Relationship between the mesoporous intermediate layer structure and the gas permeation property of an amorphous silica membrane synthesized by counter diffusion chemical vapor deposition

Article abstract of DOI:10.1111/j.1551-2916.2007.02073.x

An amorphous silica membrane with an excellent hydrogen/nitrogen (H ₂/N₂) permselectivity of >10 000 and a He/H₂ permselectivity of 11 was successfully synthesized on a γ-alumina (γ-Al₂O₃)-coated α-alumina (α-Al ₂O₃) porous support by counter diffusion chemical vapor deposition using tetramethylorthosilicate and oxygen at 873 K. An amorphous silica membrane possessed a high H₂ permeance of >1.0 × 10^-7 mol·(m²·s·Pa)^-1 at ≥773 K. The dominant permeation mechanism for He and H₂ at 373-873 K was activated diffusion. On the other hand, that for CO₂, Ar, and N₂ at 373-673 K was a viscous flow. At ≥673 K, that for CO ₂, Ar, and N₂ was activated diffusion. H₂ permselectivity was markedly affected by the permeation temperature, thickness, and pore size of a γ-Al₂O₃ mesoporous intermediate layer.

Full text of DOI:10.1111/j.1551-2916.2007.02073.x

J. Am. Ceram. Soc., 91 [1] 71–76 (2008)
DOI: 10.1111/j.1551-2916.2007.02073.x
r 2007 The American Ceramic Society
ournal
J
Relationship between the Mesoporous Intermediate Layer Structure
and the Gas Permeation Property of an Amorphous Silica Membrane
Synthesized by Counter Diffusion Chemical Vapor Deposition
w
Takayuki Nagano, Shinji Fujisaki, Koji Sato, Koji Hataya, and Yuji Iwamoto
Japan Fine Ceramics Center, Nagoya 456-8587, Japan
Mikihiro Nomura
Shibaura Institute of Technology, Tokyo 135-8548, Japan
Shin-Ichi Nakao
University of Tokyo, Tokyo 113-8656, Japan
An amorphous silica membrane with an excellent hydrogen/
nitrogen (H /N ) permselectivity of 410000 and a He/H perm-
bining these processes into a single step represent technical
1
–3
problems.
A membrane reactor has been investigated as a device for H
2
2
2
selectivity of 11 was successfully synthesized on a c-alumina
2
(
c-Al O )-coated a-alumina (a-Al O ) porous support by
mass production. It is well known that the application of high-
temperature membrane reactors to this steam-reforming step
has the potential of achieving the same conversion efficiencies as
those attained using conventional reactors at a significantly low
temperature of approximately 773 K. If a membrane with a high
durability under a steam atmosphere at elevated temperatures is
developed, these technical problems can be solved concurrently.
2
counter diffusion chemical vapor deposition using tetra-
3
2
3
methylorthosilicate and oxygen at 873 K. An amorphous silica
membrane possessed a high H permeance of 41.0 ꢀ 10
ꢁ
7
2
2
mol (m s Pa)
ꢁ1
.
mechanism for He and H at 373–873 K was activated diffu-
. .
at ꢂ 773 K. The dominant permeation
2
sion. On the other hand, that for CO , Ar, and N at 373–673 K
2
2
was a viscous flow. At ꢂ 673 K, that for CO , Ar, and N was
A palladium-based membrane with a high H permselectivity is
2
a candidate membrane reactor. However, there are some prob-
lems associated with the total cost of a resource.
An inorganic membrane with a high strength at elevated
temperatures and a high chemical stability is used under a high
pressure and corrosion atmosphere at elevated temperatures.
2
2
activated diffusion. H permselectivity was markedly affected by
2
the permeation temperature, thickness, and pore size of a
c-Al O mesoporous intermediate layer.
