In vacuo XPS investigation of Cu(In,Ga)Se2 surface after RbF post-deposition treatment

Abstract

Latest record efficiencies of Cu(In,Ga)Se2 (CIGSe) solar cells were achieved by means of a rubidium fluoride (RbF) post-deposition treatment (PDT). To understand the effect of the RbF-PDT on the surface chemistry of CIGSe and its interaction with sodium that is generally present in the CIGSe absorber, we performed an X-ray photoelectron spectroscopy (XPS) study on CIGSe thin films as-deposited by a three-stage co-evaporation process and after the consecutive RbF-PDT. The sample transfer from the deposition to the XPS analysis chamber was performed via an ultra-high vacuum transfer system. This allows to minimize air exposure, avoiding oxide formation on the CIGSe surface, especially for alkali-treated absorbers. Beside an expected reduction of Cu-and Ga-content at the surface of RbF-treated CIGSe films, we find that Rb penetrates the CIGSe and, contrary to fluorine, it is not completely removed by subsequent ammonia etching. The remaining Rb contribution at 110.0 eV binding energy, which appears after the RbF-PDT is similar to the one detected on a coevaporated RbInSe2 reference sample and together with a new Se 3d contribution may hence belong to an Rb-In-Se secondary phase on the CIGSe surface. In addition, Na is driven towards the surface of the CIGSe absorber as a direct result of the RbF-PDT. This proves the ion-exchange mechanism in the absence of moisture and air/oxygen between heavy Rb atoms incorporated via PDT and lighter Na atoms supplied by the glass substrate. A remaining XPS signal of Na 1s is observed after etching the vacuum transferred RbF-CIGSe sample, indicating that Rb and/or F are not as much a driving force for Na as oxygen usually is.

Properties of Co‐Evaporated RbInSe2Thin Films

“…With the methods utilized in this study we could not detect the formation of an RbInSe 2 phase after an RbF‐PDT of an In 2 Se 3 layer. However, a more surface‐sensitive comparison of the chemical environment of the Rb‐atoms in our RbInSe 2 sample with the one of an RbF‐treated CIGS thin film shows good agreement . The fact that both theoretical and experimental results show a higher bandgap energy for RbInSe 2 compared to the one of In 2 Se 3 and CIGS is supporting the hypothesis that interface recombination at the CIGS/CdS junction could be reduced due to the formation of RbInSe 2 during an RbF‐PDT.…”

“…However, based on these results, no direct conclusion can be drawn on the formation of RbInSe 2 during the RbF‐PDT of CIGS. Nonetheless, as a result of an earlier X‐Ray Photoelectron Spectroscopy (XPS) investigation a good match of the Rb 3d core level binding energies was already shown between the same RbInSe 2 sample discussed here to those of RbF‐treated CIGS . We also investigated the formation of intrinsic defects in RbInSe 2 .…”

“…Nonetheless, as a result of an earlier X-Ray Photoelectron Spectroscopy (XPS) investigation a good match of the Rb 3d core level binding energies was already shown between the same RbInSe 2 sample discussed here to those of RbF-treated CIGS. [24] Figure 1. XRD-diffractogram of our In-Se, In-Se þ RbF, and Rb-In-Se samples (solid lines) in comparison to reference diffractograms of In 2 Se 3 (used for In-Se and In-Se þ RbF), [17] RbInSe 2 (dashed lines), [16] and markers showing the peaks related to the molybdenum back contact.…”

Limitation of Current Transport across the Heterojunction in Cu(In,Ga)Se2Solar Cells Prepared with Alkali Fluoride Postdeposition Treatment

“…[43] We have shown that even a small change in the spike at the CIGS/CdS interface, when combined with the CdS/ZnO cliff, can also significantly affect the current transport. Accumulation of alkali metals and fluoride [22,44] could also change this band lineup via the formation of dipoles. A formation of an alkali-In-Se layer at the surface of the absorber after PDT [5,7,21,22] might also lead to a change in the band lineup to the buffer layer.…”

“…The formation of secondary phases such as RbInSe 2 has been described at the CIGS surface during PDT and interpreted as the layer responsible for the surface bandgap increase in the absorber. [5,[20][21][22] As the goal here is to understand the principle mechanisms of limited current transport across the interface, no such layer has been explicitly added to the models. However, the main effects of such a layer would be a change in the band offsets and local doping which are considered in our model.…”