Bismuth sulfide absorbs light strongly but holds on to charges, study finds

A semiconductor called bismuth sulfide absorbs light strongly yet loses its light-generated charges unusually slowly, physicists at the University of Oxford and collaborators reported on October 28 in Nature. The two properties normally pull against each other in solar materials, and the authors say the mechanism behind the result could guide the search for better solar absorbers.

The trade-off has deep roots. The paper’s abstract notes that Kirchhoff demonstrated the reciprocity between absorptivity and emissivity for blackbody radiation in 1860: “a good absorber must also be a good emitter.” In practice, according to the University of Oxford, solar cell materials that absorb light strongly tend to lose their captured charges quickly, as electrons and holes recombine, which limits efficiency.

Bismuth sulfide (Bi₂S₃) appears to avoid this. It absorbs light mainly through one strong, direct energy transition, with a weaker, lower-energy indirect transition sitting just below it. The team’s fits put that gap at about 38 meV, and first-principles calculations give about 40 meV. Charges settle into the lower state, where they are lost slowly, so it acts as a temporary reservoir.

The paper reports a bimolecular recombination coefficient of about 5 × 10⁻¹¹ cm³ s⁻¹ at room temperature. For comparison, the authors cite values of about 10⁻⁹ cm³ s⁻¹ for lead halide perovskites and 10⁻¹⁰ cm³ s⁻¹ for gallium arsenide, both direct semiconductors. The team estimates that at room temperature this arrangement cuts the rate of charge loss by about 75% compared with conventional materials that absorb light as strongly. The material’s absorption begins at about 1.4 eV.

To make the measurements, the researchers thermally evaporated bismuth sulfide onto quartz to form films about 150 nm thick, then annealed them at 250 °C for 30 minutes in a nitrogen-filled glove box. They tracked the charges with ultrafast terahertz spectroscopy at temperatures from 4 K to 295 K, or from just above absolute zero to room temperature. Charge loss slowed as the material was cooled, the opposite of the trend seen in most semiconductors. Marina Filip’s group in Oxford’s Department of Physics carried out the electronic structure calculations.

“Our work shows that this trade-off is not inevitable,” said Laura Herz, a professor in Oxford’s Department of Physics and the paper’s corresponding author. The 13 authors are based at the University of Oxford, the University of Padova in Italy, Wuhan University in China and King’s College London.

Why it matters

Losing charges quickly limits how efficiently a solar material can work, according to the University of Oxford, so a strong absorber that holds on to its charges addresses a long-standing constraint. The authors describe the combination of strong absorption and delayed recombination as offering “the unprecedented possibility of overcoming limitations associated with Planck’s law,” with “far-reaching potential for solar energy conversion.” In the paper they call this combination a long-sought goal for solar absorbers used in photovoltaics and solar fuels.

The finding is framed as a design principle rather than a single material. According to the University of Oxford, the authors say other materials with the same arrangement of energy transitions could be found by screening materials databases, or engineered deliberately. That gives materials scientists a specific feature to look for in candidate absorbers.

What to watch

Two questions are left open by the sources. One is whether the slow recombination measured in bismuth sulfide films translates into better performance in an actual solar cell, which has not yet been demonstrated. The other is whether screening materials databases, as the authors propose, turns up other semiconductors with a similar pair of closely spaced transitions.

The caveats

The study is peer reviewed. It was received by Nature on March 4, 2026, accepted on September 29 and published on October 28. It demonstrates a physical mechanism in thin laboratory films, not a working solar cell, and the University of Oxford states that improved solar cell performance has not yet been shown. The 75% figure is the team’s estimate relative to comparably absorbing conventional materials.

The work was funded by the UK Engineering and Physical Sciences Research Council (EPSRC) and the Italian Ministry of University and Research, with computing time on the Frontera system at the Texas Advanced Computing Center. One author was supported by an EPSRC Industrial CASE studentship with Oxford Photovoltaics, a company. The authors declare no competing interests.

Corrections and tips: email the editor using the address on the Contact page.