A Combined a-SAXS and XAS Study of the Operando Compositional and Morphological Changes Undergone By High Surface Area IrO2 Oxygen Evolution Electrocatalysts

Abstract

Sluggish kinetics of the anodic process in polymer electrolyte water electrolyzers (PEWEs) - oxygen evolution reaction (OER) - is the main contributor to the efficiency losses.1 Iridium oxide is the state-of-the-art OER catalyst, however its high price and low abundance2 require minimizing its content in the anode while keeping high OER activity. This can be achieved by maximizing the mass specific surface area (SA) of IrO2, e.g. by synthesizing small IrO2 particles.3 On the other hand, high-SA IrO2 catalysts tend to degrade when operating under the dynamic conditions typical for PEWE’s coupled with intermittent renewable energy source such as wind or solar, and the mechanisms causing this instability remain poorly understood.4 In order to study the degradation mechanisms in detail, we used a modified, chlorine-free Adams fusion method to produce a high-SA IrO2 catalyst3 (referred as ‘IrO2-AS’ in what follows) that displayed a SA of 350 m2∙g-1 based on N2 adsorption / desorption measurements. A fraction of this as-synthetized catalyst was heat treated for 1 h at 400 °C in air to yield a second sample (‘IrO2-HT’) with a SA of 250 m2∙g-1. These two samples were electrochemically tested in 0.1 M HClO4 electrolyte measuring the polarization curve and applying an accelerated stress test (AST) in which the potential is cycled 500 times between 1.0 and 1.6 V vs. the reversible hydrogen electrode (RHE) holding the potential for 10 s at each value, thus simulating the dynamic operation conditions discussed above. The polarization curve measurement indicate an initial activity of 1.490 V and 1.509 V (at i= 10 A∙gox -1) for IrO2-AS and IrO2-HT, respectively. While, the latter AST led to the loss of 50 % vs. 30 % of the initial OER-activity displayed by IrO2-AS vs. IrO2-HT, respectively. To shed light on the reasons for this difference in stability, both catalysts were characterized using a novel setup available at the SuperXAS beamline of the Swiss Light Source and that allows to study the morphological and compositional properties of an (electro)catalyst by combining anomalous small angle X-ray scattering (A-SAXS) and X-ray absorption spectroscopy (XAS), respectively.5 Base on the X-ray absorption fine structure (XANES) and the Extended X-ray absorption fine structure analysis the operando XAS measurements (at 1.0 V vs. RHE) proved that neither of the catalysts suffered from significant changes in their chemical composition in the course of the AST. Complementarily, the A-SAXS curves of the IrO2-AS sample did change during the AST (see Fig. 1a), and were successfully fitted using a model that implies particles with a disk shape assembled in a 2D-agglomerated structure. The model suggests that during the AST the disks’ diameter grew and their thickness became thinner (see Figs. 1c and 1d, respectively), while the 2D-agglomerated structure remained stable. This allowed us determining the variation of SA and corresponding iridium mass, leading to the conclusion that the sample’s loss of OER-performance mainly stems from Ir-dissolution. On the contrary, for IrO2-HT the A-SAXS curves did not change during the AST (cf. Fig. 1b), thus proving that this catalyst displays a stable morphology that is consistent with its enhanced stability in the AST. Chiefly, these A-SAXS-derived conclusions were validated through high-resolution transmission electron microscopy measurements of the initial and post-mortem materials. In conclusion, the combined use of A-SAXS and XAS unveiled the reasons for the different stability behavior of pristine and calcined iridium oxides. Specifically, while XAS showed that the two samples did not change composition in the course of the AST, the A-SAXS results obtained for IrO2-AS indicated the changes in its disk morphology driven by Ir-dissolution. Comparatively, the enhanced stability displayed by IrO2-HT can be attributed to the preliminary agglomeration of its oxide particles caused by the heat-treatment step, which prevents drastic shape variation and limits the loss of Ir during the AST. M. Bernt and H. A. Gasteiger, J. Electrochem. Soc., 163, 3179–3189 (2016).P. C. K. Vesborg and T. F. Jaramillo, RCS Adv., 2, 7933–7947 (2012).D. F. Abbott et al., Chem. Mater., 28, 6591–6604 (2016).C. Rakousky et al., J. Power Sources, 326, 120–128 (2016).M. Povia et al., ACS Catal., 8, 7000–7015 (2018). Figure 1