Accurate and Efficient Behavioral Modeling of GaN HEMTs Using An Optimized Light Gradient Boosting Machine

Purpose Gallium nitride high electron mobility transistors (GaN HEMTs) are representative wide-bandgap semiconductor devices characterized by their superior high-frequency switching characteristics. This paper aims to review their modeling methods and drive control technologies, systematically summarize the research progress and predict the future development trends. Design/methodology/approach This paper first introduces the background and advantages of GaN HEMTs and then reviews four modeling methods: physical model, behavioral model, analytical model and hybrid model. It also summarizes their respective characteristics and application scopes. Furthermore, key drive control techniques are discussed in depth. Finally, the development trend of modeling technology is predicted by combining the characteristics of typical devices. Findings Switching modeling methods for GaN HEMTs have advanced to better address high-frequency and high-power application requirements. Each model type offers distinct advantages regarding accuracy, complexity and physical interpretability. Integrating appropriate drive control strategies further enhances system performance. Originality/value This paper summarizes four switching models for GaN HEMTs and compares their characteristics and applicable scenarios. Subsequently, it further explores the driving characteristics and key technologies of GaN HEMTs, including dead time control, segmented driving and level shifting techniques. Finally, future development trends are discussed in light of practical device effects, with a particular focus on multiphysics coupling, emerging artificial intelligence (AI) modeling and adaptive driving technologies.

Resonant driving scheme for p-doped gallium nitride high electron mobility transistor to reduce driving power loss

The Gallium Nitride high electron mobility transistor (GaN HEMT) has been considered as a potential power semiconductor device for high switching speed and high power density application since its commercialization. Compared with the traditional Si transistors, GaN HEMT has faster switching speed and lower on-off loss. As a result, it is more sensitive to the nonlinear parameters due to the fast switching speed. The subsequent voltage and current overshooting will affect the efficiency and safety of the GaN HEMT and power electronic systems. In this paper, an accurate switching transient analytical model for GaN HEMT is proposed, which considers the effects of parasitic inductances, nonlinear junction capacitances and nonlinear transconductance. The model characteristic of turn-ON process and turn-OFF process is illustrated in detail, and the equivalent circuits are derived for each switching transition. The accuracy of the proposed model can be verified by comparing the predicted switching waveform and switching loss with that of the experimental results based on the double pulse test (DPT) circuit. Compared with the conventional model, the proposed model is more accurate and matches better with the experimental results than the conventional model. Finally, this model can be used for analyzing the influences of gate resistance, nonlinear junction capacitances, and parasitic inductances on switching transient waveform and refining calculation switching loss.

The self-heating effect has long been a persistent issue for high electron mobility transistors based on gallium nitride due to their inherently poor heat dissipation capability. Although a wide variety of thermal management solutions has to date been proposed, the problem of the extremely non-uniform heat dissipation at the micrometer scale is still challenging. It has recently been demonstrated, however, that the performance of gallium nitride high electron mobility transistors can be substantially improved by the introduction of various heat-spreading elements based on graphene, boron nitride or diamond. In this paper, using numerical simulation, we carried out a design optimization procedure for a normally-off gallium nitride high electron mobility transistor containing both graphene and cubic boron nitride layers. First, a screening experiment based on a very economical Plackett-Burman design was performed in order to find the most critical geometric parameters that influence the dc characteristics. After that, a full two-level factorial experiment consisting of three factors was implemented and an optimized parameter set was yielded. By applying this set, the output power was increased by 11.35%. The combination of the most significant parameters does not include any factors related to the heat-spreading layers. Keywords: gallium nitride, high electron mobility transistor, optimization, Plackett-Burman design, screening experiment, self-heating.

This article proposes a high-precision simulation method for transient electro-thermal coupling to characterize the self-heating effect in gallium nitride high electron mobility transistors (GaN HEMTs). The method is based on governing coupled drift-diffusion and heat conduction equations. A time scaling factor is introduced for the first time to accelerate the simulation process of electro-thermal coupling in GaN HEMTs. The technique addresses the cross-scale issue associated with the electrical and thermal response times encountered during the solution of the nonlinear coupling equations for GaN HEMTs. An accurate spectral element time-domain (SETD) method was utilized for the numerical realization of the nonlinear equations. The selection rules for the time scaling factor are discussed and exemplified through numerical simulations. By compared with commercial software, the advantages of the proposed method are demonstrated in terms of efficiency, accuracy, and random access memory (RAM) requirements. Based on its ability to accurately take into account electro-thermal coupling, this technique represents a powerful simulation tool for the design of high-performance devices and heat management.

