Microdroplets make hydrogen peroxide, but where it forms is still unresolved

Aqueous microdroplets have become one of the most debated reaction environments in physical chemistry. Reports stretching back more than a decade describe measurable hydrogen peroxide forming when water is broken into micron-sized droplets, apparently without any added catalyst. A new Analysis published in the journal Catal argues that the accumulated evidence is strong enough to treat spontaneous hydrogen peroxide and reactive oxygen species formation as a genuine capability of gas–liquid interfaces in aqueous microdroplets, rather than a pure artefact of any single experimental platform. At the same time, the authors, a team including Tong Dong, Xiaoxue Song, Yuqiao Zhang, Miaomiao Chen, Jianming Zhang, Shun Li and Long Zhang, insist that the field’s central mechanistic questions remain open, because similar observations are still being interpreted through platform-specific models that are not directly comparable.

The core problem, according to the Analysis, is that the phrase microdroplet interface does not describe the same experimental object from one study to the next. Droplets produced by spraying, condensation, ultrasonication, levitation or emulsification differ in how much solid surface they touch, how much charge they carry, how much shear or cavitation energy they experience, what impurities they pick up, and how long they survive before analysis. These factors change both the thermodynamic environment, including solvation, interfacial electrostatics and ion organization, and the kinetic expression of any chemistry, including transport, radical lifetime and reactions that continue after the droplet is collected. A nominal water microdroplet can therefore contain a gas–liquid boundary, an electrical double layer created by preferential ion adsorption, and a liquid–solid interface introduced by nozzles, tubing, collectors or vessel walls, all coupled together in a single measured outcome.

To bring order to this situation, the authors propose an evidence ladder with three rungs. Level A is detection: a signal consistent with hydrogen peroxide or a reactive-oxygen-derived product appears above the method’s limit, under workflow-matched blanks and matrix-validated calibration. Level B is net production: the measured species must accumulate during the droplet-processing window itself, rather than during storage, transfer, collection or assay development, and the signal should scale reproducibly with a droplet-relevant variable such as residence time, droplet size, curvature, humidity or controlled gas composition. Level C is interfacial attribution, the strongest claim, reserved for work that assigns formation to a specific interface or mechanistic class while experimentally bounding the competing contributions of dissolved oxygen, wetted solid contact, source-region effects, collection conditions and post-collection reactions.

Strikingly, the Analysis concludes that no single study has yet provided a fully decisive Level C demonstration that uniquely assigns hydrogen peroxide formation to a pristine air–water interface while simultaneously excluding all of those alternatives. The authors treat Level C as a prospective benchmark rather than a label automatically earned by existing positive reports. This framing protects both sides of the current debate: it avoids dismissing positive hydrogen peroxide findings as artefacts simply because the mechanism is uncertain, while preventing an endpoint detection from being overinterpreted as proof of pristine interfacial origin. A collected-sample concentration, the paper stresses, is an integrated outcome of droplet generation, flight, evaporation, deposition, collection and analysis, not a direct record of a single interfacial event.

The mechanistic landscape the Analysis maps out contains several related but non-equivalent regimes. One class of proposals invokes oxygen: when dissolved or interfacial O2 is available, it can serve as an electron acceptor, with reduced oxygen species such as superoxide or hydroperoxyl radicals participating before disproportionation or further reduction yields hydrogen peroxide. This interpretation is particularly relevant for workflows containing wetted solids, active electrodes or metal surfaces prone to oxidation, where apparent microdroplet peroxide formation may reflect coupling between oxygen reduction and oxidation of a contacting surface. Under other workflows, however, hydrogen peroxide remains detectable without added oxygen or is even enhanced under inert atmospheres, supporting mechanisms in which interfacial water, hydroxide ions or charge-separated states play the larger role. The authors argue these observations are not contradictory: oxygen can act as a reactant, a competing electron acceptor, a transport-limited species, or a factor controlling product survival, and its role must be established within each platform rather than assumed across the field.

A second class of mechanisms assigns the chemistry to water or hydroxide at the gas–liquid interface itself, motivated by incomplete hydration, ion enrichment or depletion, altered hydrogen-bond structure and strong local electric fields. Proposed pathways include field-assisted electron transfer from hydroxide, formation of water radical-cation or radical-anion-like species, production of hydroxyl radicals or hydrated electrons, and subsequent recombination to peroxide. The Analysis is blunt that this is the class most easily overstated: evidence for field-sensitive chemistry or peroxide formation under contact-minimized conditions supports the plausibility of gas–liquid interfacial redox pathways, but does not establish that such pathways dominate in every workflow. Because different initiating events, from oxygen reduction to hydroxide electron loss to solid-contact electron transfer, can converge into a common radical network that interconverts intermediates and generates peroxide, the same endpoint signal cannot adjudicate the mechanism on its own.

