Hao Jihua
(Department of Chemical Engineering, Tsinghua University, Beijing 100084, China )

Received Apr. 13,  2000.

Abstract Polymeric membrane materials for CO₂-selective separation have been reviewed. On the basis of the introduction to solution-diffusion transport mechanism of gases in solid polymeric membranes, transport characteristics of CO₂ were summarized. CO₂-selective polymeric membrane can be divided into four types, including the chemical and physical modification of the traditional membrane materials, polymer blending, the design and synthesis of new membrane materials and organic-inorganic hybrid membrane. Finally, the research topics in future were also suggested.
Keywords membrane; carbon dioxide;

CO₂-selective separation is a very important problem in many fields, such as environment, petrochemical engineering, agriculture and other relative industry.^[1-16] One of the promising techniques is membrane processes. Because of the potential economy of the membrane technology, many governments have initiated and extensively focused in the projects on the CO₂-selective polymeric membranes and membrane processes. Membranes and membrane processes for the separation of CO₂ and H₂S from natural gases have been a project under the "BRITE"(Basic Research in Industrial Technology for Europe) formed as a European Community R&D program in the middle of 1980s.^[5] In Japan, membrane processes for CO₂-selective separation has been an important project under C1 Chemistry "Sunshine" plan.^[6] In China, membrane processes for CO₂/CH₄ separation have been studied because of the existence of several large sour natural gas fields and application in tertiary oil recovery.^[7] A number of large American companies first applied the membrane technology for separation of CO₂/CH₄ and continued efforts on new membrane materials and membrane process design for CO₂ separation.^[8] The membranes for CO₂-selective separation have made much progress during this period.
    Because earlier cellulose acetate membrane for reverse osmosis could be easily adapted to form dry asymmetric structures using solvent-exchange dehydration for gas separation, especially Monsanto invented "caulked membrane" successfully for recovery of hydrogen from ammonia purge gas, solid polymer-based membranes for gas separation have been paid attention to in the membrane field over the past twenty years. New polymeric membrane formation with integrally-skinned dense thin asymmetric structure have been the main research topics.
    There have been two kinds of polymeric membrane for CO₂-selective separation processes. One is the solid polymeric membrane based on the solution-diffusion mechanism, another is the facilitated transport membrane based on the solution-reaction-diffusion mechanism.^[17]
     In this review we limited only to membrane materials based on the solution-diffusion mechanism.

1 TRANSPORT MECHANISM OF GASES IN SOLID POLYMERIC MEMBRANES AND CHARACTERISTICS OF CO₂
The permeation of a gas molecule through a dense polymer membrane proceeds by a solution-diffusion mechanism. In this model, gas molecules first dissolve into the high pressure face of the membrane, diffuse across the membrane to the low pressure side, and desorb from the face.^{[18, 19]}
    The permeability coefficient P of a membrane in solution-diffusion processes can be represented as the product of a diffusivity [D] and solubility [S] coefficient as shown in the following equation.
P=[D][S]
The ideal selectivity can be expressed as
α_AB=P_A/P_B=[D_A/D_B][S_A/S_B]
Where, [D_A/D_B]---------the diffusivity selectivity
             [S_A/S_B]----------the solubility selectivity
    In addition to operating conditions (i.e. temperature, pressure and composition), gas solubility also depends on factors such as penetrant interactions, and polymer morphology(i.e. crystallinity, orientation, etc.).^[20] Gas diffusivity depends on the ability of small gas molecules to undergo diffusive jumps which occur when thermally driven, random, cooperative polymer segmental dynamics from transient gaps, large enough to accommodate penetrants, in the immediate vicinity of the gas molecules.^[21] Like solubility, diffusivity depends on membrane operating conditions; moreover, gas diffusivity is also sensitive to properties such as penetrant size, polymer morphology and polymer segmental dynamics.

2 CHARACTERISTICS OF CO₂ AND CH₄
2.1 Gas condensibility
In most polymers, CO₂(T_c=31°C) is more soluble than CH₄ (T_c=-82.1°C). Gas critical temperature, T_c, normal boiling point, T_b, and Lennard-Jones force constant are all measures of condensibility which correlate well with the solubility coefficients of a range of penetrants in polymers. Gas solubility in polymers generally increases with increasing gas condensibility.^{[22, 23]}
2.2 Penetrant size and shape 
Diffusion coefficients in polymers are sensitive to penetrant size and shape. The diffusivity of linear or oblong penetrant molecules such as CO₂ is higher than the diffusivities of spherical molecules of equivalent molecular volume. The van der Waals volumes of CO₂ and CH₄ are estimated to be 17.5 and 17.2 cm³/mole.^{[22, 24]} These molecular volumes yield equivalent spherical diameters of 3.33 and 3.31Å for CO₂ and CH₄, respectively. But the diffusion coefficient of CO₂ is typically greater than that of CH₄, despite the larger van der Waals volume of CO₂. The kinetic diameter of the asymmetric molecule CO₂ (3.30Å) is smaller than that of methane, a symmetric molecule (3.80Å). Based on such results, the transport of small, asymmetric molecules is understood to proceed with diffusion jumps of the penetrant through a polymer occurring principally parallel to the long axis of the penetrant.

