Journal of Power Sources 167 (2007) 281–287

Direct methanol fuel cell with extended reaction zone anode:

PtRu and PtRuMo supported on graphite felt
Alex Bauer, Előd L. Gyenge ∗ , Colin W. Oloman
Department of Chemical and Biological Engineering, The University of British Columbia, 2360 East Mall, Vancouver, BC, Canada V6T 1 Z3
Received 2 February 2007; received in revised form 22 February 2007; accepted 23 February 2007
Available online 2 March 2007

Abstract
Pressed graphite felt (thickness ∼350 ␮m) with electrodeposited PtRu (43 g m−2 , 1.4:1 atomic ratio) or PtRuMo (52 g m−2 , 1:1:0.3 atomic ratio)
nanoparticle catalysts was investigated as an anode for direct methanol fuel cells. At temperatures above 333 K the fuel cell performance of the
PtRuMo catalyst was superior compared to PtRu. The power density was 2200 W m−2 with PtRuMo at 5500 A m−2 and 353 K while under the
same conditions PtRu yielded 1925 W m−2 . However, the degradation rate of the Mo containing catalyst formulation was higher. Compared to
conventional gas diffusion electrodes with comparable PtRu catalyst composition and load, the graphite felt anodes gave higher power densities
mainly due to the extended reaction zone for methanol oxidation.
© 2007 Elsevier B.V. All rights reserved.

Keywords: Nanoparticles; Methanol oxidation; Electrodeposition; Three-dimensional electrodes

1. Introduction uid/gas) counter-current flow of CH3 OH(aq) toward and CO2(g)

away from the catalytically active sites in the anode structure
Methanol has been proposed the key energy carrier and [11–13]. These problems are expected to worsen with scale-
chemical raw material of the future, with production shift- up to high power output (e.g. 100 kW range) and larger stacks,
ing increasingly toward biomass and chemical/electrochemical such as those required for transportation. Therefore, at present
conversion of atmospheric CO2 as opposed to present day tech- portable power applications are deemed the most feasible for
nologies based on syn-gas obtained from natural gas and coal commercialization of direct methanol fuel cells [14].
[1]. Closing the CO2 emission loop by employing an electro- In order to solve these challenges it is important to recognize
chemical reactor producing CH3 OH and/or HCOO− /HCOOH the interaction among the above mentioned effects such as a
at the cathode [2] to be fed directly into a fuel cell anode with low methanol oxidation rate at the anode, which will lead to a
an overall net energy production, while still a distant goal, could high methanol concentration adjacent to the membrane causing
represent an important advancement toward sustainable energy an enhanced crossover flux to the cathode. Moreover, CO2 gas
generation. entrapment and accumulation in the porous anode increases the
The issues that need to be addressed associated with direct pressure on the anode side and could induce both CH3 OH and
methanol fuel cells (DMFC) are well-documented in the lit- CO2 crossover.
erature. These are (selected references are also indicated): Recently it has been proposed to replace the traditionally
electrocatalysis of CH3 OH oxidation (e.g. on PtRu with load employed gas diffusion electrode relying on a thin (∼20 ␮m)
typically above 10 g m−2 ) [3–8], CH3 OH permeation through catalyst layer, with a three-dimensional anode (also referred to as
the polymer electrolyte-membrane causing a mixed cathode extended reaction zone anode), having uniformly distributed cat-
potential and bringing about a need for methanol tolerant O2 alyst nanoparticles throughout its thickness (i.e. approximately
electroreduction catalysts [9,10] and inefficient two-phase (liq- between 200 and 1000 ␮m depending on the type of electrode
material and compression in the fuel cell) [15–17].
Most of the research to date in this area has focused on
∗ Corresponding author. Tel.: +1 604 822 3217; fax: +1 604 822 6003. developing novel methods for nanostructured catalyst synthesis
E-mail address: egyenge@chml.ubc.ca (E.L. Gyenge). and deposition onto various three-dimensional electrodes such

