Narrow band gap Cu(In,Ga)Se2 for tandem solar cell application

Abstract

Photovoltaic (PV) energy generation has become one of the key pillars of the shift to a renewable energy future. Current devices, under favorable conditions, can already undercut the price per kWh electricity of other technologies on the market. Further reduction in the cost of installed PV systems and increase in solar module conversion efficiency will improve the affordability even more and will substantially aid in wider market penetration and enhance the volume of PV installations. Currently the PV market is dominated by silicon wafer based solar cells, but alternative technologies offer some distinctive advantages, making them interesting for numerous applications. Thin film technologies, as for example based on Cu(In,Ga)Se2 (CIGS) compound semiconductors with high optical absorption coefficient, are becoming important due to lower material and energy requirements for processing of high conversion efficiency solar cells. Inherent advantages are large area depositions with low production costs, and the possibilities for construction of lightweight, flexible devices with roll-to-roll manufacturing processes. The highest efficiency of single-junction CIGS solar cells is approaching the thermodynamic limit, making the use of alternative concepts such as concentration or multijunction (tandem-) devices the next logical step for further increase in efficiency beyond the Shockley-Queisser limit (S-Q limit). Especially the multi-junction technology, in which the thermodynamic losses are reduced by stacking of solar cells with different band gaps, decreasing thermalization of charge carriers excited with energies above the band gap, is a promising approach for enhanced utilization of the solar spectrum, yielding improved efficiency. Such devices, based on epitaxial layers of III-V compounds have already demonstrated remarkably high efficiencies beyond the S-Q limit. However, these devices grown on rather expensive single crystal wafers and with small size are prohibitively pricey for low cost terrestrial solar electricity generation. On the other hand, multi-junction solar cell technology based on polycrystalline thin films is an attractive option for large area, low cost production, provided adequately high efficiencies are achieved. In this context, two-junction tandem devices, developed by stacking a semitransparent large band gap solar cell of 1.6-1.7 eV on top of a low band gap (~1.0 eV) bottom cell, is a viable option. Earlier attempts in this direction were not so successful, but with the rise of perovskite thin film solar cells as a compatible high efficiency wide band gap (>1.6 eV) top cell and CIGS with a tunable band gap as bottom cell, the prospect for all thin film tandem devices with efficiencies beyond the single-junction limitations has opened. Such all thin film devices hold the potential for the low cost production necessary for large scale terrestrial application. This thesis focuses on the development of high efficiency narrow bandgap (1.0 eV) CIGS solar cells for application in all thin film tandem devices. While for CIGS with band gap of around 1.15 eV efficiencies of over 23 % have been demonstrated, cells with a narrow band gap close to 1.0 eV only reach 15.0 %. The efficiency of these narrow band gap cells are limited by charge carrier recombination, leading to low open circuit voltage (VOC) and reduced fill factor. For solar cell efficiency enhancement it is necessary to investigate the underlying reasons contributing to the deficits in PV parameters and develop processes to overcome the limiting factors. An option to reduce recombination within the solar cell is the implementation of a band gap grading as discussed in Chapter 3. The increase of the band gap at the location of highest recombination leads to a reduction in diode current, and therefore an increase in VOC. To keep the band gap of 1.0 eV a substantial part of the absorber needs to be Ga free. As the primary source of recombination is not obvious, different gradings (realized by a change in the Ga to In ratio) are implemented and compared. A single grading with increased band gap (higher Ga/In ratio) towards the front of the absorber shows no significant improvement on photovoltaic parameters. Any gain in VOC is offset by losses in current due to reduced charge collection, mainly visible for long wavelength photons and probably a result of the upwards bending in the conduction band. A single backgrading (higher Ga/In ratio towards the back electric contact) on the other hand leads to substantial improvements in performance ( from 12.0 % to 16.1 %). It is shown that the collection of photo-generated charge carriers improves and recombination is reduced. Measurements of the effective lifetime by time resolved photo-luminescence are carried out, showing an increase from approximately 20 ns to 100 ns when comparing ungraded with back-graded absorbers. By selectively changing the recombination speed at the back contact, strong differences in the behavior of cells with and without a band gap widening towards the back are observed. The results support that considerable recombination at the back contact is present in pure CIS solar cells, and that the single Ga back-grading approach is effective at suppressing this loss channel. In Chapter 4 the alkali treatment of CIS based solar cells is investigated. Alkali elements are known to strongly influence doping and passivation in CIGS solar cells. It is shown that the amount of sodium necessary to reach sufficient doping levels for high performance CIS solar cells is not achieved using the processes developed for CIGS. This may be based on insufficient Na diffusion into the grain, as those cells generally show larger grains than their CIGS counter parts, and since alkali migration energies in CIS are reported to be higher compared to those in CGS. If CIS cells are grown on soda lime glass without any diffusion barrier and additionally receive post deposition treatment (PDT) with NaF they still show low apparent doping concentration and poor PV performance ( = 10.9 %). However, additional annealing at ~ 370 C substrate temperature after PDT is shown to solve this problem, leading to an increase in apparent doping levels close to 1016 cm−3 and cell efficiency of 15.0 %. The application of an additional heavy alkali PDT, specifically RbF, is shown to lead to further improvements in cell efficiency. Changes at the front interface due to the PDT allow a decrease of buffer layer thickness, leading to a higher photo current (approximately + 1.0 mAcm−2). In addition, reduced recombination and the resulting increase in lifetime leads to additional gains in VOC, resulting in considerably improved device performance, up to an efficiency of 18.0 %. Further efficiency improvement is achieved by investigating the effect of close to stoichiometric compositions of Cu to group III elements as described in Chapter 5. The sub-stoichiometric Cu composition of state-of-the-art CIGS absorbers leads to a high concentration of detrimental defects. The defect density within the absorbers is reduced by approaching a stoichiometric Cu composition. Improvements in the defect density are identified by the decrease of Urbach energy from 20 to 16 mV and an increase in doping is observed for cells with almost stoichiometric Cu content. Cells with high, and especially stoichiometric Cu composition tend to be limited by recombination at the front interface, leading to a decrease of VOC of about 20 mV. Using the modified absorber surface after heavy alkali PDT, these losses are suppressed. Based on these improvements, a narrow band gap cell with record breaking 19.2 % efficiency and an open circuit voltage of 609 mV is achieved. Throughout the whole thesis the suitability of these cells for tandem devices with semitransparent perovskite top cells is investigated by 4-terminal tandem measurements. The improvements achieved in this work led to CIS based solar cells that not only show outstanding single cell performance, but also enable highly efficient tandem devices up to 25.0 %. They outperform state-of-the-art single junction CIGS and perovskite cells while showing prospects for further efficiency improvement. Due to the low band gap of the CIS absorber the current density from the bottom cell is high enough to produce current matched tandem devices with high efficient perovskite top cells (19.2 to 18.6 mAcm−2 in 4-terminal configuration), and also monolithic two-terminal configurations are feasible in the future.