Oxygen Evolution Reaction
The oxygen evolution reaction (OER) is the anodic half reaction of water electrolysis, in which water is oxidized to O2. It is kinetically sluggish, and rutile RuO2 and IrO2 are among the most active known catalysts. In a widely used computational model, the reaction proceeds through four proton-electron transfer steps via adsorbed O, OH and OOH, and the theoretical overpotential is set by the largest free energy step computed from their adsorption energies relative to the standard hydrogen electrode. Because these adsorption energies scale with one another, activity can be plotted on a volcano plot against a single descriptor. See also Adsorption scaling relations.
Oxygen evolution in carbonate and bicarbonate electrolytes for CO2 release (2010)
In an electrochemical CO2 pump the anode evolves O2, and the protons it generates decompose bicarbonate or carbonate ions to release CO2. Pennline et al. screened metal wires by cyclic voltammetry in 0.1 M KHCO3 and found Ni more active than Pd and Pt in the 0.8 to 1.0 V vs. SCE oxygen evolution range, with Cu, Ag, Pb and Sn corroding and Au inactive; polyol-synthesized Ni particles on carbon paper gave higher activity than Ni wire (Pennline et al. 2010). CO2 evolution was faster from bicarbonate than carbonate at equal current because bicarbonate needs only one proton to decompose. Summary: pennline2010.
Oxygen evolution as the release step in electrochemical CO2 separation (2010)
In the anion exchange membrane CO2 separator of Landon and Kitchin, a polyol-synthesized Ni anode oxidizes hydroxide to O2 while bicarbonate arriving through the membrane is neutralized to release CO2, giving a CO2:O2 ratio of 3.56:1 (Landon 2010). Because the membrane added only about 30 mV at 1.2 V, the authors identified slow oxygen evolution and reduction kinetics as the main barrier and called for more active electrocatalysts. Summary: landon2010.
Universal OER volcano on oxide surfaces (2011)
Using a DFT (RPBE) database of HO*, O* and HOO* adsorption energies on rutile, anatase, perovskite, spinel, rock-salt and manganese oxide surfaces, with free energies from the computational hydrogen electrode, Man, Kitchin, Rossmeisl and coworkers showed that the OER on all these oxides is described by one volcano in a single descriptor, the free energy difference between O* and HO*. Either HO* oxidation or HOO* formation is potential-determining, and the predicted orderings (for example SrCoO3 > LaNiO3 > SrNiO3 > SrFeO3 > LaCoO3 > LaFeO3 > LaMnO3 among perovskites, with Co3O4 and RuO2 near the top among other oxides) agree in trend with alkaline and acidic experiments (Man et al. 2011).
Because HOO* and HO* are separated by about 3.2 eV rather than the ideal 2.46 eV, the paper concludes that planar oxides have a minimum theoretical overpotential of roughly 0.4 V and cannot be made much better than RuO2 by tuning binding alone; breaking the limit requires stabilizing HOO* relative to HO* (Man et al. 2011). Summary: man2011.
Polymorph structure as an OER tuning handle (2015)
Xu and Kitchin used this framework on anatase, brookite, columbite and pyrite polymorphs of RuO2, RhO2, IrO2 and PtO2, with adsorption at five-fold coordinatively unsaturated cation sites. O, OH and OOH adsorption energies on the polymorphs fell on the same scaling lines as on rutile, so the polymorphs sit on the same volcano but at different positions. Anatase and columbite IrO2 and brookite RhO2 and PtO2 were predicted to have lower theoretical overpotentials than the rutile forms, and columbite IrO2 was singled out as a synthesis target. The authors present this as a test of whether structural distortions can change activity substantially, not as a complete assessment of any one surface (Xu 2015). Summary: xu2015a; see also Oxide polymorphs.
A common descriptor for OER intermediates on metals and oxides (2013)
Calle-Vallejo and co-workers proposed the number of outer electrons on the active metal as a descriptor for *O, *OH and *OOH adsorption across 3d metals, monoxides and perovskites. Adsorption weakens with more outer electrons along the 3d series and with higher oxidation state for a given metal, giving a predictive adsorption-energy grid. The *OOH versus *OH scaling intercept of 3.2 eV that falls out of the grids implies an intrinsic overpotential for ORR-OER electrocatalysis (Calle-Vallejo et al. 2013). Summary: callevallejo2013.