0378-7753/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.jpowsour.2007.02.053
282 A. Bauer et al. / Journal of Power Sources 167 (2007) 281–287

as pressed and uncompressed graphite felts [15,17] and retic- deionized water at 333 K, followed by drying in air at 333 K.
ulated vitreous carbon [16]. Therefore, the extended reaction To reduce the surface oxides formed during sample treatment,
zone concept has not been thoroughly validated yet by fuel cell electrochemical reductive cleaning of the PtRu(Mo) electrode-
experiments. posited pressed felt was performed for 10 min in 0.5 M H2 SO4
In this context, the objective of the present work was to follow at a constant potential of −0.8 V versus Hg/Hg2 SO4 , K2 SO4std.
up on our previous study, where we presented a novel surfactant reference electrode (abbreviated as MSE) for 10 min [E (V versus
assisted method to electrodeposit PtRu nanoparticles on pressed SHE) = 0.64 + E (V versus MSE)].
graphite felt (thickness ∼350 ␮m) [15], by investigating the per- Scanning electron microscopy was carried out to study the
formance of the graphite felt anode in DMFCs in comparison deposit morphology. The catalyst loading and bulk atomic ratio
with commercial gas diffusion electrodes with carbon black sup- were determined by digesting a sample of deposited felt in aqua
ported PtRu catalyst. Furthermore, the nanoparticle preparation regia at 363 K for 3 h followed by inductively coupled plasma
procedure was extended to the ternary system PtRuMo and the atomic emission spectroscopy (ICP-AES) analysis. The effec-
role of Mo in the electrocatalyst formulation was studied in both tive Pt surface area was estimated by the Cu underpotential
half-cell and fuel cell experiments. deposition and anodic stripping method [15,16,22].
Interestingly, while there are a number of publications reveal-
ing the positive effect of Mo on methanol electrooxidation 2.2. Electrochemical half-cell experiments
kinetics in half-cell experiments [18–21], there is generally a
lack of longer-term fuel cell studies involving PtRuMo. Neto A PARSTAT 2263 potentiostat controlled by PowerSuite®
et al. showed by cyclic voltammetry at 293 K that PtRuMo software (Princeton Applied Research) was employed. The
nanoparticles prepared by a colloidal method with 1:1:1 atomic working electrode was 1 cm2 of pressed felt with PtRu or
ratio gave a 100 mV lower methanol oxidation onset potential PtRuMo electrodeposited on it, immersed in 0.1 M H2 SO4 . Two
compared to PtRu (1:1) [19]. This observation is corroborated graphite rods of 20 cm2 total geometric area served as counter
by the findings of Lima et al., who identified by in situ IR electrodes and an MSE was employed via a Luggin capillary as
reflectance spectroscopy the presence of CO2 from CH3 OH the reference electrode. Cyclic voltammetry was carried out in
oxidation at a 100 mV lower potential when Mo was present 0.1 M H2 SO4 at 5 mV s−1 to provide a blank scan. Afterwards
in a PtRu/polyaniline catalyst. Furthermore, the same authors 0.5 M CH3 OH was added to the electrolyte and cyclic voltam-
