Recent developments in membranes for efficient hydrogen purification

Recent developments in membranes for efficient hydrogen purification

Author’s Accepted Manuscript Recent developments in membranes for efficient Hydrogen purification Panyuan Li, Zhi Wang, Zhihua Qiao, Yanni Liu, Xiaoch...

3MB Sizes 0 Downloads 102 Views

Author’s Accepted Manuscript Recent developments in membranes for efficient Hydrogen purification Panyuan Li, Zhi Wang, Zhihua Qiao, Yanni Liu, Xiaochang Cao, Wen Li, Jixiao Wang, Shichang Wang www.elsevier.com/locate/memsci

PII: DOI: Reference:

S0376-7388(15)30105-8 http://dx.doi.org/10.1016/j.memsci.2015.08.010 MEMSCI13897

To appear in: Journal of Membrane Science Received date: 9 April 2015 Revised date: 31 July 2015 Accepted date: 2 August 2015 Cite this article as: Panyuan Li, Zhi Wang, Zhihua Qiao, Yanni Liu, Xiaochang Cao, Wen Li, Jixiao Wang and Shichang Wang, Recent developments in membranes for efficient Hydrogen purification, Journal of Membrane Science, http://dx.doi.org/10.1016/j.memsci.2015.08.010 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Recent developments in membranes for efficient hydrogen purification Panyuan Li, Zhi Wang1, Zhihua Qiao, Yanni Liu, Xiaochang Cao, Wen Li, Jixiao Wang, Shichang Wang Chemical Engineering Research Center, School of Chemical Engineering and Technology, Tianjin University, Tianjin 30072, PR China Tianjin Key Laboratory of Membrane Science and Desalination Technology, State Key Laboratory of Chemical Engineering, Collaborative Innovation Center of Chemical Science and Engineering, Tianjin University, Tianjin 30072, PR China

Abstract Hydrogen has been extensively accepted as a clean and efficient energy carrier to Abbreviations: WGS, water gas shift; AIPO, aluminophosphate; SAPO, silicoaluminophosphate; PES, polyethersulfone; APTES, 3-aminopropyltriethoxysilane; DIC-4, 1,4-diisocyanate; CPTMS, 3-chlropropylyrimethoxysilane; CCD, catalytic cracking deposition; MDES, methyldiethoxysilane; TEOS, tetraethylorthosilicate; CMS, carbon molecular sieve;

OMs,

oxometalates;

POMs,

polyoxometalates;

PSf,

polysulfone;

PAN,

polyacrylonitrile;

PVDF,

polyvinylidenefluoride; DOP, dopamine; PDA, polydopamine; LDH, layered double hydroxide; PI, polyimide; PBI, polybenzimidazole;

6FDA,

hexafluorisopropylidene-diphtatic

anhydride;

TR-PBO,

TR-polybenzoxazole;

PAHs,

poly(hydroxylamide)s; PAA, poly(acrylic acid); PEI, polyethylenimine; PMMA, poly(methyl methacrylate); PVP, poly(vinylpyrrolidone); SBI, sprobisindane; SBF, spirobifluorene; EA, ethanoanthracene; Trip, triptycene; TB, Troger’s base; PAF, porous aromatic framework; PTMSP, poly(trimethylsilylpropyne); PMP, poly(4-methyl-2-pentyne); PEO, poly(ethylene oxide); PEG, poly(ethylene glycol); PEGDA, poly(ethylene glycol) diacrylate; PEGA, poly(ethylene glycol) acrylate; PEGMEA, poly(ethylene glycol) methyl ether acrylate; PEGDMA, poly(ethylene glycol) dimethacrylate; PEO-PI, PEO-polyimide; PEO-PA, PEO-polyamide; PEO-PBT, PEO-poly(butylene terephthalate); PEO-PTT, PEO-poly(trimethylene terephthalate); T6T6T, tetra-amide; PEO-PSf, PEO-polysulfone; PEG-DME, polyethylene glycol dimethyl ether, GTA, glycerol triacetate; PEG-DBE, polyethylene glycol dibutylether; TMC, trimesoyl chloride; DGBAmE, diethylene glycol bis(3-aminopropyl) ether; DAmPEG, diaminopolyethylene glycol; RTILs, room temperature ionic liquids; PILs, poly(ionic liquid)s; TSILs, task-specific ionic liquids; PVAm, polyvinylamine; PAAm, polyallylamine; PAMAM, polyamidoamine; EDA, ethylenediamine; PIP, piperazine; MEA, monoethanolamine; MC, methylcarbamate; PANI, polyaniline; POSS, polyhedral oligomeric silsesquioxane; MWNTs, multi-walled carbon nanotubes; HT, hydrotalcite. 1

Corresponding author at: Chemical Engineering Research Center, School of Chemical Engineering

and Technology, Tianjin University, Weijin Road 92#, Nankai District, Tianjin 300072, PR China. Tel.: +86-22-27404533; fax: +86-22-27404496. E-mail address: [email protected] (Z. Wang). 1

alleviate the mounting global energy and environmental crisis. Therefore, an ever-increasing demand for high-quality hydrogen provides a strong driving force towards developing efficient hydrogen purification technologies. Membrane-based gas separation technology for hydrogen purification has attracted considerable attention owing to the inherent advantages over other conventional separation techniques. Benefited from the booming development of chemical science, materials science and membrane science, an increasing number of advanced membrane materials and membranes have been developed for hydrogen purification in recent years. This review primarily focuses on the latest developments in design and fabrication of H2-selective membranes and CO2-selective membranes for hydrogen purification, and the comparison of H2-selective membranes and CO2-selective membranes will be briefly discussed. In addition, future direction to further explore energy-efficient membranes for hydrogen purification will be presented for discussion. It is anticipated that the present review will provide the guidance for the future research and development of membrane materials and membranes for hydrogen purification, and hence promote the development of sustainable and clean hydrogen energy. Keywords:Gas separation, Hydrogen purification, H2-selective membranes, CO2-selective membranes

1. Introduction Sustainable and clean energy development has become a major global issue in terms of the world’s energy shortage and environmental problem. However, fossil fuels are still expected to be the predominant resource of energy by preference in the near term (next 5~20 2

years or even more) despite that an increasing number of renewable energies have received considerable attention over the several decades. Therefore, it is of great importance and impendency to develop more efficient ways to utilize these limited fossil fuels for sustainable development. Hydrogen has been widely considered to be an attractive energy carrier and storage medium with high efficiency for developing a cost-effective, environmental-benign and sustainable energy system, because it possesses distinct advantages of high gravimetric energy density (1.43×108 J/kg) and low greenhouse gas emission [1]. In addition, hydrogen is an important feedstock with increasing demands for the chemical industries. Hence, about 53 million metric tons of hydrogen worldwide was produced annually, and the hydrogen market valued at $88 billion in 2010 [2]. So far, hydrogen production is dominated by thermochemical processes, and about 96% hydrogen is generated from fossil fuels. In these hydrogen production processes, synthesis gas production is an intermediate step, and CO in synthesis gas could further react with water vapor via the water gas shift (WGS) reaction for enhancing H2 yield. The shifted synthesis gas mainly consists of H2 and CO2, along with some minor contaminants such as CO, H2S and CH4. Generally, H2 content in shifted synthesis gas varies from 60 vol% to 80 vol%, which depends on the quality of feedstock and process conditions. Moreover, biohydrogen production such as dark fermentation is a very promising alternative method to generate hydrogen, even if it contributes very limitedly to global hydrogen supply nowadays [3]. Similar to conventional thermochemical processes, the hydrogen product generated by biotechnological technique also contains some impurities (mainly CO2). The produced H2-rich 3

gaseous mixture via various aforementioned production techniques is considered as raw H2 product. However, it is difficult for raw H2 product to meet the demands for purity in most cases [4]. For example, high-purity hydrogen (> 99.99 vol%) supply is a prerequisite for the success of fuel cell technology. Therefore, hydrogen purification is essential to satisfy the purity requirements of various potential applications, and it is an important issue for efficient hydrogen supply. Moreover, H2-CO2 separation is a key process in pre-combustion CO2 capture for integrated gasification combined cycle (IGCC) power plants, even though the required hydrogen purity for subsequent electricity generation is generally lower than that of aforementioned applications [2]. It should be noted that hydrogen possesses a very low volumetric energy density despite its aforementioned advantages over conventional liquid fuels. Hence, hydrogen storage is vital for the widespread utilization of hydrogen as well as hydrogen purification [5]. As a relatively new and rapidly developing technology, membrane technology exhibits inherent advantages of energy-efficiency, cost-effective and environmental compatibility compared to conventional separation techniques. Moreover, membrane technology can be facilely coupled with other separation techniques to enhance the efficiency and economics of separation process. Nowadays, membrane technology has been widely used in water treatment, meanwhile it has also commercialized for air separation, natural gas sweetening and hydrogen recovery from ammonia purge gas [6]. With the rapid development of hydrogen economy and membrane science, membrane-based gas separation technology shows great potential for the hydrogen purification market as well. Great demands for high-quality hydrogen products 4

provide the driving force for research and development of advanced membrane materials and membranes for hydrogen purification. Benefited from the remarkable progress in materials science over the past several decades, some conventional membrane materials have exhibited notably improved performances via structural optimization. In addition, an increasing number of advanced materials, such as metal organic frameworks (MOFs), graphene-based materials, thermal rearranged (TR) polymers, polymers of intrinsic microporosity (PIMs), ionic liquids (ILs) and functionalized polymers, have been developed and subsequently used for fabrication of novel membranes for the purification of hydrogen produced by thermochemical and biotechnological approaches. For the separation of H2-CO2 system, membranes could be classified as H2-selective membranes and CO2-selective membrane according to the preferential permeation of gas species. The schematic diagrams of two types of membranes are illustrated in Fig. 1. Alternatively, based on the type of membrane materials, membranes could also be roughly divided as inorganic membranes, polymeric membranes and mixed matrix membranes.

5

Fig. 1. Schematic diagrams of (a) H2-selective membranes and (b) CO2-selective membranes.

In addition to the development of advanced membrane materials for H2-CO2 separation, membrane performance requirements for pre-combustion CO2 capture and hydrogen production have been extensively investigated via modeling study [7-10]. It should be noted that the predicted requirements for membrane performance are different in various studies, because actual demands are highly system dependent. Generally, H2-selective membranes are suitable for achieving high H2 recovery accompanied by moderate H2 purity, and CO2-selective membranes are favorable for high H2 purity combined with moderate H2 recovery. According to our recent modeling study using a two-stage membrane process, the demand (99.9 % H2 recovery & 95 % H2 purity) for power generation could be met when the 6

H2/CO2 selectivity reaches 11.8 at 3.0 MPa feed pressure, while the CO2/H2 selectivity of 108.5 is necessary. In contrast, the membrane with the CO2/H2 selectivity of 9.2 could achieve the target (99.99 % H2 purity & 90 % H2 recovery) set forth by the U.S. Department of Energy, while the selectivity of 368 is needed for H2-selective membranes [11]. Several review articles presented on membrane technology for hydrogen purification merely introduced H2-selective membranes or polymeric membranes. Furthermore, the majority of these reviews were involved in the research achievements before 2010 besides an in-depth review about the application of polymeric and ionic liquid based membranes for biohydrogen purification in 2013 [4, 12-15]. To the best of our knowledge, there is no comprehensive review available so far to focus specially on recent advances in membrane-based technology for hydrogen purification. Accordingly, this paper will give an overview on the current research in membrane materials and membranes for hydrogen purification. It should be mentioned that WGS membrane reactors were not discussed in this review, the detailed introduction about which could be found elsewhere [16, 17]. This review is arranged as follows. Section 2 gives the general principles of membrane gas separation technology. The recent advances in membranes for hydrogen purification including H2-selective membranes and CO2-selective membranes are described based on the type of membrane materials in section 3 and 4, respectively. The comparison and potential application of H2-selective membranes and CO2-selective membranes are briefly discussed in section 5. Last, our suggestions about the further development in this research field are presented for discussion in section 6. We sincerely wish that this review will give the inspiration to 7

researchers in the design and fabrication of high-performance membrane materials and membranes for hydrogen purification.

2. General background of gas separation membranes Simply, membrane is a selective barrier that allows the permeation of certain constituents and retains other constituents. For gas separation membranes, permeability and selectivity are two important parameters to evaluate membrane performance. The permeance (R) is usually applied to assess the permeability for composite membranes and asymmetric membranes, and it can be expressed as follows:

where Qi is the flux of gas species i at standard temperature and pressure (STP), △Pi is the transmembrane partial pressure difference of gas species i, and A is the effective membrane area. The unit of permeance is Gas Permeance Units (GPU), where 1 GPU = 10-6 cm3 (STP) cm-2 s-1 cmHg-1 = 3.35×10-10 mol m-2 s-1 Pa-1. While for homogeneous membranes, the permeation coefficient (P) is often used, the unit of which is Barrer, where 1 Barrer = 10-10 cm3 (STP) cm-1 s-1 cmHg-1. The permeation coefficient is the permeance normalized by the membrane thickness (l), and the relationship is illustrated in eq. (2).

Selectivity is used to quantify the permeability difference of two components. Ideal selectivity is defined as the permeability coefficient or permeance ratio of two pure gases shown in eq. (3).

8

where Pi, Ri, Pj and Rj are the permeability coefficient and permeance of gas species i and j in the membrane, respectively. Real selectivity, which is also known as separation factor, is calculated from the ratio of the composition of feed gas to permeant gas shown in eq. (4). ⁄ ⁄ where yi and yj are the molar fraction of gas species i and j in the permeate side, while xi and xj are the molar fraction of gas species i and j in the feed side. It should be noted that ideal selectivity is an intrinsic characteristic of a membrane material, while real selectivity is used to assess the ability to separate a particular gas mixture for a membrane [18]. When the pressure in the feed side is much larger than that in the permeate side, and the pressure in the permeate side approaches to zero, real selectivity will be in accordance with ideal selectivity. The prevalent transport mechanisms for gas separation membranes include (1) Poiseuille flow, (2) Knudsen diffusion, (3) molecular sieving, (4) capillary condensation, (5) surface diffusion, (6) solution-diffusion, and (7) facilitated transport. The schematic diagrams of these gas transport mechanisms are shown in Fig. 2. One and a combination of aforementioned gas transport mechanisms could be achieved based on various properties (e.g., size, shape, condensability and reactivity) of permeant gases, physical and chemical structures of membranes, as well as the interactions between permeant gases and membranes. The comparisons in various physicochemical properties between CO2 and H2 are shown in Table 1. If Poiseuille flow governs the mass transfer in membrane pores, the gas transport rate is independent of gas species. This phenomenon is unfavorable for gas separation membranes. Knudsen diffusion takes place when the ratio of pore radius to the mean path of gas molecules is smaller than one. The ideal 9

selectivity for H2/CO2 is only 4.7 when Knudsen diffusion accounts fully for gas transport, which could hardly achieve an effective separation. For molecular sieving mechanism, the smaller H2 molecule could permeate through the membrane while the diffusion of the larger CO2 molecule is blocked. Moreover, capillary condensation, surface diffusion and facilitated transport mechanisms could achieve CO2 preferential transport in most cases due to higher condensability and acidic nature of CO2. Either CO2 or H2 preferential permeation could be achieved based on solution-diffusion mechanism in dense membranes, when gas transport is dominated by solubility selectivity or diffusivity selectivity. In addition, it should be noted that a unique solution-diffusion mechanism could achieve high-efficiency H2 preferential permeation in dense metallic membranes, the detailed information about which could be found in section 3.1. Currently, the dominant mechanisms for H2-selective membranes are molecular sieving mechanism and solution-diffusion mechanism. For example, ProteusTM polymeric membrane based on solution-diffusion mechanism (determined by diffusivity selectivity) displays a promising performance for H2/CO2 separation in the pilot-scale trial. As for CO2-selective membranes, the predominant ones are facilitated transport mechanism and solution-diffusion mechanism. For instance, PolarisTM polymeric membrane based on solution-diffusion mechanism (determined by solubility selectivity) exhibits a favorable performance for CO2/H2 separation in the membrane demonstration system.

10

Fig. 2. Schematic diagrams of gas transport mechanisms in gas separation membranes.

Table 1. Distinction between CO2 and H2 in various properties relevant to separation [19] gas

molecular

kinetic diameter

critical temperature

polarizability

quadrupole moment

others

weight

(Å)

(K)

(10-25 cm-3)

(10-27 esu-1 cm-1)

CO2

44

3.30

304.2

29.1

43.0

acid

H2

2

2.89

33.3

8.04

6.62

neutral

3. Advances in H2-selective membranes Currently, most gas separation membranes reported for H2-CO2 separation are H2-selective membranes. This phenomenon is probably due to the fact that H2 molecule possesses the smallest kinetic diameter among gas molecules besides He and H2O molecules, thereby H2 molecule normally has the biggest diffusion coefficient, achieving its preferential permeation through membranes. In this section, H2-selective membranes are highlighted and

11

divided into five major classes: dense metallic membranes, microporous inorganic membranes, MOF membranes, polymeric membranes and mixed matrix membranes.

3.1 Dense metallic membranes Dense metallic membranes are the most common membranes for hydrogen purification, and the metals in Groups Ⅲ-Ⅴare all permeable to hydrogen. In particular, the membranes based on Palladium (Pd) have been intensively investigated, because they are highly permeable to hydrogen. Most previous reviews about membranes for hydrogen purification principally introduced dense metallic membranes [4, 16]. Moreover, several special reviews about Pd and Pd-alloy membranes have been published [20, 21]. Here a general introduction about dense metallic membranes was shown. Hydrogen permeation through dense metallic membranes obeys the unique solution-diffusion mechanism which mainly involves three steps: (1) the chemical dissolution of molecular hydrogen on the membrane surface of feed side to produce atomic hydrogen, (2) the diffusion of atomic hydrogen through the bulk metal, (3) the association of atomic hydrogen on the membrane surface of permeate side to reproduce molecular hydrogen. This gas transport mechanism is different from the widely known solution-diffusion mechanism, while it is somewhat similar to facilitated transport mechanism for CO2 preferential permeation in section 4.4. Because only hydrogen can transport through dense metallic membranes based on this mechanism, the membranes should theoretically have an infinite selectivity for hydrogen over any other gas species. Though dense metallic membranes possess an extremely high selectivity towards H2, the corresponding low gas permeance hampers their practical application. Development of composite membranes in 12

which a thin metallic layer is deposited on a porous polymeric or inorganic support, has been considered as an effective method to improve gas permeance of dense metallic membranes [22]. Currently, there are diverse dense metallic layer deposition technologies available, such as chemical vapor deposition, physical vapor deposition, electroplating and electroless plating. Among them, electroless plating is the most extensively used method owing to its cost efficiency and ability of covering supports with complex geometries [23]. It should be noted that the selectivity for H2 of metallic composite membranes is commonly measurable, which is different from dense metallic homogeneous membranes. It is because that a thin supported metallic layer is difficult to be strictly dense, which could result in a small number of other gas molecules permeation through the membranes. Nevertheless, the H2/gas selectivity of metallic composite membranes remains at a high level and it is generally higher than other styles of membranes. There are some serious limitations for application of pure Pd membrane. First, Pd membrane undergoes phase change when operated at a temperature below 571 K and a pressure below 20 bar, so called hydrogen-induced embrittlement phenomenon. Second, surface poisoning by contaminants, especially sulfur species, could result in an obvious decrease in H2 permeance. Last but not least, Pd is a kind of precious metal, and Pd membranes often have the limited lifetime of 2-3 years. The resultant substantial investment and operational costs are unfavorable for commercial application. To overcome these drawbacks above, Pd has been suggested to be alloyed with a series of other metallic elements such as Ag, Cu, Ni, Pt and Au [24, 25]. Influences of alloy content, membrane thickness, test 13

conditions and potential contaminants (CO and H2O) on performances of Pd-alloy membranes have been widely and intensively investigated via experimental study. Moreover, detailed density functional theory (DFT) and Monte Carlo techniques have been used to predict the hydrogen permeability and the resistance to poisoning of the membranes based on diverse alloys, giving the guideline for the design of Pd-alloy membranes [26, 27]. To sum up, the investigation of dense metallic membranes for hydrogen purification has been carried out over more than half a century. However, they have found very limited industrial application despite their excellent separation performances. Very high fabrication cost implies that dense metallic membranes might be only acceptable for small-scale applications such as the electronics industry where only a small amount of ultrahigh-purity hydrogen is required.

3.2 Microporous inorganic membranes Porous inorganic membranes that are suitable for light gas separation are generally microporous inorganic membranes with a pore diameter smaller than 2 nm and even less. Currently, microporous inorganic membranes mainly rely on the molecular sieving mechanism to achieve H2 preferential permeation. Therefore, the size and shape of pores are key elements to determine the separation performance of microporous inorganic membranes for hydrogen purification. H2 exhibits temperature-activated transport through microporous membranes regimented by molecular sieving mechanism, while CO2 shows the opposite trend. As a result, microporous inorganic membranes generally display a higher H2/CO2 selectivity at higher temperature, which indicates that they are suitable for high temperature application., 14

The H2/CO2 selectivity of microporous membranes is not competitive at low temperature, and even the phenomenon of CO2 preferential permeation may occur [28]. Regarding the type of membrane materials, microporous inorganic membrane can be classified into zeolite membranes, silica membranes and carbon-based membranes. Very recently, a comprehensive review article about porous inorganic membranes for CO2 capture has been published [29]. Here, there was a brief introduction about conventional microporous inorganic membranes, and the latest development of graphene-based membranes was shown in detail. 3.2.1 Zeolite membranes Zeolite membranes are the most widely investigated inorganic membranes due to their excellent mechanical, thermal and chemical stability. They are mainly derived from silicate, aluminophosphate (AIPO) and silicoaluminophosphate (SAPO) [30]. Considering the structures of zeolites, various zeolite membranes with different types such as MFI, DDR, LTA, CHA and FAU, have been developed. Currently, zeolite membranes are usually supported membranes on porous supports such as alumina and stainless steel to obtain a higher gas permeance and better mechanical stability [31]. In addition to conventional inorganic supports, a defect-free zeolite layer was recently successfully grown on porous polyethersulfone (PES) polymeric supports within about an hour [32]. The rapid crystal growth was owing to an extensive nucleation via manipulating the supersaturation during the crystallization. This pioneering work may provide an alternative method for reducing the fabrication cost of zeolite membranes. The predominant fabrication technologies for zeolite membranes are in-situ growth and secondary growth, both of which belong to the hydrothermal synthesis. In recent 15

years, other advanced synthesis techniques have also been attempted such as ionothermal synthesis, in which various ionic liquids could be selected to act as reaction media and structure-directing agent, adjusting the structure and property of zeolite. Moreover, ionothermal synthesis could proceed under ambient conditions, which is energy-saving and secure. In 2012, Li et al. for the first time reported the ionothermal synthesis method for fabricating different types of AIPO-based molecular sieving membranes on δ-Al2O3 substrates via substrate-surface conversion for gas separation [33]. In this process, the δ-Al2O3 substrates acted as both supports and Al source, which not only guaranteed the formation of continuous membrane but also strengthened the attachment between zeolite layers and substrates. The synthesized CHA-type membrane exhibited the H2 permeance of about 2090 GPU at 1 bar and 293 K, and the H2/CO2 ideal selectivity of 4.8 was slight higher than Knudsen selectivity. No matter which method was chosen, how to form a continuous defect-free zeolite layer on porous supports is generally a big challenge for supported zeolite membranes. Therefore, many strategies such as surface modification of supports, external force assisted synthesis and stepwise synthesis, have been suggested to improve the growth of zeolite layers. LTA-type zeolite membranes may be the most common zeolite membranes, and they are quite effective for dehydration of alcohol/water system. The first industrial application of zeolite membranes is the dehydration of bioethanol using LTA-type zeolite membranes. Theoretically, LTA-type zeolite membranes should be suitable for molecular sieving separation because they possess a small pore size of 4 Å (Na-LTA). Moreover, the pore size could be further optimized via cation exchange treatment such as about 3 Å for K-LTA, which 16

is suitable for H2/CO2 separation. Unfortunately, there were few reports about LTA-type zeolite membranes as high-efficiency molecular sieving membranes for gas separation, due to the existence of many inter-crystalline defects. The size of these inter-crystalline defects is generally 1-2 nm, which is large enough to allow fast transport of CO2 molecules. To address this

problem,

Huang

et

3-aminopropyltriethoxysilane

al.

used

a

series

(APTES),

of

covalent

1,4-diisocyanate

linkers (DIC-4)

such

as and

3-chlropropylyrimethoxysilane (CPTMS) between zeolite layers and porous supports, promoting the attachment of zeolite precursor and further the nucleation and growth (Fig. 3) [34-36]. All the resultant LTA-type zeolite membranes exhibited the enhanced H2/CO2 selectivity that exceeded the corresponding Knudsen selectivity (~4.7). Then a more simple method for surface modification of supports was suggested to coat a polycation and even with polyethylene glycol and subsequent calcination [37]. An improved layer continuity and crystal intergrowth were achieved owing to the formation of monolayer by carbon-oxygen species.

Fig. 3. Scheme of the synthesis of zeolite LTA membranes on covalently functionalized supports by using 1,4-diisocyanate (DIC-4) as molecular binders to in situ anchor LTA nutrients during hydrothermal synthesis [35].

17

Some multi-layer zeolite membranes have also been developed via stepwise synthesis to further improve separation property [38, 39]. The double-layer FAU-LTA zeolite membranes and triple-layer LTA-type membranes were fabricated by the same group (Fig. 4). In both cases, APTES acted as a protective barrier to prevent the dissolution of the as-synthesized zeolite layer during the following hydrothermal process, in addition to being a covalent linker between zeolite layers. The results showed that the H2/CO2 ideal selectivities of single-layer, double-layer and triple-layer LTA-type zeolite membranes were 7.0, 17.1 and 19.3, respectively. The gradually enhanced separation performance was probably attributed to that the upper zeolite layer could seal the defects of underlying zeolite layer.

Fig. 4. (a) Scheme of the layer-by-layer hydrothermal synthesis of sandwich-like multi-layer LTA membranes by using 3-aminopropyltriethoxysilane (APTES) as self-assembled interlayer. (b) Single gas permeances of different gases through the single/multi-layer zeolite LTA membranes at 373 K as a function of the gas kinetic diameter [38]. 18

MFI-type zeolite membranes are another kind of common zeolite membranes, and they have been widely used for the separation of xylene isomer system. However, it is difficult for pristine MFI-type zeolites to develop high-performance molecular sieving membranes for H2/CO2 separation because of their relatively large pore size of 5.6 Å. Some measures have been taken to reduce the pore size of MFI-type zeolites to enhance the separation performance. Catalytic cracking deposition (CCD) of methyldiethoxysilane (MDES) was the most widely adopted method to narrow the pore size, and consequently improve H2/CO2 selectivity of MFI-type zeolite membranes. Briefly, the formed SiO2 units through adsorption and subsequent calcination process were deposited on the pore walls for pore size reduction. However, conventional CCD modification often resulted in a selectivity improvement at the expense of more than one order of magnitude decrease in H2 permeance, which was due to excessive SiO2 deposition over the entire membrane. Recently, many strategies have been adopted to address this problem. Tang et al. reported the controlled partial modification of pores in MFI-type zeolite membranes [40]. Following this method, CCD modification of the MFI-type zeolite with high SI/Al ratio was done twice with 10 h annealing period in between, and the remaining chemisorbed organosilyl species in the pore decomposed during the annealing process, which resulted in the increase in membrane porosity. Therefore, the as-modified membrane displayed a high H2/CO2 ideal selectivity of 141 combined with a high H2 permeance of 1182 GPU at 1 bar and 723 K. Moreover, Wang et al. prepared a thin ZSM-5 layer on the thick silicate layer before CCD modification, thus CCD modification only occurred in the ZSM-5 layer instead of the entire membrane [41]. The resultant H2/CO2 ideal 19

selectivity was 25.3 (over five times improvement) accompanied by only 31% reduction in H2 permeance at 723 K. Then the same group fabricated bilayer MFI zeolite membranes consisting of a 2 μm thick ZSM-5 layer (as CCD modified) on an 8 μm thick silicalite on Al2O3 supports with an YSZ intermediate layer via a multiple step synthesis method [42]. The bilayer membrane exhibited a H2 permeance of 358 GPU with the H2/CO2 selectivity of about 23 at 1 bar and 773 K. Hong et al. developed a novel method that on-stream CCD was used to modify MFI hollow fiber membranes when membrane performance was measured at high temperature [43]. In this case, the separation performance could be tested in real time to control modification time. The as-modified zeolite membrane showed a H2 permeance of 30 GPU with the H2/CO2 selectivity of 45.6 at 773 K. The quality of original unmodified MFI-type zeolite membranes had a great influence on separation performances after CCD modification besides CCD modification technology. Lin et al. pointed out a phenomenon that on-stream CCD modification of MDES was effective in enhancing H2/CO2 selectivity merely for the membranes with lower initial H2/CO2 selectivity at room temperature, which seemed a little counterintuitive [44]. In fact, lower initial H2/CO2 selectivity indicated that there were limited inter-crystalline defects in the zeolite membranes and H2 permeation was blocked by adsorbed CO2 to a certain extent, while relatively higher initial H2/CO2 selectivity implied that gas molecules transport through defects followed the Knudsen diffusion. Zeolites generally have anisotropic pore structure with pores of various sizes or structures in different orientation. In most cases, oriented zeolite membranes possess more excellent permeation or separation performances than randomly oriented zeolite membranes [31]. For 20

example, Huang et al. applied secondary growth method to develop a highly oriented, neutral and cation-free AIPO4 LTA-type zeolite membrane via assembling a highly oriented zeolite seeds monolayer [45]. The H2/CO2 ideal selectivity of resultant oriented membrane was 10.9, which was higher than randomly oriented one (~7.6). Therefore, there were great efforts have been made to develop high–quality zeolite membranes with oriented structures. Pham et al. creatively develop a facile method for the growth of uniformly oriented MFI-type zeolite films using oriented seed monolayer and synthesis gel with adjusted composition and temperature (Fig. 5) [46]. In this method, there was no new MFI crystal formed in the precursor gel during the growth process, which may guarantee the uniform oriented growth. However, the gel preparation step was inevitable, which was extremely time-consuming and chemical-wasting. Therefore, the same group reported a “gel-free secondary growth” approach to fabricate oriented zeolite membranes, and this approach was more economic and environmental friendly [47]. As a result, the as-prepared oriented membrane exhibited a remarkable permeance and selectivity for xylene isomer system. Moreover, rapid thermal processing was adopted to replace conventional slow calcination process for fabrication of oriented membranes, which not only could save time and energy but also could reduce inter-crystalline defects [48]. Das et al. reported the fabrication of highly oriented SAPO-34 zeolite membranes on silica modified clay-Al2O3 supports via selective deposition of oriented seed crystals [49]. Compared to randomly oriented SAPO-34 ones, the synthesized membranes displayed a higher H2/CO2 selectivity of 16.66 with the corresponding H2 permeance of about 26900 GPU at 1 bar and room temperature. In order to fully achieve the benefits of oriented zeolite membranes, the 21

synthesized zeolite layer should be as thin as possible. Moreover, the study of Pham et al. indicated that oriented membranes with thinner thickness were more favorable for high permeance and selectivity for xylene separation [47]. Unfortunately, there are limited studies about the separation performance of current oriented zeolite membranes for gas separation, which is probably due to the fact that unmodified pore size is too large for small gas molecules. It is expected that a highly oriented membrane will exhibit an excellent performance for H2/CO2 separation via tailoring and optimizing pore size based on the aforementioned modification methods.

