Avalanche Breakdown in Gallium Arsenidep−nJunctions

Abstract

Charge multiplication and avalanche breakdown have been studied in relatively narrow diffused $p\ensuremath{-}n$ junctions in GaAs. For these studies both junctions with and without microplasma effects were available. From the charge multiplication studies in microplasma-free junctions, the ionization rate, $\ensuremath{\alpha}$, has been determined as a function of the field, $\mathcal{E}$. These results are unusual in that they give a decidedly better fit to Wolff's theory than to Shockley's; this is found to be consistent with the criterion, $e\mathcal{E}\ensuremath{\lambda}g(\mathrm{o}\mathrm{p}\mathrm{t}\mathrm{i}\mathrm{c}\mathrm{a}\mathrm{l}\ensuremath{-}\mathrm{p}\mathrm{h}\mathrm{o}\mathrm{n}\mathrm{o}\mathrm{n}\phantom{\rule{0ex}{0ex}}\mathrm{e}\mathrm{n}\mathrm{e}\mathrm{r}\mathrm{g}\mathrm{y})$, discussed in a more recent theory due to Baraff which combines the approaches of both Wolff and Shockley. From detailed comparisons with Baraff's theory, the mean free path, $\ensuremath{\lambda}$, for the hot carriers between collisions involving optical phonon emission is 15\ifmmode\pm\else\textpm\fi{}2 \AA{} and the threshold energy for pair production is 1.7\ifmmode\pm\else\textpm\fi{}0.3 eV.Strong correlations were found between the quality of the breakdown (i.e., formation or absence of microplasmas) and the quality of the substrate crystal. Those junctions that exhibited microplasma noise and several multiplication peaks also had soft reverse characteristics, whereas the junctions that appeared to be free from microplasmas had hard breakdown characteristics. With the soft junctions, quantitative comparisons were made between the dark current and the number of microplasmas.