Correlation between modeled and experimental device performance results for CIGS solar cells

The performance of copper indium gallium diselenide (CIGS) solar cells varies widely among the various laboratories and production facilities that make such devices. In a recent study performed by members of NREL´s Thin Film Partnership Program, CIGS and copper indium gallium sulfoselenide (CIGSS) films made by several facilities were sent to a single laboratory where they were identically processed into small-area solar cells. The films and devices made from these films were extensively characterized by a variety of techniques. The data from this study provide a unique opportunity for investigating how the observed differences in performance between various CIGS and CIGSS solar cells correlate with measurable material properties. In the present study, device modeling was used to determine whether the variations in measured material properties of the absorber layer are consistent with observed variations in device performance. The modeling shows that observed differences in carrier concentration and midgap state density, as measured by drive-level capacitance profiling (DLCP), are of sufficient magnitude to dominate the observed differences in device efficiency. The best correlation between modeled and measured device performance was obtained when the modeling accounted for differences in bandgap, carrier concentration, and midgap defect state density.