Chalcogenides with stoichiometry ABCh₃ (Ch = S, Se) form in competing, complex crystal structures and are challenging to synthesize. These aspects may be obstacles for solar cell applications, but are opportunities for fundamental materials science because the distinct structures have distinct (and possibly useful) properties, and because the kinetic barriers to synthesis may enable long-lived metastable phases. The most widely-studied chalcogenide perovskite BaZrS₃, with band gap 1.9 eV, sits near the edge of phase stability. Alloying to reduce the band gap usually results in destabilizing the corner-sharing perovskite structure in favor of lower-dimensional structures with highly-anisotropic optoelectronic properties. Epitaxial film growth presents opportunities to stabilize structures that are thermodynamically unstable, and to explore the limits of phase stability and properties. I will present how we use methods of molecular beam epitaxy (MBE) to select between competing phases of BaZrSe₃: the perovskite with band gap near 1.5 eV, and a face-sharing hexagonal structure (h-BaZrSe₃) with large infrared birefringence. We can direct film growth towards one phase or the other by varying temperature, selenium potential, and/or the growth substrate. By choosing a substrate with rectangular symmetry in-plane, we demonstrate first-of-a-kind growth of fully-oriented h-BaZrSe₃ thin films with giant birefringence that can be measured directly polarized transmission spectrophotometry.

Chalcogenide perovskites are also distinguished by their vibrational properties. Their polar optical phonon frequencies are low, and the ionic contribution to their dielectric polarizability is high - they are among the most polarizable of all dielectric semiconductors with comparable band gap. I will present measurements of dielectric response and Hall mobility that demonstrate effective screening of charge defects. However, the phonon frequencies are not so low as for the halide perovskites. I will present Raman and photoluminescence spectroscopy measurements that suggest how and why phonon-assisted, non-radiative recombination (i.e., Shockley-Read-Hall) is much faster in chalcogenide perovskites than in their halide counterparts.

The relative dearth of chalcogenides that are stable in the perovskite structure points to a basic underlying fact, that the perovskite structure is most often found in ionic compounds (e.g., oxides and halides), whereas transition metal chalcogenides have a strong component of covalent bonding. Chalcogenide perovskites can be thought of as covalent semiconductors trapped in ionic structures. I will present explorations of the consequences of this tension, including X-ray spectroscopic measurements of directional bonding, and scanning transmission electron microscopy (STEM) data that suggest substantial band gap fluctuations throughout thin film samples. 

I will end by highlighting exciting directions for future research on complex chalcogenide semiconductors, including synthesis science, defect control, and advanced spectroscopy, that may enable future applications in photovoltaics and infrared photonics.