Influence of the microporous layer on carbon corrosion in the catalyst layer of a polymer electrolyte membrane fuel cell

Abstract

Corrosion of the catalyst support reduces PEM fuel cell performance via catalyst layer (CL) degradation (loss of porosity, catalyst connectivity, and active catalyst surface area). Carbon corrosion was investigated in a segmented cell for cathode gas diffusion layers (GDLs) with and without a microporous layer (MPL) to investigate the spatial aspects of GDL effect on corrosion. The cells were aged in situ using an accelerated stress test (AST) for carbon-support corrosion consisting of consecutive holds at 1.3 V. Carbon corrosion was quantified by measuring CO₂ evolution during the AST.

Performance degradation was substantial both with and without cathode MPL, but the degradation of the CL after prolonged corrosion was lower in the presence of an MPL. This was corroborated by better cell performance, higher remaining Pt active area, lower kinetic losses and smaller Pt particle size. The cell with an MPL showed increasingly nonuniform current distribution with corrosion time, which is correlated to the distribution of the Pt particle growth across the active area. This cell also showed an increase in mass-transport resistance due to MPL degradation. Without an MPL, GDL carbon fibers caused localized thinning in the cathode CL, originating from the combined effects of compression and corrosion.

Introduction

Polymer electrolyte membrane (PEM) fuel cells are energy conversion devices which have shown great promise for automotive, stationary and portable power applications. Cost-effective commercialization of PEM fuel cells requires reductions in fuel cell cost and improvements in durability, however with a minimal impact on the state-of-the-art initial fuel cell performance. Thus, over the past decade, more research has been focused on improving the durability of the fuel cell systems and components [1], [2], [3], [4], [5], [6]. Durability of the catalyst layer (CL) remains one of the primary challenges in developing PEM fuel cell systems with acceptable lifetimes.

Current US DOE durability targets for fuel cell systems (with defined acceptable degradation rate) are 5000 h (∼7 months) for automotive and 40,000 h (∼4.6 years) for stationary applications. Considering the time and cost required for real-time testing of system lifetimes (so called ‘life tests’), there is an urgent need for establishing standardized protocols that allow screening the durability characteristics of component materials in a reasonable turnaround time. The goal is to decouple the degradation mechanisms and separately evaluate the degradation rates for individual cell components. Such an approach requires fundamental understanding of the degradation mechanisms. Several accelerated stress tests (ASTs) have been developed to enable rapid evaluation of degradation rates for individual components of a fuel cell system [1], [4], [6], [7], [8]. Although these durability protocols are not yet fully standardized, four ASTs have been proposed to characterize the (i) catalyst (potential cycling); (ii) catalyst support (elevated voltage hold); (iii) membrane-electrode assembly (MEA) chemical stability (open-circuit operation); and (iv) membrane mechanical stability (relative humidity cycling). The ASTs are crucial for quantifying tradeoffs between the cost, durability and performance. In order to quantitatively correlate the degradation rates in real system testing (i.e. field data) to the accelerated stress testing [8], one needs to take into account the effects of additional stressors (e.g. start/stop, temperature, and relative humidity).

In this work we consider a catalyst layer commonly used in PEM fuel cells; a porous structure consisting of carbon-supported platinum catalyst in an ionomer matrix [9], [10], [11], [12]. Corrosion of the catalyst support in PEM fuel cells reduces the overall cell performance as it leads to the loss of the active catalyst surface area and catalyst connectivity, and lowers the porosity and hydrophobicity of the catalyst layer [1], [2], [3], [4], [9], [13], [14]. Thus, carbon-support corrosion is a key degradation mechanism in PEM fuel cells.

Carbon corrosion in aqueous acid electrolyte (such as a proton-exchange membrane) is described by [15]:

$$\text{C}+2{\text{H}}_2\text{O}\to {\text{CO}}_2+4{\text{H}}^{+}+4{\text{e}}^{-},{\text{E}}_0=0.207{\text{V}}_{\text{RHE}}$$

At sufficient anodic potential, carbon dioxide is produced from the complete oxidation of elemental carbon by water. Although the corrosion is accelerated by Pt catalyst [16], in the typical range of cathode potentials this reaction is slow due to low operating temperature of PEM fuel cells. However, certain areas of the cathode CL may experience high potentials (versus reference hydrogen electrode, RHE) during cell operation, yielding high corrosion rates which cause a major damage to the catalyst support. Such situations occur when the anode compartment of a single cell contains both hydrogen-starved and hydrogen-rich regions. Accelerated carbon corrosion proceeds locally in the cathode region corresponding to the hydrogen-starved region of the anode [17]. The proposed mechanism was termed ‘the reverse-current mechanism’ [18] as the H₂/air portion of the cell drives the reverse current in the hydrogen-starved region. During the steady-state operation, localized fuel starvation may occur due to anode flooding or flow maldistribution [17], [19], [20], [21]. Patterson and Darling [17] simulated such a condition by restricting the hydrogen access in a small portion of the anode catalyst layer in a cell operating at 0.4 A cm⁻² for 100 h. They demonstrated that the cathode CL corresponding to the hydrogen-depleted area suffered the loss of carbon and significant thinning.

