Logo: to the web site of Uppsala University

We studied the interaction of four different window layer combinations in Cu(In,Ga)Se-2 solar cells. Intrinsic ZnO (i-ZnO) layers were grown on CdS by either chemical vapor deposition (CVD) or magnetron sputtering. These were combined with sputtered ZnO:Al or In2O3:H grown by atomic layer deposition as transparent conducting oxides (TCO). It was found that the thickness of the CVD i-ZnO layer affects the open circuit voltage (V-oc) significantly when using In2O3:H as TCO. The V-oc dropped by roughly 30 mV when the i-ZnO thickness was increased from 20 to 160 nm. This detrimental effect on V-oc was not as prominent when a ZnO:Al TCO was used, where the corresponding decrease was in the range of 5 to 10 my. In addition, the V-oc drop for the CVD i-ZnO/In2O3:H structure was not observed when using the sputtered i-ZnO layer. Furthermore, large fill factor variations were observed when using the In2O3:H TCO without an i-ZnO layer underneath, where already a thin (20 nm) CVD i-ZnO layer mitigated this effect. Device simulations were applied to explain the experimentally observed Voc trends.

Tandem solar cell structures require a high‐performance wide band gap absorber as top cell. Apossible candidate is CuGaSe2, with a fundamental band gap of 1.7 eV. However, a significantopen‐circuit voltage deficit is often reported for wide band gap chalcopyrite solar cells likeCuGaSe2. In this paper, we show that the open‐circuit voltage can be drastically improved in wideband gap p‐Cu(In,Ga)Se2and p‐CuGaSe2devices by improving the conduction band alignment tothe n‐type buffer layer. This is accomplished by using Zn1−xSnxOy, grown by atomic layer deposi-tion, as a buffer layer. In this case, the conduction band level can be adapted to an almost perfectfit to the wide band gap Cu(In,Ga)Se2and CuGaSe2materials. With an improved buffer bandalignment for CuGaSe2absorbers, evaporated in a 3‐stage type process, we show devicesexhibiting open‐circuit voltages up to 1017 mV, and efficiencies up to 11.9%. This is to the bestof our knowledge the highest reported open‐circuit voltage and efficiency for a CuGaSe2device.Temperature‐dependent current‐voltage measurements show that the high open‐circuit voltageis explained by reduced interface recombination, which makes it possible to separate theinfluence of absorber quality from interface recombination in future studies.

We investigate the possibility to combine Zn(O,S) buffer layers grown by atomic layer deposition (ALD) with KF post-deposition treated Cu(In,Ga)Se-2 (CIGS-KF) in solar cells. It is shown that the beneficial effect on open-circuit voltage from the post-deposition treatment is essentially independent of buffer layer material. However, a wet chemical surface treatment is required prior to ALD in order to achieve competitive fill factor values. A water rinse is sufficient to create an absorber surface similar to the one formed during a conventional CdS chemical bath deposition process. However, it is observed that CIGS-KF/Zn(O,S) devices made with water-rinsed absorbers systematically result in lower fill factor values than for the corresponding CIGS-KF/CdS references. This effect can be mitigated by decreasing the H2S:H2O precursor ratio during ALD initiation, indicating that the fill factor limitation is linked to the initial Zn(O,S) growth on the modified CIGS-KF surface. The best CIGS-KF/Zn (O,S) devices were fabricated by etching away the KF-modified surface layer prior to ALD, followed by a low temperature anneal. The thermal treatment step is needed to increase the open-circuit voltage close to the value of the CdS devices. The results presented in this contribution indicate that the main beneficial effects from KFPDT in our devices are neither associated with the CdS CBD process nor due to the formation of a K-In-Serich phase on the CIGS surface.

A wide range of applications benefit from transparent semiconducting oxides with tunable electronic properties, for example, electron transport layers in solar cell devices, where the electron affinity is a key parameter. Presently, a few different ternary oxides are used for this purpose, but the attainable electron affinity range is typically limited. In this study, the authors develop a low-temperature atomic layer deposition (ALD) process to grow amorphous Sn1-xGaxOy thin films from dimethylamino-metal complexes and water. This oxide is predicted to provide a wide selection of possible electron affinity values, from around 3 eV for pure Ga2O3 to 4.5 eV for pure SnO2. The ALD process is evaluated for deposition temperatures in the range of 105-195 degrees C by in situ quartz crystal microbalance and with ex situ film characterization. The growth exhibits an ideal-like behavior at 175 degrees C, where the film composition can be predicted by a simple rule of mixture. Depending on film composition, the growth per cycle varies in the range of 0.6-0.8 angstrom at this temperature. Furthermore, the film composition for a given process appears insensitive to the deposition temperature. From material characterization, it is shown that the deposited films are highly resistive, fully amorphous, and homogeneous, with moderate levels of impurities (carbon, nitrogen, and hydrogen). By tailoring the metal cation ratio in films grown at 175 degrees C, the optical bandgap can be varied in the range from 2.7 eV for SnO2 to above 4.2 eV for Ga2O3. The bandgap also varies significantly as a function of deposition temperature. This control of properties indicates that Sn1-xGaxOy is a promising candidate for an electron transport layer material in a wide electron affinity range. Published by the AVS.

Amorphous tin-gallium oxide (a-SGO) grown with atomic layer deposition was evaluated as a buffer layer in (Ag,Cu)(In,Ga)Se₂ thin-film solar cells in search for a new material that is compatible with a variety of absorber band gaps. Hard and soft X-ray photoelectron spectroscopy on absorber/a-SGO stacks combined with J–V characterization of solar cells that were fabricated, showed that the conduction band alignment at the absorber/a-SGO interface can be tuned by varying the cation composition and/or growth temperature. Here, the surface band gap was 1.1 eV for the absorber. However, optical band gap data for a-SGO indicate that a suitable conduction band alignment can most likely be achieved even for wider absorber band gaps relevant for tandem top cells. A best efficiency of 17.0% was achieved for (Ag,Cu)(In,Ga)Se₂/a-SGO devices, compared to η = 18.6% for the best corresponding CdS reference. Lower fill factor and open-circuit voltage values were responsible for lower cell efficiencies. The reduced fill factor is explained by a larger series resistance, seemingly related to interface properties, which are yet to be optimized. Some layer constellations resulted in degradation in fill factor during light soaking as well. This may partly be explained by light-induced changes in the electrical properties of a-SGO, according to analysis of Al/SGO/n-Si metal-oxide-semiconductor capacitors that were fabricated and characterized with J–V and C–V. Moreover, the introduction of a 1 nm thick Ga₂O₃ interlayer between the absorber and a-SGO improved the open-circuit voltage, which further indicates that the absorber/a-SGO interface can be improved.