Modified Porous SiO2-Supported Cu3(BTC)2 Membrane with High Performance of Gas Separation

The structures and applications of metal-organic framework materials (MOFs) have been attracting great interest due to the wide variety of possible applications, for example, chemical sensing, separation, and catalysis. N-[3-(Trimethoxysilyl)propyl]ethylenediamine is grafted on a porous SiO2 disk to obtain a modified porous SiO2 disk. A large-scale, continuous, and compact Cu3(BTC)2 membrane is prepared based on a modified porous SiO2 disk. The chemical structure, surface morphology, thermal stability, mechanical stability, and gas separation performance of the obtained Cu3(BTC)2 membrane is analyzed and characterized by means of X-ray diffraction (XRD), scanning electron microscopy (SEM), thermogravimetric analysis (TGA) and a gas separation experiment. The results show that the prepared Cu3(BTC)2 membrane has an intact morphology with its crystal. It is continuous, compact, and intact, and has good thermal stability and mechanical stability. The result of the gas separation experiment shows that the Cu3(BTC)2 membrane has a good selectivity of hydrogen and can be used to recover and purify hydrogen.


Introduction
During the past few decades, the development of membrane materials has drawn research interests in both research-oriented and industrial applications [1][2][3]. The membrane is a barrier, which can selectively control some materials to pass through and make other substances not. As a result, the membrane can be used to separate mixtures [4][5][6]. Compared with traditional separation methods, the separation-based membrane has the advantage of energy saving and efficiency. Traditionally, the development and application of membranes involve documentation of the polymer membrane because it has disadvantages of short life, low working temperature, and lower chemical stability and selectivity, but researchers need to explore some new membrane materials that have stable performance and are more conducive to separation [7][8][9]. Metal-organic framework materials (MOFs) have become excellent candidates for membrane fabrication because they have diverse structures, uniform pore size, permanent porosity, and high thermal and chemical stability. The membrane separation is mainly based on the molecular-size sieve, which also can separate some materials that can react with the membrane through the method of adsorption and diffusion. The MOF membrane has developed

Pretreatment of the SiO 2 Disk
The porous SiO 2 disk (diameter = 2.0 cm) was soaked into the mixed solution of the concentrated sulfuric acid and the hydrogen peroxide with the volume ratio of 6:4 for 5 h making the surface completely oxidized. To follow, the porous SiO 2 disk was taken out and placed into the beaker of 50 mL. 30 mL of deionized water was added into the beaker. The solution in the beaker was conducted by ultrasonic treatment for 10 minutes and poured out. The washing process was repeated three times. The porous SiO 2 disk was dried at 120 • C for 2 h.

Surface Modification of the SiO 2 Disk with N-[3-(Trimethoxysilyl)propyl]ethylenediamine
N- propyl]ethylenediamine (1 mL) and alcohol (50 mL) were added to the beaker and hydrolyzed for ten minutes. The oxidized porous SiO 2 disk slice was placed horizontally at the bottom of the beaker and stirred for 24 h at 25 • C. Following the reaction, the SiO 2 disk was washed with anhydrous ethanol repeatedly to remove the N- propyl]ethylenediamine which was not functionalized. Then the functionalized porous SiO 2 disk was dried in a vacuum.

Synthetic Cu 3 (BTC) 2 Membrane with the Modified SiO 2 Disk
The MOF membranes were prepared by means of the hydrothermal method. The MOF membrane chosen was Cu 3 (BTC) 2 . Then, 0.7 g of cupric nitrate was dissolved in 19.2 mL of distilled water and solution A was obtained. Similarly, 0.336 g trimesic acid was dissolved in 19.2 mL ethanol and solution B was obtained. Solution B was then poured into solution A and stirred for 1 h. The mixed solution was poured into the Teflon-lined autoclave. The substrate of the functional porous SiO 2 disk was placed in the Teflon-lined autoclave with tweezers at 100 • C for 3 d [41]. Then, the membrane was washed several times with ethanol and dried at 25 • C. The Cu 3 (BTC) 2 crystals adhered to the membrane surface were washed away with ethanol and the dried MOF membrane based on the functional porous SiO 2 disk was prepared by air-drying. A schematic diagram of the whole synthesis process is shown in Figure 1.

