Cd and Cu Interdiffusion in Cu(In, Ga)Se₂/CdS Hetero-Interfaces

Abstract

We report a detailed characterization of an industrylike prepared Cu(In,Ga)S e2 (CIGS)/CdS heterojunction by scanning transmission electron microscopy (S TEM) and photoluminescence (PL). Energy dispersive x-ray spectroscopy (EDS) shows the presence of several regions in the CIGS layer that are Cu deprived and Cd enriched, suggesting the segregation of Cd-S e. Concurrently, the CdS layer shows Cd-deprived regions with the presence of Cu, suggesting a segregation of Cu-S. The two types of segregations are always found together, which, to the best of our knowledge, is observed for the first time. The results indicate that there is a diffusion process that replaces Cu with Cd in the CIGS layer and Cd with Cu in the CdS layer. Using a combinatorial approach we identified that this effect is independent of focused-ion beam sample preparation and of the TEM-grid. Furthermore, photoluminescence measurements before and after an HCl etch indicate a lower degree of defects in the post-etch sample, compatible with the segregates removal. We hypothesize that Cu2-xSe nanodomains react during the chemical bath process to form these segregates since the chemical reaction that dominates this process is thermodynamically favourable. These results provide important additional information about the formation of the CIGS /CdS interface.

Giant Voc Boost of Low‐Temperature Annealed Cu(In,Ga)Se2 with Sputtered Zn(O,S) Buffers

“…In analogy to the room temperature CIGS/CdS interface, our XPS analysis of the proxy CIGS/Zn(O,S) junction is consistent with elemental diffusion upon annealing and conversion of some of the sulfide into sulfate. This has two consequences.…”

Evidence for Chemical and Electronic Nonuniformities in the Formation of the Interface of RbF-Treated Cu(In,Ga)Se2 with CdS

“…6 Other studies using energy-dispersive X-ray spectroscopy (EDS) suggest that the CBD process induces a chemical reaction at the chalcopyrite surface which, besides other effects (i.e., surface cleaning), creates CdSe in the form of segregated clusters. 37,38 For CBD times exceeding the nucleation period of the CBD CdS process, we speculate that these clusters could act as nucleation centers resulting in a preferred deposition of CdS. These clusters alone, however, cannot explain the observed attenuation of the HAXPES absorber signals discussed above, and so, a buffer layer conformally covering the CIGSe absorber must also be formed in the CBD.…”

Effect of KF absorber treatment on the functionality of different transparent conductive oxide layers in CIGSe solar cells

“…Because a copper sample holder was used here, the Cu signal is overestimated in both scans as well. The high Cu content inside the CdS buffer layer has been observed in many publications before, when FIB‐prepared samples are analyzed by TEM or atom probe tomography (APT) . It is suggested here, that this is likely an artifact of FIB preparation, because the GDOES measurements on the same samples showed no Cu inside the CdS layer.…”