Electronic Defects and Device Performance in CuGaSe 2 Solar Cells J. Jedediah Rembold 1 , Todd W. Curtis 1 , Jennifer T. Heath 1 , David L. Young 2 Steve W. Johnston 2 , and William N. Shafarman 3 1 Linfield College, McMinnville, OR, 97128 2 National Renewable Energy Laboratory, Golden, CO, 80401 3 Institute of Energy Conversion, University of Delaware, Newark, DE, 19716 ABSTRACT The electronic and materials properties of two series of wide-bandgap solar cells with Cu- poor CuGaSe 2 (CGS) absorbers have been studied, to better understand limitations on the device performance. One series of samples displayed distinct lateral non-uniformities in Cu/Ga ratio, Na content, and thickness. The C-V data on these samples suggested the presence of a distinct band of near-interface defect states. Although the second series of samples appeared uniform and C-V results did not indicate the presence of near-interface defect states, J-V-T results show V oc (0T) <E g in all samples, which typically is thought to indicate that a dominant recombination current at the interface limits V oc . The device performance did not appear to be correlated to defect densities in the bulk CGS film. INTRODUCTION The series of chalcopyrite CuIn 1-x Ga x Se 2 (CIGS) materials have been investigated for photovoltaic applications for some time now, due to the high efficiency of devices based on lower bandgap material with x 0.3, and the tunable bandgap. Other alloys containing S and Al have also been studied. While devices based on lower gap material have been extremely successful, achieving efficiencies of 19.5% [1], the higher gap alloys have lagged behind. Yet, these high bandgap alloys are desirable for several reasons. Photovoltaic devices with higher bandgaps have larger V oc and reduced currents, allowing module operation with reduced power losses related to lower series resistance. Additionally, the vision of an all-chalcopyrite tandem device has long been discussed, and progress is being made towards its implementation [2]. In this study, we investigate two sets of CGS samples prepared in different labs by slightly different techniques, to gain further insight into mechanisms limiting the device performance. EXPERIMENTAL METHODS The samples were prepared on soda-lime glass, which was coated with a dc-sputtered, 1 µm Mo layer. The approximately 2 µm thick CGS layers were deposited by elemental evaporation. Samples were created in two different labs using slightly different techniques, namely a uniform single-step process and a three-step process, yielding sample series A and B, respectively. These deposition processes are described in more detail elsewhere [3,4]. All CGS films were grown Cu-poor. A thin CdS layer was deposited using a chemical bath, followed by rf sputtering of 50 nm of ZnO with a resistivity of 50 Ω-cm. The devices were completed by sputtering a 150 nm ITO layer with sheet resitistance 25 /sq (series A) or a 500 nm ZnO:Al layer with sheet resistance 20 Ω/sq, (series B) and Ni-Al grids. Mater. Res. Soc. Symp. Proc. Vol. 1012 © 2007 Materials Research Society 1012-Y12-27 ,