Accelerating Green Hydrogen: IrO2 Nanoparticle Models for Electrocatalytic Water Splitting

The design of active and stable electrocatalysts for the oxygen evolution reaction (OER) remains a critical bottleneck in water electrolysis for hydrogen fuel generation, primarily due to the sluggish kinetics and high overpotentials required at the anode. This study focuses on constructing IrO₂ nanoparticles, recognized as state-of-the-art OER catalysts due to their robust catalytic activity and stability. Through density functional theory (DFT), we calculate the surface energies of various IrO₂ facets and terminations to inform Wulff constructions, which model the most thermodynamically favorable geometries of these nanoparticles. Additionally, we explore how their structural properties change under different electrochemical potentials, shedding light on the relationship between nanoparticle morphology and catalytic stability.

Our ab initio thermodynamics analysis reveals that IrO₂ nanoparticles undergo a transformation in shape and stability under operational conditions. Specifically, we find a thermodynamic instability in the rutile crystal structure caused by the stabilization of highly oxidized oxygen species at the surface at OER onset potentials. As the potential increases, this instability leads to a transformation in the equilibrium shape, shifting from the traditionally studied IrO₂ (110) facets to the predominant IrO₂ (111) facets. Additionally, our findings align with experimentally synthesized nanoparticles in both shape and facet ratios, further validating our computational models.

Future steps will focus on enhancing performance through doping with 3d and 4d transition metals, which has been shown to help mitigate the trade-off between activity and stability. Additionally, extending detailed characterization studies to higher potential ranges for RuO₂, a more reactive but less stable metal oxide, is essential due to its higher natural abundance compared to iridium, making it advantageous for large-scale applications. We are also training machine learning models on our data to circumvent the computational cost of DFT calculations in predicting Pourbaix and surface phase diagrams. Our approach contributes to the ongoing effort to develop sustainable and efficient hydrogen production technologies.