Explaining the Different Time Constants Extracted from Low Frequency Y22 and $I_{DS}$-DLTS on GaN HEMTs

This paper presents an explanation for the observed differences on the trapping time constants extracted from low frequency
$Y_{22}$
and
$I_{DS}$
-Deep-Level Transient Spectroscopy (DLTS) on GaN high electron mobility transistors (HEMTs). It is shown that the trapping time constant, not being a physical identity but a parameter of a model, can vary according to the conditions of extraction. This dependence is thus clearly seen in
$Y_{22}$
and
$I_{DS}$
-DL TS measurements, as they are made respectively under small-signal steady-state and large-signal transient conditions. A theoretical explanation supported by the Shockley-Read-Hall (SRH) model is proposed, showing how time-constants extracted from
$I_{DS}$
-DLTS, i.e. under charge carrier emission, must be longer than the ones obtained from
$Y_{22}$
, i.e. under equilibrium and thus where both emission and much faster capture processes play a role.