Pretreatment of the SiO2 Disk
The porous SiO2 disk (diameter = 2.0 cm) was soaked into the mixed solution of the concentrated sulfuric acid and the hydrogen peroxide with the volume ratio of 6:4 for 5 h making the surface completely oxidized. To follow, the porous SiO2 disk was taken out and placed into the beaker of 50 mL. 30 mL of deionized water was added into the beaker. The solution in the beaker was conducted by ultrasonic treatment for 10 minutes and poured out. The washing process was repeated three times. The porous SiO2 disk was dried at 120 °C for 2 h.

Surface Modification of the SiO2 Disk with N-[3-(Trimethoxysilyl)propyl]ethylenediamine
N- propyl]ethylenediamine (1 mL) and alcohol (50 mL) were added to the beaker and hydrolyzed for ten minutes. The oxidized porous SiO2 disk slice was placed horizontally at the bottom of the beaker and stirred for 24 h at 25 °C. Following the reaction, the SiO2 disk was washed with anhydrous ethanol repeatedly to remove the N-[3-(Trimethoxysilyl)propyl]ethylenediamine which was not functionalized. Then the functionalized porous SiO2 disk was dried in a vacuum.

Synthetic Cu3(BTC)2 Membrane with the Modified SiO2 Disk
The MOF membranes were prepared by means of the hydrothermal method. The MOF membrane chosen was Cu3(BTC)2. Then, 0.7 g of cupric nitrate was dissolved in 19.2 mL of distilled water and solution A was obtained. Similarly, 0.336 g trimesic acid was dissolved in 19.2 mL ethanol and solution B was obtained. Solution B was then poured into solution A and stirred for 1 h. The mixed solution was poured into the Teflon-lined autoclave. The substrate of the functional porous SiO2 disk was placed in the Teflon-lined autoclave with tweezers at 100 °C for 3 d [41]. Then, the membrane was washed several times with ethanol and dried at 25 °C. The Cu3(BTC)2 crystals adhered to the membrane surface were washed away with ethanol and the dried MOF membrane based on the functional porous SiO2 disk was prepared by air-drying. A schematic diagram of the whole synthesis process is shown in Figure 1.

Characterization of the Cu3(BTC)2 Membranes
The thermogravimetric analysis (TGA) was performed using a DTG-60 thermal analyzer system (Shimadzu Corporation, Kyoto, Japan) at the heating rate of 10 °C min −1 to 900 °C in a dried air atmosphere. The air flow rate was 30 mL min −1 . Samples were loaded in a platinum pan. The FTIR spectra (KBr Sigma, Aldrich, St. Louis, MO, USA) were measured using a IRAFFINITY-1 Fourier transform infrared spectrometer (Shimadzu Corporation, Nakagyo-ku, Kyoto, Japan). Samples were

Characterization of the Cu 3 (BTC) 2 Membranes
The thermogravimetric analysis (TGA) was performed using a DTG-60 thermal analyzer system (Shimadzu Corporation, Kyoto, Japan) at the heating rate of 10 • C min −1 to 900 • C in a dried air atmosphere. The air flow rate was 30 mL min −1 . Samples were loaded in a platinum pan. The FTIR spectra (KBr Sigma, Aldrich, St. Louis, MO, USA) were measured using a IRAFFINITY-1 Fourier transform infrared spectrometer (Shimadzu Corporation, Nakagyo-ku, Kyoto, Japan). Samples were packed firmly to obtain transparent films. PXRD studies were performed using a D/MAX2550 diffractometer (Riguku Corporation, Akishima, Tokyo, Japan) using Cu-Ka radiation, 40 kV, 200 mA with a scanning rate of 0.3 • min −1 (2θ). Scanning Electron Microscopy (SEM) analysis was performed on a JSM 6700 (JEOS Corporation, Akishima, Tokyo, Japan).

