Measurement and Modeling of a-Si,Ge:H Solar Cell Performance After High-Intensity Light Soaking

  • PDF / 349,607 Bytes
  • 6 Pages / 414.72 x 648 pts Page_size
  • 118 Downloads / 161 Views

DOWNLOAD

REPORT


MEASUREMENT AND MODELING OF a-Si,Ge:H SOLAR CELL PERFORMANCE AFTER HIGH-INTENSITY LIGHT SOAKING 2 Vasanth1, M. Nakata, S. Wagner and M. Bennett K. 1 Department of Electrical Engineering, Princeton University, Princeton, NJ 08544 2 Solarex Thin Film Division, Newtown, PA 18940

ABSTRACT For a-Si:H solar cells a clear correlation exists between the defect density in the i-layer and cell performance. For cells with a-Si,Ge:H alloy i-layers, on the other hand, the effect of the properties of the i-layer and its interfaces on cell performance is under debate. To address these questions we light-soaked to saturation and measured a-Si,Ge:H films and cells, and numerically modeled the cell performance using film properties as model inputs. We study three cells with the structure Glass/textured SnO 2 /p+ a-Si,C:H/i a-Si,Ge:H/n+a-Si:H/ZnO/Ag. The i-layers are of uniform composition, 150 nm thick and have Tauc gaps of .1.40, 1.46 and 1.50 eV. We modeled the cells with AMPS, a numerical model solving the Poisson equation and the continuity equations for electrons and holes. The defect density in the 1.40 eV film stays constant during light soaking, while the densities in the 1.46 eV and 1.50 eV films rise. The efficiency of the 1.40 eV i-layer cell stays constant, while it drops in the cells with the 1.46 eV and 1.50 eV i-layers. Using the currently accepted input parameters to the model, we obtain cell performances substantially above the experimentally observed values.

INTRODUCTION

The quantitative understanding of p-i-n cell performance is a key to the industrialization of aSi:H based photovoltaics. Such understanding requires a reliable device model that can interpret and predict device performance in terms of the physical parameters of the cell materials and of features specific to the device, for example, interfaces. Because the distribution of dangling-bond defects in a-Si:H responds to the history and condition of device operation, a separate defect model must be folded into the device model to cover this material parameter. Such a comprehensive device model becomes an important tool for continuously increasing the cell efficiency via loss minimization. We report our first application of the device model AMPS to a-Si,Ge:H pin solar cells. This work follows the simulation of a-Si:H cells which we measured at stages during light-soaking, simultaneously with i-layer material 1' 2 . This simulation showed that cells with 700-nm thick ilayers could be modeled well using the defect density separately measured on i-layer material, but that an extra loss mechanism appears during the light-soaking of cells with -300-nm thick i-layers. We tentatively assigned this loss to defects developing at the p/i interface. Many material parameters required for device simulation are not known with sufficient accuracy. Therefore the application of the model goes hand in hand with providing feedback to the model on improved material parameters. We have been taking advantage of our experience with accelerated light-induced defect generation