Water oxidation watch: O-O bond pathways stay unsettled

Coverage note: the current-week search did not surface enough fetchable primary papers to make a useful issue. For this launch read, I widened the window to recent mechanism-heavy papers from 2025-2026; regular issues should tighten back to the prior week.

The strongest signal this week is not a new record current density or a larger catalyst library. It is a set of papers asking the same uncomfortable question from different sides: when water oxidation makes the O-O bond, are we looking at radical coupling, nucleophilic attack, protein-steered substrate selection, or a surface-induced proton-coupled electron-transfer state?

That distinction matters. If radical coupling is right for a class of catalysts, design pressure falls on stabilizing high-valent oxo/oxyl pairs close enough to couple. If nucleophilic attack is right, the local base, orbital accessibility, and solvent structure become design variables. If the protein or semiconductor surface pre-organizes the substrate, then the catalyst is not the whole mechanism; the surrounding hydrogen-bond or surface-electronic environment is part of the active site.

The OEC papers are moving in two complementary directions. Pushkar's 2026 proposal tries to change the identity of the coupling oxygens: O3, dynamically hydrogen-bonded to His337, couples with O6 generated at Mn1 during the S2-to-S3 transition. The point is not just a different atom label. It says the protein matrix may select the reactive pair by charge control, hydrogen bonding, and open coordination at Mn1. 1

Guo et al. tackle the other side of the cycle: after O2 leaves, how does the Mn4Ca cluster become catalytically ready again? Their simulations start from an O2-release cavity and find water insertion through closed-cubane intermediates, followed by proton release, water dissociation, and ligand transfer toward S0. They also flag the W2 protonation state as unresolved because different spectroscopies and calculations favor different assignments. 2

Read together, these papers make a useful correction to the usual cartoon of the Kok cycle. The O-O step is not the only unknown. Substrate identity, proton placement, ligand rearrangement, and the reset step are all still live mechanistic variables.

The Fe paper is the cleanest experimental mechanism entry in this set. Patel et al. start with a μ-oxo diiron complex, but in situ XAS indicates the dimer breaks into monomers under catalytic conditions. The evidence package then points toward radical coupling: second-order dependence on catalyst, a minimal isotope effect near kH/kD = 1, detected FeV=O, and DFT preference for the radical-coupling route. 3

That does not make radical coupling the default answer for first-row metals. Huo et al.'s Mn paper argues the opposite direction for a mononuclear manganese catalyst: free hydroxide nucleophilic attack controls O-O bond formation, and the barrier depends on both hydroxide nucleophilicity and access to the Mn dz2 orbital. 4

The practical takeaway is simple: "earth-abundant WOC" is too broad a mechanism category. Fe dimers/monomers, mononuclear Mn complexes, and Mn-Ca OEC mimics may all oxidize water, but the rate-limiting O-O chemistry can be governed by different local variables.

The new Mn2(III)Ca(II) complex is useful because it puts water oxidation activity, magnetic coupling, and OEC-inspired Mn-Ca composition in the same experimental object. The reported TOF values, 2.41 s-1 with H2O and 3.5 s-1 with OH-, give readers a concrete activity scale, while the magnetic data show antiferromagnetic exchange between the Mn(III) centers. 5

The caution is just as important. A Mn-Ca core can teach us about structural and electronic motifs that resemble the OEC. It does not automatically tell us whether the catalyst forms O-O by oxo-oxyl coupling, hydroxide attack, water nucleophilic attack, or a decomposition-derived active phase. For this channel, biomimetic papers should be treated as mechanism candidates until spectroscopy, kinetics, isotope effects, and computational alternatives line up.

Cruz et al. show why heterogeneous artificial-photosynthesis mechanisms are hard to reduce to a single active site. On polymeric carbon nitride, D2O adsorption shifts nitrogen and carbon XPS features and moves the valence-band position by 1.04 eV to higher binding energy. Under illumination, the authors interpret the electronic shifts as support for a proton-coupled electron-transfer route in which water adsorption helps create the reactive surface state. 6

That is a useful contrast with molecular WOCs. In a molecular complex, the argument often centers on a metal-oxo, metal-oxyl, or incoming nucleophile. In carbon nitride, the water adsorbate and semiconductor surface jointly change the electronic starting point. The mechanism is not just what the catalyst does to water; it is what water does to the catalyst before the photon-driven step.

Three follow-up tests would make the next few months easier to read.

First, OEC papers need experiments that can distinguish O3-O6 from O5-O6 coupling without relying on a single structural snapshot. Substrate-water exchange, isotope labeling, and time-resolved spectroscopy are the obvious pressure points.

Second, first-row molecular WOCs need mechanism packages, not just activity metrics. The Fe paper is a good model because kinetics, isotope effects, spectroscopy, and DFT all bear on the O-O step. 3

Third, heterogeneous photocatalyst papers should state whether water adsorption changes the catalyst's electronic structure before claiming a water-oxidation mechanism. The carbon nitride study shows that this pre-reaction state can be experimentally visible. 6

For now, the debate is not "which catalyst is best." It is which local environment makes O-O formation chemically plausible, and how much evidence is needed before a proposed pathway becomes more than a tidy scheme.