Low-Pressure N 2 Sorption Measurements
Nitrogen sorption experiments were performed at 77 K up to 1 bar using a manometric sorption analyzer Autosorb iQ MP (Quantachrome Instruments, Boynton Beach, FL, USA). Prior to sorption analysis, the sample was evacuated at 150 • C for 10 h using a turbomolecular vacuum pump. Specific surface areas were calculated from nitrogen adsorption data by multipoint Brunauer-Emmett-Teller (BET) analysis. Pore size distributions were calculated from the N 2 adsorption isotherms using a quenched solid density functional theory (nitrogen on carbon, slit pore) method which gave the least fitting error.

The Gas Separation Test
Prior to gas permeation measurements, the membranes were sealed in modules and swept by using Ar (sweep gas) and detecting gas. Meanwhile, the modules were heated to 80 • C and held for 1 h, then cooled to room temperature. Regarding both single component and mixture permeation, the permeate side and the feed side pressure were both set to 1 bar. Concerning mixture permeation, both feed gas and sweep gas rates were 80 mL min −1 . A soap-film flow meter was used to measure the flux of the gas and the volume ratio of the mixture gas. This assembly is shown in Figure 2. with a scanning rate of 0.3° min −1 (2θ). Scanning Electron Microscopy (SEM) analysis was performed on a JSM 6700 (JEOS Corporation, Akishima, Tokyo, Japan).

Low-Pressure N2 Sorption Measurements
Nitrogen sorption experiments were performed at 77 K up to 1 bar using a manometric sorption analyzer Autosorb iQ MP (Quantachrome Instruments, Boynton Beach, FL, USA). Prior to sorption analysis, the sample was evacuated at 150 °C for 10 h using a turbomolecular vacuum pump. Specific surface areas were calculated from nitrogen adsorption data by multipoint Brunauer-Emmett-Teller (BET) analysis. Pore size distributions were calculated from the N2 adsorption isotherms using a quenched solid density functional theory (nitrogen on carbon, slit pore) method which gave the least fitting error.

The Gas Separation Test
Prior to gas permeation measurements, the membranes were sealed in modules and swept by using Ar (sweep gas) and detecting gas. Meanwhile, the modules were heated to 80 °C and held for 1 h, then cooled to room temperature. Regarding both single component and mixture permeation, the permeate side and the feed side pressure were both set to 1 bar. Concerning mixture permeation, both feed gas and sweep gas rates were 80 mL min −1 . A soap-film flow meter was used to measure the flux of the gas and the volume ratio of the mixture gas. This assembly is shown in Figure 2.

The FTIR of the Modified SiO2 Disk
The FTIR spectra of the produced porous SiO2 disk and the modified porous SiO2 disk, demonstrated that the N- propyl]ethylenediamine groups were grafted onto the porous SiO2 disk surface, as presented in Figure 3. Regarding the case of the porous SiO2 disk, the sharp band at 3450 cm −1 corresponded to the presence of silanol groups (Si-OH) on the silica surface. The absorption bands at 1645 cm and 1080 cm −1 were related to the bending vibration of H2O and the isolated terminal silanol (Si-OH) groups, respectively. Following modification with N-[3-(Trimethoxysilyl)propyl]ethylenediamine, the absorption of water and the Si-OH absorption peak intensity decreased, which was due to the surface of the porous SiO2 hydroxyl (-OH) and N-[3-(Trimethoxysilyl)propyl]ethylenediamine condensation reaction reducing the number. This changed the degree of bonding of the porous SiO2 surface to water, that is, the bonding density with hydrogen