2
3
I. Introduction
Therefore, H production processes can be combined to sup-
2
press energy loss. An amorphous silica membrane is well known
as a H -permselective membrane. Amorphous silica membranes
YDROGEN (H
2
) has been attracting considerable attention as
2
H
a clean alternative energy source for fossil hydrocarbon
fuels. It is expected that H will be produced using natural
energies. However, it is difficult to meet the demand for H , and
H production through natural gas reforming has been investi-
2
gated. For example, methane (CH ), the main constituent of
4
4
are produced by sol–gel coating, chemical vapor deposition
,5
2
6
–14
(
CVD) and chemical vapor infiltration (CVI),
and precursor
In CVD and CVI, the surfaces of large
pores tend to be preferentially modified. Therefore, controlling
the formation of pinhole defects is easy, and a high H per-
meance to nitrogen (N ) (H /N permselectivity) has been
reported in membranes synthesized by CVD and CVI. For
2
1
5,16
polymer pyrolysis.
2
2
natural gas, is converted to H and carbon monoxide (CO)
through steam reforming as follows:
2
2
2
8
example, Tsapatsis et al. oxidized SiCl
2
glass by supplying H O gas and reported a H
4
gas on porous Vycor
/N permselectiv-
CH4 þ H2O ! 3H2 þ CO; DE ¼ 206 kJ=mol
(1)
2
2
1
2
ity of 5000 at 723 K. Nijimeijer et al. prepared a microporous
silica membrane on a mesoporous g-alumina (g-Al O )-coated
2
3
This reaction is endothermic and is induced at approximately
073 K. In this process, the synthesis gas then proceeds to a
a-alumina (a-Al
2
O
3
) porous support by low-temperature CVI
/N permselectivity of 440 at 523 K. Hwang
et al. prepared amorphous silica on a-Al porous tubes by
CVD using tetraethylorthosilicate and reported a H /N perm-
1
and reported a H
2
2
water–gas shift reactor where steam and CO can be converted
into H and carbon dioxide (CO ). Finally, H is separated from
1
3
2 3
O
2
2
2
2
2
the gas mixture using one of several adsorption technologies,
such as pressure swing adsorption. Owing to a high operation
temperature and several complex purification steps, the cost of
producing H with these current technologies is too high for it to
2
be used as an alternative source for conventional hydrocarbon
fuels. Therefore, decreasing the operation temperature and com-
1
4
selectivity of 160 at 873 K. Nomura et al. reported that a stable
amorphous silica membrane synthesized by counter diffusion
chemical vapor deposition (CDCVD) showed a H /N permse-
2
2
lectivity of 41000.
Inorganic membranes are composed of an active layer for
separating gas, an intermediate layer for suppressing the forma-
tion of pinhole defects on the active layer, and a porous support.
In the preparation of a microporous membrane by CVD and
CVI, the gas separation property strongly depends on the struc-
ture of the mesoporous intermediate layer. However, the rela-
tionship between the intermediate layer structure and gas
separation property has been reported in only a few papers.
T. Bessmann—contributing editor
Manuscript No. 23078. Received April 13, 2007; approved August 14, 2007.
This work was carried out as a part of the R&D Project on ‘‘Highly Efficient Ceramic
Membranes for High-Temperature Separation of Hydrogen’’ supported by the New Energy
and Industrial Technology Development Organization (NEDO), Japan.
13
Hwang et al. synthesized amorphous silica membranes on
w
Author to whom correspondence should be addressed. e-mail: t_nagano@jfcc.or.jp
a-Al O with an average pore size of 100 nm and g-Al O
3
2
3
2
7
1

7
2
Journal of the American Ceramic Society—Nagano et al.
Vol. 91, No. 1
(
average pore size of 10 nm)-coated a-Al
He permeance and He/N permselectivity in amorphous silica
on g-Al O -coated a-Al O were higher than those in amor-
2 3
O . They reported that
Table I. Preparation Conditions of g-Al O -Coated a-Al O
2
3
2
3
2
Capillaries
2
3
2
3
17
phous silica on a-Al
phous silica membranes on g-Al
3 nm)-coated a-Al O by CDCVD. They reported that
2 3
O . Nomura et al. synthesized amor-
Sample
Number of g-Al
O
2 3
coatings
Calcination temperatures
2 3
O
(average pore sizes of 4 and
1
2
1
2
A
A
B
B
1
2
1
2
873 K 1 h
873 K 1 h
1073 K 1 h
1073 K 1 h
1
2
3
H permeance and H /N permselectivity in amorphous silica on
g-Al
phous silica on g-Al
2
2
2
O
2 3
2 3
(4 nm)-coated a-Al O were higher than those in amor-
O
2 3
(13 nm)-coated a-Al O .
