Ceramics International 47 (2021) 17768–17777

Contents lists available at ScienceDirect

Ceramics International
journal homepage: www.elsevier.com/locate/ceramint

Ce0.9Gd0.1O2-δ-based multi-phase membranes with high CO2-tolerance

Yuliang Liu, Martin van Sint Annaland, Fausto Gallucci *
Inorganic Membranes and Membrane Reactors, Sustainable Process Engineering, Department of Chemical Engineering and Chemistry, Eindhoven University of
Technology, the Netherlands

A R T I C L E I N F O A B S T R A C T

Keywords: In this work, the oxygen ionic conductor Ce0.9Gd0.1O2-δ was paired with the electronic/mixed conductive oxides,
Oxygen permeable membrane including NiFe2O4 (NFO), La2NiO4+δ (LNO) and Ba0⋅5Sr0⋅5Co0⋅8Fe0⋅2O3-δ (BSCF), forming three kinds of dual-
Dual-phase membrane phase membranes. The performance of the membranes was compared in terms of oxygen flux and CO2 toler­
Triple-phase membrane
ance. The CGO/BSCF provided the highest permeation flux among the three composites with either helium or
CO2-Tolerance
CO2 fed as the sweep gas, although a sharp decrease of the permeation value was observed for the CGO/BSCF
when CO2 was in contact with the membrane. Based on the results for the dual-phase membranes, some CGO/
BSCF/NFO triple-phase membranes were developed. The triple-phase membranes were found be less sensitive to
the presence of CO2 in the sweep gas than the CGO/BSCF, and the application of the two triple-phase composites
instead of the dual-phase ones promoted the oxygen permeation fluxes when the membranes were swept with
pure CO2. The results provide valuable information for the material selection and performance optimization of
CGO-based multi-phase membranes for oxygen permeation.

1. Introduction remains a challenge for the industrial application of the membranes.

Reduction of anthropogenic greenhouse gas emissions to the atmo­
sphere can be realized by carbon capture from fossil-fuel power plants
and from other energy intensive plants such as steel, iron or cement
facilities. Oxyfuel combustion, i.e. using oxygen instead of air to
combust the fuels, represents an efficient way for carbon capture as it
enables the production of highly CO2 enriched exhaust gas due to the
absence of N2 in the combustion, which allows a more direct CO2 sep­
aration process [1]. Industrial oxygen is generally supplied by cryogenic
distillation and pressure swing adsorption, which both require high in­
vestments and operating costs. As a potential alternative, air separation
using ceramic membranes with mixed electronic-ionic conductivity
gained great interest in these years [2–6]. With efficient process inte­
gration, lower energy cost can be achieved by the membrane-assistant
air separation compared to most of the conventional oxygen produc­
tion processes. Additionally, membrane separation can be easily scaled
down to smaller scales compared to cryogenic distillation. The operation
conditions of the membranes include oxygen partial-pressure difference,
high temperature and in many cases the exposure to CO2-containing
exhaust gases, which means that, apart from high permeability, good
operation stability is also necessary for an oxygen transport membrane
[7–9]. Identifying suitable materials with all the desired characteristics
Numerous single-phase materials has been studied since the pio­
neering application of mixed-conducting oxides to air separation, and
several perovskite materials, such as Sr(Co,Fe)O3-δ, (Ba,Sr) (Co,Fe)O3-δ
and (La,Sr) (Co,Fe)O3-δ, are found to present superior oxygen perme­
ability. However, irreversible permeation losses were observed for the
alkaline-earth-metal-based perovskites when the membranes were
exposed to a CO2-containing atmosphere [10–15]. The threshold be­
tween permeability and stability of single-phase materials motivates the
application of dual-phase membranes to achieve combined good
CO2-tolerance and high permeation fluxes. For this composite mem­
brane material, one phase should allow the percolation of oxygen ions
while the other phase acts as the conductor of electrons. Similar idea has
already been tested in the early stage of membrane development, which
is known as the cermet membrane [16–20]. Precious metals provided
the electron conducting path in the cermet membranes, while ceramics,
such as stabilized bismuth oxide (SBO) and yttria-stabilized zirconia
(YSZ), served as a pure oxygen ionic-conductive phase. No less than 28
vol% precious metal is normally required in those membranes to form a
continuous path for the diffusion of electrons [21]. The high cost of the
precious metals, together with the chemical incompatibility of the two
phases and the relatively low oxygen permeability hampered applica­
tion of the cermet membranes. Cercer (ceramic-ceramic) composites

* Corresponding author. Address: P.O. Box 513, STW 1MB, Eindhoven, the Netherlands.
E-mail address: F.Gallucci@tue.nl (F. Gallucci).

