The last decade has witnessed significant progress in the development of solar water splitting devices, with solar-to-hydrogen (STH) efficiency as high as 30% already demonstrated. However, two major challenges remain. First, the high-efficiencies (> 15%) have only been achieved using devices based on expensive and non-scalable III-V semiconductors. On the other hand, low-cost metal-oxide based devices, mainly using BiVO₄ as the absorber, have only achieved STH efficiency of < 10%. Due to stability limitations, many of these metal-oxide based devices are operated in near-neutral pH electrolytes, which presents an additional mass transport challenge. Second, the majority of the demonstrated devices are still at the laboratory scale. Reports on large-area devices start to emerge, but they typically show much lower efficiencies. This is best illustrated in a recent review:^[1] even when III-V semiconductor-based devices are considered, there is no report of devices with a semiconductor absorber area larger than 10 cm² and STH efficiency > 10%.

In this talk, we will discuss the scale-up of our photoelectrochemical water splitting devices based on a complex metal oxide photoabsorber. Factors other than the semiconductor photoabsorber itself are found to be responsible for a total voltage loss of > 500 mV and therefore limit the overall performance of the large-area device.^[2] To properly address this limitation, we quantify and break down the different loss mechanisms associated with the device scale-up and the practical operational conditions.^[3] Concentration overpotential due to pH gradient is found to be a major contributor to the performance loss, and we show using multiphase multiphysics simulations and in-situ fluorescence measurements that careful control of natural and forced convection can overcome this limitation.^[3-5] In addition, we also explore the possibility to achieve efficient product separation in devices with and without separators. The product crossover, optical and Ohmic losses are quantified using a combination of experiments and simulations, and the optimization of the device working parameters and/or separator properties to achieve the minimum overall loss will be discussed.^[6,7]

References

1. J. H. Kim et al., Chem. Soc. Rev. 48, 2019, 1908
2. I. Y. Ahmet et al., Sust. Energy Fuels 3, 2019, 2366
3. F. F. Abdi et al. Sust. Energy Fuels 4, 2020, 2734
4. K. Obata et al. Energy Environ. Sci. 13, 2020, 5104
5. K. Obata & F. F. Abdi, Sust. Energy Fuels, 5, 2021, 3791
6. K. Obata et al. Cell Rep. Phys. Sci. 2, 2021, 100358
7. C. Özen et al. in revision