Available online at www.sciencedirect.com
Ceramics International 36 (2010) 643–651
www.elsevier.com/locate/ceramint
Preparation of asymmetric perovskite-type membranes by
a settlement method
Maki Matsuka *, Igor E. Agranovski, Roger D. Braddock
Griffith School of Engineering, Nathan Campus, Griffith University, 170 Kessels Road, Nathan, Brisbane, Qld 4111, Australia
Received 10 June 2009; received in revised form 18 June 2009; accepted 29 September 2009
Available online 4 November 2009
Abstract
Asymmetric SrCe0.95Yb0.05O3a (SCYb) membranes were prepared by using a settlement method to deposit a thin dense SCYb layer on the
green porous SCYb support, followed by conventional dry pressing and co-sintering. SCYb powder was mixed in a liquid media (isopropanol) and
directly poured onto the pre-pressed green porous SCYb support placed in a beaker. After the powder fully settled onto the support, the green
asymmetric sample was dried and co-sintered. The cross-sectional images of the asymmetric systems indicate that thin dense SCYb films in the
asymmetric membranes had uniform thicknesses at this level of analysis and show excellent adhesion to the porous support without any undesirable
defects, delamination or cracking. Further, different dense layer thicknesses (80–20 mm) were readily obtained by altering the amount of the SCYb
powders in the liquid media. The minimum dense layer thickness obtained by the settlement method, with the synthesis conditions utilised in this
study, was around 20 mm.
# 2009 Elsevier Ltd and Techna Group S.r.l. All rights reserved.
Keywords: A. Pressing; D. Perovskite; E. Membranes; Settlement method
1. Introduction
More than 85% of the current world energy sources are from
non-renewable sources, which have led to significant environmental impacts such as greenhouse gas emissions. The
greenhouse gas emissions from the stationary and transport
energy sectors, and fugitive emissions, are the largest sources of
greenhouse gases in Australia [1]. These data strongly indicate
the necessity of developing alternative renewable, clean and
reliable energy sources, such as hydrogen fuel.
One of the main difficulties in the production of hydrogen
fuel is the cost of the separation of hydrogen from gas mixtures
such as gasification syngas. There are three major hydrogen
separation processes, namely pressure swing adsorption,
cryogenic distillation, and membrane separation. Hydrogen
separation membranes can include polymer, dense metal, nanoporous inorganic, dense cermet and dense ceramic membranes
[2,3]. Dense ceramic membranes with mixed protonic–
electronic conduction seem to be one of the most promising
* Corresponding author. Tel.: +61 0 7 3735 5259; fax: +61 0 7 3735 7459.
E-mail address: m.matsuka@griffith.edu.au (M. Matsuka).
technologies for high temperature hydrogen separation. It
offers advantages over the alternative technologies, such as low
energy consumption, relatively inexpensive material, and
simple and compact equipment and operation, and no
requirement for an external electric supply [2,4]. However,
the attainment of high hydrogen separation rates of such
membranes is still the greatest barrier to commercially viable
hydrogen separation membranes. There are several approaches
that can be employed to improve hydrogen permeation flux
through mixed protonic–electronic conducting dense ceramic
membranes: (1) optimisation of membrane compositions to
increase proton and electron conductivities; (2) deposition of a
thin film on a porous support to decrease the dense membrane
thickness [4]; (3) coating of a dense membrane with a porous
layer to increase effective surface area for hydrogen dissociation/re-association reactions; (4) coating of a dense membrane
with catalyst to increase the surface exchange reactions of
hydrogen [5]; and (5) reduction of the grain boundary
resistance.
When bulk diffusion is the limiting factor in the hydrogen
permeation process, reducing the dense membrane thickness
may be one of the most simple but effective techniques to obtain
improved hydrogen permeation flux. For example, Hamakawa
0272-8842/$36.00 # 2009 Elsevier Ltd and Techna Group S.r.l. All rights reserved.
doi:10.1016/j.ceramint.2009.10.007
644
M. Matsuka et al. / Ceramics International 36 (2010) 643–651
et al. [6] verified that hydrogen permeation rates through a
2 mm-thick SrCe0.95Yb0.05O3a dense layer at 677 8C were
approximately 500 times larger than through 1 mm-thick
SrCe0.95Yb0.05O3a membrane. This can be described with the
Wagner equation where the hydrogen permeation flux is
proportional to the reciprocal of the thickness of the membrane.