https://doi.org/10.1016/j.ceramint.2021.03.099
Received 30 December 2020; Received in revised form 26 February 2021; Accepted 10 March 2021
Available online 13 March 2021
0272-8842/© 2021 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Y. Liu et al.
Fig. 1. Oxygen-ionic and electronic percolation in the (a, b) dual-phase membranes and (c) triple-phase membrane.
were used to prepare the dual-phase membranes since the early 2000’s
[22]. Application of the electronic conducting ceramics instead of the
precious metals may reduce the price of the membrane production.
For a cercer dual-phase membrane, compatible materials need to be
selected to avoid reactions between the two phases and to minimize the
chemical-thermal expansion mismatch between both ceramics. Several
combinations of crystalline materials, including fluorite, perovskite,
spinel and Ruddlesden–Popper-type oxides, have been reported in the
literature-. The materials with a fluorite crystalline structure are
generally applied as the oxygen-ionic phase in the composites since the
ionic conductivity is dominant, and the doped ceria is a popular fluorite
oxide which shows ultra-high chemical stability and mechanical
strength. The spinel, Ruddlesden–Popper-type oxides and the perov­
skites can serve as the electronic phase. Because of the much higher
mobility of electrons than the oxygen ions, an increase in the oxygen-
ionic conductivity of the membrane may result in an improvement of
the total oxygen permeation flux. Following this strategy, the mixed
conducting perovskites instead of the pure electronic conductors were
combined with the fluorites by some researchers for the fabrication of
dual-phase membranes [23–31]. Such composites ensure that both
phases can provide the paths for the oxygen-ionic percolation, as illus­
trated in Fig. 1. A fluorite-perovskite membrane was fabricated by Xue
et al. [27] using Ce0.9Gd0.1O2-δ(CGO) and Ba0⋅5Sr0⋅5Co0⋅8Fe0⋅2O3-δ
(BSCF) powders, which exhibited an oxygen permeation flux of 2.1
mL/(min⋅cm2) at 975 ◦ C with a membrane thickness of 0.5 mm. Kim
et al. [25] combined the CGO with another perovskite material
Sr0⋅95Fe0⋅5Co0⋅5O3-δ (SFC) and found good compatibility of the two
phases. Moreover, the SFC phase in this composite showed high stability
when exposed to the same condition where phase transformation was
observed for the single-phase SFC membranes. CGO was also paired with
La0⋅6Sr0⋅4Co0⋅2Fe0⋅8O3− δ (LSCF) by Nam et al. [32]. The resulted
dual-phase membrane was coated with SrCo0⋅1Fe0⋅8Nb0⋅1O3− δ (SCFN)
active layer and a high permeation flux of 10.41 ml/min⋅cm2 was
observed for this membrane at 1000 ◦ C. Besides, a group of
Ce0.9Ln0.1O2-δ-Ln0.6Ca0⋅4FeO3-δ (Ln = La, Pr, Nd, Sm) membranes were
fabricated and compared by Wang et al. [33] and the
Ce0.9Pr0.1O2-δ-Pr0.6Ca0⋅4FeO3-δ composite was found to provide superior
oxygen permeability and CO2 tolerance among the tested membranes.
However, the presence of alkaline-earth-metal-containing perovskites in
those membranes would cause the formation of carbonate and imme­
diate reduction of the oxygen permeation when the cercer membranes
was exposed to CO2, although this reduction was found to be reversable.
Therefore, CO2-stable electronic conductors are preferred by some re­
searchers. For example, several fluorite-spinel composites, such as
CGO-NiFe2O4 (NFO), Ce0⋅8Tb0⋅2O2-δ (CTO)-NFO and Mn1⋅5Co1⋅5O4-δ
(MCO)-Ce0.9Pr0.1O2-δ (CPO) membranes were tested in CO2-containing
atmospheres and excellent results were obtained in terms of the mem­
brane stability [34–36]. Besides, a Ruddlesden–Popper-type oxide of
La2NiO4+δ was combined with CGO by Han and his colleagues [37] in
17769
Ceramics International 47 (2021) 17768–17777
order to form a robust membrane with high CO2 resistance. However,
the studies of different cercer composites were mostly based on varied
membrane thicknesses, geometries (tubular, hollow fibre and planar)
and surface modifications, and were tested under different test condi­
tions, which bring difficulties to the comparation between the perfor­
mance of the fluorite-perovskites and other dual-phase composites with
pure electronic and ionic conductors.
Selection of proper material is the initial step for the production of
highly performed membranes. In this work, the strategies to composite
the cercer membranes with CGO as the oxygen-ionic conductor are
further investigated. The two strategies, using mixed-conductive/ionic
conductors and pure electronic/ionic conductors to composite cercer
membranes are compared at the same membrane dimensions and test
conditions, with the purpose to provide a reference for the material
selection of multi-phase membranes. A modified phase-inversion casting
method, which has been reported in our previous work [38], was
applied for the production of dead-end tubular membranes. This mem­
brane geometry allows easy sealing and handling in the reactors. The
permeability of the fabricated CGO-BSCF, CGO-NFO and CGO-LNO
membranes is tested and compared in detail, together with an investi­
gation in their permeation behavior in a CO2 atmosphere. Based on the
test results of the dual-phase membranes, some CGO/BSCF/NFO
triple-phase membranes are developed (Fig. 1c), with the purpose of
optimizing the membrane performance in CO2-containing atmospheres.
A large amount of each phase is required for the formation of continuous
ionic/electronic paths in the dual-phase membranes. However, in the
triple-phase membranes, the concentration of BSCF can be reduced
while maintaining a continuous electronic network by adding the elec­
tronic conductor NFO. The oxygen permeability and CO2-tolerance of
the triple-phase membranes are compared with that of the CGO-BSCF
and CGO-NFO membranes.
2. Experimental procedure
Commercially available powders were applied to prepare the multi-
phase membranes. CGO powders purchased from Ningbo Sofcman En­
ergy were milled for 3 h in a ball mill (Pulverisette 6, Fritsch) before
being used as the oxygen-ionic conductive phase. The NFO powders for
the electronic conductive phase was supplied by Alfa Aesar, while the
BSCF and LNO powders were ordered from Cerpotech. Polymers in the
casting slurries for phase inversion process include polyethersulfone
(PES, Veradel 3000P) from Solvay Advanced Polymers and Poly­
vinylpyrrolidone (PVP) from Sigma-Aldrich, and the solvent was 1-
methyl-2-pyrrolidone (NMP, 99.5%) ordered from Sigma-Aldrich.
The fabrication of the dead-end membranes has been reported in our
previous work [38], which started with the preparation of a solution by
adding 5 g PES and 1 g PVP in to 20 g NMP and mixing with a centrifugal
mixer (ARE-250, Thinky) until the full dissolution of the polymers.
Subsequently, ceramic powders were mixed with the solution for 90 min
Y. Liu et al. Ceramics International 47 (2021) 17768–17777

Table 1
Weight percentage of each phase and sintering temperatures for the multi-phase membranes under study.
Membrane CGO concentration, wt% BSCF concentration, wt% NFO concentration, wt% LNO concentration, wt% Sintering temperature, ◦ C