Fig. 5. Schematic illustrations of (A) leaflet-shaped and (B) coffin-shaped silicalite-1 (SL) crystals and (C) truncated bipyramidal beta (BEA) crystals and their channel systems, as well as their respective (D) a-oriented, (E) b-oriented, and (F) a oriented monolayers. (G to I) Secondary growth on these monolayers produces uniformly oriented films [46]. 22

3.2.2 Silica membranes Microporous silica membranes are another kind of common porous inorganic membranes. Different from zeolite membranes, silica membranes are generally amorphous, and they are more easily fabricated into an ultra-microporous thin layer that exhibits an excellent molecular sieving property. Duke et al. suggested that the pore size of amorphous silica network originated from the most conventional precursor tetraethylorthosilicate (TEOS) was about 3.0 Å based on Position Annihilation Lifetime Spectroscopy (PALS) study [50]. Therefore, silica membranes have been widely investigated for hydrogen purification, some of which exhibited superior separation performances at high temperature. For instance, Gu et al. developed a ultrathin (20~30 nm) silica layer on an intermediate multilayer Al2O3 support with a graded structure [51]. The resultant silica composite membrane exhibited a high H2/CO2 ideal selectivity of 1500 with the H2 permeance of about 1500 GPU at 1 bar and 873 K. The conventional fabrication methods for microporous silica membranes are sol-gel method and chemical vapor deposition method. The sol-gel method is the oldest method for porous inorganic membranes, which usually includes dip-coating of silica precursors on a porous support, followed by controlled drying and firing at elevated temperature. The chemical vapor deposition method consists of thermal deposition of precursors, followed by chemical reaction with an oxidizing agent. Generally, the membrane derived from the sol-gel method yields a relatively higher gas permeance with lower gas selectivity owing to more open structures, while the membrane based on the chemical vapor deposition method exhibits a higher selectivity but lower gas permeance. The requirements of industrial application to 23

maximum production could be likely favorable for high permeance than high selectivity, though it will be dependent on product specification. Therefore, the development drive has caused sol-gel derived silica membranes with higher selectivity while chemical vapor deposition derived membranes with higher permeance. Effects of silica precursors, technological parameters and support properties on membrane separation performances have been extensively investigated [52-54]. Although many lab-scale microporous silica membrane have shown excellent separation performances, the most serious problem that limits large-scale application of microporous silica membranes is the lack of stability at high temperature, especially in the presence of steam. This is mainly due to the condensation of silanol groups in the silica layer catalyzed by water, being accelerated at high temperature and under humid conditions. This phenomenon could not only result in an obvious decrease in gas permeance but also lead to the embrittlement of silica layer. Therefore, considerable efforts have been made to improve the hydrothermal stability of silica membranes. One approach is to incorporate hydrophobic organic groups into the silica structure [55, 56]. Hydrophobic silica membranes could possess improved hydrothermal stability to a certain extent. However, there still exists the potential problem of hydrothermal stability above decomposition temperature of organic groups. Another modification strategy is the introduction of inorganic additives including metal and metal oxide nanoparticles into the silica matrix [57-61]. The improved hydrothermal stability was attributed to that incorporated nanoparticles might reduce the thermal-induced molecular motion of microporous silica networks at elevated temperature. In addition to improving 24

hydrothermal stability, these metal-based additives could potentially enhance the separation performances of silica membranes at high temperature owing to reversible hydrogen adsorption properties of metal and metal oxide nanoparticles [62-64]. It should be noted that the selectivity in mixed gas test was much smaller than that in pure gas test for majority microporous silica membranes, which was mainly due to the strong adsorption of CO2 in pores that partially blocked H2 transport [63]. However, the reported performances of most silica membranes were based on single gas permeation test. Pure gas measurement is a necessary step to study gas transport property in the new material screening process but must be followed by gas mixture measurements under realistic conditions, which could provide engineering parameters to predict the separation for industrial-scale system [60]. 3.2.3 Carbon-based membranes The most conventional carbon-based membranes are carbon molecular sieve (CMS) membranes with amorphous microporous structures, the study of which has a history of more than half a century. Generally, CMS membranes are fabricated under an inert atmosphere or vacuum via the carbonization or pyrolysis of various polymeric precursors such as poly(furfuryl alcohol), polyacrylonitrile, phenolic resins,poly(vinylidene chloride-co-vinyl chloride), polyimide and its derivatives. Among them, polyimide and its derivatives are the most common carbon precursors owing to the structural variability via the combination of different diamines and dianhydrides. The carbonization behaviors of a series of aromatic polyimides, the relationship between structure and property of synthesized carbon materials have been summarized [65]. The pore system of CMS membranes is created with the 25

degradation of polymers, which commonly consists of wide openings with relatively narrow constrictions. The larger pores (0.6-2.0 nm) are responsible for adsorption capacity, while the smaller pores (< 0.6 nm) are responsible for molecular sieving properties [66]. This structural characteristic could enable high permeability and high selectivity simultaneously. Many studies have intensively investigated and demonstrated that polymeric precursors’ properties, pretreatment, pyrolysis conditions (pyrolysis temperature, soak time, pyrolysis atmosphere) and post-treatment primarily influenced the formation of micropore structures and then gas transport properties, finally determined gas separation performances of the synthesized CMS membranes [67, 68]. Currently, CMS membranes can be mainly grouped into two categories: one is supported membranes with the relatively high gas permeance and good mechanical strength, the other is unsupported membranes such as carbon hollow fiber membranes. For supported CMS membranes, the properties of supports such as surface roughness and pore structures play an important role in the formation of a defect-free thin CMS layer. Several coating-pyrolysis cycles are often required to avoid the defects, which could result in the decrease in gas permeance. Therefore, a mesoporous intermediate layer is commonly necessary between macroporous supports and microporous CMS layer to guarantee the formation of a thin coating without pinholes [69]. Moreover, a plasma-enhanced chemical vapor deposition method was used to form an ultrathin defect-free coating instead of conventional spin-coating process which was beneficial for reproducibility improvement [70]. Being a typical example of unsupported CMS membranes, carbon hollow fiber membranes have received considerable attention in recent years due to the high packing density. Several 26

newly-developed CMS hollow fiber membranes exhibited superior gas separation performances, and the corresponding modules have also been successfully fabricated [71]. However, the gas permeance through CMS hollow fiber membranes was usually lower than expected when translating from dense flat film to hollow fiber, which was mainly due to the densification of microporous structure morphology. Some treatments have been suggested to restrict the morphology collapse [72]. CMS membranes have been widely studied for application in air separation (O2/N2), natural gas purification (CO2/CH4 and N2/CH4), CO2 capture (CO2/N2) and hydrogen recovery (H2/N2 and H2/CH4) [67, 73, 74]. However, few studies focused on H2/CO2 separation. According to pure gas data of different CMS membranes reported, the H2/CO2 ideal selectivity of most CMS membranes was not more than 20, and it is expected that the selectivity for H2/CO2 gas mixtures might be lower. Moreover, one major disadvantage that limits practical application of CMS membranes is their brittleness, even though it could be restricted to a certain degree via optimizing precursor structures and technological parameters for pyrolysis process. Therefore, the mechanical strength of CMS membrane under high pressure conditions should be enhanced to meet the demand for practical application. In addition to traditional carbon-based membranes, graphene-based membranes, a new kind of carbon-based membrane, have been developed in recent years. In the field of separation science, graphene-based materials such as graphene and graphene oxide have attracted considerable attentions in water desalinization, ion-selective transport and gas separation over the past five years [75-78]. Very recently, a comprehensive review about 27

graphene-based membranes and their implications in molecular separation has been published [79]. Theoretically, graphene-based materials are a class of excellent candidates for fabricating separation membranes owing to their inherent characteristics of single-atom thickness, great mechanical strength and good chemical stability [80-82]. However, the study has demonstrated that perfect graphene is impermeable to all gases and liquids [83]. Therefore, it is necessary to drill pores with suitable size and shape to achieve selective gas permeation. Over the past years, many physical and chemical methods such as laser irradiation, helium ion bombardment, electron beam irradiation and steam etching, have been used to generate pores on graphene sheets [84-87]. Generally, the pore size of porous graphene prepared via physical methods could be tuned by changing the ion or electron doses, while porous graphene through chemical methods possess a relatively narrow size distribution [88]. Very recently, a general and scalable fabrication approach for porous graphene was developed via the carbonization reaction of graphene oxide etched by metal oxide particles generated from oxometalates (OMs) and polyoxometalates (POMs) [89]. In this method, the pore size could be tuned within the range of 1 nm to 50 nm via controlling the size of metal oxide particles. Similarly, nitrogen-doped porous graphene could also be prepared using ammonium group-containing OMs and POMs. Furthermore, Celebi et al. engineered large-scale porous graphene membranes with high porosity and narrow pore size distribution [90]. These features imply that porous graphene-based membranes could exhibit an ultimate permeation property of gas and liquid. Unfortunately, the resultant pore size based on these two scale-up methods may be too large to be applicable for small gas separation. In the viewpoint of high selectivity for 28

H2/CO2 separation, the pore size should be accurately controlled within the angstrom scale. Over the past five years, many computational studies have suggested that porous graphene-based membranes with desirable pore geometry could achieve an excellent permeance and selectivity [75, 91-93]. For instance, Jiang et al. predicted that both the nitrogen-functionalized pore (3.0 Å × 3.8 Å) and the all-hydrogen passivated pore (2.5 Å × 3.8 Å) on graphene sheets could achieve extremely high H2/CH4 selectivity based on first principles density functional theory calculations [75]. Then Yao et al. investigated the separation performances of porous graphene membranes with different pore sizes via first principles density functional theory calculations and molecular dynamic simulations [93]. An appropriate pore size of 3.2375 Å was suggested to be suitable for hydrogen purification. Unfortunately, there was no experimental data to verify the aforementioned simulation results. In 2012, Koenig et al. for the first time demonstrated the molecular sieving effects of porous graphene membranes via experimental study [84]. In this study, the micrometer-scale membranes with sub-nanometer-sized pores were created via ultraviolet oxidative etching method, and the gas leak experiments indicated that the gas molecules with the kinetic diameter smaller than 3.4 Å were much more permeable than those with the kinetic diameter larger than 3.4 Å. Consequently, the H2/CO2 ideal selectivity was about 1.7, which was likely due to the larger pore size than CO2 and H2 molecules. Besides, very limited experimental data have been reported for porous graphene-based membranes for gas separation, which could probably arise from an enormous challenge of reproductively developing high-quality porous graphene-based membranes possessing high porosity and uniform pore size within the 29

sub-nanometer scale [86]. Intrinsic structural defects are generally inevitable for macroscopic graphene membranes. Therefore, it is vital to understand gas transport property through these defects to achieve the full potential of graphene-based membranes for separation. Boutilier et al. quantified the effects of intrinsic defects on the performance of graphene membranes deposited on a series of porous polycarbonate track-etched supports with various pore sizes [94]. When the pore diameter of the support was much smaller than the average distance between defects, it was possible to reduce and even eliminate the adverse effects of defects on separation performance via independently stacking several layers of graphene. Different from most membranes where defects could result in a serious deterioration of separation performances, intrinsic structural defects within graphene oxide flakes could serve as molecular-sieving transport channels. Li et al. developed the thin graphene oxide membranes on anodic aluminum oxide supports via vacuum filtration method, and the membrane thickness could be tuned by changing the concentration of graphene oxide dispersion solutions [95]. The extremely high H2/CO2 selectivity of 3400 with the H2 permeance of about 300 GPU could be obtained from 9-nm-thick graphene oxide membranes at 293 K. The author suggested that the predominant transport pathway for H2 was selective structural defects within graphene oxide flakes where CO2 could not permeate through, and the gas transport was negligible in interlayer spacing between graphene oxide flakes. However, an opposite conclusion that gas permeated through not only defects but also spacing between graphene interlayers was speculated by Kim et al. [96]. In addition, he prepared few-layered graphene oxide membranes on PES supports via 30

two different coating methods. The first method was to contact the support surface to the air-liquid interface of graphene oxide solution, followed by spin-coating. The second one was direct spin-casting of graphene oxide solution on the support surface. The resultant two membranes showed different gas transport behaviors. The membrane prepared using the first method exhibited a H2/CO2 selectivity of 30 at 1 bar and 308 K under dry state. However, it became CO2-selective when the humid feed gas was used. This phenomenon was probably owing to the high CO2 solubility in water that improved CO2 permeance by 50 times. Therefore, the membrane displayed the CO2/H2 selectivity of about 5 with the CO2 permeance of about 50 GPU at 1 bar and 308 K. The membrane prepared by the second method was CO2-selective membrane. It showed the CO2/H2 selectivity of about 10 with the calculated CO2 permeability of about 8500 barrer under humid conditions. To explain the significant differences of two membranes, the author suggested that it was due to different stacked graphene oxide structures: the first one was less-interlocked, and there existed nano-scale pores by the edges of non-interlocked graphene oxide sheets, while the second one was highly interlocked and closed-packed. Unfortunately, gas transport mechanisms in these two membranes are still unclear and confusing. It should be noted that the performances of both membranes could be improved with the aid of water, and the similar phenomenon was also observed in other graphene oxide membranes [97]. However, the role of water was not intensively studied. Recently, Kim et al. investigated the effects of water on the intercalation of gas molecules on graphene oxide interlayers [98]. The results revealed that the degree of intercalation was strongly influenced by the interaction between gas molecules and 31

hydrophilic surface of graphene oxide, and only CO2 could be intercalated in the dried interlayer of graphene oxide. When interlayers were water-swollen, the amount of intercalated CO2 could be remarkably enhanced owing to retarded dynamic of intercalated water. Furthermore, the storage amount of CO2 could obviously increase with increasing injected feed pressure. This finding indicates that water-assisted graphene oxide membranes could possess a higher CO2 storage capacity, namely an improved CO2 solubility. A great progress in graphene-based membranes for gas separation has been made over the past years. However, gas transport mechanism through porous graphene-based membranes remains elusive. Several researchers have attempted to explore gas transport mechanism using different simulation methods [92, 99-101]. The gas transport mechanisms of conventional gas separation membranes may be not suitable for porous graphene membranes with single-atom thickness, while the widely accepted viewpoint is that the permeation of gas molecules through porous graphene membranes is closely related to both transport rate to the surface and molecular adsorption on the graphene sheet surface, as well as the size and functionalization of pores. Till now, there is still no perfect theory to explain gas transport properties in graphene-based membranes, which may result in the deficiency of the guideline to design high-performance graphene-based membranes. In all, fabrication of high-quality porous graphene, generation of highly selective defects, and exploration of gas transport mechanism in graphene-based membranes are the key issues for development of graphene-based membranes for gas separation.

32

3.3 MOF membranes Metal organic frameworks (MOFs) have attracted considerable attention in fields such as materials science, chemistry and chemical engineering, and the research field of MOFs has exhibited tremendous growth over the past two decades [102]. MOFs are generally a new class of organic-inorganic hybrid porous solid materials with geometrically and crystallographically well-defined structures, which are constructed from metal ions or metal ion clusters that are connected by organic linkers. Compared to aforementioned traditional porous materials used for fabrication of microporous inorganic membranes, MOFs possess inherent advantages of structure variability, ultrahigh porosity, uniform but tunable pore size and adjustable internal surface properties. Thanks to these unique features, MOFs shows great potential in gas storage and separation, especially for H2 storage and CO2 capture [19, 103, 104]. Early investigations on MOFs concentrated more on the fabrication of gas adsorbents. In 2009, the Lai’s group prepared a MOF-5 membrane on the porous Al2O3 support which was the first reported MOF membrane for gas separation [105]. After that, MOFs have been widely suggested to be assembled into MOF membranes for gas separation during the past five years. Very recently, several excellent comprehensive reviews focused on MOF membranes for gas and liquid separation have been published [106, 107]. It should be mentioned that MOF membranes refer to supported membranes here, while self-supported MOF membranes, so called MOF films are not included in this review. From the viewpoint of large-scale practical application that pursues low cost and high throughput, MOF membranes may be a more suitable candidate compared to MOF films, even though the fabrication of defect-free MOF membranes is 33

generally more challenging. Many innovative synthetic methods have been explored for the fabrication of MOF membranes such as in situ (direct)growth, secondary growth, layer-by-layer growth, liquid phase epitaxy and rapid thermal deposition. Among them, the former two technologies are widely used and relatively mature. It should be noted that the study of MOF membranes resembles zeolite membranes in many aspects. 3.3.1 In situ growth In situ growth means that the support is directly immersed in the growth solution without any crystals attached to the surface previously. In this method, nucleation and growth occur simultaneously. The aforementioned first reported MOF membrane for gas separation was fabricated via this method. Bux et al. directly synthesized the ZIF-8 layer on a porous TiO2 support by a microwave-assisted solvothermal method, and the membrane showed a H2 permeance of 180 GPU with the H2/CO2 ideal selectivity of 4.5 at 1 bar and 298 K, lower than Knudsen selectivity of 4.7 [108]. In theory, a perfect ZIF-8 layer should have an excellent molecular sieving property owing to its narrow pore size of 3.4 Å. However, the poor selectivity indicated that molecular sieving was scarcely observed, which implied that the synthesized MOF layer was not well-intergrown and there existed a few defects in ZIF-8 crystals. Indeed, it is a significant challenge to develop a continuous and defect-free MOF membrane on native porous inorganic supports using in situ growth method, which is due to the fact that the heterogeneous nucleation of MOF crystals on the support is not efficient. Kang et al. developed a novel method to prepare MOF membranes on a nickel screen [109]. 34

The nickel screen was not only as the support but also as the only metal source for the coordination reaction system, which guaranteed the formation of a defect-free MOF layer. Moreover, the growth process in this method was self-limiting, making the final membrane very thin and continuous. In addition to conventional inorganic supports, commercial polymeric porous supports, such as polysulfone (PSf), PES, polyacrlonitrie (PAN) and polyvinylidenefluoride (PVDF) ultrafiltration flat sheet or hollow fiber membranes have been used as supports for fabrication of MOF membranes by several research groups in recent years [110-113]. Cacho-Bailo et al. and Nagaraju et al. demonstrated the synthesis of ZIF-8 membranes on flexible PSf porous supports by in situ growth [110, 111]. In particular, Brown et al. described a sophisticated route to fabricate ZIF-8 membranes in situ on porous hollow fiber membrane modules via interfacial microfluidic membrane processing method [113]. This work demonstrates the feasibility for large-scale fabrication of MOF membranes. The successful growth of MOF on low-cost polymeric supports exhibits great potential for the commercialization of MOF membranes in the near future. In order to to improve the heterogeneous nucleation on various supports and strengthen the binding between porous supports and polycrystalline MOF layers, surface modification of supports was considered as an effective strategy. Caro and Huang have made significant advances in this field. A series of covalent linkers, such as APTES and dopamine (DOP), were used to promote the nucleation and growth of ZIFs including ZIF-90, ZIF-95, ZIF-22 and ZIF-8 on porous supports [114-117]. For example, Huang et al. developed a highly permeable 35

ZIF-95 membrane on APTES-modified α-Al2O3 supports [115]. The membrane exhibited a H2 permeance of about 5800 GPU with the H2/CO2 selectivity of 25.7 at 1 bar and 598 K. The high selectivity was because that a large number of CO2 molecules were trapped in huge cavities (~2.4 nm) of ZIF-95. Moreover, inspired by bioadhesion property of mussels, polydopamine (PDA) was for the first time suggested to deposit on porous Al2O3 supports via oxidative polymerization of DOP, and it could anchor the ZIF nutrients on the support surface owing to the covalent reaction and non-covalent adsorption ability [116]. This membrane showed the H2 permeance of 573 GPU with the H2/CO2 selectivity of 8.9 at 1 bar and 423 K, as well as high thermal stability. Most recently, the same group further prepared bicontinuous ZIF-8@Graphene oxide membranes on PDA-modified supports via layer-by-layer deposition, which was based on the treatment of ZIF-8 layer by a graphene oxide solution (Fig. 6) [118]. In this case, the gaps between the ZIF-8 crystals could be completely filled with graphene oxide via covalent bonds and capillary forces, eliminating the non-selective transport zones. As a result, molecular sieving pores of ZIF-8 crystals were the unique gas transport pathways, thus the membrane exhibited the enhanced H2/CO2 selectivity of 14.9 with the H2 permeance of about 390 GPU at 1 bar and 523 K. In addition, ZIF-8 membranes were developed on PDA-functionalized macroporous stainless steel nets to improve H2 permeance instead of mesoporous Al2O3 supports [119]. Compared to ZIF-8/Al2O3 membranes, they exhibited a remarkably high H2 permeance of about 71000 GPU with the H2/CO2 selectivity of 8.1 at 1 bar and 373 K, owing to lower gas transport resistance through stainless steel nets than Al2O3 supports. 36

Fig. 6. Scheme of preparation of bicontinuous ZIF-8@GO membranes through layer-by-layer deposition of graphene oxide on the semicontinuous ZIF-8 layer which was synthesized on a polydopamine-modified alumina disk [118].

In addition to organic covalent linkers, another protocol was to introduce an inorganic buffer layer between supports and MOF layers. Very recently, Liu et al. found that well-intergrown ZIF-8 membranes could be directly prepared on the ZnAl-CO3 layered double hydroxide (LDH) buffer layer-modified Al2O3 supports, because Zn2+ ions from LDHs could provide active sites for the nucleation and growth of ZIF-8 crystals (Fig. 7) [120]. Zhang et al. introduced vertically aligned, single crystal ZnO nanorods as a buffer layer for fabrication of ZIF-8 membranes [121]. The activated nanorods could initiate the uniform nucleation of ZIF nuclei and further facilitate the growth of continuous ZIF-8 membranes. The resultant membrane displayed a H2 permeance of 472 GPU with the H2/CO2 selectivity of 4.6 at 1 bar and 303 K. Furthermore, Zhang et al. synthesized a non-activation ZnO array as a buffer layer on PVDF hollow fiber supports for ZIF-7 membranes [122]. The results indicated that the ZIF-7 membrane exhibited a significantly high H2 permeance of about 7000 GPU with the H2/CO2 ideal selectivity of 18.43 at 1 bar and 298 K. 37

Fig. 7. Schematic illustration of in-situ solvothermal growth of ZIF-8 membrane on a ZnAl-LDH buffer layer-modified γ-Al2O3 substrate [120].

Considering the aforementioned strategies including surface modification of supports and introduction of buffer layers, the process of preparing MOF membranes often requires many additional steps. Recently, a simple in situ growth method to fabricate MOF membranes based on counter diffusion concept was developed (Fig. 8) [123]. Kwon et al. prepared well-intergrown ZIF-8 membrane via this method. The porous support was first immersed into a metal source solution, and then subjected to growth in an organic linker solution. The metal ion and organic linker diffused along concentration gradient and further contacted near the support surface. This process is quite similar to interfacial polymerization that has been widely used for fabrication of reverse osmosis membranes, nanofiltration membranes and gas separation membranes. Following this strategy, Xie et al. developed a ‘two-in-one” method to construct a defect-free thin ZIF-8 membrane [124]. In this method, APTES-modified Al2O3 particles were deposited on a macroporous support by hot dip-coating method to reduce the 38

pore size of supports and promote a high density of heterogeneous nucleation sites. The thickness of the synthesized ZIF-8 layer was merely 2 μm, thus the resultant membrane displayed an unprecedented H2 permeance of about 171,000 GPU with the H2/CO2 ideal selectivity of 17 at room temperature.

Fig. 8. Schematic illustration of membrane synthesis using the counter-diffusion-based in situ method [123].

3.3.2 Secondary growth There are serious limitations for selection of support materials using in situ growth method. Furthermore, additional surface modification processes are generally indispensable to facilitate the nucleation of MOF crystals, which inevitably increases the complexity of fabrication technology. In contrast, it has been demonstrated that secondary growth is more convenient to obtain a continuous MOF membrane without defects. Moreover, it allows better microstructural control of the synthesized membranes. Therefore, relatively more efforts are having been made to accelerate the development of this technology. The nucleation and the growth steps are decoupled in secondary growth method. Seeding procedure is critical for fabrication of continuous defect-free MOF membranes. In addition, it could determine the 39

binding strength between MOF crystal layers and supports. There are various seeding technologies have been developed such as microwave-assisted seeding, reactive seeding, thermal seeding, step-by-step seeding and surface modification of porous supports for seeding. Yoo et al. synthesized crystals which could strongly attach on the support with the aid of microwave irradiation in a couple of minutes [125]. High microwave energy could rapidly increase the temperature near the support surface (reaction zone), resulting in the rapid nucleation of MOF crystals. Following this method, Li et al. prepared a highly stable ZIF-7 molecular sieve membrane [126]. The H2/CO2 selectivity was up to 13.6 at 1 bar and 493 K because of the small pore size of 3.0 Å. Moreover, Hu et al. introduced a facile reactive seeding method to develop a continuous MIL-53 membrane on Al2O3 supports (Fig. 9) [127]. In this method, Al2O3 support acted as the metal source to react with organic linker to form a seeding layer. The gas molecules permeated through the resultant membrane based on Knudsen diffusion, which was probably due to the larger pore size of MIL-53 (7.3 Å × 7.7 Å). Furthermore, the suggested growth mechanism of MOF membranes via reactive seeding method was that the Al2O3 support surface was first transformed to AlO(OH) via reacting with H2O, and then AlO(OH) interacted with organic linkers to generate MOF crystals [128].

40

Fig. 9. Schematic diagram of preparation of the MIL-53 membrane on alumina support via the reactive seeding method [127].

Moreover, a step-by-step seeding procedure was for the first time adopted to fabricate a seeding layer by the same group (Fig. 10) [129]. This process resembles the layer-by-layer assembly using polyanions and polycations. Lee et al. prepared a Ni-MOF-74 membrane via the step-by-step seeding, and the membrane showed an extremely high H2 permeance of 37920 GPU with the H2/CO2 ideal selectivity of 9.1 at 1 bar and 298 K [130].

Fig. 10. Schematic diagram of step-by-step deposition of btc3- and Cu2+ on alumina support [129]. 41

3.3.3 Structure optimization during the synthesis and post-synthesis modification Even though a perfect MOF layer without defects could be fabricated, it is generally difficult to achieve high efficiency separation for H2 and CO2 gas mixtures based on original pore features of most MOF crystals, because there is a very limited size difference of two gas molecules. In order to improve the selectivity of MOF membranes, the design and modification of MOFs at multiple levels (atomic, groups, molecular and macroscopic level) is necessary. More specifically, the enhanced selectivity could be achieved via tailoring the pore size and shape as well as functionalizing the pore wall. Huang et al. developed a highly H2-selective ZIF-90 membrane by covalent post-modification using ethanolamine (Fig. 11) [131]. The H2/CO2 selectivity of the resultant membrane could reach 62.5 after complete imine reaction, because the imine functionality constricted the pore aperture and thus improved molecular sieving property. However, the significant increase in H2/CO2 selectivity was at the expense of the notable decrease in H2 permeance, and the resultant H2 permeance decreased to 45 GPU. To weaken the adverse effects on H2 permeance, the same group prepared a covalent-modified ZIF-90 membrane using organosilica APTES instead of ethanolamine as a modifier [132]. The modification only occurred on the ZIF-90 layer surface because the large APTES molecules could not enter the ZIF-90 layer. Consequently, there was nearly no variation of H2 permeance before and after modification, and the as-modified membrane possessed a relatively high H2 permeance of 842 GPU with the H2/CO2 selectivity of 20 at 1 bar and 498 K.

42

Fig. 11. Covalent post-functionalization of a ZIF-90 molecular sieve membrane by imine condensation with ethanolamine to enhance H2/CO2 selectivity [131].

Based on the aforementioned Ni2(L-asp)2(bipy) membrane, Kang et al. chose a pillar shorter organic ligand pyrazine (pz) instead of 4,4’-bipyridine (bipy) to prepare a new Ni2(L-asp)2(pz) (JUC-150) membrane on nickel screen via secondary growth [133]. The membrane exhibited a very high H2/CO2 selectivity of 38.7 with the H2 permeance of about 900 GPU at 1 bar and 298 K owing to its ultra-microporous narrow pore size of 2.5 Å×4.5 Å. Moreover, Zhang et al. fabricated an amino group-functionalized MOF (NH2-MIL-53(Al)) membrane using amino group containing organic ligand via colloidal assembly of MOF seeds method [134]. In the NH2-MIL-53(Al) membrane, an obviously stronger adsorption ability of CO2 resulted in a longer retention time compared to H2 because of the specific interaction between CO2 molecules and amino groups. Furthermore the adsorbed CO2 could hardly block H2 transport due to the large pore size (~7.5 Å) of NH2-MIL-53(Al). Therefore, this membrane 43

showed a high H2/CO2 selectivity of 30.9 with the H2 permeance of about 6000 GPU at 1 bar and 288 K. Very recently, Peng et al. innovatively developed a high-efficiency molecular sieving membrane by using MOF nanosheets as building blocks [135]. Owing to an ultrathin membrane thickness (~1 nm) and suppressed lamellar stacking of MOF nanosheets, the resultant membrane exhibited a remarkably high H2 permeance of 2700 GPU with the H2/CO2 selectivity of 291 at 1 bar and room temperature. This developed membrane resembles the aforementioned GO nanosheet membrane very much, however, MOF nanosheet possesses a much higher density of H2 transport channels than that of GO nanosheet, and consequently a higher gas permeance. Moreover, it should be noted that this membrane displayed good thermal and hydrothermal stability, suggesting great potential in hydrogen purification

3.4 H2-selective polymeric membranes Polymeric membranes are commercially more attractive with respect to low cost as well as the ease of manufacture and processability. Currently, the predominant commercial membranes are polymeric membranes even though some inorganic membranes have exhibited superior performances [136, 137]. Conventional polymeric membranes for gas separation are mainly dense membrane, and gas molecules permeate through these membranes following the solution-diffusion mechanism. In this case, the permeability coefficient (P) is determined by the diffusion coefficient (D) and the solubility coefficient (S) of gas in membranes, and the relationship among these parameters is based on solution-diffusion model shown in eq. (5). (5) 44

where D is a kinetic parameter which generally corresponds to the size of gas molecule, and S is a thermodynamic parameter which is closely related to the condensability of permeating gas and the interaction between permeating gas and membrane. For a binary gas mixture, the ideal selectivity is given as the product of diffusivity selectivity and solubility selectivity shown in eq. (6).

where Di, Si, Dj and Sj are the diffusivity coefficient and solubility coefficient of gas species i and j in the membrane. Compared with the properties of H2 and CO2 indicated in Table 1, H2 is kinetically more favorable due to the smaller kinetic diameter, while CO2 is thermodynamically more favorable owing to the higher condensability [138]. These competing effects result in the unprecedented difficulties in developing highly selective dense polymeric membranes based on the solution-diffusion mechanism for H2-CO2 separation, regardless of H2-selective membranes or CO2-selective membranes. For H2-selective polymeric membranes, the combination of enhancing the diffusivity selectivity and weakening the solubility selectivity is an effective strategy to obtain high H2/CO2 selectivity. Glassy polymers are a class of outstanding membrane materials for H2/CO2 separation, because they commonly possess a rigid structure and narrow free volume distribution. Polymeric membranes with these structural characteristics somewhat resemble molecular sieving inorganic membranes. Currently, polyimide (PI), polybenzimidazole (PBI) and their derivatives are representative membrane materials for H2/CO2 separation. Moreover, as a relatively new membrane material, TR polymers are having received increasing interest in 45

hydrogen purification. 3.4.1 Polyimide-based membranes PI is one of the most frequently used polymers in membrane technology, and many commercial polyimides have been used for pervaporation, nanofiltration and gas separation membranes [139, 140]. Compared to other commercial membranes based on polysulfone and cellulose

acetate,

polyimide-based

membranes

generally

exhibit

better

separation

performances combined with more favorable stability under harsh conditions. Over the past decades, different dianhydrides and different diamines (or diisocyanates) have been conjugated to synthesize various novel polyimides as membrane materials. Among them, polyimides based on fluorinated hexafluorisopropylidene-diphtatic anhydride (6FDA) are the most extensively applied materials for gas separation. However, the intrinsic H2/CO2 selectivity of most polyimides is generally not more than 10, which could not meet the requirement for hydrogen purification. Therefore, some post-modification strategies have been explored to improve H2/CO2 selectivity while maintaining relatively high gas permeability. Crosslinking has been demonstrated as an effective method to optimize material structures and properties. A comprehensive review about crosslinking polyimides for membrane application have been presented by Vankelecom et al. [139]. Crosslinked polyimides were initially developed to improve CO2-induced plasticization resistance for natural gas purification. Recently, crosslinking has also been used to increase the diffusivity selectivity of polyimide membranes for H2/CO2 separation. A series of diamines and triamines, such as diaminopropane, ethylenediamine and diethylenetriamine, have been employed by the Chung’s group to crosslink 46

polyimide for constructing an amide polymeric network [141-144]. Solution-immersion crosslinking was first adopted in the early works where the films were immersed into diamine-containing solutions. The resultant remarkable improvement in H2/CO2 selectivity was at the expense of a significant decrease in H2 permeability. For example, the H2/CO2 ideal selectivity increased from 1.8 to 120 after crosslinking modification, while the corresponding H2 permeability decreased sharply from 78 barrer to 12 barrer for 6FDA-based membrane [141]. Moreover, membrane morphology might be damaged due to uneven swelling by solvents. In order to enhance H2 permeability and mechanical property, vapor-phase crosslinking was then suggested which rendered crosslinking reaction only occur on the membrane surface rather than the entire membrane [142, 144, 145]. Effects of vapor-phase crosslinking parameters (temperature and duration time) on the modified membrane performance were intensively investigated [145]. As shown in Fig. 12, the H2 permeability of 125 barrer accompanied by the H2/CO2 ideal selectivity of 135 could be obtained under the optimal modification condition. In addition, UV crosslinking and thermal crosslinking have been adopted to crosslink various polyimides [139].

