Chalcogenide perovskites have much to recommend them for photovoltaics (PV). They absorb light strongly, with direct band gap tunable (at least) over the range 1.4 – 1.9 eV. They have limited polymorphism and are stable in air, water, and at high temperature. They are made of Earth-abundant, and (mostly) non-toxic elements, and have isotropic properties. However, there are substantial challenges facing the development of chalcogenide perovskite PV. Synthesis requires aggressive conditions - oxygen-free sulfurization or selenization at high temperature – that severely constrain thin-film growth. The few published reports of transport properties describe n-type material with high electron concentration, undesirable for thin-film PV. Photoluminescence (PL) has been reported, but the quantum yield is often low, and sample-to-sample variability is high. The defects that limit performance are not yet understood, and even less is known about interface and heterojunction design. Clearly, it will be a long road to chalcogenide perovskite PV technology.

I will motivate why, despite these challenges, research on chalcogenide perovskites for PV is worthwhile and exciting. I will then describe our own efforts, which center on the processing and properties of thin films. We have achieved a number of synthesis milestones, including growing thin films of BaZrS₃ and BaZr(S,Se)₃ alloys with tunable band gap, in epitaxial and polycrystalline forms. Selenium alloying can produce films with band gap suitable for single- and dual-junction PV, but the vast majority of synthesis procedures reported to-date focus on pure sulfides. I will discuss our finding of rapid alloying by post-growth selenization of sulfide thin films. This recalls the sulfurization-after-selenization process in CIGS manufacturing, and may make alloy studies more widely accessible. I will also present findings on how variations in cation composition affect crystallization kinetics. These results bolster evidence for BaS₃-liquid-assisted crystal growth, and may be useful for lowering the temperature of thin film synthesis.

I will then discuss our ongoing studies of photoluminescence (PL) and electronic transport. We have previously reported long excited-state PL lifetimes for BaZrS₃ and Ba₃Zr₂S₇, and others have reported band-edge PL even from powder samples. However, PL emission is highly variable, sample-to-sample, and many samples have no measurable band-edge emission. To understand this variability, we carry out a quantitative comparison of temperature-dependent PL of BaZrS₃ and a prototypical halide perovskite, CsPbBr₃. The halide has PL yield between 100 and 10,000 times larger than the chalcogenide. By comparing the vibrational properties of the chalcogenide and the halide, we suggest why defect-assisted recombination may be faster in the chalcogenide. On the other hand, the variability between chalcogenide samples suggests a substantial upside, if the recombination-active defect(s) can be identified and diminished. Our temperature-dependent Hall transport studies find that mobility at room temperature is limited by electron-phonon scattering, even in highly-doped samples; this may be related to our previous finding that chalcogenide perovskites have exceptional dielectric polarizability. At cryogenic temperature, the role of ionized defect scattering varies sample-to-sample. We also find that photoconductive responsivity varies tremendously from sample-to-sample, apparently due to variations in processing that affect the concentration of extended defects. All films are n-type as grown, but with tremendous variability in electron concentration. Studies of post-growth annealing support the hypothesis that the predominant intrinsic shallow donors are sulfur vacancies; we use this understanding to vary electron concentration by over a million-fold.

I will end by highlighting exciting next-steps including alternative methods of thin film deposition to make thicker films at lower temperature, studies of device semi-fabricates including detailed investigation of Mo/BaZrS₃ interfaces, and controlling carrier concentration.