Membrane Engineering for the Treatment of Gases

Volume 1: Gas-separation Problems with Membranes

By Enrico Drioli, Giuseppe Barbieri

The Royal Society of Chemistry

Copyright © 2011 Royal Society of Chemistry
All rights reserved.
ISBN: 978-1-84973-171-3

Contents

Volume 1,
Chapter 1 Multi-scale Molecular Modeling Approaches for Designing/ Selecting Polymers used for Developing Novel Membranes Elena Tocci and Pluton Pullumbi, 1,
Chapter 2 Simulation of Polymeric Membrane Systems for CO2 Capture Eric Favre, 29,
Chapter 3 Physical Aging of Membranes for Gas Separations B.W. Rowe, B.D. Freeman and D.R. Paul, 58,
Chapter 4 Recent High Performance Polymer Membranes for CO2 Separation S.H. Han and Y.M. Lee, 84,
Chapter 5 Design of Membrane Modules for Gas Separations M. Scholz, M. Wessling and J. Balster, 125,
Chapter 6 Gas/Vapor Permeation Applications in the Hydrocarbon-processing Industry Arnaud Baudot, 150,
Chapter 7 Membrane Gas Separation Processes for Post-combustion CO2 Capture Peter Michael Follmann, Christoph Bayer, Matthias Wessling and Thomas Melin, 196,
Chapter 8 Commercial Applications of Membranes in Gas Separations Pushpinder S. Puri, 215,
Chapter 9 Novel Hybrid Membrane/Pressure Swing Adsorption Processes for Gas Separation Applications Isabel A.A.C. Esteves and José P.B. Mota, 245,
Subject Index, 276,
Volume 2,
Chapter 10 Modeling of Membrane Reactors for Hydrogen Production and Purification F. Gallucci, M. van Sint Annaland and J.A.M. Kuipers, 1,
Chapter 11 Palladium-based Membranes in Hydrogen Production Rune Bredesen, Thijs A. Peters, Marit Stange, Nicla Vicinanza and Hilde J. Venvik, 40,
Chapter 12 Membrane Reactors in Hydrogen Production A. Brunetti, G. Barbieri and E. Drioli, 87,
Chapter 13 Palladium-based Selective Membranes for Hydrogen Production G. Iaquaniello, M. De Falco and A. Salladini, 110,
Chapter 14 Polarization and Inhibition by Carbon Monoxide in Palladium-based Membranes Giuseppe Barbieri, Alessio Caravella and Enrico Drioli, 137,
Chapter 15 Carbon Molecular Sieve Membranes for Gas Separation May-Britt Hägg and Xuezhong He, 162,
Chapter 16 Perovskite Membranes for High Temperature Oxygen Separation F. Liang and J. Caro, 192,
Chapter 17 Zeolite Membranes for Gas Separations C. Algieri, G. Barbieri and E. Drioli, 223,
Chapter 18 Engineering Aspects of MIEC Hollow Fiber Membranes for Oxygen Production X. Tan and K. Li, 253,
Chapter 19 New Metrics in Membrane Gas Separation A. Brunetti, G. Barbieri and E. Drioli, 279,
Subject Index, 302,

CHAPTER 1

Multi-scale Molecular Modeling Approaches for Designing/ Selecting Polymers used for Developing Novel Membranes

ELENA TOCCI AND PLUTON PULLUMBI

1.1 Introduction

During the last decade computational chemistry and numerical simulations have had a favorable impact in almost all branches of materials research ranging from phase determination to structural characterization and property prediction. An important effort has been focused on developing simulation tools to describe thermodynamic and transport properties of confined fluids. The present contribution illustrates the benefit of coupling experiment to molecular modeling for selecting novel membrane materials with better separation properties for given gas mixtures as well as the limitations of the existent computational methodologies. New modeling and simulation tools based on multi-scale hierarchical modeling are needed to cope with the complexity of materials and associated phenomena at different length and time scales.

Transport properties of small molecules in amorphous polymer matrices play an important role in many industrial applications such as gas separation of mixtures, packaging applications ranging from food conservation to controlled drug and cosmetics release, to special coatings for protecting specific substrates from gases.

