Temperature accelerated Degradation of GaN HEMTs under High power - PowerPoint PPT Presentation

Temperature ‐ accelerated Degradation of GaN HEMTs under High ‐ power Stress: Activation Energy of Drain Current Degradation Yufei Wu, Chia ‐ Yu Chen and Jesús A. del Alamo Microsystems Technology Laboratory Acknowledgement: DRIFT ‐ MURI, TriQuint Semiconductor 1

Outline 1. Motivation 2. High ‐ power and high ‐ temperature stress experiments 3. An improved approach 4. Conclusions 2

Motivation • Activation energy, E a : N. Malbert, IRPS 2010 essential in predicting lifetime • Conventionally: high temperature accelerated life test 3

Motivation • Activation energy, E a : N. Malbert, IRPS 2010 essential in predicting lifetime • Conventionally: high temperature accelerated life test Problems: • Requires multiple devices • Carrier trapping not properly dealt with • Different degradation mechanisms can emerge at different temperatures 4

Motivation • Activation energy, E a : N. Malbert, IRPS 2010 essential in predicting lifetime • Conventionally: high temperature accelerated life test Desirable: E a extraction from T j measurements on a single device Step ‐ temperature stress time 5

Outline 1. Motivation 2. High ‐ power and high ‐ temperature stress experiments 3. An improved approach 4. Conclusions 6

Setup for DC reliability studies Accel ‐ RF System Devices: Prototype GaN Power Amplifier Hardware DC/Pulsed MMIC from industry Characterization DUT Switching Matrix ‐ KeithleySources ‐ Agilent B1500A Heater RF/DC Units Accel ‐ RF AARTS RF10000 ‐ 4/S system T base Windows ‐ based PC MIT RF/DC Accel ‐ RF Software Characterization Suite ‐ RF measurement ‐ DC FOMs ‐ Temperature control ‐ Current collapse ‐ Stressing Augmented with: • external instrumentation for DC/pulsed characterization • software to control external instrumentation and extract DC FOMs 7

High ‐ power DC Experiment Flowchart • Detrapping : T base = 250 °C for 7.5 hours Start • Full characterization Detrapping o At T base = 50 °C o Full DC I ‐ V sweep Full Characterization o Current collapse DC and Temperature Stress Inner loop Short Characterization (DC) End: detrapping + full characterization 8

High ‐ power DC Experiment Flowchart • Detrapping : T base = 250 °C for 7.5 hours Start • Full characterization Detrapping o At T base = 50 °C o Full DC I ‐ V sweep Full Characterization o Current collapse • Stress: DC and Temperature Stress Inner loop o High ‐ power condition o Base temperature stepped up Short Characterization (DC) • Short characterization o Every 30 minutes at T base = 50 °C End: detrapping + o DC FOMs: I Dmax , I Goff , R D , R S , V T , … full characterization 9

High ‐ power DC Experiment Flowchart • Detrapping: T base = 250 °C for 7.5 hours Start • Full characterization Detrapping o At T base = 50 °C o Full DC I ‐ V sweep Full Characterization o Current collapse • Stress: DC and Temperature Stress Inner loop o High ‐ power condition o Base temperature stepped up Short Characterization (DC) • Short characterization o Every 30 minutes at T base = 50 °C End: detrapping + o DC FOMs: I Dmax , I Goff , R D , R S , V T , … full characterization 10

Definitions of Various Figures of Merit Parameter Definition I Dmax I D at V GS =2 V, V DS =5 V I Goff I G at V GS = ‐ 5 V, V DS =0.1 V R D Drain resistance measured with I G = 20 mA/mm R S Source resistance measured with I G = 20 mA/mm V T V GS – 0.5V DS when I D = 1 mA/mm at V DS = 0.1 V Current Collapse Percentage change in I Dmax after 1 sec. V DS = 0 V, V GS = ‐ 10 V pulse 11

High ‐ power DC Experiment High ‐ power stress: V DS = 40 V, I D = 100 mA/mm, T base = 50 °C – 230 °C, 600 min/step outer loop data outer loop data 1.1 120 50 140 160 180 1.0 1 90 130 150 170 0.9 220 0.8 190 |I Goff | (mA/mm) 190 I Dmax /I Dmax (0) 0.7 0.1 180 200 0.6 0.5 210 50 0.4 0.01 0.3 170 T base (°C) = 220 0.2 160 0.1 1E-3 0 5000 10000 15000 0 5000 10000 15000 time (min) time (min) 12

High ‐ power DC Experiment High ‐ power stress: V DS = 40 V, I D = 100 mA/mm, T base = 50 °C – 230 °C, 600 min/step outer loop data outer loop data 1.1 120 50 140 160 180 1.0 1 90 130 150 170 0.9 220 0.8 190 |I Goff | (mA/mm) 190 I Dmax /I Dmax (0) 0.7 0.1 180 200 0.6 0.5 210 50 0.4 0.01 0.3 170 T base (°C) = 220 0.2 160 0.1 1E-3 0 5000 10000 15000 0 5000 10000 15000 time (min) time (min) • |I Goff | increases from T base =170 to 190°C; then saturates • Significant I Dmax degradation for T base > 180 °C • Thermally activated I Dmax degradation rate shown 13