47

Fig. 12. (a) Possible scheme for reaction between 6FDA-durene and diethylenetriamine. (b) Comparison of H2/CO2 separation performances of the unmodified and diethylenetriamine-modified 6FDA-durene films with the Robeson’s upper bound [145].

3.4.2 Polybenzimidazole-based membranes PBI is another kind of polymer with an extremely high thermal and mechanical stability 48

owing to the rigid rod like molecular structure. These properties render it a perfect membrane material

for

application

under

harsh

conditions.

Among

various

PBI,

poly

(2,2’-(m-phenylene)-5,5’ bibenzimidazole) is the most typical one, and it is the first commercial one in the PBI family. Many studies have demonstrated that PBI membranes could effectively separate H2 from various gas mixtures owing to high diffusivity selectivity that arises from strong intermolecular hydrogen bonding and chain rigidity. For instance, Berchtold et al. prepared PBI composite membranes on a porous stainless steel support with a zirconia intermediate layer [146]. Then the performance of a single tube laboratory scale membrane module was tested under industrially relevant operating conditions. The membrane module exhibited a high H2/CO2 selectivity of 47 with the H2 permeance of 7 GPU at 3.45 bar and 523 K. It was obvious that the unfavorably low gas permeation performance was shown. Molecular structure modification was regarded as an effective strategy to tune chain packing and free volume for ultimately enhancing gas permeability properties of PBI-based membrane materials. Li et al. synthesized a series of PBI derivatives with different backbone structures, and effects of main chain structures on gas transport properties were studied [147]. The results indicated that the introduction of bulky, flexible and frustrated functional units into the PBI backbone all could disrupt the chain packing and increase fractional free volume, thereby the optimum membrane based on modified PBI showed a much higher H2 permeability of 997.2 barrer compared to pristine commercial available PBI (~77 barrer) at 3.4 bar and 523 K, while the corresponding gas selectivity deteriorated expectedly owing to the decreased diffusivity selectivity. Therefore, a balance between permeability and selectivity is necessary in the light 49

of specific applications. It should be noted that PBI is not well soluble in common organic solvents such as methanol, acetone, tetrahydrofuran and dimethylformamide at room temperature, which restricts the preparation of composite membranes on conventional polymeric supports via solution coating, thereby the fabrication cost of PBI composite membranes may be not economically attractive. To circumvent this problem, PBI-based asymmetric hollow fiber membranes were developed via traditional dry-jet wet spinning technique [148]. The resultant hollow fiber membranes exhibited a H2 permeance of 2.6 GPU with the H2/CO2 selectivity of 27 at 5-8 bar and 673 K. 3.4.3 Thermally rearranged (TR) polymer-based membranes In addition to conventional glassy polymers for gas separation introduced above, various advanced microporous polymers such as conjugated microporous polymers (CMPs), hypercrosslinked polymers (HCPs) and covalent organic frameworks (COFs), TR polymers and PIMs,have received much attention in gas separation over the past years [149, 150]. These microporous polymers possess interconnected pores on a nanometer or even sub-nanometer scale with high porosity, and these pores generally generated by a rigid polymer chain structure or continuous network of interconnected intermolecular voids. Therefore, various microporous polymers have been designed and synthesized as gas storage materials [151]. However, there are relatively limited studies about their application as gas separation membrane materials. Most microporous polymers have formed crosslinked networks, thus their insolubility brings out a great difficulty in membrane manufacturing 50

process [152]. Hence, solution processability is important to develop membranes based on microporous polymers from the viewpoint of the fabrication process. Fortunately, polymer precursors (mainly functionalized polyimides) of TR polymers are soluble processable, though the resultant TR polymers are generally insoluble. Thus TR polymers have been widely used for gas separation membranes for valuable gas pairs in recent years. A stiff or rigid backbone not only could generate the molecular windows of controlled size to increase diffusivity selectivity but also could disrupt the chain packing to improve gas permeability for microporous polymers, hence great efforts have been made to enhance the rigidity of entire microporous polymer structure for separation performance improvement [150]. The Lee’s group creatively proposed a thermal rearrangement concept that is basically a thermal conversion in a solid state where the precursor is converted to TR polymer shown in Fig. 13 [153]. The cavity size and cavity size distribution of TR polymers could be controlled by spatial arrangement of the rigid polymer chain segments in the glassy phase. This tailoring of free volume element could provide a strategy to develop high-performance polymeric membranes for molecular-scale separation. TR polymers generally possess bimodal cavity size distribution where the large cavities are responsible for fast gas transport, while the smaller cavities are appropriate for gas separation based on size difference. This structural characteristic of TR polymers is similar to carbon molecular sieve membranes derived from polymeric precursors via calcination. It should be noted that gas transport properties of TR polymer-based membranes are very sensitive to cavity size, and a slight difference in cavity size could result in an obvious change in gas permeability. Therefore, the cavity size should be 51

accurately controlled to develop high-performance TR polymer-based membranes for small gas separation.

Fig. 13. Two major factors contributing structural change during thermal rearrangement of polyimides. (A) Change of chain conformation–polymer chains consisting of meta- and/or para-linked chain conformations can be created via rearrangement. (B) Spatial relocation due to chain rearrangement in confinement, which may lead to the generation of free-volume elements [153].

One of the great benefits of TR polymers is the ability to tune the cavity size via designing and synthesizing polymer precursors with different structures and controlling thermal rearrangement conditions. Currently, common polymer precursors for TR polymers are functionalized polyimides and polyamides such as polyimides with hydroxyl groups, polyimides with thermally labile units and polyamides with hydroxyl groups. The most notable advantage for microporous polymers is the considerable high gas permeability due to the ultra-microporous structures compared to conventional membrane materials. In particular, TR polymers with an extremely high CO2 permeability could be obtained from polyimides 52

containing various thermally liable units such as β-cyclodextrin [154, 155]. Owing to the decomposition of β-cyclodextrin during post-thermal treatment, the TR polymer had a very high fraction free volume, and consequently the CO2 permeability of this TR polymer-based membrane could reach about 8000 barrer at 308 K. So far, many studies have proposed that TR polymers derived from various polyimides displayed excellent gas permeability and favorable selectivity for H2/CH4, H2/N2 and CO2/CH4 mixed gas [150]. In these binary mixtures, the difference of gas molecules in kinetic diameter is relatively obvious, while the size difference is only 0.41 Å between H2 and CO2 molecules. Furthermore, the kinetic diameter of the bigger CO2 molecule is generally larger than the smaller cavity size of TR polymers, thus the size exclusion effect is very limited. Consequently, even though the H2 permeability could reach thousands of barrer for TR polymer-based membranes, the corresponding H2/CO2 ideal selectivity generally varied only from 1.1 to 2. To date, only a handful of high-performance membranes based on TR polymers have been reported for H2/CO2 separation. Han et al. suggested that polymer precursors with more rigid main chain structures should be chosen and thermal rearrangement should occur at a lower temperature so as to obtain smaller cavity size that is suitable for H2/CO2 separation [156]. Following this strategy, TR-polybenzoxazole (TR-PBO) membranes were prepared from three representative rigid

poly(hydroxylamide)s

(PAHs)

including

meta-phenylene,

para-phenylene

and

hexafluoroisopropylidene bisphenylene moieties as precursors, to investigate the effects of precursor structures on physicochemical properties as well as gas separation performances of the resultant membranes. The results indicated that TR-PBO membrane derived from 53

meta-phenylene exhibited the most excellent H2/CO2 ideal selectivity of 6.2 with the corresponding H2 permeability of 206 barrer at 483 K. Moreover, not only the H2 permeability but also the H2/CO2 selectivity increased with temperature, which was favorable for hydrogen purification from shifted synthesis gas which is generally operated at high temperature. Then the same group introduced TR

poly(benzoxazole-amide) membranes

by thermal

rearrangement of poly(hydroxyamide-co-amide) with rigid biphenyl moiety and flexible ether chain [157]. The rigid biphenyl group improved gas selectivity and the flexible ether group enhanced gas permeability. Therefore, the membrane performance could be optimized via tuning the ratio of hydroxyamide to amide units in polymer precursors. Consequently, the best performance was H2/CO2 ideal selectivity of 8 with the H2 permeability of 26.8 barrer obtained at 483 K. It was obvious that the enhanced gas selectivity unfavorably caused the significant decrease in gas permeability, which offset the advantage of TR polymers to a great extent. In addition to experimental studies, various simulation methods have been adopted to reveal the relationship between polymer precursor structures and resultant cavity size, as well as gas transport properties of TR polymers, which could give a guideline to design novel TR polymers for specific separation. For instance, molecular dynamic (MD) simulation on a nanosecond scale could characteristic the structural property variation from polymer precursors to TR polymers during thermal rearrangement in terms of free volume and molecular conformation [158]. Furthermore, the calculated gas solubility, diffusivity and permeability of TR polymers based on the simulation approach showed a favorable correlation 54

to experimental results, indicating the reliability of simulation studies [159]. To further improve the separation performance, it could be concluded that how to enhance polymer rigidity that achieves a smaller cavity size and a narrower size distribution while maintaining relatively high fractional free volume is a key issue for fabrication of high-performance TR polymer-based membranes. Besides the development of novel TR polymers, TR polymer hollow fiber membranes have been successfully fabricated, and the membrane exhibited an excellent CO2 permeance of about 2000 GPU [160]. Polymers of intrinsic microporosity (PIMs) are another kind of common microporous polymers. Generally, PIMs membranes are CO2-selective membranes for CO2/H2 separation (see section 4.2). However, some modified PIMs membranes with improved diffusivity selectivity could possibly achieve H2 preferential permeation [161, 162]. For example, the chains of PIM-1 underwent 1,2-migration reaction and transformed to close-to-planar like rearranged structure via ultraviolet (UV) radiation. Thus the free volume and the size of microporous pores both notably decreased based on PALS and XRD analyses, which resulted in the significantly enhanced H2/CO2 diffusivity selectivity [161]. Consequently, the UV-rearranged PIM-1 membrane exhibited the H2/CO2 ideal selectivity of 7.3 with the H2 permeability of 452 barrer at 3.5 bar and 308 K. 3.4.4 Polyelectrolyte multilayers membranes In recent years, polyelectrolyte multilayers that are formed via alternately depositing polycations and polyanions, have been explored to exhibit an exceptional molecular sieving property. Regen et al. first fabricated single Langmuir-Blodgett bilayers derived from 55

calix[n]arene-based surfactants (n=4, 5, 6) and poly(acrylic acid) (PAA) that exhibited the excellent H2/CO2 ideal selectivity of about 70 [163]. Then the same group prepared a 7 nm thick polymeric membrane via the similar method, and the only difference was that cationic surfactants

was

replaced

by

quaternary

ammonium

derivative

of

poly(maleic

anhydride-alt-1-octadecene) [164]. The membrane displayed an extremely high H2/CO2 selectivity of 200. The author claimed that the notably enhanced selectivity was owing to the increased number of ionic sites which resulted in more concentrated ionic crosslinking. Then a highly ionically crosslinked polyelectrolyte multilayer membrane was prepared via layer-by-layer (LBL) assembly of branched polyethylenimine (PEI) and PAA on Al2O3 coated porous stainless steel tubes (Fig. 14) [165]. Titrated by the two oppositely charged polyelectrolytes, PEI and PAA became more ionized and required charge overcompensation, resulting in a highly interpenetrating network. As a result, the CO2 permeance was below the detection limit and the estimated minimum H2/CO2 selectivity was 190. It should be noted that these as-developed polyelectrolyte multilayer membranes showed the exceptional H2/CO2 selectivity combined with the extremely low gas permeance, even though the membrane thickness was only a dozen or even a few nanometers [166]. Therefore, these membranes show great potential in gas barrier applications, however, they are generally unattractive in large-scale hydrogen purification. How to significantly improve gas permeance is a key issue for polyelectrolyte multilayer membranes for practical application.

56

Fig. 14. Schematic of PEI/PAA layer-by-layer gas separation membrane supported on an alumina coated porous stainless steel tube [165].

3.5 H2-selective mixed matrix membranes Mixed matrix membranes (MMMs) are probably the most popular membrane morphology for various separation applications over the past ten years, especially for gas separation. A number of review articles about MMMs for gas separation have been presented [167-172]. As is widely accepted, MMMs, consisting of organic polymers as the continuous phase and inorganic fillers as the disperse phase (Fig. 15), not only could circumvent the limitations of polymeric membranes and inorganic membranes used alone, but also could combine their distinct advantages such as good processability of polymers and excellent gas separation property of inorganic materials. Moreover, owing to the diversity of polymers and inorganic particles, a variety of MMMs could be designed and fabricated theoretically. Meanwhile, some advanced predictive models for gas transport properties of MMMs have been developed, which could provide the guidance for the rational design and fabrication of MMMs [173]. Therefore, MMMs is the promising membranes for efficient separation. An idealized MMM should have well-dispersed inorganic fillers with as high a loading as possible. However, there are some phenomena such as non-selective interface voids, chain rigidification and pore blockage, which restrict achieving the high inorganic phase loading in MMMs. The 57

detailed information about development and challenge of MMMs could be found elsewhere [167-169, 171]. Here some representative H2-selectivie MMMs for H2/CO2 separation were introduced.

Fig. 15. Schematic diagram of mixed matrix membranes that inorganic fillers are embedded into the polymer matrix.

Similar to other H2-selective membranes introduced above, the guideline to design H2-selective MMMs is to enhance the diffusivity selectivity. Therefore, glassy polymers and inorganic fillers with molecular sieving property should be a perfect combination. Accordingly, a series of novel MMMs have been developed to improve the performance of H2-selective membranes. In these MMMs, the most extensively used polymers were PI and PBI as the membrane matrix. These polymers intrinsically exhibit a moderate H2/CO2 selectivity combined with the relatively low gas permeability. As for the dispersed phase, the early works mainly concentrated on conventional inorganic materials such as nonporous silica and zeolites. For instance, a layered material with nanoporous layers was suggested to blend with PBI. The resultant membrane exhibited a notable increase in H2/CO2 ideal selectivity, more than by a 58

factor of 2 compared to pristine PBI membrane [174]. In recent years, MOFs have emerged as a promising candidate for molecular sieving fillers due to the tunability and uniformity of pore size. Ordoňez et al. first prepared the molecular sieving MMMs that consisted of Matrimid® and ZIF-8 particles [175]. The molecular sieving effect of ZIF-8 in MMMs was demonstrated via control experiments using ZIF-8 with filled pores as the dispersed phase. Compared to the pristine Matrimid® membrane, the resultant H2/CO2 selectivity of MMMs increased 172 % from 2.58 to 7.01 when the ZIF-8 loading was up to 60 wt%. Similar to modification of pristine PI membranes, crosslinking method was suggested to further improve the H2/CO2 selectivity of MMMs based on PI. Following this strategy, Wijenayake et al. developed surface-crosslinked PI/ZIF-8 MMMs with ethylenediamine vapor as crosslink agent [176]. The diffusivity selectivity of the continuous phase was enhanced via crosslinking, thus the enhanced H2/CO2 ideal selectivity (~12) of MMMs was observed. Based on the above work, the same group recently developed a novel composite membrane that a thin crosslinked polyimide layer was coated on the Matrimid®/ZIF-8 MMM [177]. The thin crosslinked polyimide layer could make less CO2 enter into the membrane, thus the resultant membrane exhibited the optimum H2/CO2 ideal selectivity of 29 with the H2 permeability of 500 barrer at 3.5 bar and 308 K. Hu et al. developed PI/MOF hollow fiber composite membranes via dry/wet-spinning method [178]. The hollow fiber MMMs displayed an encouraging CO2 permeance of 1270 GPU with the H2/CO2 ideal selectivity of 27.8 at 10 bar and 298 K. Besides PI, PBI was regarded as a more suitable choice as H2-selective membrane material. The Chung’s group fabricated a series of PBI-based MMMs where various MOF nanoparticles 59

including ZIF-7, ZIF-8 and ZIF-90, were suggested as the dispersed phase for H2/CO2 separation [179-182]. First, ZIF-7 was introduced into the PBI matrix, because it possesses a small pore size of 3.0 Å that is between the kinetic diameters of H2 and CO2. Two strategies were adopted to facilitate the dispersion of ZIF-7 and enhance the compatibility between PBI and ZIF-7. On one hand, ZIF-7 was synthesized using excess benzimidazole, thus similar structural chains between ZIF-7 and PBI guaranteed the good compatibility. On the other hand, the unconventional drying procedure avoided the agglomeration of nanoparticles to a great extent. Compared to the pristine PBI membrane, an obvious enhancement in H2/CO2 ideal selectivity (from 8.7 to 14.9) was shown when 50 wt% ZIF-7 was added. However, the corresponding CO2 permeability was only 26.2 barrer due to the small pore size of ZIF-7. Then ZIF-8 with a larger pore size was introduced into the polymer matrix to improve gas permeability instead of ZIF-7 [180]. Consequently, the H2 permeability increased to 105.4 barrer combined with a slight decrease in H2/CO2 ideal selectivity (~12.3) compared to PBI/ZIF-7 MMMs. In addition, potential contaminants such as water vapor and carbon monoxide in shifted synthesis gas did not significantly influence the performance of PBI/ZIF-8 MMMs [182]. In addition to PI and PBI, other glassy polymers such as PSf and poly(methyl methacrylate) (PMMA), have been used as the matrix of H2-selective MMMs [183-185]. For instance, Guiver et al. developed PSf/APTES-modified zeolite 3A MMMs [183]. The good compatibility was achieved because organic-modified zeolite particles were covalently bounded to PSf, hence the membrane exhibited a very high H2/CO2 ideal selectivity of 72 with the H2 permeability of 7.1 barrer owing to the small pore size of zeolite. Moreover, 60

Cao et al. fabricated a highly permeable MMM consisting of PMMA and amino group-functionalized MOF [185]. The MOF could be well dispersed in the PMMA matrix owing to hydrogen-bonding interactions between amino groups of MOF and carbonyl groups of PMMA. The resultant membrane exhibited the optimum H2 permeability of 11000 barrer with the H2/CO2 ideal selectivity of 13 when the MOF loading was 15 wt%. Different from common fillers in MMMs, metallic nanoparticles, especially Pd-based nanoparticles, have been suggested as fillers in H2-selective MMMs, because they show a specific “affinity” towards H2. It should be mentioned that metallic nanoparticle-embedded MMMs must be operated at elevated temperature (~573 K) to guarantee the efficiency of metallic nanoparticles. Suhaimi et al. incorporated Pd nanoparticles into the PSf matrix, and polyvinylpyrrolidone (PVP) as stabilizer was used to avoid the agglomeration of nanoparticles [186]. As a result, the CO2 permeance of 3124 GPU with the H2/CO2 ideal selectivity of 6.2 was obtained. Pd nanoparticles were also incorporated into silica matrix [187]. The author claimed that H2 transport mechanism in the membrane changed from molecular sieving to solution-diffusion due to the existence of Pd nanoparticles. Therefore, the membrane exhibited an obvious higher H2 permeability and H2/gas selectivity.

3.6 Comparison of H2-selective membranes Fig. 16 presents the separation performances of representative H2-selective membranes for H2/CO2 separation plotted on the 2008 Robeson upper bound [188]. Inorganic membranes, especially dense metallic membranes, generally exhibit superior separation performances for hydrogen purification. Moreover, they are applicable for high temperature separation, which is 61

favorable for pre-combustion CO2 capture and high-temperature WGS reaction. However, every inorganic membrane possesses its own disadvantages. Dense metallic membranes are very sensitive to application conditions including temperature and contaminants. Moreover, one serious limitation of dense metallic membranes is the high fabrication cost, thus they may be merely suitable for small-scale application for ultrahigh-purity hydrogen production. Microporous inorganic membranes including zeolite membranes, silica membranes and carbon molecular sieve membranes, show good mechanical property and chemical stability. However, water vapor in feed gas could result in a significant deterioration of membrane performance. In addition, it is difficult for almost all inorganic membranes to be transformed into a high surface area membrane module for large-scale application. As for H2-selective polymeric membranes from glassy polymers, they could withstand a moderately high temperature because these polymers possess high glassy transition temperature. However, there is a well-known trade-off phenomenon between H2 permeability and H2/CO2 selectivity which restricts the fabrication of high-performance H2-selective polymeric membranes. Furthermore, CO2-induced plasticization could cause an obvious decrease in H2/CO2 selectivity at high CO2 partial pressure. For some newly-developed membrane materials, MOF membranes and TR polymer membranes generally show higher gas permeability than conventional membranes and some special ones have exhibited the favorable H2/CO2 selectivity, the studies about which are however laboratory-scale. Therefore, the systematic investigation under realistic conditions is necessary to evaluate the potential for practical application. Moreover, the commercial manufacture of MOFs and TR polymers is vital for large-scale application. To the 62

best of our knowledge, there is no H2-selective membrane that has been brought to commercial reality at a large scale to date, every membrane should overcome its fatal weakness when competes with commercial available pressure swing adsorption and absorption processes for hydrogen purification.

Fig. 16. Separation performances of representative H2-selective membranes for H2/CO2 separation (square: microporous inorganic membranes, triangle: MOF membranes, star: polymeric membranes, circle: mixed matrix membranes). The data could be seen in Table S1 in supporting information.

4. Advances in CO2-selective membranes Reverse selective membranes have received much attention because of their inherent advantages for some specific separation applications, even though the fabrication of reverse selective membranes is generally more challenging [189]. For hydrogen purification from 63

shifted synthesis gas, CO2-selective membranes are reverse selective membranes where larger CO2 molecule can permeate faster than smaller H2 molecule [2, 14]. In order to achieve CO2 preferential permeation, the membrane should possess enough strong CO2 affinity. Considering the high condensability and acid nature of CO2 as shown in Table 1, an effective strategy to enhance the affinity towards CO2 is suggested to improve the polarity and basicity of membrane materials. In this section, CO2-selective membranes are divided into five categories, microporous membranes based on CO2 preferential sorption, polymers of intrinsic microporosity (PIMs)-based membranes, CO2-philic polymeric membranes, facilitated transport membranes and mixed matrix membranes.

4.1 Microporous membranes based on CO2 preferential sorption As described in Section 3.2 and 3.3, the overwhelming majority of microporous membranes are H2-selective membranes where gas molecules transport is mainly based on Knudsen diffusion or molecular sieving mechanism. In the case that the pore size is big enough to allow CO2 molecules diffuse into the pores, CO2 molecules are commonly preferentially adsorbed on the pore walls because CO2 has much higher condensability than H2. Therefore, if there is enough strong CO2 affinity of pore walls and the pore size is suitable, CO2 molecules adsorbed on the pore walls could seriously restrict and even block H2 diffusion through the membranes, achieving CO2 preferential permeation. For example, Hong et al. developed a CO2-selective SAPO-34 membrane at low temperature [28]. The H2 permeance in mixed gas test was less than 1% of that in pure gas test due to the inhibition of CO2, while H2 did not affect CO2 permeance much. Therefore, an excellent CO2/H2 selectivity of about 140 64

was achieved at 253 K. In these CO2-selective microporous membranes, CO2 transport generally obeys capillary condensation or surface diffusion mechanism. Capillary condensation is the extension of surface diffusion, and there are many rigorous prerequisites such as temperature and pressure for the occurrence of capillary condensation. To the best of our knowledge, there has been no article published about gas separation membranes where CO2 transport is fully based on capillary condensation mechanism till now. As for surface diffusion, CO2 molecules could essentially adsorb and then diffuse along the walls of the pores at a faster rate than they move in bulk. Unfortunately, it is difficult to accurately define the conditions that surface diffusion mechanism dominates gas transport through the membrane, which is closely related to test condition such as temperature, pore size and interactions between pore walls and gas molecules [190, 191]. Lee et al. suggested that the pore size should be less than 3-4 times of the kinetic diameter of CO2 molecule to maximize the efficiency of surface diffusion mechanism [192]. There are very few studies have been performed on hydrogen purification using CO2-selective microporous membranes. It could be concluded that it is a very challenging issue to reproductively develop high-performance CO2-selective microporous membranes for hydrogen purification. The importance of surface properties has encouraged the introduction of specific functional groups for enhancing the attachment of CO2 molecules to the pore walls and then promoting the diffusion through the membranes. In particular, amino group functionalization is the most widely adopted method owing to the strong interactions between CO2 and amino groups. Lindmark et al. fabricated the functionalized MFI-type zeolite membranes with high 65

Si/Al ratio using methylamine to enhance the CO2 affinity [193]. After modification, the H2 permeance decreased more significantly than CO2, thereby the modified membranes showed a CO2/H2 selectivity of 6.5 with the CO2 permeance of about 2986 GPU at 1 bar and 333 K. Maturin et al. suggested the post-synthesis ammonia treatment of porous carbon membranes at elevated temperature, which not only increased the porosity but also achieved the surface functionalization [194]. The resultant CO2 permeance and CO2/gas selectivity were obviously enhanced simultaneously. Similar to zeolite membranes, silica membranes have been successfully modified by various aminosilanes as well. For example, the CO2 affinity of silica membranes was enhanced by covalent bonding of APTES to the pore walls (Fig. 17) [195, 196]. Besides the amino group surface modification, cation exchange treatment of zeolite membranes was another method to improve CO2 sorption capacity. Many studies have indicated that the basicity of cation strongly affects the CO2 sorption rate and capacity in the zeolite framework, however, the cation of raw zeolite membranes is usually H+ that is unfavorable for weakly acidic CO2. Therefore, some alkaline cations (Na+, K+, Cs+) and alkaline earth cations (Mg2+, Ca2+, Ba2+) were selected to replace H+ for enhancing CO2 sorption capacity. For example, Lindmark et al. and Chew et al. developed a series of ion exchange treated ZSM-5 and SAPO-34 membranes [197, 198]. It should be noted that the cation exchange could also affect pore size. Generally, the larger radius of the cation means the higher basicity among the cations of the same class. Thereby the change of membrane performance depends on the competitive effects of changes in gas sorption and gas diffusivity after cation exchange treatment. The CO2 permeance of 3900 GPU and the CO2/H2 selectivity 66

of 6.2 were obtained at 1 bar and 295 K for Ba-ZSM-5 membrane when the feed contained a few amount of water.

Fig. 17. Schematic of the brush-like structure in the silica pore after APTES modification (not to scale) [196].

Sandstrom et al. prepared a MFI-type zeolite membrane with a very high CO2 permeance, and the membrane exhibited the CO2 permeance of about 28000 GPU with the CO2/H2 selectivity of about 16.2 at 8 bar and 296 K [199]. Effects of temperature and feed pressure on membrane performances were also intensively studied. In particular, the maximum CO2/H2 selectivity could reach 32.1 at the low temperature of 275 K because the low temperature benefits CO2 sorption. The same group further investigated the influence of impurities (CO, H2S and H2O) on membrane separation performances, and the result verified that the membrane displayed a good stability under industrially relevant conditions [200]. Recently, Zhou et al. synthesized a thin and uniform b-oriented MFI-type zeolite membrane in a fluoride media [201]. At the extremely low temperature of 238 K, the membrane exhibited a 67

remarkably high CO2/H2 selectivity of 109 with the high CO2 permeance of more than 15000 GPU at 8 bar due to the strong CO2 adsorption and the blockage of H2 permeation, while the selectivity dropped rapidly to 31 at 298 K, indicating that surface diffusion mechanism tends to be more effective at lower temperature. In addition to conventional microporous inorganic membranes, MOF membranes could also become CO2-selective membranes. Takamizawa et al. first developed a single-crystal MOF membrane of [Cu2(benzoate)4(pyrazine)] that could achieve CO2 permeation prior to H2 [202]. Moreover, Zhao et al. found that MOF-5 membrane could achieve CO2 preferential permeation at high CO2 feed composition (~82 vol%) [203]. In this case, a large number of CO2 molecules adsorbed on the pore walls strongly suppressed the adsorption of H2 molecules, hence the CO2/H2 selectivity was close to 5 at about 3.4 bar and 298 K. The slight increase in CO2/H2 selectivity was observed with increasing feed pressure, which was different from other microporous inorganic membranes. Very recently, Al-Maythalomy et al. first fabricated a continuous

CO2-selective

MOF

membrane

using

In3+

as

metal

resource

and

imidazoledicarboxylate as linker [204]. This membrane showed the anionic characteristic due to the presence of extra-framework imidazolium cations. Therefore, the membrane displayed a reverse-selectivity property that CO2/H2 selectivity reached 5.2 at 3.4 bar and 308 K. In all, there have been a very limited number of reports about CO2-selective microporous inorganic membranes until now. Fabrication of defect-free microporous membranes possessing the pores that could achieve CO2 preferential transport is quite troublesome even if every inorganic membrane could be a CO2-selective one theoretically. Moreover, the majority 68

of CO2-selective microporous membranes exhibit a high CO2/H2 selectivity just at low temperature, which tends to counteract the significant advantage of inorganic membranes that they are suitable for high temperature application.

