Spectroelectrochemical study of water oxidation on nickel and iron oxyhydroxide electrocatalysts

Ni/Fe oxyhydroxides are the best performing Earth-abundant electrocatalysts for water oxidation. However, the origin of their remarkable performance is not well understood. Herein, we employ spectroelectrochemical techniques to analyse the kinetics of water oxidation on a series of Ni/Fe oxyhydroxide films: FeOOH, FeOOHNiOOH, and Ni(Fe)OOH (5% Fe). The concentrations and reaction rates of the oxidised states accumulated during catalysis are determined. Ni(Fe)OOH is found to exhibit the fastest reaction kinetics but accumulates fewer states, resulting in a similar performance to FeOOHNiOOH. The later catalytic onset in FeOOH is attributed to an anodic shift in the accumulation of oxidised states. Rate law analyses reveal that the rate limiting step for each catalyst involves the accumulation of four oxidised states, Ni-centred for Ni(Fe)OOH but Fe-centred for FeOOH and FeOOHNiOOH. We conclude by highlighting the importance of equilibria between these accumulated species and reactive intermediates in determining the activity of these materials.


Elemental Analysis
The elemental analyses of the electrocatalysts, both as prepared and after electrochemical tests, were carried out by dissolving the electrocatalyst films in 10 mL of 10 wt. % nitric acid (Fisher Chemical, TraceMetal Grade) and the amount of Fe and Ni in the resulting solutions quantified using inductively coupled plasma optical emission spectroscopy (ICP-OES). All disposable pipette tips, polypropylene tubes, and beakers were pre-rinsed with ~5 wt. % nitric acid, followed by rinsing with DI water and drying in air prior to use to remove any residual Fe. ICP standard solutions of Fe and Ni with various concentrations (10, 50, 100, 300, and 500 ppb) were used for calibration and 100 ppb yttrium was used as an internal standard to compensate for the signal drift among samples (Sigma Aldrich 43149 (Fe), 28944 (Ni), and 01357 (Y)). The atomic % of Fe in Ni(Fe)OOH and FeOOHNiOOH samples was determined using Eq. 1. The ICP-OES data is summarized in Supplementary Table 1.

Electrocatalyst Activation Process
The anodic electrodeposition methods used in this study produced FeOOH and NiOOH films. While FeOOH is stable in the air, black NiOOH gradually changes to transparent Ni(OH)2 in the air when it is not anodically protected. Therefore, Ni(Fe)OOH and FeOOHNiOOH samples were activated before optical and electrochemical experiments to convert Ni(OH)2 in these samples to NiOOH. This process consists with the oxidation of the Ni II (OH)2 to the higher valence NiOOH, leading to strong absorbance assigned to metal-to-oxygen charge transfer transition. To do this, the potential was swept from the open circuit potential of the sample to 1.5 V vs. the Ag/AgCl RE (scan rate = 10 mV/s) five times. The conversion of Ni(OH)2 to NiOOH could be confirmed visually by the return of the black color.
The percentage of Ni centres oxidised in the activation process for Ni(Fe)OOH and FeOOHNiOOH films was calculated from the spectroelectrochemistry data in Figure 2c in the main article and the ICP-OES results.

Extinction Coefficient Estimation
Two different approaches have been used to estimate the extinction coefficient (Ɛ) of all the species.
(1) Ni(Fe)OOH(+): This extinction coefficient was estimated using the Lambert-Beer law combining linear sweep voltammetry experiments and optical measurements. We integrated the current under the redox wave in Supplementary Figure 4 to estimate the number of electrons involved in this process. The absorbance at 528 nm is taken from the UV-Vis spectrum presented in Supplementary Figure 5b.

Equation 2
= × where A = Absorbance at a particular wavelength, Ɛ is the extinction coefficient and c is the concentration of electrons per cm 2 . (2) FeOOH(++), FeOOHNiOOH(++) and Ni(Fe)OOH(++): The extinction coefficient of the second oxidized species in each case was estimated using the SP-SEC technique. We monitored the changes in absorbance when a voltage step was applied. The recorded optical data is proportional to the population of the oxidized states at the applied potential ( Supplementary Figures 10a-12a).
Simultaneously, the corresponding current is measured ( Supplementary Figures 10b-12b) with an increase in current at the high applied potential and a reductive spike observed once the potential is switched to a lower one. The latter spike corresponds to the reduction of the accumulated oxidized states at the high potential region. The integration of these reductive currents (Inset Supplementary  Figures 10b-12b) allow us to quantify the electrons used to reduce the oxidized states. Using the Lambert-Beer law (equation 2) we can estimate the corresponding extinction coefficients by plotting ΔO.D. against the electrons extracted ( Supplementary Figures 10c-12c). The slope of the corresponding graphs yields the Ɛ in cm 2 /number of e -, which can be transformed to M -1 cm -1 as previously detailed. 5

TOF (s -1 ) and τ (s) calculation from measured [++] state densities and electrochemically measured water oxidation current densities.
In order to calculate these parameters we need the number of accumulated [++] (cm -2 ), which come from the conversion of the obtained ΔO.D. using the extinction coefficient. On the other hand the current density data (A/cm 2 ) can be transformed into number of electrons/s using the electron charge (1.60 x 10 -19 Coulombs).
For TOF calculation we consider that to produce one molecule of oxygen 4 oxidized species are needed Equation 3.
τ is the lifetime of the oxidized species MOOH(++)    (orange) is also shown (scan rate = 10 mV/s). The distorted slope for the FeOOHNiOOH is not present on the activated sample as can be observed in Figure 1 in the main article.

Optical Measurements in Organic Solvent
In order to determine if water is involved in the second oxidation of FeOOH, we carried out spectroelectrochemical experiments in propylene carbonate. As can be observed in Supplementary  Figure 13c, in the absence of water the catalytic current is very low and shifted to higher potentials. The addition of 2.5% water induces an oxidative process seen around 1.2 V vs Ag/AgCl. This process can be tentatively assigned to the generation of the catalytic FeOOH(++) species, which cannot be generated in the absence of water. This is confirmed by spectroelectrochemical experiments in water (Supplementary Figure 13a) and in the absence of water (Supplementary Figure 13b). As shown, the change in absorbance, and thus the formation of the (++) species, is dependent on contact with water and does not form in its absence.  Schematic representation of a possible order four mechanism considering a dimer as a reactive cluster. The spheres highlighted in yellow would be those observed under the steady state of the reaction. This merely acts as a visual aid to demonstrate how a 4 th order reaction could arise in an example system.