Therefore, there have been many studies on fabrication of
thin dense layers on porous supports (not limited to hydrogen
separation membranes). Ma et al. [7] prepared 60–150 mmthick dense BaCe0.9xZrxYb0.1O3a membranes on porous Ni–
BaCeO3 supports by air plasma spraying. Etchegoyen et al. [8]
used tape casting to obtain asymmetric membranes with a
120 mm dense La0.6Sr0.4Fe0.9Ga0.1O3a layer on a thick porous
support layer of the same material. Other studies include for
example, dip coating (20 mm thin La0.2Sr0.8CoO3a films on
porous MgO supports) [9], pulsed laser deposition (0.6–1.5 mm
yttria-stabilised zirconia layers on platinum and sapphire
supports) [10], and spin coating (2–140 mm dense
SrCe0.95Yb0.05O3 thin films on SrZr0.95Y0.05O3 support) [6].
Despite there being a number of sophisticated techniques
available for the deposition of thin dense layers on porous
supports, some of the techniques can be very complex and
difficult to replicate. In contrast, conventional dry pressing is
known to be a simple, reproducible and cost-effective method
to fabricate asymmetric membranes [11,12]. For example,
Cheng et al. [12] fabricated asymmetric membranes consisting of 150–800 mm dense SrCe0.95Tm0.05O3a layers on
SrCe0.95Tm0.05O3a porous supports by conventional dry
pressing, which is the currently achievable thickness using
such simple techniques. Xia and Liu [11] prepared an 8 mm
Gd0.1Ce0.9O1.95 (GDC) dense film on a porous NiO-GDC
substrate by traditional dry pressing with unique ‘foam’ (highly
porous) GDC powder. However, they reported that the GDC
‘foam’ powder had very poor flow behaviour due to its high
porosity and particle sizes, resulting in non-uniform distribution of powder and low packing density. They suggested that
improvements can be made by improving the flow behaviour of
the powder or by improving the filling techniques to obtain a
uniform and reproducible filling. Therefore the focus of the
present work is the improvement of the filling techniques to aid
the conventional dry pressing, by developing a method of
depositing a thin uniform dense layer on a green porous
support. The layers so generated will need to overcome
difficulties of cracking and non-uniformity, and also need to
be readily reproducible.
2. Experimental
Preparation of asymmetric membranes can be complicated
due to differences in shrinkage or thermal expansion rates in the
dense and porous layers, resulting in cracks, reactions and/or
delamination. Etchegoyen et al. [8] co-sintered (i.e. no presintering of porous supports) asymmetric perovskite membranes prepared by tape casting and lamination, which
contained cornstarch particles as the pore forming agent in
the porous supports. In that study, the shrinkage behaviour of a
dense membrane and of a porous support containing cornstarch
particles was measured by dilatometry during the co-sintering
of the asymmetric system. The result showed identical
shrinkage behaviour of the asymmetric system, indicating
successful co-sintering of the dense and porous layers.
Scanning electron microscope images of the co-sintered
membranes also confirmed flat and crack-free asymmetric
membranes with no delamination or interfacial reaction
between the two layers. This suggests that there will be less/
no complications with the control of shrinkage behaviours in
preparation of asymmetric membranes [8,13]. Therefore, green
dense SCYb membranes and green porous SCYb supports
containing cornstarch particles were co-sintered in this study.
2.1. Precursor powder synthesis
The SrCe0.95Yb0.05O3a (SCYb) powder was synthesised by
solid-state reaction. Appropriate amounts of oxide powders,
SrCO3 (99.9+%, Sigma–Aldrich) and CeO2 (99.9%, Strem
Chemicals) and Yb2O3 (99.9%, Alfa Aesar) were mixed in a
ball mill. The mixture was calcined in an alumina crucible in
air at 1300 8C for 12 h with the heating and cooling rates of
5 8C/min. The calcined powder was then ground with a mortar
and pestle, and sieved. The calcination, grinding and sieving
were repeated.