High ‐ power DC Experiment High ‐ power stress: V DS = 40 V, I D = 100 mA/mm, T base = 50 °C – 230 °C, 600 min/step 3.00 230 2.75 2.50 220 2.25 R D R/R(0) 2.00 1.75 210 1.50 200 T base (°C)= 1.25 180190 50 160 120 140 1.00 R S 150 170 90 130 0.75 0 5000 10000 15000 time (min) • R D increases significantly, consistent with I Dmax decrease • R S increases much less 14

Activation Energies of Degradation Rates Inner loop data Outer loop data (device detrapped) 12 12 10 10 R D : E a =0.91 eV ln(1/|slope|) 8 R D : E a =1.00 eV ln(1/|slope|) 8 6 6 4 4 I Dmax : E a =0.94 eV I Dmax : E a =0.58 eV 2 2 25 26 27 28 29 30 31 25 26 27 28 29 30 31 -1 ) 1/kT channel (eV -1 ) 1/kT channel (eV T channel obtained from thermal model of MMICs 15

Activation Energies of Degradation Rates Inner loop data Outer loop data (device detrapped) 12 12 10 10 R D : E a =0.91 eV ln(1/|slope|) 8 R D : E a =1.00 eV ln(1/|slope|) 8 6 6 4 4 I Dmax : E a =0.94 eV I Dmax : E a =0.58 eV 2 2 25 26 27 28 29 30 31 25 26 27 28 29 30 31 -1 ) 1/kT channel (eV -1 ) 1/kT channel (eV T channel obtained from thermal model of MMICs • Inner loop data : Large difference between E a for I Dmax and R D 16

Activation Energies of Degradation Rates Inner loop data Outer loop data (device detrapped) 12 12 10 10 R D : E a =0.91 eV ln(1/|slope|) 8 R D : E a =1.00 eV ln(1/|slope|) 8 6 6 4 4 I Dmax : E a =0.94 eV I Dmax : E a =0.58 eV 2 2 25 26 27 28 29 30 31 25 26 27 28 29 30 31 -1 ) 1/kT channel (eV -1 ) 1/kT channel (eV T channel obtained from thermal model of MMICs • Inner loop data : Large difference between E a for I Dmax and R D • Outer loop data : Thermally activated behavior 17

Activation Energies of Degradation Rates Inner loop data Outer loop data (device detrapped) 12 12 10 10 R D : E a =0.91 eV ln(1/|slope|) 8 R D : E a =1.00 eV ln(1/|slope|) 8 6 6 4 4 I Dmax : E a =0.94 eV I Dmax : E a =0.58 eV 2 2 25 26 27 28 29 30 31 25 26 27 28 29 30 31 -1 ) 1/kT channel (eV -1 ) 1/kT channel (eV T channel obtained from thermal model of MMICs • Inner loop data : Large difference between E a for I Dmax and R D • Outer loop data : Close E a values for I Dmax and R D  common physical origin 18

Conclusions Drawn from the Experiment 1 • I G degradation: 220 |I Goff | (mA/mm) 190 o Increases fast at first 0.1 180 o Eventually saturates 50 0.01 170 • I D degradation: 160 1E-3 o Significant degradation only 0 5000 10000 15000 time (min) after I G degradation is 1.1 saturated 120 50 140 160 180 1.0 90 130 150 170 0.9 o Thermally activated 0.8 190 I Dmax /I Dmax (0) 0.7 200 0.6 0.5 210 0.4 0.3 T base (°C) = 220 0.2 0.1 0 5000 10000 15000 19 time (min)

Conclusions Drawn from the Experiment • I G degradation: o Increases fast at first o Eventually saturates • I D degradation: o Significant degradation only after I G degradation is saturated o Thermally activated • Desirable: separate I G and I D degradation • Key idea: short stress to degrade I G without I D degradation, then long stress to degrade I D 20

Outline 1. Motivation 2. High ‐ power and high ‐ temperature stress experiments 3. An improved approach 4. Conclusions 21

DC Experiment : Improved Approach  Phase 1 : degrade I G without significant I D degradation • Short stress period o T base = 50 ‐ 220 °C, in 20 °C steps o Stress time: 6 minutes 22

DC Experiment : Improved Approach  Phase 1 : degrade I G without significant I D degradation • Short stress period o T base = 50 ‐ 220 °C, in 20 °C steps o Stress time: 6 minutes  Phase 2 : degrade I D without further I G degradation • Longer stress period o T base : from 120 °C, increase in steps o Stress time: 24 hours 23

A Typical Experiment (Phase 2) High ‐ power stress: V DS = 40 V, I D = 100 mA/mm, T base = 120 °C – 215 °C, 24 hours/step After detrapping After detrapping 1.00 1  I Dmax 0.95  |I Goff | 120 150 170 205 210 215 |I Goff | (mA/mm) 120 150 170 185 200 185 0.90 I Dmax /I Dmax (0) 0.85 200 T base (°C) 0.1 0.80 205 0.75 210 0.70 T base (°C) = 215 0.01 0.65 0 5000 10000 15000 0 5000 10000 15000 time (min) time (min) During phase 1: |I Goff |increases by 2 orders of magnitude; I Dmax decreases by 3% 24