1 Development of Innovative Gas Separation Membranes through Sub- Nanoscale Material Control Investigators Research Institute of Innovative Technology for the Earth (RITE) Yuichi Fujioka Group Leader, Chief Researcher (Dr. Eng.), Chemical Research Group Koichi Yamada Vice General Director (Dr. Eng.), Chemical Research Group Shingo Kazama Senior Researcher (Dr. Eng.), Chemical Research Group Katsunori Yogo Senior Researcher (Ph.D.), Chemical Research Group Teruhiko Kai Researcher (Ph.D.), Chemical Research Group Shigetoshi Matsui Researcher (Ph.D), Chemical Research Group Takayuki Kouketsu Researcher (Ph.D), Chemical Research Group Naoki Yamamoto Researcher (Ph.D), Chemical Research Group Yuzuru Sakamoto Researcher (Ph.D), Chemical Research Group Kousuke Uoe Researcher (Ph.D.), Chemical Research Group Manabu Miyamoto Researcher (Ph.D.), Chemical Research Group Abstract New membrane types with well-controlled sub-nanostructures were developed from both polymeric and inorganic materials: carbon-based membranes, functionalized mesoporous oxide membranes and molecular sieving zeolite membranes. Regarding the carbon membranes, a novel carbon membrane with enhanced CO 2 affinity was developed. The prepared membrane had a higher separation performance than that of the original carbon membrane in the separation of CO 2 /N 2 mixtures under humid conditions. In addition, a method to control the sizes of subnano pores of the carbon membrane was developed. Regarding the inorganic membranes, various kinds of high silica zeolites were prepared. It was found that zeolites with a high pore volume had high CO 2 adsorption capacity. We proposed a novel method for the synthesis of the composite layer of zeolite and the porous substrate; i.e., a melt-filling method. It was revealed that CO 2 separation selectivity increased after the treatment. From the XRD measurement, it was revealed that the composite layer of zeolite and the porous substrate was successfully prepared without experiencing phase transition or byproduct formation after the treatment. In addition, we found the functionalization of the pore walls of mesoporous silicas by compounds with high amine density is effective in the separation of CO 2 . Introduction Carbon dioxide capture and storage (CCS) could allow the utilization of abundant fossil fuel reserves, while significantly decreasing emissions of CO 2 to the atmosphere. However, the cost of CCS, especially of CO 2 capture, is still too expensive for CCS to be considered as a cost effective technique. This project intends to develop a variety of efficient, low-cost polymeric and inorganic CO 2 separation membranes as a game changing technology. Material structure engineering on the scale of gas molecules will be used to increase the permeability and selectivity of the membrane. Regarding the polymeric based materials, a cardo polymer-based carbon membrane incorporating
Transcript
1
Development of Innovative Gas Separation Membranes through Sub-Nanoscale Material Control
Investigators Research Institute of Innovative Technology for the Earth (RITE) Yuichi Fujioka Group Leader, Chief Researcher (Dr. Eng.), Chemical Research
Group Koichi Yamada Vice General Director (Dr. Eng.), Chemical Research Group Shingo Kazama Senior Researcher (Dr. Eng.), Chemical Research Group Katsunori Yogo Senior Researcher (Ph.D.), Chemical Research Group Teruhiko Kai Researcher (Ph.D.), Chemical Research Group Shigetoshi Matsui Researcher (Ph.D), Chemical Research Group Takayuki Kouketsu Researcher (Ph.D), Chemical Research Group Naoki Yamamoto Researcher (Ph.D), Chemical Research Group Yuzuru Sakamoto Researcher (Ph.D), Chemical Research Group Kousuke Uoe Researcher (Ph.D.), Chemical Research Group Manabu Miyamoto Researcher (Ph.D.), Chemical Research Group Abstract
New membrane types with well-controlled sub-nanostructures were developed from both polymeric and inorganic materials: carbon-based membranes, functionalized mesoporous oxide membranes and molecular sieving zeolite membranes. Regarding the carbon membranes, a novel carbon membrane with enhanced CO2 affinity was developed. The prepared membrane had a higher separation performance than that of the original carbon membrane in the separation of CO2/N2 mixtures under humid conditions. In addition, a method to control the sizes of subnano pores of the carbon membrane was developed. Regarding the inorganic membranes, various kinds of high silica zeolites were prepared. It was found that zeolites with a high pore volume had high CO2 adsorption capacity. We proposed a novel method for the synthesis of the composite layer of zeolite and the porous substrate; i.e., a melt-filling method. It was revealed that CO2 separation selectivity increased after the treatment. From the XRD measurement, it was revealed that the composite layer of zeolite and the porous substrate was successfully prepared without experiencing phase transition or byproduct formation after the treatment. In addition, we found the functionalization of the pore walls of mesoporous silicas by compounds with high amine density is effective in the separation of CO2.
