UNUSUAL LOCK-IN THERMOGRAPHY SIGNALS: SCHOTTKY-TYPE GRID CONTACTS, PELTIER EFFECTS, AND THERMAL WAVE INTERFERENCE O. Breitenstein, J.P. Rakotoniaina Max Planck Institute of Microstructure Physics, Weinberg 2, D-06120 Halle, Germany ABSTRACT The technique of lock-in thermography (LIT) provides a number of well-established methods to image the lateral homogeneity of electrical parameters of solar cells. Using these methods, linear (ohmic) and nonlinear (diode-like) local leakage currents (edge currents, shunts), local Joule heating, and inhomogeneities of the series resistance and of the minority carrier lifetime can be imaged. However, in some cases, unusual LIT signals appear, which cannot be interpreted in the frame of generally accepted LIT methods. This contribution explains LIT signals connected with Schottky-type grid contacts, Peltier effects, and thermal wave interference, which is important to avoid any misinterpretation of LIT results. INTRODUCTION Lock-in thermography (LIT) is meanwhile standing for a whole group of solar cell and solar material characterization techniques, which are all aiming to image lateral inhomogeneities of the electronic parameters of solar cells or solar materials S1-S3. Until now, a number of assumptions were always implicitely made for interpreting LIT results, which are e.g. the assumption of good ohmic contacts both at the emitter and at the base, and the neglection of Peltier effects at these contacts. LIT experiments performed in the dark (DLIT S23) are usually interpreted based on the energy conservation law, hence the local T modulation amplitude has been interpreted as a measure of the local dissipated power density. However, in some cases, certain signal features are observed in LIT investigations on solar cells, which cannot be interpreted in the frame of the pervious generally accepted models. For example, LIT experiments performed by pulsed illumination under short circuit conditions (Jsc-ILIT) sometimes show bright spots below grid lines, which disappear if the measurement is performed under forward bias. It will be shown that these are due to Schottky-type grid contacts. In other cases, the structure of the ohmic base contact becomes visible in LIT images. This is often due to the Peltier effect, which appears if the current is passing between the base metal and the silicon base material. Finally, in some cases, local minima of the LIT signal may appear, which are not due to local minima of the dissipated heat, but have to be interpreted as interferences between the thermal waves originating from different heat sources in the sample. This contribution explains the physical origin of such unusual LIT signals. THE USUAL INTERPRETATION OF LIT SIGNALS In non-illuminated (dark) lock-in thermography (DLIT) the amplitude signal is a measure of the locally dissipated power density S13. Hence, a homogeneous bright region usually indicates homogeneous carrier injection, and bright lines or spots are indicating shunts or Joule heating. In illuminated LIT (ILIT) performed by pulsed illumination, the whole area of a solar cell appears bright because of the essentially homogeneous carrier thermalization S23. If this measurement is performed under short circuit conditions (Jsc-ILIT), dark regions are indicating non- contacted regions, and even brighter regions are indicating Joule losses due to horizontal current flow S23. Also strong shunts may appear in Jsc-ILIT as bright spots, since the local bias is never exactly zero. However, bright spots, which are not visible in DLIT, are not expected to appear in Jsc-ILIT. If LIT is performed under pulsed illumination, and the bias is also pulsed to maximum power point conditions (mpp-ILIT), all energy losses in operation are visible. Hence here a bright region is caused by carrier thermalization losses or homogeneous injection, and bright spots or lines are caused by shunts or Joule heat. Rs-ILIT means that the illumination is performed continuously and the bias is pulsed between short circuit and mpp (usually close to 0.5 V). In this operation mode, well-contacted regions in a cell are appearing dark, and non-contacted regions and shunts are appearing bright. Dark spots are not expected to appear in Rs-ILIT. SCHOTTKY-TYPE GRID CONTACTS The emitter of silicon solar cells is highly doped so that the metallic (usually silver) grid lines yield a good ohmic contact. However, if the emitter doping concentration should be too low, or if the grid firing conditions should be inappropriate, the contact resistance may be too high, leading to an increased series resistance, which degrades the solar cell parameters. Such poorly contacted regions can easily be imaged by methods like Corescan, CELLO, Rs-ILIT or Jsc-ILIT SS3. Regions of high series resistance appear dark in Jsc-ILIT and bright in Rs-ILIT. However, in some cases bright spots are appearing in Jsc-ILIT, corresponding to dark spots in Rs-ILIT, which are definitely no regions of especially low series resistance. As Fig. 1 shows, these spots are vertically arranged, hence they are lying in positions of grid lines, and they are visible neither in dark lock-in thermography (DLIT) nor in mpp-ILIT. These measurements were performed at a relatively high lock-in frequency of 48 Hz, leading to a good spatial