Recovery after PID-degradation of CIGS-solar cells

Read the original article here

For solar cells that have been connected in series, the potential for the system can be very high. This could give rise to a type of degradation of the solar cells, which is called potential-induced degradation (PID) and leads to a decrease of the module efficiency. This could be avoided by proper design of the electric circuitry, but it could also be interesting to investigate the reasons for the phenomenon and possible ways to mitigate it. PID has been well researched already for the conventional silicon solar cells, but it is relatively unexplored for thin film solar cells. In this study, the dynamics of PID for CIGS-solar cells with two different buffer layers was investigated, as well as the possibility for recovery from the degradation using three different methods.

Solar cells in a stack of Mo/CIGS/buffer layer/i-ZnO/ZnO:Al/Ni-Al-Ni contact were manufactured on a glass substrate with a buffer layer of CdS and Zn(O,S) respectively. These were then degraded by an electrical field between the back side of the glass and the interface between the glass and Mo. In the first recovery method, etching, the top layers of the solar cell down to the CIGS was etched off and replaced with new layers. In the second, accelerated method, the solar cells were places in a reverse electrical field for some time. In the third, unaccelerated method, the samples were stored in darkness at 22°C. The solar cells with Zn(O,S) as buffer layer were only tested with the accelerated method. The samples were analyzed before and after PID as well as after recovery.

After the PID-treatment, all samples had degraded to zero or near-zero efficiency. The samples with CdS as buffer layer improved their performance with both the accelerated and unaccelerated methods. The fastest method for recovery, at least initially, was etching. For samples with Zn(O,S) as buffer layer, no improvement of the efficiency was seen. The depth profile showed that the degraded solar cells contained increased levels of sodium (Na) in the CdS layer and the top third of the CIGS. After acceleration, the Na level decreased in the CdS layer, although it was still higher than before the degradation. For the Zn(O,S)-samples the Na level was higher in the whole CIGS layer, especially in the top third. After acceleration the Na level had decreased, although it was still higher than initially as well as compared to the accelerated CdS-samples.

The buffer layer was therefore demonstrated to be of significant importance for the processes in PID and recovery. For CdS-samples, the buffer layer seems to act as a sink for Na and the degradation originate from either the buffer layer, the interface to the CIGS or the uppermost part of the CIGS layer. This is supported by the improved performance by the etching where the CdS layer was replaced and the interface with the CIGS layer refreshed. For the Zn(O,S)-samples the buffer layer acted as a barrier for Na, which led to a significantly higher Na level in the CIGS layer. Very little improvement was seen in the performance for these samples, which therefore seemed to be permanently degraded. Since the Na level was still high in the accelerated CdS-samples, which had been recovered to almost initial efficiency, the degradation cannot be unambiguously linked to the absolute level of Na. Further experiments could be needed to analyze this.