Kyle G. Weideman

The enargite phase of Cu3AsS4 (ENG) is an emerging photovoltaic material with a ∼1.4 eV bandgap and is composed of earth abundant elements with favorable defect properties arising from the differing ionic radii of the constituent elements. Unfortunately, ENG-based photovoltaic devices have experimentally been shown to have low power conversion efficiencies, possibly due to defects in the material. In this joint computational and experimental study, we explore the defect properties of ENG and employ synthesis approaches, such as silver alloying, to reduce the density of harmful defects. We show that shallow copper vacancies (VCu) are expected to be the primary defects in ENG and contribute to its p-type character. However, as shown through photoluminescence (PL) measurements of synthesized ENG, a large mid-bandgap PL peak is present at ∼0.87 eV from a band edge, potentially caused by a copper- or sulfur-related defect. To improve the properties of ENG films and mitigate the mid-bandgap PL, we employed an amine-thiol molecular precursor-based synthesis approach and utilized silver alloying of ENG films. While silver alloying did not affect the mid-bandgap PL peak, it increased grain size and lowered film porosity, improving device performance. In conclusion, we found that incorporating silver such that [Ag]/([Ag] + [Cu]) is 0.05 in the film using an amine-thiol based molecular precursor route with As2S3 as the arsenic source resulted in improved photovoltaic device performance with a champion device of efficiency 0.60%, the highest reported efficiency for an Cu3AsS4 (ENG)-based device to date.

The amine–thiol solvent system has been used extensively to synthesize metal chalcogenide thin films and nanoparticles because of its ability to dissolve various metal and chalcogen precursors. While previous studies of this solvent system have focused on understanding the dissolution of metal precursors, here we provide an in-depth investigation of the dissolution of chalcogens, specifically Se and Te. Analytical techniques, including Raman, X-ray absorption, and NMR spectroscopy and high-resolution tandem mass spectrometry, were used to identify pathways for Se and Te dissolution in butylamine–ethanethiol and ethylenediamine–ethanethiol solutions. Se in monoamine–monothiol solutions was found to form ionic polyselenides free of thiol ligands, while in diamine–monothiol solutions, thiol coordination with polyselenides was predominately observed. When the relative concentration of thiol is increased to that of Se, the chain length of polyselenide species was observed to shorten. Analysis of Te dissolution in diamine–thiol solutions also suggested the formation of relatively unstable thiol-coordinated Te ions. This instability of Te ions was found to be reduced by codissolving Te with Se in diamine–thiol solutions. Analysis of the codissolved solutions revealed the presence of atomic interaction between Se and Te through the identification of Se–Te bonds. This new understanding then provided a new route to dissolve otherwise insoluble Te in butylamine–ethanethiol solutions by taking advantage of the Se^2– nucleophile. Finally, the knowledge gained for chalcogen dissolutions in this solvent system allowed for controlled alloying of Se and Te in PbSe_nTe_1–n material and also provided a general knob to alter various metal chalcogenide material syntheses.