Microanalysis of post-deposition annealing of Cu(In,Ga)Se 2 solar cells J¨ orn Timo W¨ atjen n , Uwe Zimmermann, Marika Edoff nn Department of Engineering Sciences, Division of Solid State Electronics, Uppsala University, Box 534, SE-751 21 Uppsala, Sweden article info Article history: Received 14 September 2011 Received in revised form 25 June 2012 Accepted 20 July 2012 Available online 11 August 2012 Keywords: CIGS Three-stage process Gradients TEM Post-deposition annealing Microscopic JV-characterization abstract The inﬂuence of selenium background pressure during post-deposition annealing of Cu(In,Ga)Se 2 (CIGS) is investigated. Solar cells made from samples post-annealed with selenium showed the same solar cell parameters as references without any annealing treatment. Dark JV measurements of microscopic devices with sizes of 10 mm  10 mm from the sample annealed with selenium showed good agreement with the corresponding macroscopic solar cells. Samples annealed without selenium showed degrada- tion in terms of open circuit voltage and ﬁll factor. Electron beam induced current (EBIC) imaging for these degraded solar cells revealed patches of reduced current. Microscopic JV measurements showed that the deterioration is not limited to these patches. Cross-sectional transmission electron microscopy analysis showed phase decomposition of the CIGS absorber in areas of the patches toward the back contact. We conclude that in addition to the local phase decomposition of the CIGS leading to patches in the EBIC image the anneal in vacuum without selenium background pressure also leads to other modiﬁcations of the CIGS layer inﬂuencing the interface region on a macroscopic scale. & 2012 Elsevier B.V. All rights reserved. 1. Introduction The compound semiconductor Cu(In,Ga)Se 2 often features gra- dients in the composition ratio x ¼ [Ga]/([Ga] þ [In]). Earlier simula- tions and experiments have concluded that Ga gradients can be engineered to reduce the recombination by a back surface ﬁeld repelling electrons from the back contact thus lowering the satura- tion current density and resulting in increased open circuit voltage [1,2]. The formation of such a gradient strongly depends on the deposition method used and is attributed to Cu diffusion in combination with different mobilities for the diffusion of In and Ga ions, respectively [3–5]. These gradients are intrinsic to the often employed three-stage or multi-stage process so far delivering the highest solar cell efﬁciencies reaching up to 20.3% [6–8]. However, to our knowledge no studies have been conducted on the effect of post-deposition annealing on these gradients which may become important for deposition methods involving increased substrate temperatures and rapid thermal annealing processes [9,10]. Furthermore, local ﬂuctuations in the solar cell material proper- ties are reported to be a limitation in reaching high efﬁciency solar cells [11,12]. Some techniques measuring such ﬂuctuations are cathodoluminescence, electron beam induced current and electron beam induced voltage [13]. However, these do not allow direct assessment of photovoltaic device parameters. An approach using photolithography to produce and characterize microscopic solar cells exists but does not offer the possibility of targeting speciﬁc regions of interest on fully processed solar cells [14]. Here, we present a method for JV characterization on the microscopic scale capable of targeting speciﬁc areas of interest on fully processed solar cells. We apply this method to study the inﬂuence of post-deposition annealing in vacuum with and with- out a selenium background pressure at high temperature on samples grown by a three-stage process in terms of device performance and composition from a microscopic as well as macroscopic perspective. 2. Methods 2.1. Device fabrication The devices were deposited on substrates of 5 cm  5 cm made from 1 mm thick sheets of soda lime glass (SLG). They were coated with a 300 nm DC sputtered Mo back contact layer according to our baseline procedure [15]. The Cu(In,Ga)Se 2 (CIGS) absorber layers were grown by a three-stage process using resistive co-evaporation of elemental Cu, In and Ga from fast acting open boat sources and a mass spectrometer to control the metal evaporation rates. A more detailed description is given in Schleussner et al. [18]. The selenium crucible source was kept at a constant temperature ensuring selenium excess during the deposition process. The substrate Contents lists available at SciVerse ScienceDirect journal homepage: www.elsevier.com/locate/solmat Solar Energy Materials & Solar Cells 0927-0248/$ - see front matter & 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.solmat.2012.07.026 n Corresponding author. nn Principal corresponding author. E-mail addresses: timo.watjen@angstrom.uu.se (J. Timo W ¨ atjen), uwe.zimmermann@angstrom.uu.se (U. Zimmermann), marika.edoff@angstrom.uu.se (M. Edoff). Solar Energy Materials & Solar Cells 107 (2012) 396–402