Method of preparing perm-selective porous membrane and method of separating gases using porous membrane prepared thereby

Abstract

The present invention relates to a method of preparing a perm-selective porous membrane and a method of separating gases using the prepared porous membrane. According to the present invention, a membrane is synthesized using a hierarchically structured alumina porous support by a counter diffusion method. During this synthesis, the diffusion rate of metal ions loaded on the porous support is controlled by controlling the pore size of the porous support, and the position at which the membrane is synthesized is controlled by synthesizing the membrane inside the support. This can increase the physical stability of the membrane and make the membrane thicker so as to ensure higher H.sub.2/CO.sub.2 separation factors.

Claims

1. A method of preparing a porous membrane, comprising: coating particles on a porous support to form a layer comprising pores smaller than pores of the porous support to synthesize a hierarchically structured porous support; and performing counter diffusion to synthesize a membrane comprising zeolitic imidazolate frameworks (ZIFs) formed inside the hierarchically structured porous support, wherein the pores of the layer have a pore size of 2 to 50 nm, and wherein the performing counter diffusion comprises immersing the hierarchically structured porous support in a metal salt solution, and then in an imidazole or imidazole derivative solution.

2. The method of preparing a porous membrane of claim 1, wherein the performing counter diffusion comprises adsorbing Zn.sup.2+ or Co.sup.2+ on the porous support by immersing the hierarchically structured porous support synthesized in the coating particles in a zinc salt solution or a cobalt salt solution, and then forming Zn.sup.2+ or Co.sup.2+ imidazole crystals inside the porous support by immersing the porous support adsorbed with Zn.sup.2+ or Co.sup.2+ in an imidazole or imidazole derivative solution.

3. The method of preparing a porous membrane of claim 1, wherein the pores of the layer are formed on one side of the hierarchically structured porous support.

4. The method of preparing a porous membrane of claim 3, wherein the porous support has a pore size of 50 to 1000 nm.

5. The method of preparing a porous membrane of claim 1, wherein the particles are at least one selected from the group consisting of -alumina, -alumina, silica, glass, mullite, zirconia, titania, yttria, ceria, vanadia, silicone, stainless steel, calcium oxide, phosphor oxide and mixtures thereof.

6. The method of preparing a porous membrane of claim 1, wherein the metal salt is at least one selected from the group consisting of zinc nitrate, zinc acetate, zinc chloride, zinc sulfate, zinc bromide and zinc iodide.

7. The method of preparing a porous membrane of claim 1, wherein the metal salt is at least one selected from the group consisting of cobalt nitrate, cobalt acetate, cobalt chloride, cobalt sulfate, cobalt bromide and cobalt iodide.

8. The method of preparing a porous membrane of claim 1, wherein the imidazole or imidazole derivative is at least one selected from the group consisting of benzimidazole, 2-methylimidazole, 4-methylimidazole, 2-methylbenzimidazole, 2-nitroimidazole, 5-nitrobenzimidazole and 5-chlorobenzimidazole.

9. The method of preparing a porous membrane of claim 1, wherein the metal salt solution, or the imidazole or imidazole derivative solution is dissolved in at least one solvent selected from the group consisting of methanol, ethanol, propanol, iso-propanol, tert-butanol, n-butanol, methoxyethanol, ethoxyethanol, dimethylacetateamide, dimethylformamide, N-methyl-2-pyrrolidone (NMP), formic acid, nitromethane, acetic acid and distilled water.

10. The method of preparing a porous membrane of claim 1, wherein the porous membrane has a thickness of 0.1 to 500 m.

11. A porous membrane prepared by the method of claim 1, wherein the ZIFs are formed inside the hierarchically structured porous support comprising a layer having pores of 2 to 50 nm, wherein the layer is formed by coating ceramic particles on one side of a ceramic support.

12. A method of separating hydrogen from a mixed gas of H.sub.2 and CO.sub.2 or syngas using the porous membrane of claim 11.

13. The method of separating hydrogen of claim 12, wherein the separating hydrogen is performed at a temperature of 100 to 500 C.

14. The porous membrane of claim 11, wherein the ceramic support is a porous support comprising pores of 50 to 1000 nm.

15. The porous membrane of claim 11, wherein the porous membrane has a thickness of 0.1 to 500 m.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The patent or application file contains a least one drawing executed in color. Copies of this patent or patent application publication with color drawing(s) will be provided by the Office upon request and payment of the necessary fee

(2) FIG. 1 schematically shows a ZIF-8 membrane synthesized using a conventional porous support (left) and a ZIF-8 membrane synthesized using a hierarchically structured porous membrane (right) according to an example of the present invention.

(3) FIG. 2 shows SEM images of an (a1)-(a2) -Al.sub.2O.sub.3 disc and (b1)-(b2) -/-Al.sub.2O.sub.3 disc according to an example of the present invention.

(4) FIG. 3 shows XRD patterns of an -Al.sub.2O.sub.3 disc (bottom) and -/-Al.sub.2O.sub.3 disc (top) according to an example of the present invention.

(5) FIG. 4 shows cross-sectional SEM images of membranes (a1)-(a2) ZIF-8_ and (b1)-(b2) ZIF-8_ according to an example of the present invention.

(6) FIG. 5 shows cross-sectional SEM images at a low magnification (1.sup.st column) and taken in the four support positions (approximately indicated in the low magnification SEM image in the 1.sup.st column) (2.sup.nd to 5.sup.th columns) for membranes (a1)-(a4) ZIF-8_ and (b1)-(b4) ZIF-8_ according to an example of the present invention.

(7) FIG. 6 shows XRD patterns for membranes ZIF-8_ and ZIF-8_ with simulated XRD patterns of ZIF-8 (bottom) and ZnO (top) added as references, according to an example of the present invention.

(8) FIG. 7 shows N.sub.2 adsorption isotherms at 77 K and pore size distribution of supports (the -Al.sub.2O.sub.3 and -/-Al.sub.2O.sub.3 discs), and supported membranes (ZIF-8_ and ZIF-8_) with the BJH analysis according to an example of the present invention. (a) and (d): N.sub.2 adsorption isotherms, (b) and (e): BJH adsorption branch, and (c) and (f): BJH desorption branch.

(9) FIG. 8 shows H.sub.2/CO.sub.2 separation performances of (a1)-(b1) intact and (a2)-(b2) intentionally damaged membranes ZIF-8_ and ZIF-8_ as a function of temperature up to 200 C. and H.sub.2/CO.sub.2 separation performances of (a3)-(b3) intact membranes ZIF-8_ and ZIF-8_ up to 300 C., according to an example of the present invention.

(10) FIG. 9 shows the contact angle of a water droplet on membranes ZIF-8_ (top) and ZIF-8_ (bottom) according to an example of the present invention.

(11) FIG. 10 shows the H.sub.2/CO.sub.2 separation performance of membranes (a1)-(a2) ZIF-8_ and (b1)-(b2) ZIF-8_ tested to investigate their thermal stabilities up to 72 h at 200 and 250 C. according to an example of the present invention.

(12) FIG. 11 shows XRD patterns of membranes (a1)-(a2) ZIF-8_ and (b1)-(b2) ZIF-8_ before and after the long-term stability tests at 200 and 250 C. according to an example of the present invention.