Electrostatics receives its own careful treatment. The authors warn that electric field, surface potential, interfacial potential gradient, net droplet charge and zeta potential are not interchangeable quantities, and that a field-based mechanism is only as strong as the observable supporting it. Spectroscopic frequency shifts are consistent with strong local fields at aqueous interfaces, with reported field magnitudes reaching around 10^9 volts per metre in charged-droplet work, but converting a shift into a field strength depends on probe calibration, molecular orientation and local microenvironment. Faraday-cup measurements integrate net charge transferred by droplet populations, while electrophoretic mobility reflects electrokinetic charge rather than a direct thermodynamic surface potential. The microelectrochemical-cell framework, in which charge separation, double layers and potential gradients allow oxidative and reductive steps to couple, is endorsed as grounded in measurable quantities, but the paper insists it is not yet a universal predictor of peroxide or radical yields and becomes quantitative only when electrostatic observables are linked to falsifiable chemical trends under stated boundary conditions.

Transient radical intermediates demand a separate standard, the authors argue, because most assignments rely on indirect readouts such as spin adducts, oxidation products, chemiluminescence, fluorescence or mass-spectrometric ions, all of which are sensitive to pH, salts, organics, gas composition, ionization conditions and probe chemistry. A radical-associated signal should be treated as a mechanistic product rather than direct proof of a radical pathway unless its chemical specificity and measurement stability have been demonstrated. If an ion signal tracks instrument conditions more strongly than droplet chemistry, it should be interpreted as source- or analysis-dependent until orthogonal evidence arrives. The most reliable assignments combine at least two independent readouts, for example mass spectrometry supported by electron paramagnetic resonance, validated fluorescence, electrochemical detection or isotope labelling, while disagreement among methods may simply indicate that a short-lived intermediate was generated but not preserved.

The practical roadmap that closes the Analysis identifies five regime-defining controls: oxygen availability, wetted solid contact, droplet-generation mode and charge-state evolution, humidity and residence time, and collection and holding conditions. For oxygen-sensitive claims, experiments should define initial dissolved-oxygen levels, measure dissolved oxygen before generation and after collection, use specified atomizing and headspace gases, and test oxygen intake through sealed, inert and oxygen-reintroduced controls. For solid-contact claims, all wetted materials should be documented and, where feasible, substituted or removed, with contact-minimized platforms such as adiabatic expansion or dry-ice-assisted fog formation serving as reference systems. The authors also chart a path from bulk collection toward single-droplet kinetics, citing recent interface-resolved work in which in situ chemiluminescence imaging visualized hydroxyl radical formation at individual gas–liquid interfaces in microbubble systems, and real-time fluorescence imaging monitored hydrogen peroxide formation in individual water microdroplets with kinetics related to curvature and interfacial electrostatics.

The overall message is a recalibration rather than a rebuttal. The accumulated evidence, the authors conclude, is difficult to reconcile with the view that all reported microdroplet redox chemistry is artifactual, and aqueous microdroplets do provide chemically active environments in which peroxide and transient reactive oxygen species can be generated without added molecular catalysts. But the central challenge is no longer whether oxidants can be detected; it is how the formation window, reaction site and operative interface can be assigned under well-defined boundary conditions. A more rigorous field, they argue, requires distinguishing detection from net production and interfacial attribution, tying electrostatic language to measured quantities and model assumptions, and moving the primary readout closer to the droplet event through contact-minimized baselines, controlled gas exposure, real-time kinetics and orthogonal validation. Under those conditions, microdroplet redox chemistry can move from platform-dependent observation toward reproducible, comparable and eventually predictive interfacial science.

Subject of Research: Hydrogen peroxide and reactive oxygen species formation in aqueous microdroplets and the evidence needed to attribute it to specific interfaces

Article Title: Evidencing microdroplet H2O2 redox chemistry: from detection to mechanistic attribution

Article References: Dong, T., Song, X., Zhang, Y., Chen, M., Zhang, J., Li, S., & Zhang, L. (2026). Evidencing microdroplet H2O2 redox chemistry: from detection to mechanistic attribution. Catal, 2(1), Article 14. https://doi.org/10.1007/s44422-026-00028-8

Image Credits: AI Generated

DOI: 10.1007/s44422-026-00028-8

Keywords: microdroplets, hydrogen peroxide, reactive oxygen species, air–water interface, redox chemistry, oxygen reduction, interfacial electrostatics, mass spectrometry, spin trapping, evidence standards, atmospheric chemistry, microelectrochemical cell