Regions of the cathode CL may experience high-potential excursions not only during steady-state operation, caused by localized fuel starvation, but also during transient operation, i.e. cell startup and shutdown (SU/SD). A fuel/air boundary on the anode side of the cell during SU/SD operation can induce local potentials at the cathode typically as high as 1.2–1.5 V [18], [22], [23], [24], [25], [26]. Although these transients are relatively short (residence time <0.1 s [27]), cumulative time at elevated potentials may cause significant damage to the CL in systems which require frequent startups and shutdowns, such as in automotive applications. The use of graphitized carbons as the catalyst support increases the CL resistance to carbon corrosion [7], [26], [28], however with the penalties of (i) higher material cost (thermal treatment in the fabrication process) and often (ii) lower initial performance (due to lower carbon surface area and consequently lower electrochemical catalyst surface area). Since the improvement in carbon-support stability will usually come at the expense of cost and/or initial performance, various system-level strategies have been proposed to mitigate the long-term effects of carbon corrosion [9].

Since the local conditions (such as humidity, temperature, and reactant composition) vary across the active area during cell operation, it is expected that different regions of the cell will experience varying rates of degradation. It is therefore important to provide spatially resolved measurements across the cell area to evaluate the degradation levels due to operational variables. Segmented cells have been used extensively as a diagnostic tool in a variety of fuel cell studies (e.g. mass-transport optimization, water and thermal management, flow field design and model validation) [29], [30], [31]. More recently, current distribution measurements have also shown to be invaluable in elucidating the degradation mechanisms [22], [27], [32], [33], [34]. By using an approach of simulated localized fuel starvation [17] in a segmented cell, Carter et al. [27], [34] measured local polarization curves across the cell area. The hydrogen-starved region exhibited significantly reduced limiting current (due to oxygen mass-transport limit) compared to the rest of the cell. This localized performance degradation was correlated to the cathode CL thinning and the loss of active area due to carbon corrosion. Lamibrac et al. [27] reported higher CL degradation in the outlet region after repeated cell startups, by measuring the internal currents and EIS (electrochemical impedance spectroscopy). A visualization technique was employed by Ishigami et al. [32] to measure the distribution of oxygen partial pressure in the flow field and record the progression of H₂/air front during SU/SD operation. The inlet and especially outlet region were identified as critical regions that suffered higher levels of performance degradation due to CL damage during SU/SD cycling.

For the most part, carbon corrosion studies have focused on the catalyst support, i.e. carbon in the catalyst layer. The gas diffusion layer (GDL) and microporous layer (MPL) are often not considered as additional sources of carbon for the carbon corrosion reaction. However, the region of the GDL (or MPL) in direct contact with the CL is also susceptible to corrosion, as it experiences similar corrosive conditions (high potential, humidity, and O₂ concentration) in the proximity of Pt catalyst and electrolyte. The MPL, typically containing amorphous carbon, is more vulnerable than the graphitized carbon fibers in the GDL substrate. Owejan et al. [35] showed that using graphitized carbon, Pureblack^® [36] instead of acetylene black in the MPL, reduced the performance degradation by more than 50% (at 0.6 V) after cumulative 25 h of AST at 1.2 V. The graphitized MPL also performed better at high current after SU/SD cycling, indicating mass-transport (MT) problems caused by corrosion in the acetylene black MPL. Note that it is challenging to distinguish between the performance degradation contributions due to mass-transport issues from carbon corrosion in the CL and carbon corrosion in the MPL.

Carbon corrosion (be it in the CL or the MPL) also makes the surfaces more hydrophilic via roughening and addition of oxide groups resulting in water management challenges [14]. In addition, pore collapse in the MPL after prolonged cell testing (∼650 h) has been suggested from porosimetry measurements (increase in the small pore volume at the expense of the large pore volume) [37]. Although the degradation mechanisms of the GDL are not yet well understood, GDL materials subjected to various ex situ aging protocols exhibited loss of hydrophobicity [1], [38], [39]. Finally, while extensive research has focused on optimizing GDL and MPL for water management [19], [20], [40], more work is needed to elucidate the influence of the GDL material on long-term cell performance and corrosion in the CL.

The present study employs a segmented cell to measure spatial variations in performance degradation as a function of GDL material and AST time. The influence of the MPL on carbon corrosion was evaluated by characterizing the degradation using the same GDL carbon-fiber substrates with and without an MPL. The materials were aged in situ in a 50 cm² cell with a 10 by 10 segmented current collector using an AST protocol for catalyst support, consisting of consecutive holds at 1.3 V in humidified H₂/N₂ at 80 °C. Carbon corrosion during potential steps was quantified by measuring the carbon dioxide content of the nitrogen exhaust stream, coming from the side that serves as the cathode during subsequent H₂/air operation (Section 3.1.1). Performance degradation was characterized in situ by cyclic voltammetry (Section 3.1.2), polarization curves, cell operation at fixed conditions, current distribution, and electrochemical impedance spectra (Section 3.1.3). Ex situ analyses were performed on new and aged MEAs to correlate the changes in Pt particle size, CL thickness, and material morphology to the performance degradation (Section 3.2). These results are useful in elucidating the influence of the MPL on carbon corrosion and CL durability.