AVS2012 Session EN+TF-MoA: Chalcogenide Solar Cells I

Given the terawatt scale of future energy needs, the most promising future photovoltaic materials should be Earth abundant with their primary mineral resources distributed across several geographic regions and their supply chains robust to reduce concerns of price volatility. In addition, the process of forming the solar cell should be scalable, low-cost, and not utilize dangerous or toxic materials. The strongest initial candidate appears to be kesterite structures of Cu₂ZnSnS₄ (CZTS) and similar materials. The presentation will review the progress in developing photovoltaics devices based on these materials and our group’s recent experimental and modeling results.

CZTS thin film solar cells have historically been synthesized by evaporating or sputtering metals (Cu, Zn, & Sn) followed by sulfurization or selenization. More recently, two potentially low-cost high-throughput approaches have been demonstrated that form the quaternary or pentenary chalcogenide directly from solution-phase processes. One is based on first synthesizing multinary sulfide nanocrystals and then sintering them to form a dense layer. The other approach utilizes molecular precursors dissolved in hydrazine. Both new approaches reach their highest device efficiencies by incorporating Se to form Cu₂ZnSn(S_x,Se_1-x)₄ devices, and each has yielded substantially higher efficiency devices than the best vacuum deposited absorbers. The hydrazine route has yielded the most efficient CZTS-based devices thus far. The presentation will highlight our recent progress in CZTS-based nanocrystal-ink devices. In particular, we have shown that germanium may be alloyed with CTZS (at least up to Ge/(Sn+Ge) ratios of 0.7) to form Cu₂Zn(Sn,Ge)S₄ nanocrystals that have an increased bandgap. The defect chemistry is serendipitous, and yields devices at with greater than 8% power conversion efficiency. This exciting prospect may be used to create a back surface field and direct carriers in a similar manner to how gallium is used in high efficiency CIGS devices. In addition, we will report recent results from high throughput experiments focused on identifying doping and passivation agents.

Cu₂ZnSnS₄ (CZTS) is an emerging low-cost solar absorber for thin film photovoltaics based on non-toxic, high earth-abundance elements. While ex situ sulfidation of Cu-Zn-Sn precursor films in S vapor is a popular synthesis route for CZTS, much remains to be understood with regard to the sulfidation mechanisms, microstructural control, and structure-property relationships. In this work, DC magnetron co-sputtered Cu-Zn-Sn films, of varying composition, were sealed with 1.0 mg of S in evacuated (10^-6 Torr) quartz ampoules, and then isothermally heat treated at sulfidation temperatures in the range 100 ^oC ≤ T_S ≤ 700 ^oC. The films were then characterized structurally by scanning electron microscopy, Raman spectroscopy, and X-ray diffraction, and electrically via resistivity measurements between 4.2 K and 300 K. We find that the phase purity of the resultant films depends strongly on T_S, with complete conversion of the precursor film to CZTS occurring only at T_S ≥ 550 ^oC. The final phase purity of ex situ sulfidized CZTS films is however remarkably insensitive to modest amounts of excess Zn and Sn in the Cu-Zn-Sn precursor film. Excess Zn or Sn in the precursor is readily ejected during sulfidation, via elemental or binary sulfide evaporation, respectively, while Sn-deficiency can also be corrected by introducing elemental Sn to the sealed quartz ampoule. However, due to the low melting point of Sn, and the absence of appropriate Sn-Zn alloys, the precursor composition does play a significant role in defining the final CZTS film microstructure, and the lateral homogeneity. In addition to detailed discussion of the physics and chemistry underlying the above observations we will also provide data on the transport properties of such films, including observations of hopping conduction.

Pyrite FeS₂ is undergoing a tremendous resurgence of interest as a candidate thin-film solar absorber based on abundant, low-cost, non-toxic elements. Historically, FeS₂-based Schottky solar cells have suffered from low open circuit voltages (~ 100 mV), and thus low efficiency, although the origins of this behavior are not entirely clear. In fact, even the electronic properties of FeS₂ are not well understood, including the conduction mechanisms and doping behavior. Understanding these issues could contribute significantly to improvements in FeS₂-based solar cells, particularly if doping can be understood and controlled. In this work, we present a comprehensive study of the conduction mechanism in FeS₂ thin films synthesized via the ex situ sulfidation of Fe films in a S vapor at sulfidation temperatures in the range 100 ^oC ≤ T_S ≤ 700 ^oC . The resultant films were characterized structurally, using X-ray diffraction, scanning electron microscopy, energy dispersive spectroscopy, and confocal Raman microscopy; electrically, via transport and magnetoresistance measurements between 4.5 and 300 K; and magnetically with high-sensitivity dc magnetometry. At T_S around 500 ° C we observe a crossover in the conduction mechanism from some form of hopping conduction to a more conventional band transport-type mechanism. Through detailed analysis of the hopping parameters, measurement of the Fe spin-state, and simple calculations based on S diffusion in Fe, we demonstrate that intergranular hopping occurs via highly conductive, S-deficient, nanoscopic grain cores separated by nominally stoichiometric FeS₂ shells. We find that the approach towards more conventional band transport as T_S is increased above 500 ^oC is due to an increase of S diffusion into the FeS₂ grains. Moreover, this conduction mechanism crossover is found to be accompanied by a sign reversal of the Hall coefficient, from hole-like (in the hopping regime) to electron-like. In addition to placing hard constraints on the conditions under which useful properties can be obtained from FeS₂ synthesized under diffusion-limited conditions, these results also highlight potential problems with prior conclusions on the dominance of p-type behavior.

The work was supported by the Initiative for Renewable Energy & the Environment, IREE ((RL-0004-11) . Part of this work was carried out in the University of Minnesota Characterization Facility, a member of the NSF-supported Material Research Facilities Network.

Pyrite (FeS₂) is a non-toxic, earth abundant chalcogenide with desirable characteristics for application as a photovoltaic absorber including a modest band gap of 0.95 eV and a large optical absorption coefficient (>10⁵ cm^-1). Although theoretically capable of >20% efficiency, to date pyrite devices have displayed poor performance. One key challenge has been the production of stoichiometric material that is free of impurity phases. Conventional approaches employ thermal sulfurization of iron-based films or precursors. These routes inherently produce contaminate phases (troilite, pyrrhotite, marcasite), that once formed, are difficult to completely remove. Thermodynamics suggests that hematite (α-Fe₂O₃) may be directly converted to pyrite in the presence of sufficiently high sulfur activity. In this work, we demonstrate pyrite synthesis using a H₂S plasma to sulfurize hematite nanorods produced using chemical bath deposition. Conversion to pyrite was achieved by exposure to a 90% Ar-10% H₂S plasma at moderate temperature (350 - 450 °C). The application of plasma dramatically enhances both the rate of conversion and the quality of the resulting material. Composition analysis using both Raman and X-ray photoelectron spectroscopy confirm that the resulting pyrite is free of common impurity phases. The degrees of sulfur incorporation could be precisely controlled by plasma exposure, and the apparent optical band gap could be systematically reduced from 2.2 to 1.0 eV. Electron microscopy images showed the surface maintained a nanostructured architecture following sulfurization, and a 150 nm thick film was sufficient to absorb 99% of incident visible light. In this presentation we discuss the kinetics of this solid state transformation, as well as report on the optoelectronic properties of these materials.