Nucleation of microcrystalline silicon: on the effect of the substrate surface nature and nano-imprint topography

Abstract

The nucleation of microcrystalline silicon thin-films has been investigated for various substrate natures and topographies. An earlier nucleation onset on aluminium-doped zinc oxide compared to glass substrates has been revealed, associated with a microstructure enhancement and reduced surface energy. Both aspects resulted in a larger crystallite density, following classical nucleation theory. Additionally, the nucleation onset was (plasma deposition) condition-dependent. Therefore, surface chemistry and its interplay with the plasma have been proposed as key factors affecting nucleation and growth. As such, preliminary proof of the substrate nature's role in microcrystalline silicon growth has been provided. Subsequently, the impact of nano-imprint lithography prepared surfaces on the initial microcrystalline silicon growth has been explored. Strong topographies, with a 5-fold surface area enhancement, led to a reduction in crystalline volume fraction of ~20%. However, no correlation between topography and microstructure has been found. Instead, the suppressed crystallization has been partially ascribed to a reduced growth flux, limited surface diffusion and increased incubation layer thickness, originating from the surface area enhancement when transiting from flat to nanostructured surfaces. Furthermore, fundamental plasma parameters have been reviewed in relation with surface topography. Strong topographies are not expected to affect the ion-to-growth flux ratio. However, the reduced ion flux (due to increasing surface area) further limited the already weak ion energy transfer to surface processes. Additionally, the atomic hydrogen flux, i.e. the driving force for microcrystalline growth, has been found to decrease by a factor of 10 when transiting from flat to nanostructured topography. This resulted in an almost 6-fold reduction of the hydrogen-to-growth flux ratio, a much stronger effect than the ion-to-growth flux ratio. Since previous studies regarding crystalline growth stated the necessity for enhanced ion-and atomic hydrogen-to-growth flux ratios, reduction of the latter is suggested to be the main cause of suppressed crystallization kinetics.