2.2. Preparation of porous supports
The same ceramic materials were utilised in the dense layer
and porous support, and the porosity of the porous support was
controlled by cornstarch particles which leaves pores after the
combustion, as mentioned above. The SCYb powder was mixed
with an appropriate amount of cornstarch particles in a ball
mill. The mixture was put into a die of 25 mm diameter and
uniaxially pressed with a pressure of approximately 80 MPa to
obtain a green porous SCYb support. The different cornstarch
contents of 10, 20, and 30 wt% in the porous supports were
preliminarily examined and for the preparation of the
asymmetric membranes, the porous supports with 20 wt%
cornstarch were adopted. This amount of cornstarch results in a
suitably porous support.
2.3. Asymmetric membrane fabrication
Green dense SCYb membranes and green porous SCYb
supports containing cornstarch particles were co-sintered in this
study. As mentioned in Section 2.2, the different cornstarch
contents in the porous supports were first examined to
determine the suitable amount of cornstarch in order to avoid
delamination. In order to deposit a thin dense SCYb layer on the
green porous SCYb support, a ‘‘settlement method’’ described
in the work by Agranovski et al. [14], was employed, instead of
tape casting and lamination utilised in the study by Etchegoyen
et al. [8]. The settlement method was then followed by
conventional dry pressing to produce the asymmetric SCYb
membranes, and co-sintering. In this study, isopropanol
(Merck, viscosity: 1.96 cP at 25 8C) was utilised as the liquid
media for the settlement method. Accurately weighed SCYb
645
M. Matsuka et al. / Ceramics International 36 (2010) 643–651
powder was added to 6 ml of isopropanol and mixed with a
magnetic stirrer for 1 h. After the powder was uniformly
dispersed in the isopropanol, the mixture was directly poured
onto the green porous SCYb support placed in a beaker. They
were left until the powder fully settled and then the excess
isopropanol was removed from the beaker. The green
asymmetric sample was dried at room temperature overnight.
The green asymmetric membrane was carefully removed from
the beaker and pressed again with a final pressure of 230 MPa.
The green membrane was sintered in air at 1500 8C for 12 h,
with the heating rate of 2 up to 550 8C, 5 8C/min from 550 to
1500 8C and the cooling rate of 5 8C/min. The heating rate of
2 8C/min was essential during the combustion of cornstarch (up
to 550 8C) to avoid the separation of the dense layer from the
porous support, due to significant temperature changes or rapid
generation of gas due to the combustion of cornstarch. The
heating rate of 5 8C/min was acceptable for heating over
550 8C, and for cooling. The samples were placed on a platinum
foil in an alumina crucible to avoid the reaction between the
samples and the alumina, as strontium can react with alumina to
form SrAl2O4 at temperatures higher than 1200 8C, possibly
leading to loss of strontium and an excess of cerium [15]. The
thickness of the dense layer was controlled by the concentration
of SCYb powder in isopropanol. Prior to the powder settlement,
it was essential to ensure that (1) the diameter of the beaker was
slightly larger than the diameter of the green support to avoid
any edge effect from the beaker wall and (2) the liquid media
provided enough height for particles to settle in order to obtain
uniform settlement of the powders. The uniformity of the
settled dense layers was verified with a scanning electron
microscopy (SEM, FEI Quanta 200).
where Wd is the weight of the dry sample (kg m/s2), Ws is the
weight of the ceramic suspended in water (kg m/s2), which is
measured as Wd minus the buoyancy force, F B (kg m/s2), W w is
the weight of the ceramic removed from water (kg m/s2). The
true density of ceramic, r, was determined from:
2.4. Characterisation
apparent porosity ¼
The elemental compositions of the synthesised powder was
analysed with inductively coupled plasma mass spectrometry
(ICP-MS). The particle sizes of SCYb powder and cornstarch
particles were determined by a volume based particle size
analyser (Malvern, Mastersizer S). Phase crystallographic
structures of the SCYb powder were determined by X-ray
diffraction analysis (XRD) using Cu Ka radiation. The
microstructure of the asymmetric membranes and the
thicknesses of the dense layers were analysed using SEM
(JEOL 840A and FEI Quanta 200) [8]. The gas tightness of the
dense layers of the asymmetric membranes was determined by
helium gas permeation at room temperature using a bubble test.
