A Theoretical Comparison of the Breakdown Behavior of $\hbox{In}_{0.52}\hbox{Al}_{0.48}\hbox{As}$ and InP Near-Infrared Single-Photon Avalanche Photodiodes

Abstract

<para xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> We study the breakdown characteristics and timing statistics of InP and <formula formulatype="inline"><tex Notation="TeX">$\hbox{In}_{0.52}\hbox{Al}_{0.48}\hbox{As}$</tex> </formula> single-photon avalanche photodiodes (SPADs) with avalanche widths ranging from 0.2 to 1.0 <formula formulatype="inline"> <tex Notation="TeX">$\mu{\hbox {m}}$</tex></formula> at room temperature using a random ionization path-length model. Our results show that, for a given avalanche width, the breakdown probability of <formula formulatype="inline"><tex Notation="TeX">$\hbox{In}_{0.52}\hbox{Al}_{0.48}\hbox{As}$</tex> </formula> SPADs increases faster with overbias than InP SPADs. When we compared their timing statistics, we observed that, for a given breakdown probability, InP requires a shorter time to reach breakdown and exhibits a smaller timing jitter than <formula formulatype="inline"><tex Notation="TeX">$\hbox{In}_{0.52}\hbox{Al}_{0.48}\hbox{As}$</tex> </formula>. However, due to the lower dark count probability and faster rise in breakdown probability with overbias, <formula formulatype="inline"><tex Notation="TeX">$\hbox{In}_{0.52}\hbox{Al}_{0.48}\hbox{As}$</tex></formula> SPADs with <formula formulatype="inline"><tex Notation="TeX">$\hbox{avalanche}\ \hbox{widths}\leq 0.5\ \mu{\hbox {m}}$</tex></formula> are more suitable for single-photon detection at telecommunication wavelengths than InP SPADs. Moreover, we predict that, in InP SPADs with <formula formulatype="inline"><tex Notation="TeX">$\hbox{avalanche}\ \hbox{widths}\leq 0.3\ \mu{\hbox {m}}$</tex></formula> and <formula formulatype="inline"> <tex Notation="TeX">$\hbox{In}_{0.52}\hbox{Al}_{0.48}\hbox{As}$</tex></formula> SPADs with <formula formulatype="inline"><tex Notation="TeX">$\hbox{avalanche}\ \hbox{widths}\leq 0.2\ \mu{\hbox {m}}$</tex></formula>, the dark count probability is higher than the photon count probability for all applied biases. </para>