Introduction Carbon dioxide capture and storage (CCS) could allow the utilization of abundant
fossil fuel reserves, while significantly decreasing emissions of CO2 to the atmosphere. However, the cost of CCS, especially of CO2 capture, is still too expensive for CCS to be considered as a cost effective technique. This project intends to develop a variety of efficient, low-cost polymeric and inorganic CO2 separation membranes as a game changing technology. Material structure engineering on the scale of gas molecules will be used to increase the permeability and selectivity of the membrane. Regarding the polymeric based materials, a cardo polymer-based carbon membrane incorporating
2
molecules or atoms having strong CO2 affinity was prepared. CO2 affinity materials were optimized and the preparation method was modified to improve the separation performance further. Regarding the inorganic materials, zeolite / functionalized mesoporous membranes were prepared. The preparation of ultra-thin, defect-free membranes was investigated.
Background Membrane separation of CO2 from other gases is an active field of study, but the best
membrane today is still considered too energy intensive and expensive to be implemented on a large scale. Gas separation in membranes is driven by a pressure difference across the membrane. To obtain a sufficiently pure stream of CO2, the selectivity for CO2 must be high. In addition, high permeability is required to produce a compact membrane facility. Many current systems require a large membrane area and cascading for the gas to permeate through multiple membrane stages to achieve the desired flow rate and purity. As such, new membrane types are required to have high permeability and selectivity, as well as long-term durability. Two areas of current gas separation membrane research are the development of polymeric and inorganic membranes. Polymeric membranes are relatively easy to manufacture and are well-suited for low temperature applications. The polymer morphology and mobility determine the gas permeability and selectivity. In addition, by carbonizing these polymeric materials it is possible to obtain a molecular sieve capability. Inorganic membranes on the other hand, have much greater thermal and chemical stability. Inorganic materials including zeolites and silicas have appropriately-sized pores that can act as molecular sieves to separate gas molecules by effective size. Surface adsorption and diffusion inside the pores can also play a role in separating gas molecules. Since the effective sizes of CO2, N2, H2, and other gases present in fossil fuel conversion systems are very similar, the membrane pore spaces must be controlled on a scale comparable to the size differences among these gas molecules. This will be achieved for a variety of membrane types using several different techniques. In this paper, we describe the development of new membrane types with well-controlled sub-nanostructures prepared from both polymeric and inorganic materials: carbon-based membranes, functionalized mesoporous oxide membranes and molecular sieving zeolite membranes.
Results (1) Carbon membrane
(1.1) Concept of the membrane structure
In the concept of molecular gate separation, originally proposed by Sirkar et al. [1], CO2 absorbed in organic materials (e.g., the poly(amidoamine) (PAMAM) dendrimer) blocks the permeation of other permanent gases such as N2 and H2; thus high separation performance is obtained. So far, we have developed new dendrimer materials that have a higher separation performance than the conventional PAMAM dendrimer has [2]. In addition, we have successfully developed dendrimer composite membrane modules using this PAMAM dendrimer. These PAMAM dendrimer composite membrane modules have high CO2 separation performance under realistic operating conditions (i.e., a pressure difference between the feed and permeate sides) [3-5]. However, the separation
3
performances of current dendrimer membranes were strongly dependent on the relative humidity (water uptake in the dendrimer). To control water uptake, excess water uptake should be suppressed by the porous matrix around the dendrimer.
