How Polycrystalline Devices Can Outperform Single-Crystal Ones: Thin Film CdTe/CdS Solar Cells


  • We thank L. Kronik (WIS) for many helpful discussions, P. De Wolf (Veeco) for guidance with SCM, S. Richter (TAU) for guidance with CP-AFM, V. Kaydanov (Col. School of Mines) for discussions, I. Bar-Yosef (WIS) for use of the Dimension microscope, USDOE (via NREL) for initial, and the Weizmann Institute's Levin fund (DC) and Feinberg Grad. School (IVF)) for further support. DC holds the Schaefer Chair in Energy Research. Most of the results that are presented here, as well as the basic model, were reported at the Spring 2003 (April 21–25) MRS meeting (S. Francisco, Symp. B; cf. also MRS Bull.2003, 28(7), 521) and at the 38th IUVSTA and ISF Workshop on Electronic Processes and Sensing on the Nano-Scale (Eilat, May 25–29, 2003). Subsequently Persson & Zunger reported electronic structure calculations (Phys. Rev. Lett. 2003, 91, 266 401), which they used to promote an explanation for the superiority of PX CuInSe2 solar cells. Their model – charge separation near, and transport at, GBs – conforms closely to ours.


Grain boundaries (GBs) participate in the photovoltaic energy conversion process in polycrystalline solar cells as efficient photocurrent collectors and transporters, as shown by high- resolution characterization of CdTe GBs in CdTe/CdS cells (see Figure). This suggests that structural defects can be advantageous for device performance, if properly designed, even in devices whose operation is based on physics of ideal, perfect solids.