reported that PtRuMo was less sensitive to COad poisoning [20]. metry and chronopotentiometry were carried out to characterize
Zhang et al. confirmed the beneficial Mo effect by chronopoten- the catalyst performance. Experiments were performed at both
tiometry at 100 A m−2 , reporting approximately a 50–100 mV 298 and 343 K.
lower anode potential in case of methanol oxidation on PtRuMo
compared to PtRu [21]. 2.3. Fuel cell experiments
Thus, in terms of the ternary catalyst composition PtRuMo,
the present work attempts to bridge a gap between fundamental The fuel cell performance of the catalyzed felt anodes was
electrochemical investigations and fuel cell experiments. evaluated employing a 5 cm2 area experimental DMFC with
gold plated stainless steel end plates having serpentine type flow
2. Experimental channels. The Nafion® 117 membrane had a 40 g m−2 Pt black
catalyst loading on the cathode side (Lynntech Inc.). As back-
2.1. Electrodeposition ing/diffusion layers untreated carbon cloth and ELAT® were
employed on the anode and cathode side, respectively. Temper-
Before deposition the pressed graphite felt (5 cm2 geomet- ature, oxidant flow rate, cathode backpressure and the electronic
ric area, 350 ␮m thickness, Test Solutions Inc.) was rinsed with load (current) were set and controlled by the FC PowerTM soft-
methanol (Fisher) and deionized water followed by air drying in ware associated with the Fideris Inc. test station. The cathode
an oven at 333 K. The electrodeposition media contained 40 wt% pressure was 2 atm (abs). Dry O2 was supplied at a flow rate of
Triton X-100 (C14 H22 O(C2 H4 O)n , n ≈ 9.5) (Aldrich) and metal 500 ml min−1 STP. The anolyte, consisting of 1 M CH3 OH in
salts: 65 mM H2 PtCl6 ·6 H2 O (99.9% Aldrich), 65 mM RuCl3 ·3 0.5 M H2 SO4 , was circulated at 5 ml min−1 and ambient pres-
H2 O (99.9% Alfa-Aesar) while in the case of the ternary cat- sure employing a peristaltic pump.
alyst, 32.5 mM MoCl5 ·5 H2 O (98% Aldrich) was also added.
The deposition was carried out twice, in galvanostatic mode at 3. Results and discussion
60 A m−2 and 333 K for 1.5 h each utilizing a sandwich type cell
with two platinized Ti plates functioning as counter electrodes. 3.1. PtRu and PtRuMo electrodeposition on graphite felt
Between the first and second deposition steps the graphite felt
electrode was washed with deionized water and methanol and As shown by high resolution scanning electron microscopy,
dried in air. The temperature for both washing and drying was the surfactant assisted electrodeposition technique produced
333 K. The second deposition step was carried out employing meso-porous PtRu and PtRuMo coatings on the fiber surface,
fresh electrodeposition media. consisting of particles and agglomerates with approximate diam-
After the two consecutive electrodeposition steps the felt was eters ranging from 10 to100 nm (Fig. 1). It must be noted that
sonicated in methanol and rinsed thoroughly with methanol and transmission electron microscopy images could not be obtained
 A. Bauer et al. / Journal of Power Sources 167 (2007) 281–287 283