The samples tested are categorised into two groups: (A)
asymmetric membranes which were sintered at 1465 8C and
(B) asymmetric membranes sintered at 1500 8C. Bulk density,
true density, theoretical density, true porosity, and apparent
porosity of porous supports with various cornstarch contents,
were calculated by using the modified Archimedes method
[16,17]. Bulk density (density of samples containing pores), rb,
was calculated from:
The fractions of pores closed were determined by
rb ¼
Wd
Ww Ws
(1)
r¼
W d rw
;
FB
(2)
where rw is the density of water. Theoretical density, rT, was
calculated from:
rT ¼
zA
V CNA
(3)
where z is the number of formula units per unit cell (4 formula
units/cell for perovskite), A is the formula weight (g/mol), VC is
the volume of the unit cell calculated from lattice parameters of
the unit cell (cm3/cell) and NA is Avogadro’s number (formula
units/mol). Percent theoretical density was determined as
theoretical density ð%Þ ¼
rT r
rT
(4)
The true porosity which includes both interconnected and
closed pores, was calculated from:
true porosity ¼
r rb
r
(5)
The apparent (interconnected) porosity which relates permeability of gases through the support, was calculated from:
Ww Wd
Ww Ws
fraction pores closed ¼
true porosityapparent porosity
:
true porosity
(6)
(7)
The downside of the settlement method may be the nonuniformity of the particle settlement, due to the fact that the
particle settlement depends on the density and diameters of
settling particles. This possibly leads to non-uniform distribution of the powders across the vertical cross-section of the dense
layers. Therefore, energy dispersive X-ray spectroscopy (EDX)
was employed in order to confirm the uniformity of the elemental compositions across the vertical cross-section of the
deposited dense layers. Approximately every 10 mm along the
vertical cross-section of the dense layers, was scanned, from the
top surface of the dense layer to the interface with the porous
support. The average atomic fractions of strontium, cerium and
ytterbium at each location were normalised to the values
obtained at the top surface of the dense layer. Non-uniformities
may also arise from differential settling near the walls of the
settling chambers. The examination of the SEM images did not
detect any evidence of variation across the asymmetric membranes.
646
M. Matsuka et al. / Ceramics International 36 (2010) 643–651
3. Results and discussion
3.1. Precursor powder synthesis
The compositions of the SCYb precursor powder
were determined by ICP-MS and the results are approximately in accordance with the desired compositions of
SrCe0.95Yb0.05O3a. Therefore the chemical formula of
SrCe0.95Yb0.05O3a is used to represent the composition of
the precursor powder utilised in this study, assuming the
valencies of Sr2+, Ce4+ and Yb3+ and no cation vacancies. The
particle size distribution of the SCYb powder and cornstarch
particles was analysed with the volume based particle size
analyser and is shown in Fig. 1. The particle size distributions of
the SCYb powders and the cornstarch particles were
determined as approximately 0.1–70 and 0.2–40 mm, respectively. The volume median diameters (the diameter where 50%
of the distribution is above and 50% is below) are about 11 mm
for the SCYb powders and around 18 mm for the cornstarch
particles. The XRD patterns of the SCYb powder exhibit
perovskite structures as shown in Fig. 2.
3.2. Sintering process
The sintering process plays an important role in the
fabrication of dense ceramic membranes, since it influences
not only the microstructures and mechanical strength but also
the transport properties of the membranes. The sintering
process generally consists of distinct stages; densification/
consolidation, removal of the pores between starting particles,
shrinkage of the components, and grain growth. It is essential to
carefully control these sintering processes, as remaining pores
can behave like grains of different materials, affecting the
uniformity and the transport properties of the membranes.
Fig. 3(A) and (B) shows the surface morphology of the dense
top layer of asymmetric membranes sintered at 1465 and
1500 8C, respectively. Comparing the membrane surfaces in
Fig. 3(A1) and (B1), it can be seen that there is a significant
difference in the extent of the sintering processes in the
samples. The sample sintered at 1465 8C (Fig. 3(A1) and (A2))
Fig. 1. Particle size distributions of the precursor SCYb powder and cornstarch
particles.
has undergone the densification process and has a smooth
surface with no distinct individual particles remaining,
however, there are still many pores remaining on the surface.
