β-Ga2O3 thin-channel metal-oxide-semiconductor field-effect transistors (MOSFETs) were evaluated using both DC and pulsed I-V measurements. The reported pulsed I-V technique was used to study self-heating effects in the MOSFET channel. The device was analyzed over a large temperature range of 23 to 200°C. A relationship between dissipated power and channel temperature was established, and it was found that the MOSFET channel was heating up to 208°C when dissipating 2.5 W/mm of power. The thermal resistance of the channel was found to be 73°C-mm/W. The results are supported with experimental Raman nano-thermography and thermal simulations and are in excellent agreement with pulsed I-V findings. The high thermal resistance underpins the importance of optimizing thermal management in future Ga2O3 devices.

The advent of acceptor-type doping of β-Ga2O3 through ion-implantation of nitrogen has opened a new design space for junction-type devices with estimated breakdown voltages in excess of a few kVs. However, the presence of deep states due to intrinsic defects in β-Ga2O3 and implantation damage could be detrimental to the performance and reliability of such devices. We give a phenomenological description and experimental demonstration of the effects of nitrogen implantation in a buried blocking layer on the performance of transistors. The partial activation of acceptor-like states in the buried implanted region has been revealed and estimated to be ∼20% through a junction spectroscopic technique involving substrate-bias and sub-bandgap illumination, which remains elusive to standard characterization techniques. The characterization technique, along with a space-charge model of the channel and band model of the buried implanted layer, has revealed the presence of photosensitive mid-bandgap (∼2.47 eV below the conduction band) and tail states near the valence band edge of nitrogen-implanted β-Ga2O3.

No abstract

The drain current temperature dependence is an efficient way to determine the channel temperature in semiconductor devices; however, it has been challenging to use due to the potential interference of trapping effects. A trapping tolerant method is proposed, illustrated here for Ga2O3 MOSFETs, making in situ temperature measurements possible, allowing a thermal resistance of 59  K·mm/W to be measured in Ga2O3 MOSFETs. However, neglecting the effect of trapping causes an error of ∼15% in the channel temperature measured using the drain current. 3D simulations show that the measured channel temperature is the average temperature value between source and drain contact.

No abstract