Different aspects of technology and industrial application of polymer membranes from materials research to permeator design to their optimal configuration to enhance processes performance have been discussed in detail in a previous lecture and recently reviewed in the literature. The potential application of a polymer as a separation membrane depends upon the selectivity towards the gas to be separated and the permeate flux. The selectivity determines the product purity and recovery whereas the permeability is related to the productivity of the membranes. This means that both the permeability and the selectivity should be as large as possible. The control of gas permeability and permselectivity of polymer membranes has become a subject of active research with worldwide participation in both industrial and academic laboratories. The design and optimization of polymer membranes used in gas separation applications would be possible if reliable predictions of transport properties could be made rapidly in advance of synthesis and experiment. The actual status of available commercial software for modeling transport phenomena in polymer membranes, does not allow the development of de novo material design approach. This is due not only to formidable time and length scales involved, but also to lack of detailed information on time evolution of the free volume and its distribution as a function of processing history during the manufacturing process. Rapid progress in computational methodology and validation of new simulation tools is improving the understanding of different facets of gas transport in polymer membranes and building the necessary tools for their effective use in materials design. The possibility to predict transport properties of small molecules through polymer matrices permits the rational selection of polymer materials used in these applications and their optimal design. Although there has been reported an increasing number of studies on this subject over the last years, the prediction of transport properties of gas molecules through glassy polymer membranes remains a difficult target. In many of the recent studies reporting molecular simulation predictions of diffusion and solubility of small gas molecules in several membrane models of the same glassy polymer a great scatter of the predicted values is observed. These results clearly indicate that the quality of the packing of the polymer chain into an amorphous cell membrane model strongly impacts the predicted gas transport properties.

The potential application of a polymer as a separation membrane depends upon the possible throughput and the purity of product. This means that both the permeability of the gas that is transported more rapidly and the selectivity should be as large as possible. The permeability coefficient, Pe, of a small molecule through a polymer membrane is defined as:

Pe = D · S (1.1)

the product of the diffusion coefficient, D (kinetic parameter), and of the solubility coefficient, S (thermodynamic parameter). The estimation of these coefficients can be done, either by molecular dynamics (MD) and grand canonical Monte Carlo simulations, or by applying the transition state theory (TST) approach provided that the quality of the membrane amorphous cells used in the calculation represent the real distribution of torsion angles, of the free volume and its distribution, as well as the structural, conformational and volumetric properties of polymer membranes. The selectivity of a polymer membrane for a pair (i, j) of gas molecules is characterized by the ideal separation factor αij defined in eqn (1.2):

αij = Pei/Pej = Di/Dj, Si/Sj (1.2)

Following this definition the selectivity of a membrane is the product of diffusion selectivity (Di/Dj) and of solubility selectivity (Si/Sj). In the case of glassy polymer membranes the overall selectivity is mainly controlled by diffusion selectivity. Two types of membranes are used commercially in gas separation technology. Glassy polymer membranes are made from stiff chain polymers and operate below their glass transition temperature. These membranes have moderately high free volume and separate gases predominantly based on differences in the sizes of the gas molecules. The smaller molecules (H2, He, O2) permeate more easily through the membrane than the larger ones (CH4, C2H6, N2). The second class of membranes is made either from highly flexible rubbery polymers or ultra-high free volume, glassy substituted polyacetylenes. These membranes separate gases principally by differences in the solubility of gas molecules in these polymers. Larger and more soluble penetrants permeate faster than smaller and less soluble ones.

In the past 20 years, the control of gas permeability and permselectivity of polymer membranes has become a subject of active research with worldwide participation in both industrial and academic laboratories. However, it has been found that simple structural modifications, which usually lead to an increase in polymer permeability, cause loss in permselectivity and vice versa. This so-called ‘trade-off’ relationship has been well described in the literature. Here the log of the separation factor α versus the log of the higher permeability gas Pe yielded a limit for achieving the desired results. The upper bound limit is not fixed in the α–Pe space but moves with time as new polymers with optimized structures become available.

Recent studies of glassy polymer membranes indicate that in addition to the free volume content, gas transport parameters depend upon the backbone chain rigidity, its segmental mobility, the inter-chain distance and the chain interactions. For example, the introduction of bulky alkyl substituents opens up the polymer matrix resulting in a greater permeability. Also the reported introduction of n-alkyl side groups on a polymer backbone increases the side-chain flexibility as well as the membrane free volume with an overall increase of the permeability. The free volume of a polymer, which corresponds to the unoccupied regions accessible to segmental motions, is an important parameter for understanding and predicting many of its characteristic properties. The free volume and the free volume distribution influence the molecular mobility and the transport properties of low molecular substances and gases in polymers. Atomistic simulations allow a detailed investigation of these geometric characteristics. A complete description of definition of the free volume of membrane atomistic model configurations as well as detailed approaches for analyzing it has been reviewed recently. Voronoi polyhedra and Delaunay simplices are used for analyzing the local arrangement of free volume in glassy polymer membranes cells. The calculated free volume and its distribution are used to describe the gas solubility and diffusion in amorphous polymer cells through a QSPR correlation as well as to control the quality of the constructed amorphous cells.

Many literature examples that use the QSPR approach associated with polymer permeability are based on the use of group contribution methods to establish a correlation between the structure of the repeat unit and some physical property of the polymer membrane such as the free volume, the mean segment distance or dielectric constant polarizability, which in its turn is used to predict permeation properties. As a result of such studies some practical criteria have emerged to guide synthetic researchers in improving the permselectivity of membranes which have evolved through extensive experimentation: (i) inhibition of inter-segmental packing meanwhile simultaneously inhibiting intra-segmental (backbone) mobility; and (ii) weakening of inter-chain interactions (reduction of charge transfer complexes). These design rules are based on phenomenological paradigms that provide guidelines for polymer selection. The QSPR approach that uses appropriate descriptors for representing at the same time the repeat unit, the unperturbed polymer chain and the packed amorphous cells, integrating data at different length scales including the some data from the processing history of the polymer membrane would lead to new criteria for materials improvement.

