Carbon Catalysts Steer Oxygen Reduction Toward Greener Hydrogen Peroxide Production

Hydrogen peroxide is one of the world’s most essential chemicals, with global production exceeding four million tons annually, yet nearly all of it is made through the energy-intensive anthraquinone process, which requires sequential hydrogenation and oxidation of alkylanthraquinones at elevated temperature and pressure, generates substantial organic waste, and depends on centralized facilities whose transportation of concentrated peroxide adds hazards and carbon footprint. A comprehensive review published in the journal Catal surveys how carbon-based electrocatalysts could replace this century-old route by driving the two-electron oxygen reduction reaction (2e– ORR), a process that needs only water, oxygen or air, and electricity, offers 100% atom efficiency, and can run at ambient temperature and pressure when powered by renewable sources such as solar or wind, enabling modular, decentralized manufacturing that bypasses the storage and transport risks of concentrated H2O2 solutions.

The central obstacle is selectivity. The oxygen reduction reaction can proceed along two competing pathways: the desired two-electron route that yields hydrogen peroxide, and a four-electron route that reduces oxygen all the way to water, which is valuable in fuel cells but wasteful for peroxide synthesis. The review emphasizes that the 2e– pathway inherently suffers from a significant theoretical overpotential compared with its 4e– counterpart, particularly in alkaline media, so the primary objective is to kinetically favor peroxide formation while suppressing the thermodynamically more favorable deeper reduction. Practical deployment additionally demands low overpotential, high current density, and long-term resistance to corrosion and deactivation, making simultaneous optimization of selectivity, activity, and stability the cornerstone of efficient electrosynthesis.

At the molecular level, everything hinges on how the oxygen molecule first binds to the catalyst. In the end-on adsorption mode, where O2 attaches through a single oxygen atom, the O–O bond is generally preserved and the two-electron pathway is promoted; in the side-on mode, where the molecule bridges two adjacent sites, conditions favor O–O bond scission and water formation. The critical branching point is the hydroperoxyl intermediate, *OOH, formed in the first proton-coupled electron transfer. If the surface binds *OOH too weakly, oxygen activation becomes rate-limiting and activity suffers; if binding is too strong, desorption is inhibited and the intermediate risks further reduction. An optimal, moderate binding strength allows the second proton-coupled electron transfer to release hydrogen peroxide in acidic media, or the hydroperoxyl ion in alkaline media.

This mechanistic picture explains why carbon materials are so attractive. Although they lack traditional metal active centers, their surfaces can be engineered through heteroatom doping with nitrogen, boron, sulfur, or phosphorus, and through deliberate creation of defects such as vacancies, edges, and distorted carbon rings. These modifications alter charge distribution, spin density, and local work function, allowing precise tuning of *OOH binding energy. Carbon’s earth-abundant reserves, low cost, tunable porous structures, excellent conductivity, and modifiable surface chemistry, together with the avoidance of metal-ion leaching that could trigger deleterious Fenton reactions, make metal-free carbons a leading platform for the reaction.

The review catalogs striking experimental demonstrations of these principles. Oxygen-rich carbon quantum dots made from glucose, containing 30.4 atomic percent oxygen, achieved nearly 100% H2O2 selectivity, a production rate of 10.06 milligrams per square centimeter per hour, and a Faraday efficiency of 97.7%, with calculations identifying ether-group carbon atoms as the active sites. A carboxylated boron nitride–graphene heterojunction on activated carbon delivered greater than 95% selectivity and Faraday efficiency and a maximum production rate of 13.4 moles per gram per hour, accumulating 2.1 weight percent H2O2 at an industrial current density of 100 milliamperes per square centimeter. A boron-doped carbon black catalyst cut the overpotential by more than 210 millivolts at 300 milliamperes per square centimeter relative to an oxidized control, maintained 85–90% selectivity, ran 30 hours without degradation, and reached a partial current density of up to 400 milliamperes per square centimeter in a solid electrolyte reactor.

Fluorine and morphology offer further levers. Fluorine-doped carbon nanotubes with fluorine on both inner and outer walls created a local alkaline microenvironment on the catalyst surface: inner-wall doping enhanced the nanoconfinement effect, outer-wall doping imparted superhydrophobicity that boosted oxygen mass transfer, and the elevated p-band center lowered the energy barrier for *OOH generation. The result was a 95.6% Faraday efficiency and a peroxide yield of 606.6 milligrams per square centimeter per hour at an industrial current density of 1.0 ampere per square centimeter in 0.05 M sulfuric acid, with stability across acidic, neutral, and alkaline conditions and compatibility with air as the oxygen source. Curvature matters too: surface-oxidized carbon nanotubes with the smallest tested diameter of 8 nanometers achieved a Faraday efficiency above 85% and a mass activity of 161 amperes per gram at 0.65 volts versus RHE, because increased curvature strains epoxy groups and activates adjacent carbon atoms.