C6L4 60 0 0 40 1400
C6N4 60 0 40 0 1450
C6B4 60 40 0 0 1170
C4B4N2 40 40 20 0 1235
C6B2N2 60 20 20 0 1245

Fig. 2. Temperature dependence of the permeation fluxes through the C6L4,
C6N4 and C6B4 dual-phase membranes, together with the corresponding acti­
vation energies.
in the same mixer. The total amount of the ceramic powders was set at
45 g for each 20 g of the NMP, while the components were varied for
different multi-phase membranes, as listed in Table 1. The formed
ceramic slurries were defoamed by the mixer to remove the bubbles
before transferred to the home-made casting device, and the device
together with the mixtures inside were cooled with ice for 30 min to
increase the slurry viscosity. Then a piston was introduced into the
slurry and push the slurry into water with the cap of the device taken off.
A low slurry fluidity and a high pushing speed of the piston are impor­
tant for obtaining the dead-end membrane shape. After the slurries got
fully solidified in the water bath, they were dried in the atmosphere for
48 h with removed from the piston. The green membrane tubes were
transferred into a horizontal tubular oven to be sintered at ultra-high
temperature. Those green tubes with varied composition required
different sintering temperatures (also listed in Table 1) to achieve
gas-tight membranes with good mechanical stability, and the optimized
sintering temperature for each kind of membrane has been determined
in a separate study. A slow heating ramp of 30 ◦ C/h was used to heat the
membranes up to the sintering temperatures and were subsequently
cooled down with a higher ramp of 60 ◦ C/h. The outer diameters of the
sintered membranes are ~8 mm and the membrane thicknesses are
~1.2 mm, while length of the membranes can be adjusted by dimension
of the casting tool.
The multi-phase membranes before or after oxygen permeation were
investigated by SEM (Quanta 3D FEG), while the element analysis was
performed with EDX. The phase structure of the membranes was
investigated via XRD (Rigaku MiniFlex 600). Gas tightness of the multi-
phase membranes after sintering process was tested in a helium setup
and only leakage-free membranes were selected for the permeation tests.
A high-temperature experimental system, which was described in our
previous work [38], was employed to measure the oxygen permeation
fluxes through the different multi-phase membranes. The open side of
the dead-end membrane was sealed to a Fe–Cr tube with the Ag–CuO
braze paste at 1020 ◦ C. The permeation cell was heated to 1020 ◦ C with
a rate of 60 ◦ C/h, kept for 30 min, and directly cooled down to 825 ◦ C for
the permeation test after the sealing procedure. He and/or CO2 were
used as sweep gas for the permeation test with air applied as feed gas.
The sweep gas was fed inside the membrane tubes with digital mass flow
controllers, while the outlet was delivered to a micro-GC (Varian,
CP-4900) for the measurement of the concentration. A film flow meter
(Horiba STEC, SF) was used to calibrate the mass flow controllers and to

Fig. 3. Permeation fluxes through the C6L4, C6N4 and C6B4 membranes with the sweep gas alternated between pure He and CO2 at (a) 850 ◦ C and (b) 950 ◦ C.

17770
Y. Liu et al.
Fig. 4. Permeation fluxes through the C6L4, C6N4 and C6B4 membranes (a) with sweep gas of different CO2 concentrations at 950 ◦ C, and (b) with pure CO2 feed as
the sweep gas at different temperatures.
measure the flowrate from permeate side.
3. Results and discussion
3.1. Oxygen permeation and CO2-tolerance of the dual-phase membranes
Measurements for the oxygen permeation as a function of operation
temperature were performed on the dual-phase membranes with fixed
sweep gas flowrate of 500 ml/min. The performance of three types of
membranes including C6L4, C6N4 and C6B4 were compared. For the
ceramic oxygen transport membranes, it was generally found that
increased operation temperatures led to a rise in the permeation fluxes,
which is because of the enhanced surface-exchange and bulk-diffusion
rate of oxygen. This trend also applies to these three dual-phase com­
posites, as shown in Fig. 2. Of the three membranes, it can be observed
that the C6B4 provided a much higher permeation rate than the other
two composites. Taking the data at 950 ◦ C as a reference, the permeation
flux through the C6B4 membrane was 0.29 ml/min⋅cm2, which is more
than 6 times the value obtained for the C6N4 membrane and 12 times
the value achieved for the C6L4 membrane at the same conditions.
These results prove the promoting effect of using mixed-conductive
materials instead of the pure electronic conductors in the dual-phase
membranes on the oxygen ion diffusion through the membrane bulk,
which results from the increased paths for oxygen ion percolation. The
activation energies (Ea) for oxygen permeation through the three
membrane samples are also included in Fig. 2. Constant activation en­
ergies were observed for all of the three dual-phase membranes, and it is
one of the factors that proves the good phase stability of the membranes
under the tested conditions [39,40]. The highest activation energy is
obtained for the C6B4 membrane which is 120.1 kJ/mol. This could be
attributed to two factors: Since both CGO and BSCF provide paths for the
oxygen ion conduction, hindered travelling of the oxygen ions through
the borders of the two phases may contribute to the high activation
energy. Besides, the sintering temperature of 1170 ◦ C for the composite
is lower than the desired sintering temperature of the CGO phase
(although a gas-tight membrane was obtained because of the presence of
BSCF), and the increased grain boundary area of CGO due to the lower
sintering temperature may have led to a higher activation energy for the
oxygen permeation [37].
The chemical stability of the three dual-phase membranes was
studied by feeding a CO2-containing sweep gas to the permeate side of
the membranes at different temperatures. The flow rate of the sweep gas
17771
Ceramics International 47 (2021) 17768–17777
was fixed at 500 ml/min. Fig. 3a reports the permeation evolution for
the three composite membranes as a function of time with the sweep gas
switched between 100% CO2 and He at 850 ◦ C. The experiment started
with pure helium fed to the permeate side, and after the permeation had
become stable for 4 h in helium, CO2 was feed instead of helium as
sweep gas. It can be observed that the permeation flux through the C6B4
membrane dropped instantly to 0.048 ml/min⋅cm2 when the sweep gas
was switched to pure CO2. Afterwards, a gradual decline was observed
and then the flux tended to be stable at ~0.032 ml/min⋅cm2. The
O2–CO2 competitive adsorption on the membranes and the carbonate
formation due to the reactions was found to be the two factors that may
have caused the permeation loss of the membranes in CO2 [41]. The
sharp decline of the oxygen permeation may be mainly attributed to the
competitive adsorption of the two gas components, while the gradual
drop afterwards may have been caused by the reaction forming the
carbonate, since the carbonate-formation process would take a longer
time. The carbonates such as BaCO3 and SrCO3 were reported to be
decomposed to oxides when exposed to a CO2-rare atmosphere at tem­
peratures above 800 ◦ C followed by the possible recovery of the
perovskite structure due to the reactions of the solid-state oxides [42,
43]. This explains the recovery of the permeation flux through the C6B4
when He was switched back as sweep gas. The performance loss with the
exposure to a CO2-containing atmosphere was also observed for some
other membranes (not reported here) e.g. the permeation value dropped
to zero when the BSCF membrane was swept with CO2. The compara­
tively high oxygen flux through the C6B4 membrane when exposed to
CO2 proves the enhancing role of CGO on the CO2-resistance of the
membrane. Fig. 3a also shows that the CO2 atmosphere brought little
poisoning effect on the C6N4 and C6L4 membranes. However, it is
noticeable that although a large decrease in the oxygen flux was
observed for the C6B4 membrane, the stable permeation value was still
higher than that of the other two composites. Similar results were ob­
tained at 950 ◦ C (Fig. 3b). The permeation performance of the three
dual-phase membranes all reached a steady state in a CO2 atmosphere
and the C6B4 composite provided the highest oxygen flux when swept
with either He or pure CO2.
The effect of the CO2 concentration in the sweep gas on the oxygen
permeation flux through the three dual-phase membranes with the
operation temperature fixed at 950 ◦ C is presented in Fig. 4a. The
permeation value was observed to decline with the rise of CO2 content in
the sweep gas for the C6B4 membrane, while the CO2 caused little
poisoning effect on the C6N4 and C6L4 membranes. This result can be
Y. Liu et al. Ceramics International 47 (2021) 17768–17777

