|
Controllable doping in semiconductor nanowires is essential for development of optoelectronic devices. Despite great progress, a fundamental challenge remains in controlling the uniformity of doping, particularly in the presence of relatively high levels of geometrical inhomogeneity in bottom-up growth. A relatively high doping level of 1E18 cm-3 corresponds to just ~1000 activated dopants in a 2µm long, 50nm diameter nanowire. High-throughput photoluminescence spectroscopy enables the collection of doping distributions across many (>10k) nanowires, but geometric variation adds additional uncertainty to the modelling. We present an approach that uses large datasets of doping and emission intensity to infer both doping and diameter across a growth, and apply Bayesian methods to study the underlying distributions in Zn-doped aerotaxy-grown GaAs nanowires. This new big-data enabled approach provides a route to exploit inherent inhomogeneity to reveal fundamental recombination mechanisms.
|
