Optoelectronic characterization of co-evaporated and low-cost Cu(In,Ga)Se 2 solar cells, a comparison Theresa Magorian Friedlmeier ⁎, Paola Mantilla Pérez 1 , Ines Klugius, Philip Jackson, Oliver Kiowski, Erik Ahlswede, Michael Powalla Zentrum für Sonnenenergie- und Wasserstoff-Forschung Baden-Württemberg (ZSW), Industriestrasse 6, 70565 Stuttgart, Germany abstract article info Available online 16 December 2012 Keywords: CIGS Optoelectronic characterization High efficiency Low cost Recombination In this investigation we selected four different Cu(In,Ga)Se 2 -based solar cells produced using different methods and with efficiencies ranging from less than 5% for the low-cost types to 20% for the highest efficiency process. These devices were intensively characterized using the optoelectronic methods of temperature-dependent current–voltage characteristics, capacitance–voltage measurements, and bias-dependent external quantum effi- ciency. By combining these methods it is possible to study the dominant recombination mechanisms in the de- vices and relate them to composition and morphology effects known through other characterization series. It was found that the highest efficiency cell has the least indication of interface recombination or tunneling effects; the temperature-dependent behavior is best described by recombination in the space-charge region. In contrast, the low-cost devices reveal high diode factors and a temperature dependence well described by tunneling- enhanced recombination at the interface. The consequences for the device behavior and recommendations for improvement will be presented. © 2012 Elsevier B.V. All rights reserved. 1. Introduction Cu(In,Ga)Se 2 (CIGS) can easily be formed with the chalcopyrite phase and the composition linked to highest efficiencies. However, the material quality can differ greatly, depending on the applied growth method. The quality of the film and most particularly the quality of the surface/interface largely determine the efficiency per- formance which can be achieved with the photovoltaic device. In order to investigate the underlying optoelectronic processes affecting the cell performance of “good” and “poor” devices, we measured temperature-dependent current–voltage (IV–T) curves, capacitance– voltage (CV), and bias-dependent external quantum efficiency (EQE-V) on a series of devices produced with CIGS from different pro- cessing routes in our lab. 2. Experimental methods 2.1. Sample preparation Of the four sample types investigated in this work, two of the devices were produced using co-evaporated CIGS with a “standard” efficiency of ca 16% [Std, Ref. [1]] and a “top” efficiency of ca 20% [Top, Ref. [2]]. These devices had a total area of 0.50 cm 2 and employ a Ni/Al contact grid. The other two devices with efficiencies b 5% were produced by a low-cost method of printing from a liquid phase, using either metal salt solutions [MS, similar to Ref. [3]] or nanoparticle-based inks [NP, similar to Ref. [4]], and subsequent selenization at ~500 °C in a graphite susceptor with elemental Se in a rapid thermal annealing system. The “MS” sam- ple has a layer of carbon between the Mo and the CIGS (see Fig. 4). The low-cost devices are scribed to a total area of 0.25 cm 2 and have no grid. All samples were prepared with average Cu/(In+Ga) ratios between 0.80 and 0.95 and Ga/(Ga+In) ratios between 0.25 and 0.35. Exact values and gradients were not determined for the specific samples. However, the co-evaporated samples are known to have increased Ga content towards both interfaces and the nanoparticle sample may con- tain some sulfur. All CIGS absorber films were grown on similar Mo-coated soda-lime glass (3-mm glass for the coevaporated samples and 1-mm glass for the low-cost samples) and completed with similar chemical-bath-deposited CdS, sputtered i-ZnO and sputtered ZnO:Al processes. As described in Ref. [2], the processes leading to top efficien- cies involve more strict control of materials purity and optimization of the layer thicknesses. Furthermore, the low-cost samples were etched with KCN prior to buffer deposition and only the 20% device had an anti-reflection coating. 2.2. Temperature-dependent measurements The IV–T curves were measured under simulated AM 1.5 global solar irradiation (from WACOM, 100 mW/cm 2 ). The samples were Thin Solid Films 535 (2013) 92–96 ⁎ Corresponding author. Tel.: +49 7117870293. E-mail address: theresa.friedlmeier@zsw-bw.de (T. Magorian Friedlmeier). 1 Present address: ICFO—The Institute of Photonic Sciences, Mediterranean Technology Park Av. Carl Friedrich Gauss, num. 3, 08860 Castelldefels, Barcelona, Spain. 0040-6090/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.tsf.2012.11.108 Contents lists available at SciVerse ScienceDirect Thin Solid Films journal homepage: www.elsevier.com/locate/tsf