On the other hand, the sample sintered at 1500 8C (Fig. 3(B1)
and (B2)) has a much denser and smoother surface with little
remaining pores and consists of bigger grains. This indicates
that the pore closing process and further grain growth occurred
between 1465 and 1500 8C. This agrees with Liu and Li [18]
who reported that fully dense membrane surface was obtained
at sintering temperature of 1500–1600 8C for their SCYb
samples.
3.3. Porous support fabrication
The porous supports with the different cornstarch contents of
10, 20, and 30 wt%, referred to as porous support A, B and C,
respectively, were prepared in order to demonstrate the porosity
control with the use of cornstarch particles. Table 1 summarises
the densities and porosities of the porous supports A, B and C.
The true densities of the supports A, B and C are 5.3, 5.5 and
5.4 g/cm3, respectively, which are on average approximately
93% theoretical density of SCYb (5.8 g/cm3). Bulk densities of
the supports are proportional to the cornstarch concentrations in
the supports. The pre-sintering porosities of the green porous
supports A, B and C and the true porosities (including both
Fig. 2. XRD patterns of the SCYb precursor powder.
647
M. Matsuka et al. / Ceramics International 36 (2010) 643–651
Fig. 3. SEM photographs of surface morphology of the dense top layer of the asymmetric membranes sintered at (A) 1465 8C and (B) 1500 8C (1 and 2 correspond to
the 3000 and 5000 magnifications, respectively).
interconnected and closed pores) of the sintered supports A, B
and C indicate that 1.7%, 12%, and 19% of the pores were lost
during the sintering process, respectively. The apparent
(interconnected) porosities of the sintered porous supports
suggest that approximately 25% of the pores in the support A
are closed pores, on the other hand, only 10% for the supports B
and C. This indicates the support B and C (containing 20 and
30 wt% cornstarch particles, respectively) would likely ensure
easy gas transport through the supports.
3.4. Asymmetric membrane fabrication
Asymmetric membranes were prepared by deposition of thin
dense SCYb layers on the green porous SCYb supports by the
settlement method, followed by conventional dry pressing and
co-sintering. Membrane microstructures of cross-sections of
co-sintered asymmetric membranes were observed using SEM
[8]. Fig. 4(A)–(C) shows the SEM cross-sectional images of the
asymmetric membranes with different dense layer thicknesses
Table 1
Density and porosity of porous supports.
Support no.
A
Pre-sintering porosity
29.3 vol% (10 wt% cornstarch)
48.4 vol% (20 wt% cornstarch)
61.6 vol% (30 wt% cornstarch)
Post-sintering
Average
Standard deviation
Average
Standard deviation
Average
Standard deviation
5.28
4.20
27.7%
20.7%
0.25
1.7%
0.029
0.06
1.1%
1.2%
0.020
–
5.51
3.67
36.8%
33.6%
0.09
11.6%
0.025
0.05
0.8%
0.8%
0.007
–
5.40
3.35
42.3%
38.1%
0.1
19.4%
0.035
0.06
1.1%
1.4%
0.012
–
3
True density (g/cm )
Bulk density (g/cm3)
True porosity (vol%)
Apparent porosity (vol%)
Fraction pores closed
Porosity lost during sintering (%)
B
C
648
M. Matsuka et al. / Ceramics International 36 (2010) 643–651
Fig. 4. SEM photographs of cross-sections of asymmetric membranes sintered at 1500 8C, showing different dense layer thickness: (A) 65 mm, (B) 40 mm and (C)
30 mm (1 and 2 correspond to the 500 and 1500 magnifications, respectively).
M. Matsuka et al. / Ceramics International 36 (2010) 643–651
649
Fig. 5. Atomic fractions of strontium, cerium, and ytterbium along the vertical cross-sections of the settled dense layers as evaluated from EDX.
of 65, 40 and 30 mm, respectively. From these figures, it is
apparent that thin dense SCYb films were successfully prepared
on the porous SCYb supports by the settlement method.
