Silicon heterojunction cells are currently the silicon-based single junction cells with the highest conversion efficiency, with record of 27.09% [1]. SHJ cells are very attractive for industrial production, and many companies are announcing/building new SHJ capacities for an estimated overall nameplate capacity > 218 GW worldwide [2]. However, with such volumes, SHJ cell manufacturing could be limited to 37-95 GW/year due to the indium (In) supply, considering the actual material consumption [3]. Indium is used in TCO (Transparent Conductive Oxides) layers, deposited on both sides of the cells, commonly in the form of tin-doped indium oxide (ITO) for its excellent opto-electrical properties (transparent and conductive). With the rapid upscaling of SHJ and other In-rich technologies, we can expect increase in In prices, or even material shortage. Thus, in order to have a more sustainable SHJ cell production, it is necessary to find solutions to decrease indium usage.

Hence, this thesis aims at finding solutions to decrease indium usage in SHJ cells, without efficiency or reliability losses in a module environment, with processes compatible with industrial production. Two solutions will be investigated here.

First, ultrathin ITO layers (< 15 nm against the 100 nm commonly used) coupled with a SiN:H layer will be experimented. We will see that, thanks to alternative selective layers on the front side ((n) nc-Si:H), it is possible to use ITO layers as thin as 15 nm on the front side without efficiency loses (-85% In on front side), with an enhanced UV reliability and a resistance against humidity comparable with reference cells at module level. However, this solution showed its limitations on the rear side, due to resistive losses. The development of p doped nc-Si:H layers could potentially allow a better integration of those ultrathin ITO layers on the rear side.

The second solution explored in this work is the use of an In-free TCO: AZO (aluminum doped zinc oxide). We will see that the lower transparency of AZO was shown to be problematic, but thanks to the use of thinner layers on the front side (30 nm) coupled with an SiN:H layer, we reached efficiency comparable to that of the reference cells (-100% In on the front side). On the rear side, the lower transparency of AZO was also shown to be troublesome, and decreasing the AZO layer thickness led to resistive losses. Moreover, AZO is know to suffer from severe degradations when exposed to humidity. We showed here that when used on the rear side of SHJ cells, AZO layer only led to minor degradation increase upon damp heat (DH) tests and that when an ultrathin 10 nm ITO layer is deposited on top of the AZO layers, the reliability against humidity is improved. However, the cells using thin AZO layers coupled with SiN:H layers on the front side showed important FF losses after 500 h of DH tests, attributed to an interaction between the AZO and SiN:H layers. Thus, it is necessary to investigate the use of alternative dielectric layers, or to change the module architecture to solve this issue.

Membres du jury/ Jury members :