(13) FIG. 12 shows the H.sub.2/CO.sub.2 separation performances of membranes (a) ZIF-8_ and (b) ZIF-8_ measured at 300 C. up to 72 h along with (c)-(d) their XRD patterns before and after the thermal stability test according to an example of the present invention.

(14) FIG. 13 shows XRD patterns of ZIF-8 particles dried at 70 C. for 24 h and heat-treated at 300 C. for 7 days in a CO.sub.2 environment according to an example of the present invention.

(15) FIG. 14 is a graph showing H.sub.2/CO.sub.2 SFs vs. H.sub.2 permeance through membranes ZIF-8_ (denoted by ) and ZIF-8_ (denoted by ) along with those through other types of ZIF membranes according to an example of the present invention.

BEST MODE FOR CARRYING OUT THE INVENTION

(16) Unless defined otherwise, all the technical and scientific terms used herein have the same meaning as those generally understood by one of ordinary skill in the art to which the invention pertains. Generally, the nomenclature used herein and the experiment methods, which will be described below, are those well known and commonly employed in the art.

(17) In the present invention, alumina having a pore size different from that of a conventional alumina porous support is coated on the conventional alumina porous support, thereby synthesizing a hierarchically structured porous support. Using the hierarchically structured porous support, a ZIF-8 membrane is synthesized by the counter diffusion method. Namely, when a porous support adsorbed with Zn.sup.2+ ions by immersion in a Zn.sup.2+ ion solution is immersed in a 2-methylimidazole solution, the Zn.sup.2+ ions are diffused from the support to the outside, and the 2-methylimidazole is diffused from the bulk solution toward the support. Thus, the two materials meet each other to form crystals. Here, the Zn.sup.2+ ion and the 2-methylimidazole have high reactivity, and thus when they meet each other, they form crystals almost at the same time. In the case of a ZIF-8 membrane synthesized using a conventional support having only large pores, Zn.sup.2+ ions are diffused quickly to the outside and meet 2-methylimidazole, and thus most of the membrane is formed on the top of the support. In the present invention, a membrane is formed inside a porous support by controlling the pore size of the porous support to reduce the diffusion rate of Zn.sup.2+ ions.

(18) Additionally, the diffusion rate of Zn.sup.2+ metal ions decreases so that an increased amount of the Zn.sup.2+ metal ions can react with 2-methylimidazole inside the porous support, thereby forming a continuous membrane. A conventional membrane is mostly synthesized on the top of a support, whereas a continuous membrane of the present invention can be formed while the inside of a porous support is filled, and the thickness of the membrane can be controlled according to the degree of filling the pores.

(19) Therefore, an aspect of the present invention is directed to a method of preparing a porous membrane, comprising: (a) synthesizing a hierarchically structured porous support by coating particles having pores smaller than those of the porous support on a porous support; and (b) synthesizing a membrane in which zeolitic imidazolate frameworks (ZIFs) are formed inside the hierarchically structured porous support by performing counter diffusion

(20) The step of (b) may comprise immersing the synthesized hierarchically structured porous support of step (a) in a zinc salt solution or a cobalt salt solution, thereby adsorbing Zn.sup.2+ or Co.sup.2+ on the porous support, and then immersing the porous support adsorbed with Zn.sup.2+ or Co.sup.2+ in an imidazole or imidazole derivative solution, thereby forming Zn.sup.2+ or Co.sup.2+ imidazole crystals inside the porous support.

(21) The hierarchically structured porous support may have 2 to 50 nm sized-pores, preferably 3 to 10 nm sized-pores, formed on one side of a wall having 50 to 1000 nm sized-pores, preferably 100 to 500 nm sized-pores.

(22) The particles having pores smaller than those of the porous support in the step of (a), have a pore size of 2 to 50 nm, preferably 3 to 10 nm. For example, the particles may be at least one selected from the group consisting of -alumina, -alumina, silica, glass, mullite, zirconia, titania, yttria, ceria, vanadia, silicone, stainless steel, calcium oxide, phosphor oxide, and mixtures thereof. Preferably, the particles that can be used in the present invention may be -alumina or -alumina, but is not limited thereto.

(23) The zinc salt may be at least one selected from the group consisting of zinc nitrate, zinc acetate, zinc chloride, zinc sulfate, zinc bromide, and zinc iodide.

(24) The cobalt salt may be at least one selected from the group consisting of cobalt nitrate, cobalt acetate, cobalt chloride, cobalt sulfate, cobalt bromide, and cobalt iodide.

(25) The imidazole or imidazole derivative may be at least one selected from the group consisting of benzimidazole, 2-methylimidazole, 4-methylimidazole, 2-methylbenzimidazole, 2-nitroimidazole, 5-nitrobenzimidazole and 5-chlorobenzimidazole. Preferably, the imidazole or imidazole derivative that can be used in the present invention may be 2-methylimidazole, but is not limited thereto.

(26) The zinc salt solution, the cobalt salt solution, or the imidazole or imidazole derivative solution may be dissolved in at least one solvent selected from the group consisting of methanol, ethanol, propanol, iso-propanol, tert-butanol, n-butanol, methoxyethanol, ethoxyethanol, dimethylacetateamide, dimethylformamide, N-methyl-2-pyrrolidone (NMP), formic acid, nitromethane, acetic acid and distilled water.

(27) The porous membrane may have a thickness of 0.1 to 500 m, preferably 5 to 100 m, more preferably 5 to 80 m.

(28) The size of pores in ZIF-8 is about 0.34 nm, which is larger than hydrogen (0.289 nm) and is slightly larger than carbon dioxide (0.33 nm). Considering that the diffusion rate of molecules changes very sensitively to the size of fine pores, H.sub.2/CO.sub.2 separation based on the difference in diffusion rate through a ZIF-8 membrane may be performed. Because H.sub.2/CO.sub.2 separation is suitable for overcoming low conversion rate which is the problem of the water gas shift reaction, it is essential to ensure the high-temperature stability of a membrane. This high-temperature stability can be ensured by synthesizing the membrane inside a porous support. In addition, although this membrane is not applicable as a membrane reactor in the water gas shift reaction, it may be used as a membrane to separate a high-temperature H.sub.2/CO.sub.2 mixture which is discharged from a water gas shift reactor.

(29) FIG. 1 is a schematic view showing that a ZIF-8 membrane can be synthesized inside a porous support by controlling the pore size of the porous support.

(30) Technology that can control the pore size of a porous support so as to control the diffusion rate of metal ions loaded on the support, thereby controlling the active site of the membrane. Since the pore size of the porous support is determined by -Al.sub.2O.sub.3 to be coated thereon, the pore size of the support may be controlled by selecting suitable coating conditions. Thus, the H.sub.2/CO.sub.2 separation factor at high temperatures can be ensured by controlling the active site of the membrane. In addition, the physical stability of the membrane may also be increased by controlling the active site of the membrane. Furthermore, when the diffusion rate of metal ions is controlled so as to increase the amount of metal ions that react inside the support, the thickness of the membrane can become thicker, thereby ensuring a higher H.sub.2/CO.sub.2 separation factor.