Fig. 2. Anodic stripping of the underpotential-deposited Cu monolayer on the

graphite felt electrode with PtRu (a) and PtRuMo (b). Scan rate 0.5 mV s−1 .
Fig. 1. HiRes SEM micrographs of PtRu (a) and PtRuMo (b) deposits on the
pressed graphite fiber surface. Fig. 2 shows the Cu stripping voltammograms used for the
effective surface area calculation. Moreover, the blank scan in
because slices of adequate thickness (∼100 nm) for this type of 0.1 M H2 SO4 conducted with the Mo containing catalyst showed
analysis could not be prepared from the catalyzed felt material. an anodic peak at −0.2 V versus MSE, indicating Mo oxida-
The properties of the binary and ternary catalyst are listed in tion (compare Fig. 2b and a). It is noteworthy that Cu did
Table 1. The total catalyst load was approximately 20% higher not underpotential-deposit on Mo, therefore, its surface area
in the case of the PtRuMo composition, while the Pt load of could not be determined by the employed method. In separate
the two catalysts was almost identical. Interestingly, the effec- experiments utilizing only Ru catalyst supported on graphite
tive Pt surface area, as determined by anodic stripping of the felt, a pronounced stripping peak at −0.26 V versus MSE was
underpotential-deposited Cu monolayer, was almost three times obtained. However, when the method was applied to either PtRu
lower in case of PtRuMo compared to PtRu (Table 1). This indi- or PtRuMo, it did not yield a distinct peak in the potential range
cates extensive surface segregation of Mo in the ternary catalyst related to Ru (i.e. between −0.34 V versus MSE and −0.17 V
formulation. versus MSE). This result is consistent with the findings of Cheng
and Gyenge regarding PtRu supported on reticulated vitreous
carbon [16]. It is therefore assumed that the area determined by
Table 1
Physico-chemical characteristics of the binary and ternary catalysts electrode-
Cu UPD and stripping reflects the active Pt area of the respective
posited on the pressed felt by the Triton X-100 assisted method binary and ternary catalyst formulations.
Physico-chemical property Pt:Ru Pt:Ru:Mo
3.2. Effect of Triton X-100 on the electrodeposition of Pt,
Atomic ratio 1.4 : 1 1:1:0.3
Ru and Mo on graphite felt
Total catalyst load [g m−2
felt ] 43 52
Pt load [g m−2
felt ] 32 32
Effective Pt surface area [m2 m−2 In order to obtain information regarding the influence of Tri-
felt ] 712 254
Mass specific Pt surface area [m2 g−1
ton X-100 on the electrodeposition of the various metal ions,
Pt ] 23 8
cathodic polarization experiments were carried out at 333 K
284 A. Bauer et al. / Journal of Power Sources 167 (2007) 281–287