The FTIR of the Modified SiO 2 Disk
The FTIR spectra of the produced porous SiO 2 disk and the modified porous SiO 2 disk, demonstrated that the N- propyl]ethylenediamine groups were grafted onto the porous SiO 2 disk surface, as presented in Figure 3. Regarding the case of the porous SiO 2 disk, the sharp band at 3450 cm −1 corresponded to the presence of silanol groups (Si-OH) on the silica surface. The absorption bands at 1645 cm and 1080 cm −1 were related to the bending vibration of H 2 O and the isolated terminal silanol (Si-OH) groups, respectively. Following modification with N- propyl]ethylenediamine, the absorption of water and the Si-OH absorption peak intensity decreased, which was due to the surface of the porous SiO 2 hydroxyl (-OH) and N- propyl]ethylenediamine condensation reaction reducing the number. This changed the degree of bonding of the porous SiO 2 surface to water, that is, the bonding density with hydrogen to produce hydrogen changes. The characteristic absorption peak after N- propyl]ethylenediamine appeared at 2980 cm −1 due to the asymmetric stretching of the C-H bond in the aminopropyl group, indicating that N- propyl]ethylenediamine had been grafted onto the porous SiO 2 surface.  Figure 4 is the XRD spectrum of the modified porous SiO2 disk-supported Cu3(BTC)2 membrane (black) and the Cu3(BTC)2 powder (red). Figure 4 shows the apex position of the XRD peak of the Cu3(BTC)2 membrane was the same as the highest position of the XRD spectrum of the Cu3(BTC)2 powder. The phenomenon illustrates that the modified porous SiO2 disk-supported Cu3(BTC)2 membrane is a pure phase composed of Cu3(BTC)2 crystal.

The TGA of the Cu3(BTC)2 Membranes
The TGA was conducted to investigate the thermal stability of the modified porous SiO2 disksupported Cu3(BTC)2 membrane. The results illustrate that the modified porous SiO2 disk-supported Cu3(BTC)2 membrane, at 63 °C, had a weight loss of 5%, which was the adsorbed water, and the Cu3(BTC)2 membrane was stable in the air to 310 °C, showing its good thermal stability. The thermogravimetric curve is shown in Figure 5.  Figure 4 is the XRD spectrum of the modified porous SiO 2 disk-supported Cu 3 (BTC) 2 membrane (black) and the Cu 3 (BTC) 2 powder (red). Figure 4 shows the apex position of the XRD peak of the Cu 3 (BTC) 2 membrane was the same as the highest position of the XRD spectrum of the Cu 3 (BTC) 2 powder. The phenomenon illustrates that the modified porous SiO 2 disk-supported Cu 3 (BTC) 2 membrane is a pure phase composed of Cu 3 (BTC) 2 crystal.  Figure 4 is the XRD spectrum of the modified porous SiO2 disk-supported Cu3(BTC)2 membrane (black) and the Cu3(BTC)2 powder (red). Figure 4 shows the apex position of the XRD peak of the Cu3(BTC)2 membrane was the same as the highest position of the XRD spectrum of the Cu3(BTC)2 powder. The phenomenon illustrates that the modified porous SiO2 disk-supported Cu3(BTC)2 membrane is a pure phase composed of Cu3(BTC)2 crystal.

The TGA of the Cu3(BTC)2 Membranes
The TGA was conducted to investigate the thermal stability of the modified porous SiO2 disksupported Cu3(BTC)2 membrane. The results illustrate that the modified porous SiO2 disk-supported Cu3(BTC)2 membrane, at 63 °C, had a weight loss of 5%, which was the adsorbed water, and the Cu3(BTC)2 membrane was stable in the air to 310 °C, showing its good thermal stability. The thermogravimetric curve is shown in Figure 5.

The TGA of the Cu 3 (BTC) 2 Membranes
The TGA was conducted to investigate the thermal stability of the modified porous SiO 2 disk-supported Cu 3 (BTC) 2 membrane. The results illustrate that the modified porous SiO 2 disk-supported Cu 3 (BTC) 2 membrane, at 63 • C, had a weight loss of 5%, which was the adsorbed water, and the Cu 3 (BTC) 2 membrane was stable in the air to 310 • C, showing its good thermal stability. The thermogravimetric curve is shown in Figure 5.