4.2 Polymers of intrinsic microporosity-based membranes Similar to TR polymers, polymers of intrinsic microporosity (PIMs) are another representative microporous polymers which possess a ladder-like backbone and a contorted structure that inhibits the freedom of chain rotation to a great extent. Compared to most microporous polymers, one significant advantage of PIMs is that the majority of PIMs are solution processable, hence PIMs can be easily processed to self-standing films or composite membranes on supports via solution coating method, suggesting them as promising membrane materials for large-scale application [149]. The first synthesized PIMs series were developed via a polycondensation reaction between tetrafluoro-monomers and tetrahydroxyl-monomers [205]. Among them, PIM-1 is the representative one as membrane materials for gas separation, the synthesis method and chemical structure of which were shown in Fig. 18. It is obvious that there is no single bond in the backbone around which rotation can occur. Moreover, there exist microporous pores near the spiro-carbon centers and ultra-microporous pores near the nitrile sites in PIM-1 matrix, which resembles the bimodal cavity size distribution of TR polymers. For PIM-1, the distinctive gas transport properties of reverse selectivity for condensable gases such as CO2 were shown that the CO2/H2 ideal selectivity was about 1.8 with the CO2 permeability of 2300 barrer at 303 K [206]. This phenomenon is mainly owing to the weak diffusivity selectivity resulted from high free volume and the enhanced solubility selectivity 69

caused by the interaction between polar nitrile groups and CO2 molecules.

Fig. 18. Synthesis of PIM-1 polymer [205].

It is worth mentioning that the membrane fabrication parameters (solvent, membrane thickness) and post treatment have a great influence on the performance of PIM membranes, even though they are originated from the same membrane material. For example, a significant increase in gas permeability of PIM-1 membrane was observed after methanol treatment for 5 days, and this change resulted from the variation in free volume size according to the PALS study. Consequently, the maximum CO2 permeability could reach 12600 barrer combined with a modest decrease in CO2/gas ideal selectivity. The further study revealed that the increase in gas permeability was mainly caused by a notable enhancement in solubility coefficient and an increase in diffusivity coefficient to a lesser extent. Most reported values of gas permeability about PIMs were based on the membranes after methanol or ethanol treatment, because methanol or ethanol treatment prior to measurement is a common method to delay the effects of physical aging for polymers with high free volume. Nowadays, the following various modification strategies have been intensively investigated to improve the membrane performance based on PIMs. Similar to TR polymers, chain rigidity is closely related to gas transport properties of PIMs. Therefore, enhancing the chain rigidity is an effective strategy to improve the 70

performances of PIM membranes especially for gas permeability. The relatively flexible dioxane chain-forming bridges and sprobisindane (SBI) linkages in PIM-1 restricted the overall rigidity. The planar dioxane has a single dihedral angle with a broad distribution, and SBI has a broad bimodal profile with two dihedral angles (Fig. 19) [207]. Bezuu et al. first used more rigid spirobifluorene (SBF) unit instead of SBI unit to synthesize PIM-SBF with enhanced gas permeability and selectivity [208]. The CO2 permeability of 13900 barrer with the CO2/H2 ideal selectivity of about 2.2 could be achieved. However, the enhanced chain rigidity was limited owing to the presence of inherently flexible spiro-centers. Then the successful strategy to improve substantial chain rigidity via replacing SBI unit and dioxane unit with stiffer ethanoanthracene (EA) unit, triptycene (Trip) unit and Troger’s base (TB) unit in the PIM-1 backbone was suggested by the Carta’s group [209, 210]. As a result, a series of new more rigid PIMs including PIM-SBI-TB, PIM-EA-TB, PIM-Trip-TB, have been developed for gas separation, and these PIMs possessed a considerably high CO2 permeability. In particular, the CO2 permeability of PIM-Trip-TB membrane approached to 10000 barrer owing to its extremely rigid structure [210]. However, different from most PIM membranes such as PIM-1, PIM-EA-TB membrane exhibited an obvious molecular sieving property that H2 permeates faster than CO2, which was probably due to the enhanced diffusivity selectivity [209]. Very recently, the intensive investigations of structures and gas transport properties of PIM-EA-TB membrane were done via the combination of experimental analysis and molecular simulation [211].

71

Fig. 19. (A) The PIM-1 polymer contains SBI and dioxane linkages, but both can bend and flex to a considerable extent. (B) The PIM-EA-TB polymer has a more rigid architecture [207].

In addition to enhancing polymer chain rigidity, functionalized PIMs have attracted increasing attention via introducing various functional groups on nitrile sites of PIM-1 to tune free volume and enhance CO2 affinity, and ultimately to optimize gas transport properties for CO2 separation [212-219]. Some representative functionalized PIMs derived from PIM-1 were illustrated in Fig. 20. Generally, functionalized PIMs showed the improved gas selectivity combined with the decreased gas permeability compared to pristine PIMs (mainly PIM-1). There are primarily the following two reasons that result in the above phenomenon. On one hand, the size of introduced functional groups is commonly larger than that of nitrile group in PIM-1, occupying the free volume available in PIM-1. On the other hand, the introduced functional groups such as carboxylate group with active hydrogen atom could promote the 72

intermolecular hydrogen-bonding interaction which also results in the decrease in free volume. For instance, Mason et al. reported the preparation of thioamide-PIM-1 through the thionation of nitrile groups to enhance the interaction between the polymer and the penetrating gas.[216] After modification, an obvious increase in CO2/N2 ideal selectivity (~38.5) was observed at the expense of the substantial decrease in CO2 permeability (~150 barrer). The CO2 permeability (~1120 barrer) was still much lower than that of PIM-1 even though ethanol treatment was done. Moreover, there was no obvious change in CO2/H2 ideal selectivity after modification, which was probably due to the limited interaction between thioamide groups and CO2. As a result, the increase in CO2/H2 solubility selectivity was balanced by the increase in H2/CO2 diffusivity selectivity. In order to improve gas selectivity while maintaining a relatively high CO2 permeability, Du et al. developed tetrazole groups containing PIM, denoted as TZ-PIM-1, via [2+3] cycloaddition reaction of nitrile groups of PIM-1 with an azide compound [215]. The introduced tetrazole groups resulted in more favorable CO2 sorption with superior affinity. The dynamics of CO2 adsorption on TZ-PIM-1 was intensively investigated via in situ

13

C nuclear magnetic resonance spectroscopy [220]. The results

indicated that the CO2 molecules interacted with tetrazole sites via physisorption, and two adsorption sites were available when the CO2 loading approached to one CO2 molecule per tetrazole unit. As a result, the TZ-PIM-1 membrane exhibited the high CO2/N2 selectivity of about 40 with the CO2 permeability of about 3000 barrer at 298 K. Moreover, the CO2/N2 selectivity from mixed gas was higher than the ideal selectivity from single gas, while the CO2/CH4 selectivity from mixed gas was similar to the ideal selectivity from single gas. This 73

difference was mainly attributed to the distinction in physicochemical properties between N2 and CH4 molecules. The preferential adsorbed CO2 blocked the transport of N2 more seriously than that of CH4. Unfortunately, the H2 permeability of TZ-PIM-1 membrane was not reported, it is expected that TZ-PIM-1 should have higher CO2/H2 selectivity than pristine PIM-1. In addition, nitrile groups in PIM-1 were converted to primary amino groups via borane complexes and amidoxime groups via the reaction with hydroxyl amine [218, 219]. Amino-PIM-1 displayed higher CO2 uptake and higher CO2/N2 sorption selectivity than PIM-1. Even though H2 permeability (~3070 barrer) was larger than CO2 (~1890 barrer) in terms of singe gas permeation, amino-PIM-1 membrane may become CO2-selective in mixed gas test. It is because that the transport of H2 molecules could be possibly hindered by a mass of adsorbed CO2 molecules [28]. Furthermore, primary amino group possesses stronger CO2 affinity than tetrazole group, it is expected that the CO2/gas selectivity of amino-PIM-1 membrane might be higher than TZ-PIM-1 membrane. Swaidan et al. investigated the pureand mixed-gas CO2/CH4 separation properties of amidoxime-functionalized PIM-1 [221]. The resultant membrane displayed a CO2 permeability of 1153 barrer with the CO2/CH4 ideal selectivity of 34. The relatively high selectivity was attributed to the size-sieving ultra-microporosity and the strong interaction between amidoxime units and CO2. In mixed gas test, not only the CO2 permeability but also the CO2/CH4 selectivity decreased to a certain extent compared to pure gas test. However, the CO2/CH4 selectivity remained above 20 at 10 bar CO2 partial pressure, demonstrating the favorable CO2-induced plasticization resistance due to an extensive intermolecular hydrogen-bonding network caused by amidoximine 74

moieties. It should be mentioned that the introduction of various CO2-philic functional groups generally could increase CO2/H2 solubility selectivity accompanied with the increase in H2/CO2 diffusivity selectivity. These competing effects between solubility selectivity and diffusivity selectivity hinder the notable improvement of CO2/H2 permeability selectivity, which is greatly different from CO2/N2 and CO2/CH4 separation. Therefore, the increased solubility selectivity should be predominant for functionalized PIM membranes to enhance CO2/H2 selectivity, which should be a guideline to design novel PIMs for fabrication of high-performance PIM membranes for CO2/H2 separation. Accordingly, in addition to the introduction of CO2-philic functional groups or units into PIM-1, the pore size of PIMs should be further accurately tailored so that preferentially adsorbed CO2 in the membranes could restrict or block H2 transport to a great extent, achieving the maximum efficiency of enhanced solubility selectivity.

Fig. 20. PIMs with various substituted pendant functional groups via post modification [212-219]. 75

Similar to the aforementioned network formed via hydrogen-bonding interaction, crosslinking that constructs a compact network with chemical bonds is an approach to improve gas selectivity. This modification strategy is quite suitable for CO2/CH4 and CO2/N2 separation. Therefore, various crosslinking methods such as thermal self-crosslinking, thermal oxidative crosslinking and ultraviolet irradiation treatment, have been suggested to tailor gas transport properties of PIMs [222-224]. As expected, these crosslinked PIMs usually exhibited a higher CO2/gas selectivity with the decreased CO2 permeability. In particular, a high CO2/gas selectivity and a high CO2 permeability could be obtained simultaneously from the thermally self-crosslinked PIM-1 membrane (Fig. 21) [222]. On one hand, the formed bulky triazine rings during the crosslinking could prevent the efficient chain packing, which could partially restrict the decrease in gas permeability resulted from crosslinking. On the other hand, various sizes of small pores could be generated for a long period of thermal treated time to increase new free volume. Consequently, the optimum CO2/N2 ideal selectivity could be more than 54 with the CO2 permeability of 4000 barrer. As for the separation of CO2 and H2, the crosslinking method generally tends to transform CO2-selective PIMs to be H2-selective ones owing to the notably enhanced diffusivity selectivity. However, if CO2-philic functional groups containing crosslinking agents could be used to crosslink PIMs, the enhanced CO2 solubility coefficient could partially restrict the decrease in CO2 permeability resulted from crosslinking. Consequently, the CO2/H2 selectivity might be improved. Moreover, crosslinking could make the membrane more stable for application under harsh conditions.

76

Fig. 21. Two-dimensional representations of the contorted PIM-1 membrane before and after thermal cross-linking reaction with the formation of triazine rings [222].

Physical aging phenomenon that could result in the dramatically reduced gas permeability over time is a fatal drawback for glassy polymers with high free volume, which seriously restricts their practical application as gas separation membrane materials. Crosslinking could weaken physical aging effects to some extent at the expense of the reduced gas permeability, which is unfavorable for gas separation membranes. Recently, Lau et al. introduced a novel and effective strategy to solve physical aging problem [225]. A very specific microporous filler, porous aromatic framework (PAF), was suggested to incorporate into three representative glassy polymers including poly(trimethyisilylpropyne) (PTMSP), poly(4-methyl-2-pentyne) (PMP) and PIM-1. PAF and glass polymers cooperatively form an interwoven nanocomposite, freezing the structure and therefore largely stopping aging process whilst increasing gas permeability and selectivity. There was only a very slight decrease (< 7%) in CO2 permeability of PIM-1/PAF membrane over time with more than 240 days, while a decrease by 62% was shown for pristine PIM-1 membrane. Meanwhile, the PIM-1/PAF-1 77

membrane exhibited an extremely high CO2 permeability of more than 10000 barrer. This innovative study would significantly boost the industrial application of glassy polymers with high free volume.

4.3 CO2-philic polymeric membranes In stark contrast to H2-selective polymeric membranes, the ultimate aim to design CO2-selective polymeric membranes is to increase CO2/H2 solubility selectivity via introducing polar functional groups and to decrease H2/CO2 diffusivity selectivity via disrupting chain packing. Rubbery polymers have been considered as suitable candidates for CO2-seletive membranes materials, because they not only show the favorable affinity towards quadrupolar CO2 molecules, amplifying the solubility selectivity of CO2/H2, but also possess the flexible chains, minimizing the diffusivity selectivity of H2/CO2. Currently, poly(ethylene oxide) (PEO), which is also referred to as poly(ethylene glycol) (PEG) at the low molecular weight, containing CO2-philic polymeric membranes are predominant CO2-philic polymeric membranes. Moreover, newly developed ionic liquid-based membranes showed promising performances for CO2 separation as well. 4.3.1 PEO-based membrane In order to enhance the interaction between polymers and CO2, different polar functional groups were suggested to introduce into the membranes. Lin et al. introduced a comprehensive material selection guideline about membranes for CO2 separation, and the systematic correlation between material structure containing different functional groups and gas solubility was suggested [226]. Ethylene oxide (EO) units were regarded to possess the optimum 78

solubility parameter and achieve the best combination of CO2 permeability and CO2/gas selectivity owing to the structural flexibility and quadruple-pole interaction between EO groups and quadrupolar CO2 molecules. Therefore, PEO-based membranes have received considerable attention in various CO2-involved separation fields including hydrogen purification from shifted synthesis gas [227, 228]. However, one serious drawback of using pure PEO as membrane materials is its strong tendency to crystalline [229]. On one hand, high crystallinity results in the decreased efficiency of EO groups, and only the functional groups resided in the amorphous phase could interact with CO2 molecules because the crystalline region is impermeable to gas [230]. On the other hand, the presence of crystalline phase reduces the chain mobility and therefore enhances the diffusivity selectivity, which is unfavorable for CO2/H2 separation. As a result, high crystallinity pure PEO membranes generally exhibited the unfavorable performance that CO2 permeability was 12 barrer combined with the CO2/H2 ideal selectivity of about 6.7 at 308 K and infinite dilution [229]. Moreover, the poor mechanical strength of PEO limits the industrial application. Therefore, the key issue to develop high-performance PEO-based membranes is to design novel PEO containing membrane materials with a relatively low crystallinity and good mechanical strength. Consequently, different smart strategies have been investigated to suppress the crystallization of PEO while maintaining high PEO content, achieving improved gas transport properties for CO2 separation. a. Crosslinked PEO membrane Crosslinking is an effective method to optimize material structures, and it has been 79

widely used to reduce the crystallinity and improve fraction free volume of PEO-based membranes. The chemical composition and structure of crosslinking agents play an important role in tuning gas transport properties of crosslinked polymers. The Freeman’s group designed and fabricated a series of crosslinked PEO membranes via the photopolymerization of poly(ethylene glycol) diacrylate (PEGDA), poly(ethylene glycol) acrylate (PEGA), poly(ethylene glycol) methyl ether acrylate (PEGMEA), poly(ethylene glycol) dimethacrylate (PEGDMA) and their mixtures [228, 231-233]. Effects of chemical composition and crosslinking density on membrane separation performances were systematically studied. In particular, Lin et al. suggested that some short non-PEO segments should be introduced into the polymer backbone to interrupt the EO repeat units for inhibiting the crystallinity, thus a highly CO2-philic crosslinked PEO membranes with 82 wt% EO units was synthesized via photopolymerizing PEGDA (30 wt%) and PEGMEA (70 wt%) (Fig. 22) [228]. The CO2 permeability of the resultant membranes increased with increasing CO2 partial pressure, meanwhile, the permeability rise was greater at lower temperature because the temperature reduction was favorable for CO2 sorption. Therefore, the maximum CO2 permeability of 440 barrer could be obtained at a feed CO2 partial pressure of 17 bar and 308 K. Moreover, there was a surprising phenomenon observed that the CO2/H2 selectivity increased with increasing CO2 partial pressure, thus the maximum CO2/H2 selectivity reached 31 at 253 K, indicating that CO2-induced plasticization enhanced the separation performance. This was because the increased free volume of highly swollen membranes caused by plasticization weakened the molecular sieving effects. Therefore, a high CO2 permeability and a high CO2/H2 selectivity 80

could be achieved simultaneously at high feed CO2 partial pressure. Then the same group used various short-chain acrylate monomers with different kinds of end groups such as hydroxyl, methoxyl and ethoxyl groups, to replace long-chain acrylate monomer PEGMEA in PEGDA/PEGMEA crosslinked network, and effects of pendant groups on membrane performance were intensively investigated [234]. The methoxyl and ethoxyl group could increase the fractional free volume that caused the increased gas permeability, while hydroxyl group played the opposite effect due to the hydrogen-bonding interaction with ether groups. Moreover, different multifunctional thiols, varying in functionality and rigidity, were suggested to incorporate into the UV-cured PEGDA membrane [235]. Recently, new PEO-containing amorphous network membranes for CO2 separation were developed via thiol-end photopolymerization by Kwisnek et al. [236]. These new crosslinked thiol-end networks provided enhanced mechanical strength than conventional acrylate networks.

Fig. 22. Schematic representation of PEGDA/PEGMEA copolymer network. Italicized and bolded parts of the network

derive

from

the

cross-linker.

R1

is

COO(CH2CH2O)14OC from PEGDA [228]. 81

CO(OCH2CH2)8OCH3

from

PEGMEA;

R2

is

b. Copolymers containing PEO In addition to developing a PEO-based crosslinked network described above, copolymerization is another common method to suppress the crystallization. An idealized copolymer for CO2 separation should have the following characteristics: weak interaction between polymer and polymer, strong interaction between polymer and CO2, and high free volume. Over the past ten years, a series of PEO-containing copolymers have been successfully synthesized, and these copolymers generally consist of a soft PEO segment and a hard segment with different compositions and molecular weights [237]. For these PEO-containing copolymers, PEO segments are responsible for high CO2 solubility selectivity, while hard segments could limit the strong crystallization tendency of PEO and enhance the mechanical strength. Several representative copolymers including PEO-polyimide (PEO-PI), PEO-polyester and PEO-polyamide (PEO-PA), have been extensively used for fabrication of membranes for CO2 separation. Effects of PEO content, PEO molecular weight of different copolymers on membrane structures and gas transport properties have been systematically investigated. PEO-PI copolymers are commonly synthesized via the reaction between different dianhydrides and ether diamines, followed by thermal and chemical imidization. Chen et al. synthesized different PEO-PI copolymers with different content and molecular weight of PEO [238]. The result suggested that the influence of PEO content on CO2 permeability was greater than that of PEO molecular weight. The optimum CO2/H2 selectivity of 22.7 with the CO2 permeability of 179 barrer could be obtained at 2 bar and 308 K when the PEO content was 82

60%. It should be noted that the selectivity from mixed gas was obviously higher than ideal selectivity (~8.5) from pure gas due to a notable decrease in H2 permeability. In mixed gas test, the restricted H2 transport was the consequence of strong interaction between polymers and CO2. PEO-polyester copolymers can be synthesized via the polycondensation of aromatic diacids and aliphatic diols. The most common PEO-polyester copolymer is the PEO-poly(butylene terephthalate) (PEO-PBT), and it has been the commercial product with the trade name of Polyactive®. Metz et al. investigated the influences of PEO length and content on the gas transport properties of PEO-PBT based membranes [239]. The performance of PEO-PBT based membranes was further optimized via tuning the content and molecular weight of PEO [240]. The CO2 permeability of 115~150 barrer and the corresponding CO2/H2 selectivity of 10.2 could be acquired at 303 K when the content and molecular weight of PEO were about 77 wt% and 1500 g/mol. Moreover, the author suggested that the highest CO2 permeability could be achieved in the case of the PEO molecular weight of 2000 or 2500 g/mol in PEO-PBT copolymer. Similarly, a series of PEO-poly(trimethylene terephthalate) (PEO-PTT) copolymers were synthesized as novel membrane materials for CO2 separation [241]. Different from PEO-PBT copolymer, the PTT segment has three methylene units while PBT has four units. The combination of experimental study and mathematical model study was adopted to optimize the PEO-PTT structure for improved gas transport properties. The results indicated that the copolymer with 70 wt% PEO and the PEO molecular weight of 2000 g/mol possessed the optimum CO2 permeability of 183~200 barrer with the CO2/H2 selectivity 83

of 10.2 at 303 K. PEO-PA copolymers are synthesized by the polycondensation of aliphatic ether diamines and aromatic diacids. The commercial products of poly(ether-amide) copolymers are marked under the trade name of Pebax®. Nowadays, Pebax® series may be the most popular PEO-based membrane materials for CO2 separation. Similar to poly(ether-imide) copolymers and poly(ether-ester) copolymers, the gas transport properties of PEO-PA copolymers could be tailored through varying the content and structure of PEO and PA segments, and the corresponding systematic research has been intensively conducted by Bondar et al. and Kim et al. [242, 243]. Moreover, Tocci et al. applied three simulation methods of Molecular dynamic, Monte Carlo and transition state theory to calculate gas transport parameters including diffusivity, solubility and permeability coefficients, which was a powerful supplement to conventional experimental studies about gas transport properties [244]. In addition to Pebax®, some other advanced PEO-PA copolymers have been synthesized. Reijerkerk et al. developed a series of block copolymers including poly(ethylene oxide)-ran-poly(propylene oxide) (PEO-ran-PPO) soft segments and monodisperse tetra-amide (T6T6T) hard segments [245, 246]. Different from traditional pure PEO segments in PEO-based copolymers, the weight ratio of PEO to PPO in soft segments is 3:1. The presence of methyl group in PPO in soft segments strongly restricted the crystallization and therefore enhanced the gas permeability. The monodisperse crystallizable T6T6T hard segments resulted in the almost complete phase separation which guaranteed high soft phase concentration. The maximum CO2 permeability could reach 470 barrer with the corresponding CO2/H2 selectivity of 10 at 308 K. 84

In addition to three series of aforementioned typical PEO-containing copolymers, other types of PEO-containing copolymers have been synthesized and used for CO2 separation. Kim et al. synthesized PEO-polysulfone (PEO-PSf) random copolymers with different PEO content. All gas permeabilities of PEO-PSf random copolymer membranes unexpectedly decreased with increasing PEO content till the content was increased to 31 wt%. This was mainly due to the fact that the good compatibility between PSf and PEO resulted in less phase-separated structures [247]. Furthermore, a series of poly(urethane-urea) copolymer containing different polyethers of PEO, PPO and their mixtures were developed [248]. As expected, the PPO-based poly(urethane-urea) showed the highest CO2 permeability of 190 barrer owing to the highest free volume, while the corresponding CO2/H2 ideal selectivity was merely 4.5. As for PEO-based poly(urethane-urea), it exhibited the lowest CO2 permeability of 69 barrer with the CO2/H2 selectivity of 5.8 because of the strong crystallization of PEO. c. Polymer blends with EO-containing additives Physical blending is a time-effective route to tailor membrane material structures and properties in comparison to designing and synthesizing new advanced crosslinked PEO networks and PEO-containing copolymers described above. Hence, physical blending has received much interest for fabricating membranes for CO2 separation in recent years. A low-molecular-weight PEG has been widely suggested to introduce into the polymer matrix for enhancing membrane separation performances, because it possesses abundant EO groups and could readily interact with CO2 under an amorphous state. The introduced PEG additive not only could increase EO content in membranes to enhance CO2 solubility, but also 85

could increase the flexibility of polymer chains to improve gas diffusivity. Therefore, an enhanced CO2-philic characteristic and a more amorphous structure could cooperatively contribute to a higher solubility and a higher diffusivity with the addition of PEG. Consequently, a notable improvement of membrane performance could be achieved via simple physical blending of PEG. Currently, polymer blends with EO-containing additives generally focused on two fields: 1) the blending of PEO-containing copolymers with PEG, 2) the blending of PEO-containing organic-inorganic hybrid materials with PEG. The latter one would be discussed in section 4.5 of PEO-based MMMs. For PEO-containing copolymers based blending system, two commercially available copolymers, Pebax® and Polyactive®, were the most extensively used as membrane matrix. Yave et al. first prepared Pebax® MH 1657/PEG200 (molecular weight: 200) membranes with different amounts of PEG200 [249]. The enhanced CO2 solubility was due to the introduced EO groups from PEG200. Meanwhile, PEG200 as a plasticizer could interrupt the interaction between polymers, resulting in the increase in fractional free volume based on PALS analysis. As a result, not only the CO2 permeability but also the CO2/H2 ideal selectivity increased with increasing PEG200 content. In particular, The CO2 permeability and CO2/H2 ideal selectivity were 151 barrer and 10.8 at 0.6 bar and 303 K when the PEG200 content reached 50 wt%. Instead of PEG200, the same group then incorporated different PEG-ethers as additives into the Pebax® MH 1657 matrix [250]. Polyethylene glycol dimethyl ether (PEG-DME) was considered as the best additive among different PEG-ethers, and the membrane separation performance for CO2/H2 separation improved with increasing 86

PEG-DME content, which was similar to that of PEG200. The membrane with 50 wt% PEG-DME showed the optimum CO2 permeability of 606 barrer with the CO2/H2 ideal selectivity of 15.1 at 0.3 bar and 303 K. Compared to the blend of PEG, a more significant improvement was observed for PEG-ethers, which was primarily due to the fact that the ether-end groups in the PEG chains caused more pronounced microphase separation structures. Very recently, other specific CO2-philic additives were suggested to incorporate into Pebax®. Ghadimi et al. developed alloyed Pebax®/PEGDA blend membranes [251]. The author claimed that the presence of PEGDA restricted the mobility of polymer chain and improved the CO2/gas selectivity, therefore the highest CO2/H2 ideal selectivity was about 16 at 40 bar and 298 K in the case of 20 wt% PEGDA added. Glycerol triacetate (GTA) was also used as a plasticizer to tune membrane structures, the optimum CO2/H2 selectivity could be more than 35 with CO2 permeability of about 1200 barrer at 24 bar and 298 K when the content of glycerol triacetate was 60 wt% [252]. The aforementioned EO-containing additives such as PEG and PEG-ether were also used to blend of another commercial CO2-philic copolymer PEO-PBT (Polyactive®). Car et al. blended PEG200 to obtain the tailor-made polymeric membranes based on Polyactive® [240]. The results indicated that a two-fold improvement in CO2 permeability (~192 barrer) was achieved by the addition of 50 wt% PEG, and the corresponding CO2/H2 ideal selectivity increased to 13.2 at 308 K. To extend the research, Yave et al. fabricated nanostructured CO2-philic polymeric membranes via blending of Polyactive® with two polyethylene glycol ethers [253]. The incorporation of polyethylene glycol dibutylether (PEG-DBE) improved 87

CO2 solubility and CO2 diffusivity in membranes simultaneously, achieving the CO2 permeability of 750 barrer and the CO2/H2 ideal selectivity of 12.4 at 0.3 bar and 303 K. To push the limits of PEO containing block copolymer membranes for CO2 separation, the most permeable block copolymer consists of PEO-ran-PPO soft segments and T6T6T hard segments, was chosen to be the membrane matrix, meanwhile, a PDMS based oligomer grafted with low molecular-weight PEG was selected as an additive [254]. Therefore, a highly permeable PEO-based blend membrane was successfully developed, which showed the maximum CO2 permeability of 896 barrer with the CO2/H2 ideal selectivity of 10.6 at 4 bar and 308 K. d. PEO-based composite membranes The PEO-based membranes introduced above were almost homogeneous dense membranes, the thickness of which was generally in the range of dozens to hundreds of microns. The homogenous membrane is an excellent choice to investigate gas transport properties of membrane materials. However, these thick membranes are uncompetitive for industrial applications due to the extremely low gas permeance. Generally, an ultrathin defect-free selective layer is essential for fabricating the membrane with a high gas permeance. Several research institutions have successfully developed various high-performance composite membranes with an ultrathin PEO-based selective layer [253, 255-258]. Membrane Technology and Research, Inc. (MTR) of USA developed a high-performance PEO-based composite membrane (PolarisTM) which exhibited the CO2 permeance of 1000 GPU with the CO2/N2 selectivity of 50 [255]. In 2012, the second generation of PolarisTM membrane 88

possessed an improved CO2 permeance of 2200 GPU with the similar CO2/N2 selectivity.[256] The membrane has been fabricated into semi-commercial spiral-wound modules and tested on a pilot-scale membrane skid treating a real shifted synthesis gas at the National Carbon Capture Center (NCCC) of USA, and the membrane module showed the CO2 permeance of 100-300 GPU with CO2/H2 selectivity of 6-10 [2]. It is encouraging that the module performance was stable over time which is of great importance for practical applications. Moreover, GKSS Research Centre Geesthact of Germany developed another PEO-based composite membrane where a thin PEO-PBT selective layer deposited on the PAN porous support coated with a crosslinked PDMS layer via dip coating method [257, 258]. In this work, the high permeable PDMS layer worked as an intermediate layer. The introduced intermediate layer prevented the phenomenon of pore blocking and rendered the membrane surface smoother, which was beneficial for the formation of a defect-free ultrathin selective layer. It should be noted that PolarisTM composite membrane consists of a PDMS intermediate layer as well. As a result, a ultrathin PEO-PBT selective layer of about 45 nm could be obtained via decreasing the concentration of coating solution. As a result, the CO2 permeance and CO2/H2 selectivity were about 1777 GPU and 9.0 at 293 K, respectively. The experimental CO2 permeance was almost two-fold higher than theoretically calculated value by the gas permeability and thickness. This unconventional behavior was mainly due to a very thin membrane thickness that was comparable to the polymer radius of gyration (Rg) (Fig. 23). In this case, the chain conformation of polymer was seriously perturbed, which resulted in the enhanced chain mobility and increased fractional free volume [258]. This membrane has also 89

been manufactured on square meter scale, and it exhibited a favorable separation performance similar to lab-scale membranes. The successful fabrication of high-performance composite membranes and membrane modules demonstrates that PEO-based membranes have a great potential in CO2 removal for hydrogen production.

Fig. 23. (a) Schematic representation of the block copolymer organization in thick films (semi-crystalline polymer and high Tg); (b) representation of a polymer chain and Rg; and (c) block copolymer organization within a super ultrathin film under the influence of the PDMS substrate (thin film mostly amorphous with high fractional free volume and low Tg of the PEO segment) [258].

Our group also prepared a PEO-based composite membrane where an EO units containing crosslinked network was developed on the porous support coated with the crosslinked PDMS layer via interfacial polymerization between diamine and trimesoyl chloride

(TMC)

(Fig.

24)

[259].

Different

diamines

named

diethylene

glycol

bis(3-aminopropyl) ether (DGBAmE, EO-3) and diaminopolyethylene glycol (DAmPEG, EO-21) were used as aqueous phase monomers, and effects of the length and concentration of 90

EO units on membrane performance were investigated. The optimum CO2 permeance of 815 GPU and CO2/H2 selectivity of 10 could be obtained at 1.1 bar and 298 K. As a mature membrane fabrication technology, interfacial polymerization has been commercially used for fabrication of reverse osmosis membranes on a large scale. Therefore, it may be convenient to achieve the scale-up manufacture of this PEO-based composite membrane.

Fig. 24. Polymerization reaction for preparation of the thin film composite membranes by interfacial polymerization [259].