(31) ZIF-8 membranes are highly suitable as H.sub.2 selective walls in membranes reactors used for water gas shift reactions because of their efficient molecular sieving properties that favor H.sub.2 transport, coupled with their high thermal stability and processability. In the present invention, an in-situ counter diffusion method is used to prepare ZIF-8 membranes. Zn sources, already placed inside a porous support, are allowed to diffuse out and react with the 2-methylimidazole (mim) ligands in the bulk phase. Because the reaction rates between the Zn source and the mim ligand are very high, their diffusion rates play a key role in determining the final properties of the membranes. To control the diffusion rate, a hierarchically structured support, i.e., a -Al.sub.2O.sub.3 layer-coated -Al.sub.2O.sub.3 disc (-/-Al.sub.2O.sub.3 disc), is used in addition to an intact -Al.sub.2O.sub.3 disc. ZIF-8 membranes in the -Al.sub.2O.sub.3 disc (membrane ZIF-8_) are primarily formed on top similar to a conventional supported-membrane, whereas those in the -/-Al.sub.2O.sub.3 disc (membrane ZIF-8_) are embedded inside the modified support. As desired, the embedded membrane ZIF-8_ shows marked H.sub.2 separation performance with a maximum H.sub.2/CO.sub.2 separation factor (SF) of 9.80.9 at 250 C. (vs. a maximum H.sub.2/CO.sub.2 SF of 8.01.1 for membrane ZIF-8_). Although both type membranes persist at 200 and 250 C. for up to 72 hours, at a higher temperature of 300 C., the membrane performance starts deteriorating after 2 hours and 10 hours for membranes ZIF-8_ and ZIF-8_, respectively. This indicates that the -Al.sub.2O.sub.3 layer serves as a protective layer for preserving the performance of the embedded ZIF-8 membrane. Their performance is completely degraded due to the eventual conversion of ZIF-8 into ZnO phases.

(32) In addition, in an example of the present invention, the hydrogen separation ability (H.sub.2/CO.sub.2 separation) of each of a ZIF-8 membrane synthesized on a conventional porous support (for convenience, referred to as membrane A) and a ZIF-8 membrane synthesized in a hierarchically structured support (for convenience, referred to as membrane B) was measured at 200 C., 250 C. and 300 C. for 3 days. As a result, it was shown that both membrane A and membrane B maintained their hydrogen separation ability at 200 C. and 250 C. for 3 days. Furthermore, it was shown that at 300 C., membrane A maintained its hydrogen separation ability for 2 hours, and membrane B maintained its hydrogen separation ability. Meanwhile, the surface of each of membrane A and membrane B was ground with sand paper, and the hydrogen separation ability of each membrane was measured. As a result, it was found that membrane A almost lost its hydrogen separation ability after grinding with sand paper, whereas membrane maintained its hydrogen separation ability even after grinding with sand paper.

(33) Therefore, an another aspect of the present invention is directed to a porous membrane which is prepared by the method of the present invention, wherein ZIFs are formed inside a hierarchically structured porous support having pores of 2 to 50 nm formed by coating ceramic particles on one side of a ceramic support having pores of 50 to 1000 nm, the porous membrane has a thickness of 0.1 to 500 m.

(34) The porous membrane according to a preferred embodiment of the present invention can allow ZIFs to be formed inside a hierarchically structured porous support having 2 to 50 nm sized-pores formed by coating ceramic particles on one side of a ceramic support having 50 to 1000 nm sized-pores, and can have a thickness of 5 to 80 m.

(35) In another aspect, the present invention is directed to a method of separating hydrogen from a mixed gas of H.sub.2 and CO.sub.2 or syngas using the above-described porous membrane.

(36) The method of separating hydrogen from a mixed gas of H.sub.2 and CO.sub.2 or syngas using the above-described porous membrane may be performed at high temperature, specifically, at a temperature of 100 to 500 C., preferably 200 to 300 C.

(37) Hereinafter, the present invention will be described in further detail with reference to examples. It will be obvious to a person having ordinary skill in the art that these examples are for illustrative purposes only and are not to be construed to limit the scope of the present invention.

Example 1: Synthesis of ZIF-8 Membrane Using Hierarchically Structured Porous Support

(38) A hierarchically structured support was synthesized by coating -Al.sub.2O.sub.3 on a conventional porous support (pore size: 150 nm), and SEM observation confirmed that the support had a hierarchical structure. XRD observation confirmed that the -Al.sub.2O.sub.3 phase appeared, and nitrogen adsorption at 77 K indicated that the hierarchically structured support had a pore size of about 5 nm. A ZIF-8 membrane was synthesized using the hierarchically structured support and analyzed by SEM and XRD. A ZIF-8 membrane was synthesized inside the porous support by controlling the pore size of the porous support to control the diffusion rate of Zn.sup.2+ ions, and the synthesized membrane was analyzed by SEM, XRD and EDX. In addition, in order to examine how the mesopores of the hierarchically structured support were filled with ZIF-8 particles, nitrogen adsorption at 77 K was performed. As a result, it was quantitatively determined that most of the ZIF-8 membrane was synthesized inside the support.

Preparation Example 1: Preparation of -Al.SUB.2.O.SUB.3 .Disc and -/-Al.SUB.2.O.SUB.3 .Disc

(39) An -Al.sub.2O.sub.3 disc (21 mm in diameter and 2 mm in thickness) was prepared according to a method disclosed in a literature (J. Choi et al., Adsorption, 2006, 12, 339-360), and a -/-Al.sub.2O.sub.3 disc was prepared by coating a boehmite sol, prepared according to a conventional method, on top of a bare -Al.sub.2O.sub.3 disc. Initially, alumina-tri-sec-butoxide (ATSB, 97%, Sigma-Aldrich) was mixed with deionized (DI) water pre-heated to 90 C. and the mixture was then stirred at 90 C. for 1 hour under reflux. After that, HCl (35 wt % in water, DAEJUNG) was added to the mixture and further stirred for 1 day under reflux so as to be homogenized. The final composition of the prepared sol was 1 ATSB: 0.07 HCl: 100 H.sub.2O. The prepared boehmite sol was stored in a refrigerator (about 5 C.) and used continually for coating -Al.sub.2O.sub.3 layers onto -Al.sub.2O.sub.3 discs. The -layer was coated on the polished surface of the -Al.sub.2O.sub.3 disc by placing the boehmite sol in contact with the surface for 10 seconds, and the remaining sol on the -Al.sub.2O.sub.3 disc was further wiped away with a paper towel. The -/-Al.sub.2O.sub.3 disc was dried at room temperature overnight and calcined at 600 C. for 3 hours with a heating ramp rate of 0.5 C./min under air flow of 200 mL/min (Pluskolab, model: CRF-M20-UP).

Preparation Example 2: Preparation of ZIF-8 Membrane by Counter Diffusion Method

(40) A ZIF-8 membrane was prepared by adopting the counter diffusion method disclosed in a literature (H. T. Kwon, H. K. Jeong, J. Am. Chem. Soc., 2013, 135, 10763-10768) with a minor difference in the horizontal position of the disc. In brief, the -Al.sub.2O.sub.3 disc or the -/-Al.sub.2O.sub.3 disc was immersed overnight in 0.18 M of Zn solution (ZnCl.sub.2 (99%, Sigma-Aldrich)). Then, the support that was already saturated by the zinc source was placed horizontally at the bottom of the Teflon-lined autoclave where the mim solution (1.58 M mim (98%, Sigma-Aldrich) in methanol) that included sodium formate (99%, Sigma-Aldrich) was already contained. The molar ratio of sodium formate to mim was 0.1. The sealed autoclave was placed in a pre-heated oven at 120 C. and the reaction proceeded for 4 hours. After the reaction, the ZIF-8 membrane was rinsed with methanol and dried in ambient conditions. For convenience, the resulting ZIF-8 membranes were referred to as membranes ZIF-8_ and ZIF-8_, respectively, where and represent the pure -Al.sub.2O.sub.3 disc or the composite -/-Al.sub.2O.sub.3 disc support, respectively. The as-synthesized ZIF-8 membrane was further activated by using a pre-heated oven at 160 C. in ambient air overnight.