In this project, the concept of molecular gates is applied to subnano/nano porous materials. The concept of the membrane structure for high separation performance is shown in Figure 1.
Figure 1. Concept of the membrane structure for high separation performance.
The pore surface of a subnano/nano porous material is modified with CO2 affinity materials to produce ideal CO2 molecular gates; i.e., high separation performance regardless of relative humidity. CO2 affinity materials can be blended with precursor, or can be inserted into the pores of the carbon membrane. The CO2 adsorbed in the porous material blocks the permeation of other permanent gases.
The design of the novel CO2 affinity-enhanced carbon membrane is shown in Figure 2.
Figure 2. Design of the novel CO2 affinity-enhanced carbon membrane
Porous matrix
Feed side
Membrane
Permeate side
CO2 affinity material
CO2H2O H2 N2
A. Pore size control
- Strong interaction between CO2 affinity materials, CO2 and water - High CO2 diffusion coefficient
Optimize pore size to ca. 0.5-5 nm- Precursor- Carbonization conditions
B. CO2 affinity control on the pore surface
Strong affinity to CO2 and water for the condensed permeation of CO2
- Treatment with alkali metal carbonate, amines etc.- Develop methods for CO2 affinity enhancement
4
To obtain ideal molecular gate separation, there are two important parameters to control: (a) pore size and (b) CO2 affinity on the pore surface. The pore size should be small enough to obtain a strong interaction between the CO2 affinity materials and CO2 even at the center of a pore.
At the same time, if CO2 affinity materials are to be inserted into the pores, the pore size should be sufficiently large. From these considerations, the optimum pore size is from 0.5 to 5 nm. To control the CO2 affinity on the pore surface, a method to incorporate CO2 affinity materials into the pores should be developed. In addition, the selection of CO2 affinity materials for high CO2 separation performance is also important.
(1.2) Membrane preparation
We examined two methods for incorporating CO2 affinity material into the pores, as shown in Figure 3.
Figure 3. Methods to incorporate CO2 affinity materials into the pores.
In method A, CO2 affinity materials were blended with precursor solution. After drying, CO2 affinity materials were dispersed in the precursor and hence dispersed also in the carbon matrix. The CO2 affinity materials must withstand high temperatures (ca. 600 °C), so heat-resistant materials were used in this method. In method B, CO2 affinity materials were incorporated into the carbon membrane by post-treatment. This method allows the use of non- heat-resistant CO2 affinity materials in addition to alkali metal carbonate. However, CO2 affinity materials might be incorporated only near the surface. A type of bis(phenyl)fluorene-based cardo polyimide, PI-BTCOOMe, was chosen as the precursor for preparing subnano/nano porous carbon membranes. The chemical structure of PI-BTCOOMe is shown in Figure 4.
Method A:
Blend CO2 affinity materials in precursor
Method B:
Post-treatment of CO2affinity materials after preparation of carbon membranes Carbon
CO2 affinity materials
Precursor
5
Figure 4. Chemical structure of PI-BTCOOMe.
A tubular-type porous alumina membrane (pore diameter of 150 nm) was used as the porous support. The precursor solution was coated on the outer surface of the alumina support by the dip-coating method. After drying, the precursor-coated membrane was carbonized under a N2 atmosphere.
Figure 5 is a photograph of a precursor-coated membrane and a carbon membrane. As seen in the figure, a yellow precursor layer was formed on the alumina support by dip-coating. Pyrolysis of the precursor layer resulted in the formation of the black carbon layer. The thickness of the carbon membrane was around 3 μm.
Figure 5. Photograph of a precursor-coated membrane and pyrolyzed carbon membrane. (1.2.1) Subnano porous carbon membranes with enhanced CO2 affinity formed using
method A Energy dispersive X-ray (EDX) spectra of the surfaces of the carbon membranes are
shown in Figure 6.
Precursor-coated membrane Carbon membrane
Substrate: porous alumina
NC
CO
C
O
O
CN
CO
O
nCOOCH3
6
Figure 6. EDX spectra of the surfaces of the carbon membranes (method A). The values in the parentheses are weight percentages of Cs2CO3 in the precursor.
