Limitations in the accuracy of photoconductance-based lifetime measurements N. Sch ¨ uler a,n , S. Anger a , K. Dornich a , J.R. Niklas a , K. Bothe b a Freiberg Instruments GmbH, Am St. Niclas Schacht 13, 09599 Freiberg, Germany b Institut f¨ ur Solarenergieforschung GmbH, Am Ohrberg 1, 31860 Emmertal, Germany article info Article history: Received 23 February 2011 Received in revised form 11 October 2011 Accepted 10 November 2011 Available online 30 November 2011 Keywords: Silicon Lifetime mPCD MDP QSSPC abstract The accuracy of photoconductance-based minority carrier lifetime measurement techniques is studied in detail. Regarding their accuracy and comparability, the quasi steady state photoconductance (QSSPC) as well as the novel steady state microwave detected photoconductivity (MDP) method are compared with the traditional microwave detected photoconductance decay (mPCD). We show that differences in measurement conditions and analyzing procedures lead to deviating results. Calculations based on a generalized rate equation system are used to model the photoconductivity and the resulting effective lifetime for different measurement conditions and defect models. The simulation results are compared to measurements on several mono- and multicrystalline silicon samples with and without surface passivation. Our results clearly show, that for low as well as high injection conditions deviations occur for the measurement and analysis techniques investigated. To allow for a comparison of lifetime data, we recommend to report exact measurement conditions and analysis procedures as well as to perform measurements in a certain injection range. Only if no trapping effects are present and the penetration depth of the applied method is significantly larger than the sample width, accurate and comparable lifetime results can be achieved. & 2011 Elsevier B.V. All rights reserved. 1. Introduction Minority carrier lifetime measurements are widely used for the characterization of the material quality of crystalline silicon wafers and ingots for PV applications. Since the carrier collection and thus the energy conversion efficiency of a solar cell strongly depends on the carrier lifetime, it is a key parameter for performance optimization. Even though the accuracy and comparability of lifetime measurements has already been discussed for several decades [1,2], no conclusive test method has been established so far. One major issue is the coexistence of two fundamentally different classes of carrier lifetime measurement techniques: steady state and dynamic techniques [3]. While for steady state techniques a suitable calibration, relating the measured physical quantity to the excess carrier density, is necessary, transient techniques allow determining the carrier lifetime directly from the time depen- dence of the measured quantity. A related issue is that lifetime measurement techniques either yield differential or absolute lifetime values [4]. In this work different photoconductance-based lifetime mea- surement methods are discussed. The well-established mPCD [5], a transient measurement method with short light pulses of only 200 ns width, is by far the most prominent technique. Here the decay of photoconductivety is analyzed according to t ¼ 1 ðð1Þ=ðDsÞÞððdsÞ=ðdtÞÞ ð1Þ Note that Eq. (1) is only valid, if the photoconductivity is proportional to the excess carrier concentration [6]. Hence it is only valid in a low injection regime (Dn 5N dot ). The most commonly used evaluation algorithm for transient measurements assumes a mono-exponential decay and the car- rier lifetime t transient is determined from the slope of a certain part of the logarithmic transient. The determined lifetime is then commonly assigned to the initial injection at the beginning of the transient. However, the assumption of a mono-exponential decay or equivalently a constant lifetime is only valid for low injection levels, which will be addressed in more detail in section 2.1. Consequently, the results of the related linear regression strongly depend on the analyzed time interval of the transient. This makes a comparison of lifetime results difficult, since for different measuring methods inconsistent procedures are used to select the appropriate transient time interval for lifetime extraction. Contents lists available at SciVerse ScienceDirect journal homepage: www.elsevier.com/locate/solmat Solar Energy Materials & Solar Cells 0927-0248/$ - see front matter & 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.solmat.2011.11.017 n Corresponding author. Tel.: þ49 3731 4195416; fax: þ49 3731 4195414. E-mail address: schueler@freiberginstruments.com (N. Sch ¨ uler). Solar Energy Materials & Solar Cells 98 (2012) 245–252