Reformate polymer electrolyte membrane (PEM) fuel cells have great potential to transform the global maritime transportation in a more environmentally friendly and CO2-neutral manner. Platinum (Pt)-ruthenium (Ru) alloy nanoparticles (NPs) supported on carbon (PtRu/C) are the most commonly used electrocatalyst material to accelerate the hydrogen oxidation reaction (HOR) of the reformate gas (H2/CO2/CO) at the anode [1]. The high CO tolerance of PtRu/C catalysts is based on the ensemble effect [2]. However, these materials suffer extremely from the Ru dissolution and Ruz+ crossover to the cathode, which are the main degradation processes in reformate PEM fuel cells. At the cathode, the Ruz+ species immediately redeposit on the Pt/C catalyst, resulting in a drastic decay of the entire performance of the reformate PEM fuel cell. Even if the Ru coverage on the Pt/C catalyst is less than 20 at. %, there is an 8-fold decrease in kinetics of the oxygen reduction reaction (ORR) [3]. Therefore, it is necessary to develop a regeneration protocols to mitigate the Ru poisoning at the cathode.In this study, we developed and tested different regeneration protocols to recover the Ru-poisoned ORR catalysts using a rotating disc electrode (RDE) set-up. To simulate the worst situation for poisoned ORR catalysts, commercially available Pt-Ru alloy NPs supported on carbon were used and characterized by TEM, XRD, XAS, and XPS. In general, our regeneration protocols can be classified into dynamic and steady-state conditions. All protocols included a pre-activation step by holding the potential at 1.4 VRHE for 5 minutes in 0.1 M HClO4. During the dynamic regeneration process, the potential was switched between 1.4 VRHE and 1.6 VRHE with a duration time of 10 s by 10 times using a pulse voltammetry technique. In contrast, the steady-state regeneration protocol consisted of only one pulse at 1.6 VRHE for 100 s. For the recovery procedure of the poisoned ORR catalysts, the experimental parameters were correlated with the changes in ORR performance and electrochemically active surface area (ECSA) via CO stripping as well as the loss of Ru content and particle structure established from μ-XRF, ICP-OES, and TEM.After the dynamic regeneration process, the ORR mass activity (MA) and specific activity (SA) of PtRu2/Vulcan increase initially from 0.07 ± 0.01 to 0.21 ± 0.02 A mgPt+Ru -1 and 62 ± 9 to 190 ± 16 μA cmPt+Ru -2 at 0.85 VRHE, iR-free, respectively. Hence, the dynamic regeneration protocol leads to an improvement of the MA by a factor of 2 and of the SA by 2.1 for PtRu2/Vulcan. In comparison, the steady-state regeneration experiment shows an enhancement by 2.2 (0.26 ± 0.03 A mgPt+Ru -1) for MA and by 3.5 (276 ± 16 μA cmPt+Ru -2) for SA, respectively. Therefore, the steady-state regeneration process is more effective compared to the dynamic regeneration protocol. Based on the µ-XRF data, we observed that the Ru content of the PtRu2/Vulcan decreases from 62 ± 2 at. % to either 5 ± 1 at. % or 8 ± 1 at. % after the steady-state or dynamic regeneration process. The loss of Ru from the particle surface can also be monitored by CO stripping, where the CO oxidation peak is shifted positively from 0.58 VRHE to 0.67 VRHE and 0.63 VRHE after steady-state and dynamic regeneration protocol, respectively. As the particle size of the PtRu2 NPs is around ~ 5.3 ± 1.2 nm, the ECSA is mainly unchanged after both regeneration protocols.Based on our results, the steady-state regeneration process with 1.6 VRHE for 100 s is more effective than the dynamic one with pulses from 1.4 to 1.6 VRHE each with 10 s by 10 times to recover the ORR activity of the PtRu2 alloy NPs. The proposed regeneration protocols can improve the performance and life-time of the reformate PEM fuel cell.
Read full abstract