The Cs peak is at 4.3 keV. No Cs peak was observed for the carbon membrane without
Cs2CO3 incorporation. On the other hand, there was a Cs peak for the carbon membrane prepared from Cs2CO3-containing precursor solution. The Cs peak became more intense as the Cs2CO3 content in the precursor solution increased. EDX analysis suggested the amount of incorporated Cs2CO3 could be controlled by the Cs2CO3 concentration in the precursor solution. The effect of the Cs2CO3 concentration in the precursor on the separation performance was examined. The results are shown in Figure 7.
Figure 7. Effect of the Cs2CO3 concentration in the precursor on the separation performance. Gas separation conditions: temperature of 40 °C, feed gas composition of CO2/N2 (5/95 vol/vol), relative humidity of ca. 100%, feed pressure of 0.1 MPa, permeate pressure of 0.1 MPa (He sweep method).
10-12
10-11
10-10
10-9
10-8
0
10
20
30
40
50
0 1 2 3 4 5
Q [m
3 (STP
) m-2
s-1
Pa-1
]
α [-]
Cs2CO
3 conc. in precursor [wt%]
α
QCO2
QN2
Cs2CO3(6.2)
Cs2CO3(0)
Cs2CO3(12)
7
As shown in Figure 7, the CO2 permeance, QCO2, and separation factor, α, with the
addition of Cs2CO3 in the precursor were higher than the values for the original carbon membrane . The separation performance under the humid condition (relative humidity: ca. 100%) was significantly improved by the addition of Cs2CO3 in the precursor. However, the separation performance became worse when the Cs2CO3 concentration in the precursor reached 4.1 wt%. Carbon membranes prepared from precursor with 6.2 and 12 wt% Cs2CO3 did not have separation ability, probably owing to the formation of pinholes (membrane defects). From these results, it is concluded the optimum Cs2CO3 concentration in precursor is less than ca. 4 wt%. (1.2.2) Subnano porous carbon membranes with enhanced CO2 affinity formed using
method B A carbon membrane was treated with Cs2CO3 using method B. The cross-sectional
distribution of carbon and cesium from the surface in a Cs2CO3–incorporated carbon membrane determined by an X-ray photoelectron spectroscopy (XPS) depth profile measurement is shown in Figure 8.
Figure 8. Cross-sectional distribution of carbon (C) and cesium (Cs) in a Cs2CO3–incorporated carbon membrane determined by an XPS depth profile measurement. The pretreatment condition was dipping in saturated Cs2CO3 aqueous solution while evacuating the permeate side at room temperature for 1 hour.
The intensity of the Cs peak was a maximum at the surface and decreased monotonically with depth from the surface, reaching the background noise level at around 300 nm. Considering the thickness of carbon membrane was around 3 μm, it is suggested that Cs2CO3 could only be incorporated into the pores of the carbon membrane near the surface using method B.
0
1
2
3
4
5
95
96
97
98
99
100
0 500 1000
Cs
[ato
mic
%]
C [a
tom
ic %
]
Depth from the surface [nm]
Cs
C
8
The relationships between the relative humidity in the feed gas and permeance, Q, and separation factor, αCO2/N2, for (a) the original carbon membrane and (b) the CO2 affinity-enhanced carbon membrane prepared using method B are shown in Figure 9.
Figure 9. Effect of humidity on the separation performance. (a) Original carbon membranes, (b) CO2 affinity-enhanced carbon membrane (method B, Cs2CO3+DAPA). Gas separation conditions: temperature of 40 °C, feed gas composition of CO2/N2 (5/95 vol/vol), relative humidity of ca. 100%, feed pressure of 0.1 MPa, permeate pressure of 0.1 MPa (He sweep method).