(41) Characterizations

(42) Scanning electron microscopy (SEM) images and energy dispersive X-ray spectroscopy (EDX) data were acquired using a field emission scanning electron microscope (FE-SEM, Hitachi S-4800). X-ray diffraction (XRD) patterns of the samples were obtained using a Rigaku Model D/Max-2500V/PC diffractometer (Japan) with Cu K.sub. radiation (40 kV, 100 mA, =0.154 nm). As a reference, the crystallographic information file (CIF) of ZIF-8 structures (deposit no. 602542) was acquired from the Cambridge Crystallographic Data Centre (CCDC, www.ccdc.cdm.ac.uk), and the corresponding XRD pattern was generated with Mercury software (CCDC, www.ccdc.cdm.ac.uk). The water droplet contact angle was measured at room temperature with an optical microscope (General Phoenix 300, Surface Electro Optics). The pore size distributions and porosities of the two kinds of supports, the -Al.sub.2O.sub.3 disc and the -/-Al.sub.2O.sub.3 disc were obtained by using an automatic mercury porosimeter (AutoPore IV 9520, Micromeritics). The specific surface areas of the supports were estimated by the Brunauer-Emmett-Teller (BET) method and the pore size distributions were calculated by using the BJH method. The micropore areas of the ZIF-8 membrane were estimated via the t-plot method. The specific surface area, pore size distribution, and micropore area were determined from the N.sub.2 adsorption isotherm (ASAP 2020, Micromeritics) at 77 K. Samples were degassed at 350 C. for the -Al.sub.2O.sub.3 and -/-Al.sub.2O.sub.3 disc, and at 200 C. for membranes ZIF-8_ and ZIF-8_. The H.sub.2 and CO.sub.2 permeation rates across the ZIF-8 membranes were measured in Wicke-Kallenbach mode with total feed and permeate side pressures of 1 atm. For the H.sub.2/CO.sub.2 permeation measurement up to 200 C., Viton O-rings were used to seal the permeation cell, whereas for the measurement up to 300 C. and long-term thermal stability tests at 200 C., 250 C., and 300 C., Kalrez O-rings were used for appropriate acquisition of the permeation rates. An approximately 100 mL/min total flow rate with 50 kPa H.sub.2/50 kPa CO.sub.2 was fed to the ZIF-8 membrane, while the permeate side was swept by an approximately 100 mL/min He flow. The H.sub.2 and CO.sub.2 permeates in the permeate side were carried by the He sweep gas to an on-line gas chromatograph (GC, YL Instrument, 6500GC System), and prior to the GC, 5 mL/min of CH.sub.4 was mixed into the permeate stream to serve as an internal standard.

(43) For detection, a pulsed discharge ionization detector (PDD) was used to quantify the composition of H.sub.2 and CO.sub.2 in the permeate stream because a thermal conductivity detector (TCD), used generally for characterizing permanent gases, could not distinguish He/H.sub.2 and Ar/CO.sub.2 pairs, apparently due to their similar thermal conductivities.

(44) FIG. 2 shows the top view and cross-sectional view SEM images of both -Al.sub.2O.sub.3 and -/-Al.sub.2O.sub.3 discs. The grains (about 250 to 700 nm) that comprise the -Al.sub.2O.sub.3 disc had an irregular shape (FIG. 2(a1)), while those of the smooth -Al.sub.2O.sub.3 layer in the -/-Al.sub.2O.sub.3 disc, which is supposed to have micropores on the order of 1 nm, were too small to be recognized under SEM resolution (FIG. 1(b1)). The cross-sectional view SEM image of the -/-Al.sub.2O.sub.3 disc in FIG. 2(b2) reveals the formation of an 8 m thick -Al.sub.2O.sub.3 layer on top of the -Al.sub.2O.sub.3 disc (FIG. 1(a2)), constituting an asymmetric structure. In addition, FIG. 3 confirms the formation of the -Al.sub.2O.sub.3 layer. The mercury porosimetry characterization in Table 1 did not distinguish any remarkable discrepancies between the two types of supports, seemingly because the 2,000 m thick -Al.sub.2O.sub.3 disc vastly outweighed the 8 m thick -Al.sub.2O.sub.3 layer in the measurement. Nevertheless, the SEM characterization shown in FIG. 1 confirmed the definite difference on the top surface between the -Al.sub.2O.sub.3 disc and the -/-Al.sub.2O.sub.3 disc.

(45) TABLE-US-00001 TABLE 1 Support Pore size (nm) Porosity (%) -Al.sub.2O.sub.3 disc 151.1 40.3 -/-Al.sub.2O.sub.3 disc 154.7 40.7

(46) FIG. 4 shows the morphology of membranes ZIF-8_ and ZIF-8_. It appears that membrane ZIF-8_ was composed of well inter-grown ZIF-8 grains with a size of about 200 to 300 nm, whereas some larger ZIF-8 particles, indicated by red arrows in the figure, were widespread throughout the surface (FIG. 4(a1)). The degree of particle aggregation was comparable to that observed in the ZIF-8 membrane prepared by the CD method disclosed in a literature (H. T. Kwon, H. K. Jeong, J. Am. Chem. Soc., 2013, 135, 10763-10768). The total thickness of membrane ZIF-8_ was estimated to be 5 m including about 2 m thick ZIF-8 phases formed below the surface (FIG. 4(a2)). This indicates that the reaction zone where the two reactants meet and further react was likely to originate below the surface of the -Al.sub.2O.sub.3 disc (H. T. Kwon, H. K. Jeong, J. Am. Chem. Soc., 2013, 135, 10763-10768). In addition, the EDX result (FIG. 4(a3)), which is effective for designating the distribution of Zn atoms, also indicates that ZIF-8 membranes were primarily fabricated on the top of the -Al.sub.2O.sub.3 disc, whereas some ZIF-8 grains were present inside the -Al.sub.2O.sub.3 disc. Finally, the contact angle measurement of a water droplet on membrane ZIF-8_ supported the formation of hydrophobic ZIF-8 phases on the -Al.sub.2O.sub.3 disc, as the corresponding contact angle of about 71 was comparable to the reported value (about 60) of the self-supported ZIF-8 pellets (X. Liu et al., Chem. Commun., 2013, 49, 9140-9142). In contrast to membrane ZIF-8_, at a first glance, the SEM images in FIG. 4(b1)-4(b2) suggest that membrane ZIF-8_ failed to grow after applying the CD method. However, a comparison of the top view SEM images in FIGS. 2(b1) and 4(b1) reveals a definite discrepancy in the surface morphology of the -/-Al.sub.2O.sub.3 disc and membrane ZIF-8_. Tiny grains, seemingly reflected by noisy features, were generated throughout the surface after conducting the CD method. In fact, the EDX analysis on membrane ZIF-8_ (FIG. 4(b3)) supports that the ZIF-8 phase rarely existed on the surface, and instead the ZIF-8 phase was distributed widely along the film thickness of the -/-Al.sub.2O.sub.3 disc. Furthermore, the contact angle of a water droplet on membrane ZIF-8_ was much smaller, approximately 15, indicating that ZIF-8 grains were not predominantly formed above the surface of the -/-Al.sub.2O.sub.3 disc.

