In the quest for active and inexpensive (photo)electrocatalysts, atomistic simulations of the oxygen evolution reaction (OER) are essential for understanding the catalytic process of water splitting at solid surfaces. In this paper, the enhancement of the OER by first-row transition-metal (TM) doping of the abundant semiconductor ZnO was studied using density functional theory (DFT) calculations on a substantial number of possible structures and bonding geometries. The calculated overpotential for undoped ZnO was 1.0 V. For TM dopants in the 3d series from Mn to Ni, the overpotentials decreased from 0.9 V for Mn and 0.6 V for Fe down to 0.4 V for Co, and rose again to 0.5 V for Ni and 0.8 V for Cu. The overpotentials were analyzed in terms of the binding to the surface of the species involved in the four reaction steps of the OER. The Gibbs free energies associated with the adsorption of these intermediate species increased in the series from Mn to Zn, but the difference between OH and OOH adsorption (the species involved in the first, respectively the third reaction step) was always in the range 3.0–3.3 eV, despite a considerable variation in possible bonding geometries. The bonding of the O intermediate species (involved in the second reaction step), which is optimal for Co, and to a somewhat lesser extend for Ni, then ultimately determined the overpotential. These results implied that both Co and Ni are promising dopants for increasing the activity of ZnO-based anodes for the OER.

Photoelectrochemical (PEC) splitting of water to make hydrogen is a promising clean-energy technology. The oxygen evolution reaction (OER) largely determines the energy efficiency in PEC water-splitting. Hematite, which is a cheap and sustainable semiconductor material with excellent chemical properties, a favourable band gap (2.1 eV) and composed of earth abundant elements is a suitable model photoanode material for studying OER. To understand the design of energy efficient anodes, it is highly desirable to have mechanistic insight into OER at an atomistic level which can be directly connected to experimentally measured quantities. We present a multiscale computational model of OER which connects the thermodynamics and kinetics of elementary charge transfer reactions in OER to kinetics of OER at laboratory length and time scales. We couple density functional theory (DFT) and DFT based molecular dynamics (DFT-MD) simulations with solvent effects at an atomistic level with kinetic Monte Carlo (kMC) simulations at a coarse-grained level in our multiscale model. The time and applied bias potential dependent surface coverage, which are experimentally not known, and the O2 evolution rate during OER at the hematite-water interface are calculated by the multiscale model. Furthermore, the multiscale model demonstrates the effect of explicitly modelling the interaction of water with the electrode surface via direct adsorption. This journal is

The photoelectrochemical (PEC) activity of microstructured electrodes remains low despite the highly enlarged surface area and enhanced light harvesting. To obtain a deeper understanding of the effect of 3D geometry on the PEC performance, well-defined WO₃/n-Si and WO₃/pn-Si micropillar arrays are fabricated and subjected to a quantitative analysis of the relationship between the geometry of the micropillars (length, pitch) and their PEC activity. For WO₃/n-Si micropillars, it is found that the photocurrent increases for WO₃/n-Si pillars, but not in proportion to the increase in surface area that results from increased pillar length or reduced pillar pitch. Optical simulations show that a reduced pillar pitch results in areas of low light intensity due to a shadowing effect. For WO₃/pn-Si micropillar photoelectrodes, the p–n junction enhances the photocurrent density up to a factor of 4 at low applied bias potential (0.8 V vs RHE) compared to the WO₃/n-Si. However, the enhancement in photocurrent density increases first and then decreases with reduced pillar pitch, which scales with the photovoltage generated by the p–n junction. This is related to an increased dead layer of the p–n junction Si surface, which results in a decreased photovoltage even though the total surface area increases.