In addition to flat sheet composite membranes above, PEO-based hollow fiber multilayer composite membranes with Pebax® MH 1657 as a selective layer were developed recently [260]. The membrane exhibited a CO2 permeance of 362 GPU with the CO2/H2 selectivity of 7.8 at 2 bar and 298 K. Compared to spiral wound membrane modules based on flat sheet membranes, hollow fiber membrane modules may be more appropriate for large-scale application owing to the high packing density.

91

4.3.2 Ionic liquid-based membranes Room temperature ionic liquids (RTILs),liquid-phase organic salts at room temperature, are another kind of materials with a strong affinity towards CO2 [261, 262]. Compared to conventional CO2 absorbents such as amine, RTILs possess inherent advantages of extremely low vapor pressure and good thermal stability. Therefore, RTILs were first suggested as absorbents for CO2 separation instead of amine solutions. Many studies have demonstrated that the combination of cation and anion modification of RTILs could allow tunable gas solubility and solubility selectivity [263]. However, the liquid nature of RTILs makes them difficult to form a solid-state film. To address this problem, one approach was to blend polymers with RTILs. For example, Chen et al. prepared poly(vinylidene fluoride)-RTIL blend membranes, and effects of RTIL content on the membrane performance were investigated [264]. When the mass ratio of PVDF to RTIL was 1 : 2, the optimum CO2 permeability of 1595 barrer with the CO2/H2 selectivity of 11.7 could be obtained at 4 bar and 308 K. Another approach was the synthesis of poly(ionic liquid)s (PILs) which are the solid phase at room temperature. The study revealed that PILs exhibited higher CO2 absorption capacity than the corresponding RTILs unexpectedly [265]. The Noble’s group developed a series of PILs-based membranes for CO2 separation [266, 267]. The result suggested that gas permeability of PILs was several orders of magnitude lower than analogous RTILs, which was probably due to a significant decrease in gas diffusion coefficient in membranes from liquid phase to solid phase. The same group then suggested a new-type membrane of PIL-RTIL blend membranes to enhance membrane performances (Fig. 25) [268]. The introduced RTILs 92

could facilely interact with penetrating CO2 molecules, meanwhile, they were relatively stable in the membranes owing to electrostatic interactions with PIL matrix. The resultant membranes displayed an obvious enhancement in CO2 permeability and a minor deterioration in CO2/gas selectivity compared to pristine PILs membranes. There are very limited data about PIL-based membranes for CO2/H2 separation, which might be due to the low CO2/H2 selectivity estimated according to the CO2/N2 and CO2/CH4 selectivity of some PIL-based membranes reported. Functionalization of ILs and PILs via introducing reactive functional groups with CO2 is an effective strategy to notably improve separation performance, the detailed information about which can be found in section 4.4.

Fig. 25. Representations of (a) poly(RTIL) framework with polymer-bound cations and (b) poly(RTIL)-RTIL composite containing 20 mol % free cations [268]

4.4 Facilitated transport membranes Generally, gas permeation through dense membranes fabricated from conventional polymeric materials, follows the solution-diffusion mechanism. The introduced CO2-philic functional groups could render the membranes solubility-controlled for CO2/H2 separation. However, there merely exist the relatively weak physical interactions between these polar 93

functional groups and CO2 molecules. As a consequence, the CO2/H2 selectivity of most CO2-philic polymeric membranes reported to date was no more than 15, except the crosslinked PEO membrane developed by Freeman et al. which showed a high CO2/H2 selectivity (~31) at extremely low temperature (~ 253 K). The relatively low CO2/H2 selectivity seriously hampers CO2-philic polymeric membranes to meet the requirements for high-purity hydrogen production. Inspired by the biological membrane that it can selectively transport specific small molecules or ions through facilitated diffusion via carrier proteins, thus facilitated transport membranes have been extensively and intensively investigated over the past decades. In terms of the mobility of carriers, facilitated transport membranes can be principally divided into two types: (1) mobile carrier membranes including liquid membranes and supported liquid membranes, (2) fixed carrier membranes. Very recently, an excellent review about energy-efficient facilitated transport membranes has been published [269]. Owing to the distinctive transport mechanism for specific molecules and ions, facilitated transport membranes offer an attractive method to simultaneously achieve high permeability and high selectivity. The high-efficiency transport of target molecules or ions is commonly based on reversible interactions of target molecules or ions with carriers. Ward et al. for the first time introduced the facilitated transport membrane for CO2 separation [270]. In this pioneering work, an immobilized film of an aqueous bicarbonate-carbonate solution was developed which was 4100 times more permeable to CO2 than to O2 due to the reversible reaction between carbonate group and CO2. After that, various facilitated transport membranes for CO2 separation have been successfully fabricated. In these membranes, carbonate group 94

(CO32-), fluorion (F-), carboxylate group (COO-) and amino groups are common carriers, and they all could reversibly react with CO2 according to the nucleophilic addition mechanism. Among them, amino groups are the most typical and widely used carriers in facilitated transport membranes for CO2 separation. The reactions between CO2 and primary or secondary amino groups are based on the zwitterion mechanism that converts CO2 into carbamate ion, and then carbamate ion can be further hydrolyzed to bicarbonate ion in the presence of water. These reactions are illustrated in the following eq. (7)-(10). RR H

2

RR H

-

RR H

RR H

-

H2

-

RR

-

RR H

H2

(7) -

RR -

RR RR H

H

RR H2

H

(8) (9)

-

(10)

where R' represents an H atom or other organic groups. Different from primary or secondary amino groups, tertiary amino group cannot directly react with CO2 in the absence of water, while it could act as a weak base catalyst for CO2 hydration in the presence of water and the reaction formula is shown in eq. (11). R

H2

2

R

H

H

-

(11)

Similar to that of tertiary amino group, the reactions of carbonate group and carboxylate group with CO2 are illustrated in eq. (12) and (13), respectively. 2-

-

2

2

H2 H2

-

2H H

(12) H

-

(13)

The early works about facilitated transport membranes for CO2 separation are primarily 95

liquid membranes and supported liquid membranes. These membranes usually exhibit an extremely high CO2/gas selectivity at the extremely low CO2 feed pressure, because highly mobile carriers could facilely interact with penetrating CO2 molecules. However, the most conventional materials containing carriers used in liquid membranes and supported liquid membranes are various amines, and they are generally volatile and very free in membranes, hence the potential loss of carriers may be a serious problem under elevated temperature and high transmembrane pressure. This drawback strongly limits the industrial application of mobile carrier membranes, even though supported liquid membranes were suggested that they could circumvent this problem to a certain extent. In recent years, some carrier-functionalized ionic liquids, so called task-specific ionic liquids (TSILs), have emerged as novel materials for CO2 separation. Compared to conventional ionic liquids introduced in section 4.3.2, TSILs generally show a higher CO2 absorption capacity owing to strong interactions between carriers and CO2. Therefore, various supported liquid membranes containing ionic liquids have been developed for facilitating CO2 transport [271-273]. The resultant membranes showed the excellent separation performances at low feed pressure. In particular, some amino acid ionic liquids containing membranes exhibited the similar performance under humid and dry conditions, which was greatly different from most facilitated transport membranes which displayed unfavorable separation performances in the absence of water. The reason for this interesting phenomenon is still unclear, and it might be related to the interaction mechanism between TSILs and CO2. However, these TSILs in supported liquid membranes could still possibly be lost at a sufficient pressure difference due to the lack of strong interactions with 96

supports. To fully solve the aforementioned problem in mobile carrier membranes, fixed carrier membranes have been developed vigorously over the past ten years. In fixed carrier membranes, carriers are generally covalently bounded with the polymer backbone, thus they show much better stability than that in mobile carrier membranes. A series of amino groups containing polymers have been synthesized and used for fabrication of fixed carrier membranes for CO2 separation. The most common ones are PEI, polyvinylamine (PVAm), polyallylamine (PAAm) and polyamidoamine (PAMAM) dendrimers, the chemical structures of which were presented in Fig. 26 [274-278]. Similar to polymers with abundant polar EO groups, amino group-rich polymers usually have a strong tendency to crystallization due to the strong hydrogen-bonding interactions among amino groups, which is unfavorable for high-performance facilitated transport membranes. High crystallinity not only could result in the decrease in gas permeability, but also could decrease the utilization level of carriers that only these carriers in the amorphous phase could interact with penetrating CO2 molecules. In addition, many commercial available polymers containing abundant amino groups show the poor thin-film-forming property on porous supports, which is probably due to the relatively low molecular weight.

97

Fig. 26. Chemical structure of representative amino groups containing polymers.

Polymer blending was regarded as a simple but effective method to address two problems above. Polyvinyl alcohol (PVA) has been selected to blend with various amino group-rich polymers for fabrication of fixed carrier facilitated transport membranes owing to its good compatibility with polyamines and good film-forming ability [274, 275, 279, 280]. For instance, Deng et al. developed PVAm/PVA composite membranes on PSf porous supports [279]. The good mechanical property was obtained due to the entanglement of different polymer chains, enabling a thin selective layer of 500 nm. The resultant membrane displayed the superior CO2 permeance and CO2/gas selectivity. Similarly, different generations of PAMAM dendrimers were immobilized in a PEG network, and effects of the type and content of PAMAM dendrimers on membrane performances were investigated [281]. 98

Especially, the membrane containing 50 wt% PAMAM dendrimer (0th generation, n=1) exhibited the optimal CO2/H2 selectivity of 500 with the CO2 permeability of 3650 barrer at 1 bar and 298 K. A significantly high CO2/H2 selectivity was obtained from the membrane with the thickness of 500 μm. However, there was a serious deterioration in CO2/H2 selectivity because of the obvious leakage of the dendrimer, when the membrane thickness reduced to less than 100 μm [282]. In most cases, the optimized structure and property of facilitated transport membrane materials via polymer blending is at the expense of reduced carrier concentration in membranes. In recent years, our group introduced a novel method that various amino group-containing small molecules were used to modify amino group-rich polymers via hydrogen-bonding crosslinking, which not only could achieve the optimization of membrane material structures at multiple levels, but also could increase the carrier concentration in the membranes [283-287]. Following this strategy, ethylenediamine (EDA), piperazine (PIP), monoethanolamine (MEA) and methylcarbamate (MC) were used as modifiers to incorporate into the PVAm matrix, and the formed hydrogen-bonding crosslinked networks were illustrated in Fig. 27 [287]. Effects of physicochemical properties and content of small molecular amines on membrane performances were intensively studied. The results suggested that the MEA-modified membrane exhibited the optimum CO2 permeance of 2660 GPU with the CO2/H2 selectivity of 63 at 1.1 bar and 323 K. It should be noted that the CO2/H2 selectivity was still 21 at the feed pressure of 30 bar, showing the favorable separation performance at high pressure. Bai et al. developed a complex membrane that both mobile 99

carriers and fixed carriers were introduced into the crosslinked PVA matrix [288]. Effects of the ratio of mobile carriers to fixed carriers on membrane performances were investigated. The mobile carriers could contribute more to facilitated transport of CO2 than fixed carriers owing to the higher mobility. However, excess mobile carriers could result in poor mechanical strength. The optimum CO2 permeability was close to 1000 barrer with the CO2/H2 selectivity higher than 35 at 379 K. The blended small molecules with carriers in these membranes showed much better stability than in supported liquid membranes owing to the good compatibility between polar small molecules and polar polymer matrix. In all, in addition to optimizing the membrane structures, the introduction of small molecules with amino groups notably increased the carrier concentration in the membranes, achieving a considerable performance improvement of facilitated transport membranes for hydrogen purification.

Fig. 27. Schematic diagram of the intermolecular hydrogen bonds in the small molecule amine-modified PVAm samples: (1) EDA-modified PVAm; (2) PIP-modified PVAm; (3) MEA-modified PVAm; (4) MC-modified PVAm [285].

Besides the carrier concentration in membranes, the physicochemical microenvironment 100

of carriers plays an important role in determining membrane separation performances, because it could influence reaction kinetic and reaction stoichiometry between carriers and CO2. Kim et al. investigated effects of pH of casting solutions on separation performances of PVAm membranes [289]. Higher pH could result in an increased number of primary amino groups available to interact with CO2, thus the membrane performance could be notably enhanced. Zhao et al. for the first time demonstrated the amine steric hindrance effect in a solid polymeric membrane [290]. A sterically hindered amine is defined as either primary amino group in which the nitrogen atom is attached to a tertiary carbon atom or a secondary amino group in which the nitrogen atom is attached to at least one secondary or tertiary carbon atom. For sterically hindered amines, the existence of alkyl group renders the carbamate ion unstable and facilitates the hydrolysis of carbamate ion to become bicarbonate ion, enhancing CO2 loading capacity of primary and secondary amino groups. The Ho’s group synthesized a series of sterically hindered polyamines derived from PAAm to investigate effects of the degree of steric hindrance on membrane performances [291]. The results indicated that the moderately hindered poly-N-isopropylallylamine (PAAm-C3H7) exhibited the greatest CO2 facilitation, achieving a dramatic 440% enhancement in CO2 permeability (~297 barrer) along with a 135% increase in CO2/H2 selectivity (~40) compared to unhindered PAAm at 1 bar and 383 K, which was mainly due to the cooperation of a higher CO2 loading capacity and a larger reaction rate constant. Then the moderately hindered PAAm-C3H7 as fixed carriers and amino acid salts as mobile carriers were incorporated into the crosslinked PVA-poly(siloxane) matrix. The CO2 permeability of 6500 barrer with the CO2/H2 selectivity of 340 could be obtained at 1 101

bar and 383 K. Compared to that of PAAm as fixed carriers, the improvements in CO2 permeability and CO2/H2 selectivity were about 66% and 111%, respectively. Similarly, our group also suggested that the carrier with higher steric hindrance and higher electronegativity should be a suitable choice for improving separation performances at high pressure [292]. For cyclic amino groups from 1,4-bis(3-aminopropyl) piperazine as carriers, there was only a slight decrease in CO2/H2 selectivity when the feed pressure increased from 1.1 bar to 30 bar. Moreover, the efficiency of carriers could be enhanced via synergistic effects of different functional groups [293]. For example, a small amount of amine could notably increase the reaction rate of carbonate groups with CO2. Therefore, the combination of different types of carriers into the membrane is an alternative method to enhance the efficiency of carriers for membrane separation performance improvement. In addition to constructing a suitable carrier microenvironment, it has been demonstrated that the optimization of carrier distribution in membranes could enhance separation performances. Blinova et al. developed surface-modified polyaniline (PANI)-based membranes where various carriers were successfully enriched on the membrane surface via grafting (Fig. 28) [294]. Compared to the carriers in dense membrane bulk, these carriers on the membrane surface could contact and further react with CO2 molecules more readily, thus the efficiency of carriers was greatly enhanced. In particular, the guanidine surface functionalized membrane showed a very high CO2 permeability of 3460 barrer with the CO2/CH4 selectivity of 540. Accordingly, it is expected that this membrane should possess an excellent separation performance for CO2/H2 separation. 102

Fig. 28. Scheme of (a) functionalization of the polyaniline membrane first photografted with glycidyl methacrylate and 2-hydroxyethylmethacrylate and then reacted with diamines; (b) additional functionalization of membrane containing hexamethylenediamine with 2-ethyl-2-thiopseudourea [294].

Most facilitated transport membranes described above were fabricated via solution coating and subsequent drying process. Interfacial polymerization is another alternative approach to develop a thin film composite membrane due to the rapid self-limiting reaction. Hence interfacial polymerization was also adopted to prepare fixed carrier facilitated transport membranes by our group in recent years [295-297]. TMC as organic phase monomer and 3,3-Diamino-N-methyldipropylamine as aqueous phase monomer were first used to form a tertiary amino groups containing crosslinked network on PSf porous supports [295]. 103

Theoretically, the membrane should possess a high gas permeance owing to a thin apparent selective polyamide layer of several hundred nanometers from scanning electron microscope (SEM) images. Nevertheless, the membrane displayed an unfavorable CO2 permeance of 173 GPU at 1.1 bar and 298 K. The resultant low gas permeance was due to serious pore blockage by the synthesized polyamide via interfacial polymerization. The pore blockage caused an increased effective selective layer thickness, and therefore an increased gas transport resistance. Similar to fabrication of ultra-thin PEO-based membranes with high permeance, a highly permeable crosslinked PDMS layer was selected to be coated on porous supports before interfacial polymerization, to avoid the formation of polyamide in the pores of supports. Following this method, our group further prepared another tertiary amino groups containing facilitated transport membrane with a notably improved CO2 permeance of 2905 GPU at 1.1 bar and 298 K [296]. In addition to conventional amino groups, Wang et al. for the first time developed the fixed carrier membranes with carboxylate groups as carriers via solution casting and interfacial polymerization, respectively [297, 298]. Though carboxylate group displays weaker basicity than amino groups, it might be more stable under realistic conditions. Especially, the membrane prepared by interfacial polymerization showed an extremely high CO2 permeance of 5693 GPU at 1.1 bar and 298 K [297]. Moreover, the membrane exhibited an excellent antioxidizability and good acid resistance. It is well known that carbonic anhydrase can catalyze CO2 hydration very quickly in various organisms. However, the easy deactivation of carbonic anhydrase limits its extensive application. Yao et al. developed a novel biomimetic 104

material of poly(N-vinylimidazole)-zinc complex [299]. This complex could possess a similar structure to active center of carbonic anhydrase via controlling the mole ratio of poly(N-vinylimidazole) to zinc. The optimum CO2 permeance could be higher than 1000 GPU combined with the good CO2/gas selectivity at 1.1 bar and 298 K. Moreover, various metal ions as carriers for facilitating CO2 transport have been reported in recent years [300-304]. Lee et al. suggested that the positively polarized copper nanoparticles as carriers could facilitate CO2 transport via the complexation between copper and CO2 [300]. Then the same group for the first time developed a salt-doped polyelectrolyte membrane [301]. The author claimed that the improved membrane performance was attributed to the favorable interaction between K+ and CO2. However, the resultant CO2/gas selectivity of these membranes containing metal ions was still at low level, implying that the contribution of facilitated transport mechanism was very limited. To enhance the efficiency of K+ as carriers, fluorion was replaced by fluorosilicate anion [304]. The enhanced separation performance was primarily due to the formation of hydrated fluorosilicate anion and the increase in binding energy of K+. Very recently, a series of alkali and alkaline-earth metal salts were doped into Pebax® MH 1657 for CO2 separation [303]. A weak facilitated transport effect was observed and the membrane performance improvement is strongly related on the hydration energy of metal ions. As a consequence, the CaCl2-doped membrane showed the optimum CO2 permeability of 2030 barrer with the high CO2/gas selectivity at 2 bar and 298 K. Moreover, effects of the content and state of water on separation performances of facilitated transport membranes were investigated. The results suggested that free water was responsible for gas permeability while 105

bounded water affected gas selectivity. However, this conclusion was merely based on a simple mathematical fitting method, and its universality remains to be proven. More efforts should be taken to explore the role of water in facilitated transport membranes. A key concern for practical application of facilitated transport membranes for CO2 separation is the stability under realistic conditions. The stability issue mainly includes the long-term stability and the resistance to potential contaminants. Real shifted synthesis gas stream usually contains some minor contaminants such as H2S, CO and CH4. These contaminants could possibly affect the performance of facilitated transport membranes via degrading carriers, competitive sorption with CO2 or competitive reaction with CO2. Therefore, effects of various potential impurities on facilitated transport membranes containing amino groups and carbonate groups have been widely investigated by several research groups [284, 291, 305-308]. According to these studies, there was a negative effect of minor contaminants on membrane performances observed. Not only the CO2 permeance but also the CO2/gas selectivity decreased to a certain extent in the presence of minor contaminants. Fortunately, the separation performance could remain stable. Meanwhile, the performance could gradually recover after contaminants removal from feed gas. Very recently, our group intensively investigated the effects of representative acidic contaminant SO2 on gas transport properties of amino groups containing facilitated transport membranes via experimental and modeling study [307]. These investigations all suggested that the developed facilitated transport membranes showed the good tolerance against potential contaminants in real gas streams. Moreover, the pilot scale testing of PVAm-based facilitated transport membranes for CO2 capture from coal 106

fired power plants were conducted in Nanoglowa EU project [309]. There was no obvious separation performance deterioration observed during more than half a year’s continuous operation under very harsh and challenging conditions, revealing great potential of facilitated transport membranes in large-scale practical application.

4.5 CO2-selective mixed matrix membranes Different from the guideline to design H2-selective MMMs, the main aim for CO2-selective MMMs is to disrupt polymer chain packing and to enhance CO2 loading capacity of membranes. Generally, the introduced inorganic fillers could effectively suppress the crystallization of polar polymeric materials for CO2-selective membranes. In addition, inorganic fillers, especially amino group-functionalized fillers, could possess higher CO2 adsorption capacity than H2. Therefore, MMMs should be an appropriate choice for CO2-selective membranes. Moreover, MMMs commonly display better mechanical property than polymeric membranes, especially CO2-selective polymeric membranes fabricated from rubbery polymers. As expected, various CO2-selective MMMs have been developed successfully over the past years. It was worth noting that porous materials have been widely used in MMMs for different separation applications, while some of them might be not suitable for CO2-selective MMMs for hydrogen purification, because most porous materials usually show molecular sieving properties that benefits H2 preferential permeation. Only porous materials with the pores that could achieve CO2 preferential transport are suitable fillers for CO2-selective MMMs. Among various CO2-selective MMMs, PEO-based MMMs and facilitated transport MMMs generally exhibited more excellent separation performances, thus 107

they were introduced in the following. 4.5.1 PEO-based MMMs Physical blending method is the most widely used method where inorganic fillers are prepared prior to membrane fabrication and then physically dispersed into polymer matrix. The most benefit of this method is that it allows an independent synthesis strategy towards inorganic fillers regardless of membrane fabrication conditions. Theoretically, almost all the inorganic fillers could be incorporated into membranes via physical blending method. Therefore, a series of inorganic fillers, such as fumed silica, carbon nanotube and polyhedral oligomeric silsesquioxane (POSS), have been used for fabrication of PEO-based MMMs for hydrogen purification [310-314]. Li et al. developed molecular-level MMMs consisted of Pebax® MH 1657 and two kinds of POSS cages (hydroxyl group-functionalized POSS and amic acid-functionalized POSS) [311]. At low filler loading, a simultaneous improvement of CO2 permeability and CO2/H2 selectivity was observed owing to the increased free volume, while the performance degraded caused by polymer chain rigidification at high loading. The highest CO2/H2 selectivity of 52.3 combined with the CO2 permeability of 136 barrer could be achieved at 8 bar and 308 K, when amic acid-functionalized POSS cage was 1 wt%. In addition, PEG-functionalized POSS cage was suggested as inorganic fillers to blend of Pebax®, a two-fold increase in CO2 permeability was shown after the incorporation of 30 wt% PEG-POSS [314]. Zhao et al. incorporated amino group-functionalized multi-walled carbon nanotubes (MWCNTs) into Pebax® MH 1657 matrix [313]. The gas permeability increased with increasing filler content, but there was 108

no obvious change in CO2/H2 selectivity (~9). Though physical blending is a very simple method to develop MMMs, there are some inherent drawbacks including filler agglomeration and non-selective interface voids, which potentially restricts the fabrication of membranes with high CO2/H2 selectivity. H2 molecule with the smaller kinetic diameter may be more sensitive to non-selective interface voids compared to CO2 molecule, thus the existence of voids could result in a more notable increase in H2 permeability. Consequently, an obvious deterioration in CO2/H2 selectivity would be shown unfavorably. To address these problems of physical blending method, sol-gel method has been suggested to develop MMMs. In this method, inorganic fillers are in situ fabricated within polymer matrix via the hydrolysis of precursors, which could promote the dispersion of inorganic phase. However, only a few types of fillers such as silica and metal oxide could be incorporated into the membranes via this method. Sforca et al. for the first time prepared MMMs for CO2 separation via sol-gel reaction between polyether diamine with epoxy silane [315]. The introduced inorganic silica phase not only resulted in the decrease in crystallinity but also contributed to the improvement in mechanical properties of membrane materials. Hence the CO2 permeability of 125 barrer with the CO2/H2 ideal selectivity of 9 could be obtained at 4 bar and 303 K. After that, the Chung’s group developed a series of MMMs for hydrogen purification using sol-gel method. Shao et al. prepared the MMMs similar to that of Sforca et al., and membrane separation performance was further optimized via tuning the ratio of organic component to inorganic component [316]. When the inorganic component was 10 wt%, the CO2 permeability was 367 barrer combined with the 109

corresponding CO2/H2 ideal selectivity of 8.95 at 3.5 bar and 308 K. Moreover, a better separation performance could be achieved via increasing the content of organic component at the expense of the reduced mechanical strength. To enhance CO2 permeability while maintaining enough mechanical strength, Lau et al. proposed to graft short PEG-based chains on the aforementioned organic-inorganic material via wet ozonolysis treatment (Fig. 29) [317]. The transformation of the inorganic phase from a well-dispersed network of finely defined fillers to rough clusters was primarily responsible for a significant improvement in CO2 permeability [318]. Therefore, the maximum CO2 permeability approached 2000 barrer with the CO2/H2 selectivity of 11 at 3.5 bar and 308 K. In addition, various PEG molecules with different molecular weights were used to further optimize the separation performance of this organic-inorganic material [319]. Effects of molecular weight and content of PEG molecules were investigated, and the highest CO2 permeability of 845 barrer could be achieved when the content of PEG1000 reached 60 wt%. Another organic-inorganic hybrid network was established via reacting polyether diamine with epoxy-functionalized POSS cage, by which could achieve more uniform dispersion of POSS cages than aforementioned physical blending method [320]. Similarly, the introduced POSS cages suppressed the crystallinity of polyether diamine, and consequently increased CO2 permeability.

110

Fig. 29. Schematic representation of the side-chain grafted organic-inorganic network [317].

4.5.2 Facilitated transport MMMs Most facilitated transport MMMs reported were developed via physical blending method. Zhao et al. synthesized a series of PANI nanoparticles with different morphologies and subsequently incorporated them into the PVAm matrix [321]. The introduced PANI nanoparticles not only could make N2 transport pathway more tortuous but also could facilitate CO2 transport owing to the specific interaction with CO2 molecules in the presence of water. Therefore, the MMMs exhibited an obvious enhancement in CO2 permeance and CO2/gas selectivity compared to pristine PVAm membrane. Furthermore, CO2-faciliated transport highway was successfully constructed in the membrane via incorporating PVP-modified PANI nanorods, by which could render CO2 transport more efficient (Fig. 30) [322]. As a result, the 111

maximum CO2 permeance of 3080 GPU with the excellent CO2/gas selectivity could be obtained at 1.1 bar and 298 K. Similarly, Liao et al. recently designed and synthesized hydrotalcite (HT) nanosheets that consist of positively charged brucite-like host layers and hydrated carbonate groups moving through the unobstructed interlayer gallery (Fig. 31) [306]. These free-movable carbonate groups could facilely interact with CO2 molecules, which was similar to mobile carrier membranes. However, the carriers could not be lost owing to strong electrostatic interactions with positively charged host layers. Then the stable high-speed facilitated transport channels were established in the membrane via grafting to the amino group-rich polymer matrix using a molecular bridge of APTES, which not only combined the advantages of mobile carrier membranes and fixed carrier membranes but also circumvented the disadvantages of them. Therefore, the membrane showed an extremely high CO2 permeance of 5693 GPU with the favorable CO2/gas selectivity at 1.1 bar and 298 K.

Fig. 30. Schematic representation of the CO2-facilitated transport highway in the PVAm-PANI selective layer, (a) intra-channel CO2 transport and (b) inter-channel CO2 transport [322]. 112

Fig. 31. Schematic diagrams of (a) hydrotalcite containing free movable hydrated carbonate groups and (b) facilitated transport membrane containing high-speed facilitated transport channels [306].

In addition to improving membrane separation performances, the incorporation of inorganic fillers could enhance high pressure resistance of polymeric membranes. Because facilitated transport membranes show better separation performance under humid conditions, they are generally operated under highly water-swollen state. However, there might be an 113

obvious membrane compaction phenomenon occurred for highly swollen membranes at elevated pressure, especially for the membrane with a thin selective layer. This phenomenon could result in a notable decrease in gas permeance. Therefore, carbon nanotubes as mechanical reinforcing fillers were suggested to blend into facilitated transport membranes by several researchers [323, 324]. The MMMs containing amines and a small amount of MWCNTs were developed by Zhao et al. A significant improvement in membrane stability was achieved under harsh conditions simulating hydrogen purification from shifted synthesis gas, and there was no change in separation performance during 450 hours at 15 bar and 380 K. Similarly, Deng et al. also developed MWCNT reinforced PVAm/PVA facilitated transport MMMs [323]. With the addition of a small amount of MWCNTs, the membrane exhibited a good durability against compaction effect in operation at high pressure up to 15 bar. Therefore, the MMM showed a nearly two-fold increase in CO2 permeance at 15 bar compared to the polymeric membrane without MWCNTs.

4.6 Comparison of CO2-selective membranes Fig. 32 presents the separation performances of representative CO2-selective membranes for CO2/H2 separation plotted on the upper bound suggested by Freeman [138, 228]. A very limited number of CO2-selective microporous membranes reported exhibit a competitive CO2/H2 selectivity merely at extremely low temperature. Due to great difficulties in fabricating CO2-selective microporous membranes, most CO2-selective membranes available are dense polymeric membranes to date, and they often need to be operated at a relatively low temperature in order to obtain a relatively high CO2/H2 selectivity. PEO-based membranes are 114

the most widely studied CO2-selective polymeric membranes, and the CO2/H2 selectivity of most PEO-based membranes is not favorable (no more than 15) besides at the extremely low temperature, which is probably due to the fact that CO2 solubility coefficient decreases substantially with increasing temperature. However, higher CO2/H2 selectivity is required for high-purity hydrogen production. PIM membranes exhibit a considerably high CO2 permeability. Nevertheless, the CO2/H2 selectivity of PIM membranes is still to be improved to meet the requirements for practical application. Facilitated transport membranes offer a choice to simultaneously achieve high permeability and high selectivity, and many facilitated transport membranes have displayed an excellent performance for CO2/H2 separation. Though the separation performances of most facilitated transport membranes deteriorate with increasing feed pressure, the CO2/H2 selectivity of some facilitated transport membranes could remain at high level and is still higher than most other CO2-selective membranes. Moreover, facilitated transport membranes exhibit good temperature resistance up to about 373 K. In addition, some representative facilitated transport membranes have been demonstrated the good long-term stability and resistance against contaminants. These results about facilitated transport membranes are generally obtained under lab-scale ideal conditions, therefore, more pilot-scale investigations under realistic conditions are necessary to accurately evaluate the feasibility of facilitated transport membranes for hydrogen purification.

115

Fig. 32. Separation performances of representative CO2-selective membranes for CO2/H2 separation (square: microporous membranes, triangle: PIM membranes, star: CO2-philic polymeric membranes, circle: facilitated transport membranes, diamond: mixed matrix membranes). The data could be seen in Table S2 in supporting information.

5. Comparison of H2-selective membranes and CO2-selective membranes Various membranes for hydrogen purification possess their prominent advantages and corresponding drawbacks, regardless of H2-selective membranes or CO2-selective membranes. For H2-selective membranes, one of the best benefits is that they can be operated at elevated temperature, because WGS reaction generally occurs in the high-temperature range. Moreover, H2 transport is based on the molecular sieving mechanism or dominated by diffusivity selectivity in H2-selective membranes except in dense metallic membranes, since H2 molecule 116

possesses the smallest kinetic diameter. This characteristic could result in the facile achievement of hydrogen enrichment in the permeate side even if there exist some minor contaminants which possess larger kinetic diameter than H2 in feed gas, provided that membranes are not damaged by contaminants. However, one obvious limitation for H2-selective membranes is that the obtained hydrogen product is nearly at atmospheric pressure in the permeate side, thus highly energy-intensive recompression is often necessary for subsequent hydrogen storage and utilization, which might offset the strength of membrane-based

gas

separation

technology.

