Two-Dimensional and Multi-Experimental Modeling of Polycrystalline Cu(In,Ga)Se 2 Solar Cells
- PDF / 122,022 Bytes
- 6 Pages / 612 x 792 pts (letter) Page_size
- 73 Downloads / 160 Views
F10.4.1
Two-Dimensional and Multi-Experimental Modeling of Polycrystalline Cu(In,Ga)Se2 Solar Cells W. K. Metzger, 1 M. Gloeckler, 2 and R. K. Ahrenkiel 1 1 National Renewable Energy Laboratory, Golden, CO 80401 2 Colorado State University, Fort Collins, CO 80523 ABSTRACT We use two-dimensional simulations to explore if charged columnar grain boundaries (GBs) in Cu(In,Ga)Se2 (CIGS) solar cells increase device performance. Although the simulations confirm that charged GBs can increase photocurrent by forming minority-carrier collection channels, this generally occurs at the expense of overall efficiency. Furthermore, improvements in photocurrent require significant GB minority collection beyond a diffusion length of the space-charge region. This collection from deep within the base can be detected by quantum efficiency (QE) spectra, electron-beam-induced current (EBIC) experiments, near-field scanning optical microscopy (NSOM), and fast photoluminescence decay. Simulations of all these experiments indicate that GB charge sufficient to significantly increase photocurrent collection is generally inconsistent with actual observations. INTRODUCTION Recent scanning Kelvin probe microscopy measurements (SKPM) indicate that CIGS GBs are positively charged and may consequently act as minority-carrier collection channels that enhance device performance [1,2]. This is intuitively plausible, but has not been tested computationally. Without GBs, one-dimensional models can describe high-efficiency (>17%) CIGS solar cells using reasonable material parameters because 90% of electron-hole pairs are generated within 0.5 µm of the CIGS/CdS interface, which is well within a diffusion length of the space-charge region. For GBs to significantly enhance the photocurrent, they must collect the relatively few carriers generated beyond this distance and/or adjust the hole concentration along GBs, thereby reducing recombination. In this article, we analyze how neutral and positively charged GBs affect CIGS solar cells and QE, EBIC, NSOM, and time-resolved photoluminescence (TRPL) experiments. Conversely, we examine if these experiments can determine if GBs significantly enhance photocurrent. SIMULATIONS Figure 1 illustrates the device layers. The CdS and ZnO are n-type with 1017 and 1018 carriers/cm3, respectively. The CIGS has a bandgap of 1.15 eV and contains 2.02x1018 cm-3 shallow acceptors, 1.95x1018 cm-3 shallow donors, and a deep acceptor that limits lifetimes to several nanoseconds over a wide range of injection conditions, as observed in experiment [3]. The net concentration is 3x1016 holes/cm3. The bottom contact forms a Schottky barrier, and the CIGS electron and hole mobilities are 100 and 25 cm2/Vs, respectively. Other basic material parameters are presented and discussed in Refs. [4] and [5]. The GB is 20 nm wide and contains the same states described for bulk CIGS. In addition, a midgap recombination center is placed along the GB to generate different values of GB surface recombination, So. GB potential
F10.4.2
Diffusion Len
Data Loading...
