In recent years, the photovoltaic market has grown significantly as module prices have continued to come down. Continued growth of the field requires higher efficiency modules at lower manufacturing costs. In particular, higher efficiencies reduce the area needed for a given power output, thus reducing the downstream balance of systems costs that scale with area such as mounting frames, installation, and soft costs. Cells and modules made from III-V materials have the highest demonstrated efficiencies to date but are not yet at the cost level of other thin film technologies, which has limited their large-scale deployment. There is a need for new materials growth, processing and fabrication techniques to address this major shortcoming of III-V semiconductors.
Chapters 2 and 3 explore growth of InP on non-epitaxial Mo substrates by MOCVD and CSS, respectively. The results from these studies demonstrate that InP optoelectronic quality is maintained even by growth on non-epitaxial metal substrates. Structural characterization by SEM and XRD show stoichiometric InP can be grown in complete thin films on Mo. Photoluminescence measurements show peak energies and widths to be similar to those of reference wafers of similar doping concentrations.
In chapter 4 the TF-VLS growth technique is introduced and cells fabricated from InP produced by this technique are characterized. The TF-VLS method results in lateral grain sizes of >500 µm and exhibits superior optoelectronic quality. First generation devices using a n-TiO2 window layer along with p-type TF-VLS grown InP have reached ~12.1% power conversion efficiency under 1 sun illumination with VOC of 692 mV, JSC of 26.9 mA/cm2, and FF of 65%. The cells are fabricated using all non-epitaxial processing. Optical measurements show the InP in these cells have the potential to support a higher VOC of ~795 mV, which can be achieved by improved device design.
Chapter 5 describes a cost analysis of a manufacturing process using an InP cell as the active layer in a monolithically integrated module. Importantly, TF-VLS growth avoids the hobbles of traditional growth: the epitaxial wafer substrate, low utilization efficiency of expensive metalorganic precursors, and high capital depreciation costs due to low throughput. Production costs are projected to be $0.76/W(DC) for the benchmark case of 12% efficient modules and would decrease to $0.40/W(DC) for the long-term potential case of 24% efficient modules.