Fig. 5. (a) Secondary electron microscopy of cross section of C6B2N2 membrane, (b, c) back-scattered electron microscopy and (d–f) EDS analysis of three points on
surface of the C6B2N2 membrane.

explained by the higher concentration of oxygen vacancies on the sur­
face of the C6B4 membrane compared to the other two membranes,
which means that more CO2 could be adsorbed on the C6B4 composite.
However, it should be noticed that, at each CO2 concentration, C6B4
provided the highest oxygen permeation rate among the three
composites.
Fig. 4b shows the temperature effect on the permeability of the three
composites when 100% CO2 was used as the sweep gas. It can be
observed that, as swept by He, an increase in the operating temperature
caused a permeation flux increase for all the three dual-phase mem­
branes, because of the improved surface exchange and bulk diffusion
rates at higher temperatures. It can also be seen from Fig. 4b that at each
temperature C6B4 provided the highest permeation flux of the three
investigated membranes which further proves the superiority of C6B4 in
terms of the performance in CO2 atmosphere compared with the other
two composites.
3.2. Phase and morphology of the triple-phase membranes
The C6B4 membrane exhibited superior permeability, but also a
larger reduction of the oxygen permeation rate than the C6N4 and C6L4
compounds when swept by CO2. Accordingly, some triple-phase mem­
branes were prepared using the CGO, NFO and the mixed conductive
BSCF oxides. Several types of triple-phase membranes can be obtained

17772
Y. Liu et al. Ceramics International 47 (2021) 17768–17777

by adjusting the ratio between the different phases, and two of these
composites were investigated and compared with the C6B4 and C6N4
dual-phase membranes in this work.
Fig. 5a–c are the SEM images of the C6BN triple-phase membrane.
With the BSD mode of SEM, three kinds of grains can be clearly observed
on the membrane surface. These uniformly distributed grains are sized
between 1 and 4 μm. According to the EDS analysis, the light grey blocks
(point 1) are the CGO phase, the ones in darker grey (point 3) are sup­
posed to be BSCF phase, and the black grains (point 2) can be assigned to
the NFO phase (Fig. 5d–f). The SEM images show a well-connected
network of the CGO phase, while the other two phases appear less
continuous. However, connected electronic pathway can be obtained
due to the collective effect of the BSCF and NFO phases. No new phases
were observed on the C6BN membrane after sintering at ultra-high
temperature, and this was confirmed by the XRD test result (Fig. 6). A
weak signal of Gd was detected on the BSCF grains, which was repeat­
able in the EDX tests. This result can be explained by the trace Gd
diffusion from CGO to the BSCF phase at high sintering temperatures. Gd
of the CGO was also found to diffuse into La0⋅6Sr0⋅4Co0⋅2Fe0⋅8O3-δ during
sintering of the dual-phase membrane by Huan et al. [44]. From the XRD
Fig. 6. XRD pattern of the C6B2N2 membrane (a) before and (b) after the
patterns of the fresh C6BN membranes (Fig. 6), it is obvious that the
permeation test with CO2 as sweep gas.
characteristic peaks of the BSCF phase all have shifted to the left, and

Fig. 7. Helium flowrate dependence of the permeation value for the (a) C6B2N2 and (b) C4B4N2 membranes; (c) Temperature dependence of the permeation fluxes
through the C6N4, C4B4N2, C6B2N2 and C6B4 membranes.