The Low-Pressure N2 Sorption Measurements and the Pore Size of the Cu3(BTC)2 Membranes
The low-pressure N2 sorption measurements and the pore size of the Cu3(BTC)2 membranes were revealed by nitrogen sorption isotherm measurement at 77 K ( Figure 6). The samples both were activated and degassed 10 h at 150 °C and measured from 0 to 1 bar (1 bar = P0). The result of the Cu3(BTC)2 membranes exhibited a type I isotherm, which is a typical feature of microporous materials. The BET surface area was evaluated, and pore diameter was consistent with those previously reported, indicating that the m as 1550 m 2 g −1 . The pore size was calculated by appropriate fitting of the density functional theory model to the isotherm yields, which was a value of 1.0 nm for the Cu3(BTC)2 membrane. BET surface area and pore diameter were consistent with those previously reported, indicating that the membrane material had the same adsorption performance as the powder material [42].

The SEM of the Cu3(BTC)2 Membranes
The characterizations of the morphology of before and after the SiO2 disk were modified and the modified porous SiO2 disk supporting the Cu3(BTC)2 membrane was conducted after defining the structural information and thermal stability of this membrane. The SEM of the modified porous SiO2 disk-supported Cu3(BTC)2 membrane is shown in Figure 7. It can be observed that there were many small holes on the surface of the SiO2 disk (Figure 7a). Results following modification of the SiO2 disk

The Low-Pressure N 2 Sorption Measurements and the Pore Size of the Cu 3 (BTC) 2 Membranes
The low-pressure N 2 sorption measurements and the pore size of the Cu 3 (BTC) 2 membranes were revealed by nitrogen sorption isotherm measurement at 77 K ( Figure 6). The samples both were activated and degassed 10 h at 150 • C and measured from 0 to 1 bar (1 bar = P 0 ). The result of the Cu 3 (BTC) 2 membranes exhibited a type I isotherm, which is a typical feature of microporous materials. The BET surface area was evaluated, and pore diameter was consistent with those previously reported, indicating that the m as 1550 m 2 g −1 . The pore size was calculated by appropriate fitting of the density functional theory model to the isotherm yields, which was a value of 1.0 nm for the Cu 3 (BTC) 2 membrane. BET surface area and pore diameter were consistent with those previously reported, indicating that the membrane material had the same adsorption performance as the powder material [42].

The Low-Pressure N2 Sorption Measurements and the Pore Size of the Cu3(BTC)2 Membranes
The low-pressure N2 sorption measurements and the pore size of the Cu3(BTC)2 membranes were revealed by nitrogen sorption isotherm measurement at 77 K ( Figure 6). The samples both were activated and degassed 10 h at 150 °C and measured from 0 to 1 bar (1 bar = P0). The result of the Cu3(BTC)2 membranes exhibited a type I isotherm, which is a typical feature of microporous materials. The BET surface area was evaluated, and pore diameter was consistent with those previously reported, indicating that the m as 1550 m 2 g −1 . The pore size was calculated by appropriate fitting of the density functional theory model to the isotherm yields, which was a value of 1.0 nm for the Cu3(BTC)2 membrane. BET surface area and pore diameter were consistent with those previously reported, indicating that the membrane material had the same adsorption performance as the powder material [42].

The SEM of the Cu3(BTC)2 Membranes
The characterizations of the morphology of before and after the SiO2 disk were modified and the modified porous SiO2 disk supporting the Cu3(BTC)2 membrane was conducted after defining the structural information and thermal stability of this membrane. The SEM of the modified porous SiO2 disk-supported Cu3(BTC)2 membrane is shown in Figure 7. It can be observed that there were many small holes on the surface of the SiO2 disk (Figure 7a). Results following modification of the SiO2 disk

