Recent Progress in Understanding the DC and RF Reliability of GaN High ‐ Electron Mobility Transistors J. A. del Alamo and J. Joh* Microsystems Technology Laboratories, MIT, Cambridge, MA *Presently with Texas Instruments, Dallas, TX Spring MRS 2012 San Francisco, April 9-13, 2012 Acknowledgements: ARL (DARPA-WBGS program), ONR (DRIFT-MURI program), J. Jimenez, C. V. Thompson, T. Palacios
Outline 1. Critical voltage for GaN degradation 2. Structural degradation of GaN HEMTs 3. Time evolution of degradation 4. Discussion 5. Tentative new model for electrical degradation 2
Critical voltage for degradation of GaN HEMTs 1.2 1.E+01 OFF-state, V GS =-10 V V DS 1.15 1.E+00 V GS =-10 V I Dmax /I Dmax (0), R/R(0) 1.1 1.E-01 R D |I Goff | (A/mm) G S D R S 1.05 1.E-02 AlGaN 1 1.E-03 2DEG 0.95 1.E-04 I Dmax I Goff 0.9 1.E-05 GaN V crit 0.85 1.E-06 Joh, EDL 2008 10 20 30 40 50 V DGstress (V) I Dmax : V DS =5 V, V GS =2 V I Goff : V DS =0.1 V, V GS =-5 V I D , R D , and I G start to degrade beyond critical voltage (V crit ) (+ increased trapping behavior – current collapse) 3
Critical voltage: a universal phenomenon? GaN HEMT on SiC GaN HEMT on SiC GaN HEMT on SiC Meneghini, IEDM 2011 Ivo, MR 2011 Liu, JVSTB 2011 GaN HEMT on Si GaN HEMT on Si GaN HEMT on sapphire Demirtas, ROCS 2009 Ma, Chin Phys B 2011 Marcon, IEDM 2010 4
Initial hypothesis: Inverse Piezoelectric Effect Mechanism Strong piezoelectricity in AlGaN G S D |V DG | ↑ tensile stress ↑ AlGaN crystallographic defects beyond critical elastic energy 2DEG GaN G G Defects: S S D D Trap electrons AlGaN AlGaN ∆Φ bi ∆Φ bi n s ↓ → R D ↑ , I D ↓ 2DEG 2DEG Strain relaxation GaN GaN I D ↓ E C E C Joh, IEDM 2006 Provide paths for I G defect defect state state Joh, IEDM 2007 I G ↑ E F E F Joh, MR 2010b AlGaN GaN AlGaN GaN 5
Structural degradation: cross section ‐ Small dimple in Gate early stages of I G degradation; ‐ I D degradation delayed AlGaN GaN 1.2 1.E+02 V DS =0 stress #2 Device #1 I Dmax /I Dmax (0), R/R(0) 1.E+01 #3 I Goff 1 |I Goff | (mA/mm) #4 1.E+00 0.8 #5 1.E-01 1.E-02 I Dmax 0.6 1.E-03 #6: unstressed 0.4 1.E-04 10 20 30 40 0 10 20 30 V DGstress (V) Joh, MR 2010 6
Correlation between pit geometry and I Dmax degradation Joh, MR 2010 50 20 Permanent I Dmax Degradation (%) Pit 40 15 depth Current collapse (%) 30 10 20 5 10 0 0 0 2 4 6 8 0 2 4 6 8 Pit depth (nm) Pit depth (nm) Pit depth and I Dmax degradation correlate: both permanent degradation and current collapse (CC) 7
Structural degradation: planar view OFF-state step-stress, V GS =-7 V, T base =150 °C Makaram, APL 2010 200 nm V DG =15 V 200 nm V DG =19 V Unstressed 200 nm (V crit ) 200 nm V DG =42 V V DG =57 V 200 nm • Continuous groove appears for V stress <V crit • Deep pits formed along groove for V stress >V crit 8
Correlation between pit geometry and I Dmax degradation Makaram, APL 2010 Cross-sectional area averaged over 1 µm 12 12 Permanent I Dmax Degradation (%) Post ‐ Stress Current Collapse (%) 10 10 8 8 6 6 4 4 2 2 0 0 0 50 100 150 0 50 100 150 Average Defect Area (nm 2 ) Average Defect Area (nm 2 ) I Dmax degradation and pit cross-sectional area correlate 9
Planar degradation: the role of time V DS =0, V GS =-40 V, T base =150 °C Joh, IWN 2010 • Very fast groove formation (within 10 s) • Delayed pit formation • Pit density/size increase with time • Good correlation between I Dmax degradation and pit area 10
-4 -4 Time evolution of 0.25 0.25 10 10 Stress: V GS = ‐ 7 V and V DS =40 V Stress: V GS = ‐ 7 V and V DS =40 V 125 °C 125 °C 0.2 0.2 degradation for -5 -5 10 10 I Goff I Goff 0.15 0.15 |I Goff | (A) |I Goff | (A) | V T | (V) | V T | (V) constant V stress > V crit -6 -6 10 10 0.1 0.1 | Δ V T | | Δ V T | -7 -7 10 10 0.05 0.05 I Goff and V T degradation: -8 -8 • fast (<10 ms) 0 0 10 10 -4 -4 -2 -2 0 0 2 2 4 4 6 6 Initial Initial 10 10 10 10 10 10 10 10 10 10 10 10 saturate after 10 4 s • Stress time (s) Stress time (s) CC degradation: • slower • hint of saturation for long time P ermanent I Dmax degradation: • much slower • does not saturate with time Joh, IRPS 2011 11
The role of temperature in time evolution Incubation time • I G : weak T dependence • CC, I Dmax : strongly T activated Joh, IRPS 2011 12
