Active Role of Phosphorus in the Hydrogen Evolving Activity of Nickel Phosphide (0001) Surfaces

Optimizing catalysts for the hydrogen evolution reaction (HER) is a critical step toward the efficient production of H₂(g) fuel from water. It has been demonstrated experimentally that transition-metal phosphides, specifically nickel phosphides Ni₂P and Ni₅P₄, efficiently catalyze the HER at a small fraction of the cost of archetypal Pt-based electrocatalysts. However, the HER mechanism on nickel phosphides remains unclear. We explore, through density functional theory with thermodynamics, the aqueous reconstructions of Ni₂P(0001) and Ni₅P₄(0001)/(0001), and we find that the surface P content on Ni₂P(0001) depends on the applied potential, which has not been considered previously. At -0.21 V ≥ U ≥ -0.36 V versus the standard hydrogen electrode and pH = 0, a PH_x-enriched Ni₃P₂ termination of Ni₂P(0001) is found to be most stable, consistent with its P-rich ultrahigh-vacuum reconstructions. Above and below this potential range, the stoichiometric Ni₃P₂ surface is instead passivated by H at the Ni₃-hollow sites. On the other hand, Ni₅P₄(0001) does not favor additional P. Instead, the Ni₄P₃ bulk termination of Ni₅P₄(0001) is passivated by H at both the Ni₃ and P₃-hollow sites. We also found that the most HER-active surfaces are Ni₃P₂+P+(7/3)H of Ni₂P(0001) and Ni₄P₃+4H of Ni₅P₄(0001) due to weak H adsorption at P catalytic sites, in contrast with other computational investigations that propose Ni as or part of the active site. By looking at viable catalytic cycles for HER on the stable reconstructed surfaces, and calculating the reaction free energies of the associated elementary steps, we calculate that the overpotential on the Ni₄P₃+4H surface of Ni₅P₄(0001) (-0.16 V) is lower than that of the Ni₃P₂+P+(7/3)H surface of Ni₂P(0001) (-0.21 V). This is due to the abundance of P₃-hollow sites on Ni₅P₄ and the limited surface stability of the P-enriched Ni₂P(0001) surface phase. The trend in the calculated catalytic overpotentials, and the potential-dependent bulk and surface stabilities explain why the nickel phosphides studied here perform almost as well as Pt, and why Ni₅P₄ is more active than Ni₂P toward HER, as is found in the experimental literature. This study emphasizes the importance of considering aqueous surface stability in predicting the HER-active sites, mechanism, and overpotential, and highlights the primary role of P in HER catalysis on transition-metal phosphides.