Accurate thermal simulation of thermal is crucial for designing reliable gallium nitride high electron mobility transistors (GaN HEMTs). Thermal analysis models that assume a uniform heat source neglect the non-uniformity and bias dependence of the heat source in the channel region, which leads to inaccurate device junction temperature and other parameter predictions. Herein, we adopt a framework of a bias-dependent non-uniform heat source model and perform detailed thermal simulations of the near-junction region of GaN HEMTs. Results show that a uniform heat source model yields an optimal GaN thickness of 2.4 μm; whereas under bias-dependent non-uniform heat sources, it remains 2.4 μm at Vg = 2 V, Vd = 3.8 V but increases to 3.1 μm at Vg = −1 V, Vd = 6.25 V. Notably, under identical power dissipation, devices on a high-thermal-conductivity diamond substrate in the saturated region exhibit a junction temperature of 377.8 K, higher than the 360.37 K observed for a low-thermal-conductivity SiC substrate in the linear region. This difference arises from heat distribution and heat flow paths related to device bias. These results provide a theoretical reference for the thermal design of high-performance GaN-based devices.

The Gallium nitride high electron mobility transistor (GaN HEMT) has become popular in the power electronics industry as it offers the possibility to reach lower switching loss and higher switching frequency compared to Si devices. This paper adopts a normally-off GaN HEMT in the wireless power transfer (WPT) system to charge a 48V battery pack on the E scooter. 178W power was delivered to the battery pack with 813 kHz switching. A simulation model of GaN HEMT is also built in LTspice to itemize the system loss. At the end some thoughts of improving the system efficiency were proposed.

The reconfigurable mixer plays an important role in both transmitter as well as receiver section in the present wireless and digital communication systems. Also such mixer requires lot of power for up-converting and down-converting the signal without any noise. This paper deals about such type of designing of reconfigurable mixer for wide band communication applications with Gallium Nitride High Electron Mobility Transistor (GaN HEMT). This GaN HEMT transistor solves the requirement of huge power over wide range without any noise. The mixer is designed in the present paper for variable gain conversion along with power consumption trade-off for a wide range of frequency (2–8 GHz) with reconfigurable concept. Applied Wave Research (AWR) simulator was used to design the reconfigurable mixer. The gain is controlled by changing the trans-conductance of the input RF stage, and the current steering in trans-impedance stage. A broadband down-converter mixer is designed that provide different conversion gain, ranging from 25–33.5 dB with two different modes. The large bandwidth of the mixer is achieved by shunting the trans-conductance stage and with various switching stages. The results also show that there is good isolation along with conversion gain of designed reconfigurable mixer, for wide range in various communication applications.

LLC resonant converter is concerned for its inherent soft switching features. What's more, Gallium Nitride high electron mobility transistors (GaN HEMT) has smaller on-state resistor and much smaller gate charge and junction capacitors compared with the silicon MOSFET. In this paper, a LLC resonant converter with high efficiency based on 600V GaN HEMT is presented. Controller design procedure is discussed based on its transfer function model that is derived from extended describing function (EDF) method. A 240W, 380V to 12V LLC converter using GaN transistors is demonstrated in this paper which achieves more than 93% efficiency without Synchronous Rectifier. Experiment results are provided to validate theoretical design and analysis.

Short-circuit protection (SCP) of gallium nitride high electron mobility transistors (GaN HEMTs) is critical to the reliability of power electronic systems. The conventional desaturation-based SCP method requires a properly configured blanking time. If the blanking time is too long, the power loss of the GaN HEMT increases during a short-circuit (SC) event. Conversely, if it is too short, false triggering of SCP may occur. To address this issue, an adaptive blanking time generation circuit is proposed. The blanking time is determined by detecting dVDS/dt, which enables adaptive activation or deactivation of VDS monitoring. With the 650 V/30 A GaN HEMT, the proposed method is validated on the laboratory test bench. Experimental results verify that effective protection is achieved under both hard switching fault (HSF) and fault under load (FUL) conditions by the proposed method. The results show that the total SCP time of the proposed method is 130 ns during HSF and 66 ns during FUL.

This paper presents a design and fabrication of Ku-band power amplifier using Gallium Nitride High Electron Mobility Transistor (GaN HEMT) die. In order to fabricate the low-cost Ku-band power amplifier, a Printed Circuit Board(PCB) was used for input/output matching circuits instead of manufacturing process to use an expensive substrate. The measured output power is 42.6 dBm, the drain efficiency is 37.7 % and the linear gain is 7.9 dB under pulse operation at the frequency of 14.8 GHz. Under the continuous wave(CW) test, the output power is 39.8 dBm, the drain efficiency is 24.1 % and the linear gain is 7.2 dB.Key words: Ku-Band, GaN HEMT Die, Low-Cost, Power Amplifier LMN62013OPQRSA8MN#(TU$%MNVIW.X YZ [\]^[ (Department of Wireless Communications Engineering, Kwangwoon University) *QRSA8(Samsung Thales) Manuscript received February 4, 2014 ; Revised March 11, 2014 ; Accepted April 16, 2014. (ID No. 20140204-016)Corresponding Author: Jin-Joo Choi (e-mail: jinchoi@kw.ac.kr)

In this article, a 1.7–2.6-GHz broadband power amplifier (PA) was proposed using the generalized quasi-Chebyshev lowpass matching approach, which was directly composed of step-impedance transmission lines (SITLs). First, the complete design theory for the introduced step-impedance matching network was provided with the detailed design equations. The impedance of the SITLs can be calculated directly based on the specified bandwidth, maximum allowable return loss, and impedance-transforming ratio. The electrical length of SITLs was utilized to adjust the harmonic impedances to perform a PA high-efficiency region. For demonstration, a PA using the 10-W gallium nitride high electron mobility transistor (GaN HEMT) with the measured drain efficiency (DE) of 65%–78% from 1.7 to 2.6 GHz was achieved. A good agreement between the simulated and measured results can be observed.