2 3
To obtain a high H permeance, the active layer should have
2
a large path area and a small thickness. However, the large pore
size of the intermediate layer leads to a pinhole defect and large
thickness of the active layer. On the other hand, to obtain high
H /N permselectivity, the pore surface of the intermediate layer
The pore size distribution of the g-Al
2
O
3
intermediate layer
on the a-Al O substrate was analyzed using a nanopermporo-
2
3
meter (TNF-3WH-110ME: Seika Sangyo Co., Tokyo, Japan).
was used as a noncondensable gas, and the liquid used as
2
2
N
2
2 2
should be modified within the kinetic diameter of N . H can
condensable vapor was water. Specimens were placed in the
apparatus after annealing at 393 K for 1 h.
The amorphous silica coating of a g-Al
substrate was performed by CDCVD, as shown in Fig. 2. The
substrate was coaxially fixed in a stainless tube and placed in an
electric tubular furnace. Tetramethylorthosilicate was supplied
permeate an amorphous silica network. Therefore, even if the
pore of the intermediate layer is completely plugged with amor-
phous silica, high H permselectivity can be expected.
2
3 2 3
O -coated a-Al O
2
In CDCVD, a membrane is synthesized at the contact point
of the source and reaction gases inside the intermediate layer, as
shown in Fig. 1. The pore size, pore filling homogeneity, silica
density gradient of the intermediate layer, and effective mem-
brane thickness can be controlled by changing the reactants
and reaction conditions. The deposition rate of the membrane
automatically decreases when the diffusion rates of the source
and reaction gases decrease with the deposition. Therefore, the
amorphous silica membrane synthesized by CDCVD exhibits a
through the outer surface of the substrate by controlling the N
flow rate at 200 mL/min. Oxygen (O ) was introduced into the
2
2
inner surface of the substrate at a flow rate of 200 mL/min. The
substrate temperature and deposition time were 873 K and 1 h,
respectively.
The top surfaces and cross-sectional structure of g-Al
2 3
O -
coated a-Al O capillaries were observed by scanning electron
2
3
high H
In this study, we investigated the relationships among the
gas separation property, pore size, and thickness of a g-Al
2
permeance.
microscopy (SEM; Model S-4500, Hitachi, Tokyo, Japan).
Transmission electron microscopy (TEM) specimens were
prepared using a focused ion beam (FIB) system (Model FB-
2 3
O
mesoporous intermediate layer in an amorphous silica mem-
2
100, Hitachi) at acceleration voltages of 40–10 kV. Tungsten
brane on g-Al O -coated a-Al O synthesized by CDCVD.
2
3
2
3
was deposited on the surface of the sampling area in the FIB
system to protect the top layer of the membrane from gallium
ion sputtering during FIB milling. The membrane was dug out
by FIB milling and small specimens (3 mm ꢀ 25 mm ꢀ 15 mm)
were prepared. The specimens were lifted out using a tungsten
needle and transferred to TEM grids. The specimens were then
fixed onto the TEM grids by FIB-assisted deposition and
thinned by FIB milling.
II. Experimental Procedure
porous capillary tube (150 nm in mean pore diam-
eter, 2.9 mm in outer diameter, 2.1 mm in inner diameter,
00 mm in length, NOK Corporation, Tokyo, Japan) was
2 3
An a-Al O
4
The cross-sectional structure of a silica membrane on a
used as a substrate. An effective membrane area with a length
of 5 cm was at the center of the substrate. The dip-coating
g-Al O -coated a-Al O capillary was observed by 200 kV
2
3
2
3
TEM (Model EM002B, TOPCON, Tokyo, Japan). The speci-
men surface was coated with carbon to prevent charge-up during
TEM observation. The composition of the top layer was ana-
lyzed by energy-dispersive X-ray spectroscopy (EDS) with TEM.
Single gas permeance (P) was evaluated by a constant-volume
manometric method. Permeance data can be measured from
solution for g-Al
g-AlOOH) with a 3.5 wt% solution of polyvinyl alcohol (mean
molecular weight5 72 000: Kanto Kagaku Co. Ltd., Tokyo,
Japan) at 363 K. The end of the a-Al substrate was plugged.
