Time‐resolved investigation of Cu(In,Ga)Se₂ growth and Ga gradient formation during fast selenisation of metallic precursors

Abstract

Ga segregation at the backside of Cu(In,Ga)Se2 solar cell absorbers is a commonly observed phenomenon for a large variety of sequential fabrication processes. Here, we investigate the correlation between Se incorporation, phase formation and Ga segregation during fast selenisation of Cu–In–Ga precursor films in elemental selenium vapour. Se incorporation and phase formation are analysed by real‐time synchrotron‐based X‐ray diffraction and fluorescence analysis. Correlations between phase formation and depth distributions are gained by interrupting the process at several points and by subsequent ex situ cross‐sectional electron microscopy and Raman spectroscopy. The presented results reveal that the main share of Se incorporation takes place within a few seconds during formation of In–Se at the top part of the film, accompanied by outdiffusion of In out of a ternary Cu–In–Ga phase. Surprisingly, CuInSe2 starts to form at the surface on top of the In–Se layer, leading to an intermediate double graded Cu depth distribution. The remaining Ga‐rich metal phase at the back is finally selenised by indiffusion of Se. On the basis of a proposed growth model, we discuss possible strategies and limitations for the avoidance of Ga segregation during fast selenisation of metallic precursors. Solar cells made from samples selenised with a total annealing time of 6.5 min reached conversion efficiencies of up to 14.2 % (total area, without anti‐reflective coating). The evolution of the Cu(In,Ga)Se2 diffraction signals reveals that the minimum process time for high‐quality Cu(In,Ga)Se2 absorbers is limited by cation ordering rather than Se incorporation. Copyright © 2014 John Wiley & Sons, Ltd.

Adjusting the Ga grading during fast atmospheric processing of Cu(In,Ga)Se2 solar cell absorber layers using elemental selenium vapor

“…Therefore, the diffusion of residual Ga close to the back contact through CIGSe to the surface must be much slower than the diffusion of residual Cu. This is in accordance with the rather low mobility of Ga in CIGSe that has already been proposed by Mainz et al .…”

“…We address our understanding of the corresponding growth dynamics of the absorber layer in comparison with processes described in the literature. We show that when a controlled amount of Se is supplied by use of the independent sources, the growth of the absorber layer may be distinctly different from that found in previous studies, where Se supply was realized from Se pellets placed next to the sample before heating up both sample and pellets simultaneously . The key finding is that the Ga grading can be adjusted to some extent.…”

“…In this case, In/Ga interdiffusion would occur only in this CISe layer, as suggested, e.g., by Hergert et al . The question is why fast In/Ga interdiffusion should occur in the present study in contrast to previous studies . This growth scenario may be analogous to the three step selenization process utilizing H 2 Se, in which a second annealing step in Se free Ar atmosphere is applied to facilitate In/Ga interdiffusion in CIGSe before the selenization is complete .…”

“…This includes the amount of provided Se as well as the Se vapor composition which is temperature dependent . In case the Se is heated simultaneously with the sample to up to about 600 °C as in previous studies, e.g., , the vapor composition with respect to molecule size and especially the Se vapor pressure can change significantly during the process and may be different above 520 °C compared with our experiment, in which the Se source temperature is kept constant. The impact of the Se source temperature as found in Figure 2 may therefore also be caused by a different Se vapor composition.…”

Real time observation of phase formations by XRD during Ga-rich or In-rich Cu(In, Ga)Se2 growth by co-evaporation

“…Clearly more peaks than the 220/ 204 double peak can be distinguished. For sequential processes, it is a commonly observed phenomenon that Garich phases segregate at the back contact [10] and we expect a similar origin for the phase segregation in this case. Interestingly, we find two different (In, Ga) 2 Se 3 modifications for the varying Ga contents.…”