(47) In an effort to investigate the location of ZIF-8 grains in both membranes in a rigorous way, cross-sectional view SEM images as well as EDX results were obtained at four different positions along the membrane thickness (right above and below the surface (top), middle, and bottom) (FIG. 5 and Table 2). In membrane ZIF-8_, most grains were formed right above or below the top surface of the -Al.sub.2O.sub.3 disc (FIGS. 5(3a1) to 5(a4)), though some ZIF-8 grains that filled the pores were sporadically present, as indicated by white arrows in FIG. 5(a2). On the contrary, in membrane ZIF-8_, most ZIF-8 grains grew to fill the pores inside the -Al.sub.2O.sub.3 disc, but near the interface between the -Al.sub.2O.sub.3 layer and -Al.sub.2O.sub.3 disc, thus forming a continuous film (FIGS. 5(b1) to 5(b4)). These grains were primarily located right below the interface, though their existence in the -Al.sub.2O.sub.3 layer could not be completely ruled out due the SEM resolution limit. In fact, similar ratios (about 8.5%) of Zn to Al near the interface and in the -Al.sub.2O.sub.3 layer indicate that ZIF-8 grains were located in the pores of the -Al.sub.2O.sub.3 layer (Table 2). Finally, it was also noted that the formation of ZIF-8 grains in the middle and bottom parts of both types of supports was negligible.

(48) TABLE-US-00002 TABLE 2 Position Membrane ZIF-8_ Membrane ZIF-8_ of -Al.sub.2O.sub.3 Weight ratio Weight ratio disc (Zn (wt %)/Al (wt %)) (Zn (wt %)/Al (wt %)) Top 7.7 4.5 8.5 1.3 (-Al.sub.2O.sub.3 layer) 8.5 1.1 Middle 0 0 Bottom 0 0

(49) The XRD patterns shown in FIG. 6 confirm the existence of the ZIF-8 phase in both membranes ZIF-8_ and ZIF-8_. In addition, these patterns indicate that both membranes were randomly oriented. At first glance, the XRD intensities suggest that the amount of ZIF-8 phases in membrane ZIF-8_ was much lower than that in membrane ZIF-8_. However, the weight gains after membrane synthesis were similar for both cases; about 11 mg for membrane ZIF-8_ and about 12 mg for membrane ZIF-8_ (Table 3). Considering the X-ray source could penetrate a finite depth into the sample, it appears that the -Al.sub.2O.sub.3 layer inhibited, and thus decreased the intensity of the XRD peak corresponding to the -Al.sub.2O.sub.3 disc. In addition, the present inventors could not rule out the possibility that the ZIF-8 grains in membrane ZIF-8_ primarily existed deeper inside the -Al.sub.2O.sub.3 disc support, as shown in the cross-sectional view SEM images (FIG. 5). Detailed analysis on the minor ZnO phase in the membranes ZIF-8_ and ZIF-8_ (indicated by red arrows in FIG. 4) was further performed.

(50) TABLE-US-00003 TABLE 3 Estimated weights of the ZIF-8 phase in membranes ZIF-8_ and ZIF-8_ obtained by weighing samples before and after membrane synthesis and by comparing micropore areas (estimated from the t-plot method). Estimated weight of Estimated ZIF-8 phase (mg) weight of obtained by ZIF-8 phase (mg) weighing samples obtained by measured before and comparing after membrane the microporous Samples synthesis surface areas ZIF-8_ 11 0.1 8.0 ZIF-8_ 12 0.1 8.4

(51) Despite the rigorous characterization via XRD, SEM, and EDX analyses, the present inventors could not determine the presence of the ZIF-8 grains in the -Al.sub.2O.sub.3 layer. In an unprecedented approach to locate the ZIF-8 grains in the -Al.sub.2O.sub.3 layer, the present inventors, for the first time, measured the N.sub.2 adsorption isotherms of the supports and supported membranes at 77 K; the bare disc supports (-Al.sub.2O.sub.3 and -/-Al.sub.2O.sub.3 discs) and the supported ZIF-8 membranes (ZIF-8_ and ZIF-8_) (FIGS. 7a and 7d). Their corresponding specific microporous surface areas (Table 4) and pore size distributions (FIGS. 5b to 5c and 7e to 7f) were further evaluated to identify and quantify the ZIF-8 grains in both ZIF-8 membranes. As a reference, the specific microporous surface area of ZIF-8 powder, synthesized in the same methanol medium, was adopted here (T. Lee et al., J. Choi, J. Phys. Chem. C, 2015, 119, 8226-8237). Specifically, the comparison between the microporous surface areas of the disc support and its supported membrane allowed for estimating the weight of ZIF-8 grains in membranes ZIF-8_ and ZIF-8_. The estimates of 8.0 mg for ZIF-8_ and 8.4 mg for ZIF-8_ were in a good agreement with the measured values of 11 mg for ZIF-8_ and 12 mg for ZIF-8_ (Table 3), which were obtained by weighing before and after membrane synthesis.

(52) TABLE-US-00004 TABLE 4 External Micropore BET surface surface area surface area area (m.sup.2/g) (m.sup.2/g) (m.sup.2/g) -Al.sub.2O.sub.3 3.88 0.008 3.84 0.036 disc -/-Al.sub.2O.sub.3 4.30 0.008 2.92 0.96 disc ZIF-8_ 10.06 0.04 2.13 7.92 ZIF-8_ 10.56 0.04 2.39 8.17 ZIF-8 1643.3 9.8 65.6 1577.7 powder.sup.1) Note.sup.1) T. Lee et al., J. Choi, J. Phys. Chem. C, 2015, 119, 8226-8237.

(53) FIGS. 7(b) and 7(c) reveal the existence of mesopores in the range of about 3 to 10 nm in the -/-Al.sub.2O.sub.3 disc, primarily due to the mesoporous -Al.sub.2O.sub.3 layer compared to the -Al.sub.2O.sub.3 disc. In particular, BJH analysis from the desorption points clearly reveals the presence of mesopores of approximately 5 nm, indicating a regular bottleneck formation in the -Al.sub.2O.sub.3 layer. In addition, FIGS. 7(e) and 7(f) show the reduced mesopores in ZIF-8_, apparently attributable to the marked mesopore filled by the ZIF-8 grains, whereas no change was observed in the mesoporous regime between the -Al.sub.2O.sub.3 disc and membrane ZIF-8_. The accumulated mesoporous volume for membrane ZIF-8_ relative to that for the -Al.sub.2O.sub.3 layer in the -/-Al.sub.2O.sub.3 disc (FIG. 5) suggests that the extent of ZIF-8 grains filling mesopores in the -Al.sub.2O.sub.3 layer was as high as about 70%. Considering the specific mesopore volume (estimated to be about 510.sup.4 cm.sup.3.Math.g.sup.1) and density (about 4 g.Math.cm.sup.3) of the -Al.sub.2O.sub.3 layer (P. Bowen et al., Powder Technol., 2005, 157, 100-107), the weight of the ZIF-8 phase present in the -Al.sub.2O.sub.3 layer was likely to be about 0.005 mg. Considering the total weight of ZIF-8 grains (about 8.4 to 12 mg; Table 3), it can be concluded that almost all ZIF-8 grains in membrane ZIF-8_ were formed in the -Al.sub.2O.sub.3 disc portion primarily right below the interface (see FIGS. 5 and 7).

