Comparison of DC measurement methods to determine GaN HEMT reliability

In this paper we compare methods to determine the life-time of AlGaN/GaN HEMTs by accelerated DC life-tests. They all base on monitoring of both, I_DSS and I_DQ during the life-test, either at room temperature of at the elevated ambient temperature. The goal is to investigate whether interruptions of the life test to measure I_DSS at room temperature can be prevented and replaced by screening this parameter during the life test at higher temperature. The investigation shows that the so called intrinsic I_DSS measurement is a quite interesting and helpful tool, but too often measurement of this parameter adds additional stress to the device. In addition we saw that intermediate measurements did not stress the devices as suspected. We also observed that for stress tests at constant P_Diss the drift of the gate voltage can be used as failure criterion, although it has no linear relationship to I_DSS and therefore the gate voltage drifts more than I_DSS and leads to more conservative life-time.