Carbon support oxidation in PEM fuel cell cathodes
Introduction
PEM fuel cell systems running on hydrogen are seen as a future alternative to conventional combustion engines for automotive application. While first fuel cell products for off-grid power supply are already commercially available, market launch of transportation applications is delayed due to tough requirements regarding costs, power density, robustness and long-term stability.
Platinum is still the only known catalyst material providing sufficiently high activity for the oxygen reduction reaction. To reduce the amount of precious metal used, platinum is usually supported on carbon in the form of nano-dispersed particles. This dispersement allows for high-catalyst surface areas at low-catalyst loadings. Carbon supported catalysts are, however, susceptible to catalyst particle agglomeration. Another major drawback of high-performance, low-loaded electrodes resides in the thermodynamic instability of carbon under PEMFC cathode conditions [1]. Above 0.207 V versus NHE carbon can be oxidized to carbon dioxide following reaction (1). Oxidation to carbon monoxide (2) is thermodynamically not favoured according to reaction (3):C + 2H2O → CO2 + 4H+ + 4e−, φ00 = 0.207 VC + H2O → CO + 2H+ + 2e−, φ00 = 0.518 VCO + H2O → CO2 + 2H+ + 2e−, φ00 = −0.103 V
Oxidation of carbon support – generally referred to as carbon corrosion – can lead to performance decrease due to accelerated loss of active surface area [2] and alteration of pore morphology and pore surface characteristics [3]. Support degradation can be initiated under conditions where the electrode is exposed to high-electrochemical potentials, as is the cathode at high-cell potentials. In automotive applications this generally corresponds to low-electric power demand, which makes up a significant portion of the fuel cells total operational time. Two distinct states of operation have to be considered separately, resulting in extremely strong performance decay due to corrosion incidents. The first state is characterized by complete fuel starvation of one or more cells in a working stack. The starved cells are driven into reversed operation with the anode potential being higher than the cathode potential. This state entails water electrolysis and carbon oxidation on the fuel cell anode in order to provide the required protons and electrons for the reduction of oxygen on the cathode. Sustaining the stack current results in massive oxidation of the anode catalyst support. Cell reversal manifests itself in negative cell voltage which can drop as low as −2.0 V [4]. The second state of irregular high-electrode potentials is characterized by partial fuel starvation of the active anode area of an individual cell in the stack. This condition may be caused either by local undersupply of hydrogen [5], [6] or a hydrogen-air front passing over the active area during start up and shut down of the fuel cell [6], [7]. Because oxygen-permeation from the cathode over the membrane into the anode compartment cannot be prevented, oxygen is present in the hydrogen deficient areas on the anode. As a consequence, potential shifts can occur caused by a drop of the in-plane membrane potential. This mechanism was already described for phosphoric acid fuel cells (PAFC) [8] and has recently been verified and simulated for PEM fuel cells [6], [9]. Complete cell reversal and local fuel starvation both result in significant performance losses within a couple of minutes and must be continuously ruled out, e.g. by an appropriate operating strategy and by choice of adequate electrode materials [9], [10], [11].
Potentiostatic electrooxidation of carbon supports under PAFC conditions has been intensely studied in the literature [12], [13], [14], [15], [16]. Gas diffusion electrodes immersed in liquid electrolyte have been used to simulate PEM fuel cell conditions [3]. Gas phase oxidation has been regarded as an accelerated test to evaluate carbon materials for use as catalyst support in low-temperature fuel cells [17]. Few literature is, however, available on carbon corrosion in real PEM fuel cell environment under potentiodynamic operating conditions. Willsau and Heitbaum made a very thorough analysis of carbon oxidation in gas diffusion electrodes in liquid electrolyte under potential sweep operation [18]. They measured an accelerating effect of platinum on corrosion rates and were able to identify several oxidation peaks by differential electrochemical mass spectroscopy (DEMS). Roen et al. got similar results with PEMFC electrodes in helium atmosphere but did not offer an explanation for all observed phenomena [19].
The scope of this paper is to evaluate carbon support oxidation under potential transients in realistic PEM fuel cell environment, i.e. in the presence of a solid ionomer electrolyte and water in both the gas and liquid phases. All experiments were therefore conducted in situ using membrane-electrode-assemblies. Carbon corrosion was measured by monitoring the evolution of carbon dioxide and carbon monoxide at the cell outlet.
Section snippets
Experimental
In-house fabricated membrane-electrode-assemblies were employed for all measurements. Catalyst and support materials to be evaluated were incorporated in the cathode catalyst layer. The results presented in this paper are based on pure Ketjenblack EC300J®-electrodes (Akzo Nobel, 736 m2 g−1 BET surface area measured, literature value 930 m2 g−1 [20]) and Ketjenblack-supported platinum-electrodes (50% Pt/C, Tanaka). Both materials were purchased in powder form and catalyst inks were prepared with 20
Theory
As already stated in Section 1, carbon is used as support for the active fuel cell catalyst component, usually platinum or platinum alloys. The surface area per unit weight of noble metal is increased by keeping it in a dispersed condition. Sp2-hybridized carbon is the most commonly used support material because of its high-electrical conductivity, its good processability and its availability in a wide variety of particle morphologies [21]. Furthermore carbon shows relatively high chemical and
Results and discussion
Dynamic measurements and measurements at constant potential were conducted at varied temperature and humidity and in different potential windows. Carbon dioxide evolution could be detected on Pt/C- as well as on pure C-electrodes. In contrast to that, carbon monoxide was only formed on pure carbon electrodes and only in concentrations of one tenth of carbon dioxide (see, e.g. Fig. 3). In presence of platinum carbon monoxide could not be detected because it is adsorbed on the metal surface at
Conclusions
The scope of the article is to present a dynamic measurement technique to evaluate carbon corrosion in PEM fuel cell cathodes. In contrast to earlier approaches, measurements were conducted in nitrogen and in synthetic air. This approach turned out to be essential to better understand corrosion mechanisms. It was shown that hydrogen peroxide formed on carbon support and platinum catalyst leads to increased corrosion at low potentials. The absolute value of the anodic potential limit strongly
Acknowledgements
Colleagues at DaimlerChrysler AG, Department of MEA and Stack Technology are acknowledged for many discussions and support. Prof. Dr.-Ing. C. Merten is thanked for scientific supervision of this work.
References (38)
- et al.
J. Power Sources
(2004) - et al.
Carbon
(1973) - et al.
J. Power Sources
(1996) - et al.
Carbon
(2005) - et al.
J. Electroanal. Chem.
(1984) - et al.
Catal. Today
(1990) - et al.
Appl. Catal. A
(1998) Carbon
(1998)- et al.
J. Electroanal. Chem.
(2001) Electrochim. Acta
(1984)