Some Challenges in Making Accurate and Reproducible Measurements of Minority Carrier Lifetime in High-Quality Si Wafers Bhushan Sopori, Srinivas Devayajanam, Prkash Basnyat, Helio Moutinho, Bill Nemeth, Vincenzo LaSalvia, and Steve Johnston, Jeff Binns2, Jesse Appee National Renewable Energy Laboratory, Golden, CO, USA 2SunEdison, St. Peters, MO, USA, 3SunEdison, Portland, OR, USA Abstract - Measurement of the minority carrier lifetime (t) of high-quality wafers (having bulk minority carrier lifetime, tb > few milliseconds) requires surface passivation with very low surface recombination velocity, typically < lcm/s. Furthermore, for mapping large (e.g., 156 x156 mm) wafers, the passivation must also be stable and uniform over the entire wafer surfaces. These are very demanding requirements and it is a common experience that they are very diicult to achieve. Yet, they are necessary for performing defect analyses of the current N-type wafers. To understand the problems associated with these measurements, we have studied effect of wafer preparation (cleaning procedures, handling) and the passivation characteristics (stability, sensitivity to light, thickness of the passivation medium required for stable passivation) for many co � monly used passivation media-iodine-ethanol (IE), qumhydrone-methanol (QHM), aluminum oxide (A 1203), amorphous-silicon (a-Si), and silicon dioxide (Si02). Here, we will discuss main factors that inluence the accuracy and repeatability of lifetime measurements. Index Terms -charge carrier lifetime, defects, oxidation, passivation, silicon. I. INTRODUCTION Silicon photovoltaic industry has developed a strong impetus towards using high-quality, mono-crystalline wafers that can yield solar cell eficiencies > 20%. Higher cell eficiencies offer potential for lower PV energy cost. Concomitantly, there is much interest in N-type CZ wafers that are currently available with minority carrier lifetime (t) in 1 to 5ms range. There is also interest in FZ wafers, which have t even > IOms, to be used for process development (because they do not have oxygen-related effects). In many cases, the high lifetime wafers have spatial non-uniformities and/or develop such non-uniformities during device processing (e.g. due to swirl defects). Hence, there is a need to have well established procedures that allow reliable measurement of lifetime and to perform high resolution lifetime mapping on high-quality, large area wafers. These methods are needed to evaluate the wafer quality and analyzing the defects. Measurement and mapping of wafers with long lifetimes poses many challenges. It requires high quality passivation with very low SRV « lcmls). Furthermore, passivation must be uniform over the entire wafer. Because it takes several hours to do a high resolution t-mapping (e.g. by Semilab tool), passivation must be stable. 978-1·4799-4398-2/14/$31.00 ©2014 IEEE A selection of wafer cleaning procedure and appropriate wafer preparation and passivation procedures are critical in making accurate and reliable measurements. This paper describes our investigations into a suitable cleaning procedure and evaluation of commonly used passivation methods such as iodine-ethanol (I-E), quinhydrone-methanol (QHM), Si0 2 , A1 2 03, and a-Si. All these methods (including Si3N4) have been used by many researchers in the past [1, 2, 3, 4, and 5]. However, to our knowledge, their uniformity, stability, compatibility with the wafer resistivity, and the ways different techniques can be combined to provide rapid mapping of lifetime, have not been discussed in detail. For example oxide can be used as a passivation layer, but is not preferred because of the concen that oxidation can change the defect structure of the wafer. Likewise, I-E gives excellent passivation but the passivation is very short lived. We have not been able to ind published work that characterizes the uniformity and quality of any of these passivation media over the entire wafer surface. An attempt has been made to ill this gap and also report best minority carrier lifetime values ever reported using the above mentioned passivation mechanisms. II. EXPERIMENTAL PROCEDURES The objective of this paper is two-fold: (i) To evaluate various passivation approaches suitable for measuring high lifetime using a single spot (area � 1 "square) measurement, and (ii) To investigate issues associated with making large area (entire wafer) lifetime mapping (e.g., to identiy non uniformities typically observed in mapping/material). We have carried out experiments on very long lifetime, N-type CZ and FZ wafers in the resistivity ranges of: 1 to 4 Q-cm, 8 - 12 Q-cm, 30-40Q-cm, 50Q-cm and 150-250 Q-cm. Wafer cleaning was found to be the most crucial step in obtaining high-quality, uniform passivation. Lifetimes were measured by two measurement systems-Sinton tool for single spot measurements, and micro-PCD mapping using Semilab tool. Passivation for large-area, single-spot measurements was done by I-E and QH, using 0.1 and 0.01 molar solutions respectively. Passivation for lifetime mapping was done by Si0 2 , AbO}, and a-Si. Thin Si0 2 , approximately 75 A, was grown in an optical unace, is very stable and typically yields 0649