The outer surface of the substrate was dipped in the solution for
0 s, whereas the inner surface of the capillary was evacuated to
obtain a pinhole-free membrane using a vacuum pump. After
dipping, the membrane was dried for 2 h in air. It was then
calcined at 873 or 1073 K for 1 h at a heating/cooling rate of
2 3
O was obtained by diluting boehmite sol
(
2 3
O
1
ꢁ
5
ꢁ12
2
ꢁ1
1
ꢀ 10 to 1 ꢀ 10
mol ꢃ (m ꢃ s ꢃ Pa) by varying the experi-
mental conditions. The outside of the membrane was filled with
pure gas under atmospheric pressure conditions. The inside of
the membrane was evacuated using a vacuum pump. After ter-
minating evacuation, the rate of pressure increase inside the
membrane was measured five times to calculate permeance. The
1
K/min. This dipping–drying–firing sequence was repeated
2 3 2 3
once or twice. The g-Al O -coated a-Al O substrates prepared
are shown in Table I.
Fig. 1. Formation mechanism of gas separation membrane in counter
diffusion chemical vapor deposition.
Fig. 2. Schematic illustration of counter diffusion chemical vapor de-
position .

January 2008
Relationship between Mesoporous Intermediate Layer Structure and Gas Permeation Property
73
Fig. 4. Scanning electron microscopic top surface image of Sample 1A.
2 3
Fig. 3. Pore size distributions of g-Al O calcined at 873 and 1073 K
for 1 h.
2 3
for gas separation was estimated to be 1/13 of the g-Al O layer
thickness.
Therefore, the synthesized sample was considered to be com-
posed of two layers, namely, an a-Al porous support and a
g-Al intermediate layer gradually modified with amorphous
silica.
average of the third, fourth, and fifth measurements was
accepted as permeance data. The gas permeance at each tem-
2 3
O
2 3
O
2 2 2
perature was evaluated in the order of He, H , CO , Ar, and N .
Permselectivity was defined as the permeance ratio of the two
2 2
gases. For example, the H /N permselectivity is given by the
(
2) Gas Permeation Property
ratio PH =PN .
2
2
The gas permeation behavior of the g-Al O intermediate layer
2
3
2 3
on the a-Al O porous capillaries is shown in Fig. 8. The ex-
perimental data at 873 K were plotted versus the square root of
molecular weight. The gas permeance of Sample 1B with an av-
erage pore size of 7.6 nm was higher than that of Sample 1A
with an average pore size of 4 nm. The permeance order of the
III. Results and Discussion
1) Microstructure of Amorphous Silica Membrane
The effect of calcination temperature on the pore size distribu-
tion of the g-Al intermediate layer is shown in Fig. 3. The
(
O
2 3
gases through
a
g-Al O -coated a-Al O
2 3 2 3
capillary was
pore sizes of the sample calcined at 873 K for 1 h were almost
entirely distributed between 2 and 5 nm. On the other hand, the
pore sizes of the sample calcined at 1073 K for 1 h were almost
distributed between 4 and 10 nm. The average pore sizes of the
specimens increased with calcination temperature. The pore size
distribution did not depend on the number of coatings. The av-
erage pore sizes of the specimens calcined at 873 and 1073 K
were 3.6–4.0 and 6.6–7.6 nm, respectively.
H 4He4N 4Ar4CO . The linear regression fit is presented
in Fig. 8. The results show a good linear dependence, confirming
that the transport of the gases is induced mainly by Knudsen
2
2
2
2
diffusion. The H permselectivities over the other gases (Sample
1A: He: 1.38, CO : 4.47, Ar: 4.09, N : 3.39; Sample 2A: He 1.37,
2
2
CO : 4.50, Ar: 4.09, N : 3.40) showed good agreement with the-
2 2
2
oretical Knudsen permselectivity values (He: 1.41, CO : 4.67,
Ar: 4.45, N : 3.73), indicating that the pores are controlled finely
2
The SEM top surface image and cross-sectional image of the
and homogeneously.
g-Al
2
O
3
-coated a-Al
2
O
3
capillary are shown in Figs. 4 and 5.
Pinholes and cracks were not observed on the specimen surface.