Fig. 1. The effect of polar carbonyl and sulfone group concentration on CO₂/CH₄ solubility^[25]

    Most glassy gas separation membrane materials achieve high permselectivity as a result of high diffusivity selectivity (i.e. by sieving penetrant molecules based on differences in molecular size).

2.3 Polarity   
Gas solubility is sensitive to specific interactions between gas and polymer molecules. Gases such as CO₂, which has a quadrupole moment, are in general, more soluble in polar polymers. Many polymeric materials for CO₂-selective membrane have been designed by use of this property. Van Amerongen determined that the solubility of quadrupolar CO₂ increased substantially as acrylonitrile content increased.^[25] Pilate suggested that CO₂ interact favorably with the sulfone group in polysulfone via induced dipole-dipole interaction.^[26] In this regard, Koros found a correlation between CO₂/CH₄ solubility selectivity and the concentration of polar carbonyl and sulfone groups in the medium in which the gases were dissolved.^[27] These data are reproduced in Fig. 1.^[25] In a related work, Chern and coworkers found that aryl nitration of polysulfone, which adds polar nitro groups to the polymer backbone, increases CO₂/CH₄.^[28]
2.4 Pressure    
Permeability of CO₂ in polymeric membrane is more sensitive to pressure. The selectivity of CO₂/CH₄ mixture in the most of the current polymeric membranes decreased at high pressure because of the membrane plasticization. This is one of the key problems to be solved for CO₂-selective polymeric membrane.
    At very high gas pressure, permeability may increase with increasing pressure. This increase is related to penetrant plasticization of the polymer matrix and is primarily due to an increase in the penetrant diffusion coefficient. The data in Fig. 2 present plasticization in the presence of CO₂.^{[22, 29]} Other condensable gases and hydrocarbon vapors can also plasticize polymers. At sufficiently high pressure the penetrant acts as an efficient plasticizer in hydrophilic polymers such as polyvinyl alcohol. The permeability decreases with pressure initially, consistent with the prediction of dual mode model, and then increases when the gas begins to plasticize the polymer. Plasticization occurs when penetrant molecules dissolve in the polymer chain segmental motion. The increased segmental mobility may be observed by a depression of the glass transition temperature of the polymer-penetrant mixture. The increased segmental mobility provides more opportunities for penetrant molecules to execute diffusive jumps, and, in turn, the penetrant diffusion coefficient increases. Plasticization is undesirable for most gas separation membranes since the polymers used in these applications depend largely on diffusivity selectivity to achieve high overall permselectivity. The relatively large, uncontrolled polymer segmental motions associated with the plasticized or rubbery polymers decrease the diffusivity selectivity and, therefore, overall permselectivity of polymer.

Fig. 2 CO₂ induced plasticization of poly(tetrabromophenolphthalein) at 35°C^[27]

3 POLYMERIC MEMBRANE MATERIALS FOR SEPARATION OF CO₂/CH₄     
Much effort has been spent to optimize CO₂-selective polymeric membranes: including the chemical and physical modification of current membranes,^{[30, 31, 32]} blending,^{[33, 34]} the design and synthesis of new membrane material^{[35, 36]} and organic-inorganic membrane.^[37]

Fig. 3 "Trade-off" curves for the CO₂/CH₄ separation system

    A typical "trade-off curve" between CO₂/CH₄ permselectivity and CO₂ permeability is pressured in Fig. 3.^[38] A crosshatched, undesirable area, below the solid line is associated with "low free volume glasses" and "rubbery polymers". The poor mobility selectivities and high diffusivities in rubbery media noted earlier produce high permeabilities and low overall permselectivities. On the other hand, the well-packed, low free volume glasses, give equally unattractively low permeabilities and high selectivities. The solids line (-^.-) represents "optimistic" tradeoff behavior typical of commercially available engineering resins.
    During the past decade many investigators attempted to synthesize entirely new polymers that exhibit both a higher permeability and a higher selectivity than presently available polymers. Or they attempted to enhance the selectivity and/or permeability of available polymers by chemical or physical modification.

