Eric R. Heller

Short-Channel Effect Limitations on High-Frequency Operation of AlGaN/GaN HEMTs for T-Gate Devices

AlGaN/GaN high-electron mobility transistors (HEMTs) were fabricated on SiC substrates with epitaxial layers grown by multiple suppliers and methods. Devices with gate lengths varying from 0.50 to 0.09 mum were fabricated on each sample. We demonstrate the impact of varying the gate lengths and show that the unity current gain frequency response (fT) is limited by short-channel effects for all samples measured. We present an empirically based physical model that can predict the expected extrinsic fT for many combinations of gate length and commonly used barrier layer thickness (tbar) on silicon nitride passivated T-gated AlGaN/GaN HEMTs. The result is that even typical high-aspect-ratio (gate length to barrier thickness) devices show device performance limitations due to short-channel effects. We present the design tradeoffs and show the parameter space required to achieve optimal frequency performance for GaN technology. These design rules differ from the traditional GaAs technology by requiring a significantly higher aspect ratio to mitigate the short-channel effects.

3.8-MV/cm Breakdown Strength of MOVPE-Grown Sn-Doped

A Sn-doped (100) β-Ga₂O₃ epitaxial layer was grown via metal-organic vapor phase epitaxy onto a single-crystal, Mg-doped semi-insulating (100) β-Ga₂O₃ substrate. Ga₂O₃-based metal-oxide-semiconductor field-effect transistors with a 2-μm gate length (L_G), 3.4-μm source-drain spacing (L_SD), and 0.6-μm gate-drain spacing (L_GD) were fabricated and characterized. Devices were observed to hold a gate-to-drain voltage of 230 V in the OFF-state. The gate-to-drain electric field corresponds to 3.8 MV/cm, which is the highest reported for any transistor and surpassing bulk GaN and SiC theoretical limits. Further performance projections are made based on layout, process, and material optimizations to be considered in future iterations.

\beta

Sn-doped gallium oxide (Ga2O3) wrap-gate fin-array field-effect transistors (finFETs) were formed by top-down BCl3 plasma etching on a native semi-insulating Mg-doped (100) β-Ga2O3 substrate. The fin channels have a triangular cross-section and are approximately 300 nm wide and 200 nm tall. FinFETs, with 20 nm Al2O3 gate dielectric and ∼2 μm wrap-gate, demonstrate normally-off operation with a threshold voltage between 0 and +1 V during high-voltage operation. The ION/IOFF ratio is greater than 105 and is mainly limited by high on-resistance that can be significantly improved. At VG = 0, a finFET with 21 μm gate-drain spacing achieved a three-terminal breakdown voltage exceeding 600 V without a field-plate.

-Ga 2 O 3 MOSFETs

AlGaN/GaN high electron mobility transistor (HEMT) device operation was modeled from the sub-micrometer scale to the substrate using a combination of an electro-thermal device model for the active device with realistic power dissipation within the device and a coupled three dimensional thermal model to account for the substrate. Temperatures for various points within a device were determined as a function of biasing conditions, substrate thickness and temperature, number of fingers, and gate length and pitch. As an example, we have used our model to show that life test results of industry-relevant devices can be significantly affected by the exact testing technique used.

Enhancement-mode Ga2O3 wrap-gate fin field-effect transistors on native (100) β-Ga2O3 substrate with high breakdown voltage

A comparative analysis of the residual stress distributions across the conductive channel of Ga-face AlGaN/GaN high electron mobility transistors (HEMTs) is presented. Stress was measured by means of micro-Raman spectroscopy and micro-photoluminescence (PL). Raman measurements probed the volume average of the stress through the GaN layer whereas the stress near the GaN surface (AlGaN/GaN heterointerface) was acquired via PL. By combining Raman, PL, and x-ray diffraction, a self-consistent method was developed to accurately determine the variation in magnitude of stress throughout the thickness of the GaN layer. Based on this framework, it is observed in AlGaN/GaN HEMTs that a depth variation in the GaN residual stress occurs near the gate and ohmic electrodes. At these regions, the stress near the AlGaN/GaN interface (or GaN surface) exhibits a tensile shift compared to the stress averaged through the entire thickness of GaN. Across the conductive channel (away from the metal pads), the bulk average stres...

Electro-thermal modeling of multifinger AlGaN/GaN HEMT device operation including thermal substrate effects.

