Correlative Characterization of Defect Clusters in GaAs Solar Cells by High-Resolution TEM and Spatially Resolved Optical Techniques

The maximum efficiency of semiconducting devices, such as a solar cell, depends in large part on the presence and character of structural defects in the active material. It is generally known that defects degrade the device performance but the effect of an individual defect on a real device, how that effect depends on the operation of the device, and how that effect correlates to the atomic-scale structure of the defect are not understood. These questions are being addressed by investigating the impact of opticallyand electrically-active defect clusters in GaAs solar cells using high-resolution TEM to determine and analyze their atomic-scale defect structure. This study was done on a GaAs solar cell comprised of many independent sub-cells. The structure of the cell, from bottom to top, was as follows: a p-type GaAs substrate, 80-nm p-type GaAs buffer layer, 50-nm p-type GaInP back-surface confinement layer, 3-μm p-type GaAs absorber layer, 40-nm n-type GaAs emitter layer, 50-nm n-type GaInP window layer, and an 80-nm n-type GaAs contact layer. An electroluminescence (EL) imaging system was used to locate isolated defects that were photoelectrically active, which appeared as dark spots in the EL image. Their location was precisely measured relative to recognizable features nearby. A Nova 200 NanoLab dualbeam system, operating the ion beam (FIB) at 30 keV, was used to prepare TEM cross-section samples. After thinning to electron transparency, the FIB was further used at 5 keV to remove remaining amorphized material. A Philips CM200-FEG high-resolution TEM, operated at 200 keV, was used for initial imaging, and atomic-resolution images were taken with an aberration-corrected JEOL ARM200F STEM at 200 keV.