Solid State Ionics 181 (2010) 453–458
Contents lists available at ScienceDirect
Solid State Ionics
j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / s s i
A new hydrophobic thin film catalyst layer for PEMFC
Wei Song a,b, Hongmei Yu a,⁎, Lixing Hao a,b, Zhili Miao a,b, Baolian Yi a, Zhigang Shao a
a
b
Laboratory of Fuel Cells, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, 457 Zhongshan Road, Dalian 116023, PR China
Graduate School of the Chinese Academy of Sciences, Beijing 100039, PR China
a r t i c l e
i n f o
Article history:
Received 25 March 2009
Received in revised form 15 December 2009
Accepted 15 January 2010
Keywords:
Fuel cell
Proton exchange membrane (PEM)
Thin film catalyst layer
PTFE
Decal method
a b s t r a c t
A new hydrophobic thin film catalyst layer (CL) was prepared by the decal method in this work.
Polytetrafluoroethylene (PTFE) was introduced to the catalyst ink with 1 wt.% Nafion® ionomer as the
hyperdispersant. Aluminum foil was adopted as the transferring medium which enabled the sintering
process of PTFE. Then an appropriate amount of Nafion® ionomer was sprayed onto the CL for proton
conduction. At last, the CL was transferred onto a Nafion® membrane and the catalyst-coated membrane
(CCM) was formed. Contact angle measurement and mercury intrusion porosimetry test were conducted to
characterize the hydrophobicity and porosity of the thin film CL. The results showed that PTFE addition
favored the CL hydrophobicity and porosity. The optimal PTFE content was also deduced by comparing the
fuel cell performance under different PTFE contents. Electrochemical analysis revealed that PTFE addition
decreased the electrochemical active area (ECA) but enhanced the diffusion process in the CL.
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
Proton exchange membrane (PEM) fuel cells have attracted much
attention due to the merits of high power density, high energy
conversion efficiency and friendly environmental adaptability. As the
core component of PEM fuel cells, the membrane electrode assembly
(MEA) directly impacts on the cell performance. During the past years,
much effort has been paid on improving the MEA preparation [1,2,9–
19]. The catalyst layer (CL) plays an important role in the PEM fuel cell
by means of reaction kinetics, three-phase boundary, mass-transport
process, water management, and electron/proton conduction. Generally, the thin film CL shows higher catalyst utilization and cell
performance [1,2].
Effective conduction of proton/electron and access of reactant gas
are essential for a catalyst site where the electrochemical reaction
takes place. Access of reactant gas, i.e. gas diffusion rate, depends on
the electrode structure. A well designed CL structure improves the gas
diffusion rate and favors the reaction process. According to the
fabrication process, CLs are usually applied to the gas diffusion layer
(GDL) or the membrane. Conventional electrodes adopt the GDL to
support the CL [3,4]. In this case, polytetrafluoroethylene (PTFE) is
commonly used to bind the catalyst particles and form a hydrophobic
structure. The impregnated Nafion® acts as the ionic transport path.
The thickness of this conventional PTFE-bonded CL is about 30–35 μm
[3]. Besides building the hydrophobic/hydrophilic path and acting as
an ionic transport path, these polymers also influence the catalyst
⁎ Corresponding author. Tel.: + 86 411 84379051; fax: + 86 411 84379185.
E-mail address: hmyu@dicp.ac.cn (H. Yu).
0167-2738/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.ssi.2010.01.022
utilization [5,6]. Furthermore, their distribution and content are
crucial to the fuel cell performance [7,8]. In 1992, M.S. Wilson
proposed a decal fabrication method of the thin film catalyst layer, in
which the CL was directly supported on the membrane [1]. In this
method, PTFE is eliminated and the dispersion of ionomer throughout
the CL is improved. The typical thickness of the thin film CL is 5–10 μm
[2,9]. However, a hydrophobic path does not form in this kind of CL,
which increases the risk of water flooding. Thereafter, much attention
has been paid on improving the fabrication process of the thin film CL.
Park [10] reported a modified decal method in 2008, the catalyst
transfer ratio during the decal process was promoted. Effects of
Nafion® content, catalyst loading and membrane thickness on the
performance of the thin film CL were investigated by Passos [11].