Effect of a linear response Hubbard U on rutile OER trends (2015)
Xu, Rossmeisl and Kitchin recomputed OER intermediates on rutile (110) surfaces with DFT + U. Adding the linear response U weakened binding by roughly 0.2 to 0.4 eV in reaction energy and slid each oxide along the volcano toward the weak-binding side. For the acid-stable oxides the predicted ordering changed from RhO2 > IrO2 > PtO2 > RuO2 at U = 0 to IrO2 > RhO2 > RuO2 > PtO2, closer to the experimental RuO2 ≈ IrO2 > RhO2 > PtO2, though no oxide crossed the volcano peak (Xu et al. 2015). Summary: xu2015; see also DFT + U and the Hubbard U.
Automated OER modeling with WhereWulff (2023)
WhereWulff computes OH*, O* and OOH* free energies, the potential-determining step and the theoretical overpotential on the most stable termination of each Wulff-relevant facet, with no manual steps after the bulk structure is supplied (Sanspeur et al. 2023). Benchmarks on O*-covered RuO2 and IrO2 gave 0.52/0.53 V and 0.49/0.39 V on (110)/(101), with O* to OOH* limiting on (110) and OH* to O* on (101). The work also shows that the assumed termination can dominate the result: BaSnTi2O6-(110) gave 2.06 V on an O*-covered surface but 0.73 V on the OH*-covered one. Summary: sanspeur2023.
Stability-guided design of Ru-based acidic OER catalysts (2024)
Abed and coworkers addressed the main weakness of RuO2 in acid, Ru dissolution, by screening rutile mixed oxides with a graph network surrogate for Pourbaix stability and then synthesizing Ru-Cr-Ti oxides. Cr raised activity (about 70 mV lower overpotential at 10 mA cm-2 for 25 at.% Cr) and Ti raised stability: Ru0.6Cr0.2Ti0.2Ox ran 200 h at 100 mA cm-2 in a PEM electrolyzer with a 267 mV overpotential, while RuOx failed within about 40 h. DFT on O*-terminated (110) surfaces gave theoretical overpotentials of 0.52 V for RuO2 and 0.39 V for RuCrTiOx, and a Ru-O-Cr dimer on RuCrO4 (101) weakened OH* binding to give 0.24 V. 18O isotope DEMS showed Ti suppressed lattice oxygen participation by about 66% (Abed et al. 2024). Summary: abed2024; see also Electrochemical stability.
An immobilized iron molecular catalyst for the OER (2014)
As an alternative to oxide catalysts, whose correlated step energetics limit the minimum overpotential, Demeter and coworkers showed that an Fe(III) TAML activator supported on carbon black with Nafion catalyzes electrochemical O2 evolution in 0.1 M HNO3, although the same complex in solution did not. Constant-current electrolysis at 5 mA on carbon paper gave about 44 μmol O2 at an apparent Faradaic efficiency of about 45% and an estimated turnover frequency of up to 0.081 s-1. The rest of the charge went largely to oxidizing the carbon support to CO2 (Demeter et al. 2014). Summary: demeter2014; see also Molecular electrocatalysts.
Mixed Ni-Fe oxides in alkaline electrolyte (2012)
Landon and coworkers made Ni-Fe oxides by three routes (evaporation-induced self-assembly, hard templating and dip-coating on Ni mesh) and found OER activity in 1 M KOH peaked near 10 mol % Fe for all of them, with Tafel slopes falling from above 50 to about 40 mV/dec and no correlation with BET surface area. XRD and Raman showed a NiFe2O4 spinel coexisting with NiO rather than Fe substituted into NiO, in situ EXAFS showed the Fe coordination changing under OER while Ni did not, and the number of basic sites probed by methanol TPRS followed the same ordering as OER activity. The spinel phase was implicated as a contributor to the enhanced activity (Landon et al. 2012). Summary: landon2012; see also Doped oxides and X-ray absorption spectroscopy.
Electrolyte cation and Fe impurity effects on NiOOH (2015)
Michael and coworkers measured NiOOH thin films in 0.1 M LiOH, NaOH, KOH and CsOH that were either purified of Fe or saturated with it. Fe saturation lowered the onset overpotential by at least 100 mV and raised current density by nearly an order of magnitude in every electrolyte, consistent with Fe being central to Ni-Fe oxyhydroxide activity as in the earlier Ni-Fe oxide work above. The cation effect was smaller but real: CsOH gave about twice the current of LiOH in purified electrolyte, and with Fe present the ordering became K+ ≈ Na+ > Cs+ > Li+ (about 7, 4.5 and 2.5 mA/cm2 for KOH or NaOH, CsOH and LiOH at 0.3 V overpotential). In situ Raman linked the CsOH enhancement to slightly longer Ni-O bonds (Michael et al. 2015). Summary: michael2015; see also Electrolyte cation effects and Raman spectroscopy.