Fig. 4 Influence of chemical modification

   Quite a lot of modification have been carried out to those polymers that are proven gas separation materials. ^{[30, 31, 32]} Stam et al. modified on-chain of polymers, like polysulfone (PSF), polyethersulfone (PES) and poly(phenylene oxide) (PPO) by substitution of aromatic protons, i.e. by means of nitration, silylation, bromination of alkyl groups and acrylation.^[30] Some results are shown in Fig. 4. Even modifications with more complex chemicals have not shown the desired result.
    Kesting et al. improved the flux of polysulfone gas separation membranes by using Lewis acid-base pairs in the aqueous phase inversion bath. The resulting membranes possessed greater free volume and hence, higher flux, than those made conventionally.³¹ In addition, Stam put additives to the polymer membrane by physical modification. The most successful was the addition of molecular sieves(Na-Y). Addition of 1% by weight of molecular sieve to the polymer membrane has shown a relative improvement of 2% on the permeability without effecting the selectivity. At contents of 15% molecular sieve maximum permeability has been reached. At higher contents selectivity was reducing again, where as permeability was not improving any further.³⁰ Barbari treated polysulfone membrane with molecular bromine. After bromine treatment, the selectivity of a polysulfone membrane for CO₂ over CH₄ was increased over 100% at 10 atm upstream pressure with only a 36% reduction in CO₂ permeability.
    Although all types of chemical and physical modification showed that selectivity or permeability could be improved, even at the expense of permeability or selectivity, the selectivities or permeabilities themselves have not exhibited remarkably attractive.
    However, the tailor-made polymers have shown that it is possible to move away from the traditional trade-off curve connecting the conventional membrane materials (-D- in Fig. 3).^[38] Although considerable data and general correlations relating structure and permeability exist, there are no truly quantitative relationships to guide detailed structure-permeability optimization. Fortunately, qualitative rules have emerged that date back to the pioneering work by Hoeln and his colleagues at DuPont in the 1970's.^{[39, 40]} Simply put, when changing the structure within a family of polymers, inhibiting intersegmental packing, while simultaneously hindering the backbone mobility tends to produce a desirable tradeoff between productivity and permselectivity changes. Currently this is the most reliable guide for understanding structure-mobility studies of a given family of polymers.
    Table 1 lists permeability coefficients for CO₂ and CH₄ in a variety of rubbery and glassy polymers, as well as the selectivities (ideal separation factors) of these polymers for CO₂/CH₄. The high selectivities of these tailor-made polymers, such as 6FDA-polyimides, polytriazoles, and polyoxadiazoles are assumed to originate from their rigid structure. Their high permeability is mostly explained in terms of their high free volume, which is due to their bulky groups preventing efficient packing. The solubility is mostly relative to the chemical composition of the polymers.^[38]
    In the following section the excellent polymer materials with extra high selectivity and/or permeability will be introduced in detail.

Table 1. Permeabilities and selectivities of CO₂/CH₄ in rubbery and glassy polymers

3.1 Improvement of the selectivity of polymers with high permeability
3.1.1 Poly(organosiloxane)
Generally, rubbery polymers exhibit high permeabilities and low selectivity, in which silicone polymers, and in particular poly(dimethylsiloxane) (PDMS), has received considerable attention. The high permeability of PDMS, [-(CH₃)₂SiO-]_n, has been attributed to its large free volume, which may be due to the flexibility of the siloxane (-SiO-) linkages of this polymer. Although the gas selectivity of PDMS is very low, the structure/permeability relationships of other silicone polymers have been systematically investigated in the last few years. These studies had the objective of finding membrane materials that exhibit higher gas selectivity than PDMS as well as high gas permeability.
    The substitution of increasingly bulkier functional groups in the backbone or side chains of PDMS, or the replacement of its flexible -SiO- linkages with stiffer -SiCH₂- linkages, raises the glass-transition temperature, T_g, of the polymers.^[21] This means the increase of chain-packing density. The effects of such substitution on the relationships between the gas permeability and selectivity of silicone polymers were listed in Table 2.

Table 2. Permeabilities and selectivities of CO₂/CH₄ in rubbery and glassy polymers

As shown in Table 2, the high CO₂/CH₄ selectivity of poly(trifluoropropyl methyl siloxane) [(CF₃C₂H₄)CH₃SiO]_n is due to an anomalously high CO₂ solubility, which results from specific interactions between this gas, whose molecules have a strong quadrupole moment, and the polar fluorine-containing groups in the polymer.^{[41, 42]}

Fig.5 Effect of differential pressure on the CO₂/CH₄ relative pure gas permeabilities in the diester functionalised membranes: diester functionality (a) 30 mol%; (b) 20mol%; (c) 10mol%;  (d) 0 mol%.

Fig.6 CO₂ permeabilities coefficient at 35C and atmospheric pressure versus composition of the CO₂:CH₄ permeant gas mixture in diester functionalised membranes: diester functionality (a) 0 mol%; (b) 10mol%; (c) 30mol%.

    In order to improve the permselectivity of rubbery polymers without unacceptable loss in overall permeability via improved solubility selectivity Ashworth et al. introduce a side-chain organofuntional groups, such as mono- or di-ester, into a PDMS matrix.^{[43, 44]} The increase of selectivity results from the interaction of CO₂ with the ester groups in the ester side-chains. The results have shown in Fig. 5, 6.
    In addition, the surface of poly(dimethylsiloxane) membrane was modified by plasma treatments using Ar, N₂, O₂ and NH₃.^[45] As a result, the selectivity of membrane has been remarkably improved by the plasma treatment. The selectivity and the permeation rate of CO₂/CH₄ obtained were 53 and 3.3×10^-4 cm³/s^.cm^2. cmHg.

