Silicon carbide and GaN power MOSFETs are very promising candidates for space power applications requiring high efficiency, fast switching and reduced mass and volume. However, to meet those requirements, such technologies have also to demonstrate a good tolerance to space radiations, in particular regarding Heavy Ion-induced Single Event Effect and Total Ionizing Dose effects. From their material properties point of view, SiC and GaN technologies are superior compared to Silicon based Power devices. Both technologies exhibit indeed higher breakdown electrical field and higher carrier saturation velocity. This raises the question about the sensitivity to radiation of those technologies. Some studies have focused on TID effects, displacement damage, electron and heavy ions. However, one cannot consider that radiation is the only stressor during space missions. In fact, the switching operating of power devices acts as a background electrical stress which might affect the overall lifetime of the power system, through mutual effects with radiation constraints. The purpose of this study is therefore to assess the impact of electrical stress on the sensitivity to TID under x-ray irradiation. Both SiC and GaN technologies will be investigated. TCAD simulation will support experimental results as well as a compact model that takes into account the eventual drift of threshold voltage after irradiation due to the creation of traps in the gate oxide for SiC power devices and in the AlGaN layer or even at the AlGaN/GHaN interface when considering GaN power technologies.
The devices used in this study were commercial power transistors manufactured by CREE (SiC MOSFET) and EPC (GaN HEMT). SiC MOSFET is a vertical device whereas the GaN HEMT is a lateral device. Maximum drain to source ratings are respectively 900 V for SiC Power MOSFET and 200 V for GaN power HEMT. Looking at the drain to source current ratings, SiC devices exhibit a 23 A maximum drain to source current where GaN devices exhibit a 3 A current rating.
Characterizations before and at each irradiation step included IDSVG and IGSVG measurements, while VGS was swept from 0V to the maximum rated gate voltage. During characterizations VDS was held at 100mV with sources grounded.
At first the variability of the batch was investigated as well as the stability of the threshold voltage for each device. Then, a constant current stress (CCS) has been performed on a first batch of devices in order to weaken them and to investigate the impact of electrical stressing on TID sensitivity. Electrical stress consisted in a 100μA constant current applied during 100 s and resulting in a final injected charge of 10 mC for all the stressed devices. Such injected charge has been chosen in order to be far below the average charge to breakdown QBD found to be about 160 mC.
TID irradiation have then been performed at room temperature, using the on-site X ray source (X-RAD 320 x-ray cabinet) with 320 keV electron gun energy and current of 12.5 mA. Devices were all irradiated at 80 % of their maximum VDS rating while source was grounded. Maximum negative bias was applied to the gate to make sure that the transistor was off, even in case of a strong VTh shift. SiC and GaN devices were respectively irradiated till 1 MRad and 500 kRad with a dose rate of 14.75rad/s. Characterization of each device was performed before and after each irradiation step allowing to investigate any variability among the batch after electrical stress and during irradiation.
Experimental results obtained after electrical stress and subsequent TID irradiations will be presented (see attached file). Results will be analyzed through TCAD simulation, using a dedicated TCAD tool allowing to consider the trapping dynamics. Then, we will detail a compact model that takes into account TID effects on static electrical characteristics.