Influence of Temperature on the Performance of Carbon- and ATO-supported Oxygen Evolution Reaction Catalysts in a Gas Diffusion Electrode Setup

State-of-the-art industrial electrocatalysts for the oxygen evolution reaction (OER) under acidic conditions are Ir-based. Considering the scarce supply of Ir, it is imperative to use the precious metal as efficiently as possible. In this work, we immobilized ultrasmall Ir and Ir0.4Ru0.6 nanoparticles on two different supports to maximize their dispersion. One high-surface-area carbon support serves as a reference but has limited technological relevance due to its lack of stability. The other support, antimony-doped tin oxide (ATO), has been proposed in the literature as a possible better support for OER catalysts. Temperature-dependent measurements performed in a recently developed gas diffusion electrode (GDE) setup reveal that surprisingly the catalysts immobilized on commercial ATO performed worse than their carbon-immobilized counterparts. The measurements suggest that the ATO support deteriorates particularly fast at elevated temperatures.


Chemicals and materials
Ultra-pure water (MilliQ-system, 2.7 ppb total organic carbon (TOC), 18.2 MΩ) was used to clean the GDE cell and to prepare the electrolyte and the catalyst ink.
For the catalyst synthesis, hydrated IrCl3 (99.8% metals basis) and RuCl3 (ReagentPlus) were purchased from Alfa Aesar and Sigma Aldrich, respectively, and stored in a glovebox. EtOH (EtOH absolute, VWR Chemicals) was used to dissolve the precursor salt, to prepare the alkaline (NaOH, Hänseler) solvent, and to disperse the support carbon Ketjen black (EC-300J, Fuel Cell Store) and ATO (NanoArc, 99.5%, Alfa Aesar). The catalyst inks were prepared using isopropanol (IPA, HPLC grade, VWR Chemicals), KOH (Hänseler) and a Nafion dispersion in H2O (D1021, Fuel Cell Store). A horn sonicator (Q500 sonicator, QSONICA sonicators) was used to disperse the support and the catalyst. A rotary evaporator (RC 600) from knf was employed to evaporate the solvent.
For conductivity measurements, a Keithley 2400 multimeter and a laser distance sensor LAR-10-5V from Waycon Präzisionstechnik GmbH were used.     Figure S4. XANES spectra (a) and Fourier transform magnitudes of the k 2 -weighted extended X-ray absorption fine structure (EXAFS) (b) and (c) data of pristine Ir/C and Ir/ATO measured at the Ir LIII-edge.

XAS of pristine Ir/C and Ir /ATOinfluence of the support on the structure of the NPs
As both the Iridium XANES spectra and the EXAFS of the NPs supported on C and ATO agree, we find no influence of the support on the structure of the metal NPs. This is further supported in the modelling of the EXAFs data, see Table S2.    All samples beside Ir0.4Ru0.6/C face a particle growth after the activation step. The reason behind the non-growth of Ir0.4Ru0.6/C is explained by an aging process. The pristine film was prepared about a year prior the actual SAXS measurement, thus the NPs had time to oxidize in air. Therefore, both pristine and activated samples demonstrate a similar particle size.
Moreover, based on the density of Ir and IrO2 (Ir: 22.6 g cm -3 , IrO2: 2.00 g cm -3 ) 2 and assuming fully reduced pristine species, a slightly higher than two-fold growth is expected if the entire NPs would be fully oxidized (from 1.8 to 4.0 nm). However, according to the obtained SAXS results, the particles grew only 1 nm (Table S5). This was explained by an incomplete oxidation of the NPs. Figure S8. PDF of activated Ir/ATO. The peaks arising from either Ir or IrO2 are difficult to distinguish due to the presence of the oxide support. Table S6. Parameters obtained for EXAFS data fitting of activated Ir/C and Ir0.4Ru0.6/C on the Ir LIII-edge and the Ru K-edge, that shows first nearest neighbour coordination shell (N), atomic bond length (R), Debye Waller factors (mean squared bond length disorder) (σ 2 ), absorption edge energy (E0), and Rf-factor as a measure of fit quality.     Activated Ru-containing samples at 60 °C were analysed to quantify any elemental leaching.

EDX data -Ru leaching
It was found that Ru was partially and not homogenously leached out of the sample. The edges of the 3 mm sample showed a decrease of the Ru content, while the centre displayed an identical ratio as the pristine one.

EDX data -Sb leaching
The suspected Sb leaching from the ATO support was determined post-mortem by EDX. It must be noted that the pristine and the post-mortem samples are different samples. In Fig. S11 and S12, the post-mortem samples measured at 30 °C and 60 °C for IrOx/ATO and Ir0.4Ru0.6Ox/ATO are represented.