Table 3. Permeability coefficients and selectivities for CO₂/CH₄ in substituted polyacetylenes

Poly[1-(trimethylsilyl)-1-propane], [-CH₃C=CSi(CH₃)₃-]_n, designated hereafter PTMSP, has the highest intrinsic gas permeability of all synthetic polymers known at present.^[46] Although it is a glassy polymer (T_g>200°C), was found to have an approximately ten-fold higher permeability for light gases than rubbery PDMS, but a somewhat lower selectivity.
    The exceptionally high permeability of PTMSP has been attributed to its large free volume, which was believe to arise from the PTMSP matrix consists of relatively stiff backbone chains separate by bulky trimethylsilyl side groups, Si(CH₃)₃. The stiffness of the backbone chains is due in part to the alternating double bonds of the polymer and in part to the fact that the bulky trimethylsilyl groups hinder intrasegmental rotation. These groups also prevent chain packing by serving as "intersegmental spacers".
    In order to study the structure/permeability relationships of these polymers and discover polymer structure that exhibit a substantially higher gas selectivity than PTMSP without a significant decline in the permeability, a large number of substituted polyacetylenes have been synthesized.^{[47, 48]} Some of the results are presented in Table 3.^[49]
     Because the rates of gas permeation through PTMSP decrease markedly with time or on thermal cycling, and their gas selectivities low, no practical uses of the polyacetylenes as membrane materials have been found so far.
3.1.3 Poly(vinyltrimethylsilane)
Poly(vinyltrimethylsilane), designated hereafter PVTMS, [-CH₂CHSi(CH₃)₃-]_n, is another glassy polymer which exhibits a high gas permeability (as shown in Table 1) due to its bulky trimethylsilyl pendant groups, as is the case also for PTMSP.^[50] The main difference between these two polymers lies in that PVTMS has flexible vinyl backbone chains whereas PTMSP has stiff backbone chains with alternating doublebonds. The gas permeability of PVTMS is 1-3 orders of magnitude lower than that of PTMSP, but the gas selectivity of PVTMS is markedly higher than that of PTMSP. Now PVTMS membranes have been used in Russia to produce oxygen-enriched air for oxygen therapy. But this material for gas separation appears to have received relatively little attention in the west.
3.2 Improvement of the permeability of polymers with high selectivity
3.2.1 Polyimide 
Polyimides have been identified as materials with high selectivities for CO₂/CH₄ (over 60-80). In addition to high selectivity, polyimides have high glass transition temperature(T_g>300°C) and can, therefore, be used at high service temperatures. A substantial insight has been gained in the last ten years into the effects of variations in the chemical structure of polyimides on gas solution and transport in these polymers. As a result, polyimides exhibiting increasingly higher gas selectivities and permeabilities are being developed in academic and industrial laboratories. In particular, those polyimides with a hexafluoro substitute carbon -C(CF₃)₂ in the backbone have been found to be considerably more CO₂-selective than other glassy polymers with comparable permabilities.^[51-60] A family of polyimides and related materials are used by DuPont and Ube for commercial gas separation membranes.Table 4 lists the permeability coefficients and selectivities of CO₂/CH₄ in the several excellent polyimides known as far.

Table 4. Permeabilities and selectivities of CO₂/CH₄ in polyimides* ^[51-60]

*6FDA:2,2-bis(3,4-dicarboxyphenyl) hexafluoro isopropane dianhydride
ODA:oxydianiline
MDA:methylenedianiline
IPDA:isopropylidenedianiline
PDA:phenylenediamine
DDS:diaminodiphenylsulfone
DMMDA:3,3'-dimethyl-4,4'-methylene dianiline
HQDPA:1,4-bis(3,4-dicarboxyphenyl) benzene dianhydride