Temperature acceleration of incubation time 15 Permanent I Dmax degradation E a =1.12 eV 10 ln( inc ) (s) Current collapse 5 E a =0.59 eV 0 I Goff , E a =0.17 eV -5 28 30 32 34 36 1/kT (eV -1 ) • Different E a for I Goff , CC, I Dmax reveal different degradation physics • E a for permanent I Dmax degradation similar to life test data * * Saunier, DRC 2007; Meneghesso, IJMWT 2010 13
Summary of degradation after OFF ‐ state stress for V stress > V crit 1. I G degradation • Fast • Voltage enhanced • Little temperature sensitivity • Tends to saturate Correlates with shallow groove • Groove continuous; on S and D side • Groove appears for V stress < V crit Mechanisms: • Groove: reduction of interfacial oxide at drain end of gate field ‐ induced oxidation? • I G rise: hopping through defects? Lower B due to gate interface reconstruction? 14
Electroluminescence correlates with I G degradation V DS =0 Zanoni, EDL 2009 Meneghini, IEDM 2011 • Gate current electrons produce EL in GaN substrate • EL spots tend to merge into a continuous line 15
Summary of degradation after OFF ‐ state stress for V stress > V crit 2. Current collapse degradation (trapping) • Slower • Enhanced by temperature, voltage • Tends to saturate for very long times Correlates with pits : • Pits randomly located on drain side • Pits grow with V stress , time and temperature • Pits eventually merge Mechanism: • Pits: relieve mechanical stress arising from inverse piezoelectric effect (but detailed process?); electrochemical oxidation? • CC: trapping associated with facets of pits? 16
Time evolution of pit growth and CC 2 10 Stress: 10000 150 °C V GS = ‐ 7 V and V DS =40 V 125 °C Current collapse (%) Average Pit Area (nm 2 ) 100 °C 75 °C 1 10 1000 Pit area~t 1/4 Slope=0.22 0 100 10 -4 -2 0 2 4 6 0 10 10 10 10 10 10 1 10 100 1000 10000 Stress Time (s) Stress Time (s) Joh, IRPS 2011 Joh, IWN 2010 Similar time dependence in current collapse and pit formation 17
Current collapse time ‐ constant spectrum Using current-transient Stress time Stress time -4 -4 methodology of Joh TED 2010 0x 10 0x 10 <1s <1s 10s 10s -0.2 -0.2 Amplitude (A.U.) Amplitude (A.U.) 100s 100s -0.4 -0.4 1000s 1000s -0.6 -0.6 V DS =0 pulse V DS =0 pulse >10ks >10ks 1s, V GS = ‐ 10 V 1s, V GS = ‐ 10 V -0.8 -0.8 T a =30 °C T a =30 °C DP1 DP1 -1 -1 -3 -3 -2 -2 -1 -1 0 0 1 1 2 2 3 3 10 10 10 10 10 10 10 10 10 10 10 10 10 10 Joh, IRPS 2011 Detrapping time constant (sec) Detrapping time constant (sec) • Dominant trap created by stress already present in virgin sample, E a =0.56 eV • CC associated with surface (?) (pit creates new surfaces right next to gate edge!) 18
Summary of degradation after OFF ‐ state stress for V stress > V crit 3. I Dmax , R D degradation • Much slower • Temperature activated • Voltage enhanced • Does not saturate Correlate with pits Mechanism: • Pit depletes electron concentration in channel below • Current loss associated with pit depth ? 19
Can degradation occur for V stress < V crit ? V crit =75 V Meneghini, IEDM 2011 • Sudden irreversible increase in I G , enhanced by V stress • No reported I D degradation • Appearance of I G noise, EL hot ‐ spots • E a =0.12 eV • Consistent with groove formation? Marcon, IEDM 2010 20
Oxygen inside pit EDX LEES Conway, Mantech 2007 • O, Si, C found inside pit • Anodization mechanism for pit formation? (Smith, ECST 2009) • Electrical stress experiments under N 2 inconclusive Park, MR 2009 21
Groove: essential for pit formation? Appearance of groove increases mechanical stress due to inverse piezoelectric effect at drain end of gate 2 nm x 3 nm groove increases • mechanical stress in AlGaN from 4.6 GPa to 13 GPa Groove has little effect in • current underneath Pit formation brings major loss • of current Ancona, SISPAD 2011 22
2D distribution of structural degradation under high ‐ power stress 4x100 µm devices SEM: finger end SEM: finger center Stress conditions: I D =250 mA/mm, V DS =40 V @120C, t=523 h AFM • Prominent pits under gate edge on drain side • Pits more prominent towards center of finger • Typical pit: 50 nm wide, 8 nm deep Lin, APL 2012 23
inner finger 2D distribution of pits characterized area outer finger • Pits evaluated along 10 µm segments • Pit cross sectional area and density increase towards center of device • Consistent with local T activating pit formation 24
Recommend
More recommend