(54) Given the same reaction rate for the synthesis of both membranes, the membrane characteristics from the CD method will be determined by the diffusion rates of the reactants. As the reaction between the Zn precursors and the mim ligands occurs rapidly towards ZIF-8 phases (J. Cravillon et al., Chem. Mater., 2009, 21, 1410-1412; J. Cravillon et al., CrystEngComm, 2012, 14, 492-498), the method used to effectively control how the two reactants meet is a key factor to determine final membrane properties. Although the nominal molar ratio of Zn.sup.2+ to mim in the synthetic precursor solution was , the actual ratio during membrane synthesis was estimated to be about 1/640. Therefore, the diffusion behavior of Zn.sup.2+ ions toward the mim ligands was critical for generation of ZIF-8 grains, and thus ZIF-8 films or membranes. In the -Al.sub.2O.sub.3 disc, the Zn.sup.2+ ions were likely to diffuse out rapidly and meet the mim ligands to produce ZIF-8 grains even below the surface of the -Al.sub.2O.sub.3 disc (about 2 m below) and continued to grow a continuous ZIF-8 membrane on and above the top surface. Further encounters between Zn.sup.2+ ions and mim ligands would be blocked by the already-synthesized ZIF-8 grains near the top surface of the -Al.sub.2O.sub.3 disc, resulting in the thin film thickness of membrane ZIF-8_ as shown in FIG. 4(a2). On the contrary, introduction of the -Al.sub.2O.sub.3 layer (about 5 nm pore size) led to alleviating the effective diffusion rate of Zn.sup.2+ ions towards the mim ligands in the bulk phase. Specifically, the mesopores in the -Al.sub.2O.sub.3 layer would prevent the Zn.sup.2+ ions from diffusing out promptly so that they would dwell longer, and thus react with mim ligands primarily inside the -Al.sub.2O.sub.3 disc. The small pore aperture of the -Al.sub.2O.sub.3 layer increased mass transfer resistance with respect to the Zn.sup.2+ ions, resulting in a wide distribution of ZIF-8 grains along the thickness of the -/-Al.sub.2O.sub.3 disc (FIGS. 4(b2) and 5(b2)) with the partial occupation (about 70%) in the -Al.sub.2O.sub.3 layer (FIGS. 7c and 7f).

(55) Subsequently, the H.sub.2 and CO.sub.2 separation performance of both membranes ZIF-8_ and ZIF-8_ was tested as a function of temperature (FIG. 8). Membrane ZIF-8_ demonstrated a H.sub.2 permeance as high as (3.91.2)10.sup.7 mol.Math.m.sup.2.Math.s.sup..Math.Pa.sup.1 and a maximum (max) H.sub.2/CO.sub.2 SF of 6.70.3 at 200 C., whereas membrane ZIF-8_ demonstrated a slightly reduced H.sub.2 permeance of (9.01.9)10.sup.8 mol.Math.m.sup.2.Math.s.sup.1.Math.Pa.sup.1 and max H.sub.2/CO.sub.2 SF of 8.20.5 at 200 C. Further increase in temperature reveals that the max H.sub.2/CO.sub.2 SFs were observed at 250 C. in both ZIF-8 membranes (FIGS. 8(a3) to 8(b3), though the SFs were still sufficiently high (about 8.9) at 300 C., which can be considered a WGS reaction temperature. Additionally, in an effort to confirm the location of ZIF-8 grains, and simultaneously compare the mechanical stabilities of both membranes, the membranes were intentionally damaged by grinding the top surface with sand paper and were tested for H.sub.2/CO.sub.2 permeation (FIGS. 8(a2) to 8(b2)). While the H.sub.2/CO.sub.2 SF of the damaged membrane ZIF-8_ was considerably lowered from about 6.7 to about 4.3 at 200 C., that of the damaged membrane ZIF-8_ was well preserved at approximately 8.5 at 200 C., strongly supporting that membrane ZIF-8_ was primarily synthesized more deeply in the interior of the -/-Al.sub.2O.sub.3 disc and desirably, showed higher mechanical stabilities.

(56) In general, the permeance of a permeating species across a membrane is inversely proportional to the membrane thickness. The H.sub.2 permeance through membranes ZIF-8_ and ZIF-8_ in FIGS. 8(a1) and 8(b1) indicates that membrane ZIF-8_ is much thicker than membrane ZIF-8_, which also implies a wide distribution of ZIF-8 grains in the -/-Al.sub.2O.sub.3 disc (FIG. 4(b3)). Based on the weight gain (about 11 mg for ZIF-8_ and about 12 mg for ZIF-8_; Table 1) after membrane synthesis and the similar porosities (about 41% in Table 1), the present inventors attempted to estimate the effective thickness of the ZIF-8 grains. Assuming a continuous phase in the -Al.sub.2O.sub.3 disc, the thickness of ZIF-8 grains below the surface was estimated to be 60 m in membrane ZIF-8_ (excluding the about 3 m thick ZIF-8 phase on top of the surface) and 80 m in membrane ZIF-8_. However, the ZIF-8 grains in membrane ZIF-8_ were not likely to be continuously formed inside, as shown in FIG. 5(b2). Therefore, the two membranes can be visualized; the 5 m thick (3 m above the surface and 2 m below the surface) membrane ZIF-8_ vs. the 80 m thick (below the interface) membrane ZIF-8_ (comparable to the EDX result shown in FIG. 4(b3)).

(57) The grain boundary defects are easily formed during thermal activation processes due to the difference in thermal expansion behaviors between the membrane and support (X. Dong et al., J. Mater. Chem., 2012, 22, 19222-19227). As the amount of the ZIF-8 grains were comparable in membranes ZIF-8_ and ZIF-8_, one could conclude that the ZIF-8 grains that formed on the -Al.sub.2O.sub.3 disc in membrane ZIF-8_ were more defective (FIG. 8). In contrast, during thermal activation through solvent removal, the ZIF-8 grains inside the -Al.sub.2O.sub.3 disc will be kept in place, and thus less-defective membrane ZIF-8_ could be fabricated. The rapid decrease in the transient measurement of the contact angle of a water droplet (up to 5 minutes) also supported that membrane ZIF-8_ was more defective (FIG. 9)), though it was seemingly continuous when viewed at SEM resolution (FIGS. 4(a1) and 4(a2)). The benefit of the continuous film formation inside porous supports, known as a confinement effect, contributed to acquiring higher H.sub.2/CO.sub.2 SFs through membrane ZIF-8_ (N. Hara et al., J. Membr. Sci., 2014, 450, 215-223; H. T. Kwon et al., Chem. Commun., 2013, 49, 3854-3856; H. T. Kwon et al., J. Am. Chem. Soc., 2013, 135, 10763-10768; Y. Pan et al., J. Membr. Sci., 2012, 390-391, 93-98; M. N. Shah et al., Langmuir, 2013, 29, 7896-7902).

