Time Evolution of Electrical Degradation under High-Voltage Stress - - PowerPoint PPT Presentation

Time Evolution of Electrical Degradation under High-Voltage Stress in GaN HEMTs

Jungwoo Joh and Jesús A. del Alamo Microsystems Technology Laboratories, MIT

Acknowledgements: ARL (DARPA WBGS program) ONR (DRIFT-MURI) TriQuint Semiconductor

2/20

Purpose

• GaN HEMT Reliability: big concern

– RF power degradation – ID↓, RD↑, IG↑, ∆VT…

• Goal: understand degradation mechanism

RF stress 10 GHz,VD=28 V IDQ=150 mA/mm Pin=23 dBm Pout=33.7 dBm

• 1
• 0.8
• 0.6
• 0.4
• 0.2

5 10 15 ∆Pout (dB) Time (hr)

3/20

Outline

• Background
• Project goal
• Experimental

– Procedure – Results

• Discussion
• Conclusions

4/20

ID, RD, and IG start to degrade beyond critical voltage (Vcrit) (+ trapping behavior – current collapse) Common physical origin in ID and IG degradation

IDmax: VDS=5 V, VGS=2 V IGoff: VDS=0.1 V, VGS=-5 V

High Voltage Degradation in GaN HEMTs

Joh, EDL 2008

G S D AlGaN GaN 2DEG

VGS VDS

1.E-06 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 0.85 0.9 0.95 1 1.05 1.1 1.15 1.2 10 20 30 40 50 |IGoff| (A/mm) IDmax/IDmax(0), R/R(0) VDGstress (V)

IDmax RS RD IGoff Vcrit

OFF-state, VGS=-10 V

=-10 V

5/20

Structural Degradation

• 1. Vstress~Vcrit:

Groove formation in GaN cap

• 2. Vstress>Vcrit:

Pit formation in AlGaN barrier

• 3. Vstress>>Vcrit:

Pit growth (to AlGaN/GaN interface) and merge + crack formation

Cross-section Plan-view Gate SiN AlGaN GaN Makaram, APL 2010 Joh, MR 2010

6/20

Trapping vs. Permanent

0.7 0.8 0.9 1 30 60 Time (min) IDmax/IDmax (0)

88 days recovery stress recovery

total (apparent) degradation permanent degradation trapping degradation

VDS=0 VGS=‐30 V

13 % permanent degradation + 15 % trapping degradation

7/20

Project Goal

• Investigate time evolution of degradation and

correlate with structural degradation

Meneghesso, IJMWT 2010 Marcon, IEDM 2010

8/20

Experimental Procedure

Full Characterization (DC, CC)

START Electrical Stress

End?

END: detrapping + Full characterization

YES

Detrapping

NO

Tstress Tbase=30°C

• Detrapping step to flush trapped

electrons quickly

• Benign device characterization:
• Full ID-VDS, ID-VGS curves
• ID transient measurement:

current collapse, detrapping time constant

• Performed at 30 °C
• Stress conditions:
• OFF-state: VDS=40 V, VGS=-7 V
• T

stress=75–200 °C

9/20

Gate Current and VT

• Very fast IGoff and VT degradation (<10 ms)

 E-field driven oxide punch-through? Electrochemical etching?

• Degradation saturates after 104 s.

10

• 4

10

• 2

10 10

2

10

4

10

6

0.05 0.1 0.15 0.2 0.25 Stress time (s) |VT| (V) 10

• 8

10

• 7

10

• 6

10

• 5

10

• 4

|IGoff| (A)

IGoff |ΔVT| Initial Stress: VGS=‐7 V and VDS=40 V 125 °C

10

• 4

10

• 2

10 10

2

10

4

10

6

0.05 0.1 0.15 0.2 0.25 Stress time (s) |VT| (V) 10

• 8

10

• 7

10

• 6

10

• 5

10

• 4

|IGoff| (A)

IGoff |ΔVT| Initial Stress: VGS=‐7 V and VDS=40 V 125 °C

10/20

ID Transient Measurement

• uncol. IDlin – col. IDlin

uncollapsed IDlin CC=

• After electrical stress:

Permanent degradation + trapping related degradation

10

• 3

10

• 2

10

• 1

10 10

1

10

2

10

3

6 7 8 9 10 t (sec) IDlin (mA)

stress time=0-1 s 10 100 1000 After 10ks permanent degradation current collapse: trapping degradation trapping pulse (1 s VGS=‐10 V, VDS=0 V) @ t=0-

uncollapsed IDlin (fresh) uncollapsed IDlin (stressed) collapsed IDlin (stressed)

10

• 3

10

• 2

10

• 1

10 10

1

10

2

10

3

6 7 8 9 10 t (sec) IDlin (mA)

stress time=0-1 s 10 100 1000 After 10ks permanent degradation current collapse: trapping degradation trapping pulse (1 s VGS=‐10 V, VDS=0 V) @ t=0-

uncollapsed IDlin (fresh) uncollapsed IDlin (stressed) collapsed IDlin (stressed)

11/20

Detrapping Time-constant Spectrum

• Sharp increase in DP1 (Ea=0.56 eV) + long time

constant slow traps beyond incubation time.

10

• 3

10

• 2

10

• 1

10 10

1

10

2

10

3

• 1
• 0.8
• 0.6
• 0.4
• 0.2

0x 10

• 4

Detrapping time constant (sec) Amplitude (A.U.)

