ALD-Deposited NiO Approaches the Performance of Platinum as a Hydrogen Evolution Cocatalyst on Carbon Nitride

The photocatalytic hydrogen evolution reaction is one of the most promising approaches to utilize solar energy and produce hydrogen as an alternative energy source to traditional fossil fuels. In this work, we focused on further exploring the cocatalyst potential of the transition metal nickel and prepared nickel oxide/carbon nitride (NiO/CN_x) heterojunctions through thermal polymerization as well as plasma-enhanced atomic layer deposition (PEALD). The optimal NiO loading thickness on the surface of the CN_x was found to be 6.5 nm, prepared from 65 cycles of ALD. This NiO(65)/CN_x had a photocatalytic hydrogen evolution rate of 3.9 μmol h^–1 in a 2 mg mL^–1 aqueous photocatalyst solution (195 μmol h^–1 g^–1), corresponding to an AQY of 3.1% and exhibiting 54% of the activity of 3 wt % Pt/CN_x under 405 nm light illumination with a sacrificial reagent. A 3 wt % Ni/CN_x prepared through a photodeposition process from a dispersion of CN_x and simple NiCl₂ had a photocatalytic hydrogen evolution rate of 2.8 μmol h^–1 (140 μmol h g^–1) and an AQY of 2.2% with 2 mg mL^–1 photocatalyst in the reactant solution, corresponding to 39% of the activity of 3 wt % Pt/CN_x. An induction period at the beginning of the hydrogen evolution process was found in all Ni-based photocatalysts, which we related to a Ni in situ photoreduction from Ni^II to Ni⁰ and the reverse reaction. In addition, the reproducibility of the induction period even in an oxygen-free environment suggests that electrons flow back from reduced Ni species to CN_x under dark conditions. Furthermore, a large population of trap states was detected in the bulk CN_x materials via photoinduced absorption spectroscopy, and the trapping severely hinders cocatalysts (even the benchmark Pt cocatalyst) from extracting charges from CN_x for subsequent interfacial redox reactions. We conclude that not only does the surface/interface engineering of the photocatalyst/catalyst heterojunctions need to be considered but also trap states in CN_x that severely limit the transfer of photogenerated charges to cocatalysts.