(58) Although the permeation rate across a microporous membrane is a coupled function of adsorption and diffusion of guest molecules, the ideal permeation selectivity, estimated by a simple product of the adsorption and diffusion selectivities, serves as an effective indicator for evaluating its separation performances (R. Krishna et al., J. Membr. Sci., 2010, 360, 323-333; A. Battisti et al., Microporous Mesoporous Mater., 2011, 143, 46-53; R. De Lange et al., J. Porous Mater., 1995, 2, 141-149; R. Krishna et al., PCCP, 2011, 13, 10593-10616). In the present invention, the adsorption of CO.sub.2 is stronger than that of H.sub.2 in ZIF-8 (or ZIF) materials, whereas the smaller H.sub.2 molecule diffuses faster in ZIF-8s than the larger CO.sub.2 molecule. Thus, the product of their diffusion coefficients and Henry's constants are likely to end up with a same order of magnitude, reflecting the challenging task of H.sub.2/CO.sub.2 separation through ZIF-8 membranes. Even, the MD simulation showed a linear correlation between the diffusion coefficients of the larger CO.sub.2 molecule (A. Battisti et al., Microporous Mesoporous Mater., 2011, 143, 46-53; D. Liu et al., J. Phys. Chem. C, 2009, 113, 5004-5009) and the pore size, suggesting difficulty in separating H.sub.2/CO.sub.2 mixtures via reduction of the pore size. In this circumstance, the fact that the max H.sub.2/CO.sub.2 SF in membrane ZIF-8_ was about 20% higher relative to that in membrane ZIF-8_ indicates a notable quality of membrane ZIF-8_. As the simplified version of the permeance can be expressed by the product of the diffusivity and adsorption constant (R. Krishna et al., Chem. Eng. J., 2007, 133, 121-131), the apparent activation energies were estimated from the permeances, which were 2.6 kJ/mol for H.sub.2 and 6.7 kJ/mol for CO.sub.2 in membrane ZIF-8_ and 3.1 kJ/mol for H.sub.2 and 7.5 kJ/mol for CO.sub.2 in membrane ZIF-8_. Assuming that adsorption behaviors of H.sub.2 and CO.sub.2 are identical for both membranes, the difference in the activation energies of H.sub.2 between membranes ZIF-8_ and ZIF-8_ was 0.5 kJ/mol, whereas the difference for CO.sub.2 diffusion was 0.8 kJ/mol. Accordingly, the resulting activation energy difference between H.sub.2 and CO.sub.2 in membrane ZIF-8_ was 0.3 kJ/mol higher than that in membrane ZIF-8_, possibly implying that the ZIF-8 grains confined inside supports were less flexible.

(59) Recognizing the discrepancy in the permeation results in membranes ZIF-8_ and ZIF-8_, the present inventors further tested their long-term thermal stabilities in an attempt to assure their validity for the ultimate usage in the WGS membrane reactor (FIGS. 10 to 12). Over a measurement period of 2 days, the ZIF-8 particles were reported to maintain their structures up to 200 C. under ambient conditions, though the ZIF-8 membranes, prepared via the seeded growth method, were recently reported to gradually lose their H.sub.2/CO.sub.2 separation ability even at 150 C. (Y. Lin et al., 7th International Zeolite Membrane Meeting, in Dalian, China. 2016). Based on the particle experiment, three temperatures, 200, 250, and 300 C., were chosen to determine the thermal stabilities of membranes ZIF-8_ and ZIF-8_ for up to 72 hours. The present inventors found that the permeation results and structures were well preserved at 200 C. and 250 C. for up to 72 hours in both membranes ZIF-8_ and ZIF-8_ (FIGS. 10 and 11). However, at the higher temperature of 300 C., the separation performance gradually deteriorated with time to a different degree for each membrane (FIG. 12). Specifically, membrane ZIF-8_ failed to sustain the H.sub.2/CO.sub.2 separation ability almost immediately after reaching the target temperature of 300 C. (about 2 hours), whereas membrane ZIF-8_ could maintain its ability for up to about 10 hours at the same temperature.

(60) This demonstrates the desirability of forming ZIF-8 grains embedded in the porous supports (FIGS. 12a and 12b) to improve both the thermal and mechanical stabilities (FIG. 8(b2)). The XRD analysis reveals that the deteriorated separation ability over time at 300 C. can be correlated with the collapse of ZIF-8 structures and the eventual phase transition into zinc oxide (FIGS. 12a and 12d). The phase transition was supported by the onset of zinc oxides from ZIF-8 particles under similar conditions (FIG. 13). The initial thermal degradation, plausibly associated with the disconnected coordination between Zn atoms and mim molecules, is analogous to the carbonization process under inert conditions (S. Gadipelli et al., Energy Environ. Sci., 2014, 7, 2232-2238).

(61) The improvement in the thermal stability of membrane ZIF-8_ can be attributed to the wide distribution of ZIF-8 grains inside the support with the effective thickness of about 80 m as mentioned above (FIG. 4(b3)). The enlarged thickness of the ZIF-8 phase in membrane ZIF-8_ and the cohesive ZIF-8 phase filling the pores were seemingly beneficial for resisting thermal degradation toward the inevitable phase transition. This suggests that ZIF-8 grains, deeply embedded in the -Al.sub.2O.sub.3 disc, were sufficiently abundant so that a longer duration was required in membrane ZIF-8_ as structural degradation would take place sequentially toward the inside of the -Al.sub.2O.sub.3 disc. Consequently, this structural difference can account for the discrepancy in thermal stability of ZIF-8 membranes (obtained by seeded growth) (Y. Lin et al., 7th International Zeolite Membrane Meeting, in Dalian, China. 2016) and membrane ZIF-8_ and ZIF-8_ (obtained by the CD method).

(62) In FIG. 8, the present inventors compared the H.sub.2/CO.sub.2 separation performances of membranes ZIF-8_ and ZIF-8_ with those of other types of ZIF membranes reported in the literature (Y. Li et al., J. Membr. Sci., 2010, 354, 48-54; Y.-S. Li et al., Adv. Mater., 2010, 22, 3322-3326; Y.-S. Li et al., Angew. Chem. Int. Ed., 2010, 49, 548-551; V. M. Aceituno Melgar et al., J. Membr. Sci., 2014, 459, 190-196; S.-J. Noh et al., J. Nanosci. Nanotechnol., 2015, 15, 575-578; K. Huang et al., Chem. Commun., 2013, 49, 10326-10328; Q. Liu et al., J. Am. Chem. Soc., 2013, 135, 17679-17682; G. Xu et al., J. Membr. Sci., 2011, 385-386, 187-193; Y. Zhu et al., Sep. Purif. Technol., 2015, 146, 68-74; A. Huang et al., Angew. Chem. Int. Ed., 2010, 49, 4958-4961; X. Dong et al., J. Mater. Chem., 2012, 22, 19222-19227; A. Huang et al., Angew. Chem. Int. Ed., 2011, 50, 4979-4982; A. Huang et al., J. Am. Chem. Soc., 2010, 132, 15562-15564; A. Huang, N. Wang, C. Kong, J. Caro, Angew. Chem. Int. Ed., 2012, 51, 10551-10555; A. Huang et al., Chem. Commun., 2012, 48, 10981-10983; H. Bux et al., Chem. Mater., 2011, 23, 2262-2269; L. Fan et al., J. Mater. Chem., 2012, 22, 25272-25276). In particular, only the permeation results obtained with the H.sub.2 and CO.sub.2 binary mixtures were collected and plotted in FIG. 14. Based on a substantial literature survey, the present inventors recognized that many researchers reported H.sub.2/CO.sub.2 separations at low temperatures. In particular, the max temperature reported for testing the performance of ZIF-8 membranes was about 150 C., though the H.sub.2/CO.sub.2 separation ability is more useful at higher temperatures. Polymeric membranes that have a high processability and an inexpensive manufacturing/operating cost are desirable for performing the H.sub.2/CO.sub.2 separation at low temperatures. As shown in FIG. 8, one may conclude that the H.sub.2/CO.sub.2 separation performance of membrane ZIF-8_ in the present invention was modest and rather inferior to that of other membranes reported in the literature. However, the present inventors would like to emphasize that the H.sub.2/CO.sub.2 SFs through membrane ZIF-8_ were highest among those reported for ZIF-8 membranes. More importantly, the H.sub.2/CO.sub.2 SFs through membrane ZIF-8_ at temperatures of 250 and 300 C. were 10.8 and 10.1, respectively, which were more than twice of the H.sub.2/CO.sub.2 SF (about 4.7) determined in the Knudsen diffusion regime.

