This page covers the degradation of the GaAs and InP transistors used in microwave and millimetre-wave circuits: MESFETs, GaAs pHEMTs and InP HEMTs, and GaAs and InP heterojunction bipolar transistors (HBTs). GaN HEMTs have their own page, and the electromigration of the gold lines follows the Black model of the Metal and Interconnect page; this page covers the gate metal, hydrogen, ohmic contacts and the current gain of HBTs.

1. Why It Happens

Trigger: Long operation at a channel or junction temperature of 100 to 150 $^{\circ} C$ , hydrogen released inside a hermetic package, high current density in HBTs and RF overdrive (Kayali; Northrop Grumman; Lin and Wang).
Mechanism, gate sinking: Gold of the gate diffuses into the GaAs, or the gate metal reacts with it, so the Schottky junction moves into the channel, which thins the channel and shifts the threshold and the current; Pt or Pd barriers slow it, and grain boundaries in them are fast paths (Kayali).
Mechanism, hydrogen: Hydrogen from the package diffuses through the nitride and reacts with the Pt of the gate, and atomic hydrogen neutralises the Si donors, so the threshold shifts by hundreds of millivolts and $g_{m}$ and gain fall (Northrop Grumman; Kayali); in InP HEMTs it also changes the breakdown voltage through the exposed InAlAs surface (Blanchard et al.).
Mechanism, HBT gain: Recombination-enhanced defect reactions in the base raise the base current, and the external base surface and metal diffusion from the emitter contact add gradual and sudden modes (Livingston; Kashio et al.).
What sets the rate: The channel or junction temperature through activation energies of 0.35 to 1.8 $eV$ (Kayali; Livingston; Kashio et al.); the hydrogen level in the package and the nitride thickness (Northrop Grumman); the collector current density in HBTs (Lin and Wang; Wuerfl et al.).
Failure criterion: A drop of the HBT current gain by 15 or 20 per cent (Kashio et al.; Lin and Wang); a threshold or $I_{D S S}$ shift, a $g_{m}$ or gain loss beyond the circuit margin, or a power slump of about 1 $dB$ (Northrop Grumman; Kayali).

2. Where to Look

Location: The gate metal and the GaAs below it (Kayali); the Pt layer of the gate and the passivation above it (Northrop Grumman); the ohmic contacts and the drain contact, where burnout starts (Kayali); the external base surface and the emitter metal of the HBT (Kashio et al.).
Signature: A slow, monotonic shift of threshold and $I_{D S S}$ (gate sinking); a positive threshold shift that depends on the package and not on the bias (hydrogen); an HBT gain that first rises or falls in a burn-in phase, then is stable and then drops quickly (Wuerfl et al.); a sudden gain drop from metal diffusion (Kashio et al.).
Confirm with: High-temperature operating life at several channel temperatures with an Arrhenius fit (Lin and Wang); exposure to forming gas or high-hydrogen atmospheres at about 200 $^{\circ} C$ (Northrop Grumman; Blanchard et al.); residual gas analysis of the package; cross sections of the gate (Kayali).
Look-alikes: Hydrogen and gate sinking both shift the threshold, but the hydrogen shift appears only in a sealed package and saturates with the hydrogen supply (Northrop Grumman); a power slump under RF drive is hot-electron trapping, which has a weak or negative temperature dependence, unlike the thermal mechanisms (Kayali).

3. Prevention

Design: Keep the junction or channel below about 125 $^{\circ} C$ , the design limit of one HBT amplifier (Lin and Wang); keep HBT current densities moderate, since the degradation grows with them (Livingston); keep gold line current densities below about $2 \times 1 0^{5}$ $A/c m^{2}$ (Kayali).
Material and process: Pt or Pd diffusion barriers in the gate (Kayali); InGaP rather than AlGaAs emitters and carbon rather than beryllium base doping (Lin and Wang; Livingston); a ledge over the external base and a tungsten-based emitter metal (Kashio et al.); a nitride passivation of the right thickness (Northrop Grumman).
Operation: For hermetic packages keep the hydrogen below 100 ppm and preferably below 10 ppm with getters, since a bake only reaches 200 to 1000 ppm (Northrop Grumman); avoid RF overdrive (Kayali).
Screening and test: A burn-in that passes the initial gain change of HBTs before delivery (Wuerfl et al.); life tests at more than one temperature, since tests at 200 to 300 $^{\circ} C$ may overestimate the life at use (Livingston).

