▪ Abstract  Combinatorial methods provide a means for accelerating the discovery of fuel cell catalysts. The first example of parallel fuel cell catalysts screening was an indirect method that used fluorescent chemosensors to detect changes in pH in proximity to electrocatalyst spots. Serial direct electrochemical methods have been developed that use voltammetry, chronoamperometry, and scanning electrochemical microscopy. An array fuel cell screens catalysts simultaneously, using high-performance fuel cell components. Heuristic models based on mechanistic and spectroscopic studies provide guidance for library development, and detailed studies of discovered catalysts can help to refine these models. The remaining challenges are the development of high throughput synthetic methods that can enable the use of discovery level and focus level screening. Until these synthetic methods are developed, a greater emphasis should be placed on smaller libraries with design of experiment strategies leveraged with informatics and data mining.

The C-O stretching frequency (nu(CO)) of atop CO/Pt in PtRu alloys is compositionally tuned in proportion to the Pt mole percent. The application of a Blyholder-Bagus type mechanism (i.e., increased back-donation from the metal d-band to the hybridized 2pi CO molecular orbitals (MOs)) to compositional tuning has been paradoxical because (1) a Pt-C bond contraction, expected with increased back-donation as the Pt mole percent is reduced, is not observed (i.e., calculated Pt-C bond is either elongated or insensitive to alloying and the binding energies of CO/Pt decrease with alloying) and (2) the lowering d-band center and increased d-band vacancies upon alloying (suggesting less back-donation to the higher energy metal hybridized 2pi CO MOs) must be reconciled with the alloy-induced red shift of the nu(CO). A library of spin-optimized Pt and Pt alloy clusters was the basis of density functional theory (DFT) calculations of CO binding energies, nu(CO) values, shifts, and broadening of 5sigma/2pi CO MO upon hybridization with the alloy orbitals and a DFT derived Mulliken electron population analysis. The DFT results, combined with FEFF8 local density of states (LDOS) calculations, validate a 5sigma donation-2pi back-donation mechanism, reconciling the direction of alloy compositional tuning with the lowering of the d-band center and increased vacancies. Although the d-band center decreases in energy with alloying, an asymmetric increase in the dispersion of the d-band is accompanied by an upshift of the metal cluster HOMO level. Concomitantly, the hybridization and renormalization of the CO 5sigma/2pi states results in a broadening of the 5sigma/2pi manifold with additional lower energy states closer to the upshifted (with respect to the pure Pt cluster) HOMO of the alloy cluster. The dispersion toward higher energies of the alloy d-density of states results in more 5sigma/2pi CO filled states (i.e., enhanced 2pi-back-donation). Finally, Mulliken and FEFF8 electron population analysis shows that the increase of the average d-band vacancies upon alloying and additional 2pi back-donation are not mutually exclusive. The d-electron density of the CO-adsorbed Pt atom increases with alloying while the average d-electron density throughout the cluster is reduced. The localized electron density is manifested as an electrostatic wall effect, preventing the Pt-C bond contractions expected with increased back-donation to the 2pi CO MOs.

Catalysts prepared by the Adams and Reetz method were ranked against a Johnson Matthey PtRu catalyst in an array fuel cell in the direct methanol fuel cell ͑DMFC͒ mode. The Reetz and Johnson Matthey PtRu catalysts were found to be equivalent. The catalysts were analyzed in membrane and electrode assemblies ͑MEA͒ before and after fuel cell performance evaluation. The X-ray diffraction ͑XRD͒ analysis showed no changes of the crystalline phases before and after fuel cell evaluation. However, the X-ray photoelectron spectroscopy ͑XPS͒ showed that in the electrode containing the Reetz catalyst, substantial loss of Ru at the MEA surface occurred during fuel cell operation. Loss of Ru also occurred at the MEA surface containing the Johnson Matthey catalyst while an enrichment of Ru on the electrode surface containing the Adams catalyst was observed. A lattice parameter analysis of all catalysts showed Pt enrichment in the alloy phase, with the largest enrichment in the Adams catalyst.

No abstract