Modelling of electron and hole ionisation in Al x Ga 1− x As 0.56 Sb 0.44 alloys using Weibull–Fréchet distributions

Abstract Accurately modelling of the exceptional low excess noise observed in Al x Ga 1− x As 0.56 Sb 0.44 avalanche photodiodes (APDs) is crucial for optimising device performance. In this study, the random path length model, incorporating the Weibull–Fréchet distribution function, was used to simulate electron and hole impact ionisation in APDs with non-uniform electric fields. The model successfully reproduces the experimentally measured multiplication gain, ⟨ M ⟩ and excess noise factor, F , in electron-initiated APDs with compositions Al 0.55 Ga 0.45 As 0.56 Sb 0.44 , Al 0.75 Ga 0.25 As 0.56 Sb 0.44 , Al 0.85 Ga 0.15 As 0.56 Sb 0.44 , and AlAs 0.56 Sb 0.44 , while also predicting a steep increase in F for hole-initiated APDs. The results demonstrate that ionisation path length distributions are strongly influenced by electric field strength and alloy composition. The model effectively captures the probability density function of ionisation path lengths, which is responsible for low excess noise. The results reveal that electron dead space increases as Al composition decreases, exhibiting an inverse trend compared to the reducing mean ionisation path length in these alloys. This behaviour is attributed to the alloy scattering effects, which become more pronounced in mid-composition alloys.