UC Berkeley

The reliable production of clean, inexpensive energy to meet the growing global demand remains one of the most significant challenges to modern society. As high-technology, high power-consumption machines and devices proliferate and as underserved populations around the globe come "on the grid," energy demand will significantly expand. Photovoltaic (PV) technologies represent one of the most direct paths to harvesting solar energy, converting solar energy directly into electricity with no moving parts, no pollution, and no greenhouse gas emissions. These technologies often utilize materials that may not exist in sufficient abundance to meet large-scale (TW) deployment demands. For instance, the highly successful Cu(In,Ga)Se2 material system has achieved module efficiencies up to 15% [1] but contains In and Se, which are relatively scarce and expensive. Shafarman reports that the availability of In would limit total installed capacity to 0.1TW [2], far below the 10 - 20 TW required to meet global energy demand. Moreover, CIGSe devices typically involve processing with highly toxic H2Se gas [2].

Therefore, significant research has recently been devoted to discovering and developing PV devices that use non-toxic, inexpensive, earth-abundant elements. Based on the success of the Cu(In,Ga)Se2 material system, the development of the electronically analogous Cu2ZnSnS4 material system represents a natural next step. CZTS replaces the scarce Group III (In/Ga) elements with more abundant Group II (Zn) and Group IV (Sn) elements. Sulfur also replaces Selenium, allowing less-toxic processing conditions. With a direct band gap of 1.5eV and high optical absorption > 104cm-1, CZTS represents an ideal absorber layer for conventional single-junction solar cells [3]. Further, it is hoped that CZTS, which shares similar crystal structure and electronic attributes with CIGSe, may also share a similar resilience in its opto-electronic properties to grain boundaries and to stoichiometric deviations that has afforded CIGSe devices much success [2].

Devices with moderately high efficiencies of 12.6% have been fabricated using Se-alloyed CZTSeS absorber layers [4]. In general, however, the fabrication of high-quality CZTSeS absorber layers has been extremely challenging, due to the narrow phase formation region for CZTSeS [3, 5], elemental and compound volatility [3, 6-8], and phase decomposition at the surface and back contacts [9, 10]. Despite the moderately high device efficiencies achieved, all aspects of CZTS device design merit continued investigation. Scalable and reliable fabrication methods that are capable of producing phase-pure, void-free CZTSeS absorber layers have yet to be demonstrated. Such advancements will require better understanding of the crystal formation processes and defect behavior of CZTSeS. The work in this dissertation aims to improve understanding of the growth processes of CZTS thin films and to provide a pathway to fabrication of high-quality absorber layers for use in photovoltaic devices. Reliable growth of high-quality CZTS thin films will enable investigation of the fundamental properties of CZTS, as well as device and interface behavior, necessary to improve device efficiency.

This dissertation investigated the growth behavior of CIGSe and CZTS thin films developed for use as photovoltaic absorber layers. The CIGSe thin films were fabricated by pulsed laser deposition, and the CZTS thin films were fabricated by pulsed laser deposition and co-electrodeposition methods. The effects of deposition and annealing parameters on the film properties were systematically investigated, critical growth parameters identified, and optimized fabrication conditions recommended. In pulsed laser deposited CIGSe thin films, annealing in sulfur background significantly improved the electronic quality of the films, reducing carrier concentrations and increasing Hall mobilities. In pulsed laser deposition of CZTS thin films, proper adjustment of the laser fluence and sputtering target composition enabled films with the desired stoichiometries to be deposited at room temperature. A two-step temperature profile involving a long dwell at low temperature and short dwell at the crystal formation temperature yielded the most-stable films with the optimum structural properties. Proposed growth mechanisms and optimized fabrication processes are presented.

To study the deposition mechanisms active in the co-electrodeposition of S-containing CZTS precursors, cyclic voltammagrams were compared with varying potential depositions. Reduction mechanisms and peak assignments are proposed. Films deposited using a fabrication process from literature demonstrated rough, powdery morphologies with Zn-poor, Sn-rich compositions. The morphology was attributed to diffusion-limited growth mechanisms with significant hydrogen co-deposition. However, removing tartaric acid from the deposition bath dramatically reduced the roughness and increased the uniformity of the films. To further improve film composition and morphology, bath component concentrations were systematically modified to yield Cu-poor, Zn-rich films. Optimum bath compositions were successfully identified, although a delayed onset of damaging hydrogen evolution reaction was observed during growth. Stopping the deposition before its onset limited the damage caused by this reaction, but films were limited to small thicknesses. Stirring, increasing bath concentration, and galvanostatic control were investigated as possible methods to increase the film thickness before the onset of this reaction. Optimized co-electrodeposition processes were identified, with linearly-swept potential and pulsed stirring methods representing the most promising processes. Similar to the pulsed laser deposited CZTS films, a two-step temperature profile yielded the most-stable films with the optimum structural properties.

Finally, co-electrodeposition of metallic CZT precursors was investigated. The as-deposited films were highly compact, uniform and free of damage, and the metal ratios in the film could be directly controlled by modifying the metal sulfate concentration ratios in the bath. Varying bath concentrations were investigated, in order to determine the effect on film thickness and composition. Optimized co-electrodeposition processes were identified. After sulfurization, films demonstrated ideal compositions and structures but also significant phase segregation, with the film laterally divided into regions of large-grained CZTS and small-grained ZnS phases. Annealing in Argon at low temperature prior to sulfurization significantly improved the film homogeneity, although a non-negligible density of pinholes remained. It is unclear if the lack of S-content in the precursors, the loss of sulfur overpressure, or both, contributed to the formation of voids and pinholes.

The work in this dissertation provides a comprehensive foundation on which to further improve the deposition of co-electrodeposited S-containing and metallic precursors. The growth of high-quality CZTS precursors was demonstrated and a pathway to additional optimization recommended. With minor adjustments in the electrodeposition process, and improved monitoring and control of the sulfur overpressure, it is believed that very high-quality CZTS films can be reliably fabricated using a two-step co-electrodeposition and sulfurization process. Notably, the co-electrodeposition method represents a simple fabrication method that utilizes low-toxicity components, complies with existing plating technologies, and provides high throughput. The demonstration of such a process would represent a significant step forward in the development of the earth-abundant, low-toxicity CZTS material system.