In this paper, we present results from the simulations of a submicrometer AlGaN/GaN highelectron-mobility transistor (HEMT) by using an in-house electro-thermal Monte Carlo simulator. We study the temperature distribution and the influence of heating on the transfer characteristics and the transconductance when the device is grown on different substrates (sapphire, silicon, silicon carbide and diamond). The effect of the inclusion of a thermal boundary resistance (TBR) is also investigated. It is found that, as expected, HEMTs fabricated on substrates with high thermal conductivities (diamond) exhibit lower temperatures, but the difference between hot-spot and average temperatures is higher. In addition, devices fabricated on substrates with higher thermal conductivities are more sensitive to the value of the TBR because the temperature discontinuity is greater in the TBR layer.
In this contribution we present the results from the simulation of an AlGaN/GaN heterostructure diode by means of a Monte Carlo tool where thermal effects have been included. Two techniques are investigated: (i) a thermal resistance method (TRM), and (ii) an advanced electro-thermal model (ETM) including the solution of the steady-state heat diffusion equation. Initially, a systematic study at constant temperature is performed in order to calibrate the electronic model. Once this task is performed, the electro-thermal methods are coupled with the Monte Carlo electronic simulations. For the TRM, several values of thermal resistances are employed, and for the ETM method, the dependence on the thermal-conductivity, thickness and die length is analyzed. It is found that the TRM with well-calibrated values of thermal resistances provides a similar behavior to ETM simulations under the hypothesis of constant thermal conductivity. Our results are validated with experimental measurements finding the best agreement when the ETM is used with a temperature-dependent thermal conductivity.
In this work, we report on Monte Carlo simulations to study the capability to generate Gunn oscillations of diodes based on InP and GaN with around 1 lm active region length. We compare the power spectral density of current sequences in diodes with and without notch for different lengths and two doping profiles. It is found that InP structures provide 400 GHz current oscillations for the fundamental harmonic in structures without notch and around 140 GHz in notched diodes. On the other hand, GaN diodes can operate up to 300 GHz for the fundamental harmonic, and when the notch is effective, a larger number of harmonics, reaching the Terahertz range, with higher spectral purity than in InP diodes are generated. Therefore, GaN-based diodes offer a high power alternative for sub-millimeter wave Gunn oscillations. V
In this work, we report on Monte Carlo simulations of InP and GaN vertical Gunn diodes to optimize their oscillation frequency and DC to AC conversion efficiency. We show that equivalent operating conditions are achieved by the direct application of a sinusoidal AC voltage superimposed to the DC bias and by the simulation of the intrinsic device coupled with the consistent solution of a parallel RLC resonant circuit connected in series. InP diodes with active region about 1 µm offer a conversion efficiency up to 5.5 % for frequencies around 225 GHz. By virtue of the larger saturation velocity, for a given diode length, oscillation frequencies in GaN diodes are higher than for InP structures. Current oscillations at frequencies as high as 675 GHz, with 0.1 % efficiency, are predicted at the sixth generation band in a 0.9 µmlong GaN diode, corroborating the suitability of GaN to operate near the THz band. At the first generation band, structures with notch in general provide lower oscillation frequencies and efficiencies in comparison with the same structures without notch.However, a higher number of generation bands are originated in notched diodes, thus typically reaching larger frequencies. Self-heating effects reduce the performance, but in GaN diodes the efficiency is not significantly degraded.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.