    To reduce undesirable effects caused by CO₂ induced swelling in CO₂/CH₄ separation processes, the polyimide structure was stabilized with crosslinks. One method for crosslinking membrane materials is the UV-irradiation of benzophenone-containing polyimides or bisphenol A based polyacrylates.^{[61, 62]} This leads, especially in the separation of gas pairs with a large difference in molecular size, to a significant improvement of permselectivity. But simultaneously the permeability is decreased by crosslinking due to the strongly reduced chain mobility is decreased packing density of the polymer chains. A disadvantage of photochemical crosslinked membrane materials is that the degree of crosslinking depends strongly on experimental conditions, i.e. irradiation time or the type of mercury lamp. Moreover, under the experimental conditions of UV-irradiation crosslinking reaction as well as photo-fries rearrangements are possible.^[62]
    Claudia et al. synthesized polyimides with strong polar associating functional groups as well as chemical crosslinked polyimides and copolyimides with simultaneous inhibition of chain mobility and intrasegmental packing by changes in backbone structure.^[63] In the polymerization reaction 6FDA was used as dianhydride monomer and mPD(m-phenylene diamine) and DABA (diamine benzoic acid) were used as diamine monomers. With copolyimides containing strong polar carboxylic acid groups (i.e. 6FDA-mPD/DABA 9:1) reduced plasticization was seen up to a pure CO₂ can be reduced at least up to a pure CO₂ feed pressure of 35atm. With increasing degree of crosslinking, increasing CO₂/CH₄ selectivity was found because of reduced swelling and polymer chain mobility. By using ethylene glycol as a crosslingking agent, CO₂ permeability was not significantly lower because the reduced chain mobility was compensated by the additional free volume caused by the crosslinks.
3.2.2 Polyaniline   
Polyaniline, (C₆H₄NH)_n, has attracted considerable attention because of the surprisingly high gas selectivity.^[64-66] Anderson et.al reported that a remarkable selectivity of CO₂/CH₄ in polyaniline, 336, was achieved by a complex process involving doping, undoping, and redoping membranes made from this polymer with select counterions, like acid solution of HF, HCl, HBr and HI. This report shows that conjugated polymer is one of the most exciting materials for gas separation membranes.^[64]
    Hachisuka et al. used an oxidized polyaniline as undoped film, use various protonic acids, e.g., methane sulphonic acid, polyvinyl sulphonic acid, polyisoprene sulphonic acid, and polyvinyl sulfate as dopants. It has shown that CO₂ permeability was increased by the formation of a quinonediimine unit in polyaniline with the oxidation. The selectivity of polyaniline membranes was found to be remarkably improved by doping. In particular, the selectivity of the polyaniline membrane for CO₂/CH₄ doped with polyvinyl sulphonic acid as a polymer dopant went up to over 2000. Table 5 show the selectivity of CO₂/CH₄ in polyaniline membranes doped with various dopants.^[65]

Table 5. Permeabilities and selectivities of CO₂/CH₄ in polyaniline

3.2.3 Polytriazoles and polyoxadiazoles    
These polymers have been stated to be chemically resistant as well as thermally stable with glass transition temperature between 260 and 360°C. Gebben et al. have measured the permeability of poly[1,3-phenyl-1,4-phenyl]-4-phenyl-1,3,4-triazole] (TIPT) to CO₂ and CH₄ at the temperatures up to 200°C.^[35] Its selectivity with a CO₂/CH₄ mixture containing 20mol% CO₂ at 25°C is over 60, a CO₂ permeability is 10-20Barrer. The T_g of 270°C of TIPT is lower than that of a number of polyimides with comparable selectivities to CO₂/CH₄ gas pair. This suggests that the intrasegmental mobility of TIPT is higher than that of the latter polymers, perhaps due to the mobility of the phenyl rings in the para positions. The high selectivity of TIPT to CO₂/CH₄ gas pairs is unexpected in view of its relatively large fractional free volume (FFV) and moderate T_g. A sharper free-volume distribution in TIPT any conceivably account for the gas selectivity of TIPT.
    An unusually high CO₂/CH₄ selectivity of 168, at an unspecified temperature, has been reported for a poly(1,3,4-oxadiazole)(POD) which incorporates a diphenyl ether (DPE) moiety in its backbone chains.^[64] However, the permeability of this polymer to CO₂ is very low, 0.3 Barrer. The high T_g of DPE-POD of 333°C indicates a low intrasegmental mobility, i.e., a high backbone-chain stiffness, in spite of the presence of potentially mobile -O- linkages in the DPE moiety. The very high CO₂/CH₄ selectivity and low permeability to CO₂ suggest a high chain packing density and low FFV, and possible a narrow free-volume distribution. Rotation around -O- linkages is probably inhibited by the high chain packing.
    By contrast, substitution in POD of either 1,1,3-trimethyl-3-phenylindane (PIDA) or 4,4',2,2'-diphenyl hexafluoropropane (HF) yielded polymers with much higher permeabilities to CO₂, 78-93 Barrer, but relatively low CO₂/CH₄ selectivities of 26-28.
3.3 Polymer blending