FET Gates, Hydrogen and Contacts

Name	Formula	Related Formulae	Variables/Constants
Gate sinking and the threshold of a MESFET or HEMT	$Δ a \approx D t, D = D_{0} e^{- E_{a} / (k_{B} T)}$	$V_{p} = \frac{q N _{D} a ^{2}}{2 ε _{s}}, Δ V_{p} \approx \frac{q N _{D} a Δ a}{ε _{s}}$	$Δ a : d e pt h b y w hi c h t h e g a t e m e t a l r e a c t i o n f r o n t m o v es in t o t h e c hann e l$ $D : e f f ec t i v e d i f f u s i v i t y o f t h e g a t e m e t a l$ $t : t im e$ $D_{0} : p r e f a c t or$ $E_{a} : a c t i v a t i o n e n er g y$ $k_{B} : B o l t z mann co n s t an t 8.617 \times 1 0^{- 5} eV/K$ $T : ab so l u t e t e m p er a t u r e$ $V_{p} : p in c h - o f f v o l t a g e o f a u ni f or m l y d o p e d c hann e l o f t hi c k n ess a$ $q : e l e m e n t a r y c ha r g e$ $N_{D} : c hann e l d o p in g$ $ε_{s} : p er mi tt i v i t y o f G a A s, 12.9 ε_{0} (t h e f r o n t f o l l o w s t h e 2 D t d i f f u s i o n l e n g t h o f t h e D i f f u s i o n F u r na ce p a g e, t h e p in c h - o f f v o l t a g e i s t ha t o f t h e M E S F E T p a g e an d i t s c han g e i s i t s d er i v a t i v e, an d 12.9 i s t h e G a A s v a l u e o f t h e C o n v e n t i o na l S e mi co n d u c t o r^{'} s M a t er ia l V a l u es p a g e); K a y a l i : A u d i f f u ses in t o G a A s an d r e d u ces t h e a c t i v e c hann e l d e pt h, t h e r a t e d e p e n d s o n d i f f u s i v i t y, t e m p er a t u r e an d co n ce n t r a t i o n g r a d i e n t, P t or P d ba r r i er l a y er s l imi t i t, an d g r ain b o u n d a r i es in t h e ba r r i er a r e f a s t p a t h s; n o a c t i v a t i o n e n er g y i s g i v e n in t h e t e x t r e a d$ $(co m p u t e d w i t h a ss u m e d N_{D} = 2 \times 1 0^{17} c m^{- 3} an d a = 100 nm : V_{p} = 1.40 V, an d a f r o n t m o v in g 5 nm in t o t h e c hann e l s hi f t s i t b y 0.14 V)$
Hydrogen poisoning of Pt-gate FETs in a hermetic package	$Δ V_{t h} = f (p_{H_{2}}, T, t), A F = e^{\frac{E _{a}}{k _{B}} (\frac{1}{T _{u}} - \frac{1}{T _{s}})}$	$c_{H_{2}} < 100 ppm (preferably < 10)$	$Δ V_{t h} : t h r es h o l d s hi f t$ $p_{H_{2}} : h y d r o g e n p a r t ia l p r ess u r e in t h e p a c k a g e$ $T : t e m p er a t u r e$ $t : t im e$ $A F : a cce l er a t i o n f a c t or b e tw ee n t h e u se an d s t r ess t e m p er a t u r es T_{u} an d T_{s}$ $E_{a} : a c t i v a t i o n e n er g y$ $k_{B} : B o l t z mann co n s t an t 8.617 \times 1 0^{- 5} eV/K$ $c_{H_{2}} : h y d r o g e n co n ce n t r a t i o n in t h e p a c k a g e g a s; N or t h r o p G r u