Material and process engineering for cost-effective silicon technologies

Conventionally, in semiconductor industry the device performance has been the main driving force while in photovoltaics the cost has been of utmost importance. This work seeks for the balance between these two approaches. The traditional device performance perspective is covered by developing high-resolution focused ion beam (FIB) patterning processes while the cost perspective is covered by studying the applicability of inexpensive quasi-mono silicon (QM-Si) material. New insights from the opposite perspective are provided by developing boron implantation gettering strategies for high-performance n-type silicon solar cells and by studying the applicability of QM­‑Si beyond photovoltaics. The first challenge that the thesis addresses is related to the low ingot yield caused by impurities present in seed-assisted QM-Si. This work implements the phosphorus diffusion gettering method to recover the scrap QM-Si material. The results show that with this method the minority carrier lifetime in the material can be increased tenfold from 17 µs to 178 µs. Secondly, the thesis reports successful demonstration of boron implantation gettering by which the bulk iron point defect concentration is reduced from 2E14 per cubic centimetre to 1E11 per cubic centimetre. This facilitates further efficiency improvement of advanced n-type solar cells. Next, a novel FIB lithography method is developed for precisely defined nanoscale pattering, both for dark and bright field masks. With the developed method, an aspect ratio of 16:1 is reached with 40 nm diameter silicon nanopillars, demonstrating that the method is capable for producing sub-100 nm resolution high aspect ratio silicon structures. Furthermore, the developed method eliminates the lattice damage and unintentional doping of the substrate. Finally, FIB lithography is applied to patterning of QM-Si. The achieved submicron resolution in a line array suggests that the material has true potential e.g. in microelectromechanical systems (MEMS).