This contribution focuses on the need to shorten the research time of novel polymer materials used for membrane fabrication by combining several computational approaches and experimental techniques. It does not pretend to be a review or cover all aspects that might be considered traits of the computational materials design. Various examples are used to illustrate the use of existing numerical simulation and modeling tools for complementing experimental work. After a brief discussion of some computational approaches used to cover different aspects of polymer membrane simulation, methodologies used to characterize gas transport through polymer membranes as well as to identify factors that control thermodynamic and kinetic properties of confined fluids in membranes will be detailed. It is noticed that the successful application of modeling approaches to gas separation by membrane technology needs the development of models dealing with multi-component gas mixture transport through model membranes. Moreover, for a given polymer membrane, both gas diffusivity and gas solubility depend strongly on process parameters such as pressure difference, feed composition, and temperature. More information on the effects of process parameters on selectivity contribution should be thoroughly considered for identifying membrane materials suitable for each application.

1.2 Computational Methods

Progress in numerical simulation techniques has followed the developments in chemical engineering science and recently has received tremendous attention in both scientific and industrial communities due to the possibility to integrate micro-scale information in investigations that can be carried out at various levels of resolution on different scales of time and length. Several multi-scale modeling methodologies that are based on the information transfer between different scales starting from the molecular level and ending up at the industrial scale have been recently reviewed in the literature indicate a clear trend towards coupling of the design of chemical engineering equipment/units with nanoscale modeling. It is generally expected that multi-scale modeling can lead to optimal unit design as well as to cost optimization of the final products. Fundamental research in materials and structure design now considers numerical simulation, as a complement to theory and experimentation. The integration of physical testing, advanced computations and system simulation would dramatically reduce the design and development time and costs.

In this contribution we do not aim to provide a detailed description of each of the numerical techniques contributing to the multi-scale modeling methodology but will only outline its basic principles, strengths and weaknesses, and potential applications. Interesting readers can refer to relevant books, reviews and research articles for details.

The terminology used for characterizing multi-scale methods often varies with the application domain. Two general approaches have been developed for integrating different models at disparate spatial and temporal scales. The hierarchical modeling extracts information from lower-scale models and transfers it as parameters to the upper-scale (coarser) models of overlapping domains. The coarse-scale model is used over the entire computational domain but a higher-resolution modeling is applied to zoom into a particular sub- domain to obtain updated parameters within the corresponding grids of the coarse model.

Mesoscale modeling uses a basic unit just above the molecular scale, and is particularly useful for studying the behavior of polymers and soft materials. It can model even larger molecular systems, but with the commensurate trade-off in accuracy.

Two methods dissipative particle dynamics (DPD) was initially devised by Hoogerbrugge and Koelman as a particle-based off-lattice simulation method for the flow of complex fluids and to tackle hydrodynamic time and space scales beyond those available with MD. Since DPD is a coarse-grained model and individual atoms or molecules are not represented directly by the particles but they are grouped together into beads, these beads represent local ‘fluid packages’ able to move independently.

The fundamentals of the DPD method are now fairly well-established, as are the technical subtleties and the coarse-grained parametrization of the DPD particles, with improvements consistently being introduced.

Furthermore, it is expected that DPD will play an ever-increasing role in multi-scale modeling approaches through bridging of the atomistic and continuum scales. In such approaches, atomistic simulations are performed to build the DPD models, followed by DPD simulations which provide the necessary input to the continuum codes.

DDFT, which was developed by Fraaije et al. in 1997, is a field-based theoretical method for studying complex fluids, their kinetics and their equilibrium structures at micrometer length and microsecond time scales. DDFT has been applied to the study of the self-assembly of block copolymers in bulk, under shear and in confinement, and to study polymer blend compatibility. Compared to the DPD method, DDFT is computationally extremely fast since larger elements can be modelled. Moreover, since the fluid elements can freely penetrate, larger time steps can be used, and furthermore it is less likely to become trapped in a local minimum. Since DPD is a particle-based method, it can provide somewhat more detailed structural information. Nonetheless, they are both powerful tools in simulating phase separated phenomena that occurs at the mesoscale and the consistency of results from the two methods for the same coarse-grain model is evaluated in this work.

(Continues…)Excerpted from Membrane Engineering for the Treatment of Gases by Enrico Drioli, Giuseppe Barbieri. Copyright © 2011 Royal Society of Chemistry. Excerpted by permission of The Royal Society of Chemistry.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.