Thèse de doctorat en Physique des plasmas
Sous la direction de Pere Roca i Cabarrocas.
Soutenue en 2010
En anglais : High deposition rates on large areas are industrial needs for mass production of microcrystalline silicon (μc-Si:H) solar cells. This doctoral work aims at exploring the usefulness of Matrix Distributed Electron Cyclotron Resonance (MDECR) plasmas to process the intrinsic layer of μc-Si:H p-i-n solar cells at high rates. With the high dissociation of silane achieved in MDECR plasmas, deposition rates as high as 6nm/s and 2. 8nm/s have been demonstrated in our lab for amorphous and microcrystalline silicon respectively, without hydrogen dilution. This technique is also promising because it can be easily scaled up on large areas, just by extending the matrix of elementary microwave applicators. This subject was a unique opportunity to cover the whole chain of this field of research : A new MDECR reactor has been specially designed and assembled during this project. Its maintenance and its improvement have been important technical challenges : for example, the addition of a load-lock enabled us to lower the oxygen concentration in our films by a factor of 10. The impact of the deposition parameters (e. G. The ion energy, the substrate temperature, different gas mixtures, the microwave power) has been explored in extensive parametric studies in order to optimize the material quality. Great efforts have been invested in the characterization of the films. Our strategy has been to develop a wide range of diagnostics (ellipsometry, Raman spectroscopy, SIMS, FTIR, XRD, electrical characterizations etc. ). Finally, p-i-n cells have been processed with the selected interesting materials. The successive successful improvements in the material quality (e. G. Diffusion lengths of holes parallel to the substrate as high as 250 nm) did unfortunately not result in high efficiency solar cells. Their limited performance is in particular due to a very poor response in the red part of the spectrum resulting in low current densities. Consequently, the potential sources of limitation of the reactor, the material and the device have been studied : e. G. The presence of “cracks” prone to post-oxidation in the highly crystallized materials and the risk of deterioration of the ZnO substrate or of the p-doped layer by a too high process temperature or by hydrogen diffusing from the plasma.
Pas de résumé disponible.