In the last few years, the development of energy-efficient electrical power management systems has received a great deal of interest. Recently, GaN Field-Effect-Transistors have emerged as a disruptive technology with great potential. The high breakdown electric field coupled with the high sheet electron density attainable in GaN heterostructures promises better than three orders of magnitude improvement in the ON-resistance/breakdown-voltage trade-off over conventional Si power-switching devices. However, in spite of recent great progress in GaN power device fabrication technologies, electrical reliability represents a great unknown in these devices. In particular, the dynamic ON-resistance (R_ON), in which the R_ON of the transistor remains high for a certain period of time after an OFF-ON switching event, is a big concern ^[1] . This phenomenon greatly affects the efficiency of electrical power management circuits based on GaN FETs.

Towards the goal of fundamental understanding on the dynamic ON resistance problem, we have developed a new dynamic R_ON measurement methodology that allows the observation of R_ON transients after an OFF-ON switching event from 200 ns to any arbitrary length of time over many decades in time ^[2] . Using this technique, we have performed a fundamental study of dynamic R_ON on high-voltage GaN FETs. Figure 1 shows two examples of R_ON transients obtained over 11 decades in time when the devices are switched from an OFF-state with V_GSQ=-5 V and V_DSQ=40 V. They are measured by a pulsed IV system (blue lines) and semiconductor device analyzer (red lines) on industrially prototyped AlGaN/GaN HEMTs fabricated on two nominally identical epitaxial wafers (labeled A and B) grown by two different commercial epitaxial vendors. In wafer A, R_ON at 200 ns is about 73% higher than in DC. For wafer B, R_ON (200 ns) is about 52% higher than R_{ON_DC}. Such a high dynamic R_ON represents a big problem for power-switching applications. The pattern of recovery of R_ON is also very different in both wafers. Wafer A exhibits a very slow transient in the s-ks time range, while wafer B recovers on a µs-s time scale. The very different transients obtained in these two wafers suggest that the details of epitaxial growth have a great impact on the dynamics of GaN devices. Figure 2 shows OFF to ON transients from different quiescent V_DS values for wafer B. As V_DSQ increases, the dynamic R_ON increases in a prominent manner in both devices, presumably as a result of the increased extension of the high-field region into the drain.

We have carried out additional studies at different temperature and pulse conditions. From all these results, we postulate that on a short time scale, dynamic R_ON arises from the fast release of electrons from border traps at the AlN/AlGaN interface right above the GaN channel through a tunneling process. In contrast, over a longer time scale, conventional thermally assisted detrapping processes from traps in the AlGaN or at the surface dominate. All of these findings provide a path for proper GaN power-switch device engineering and operation that minimize dynamic R_ON.

1. J. A. del Alamo and J. Joh, “GaN HEMT reliability,” Microelectronics Reliability, vol. 49, pp. 1200-1206, September 2009.
2. D. Jin and J. A. del Alamo, “Mechanisms responsible for dynamic ON-resistance in GaN high-voltage HEMTs,” International Symposium on Power Semiconductor Devices and ICs Technical Digest, June 2012.