In the case of the original carbon membrane (Figure 9(a)), the CO2 permeance and separation factor were 9.3 × 10-11 m3(STP)m-2s-1Pa-1 and 21, respectively, when the dry feed gas was used (i.e., relative humidity of 0%). When the feed gas mixture was humidified (20RH%), the CO2 permeance and separation factor dropped to 2.3 × 10-11 m3(STP)m-2s-1Pa-1 and 10, respectively. The CO2 permeance and separation factor did not change significantly when the relative humidity in the feed gas was increased from 20 to 100%, having values of 2.4 × 10-11 m3(STP)m-2s-1Pa-1 and 11 at 100RH%, respectively. Therefore, the separation performance of the original carbon membrane became worse under the humidified conditions. On the other hand, in the case of the CO2 affinity-enhanced carbon membrane (Figure 9(b)), the CO2 permeance and separation factor were 1.5 × 10-12 m3(STP)m-2s-1Pa-1 and 6.7, respectively, when the dry feed gas was used (i.e., relative humidity of 0%). Both CO2 permeance and N2 permeance increased as the relative humidity in the feed gas increased and reached constant values for 60-100 RH%. Since CO2 permeance increased much more than N2 permeance did, the separation factor increased as the relative humidity in the feed gas increased and reached a constant value for 60-100 RH%. The CO2 permeance and separation factor were 1.8 × 10-10 m3(STP)m-
2s-1Pa-1 and 51, respectively. As designed, both the CO2 permeance and separation factor of the CO2 affinity-enhanced carbon membrane were higher than values for the original carbon membrane under humidified conditions.
10-13
10-12
10-11
10-10
10-9
10-8
0
10
20
30
40
50
60
0 20 40 60 80 100 120
Q [m
3 (STP
) m-2
s-1
Pa-1
]
αC
O2/
N2 [-
]
Relative humidity in feed [RH%]
α
QCO2
QN2
10-13
10-12
10-11
10-10
10-9
10-8
0
10
20
30
40
50
60
0 20 40 60 80 100 120
Q [m
3 (STP
) m-2
s-1
Pa-1
]
αC
O2/
N2 [-
]
Relative humidity in feed [RH%]
α
QCO2
QN2
9
(1.2.3) Control of subnano/nano pores in carbon membranes We developed a method to control the subnano/nano pore size in carbon membranes, by incorporating the controlled amount of Cs2CO3 in the precursor. The relationship between the Cs2CO3 concentration in the precursor and the Cs concentration in the carbon of carbon free-standing films and carbon-coated membranes is shown in Figure 10.
Figure 10. Relationship between the Cs2CO3 concentration in the precursor and the Cs concentration in the carbon of the carbon free-standing films and carbon-coated membranes.
The relationship between the Cs2CO3 concentration in the precursor and the Cs concentration in the carbon was similar for carbon free-standing films and carbon-coated membranes, and the Cs concentration in the carbon was proportional to the Cs2CO3 concentration in the precursor. The Cs concentration in the carbon without Cs2CO3 (i.e. Cs2CO3(0) film) was below the detection limits of AAS and ICP-MS. Considering the weight loss of the precursor by carbonization, it was suggested that most of the Cs blended in the precursor remained in the carbon after the carbonization. From this result, it was suggested that Cs can be incorporated into the carbon by blending Cs2CO3 with precursor solutions, and that the concentration of Cs in the carbon can be controlled by the concentration of Cs2CO3 in the precursor.
EPMA line-mappings of Cs over the cross-sections of carbon free-standing films are shown in Figure 11.
0
1
2
3
4
5
6
0 1 2 3 4 5
Carbon free-standing filmsCarbon-coated membranes
Cs
conc
entra
tion
in c
arbo
n [w
t%]
Cs2CO
3 concentration in precursor [wt%]
10
Figure 11. EPMA line-mapping of Cs over the cross-section of carbon free-standing films.
Uniform distributions of Cs over the cross-sections were observed for Cs2CO3(0.71), Cs2CO3(1.4) and Cs2CO3(4.1) films. The relative intensities of Cs in Cs2CO3(0.71), Cs2CO3(1.4), and Cs2CO3(4.1) films were ca. 35, 60, and 180, respectively, and it increased as the Cs2CO3 concentration in the precursor increased.
The results of water vapor sorption experiments using carbon free-standing films at 40 °C are shown in Figure 12.
Figure 12. Relationship between relative humidity and sorbed amount of water, using carbon free-standing films at 40 °C.