Stress time <1s 10s 100s 1000s >10ks DP1 VDS=0 pulse 1s, VGS=‐10 V Ta=30 °C

10

• 3

10

• 2

10

• 1

10 10

1

10

2

10

3

• 1
• 0.8
• 0.6
• 0.4
• 0.2

0x 10

• 4

Detrapping time constant (sec) Amplitude (A.U.)

Stress time <1s 10s 100s 1000s >10ks DP1 VDS=0 pulse 1s, VGS=‐10 V Ta=30 °C

12/20

Drain Current Degradation

10

• 4

10

• 2

10 10

2

10

4

10

6

1 2 3 4 5 6 Stress time (s) Permanent IDmax Degradation (%) 5 10 15 20 25 30 35 Current collapse (%)

Current collapse IDmax degradation Initial Stress: VGS=‐7 V and VDS=40 V 125 °C

10

• 4

10

• 2

10 10

2

10

4

10

6

1 2 3 4 5 6 Stress time (s) Permanent IDmax Degradation (%) 5 10 15 20 25 30 35 Current collapse (%)

Current collapse IDmax degradation Initial Stress: VGS=‐7 V and VDS=40 V 125 °C

Incubation time

• For current collapse and permanent IDmax

degradation, incubation time is observed.

• uncol. IDlin – col. IDlin

uncollapsed IDlin CC=

13/20

Temperature Dependence: IG

• Weak temperature dependence

10

• 4

10

• 2

10 10

2

10

4

10

6

10 10

1

10

2

10

3

10

4

Stress Time (s) |IGoff/IGoff(0)|

150 °C 100 °C 75 °C 125 °C Stress: VGS=‐7 V and VDS=40 V

Normalized |IGoff|

14/20

Temperature Dependence: VT

10

• 4

10

• 2

10 10

2

10

4

10

6

• 0.35
• 0.3
• 0.25
• 0.2
• 0.15
• 0.1
• 0.05

Stress Time (s)

VT (V)

150 °C 100 °C 75 °C 125 °C Stress: VGS=‐7 V and VDS=40 V

10

• 4

10

• 2

10 10

2

10

4

10

6

• 0.35
• 0.3
• 0.25
• 0.2
• 0.15
• 0.1
• 0.05

Stress Time (s)

VT (V)

150 °C 100 °C 75 °C 125 °C Stress: VGS=‐7 V and VDS=40 V

• No dependence during initial negative VT shift
• Positive turn-around seems to occur earlier at high T

15/20

Permanent IDmax Degradation

10

• 4

10

• 2

10 10

2

10

4

10

6

0.92 0.94 0.96 0.98 1 1.02 Stress Time (s) IDmax (norm)

150 °C 100 °C 75 °C 125 °C Stress: VGS=‐7 V and VDS=40 V

10

• 4

10

• 2

10 10

2

10

4

10

6

0.92 0.94 0.96 0.98 1 1.02 Stress Time (s) IDmax (norm)

150 °C 100 °C 75 °C 125 °C Stress: VGS=‐7 V and VDS=40 V

• Shorter incubation time at high T
• No saturation behavior up to >105 s

16/20

Current Collapse

10

• 4

10

• 2

10 10

2

10

4

10

6

10 20 30 40 Stress Time (s) Current collapse (%)

150 °C 100 °C 75 °C 125 °C VGS=‐7 V and VDS=40 V

10

• 4

10

• 2

10 10

2

10

4

10

6

10 20 30 40 Stress Time (s) Current collapse (%)

150 °C 100 °C 75 °C 125 °C VGS=‐7 V and VDS=40 V

• Shorter incubation time at high T
• More degradation at high T

17/20

Temperature Acceleration of Incubation Time

28 30 32 34 36

• 5

5 10 15 1/kT (eV-1) ln(inc) (s)

Permanent IDmax degradation Ea=1.12 eV Current collapse Ea=0.59 eV IGoff, Ea=0.17 eV

• Different level of temperature acceleration for incubation time.
• Ea for permanent IDmax degradation is similar to life test data*.

* Saunier, DRC 2007; Meneghesso, IJMWT 2010

18/20

Discussion: Time Evolution of Structural Degradation

VDS=0, VGS=-40 V, Tbase=150 °C

• Very fast groove formation (10 s) on gate edge.

 Related to gate current degradation

• Pit density/size gradually increase with time.

Joh, IWN 2010

19/20

Electrical vs. Structural Degradation

100 1000 10000 1 10 100 1000 10000 Average Pit Area (nm2) Stress Time (s)

Pit area~t1/4

10

• 4

10

• 2

10 10

2

10

4

10

6

10 10

1

10

2

Stress Time (s) Current collapse (%)

150 °C 100 °C 75 °C 125 °C Slope=0.22 Stress: VGS=‐7 V and VDS=40 V

Similar time dependence in current collapse and pit formation.

Joh, IWN 2010

20/20

Conclusion

• Investigated time evolution of electrical

degradation in GaN HEMTs

• Fast IG degradation ~ 10-100 ms

– Weak temperature dependence – Oxide punch through / groove formation?

• Current collapse degradation ~ 10-100 s

– Related to pit formation

• Permanent ID degradation >100 s

– Strong thermal activation (Ea=1.1 eV)