Nathan Wells2 Christopher O'Rourke2 Olivia Hendricks1 Robert Tang-Kong1 Andrew Mills2 Paul McIntyre1

2, Queens University Belfast, Belfast, Northern Ireland, United Kingdom
1, Stanford University, Stanford, California, United States

Photoelectrochemical (PEC) cells are a promising approach for solar energy storage. In traditional PEC cells, solar energy drives water oxidation at the anode and proton reduction at the cathode, “splitting” water to form O2 and H2 gases. The efficiency, however, is limited by the high overpotentials required for water oxidation.1 Splitting seawater to generate H2, NaOH, and Cl2 (equivalently H2 and NaOCl) is a promising alternative to overcome the efficiency limits of traditional PEC cells. In this case, chloride oxidation occurs at the anode to generate Cl2, and proton reduction occurs at the cathode to generate hydroxide ions and H2. While chloride oxidation occurs at similarly positive potentials to water oxidation, its overpotential is significantly lower.2 Moreover, NaOCl is a valuable chemical oxidant used extensively for water purification,3 which remains a significant worldwide problem.4

Alloy films of TiO2-RuO2 or TiO2-IrOx synthesized by atomic layer deposition (ALD) represent an ultra-thin analogue to the dimensionally stable anode (DSA), the current industry standard for chloride oxidation. While the DSA is 10s of microns thick, we deposited alloy films ranging from 10-45 nm, thus reducing noble metal usage by a factor of 100 to 1000 per unit area. An all-ALD process for alloy deposition also enables precise control of the noble metal content in the film by altering the ratio of TiO2 to RuO2 or IrOx cycles. Despite using less noble metal, ALD alloys performed competitively with a commercial DSA tested under the same conditions. For example, a 45 nm TiO2-RuO2 alloy composed with 46:54 Ru:Ti ratio had a slightly higher overpotential (74 mV vs. 58 mV at 1 mA cm-2) but an equivalent Tafel slope of 52 mV dec-1 and a higher Cl2 yield than the DSA (93% vs 87%). This 46% Ru alloy also maintained a higher current density than the DSA over a 10 hour stability test. The overpotentials for TiO2-IrOx alloys with 49% Ir varied from 41 mV to 111 mV at 1 mA cm-2, but adding an additional ALD IrOx coating did not reduce the overpotentials further. Similar experiments with PVD Ir coatings showed comparable Tafel slopes of 55mV dec-1 despite slightly lower overpotentials for Cl2 evolution (30 mV). This suggests that we do not sacrifice catalytic activity by diluting noble metal ions in a TiO2 matrix. Finally, ALD TiO2-RuO2 alloys possess a work function > 5 eV, capable of generating photovoltages > 500 mV on n-type silicon in a photoelectrochemical cell.[5] Overall, this work demonstrates the potential of ALD to tune the composition of electrode materials for practical application in solar-driven saltwater splitting.

[1] M.G. Walter, et al. Chem. Rev. 110, 6446–73 (2010).
[2] N.S. Lewis, D.G. Nocera, Proc. Natl. Acad. Sci. 103, 15729–35 (2006).
[5] O. L. Hendricks et al. ACS Applied Materials & Interfaces, 2016, 8, 23763-23773