17773
Y. Liu et al. Ceramics International 47 (2021) 17768–17777

Fig. 8. Oxygen fluxes through the C6B2N2 and C4B4N2 membranes with the sweep gas alternated between helium and 100% CO2 at 850 ◦ C.

this can be explained by the large amount of lattice oxygen release
during the sintering process.
3.3. Performance comparation between the triple-phase and dual-phase
membranes
Fig. 7a/b give the oxygen permeation fluxes through the C6B2N2
and C4B4N2 membranes when swept by helium with different flow rates
at temperatures ranging from 825 to 950 ◦ C. A positive correlation was
observed between the permeation value and the flowrate of helium for
both of the membranes e.g. at 950 ◦ C the oxygen flux through the
C6B2N2 and C4B4N2 incteased from 0.13 to 0.15 and from 0.10 to 0.12
ml/min⋅cm2, respectively, when the flow rate of the sweep gas was
increased from 250 to 750 ml/min. The rise in the oxygen permeation
flux can be attributed to the reduced oxygen partial pressure on
permeate side of the membranes. Besides, the permeation was found to
be less sensitive to the helium flowrate increase at higher feeding rate of
the sweep gas.
The temperature dependence of the permeation values for the two
triple-phase membranes together with the C6B4 and C6N4 dual-phase
membranes are shown in Fig. 7c. Apparently, the permeation fluxes of
both triple-phase membranes are higher than that of the C6N4 but lower
when compared with that of the C6B4 membrane at the same
temperature. The activation energies of the four composites are also
given in Fig. 7c. Constant activation energies for the oxygen permeation
through the two triple-phase membranes were observed over the test
temperature range, and similar as their permeation fluxes activation
energies of the C6B2N2 and C4B4N2 membranes are between those of
the two dual-phase composites. The C6B2N2 membrane was observed to
show a higher activation energy than the C4B4N2, which can be
attributed to the low activation energy of the oxygen-ion percolation in
BSCF compared to that of the ionic conduction in CGO [45].
Fig. 8 shows the permeation evolution trend for the C6B2N2 and
C4B4N2 membranes as a function of time on stream with the sweep gas
switched between pure CO2 and He at 850 ◦ C. Similar as the C6B4
membrane, an instant breakdown of the permeability was observed for
both triple-phase membranes when CO2 was fed as sweep gas instead of
helium. However, in the following 24 h, the permeation fluxes through
C6B2N2 and C4B4N2 were found to rise gradually to 0.041 ml/min⋅cm2
and 0.034 ml/min⋅cm2, respectively, as illustrated by curve-2. When
helium was switched back as the sweep gas, the oxygen permeation
through the C6B2N2 membrane restored to 0.092 ml/min⋅cm2 imme­
diately. It is worth noting that the restored permeation value is much
higher than the original level of curve-1, which is 0.047 ml/min⋅cm2.
This upgrade of the oxygen permeation flux can be explained by the
postulation that the formation and decomposition process of the

Fig. 9. SEM images of the C6B2N2 membrane tube: (a) fresh inner surface and (b) inner surface after the permeation test with pure CO2 as sweep gas at 850 ◦ C.

