Abstract :
[en] Current commercial photovoltaic technologies are close to their practical limits, and enhancing their power conversion efficiency (PCE) requires a paradigm shift to tandem approaches. Tandem solar cells can exceed the single junction practical and thermodynamic limits. The desired top cell bandgap to enhance PCE of current photovoltaic technologies is ~1.6 1.7 eV. The bandgap tunability from 1.5 eV to 2.5 eV positions Cu(In,Ga)S2 as a prime top cell candidate for next generation low-cost tandem cells. However, they are limited by a low external open circuit voltage (VOC,ex). In this thesis, we have studied the interface recombination and found it to cause a difference between VOC,ex and internal open-circuit voltage (VOC,in) in Cu(In,Ga)S2 solar cell. We have introduced a quantifiable metric that has not been used before for Cu(In,Ga)S2, to evaluate VOC disparity in terms of “interface VOC deficit” defined as (VOC,in – VOC,ex). The temperature dependent current-voltage measurement allows to investigate the activation energy (Ea) of the dominating recombination path in the device, uncovering the cause of interface VOC deficit in Cu poor and Cu-rich Cu(In,Ga)S2 devices. We find that negative conduction band offset (CBO) at the absorber/buffer interface results in interface VOC deficit in Cu poor Cu(In,Ga)S2 devices.
Although the interface VOC deficit can be reduced by replacing the buffer for favorable band alignment at the absorber/buffer interface, a substantial deficiency still exists. We observe that the CBO not only at the absorber/buffer interface but also at the buffer/i-layer interface leads to an interface VOC deficit in devices. This, in general, is not an issue in Cu(In,Ga)Se2 devices. By optimizing buffer and i-layer, we mitigate and overcome buffer/i-layer losses to get Cu poor Cu(In,Ga)S2 devices with consistently low interface VOC deficit. As a result, an in-house PCE of 15.1 % is achieved together with an externally certified PCE of 14 %. This is, by far, the best Cu(In,Ga)S2 device performance except for the record PCE device.
In contrast, the interface VOC deficit and the interface recombination persists in Cu-rich Cu(In,Ga)S2 devices and is not resolved by alternative buffers. To identify the possible origin of the interface VOC deficit, we characterize two sister systems CuInS2 and CuInSe2, which offer reduced complexity due to Ga exclusion. The Cu-rich devices of these systems are also known to suffer from interface recombination, and for CuInSe2, it has been linked to the “200 meV” defect. However, the underlying mechanism of how this defect leads to interface recombination remains unknown. Through results obtained from photoelectron spectroscopic measurements, we exclude the possibility of two commonly evoked causes of interface recombination: negative CBO and Fermi-level pinning.
Sulfur-based post-deposition treatments on KCN etched Cu-rich CuInS2 absorbers reveal near interface defects as a possible alternative cause of interface VOC deficit. The treatment increases the VOC,ex, which originates from improved Ea and interface VOC deficit in treated devices. The capacitance transient measurements further reveal that slow metastable defects are present in the untreated sample. The treated samples show that the slow transient is suppressed, suggesting the passivation of slow metastable defects. The treatment adapted to Cu rich CuInSe2 displays a reduction in the deep defect signature in admittance spectra, which explains the observed improvement in interface VOC deficit. This indicates that the defects near the absorber/buffer interface, acting as non-radiative recombination centers, as the source of interface VOC deficit.
Finally, to understand how the defect leads to interface recombination, a new model based on near interface defects is offered using the holistic analysis and evaluation of the defect characteristics. We can reproduce an interface VOC deficit with all the signatures of an interface recombination-dominated device using numerical simulations. This model provides a solution for the consideration of interface recombination by defects distributed in a thin layer within the bulk absorber, an explanation beyond classical models. The near interface defect model finally explains why Cu rich chalcopyrite solar cells are limited in their VOC,ex despite a good VOC,in, which was not discovered before. The model thus forms a new third explanation for interface recombination signature in devices and is applicable to any device with highly recombinative defects near the interface.