mman : h y d r o g e n d i f f u ses t h r o ug h t h e S i N an d r e a c t s w i t h t h e P t o f t h e g a t e, t h e p os i t i v e t h r es h o l d s hi f t i s o f t e n 300 t o 500 mV a f t er m or e t han 100, 000 pp m a t 200^{\circ} C f or m or e t han 100 h, E_{a} ab o u t 0.6 or 0.7 eV, g_{m} an d l in e a r g ain a r e m os t se n s i t i v e, a bak e b r in g s t h e l e v e l o n l y t o 200 t o 1000 pp m, g e tt er s b e l o w 10 pp m, an d t h e i r P t - co n t ainin g H B T s s h o w n o h y d r o g e n d e g r a d a t i o n; K a y a l i : a t o mi c h y d r o g e n f or m s S i - H co m pl e x es t ha t n e u t r a l i se t h e S i d o n or s, an d in p H E M T s t h e d r ain c u r r e n t c an a l so in cr e a se$ $(co m p u t e d : f r o m a 200^{\circ} C t es t t h e a cce l er a t i o n t o 60^{\circ} C u se i s 485 f or 0.6 eV an d 1360 f or 0.7 eV, an d t o 85^{\circ} C 113 an d 248)$
Ohmic contacts, electromigration and burnout	$M T T F \propto e^{E_{a} / (k_{B} T)}, E_{a} = 0.5 to 1.8 eV$	$j < j_{t h} \approx 2 \times 1 0^{5} A/c m^{2}$	$M T T F : m e d ian l i f e o f t h e A u G e / N i o hmi c co n t a c t$ $E_{a} : a c t i v a t i o n e n er g y$ $k_{B} : B o l t z mann co n s t an t 8.617 \times 1 0^{- 5} eV/K$ $T : c hann e l t e m p er a t u r e$ $j : c u r r e n t d e n s i t y in t h e g o l d in t er co nn ec t$ $j_{t h} : t h r es h o l d f or e l ec t r o mi g r a t i o n d ama g e; K a y a l i : A u G e / N i co n t a c t s a r e a l l oy e d ab o v e 400^{\circ} C w i t h 0.1 t o 0.5 Ω mm, t h e o hmi c a c t i v a t i o n e n er g y i s 0.5 t o 1.8 eV (r e a so nab l e l i f e b e l o w 100^{\circ} C, r a p i d w e a r ab o v e 150^{\circ} C), e l ec t r o mi g r a t i o n v o i d s g r o w in a se l f - a cce l er a t in g r u na w a y, a t es t e x am pl e i s 5000 h a t 200^{\circ} C an d 5.3 \times 1 0^{5} A/c m^{2}, so u r ce - d r ain b u r n o u t s t a r t s w h e n t h e b u f f er or s u b s t r a t e l oc a l l y e x cee d s 550^{\circ} C, an d R F o v er d r i v e g i v es a p o w er s l u m p o f ab o u t 1 dB in 1000 h w i t h a w e ak or n e g a t i v e t e m p er a t u r e d e p e n d e n ce (h o t - e l ec t r o n t r a pp in g)$ $(co m p u t e d : 50 K f r o m 100 t o 150^{\circ} C s h or t e n s t h e co n t a c t l i f e b y 6.3 f or 0.5 eV an d b y 746 f or 1.8 eV, so t h e r an g e o f a c t i v a t i o n e n er g y i s t h e main u n cer t ain t y; t h e B l a c k m o d e l o f t h e M e t a l an d I n t er co nn ec t p a g e a ppl i es t o t h e g o l d l in es)$