Fig. 7 Permeability of CO₂ for various PSF/PI blends at 40°C

Although polyimide show excellent mechanical properties, high temperature and chemical resistance, and separation performances for CO₂, the polyimides are seriously affected by highly soluble penetrants CO₂, which above a given partial pressure can plasticize the polymer matrix.^[68] The critical partial pressure of plasticization for polyimide is relatively low, varying between 10-20 bar.^[69] For this reason polyimide membranes have found little application in the separation of CO₂ from light hydrocarbons in enhanced oil recovery or in landfill gas upgrading programs.
    Polymer blending has been taken as optimization of membrane properties. If one could achieve equivalent segmental composition by either blending or copolymerization, it is currently thought that equivalent solution/diffusion separation properties would result. Although earlier extensive work on polymer blend membranes was attempted to do, it is very difficult to find miscible polymer blends with satisfied separation performances.^{[33, 34]}
    However, Sakellaropoulos et al. prepared the blend membrane from the mixtures of the polysulfone and the aromatic polyimide.^{[69, 70]} Because the critical pressure of plasticization for polysulfone exceed 55Bar, blend membranes were considerably more resistant to the plasticization phenomenon compared with those of pure polyimide, shown as in Fig. 7.^[71] In addition, the membranes showed significant permeability improvements, compared to pure polyimide, with minor change in their selectivity. Polysulfone and polyimide proved to be completely miscible polymers as confirmed from optical microscopy, glass transition temperatures and spectroscopy analyses of the prepared mixtures. The complete miscibility permits the preparation of symmetric and asymmetric blend membranes in any proportion (1-99%) of polysulfone and polyimide.
    Blending of suitably selected polymers is generally considered as a viable method to modify the properties of polymers used for preparation of gas separation membranes. But the candidate polymer pairs for preparation of miscible blends are few.
3.4 Organic-inorganic hybrid membranes

Table 6. Permeabilities and selectivities of CO₂/CH₄ in the 6FPAI and 6FPAI/TiO₂

Because of the limitations of purely organic polymers, especially to plasticization, organic-inorganic hybrid materials may offer new possibilities to gas separation application.^{[35, 72]} Hybrid organic-inorganic materials can be formed utilizing a sol-gel process via a number of approaches. One method involves treating an orthosilicate directly with the organic polymer or an oligomer, which contains functional groups capable of a cross-reaction with the inorganic oxide, thus provide connectivity between the organic phase and inorganic network.^[73] Alternatively, alkoxysilane monomers with a polymerizable organic moiety covalently attached to the silicon can be precipitated, followed by polymerization of the organic moiety.^[72] Finally, in situ polymerization of the alkoxide within a swollen polymer network can be used to form nano- or micro composite materials without covalent crosslinks.^[72]
    Hu et al. have successfully fabricated a fluorinated poly(amide-imide)/TiO₂ nano-composite membranes by a sol-gel method. The nano-composite membrane has a denser and a more rigid structure compared to the corresponding pure poly(amide-imide) membrane. Furthermore, there appear to be specific intermolecular interactions between CO₂ and the TiO₂ domains. Finally, the composite membrane has shown higher permselectivities for CO₂/CH₄, as shown in Table 6, even at very low volume concentration of the TiO₂ component. Such results are encouraging, because they suggest that possibly higher selectivities could be achieved at increasingly higher concentrations of the ceramic component.

4 CURRENT PROBLEMS AND RESEARCH TOPICS IN FUTURE  
These "tailor-made" polymers allow both higher selectivities of CO₂/CH₄ and higher permeabilities of CO₂ compared to the traditional polymers. Although a very substantial amount of data on the solution, diffusion, and permeation of CO₂/CH₄ in a variety of rubbery and glassy polymers was already available, the relationships between the chemical structure of polymers and their gas permeability and selectivity were only poorly understood. The new polymers with good gas selectivity or permeability, like 6FDA-ODA, polyaniline, PTMSP and PDMS etc., was occasionally synthesized or found. But it is largely based on prior experience and considerable trial and error. And these "tailor-made" polymers themselves exist some problems for CO₂-selective separation processes. Polyimide is sensitive to plasticize by CO₂. Polyaniline with highest selectivity for CO₂/CH₄ known so far is not sufficiently stable in the presence of humidity, prolonged oxygen exposure and physical aging.
    Systematic primary chemical structure modificatioin of the polymer backbone which simultaneously increase chain rigidity and decrease chain packing density have been shown to lead to membranes with improved permeability without significant losses in gas permselectivity. But as chain rigidity increases, it becomes more difficult to find acceptable solvents in which to dissolve the material. Some of the lower free volume imides must also be cast in amide acid form followed by thermal ring closure. Some such materials, in fact are used in commercial membranes; however, they are clearly more complex and expensive.
    A large increase in solubility might induce plasticization, which would decrease membrane permselectivity. Another deleterious effect of strong polymer-penetrant interactions is the loss of mobility selectivity due to excessive binding between the gas and the polymer. Clearly, the strategy of augmenting permselectivity by increasing solubility selectivity is a delicate balance. An increase in polymer polarity with a simultaneous increase polymer free volume might be expected to increase CO₂/CH₄ solubility selectivity, permeability and permselectivity. However, studies involving the introduction of substitutents, which simultaneously interact with one component in a gas mixture and frustrate chain packing have not been reported.
    From a practical standpoint of processing advance materials, most current generation commercial membranes have asymmetric forms with an open porous support on top of which is an integral skin formed during the cast of a solution of the polymer into a nonsolvent bath. Besides high permeability and selectivity values, processability of the membrane material into desired membrane morphology is of utmost importance.
    In extreme cases, the gaseous feed stream can be composed of over 50% CO₂ at an elevated temperature and up to a feed pressure of 60atm. These extreme operation conditions result in the swelling and plasticization of most membrane materials by the CO₂ present in the feed stream. For the application of membrane based gas separation in EOR processes, it is important to develop new membrane materials with reduced plasticization effects due to higher CO₂ partial pressure.
    Materials research on membranes for CO₂/CH₄ will fouse on these two fields: (a) development of new polymer membrane materials with improved selectivities and permeabilities, simultaneously with unplasticization to higher pressure of CO₂. (b) fabrication of new membrane morphologies allowing higher fluxes(e.g., asymmetric membranes, hollow fiber membranes, and composite membranes) for a given polymer with high permeability of CO₂ and high selectivity of CO₂/CH₄.