Except for the decal method, the thin film CL can also be fabricated by
a direct spraying method. Catalyst ink is directly sprayed onto the
membrane to form the catalyst-coated membrane (CCM). However,
the Nafion® membrane swells severely when it encounters organic
solvents in this method. To avoid the swelling phenomenon and make
sure the CL is homogeneous, Sun et al. proposed an improvement
method [9]. According to their report, the Nafion® membrane was
soaked in ethylene-glycol (EG) and a small amount of EG was added
into the catalyst ink during the fabrication process. Furthermore,
some research relevant to the depositing process of Pt and the
enhancement of the interfacial properties was carried out [12,13]. No
matter if direct spraying method or decal method is adopted, the gas
diffusion path in the thin film CL is necessary for the cell performance.
Some pore forming methods were also reported aiming at decreasing
the gas diffusion resistance of the thin film CL [14–16]. But the
constructed pores were easy to be compressed. Besides the gas
diffusion resistance, water management in the thin film CL is also
454
W. Song et al. / Solid State Ionics 181 (2010) 453–458
important. A hydrophilic CL can have a negative effect on the water
removal rate during cell operation [20]. To improve the gas
penetration and water management, the fabrication process of
MEAs still can be improved.
In this work, a decal fabrication process for the hydrophobic thin
film CL was proposed. By adopting 1 wt.% Nafion® ionomer as the
hyperdispersant, PTFE was mixed homogeneously in the catalyst ink.
Five relative low PTFE contents were investigated. The physical
characteristics of the prepared catalyst-coated membrane (CCM)
were measured. The electrochemical active area (ECA) and the fuel
cell performance were also studied.
2. Experimental
2.1. Preparation of the CCMs
Nafion® ionomer was introduced into a 10 wt.% PTFE solution in an
ultrasonic environment to form a PTFE-Nafion® solution, in which the
Nafion® content was 1 wt.%. A mixture composed of Pt/C catalyst
(50 wt.% Pt from Johnson Matthey Inc.), distilled water and isopropyl
alcohol was agitated with ultrasonic for 30 min. Then a certain PTFENafion® solution was added, and another 30 min ultrasound treatment was carried out. Finally, the catalyst ink was formed.
The catalyst ink was sprayed onto a smooth aluminum foil, which
was sintered at 240 °C for 30 min and 340 °C for 30 min under a
nitrogen atmosphere. This high temperature process destroyed the
surfactant and made the PTFE hydrophobic. It was reported that the
sulphonate functionality on the Nafion® polymer would be lost in the
temperature range of 275–380 °C [21]. In this experiment, the 1 wt.%
Nafion® from the PTFE-Nafion solution decomposed due to the 30 min
treatment at 340 °C.
A 5 wt.% Nafion® solution from DuPont was diluted in isopropyl
alcohol, which was sprayed onto the catalyst layer. During this
process, Nafion® penetrated the catalyst layer to form the proton
transfer network. After all the residual solvent was evaporated, the
catalyst layer was hot-pressed onto a Nafion® 212 (N212) membrane
on both sides. Then the CCM was prepared. The Pt loading of the
prepared CCM was 0.4 mg cm− 2 on each side.
Fig. 1 shows the whole fabrication process of the CCM. A patent
referring to this fabrication method has been registered in the Chinese
State Patent Office [22]. Thin film hydrophobic CLs with 0 wt.%, 1 wt.%,
3 wt.%, 5 wt.% and 7 wt.% PTFE were fabricated. In these CLs, the mass
ratios of PTFE/carbon were 0, 0.025, 0.088, 0.15 and 0.21, respectively.
The mass ratios of Nafion® to carbon were always 0.8.
2.2. Physical characterization of the CCMs
The cross-section morphology of the CCM was investigated with
the Super Depth Surface Profile Measurement Microscope VK-8550
from KEYENCE. The image was obtained by zooming in for 1000 times.
The contact angle of the CCM surface was measured by the KRŰSS
DSA100 Drop Shape Analysis System. 3 μl distilled water was dropped
onto the surface of the CCM at room temperature. Ten seconds after
the water drop was placed on the surface, the contact angle image was
taken. The contact angle was analyzed with the tangent method of
Sessile Drop Fitting.