The gas permeation properties through amorphous silica on
Samples 1A, 2A, 1B, and 2B at 798 K are shown in Fig. 9. After
CDCVD coating, permeance tended to increase with decreasing
The surface morphology did not change with calcination tem-
perature and the number of coatings. a-Al
coated with the g-Al layers, and the thicknesses of the
g-Al layers in Samples 1A, 2A, 1B, and 2B were approxi-
2
O
3
was uniformly
kinetic diameter. The He and H
affected by the changes in the pore size and thickness of the
g-Al layer. However, the CO , Ar, and N permeance
decreased in the specimens of amorphous silica on Sample 2A
and amorphous silica on Sample 2B coated with g-Al twice.
2
permeance was not markedly
2 3
O
O
2 3
O
2 3
2
2
mately 1.8, 2.6, 2.0, and 3.0 mm, respectively.
The TEM cross-sectional images and diffraction patterns of
the synthesized sample of amorphous silica on Sample 2B are
shown in Fig. 6. The top layer was observed as a defect-free thin
layer having a thickness of approximately 2.3 mm. The thickness
of the top layer was almost equal to that of the g-Al O inter-
2 3
mediate layer, as shown in Fig. 5. The selected area diffraction
patterns of the top layer were identical to those of amorphous
2 3
O
Therefore, the increase in intermediate layer thickness has an
effect of decreasing the CO , Ar, and N permeance.
2
2
Single gas permeance through amorphous silica on Sample
2B at 373–873 K is shown in Fig. 10. Permeance tended to
increase with decreasing kinetic diameter at all permeation tem-
peratures. This membrane showed molecular sieving behavior.
silica and g-Al
2
O
3
. The diffuse ring patterns were assigned to
The He and H
perature. On the other hand, the CO
was the lowest at 573 K.
The dependence of temperature on each gas permeance for
amorphous silica on Sample 2B is shown in Fig. 11. The He,
2
permeance increased with the permeation tem-
g-Al , suggesting that each grain was less crystalline g-Al
O
2 3
2 3
O .
, Ar, and N permeance
2 2
The TEM sample was also analyzed by EDS. EDS showed that
the entire top layer was composed of silicon (Si), aluminum (Al),
and O
2
.
ꢁ
ꢁ12
7
The Si/Al atomic ratios of the g-Al
silica on Sample 2B are shown in Fig. 7. The x-axis (distance
from the surface) is zero, indicating the surface of the g-Al O
shown in Fig. 6. The Si/Al ratios near the surface (distance from
membrane surface approximately o0.18 mm) of the g-Al
layer were increased sharply. The effective membrane thickness
O
2 3
layer for amorphous
H
2
, CO , Ar, and N
2
2
permeance at 798 K was 2.98 ꢀ 10
,
ꢁ
7
ꢁ11
ꢁ11
1.21 ꢀ 10
,
4.10 ꢀ 10
,
1.99 ꢀ 10
2
,
and 9.92 ꢀ 10
2
ꢁ1
mol ꢃ (m ꢃ s ꢃ Pa) , respectively. The H permeance at 798 K
2
3
was four orders of magnitude higher than the N
at 798 K, and the H /N permselectivity was calculated to
be 12 200. The He and H2 permeance increased with the
2
permeance
O
2 3
2
2

7
4
Journal of the American Ceramic Society—Nagano et al.
Vol. 91, No. 1
Fig. 5. Scanning electron microscopic cross-sectional image of Sample 1A.
permeation temperature. Accordingly, the dominant permeation
mechanism for He and H was considered to be activated diffu-
sion. On the other hand, the CO , Ar, and N permeance de-
creased with increasing permeation temperature at 373–673 K.
According to the Knudsen diffusion and viscous flow mecha-
nisms, the permeation of a gas molecule in a porous membrane
leads to a decrease in permeance with increasing temperature.
The permeance order of the gases in the Knudsen diffusion
2
mechanism is H
dependence on the molecular weight of the gases. Therefore,
the dominant permeation mechanism for CO , Ar, and N was
2 2 2
4He4N 4Ar4CO , showing an inverse
2
2
2
2
considered to be viscous flow at 373–673 K. These gases were
considered to permeate through defects rather than the micro-
pores of the amorphous silica network. However, the CO , Ar,
2
Fig. 6. Transmission electron microscopic cross-sectional images of amorphous silica membrane on Sample 2B.