Moreover,

CO2-induced

plasticization

phenomenon could possibly cause a notable decrease in H2/CO2 selectivity for H2-selective polymeric membranes at a high operational pressure. In terms of CO2-selective membranes, the most notable advantage is the circumvention of hydrogen recompression, since hydrogen is obtained at high-pressure retentate side. Moreover, different from H2-selective polymeric membranes, CO2-induced plasticization could potentially improve the separation performance of CO2-selective polymeric membranes [228]. Nevertheless, the cooling treatment of gas streams from WGS reactors is inevitability required, because CO2-selective membranes usually exhibit more excellent separation performances at a relatively low temperature. In addition, the potential contaminants in real gas streams could possibly affect the performance of CO2-selective membranes owing to the competitive sorption or/and competitive reaction with CO2, especially for facilitated transport membranes. Till now, most studies about membrane process for H2-CO2 separation focused on the application in pre-combustion CO2 capture [7-9]. In pre-combustion CO2 capture, the purity 117

and recovery of CO2 are the key parameters to evaluate the feasibility of separation process, while high-purity hydrogen is not necessary for subsequent combustion to produce power. Many researchers have suggested that pre-combustion CO2 capture based on membrane system is technically possible, and membrane selectivity should be further improved to compete with a conventional absorption process. Franz et al. proposed that the H2/CO2 selectivity of at least 50 or the CO2/H2 selectivity of at least 60 was required for a competitive membrane process for pre-combustion CO2 capture [7]. Different from pre-combustion CO2 capture, the purity and recovery of H2 are however vital in terms of hydrogen production, which directly determines the quality and cost of hydrogen products. Therefore, the systematic evaluation of membrane technology for high-efficiency hydrogen production should be conducted based on different membrane systems using H2-selective membranes and CO2-selective membranes. Very recently, we intensively discussed the membrane performance requirements for hydrogen production with a two-stage membrane process via modeling study [11]. When the selectivity values of H2-selective membranes and CO2-selective membranes are identical, high-purity hydrogen could be more conveniently obtained using CO2-selective membranes compared to H2-selective membranes. Moreover, it is difficult to acquire ultrahigh-purity hydrogen accompanied by a relatively high hydrogen recovery using H2-selective membranes. A slight increase in hydrogen purity could result in an obvious decrease in hydrogen recovery when the purity is at high level (> 99.9 vol%). These results match the characteristic of membrane separation process that high-purity product is difficult to be obtained in the permeate side, because the membrane selectivity is generally a 118

finite value [7]. Therefore, much higher H2/CO2 selectivity for H2-selective membranes is required for hydrogen production system than IGCC system because of high-purity hydrogen requirement. It could be concluded that CO2-selective membranes is more favorable than H2-selective membranes for high-purity hydrogen production from the point of technological feasibility. In addition, it has been suggested that the combination of H2-selective membranes and CO2-selective membranes may be a more promising strategy via designing advanced membrane process [2]. However, it should be noted that it is very challenging to simultaneously achieve high hydrogen purity & recovery based on current available membrane modules regardless of H2-selective or CO2-selective membranes, even if some reported lab-scale membranes display enough good separation performances. Nowadays, membrane process used alone for hydrogen production from shifted synthesis gas is therefore less competitive. A feasible strategy is that membrane technology could couple with other separation techniques to take advantage of the benefits and avoid the drawbacks of every technique. As a rapidly emerging opportunity, biohydrogen production presents a hot research topic of hydrogen energy. However, the intensive investigation about membrane performance requirements for biohydrogen enrichment is scare. Compared with conventional hydrogen production techniques, the notable benefit of biohydrogen production is that it could proceed under ambient conditions (nearly ambient temperature and pressure) where most membranes could be well operated. Therefore, the integration of hydrogen production and purification is a promising approach to enhance the efficiency of hydrogen production. It is expected that the 119

required membrane performance for biohydrogen enrichment could be lower than that of hydrogen purification from shifted synthesis gas, especially when the integration of biohydrogen production and purification is realized. Some commercial available membrane modules have been used for biohydrogen production [325-327]. Therefore, membrane-based gas separation technology may achieve widespread application in biohydrogen production in advance, promoting the increased biohydrogen share in the global hydrogen market.

6. Conclusion and outlook Hydrogen energy has been widely considered as a promising energy in the future, which greatly pushes forwards the innovation of separation techniques for high-quality hydrogen production. Membrane-based gas separation technology has exhibited inherent advantages over conventional separation techniques for hydrogen purification. Therefore, an increasing number of advanced membranes made from various materials including inorganic materials, polymers and organic-inorganic hybrid materials, have been successfully developed via different novel membrane fabrication methods, some of which have already showed favorable performances for H2-CO2 separation expectedly. Several representative low-cost membrane materials (e.g. PEO-based copolymers and polyvinylamine) with excellent solution processing properties have been fabricated into large-area composite membranes with high performances. Moreover, carbon molecular sieve hollow fiber membranes and TR polymer hollow fiber membranes have been prepared. Furthermore, the corresponding membrane modules have been manufactured. In particular, ProteusTM and PolarisTM membrane modules have been successfully used on the membrane-based demonstration system for CO2 removal from shifted 120

synthesis gas. Despite the achievements described above, there is still a long way to go in achieving large-scale industrial application of membrane-based gas separation technology for hydrogen purification. The performance of most reported membranes for H2-CO2 separation, especially the selectivity, still could not fully meet the requirement for high-quality hydrogen production. In order to enhance the technological and economic competiveness of membrane process, some challenges which are to address for both the scientific community and the engineering community, are highlighted as follows. For H2-selective membranes, the sole high-efficiency gas transport mechanism is the molecular sieving mechanism for microporous membranes. The accurate control of pore structures including pore size, shape and pore size distribution is a key issue to completely exert molecular sieving effects. Fortunately, this demand is exactly the most notable characteristic of MOFs. Accordingly, MOFs may be the most appropriate candidate for molecular-sieving H2-selective membranes compared to conventional porous materials. It is expected that highly molecular-sieving MOF membranes could come out via exploring new fabrication methods. Moreover, porous graphene-based membrane is another kind of promising H2-selective microporous membranes if enough uniform sub-nanometer pores could be conveniently generated. As for H2-selective polymeric membranes, membrane separation performance is mainly dominated by diffusivity selectivity. Therefore, microporous polymers with high fractional free volume but narrow free volume distribution might be best choice via further enhancing polymer chain rigidity. In addition, how to substantially cut down the 121

consumption of precious metal is vital for large-scale industrial application of metallic membranes for hydrogen purification. For CO2-selective membranes, the most common ones are CO2-philic polymeric membranes. In theory, these membranes could simultaneously achieve high CO2 permeability and high CO2/H2 selectivity shown in the suggested CO2/H2 upper bound, because their performances are commonly dominated by solubility selectivity. Therefore, new functional groups or units with stronger CO2 affinity should be explored to further enhance CO2 solubility coefficient and CO2/H2 solubility selectivity via the combination of experimental and simulation study. In addition, great efforts should be taken to render diffusivity selectivity close to unity, maximizing the efficiency of solubility selectivity. It is anticipated that a perfect material perhaps should possess the property of liquid-like permeability and selectivity in a solid state. Facilitated transport membrane seems a more promising one owing to high selectivity towards CO2, especially at low pressure. Facilitated transport effects are expected to be further enhanced through the following two strategies so as to improve pressure resistance. One is to enhance the efficiency of carriers via designing new carriers possessing high reaction rate constant and reaction equilibrium constant with CO2, and the other one is to investigate the effects of concentration and distribution of carriers in membranes on separation performances with the aid of simulation study, and then to develop new advanced membrane fabrication techniques for achieving the optimized carrier concentration and distribution. In addition, a promising strategy is particularly proposed that different gas transport mechanisms could be combined into a membrane, achieving the synergistic effects towards 122

membrane performance improvement, which has been primarily demonstrated by our group [328]. This strategy could be applicable for both H2-selective and CO2-seletive membranes. For H2-selective membranes, Pd or Pd alloy nanoparticles could be incorporated into microporous networks including inorganic membranes and TR polymer membranes, which not only could achieve the combination of molecular-sieving mechanism and the unique solution-diffusion mechanism of Pd towards H2, but also could solve the problems of dense metallic membranes to a certain extent. It should be noted that H2 transport based on both the aforementioned gas transport mechanisms is temperature-activated, thus the membranes with these two gas transport mechanisms are expected to be very suitable for high-temperature shifted synthesis gas separation. As for CO2-selective membranes, surface diffusion mechanism and facilitated transport mechanism could achieve very high CO2/H2 selectivity. However, there exist some limitations for fabrication of high-performance CO2-selective membranes based on the single mechanism. It is challenging to develop defect-free microporous membranes where CO2 transport obeys surface diffusion mechanism completely. However, the synthesis of microporous nanoparticles with high CO2 adsorption capacity is relatively convenient through surface functionalization of pore walls. Moreover, the pore size should be optimized to make preferentially absorbed CO2 restrict or block H2 transport through the pores. The microporous nanoparticles with the aforementioned properties as fillers could be incorporated into amino group-rich polymer matrix to develop CO2-selective MMMs with different gas transport mechanisms. It is expected that these MMMs would display excellent separation performances for CO2/H2 separation. Along with experimental study, 123

simulation study about the transport mechanisms of CO2 and H2 in different types of membranes with various microscopic structures, should be further conducted, which may be a more saving-time and powerful approach to guide the design and fabrication of membranes with more sophisticated structures. From the viewpoint of practical applications, there are the following two aspects that should be considered for future development of membrane materials. (1) The ability of materials to form a defect-free thin layer on porous supports or a self-standing thin film with enough mechanical strength, is the key to achieve practical application, which has been however rarely investigated. If a membrane material cannot be easily fabricated into a thin selective layer, it is unlikely to achieve industrial application even if it possesses relatively high gas permeability. Therefore, the developments of advanced membrane materials and the corresponding large-area thin membrane fabrication methods should be conducted in parallel. (2) The ability of materials to possess stable separation performances under practical conditions (pressure, temperature and potential contaminants) is a prerequisite for commercial application. However, the overwhelming majority of membranes reported were tested under relatively ideal conditions such as low pressure and pure gas test. In addition, the systematic investigations of effects of temperature, pressure, contaminants on membrane performances were rare. Therefore, the feasibility study of membrane materials and the corresponding membranes for hydrogen purification should be carried out to push forward the industrialization. Altogether, a variety of advanced membrane materials and membranes have significantly 124

promoted the development of membrane-based technology for hydrogen purification. We believe that the literatures published to date donate excellent examples for future exploitation of membrane materials and membranes for hydrogen purification. With the rapid development of chemical science, materials science and membrane science, membranes with excellent separation performances and favorable operational stability which could meet the requirements of practical application in hydrogen purification, are expected to be developed in the near future.

Acknowledgements The authors are grateful for the financial support from the National Natural Science Foundation of China (No. 21436009), the National High Technology Research and Development Program of China (No. 2012AA03A611) and the Program of Introducing Talents of Discipline to Universities (No. B06006).

References

[1] J.A. Turner, Sustainable hydrogen production, Science 305 (2004) 972-974. [2] H. Lin, Z. He, Z. Sun, J. Vu, A. Ng, M. Mohammed, J. Kniep, T.C. Merkel, T. Wu, R.C. Lambrecht, CO2-selective membranes for hydrogen production and CO2 capture – Part I: Membrane development, J. Membr. Sci. 457 (2014) 149-161. [3] X.M. Guo, E. Trably, E. Latrille, H. Carrère, J.-P. Steyer, Hydrogen production from agricultural waste by dark fermentation: A review, Int. J. Hydrogen Energy 35 (2010) 10660-10673. [4] N.W. Ockwig, T.M. Nenoff, Membranes for hydrogen separation, Chem. Rev. 107 (2007) 4078-4110.

125

[5] L.J. Murray, M. Dinca, J.R. Long, Hydrogen storage in metal-organic frameworks, Chem. Soc. Rev. 38 (2009) 1294-1314. [6] R.W. Baker, Future directions of membrane gas separation technology, Ind. Eng. Chem. Res. 41 (2002) 1393-1411. [7] J. Franz, V. Scherer, An evaluation of CO2 and H2 selective polymeric membranes for CO2 separation in IGCC processes, J. Membr. Sci. 359 (2010) 173-183. [8] A.Y. Ku, P. Kulkarni, R. Shisler, W. Wei, Membrane performance requirements for carbon dioxide capture using hydrogen-selective membranes in integrated gasification combined cycle (IGCC) power plants, J. Membr. Sci. 367 (2011) 233-239. [9] T.C. Merkel, M. Zhou, R.W. Baker, Carbon dioxide capture with membranes at an IGCC power plant, J. Membr. Sci. 389 (2012) 441-450. [10] V. Vakharia, K. Ramasubramanian, W.S. Winston Ho, An experimental and modeling study of CO2-selective membranes for IGCC syngas purification, J. Membr. Sci. 488 (2015) 56-66. [11] J. Xu, Z. Wang, C. Zhang, S. Zhao, Z. Qiao, P. Li, J. Wang, S. Wang, Parametric analysis and potential prediction of membrane processes for hydrogen production and pre-combustion CO2 capture, Chem. Eng. Sci. doi:10.1016/j.ces.2015.04.033. [12] S. Adhikari, S. Fernando, Hydrogen membrane separation techniques, Ind. Eng. Chem. Res. 45 (2006) 875-881. [13] G.Q. Lu, J.C. Diniz da Costa, M. Duke, S. Giessler, R. Socolow, R.H. Williams, T. Kreutz, Inorganic membranes for hydrogen production and purification: A critical review and perspective, J. Colloid Interface Sci. 314 (2007) 589-603. [14] L. Shao, B.T. Low, T.-S. Chung, A.R. Greenberg, Polymeric membranes for the hydrogen economy: Contemporary approaches and prospects for the future, J. Membr. Sci. 327 (2009) 18-31. [15] P. Bakonyi, N. Nemestóthy, K. Bélafi-Bakó, Biohydrogen purification by membranes: An overview on the operational conditions affecting the performance of non-porous,

126

polymeric and ionic liquid based gas separation membranes, Int. J. Hydrogen Energy 38 (2013) 9673-9687. [16] F. Gallucci, E. Fernandez, P. Corengia, M. van Sint Annaland, Recent advances on membranes and membrane reactors for hydrogen production, Chem. Eng. Sci. 92 (2013) 40-66. [17] K. Babita, S. Sridhar, K.V. Raghavan, Membrane reactors for fuel cell quality hydrogen through WGSR – Review of their status, challenges and opportunities, Int. J. Hydrogen Energy 36 (2011) 6671-6688. [18] D.F. Sanders, Z.P. Smith, R. Guo, L.M. Robeson, J.E. McGrath, D.R. Paul, B.D. Freeman, Energy-efficient polymeric gas separation membranes for a sustainable future: A review, Polymer 54 (2013) 4729-4761. [19] J.-R. Li, J. Sculley, H.-C. Zhou, Metal–organic frameworks for separations, Chem. Rev. 112 (2011) 869-932. [20] Ø. Hatlevik, S.K. Gade, M.K. Keeling, P.M. Thoen, A.P. Davidson, J.D. Way, Palladium and palladium alloy membranes for hydrogen separation and production: History, fabrication strategies, and current performance, Sep. Purif. Technol. 73 (2010) 59-64. [21] S. Yun, S. Ted Oyama, Correlations in palladium membranes for hydrogen separation: A review, J. Membr. Sci. 375 (2011) 28-45. [22] A.E. Lewis, D.C. Kershner, S.N. Paglieri, M.J. Slepicka, J.D. Way, Pd–Pt/YSZ composite membranes for hydrogen separation from synthetic water–gas shift streams, J. Membr. Sci. 437 (2013) 257-264. [23] S. Yun, J.H. Ko, S.T. Oyama, Ultrathin palladium membranes prepared by a novel electric field assisted activation, J. Membr. Sci. 369 (2011) 482-489. [24] T.A. Peters, T. Kaleta, M. Stange, R. Bredesen, Development of ternary Pd–Ag–TM alloy membranes with improved sulphur tolerance, J. Membr. Sci. 429 (2013) 448-458. [25] Y. She, S.C. Emerson, N.J. Magdefrau, S.M. Opalka, C. Thibaud-Erkey, T.H. Vanderspurt, Hydrogen permeability of sulfur tolerant Pd–Cu alloy membranes, J. Membr. Sci. 452 (2014) 203-211. 127

[26] K.M. Nicholson, N. Chandrasekhar, D.S. Sholl, Powered by DFT: Screening methods that accelerate materials development for hydrogen in metals applications, Acc. Chem. Res. 47 (2014) 3275-3283. [27] B.-M. Lee, J.-H. Shim, J.-Y. Suh, B.-J. Lee, A semi-empirical methodology to predict hydrogen permeability in amorphous alloy membranes, J. Membr. Sci. 472 (2014) 102-109. [28] M. Hong, S. Li, J.L. Falconer, R.D. Noble, Hydrogen purification using a SAPO-34 membrane, J. Membr. Sci. 307 (2008) 277-283. [29] M. Pera-Titus, Porous inorganic membranes for CO2 Capture: Present and prospects, Chem. Rev. 114 (2013) 1413-1492. [30] J. Gascon, F. Kapteijn, B. Zornoza, V. Sebastián, C. Casado, J. Coronas, Practical Approach to zeolitic membranes and coatings: State of the art, opportunities, barriers, and future perspectives, Chem. Mater. 24 (2012) 2829-2844. [31] Y.S. Lin, M.C. Duke, Recent progress in polycrystalline zeolite membrane research, Current Opinion in Chemical Engineering 2 (2013) 209-216. [32] M. Severance, B. Wang, K. Ramasubramanian, L. Zhao, W.S.W. Ho, P.K. Dutta, Rapid Crystallization of faujasitic zeolites: Mechanism and application to zeolite membrane growth on polymer supports, Langmuir 30 (2014) 6929-6937. [33] K. Li  , Z. Tian, X. Li, R. Xu, Y. Xu, L. Wang, H. Ma, B. Wang, L. Lin, Ionothermal synthesis of aluminophosphate molecular sieve membranes through substrate surface conversion, Angew. Chem. Int. Ed. 51 (2012) 4397-4400. [34] A. Huang, F. Liang, F. Steinbach, J. Caro, Preparation and separation properties of LTA membranes by using 3-aminopropyltriethoxysilane as covalent linker, J. Membr. Sci. 350 (2010) 5-9. [35] A. Huang, J. Caro, Facile synthesis of LTA molecular sieve membranes on covalently functionalized supports by using diisocyanates as molecular linkers, J. Mater. Chem. 21 (2011) 11424-11429.

128

[36] A. Huang, Q. Liu, N. Wang, X. Tong, B. Huang, M. Wang, J. Caro, Covalent synthesis of dense zeolite LTA membranes on various 3-chloropropyltrimethoxysilane functionalized supports, J. Membr. Sci. 437 (2013) 57-64. [37] S. Aguado, J. Gascon, D. Farrusseng, J.C. Jansen, F. Kapteijn, Simple modification of macroporous alumina supports for the fabrication of dense NaA zeolite coatings: Interplay of electrostatic and chemical interactions, Microporous Mesoporous Mater. 146 (2011) 69-75. [38] A. Huang, N. Wang, J. Caro, Synthesis of multi-layer zeolite LTA membranes with enhanced gas separation performance by using 3-aminopropyltriethoxysilane as interlayer, Microporous Mesoporous Mater. 164 (2012) 294-301. [39] A. Huang, N. Wang, J. Caro, Stepwise synthesis of sandwich-structured composite zeolite membranes with enhanced separation selectivity, Chem. Commun. 48 (2012) 3542-3544. [40] Z. Tang, J. Dong, T.M. Nenoff, Internal surface modification of MFI-type zeolite membranes for high selectivity and high flux for hydrogen, Langmuir 25 (2009) 4848-4852. [41] H. Wang, Y.S. Lin, Synthesis and modification of ZSM-5/silicalite bilayer membrane with improved hydrogen separation performance, J. Membr. Sci. 396 (2012) 128-137. [42] H. Wang, X. Dong, Y.S. Lin, Highly stable bilayer MFI zeolite membranes for high temperature hydrogen separation, J. Membr. Sci. 450 (2014) 425-432. [43] Z. Hong, F. Sun, D. Chen, C. Zhang, X. Gu, N. Xu, Improvement of hydrogen-separating performance by on-stream catalytic cracking of silane over hollow fiber MFI zeolite membrane, Int. J. Hydrogen Energy 38 (2013) 8409-8414. [44] X. Zhu, H. Wang, Y.S. Lin, Effect of the Membrane quality on gas permeation and chemical vapor deposition modification of MFI-type zeolite membranes, Ind. Eng. Chem. Res. 49 (2010) 10026-10033. [45] A. Huang, J. Caro, Highly oriented, neutral and cation-free AlPO4 LTA: from a seed crystal monolayer to a molecular sieve membrane, Chem. Commun. 47 (2011) 4201-4203. [46] T.C.T. Pham, H.S. Kim, K.B. Yoon, Growth of uniformly oriented silica MFI and BEA zeolite films on substrates, Science 334 (2011) 1533-1538.

129

[47] T.C.T. Pham, T.H. Nguyen, K.B. Yoon, Gel-free secondary growth of uniformly oriented silica MFI zeolite films and application for xylene separation, Angew. Chem. Int. Ed. 52 (2013) 8693-8698. [48] T. Lee, J. Choi, M. Tsapatsis, On the performance of c-oriented MFI zeolite membranes treated by rapid thermal processing, J. Membr. Sci. 436 (2013) 79-89. [49] J.K. Das, N. Das, S. Bandyopadhyay, Highly oriented improved SAPO 34 membrane on low cost support for hydrogen gas separation, J. Mater. Chem. A 1 (2013) 4966-4973. [50] M.C. Duke, S.J. Pas, A.J. Hill, Y.S. Lin, J.C.D. da Costa, Exposing the molecular sieving architecture of amorphous silica using positron annihilation spectroscopy, Adv. Funct. Mater. 18 (2008) 3818-3826. [51] Y. Gu, S. Ted Oyama, Ultrathin, hydrogen-selective silica membranes deposited on alumina-graded structures prepared from size-controlled boehmite sols, J. Membr. Sci. 306 (2007) 216-227. [52] H.R. Lee, M. Kanezashi, Y. Shimomura, T. Yoshioka, T. Tsuru, Evaluation and fabrication of pore-size-tuned silica membranes with tetraethoxydimethyl disiloxane for gas separation, AIChE J. 57 (2011) 2755-2765. [53] H. Qi, J. Han, N. Xu, Effect of calcination temperature on carbon dioxide separation properties of a novel microporous hybrid silica membrane, J. Membr. Sci. 382 (2011) 231-237. [54] T. Van Gestel, F. Hauler, M. Bram, W.A. Meulenberg, H.P. Buchkremer, Synthesis and characterization of hydrogen-selective sol-gel SiO2 membranes supported on ceramic and stainless steel. supports, Sep. Purif. Technol. 121 (2014) 20-29. [55] Q. Wei, F. Wang, Z.-R. Nie, C.-L. Song, Y.-L. Wang, Q.-Y. Li, Highly hydrothermally stable microporous silica membranes for hydrogen separation, J. Phys. Chem. B 112 (2008) 9354-9359. [56] Q. Wei, Y.-L. Ding, Z.-R. Nie, X.-G. Liu, Q.-Y. Li, Wettability, pore structure and performance of perfluorodecyl-modified silica membranes, J. Membr. Sci. 466 (2014) 114-122. 130

[57] S. Battersby, S. Smart, B. Ladewig, S. Liu, M.C. Duke, V. Rudolph, J.C.D.d. Costa, Hydrothermal stability of cobalt silica membranes in a water gas shift membrane reactor, Sep. Purif. Technol. 66 (2009) 299-305. [58] Y. Gu, S.T. Oyama, Permeation properties and hydrothermal stability of silica–titania membranes supported on porous alumina substrates, J. Membr. Sci. 345 (2009) 267-275. [59] H. Qi, H. Chen, L. Li, G. Zhu, N. Xu, Effect of Nb content on hydrothermal stability of a novel ethylene-bridged silsesquioxane molecular sieving membrane for H2/CO2 separation, J. Membr. Sci. 421–422 (2012) 190-200. [60] C. Yacou, S. Smart, J.C. Diniz da Costa, Long term performance cobalt oxide silica membrane module for high temperature H2 separation, Energ Environ. Sci. 5 (2012) 5820-5832. [61] A. Darmawan, J. Motuzas, S. Smart, A. Julbe, J.C. Diniz da Costa, Binary iron cobalt oxide silica membrane for gas separation, J. Membr. Sci. 474 (2015) 32-38. [62] Y.H. Ikuhara, H. Mori, T. Saito, Y. Iwamoto, High-temperature hydrogen adsorption properties of precursor-derived nickel nanoparticle-dispersed amorphous silica, J Am Ceram Soc 90 (2007) 546-552. [63] S. Battersby, T. Tasaki, S. Smart, B. Ladewig, S. Liu, M.C. Duke, V. Rudolph, J.C. Diniz da Costa, Performance of cobalt silica membranes in gas mixture separation, J. Membr. Sci. 329 (2009) 91-98. [64] S. Smart, J.F. Vente, J.C. Diniz da Costa, High temperature H2/CO2 separation using cobalt oxide silica membranes, Int. J. Hydrogen Energy 37 (2012) 12700-12707. [65] M. Inagaki, N. Ohta, Y. Hishiyama, Aromatic polyimides as carbon precursors, Carbon 61 (2013) 1-21. [66] M. Kiyono, P.J. Williams, W.J. Koros, Effect of pyrolysis atmosphere on separation performance of carbon molecular sieve membranes, J. Membr. Sci. 359 (2010) 2-10. [67] S.M. Saufi, A.F. Ismail, Fabrication of carbon membranes for gas separation––a review, Carbon 42 (2004) 241-259.

131

[68] W.N.W. Salleh, A.F. Ismail, T. Matsuura, M.S. Abdullah, Precursor selection and process conditions in the preparation of carbon membrane for gas separation: A Review, Separation & Purification Reviews 40 (2011) 261-311. [69] L. Li, C. Song, H. Jiang, J. Qiu, T. Wang, Preparation and gas separation performance of supported carbon membranes with ordered mesoporous carbon interlayer, J. Membr. Sci. 450 (2014) 469-477. [70] L.-H. Cheng, Y.-J. Fu, K.-S. Liao, J.-T. Chen, C.-C. Hu, W.-S. Hung, K.-R. Lee, J.-Y. Lai, A high-permeance supported carbon molecular sieve membrane fabricated by plasma-enhanced chemical vapor deposition followed by carbonization for CO2 capture, J. Membr. Sci. 460 (2014) 1-8. [71] X. He, M.-B. Hägg, Hollow fiber carbon membranes: From material to application, Chem. Eng. J. 215–216 (2013) 440-448. [72] N. Bhuwania, Y. Labreche, C.S.K. Achoundong, J. Baltazar, S.K. Burgess, S. Karwa, L. Xu, C.L. Henderson, P.J. Williams, W.J. Koros, Engineering substructure morphology of asymmetric carbon molecular sieve hollow fiber membranes, Carbon 76 (2014) 417-434. [73] W.N.W. Salleh, A.F. Ismail, Fabrication and characterization of PEI/PVP-based carbon hollow fiber membranes for CO2/CH4 and CO2/N2 separation, AIChE J. 58 (2012) 3167-3175. [74] A.K. Itta, H.-H. Tseng, M.-Y. Wey, Fabrication and characterization of PPO/PVP blend carbon molecular sieve membranes for H2/N2 and H2/CH4 separation, J. Membr. Sci. 372 (2011) 387-395. [75] D.-e. Jiang, V.R. Cooper, S. Dai, Porous graphene as the ultimate membrane for gas separation, Nano Lett. 9 (2009) 4019-4024. [76] D. Cohen-Tanugi, J.C. Grossman, Water desalination across nanoporous graphene, Nano Lett. 12 (2012) 3602-3608. [77] R.R. Nair, H.A. Wu, P.N. Jayaram, I.V. Grigorieva, A.K. Geim, Unimpeded permeation of water through helium-leak–tight graphene-based membranes, Science 335 (2012) 442-444. [78] D.R. Paul, Creating new types of carbon-based membranes, Science 335 (2012) 413-414.

132

[79] G. Liu, W. Jin, N. Xu, Graphene-based membranes, Chem. Soc. Rev. 44 (2015) 5016-5030. [80] J.C. Meyer, A.K. Geim, M.I. Katsnelson, K.S. Novoselov, T.J. Booth, S. Roth, The structure of suspended graphene sheets, Nature 446 (2007) 60-63. [81] C. Lee, X. Wei, J.W. Kysar, J. Hone, Measurement of the elastic properties and intrinsic strength of monolayer graphene, Science 321 (2008) 385-388. [82] Ç.Ö. Girit, J.C. Meyer, R. Erni, M.D. Rossell, C. Kisielowski, L. Yang, C.-H. Park, M.F. Crommie, M.L. Cohen, S.G. Louie, A. Zettl, Graphene at the edge: Stability and dynamics, Science 323 (2009) 1705-1708. [83] J.S. Bunch, S.S. Verbridge, J.S. Alden, A.M. van der Zande, J.M. Parpia, H.G. Craighead, P.L. McEuen, Impermeable atomic membranes from graphene sheets, Nano Lett. 8 (2008) 2458-2462. [84] S.P. Koenig, L. Wang, J. Pellegrino, J.S. Bunch, Selective molecular sieving through porous graphene, Nat Nano 7 (2012) 728-732. [85] R. Zan, Q.M. Ramasse, U. Bangert, K.S. Novoselov, Graphene reknits its holes, Nano Lett. 12 (2012) 3936-3940. [86] S. .

’Hern, M.S.H. Boutilier, J.-C. Idrobo, Y. Song, J. Kong, T. Laoui, M. Atieh, R.

Karnik, Selective ionic transport through tunable subnanometer pores in single-layer graphene membranes, Nano Lett. 14 (2014) 1234-1241. [87] T.H. Han, Y.-K. Huang, A.T.L. Tan, V.P. Dravid, J. Huang, Steam etched porous graphene oxide network for chemical sensing, J. Am. Chem. Soc. 133 (2011) 15264-15267. [88] P. Xu, J. Yang, K. Wang, Z. Zhou, P. Shen, Porous graphene: Properties, preparation, and potential applications, Chinese Sci Bull 57 (2012) 2948-2955. [89] D. Zhou, Y. Cui, P.-W. Xiao, M.-Y. Jiang, B.-H. Han, A general and scalable synthesis approach to porous graphene, Nat. Commun. 5 (2014) 4716. [90] K. Celebi, J. Buchheim, R.M. Wyss, A. Droudian, P. Gasser, I. Shorubalko, J.-I. Kye, C. Lee, H.G. Park, Ultimate permeation across atomically thin porous graphene, Science 344 (2014) 289-292. 133

[91] J. Schrier, Helium separation using porous graphene membranes, J Phys Chem Lett 1 (2010) 2284-2287. [92] H. Du, J. Li, J. Zhang, G. Su, X. Li, Y. Zhao, Separation of hydrogen and nitrogen gases with porous graphene membrane, J. Phys. Chem. C 115 (2011) 23261-23266. [93] Y. Tao, Q. Xue, Z. Liu, M. Shan, C. Ling, T. Wu, X. Li, Tunable hydrogen separation in porous graphene membrane: First-principle and molecular dynamic simulation, ACS Appl. Mater. Interfaces 6 (2014) 8048-8058. [94] M.S.H. Boutilier,

. Sun, S. .