Carbon hybridization itself has emerged as a decisive descriptor. Graphdiynes tuned to 50 atomic percent sp-carbon delivered over 95% H2O2 selectivity, a Faraday efficiency exceeding 99%, and a yield of 83.3 nanomoles per second per square centimeter, since the electron-rich π orbitals of sp-carbon favor electron transfer to oxygen. Nitrogen-doped diamond carbon dots, built on an sp3-hybridized matrix, reached up to 99% selectivity with a production rate of 1302.14 millimoles per gram per hour. One re-evaluation using partial least squares regression concluded that the conventional emphasis on oxygen functional groups is misplaced, identifying the sp3/sp2 hybridization ratio as the key performance descriptor for near-100% selectivity between 0.4 and 0.6 volts versus RHE, with an optimized onset potential of 0.766 volts matching the theoretical thermodynamic limit. Edge engineering adds another dimension: three-dimensional fuzzy graphene with tailored edge density achieved an onset potential of 0.79 volts versus RHE and 94% selectivity, with carbonyl and hydroxyl configurations at armchair and zigzag edges identified as the active sites.

When metal sites are introduced, precision increases further. Carbon-supported platinum nanoparticles smaller than 5 nanometers, rich in edge sites that inhibit O–O bond cleavage, reached 92% selectivity in acidic media. Palladium single atoms on nitrogen, sulfur-doped hollow carbon spheres leveraged sulfur to boost selectivity to roughly 90% and curb leaching to only 0.2%, retaining more than 85% of current after stability testing. A cobalt single-atom catalyst with boron in the second coordination sphere achieved an onset potential of 0.724 volts versus RHE, about 94% selectivity in acidic electrolyte, a record productivity of 202.7 milligrams per square centimeter per hour in a flow cell, and 60 hours of stability at 200 milliamperes per square centimeter. A nickel wastewater-derived single-site catalyst with square-planar Ni-S4 coordination exceeded 95% Faraday efficiency across 200–800 milliamperes per square centimeter, with a record productivity of 70.75 moles per gram per hour at 781 milliamperes per square centimeter. Dual-atom designs push stability further: a manganese dual-atom catalyst ran 240 hours at 300 milliamperes per square centimeter, while a cobalt/molybdenum pair anchored in mesoporous carbon hollow spheres delivered 90–95% selectivity in acid with 150 hours of stability.

Framework-derived materials round out the portfolio. A chromium-based metal-organic framework whose Lewis acidity accumulates hydroxide ions on its surface achieved Faraday efficiencies of 96–98% at current densities of 50–150 milliamperes per square centimeter in neutral electrolyte. A dithiine-linked cobalt phthalocyanine covalent organic framework produced 0.48 weight percent H2O2 at 125 milliamperes per square centimeter with selectivity above 95% for 20 hours, and a nickel phthalocyanine COF with enhanced π-conjugation sustained an ultrahigh current density of 530 milliamperes per square centimeter at nearly 100% Faraday efficiency and 3.5 volts, generating medical-grade 3.5 weight percent peroxide without electrolytes at a concentration 3.5 times higher than previously reported electrocatalysts and 1.5 times that of the anthraquinone process, with a 27% cost reduction. The review’s authors caution that the field still lacks high-throughput screening of metal-dopant combinations, reproducible coordination control, operando mechanistic insight, and quantified synergies among design factors, and they point toward machine-learning-driven discovery, in situ spectroscopy, and practical device integration with renewable energy as the routes that could turn these laboratory benchmarks into a genuinely decentralized peroxide supply chain.

Subject of Research: Carbon-based electrocatalysts for selective hydrogen peroxide production via the two-electron oxygen reduction reaction

Article Title: Carbon-based electrocatalysts for selective two-electron oxygen reduction

Article References: Wei, X., Cao, Y., Ma, Y., & Cao, R. (2026). Carbon-based electrocatalysts for selective two-electron oxygen reduction. Catal, 2(1), Article 5. https://doi.org/10.1007/s44422-026-00019-9

Image Credits: AI Generated

DOI: 10.1007/s44422-026-00019-9

Keywords: hydrogen peroxide, electrocatalysis, oxygen reduction reaction, carbon materials, heteroatom doping, defect engineering, single-atom catalysts, metal-organic frameworks, covalent organic frameworks, green chemistry, 2e- ORR, sustainable synthesis