January 2008
Relationship between Mesoporous Intermediate Layer Structure and Gas Permeation Property
75
Fig. 9. Effects of number of coatings and pore size on gas permeation
behavior.
Fig. 7. Distribution of Si/Al ratios for amorphous silica membrane on
Sample 2B by energy-dispersive X-ray spectroscopy analysis.
2
and N permeance increased with the permeation temperature
at ꢂ673 K. This result shows that the dominant permeation
mechanism for CO , Ar, and N changed to activated diffusion
2
2
at ꢂ673 K.
The activation energies of permeation were obtained by
fitting the experimental gas permeance data to the Arrhenius
expression
Q ¼ Q0 expðꢁEa=RTÞ
(2)
where Q is the permeance, Q
(
gas constant (8.314 J ꢃ (molꢃ K) ), and T the temperature (K).
The apparent activation energies in the amorphous silica mem-
brane on Sample 2B were roughly estimated from the experi-
0
the pre-exponential factor
the activation energy (Jꢃ mol ), R the
2
ꢁ1
ꢁ1
mol ꢃ (m ꢃ s ꢃ Pa) ), E
a
ꢁ
1
mental data. The apparent activation energies for the He and H
2
permeance in amorphous silica on Sample 2B were 8.1 and 16.8
kJ/mol in the temperature range from 373 to 873 K, respectively.
The apparent activation energy of amorphous silica on Sample
2B for the H permeance was slightly higher than that of amor-
2
phous silica on g-Al O -coated a-Al O synthesized by CVD
2
3
2
3
18
1
3
(
Hwang et al. : 15.3 kJ/mol, Lee et al. : 14.8 kJ/mol). On the
Fig. 10. Gas permeation behavior of amorphous silica on Sample 2B in
the temperature range from 373 to 873 K.
other hand, those for the CO , Ar, and N permeance were 26.2,
2
2
Fig. 11. Dependence of temperature on gas permeance in amorphous
silica on Sample 2B.
Fig. 8. Gas permeation behavior of Samples 1A and 1B at 873 K.

7
5
6
Journal of the American Ceramic Society—Nagano et al.
Vol. 91, No. 1
6
2 2
S. W. Nam and G. R. Gavalas, ‘‘Stability of H -Permselective SiO Films
2.9, and 25.0 kJ/mol in the temperature range from 673 to 873
Formed by Chemical Vapor Deposition,’’ AIChE Symp. Ser., 85 [268] 68–74
(1989).
K, respectively.
As discussed previously, He, H , CO , Ar, and N through
7
2
2
2
2
G. R. Gavalas, C. E. Megiris, and S. W. Nam, ‘‘Deposition of H -Permselec-
the specimen of amorphous silica on Sample 2B mainly perme-
ated through the amorphous Si–O network at ꢂ673 K. Their
permeance depended on the effective area and thickness of the
amorphous silica film. Accordingly, the thickness and pore size
tive SiO
2
Films,’’ Chem. Eng. Sci., 44 [9] 1829–35 (1989).
8
M. Tsapatsis, S. Kim, S. W. Nam, and G. R. Gavalas, ‘‘Synthesis of Hydrogen
, TiO , Al , and B Membranes from the Chloride Pre-
cursors,’’ Ind. Eng. Chem. Res., 30 [9] 2152–9 (1991).
Permselective SiO
2
2
O
2 3
2 3
O
9
C. E. Megiris and J. H. E. Glezer, ‘‘Synthesis of
2
H -Permselective
of the g-Al
proving the H
2
O
3
intermediate layer are important factors for im-
permeance to CO , Ar, and N
Membranes by Modified Chemical Vapor Deposition, Microstructure and Perm-
selectivity of SiO /C/Vycor Membranes,’’ Ind. Eng. Chem. Res., 31 [5] 1293–9
1992).
2
.
2
2
2
(
10
H. Y. Ha, S. W. Nam, S.-A. Hong, and W. K. Lee, ‘‘Chemical Vapor De-
position of Hydrogen-Permselective Silica Films on Porous Glass Supports from
Tetraethylorthosilicate,’’ J. Membr. Sci., 85 [3] 279–90 (1993).
