Reaction-level identification of viable five-atom 2D MxSey allotropes for solar photocatalytic overall water splitting

Efficient solar hydrogen production through photocatalytic overall water splitting requires more than appropriate band gaps and favorable band-edge alignment, yet this requirement is often underestimated in computational screening of two-dimensional photocatalysts. Here, we combine USPEX global structural search and first-principles calculations to identify viable five-atom 2D M x Se y allotropes (M = Si, Ge, Sn; x + y = 5) for solar photocatalytic overall water splitting. Starting from 11,280 converged structures, 91 semiconducting monolayers were identified, among which 16 electronically promising allotropes were retained after HSE06 band-edge alignment and corrected solar-to-hydrogen efficiency screening. Family-wide analysis shows that Se-rich stoichiometries and low-symmetry surface-inequivalent allotropes enlarge the pool of potential candidates by promoting intrinsic polarity, band-edge charge separation, and anisotropic carrier transport. However, these favorable electronic features do not by themselves ensure overall-water-splitting feasibility. The decisive factor is the thermodynamic compatibility of hydrogen and oxygen evolution within a common pH window, followed by structural robustness under operating conditions. Under this reaction-level screening framework, GeSe 4 -4 emerges as the most viable candidate, operating within pH 3–5 with a predicted corrected solar-to-hydrogen efficiency of 17.41–20.86%, which can be further increased to 22.10% under + 4% biaxial tensile strain. Phonon calculations and ab initio molecular dynamics simulations, including solvent-assisted simulations, further support its stability. These results establish composition-structure–reactivity criteria for identifying viable 2D photocatalysts for solar hydrogen production and clarify why many electronically attractive candidates fail at the reaction level.