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

degradation in the device characteristics

extreme temperature free carrier mobility decreases

Si GaN

Advantages of GaN

Properties Si GaAs GaN Energy Gap (eV) 1.12 1.43 3.5 Hole Mobility (cm2Vs-1) 600 400 200 Electron Mobility (cm2Vs-1) 1400 8500 1250 Breakdown Voltage (×106

6 Vcm-1)

0.3 0.4 3 Thermal Conductivity(W cm-1 K-1) 1.5 0.5 1.3 Saturation Drift Velocity (×107 cm s-1) 1 2 2.7 high temperature, high power applications

high breakdown voltage wide band gap high saturation velocity

Information for Physical properties of Si, GaAs and GaN

Current Issue

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

device

• peratio

n at high temperat ure

efficient thermal management improve device reliability

limited numerical analysis inefficient device modeling

indepth study of device operation at high temperature

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

• perating 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 13cm-2)

polarisation in AlGaN and GaN layers using

an un-doped hetero-interface

energy band tilts towards interface

triangular quantum well electrons are confined due to

• 1. large polarisation difference and
• 2. large conduction band difference

increased electron mobility high frequency and high power devices

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

Output characteristics (Id-Vd) of GaN HEMTs at T=473K Diagram of a Schottky-contact gate

knee voltage current decreases due to self-heating current saturates lattice heating due to inefficient heat dissipation additional phonon scattering degrades eletron mobility pinch-off voltage= -3V

• depletion region modulates
• more than proportionate change in Id

High Temperature Performance of AlGaN/GaN HMETs on Si Substrates

Id-Vd graph at Vg=1V for temperatures ranging from 273K to 473K

• Id shown in the curve is

normalised.

• Id characteristics of

AlGaN/GaN HEMTs are inversely proportional to temperature.

473K 423K 373K 323K 298K 273K

where T0=273K B is temperature coefficient

Linear relationship between Id and Temperature

)] ( 1 [

) ( ) (

T T B P P

T T

  

Explanation of current degradation

lattice scatteri ng& impurit y scatteri ng

5 . 1 

T I

 E v 

nevA I 

High Temperature Performance of AlGaN/GaN HMETs on Si Substrates

Id-Gm graph for temperatures from 273K to 473K at Vd=1V

• After threshold value (Vg= -3V),

Gm increases exponentially

• After Vg reaches around -1V, it

starts to decrease, which is called Gm collapse

Vg Id Gm   

Linear relationship between Gm and Temperature

Gmmax-T graph

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 BP(T0)

• 2.91E-01

B

• 0.001070968

T0/K 273 Equation Gmmax=271.25[1-0.000107(T-273)]

T I T Vg I T Gm          / Vg Id Gm   

Conclusion and Future work

Conclusion

Device performance degradation with temperature.

• Due to

decreas ed device channel electron mobility Empirical models of AlGaN/GaN HEMTs device performance with increasing ambient temperature

• Used to

model device performa nce at high temperat ure

• peratin

g condition s, such

Recommendations for Future work

Develop a nonlinear model for the temperature dependence of GaN HEMTs D-C characteristics

• Using machine

learning method, Artificial Neural Network

Better predict HEMT electrical characteristics at higher temperatures

• To test the integrity
• f designed circuits

and to predict their performance before fabrication

Thank you!