Fig. 4. Effect of Triton X-100 on the voltammogram of PtRu (a) and PtRuMo
(b) codeposition on graphite felt. Scan rate 5 mV s−1 .

MSE where the metal deposition occurs with high current effi-
ciency, the largest cathodic current was obtained for PtCl6 2− .
Thus, as discussed also previously [15], in the absence of Triton
X-100 the electrodeposition of Pt is favored forming a Pt-rich
catalyst on the graphite felt.
The gradual increase of Triton X-100 concentration from 0
to 40 wt% resulted in a significant decrease of the Pt deposi-
tion current density (Fig. 3a). At 40 wt% Triton X-100 the Pt
deposition polarization curve became linear up to −0.8 V with a
large dE/di ratio (Fig. 3a). This is due to the large crystallization
Fig. 3. Effect of Triton X-100 on the voltammogram of Pt (a), Ru (b) and Mo overpotential as a result of low ad-atom surface diffusivity on
(c) electrodeposition on graphite felt. Scan rate 5 mV s−1 . the surfactant-covered surface [23], leading to isolated nuclei
formation followed by restricted growth of nuclei to three-
and 5 mV s−1 on 1 cm2 pressed felt in the metal salt solutions, dimensional crystallites and possible coalescence into larger
H2 PtCl6 , RuCl3 and MoCl5 both individually (Fig. 3) and in aggregates (nucleation-coalescence mechanism for electrode-
combination (Fig. 4). The concentration of each metal species position).
was 1 mM. The effect of surfactant on the electrodeposition current of
Fig. 3 shows the scans for the individual metal depositions. either Ru or Mo was less pronounced compared to Pt (compare
In the absence of surfactant, for all three species essentially two Fig. 3a–c). This indicates that the electroreduction kinetics of
cathodic waves can be distinguished. At potentials more pos- both Ru and Mo ions are slow, hence, their deposition current
itive than −0.8 V versus MSE, primarily the electrodeposition density was less affected by ad-atom diffusion limitation and
of Pt, Ru and Mo takes place, while at more negative poten- restricted growth effects.
tials the secondary reaction of H2 evolution gains significance, Fig. 4 shows the voltammograms obtained for codeposition of
thereby, lowering the deposition current efficiency. In the poten- PtRu (Fig. 4a) and PtRuMo (Fig. 4b), respectively. In the case of
tial domain of interest, i.e. between −0.4 and −0.8 V versus both PtRu and PtRuMo, without surfactant present in the depo-
 A. Bauer et al. / Journal of Power Sources 167 (2007) 281–287 285