The original carbon film (i.e. Cs2CO3(0)) had a sigmoid sorption isotherm. This is the
typical water sorption isotherm of a carbon and such behavior is believed to be due to the combination of weak carbon-water dispersive attractions and strong water-water associative interactions. On the other hand, the shape of sorption isotherm changed as Cs2CO3 concentration in precursor solution increased. From these results, it was suggested that the pore surface properties of the carbon became more hydrophilic and the total pore volume of carbon decreased to some extent by Cs incorporation into carbon.
The results of through-pore size distribution measurement by nanopermporometry using the carbon-coated membranes are shown in Figure 13.
Figure 13. Through-pore size distribution of carbon-coated membranes by nanopermporometry.
The principle of nanopermporometry is as follows. He gas permeance is measured under a constant operating pressure, while the vapor pressure of a condensable gas (H2O) is varied in the He carrier gas. He permeance decreases as the H2O vapor pressure increases, because the H2O condenses and plugs some of the membrane pores by capillary condensation. Since the relationship between the pore diameter and H2O vapor pressure when capillary condensation occurs is described by the Kelvin equation, the pore diameter distribution can be evaluated. In the case of original carbon membranes, the pore diameter of the pores contributing to the He permeation is limited to the sub-nano order. Therefore, relatively large dendrimers (molecular size > 1 nm) cannot be inserted into the original carbon membrane. On the other hand, the new carbon membranes have nanopores ranging 1-5 nm in size, in addition to the subnano pores, and the dendrimer can be inserted into these pores.
0
1 10-10
2 10-10
3 10-10
4 10-10
5 10-10
6 10-10
7 10-10
8 10-10
0 2 4 6 8 10
Cs2CO
3(0)
Cs2CO
3(0.014)
Cs2CO
3(0.14)
Cs2CO
3(0.71)
Cs2CO
3(1.4)
Cs2CO
3(4.1)
ΔQ
He [m
3 (STP
) m-2
s-1
Pa-1
nm
-1]
Kelvin diameter [nm]
12
From these results, it was revealed that pore size of the carbon membranes can be controlled by incorporating controlled amount of Cs2CO3 into the precursor.
13
(2) Inorganic membrane (2.1) Synthesis of novel zeolite membranes for CO2 separation
Based on the simulation results [6], we have selected a number of candidate zeolite structures for use in CO2 separation studies and synthesized of new zeolite membranes. In addition, we are continuing to prepare several zeolite seed crystals, from which several new zeolite membranes are anticipated. For example, defect free pure silica zeolites are regarded as hydrophobic and are therefore expected to be a candidate material for the CO2 separation membrane. In this study, we newly prepared various high silica zeolites. These have not previously been reported as a membrane (Figure 14).
Figure 14. SEM images of prepared zeolite crystals
(2.2) Challenging development of zeolite membrane without defects The preparation method of ultra-thin zeolite membranes was studied using two types of seeding method; i.e., an ultrasonication-based coating method and a rubbing-based seeding method. It was found that the performance and denseness of the membrane prepared by the rubbing-based seeding method was higher than that of the membrane produced by the ultrasonication-based coating method. The zeolite Y membrane synthesized by the rubbing-based seeding method possessed a complex layer consisted of zeolite and porous substrate particles, and yielded a separation factor of 69.3. Based on these results, we have demonstrated the novel synthesis of a defect-free zeolite membrane suitable for CO2 separation. Here, the porous substrate has been filled with zeolite crystals. Figure 15 shows a scheme illustrating the ‘melt-filling method’. We found that
14
the chemical and physical conditions must be optimized for the melting, plasticization, and re-crystallization of the zeolite membrane. As shown in Figure 16, SEM images of the zeolite DDR membrane reveal that the membrane has moved into the porous substrate as a result of the melt-filling method.
Figure 15. Schematic illustration of the melt-filling method involving the melting and re-crystallization of the zeolite membrane.
15
From the XRD measurement, it was revealed that these composites had the same diffraction pattern; there was no phase transition and no byproducts before and after the melt-filling synthesis. The gas permeability measurements of the zeolite membrane prepared by the melt-filling method are currently in the process of evaluation. For example, while the CO2/N2 separation selectivity of the zeolite Y membrane was 7, the selectivity became 25 with the melt-filling synthesized membrane (Figure 17).