The different dense layer thicknesses were easily obtained
by altering the amount of the SCYb powders suspended in the
liquid media. The use of the same ceramic material in the dense
layer and green porous support and co-sintering have been
demonstrated as effective in producing crack-free asymmetric
membranes, as shown in Fig. 4. The dense layer and the porous
support in the asymmetric system densified at the same, or at a
similar, rate during the sintering and there was no reaction
between the two layers. As can be seen in Fig. 4, the
asymmetric systems are distinctive and the transition from the
dense membranes to porous supports does not show any signs of
delamination. The creation and control of the pores in the
support have also been effective through the use of cornstarch
particles. Cornstarch particles did not interfere with the
sintering behaviour of the asymmetric system, and the cosintering of the green asymmetric membranes resulted in no
delamination or cracks.
The downside of the settlement method may be the nonuniformity of the particle settlement, due to the fact that the
particle settlement depends on the density and diameters of
settling particles. This possibly leads to non-uniform microstructures of the dense layers. However, these layers in
Fig. 4(A)–(C) shows uniform thicknesses at this level of
analysis, homogenous microstructures, and excellent adhesion
to the porous supports with no signs of undesirable defects,
cracks or delamination.
The elemental compositions across the vertical cross-section
of the deposited dense layers were examined with EDX, in
order to determine the uniformity of the SCYb powder
distributions. The normalised atomic fractions of strontium,
cerium, and ytterbium along the vertical cross-sections are
displayed in Fig. 5, for the 40 mm layer. Atomic fractions at
each section of the dense layers, for each element, are
normalised to one at the top surface of the dense layer. The
result indicates consistency in the settled dense layers, with a
standard deviation of 5% for strontium, 6% for cerium and 9%
for ytterbium. This suggests the overall uniform distribution of
the SCYb powders across the deposited dense layers.
Fig. 6. SEM photographs of cross-sections of asymmetric membranes sintered
at 1500 8C, showing the minimum thickness (20 mm) (shown at two magnifications, 500 and 1000 magnifications).
650
M. Matsuka et al. / Ceramics International 36 (2010) 643–651
Fig. 7. SEM photographs of surface morphology of the thinner dense top layer (7–18 mm, below the minimum thickness), sintered at 1500 8C (shown at two
magnifications, 1500 and 3000 magnifications). The arrows show non-uniformities in the dense layer.
The minimum thickness of the dense layer obtained by the
settlement method, with the synthesis conditions utilised in
this study, is around 20 mm, as shown in Fig. 6. From the figure,
it is apparent that the thin 20 mm dense layer is still distinctive
and has a uniform thickness at this level of analysis with no
delamination or cracks on the interface. Further, the SEM
cross-sectional images of the edges of the asymmetric
membranes, indicates that there is no evidence of edge effects
from the settling process. Although the SCYb powder utilised
in this study included particles with sizes up to 70 mm, the thin
20 mm dense layer was successfully prepared on the porous
support by the settlement method, conventional dry pressing
and sintering.
By further reducing the amount of SCYb powders in the
liquid media, thinner dense layers (7–18 mm) were prepared on
the porous supports. However the 7–18 mm dense layers
showed less uniform thickness and occasional gaps on the
surface of the top dense layers, as shown in Fig. 7. The surface
morphologies shown in Fig. 7 indicate that most of the surface
is dense and smooth, however there are apparent gaps on the
surfaces due to an insufficient amount of SCYb powder being
deposited on the surface. Therefore, it is concluded that the
minimum thickness which can be prepared by the settlement
method with the synthesis conditions utilised in this study is
approximately 20 mm.
The gas tightness of the dense layers in the asymmetric
membrane samples was determined by helium gas permeation
at room temperature using a bubble test and the results are
shown in Table 2. It should be noted that Table 2 does not
include the result for the samples sintered at 1465 8C (i.e. group
A), as all of the samples sintered at 1465 8C are not helium gas
tight, regardless of the dense layer thicknesses. This agrees with
the findings in Section 3.2. The visual inspection of the surface
morphology of the dense top layer of asymmetric membranes
sintered at 1465 and 1500 8C (Fig. 3(A) and (B), respectively)
indicates that the pore closing process and further grain growth
occurred between 1465 and 1500 8C and that a sintering
temperature of at least 1500 8C is required to obtain fully dense
membrane surface for SCYb samples. This confirms the results
of Liu and Li [18].