sition bath the codeposition current at potentials between −0.6

and −0.8 V (Fig. 4) is fairly close to the sum of the individ-
ual deposition currents of the constituent elements, therefore, a
Pt-rich catalyst would be generated (Fig. 3a–c). With 40 wt%
Triton X-100, while the codeposition current is again approx-
imately equal to the sum of the individual deposition currents
of the respective elements, due to the low Pt deposition cur-
rent (Fig. 3a), the ratio of the co-deposited elements is different.
Therefore, the presence of surfactant was crucial to control the
Pt:Ru and Pt:Ru:Mo atomic ratio in the deposit, by selectively
lowering the Pt deposition current density compared to Ru and
Mo (Fig. 3a–c).
In the case of the PtRu codeposition without Triton X-100
(Fig. 4a) a nucleation peak was observed at −0.2 V as well as
a shoulder wave between −0.45 and −0.55 V, which represents
crystal growth. When the surfactant concentration was increased
to 1 and 10 wt% nucleation became more difficult as indicated by
the shift of the nucleation wave to more negative potentials. At Fig. 6. Chronopotentiometry of methanol electro-oxidation using PtRu and
PtRuMo catalysts deposited on compressed felt: 0.5 M CH3 OH–0.1 M H2 SO4 ,
40 wt% Triton X-100 both the nucleation and growth peaks dis-
50 A m−2 .
appeared due to isolated nucleation sites, low surface diffusivity
of ad-atoms and restricted growth.
adsorbs on Mo, which is thought to have bifunctional properties
similar to Ru, hence, facilitating the formation of OHad [25].
3.3. Methanol electro-oxidation experiments
Furthermore, the role of Mo has been also explained based on
the assumption that the activation barrier for the oxidation of
Cyclic voltammograms for methanol oxidation on PtRu and
COad is lowered due to an oxygen spillover effect [26].
PtRuMo are presented in Fig. 5, at 298 and 343 K, respectively.
Increasing the temperature from 298 to 343 K had the most
The forward (anodic) scans were virtually identical with the
pronounced effect on the Mo containing catalyst, enhancing the
reverse (cathodic) scans, therefore, only the former are pre-
CH3 OH oxidation superficial current density by up to five or
sented.
six times at potentials more positive than −0.4 V versus MSE
At 298 K the PtRu catalyst yielded significantly higher oxida-
(Fig. 5).
tion currents than PtRuMo. This can be explained in part by the
Additionally, chronopotentiometry of methanol oxidation
larger Pt surface area of the PtRu catalyst (Table 1) available for
was performed on the two catalysts at 298 and 343 K, respec-
methanol adsorption. It has been established that at low temper-
tively (Fig. 6). The latter experiment is more relevant for fuel
atures (e.g. 298 K) methanol adsorption and dehydrogenation is
cell operation compared to cyclic voltammtery, since it simu-
not favored on Ru. This was elegantly shown by first principle
lates the anode potential variation in time at a constant current
quantum mechanics calculation of binding energies and heats
density. In accordance with the cylic voltammetry data, at 298 K
of formation [24]. It is unknown, however, whether methanol
and 50 A m−2 the anode potential of the PtRuMo catalyst was
more positive by about 60 mV. However, at 343 K an approxi-
mately 10–30 mV lower anode potential was obtained with the
Mo containing catalyst (Fig. 6), while having the same Pt load
for both catalyst formulations (Table 1). Thus, there is a strong
interaction effect between the presence of Mo and temperature.
Addition of Mo to the PtRu catalyst formulation is beneficial
only at higher temperatures such as 343 K and above. The pos-
itive interaction effect between temperature and Mo content
seems to follow the observation of Dickinson et al. regarding
Ru, in that a higher Ru content (i.e. PtRu ratio of 1:1 versus
1.5:1) enhanced the anode performance at high temperatures
(e.g. 338 K) [27]. This is related to the rate determining step shift
from CH3 OH adsorption/dehydrogenation at low temperatures
to the reaction of COad with OHad at high temperatures.

