Solar water splitting (SWS) has been widely studied as a promising technology for generating carbon-free hydrogen. In overall water splitting, the cathode and the anode produce simultaneously H₂ and O₂, respectively. This process requires a bias of 1.23 V between the two electrodes. Considering the various intrinsic losses in the system (such as the catalytic overpotentials), the bias is reported as 1.6-2.4 V.¹ In SWS, the active core of the system absorbs sunlight to generate non-equilibrium electron-hole pairs producing a bias. Considering an absorber made of a single material, this bias is inversely proportional to the amount of sunlight absorbed, leading to the Shockley–Queisser limit. To achieve high bias while maximising the absorption, one can use a type II heterojunction of two absorbers in a Z-scheme configuration.² Moreover, to improve the overall water splitting process, the electron-hole pairs must be generated as close to the surface as possible. Here, we present a complete design of a direct Z-scheme system based on two-dimensional (2D) transition metal dichalcogenide (TMDC). We also implement a multi-physics model using density functional theory (DFT), the detailed balance method and the Butler-Volmer kinetics to compute the STH efficiency. TMDCs and their van der Waals heterojunctions (vdWH) are attractive for photovoltaic³ and catalytic^4,5 applications and the production of 2D TMDC is easy and low-cost thanks to the exfoliation process.⁶ The active core of the presented system is a MoS₂/WSe₂ vdWH where MoS₂ and WSe₂ are the anode and the cathode, respectively.

The scheme of the proposed cell is presented in Fig. 1(a) where two distinct regions form its active core: with and without a 2D hexagonal boron nitride (hBN) layer isolating the MoS₂ and WSe₂ layers. Fig. 1(b) represents the band diagram of region 1, where the two layers are isolated. In this region 1, electrons generated in WSe₂ and holes generated in MoS₂ produce the electrochemical reactions. As confirmed by our DFT calculations, the band edges positions are suitable with χ_O2 and χ_H2 equal to 0.66 eV and 0.50 eV respectively at pH 7 (see Fig 1(b)). The extra carriers generated in the first region (i.e., the electrons of MoS₂ and the holes of WSe₂) must recombine to avoid an accumulation of carriers which would stop the SWS process. They have to diffuse into region 2, without hBN. As depicted in Fig. 1(c), which presents the band diagram of this corresponding region, the recombination is ensured by the hybridisation of the electronic states in the valence band of the TMDCs. This hybridisation, which is confirmed by our DFT calculations, has also been experimentally observed⁷. Our ab initio calculations show very efficient recombination. Consequently, a small fraction of region 2 suffices to enable the recombination of all extra electron-hole pairs. This is advantageous for the design of the device since we can imagine stacking exfoliated flakes of MoS₂, hBN and WSe₂ randomly to produce the active core. The non-ideal stacking will inevitably create some regions 2 without hBN. Finally, we propose that the ultra-thin active core is protected and supported by mesoporous transparent oxide, enabling the free circulation of gases and water.

Since some parameters of the considered materials are not well known, we calculate the STH efficiency using our multi-physics model considering 3 different cases: ideal, high-performance, and standard quality. In the latter case, with an optimized ratio of regions 1 and 2 and considering a high optical absorption, we find a 14 % efficiency Since the ultimate thinness of the active region limits the optical absorption, we propose that the mesoporous oxide serves as a support for absorption enhancement systems, such as photonic crystals, resonant cavities, nanoparticles.

This work, therefore, proposes the design of a system that exceeds the critical 10% efficiency required to make SWS economically viable.⁸ Since our system is wireless and requires simple manufacturing processes (exfoliation), this result is remarkable.

References

1. M. G. Walter et al., Chem. Rev., 110, 6446–6473 (2010).
2. Q. Xu et al., Materials Today, 21, 1042–1063 (2018).
3. S. Das, D. Pandey, J. Thomas, and T. Roy, Adv. Mater., 31, 1802722 (2019).
4. J. Wu et al., Adv. Mater. Interfaces, 3, 1500669 (2016).
5. X. Yin et al., Chem. Soc. Rev., 50, 10087–10115 (2021).
6. K. Si et al., Appl. Surf. Sci., 507, 145082 (2020).
7. O. Karni et al., Phys. Rev. Lett., 123, 247402 (2019).
8. Y. Fan, J. Wang, and M. Zhao, Nanoscale, 11, 14836–14843 (2019).