Gallium nitride high electron mobility transistors (HEMTs) have shown the potential to be highly resistant to radiation damage, making them ideal for use in microwave power amplifiers and DC/DC converters in space-based applications. To investigate the mechanisms of radiation-induced degradation in AlGaN/GaN HEMTs, 2MeV protons were used to simulate the space radiation environment, and to probe the second order effect of inherent material defects on the response the initial material quality was intentionally varied by studying HEMTs on sapphire, Si, and SiC substrates. Irradiation was performed incrementally up to a fluence of 6x1014 cm-2. The Hall mobility and 2DEG density was measured on Van der Pauw structures before irradiation and at each dose. The decrease in mobility can be attributed to increased carrier scattering in the 2DEG as a result of radiation-induced defects while the decrease in sheet carrier density is attributed to screening of the 2DEG from charged trap formation [1]. The magnitude of change in the 2DEG density is the same for all substrate materials (1x1012 cm-2), though the percent change varies due to different initial values. Note that mobility begins to degrade at lower fluence than sheet carrier density. FET I-V analysis indicated increased ON-resistance and decreased saturation current, consistent with the reduced mobility and sheet carrier density observed in the Hall measurements. The off-state leakage current decreased and threshold voltage was shifted slightly positive, again due to the reduced 2DEG density. The trapping mechanisms were further probed using pulsed I-V measurements at varying quiescent bias conditions. While the dynamic ON-resistance degradation under quasistatic conditions was consistent with DC measurements, the degradation under off-state stress conditions initiated at a lower fluence, and a culminated with a 10X higher degradation at the final fluence. This implies that dynamic ON-resistance, a parameter related to charge trapping, is much more sensitive to radiation than DC characteristics and must be considered when evaluating the radiation hardness of HEMT devices. TEM imaging was employed to directly probe the mechanisms suggested by the electrical measurements. Process-induced defects were visible in both reference and irradiated samples, notably an overflow of Au in the metal stack at the gate edge, possibly related to PECVD deposition of passivation layers. In addition a new, radiation-induced void at both edges of the gate in the Ni region of the Ni/Au gate metallization was revealed [2]. EDS line scans indicate that, after radiation, the 20 nm Ni layer diffused up into the Au, and to a lesser extent, into the AlGaN, leaving voids protruding ~300 nm under the gate edges. The mechanism of the void formation shows all features of Ni/Au inter-diffusion through vacancy exchange, known as the Kirkendall effect. The exact process is still under investigation, but the voids result in reduced gate area, which is responsible for the degraded off-state performance. It is particularly rare that such pronounced diffusion occurs at room temperature. Since the voids only occur at the gate edges, an additional electrochemical or strain-induced driving force is likely present. Therefore the role of the SiNx passivation layer, as well as strain, at the gate edge will be factor in this ongoing research. 1. T.J. Anderson, et. Al. IEEE Electron Dev. Lett. 35 (8), 826-828 (2014) 2. A.D. Koehler, et. Al. IEEE Electron Dev. Lett. 35 (10) (2014)

The motivation of this work is the characterization of semiconductor switches in DC/DC Wide Band-Gap based converters for electrical vehicle applications. The paper studies Gallium Nitride High Electron Mobility (GaN HEMT) Transistors, which present very good features for high-frequency high efficiency applications. In order to obtain an efficient DC/DC converter design, a full characterization of these devices must be carried out. A double pulse test is applied to the GaN device, in order to study the switching behavior of these transistors. The test will be performed experimentally and in simulation under different operating conditions, in order to have the ability to understand the difference in the transistor behavior under these different operating points. Finally, the obtained results will be presented, commented and compared to identify the particularities of these devices.

Accurately and efficiently modeling the drain-lag effects is crucial in nonlinear large-signal modeling for Gallium Nitride high electron mobility transistors. In this paper, a simplified yet accurate drain-lag model based on an industry standard large-signal model, i.e., the Chalmers (Angelov) model, extracted by means of pulsed S-parameter measurements, is presented. Instead of a complex nonlinear drain-lag description, only four constant parameters of the proposed drain-lag model need to be determined to accurately describe the large impacts of the drain-lag effects, e.g., drain-source current slump, typical kink observed in pulsed IV curves, and degradation of the output power. The extraction procedure of the parameters is based on pulsed S-parameter measurements, which allow to freeze traps and isolate the trapping effects from self-heating. It is also shown that the model can very accurately predict the load pull performance over a wide range of drain bias voltages. Finally, the large-signal network analyzer measurements at low frequency are used to further verify the proposed drain-lag model in the prediction of the output current in time domain under large-signal condition.