REFERENCES   
[1] Hao J H. The preparation of Integrally-skinned asymmetric membrane for separation of CO₂/CH₄ and studies on its formation mechanism, PhD Dissertation, Tianjin University, 1995.
[2] Schell W J. CEP, 1982, 78: 45.
[3] David P. Oil & Gas Journal, Technology, 1984: 85.
[4] Goddin C S. Hydrocarbon Processing, 1982, 61 (5): 125.
[5] Bhide B D, Stern S A. J. Membrane Sci., 1993, 81: 209.
[6] Cooley T E, Dethlff W L. Chem. Eng. Progress, 1985, 81: 45.
[7] Fournie F J C, Agostini T P. J. Petroleum Tech., 1987, 39: 707.
[8] Mc Kee R L, Changela M K, Reading G J. Hydrocarbon Processing, 1991, 70 (4): 63.
[9] Maddox R N, Morgan D J. Hydrocarbon Processing, 1986, 64 (8): 59.
[10] Babcoek R E, Spillman R W, Goddin G S et al. Energy Progress, 1988, 8 (3): 135.
[11] Schell W J. J. Membrane Sci. 1985, 22: 217.
[12] Vansant E F, Dewolfs R. Gas Separation Technology. Amsterdam: Elsevier Science Publishers B. V., 1990: 457.
[13] Rautenbach R, Welsch K J. membrane Sci., 1994, 87 (1): 107.
[14] Cecille L, Toussaint J C. Future Industrial Prospects of Membrane Processes, New York: Elsevier, 1989.
[15] Koros W J, Fleming G K. J. Membrane Sci., 1993, 83 (1): 1-80.
[16] Bhide B D, Stern S A. J. Membrane Sci., 1993, 81 (3): 238.
[17] Teramoto M, Nakai K, Ohnishi N et al. Ind. Eng. Chem. Res. 1996, 35 (2): 538-45.
[18] Koros W J, Fleming G K, Jordan S M et al. Prog. Polym. Sci., 1988, 13: 339-401.
[19] LeBlanc O H, Ward W J, Matson S L et al. J. Membrane Sci., 1980, 6: 339-343.
[20] Weinkauf D H, Paul D R. in Barrier Polymers and Barrier Structures, W. J. Koros, ed., American chemical Society, 1990.
[21] Koros W J, Hellums M W. Encyclopedia of polymer Science, 2nd edn., New York:Wiley-Interscience Publishers, Supplement Volume, 1989: 724.
[22] Ghosal K, Freeman B D. Polymers for Advanced Technologies, 1994, 5: 673-697.
[23] Chern R T, Koros W J, Hopfenberg H B. Material Science of Synthetic membranes, D. R. Lloyd, ed., Washington, DC: American Chemical Society, 1985: Chapter 2.
[24] Chern R T, Brown N F. Macromolecules, 1990, 23: 2370.
[25] van Amerongen G J. J. Polym. Sci., 1950, 5: 307.
[26] Pilato L, Litz L, Hargitay B et al. Polym. Preprint, Am. Chem. Soc. Div. Polym. Sci., 1950, 5: 307.
[27] Koros W J. J. Polym. Sci.:Polym. Phys. Ed., 1985, 23: 1611.
[28] Ghosal K, Chern R T, Freeman B D et al. J. Polym. Sci., Part B: Polym. Phys. 1995, 33 (4): 657-66.
[29] Chern R T, Provan C N. Macromolecules, 1991, 24: 2203.
[30] Stam H. Future Industrial Prospects of Membrane Processes, Ed. By Lecille L and Toussaint J C, NY: Elsevier, 1989.
[31] Kesting R E, Fritzsche A K, Murphy M K et al. J. Applied Polym. Sci., 1990, 40: 1557.
[32] Barbari T A, Datwani S S. J. Membrane Sci., 1995, 107: 263-266.
[33] Paul D R. Polymer Compatibility and Incompatibility: Principle and Practice, K. Solc Ed., MMI Press Symp. Ser., 2, New York: Harwood Academic Publishers, 1982.
[34] Chiou J S, Paul D R. J. Appl. Polym. Sci., 1987, 33: 2935.
[35] Gebben B, Mulder N H V, Smolders C A. J. Membrane Sci., 1989, 46: 29-41.
[36] Li Y, Ding M, Xu J. Polymer International, 1997, 42 (1): 121-126.