The SEM of the Cu 3 (BTC) 2 Membranes
The characterizations of the morphology of before and after the SiO 2 disk were modified and the modified porous SiO 2 disk supporting the Cu 3 (BTC) 2 membrane was conducted after defining the structural information and thermal stability of this membrane. The SEM of the modified porous SiO 2 disk-supported Cu 3 (BTC) 2 membrane is shown in Figure 7. It can be observed that there were many small holes on the surface of the SiO 2 disk (Figure 7a). Results following modification of the SiO 2 disk (Figure 7b) were the same as before modification (Figure 7a). The modification of the SiO 2 disk with the organosiloxane agent did not affect its permeability. The results show that the modified SiO 2 disk had the same permeability for H 2 , CO 2 , N 2 and CH 4 , and were all 1.90 × 10 −6 mol m −2 s −1 Pa −1 . Thus, the modified base without Cu 3 (BTC) 2 membrane was not selective for H 2 , CO 2 , N 2 and CH 4 . The obtained Cu 3 (BTC) 2 membrane was composed of numerous octahedron crystals inlaid and stacked to form a uniform and dense continuous defect-free membrane structure. The scale of the microscope was 200 µm, and the membrane was continuously dense. When the scale was made gradually smaller/the magnification was gradually increased, the positive octahedral structure of the Cu 3 (BTC) 2 membrane became increasingly obvious but the Cu 3 (BTC) 2 membrane was composed of several positive octahedral crystals intercalated and stacked to form a uniformly dense continuous defect-free membrane. Cu3(BTC)2 membrane was composed of numerous octahedron crystals inlaid and stacked to form a uniform and dense continuous defect-free membrane structure. The scale of the microscope was 200 µm, and the membrane was continuously dense. When the scale was made gradually smaller/the magnification was gradually increased, the positive octahedral structure of the Cu3(BTC)2 membrane became increasingly obvious but the Cu3(BTC)2 membrane was composed of several positive octahedral crystals intercalated and stacked to form a uniformly dense continuous defect-free membrane.

The Gas Separation Test of Cu3(BTC)2 Membrane
The permeability of the modified porous SiO2 disk-supported Cu3(BTC)2 membrane to four components of H2, N2, CO2 and CH4 was evaluated. The dynamic diameter of the four gas molecules of H2, CO2, N2 and CH4 and the specific results of the membrane permeability to these four gases are listed in Table 1. The separation experiments of the H2/CO2, H2/CH4 and H2/N2 mixed gases by Cu3(BTC)2 membrane were conducted to investigate the two-component gas separation of the modified porous SiO2 disk-supported Cu3(BTC)2 membrane. Separation test results of various mixed gases at 298 K

The Gas Separation Test of Cu 3 (BTC) 2 Membrane
The permeability of the modified porous SiO 2 disk-supported Cu 3 (BTC) 2 membrane to four components of H 2 , N 2 , CO 2 and CH 4 was evaluated. The dynamic diameter of the four gas molecules of H 2 , CO 2 , N 2 and CH 4 and the specific results of the membrane permeability to these four gases are listed in Table 1. The separation experiments of the H 2 /CO 2 , H 2 /CH 4 and H 2 /N 2 mixed gases by Cu 3 (BTC) 2 membrane were conducted to investigate the two-component gas separation of the modified porous SiO 2 disk-supported Cu 3 (BTC) 2 membrane. Separation test results of various mixed gases at 298 K and 0.1 MPa are shown in Table 2. These are test results of single component permeable flow and test results of two-component permeable flow. The separation factor and the ideal separation factor calculated according to the results are also shown in Table 2.

Mechanical Stability of Cu 3 (BTC) 2 Membrane
To study the mechanical properties of the synthesized Cu 3 (BTC) 2 membrane, the gas separation performance of H 2 /CO 2 (red), H 2 /N 2 (black), H 2 /CH 4 (blue) with the Cu 3 (BTC) 2 membrane was tested repeatedly under 298 K and 0.1 MPa. Among them, red is powder, black is membrane, and blue is methane. The results show that the separation factor of the Cu 3 (BTC) 2 membrane, the synthesized membrane reproducibility, was not obviously changed after 24 h of repeated tests. The mechanical properties were strong, and the utilization rate was high ( Figure 8).

Mechanical Stability of Cu3(BTC)2 Membrane
To study the mechanical properties of the synthesized Cu3(BTC)2 membrane, the gas separation performance of H2/CO2 (red), H2/N2 (black), H2/CH4 (blue) with the Cu3(BTC)2 membrane was tested repeatedly under 298 K and 0.1 MPa. Among them, red is powder, black is membrane, and blue is methane. The results show that the separation factor of the Cu3(BTC)2 membrane, the synthesized membrane reproducibility, was not obviously changed after 24 h of repeated tests. The mechanical properties were strong, and the utilization rate was high ( Figure 8).