IV. Conclusion
11
S. Yan, H. Maeda, K. Kusakabe, S. Morooka, and Y. Akiyama, ‘‘Hydrogen-
Permselective SiO Membrane Formed in Pores of Alumina Support Tube by
An amorphous silica membrane with molecular-sieving behav-
ior was synthesized on g-Al O -coated a-Al O by CDCVD.
The thickness and pore size distribution of a g-Al O mesopo-
2
2
3
2
3
Chemical Vapor Deposition with Tetraethyl Orthosilicate,’’ Ind. Eng. Chem. Res.,
33 [9] 2096–101 (1994).
2
3
12
A. Nijimeijer, B. J. Bladergroen, and H. Verweij, ‘‘Low-Temperature CVI
Modification of g-Alumina Membranes,’’ Microporous Mesoporous Mater., 25 [1/
2
rous intermediate layer were significant factors in H permse-
lectivity. The optimization of the intermediate layer structure is
very important for decreasing the permeance of gases with large
kinetic diameters.
3
] 179–84 (1998).
G.-J. Hwang, K. Onuki, and S. Shimizu, ‘‘Studies on Hydrogen Separation
13
Membrane for IS Process, Membrane Preparation with Porous a-Alumina Tube,’’
JAERI Res., 98-002, 1–8 (1998).
14
M. Nomura, K. Ono, S. Gopalakrishnan, T. Sugawara, and S.-I. Nakao,
‘Preparation of a Stable Silica Membrane by a Counter Diffusion Chemical Va-
por Deposition Method,’’ J. Membr. Sci., 251 [1–2] 151–8 (2005).
‘
References
1
15
E. Kikuchi, ‘‘Membrane Reactor Application to Hydrogen Production,’’ Catal.
Today, 56 [1/3] 97–101 (2000).
D. Li and S.-T. Hwang, ‘‘Preparation and Characterization of Silicon
Based Inorganic Membrane for Gas Separation,’’ J. Membr. Sci., 59, 331–52
(1991).
2
E. Kikuchi, Y. Nemoto, M. Kajikawa, S. Uemiya, and T. Kojima, ‘‘Steam
Reforming of Methane in Membrane Reactors: Comparison of Electroless-Plating
16
Y. Iwamoto, K. Sato, T. Kato, T. Inada, and Y. Kubo, ‘‘A Hydrogen-Perm-
and CVD Membranes and Catalyst Packing Modes,’’ Catal. Today, 56 [1/3] 75–81
(
selective Amorphous Silica Membrane Derived from Polysilazane,’’ J. Eur. Ceram.
Soc., 25 [2–3] 257–64 (2005).
M. Nomura, S. Gopalakrishnan, H. Aida, T. Sugawara, S.-I. Nakao, S. Yam-
2000).
3
17
K. Jarosch and H. I. de Lasa, ‘‘Novel Riser Simulator for Methane Reforming
Using High Temperature Membranes,’’ Chem. Eng. Sci., 54 [10] 1455–60 (1999).
azaki, Y. Iwamoto, R. Kojima, and A. Nakao, ‘‘Investigation of a Hydrothermal
Stable Hydrogen Permselective Silica Membrane for a Large Membrane Area
Module,’’ Membrane, 30 [5] 275–81 (2005).
4
T. Okubo and H. Inoue, ‘‘Single Gas Permeation through Porous Glass Mod-
ified with Tetraethoxysilane,’’ AIChE J., 35 [5] 845–8 (1989).
5
18
R. J. R. Uhlhorn, K. Keizer, and A. J. Burggraaf, ‘‘Gas Transport and Sep-
aration with Ceramic Membranes, Part II: Synthesis and Separation Properties of
Microporous Membranes,’’ J. Membr. Sci., 66 [2/3] 271–87 (1992).
D. Lee, L. Zhang, S. Niu, and R. F. Saraf, ‘‘Synthesis, Characterization, and
Gas Permeation Properties of a Hydrogen Permeable Silica Membrane Supported
on Porous Alumina,’’ J. Membr. Sci., 231 [1–2] 117–26 (2004).
&