17774
Y. Liu et al.
Fig. 10. Permeation fluxes through the C6N4, C4B4N2, C6B2N2 and C6B4 membranes (a) with sweep gas of different CO2 concentrations at 950 ◦ C and (b) with pure
CO2 feed as the sweep gas at different temperatures.
carbonates on the membrane could increase the roughness of the
membrane surface, and the resulted larger surface area would facilitate
the surface-exchange reactions of oxygen. The postulation is supported
by the SEM images of the C6B2N2 before and after the permeation test in
CO2 atmosphere (Fig. 9). It is obvious that the surface of the membrane
tube which was exposed to the sweep gas became rough and porous after
the exposure to CO2, while a smooth surface was observed for the fresh
C6B2N2 membrane. However, the XRD pattern of the membrane after
exposure to CO2 shows no new phases (Fig. 6), which can be explained
by the decomposition of the carbonate and recovery of the perovskite
structure in the CO2-rare atmosphere and that the trace amounts of the
carbonate is very hard to be detected by XRD.
When CO2 was fed again to the permeate side of the C6B2N2
membrane, the permeation flux instantly dropped to 0.052 ml/min⋅cm2
and then gradually reduced to 0.043 ml/min⋅cm2 (stage-4 in Fig. 8).
Then, as helium was switch back, the permeation value restored to
0.093 ml/min⋅cm2, which is at a similar level of curve-3, and this is
because of the already-formed surface microstructure of the membrane.
A similar trend was observed for the oxygen permeation of C4B4N2
when CO2 and helium was fed alternatively to the permeate side, but the
permeability of the C4B4N2 seems to be less sensitive to the micro­
structure changes, that resulted from the CO2 corrosion, compared to the
C6B2N2 membrane.
The permeation rate of the two triple-phase membranes were
compared with that of the C6B4 and C6N4 membranes at 950 ◦ C by
Fig. 11. Effect of carbonate formation on the dissociation of O2 from the surface of (a) CGO-BSCF membrane and (b) the triple-phase membrane.
17775
Ceramics International 47 (2021) 17768–17777
feeding sweep gas with different CO2 concentrations. As can be seen in
Fig. 10a, although the oxygen fluxes through the C6B2N2 and C4B4N2
also declined with an increased CO2 concentration in the permeate side,
the slops of the permeation rates were much smaller than that of the
C6B4 membrane. For instance, with 25% CO2 in the sweep gas, the
permeation flux through the C6B4 membrane reached 0.23 ml/min⋅cm2,
but reduced by 74% to 0.06 ml/min⋅cm2 when feeding pure CO2 to the
permeate side, whereas the permeation rate of the C6B2N2 membrane
decreased from 0.15 ml/min⋅cm2 to 0.14 ml/min⋅cm2 with the CO2
concentration in the sweep gas rising from 25% to 100%. The applica­
tion of the triple-phase composites instead of the C6B4 and C6N4
improved the oxygen permeation when a high CO2-containing sweep gas
was fed to permeation side of the membranes.
Fig. 10b gives the oxygen permeation flux through the C6N4,
C6B2N2, C4B4N2 and C6B4 membranes with pure CO2 fed as sweep gas
at different temperatures. The permeation rates of the four membranes
all exhibited positive correlation with the temperatures, and the triple-
phase membranes showed a higher permeability than both the C6N4
and C6B4 membranes at each temperature. The higher permeability of
the triple-phase membranes compared to the C6B4 membrane under the
Air/CO2 gradient may be explained by Fig. 11. When the permeate side
of C6C4 membrane was exposed to the sweep gas of CO2, carbonate
formation on the membrane surface would cause a large reduction of the
effective surface area. At this time the surface exchange on the permeate
side is supposed to become the controlling step for oxygen permeation.
Y. Liu et al.
For a dual-phase membrane, the volume fraction of each phase should
be no less than 30% for the formation of continuous ionic/electronic
paths. However, in the triple-phase membranes, amount of BSCF can be
reduced with maintaining the continuous electronic network by adding
the electronic conductor of NFO. Less BSCF exposed to CO2 on the
membrane surface means less carbonate formation and smaller reduc­
tion of the surface exchange speed. In contrast, although the exposure to
a CO2 atmosphere didn’t cause an obvious reduction of the oxygen flux
through the C6N4 membrane, the C6N4 membrane still provided lower
permeability compared to the triple-phase membranes. This is because
the oxygen ion percolation trough the C6N4 membrane is blocked by the
NFO phase, while, the triple-phase membranes provide more paths for
the oxygen-ionic condition, due to the partial replacement of NFO with
BSCF, which promoted the oxygen permeability with either He or CO2 as
sweep gas.
4. Conclusions and outlook
CGO-based multi-phase membranes with the dead-end tubular ge­
ometry were fabricated with a modified phase-inversion casting tech­
nique. The performance of the membranes were compared at different
conditions. Three dual-phase membranes were obtained by paring BSCF,
NFO and LNO oxides with CGO. Results of the oxygen permeation tests
exhibited that the C6B4 membrane provided the highest permeation
flux, and the constant activation energies reflected good phase stabilities
of the three composites over the operation temperature range. The CO2-
tolerance of the membranes was investigated by feeding CO2 and helium
alternatively as sweep gas. A large drop of the oxygen flux was observed
for the C6B4 composites when CO2 was fed instead of helium to the
permeate side, but eventually a stable permeation was obtained which
was still higher than that of the other two membranes. Two triple-phase
membranes, C6B2N2 and C4B4N2, were fabricated with CGO, BSCF and
NFO powders. The permeation rates and corresponding activation en­
ergies obtained for the triple-phase composites were higher than that of
the C6N4 membrane but lower compared with that of the C6B4 mem­
brane. When CO2 was fed instead of helium as sweep gas, the perme­