Preparation of the MOF Membrane
Here, we report a convenient and universal method to prepare MOF membranes by

Preparation of the MOF Membrane
Here, we report a convenient and universal method to prepare MOF membranes by hydrothermal method. First, the porous SiO 2 disk is soaked in the mixed solution of the concentrated sulfuric acid and the hydrogen peroxide with a volume ratio of 6:4 for 5 h, which gets hydroxyl on the surface of the completely oxidized porous SiO 2 disk. Then the oxidized porous SiO 2 disk was modified by N- propyl]ethylenediamine. The FTIR spectra of the modified porous SiO 2 disk demonstrated that many amino groups existed on the surface of the modified porous SiO 2 disk, which could be used to grow MOF membrane. The MOF membrane used was Cu 3 (BTC) 2 . The stable 3D structure of the Cu 3 (BTC) 2 was formed by these secondary structural units interlaced with each other, and the 3D structure had a square aperture with a regular aperture of about 1 nm. The results of N 2 adsorption show that the specific surface area of BET was about 1550 m 2 /g [42]. The XRD illustrates that the modified porous SiO 2 disk-supported Cu 3 (BTC) 2 membrane was a pure phase composed of Cu 3 (BTC) 2 crystals (Figure 2).

The Morphology and the Stability of the Cu 3 (BTC) 2 Membrane
The TGA was conducted to investigate the thermal stability of the modified porous SiO 2 disk-supported Cu 3 (BTC) 2 membrane. The results show that the modified porous SiO 2 disk-supported Cu 3 (BTC) 2 membrane was at 63 • C with a weight loss of 5% (the adsorbed water), and the Cu 3 (BTC) 2 membrane was stable in the air to 300 • C, showing its good thermal stability. The thermogravimetric curve is shown in Figure 5. The characterizations of the morphology of the SiO 2 disk and the modified porous SiO 2 disk-supported Cu 3 (BTC) 2 membrane was conducted after defining the structural information and thermal stability of this membrane. The SEM of the modified porous SiO 2 disk-supported Cu 3 (BTC) 2 membrane is shown in Figure 7. It can be observed that the modified porous SiO 2 disk-supported Cu 3 (BTC) 2 membrane was a thin, compact, and continuous membrane, closely attached to the modified SiO 2 substrate. Viewed through the scanning electron microscope, the modified porous SiO 2 disk-supported Cu 3 (BTC) 2 membrane also showed that the intergrowth-crystallized octahedral architectures merged tightly. To study the mechanical properties of the synthesized Cu 3 (BTC) 2 membrane, the gas separation performance of the Cu 3 (BTC) 2 membrane was tested repeatedly under 298 K and 0.1 MPa. The results show that the separation factor of the Cu 3 (BTC) 2 membrane, the synthesized membrane reproducibility, was not obviously changed after 24 h of repeated testing. The mechanical properties were strong, and the utilization rate was high ( Figure 8).

