EEE20B EEE20B-Temperature Dependent Electrical Performance of GaN - PowerPoint PPT Presentation

EEE20B EEE20B-Temperature Dependent Electrical Performance of GaN High Electron Mobility Transistors by Numerical Analysis Ge Shu

Background and Introduction

Problems with Si GaN Si free carrier extreme mobility temperature decreases degradation in the device characteristics

Advantages of GaN Properties Si GaAs GaN Energy Gap ( eV ) 1.12 1.43 3.5 Hole Mobility ( cm 2 Vs -1 ) 600 400 200 wide band Electron Mobility ( cm 2 Vs -1 ) 1400 8500 1250 gap Breakdown Voltage ( × 10 6 6 Vcm -1 ) 0.3 0.4 3 Thermal Conductivity ( W cm -1 K -1 ) 1.5 0.5 1.3 high high Saturation Drift Velocity ( × 10 7 cm s -1 ) 1 2 2.7 breakdown saturation voltage velocity high temperature, high power applications Information for Physical properties of Si, GaAs and GaN

Current Issue junction and electron mobility device channel & saturation performance temperature velocity degradation

efficient thermal management limited numerical analysis device indepth study of operatio device operation n at high at high temperat temperature ure inefficient device modeling improve device reliability

Aims and Objective To investigate the impact of temperature on AlGaN/ GaN HEMT device electrical D-C characteristics, drain current (Id) and transconductance (Gm); To build empirical models of Idmax and Gmmax with the external temperature, which can be used for device modeling at high temperature operating conditions.

Methodology and Materials

Materials AlGaN/GaN HEMTs grown on a Silicon substrate with dimensions: gate width (Wg)=80μm, gate length (Lg)=2μm, gate to drain length (Lgd)=4μm and gate to source length (Lgs)=2μm

Formation of 2DEG (10 13 cm -2 ) electrons are confined due to 1. large polarisation difference and 2. large conduction band difference polarisation in AlGaN and GaN layers using an un-doped hetero-interface increased electron mobility high frequency and high power devices triangular quantum well energy band tilts towards interface

Experiment Materials Photographs of Agilent B1505A power device analyzer and CASCADE MICROTECH Summit 11000M probe station , probing levels of up to 3,000V and 100W/cm2 with varying temperatures, from 273K to 473K .

Results and Discussion

3-Terminal Device D-C Measurements at 473K lattice heating due to knee current decreases inefficient heat dissipation voltage due to self-heating additional phonon scattering pinch-off voltage= -3V degrades eletron mobility current saturates Diagram of a Schottky-contact gate • Output characteristics (Id-Vd) of GaN HEMTs at T=473K depletion region modulates • more than proportionate change in Id

High Temperature Performance of AlGaN/GaN HMETs on Si Substrates • Id shown in the curve is normalised. • Id characteristics of AlGaN/GaN HEMTs are inversely proportional to 473K 423K temperature. 373K 323K 298K 273K Id-Vd graph at Vg=1V for temperatures ranging from 273K to 473K

Linear relationship between Id and Temperature    P P [ 1 B ( T T )] ( T ) ( T ) 0 0 where T 0 =273K B is temperature coefficient

Explanation of current degradation lattice scatteri ng& impurit y scatteri  T  1 . 5 I ng v   E I  nevA

High Temperature Performance of AlGaN/GaN HMETs on Si Substrates • After threshold value (Vg= -3V), Gm increases exponentially • After Vg reaches around -1V, it starts to decrease, which is called Gm collapse  Id  Gm  Vg Id-Gm graph for temperatures from 273K to 473K at Vd=1V

Linear relationship between Gm and Temperature Temp[k] Gmmax[mS/mm] 273 194.875 298 188.21875 323 173.2395833 373 163.8958333 423 134.8125 473 144.1145833 expression y = -0.2905x + 271.25 R^2 0.884 P(T0) 271.25 Gmmax-T graph BP(T0) -2.91E-01 B -0.001070968 273 T0/K  Id  Gm Equation Gmmax=271.25[1-0.000107(T-273)]  Vg     Gm I / Vg I      T T T

Conclusion and Future work

• Due to Conclusion • Used to Device decreas model performance ed device degradation device performa with channel Empirical nce at temperature. electron models of high mobility AlGaN/GaN temperat HEMTs device ure performance operatin with g increasing condition ambient s, such temperature

Recommendations for Future work Develop a nonlinear model for the • Using machine temperature learning method, dependence of GaN Artificial Neural Network HEMTs D-C characteristics • To test the integrity Better predict HEMT of designed circuits electrical and to predict their characteristics at performance higher temperatures before fabrication

Thank you!