The porosity of the CCMs was measured by the mercury intrusion
method with a PoremasterGT60 (Quantachrome). The pressure was
from 1.38 × 103 to 4.13 × 107 Pa (0.2–0.6 × 104 psi) and the mercury
contact angle was 140°.
2.3. Single cell performance and electrochemical tests
Toray carbon paper (TGP-H-060) was chosen as the diffusion layer
for anode/cathode. Two graphite polar plates machined with parallel
flow channels acted as the current collectors and gas flow channels.
Fig. 1. The flow chart of the hydrophobic CCM fabrication process.
With two organic glasses as the end plates, a single cell was
assembled. The effective area of the single cell was 4 cm2. Cycling
water was introduced between the graphite polar plates and the end
plates to control the cell temperature.
The fuel cell performance was tested at H2/O2 and H2/Air conditions.
When testing the H2/O2 performance, pure hydrogen was used as the
fuel and oxygen as the oxidant at atmospheric pressure. The flow rates
were set at 40 ml min− 1 and 100 ml min− 1 for H2 and O2, respectively.
The cell temperature was kept at 60 °C and the humidification
temperatures were set at 65 °C/60 °C for H2/O2. When testing the H2/
Air performance, the flow rates were set at 40 ml min− 1 and
600 ml min− 1 for H2 and air, respectively. The high gas flow rate
removed more water out. As a result, the cell temperature was set at
65 °C and the humidification temperature was set at 70 °C/70 °C for H2/
Air. The utilization of O2/Air at 0.5A cm− 2 was 6.7%/5.5%, and that at
1 A cm− 2 was 13.9%/11.1% for O2/Air.
The fuel cell was firstly activated until the steady state. Then the
polarization curve of the cell was measured with a KFM 2030
Impedance Meter (Kikusui, Japan) using the Fuelcell-Load & Impedance Meter software (Kikusui, Japan). The i–V curve was obtained by
measuring the voltage in intervals of 0.05 A cm− 2 between 0 and
1 A cm− 2. At each current step the equilibration time was 8 s. Along
with each cell voltage, the corresponding ohmic resistance was also
recorded. To ensure the measurement reproducibility, this testing
process was repeated 3 times. The obtained three curves were almost
consistent and the third curve was used. Then the EIS at two current
densities were measured.
After the cell shut down, high-purity nitrogen was fed to the
cathode with a flow rate of 50 ml min− 1. The cyclic voltammetry (CV)
measurement was carried out using BI-STAT 2 channel Potentiostat
W. Song et al. / Solid State Ionics 181 (2010) 453–458
Fig. 2. Micrograph of the CCM cross-section,×1000.
(USA, Princeton). The scan rate was 50 mV s− 1 and the voltage range
was 0–0.8 V referring to the hydrogen electrode.
3. Results and discussions
3.1. Physical characteristics of the CCMs
According to Mao's results [23], the Nafion® ionomer can act as a
hyperdispersant for the PTFE water dispersion when preparing the
gas diffusion layer. In this work, the Nafion® hyperdispersant is
455
introduced into the PTFE dispersion for the CL fabrication. Since the
investigated PTFE contents are relatively low (0 wt.%–5 wt.%) and the
Nafion® hyperdispersant is adopted, the aggregation phenomenon of
PTFE is improved. The cross-section morphology of the CCM is shown
in Fig. 2. The picture clearly shows the three layers of the CCM. The
central layer is the N212 membrane, neighboring to which are the
catalyst layers. The thicknesses of the hot-pressed membrane and the
catalyst layer are about 50 μm and 10 μm, respectively. As shown in
the image, the transfer process does not affect the membrane
thickness severely.
The contact angles of the CCM surface are measured to analyze the
hydrophobicity of the CLs with different PTFE contents. The images of
the contact angle measurements are shown in Fig. 3, and the results
are listed in Table 1. Here, the five CCMs with 0 wt.%, 1 wt.%, 3 wt.%,
5 wt.% and 7 wt.% PTFE are named CCM0, CCM1, CCM3, CCM5 and
CCM7, respectively. Their corresponding contact angles are 126.7°,
139.9°, 146.1°, 146.1° and 158.9°. Obviously, the addition of PTFE
improves the hydrophobicity of the CL surface. However, there is no
effective method for measuring the internal hydrophobicity of the
inhomogeneous material by now. The optimal PTFE content should be
further confirmed with the cell performance measurements in the
following section.