The Gas Separation Performance of Cu 3 (BTC) 2 Membrane
The dynamic diameter of the four gas molecules of H 2 , N 2 , CO 2 and CH 4 and the specific results of the membrane permeability to these four gases are listed in Table 1. It can be observed that the order of the flow of the four gases through the modified porous SiO 2 disk-supported Cu 3 (BTC) 2 membrane was H 2 > N 2 > CH 4 > CO 2 , and the flow rate was 1.61 × 10 −7 mol m −2 s −1 Pa −1 , 1.84 × 10 −8 mol m −2 s −1 Pa −1 , 1.98 × 10 −8 mol m −2 s −1 Pa −1 and 1.69 × 10 −8 mol m −2 s −1 Pa −1 , respectively. Based on the ideal separation coefficient formula a = J A /J B , the ideal separation factor of H 2 /CO 2 , H 2 /N 2 , H 2 /CH 4 were 9.53, 8.75 and 8.13, respectively. It was higher than the corresponding Knudsen diffusion coefficient (4.69 H 2 /CO 2 , 3.74 H 2 /N 2 , and 2.83 H 2 /CH 4 ) and was also much larger than the ideal separation factor that was reported to separate the same gas through the Cu 3 (BTC) 2 membrane [42]. This preliminarily determines that the Cu 3 (BTC) 2 membrane synthesized in this study are suitable for the separation of H 2 in the mixed components of H 2 /CO 2 , H 2 /N 2 , H 2 /CH 4 . Since the pore size of the Cu 3 (BTC) 2 membrane was about 1 nm, which was bigger than the dynamic diameters of H 2 , CO 2 , N 2 and CH 4 molecules (Table 1). Thus, there was no effect on H 2 by the molecular sieve points from the other gases. The Cu 3 (BTC) 2 structure contained many Cu elements for the adsorption of CO 2 , N 2 and CH 4 gases providing the active site and the Cu 3 (BTC) 2 membrane by chemical adsorption to gas diffusion the effect of separation, as far as the authors are aware. According to reports in the literature, Cu 3 (BTC) 2 for CO 2 , N 2 and CH 4 adsorption enthalpy is far greater than H 2 , and the preparation of the Cu 3 (BTC) 2 membrane nitrogen adsorption performance showed the same as the previous preparation of fission material adsorption performance. Thus, in this study, some adsorption properties of the MOF powder could be on behalf of the MOF-related adsorption properties of the membrane [43][44][45][46]. Therefore, the separation of two-component gas is investigated. The separation experiments of the H 2 /CO 2 , H 2 /N 2 , H 2 /CH 4 mixed gases by Cu 3 (BTC) 2 membrane was conducted to investigate the two-component gas separation of the modified porous SiO 2 disk-supported Cu 3 (BTC) 2 membrane. Separation test results of various mixed gases at 298 K and 0.1 MPa are listed in Table 2. There are test results of single component permeable flow, and two-component permeable flow. The separation factor and the ideal separation factor calculated according to the results are also listed in Table 2. Looking at Table 2, it can be observed that the flow of H 2 in the mixed gas is 1.61 × 10 −7 mol m −2 s −1 Pa −1 , many times higher than the flow of other gases. This phenomenon illustrates that the modified porous SiO 2 disk-supported Cu 3 (BTC) 2 membrane has the function of separating and purifying H 2 and can be used to separate and purify H 2 in the mixture of H 2 /CO 2 , H 2 /N 2 , H 2 /CH 4 . The flow of H 2 in the two-component gas is also much higher than that of the other components, it can be obtained through the calculation of the separation factors of the modified porous SiO 2 disk-supported Cu 3 (BTC) 2 membrane for the H 2 /CO 2 , H 2 /N 2 , H 2 /CH 4 components are 10.07, 10.20 and 11.34 in the condition of 298 K and 0.1 MPa. These values are higher than the corresponding Knudsen values. This phenomenon illustrates that the modified porous SiO 2 disk-supported Cu 3 (BTC) 2 membrane can be used in gas separation and has a good performance of gas separation.

Conclusions
This work is the first report of the synthesis of modified porous SiO 2 disk-supported Cu 3 (BTC) 2 membranes. The obtained functioned porous SiO 2 disk-supported Cu 3 (BTC) 2 membranes have crystal phases that coincide with Cu 3 (BTC) 2 crystals with a high thermal stability and intact morphology. Additionally, the performances of the modified porous SiO 2 disk-supported Cu 3 (BTC) 2 membrane for the separation of hydrogen and other gases were evaluated and the separation factor of each group of experience was calculated in detail. It was found that the membrane has a good separation performance for hydrogen and can be used in hydrogen recovery in industry. This Cu 3 (BTC) 2 membrane fabrication method is simple and convenient and can be readily applied to a variety of other material compositions to produce functional membranes with diverse micropore structures, thus opening a host of opportunities for the development of new functional MOFs.