Beyond Presumptions: Toward Mechanistic Clarity in Metal-Free Carbon Catalysts for Electrochemical H2O2 Production via Data Science.

Electrochemical synthesis of hydrogen peroxide (H2O2) via the two-electron oxygen reduction reaction (2e- ORR) has emerged as an environmentally friendly alternative to the traditional anthraquinone process. Metal-free carbon catalysts, featuring tunable structures, readily available precursors, and excellent stability, have garnered significant attention for sustainable H2O2 production. However, despite extensive investigations, the precise mechanisms underlying catalytic selectivity on these carbon materials remain unclear and highly debated. Previous mechanistic interpretations frequently attribute catalytic activity to specific oxygen functional groups or heteroatom dopants through correlation-driven hypotheses and simplified theoretical models. Such approaches often overlook the intrinsic complexity of carbon surfaces, where multiple variables, including dopant types, defect structures, surface groups, and hybridization states, coexist and interact simultaneously, leading to contradictory conclusions. This review critically examines the limitations of these traditional approaches and emphasize the need of systematic experimental designs that independently vary structural parameters, along with advanced analytical methods capable of resolving active-site ambiguity, are critically reviewed. Recent developments employing orthogonal material libraries, rigorous experimental controls, catalyst passport metadata, and advanced multivariate and meta-analytical tools have emerged as robust frameworks for bias-resistant catalyst design. Integrating explainable and generative machine learning models with operando spectroscopy provides a robust, end-to-end approach for identifying and validating accurate catalytic descriptors.