ation fluxes through the C6B2N2 and C4B4N2 membranes exhibited
only a small reduction compared to the C6B4 membrane, and the
C6B2N2 membrane was found to provide the highest oxygen perme­
ability among all the dual-phase and triple-phase membranes investi­
gated in this work, when 100% CO2 was fed as the sweep gas. The main
purpose of this work was to select robust material for oxygen produc­
tion, so a single-layer membrane structure with a relatively high mem­
brane thickness was applied. This membrane structure allows an easier
fabrication process but results in a higher resistance for oxygen
permeation. In future work, some dual-phase porous support with
similar material as the selective layer can be prepared with the same
phase inversion casting method, and phase membrane layers can be
coated onto the support with dip-coating method. With this strategy
supported dual-phase membrane with improved permeability, good
chemical stability and sufficient mechanical strength can be
manufactured.
Declaration of competing interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
Acknowledgment
The China Scholarship Council is gratefully acknowledged for sup­
porting this work by providing PhD scholarship (NO. 201606450024) to
Yuliang Liu.
17776
Ceramics International 47 (2021) 17768–17777
References
[1] N. MacDowell, N. Florin, A. Buchard, J. Hallett, A. Galindo, G. Jackson, C.
S. Adjiman, C.K. Williams, N. Shah, P. Fennell, An overview of CO2 capture
technologies, Energy Environ. Sci. 3 (2010) 1645–1669, https://doi.org/10.1039/
c004106h.
[2] S.S. Hashim, A.R. Mohamed, S. Bhatia, Oxygen separation from air using ceramic-
based membrane technology for sustainable fuel production and power generation,
Renew. Sustain. Energy Rev. 15 (2011) 1284–1293, https://doi.org/10.1016/j.
rser.2010.10.002.
[3] M.A. Habib, H.M. Badr, S.F. Ahmed, R. Ben-Mansour, K. Mezghani, S. Imashuku, G.
J. la O’, Y. Shao-Horn, N.D. Mancini, A. Mitsos, A review of recent developments in
carbon capture utilizing oxy-fuel combustion in conventional and ion transport
membrane systems, Int. J. Energy Res. 35 (2011) 741–764, https://doi.org/
10.1002/er.
[4] W. Chen, C. sheng Chen, H.J.M. Bouwmeester, A. Nijmeijer, L. Winnubst, Oxygen-
selective membranes integrated with oxy-fuel combustion, J. Membr. Sci. 463
(2014) 166–172, https://doi.org/10.1016/j.memsci.2014.03.063.
[5] W. Chen, L. van der Ham, A. Nijmeijer, L. Winnubst, Membrane-integrated oxy-fuel
combustion of coal: process design and simulation, J. Membr. Sci. 492 (2015)
461–470, https://doi.org/10.1016/j.memsci.2015.05.062.
[6] R. Falkenstein-Smith, P. Zeng, J. Ahn, Investigation of oxygen transport membrane
reactors for oxy-fuel combustion and carbon capture purposes, Proc. Combust. Inst.
36 (2017) 3969–3976, https://doi.org/10.1016/j.proci.2016.09.005.
[7] H. Luo, K. Efimov, H. Jiang, A. Feldhoff, H. Wang, J. Caro, CO2-stable and Cobalt-
free Dual-phase Membrane for Oxygen Separation, 2011, pp. 759–763, https://doi.
org/10.1002/anie.201003723.
[8] C. Zhang, J. Sunarso, S. Liu, Designing CO2-resistant oxygen-selective mixed ionic-
electronic conducting membranes: guidelines, recent advances, and forward
directions 46 (2017) 2941–3005, https://doi.org/10.1039/c6cs00841k.
[9] L. Shi, S. Wang, T. Lu, Y. He, D. Yan, Q. Lan, Z. Xie, High CO2-tolerance Oxygen
Permeation Dual-phase Membranes Ce0.9Pr0.1O2-δ-Pr0.6Sr0.4Fe0.8Al0.2O3-δ,
vol. 806, 2019, pp. 500–509, https://doi.org/10.1016/j.jallcom.2019.07.281.
[10] T. Nagai, W. Ito, T. Sakon, Relationship between cation substitution and stability of
perovskite structure in SrCoO3-δ-based mixed conductors, Solid State Ionics 177
(2007) 3433–3444, https://doi.org/10.1016/j.ssi.2006.10.022.
[11] J. Yi, M. Schroeder, T. Weirich, J. Mayer, Behavior of Ba(Co, Fe, Nb)O3-δ perovskite
in CO2-containing atmospheres: degradation mechanism and materials design,
Chem. Mater. 22 (2010) 6246–6253, https://doi.org/10.1021/cm101665r.
[12] J. Yi, M. Schroeder, High temperature degradation of Ba0.5Sr0.5Co0.8Fe0.2O3-δ
membranes in atmospheres containing concentrated carbon dioxide, J. Membr. Sci.
378 (2011) 163–170, https://doi.org/10.1016/j.memsci.2011.04.044.
[13] K. Efimov, T. Klande, N. Juditzki, A. Feldhoff, Ca-containing CO2-tolerant
perovskite materials for oxygen separation, J. Membr. Sci. 389 (2012) 205–215,
https://doi.org/10.1016/j.memsci.2011.10.030.
[14] T. Klande, O. Ravkina, A. Feldhoff, Effect of A-site lanthanum doping on the CO2
tolerance of SrCo0.8Fe0.2O3-δ oxygen-transporting membranes, J. Membr. Sci. 437
(2013) 122–130, https://doi.org/10.1016/j.memsci.2013.02.051.
[15] W. Fang, F. Steinbach, C. Chen, A. Feldhoff, An approach to enhance the CO2
tolerance of fluorite-perovskite dual-phase oxygen-transporting membrane, Chem.
Mater. 27 (2015) 7820–7826, https://doi.org/10.1021/acs.chemmater.5b03823.
[16] C.S. Chen, B.A. Boukamp, H.J.M. Bouwmeester, G.Z. Cao, H. Kruidhof, A.J.
A. Winnubst, A.J. Burggraaf, Microstructural development, electrical properties
and oxygen permeation of zirconia-palladium composites, Solid State Ionics 76
(1995) 23–28, https://doi.org/10.1016/0167-2738(94)00253-O.
[17] J.E. ten Elshof, Oxygen permeation properties of dense Bi1.5Er0.5O3-Ag cermet
membranes, J. Electrochem. Soc. 144 (1997) 4361, https://doi.org/10.1149/
1.1838194.
[18] C.S. Chen, H. Kruidhof, H.J.M. Bouwmeester, H. Verweij, A.J. Burggraaf, Oxygen
permeation through oxygen ion oxide-noble metal dual phase composites, Solid
State Ionics 86–88 (1996) 569–572, https://doi.org/10.1016/0167-2738(96)
00206-8.
[19] A. Löfberg, S. Boujmiai, E. Capoen, M.C. Steil, C. Pirovano, R.N. Vannier,
G. Mairesse, E. Bordes-Richard, Oxygen permeation versus catalytic properties of
bismuth-based oxide ion conductors used for propene oxidation in a catalytic dense
membrane reactor, Catal. Today 91–92 (2004) 79–83, https://doi.org/10.1016/j.
cattod.2004.03.013.
[20] J. Kim, Y.S. Lin, Synthesis and oxygen permeation properties of ceramic-metal
