Assessing Reactive Oxygen Species Generation in Heterogeneous Photocatalysis: Methods, Metrics, and Methodological Pitfalls

Reactive oxygen species are detected in almost every study of heterogeneous photocatalysis and are often said to drive the reaction. But detecting them does not prove that they are responsible; each claim requires different evidence. This review examines the four routes by which attribution is attempted—direct spectroscopy, chemical probes, scavenger suppression and kinetic analysis—and asks of each what it can and cannot establish. The case against uncritical practice is quantitative. Spin-trapping measurements report adduct concentrations, not radical fluxes, and adducts form by non-radical routes, decay on the time scale of the experiment (DMPO-•OOH half-life ≈ 66 s) and arise from trace-metal chemistry; the hydroxyl radical accounts for well under one percent of transferred charge where it has been measured electrochemically. Probe yield factors are treated as constants although the coumarin conversion factor inherited from radiolysis (6.1%) differs threefold from values measured in photocatalysis (18 ± 4%), and the hydroxyterephthalate yield varies with pH and temperature. Tetrazolium salts over-report superoxide under ultraviolet illumination, while spin trapping under-reports it. Scavenger inhibition is the weakest common evidence: several standard reagents generate their own reactive radicals, and p-benzoquinone increases hydroxyl radical production while suppressing the observed rate. Because these errors have known signs, methods whose biases run in opposite directions are the ones worth pairing. Since apparent rate constants measured under unreported illumination are not comparable quantities, the review also argues that only photon-normalised yields permit comparison between laboratories. It closes with a minimum defensible protocol, a control matrix and four reinterpreted case configurations, including activity rankings that invert once dark-stage conversion is separated from the illuminated rate.