In this talk, we discuss a lead-free photovoltaic absorber candidate, CsBiSCl₂, which has recently attracted attention following a report of >10% power conversion efficiency in solution-processed thin-film solar cells [1]. If repeatable, this would represent a highly promising breakaway material in the field of lead-free photovoltaics. While other works speculated CsBiSCl₂ may adopt a perovskite structure, its true structure remains unverified, and the synthetic route used to produce the reported films has not been verified. We conduct a comprehensive investigation into the Cs–Bi–S–Cl system, integrating global structure search with targeted experimental synthesis to determine whether CsBiSCl₂ is a realistic photovoltaic candidate.

Using ab initio Random Structure Searching accelerated by a bespoke Ephemeral Data Derived Potential, we explore the Cs-Bi-S-Cl potential-energy surface and identify a four-formula-unit CsBiSCl₂ orthorhombic Pnma structure lying 2.4 meV per atom above the convex hull. This phase is dynamically stable and represents an energetically isolated and plausible CsBiSCl₂ polymorph. Perovskite-derived structures (e.g. cubic, tetragonal) were found to be energetically implausible, and simulated diffraction patterns bear little resemblance to those attributed to CsBiSCl₂ in previous reports.

Attempting to repeat the previously reported solution synthesis route for thin-film fabrication, we find that the intermediate “DMABiS₂” precursor does not readily form. Instead, the reaction yields Bi₂S,, alongside persistent iodide residues that cannot be removed through solvent washing. Thin films prepared from this intermediate contain mixtures of Cs₃Bi₂I₉ and Bi₂S₃, and show no evidence of chloride incorporation. Parallel solid-state reactions performed in the absence of iodine likewise fail to produce CsBiSCl₂, instead forming only binary and ternary phases that lie on the convex hull.

Taken together, these results demonstrate that CsBiSCl₂ is difficult to access synthetically, metastable with respect to competing phases, and unlikely to be an effective absorber in a solar cell. We therefore caution against further investigation into this material. This work has recently been published in EES Solar [2]

[1] The Journal of Physical Chemistry Letters 2024 15 (12), 3383-3389 - DOI: 10.1021/acs.jpclett.4c00310

[2] EES Sol., 2025, Advance Article - DOI: 10.1039/D5EL00157A