Although there are some variations in the results shown in
Table 2, it can generally be concluded that thicknesses greater
than 30 and less than 20, are not helium gas tight. Dense layers
thinner than 20 mm are not helium gas tight, which can be
explained by the findings in Fig. 7. Fig. 7 shows the surface
morphologies of the 7–18 mm-thick dense layers where there
are apparent gaps on the surfaces due to insufficient amounts of
SCYb powders deposited on the surface. Those gaps obviously
result in a lack of gas tightness of the dense layers.
The dense layers which are thicker than 30 mm, are also
found to not gas tight for helium. This may be seen in Fig. 4(A)–
(C), which show the SEM cross-sectional images of the
Table 2
Helium gas tight test results for group B (samples sintered at 1500 8C).
Sample no.
Areal concentration
(g/cm2)
Estimated dense
layer thickness (mm) a
Helium gas
tightness
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
0.0566
0.0448
0.0286
0.0285
0.0283
0.0251
0.0212
0.0197
0.0170
0.0157
0.0142
0.0141
0.0143
0.0129
0.0129
0.0114
0.0114
80.7
65.0b
41.3b
41.2b
41.5
36.4
30.9
30.4
24.9
20.6
20.2
20.1b
20.5
20.5
18.3b
16.2b
16.1b
No
No
No
No
No
No
Tight
No
Tight
Tight
Tight
Tight
Tight
No
Tight
No
No
a
The thicknesses of the dense layers were estimated either from the SEM
images when the dense layers in the SEM images are clear.
b
Otherwise the thicknesses were estimated from the equation for the trendline in Fig. 8, y = 1463.5x 0.5415, where x is the areal concentration (g/cm2).
M. Matsuka et al. / Ceramics International 36 (2010) 643–651
651
require pre-sintering of supports or sequences of coatings, and
hence is less labour consuming; (3) the liquid media can be any
liquid which does not interfere or react with the powder; (4) the
method requires simple mixing of powder and the liquid media,
and does not require any additions such as dispersant; and (5)
the method includes only a few process variables making it less
complex and suitable for a larger scale application. So in
conclusion, this study has demonstrated settlement methods as
the effective filling technique which aids the conventional dry
pressing, in order to prepare a thin uniform dense layer on a
porous support.
Fig. 8. Relationship between the concentrations of the SCYb powder in the
liquid media and the corresponding dense layer thicknesses in the sintered
asymmetric membranes.
asymmetric membranes with different dense layer thicknesses
of 65, 40 and 30 mm, respectively. The 65 mm- and 40 mmthick dense layers (Fig. 4(A) and (B), respectively) contain
more pores than the 30 mm-thick dense layer (Fig. 4(C)). It is
suggested that the non-gas tightness may result from
insufficient sintering temperature or time for the gas seal to
be attained for the thicker samples.
Fig. 8 summarises the thicknesses of the dense thin layers
prepared on the porous supports by the settlement method. It
plots the areal concentration (g/cm2) of the SCYb powders in
the liquid media against the corresponding thickness of the
dense layer on the porous support obtained after sintering. It
should be noted that the areal concentration, rather than
volumetric concentration (g/cm3), is utilised in Fig. 8. This
metric is used because the reason that the critical factor is the
amount of SCYb powder in the liquid media per unit area of the
support (for example, 1 cm3 of 1 g/cm3-solution provides 1 g of
SCYb powder for settling on the support, while 2 cm3 of 0.5 g/
cm3-solution also provides 1 g of SCYb powder). The
coefficient of determination (R2) value is 0.99 which shows
an excellent fit. This suggests the effective control of the dense
layer thickness by this method and the layers so generated are
readily reproducible. This study has verified that settlement
methods are an effective filling technique which aids the
conventional dry pressing, in order to prepare a thin uniform
dense layer on a porous support.
