Recent research indicates a severe discrepancy between oxygen evolution reaction catalysts dissolution in aqueous model systems and membrane electrode assemblies. This questions the relevance of the widespread aqueous testing for real world application. In this study, we aim to determine the processes responsible for the dissolution discrepancy. Experimental parameters known to diverge in both systems are individually tested for their influence on dissolution of an Ir-based catalyst. Ir dissolution is studied in an aqueous model system, a scanning flow cell coupled to an inductively coupled plasma mass spectrometer. Real dissolution rates of the Ir OER catalyst in membrane electrode assemblies are measured with a specifically developed, dedicated setup. Overestimated acidity in the anode catalyst layer and stabilization over time in real devices are proposed as main contributors to the dissolution discrepancy. The results shown here lead to clear guidelines for anode electrocatalyst testing parameters to resemble realistic electrolyzer operating conditions.

The structure of IrO2 during the oxygen evolution reaction (OER) was studied by operando X-ray absorption spectroscopy (XAS) at the Ir L3-edge to gain insight into the processes that occur during the electrocatalytic reaction at the anode during water electrolysis. For this purpose, calcined and uncalcined IrO2 nanoparticles were tested in an operando spectroelectrochemical cell. In situ XAS under different applied potentials uncovered strong structural changes when changing the potential. Modulation excitation spectroscopy combined with XAS enhanced the information on the dynamic changes significantly. Principal component analysis (PCA) of the resulting spectra as well as FEFF9 calculations uncovered that both the Ir L3-edge energy and the white line intensity changed due to the formation of oxygen vacancies and lower oxidation state of iridium at higher potentials, respectively. The deconvoluted spectra and their components lead to two different OER modes. It was observed that at higher OER potentials, the well-known OER mechanisms need to be modified, which is also associated with the stabilization of the catalyst, as confirmed by in situ inductively coupled plasma mass spectrometry (ICP-MS). At these elevated OER potentials above 1.5 V, stronger Ir–Ir interactions were observed. They were more dominant in the calcined IrO2 samples than in the uncalcined ones. The stronger Ir–Ir interaction upon vacancy formation is also supported by theoretical studies. We propose that this may be a crucial factor in the increased dissolution stability of the IrO2 catalyst after calcination. The results presented here provide additional insights into the OER in acid media and demonstrate a powerful technique for quantifying the differences in mechanisms on different OER electrocatalysts. Furthermore, insights into the OER at a fundamental level are provided, which will contribute to further understanding of the reaction mechanisms in water electrolysis.

The increasing scarcity of iridium (Ir) and its rutile-type oxide (IrO 2 ), the current state-of-the-art oxygen evolution reaction (OER) catalysts, is driving the transition toward the use of mixed Ir oxides with a highly active yet inexpensive metal (Ir x M 1−x O 2 ). Ruthenium (Ru) has been commonly employed due to its high OER activity although its electrochemical stability in Ir-Ru mixed oxide nanoparticles (Ir x Ru 1−x O 2 NPs), especially at high relative contents, is rarely evaluated for long-term application as water electrolyzers. In this work, we bridge the knowledge gap by performing a thorough study on the composition-and phase-dependent stability of welldefined Ir x Ru 1−x O 2 NPs prepared by flame spray pyrolysis under dynamic operating conditions. As-prepared NPs (Ir x Ru 1−x O y ) present an amorphous coral-like structure with a hydrous Ir-Ru oxide phase, which upon post-synthetic thermal treatment fully converts to a rutile-type structure followed by a selective Ir enrichment at the NP topmost surface. It was demonstrated that Ir incorporation into a RuO 2 matrix drastically reduced Ru dissolution by ca. 10-fold at the expense of worsening Ir inherent stability, regardless of the oxide phase present. Hydrous Ir x Ru 1−x O y NPs, however, were shown to be 1000-fold less stable than rutile-type Ir x Ru 1−x O 2 , where the severe Ru leaching yielded a fast convergence toward the activity of monometallic hydrous IrO y . For rutiletype Ir x Ru 1−x O 2 , the sequential start-up/shut-down OER protocol employed revealed a steady-state dissolution for both Ir and Ru, as well as the key role of surface Ru species in OER activity: minimal Ru surface losses (<1 at. %) yielded OER activities for tested Ir 0.2 Ru 0.8 O 2 equivalent to those of untested Ir 0.8 Ru 0.2 O 2 . Ir enrichment at the NP topmost surface, which mitigates selective subsurface Ru dissolution, is identified as the origin of the NP stabilization. These results suggest Ru-rich Ir x Ru 1−x O 2 NPs to be viable electrocatalysts for long-term water electrolysis, with significant repercussions in cost reduction.