Regulating Ru-Ru bond length enables a universal dual-site oxygen evolution pathway

Why acidic OER needed a third pathway, and why it stayed elusive

In acidic PEM water electrolysis, the anode is the bottleneck: the adsorbate evolution mechanism is capped by linear scaling among *OH, *O and *OOH, while lattice oxygen mechanisms can bypass that cap but risk lattice oxygen loss and structural degradation [1]. The dual-site oxide path mechanism is attractive because adjacent *O species on neighboring metal sites couple directly to form O-O, skipping the high-energy *OOH intermediate and potentially combining activity with durability [1]. The concept is not new, but activating it across an entire surface has been hard: it requires densely packed, geometrically aligned dual-metal sites at an optimal interatomic distance, and prior strategies that deposit Ru or Ir on smaller-lattice substrates such as α-MnO2 or Co3O4 induce compressive strain only locally, because strain relaxation in one region expands adjacent regions and limits how far dual-site activation spreads [1].

That localization problem is the baseline the new paper attacks. Earlier work on ordered oxide model catalysts in alkaline media had already shown that dopants reshape the active oxyhydroxide skin and its disorder in ways that change activity, for example iron suppressing nickel redox and increasing NiOOH disorder in Ni0.75Fe0.25Ox(001) thin films [2]. The acidic Ru-Ir case adds a harsher constraint: the same compressive strain that shortens metal-metal distances can also destabilize the framework, so a design must compress uniformly without accumulating excess surface energy that drives atomic migration and agglomeration [1].

A narrow bond-length window, and the Y-doped route into it

The anchor paper first maps the relationship between the Ru-Ru1 distance and the free-energy barrier for bridging O-O formation. Within 3.2-3.0 Å, shortening the bond lowers the barrier, but below 3.0 Å the barrier rises again, defining a window rather than a monotonic trend [1]. Screening dopants by ionic radius, the authors find that YRuIrOx relaxes to the shortest Ru-Ru1 distance inside that window, about 3.075 Å, whereas larger-radius Sr or smaller-radius Mg cannot push the distance further because the framework tends to preserve its original structure [1]. Experimentally, EXAFS peak fitting gives an average Ru-Ru1 bond length of 3.07 Å in Y0.1Ir0.2Ru0.7Ox versus 3.18 Å in Ir0.2Ru0.8Ox, with the Ru-O bond slightly extended from 1.97 to 2.04 Å, consistent with stress being dispersed through Ru-O elongation [1].

The structural trick is not just the dopant but the morphology. The catalyst is built as interconnected mesoporous crystalline oxide nanoparticles with non-epitaxial interparticle contacts, which dissipate local lattice distortions from high dopant incorporation; aberration-corrected TEM shows short-range atomic order with staggered crystallographic planes, and the long-range region beyond 4 Å is disordered, unlike the periodic features retained by Ir0.2Ru0.8Ox [1]. This locally ordered, long-range disordered architecture is what allows the compression to be homogeneous across a dense array of active sites rather than concentrated at an interface. The trade-off is that the optimal distance is defined computationally and confirmed for one composition; whether the same 3.0-3.2 Å window transfers to other Ru-based frameworks is not established here [1].

Mechanism evidence: pathway uniformity, not just pathway presence

The mechanistic case rests on operando infrared spectroscopy and 18O-labeled online mass spectrometry. In the 18O DEMS experiment, the labeled catalyst evolves a dominant signal consistent with direct coupling of adjacent adsorbed oxygen species, and the paper reports that Y0.1Ir0.2Ru0.7Ox consistently follows the dual-site pathway while Ir0.2Ru0.8Ox remains AEM-dominated [1]. The authors frame this as site-independent behavior of OPM-active sites across the surface, which is a stronger claim than detecting OPM at a few sites [1]. Independent work on an alkaline Ir-Ni dual-site electrode reaches a compatible conclusion using the same logic: 18O labeling produced 36O2 from adjacent 18O adsorbates, which the authors call the most direct evidence for OPM, and DFT plus DEMS supported Ir-Ni dual-site coupling [5]. That agreement across a different material, electrolyte and metal pair strengthens the plausibility of dual-site O-O coupling as a real pathway, but it also shows that DEMS signatures alone do not identify which specific site pair is responsible.