4. Conclusions
Thin dense SCYb films were successfully prepared on the
porous SCYb supports by the settlement method. These dense
layers have uniform thicknesses at this level of analysis, and
show excellent adhesion to the porous support without any
undesirable defects or crack. The settlement method provides a
uniform distribution across the membranes without any major
edge effects on the deposited surfaces. Further, different dense
layer thicknesses were easily controlled by altering the amount
of the SCYb powders in the liquid media. The settlement
method can provide the advantages that (1) the control of the
dense layer thickness is very easy and effective; (2) it does not
References
[1] Australian Greenhouse Office (Department of the Environment and
Heritage, Ed.), Commonwealth of Australia, 2006.
[2] S. Adhikari, S. Fernando, Hydrogen membrane separation techniques, Ind.
Eng. Chem. Res. 45 (3) (2006) 875–881.
[3] J.W. Phair, S.P.S. Badwal, Review of proton conductors for hydrogen
separation, Ionics 12 (2) (2006) 103–115.
[4] S.E. Roark, R. Mackay, A.F. Sammells, Vision 21 Program Review
Meeting, Eltron Research, Inc., National Energy Technology Laboratory,
2001.
[5] S.J. Song, E.D. Wachsman, J. Rhodes, S.E. Dorris, U. Balachandran,
Hydrogen permeability of SrCe1xMxO3-[delta] (x = 0.05, M = Eu, Sm),
Solid-State Ionics 167 (1–2) (2004) 99–105.
[6] S. Hamakawa, L. Li, A. Li, E. Iglesia, Synthesis and hydrogen permeation
properties of membranes based on dense SrCe0.95Yb0.05O3-[alpha] thin
films, Solid-State Ionics 148 (1–2) (2002) 71–81.
[7] X. Ma, J. Dai, H. Zhang, D.E. Reisner, Protonic conductivity nanostructured ceramic film with improved resistance to carbon dioxide at elevated
temperatures, Surf. Coat. Technol. 200 (5–6) (2005) 1252–1258.
[8] G. Etchegoyen, T. Chartier, P. Del-Gallo, An architectural approach to the
oxygen permeability of a La0.6Sr0.4Fe0.9Ga0.1O3-[delta] perovskite membrane, J. Eur. Ceram. Soc. 26 (13) (2006) 2807–2815.
[9] L. Hong, X. Chen, Z. Cao, Preparation of a perovskite La0.2Sr0.8CoO3x
membrane on a porous MgO substrate, J. Eur. Ceram. Soc. 21 (12) (2001)
2207–2215.
[10] J.H. Joo, G.M. Choi, Electrical conductivity of YSZ film grown by pulsed
laser deposition, Solid-State Ionics 177 (11–12) (2006) 1053–1057.
[11] C. Xia, M. Liu, A simple and cost-effective approach to fabrication of
dense ceramic membranes on porous substrates, J. Am. Ceram. Soc. 84 (8)
(2001) 1903–1905.
[12] S. Cheng, V.K. Gupta, J.Y.S. Lin, Synthesis and hydrogen permeation
properties of asymmetric proton-conducting ceramic membranes, SolidState Ionics 176 (35–36) (2005) 2653–2662.
[13] W. Jin, S. Li, P. Huang, N. Xu, J. Shi, Preparation of an asymmetric
perovskite-type membrane and its oxygen permeability, J. Membr. Sci.
185 (2) (2001) 237–243.
[14] I.E. Agranovski, A.Y. Ilyushechkin, I.S. Altman, T.E. Bostrom, M. Choi,
Methods of introduction of MgO nanoparticles into Bi–2212/Ag tapes,
Physica C: Superconduct. 434 (1) (2006) 115–120.
[15] F.L. Chen, O.T. Sorensen, G.Y. Meng, D.K. Peng, Preparation of Nddoped BaCeO3 proton-conducting ceramic and its electrical properties in
different atmospheres, J. Eur. Ceram. Soc. 18 (10) (1998) 1389–1395.
[16] S.W. Hughes, Archimedes revisited: a faster, better, cheaper method of
accurately measuring the volume of small objects, Phys. Educ. 40 (2005)
468–474.
[17] D.R. Askeland, P.P. Phule, The Science and Engineering of Materials,
Thomson, Australia, 2006.
[18] Y. Liu, K. Li, Preparation of SrCe0.95Yb0.05O3-[alpha] hollow fibre
membranes: study on sintering processes, J. Membr. Sci. 259 (1–2)
(2005) 47–54.