3.4. Fuel cell experiments

Fig. 5. Voltammogram of methanol electro-oxidation using PtRu and PtRuMo Fuel cell polarization curves were obtained for the two
catalysts deposited on compressed felt: 0.5 M CH3 OH–0.1 M H2 SO4 , 5 mV s−1 . pressed graphite felt supported catalysts (PtRu and PtRuMo,
286 A. Bauer et al. / Journal of Power Sources 167 (2007) 281–287

Fig. 7. Fuel cell polarization experiments at 333 K. Anode comparison: com- Fig. 8. Fuel cell polarization experiments at 343 and 353 K using the pressed felt
mercial gas diffusion electrode vs. PtRu and PtRuMo deposited onto pressed supported PtRu and PtRuMo catalysts. Anode feed: 1 M CH3 OH–0.5 M H2 SO4 ,
felt. Anode feed: 1 M CH3 OH–0.5 M H2 SO4 , 5 ml min−1 , ambient pressure; 5 ml min−1 , ambient pressure; cathode feed: dry O2 , 500 ml min−1 STP, 2 atm.
cathode feed: dry O2 , 500 ml min−1 STP, 2 atm.

trode, having a power output of 600 W m−2 at 3000 A m−2 and

respectively) as well as for a commercially available PtRu gas 363 K [31].
diffusion electrode (GDE) of comparable load and composition Fig. 8 shows that the Mo presence in the catalyst formula-
(40 g m−2 , 1:1 atomic ratio) (Lynntech Inc.). tion of the pressed felt anode was beneficial for the fuel cell
Fig. 7 compares the performance of the gas diffusion elec- power output at higher temperatures (i.e. especially at 353 K)
trode with pressed felt at 333 K. At the latter temperature there in agreement with the voltammetric and chronopotentiometric
was virtually no difference between the PtRu and PtRuMo cata- experiments.
lysts electrodeposited on graphite felt, supporting therefore, the At 343 and 353 K the open circuit voltages were 0.72 and
conclusions of the voltammetry and chronopotentiometry exper- 0.74 V without Mo and 0.74 and 0.77 V when using PtRuMo,
iments. However, the pressed felt electrodes gave significantly respectively. At a current density of 5500 A m−2 and 353 K the
better performance compared to the gas diffusion electrode. The power density with the PtRuMo catalyst was 2200 W m−2 while
open circuit voltage was 0.67 V for the commercial electrode and the binary PtRu catalyst yielded 1925 W m−2 (Fig. 8).
0.70–0.71 V for the graphite felt supported PtRu and PtRuMo We could find only one literature report on the activity of
containing anodes. Furthermore, at 3000 A m−2 and 333 K the PtRuMo utilized in a DMFC. Employing 20 g m−2 PtRuMo dis-
fuel cell power density was enhanced by 38% with the pressed persed on polyaniline, Lima et al. obtained a peak power output
felt anode, from 870 W m−2 (i.e. peak power for the GDE) to of only about 200 W m−2 at 383 K [18], i.e. about an order of
1200 W m−2 . The maximum power output of the novel extended magnitude lower than in the present work carried out at 353 K.
reaction zone anode was about 1500–1600 W m−2 . Thus, it can be stated that the interaction between the pressed
It must be noted that the performance of the reference graphite felt support and either PtRu or PtRuMo points toward a
GDE can be considered representative of the state-of-the-art promising direction in the effort of improving the power output
of these electrode types. In the literature, for GDEs with car- of DMFCs. Moreover, the three-dimensional electrode design
bon black (Vulcan XC-72) supported PtRu anode catalysts of presented here could be advantageous for flow-through mixed
40–50 g m−2 load, peak power outputs at 333 K of 250 W m−2 reactant DMFCs [32].
[28], 600–750 W m−2 [29] and 1000 W m−2 [30] were reported.
Therefore, it can be stated that the extended reaction zone anode 3.5. Fuel cell anode durability
gave clearly a better performance. This can be most likely
attributed to the higher utilization of the nanoparticle catalyst, Generally, very few literature reports show the performance
as shown also by Lycke and Gyenge for direct ethanol fuel cells of direct fuel cells over extended periods of time. Performance
[17]. Moreover, although not proven experimentally yet, it is evaluations are typically made based on polarization curves,
proposed that the extended reaction zone lowers the methanol which are recorded over a time frame of less than an hour. To
crossover rate. Thereby, the negative effect of the mixed cathode address this issue in the present work, a constant superficial
potential is to some extent mitigated. current density of 4000 A m−2 was applied for 3 h to moni-
Due to the novelty of the employed anode design, there are tor the anode catalyst deactivation during fuel cell operation
very few literature results that could be used for comparison. at 353 K (Fig. 9). While the PtRuMo catalyst performed bet-
Scott and co-workers reported DMFC polarization curves using ter during polarization experiments, its cell voltage decreased
PtRu (10 g m−2 , 1:1 atomic ratio) supported on Ti mesh, which at a higher rate. Assuming an approximately linear correla-
could be also considered a three-dimensional extended reac- tion between voltage loss and time, the degradation rate was
tion zone electrode type. The Ti mesh anode gave virtually no 26 mV h−1 in the case of PtRuMo and 13 mV h−1 for PtRu,
improvement compared to a conventional gas diffusion elec- respectively.
 A. Bauer et al. / Journal of Power Sources 167 (2007) 281–287 287