’Hern, H. Au,

.G. Hadjiconstantinou, R. Karnik,

Implications of permeation through intrinsic defects in graphene on the design of defect-tolerant membranes for gas separation, ACS Nano 8 (2014) 841-849. [95] H. Li, Z. Song, X. Zhang, Y. Huang, S. Li, Y. Mao, H.J. Ploehn, Y. Bao, M. Yu, Ultrathin, molecular-sieving graphene oxide membranes for selective hydrogen separation, Science 342 (2013) 95-98. [96] H.W. Kim, H.W. Yoon, S.M. Yoon, B.M. Yoo, B.K. Ahn, Y.H. Cho, H.J. Shin, H. Yang, U. Paik, S. Kwon, J.Y. Choi, H.B. Park, Selective gas transport through few-layered graphene and graphene oxide membranes, Science 342 (2013) 91-95. [97] H.W. Kim, H.W. Yoon, B.M. Yoo, J.S. Park, K.L. Gleason, B.D. Freeman, H.B. Park, High-performance CO2-philic graphene oxide membranes under wet-conditions, Chem. Commun. 50 (2014) 13563-13566. [98] D. Kim, D.W. Kim, H.-K. Lim, J. Jeon, H. Kim, H.-T. Jung, H. Lee, Intercalation of gas molecules in graphene oxide interlayer: The role of water, J. Phys. Chem. C 118 (2014) 11142-11148. [99] L.W. Drahushuk, M.S. Strano, Mechanisms of gas permeation through single layer graphene membranes, Langmuir 28 (2012) 16671-16678. [100] A. Ambrosetti, P.L. Silvestrelli, Gas separation in nanoporous graphene from first principle calculations, J. Phys. Chem. C 118 (2014) 19172-19179.

134

[101] C. Sun, M.S.H. Boutilier, H. Au, P. Poesio, B. Bai, R. Karnik, N.G. Hadjiconstantinou, Mechanisms of molecular permeation through nanoporous graphene membranes, Langmuir 30 (2014) 675-682. [102] J.R. Long, O.M. Yaghi, The pervasive chemistry of metal-organic frameworks, Chem. Soc. Rev. 38 (2009) 1213-1214. [103] S. Chaemchuen, N.A. Kabir, K. Zhou, F. Verpoort, Metal-organic frameworks for upgrading biogas via CO2 adsorption to biogas green energy, Chem. Soc. Rev. 42 (2013) 9304-9332. [104] J.-R. Li, R.J. Kuppler, H.-C. Zhou, Selective gas adsorption and separation in metal-organic frameworks, Chem. Soc. Rev. 38 (2009) 1477-1504. [105] Y. Liu, Z. Ng, E.A. Khan, H.-K. Jeong, C.-b. Ching, Z. Lai, Synthesis of continuous MOF-5 membranes on porous α-alumina substrates, Microporous Mesoporous Mater. 118 (2009) 296-301. [106] S.R. Venna, M.A. Carreon, Metal organic framework membranes for carbon dioxide separation, Chem. Eng. Sci. 120 (2014) 174-190. [107] S. Qiu, M. Xue, G. Zhu, Metal-organic framework membranes: from synthesis to separation application, Chem. Soc. Rev. 43 (2014) 6116-6140. [108] H. Bux, F. Liang, Y. Li, J. Cravillon, M. Wiebcke, J. Caro, Zeolitic imidazolate framework membrane with molecular sieving properties by microwave-assisted solvothermal synthesis, J. Am. Chem. Soc. 131 (2009) 16000-16001. [109] Z. Kang, M. Xue, L. Fan, J. Ding, L. Guo, L. Gao, S. Qiu, "Single nickel source" in situ fabrication of a stable homochiral MOF membrane with chiral resolution properties, Chem. Commun. 49 (2013) 10569-10571. [110] D. Nagaraju, D.G. Bhagat, R. Banerjee, U.K. Kharul, In situ growth of metal-organic frameworks on a porous ultrafiltration membrane for gas separation, J. Mater. Chem. A 1 (2013) 8828-8835. [111] F. Cacho-Bailo, B. Seoane, C. Tellez, J. Coronas, ZIF-8 continuous membrane on porous polysulfone for hydrogen separation, J. Membr. Sci. 464 (2014) 119-126. 135

[112] L. Ge, W. Zhou, A. Du, Z. Zhu, Porous polyethersulfone-supported zeolitic imidazolate framework membranes for hydrogen separation, J. Phys. Chem. C 116 (2012) 13264-13270. [113] A.J. Brown, N.A. Brunelli, K. Eum, F. Rashidi, J.R. Johnson, W.J. Koros, C.W. Jones, S. Nair, Interfacial microfluidic processing of metal-organic framework hollow fiber membranes, Science 345 (2014) 72-75. [114] A. Huang, H. Bux, F. Steinbach, J. Caro, Molecular-sieve membrane with hydrogen permselectivity: ZIF-22 in LTA topology prepared with 3-aminopropyltriethoxysilane as covalent linker, Angew. Chem. Int. Ed. 49 (2010) 4958-4961. [115] A. Huang, Y. Chen, N. Wang, Z. Hu, J. Jiang, J. Caro, A highly permeable and selective zeolitic imidazolate framework ZIF-95 membrane for H2/CO2 separation, Chem. Commun. 48 (2012) 10981-10983. [116] Q. Liu, N. Wang, J. Caro, A. Huang, Bio-inspired polydopamine: A versatile and powerful platform for covalent synthesis of molecular sieve membranes, J. Am. Chem. Soc. 135 (2013) 17679-17682. [117] A. Huang, W. Dou, J. Caro, Steam-stable zeolitic imidazolate framework ZIF-90 membrane with hydrogen selectivity through covalent functionalization, J. Am. Chem. Soc. 132 (2010) 15562-15564. [118] A. Huang, Q. Liu, N. Wang, Y. Zhu, J. Caro, Bicontinuous zeolitic imidazolate framework ZIF-8@GO membrane with enhanced hydrogen selectivity, J. Am. Chem. Soc. 136 (2014) 14686-14689. [119] A. Huang, Q. Liu, N. Wang, J. Caro, Highly hydrogen permselective ZIF-8 membranes supported on polydopamine functionalized macroporous stainless-steel-nets, J. Mater. Chem. A 2 (2014) 8246-8251. [120] Y. Liu, N. Wang, J.H. Pan, F. Steinbach, J. Caro, In-situ synthesis of MOF membranes on ZnAl-CO3 LDH buffer layer-modified substrates, J. Am. Chem. Soc. 136 (2014) 14353-14356. [121] X. Zhang, Y. Liu, S. Li, L. Kong, H. Liu, Y. Li, W. Han, K.L. Yeung, W. Zhu, W. Yang, J. Qiu, New membrane architecture with high performance: ZIF-8 membrane supported on 136

vertically aligned ZnO nanorods for gas permeation and separation, Chem. Mater. 26 (2014) 1975-1981. [122] W. Li, Q. Meng, X. Li, C. Zhang, Z. Fan, G. Zhang, Non-activation ZnO array as a buffering layer to fabricate strongly adhesive metal-organic framework/PVDF hollow fiber membranes, Chem. Commun. 50 (2014) 9711-9713. [123] H.T. Kwon, H.-K. Jeong, In situ synthesis of thin zeolitic–imidazolate framework ZIF-8 membranes exhibiting exceptionally high propylene/propane separation, J. Am. Chem. Soc. 135 (2013) 10763-10768. [124] Z. Xie, J. Yang, J. Wang, J. Bai, H. Yin, B. Yuan, J. Lu, Y. Zhang, L. Zhou, C. Duan, Deposition of chemically modified [small alpha]-Al2O3 particles for high performance ZIF-8 membrane on a macroporous tube, Chem. Commun. 48 (2012) 5977-5979. [125] Y. Yoo, Z. Lai, H.-K. Jeong, Fabrication of MOF-5 membranes using microwave-induced rapid seeding and solvothermal secondary growth, Microporous Mesoporous Mater. 123 (2009) 100-106. [126] Y. Li, F. Liang, H. Bux, W. Yang, J. Caro, Zeolitic imidazolate framework ZIF-7 based molecular sieve membrane for hydrogen separation, J. Membr. Sci. 354 (2010) 48-54. [127] Y. Hu, X. Dong, J. Nan, W. Jin, X. Ren, N. Xu, Y.M. Lee, Metal-organic framework membranes fabricated via reactive seeding, Chem. Commun. 47 (2011) 737-739. [128] J. Nan, X. Dong, W. Wang, W. Jin, Formation mechanism of metal–organic framework membranes derived from reactive seeding approach, Microporous Mesoporous Mater. 155 (2012) 90-98. [129] J. Nan, X. Dong, W. Wang, W. Jin, N. Xu, Step-by-step seeding procedure for preparing HKUST-1 membrane on porous α-alumina support, Langmuir 27 (2011) 4309-4312. [130] D.-J. Lee, Q. Li, H. Kim, K. Lee, Preparation of Ni-MOF-74 membrane for CO2 separation by layer-by-layer seeding technique, Microporous Mesoporous Mater. 163 (2012) 169-177.

137

[131] A. Huang, J. Caro, Covalent post-functionalization of zeolitic imidazolate framework ZIF-90 membrane for enhanced hydrogen selectivity, Angew. Chem. Int. Ed. 50 (2011) 4979-4982. [132] A. Huang, N. Wang, C. Kong, J. Caro, Organosilica-functionalized zeolitic imidazolate framework ZIF-90 membrane with high gas-separation performance, Angew. Chem. Int. Ed. 51 (2012) 10551-10555. [133] Z. Kang, M. Xue, L. Fan, L. Huang, L. Guo, G. Wei, B. Chen, S. Qiu, Highly selective sieving of small gas molecules by an ultra-microporous metal-organic framework membrane, Energ Environ. Sci. 7 (2014) 4053-4060. [134] F. Zhang, X. Zou, X. Gao, S. Fan, F. Sun, H. Ren, G. Zhu, Hydrogen selective NH2‐ MIL‐53 (Al) MOF membranes with high permeability, Adv. Funct. Mater. 22 (2012) 3583-3590. [135] Y. Peng, Y. Li, Y. Ban, H. Jin, W. Jiao, X. Liu, W. Yang, Metal-organic framework nanosheets as building blocks for molecular sieving membranes, Science 346 (2014) 1356-1359. [136] P. Pandey, R.S. Chauhan, Membranes for gas separation, Prog. Polym. Sci. 26 (2001) 853-893. [137] P. Bernardo, E. Drioli, G. Golemme, Membrane gas separation: A review/state of the art, Ind. Eng. Chem. Res. 48 (2009) 4638-4663. [138] B.D. Freeman, Basis of permeability/selectivity tradeoff relations in polymeric gas separation membranes, Macromolecules 32 (1999) 375-380. [139] K. Vanherck, G. Koeckelberghs, I.F.J. Vankelecom, Crosslinking polyimides for membrane applications: A review, Prog. Polym. Sci. 38 (2013) 874-896. [140] P. Vandezande, L.E.M. Gevers, I.F.J. Vankelecom, Solvent resistant nanofiltration: separating on a molecular level, Chem. Soc. Rev. 37 (2008) 365-405. [141] B.T. Low, Y. Xiao, T.S. Chung, Amplifying the molecular sieving capability of polyimide membranes via coupling of diamine networking and molecular architecture, Polymer 50 (2009) 3250-3258. 138

[142] L. Shao, C.-H. Lau, T.-S. Chung, A novel strategy for surface modification of polyimide membranes by vapor-phase ethylenediamine (EDA) for hydrogen purification, Int. J. Hydrogen Energy 34 (2009) 8716-8722. [143] L. Shao, L. Liu, S.-X. Cheng, Y.-D. Huang, J. Ma, Comparison of diamino cross-linking in different polyimide solutions and membranes by precipitation observation and gas transport, J. Membr. Sci. 312 (2008) 174-185. [144] C.H. Lau, B.T. Low, L. Shao, T.-S. Chung, A vapor-phase surface modification method to enhance different types of hollow fiber membranes for industrial scale hydrogen separation, Int. J. Hydrogen Energy 35 (2010) 8970-8982. [145] H. Wang, D.R. Paul, T.-S. Chung, Surface modification of polyimide membranes by diethylenetriamine (DETA) vapor for H2 purification and moisture effect on gas permeation, J. Membr. Sci. 430 (2013) 223-233. [146] K.A. Berchtold, R.P. Singh, J.S. Young, K.W. Dudeck, Polybenzimidazole composite membranes for high temperature synthesis gas separations, J. Membr. Sci. 415–416 (2012) 265-270. [147] X. Li, R.P. Singh, K.W. Dudeck, K.A. Berchtold, B.C. Benicewicz, Influence of polybenzimidazole main chain structure on H2/CO2 separation at elevated temperatures, J. Membr. Sci. 461 (2014) 59-68. [148] S.C. Kumbharkar, Y. Liu, K. Li, High performance polybenzimidazole based asymmetric hollow fibre membranes for H2/CO2 separation, J. Membr. Sci. 375 (2011) 231-240. [149] N.B. McKeown, P.M. Budd, Polymers of intrinsic microporosity (PIMs): organic materials for membrane separations, heterogeneous catalysis and hydrogen storage, Chem. Soc. Rev. 35 (2006) 675-683. [150] S. Kim, Y.M. Lee, Rigid and microporous polymers for gas separation membranes, Prog. Polym. Sci. 43 (2015) 1-32. [151] R. Dawson, A.I. Cooper, D.J. Adams, Nanoporous organic polymer networks, Prog. Polym. Sci. 37 (2012) 530-563. 139

[152] G. Cheng, T. Hasell, A. Trewin, D.J. Adams, A.I. Cooper, Soluble conjugated microporous polymers, Angew. Chem. Int. Ed. 51 (2012) 12727-12731. [153] H.B. Park, C.H. Jung, Y.M. Lee, A.J. Hill, S.J. Pas, S.T. Mudie, E. Van Wagner, B.D. Freeman, D.J. Cookson, Polymers with cavities tuned for fast selective transport of small molecules and ions, Science 318 (2007) 254-258. [154] Y.C. Xiao, T.S. Chung, Grafting thermally labile molecules on cross-linkable polyimide to design membrane materials for natural gas purification and CO2 capture, Energ Environ. Sci. 4 (2011) 201-208. [155] M.L. Chua, Y.C. Xiao, T.-S. Chung, Modifying the molecular structure and gas separation performance of thermally labile polyimide-based membranes for enhanced natural gas purification, Chem. Eng. Sci. 104 (2013) 1056-1064. [156] S.H. Han, H.J. Kwon, K.Y. Kim, J.G. Seong, C.H. Park, S. Kim, C.M. Doherty, A.W. Thornton, A.J. Hill, A.E. Lozano, K.A. Berchtold, Y.M. Lee, Tuning microcavities in thermally rearranged polymer membranes for CO2 capture, Phys. Chem. Chem. Phys. 14 (2012) 4365-4373. [157] Y.S. Do, J.G. Seong, S. Kim, J.G. Lee, Y.M. Lee, Thermally rearranged (TR) poly(benzoxazole-co-amide) membranes for hydrogen separation derived from 3,3 ′ -dihydroxy-4,4 ′ -diamino-biphenyl (HAB), 4,4 ′ -oxydianiline (ODA) and isophthaloyl chloride (IPCl), J. Membr. Sci. 446 (2013) 294-302. [158] C.H. Park, E. Tocci, Y.M. Lee, E. Drioli, Thermal treatment effect on the structure and property change between hydroxy-containing polyimides (HPIs) and thermally rearranged polybenzoxazole (TR-PBO), J. Phys. Chem. B 116 (2012) 12864-12877. [159] C.H. Park, E. Tocci, S. Kim, A. Kumar, Y.M. Lee, E. Drioli, A simulation study on OH-containing polyimide (HPI) and thermally rearranged polybenzoxazoles (TR-PBO): Relationship between gas transport properties and free volume morphology, J. Phys. Chem. B 118 (2014) 2746-2757. [160] S. Kim, S.H. Han, Y.M. Lee, Thermally rearranged (TR) polybenzoxazole hollow fiber membranes for CO2 capture, J. Membr. Sci. 403–404 (2012) 169-178. 140

[161] F.Y. Li, Y. Xiao, Y.K. Ong, T.S. Chung, UV-rearranged PIM-1 polymeric membranes for advanced hydrogen purification and production, Adv Energy Mater 2 (2012) 1456-1466. [162] B.S. Ghanem, R. Swaidan, X. Ma, E. Litwiller, I. Pinnau, Energy-efficient hydrogen separation by AB-type ladder-polymer molecular sieves, Adv. Mater. 26 (2014) 6696-6700. [163] M. Wang, V. Janout, S.L. Regen, Hyper-thin organic membranes with exceptional H2/CO2 permeation selectivity: importance of ionic crosslinking and self-healing, Chem. Commun. 47 (2011) 2387-2389. [164] M. Wang, S. Yi, V. Janout, S.L. Regen, A 7 nm thick polymeric membrane with a H2/CO2 selectivity of 200 that reaches the upper bound, Chem. Mater. 25 (2013) 3785-3787. [165] D. Kim, P. Tzeng, K.J. Barnett, Y.-H. Yang, B.A. Wilhite, J.C. Grunlan, Highly size-selective ionically crosslinked multilayer polymer films for light gas separation, Adv. Mater. 26 (2013) 746-751. [166] M. Wang, V. Janout, S.L. Regen, Gas transport across hyperthin membranes, Acc. Chem. Res. 46 (2013) 2743-2754. [167] R. Mahajan, W.J. Koros, Factors controlling successful formation of mixed-matrix gas separation materials, Ind. Eng. Chem. Res. 39 (2000) 2692-2696. [168] T.S. Chung, L.Y. Jiang, Y. Li, S. Kulprathipanja, Mixed matrix membranes (MMMs) comprising organic polymers with dispersed inorganic fillers for gas separation, Prog. Polym. Sci. 32 (2007) 483-507. [169] P.S. Goh, A.F. Ismail, S.M. Sanip, B.C. Ng, M. Aziz, Recent advances of inorganic fillers in mixed matrix membrane for gas separation, Sep. Purif. Technol. 81 (2011) 243-264. [170] G. Dong, H. Li, V. Chen, Challenges and opportunities for mixed-matrix membranes for gas separation, J. Mater. Chem. A 1 (2013) 4610-4630. [171] M. Rezakazemi, A. Ebadi Amooghin, M.M. Montazer-Rahmati, A.F. Ismail, T. Matsuura, State-of-the-art membrane based CO2 separation using mixed matrix membranes (MMMs): An overview on current status and future directions, Prog. Polym. Sci. 39 (2014) 817-861.

141

[172] B. Seoane, J. Coronas, I. Gascon, M.E. Benavides, O. Karvan, J. Caro, F. Kapteijn, J. Gascon, Metal-organic framework based mixed matrix membranes: a solution for highly efficient CO2 capture?, Chem. Soc. Rev. 44 (2015) 2421-2454. [173] H. Vinh-Thang, S. Kaliaguine, Predictive models for mixed-matrix membrane performance: A review, Chem. Rev. 113 (2013) 4980-5028. [174] S. Choi, J. Coronas, Z. Lai, D. Yust, F. Onorato, M. Tsapatsis, Fabrication and gas separation properties of polybenzimidazole (PBI)/nanoporous silicates hybrid membranes, J. Membr. Sci. 316 (2008) 145-152. [175] M.J.C. Ordoñez, K.J. Balkus Jr, J.P. Ferraris, I.H. Musselman, Molecular sieving realized with ZIF-8/Matrimid® mixed-matrix membranes, J. Membr. Sci. 361 (2010) 28-37. [176] S.N. Wijenayake, N.P. Panapitiya, S.H. Versteeg, C.N. Nguyen, S. Goel, K.J. Balkus, I.H. Musselman, J.P. Ferraris, Surface cross-linking of ZIF-8/polyimide mixed matrix membranes (MMMs) for gas separation, Ind. Eng. Chem. Res. 52 (2013) 6991-7001. [177] S.N. Wijenayake, N.P. Panapitiya, C.N. Nguyen, Y. Huang, K.J. Balkus Jr, I.H. Musselman, J.P. Ferraris, Composite membranes with a highly selective polymer skin for hydrogen separation, Sep. Purif. Technol. 135 (2014) 190-198. [178] J. Hu, H. Cai, H. Ren, Y. Wei, Z. Xu, H. Liu, Y. Hu, Mixed-matrix membrane hollow fibers of Cu3(BTC)2 MOF and polyimide for gas separation and adsorption, Ind. Eng. Chem. Res. 49 (2010) 12605-12612. [179] T. Yang, Y. Xiao, T.-S. Chung, Poly-/metal-benzimidazole nano-composite membranes for hydrogen purification, Energ Environ. Sci. 4 (2011) 4171-4180. [180] T. Yang, G.M. Shi, T.-S. Chung, Symmetric and asymmetric zeolitic imidazolate frameworks (ZIFs)/polybenzimidazole (PBI) nanocomposite membranes for hydrogen purification at high temperatures, Adv Energy Mater 2 (2012) 1358-1367. [181] T. Yang, T.-S. Chung, Room-temperature synthesis of ZIF-90 nanocrystals and the derived nano-composite membranes for hydrogen separation, J. Mater. Chem. A 1 (2013) 6081-6090.

142

[182] T. Yang, T.-S. Chung, High performance ZIF-8/PBI nano-composite membranes for high temperature hydrogen separation consisting of carbon monoxide and water vapor, Int. J. Hydrogen Energy 38 (2013) 229-239. [183] M.D. Guiver, H.N. Le Thi, G.P. Robertson, Composite gas separation membranes, in, US Patent 20,020,062,737, 2002. [184] A.L. Khan, A. Cano-Odena, B. Gutiérrez, C. Minguillón, I.F.J. Vankelecom, Hydrogen separation and purification using polysulfone acrylate–zeolite mixed matrix membranes, J. Membr. Sci. 350 (2010) 340-346. [185] L. Cao, K. Tao, A. Huang, C. Kong, L. Chen, A highly permeable mixed matrix membrane containing CAU-1-NH2 for H2 and CO2 separation, Chem. Commun. 49 (2013) 8513-8515. [186] H.S.M. Suhaimi, M.N.I.M. Khir, C.P. Leo, A.L. Ahmad, Preparation and characterization of polysulfone mixed-matrix membrane incorporated with palladium nanoparticles dispersed in polyvinylpyrrolidone for hydrogen separation, J. Polym. Res. 21 (2014) 1-8. [187] M. Kanezashi, M. Sano, T. Yoshioka, T. Tsuru, Extremely thin Pd-silica mixed-matrix membranes with nano-dispersion for improved hydrogen permeability, Chem. Commun. 46 (2010) 6171-6173. [188] L.M. Robeson, The upper bound revisited, J. Membr. Sci. 320 (2008) 390-400. [189] C.H. Lau, P. Li, F. Li, T.-S. Chung, D.R. Paul, Reverse-selective polymeric membranes for gas separations, Prog. Polym. Sci. 38 (2013) 740-766. [190] M.B. Rao, S. Sircar, Nanoporous carbon membranes for separation of gas mixtures by selective surface flow, J. Membr. Sci. 85 (1993) 253-264. [191] M.B. Rao, S. Sircar, Performance and pore characterization of nanoporous carbon membranes for gas separation, J. Membr. Sci. 110 (1996) 109-118. [192] K.-H. Lee, S.-T. Hwang, The transport of condensible vapors through a microporous vycor glass membrane, J. Colloid Interface Sci. 110 (1986) 544-555.

143

[193] J. Lindmark, J. Hedlund, Modification of MFI membranes with amine groups for enhanced CO2 selectivity, J. Mater. Chem. 20 (2010) 2219-2225. [194] S.M. Mahurin, J.S. Lee, X. Wang, S. Dai, Ammonia-activated mesoporous carbon membranes for gas separations, J. Membr. Sci. 368 (2011) 41-47. [195] Y. Sakamoto, K. Nagata, K. Yogo, K. Yamada, Preparation and CO2 separation properties of amine-modified mesoporous silica membranes, Microporous Mesoporous Mater. 101 (2007) 303-311. [196] M. Ostwal, R.P. Singh, S.F. Dec, M.T. Lusk, J.D. Way, 3-aminopropyltriethoxysilane functionalized inorganic membranes for high temperature CO2/N2 separation, J. Membr. Sci. 369 (2011) 139-147. [197] J. Lindmark, J. Hedlund, Carbon dioxide removal from synthesis gas using MFI membranes, J. Membr. Sci. 360 (2010) 284-291. [198] T.L. Chew, A.L. Ahmad, S. Bhatia, Ba-SAPO-34 membrane synthesized from microwave heating and its performance for CO2/CH4 gas separation, Chem. Eng. J. 171 (2011) 1053-1059. [199] L. Sandström, E. Sjöberg, J. Hedlund, Very high flux MFI membrane for CO2 separation, J. Membr. Sci. 380 (2011) 232-240. [200] E. Sjöberg, L. Sandström, O.G.W. Öhrman, J. Hedlund, Separation of CO2 from black liquor derived syngas using an MFI membrane, J. Membr. Sci. 443 (2013) 131-137. [201] M. Zhou, D. Korelskiy, P. Ye, M. Grahn, J. Hedlund, A uniformly oriented MFI membrane for improved CO2 separation, Angew. Chem. Int. Ed. 53 (2014) 3492-3495. [202] S. Takamizawa, Y. Takasaki, R. Miyake, Single-crystal membrane for anisotropic and efficient gas permeation, J. Am. Chem. Soc. 132 (2010) 2862-2863. [203] Z. Zhao, X. Ma, A. Kasik, Z. Li, Y.S. Lin, Gas separation properties of metal organic framework (MOF-5) membranes, Ind. Eng. Chem. Res. 52 (2012) 1102-1108. [204] B.A. Al-Maythalony, O. Shekhah, R. Swaidan, Y. Belmabkhout, I. Pinnau, M. Eddaoudi, Quest for anionic MOF membranes: Continuous sod-ZMOF membrane with CO2 adsorption-driven selectivity, J. Am. Chem. Soc. 137 (2015) 1754-1757. 144

[205] P.M. Budd, K.J. Msayib, C.E. Tattershall, B.S. Ghanem, K.J. Reynolds, N.B. McKeown, D. Fritsch, Gas separation membranes from polymers of intrinsic microporosity, J. Membr. Sci. 251 (2005) 263-269. [206] P.M. Budd, N.B. McKeown, B.S. Ghanem, K.J. Msayib, D. Fritsch, L. Starannikova, N. Belov, O. Sanfirova, Y. Yampolskii, V. Shantarovich, Gas permeation parameters and other physicochemical properties of a polymer of intrinsic microporosity: Polybenzodioxane PIM-1, J. Membr. Sci. 325 (2008) 851-860. [207] M.D. Guiver, Y.M. Lee, Polymer rigidity improves microporous membranes, Science 339 (2013) 284-285. [208] C.G. Bezzu, M. Carta, A. Tonkins, J.C. Jansen, P. Bernardo, F. Bazzarelli, N.B. McKeown, A spirobifluorene-based polymer of intrinsic microporosity with improved performance for gas separation, Adv. Mater. 24 (2012) 5930-5933. [209] M. Carta, R. Malpass-Evans, M. Croad, Y. Rogan, J.C. Jansen, P. Bernardo, F. Bazzarelli, N.B. McKeown, An efficient polymer molecular sieve for membrane gas separations, Science 339 (2013) 303-307. [210] M. Carta, M. Croad, R. Malpass-Evans, J.C. Jansen, P. Bernardo, G. Clarizia, K. Friess, M. Lanč,

.B. McKeown, Triptycene induced enhancement of membrane gas selectivity for

microporous tröger's base polymers, Adv. Mater. 26 (2014) 3526-3531. [211] E. Tocci, L. De Lorenzo, P. Bernardo, G. Clarizia, F. Bazzarelli, N.B. McKeown, M. Carta, R. Malpass-Evans, K. Friess, K. Pilnáček, M. Lanč, Y.P. Yampolskii, L. Strarannikova, V. Shantarovich, M. Mauri, J.C. Jansen, Molecular modeling and gas permeation properties of a polymer of intrinsic microporosity composed of ethanoanthracene and tröger’s base units, Macromolecules 47 (2014) 7900-7916. [212] N. Du, G.P. Robertson, J. Song, I. Pinnau, S. Thomas, M.D. Guiver, Polymers of intrinsic microporosity containing trifluoromethyl and phenylsulfone groups as materials for membrane gas separation, Macromolecules 41 (2008) 9656-9662. [213] N. Du, G.P. Robertson, I. Pinnau, M.D. Guiver, Polymers of intrinsic microporosity derived from novel disulfone-based monomers, Macromolecules 42 (2009) 6023-6030. 145

[214] N. Du, G.P. Robertson, J. Song, I. Pinnau, M.D. Guiver, High-performance carboxylated polymers of intrinsic microporosity (PIMs) with tunable gas transport properties, Macromolecules 42 (2009) 6038-6043. [215] N.Y. Du, H.B. Park, G.P. Robertson, M.M. Dal-Cin, T. Visser, L. Scoles, M.D. Guiver, Polymer nanosieve membranes for CO2-capture applications, Nat Mater 10 (2011) 372-375. [216] C.R. Mason, L. Maynard-Atem, N.M. Al-Harbi, P.M. Budd, P. Bernardo, F. Bazzarelli, G. Clarizia, J.C. Jansen, Polymer of intrinsic microporosity incorporating thioamide functionality: Preparation and gas transport properties, Macromolecules 44 (2011) 6471-6479. [217] N. Du, G.P. Robertson, M.M. Dal-Cin, L. Scoles, M.D. Guiver, Polymers of intrinsic microporosity (PIMs) substituted with methyl tetrazole, Polymer 53 (2012) 4367-4372. [218] H.A. Patel, C.T. Yavuz, Noninvasive functionalization of polymers of intrinsic microporosity for enhanced CO2 capture, Chem. Commun. 48 (2012) 9989-9991. [2 9]

.R. Mason, L. Maynard-Atem, K. .J. Heard, B. Satilmis, P.M. Budd, K. Friess, M.

Lanc , P. Bernardo, G. Clarizia, J.C. Jansen, Enhancement of CO2 affinity in a polymer of intrinsic microporosity by amine modification, Macromolecules 47 (2014) 1021-1029. [220] J.K. Moore, M.D. Guiver, N. Du, S.E. Hayes, M.S. Conradi, Molecular motions of adsorbed CO2 on a tetrazole-functionalized PIM polymer studied with 13C NMR, J. Phys. Chem. C 117 (2013) 22995-22999. [221] R. Swaidan, B.S. Ghanem, E. Litwiller, I. Pinnau, Pure- and mixed-gas CO2/CH4 separation properties of PIM-1 and an amidoxime-functionalized PIM-1, J. Membr. Sci. 457 (2014) 95-102. [222] F.Y. Li, Y. Xiao, T.-S. Chung, S. Kawi, High-performance thermally self-cross-linked polymer of intrinsic microporosity (PIM-1) membranes for energy development, Macromolecules 45 (2012) 1427-1437. [223] Q. Song, S. Cao, P. Zavala-Rivera, L. Ping Lu, W. Li, Y. Ji, S.A. Al-Muhtaseb, A.K. Cheetham, E. Sivaniah, Photo-oxidative enhancement of polymeric molecular sieve membranes, Nat. Commun. 4 (2013) 1918.

