Formation of the physical vapor deposited CdS∕Cu(In,Ga)Se2 interface in highly efficient thin film solar cells

Abstract

We report on the buffer/absorber interface formation in highly efficient ͑14.5%, air mass 1.5͒ ZnO/CdS/Cu͑In, Ga͒Se 2 solar cells with a physical vapor deposited CdS buffer. For Se-decapped Cu͑In, Ga͒Se 2 ͑CIGSe͒ absorbers we observe sulfur passivation of the CIGSe grain boundaries during CdS growth and at the interface a thermally stimulated formation of a region with a higher band gap than that of the absorber bulk, determining the height of the potential barrier at the CdS / CIGSe interface. For air-exposed CIGSe samples the grain boundary passivation is impeded by a native oxide/adsorbate layer at the CIGSe surface determining the thermal stability of the potential barrier height. © 2006 American Institute of Physics. ͓DOI: 10.1063/1.2190768͔ Cu͑In, Ga͒Se 2 ͑CIGSe͒ thin film solar cells achieve efficiencies up to 19.5% when using a "wet" chemical bath deposited ͑CBD͒ CdS buffer layer between the p-type absorber and the n-type ZnO window. 1 Surprisingly, CIGSe polycrystalline thin film solar cells exhibit higher internal quantum efficiencies ͑QEs͒ than devices from single crystalline CIGSe films. 2 Recent theoretical 3 and experimental 4 studies propose an explanation based on a model where Cu depleted polar grain boundary ͑GB͒ interfaces lead to a valence band offset between GB and grain interior ͑GI͒, repeling holes from the GB and thus diminishing GB recombination. Other experiments observed light induced changes in the conduction band at GBs. 5 Nevertheless, when a CdS buffer is deposited by a "dry" physical vapor deposition ͑PVD͒ the CI͑G͒Se solar cells show significantly lower efficiency. 6 The highest reported efficiency using a PVD-CdS is ϳ12.4%. 7 This shows that differently deposited CdS buffers provide differences at the related interfaces and finally in the device operation.In this Letter we show that ͑1͒ not only the initial state of the absorber GBs has a decisive role but that additionally a GB passivation process, which occurs during the CdS buffer deposition, enhances the absorber electronic properties; ͑2͒ the condition of the absorber surface strongly influences the passivation process and the potential barrier of recombination at the CdS / CIGSe interface; and ͑3͒ the preparation of highly efficient solar cells with PVD-CdS buffers is possible.CIGSe thin films ͓Ga/ ͑Ga+ In͒ =24%͔ were deposited by a three-stage coevaporation process 8 on Mo-coated soda lime glass substrates. An ϳ300 nm Se cap protection layer was deposited on top of the CIGSe absorber. Se decapping and in situ thermal deposition of a CdS layer from a single source were performed in an ultrahigh vacuum ͑UHV͒ ͑p ϳ 10 −9 mbar͒ system. Samples comparable to those in standard solar cell production were prepared by exposure to air for 40 min prior to CdS deposition. This results in the formation of Na 2 CO 3 and also of In 2 O 3 , Ga 2 O 3 , and SeO x native oxides at the absorber surface. 9 For the deposition of the CdS layers, the CdS source and substrate temperatures were kept constant at 680 and 100°C, respectively, providin...

Surface potential of chalcopyrite films measured by KPFM

“…We have followed this route using a co-evaporated Cu(In, Ga)Se 2 thin film. Topography and work function images (not shown) obtained by KPFM show similar results to the previous samples [25] and the histogram of the work function image is presented in Fig. 3 (c).…”

Microscopic characterization of individual grain boundaries in Cu-III–VI2 chalcopyrites

“…However, in the final solar cell device, the state of the GB could be modified by the further deposition processes during the device fabrication, especially when considering that diffusion along GBs is considerably facilitated with respect to bulk diffusion. Rusu et al [29] have presented a study of UHV-clean Cu(In,Ga)Se 2 films onto which CdS was evaporated under UHV conditions. In vacuum transfer into a KPFM system allowed to study the GB properties as a function of the deposited CdS layer thickness.…”

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

“…11 Basic knowledge related to the chemical and electronic properties of CIGSe/ CdS interfaces has been gained previously, 12−14 but many aspects related to the CdS growth mechanism and its effect on the interface quality, especially for CIGSe subjected to an alkali PDT, 15−18 are still not well understood, in particular taking spatial inhomogeneities into consideration. 19,20 In this work, we investigate the early stages of the CdS growth on RbF-PDT CIGSe absorbers, which result in efficiencies around 20% (w/ ARC) for related reference cells, by a combination of spatially resolved and integrating techniques. Several samples, with CBD treatment times varying between 1 s and 5 min, are investigated by atomic (AFM) and Kelvin probe (KPFM) force microscopy, hard X-ray photoelectron spectroscopy (HAXPES), and soft Xray emission spectroscopy (XES).…”