To improve the porosity of the CL and enhance the mass-transport
properties during the oxygen reduction reaction, PTFE is added into
the CL. To investigate the effect of the PTFE content on the CL pore size,
mercury intrusion porosimetry is carried out. During the measurement, larger pores are firstly intruded with mercury. As the pressure
increases, mercury gradually enters the smaller pores. Different
pressures correspond to different pore sizes. So the pore size
Fig. 3. Images of the contact angle measurement.
456
W. Song et al. / Solid State Ionics 181 (2010) 453–458
Table 1
Results of the contact angle measurement and the mercury intrusion porosimetry test.
0 wt.%
1 wt.%
3 wt.%
5 wt.%
7 wt.%
Contact
angle (°)
Mean pore
diameter (nm)
Medium pore
diameter (nm)
126.7
139.9
146.1
151.6
158.9
5.68 [20]
8.24
9.22
10.63
11.62
5.70 [20]
8.09
9.19
16.22
22.24
distribution is obtained basing on the relationship between pore size
and intruded mercury amount. Fig. 4 shows the integrated pore size
distribution of the four CCMs with PTFE and the detailed results are
listed in Table 1. In this work the range of pore size is between
0.0036 μm and 0.1 μm. Since the pore size of the membrane is small
enough to be neglected [24,25], the calculated pore volume accounts
for the catalyst layer. The pore size of the CCM with 0 wt.% PTFE has
been tested before, which is referred to in Table 1 [20]. Here, the mean
pore diameter represents the average pore size of the whole CL, which
is obtained by dividing the four times specific pore volume with the
specific surface area. The medium pore diameter stands for the pore
size when the mercury intrusion percentage is 50%. As the PTFE
content increases from 0 wt.% to 7 wt.%, the mean pore size increases
by 5.94 nm and the medium pore size increases by 16.54 nm.
Obviously, more PTFE results in larger mean pore size. It has been
reported by both Watanabe [26] and Uchida [17] that there exists a
pore size boundary in the CL. In this experiment, there are also two
distinctive pore distributions with the boundary at about 0.01 μm. As
for CCM1, the percentage of the pores smaller than 0.01 μm is 51.8%,
and those of CCM3, CCM5, and CCM7 are 53.2%, 41.7% and 37.5%,
respectively. Simultaneously, the percentages of the pores between
0.01 μm and 0.1 μm are 48.2% for CCM1, 46.8% for CCM3, 58.3% for
CCM5 and 62.5% for CCM7. As a result, the addition of PTFE increases
the percentage of larger pores which may act as the main gas
transport channels [26].
the polarization curves which can distinguish the kinetic effects,
ohmic effects and gas diffusion, such as the model of Kim et al. [27]:
E = E0 −b log i−Ri−m exp ðniÞ
ð1Þ
In this model, the exponential term characterizes the masstransport region of the polarization curve. Besides that Lee et al. [28],
Squadrito et al. [29], and L. Pisani et al.[30] have also developed some
important models to estimate the polarization curves. Then Santarelli
and his partners developed a parameter estimation method for the
polarization curve in 2006 [31]. This equation is adopted in this work
as follows:
V=
0:5
RT
RT
−1 i + in
−1 i + in
−Δ ―
g ðTÞ
RT pH2 pO2
− c Sinh ð
Þ− a Sinh ð
Þ
+
ln
αF
α F
P
2i0;c
2i0;a
2F
2F
H2 O
−rði + in Þ +
RT
i + in
RT
i + in
lnð1−
lnð1−
Þ+
Þ
4F
2F
il;c
il;a
ð2Þ
where ḡ is the molar Gibbs free energy (J mol− 1), αc/αa is the anode/
cathode transfer coefficient, in is the internal current density (A cm− 2),
il,a/il,c is the anode/cathode limiting current density (A cm− 2), i0,a/i0,c is
the anode/cathode exchange current density (A cm− 2). Since the
temperature and other reactant condition are all the same in our
experiment, the temperature and reactant gas pressure relevant term
can be written into a constant. Since the anode takes a very small effect
to the polarization, the anode term is omitted in this paper. The internal
current density is also neglected. Then the equation is rewritten as
V = C−
RT
i
RT
i
−1
Sinh ð
lnð1− Þ
Þ−ri +
0:5F
2i0;c
4F
il;c
ð3Þ
3.2. Fuel cell performance
The CCMs with different PTFE contents are assembled into fuel
cells. Fig. 5 shows the fuel cell performances and the corresponding
ohmic resistances (RΩ) under H2/O2 and H2/Air conditions. From the
i–V curves, the PTFE content obviously influences the cell performance. Comparing with CCM0, the addition of 1 wt.%, 3 wt.% and 5 wt.
