Designing multinary noble metal-free catalyst for hydrogen evolution reaction

The hydrogen evolution reaction (HER), the key reaction for electrocatalytic production of hydrogen, is of fundamental importance due to its simplicity yet is very important for renewable energy. Notwithstanding, Pt is still the main catalyst for this reaction, which is not practical for the industrial deployment of this technology owing to the high cost and scarcity of Pt. The successful synthesis of high entropy alloy (HEA) nanoparticles opens a new frontier for the development of new catalysts. Herein we investigate the design of a multinary noble metal-free HER catalyst based on earth-abundant elements Co, Mo, Fe, Ni, and Cu. Using a machine learning (ML) approach in conjunction with first-principles methods, we build a model that can rapidly compute the hydrogen adsorption energy on the alloyed surfaces with high fidelity. Within the large composition space of the CoMoFeNiCu HEA, a large number of alloy combinations are shown to optimally bind hydrogen with a high probability. Further, most of these alloy compositions are found stable against dissociation into intermetallics, and hence synthesizable as a solid solution, by virtue of a large mixing entropy compared to mixing enthalpy and a small lattice mismatch between the elements. This finding is partly consistent with recent experimental results that synthesized five different CoMoFeNiCu HEA compositions. Our study underscores the significant impact that computational modeling and ML can have on developing new cost-effective electrocatalysts in the nearly-infinite materials design space of HEAs, and calls for experimental validation.