First-principles study on pyrites and marcasites for photovoltaic application

This thesis deals with first-principles calculations for pyrite and marcasite compounds, with a particular focus on their suitability for photovoltaic applications. Their electronic structure and their optical properties are thoroughly investigated within density-functional theory (DFT) using various exchange-correlation functionals, among them sophisticated hybrid functionals. To account for electronic excitations the many-body perturbation theory in the GW approximation has also been exploited. The investigation of the electronic and optical properties of iron pyrite (FeS2) covers an essential part of this thesis, since iron pyrite is reported to be a promising material for photovoltaic applications due to its large optical absorption, a suitable band gap and large photocurrents. Furthermore, iron pyrite consists of abundant elements, and thus would allow for a large-scale and long-term utilization. However, iron pyrite solar cells exhibit only an open-circuit voltage of merely 200 mV, leading to a small conversion efficiency of 3%, which disqualifies iron pyrite for photovoltaic applications at present. This thesis exposes that the question about the size of the fundamental and optical band gap of iron pyrite, both, theoretically and experimentally, might not be settled yet. Low-intensity contributions in the optical absorption might complicate the determination of the optical band gap, and the GW results show that the fundamental band gap might be much smaller than expected. Therefore, the small fundamental band gap of pristine iron pyrite in the bulk phase might be already responsible for the low open-circuit voltage. Since interfaces and surfaces play an important role for the photovoltaic performance, the electronic structure of the most stable iron pyrite surfaces is also discussed, revealing that surface states of Fe 3d character might act as charge recombination centers. First attempts to passivate these surface states indicate that heavier adatoms are more suitable than light adatoms. The application of the GW approximation on iron pyrite yields an unconventional reduction of the band gap compared to the “plain” DFT results, whereas largely overestimated band gaps are obtained using hybrid functionals. By extending the calculations to other pyrite compounds (RuS2, OsS2, NiP2 and ZnS2) and to the structurally closely related marcasite compounds (FeS2, FeSe2 and FeTe2), it is shown that the interplay of transitions between p and d states and the screening caused by the d states is responsible for this peculiar behavior. Finally, a particular focus is set on FeS2 marcasite, which is reported to coexist with the pyrite phase, but is presumed to degrade the photovoltaic performance. However, the results in this thesis indicate that iron marcasite might be better suited for photovoltaic applications than iron pyrite.