% PTFE is favorable for the cell performance, among which CCM5
shows the best results. There are several good models for analyzing
Fig. 4. Pore size distribution curves of the four hydrophobic CCMs.
Fig. 5. Steady-state fuel cell polarization curves: (A) H2/O2 performance (B) H2/Air
performance.
W. Song et al. / Solid State Ionics 181 (2010) 453–458
457
Table 2
The fitting parameters of the polarization curves.
Samples
CCM0
CCM1
CCM3
CCM5
CCM7
C (mV)
i0,c (A cm− 2)
r (Ω cm2)
Il,c (A cm− 2)
H2/O2
H2/Air
H2/O2
H2/Air
H2/O2
H2/Air
H2/O2
H2/Air
881.4
845.7
837.3
856.9
839.7
874.3
841.8
864.8
863.2
860.3
0.224
0.255
0.222
0.263
0.271
0.255
0.279
0.258
0.265
0.330
0.489
0.412
0.380
0.375
0.450
0.736
0.598
0.640
0.596
0.735
1.01
1.08
1.01
0.95
1.00
0.878
0.808
0.903
0.943
0.872
Here, C, i0,c, il,c, and r are the parameters to be estimated. αc is
assumed to be 0.5 [31]. The exchange current density is mainly a
function of the electrode characteristics and the limiting current
density is a parameter linked to the concentration overpotential
which is generally caused by a change in porosity, tortuosity in the
diffusion layer (DL) and the presence of water droplets or films. r
represents the total contribution of the linear polarization components including the ohmic losses, the charge transfer resistance (Rct)
of the hydrogen oxidation reaction. The fitted curves are shown in
Fig. 5 as solid lines and the results are listed in Table 2. Since all the
MEAs are made of the same gas diffusion layer and membrane, the
difference of these parameters should attribute to the different
structure of the CLs. Under the H2/O2 or H2/Air condition, the CL with
5 wt.% PTFE shows the smallest r which acts as the optimal PTFE
content. As it is known, the H2/Air cell performance is more related
with the concentration polarization. Thus, the last column of Table 2
indicates that adding a certain PTFE can improve the concentration
polarization of the thin film CL.
3.3. ECA
PTFE can affect the continuity of the electron/proton transfer
process in the CL. ECA stands for the active surface area where the
catalyst contacts with the proton conductor and the electron
conductor. Then ECA of the CL can directly reflect the influence of
PTFE addition. Fig. 6 shows the CV scan curves of the CCMs. Obviously,
a larger PTFE content results in a lower ECA. But considering the cell
performance discussed above, the addition of an appropriate amount
of PTFE in the CCM is acceptable despite of the loss in ECA.
3.4. Electrochemical impedance spectra (EIS)
The polarization resistance (Rp) is analyzed in Section 3.1 based on
the polarization curves. In this section, the Rp in the CL is further
discussed using EIS, which is quite an effective diagnostic tool. The
cathode acts as the working electrode and the anode acts as the
reference and counter electrode. EIS of the cells under the H2/O2
Fig.6. Cyclic voltammogram curves of different CCMs.
Fig. 7. EIS of the fuel cells at 0.1 A cm− 2 and 0.5 A cm− 2.
condition were measured at 0.1 A cm− 2 and 0.5 A cm− 2, as shown in
Fig. 7.
The Nyquist plots from different references show different
patterns. Generally, it contains one or two or three semicircles
which indicate different electrode processes [32,33]. The EIS results in
this paper are shown in Fig. 8, which clearly show only one semicircle
in the experimental frequency range. Quantitative analyses and
extraction of reliable system-dependant parameters require the
assumption of an electrochemical equivalent circuit (EC). The Randles
cell, depicted in Fig. 8, is used to model the system and fit the spectra.
