Quality control of as-cut multicrystalline silicon wafers using photoluminescence imaging for solar cell production Jonas Haunschild n , Markus Glatthaar, Matthias Demant, Jan Nievendick, Markus Motzko, Stefan Rein, Eicke R. Weber Fraunhofer Institute for Solar Energy Systems ISE, Heidenhofstr. 2, 79110 Freiburg, Germany article info Article history: Received 7 April 2010 Received in revised form 6 May 2010 Accepted 6 June 2010 Keywords: Multicrystalline silicon Silicon solar cells Lifetime measurement Photoluminescence imaging Material quality abstract Measuring the lifetime of excess charge carriers gives the opportunity to access the electric quality of the material. However, on as-cut wafers before production this quantity is strongly limited by the surface of the material. On a batch of solar cells we show that the open circuit voltage of the finished cells only scales with the lifetime, measured on as-cut wafers, if the material quality is very low. The difference between moderate and high material quality cannot be resolved. Using photoluminescence imaging the lifetime can be acquired with high spatial resolution. We show that by analyzing crystallization-related features in the images, certain defects can be identified: Such features are defects of crystal growth (e.g. dislocations) and areas of reduced lifetime that form at the edges of the crystallization crucible or near the top or bottom of a brick. Those features can be detected before production and we show their influence on cell parameters. By recognizing and rating these features, a more accurate quality control for wafers can be introduced. & 2010 Elsevier B.V. All rights reserved. 1. Introduction Photovoltaic industries are growing rapidly worldwide. Demands on systems for quality control are growing as well, especially for systems that are already applicable on as-cut wafers before solar cell production. A meaningful rating of the material quality is very important for cell manufacturers, because low solar cell efficiencies can be directly attributed to low material quality and process- related problems can be excluded. Some manufacturers of wafer inspection systems already integrate different tools for measuring the effective lifetime of excess charge carriers. However, we will show that it is of little benefit to use this quantity, as it does not necessarily scale with later cell performance. Only very low material quality can be identified. Photoluminescence (PL) imaging [1] has been introduced as a fast tool to image the lifetime of a wafer. In recent years, it has been continuously improved to give access to different properties (e.g. diffusion length [2] or iron contamination [3]) and to be applicable in different steps of the production line [4,5]. In this publication we show in a first step experimentally and theoretically that lifetime measurements, yielding a global value, cannot be taken as measure for material quality (except in the case of very poor material). In a second step, we show that efficiency limiting defects, which can occur at crystallization, can be seen in the PL images. Via correlation with the open circuit voltage of the finished cells, we prove how these features can be used for quality control on as-cut wafers. 2. Setup and experiment Trupke et al. [1] described the experimental setup needed to perform photoluminescence imaging. In the system developed at Fraunhofer ISE, the sample is positioned on a chuck, which is temperature-stabilized to 25 1C. The whole wafer area is homo- genously illuminated by a laser radiating at 790 nm, with an intensity equivalent to a total charge carrier generation of up to two suns. From luminescence radiation with its peak at 1150 nm, only the short-wavelength part is detectable with silicon charge coupled device (Si-CCD) cameras and therefore – depending on the camera’s quantum efficiency – rather long exposure times are necessary. To achieve data acquisition times o1 s, we use a cooled Si-CCD camera and a lens that have been optimized for infrared light beyond 1000 nm. Please note that exposure times are strongly dependent on experimental conditions and used hardware. The camera is mounted above the sample and can be moved closer to it to image a section with higher resolution. Reflected laser light is suppressed by a stack of long pass filters in front of the camera lens. QSSPC [6] measurements are performed on a Sinton Consulting WCT-100 tool and base resistance is measured on a Kitec PV-R system. Contents lists available at ScienceDirect journal homepage: www.elsevier.com/locate/solmat Solar Energy Materials & Solar Cells 0927-0248/$ - see front matter & 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.solmat.2010.06.003 n Corresponding author. E-mail address: Jonas.Haunschild@ise.fraunhofer.de (J. Haunschild). Please cite this article as: J. Haunschild, et al., Sol. Energy Mater. Sol. Cells (2010), doi:10.1016/j.solmat.2010.06.003 Solar Energy Materials & Solar Cells ] (]]]]) ]]]–]]]