparticles 374 ZIF-8/PEI hybrid NF membrane, 393 ZIF-8/PMPS membrane 374

zirconium(IV)-carboxylate MOFs (Zr-MOFs) 369 Zn-based ZIF membranes 369 Zn2 (d/l-cam)2 dabco (diazabicyclo[2.2.2]-octane, dabco) 394 ZnO incorporated PVDF membrane 150 Zr-based MOF 338, 369, 372 ZSM-5 195 MFI-type zeolite 170 ZSM-5/Matrimid MMMs 171 ZSM-5/silicalite bilayer membranes 273 Zyvex Performance Materials 10