The highest frequency intercept with the real axis of this semicircle
stands for the RΩ. The diameter of the semicircle is related with the Rct
for the oxygen reduction reaction and the catalyst layer capacitance
(C) [34–36]. Here, C is replaced with a constant phase element (CPE,
defined as Z(CPE) = 1 / Q(jw)− n, where Q has the numerical value of
the admittance (1/|Z|) at ω = 1 rad s− 1 and n is the exponent value)
because the capacitance due to the double layer charging is
distributed along the length of the pores in the porous electrode
[37]. The ZSimpWin fitting soft is used to fit the data by applying the
EC mentioned above, and the Rct results for the cells are listed in
Table 3. As it is known, the diameter of the semicircle depends on the
electrode potential. The two investigated current densities result in
two Rct values. Since there is only one loop in the EIS, the diameter
may also be caused by the mass transport through the pores and the
Nafion® polymers in the CL [38,39]. This contribution should be more
obvious at a higher current density.
Comparing with CCM0 (Table 3), there are no obvious changes in
Rct at 0.1 A cm− 2 of CCM1–CCM7. This means the PTFE addition to the
CL does not affect the charge transfer process in the catalyst/
Fig. 8. The applied equivalent circuit for fitting the EIS.
458
W. Song et al. / Solid State Ionics 181 (2010) 453–458
Table 3
The fitting results of the charge transfer resistance for the cells.
CCMs
Rct at 0.1 A cm− 2/Ω cm2
Rct at 0.5 A cm− 2/Ω cm2
CCM0
CCM1
CCM3
CCM5
CCM7
0.493
0.498
0.521
0.473
0.536
0.264
0.242
0.176
0.147
0.191
electrolyte interface. On the other hand, the Rp reflected by the
semicircle at 0.5 A cm− 2 obviously improved after adding PTFE to the
CL. When the content is 5 wt.%, Rp at 0.5 A cm− 2 decreases by 44.3%
compared with that without PTFE. From the different changing rate of
Rp at the two current densities, it is confirmed that Rp is related with
the mass transport in the CL. And it is concluded that the addition of
an appropriate amount of PTFE in the CCM can decrease the diffusion
resistance in the CL. This means that the hydrophobic thin film CL can
favor the mass-transport process.
4. Conclusions
A new hydrophobic thin film catalyst layer was fabricated by decal
method. By introducing the Nafion® hyperdispersant to the PTFE
solution, the aggregation of PTFE in the ink was improved. Due to the
addition of PTFE, the hydrophobicity of the CL was enhanced and the
pore volume of the CL was expanded. The cell performance of the
CCMs proved that the optimal PTFE content was 5 wt.%. The EIS results
at 0.1 A cm− 2 and 0.5 A cm− 2 showed that adding PTFE to the CL did
not affect the charge transfer resistance, while the internal mass
diffusion of the CL was improved.
Acknowledgement
This work was financially supported by the National High Technology Research and Development Program of China (No. 2007AA05Z123)
and the National Natural Science Foundations of China (No. 20636060,
No. 20876154). We also appreciate Kikusui Electronics Corp. for
supplying the electronic load.
References
[1] M.S. Wilson, S. Gottesfeld, J. Applied Electrochem. 22 (1992) 1.
[2] M. Prasanna, E.A. Cho, T.-H. Lim, I.-H. Oh, Electrochim. Acta 53 (2008) 5434.
[3] Y.G. Chun, C.S. Kim, D.H. Peck, D.R. Shin, J. Power Sources 71 (1998) 174.
[4] S.J. Lee, S. Mukerjee, J. McBreen, Y.W. Rho, Y.T. Kho, T.H. Lee, Electrochim. Acta 43
(1998) 3683.
[5] C.K. Subramaniam, N. Rajalakshmi, K. Ramya, K.S. Dhathathreyan, Bull. Electrochem. 16 (2000) 350.
[6] S. Kamaguchi, M. Kamada, A. Yamada, M. Umeda, Bunseki Kagaku 53 (2004) 981.
[7] K.M. Yin, J. Applied Electrochem. 37 (2007) 971.
