The Si substrates that are often used for the fabrication of solar cells to reduce cost give rise to defect issues that must be addressed. Hydrogen is commonly introduced into silicon solar cells to reduce the deleterious effects of defects and to increase cell efficiency [1]. A process that is used by industry to introduce hydrogen is by the post-deposition annealing of a hydrogen-rich SiNx layer that is used as an antireflection coating [2]. A number of questions about this hydrogen introduction process and hydrogen’s subsequent interactions with defects have proved difficult to address because of the low concentration of hydrogen that is introduced into the Si bulk.Fundamental studies of hydrogen-containing defects in silicon provide a foundation for addressing issues of interest to the Si solar-cell community. Strategies have been developed by which hydrogen in silicon can be detected by IR spectroscopy with high sensitivity [3,4]. The introduction of hydrogen into Si by the post-deposition annealing of a SiNx coating has been investigated to reveal hydrogen’s concentration, diffusivity, and reactions with defects. The effect of processing variations on the concentration of hydrogen that is introduced into the Si bulk has also been studied. The contributions of F. Jiang, S. Kleekajai, V. Yelundur, A. Rohatgi, L. Carnel, J. Kalejs, and G. Hahn to our studies are gratefully acknowledged. This work has been supported by the Silicon Solar Research Center SiSoC Members through NCSU Subaward No. 2008-0519-02 and NSF Grant No. DMR 0802278.[1] J. I. Hanoka, C. H. Seager, D. J. Sharp, and J. K. G. Panitz, Appl. Phys. Lett. 42,618 (1983).[2] F. Duerinckx and J. Szlufcik, Sol. Energy Mater. Sol. Cells 72, 231 (2002).[3] F. Jiang et al., Appl. Phys. Lett. 83, 931 (2003).[4] S. Kleekajai et al., J. Appl. Phys. 100, 093517 (2006).

Average dislocation density and spatial distribution of dislocations are routinely used as a measure of crystal quality of single- and multicrystalline Si (mc-Si) wafers. A variety of techniques have been developed to generate dislocation maps, including X-ray imaging, Cu decoration, and chemical delineation. The most common method is to defect etch the wafer with a suitable chemical etchant and then count the etch pits using an optical microscope. Commercial camera systems, with image analysis software, are available as microscope attachments that can count etch pits within the field of view and combine that information to produce maps of dislocation distribution over a wafer. An improved technique uses light scattered by etch pits to statistically count dislocations. The wafer is illuminated by a laser beam and the total scattered light, which is proportional to the number of etch pits in the illuminated region, is measured. An instrument based on this technique takes 30–60 minutes to map a 6-in x 6-in wafer.This paper describes a new technique that uses scattering from a defect-etched wafer to map dislocation distribution of the entire wafer in a single image. The measurement is very fast and compatible with large-area wafers. In this technique, the single- or multicrystalline wafer is polished to produce a damage-free polished surface. The wafer is then defect etched using Sopori etch (HF:CH3COOH:HNO3 in a 36:15:1 ratio) for 30 s to produce etch pits at dislocation sites. The shape of the etch pit depends on the direction of dislocation at the surface and does not depend on the orientation of the wafer or grain (in mc-Si). The wafer is then placed in a reflectometer where a set of lights, symmetrically placed around the wafer, illuminate it at an oblique incidence. The light scattered normal to the wafer is collected by a camera and imaged. The image corresponds to the local reflectance of the defect-etched wafer. Because local scattering is proportional to the density of etch pits, the camera image is proportional to the local variation in the dislocation density of the wafer. The system is calibrated by using a reference sample to convert the reflectance map into a dislocation map. This technique allows a fast (< 1 s) map