(63) Among the many ZIF membranes, ZIF-95 membranes showed the highest H.sub.2/CO.sub.2 separation performance with a H.sub.2/CO.sub.2 SF of about 25.7 at about 325 C. (FIG. 14). This is the highest temperature with the highest H.sub.2/CO.sub.2 separation performance for ZIF type membranes. Interestingly, although the pore size (0.37 nm) of ZIF-95 is larger than that (0.34 nm) of ZIF-8, the resulting ZIF-95 membrane showed much superior H.sub.2/CO.sub.2 performance. Ironically, this implies that the performance of membrane ZIF-8_ can be further ameliorated mainly via reducing non-ZIF defective structures. Despite the high performance of ZIF-95 membranes, a sealing issue needs to be addressed because silicone O-rings, which are only effective up to about 200 C., were used as sealants in the literature (A. Huang et al., Chem. Commun., 2012, 48, 10981-10983). In addition, the use of a very expensive ligand (5-chlorobenzimidazol) for the synthesis of ZIF-95 should be overcome for practical use. In addition to ZIF-95, ZIF-90 membranes treated with 3-aminopropyltrimethoxysilane (APTES) also demonstrated high H.sub.2/CO.sub.2 separation performance with a H.sub.2/CO.sub.2 SF of 62.5 at about 225 C., though the effects of the silicone O-rings on performance should be confirmed via a long-term permeation measurement. Furthermore, the surface modified with APTES, which causes a complicated manufacturing process, was required for the synthesis of ZIF-90 membranes. In addition, the ZIF-7 membrane with a pore size of about 0.29 nm was shown to be effective for H.sub.2/CO.sub.2 separations, though their flexible structural transition during removal of the solvent (DMF; 0.55 nm) should be addressed appropriately. Coincidentally, the synthesis of both ZIF-90 and -95 requires the use of DMF as solvent. Considering the pore sizes of ZIF-90 (0.35 nm) and 95 (0.37 nm), a careful investigation into activation processes that are directly relevant to stress generation during the removal of occluded DMFs should be conducted. In fact, it was addressed that ZIF-7 and -78 membranes suffered from defects, which were seemingly generated during DMF removal, requiring the introduction of an alternative approach. In this aspect, ZIF-8 membranes are highly desirable because smaller and more volatile solvents (such as water (0.268 nm) and methanol (0.36 nm)) than DMF can be used (Y. Pan et al., Chem. Commun., 2011, 47, 10275-10277; Y. Pan et al., J. Membr. Sci., 2012, 390-391, 93-98; H. Bux et al., J. Am. Chem. Soc., 2009, 131, 16000-16001; H. Bux et al., Chem. Mater., 2011, 23, 2262-2269). Therefore, the present inventors would like to emphasize that membrane ZIF-8_, mainly formed inside porous supports via introduction of the diffusion barrier, is highly promising for use in mass production due to its simple synthesis protocol, robust separation performance, and high chemical and mechanical stabilities.

(64) In the present invention, a simple yet reliable method to adjust the diffusion rate of ZIF-8 precursors via a pore size reduction of supports with a -Al.sub.2O.sub.3 layer (pore size: 5 nm) on top of an -Al.sub.2O.sub.3 disc (pore size: 150 nm) allowed for embedding ZIF-8 grains inside the porous support, and thus manufacturing continuous ZIF-8 membranes. Based on the direct measurement of N.sub.2 physisorption, it was found that the ZIF-8 grains in membrane ZIF-8_ was pre-dominantly formed inside the -Al.sub.2O.sub.3 disc, while filling about 70% of the mesopores in the -Al.sub.2O.sub.3 layer. The resulting membrane ZIF-8_ showed a max H.sub.2/CO.sub.2 SF of 9.80.9 at 250 C. compared to 8.01.1 for membrane ZIF-8_. More importantly, membrane ZIF-8_ showed highly improved mechanical and thermal stabilities, as evidenced by preserved separation performance after intentional surface grinding and by sustaining H.sub.2/CO.sub.2 separation performance at 300 C. for a longer duration, respectively. The high stabilities and H.sub.2/CO.sub.2 separation performance make membrane ZIF-8_ highly desirable for reliable H.sub.2/CO.sub.2 separations at high temperatures. Currently, the present inventors are attempting to manufacture ZIF-8 membranes on -Al.sub.2O.sub.3 layer-coated -Al.sub.2O.sub.3 tubes in an effort to use them to carry out WGS reactions to achieve the optimal membrane reactor configuration.

INDUSTRIAL APPLICABILITY

(65) The membrane according to the present invention is positioned inside a hierarchically structured porous support by the counter diffusion method so that the membrane can have an increased stability against physical changes caused by external impact or the like, and is not exposed directly to high-temperature gases, unlike a conventional membrane that shows low thermal stability due to direct contact with high-temperature H.sub.2 and CO.sub.2 gases, and thus can ensure higher thermal stability than the conventional membrane.

(66) Additionally, according to the present invention, the diffusion rate of Zn.sup.2+ metal ions decreases so that an increased amount of the Zn.sup.2+ metal ions can react with 2-methylimidazole inside the porous support, thereby forming a continuous membrane. A conventional membrane is mostly synthesized on the top of a support, whereas the present invention has features in that a continuous membrane can be formed while filling the inside of a porous support, and the thickness of the membrane can be controlled according to the degree of filling the pores. In addition, the membrane of the present invention is formed while filling the pores so that membrane defects that may occur incidentally in the thermal activation process can be reduced, thereby ensuring relatively high hydrogen separation ability (e.g., separation from H.sub.2/CO.sub.2 or syngas) compared to a conventional membrane.

(67) Although the present invention has been described in detail with reference to the specific features, it will be apparent to those skilled in the art that this description is only for a preferred embodiment and does not limit the scope of the present invention. Thus, the substantial scope of the present invention will be defined by the appended claims and equivalents thereof.