Statistical Benchmarking of MnOx Thin-Film Catalysts for Alkaline OER: A Medium-Throughput Approach

The lack of statistical rigor in electrocatalytic measurements, where catalyst performance is traditionally determined based on a few representative measurements of thin catalyst films, limits the reliability of benchmarking reactions such as the oxygen evolution reaction (OER). This study reports a medium-throughput, statistically rigorous benchmarking framework that converts anecdotal claims of catalytic activity into quantitatively validated evidence. Two MnOₓ-based OER catalysts synthesized via different hydrothermal routes were evaluated across 144 identically prepared thin-film electrodes to explicitly resolve film-to-film variability. The data distribution, intrinsic dispersion, and required sampling size for robust benchmarking were quantified statistically. The HCl-derived catalyst, MnOₓ(H), exhibits narrow, near-Gaussian activity distributions with low relative standard error of the mean (RSEM < 8%) and tight confidence intervals, indicating statistically converged performance at moderate sample sizes. In contrast, the H₂SO₄-derived catalyst, MnOₓ(S), displays broad, right-skewed, non-Gaussian distributions with persistent high-activity outliers and wide confidence intervals (±25-50%), reflecting intrinsic catalyst-film heterogeneity. Importantly, population-level statistics demonstrated that reliance on a small number of selected measurements can substantially overestimate apparent activity and obscure statistically relevant variability. The findings demonstrate that statistical robustness is as essential as electrochemical metrics for meaningful comparison of OER activity data and that variability can arise from both intrinsic material heterogeneity and measurement artefacts. Taken together, this study highlights the limitations of conventional representative-curve reporting and provides a generalizable framework for statistically robust benchmarking of electrocatalysts.