Acknowledgments

The authors gratefully acknowledge the financial support of

the BC Advanced Systems Institute and the Natural Sciences
and Engineering Research Council of Canada.

References

[1] G. Olah, A. Goeppert, G.K. Surya Prakash, Beyond Oil and Gas: The
Methanol Economy, Wiley-VCH Verlag GmbH & Co., Weinheim, 2006.
[2] H. Li, C. Oloman, J. Appl. Electrochem. 36 (2006) 1105.
[3] H.A. Gasteiger, N. Markovic, P.N. Ross, E.J. Cairns, J. Phys. Chem. 97
(1993) 12020.
[4] T. Iwasita, H. Hoster, A. John-Anacker, W.F. Lin, W. Vielstich, Langmuir
16 (2000) 522.
[5] D. Cao, G.-Q. Lu, A. Wieckowski, S.A. Wasilevski, M. Neurock, J. Phys.
Chem. B 109 (2005) 11622.
[6] C. Bock, B. MacDougall, Y. LePage, J. Electrochem. Soc. 151 (2004)
A1269.
Fig. 9. Longer-term fuel cell polarization experiments at 4000 A m−2 and 353 K. [7] H. Wang, C. Wingender, H. Baltruschat, M. Lopez, M.T. Reetz, J. Elec-
Pressed felt anode with PtRuMo and PtRu. Conditions identical to Fig. 8. troanal. Chem. 509 (2001) 163.
[8] J. Jiang, A. Kucernak, J. Electroanal. Chem. 543 (2003) 187.
[9] B. Gurau, E.S. Smotkin, J. Power Sources 112 (2002) 339.
[10] J.L. Fernandez, V. Raghuveer, A. Manthiram, A.J. Bard, J. Am. Chem. Soc.
After 3 h of continuous operation the cell voltages for
127 (2005) 13100.
PtRuMo and PtRu became almost identical due to degradation [11] P. Argyropoulos, K. Scott, W.M. Taama, J. Appl. Electrochem. 29 (1999)
of the PtRuMo anode performance, whilst the cell voltage with 661.
the PtRu catalyst stabilized after 2 h (Fig. 9). The lower stabil- [12] W.M. Yang, S.K. Chou, C. Shu, J. Power Sources 164 (2007) 549.
ity of PtRuMo could be explained by possible Mo oxidation to [13] T. Schultz, U. Krewer, T. Vidakovic, M. Pfafferodt, M. Christov, K. Sund-
macher, J. Appl. Electrochem. 37 (2007) 111.
Mo3+ as suggested also by the voltammogram in 0.1 M H2 SO4
[14] C.Y. Chen, J.Y. Shiu, Y.S. Lee, J. Power Sources 159 (2006) 1042.
(Fig. 2b). [15] A. Bauer, E.L. Gyenge, C.W. Oloman, Electrochim. Acta 51 (2006) 5356.
[16] T.T. Cheng, E.L. Gyenge, Electrochim. Acta 51 (2006) 3904.
4. Conclusions [17] D.R. Lycke, E.L. Gyenge, Electrochim. Acta 52 (2007) 4287.
[18] A. Lima, C. Coutanceau, J.-M. Leger, C. Lamy, J. Appl. Electrochem. 31
(2001) 379.
Surfactant assisted galvanostatic electrodeposition of PtRu
[19] A.O. Neto, E.G. Franco, E. Arico, M. Linardi, E.R. Gonzales, J. Eur. Ceram.
and PtRuMo onto pressed graphite felt (thickness 350 ␮m) Soc. 23 (2003) 2987.
yielded particles and agglomerates of ∼10–100 nm diame- [20] A. Lima, F. Hahn, J.-M. Leger, Russ. J. Electrochem. 40 (2004) 326.
ter. Fuel cell polarization experiments revealed a significant [21] X. Zhang, F. Zhang, K.-Y. Chan, J. Mater. Sci. 39 (2004) 5845.
improvement in power output with the extended reaction zone [22] C.L. Green, A. Kucernak, J. Phys. Chem. B 106 (2002) 1036.
[23] H. Fischer, Angew. Chem. Int. Ed. Engl. 8 (1969) 108.
graphite felt anode compared to a conventional gas diffu-
[24] J. Kua, W.A. Goddard III, J. Am. Chem. Soc. 121 (1999) 10928.
sion electrode having approximately the same Pt loading. At [25] H. Bolivar, S. Izquierdo, R. Tremont, C.R. Cabrera, J. Appl. Electrochem.
3000 A m−2 and 333 K the fuel cell power density was enhanced 33 (2003) 1191.
by 38% with the novel anode. The PtRuMo catalyst showed a [26] G. Samjeske, H. Wang, T. Löffler, H. Baltruschat, Electrochim. Acta 47
strong positive interaction effect with temperature, performing (2002) 3681.
[27] A.J. Dickinson, L.P.L. Carrette, J.A. Collins, K.A. Friedrich, U. Stimming,
better than PtRu at higher temperatures, such as 343 and 353 K,
J. Appl. Electrochem. 34 (2004) 975.
respectively. The power density obtained with PtRuMo sup- [28] C. Coutanceau, A.F. Rakotondrainibe, A. Lima, E. Garnier, S. Pronier,
ported on pressed graphite felt was 2200 W m−2 at 3000 A m−2 J.-M. Leger, C. Lamy, J. Appl. Electrochem. 34 (2004) 61.
and 353 K. However, a higher anode catalyst degradation rate [29] V. Baglio, A. Di Blasi, E. Modica, P. Creti, V. Antonucci, A.S. Arico, Int.
was observed for the Mo containing catalyst. Therefore, this J. Electrochem. Sci. 1 (2006) 71.
[30] http://www.etek-inc.com/pdfs/MEASeries12D-W.pdf.
work showed that Mo based anode catalysts could be practi-
[31] R.G. Allen, C. Lim, L.X. Yang, K. Scott, S. Roy, J. Power Sources 143
cal in DMFCs only if an adequate catalyst regeneration method (2005) 142.
could be devised or a more stable alloy formulation could be [32] M.A. Priestnall, V.P. Kotzeva, D.J. Fish, E.M. Nilsson, J. Power Sources
identified. 106 (2002) 21.