[37] Hu Q, Marand E, Dhingra S et al. J. Membrane Sci., 1997, 135: 5-79.
[38] Hensema E R. Adv. Mater., 1994, 6 (4): 269-279.
[39] Pye D G, Hoehn M H, Panar M. J. Applied Polym. Sci., 1976, 20: 287-301.
[40] Hoehn H H. Material Science of Synthetic membranes, D. R. Lloyd, ed., Washington, DC: American Chemical Society, 1985: 81-98.
[41] Stern S A, Shah V M, Hardy B J. J. Polym. Sci., Part B: Polym. Phys., 1987, 25: 1263-1298.
[42] Shah V M, Hardy B J, Stern S A. J. Polym. Sci., Part B: Polym. Phys., 1986, 24: 2033-2047.
[43] Ashworth A J, Brisdon B J, England R et al. Br. Polym. J., 1989, 21: 491.
[44] Ashworth A J, Brisdon B J, England R et al. J. Membrane Sci., 1991, 56: 217.
[45] Matsuyama H, Teramoto M, Hirai K. J. Membrane Sci., 1995, 99: 139-147.
[46] Masuda T, Isobe E, Higashimura T. J. Am. Chem. Soc., 1983, 105: 7473-7474.
[47] Takada K, Matsuya H, Masuda T et al. J. Applied Polym. Sci., 1985, 30: 1605-1616.
[48] Masuda T, Iguchi T, Tang B Z. Polymer, 1988, 29: 2141-2149.
[49] Odani H, Masuda T. Design of polymer membranes for gas separation, in N. Toshima(Ed.), New York, NY, 1992, 107-144.
[50] Davydova M B, Yampol'skii Yu P, Durgar'jan S G. Poly. Sci. USSR, 1988, 30: 1501-1507.
[51] O'Brien K C, Koros W J, Husk G R. J. Membrane Sci., 1988, 35: 217-230.
[52] Kim T H, Koros W J, Husk G R et al. J. Membrane Sci., 1988, 37: 45-62.
[53] Stern S A, Mi Y, Yamamoto H et al. J. Polym. Sci., Polym. Phys. Ed., 1989, 27: 1887-1909.
[54] Tanaka K, Kita H, Okano M et al. Polymer, 1992, 33: 585-592.
[55] Matsumoto K, Xu P. J. Membrane Sci., 1993, 81: 23-30.
[56] Tanaka K, Kita H, Okamoto K. J. Polym. Sci., Part B: Polym. Phys., 1993, 31: 1127-1133.
[57] Kawakami H. J. Applied Polym. Sci., 1996, 62: 965-971.
[58] Rene A, Canada, CA2133046, 1995.
[59] Hiroyoshi K, Masato M, Shoji N. J. Applied Polm. Sci., 1996, 62 (7): 965-970.
[60] Hiroyoshi K, Masato M, Shoji N. J. Membrane Sci., 1996, 118 (2): 223-230.
[61] Kita H, Inada T, Tanaka K et al. J. Membrane Sci., 1994, 87: 139.
[62] Wright C T, Paul D R. J. Membrane Sci., 1997, 124: 161 .
[63] Staudt-Bickel C, Koros W J. J. Membrane Sci., 1999, 155 (1): 145-154.
[64] Anderson M R, Mattes B R, Reiss H et al. Science, 1991, 252: 1412-1415.
[65] Hachisuka H, Ohara T, Ikeda K I et al. J. Applied Polym. Sci., 1995, 56: 1479-1485.
[66] Chang M J, Liao Y H, Myerson A S et al. J. Applied Polym. Sci., 1996, 62 (9): 1427-1436.
[67] Hensema E R, Ph.D. Dissertation, University of Twenty, Enschede, 1991.
[68] Koros W J, Coleman M R, Walker D R B, Ann. Rev. Sci., 1992, 22: 47.
[69] Sakellaropoulos G P, Kaldis S P, Kapantaidakis G C et al. European patent, EP 778077, 1997.
[70] Kapantaidakis G C, Kaaldis S P, Dabou X S et al. J. Membrane Sci., 1996, 110: 239-247.
[71] Sander E S, Jordan S M, Sabrmanian R. J. Membrane Sci., 1992, 74: 29.
[72] Okazaki I, Ohya H, Semenova S I et al. J. Membrane Sci., 1998, 141 (1): 65-78.
[73] Smaihi M, Jermoumi T, Marigan J et al. J. Membrane Sci., 1996, 116: 211-220.

[ Back ] [ Home ] [ Up ] [ Next ] Mirror Site in USA  Europe  China  CSTNet ChinaNet