146

[224] Q. Song, S. Cao, R.H. Pritchard, B. Ghalei, S.A. Al-Muhtaseb, E.M. Terentjev, A.K. Cheetham, E. Sivaniah, Controlled thermal oxidative crosslinking of polymers of intrinsic microporosity towards tunable molecular sieve membranes, Nat. Commun. 5 (2014) 4813. [225] C.H. Lau, P.T. Nguyen, M.R. Hill, A.W. Thornton, K. Konstas, C.M. Doherty, R.J. Mulder, L. Bourgeois, A.C.Y. Liu, D.J. Sprouster, J.P. Sullivan, T.J. Bastow, A.J. Hill, D.L. Gin, R.D. Noble, Ending aging in super glassy polymer membranes, Angew. Chem. Int. Ed. 126 (2014) 5426-5430. [226] H.Q. Lin, B.D. Freeman, Materials selection guidelines for membranes that remove CO2 from gas mixtures, Journal of Molecular Structure 788 (2006) 250-250. [227] H. Lin, E. Van  Wagner, R. Raharjo, B.D. Freeman, I. Roman, High-performance polymer membranes for natural-gas sweetening, Adv. Mater. 18 (2006) 39-44. [228] H. Lin, E. Van Wagner, B.D. Freeman, L.G. Toy, R.P. Gupta, Plasticization-enhanced hydrogen purification using polymeric membranes, Science 311 (2006) 639-642. [229] H. Lin, B.D. Freeman, Gas solubility, diffusivity and permeability in poly(ethylene oxide), J. Membr. Sci. 239 (2004) 105-117. [230] N.P. Patel, A.C. Miller, R.J. Spontak, Highly CO2-permeable and -selective membranes derived from crosslinked poly(ethylene glycol) and its nanocomposites, Adv. Funct. Mater. 14 (2004) 699-707. [231] H. Lin, B.D. Freeman, Gas and vapor solubility in cross-linked poly(ethylene glycol diacrylate), Macromolecules 38 (2005) 8394-8407. [232] H. Lin, T. Kai, B.D. Freeman, S. Kalakkunnath, D.S. Kalika, The Effect of cross-linking on gas permeability in cross-linked poly(ethylene glycol diacrylate), Macromolecules 38 (2005) 8381-8393. [233] H. Lin, E.V. Wagner, J.S. Swinnea, B.D. Freeman, S.J. Pas, A.J. Hill, S. Kalakkunnath, D.S. Kalika, Transport and structural characteristics of crosslinked poly(ethylene oxide) rubbers, J. Membr. Sci. 276 (2006) 145-161.

147

[234] V.A. Kusuma, S. Matteucci, B.D. Freeman, M.K. Danquah, D.S. Kalika, Influence of phenoxy-terminated short-chain pendant groups on gas transport properties of cross-linked poly(ethylene oxide) copolymers, J. Membr. Sci. 341 (2009) 84-95. [235] L. Kwisnek, S. Heinz, J.S. Wiggins, S. Nazarenko, Multifunctional thiols as additives in UV-cured PEG-diacrylate membranes for CO2 separation, J. Membr. Sci. 369 (2011) 429-436. [236] L. Kwisnek, J. Goetz, K.P. Meyers, S.R. Heinz, J.S. Wiggins, S. Nazarenko, PEG containing thiol–ene network membranes for CO2 separation: Effect of cross-linking on thermal, mechanical, and gas transport properties, Macromolecules 47 (2014) 3243-3253. [237] S.P. Nunes, A. Car, From charge-mosaic to micelle self-assembly: Block copolymer membranes in the last 40 years, Ind. Eng. Chem. Res. 52 (2012) 993-1003. [238] H. Chen, Y. Xiao, T.-S. Chung, Synthesis and characterization of poly (ethylene oxide) containing copolyimides for hydrogen purification, Polymer 51 (2010) 4077-4086. [239] S.J. Metz, M.H.V. Mulder, M. Wessling, Gas-permeation properties of poly(ethylene oxide) poly(butylene terephthalate) block copolymers, Macromolecules 37 (2004) 4590-4597. [240] A. Car, C. Stropnik, W. Yave, K.V. Peinemann, Tailor-made polymeric membranes based on segmented block copolymers for CO2 separation, Adv. Funct. Mater. 18 (2008) 2815-2823. [241] W. Yave, A. Szymczyk, N. Yave, Z. Roslaniec, Design, synthesis, characterization and optimization of PTT-b-PEO copolymers: A new membrane material for CO2 separation, J. Membr. Sci. 362 (2010) 407-416. [242] V.I. Bondar, B.D. Freeman, I. Pinnau, Gas transport properties of poly(ether-b-amide) segmented block copolymers, J. Polym. Sci., Part B: Polym. Phys. 38 (2000) 2051-2062. [243] J.H. Kim, S.Y. Ha, Y.M. Lee, Gas permeation of poly(amide-6-b-ethylene oxide) copolymer, J. Membr. Sci. 190 (2001) 179-193. [244] E. Tocci, A. Gugliuzza, L. De Lorenzo, M. Macchione, G. De Luca, E. Drioli, Transport properties of a co-poly(amide-12-b-ethylene oxide) membrane: A comparative study between experimental and molecular modelling results, J. Membr. Sci. 323 (2008) 316-327.

148

[245] S.R. Reijerkerk, A. Arun, R.J. Gaymans, K. Nijmeijer, M. Wessling, Tuning of mass transport properties of multi-block copolymers for CO2 capture applications, J. Membr. Sci. 359 (2010) 54-63. [246] S.R. Reijerkerk, A.C. Ijzer, K. Nijmeijer, A. Arun, R.J. Gaymans, M. Wessling, subambient temperature CO2 and light gas permeation through segmented block copolymers with tailored soft phase, ACS Appl. Mater. Interfaces 2 (2010) 551-560. [247] H.W. Kim, H.B. Park, Gas diffusivity, solubility and permeability in polysulfone– poly(ethylene oxide) random copolymer membranes, J. Membr. Sci. 372 (2011) 116-124. [248] H. Li, B.D. Freeman, O.M. Ekiner, Gas permeation properties of poly(urethane-urea)s containing different polyethers, J. Membr. Sci. 369 (2011) 49-58. [249] W. Yave, A. Car, K.-V. Peinemann, M.Q. Shaikh, K. Rätzke, F. Faupel, Gas permeability and free volume in poly(amide-b-ethylene oxide)/polyethylene glycol blend membranes, J. Membr. Sci. 339 (2009) 177-183. [250] W. Yave, A. Car, K.V. Peinemann, Nanostructured membrane material designed for carbon dioxide separation, J. Membr. Sci. 350 (2010) 124-129. [251] A. Ghadimi, M. Amirilargani, T. Mohammadi, N. Kasiri, B. Sadatnia, Preparation of alloyed poly(ether block amide)/poly(ethylene glycol diacrylate) membranes for separation of CO2/H2 (syngas application), J. Membr. Sci. 458 (2014) 14-26. [252] H. Rabiee, M. Soltanieh, S.A. Mousavi, A. Ghadimi, Improvement in CO2/H2 separation by fabrication of poly(ether-b-amide6)/glycerol triacetate gel membranes, J. Membr. Sci. 469 (2014) 43-58. [253] W. Yave, A. Car, S.S. Funari, S.P. Nunes, K.V. Peinemann, CO2-philic polymer membrane with extremely high separation performance, Macromolecules 43 (2010) 326-333. [254] S.R. Reijerkerk, M. Wessling, K. Nijmeijer, Pushing the limits of block copolymer membranes for CO2 separation, J. Membr. Sci. 378 (2011) 479-484. [255] T.C. Merkel, H. Lin, X. Wei, R. Baker, Power plant post-combustion carbon dioxide capture: An opportunity for membranes, J. Membr. Sci. 359 (2010) 126-139.

149

[256] T.C. Merkel, X. Wei, Z. He, L.S. White, J.G. Wijmans, R.W. Baker, Selective exhaust gas recycle with membranes for CO2 capture from natural gas combined cycle power plants, Ind. Eng. Chem. Res. 52 (2012) 1150-1159. [257] W. Yave, A. Car, J. Wind, K.V. Peinemann, Nanometric thin film membranes manufactured on square meter scale: ultra-thin films for CO2 capture, Nanotechnology 21 (2010) 395301. [258] W. Yave, H. Huth, A. Car, C. Schick, Peculiarity of a CO2-philic block copolymer confined in thin films with constrained thickness: "a super membrane for CO2-capture", Energ Environ. Sci. 4 (2011) 4656-4661. [259] S. Li, Z. Wang, C. Zhang, M. Wang, F. Yuan, J. Wang, S. Wang, Interfacially polymerized thin film composite membranes containing ethylene oxide groups for CO2 separation, J. Membr. Sci. 436 (2013) 121-131. [260] H.Z. Chen, Z. Thong, P. Li, T.-S. Chung, High performance composite hollow fiber membranes for CO2/H2 and CO2/N2 separation, Int. J. Hydrogen Energy 39 (2014) 5043-5053. [261] J.E. Bara, D.E. Camper, D.L. Gin, R.D. Noble, Room-temperature ionic liquids and composite materials: Platform technologies for CO2 capture, Acc. Chem. Res. 43 (2010) 152-159. [262] X. Zhang, X. Zhang, H. Dong, Z. Zhao, S. Zhang, Y. Huang, Carbon capture with ionic liquids: overview and progress, Energ Environ. Sci. 5 (2012) 6668-6681. [263] Y.-F. Hu, Z.-C. Liu, C.-M. Xu, X.-M. Zhang, The molecular characteristics dominating the solubility of gases in ionic liquids, Chem. Soc. Rev. 40 (2011) 3802-3823. [264] H.Z. Chen, P. Li, T.-S. Chung, PVDF/ionic liquid polymer blends with superior separation performance for removing CO2 from hydrogen and flue gas, Int. J. Hydrogen Energy 37 (2012) 11796-11804. [265] J. Tang, W. Sun, H. Tang, M. Radosz, Y. Shen, Enhanced CO2 absorption of poly(ionic liquid)s, Macromolecules 38 (2005) 2037-2039.

150

[266] J.E. Bara, S. Lessmann, C.J. Gabriel, E.S. Hatakeyama, R.D. Noble, D.L. Gin, Synthesis and performance of polymerizable room-temperature ionic liquids as gas separation membranes, Ind. Eng. Chem. Res. 46 (2007) 5397-5404. [267] J.E. Bara, C.J. Gabriel, E.S. Hatakeyama, T.K. Carlisle, S. Lessmann, R.D. Noble, D.L. Gin, Improving CO2 selectivity in polymerized room-temperature ionic liquid gas separation membranes through incorporation of polar substituents, J. Membr. Sci. 321 (2008) 3-7. [268] J.E. Bara, R.D. Noble, D.L. Gin, Effect of “Free” ation substituent on gas separation performance of polymer-room-temperature ionic liquid composite membranes, Ind. Eng. Chem. Res. 48 (2009) 4607-4610. [269] Y. Li, S. Wang, G. He, H. Wu, F. Pan, Z. Jiang, Facilitated transport of small molecules and ions for energy-efficient membranes, Chem. Soc. Rev. 44 (2015) 103-118. [270] W.J. Ward, W.L. Robb, Carbon dioxide-oxygen separation: Facilitated transport of carbon dioxide across a liquid film, Science 156 (1967) 1481-1484. [271] S. Kasahara, E. Kamio, T. Ishigami, H. Matsuyama, Amino acid ionic liquid-based facilitated transport membranes for CO2 separation, Chem. Commun. 48 (2012) 6903-6905. [272] S. Kasahara, E. Kamio, T. Ishigami, H. Matsuyama, Effect of water in ionic liquids on CO2 permeability in amino acid ionic liquid-based facilitated transport membranes, J. Membr. Sci. 415–416 (2012) 168-175. [273] P. Cserjési, N. Nemestóthy, K. Bélafi-Bakó, Gas separation properties of supported liquid membranes prepared with unconventional ionic liquids, J. Membr. Sci. 349 (2010) 6-11. [274] H. Matsuyama, A. Terada, T. Nakagawara, Y. Kitamura, M. Teramoto, Facilitated transport of CO2 through polyethylenimine/poly(vinyl alcohol) blend membrane, J. Membr. Sci. 163 (1999) 221-227. [275] Y. Cai, Z. Wang, C. Yi, Y. Bai, J. Wang, S. Wang, Gas transport property of polyallylamine–poly(vinyl alcohol)/polysulfone composite membranes, J. Membr. Sci. 310 (2008) 184-196. [276] T.-J. Kim, B. Li, M.-B. Hägg, Novel fixed-site–carrier polyvinylamine membrane for carbon dioxide capture, J. Polym. Sci., Part B: Polym. Phys. 42 (2004) 4326-4336. 151

[277] A.S. Kovvali, H. Chen, K.K. Sirkar, Dendrimer membranes: A CO2-selective molecular gate, J. Am. Chem. Soc. 122 (2000) 7594-7595. [278] P. Li, Z. Wang, W. Li, Y. Liu, J. Wang, S. Wang, High-performance multilayer composite membranes with mussel-inspired polydopamine as a versatile molecular bridge for CO2 separation, ACS Appl. Mater. Interfaces 7 (2015) 15481-15493. [279] L.Y. Deng, T.J. Kim, M.B. Hagg, Facilitated transport of CO2 in novel PVAm/PVA blend membrane, J. Membr. Sci. 340 (2009) 154-163. [280] S. Duan, T. Kai, I. Taniguchi, S. Kazama, Development of poly(amidoamine) dendrimer/poly(vinyl alcohol) hybrid membranes for CO2 separation, Desalin Water Treat 51 (2013) 5337-5342. [281] I. Taniguchi, S. Duan, S. Kazama, Y. Fujioka, Facile fabrication of a novel high performance CO2 separation membrane: Immobilization of poly(amidoamine) dendrimers in poly(ethylene glycol) networks, J. Membr. Sci. 322 (2008) 277-280. [282] I. Taniguchi, S. Duan, T. Kai, S. Kazama, H. Jinnai, Effect of the phase-separated structure on CO2 separation performance of the poly(amidoamine) dendrimer immobilized in a poly(ethylene glycol) network, J. Mater. Chem. A 1 (2013) 14514-14523. [283] S. Yuan, Z. Wang, Z. Qiao, M. Wang, J. Wang, S. Wang, Improvement of CO 2/N2 separation characteristics of polyvinylamine by modifying with ethylenediamine, J. Membr. Sci. 378 (2011) 425-437. [284] Z. Qiao, Z. Wang, C. Zhang, S. Yuan, Y. Zhu, J. Wang, S. Wang, PVAm–PIP/PS composite membrane with high performance for CO2/N2 separation, AIChE J. 59 (2013) 215-228. [285] Z. Qiao, Z. Wang, S. Zhao, S. Yuan, J. Wang, S. Wang, High adsorption performance polymers modified by small molecules containing functional groups for CO2 separation, RSC Adv. 3 (2013) 50-54. [286] Z. Ma, Z. Qiao, Z. Wang, X. Cao, Y. He, J. Wang, S. Wang, CO2 separation enhancement of the membrane by modifying the polymer with a small molecule containing amine and ester groups, RSC Adv. 4 (2014) 21313-21317. 152

[287] Z. Qiao, Z. Wang, S. Yuan, J. Wang, S. Wang, Preparation and characterization of small molecular amine modified PVAm membranes for CO2/H2 separation, J. Membr. Sci. 475 (2015) 290-302. [288] H. Bai, W.S. Winston Ho, Carbon dioxide-selective membranes for high-pressure synthesis gas purification, Ind. Eng. Chem. Res. 50 (2011) 12152-12161. [289] T.J. Kim, H. Vrålstad, M. Sandru, M.B. Hägg, Separation performance of PVAm composite membrane for CO2 capture at various pH levels, J. Membr. Sci. 428 (2013) 218-224. [290] Y. Zhao, W.S. Winston Ho, Steric hindrance effect on amine demonstrated in solid polymer membranes for CO2 transport, J. Membr. Sci. 415–416 (2012) 132-138. [291] Y. Zhao, W.S.W. Ho, CO2-selective membranes containing sterically hindered amines for CO2/H2 separation, Ind. Eng. Chem. Res. 52 (2013) 8774-8782. [292] W. He, Z. Wang, W. Li, S. Li, Z. Bai, J. Wang, S. Wang, Cyclic tertiary amino group containing fixed carrier membranes for CO2 separation, J. Membr. Sci. 476 (2015) 171-181. [293] P. Li, Z. Wang, Y. Liu, S. Zhao, J. Wang, S. Wang, A synergistic strategy via the combination of multiple functional groups into membranes towards superior CO2 separation performances, J. Membr. Sci. 476 (2015) 243-255. [294] N.V. Blinova, F. Svec, Functionalized polyaniline-based composite membranes with vastly improved performance for separation of carbon dioxide from methane, J. Membr. Sci. 423–424 (2012) 514-521. [295] X. Yu, Z. Wang, Z. Wei, S. Yuan, J. Zhao, J. Wang, S. Wang, Novel tertiary amino containing thin film composite membranes prepared by interfacial polymerization for CO 2 capture, J. Membr. Sci. 362 (2010) 265-278. [296] F. Yuan, Z. Wang, S. Li, J. Wang, S. Wang, Formation–structure–performance correlation of thin film composite membranes prepared by interfacial polymerization for gas separation, J. Membr. Sci. 421–422 (2012) 327-341.

153

[297] M. Wang, Z. Wang, S. Li, C. Zhang, J. Wang, S. Wang, A high performance antioxidative and acid resistant membrane prepared by interfacial polymerization for CO2 separation from flue gas, Energ Environ. Sci. 6 (2013) 539-551. [298] M. Wang, Z. Wang, J. Wang, Y. Zhu, S. Wang, An antioxidative composite membrane with the carboxylate group as a fixed carrier for CO2 separation from flue gas, Energ Environ. Sci. 4 (2011) 3955-3959. [299] K. Yao, Z. Wang, J. Wang, S. Wang, Biomimetic material-poly(N-vinylimidazole)-zinc complex for CO2 separation, Chem. Commun. 48 (2012) 1766-1768. [300] J.H. Lee, J. Hong, J.H. Kim, Y.S. Kang, S.W. Kang, Facilitated CO2 transport membranes utilizing positively polarized copper nanoparticles, Chem. Commun. 48 (2012) 5298-5300. [301] J.H. Oh, Y.S. Kang, S.W. Kang, Poly(vinylpyrrolidone)/KF electrolyte membranes for facilitated CO2 transport, Chem. Commun. 49 (2013) 10181-10183. [302] F. Li, Y. Li, T.-S. Chung, S. Kawi, Facilitated transport by hybrid POSS®–Matrimid®– Zn2+ nanocomposite membranes for the separation of natural gas, J. Membr. Sci. 356 (2010) 14-21. [303] Y. Li, Q. Xin, H. Wu, R. Guo, Z. Tian, Y. Liu, S. Wang, G. He, F. Pan, Z. Jiang, Efficient CO2 capture by humidified polymer electrolyte membranes with tunable water state, Energ Environ. Sci. 7 (2014) 1489-1499. [304] I.S. Chae, M. Kim, Y.S. Kang, S.W. Kang, Enhanced CO2 carrier activity of potassium cation with fluorosilicate anions for facilitated transport membranes, J. Membr. Sci. 466 (2014) 357-360. [305] M. Washim Uddin, M.-B. Hägg, Natural gas sweetening—the effect on CO2–CH4 separation after exposing a facilitated transport membrane to hydrogen sulfide and higher hydrocarbons, J. Membr. Sci. 423–424 (2012) 143-149. [306] J. Liao, Z. Wang, C. Gao, S. Li, Z. Qiao, M. Wang, S. Zhao, X. Xie, J. Wang, S. Wang, Fabrication of high-performance facilitated transport membranes for CO2 separation, Chem. Sci. 5 (2014) 2843-2849. 154

[307] S.C. Li, Z. Wang, W.J. He, C.X. Zhang, H.Y. Wu, J.X. Wang, S.C. Wang, Effects of minor SO2 on the transport properties of fixed carrier membranes for CO2 capture, Ind. Eng. Chem. Res. 53 (2014) 7758-7767. [308] T.-J. Kim, M.W. Uddin, M. Sandru, M.-B. Hägg, The effect of contaminants on the composite membranes for CO2 separation and challenges in up-scaling of the membranes, Energy Procedia 4 (2011) 737-744. [309] M. Sandru, T.-J. Kim, W. Capala, M. Huijbers, M.-B. Hägg, Pilot scale testing of polymeric membranes for CO2 capture from coal fired power plants, Energy Procedia 37 (2013) 6473-6480. [310] N.P. Patel, A.C. Miller, R.J. Spontak, Highly CO2-permeable and selective polymer nanocomposite membranes, Adv. Mater. 15 (2003) 729-733. [311] Y. Li, T.-S. Chung, Molecular-level mixed matrix membranes comprising Pebax® and POSS for hydrogen purification via preferential CO2 removal, Int. J. Hydrogen Energy 35 (2010) 10560-10568. [312] B. Yu, H. Cong, Z. Li, J. Tang, X.S. Zhao, Pebax-1657 nanocomposite membranes incorporated with nanoparticles/colloids/carbon nanotubes for CO2/N2 and CO2/H2 separation, J. Appl. Polym. Sci. 130 (2013) 2867-2876. [313] D. Zhao, J. Ren, H. Li, X. Li, M. Deng, Gas separation properties of poly(amide-6-b-ethylene oxide)/amino modified multi-walled carbon nanotubes mixed matrix membranes, J. Membr. Sci. 467 (2014) 41-47. [314] M.M. Rahman, V. Filiz, S. Shishatskiy, C. Abetz, S. Neumann, S. Bolmer, M.M. Khan, V. Abetz, PEBAX® with PEG functionalized POSS as nanocomposite membranes for CO2 separation, J. Membr. Sci. 437 (2013) 286-297. [315] M.L. Sforça, I.V.P. Yoshida, S.P. Nunes, Organic–inorganic membranes prepared from polyether diamine and epoxy silane, J. Membr. Sci. 159 (1999) 197-207. [316] L. Shao, T.-S. Chung, In situ fabrication of cross-linked PEO/silica reverse-selective membranes for hydrogen purification, Int. J. Hydrogen Energy 34 (2009) 6492-6504.

155

[317] C.H. Lau, S. Liu, D.R. Paul, J. Xia, Y.-C. Jean, H. Chen, L. Shao, T.-S. Chung, Silica nanohybrid membranes with high CO2 affinity for green hydrogen purification, Adv Energy Mater 1 (2011) 634-642. [318] C.H. Lau, T.-S. Chung, Effects of Si–O–Si Agglomerations on CO2 transport and separation properties of sol-derived nanohybrid membranes, Macromolecules 44 (2011) 6057-6066. [319] J. Xia, S. Liu, C.H. Lau, T.-S. Chung, Liquidlike Poly(ethylene glycol) supported in the organic–inorganic matrix for CO2 removal, Macromolecules 44 (2011) 5268-5280. [320] M.L. Chua, L. Shao, B.T. Low, Y. Xiao, T.-S. Chung, Polyetheramine–polyhedral oligomeric silsesquioxane organic–inorganic hybrid membranes for CO2/H2 and CO2/N2 separation, J. Membr. Sci. 385–386 (2011) 40-48. [321] J. Zhao, Z. Wang, J. Wang, S. Wang, High-performance membranes comprising polyaniline nanoparticles incorporated into polyvinylamine matrix for CO2/N2 separation, J. Membr. Sci. 403–404 (2012) 203-215. [322] S. Zhao, Z. Wang, Z. Qiao, X. Wei, C. Zhang, J. Wang, S. Wang, Gas separation membrane with CO2-facilitated transport highway constructed from amino carrier containing nanorods and macromolecules, J. Mater. Chem. A 1 (2013) 246-249. [323] L. Deng, M.-B. Hägg, Carbon nanotube reinforced PVAm/PVA blend FSC nanocomposite membrane for CO2/CH4 separation, Int. J. Greenhouse Gas Control 26 (2014) 127-134. [324] Y.A. Zhao, B.T. Jung, L. Ansaloni, W.S.W. Ho, Multiwalled carbon nanotube mixed matrix membranes containing amines for high pressure CO2/H2 separation, J. Membr. Sci. 459 (2014) 233-243. [325] J.E. Ramírez-Morales, E. Tapia-Venegas, N. Nemestóthy, P. Bakonyi, K. Bélafi-Bakó, G. Ruiz-Filippi, Evaluation of two gas membrane modules for fermentative hydrogen separation, Int. J. Hydrogen Energy 38 (2013) 14042-14052.

156

[326] P. Bakonyi, G. Kumar, N. Nemestóthy, C.Y. Lin, K. Bélafi-Bakó, Biohydrogen purification using a commercial polyimide membrane module: Studying the effects of some process variables, Int. J. Hydrogen Energy 38 (2013) 15092-15099. [327] P. Bakonyi, N. Nemestóthy, J. Lankó, I. Rivera, G. Buitrón, K. Bélafi-Bakó, Simultaneous biohydrogen production and purification in a double-membrane bioreactor system, Int. J. Hydrogen Energy 40 (2015) 1690-1697. [328] S. Li, Z. Wang, X. Yu, J. Wang, S. Wang, High-performance membranes with multi-permselectivity for CO2 separation, Adv. Mater. 24 (2012) 3196-3200.

157

Captions of figures and tables Figures Fig. 1. Schematic diagrams of (a) H2-selective membranes and (b) CO2-selective membranes. Fig. 2. Schematic diagrams of gas transport mechanisms in gas separation membranes. Fig. 3. Scheme of the synthesis of zeolite LTA membranes on covalently functionalized supports by using 1,4-diisocyanate (DIC-4) as molecular binders to in situ anchor LTA nutrients during hydrothermal synthesis [35]. Fig. 4. (a) Scheme of the layer-by-layer hydrothermal synthesis of sandwich-like multi-layer LTA membranes by using 3-aminopropyltriethoxysilane (APTES) as self-assembled interlayer. (b) Single gas permeances of different gases through the single/multi-layer zeolite LTA membranes at 373 K as a function of the gas kinetic diameter [38]. Fig. 5. Schematic illustrations of (A) leaflet-shaped and (B) coffin-shaped silicalite-1 (SL) crystals and (C) truncated bipyramidal beta (BEA) crystals and their channel systems, as well as their respective (D) a-oriented, (E) b-oriented, and (F) a oriented monolayers. (G to I) Secondary growth on these monolayers produces uniformly oriented films [46]. Fig. 6. Scheme of preparation of bicontinuous ZIF-8@GO membranes through layer-by-layer deposition of graphene oxide on the semicontinuous ZIF-8 layer which was synthesized on a polydopamine-modified alumina disk [118]. Fig. 7. Schematic illustration of in-situ solvothermal growth of ZIF-8 membrane on a ZnAl-LDH buffer layer-modified γ-Al2O3 substrate [120]. Fig. 8. Schematic illustration of membrane synthesis using the counter-diffusion-based in situ method [123].

158

Fig. 9. Schematic diagram of preparation of the MIL-53 membrane on alumina support via the reactive seeding method [127]. Fig. 10. Schematic diagram of step-by-step deposition of btc3- and Cu2+ on alumina support [129]. Fig. 11 Covalent post-functionalization of a ZIF-90 molecular sieve membrane by imine condensation with ethanolamine to enhance H2/CO2 selectivity [131]. Fig. 12. (a) Possible scheme for reaction between 6FDA-durene and diethylenetriamine. (b) Comparison of H2/CO2 separation performances of the unmodified and diethylenetriamine-modified 6FDA-durene films with the Robeson’s upper bound [145]. Fig. 13. Two major factors contributing structural change during thermal rearrangement of polyimides. (A) Change of chain conformation–polymer chains consisting of meta- and/or para-linked chain conformations can be created via rearrangement. (B) Spatial relocation due to chain rearrangement in confinement, which may lead to the generation of free-volume elements [153]. Fig. 14. Schematic of PEI/PAA layer-by-layer gas separation membrane supported on an alumina coated porous stainless steel tube [165]. Fig. 15. Schematic diagram of mixed matrix membranes that inorganic fillers are embedded into the polymer matrix. Fig. 16. Separation performances of representative H2-selective membranes for H2/CO2 separation (square: microporous inorganic membranes, triangle: MOF membranes, star: polymeric membranes, circle: mixed matrix membranes). Fig. 17. Schematic of the brush-like structure in the silica pore after APTES modification (not to scale) [196]. Fig. 18. Synthesis of PIM-1 polymer [205].

159

Fig. 19. (A) The PIM-1 polymer contains SBI and dioxane linkages, but both can bend and flex to a considerable extent. (B) The PIM-EA-TB polymer has a more rigid architecture [207]. Fig. 20. PIMs with various substituted pendant functional groups via post modification [212-219]. Fig. 21. Two-dimensional representations of the contorted PIM-1 membrane before and after thermal cross-linking reaction with the formation of triazine rings [222]. Fig. 22. Schematic representation of PEGDA/PEGMEA copolymer network. Italicized and bolded parts of the network

derive

from

the

cross-linker.

R1

is

CO(OCH2CH2)8OCH3

from

PEGMEA;

R2

is

COO(CH2CH2O)14OC from PEGDA [228]. Fig. 23. (a) Schematic representation of the block copolymer organization in thick films (semi-crystalline polymer and high Tg); (b) representation of a polymer chain and Rg; and (c) block copolymer organization within a super ultrathin film under the influence of the PDMS substrate (thin film mostly amorphous with high fractional free volume and low Tg of the PEO segment) [258]. Fig. 24. Polymerization reaction for preparation of the thin film composite membranes by interfacial polymerization [259]. Fig. 25. Representations of (a) poly(RTIL) framework with polymer-bound cations and (b) poly(RTIL)-RTIL composite containing 20 mol % free cations [268]. Fig. 26 Chemical structure of representative amino groups containing polymers. Fig. 27. Schematic diagram of the intermolecular hydrogen bonds in the small molecule amine-modified PVAm samples: (1) EDA-modified PVAm; (2) PIP-modified PVAm; (3) MEA-modified PVAm; (4) MC-modified PVAm [285].

160

Fig. 28. Scheme of (a) functionalization of the polyaniline membrane first photografted with glycidyl methacrylate and 2-hydroxyethylmethacrylate and then reacted with diamines; (b) additional functionalization of membrane containing hexamethylenediamine with 2-ethyl-2-thiopseudourea [294]. Fig. 29. Schematic representation of the side-chain grafted organic-inorganic network [317]. Fig. 30. Schematic representation of the CO2-facilitated transport highway in the PVAm-PANI selective layer, (a) intra-channel CO2 transport and (b) inter-channel CO2 transport [322]. Fig. 31. Schematic diagrams of (a) hydrotalcite containing free movable hydrated carbonate groups and (b) facilitated transport membrane containing high-speed facilitated transport channels [306]. Fig. 32. Separation performances of representative CO2-selective membranes for CO2/H2 separation (square: microporous membranes, triangle: PIM membranes, star: CO2-philic polymeric membranes, circle: facilitated transport membranes, diamond: mixed matrix membranes).

Tables Table 1. Distinction between CO2 and H2 in various properties relevant to separation [19]

Highlights  Summarize recent advances in membranes for hydrogen purification comprehensively.  H2-selective and CO2-selective membranes were reviewed based on materials.  H2-selective and CO2-selective membranes were compared.  Discuss future directions in developing membranes for hydrogen purification.

161

162