Figure 16. SEM images of a) initial DDR zeolite membrane (top view), b) intermediate of the melt-filling treatment (top view), c) filled zeolite crystals in porous substrate after the melt-filling treatment (top view), d) initial zeolite membrane (cross-section), e) intermediate of the melt-filling treatment (cross-section), and f) filled zeolite crystals in porous substrate after the melt-filling treatment (cross-section).
16
Figure 17. SEM images of the initial zeolite Y membrane (left) and filled zeolite crystals in porous substrate after melt-filling treatment (right). However, the separation selectivity of the membranes formed by this method tends to be low. We assumed that this is due to defects in the formed zeolite membranes. Therefore, the difference in the coefficiency of thermal expansion between the zeolite membrane and substrate needs to be taken into consideration to solve the defect in the membranes. So far, we have been attempting to develop a structure with a porous substrate that forms a strain relaxation layer (Figure 18). Optimization of the a strain relaxation layer is still in progress.
Figure 18. SEM cross-section image of the zeolite membrane and a porous substrate which is effective for thermal stress relaxation.
17
(2.3) Development of a synthetic method of defect-free mesoporous thin films.
Self-assembly of surfactants is considered as a popular template to tailor the pore sizes of inorganic nanoporous materials. Mesoporous silica, synthesized by this technique, has a sharp pore size distribution, and its pore size is larger than zeolites (zeolites: 0.3-1 nm, mesoporous silica: 2-10 nm). Therefore, it is possible to modify the internal pore surface with various compounds. The functionalization of the pore walls of mesoporous silicas is effective for selective CO2 adsorption [7].So far mesoporous silica thin membranes on the porous alumina substrate with uniform pore structures have been prepared by templating silicates with various surfactant micelles and using dip-coating/spin-coating techniques and hydrothermal treatment.
Figure 19. Preparation of amine-modified silica membrane.
In this study, various functionalized mesoporous silica membranes for CO2
separation were successfully prepared on porous alumina supports by hydrothermal treatment and spin-coating of silica sol. Figure 19 shows the images of amine-modified mesoporous silica membrane. SEM images showed that these techniques deposited dense mesoporous silica layers of 200 - 500 nm, on the alumina supports. From the TEM and XRD observations, it was shown that these membranes have a highly ordered cubic structure with a pore diameter of ca. 2 nm. The gas permeation properties of these mesoporous silica membranes were governed by the Kundsen diffusion mechanism. Surface modification of the pore walls of these mesoporous silica membranes by grafting amino-silane greatly improved the CO2 permselectivities. After amine modification, the thickness of the dense mesoporous silica membrane layer prepared by the hydrothermal method increased to 2000 nm, while no obvious change was observed for the membrane prepared by the spin-coating method. Immersion of the porous alumina substrate into liquid paraffin before spin-coating of the silica sol was effective at preventing the sol from percolating into the pores of the substrate, causing the formation of a dense layer. The amine-modified mesoporous silica membrane prepared by the spin-coating method showed higher CO2 permeability and selectivity than the membrane prepared by the hydrothermal method, because of the thinner separation layer.
18
Both of the amine-modified membranes showed extremely high CO2 permselectivity and their α(CO2) values at 373 K were 50 and 800, respectively.
It was confirmed that amine-modified mesoporous silica layers work as an effective CO2 separation layer, and this result showed a possibility of CO2 separation at high temperature. However, the CO2 permeability of the obtained modified mesoporous silica membrane is still insufficient.（CO2 permeance = 2.2×10-11 m3(STP)·m-2·s-1·Pa-1). Therefore, in order to improve the CO2 permeance, various mesoporous silica membranes with different pore size/structures on porous alumina and glass substrates were prepared by hydrothermal, sol-gel spin-coating and sol-gel dip-coating techniques. It was found that hydrophobizing treatment of the porous substrates was effective for the preparation of thin separation layers. Based on this, we have prepared SBA-16 membrane with high gas permeability (CO2 permeance = 5.0×10-9 m3(STP)·m-2·s-1·Pa-1). From these results,we comfirmed that it is necessary to optimize the guest material for surface functionalization and pore size, as well as the modification method of the pore walls. Conclusions (1) Carbon membranes
(1.1) Sub-nanoporous carbon membranes with enhanced CO2 affinity
Novel carbon membranes with enhanced CO2 affinity have been developed. In both methods A and B, the prepared membranes had higher separation performances than did the original carbon membranes in the separation of CO2/N2 mixtures at high relative humidity.
