Non-Uniform Degradation Behavior Across Device Width in RF Power - PowerPoint PPT Presentation

Non-Uniform Degradation Behavior Across Device Width in RF Power GaAs PHEMTs A. A. Villanueva 1 , J. A. del Alamo 1 , T. Hisaka 2 , K. Hayashi 2 and M. Somerville 3 1 Massachusetts Institute of Technology 2 Mitsubishi Electric 3 Olin College of Engineering Sponsor: Mitsubishi Electric

Motivation • Electrical degradation is serious concern in RF power GaAs PHEMTs – Under stressing: R D � and I max � → P out � • Degradation mechanisms identified [1]- [3], but no studies of uniformity • This study : investigate degradation across device width G S D W [1] del Alamo et al (IEDM 2004) [2] Meneghesso et al (1996) [3] Hisaka et al (GaAs IC 2003)

Outline • Introduction • Experimental • PHEMT Degradation – Light Emission • TLM Degradation – Light Emission – Materials Analysis • Conclusions

Introduction Stressing: I D = 400 mA/mm, step V DGo +V T . In air @ 300 K. Drain Source Gate 1.10 7.4 I max 7.2 R D /R D (0), I max /I max (0) ohmics 1.05 n + GaAs 7.0 cap V DGo +V T [V] V DGo +V T n - GaAs 6.8 etch-stop 1.00 AlGaAs supply 6.6 channel InGaAs 6.4 R D 0.95 supply AlGaAs 6.2 GaAs buffer 0.90 6.0 0 100 200 300 400 500 600 700 time [min] With step-stress : • Experimental RF power PHEMTs • R D � • L g = 0.25 μ m, W g = 50 μ m f t ~ 40-50 GHz, BV DG,off ~ 12-15 V • I max � (from V T � )

Light Emission & Degradation • R D degradation due to G S D surface corrosion [3], high E involved h ν h ν • High E � impact ionization (II) � - + recombination � light + + - emission - - • Light-emission picture: spatial view of II, E [3] Hisaka et al (GaAs IC 2003)

Light-emission: Experimental camera • Astronomical-grade CCD sensor microscope • Stressing: constant V GS & constant V DS probes – V DS stepped DUT • Photographs taken: 8.2 8.0 – at frequent intervals V GS =0.3 V 7.8 – at fixed (low) value of 7.6 V DS [V] 7.4 V DS 7.2 7.0 6.8 6.6 6.4 0 100 200 300 400 500 600 700 800 900 time [min]

Light-Emission vs. Stressing (1) V GS =0.3 V • V DS � → I h ν � 5 x 10 8.2 V DS 7.8 V DS [V] 7.4 • For constant V DS , 7.0 I h ν constant, but 6.6 eventually � 8 I hv / I D [a.u.] 6 � R D � → V DGo � I h ν /I D 4 2 0 0 0 200 200 400 400 600 600 800 800 time [min]

Light-Emission vs. Stressing (1) V GS =0.3 V • V DS � → I h ν � 5 x 10 8.2 V DS 7.8 V DS [V] 7.4 • For constant V DS , 7.0 I h ν constant, but 6.6 eventually � 8 I hv / I D [a.u.] 6 � R D � → V DGo � I h ν /I D 4 2 0 0 0 200 200 400 400 600 600 800 800 time [min]

Light-Emission vs. Stressing (2) V GS = 0.3 V S D V DS = 6.6 V W=50 µm 45 µm 30 µm 45 µ m t = 0 min 218 min 428 min 638 min 849 min • Initially, light concentrates in center ~30 µm of width • With stressing: (1) light spreads out along width (2) weakens in intensity

Light Emission vs. Width V GS = 0.3 V V GS = 0.3 V light from source side V DS = 6.6 V V DS = 6.6 V 14 6 t = 428 min 12 5 10 i hv / I D [a.u.] 4 I h ν / I D [a.u.] I hv / I D [a.u.] I hv / I D [a.u.] 8 t =849 min 3 6 2 4 t = 0 min 1 2 0 0 0 10 20 30 40 50 60 0 200 400 600 800 width [ μ m] width (um) time [min] • 1 st half: light spreads out → I hv � • 2 nd half: intensity decreases → I hv �

Light Emission During Stressing device width 5 x 10 8.2 160 t = 639 min, V DS = 7.8 V 7.8 V DS [V] 140 7.4 120 7.0 t = 428 min, V DS = 7.4 V i hv / I D [a.u.] I hv / I D [a.u.] 100 6.6 8 80 I hv / I D [a.u.] 6 60 t = 218 min, V DS = 7.0 V 4 40 2 t = 0 min, V DS = 6.6 V 20 0 0 0 0 200 200 400 400 600 600 800 800 0 10 20 30 40 50 60 time [min] width [ μ m] width [um] • During stressing, at high bias: • early stages: degradation peaks in center • advanced stages: degradation peaks at edges

Light-Emission of TLMs (1) 5 x 10 TLM: same structure as 6.00 V D PHEMT, but no gate 5.75 V D [V] 5.50 h ν 5.25 h ν 5.00 3 I hv / I D [a.u.] - + 2 - + + - - I h ν / I D 1 0 0 0 50 50 100 100 150 150 200 200 time [min] • V D � → I hv � • Constant V D → I hv � � R � → II �

Light Emission of TLMs (2) V D = 5.0 V 5.0 V 5.25 V 5.50 V 5.75 V 6.0 V • Light initially concentrated in center • With stressing: W=100 µm 90 µm 55 µm – Light spreads out over width of TLM – I h ν � (for constant voltage) t = 0 min 67 min 100 min 133 min 167 min 201 min (Similar to PHEMT light emission behavior)

Light Emission vs. Width (TLMs) device width 3.0 t = 0 min, V D = 5 V 2.5 2.0 i hv / I D [a.u.] I hv / I D [a.u.] t = 100 min, V D = 5.25 V 1.5 1.0 t = 201 min, V D = 6 V 0.5 0 0 20 40 60 80 100 width ( μ m) width [ μ m] • During stressing, at high bias, light “peaks” at edges → similar behavior in PHEMTs

Origin of Non-Uniform II 3 possible causes for non-uniform II across device width: •Non-uniform I D •Non-uniform T •Non-uniform E-field

Non-Uniform Drain Current ? G • Non-uniform I D S D – but II ∝ I D , so I D requires edges be “shut off”

Non-Uniform Temperature? G • Non-uniform T S D – but edges should be cooler → more II

Non-Uniform Electric Field? G • Non-uniform E-field S D – II ∝ exp(-1/E), small Δ E → large Δ II – from non-uniform recess

Recess Non-Uniformity (TLMs) L=2.4 µm x y narrower wider W = 60 μ m Examined top view of entire recess area � recess is shorter in the center

Recess vs. Width (TLMs) W = 60 µm • Nominal recess: 0.7 µm 1.0 • Actual recess varies: 0.9 – Center: ~0.6-0.7 µm ] length [um 0.8 – Edges: ~0.8-0.9 µm 0.7 0.6 AFM 0.5 0 10 20 30 40 50 60 width [ μ m] x [um] In center: electric field � → II � → degradation � � Same phenomenon likely happening in PHEMTs

Conclusions • Non-uniform recess geometry � non-uniform E • Areas of higher E � areas more susceptible to degradation • To improve long-term device reliability: must identify & minimize non-uniformities in device geometry