Keywords :
Chacopyrite; Solar cells; Thin films; Gallium gradient; Absolute photoluminescence; Time-resolved photoluminescence; Cu(In,Ga)(S,Se)2
Abstract :
[en] It is an accepted fact, the Earth temperature is raising at an alarming rate, having devastating repercussions on the life on the planet. To mitigate this effect, it is paramount to reduce the emission of greenhouse gases, major responsible of the global warming. Among other technologies, solar cells are active players in reducing the emission of greenhouse gases. In particular, chalcopyrite-based solar cells, with a low carbon footprint and current record power conversation efficiency of 23.6%, participate to accelerate the energy transition. Despite the already good performance of Cu(In,Ga)Se2 solar cells, higher theoretical efficiency could potentially be achieved (nearly 34% for a material of 1.34 eV band gap), indicating a margin for further improvement. The aim of the current work is to characterize high quality industrial Cu(In,Ga)(S,Se)2 (CIGSSe) absorbers in order to understand the underlying efficiency limitations and their physical origin. Specifically, the intentional band gap gradient formed towards the metallic back contact is investigated. Such an absorber architecture is traditionally used to reduce the back surface recombination, responsible for significant non-radiative losses. Furthermore, the impact of silver (Ag) alloying on the band gap gradient and the performance of the absorbers and corresponding submodules is examined. Photoluminescence (PL) spectroscopy represents the primary investigation method used during this thesis at the Laboratory for Photovoltaics. On one hand, absolute PL measurements allow for quantitative analysis of the absorbers, yielding values for the quasi-Fermi level splitting, the non-radiative losses and the optical diode factor. Combined with a gradual etching of the absorbers, a depth-resolved investigation of the radiative recombination activity is achieved. On the other hand, time-resolved PL (TRPL) leads to an estimation of the charge carrier lifetime and doping density in various absorbers. Collaborations with several European research groups provided access to other investigation techniques such as cathodoluminescence (CL) spectroscopy or Raman spectroscopy, both contributing greatly to the findings of this work. From a combination of the diverse mentioned measurement techniques, a novel model for the band gap gradient is proposed. Instead of the gradual band gap variation suggested by compositional analysis (e.g. GDOES), it is found that mostly two 2 chalcopyrite phases of low (~ 1.04 eV) and high (~ 1.5 − 1.6 eV) band gap form respectively towards the front and the back sides of the absorber and interlace in the bulk of the material, leading to an apparent gradient. Furthermore, alloying the CIGSSe absorbers with low amounts of Ag leads to a decreased number of band gap jumps as observed in the Ag-free samples. The resulting smoother band gap gradient is attributed to an enhancement of the Ga and In interdiffusion in the ACIGSSe absorbers. Moreover, PL decay measurements indicate that the minority carrier lifetime increases upon Ag-alloying. Additionally, a reduction of the nonradiative recombination losses by as much as 30 meV is measured for the best ACIGSSe absorber compared to the Ag-free reference. Furthermore, it is found that improved performance can be achieved in ACIGSSe absorbers grown at a lower process temperature, what could lead eventually to a reduction of the production costs. Nevertheless, a comparison between absorbers and corresponding finished submodules, which showed only a relative improvement of 1.6% in terms of efficiency in the best case, suggests that the cell-making process could be optimized further to fully benefit from the improved ACIGSSe absorbers. All in all, this thesis provides new insights about the CIGSSe absorbers grown at AVANCIS. It challenges the current view of the band gap gradient in graded absorbers and demonstrates that improvement of both the absorbers and the resulting submodules is possible upon Agalloying. Finally, it is established how TRPL investigation may be utilized to determine reliable charge carrier lifetime and doping densities.