In this paper, we utilize micro-Raman spectroscopy to measure temperature and stress in state-of-the-art AlGaN/GaN HEMTs. A rigorous discussion on the physical accuracy, precision, and precautions for diverse Raman thermometry methods is developed. Thermometry techniques utilizing shifts in a single Raman Stokes peak position underpredict the channel temperature due to induction of operational thermoelastic stress in operating devices. Utilizing the change in phonon linewidth by employing a proper reference condition gives true temperature results. Making use of frequency shifts in both the E2(high) and A1(LO) phonon modes offers accurate and time-efficient means to determine the state of temperature and thermal stress in operating AlGaN/GaN HEMTs presuming that linear relations between phonon frequencies and temperature/stress are well determined. Useful applications of this method such as monitoring stress in GaN wafers between fabrication steps and Raman thermography on AlGaN/GaN HEMTs are demonstrated.

Analysis of the residual stress distribution in AlGaN/GaN high electron mobility transistors

We report on Sn-doped β-Ga2O3 MOSFETs grown by molecular beam epitaxy with as-grown carrier concentrations from 0.7 × 1018 to 1.6 × 1018 cm−3 and a fixed channel thickness of 200 nm. A pulsed current density of >450 mA/mm was achieved on the sample with the lowest sheet resistance and a gate length of 2  μm. Our results are explained using a simple analytical model with a measured gate voltage correction factor based on interface charges that accurately predict the electrical performance for all doping variations.

Thermometry of AlGaN/GaN HEMTs Using Multispectral Raman Features

The thermal response of AlGaN/GaN high electron mobility transistors directly correlates with the overall performance and reliability of these devices. In general, a hot spot develops near the drain end of the gate electrode during power dissipation. The device channel temperature was examined via micro-Raman spectroscopy under various bias conditions where power dissipation levels were identical. Under these bias conditions, difference in internal states (sheet carrier density and electric held distribution) within the device alters the heat generation profile across the channel. High Vds conditions lead to significantly higher channel temperature compared to that for low Vds conditions although the power dissipation is kept constant. Experimental results show ~13°C deviation between Vds = 45 V and Vds = 7 V cases when the power dissipation is 4.5 W/mm. This suggests that bias conditions may have a relatively signihcant impact on device reliability and that this effect must be considered when building thermal models of devices under operation or undergoing accelerated life testing.

High pulsed current density β-Ga2O3 MOSFETs verified by an analytical model corrected for interface charge

Abstract Understanding the distribution of the considerable heat generated in the active region of high power AlGaN/GaN high electron mobility transistors (HEMTs) at the sub-micron length scales relevant to the failures being observed in these devices is crucial for understanding device performance and reliability. In addition, electrical bias conditions and structural characteristics such as field plates alter the electric field distribution and thermal path within the device leading to changes in the heat generation profile across the channel. This in turn influences the value and location of the device peak temperature and the channel to ambient (or case or base-plate) thermal resistance. The channel temperature distribution of AlGaN/GaN HEMT structures with and without source connected field plates were examined via micro-Raman spectroscopy and coupled electro-thermal simulation. For both type of structures, high V ds conditions lead to significantly higher channel temperature compared to that for low V ds conditions for the same power dissipation level . This is important because the industry standard Arrhenius relation assumes the total power is sufficient to describe the device channel temperature and that the bias condition is irrelevant [1] . We explore the level of agreement between modeling and experiment, and also the extent to which variability in input parameters for the modeling affects model results. We show that operating bias condition has a significant role in device reliability by altering value and location of the peak temperature, which then alters the type and rate of thermally induced degradation taking place at critical locations such as the drain side corner of the gate. Specifically, care must be taken when extrapolating results of an accelerated life test to usage conditions at dissimilar bias conditions to consider if the results will be applicable.

The Impact of Bias Conditions on Self-Heating in AlGaN/GaN HEMTs

Coupled electro-thermo-mechanical simulation and Raman thermometry were utilized to analyze the evolution of mechanical stress in AlGaN/GaN high electron mobility transistors (HEMTs). This combined analysis was correlated with electrical step stress tests to determine the influence of mechanical stress on the degradation of actual devices under diverse bias conditions. It was found that the total stress as opposed to one dominant stress component correlated the best with the degradation of the HEMT devices. These results suggest that minimizing the total stress as opposed to the inverse piezoelectric stress in the device is necessary in order to avoid device degradation which can be accomplished through various growth methods.