Stability of CoPx Electrocatalysts in Continuous and Interrupted Acidic Electrolysis of Water

Abstract Cobalt phosphides are an emerging earth‐abundant alternative to platinum‐group‐metal‐based electrocatalysts for the hydrogen evolution reaction (HER). Yet, their stability is inferior to platinum and compromises the large‐scale applicability of CoPx in water electrolyzers. In the present study, we employed flat, thin CoPx electrodes prepared through the thermal phosphidation (PH3) of Co3O4 films made by plasma‐enhanced atomic layer deposition to evaluate their stability in acidic water electrolysis by using a multi‐technique approach. The films were found to be composed of two phases: CoP in the bulk and a P‐rich surface CoPx (P/Co>1). Their performance was evaluated in the HER and the exchange current density was determined to be j 0=−8.9 ⋅ 10−5 A/cm2. The apparent activation energy of HER on CoPx (E a=81±15 kJ/mol) was determined for the first time. Dissolution of the material in 0.5 M H2SO4 was observed, regardless of the constantly applied cathodic potential, pointing towards a chemical instead of an electrochemical origin of the observed cathodic instability. The current density and HER faradaic efficiency (FE) were found to be stable during chronoamperometric treatment, as the chemical composition of the HER‐active phase remained unchanged. On the contrary, a dynamic potential change performed in a repeated way facilitated dissolution of the film, yielding its complete degradation within 5 h. There, the FE was also found to be changing. An oxidative route of CoPx dissolution has also been proposed.

. a) Au 4f XP spectrum of a fresh CoPx (400 °C) film (substrate); b) XRD patterns of the substrate, Co3O4 and CoPx films; c) Raman spectrum of the bare cleaned Au/Ti/Si substrate.
Peaks at 180, 200, 355 and 370 cm -1 correspond to HeNe laser impurities such as: contribution of additional laser modes and Raman lasing (Figure S2 c). [1,2] The peak at 520 cm -1 can be attributed to Si of the Au/Ti/Si substrate.  Figure S3. a) Comparison of activity and stability between CoPx cathodes prepared at different temperatures; b) potential-dependent m/z = 2 (H2 + ) and m/z = 34 (PH3 + ) ion currents recorded on CoPx during cathodic potential scan in 0.5 M H2SO4.

DFT calculation of CoP structure and Raman spectrum:
To calculate the Raman spectrum of the CoP crystal structure (orthorhombic, Pnma (62)), a twostep procedure was followed. At first, spin-polarized optimization of the CoP crystal structure was performed using periodic density functional theory (DFT) with the PBE exchange-correlation functional [3] as implemented in the Vienna Ab-Initio Simulation Package (VASP). [4][5][6][7][8] The electronic wavefunctions were expanded following the projected-augmented-wave scheme (PAW) to describe the electron-ion interactions. Integration in the first Brillouin zone was performed using a 7x7x7 Monkhorst-Pack k-point mesh. The total energies were computed with a cut-off energy of 400 eV and a root-mean-square (RMS) force convergence criterion of 0.01 eV/Å. Next, the unit cell was expanded in the b-direction such that a 1x2x1 supercell with the chemical composition Co8P8 was obtained. The atom positions were used directly in Gaussian 09 Rev. A0.2 [9] to compute vibrational frequencies and intensities of the Co8P8 cluster without prior geometry optimization.
The optimization step was omitted in order to avoid deformation of the Co8P8 cluster due to it not being part of a periodic structure anymore. The vibrational frequencies and intensities were calculated using the PBE exchange-correlation functional and the 6-311g all-electron basis set.
The exact procedure to obtain the Raman spectrum is outlined elsewhere. [10] Calculated intensities should be considered qualitative.  A fresh CoPx electrode was subjected to a constant CA treatment at E = -0.12 V. In order to evaluate the temporal evolution of Co dissolution, the electrolyte was sampled after 2.5 h, 5.0 h and 7.5 h of electrolysis. Dissolution rates were evaluated by ICP-OES. Initial dissolution rates (CA < 2.5 h) could not be determined by the available instrumentation due to the low concentration of Co ions. However, a more accurate evaluation of initial dissolution rates could be done by utilizing for instance online ICP methods such as online ICP-MS connected to a flow cell. [11]  Influence of exposure to air on CoPx composition: Figure S12. P 2p XP spectra of a fresh CoPx electrode before and after 60 min of exposure to ambient air.
CoPx was subjected to ambient air oxidation for 60 min. The duration of air exposure was deliberately chosen to be significantly longer (60 min) than in the actual experiments (~15 min). The increase of InPOx content (BE = 138-131 eV) is due to surface oxidation upon air exposure.