Data-Guided Discovery of Cu-Host Single-Atom Alloys for Selective Electrohydrogenation via Intermediate Hydrogen...

Electrocatalytic hydrogenation can upgrade oxygenated biomass molecules without pressurized H2 using renewable electricity, but selectivity and energy efficiency are limited by competition with the hydrogen evolution reaction through the shared surface-hydrogen intermediate (H*). Here we develop a data-guided and interpretable strategy to design Cu-based catalysts that balance H* supply and utilization by targeting an intermediate H* binding regime. A density functional theory (DFT) data set of H* adsorption energies and descriptors trains an interpretable machine-learning model and identifies Pt1Cu as an intermediate-binding catalyst, which is synthesized on a Cu nanowire scaffold. The catalyst delivers Faradaic efficiency above 90% for selective hydrogenation of 5-hydroxymethylfurfural over a wide potential window while suppressing hydrogen evolution and byproduct formation. The design principle is further validated in the furfural hydrogenation. Operando spectroscopy and DFT calculations link the performance to adequate H* availability and moderated stabilization of hydrogenated intermediates, providing a transferable framework for electricity-efficient biomass upgrading.