[8] S. Wang, G. Sun, Z. Wu, Q. Xin, J. Power Sources 1 (2007) 128.
[9] L. Sun, R. Ran, G. Wang, Z.P. Shao, Solid State Ionics 179 (2008) 960.
[10] H.-S. Park, Y.-H. Cho, Y.-H. Cho, I.S. Park, N. Jung, M. Ahn, Y.E. Sung, J. Electrochem.
Soc. 15 (2008) B455.
[11] R.R. Passos, V.A. Paganin, E.A. Ticianelli, Electrochim. Acta 51 (2006) 5239.
[12] S. Towne, V. Viswanathan, J. Holbery, P. Rieke, J. Power Sources 171 (2007) 575.
[13] Y.H. Cho, S.J. Yoo, Y.H. Cho, H.-S. Park, I.S. Park, J.K. Lee, Y.E. Sung, Electrochim. Acta
53 (2008) 6111.
[14] Z.G. Shao, B.L. Yi, M. Han, Hydrogen Energy Progress XIII 1–2 (2000) 815.
[15] A. Fischer, J. Jindra, H. Wendt, J. Applied Electrochem. 28 (1998) 277.
[16] Y. Song, Y. Wei, H. Xu, M. Williams, Y.X. Liu, L.J. Bonville, H.R. Kunz, J.M. Fenton,
J. Power Sources 141 (2005) 250.
[17] M. Uchida, Y. Aoyama, N. Eda, A. Ohta, J. Electrochem. Soc. 142 (1995) 4143.
[18] S. Hirano, J. Kim, S. Srinivasan, Electrochim. Acta 42 (1997) 1587.
[19] X. Zhang, P. Shi, Electrochem. Commun. 8 (2006) 1229.
[20] W. Song, J.B. Hou, Appl. Electrochem. 39 (2009) 609.
[21] S.R. Samms, S. Wasmus, R.F. Savinell, J. Electrochem. Soc. 143 (1996) 1498.
[22] L. Yi, H.M. Yu, W. Song, Chinese state patent office, Patent application number:
200810228028.5
[23] Q. Mao, G. Sun, S. Wang, H. Sun, Y. Tian, I. Tian, O. Xin, J. Power Sources 175 (2008)
826.
[24] J. Xie, Electrochem. Soc. 151 (2004) A1841.
[25] S. Koter, J. Membr. Sci. 166 (2000) 127.
[26] M. Watanabe, M. Tomikawa, S. Motoo, J. Electroanal. Chem. 195 (1985) 81.
[27] J. Kim, Electrochem. Soc. 142 (1995) 2670.
[28] J. Lee, T. Lalk, A. Appleby, J. Power Sources 70 (1998) 258.
[29] G. Squadrito, G. Maggio, E. Passalacqua, F. Lufrano, A. Patti, J. Appl. Electrochem. 29
(1999) 1449.
[30] L. Pisani, G. Murgia, M. Valentini, B. D'Aguanno, J. Power Sources 108 (2002) 192.
[31] M.G. Santarelli, M.F. Torchio, P. Cochis, J. Power Sources 159 (2006) 824.
[32] Y. Bultel, K. Wiezell, F. Jaouen, P. Ozil, G. Lindbergh, Electrochim. Acta 51 (2005)
474.
[33] X. Yuan, J.C. Sun, M. Blanco, J. Power Sources 161 (2006) 920.
[34] F. Jaouen, G. Lindbergh, G. Sundholm, J. Electrochem. Soc. 149 (2002) A437.
[35] J. Ihonen, F. Jaouen, G. Lindbergh, A. Lundblad, G. Sundholm, J. Electrochem. Soc.
149 (2002) A448.
[36] V.A. Paganin, C.L.F. Oliveira, E.A. Ticianelli, T.E. Springer, E.R. Gonzalez, Electrochim. Acta 43 (1998) 3761.
[37] T. Romero-Castanon, L.G. Arriaga, J. Power Sources 118 (2003) 179.
[38] J.B. Hou, W. Song, H.M. Yu, Y. Fu, Z.G. Shao, B.L. Yi, J. Power Sources 171 (2007)
610.
[39] M. Ciureanu, R. Roberge, J. Phys. Chem. B 105 (2001) 3531.