dual-phase membranes, J. Membr. Sci. 167 (2000) 123–133, https://doi.org/
10.1016/S0376-7388(99)00273-2.
[21] Z. Zhang, W. Zhou, Y. Chen, D. Chen, J. Chen, S. Liu, W. Jin, Z. Shao, Novel
Approach for developing dual-phase ceramic membranes for oxygen separation
through beneficial phase reaction, ACS Appl. Mater. Interfaces 7 (2015)
22918–22926, https://doi.org/10.1021/acsami.5b05812.
[22] U. Nigge, H. Wiemho, E.W.J. Ro, Permeable membranes for exhaust gas sensors
146 (2002) 163–174.
[23] S. Huang, W. Li, Z. Cao, H. Li, H. Ma, X. Zhu, W. Yang, Effect of Bi doping on the
performance of dual-phase oxygen-permeable membranes, J. Membr. Sci. 579
(2019) 342–350, https://doi.org/10.1016/j.memsci.2019.03.005.
[24] H. Li, X. Zhu, Y. Liu, W. Wang, W. Yang, Comparative investigation of dual-phase
membranes containing cobalt and iron-based mixed conducting perovskite for
oxygen permeation, J. Membr. Sci. 462 (2014) 170–177, https://doi.org/10.1016/
j.memsci.2014.03.047.
[25] S.K. Kim, M.J. Shin, J. Rufner, K. van Benthem, J.H. Yu, S. Kim, Sr0.95Fe0.5Co0.5O3-
δ-Ce0.9Gd0.1O2-δ dual-phase membrane: oxygen permeability, phase stability, and
Y. Liu et al.
chemical compatibility, J. Membr. Sci. 462 (2014) 153–159, https://doi.org/
10.1016/j.memsci.2014.03.030.
[26] X. Zhu, W. Yang, Composite membrane based on ionic conductor and mixed
conductor for oxygen permeation, AIChE J. 54 (2008) 665–672, https://doi.org/
10.1002/aic.
[27] J. Xue, Q. Zheng, Y. Wei, K. Yuan, Z. Li, H. Wang, Dual phase composite oxide of
Ce0.9Gd0.1O2-δ-Ba0.5Sr0.5Co0.8Fe0.2O3-δ with excellent oxygen permeation, Ind. Eng.
Chem. Res. 51 (2012) 4703–4709, https://doi.org/10.1021/ie300083r.
[28] X. Zhu, H. Liu, Q. Li, Y. Cong, W. Yang, Unsteady-state permeation and surface
exchange of dual-phase membranes, Solid State Ionics 185 (2011) 27–31, https://
doi.org/10.1016/j.ssi.2011.01.005.
[29] X. Zhu, Q. Li, Y. He, Y. Cong, W. Yang, Oxygen permeation and partial oxidation of
methane in dual-phase membrane reactors, J. Membr. Sci. 360 (2010) 454–460,
https://doi.org/10.1016/j.memsci.2010.05.044.
[30] X. Zhu, Q. Li, Y. Cong, W. Yang, Syngas generation in a membrane reactor with a
highly stable ceramic composite membrane, Catal. Commun. 10 (2008) 309–312,
https://doi.org/10.1016/j.catcom.2008.09.014.
[31] Q. Li, X. Zhu, W. Yang, Single-step fabrication of asymmetric dual-phase composite
membranes for oxygen separation, J. Membr. Sci. 325 (2008) 11–15, https://doi.
org/10.1016/j.memsci.2008.08.002.
[32] G.D. Nam, G. Lee, S. Choi, J. Lee, S.J. Song, J.H. Joo, Chemically and mechanically
stable dual-phase membrane with high oxygen permeation flux, J. Mater. Chem.
(2020) 23884–23893, https://doi.org/10.1039/d0ta07680e.
[33] S. Wang, L. Shi, M. Boubeche, X. Wang, L. Zeng, H. Wang, Z. Xie, W. Tan, H. Luo,
Influence of Ln elements (Ln = La, Pr, Nd, Sm) on the structure and oxygen
permeability of Ca-containing dual-phase membranes, Separ. Purif. Technol. 251
(2020) 117361, https://doi.org/10.1016/j.seppur.2020.117361.
[34] M. Balaguer, J. García-Fayos, C. Solís, J.M. Serra, Fast oxygen separation through
SO2- and CO2-stable dual-phase membrane based on NiFe2O4-Ce 0.8Tb0.2O2-δ,
Chem. Mater. 25 (2013) 4986–4993, https://doi.org/10.1021/cm4034963.
[35] H. Luo, H. Jiang, K. Efimov, J. Caro, Influence of the preparation methods on the
microstructure and oxygen permeability of a CO2-stable dual phase membrane,
AIChE J. 57 (2011) 2738–2745, https://doi.org/10.1002/aic.
[36] H. Luo, H. Jiang, T. Klande, F. Liang, Z. Cao, H. Wang, J. Caro, Rapid glycine-
nitrate combustion synthesis of the CO2-stable dual phase membrane
17777
Ceramics International 47 (2021) 17768–17777
40Mn1.5Co1.5O4-δ-60Ce0.9Pr0.1O2-δ for CO2 capture via an oxy-fuel process,
J. Membr. Sci. (2012) 423–424, https://doi.org/10.1016/j.memsci.2012.08.046,
450–458.
[37] N. Han, Q. Wei, H. Tian, S. Zhang, Z. Zhu, J. Liu, S. Liu, Highly stable dual-phase
membrane based on Ce0.9Gd0.1O2-δ-La2NiO4+δ for oxygen permeation under pure
CO2 atmosphere, Energy Technol. 7 (2019) 1–10, https://doi.org/10.1002/
ente.201800701.
[38] Y. Liu, M.V.S. Annaland, F. Gallucci, Performance control of dead-end tubular
membranes fabricated with a modified phase inversion casting method, Ceram. Int.
46 (2020) 22429–22437, https://doi.org/10.1016/j.ceramint.2020.05.325.
[39] J. Tong, W. Yang, B. Zhu, R. Cai, Investigation of ideal zirconium-doped
perovskite-type ceramic membrane materials for oxygen separation, J. Membr. Sci.
203 (2002) 175–189, https://doi.org/10.1016/S0376-7388(02)00005-4.
[40] H. Luo, H. Jiang, K. Efimov, F. Liang, H. Wang, J. Caro, CO2-tolerant oxygen-
permeable Fe2O3-Ce0.9Gd0.1O2-δ dual phase membranes, Ind. Eng. Chem. Res. 50
(2011) 13508–13517, https://doi.org/10.1021/ie200517t.
[41] H. Cheng, L. Luo, W. Yao, X. Lu, X. Zou, Z. Zhou, Novel cobalt-free CO2-tolerant
dual-phase membranes of Ce0.8Sm0.2O2-δ-Ba0.95La0.05Fe1-xZrxO3-δ for oxygen
separation, J. Membr. Sci. 492 (2015) 220–229, https://doi.org/10.1016/j.
memsci.2015.05.057.
[42] S. Liu, X. Tan, K. Li, R. Hughes, Synthesis of strontium cerates-based perovskite
ceramics via water-soluble complex precursor routes, Ceram. Int. 28 (2002)
327–335, https://doi.org/10.1016/S0272-8842(01)00098-0.
[43] O. Czuprat, M. Arnold, S. Schirrmeister, T. Schiestel, J. Caro, Influence of CO2 on
the oxygen permeation performance of perovskite-type BaCoxFeyZrzO3-δ hollow
fiber membranes, J. Membr. Sci. 364 (2010) 132–137, https://doi.org/10.1016/j.
memsci.2010.08.019.
[44] H. Huang, S. Cheng, J. Gao, C. Chen, J. Yi, Phase-inversion tape-casting
preparation and significant performance enhancement of Ce0.9Gd0.1O1.95-
La0.6Sr0.4Co0.2Fe0.8O3-δ dual-phase asymmetric membrane for oxygen separation,
Mater. Lett. 137 (2014) 245–248, https://doi.org/10.1016/j.matlet.2014.09.016.
[45] W.K. Hong, G.M. Choi, Oxygen permeation of BSCF membrane with varying
thickness and surface coating, J. Membr. Sci. 346 (2010) 353–360, https://doi.org/
10.1016/j.memsci.2009.09.056.