Synthesis, Characterization and Electronic Transport Properties of Thin Film Iron Pyrite for Photovoltaic Applications

Abstract

The pyrite form of FeS2 has long been recognized as an earth-abundant and non-toxic material with exceptional properties as a solar absorber for inexpensive photovoltaic devices. However, a significant research effort from the mid 1980’s achieved power conversion efficiencies of only less than 3 %. The reasons for such low efficiencies have not been fully elucidated yet, primarily because the electronic transport and doping mechanisms of pyrite are poorly understood. One classic example is well-known puzzle remaining in pyrite, where bulk single crystals are almost exclusively n-type based on Hall effect measurements, whereas polycrystalline thin films are typically deduced to be p-type, mostly from thermopower measurements. The fundamental reason(s) for this are not understood, and identifying the unintentional dopants in FeS2 remains an outstanding challenge. In this work we address, using ex situ sulfidation synthesis, this long-standing problem of understanding conduction mechanisms and doping in FeS2 films. This is done by systematically exploring the effects of film synthesis conditions on microstructure, surface morphology, chemical stoichiometry, electronic transport mechanisms, charge carrier mobility and charge density. More than a hundred of FeS2 thin films and synthetic crystals were probed in this study. In addition to conventional diffusive transport, hopping transport was also frequently observed in FeS2 thin films. This hopping transport was discovered to be caused by nanoscale inhomogeneity (e.g. nanoscale Fe or FeS clusters), which has been overlooked by the pyrite community until now. This hopping transport may explain the poor performance of some FeS2-based solar cells, since the carrier mobility and lifetime are significantly reduced in hopping. More importantly, accompanying the crossover from diffusive to hopping transport, we find significant suppression, and sign inversion from electron-like to hole-like, of Hall and themopower signals in FeS2 thin films. The results indicate that thin films with diffusive transport show n-type conduction, just like single crystals, which implies that the major n-type dopants may be the same for both FeS2 thin films and single crystals. As the transport crosses over to hopping, both Hall and thermopower measurements indicate sign inversions, which are not caused by real p-type doping, but are rather an artifact of hopping conduction. These findings provide the first potential resolution for the “doping puzzle” in FeS2, and emphasize that understanding the electronic transport mechanisms is mandatory for interpreting the sign of Hall and thermopower coefficients in FeS2. In the last part of this work, some preliminary results for identifying the unintentional dopant(s) in FeS2 are presented. The results suggest the major n-type dopants in FeS2 are unlikely to be metal impurities or oxygen. S vacancies are a genuine possibility however, although further study is still required to settle this issue. These findings answer several critical questions for understanding the electronic transport and doping mechanisms in pyrite FeS2 thin films. They also have important implications for FeS2 solar cell development, emphasizing the need for (a) nanoscale chemical homogeneity, (b) caution in interpreting carrier types and densities, and (c) doping control in pyrite FeS2 films.