Carbon corrosion is an important degradation mechanism that can impair PEMFC performance through the destruction of catalyst connectivity, collapse of the electrode pore structure, loss of hydrophobic character, and an increase of the catalyst particle size. In this study, carbon corrosion was quantified in situ by measurement of carbon dioxide in the fuel cell exhaust gases through non-dispersive infrared spectroscopy during simulated drive cycle operations consisting of potential cycling with varying upper and lower potential limits. These studies were conducted for three different types of carbon supports. A reduction in the catalyst layer thickness was observed during a simulated drive cycle operation with a concomitant decrease in catalyst layer porosity, which led to performance losses due to increased mass transport limitations. The observed thickness reduction was primarily due to compaction of the catalyst layer, with the actual mass of carbon oxidation (loss) contributing only a small fraction (< 20%). The dynamics of carbon corrosion are presented along with a model that simulates the transient and dynamic corrosion rates observed in our experiments. Accelerated carbon corrosion stress tests are presented and their effects are compared to those observed for the drive cycle test.

An electrochemical flow cell system with catalyst-ionomer ink deposited on glassy carbon is used to investigate the aqueous stability of commercial PtCo alloys under cyclic potentials. An on-line inductively coupled plasma-mass spectrometer, capable of real-time measurements, is used to resolve the anodic and cathodic dissolution of Pt and Co during square-wave and triangle-wave potential cycles. We observe Co dissolution at all potentials, distinct peaks in anodic and cathodic Pt dissolution rates above 0.9 V, and potential-dependent Pt and Co dissolution rates. The amount of Pt that dissolves cathodically is smaller than the amount that dissolves anodically if the upper potential limit (UPL) is lower than 0.9 V. At the highest UPL investigated, 1.0 V, the cathodic dissolution greatly exceeds the anodic dissolution. A non-ideal solid solution model indicates that the anodic dissolution can be associated with the electrochemical oxidation of Pt and PtOH to Pt 2+ , and the cathodic dissolution to electrochemical reduction of a higher Pt oxide, PtO x (x > 1), to Pt 2+ . Pt also dissolves oxidatively during the cathodic scans but in smaller amounts than due to the reductive dissolution of PtO x . The relative amounts Pt dissolving oxidatively as Pt and PtOH depend on the potential cycle and UPL.

The durability of Pt-Co alloy cathode catalysts supported on high surface area carbon is investigated by subjecting them to accelerated stress tests (ASTs). The catalysts had different initial Co contents and nanoparticle morphologies: a "spongy" porous morphology for the high-Co (H) content catalyst, and a fully alloyed crystalline morphology for the medium-Co (M) and low-Co (L) content catalysts. The specific activity of the catalysts depends on their initial Co content, morphology and nanoparticle size, and remained higher than 1000 μA/cm 2 -Pt after 27-50% Co loss. The H-catalyst electrode showed the smallest kinetic overpotentials (η c s ) due to higher initial Pt loading than the other two electrodes, but it had the fastest increase in η c s with AST cycling due to lower Co retention; the L-catalyst electrode showed higher η c s due to a lower initial Pt loading, but had a smaller increase in η c s with aging due to higher Co retention; the M-catalyst electrode showed a similar increase in η c s with aging, but this increase was due to the combined effects of Co dissolution and electrochemically active surface area (ECSA) loss. The modeled increase in mass transfer overpotentials with aging correlates with the initial Pt loading, ECSA loss and the initial catalyst morphology.

No abstract

One of the major contributors to degradation involves the electrocatalyst, including the corrosion of the carbons used as catalyst supports, which leads to changes in the catalyst layer structure. We have measured and quantified carbon corrosion during drive cycle operation and as a variation of the upper and lower potential limits used during drive cycle operation. The amount of carbon corrosion is exacerbated by the voltage cycling inherent in the drive cycle compared with constant potential operation. The potential gap between upper and lower potentials appears to be more important than the absolute operating potentials in the normal operating potential regime (0.40V to 0.95V) as changes in the measured carbon corrosion are similar when the upper potential was lower compared to raising the lower potential. Catalyst layer thinning was observed during the simulated drive cycle operation which had an associated decrease in catalyst layer porosity. This catalyst layer thinning is not due solely to carbon corrosion, although carbon corrosion likely plays a role; much of this thinning must be from compaction of the material in the catalyst layer. The decrease in catalyst layer porosity leads to additional performance losses due to mass transport losses.