(1.2) Control of subnano/nano pores in carbon membranes
It was revealed by a water vapor sorption experiment and nanopermporometry that carbon pores became more hydrophilic and that pore size distribution was shifted to larger pore size by the incorporation of Cs into carbon membranes. It was suggested that the change in pore size and pore surface properties played an important role to improve CO2 separation performance under humid conditions.
(2) Inorganic membranes (2.1) It was revealed that the formation of a complex layer of zeolite and the substrate improved the permeation performance when compared with the membrane prepared by arranging crystals on the substrate. A novel method for the preparation of dense complex zeolite membranes was proposed and shown to be effective. (2.2) The functionalization of the pore walls of mesoporous silicas by compounds with high amine density is effective in the separation of CO2. To improve the CO2 permeance
19
of the mesoporous silica membrane, it was deemed necessary to pre-treat the porous substrate and optimize the mesopore pore structure for CO2 separation. Publications
1. Teruhiko Kai, Firoz Alam Chowdhury, Shingo Kazama and Yuichi Fujioka, Material, manufacture and application of novel carbon membrane for CO2 capture application, October 1, '07, Number: Jpn Appl Number 2007-257493.
2. Teruhiko Kai, Shingo Kazama and Yuichi Fujioka, "Development of cesium-incorporated carbon membranes for CO2 separation under humid conditions", submitted to Journal of Membrane Science.
3. “A new preparation method of dense zeolite membrane” Japanese patent application 2007-179072, Katsunori YOGO, Kousuke UOE, Naoki YAMAMOTO, Manabu MIYAMOTO, Yuichi Fujioka
4. “A new synthetic method of pure silica zeolites” Japanese patent application 2007-286783 Kousuke UOE, Katsunori YOGO, Manabu MIYAMOTO, Yuichi Fujioka
5. “Preparation and CO2 separation properties of amine modified mesoporous silica membrane”, Yuzuru Sakamoto, Kensuke Nagata, Katsunori Yogo and Koichi Yamada, Microporous and Mesoporous Materials,vol101(1-2) 303-311(19 April 2007).
References
1. A. S. Kovvali, H. Chen, and K. K. Sirkar, Dendrimer membranes: A CO2-Selective Molecular Gate, J. Am. Chem. Soc. 122 (2000) 7594-7595.
2. F. A. Chowdhry, Y. Shimada, H. Oku, S. Kazama and K. Yamada, CO2 Separation Membrane of Modified PAMAM Dendrimer, International Congress on Membranes and Membrane Processes (ICOM), Seoul, Korea (2005)
3. S. Duan, T. Kouketsu, S. Kazama and K. Yamada, Development of PAMAM dendrimer composite membrane for CO2 separation, J. Membr. Sci., 283 (2006) 2-6.
4. T. Kouketsu, S. Duan, T. Kai, S. Kazama and K. Yamada, PAMAM dendrimer composite membrane for CO2 separation: Formation of a chitosan gutter layer, J. Membr. Sci., 287 (2007) 51-59.
5. T. Kai, T. Kouketsu, S. Duan, S. Kazama, K. Yamada, Development of commercial-sized dendrimer composite membrane modules for CO2 removal from flue gas, Sep. Purif. Tech., 63 (2008) 524-530.
6. H.Takaba et. al, ZMPC2006, OC101 (2006). 7. N. Hiyoshi, K. Yogo and T. Yashima, “Adsorption Characteristics of Carbon Dioxide on
Organically Functionalized SBA-15”, Microporous and Mesoporous Materials, 84(2005) 357-365.