A separate line of evidence complicates any simple structure-activity story. Using surface interrogation scanning electrochemical microscopy on amorphous IrOx and crystalline IrO2, Kim and colleagues quantified electrolyte-accessible active oxygen species and found that activity was governed mainly by the number of accessible active sites, not by the charge-storage capacity that is often invoked to lower the OER barrier [7]. In that framework, the Y0.1Ir0.2Ru0.7Ox improvement could partly reflect a larger accessible active-site population created by the mesoporous, locally disordered architecture, in addition to any intrinsic barrier reduction from the shortened Ru-Ru1 distance. The anchor paper's DFT barrier comparison (OPM 1.53 eV versus AEM 1.97 eV for the Y-Ru-Ir model) supports an intrinsic effect, but the two explanations are not fully separated experimentally [1][7].

Device numbers, durability, and where the claim stops

In a PEM electrolyzer with a Nafion 115 membrane, the Y0.1Ir0.2Ru0.7Ox anode reaches 1.75 V at 3 A cm⁻² with low platinum-group metal loading (0.1 mg Ir cm⁻² and 0.3 mg Ru cm⁻²) and operates for more than 2800 hours at 3 A cm⁻² with a decay rate of 0.0625 mV h⁻¹ [1]. The Pourbaix analysis offers a thermodynamic rationale for the durability: within the PEMWE operating region (pH 0-3), Y0.1Ir0.2Ru0.7Ox has a 0.05 eV dissolution free-energy advantage over Ir0.2Ru0.8Ox and 0.15 eV over RuO2, and Y is predicted to remain incorporated in oxide phases rather than dissolve [1]. These are meaningful numbers because they combine high current density, low PGM loading and long duration in one device, which is the combination PEMWE deployment actually requires.

The boundaries are equally important. The mechanism and stability conclusions are tied to this Y0.1Ir0.2Ru0.7Ox composition, its synthesis route and these PEM operating conditions; the paper itself calls the design principle generalizable but does not demonstrate it across other Ru-based systems [1]. Competing approaches to acidic OER durability exist, including low-valent RuIr oxides designed for reversible valence dynamics and stable solid solutions, and electrodeposited Ti@Pt@RuIrO2 architectures, but the supplied evidence for those is abstract-level and does not allow a direct performance or mechanism comparison [3][4]. Component-level degradation can also dominate in practice: Pt coatings on the porous transport layer have been reported to delaminate and accelerate catalyst-layer degradation, a failure mode outside the catalyst's own stability [6]. Finally, the anchor paper's own durability test used an automatic start-stop setting with a 20-second start-stop every hour, so the 2800-hour figure should not be read as a full intermittent-renewable duty cycle [1].

What would turn a single-composition result into a design rule

The most useful next test is transferability: applying the same interatomic-distance logic to other Ru-based or Ir-based frameworks and checking whether the 3.0-3.2 Å barrier window and the locally ordered/long-range disordered morphology still coincide with dual-site behavior [1]. A second test is separating intrinsic barrier effects from active-site accessibility, since SI-SECM work shows that accessible site count can dominate activity trends even when intrinsic reactivity is comparable [7]. Third, the field needs operando methods that identify the specific coupling site pair, because 18O DEMS establishes O-O coupling but not which neighboring metals are involved, as the Ir-Ni study also illustrates [5]. Until those pieces are in place, the honest reading is that this paper converts the oxide path mechanism from a sporadic observation into a tunable structural target, while leaving its universality an open, testable question [1].

About These Sources

This research page is built on 7 peer-reviewed studies — published from 2024 to 2026, 7 from 2024 or later — selected as the most relevant from 7 studies that passed quality screening, drawn from 64 papers retrieved from a database of over 500 million.

Sources used in this answer