HBT Current Gain and Life

Name	Formula	Related Formulae	Variables/Constants
Current gain of an HBT and the periphery recombination	$\frac{1}{β} = \frac{J _{bi}}{J _{c}} + \frac{J _{L}}{J _{c}} \frac{L _{E}}{S _{E}}$	$\frac{β ( t )}{β _{0}} \leq 0.80 or 0.85 \Rightarrow failure$	$β : co mm o n - e mi tt er c u r r e n t g ain$ $J_{bi} : in t er na l (a r e a) r eco mbina t i o n c u r r e n t d e n s i t y in t h e ba se$ $J_{c} : co l l ec t or c u r r e n t d e n s i t y$ $J_{L} : p er i p h er y r eco mbina t i o n c u r r e n t p er u ni t l e n g t h o f t h e e mi tt er e d g e$ $L_{E} : e mi tt er p er i p h er y l e n g t h$ $S_{E} : e mi tt er a r e a$ $β_{0} : ini t ia l g ain; K a s hi o e t a l . (N T T, I n P H B T) : E q . (1) a s w r i tt e n, a l e d g e c u t s J_{L} t o 0.3 μ A/ μ m a t 1 mA/ μ m^{2}, ab o u t t e n t im es l ess t han w i t h o u t, t h e g r a d u a l m o d e co m es f r o m t h e e x t er na l ba se s u r f a ce b e tw ee n t h e l e d g e an d t h e ba se m e t a l, t h e s u dd e n m o d e f r o m T i / A u d i f f u s i o n, w hi c h a t u n g s t e n - ba se d e mi tt er m e t a l s t o p s, an d t h e cr i t er i o n i s a 15 p er ce n t d r o p o f β / β_{0}; L in an d W an g : a 20 p er ce n t d r o p o f β i s t h e u s u a l cr i t er i o n; W u er f l e t a l . : d e g r a d a t i o n s h o w s a b u r n - in p ha se, a s t ab l e p ha se an d a r a p i d e n d - o f - l i f e d ec l in e$ $(co m p u t e d w i t h an a ss u m e d 0.5 b y 3 μ m e mi tt er (L_{E} = 7 μ m, S_{E} = 1.5 μ m^{2}) an d J_{bi} / J_{c} = 0.005 (a ss u m e d) : t h e p er i p h er y t er m i s 0.0014 an d β = 156; a d o u b l e d J_{L} g i v es β = 128, an 18 p er ce n t d r o p, n e a r t h e e n d - o f - l i f e cr i t er i o n)$
Life of GaAs and InP HBTs against junction temperature	$M T T F (T_{u}) = M T T F (T_{s}) e^{\frac{E _{a}}{k _{B}} (\frac{1}{T _{u}} - \frac{1}{T _{s}})}$	$T_{j} = T_{a} + R_{t h} I_{C} V_{C E}$	$M T T F : m e d ian t im e t o t h e g ain cr i t er i o n$ $T_{u}, T_{s} : u se an d s t r ess j u n c t i o n t e m p er a t u r es in K$ $E_{a} : a c t i v a t i o n e n er g y$ $k_{B} : B o l t z mann co n s t an t 8.617 \times 1 0^{- 5} eV/K$ $T_{j} : j u n c t i o n t e m p er a t u r e$ $T_{a} : ambi e n t$ $R_{t h} : t h er ma l r es i s t an ce$ $I_{C} V_{C E} : d i ss i p a t e d p o w er (t h e seco n d f or m i s t h e j u n c t i o n t e m p er a t u r e o f t h e T h er ma l S t r ess p a g e); L in an d W an g (I n G a P / G a A s) : ab o u t 3200 h a t 75 kA/c m^{2} an d a 330^{\circ} C j u n c t i o n, ab o v e 7.4 \times 1 0^{6} h p r o j ec t e d a t 125^{\circ} C w i t h a co n ser v a t i v e 0.8 eV, an d m or e t han 6500 h w i t h o u t m os t f ai l u r es a t 25 an d 50 kA/c m^{2} an d 315^{\circ} C; W u er f l e t a l . (G a I n P / G a A s) : t y p i c a l l y 3 \times 1 0^{7} h a t 1 \times 1 0^{5} A/c m^{2} an d 125^{\circ} C; K a s hi o e t a l . (I n P) : ab o v e 1 \times 1 0^{8} h a t 125^{\circ} C w i t h E_{a} = 1.5 eV, s t r esse d a t j u n c t i o n s o f 284 an d 289^{\circ} C; L i v in g s t o n : a c t i v a t i o n e n er g i es o f r eco mbina t i o n - e nhan ce d d e f ec t r e a c t i o n s a s l o w a s 0.35 eV, I n G a P e mi tt er s ab o u t t e n t im es t h e l i f e o f A l G a A s, l o w er - V_{B E} w a f er s s u r v i v e f i v e t im es l o n g er, an d l i f e t es t s a t 200 t o 300^{\circ} C ma y o v er es t ima t e t h e l i f e a t u se$ $(co m p u t e d : f r o m 330 t o 125^{\circ} C t h e a cce l er a t i o n i s 2770 f or 0.8 eV, so 3200 h b eco m es 8.8 \times 1 0^{6} h, co n s i s t e n t w i t h ab o v e 7.4 \times 1 0^{6}; f or 0.35 an d 1.5 eV i t i s 32 an d 2.8 \times 1 0^{6}, so t h e s am e t es t p r e d i c t s 1 \times 1 0^{5} t o 9 \times 1 0^{9} h)$

Important Questions

Ques: Which activation energy should be used for GaAs and InP HBT life: 0.35, 0.8 or 1.5 electronvolts?

The sources report different values for different mechanisms. Livingston notes recombination-enhanced defect reactions with activation energies as low as 0.35 $eV$ in GaAs HBTs, Lin and Wang project InGaP/GaAs life with a conservative 0.8 $eV$ , and Kashio et al. find 1.5 $eV$ for InP HBTs with a surface-related gradual mode. From a 330 $^{\circ} C$ test to 125 $^{\circ} C$ these give accelerations of 32, 2770 and $2.8 \times 1 0^{6}$ (computed), so one test result spans five orders of magnitude of projected life. The value must belong to the mechanism that actually fails at use, which is why Livingston warns that tests at 200 to 300 $^{\circ} C$ can overestimate the life. For a projection without mechanism data the lowest value is the safest, and 0.8 $eV$ is a reasonable middle for InGaP devices (my reading). Reference: H. Livingston (BAE Systems), ‘A Survey of Heterojunction Bipolar Transistor (HBT) Device Reliability’, updated 3 June 2003 (fetch-tool summary); B. Lin and L. Wang, ‘InGaP emitters make HBTs more reliable’, Compound Semiconductor, May 2001 (fetch-tool summary); N. Kashio, K. Kurishima, Y. K. Fukai and S. Yamahata, ‘High-speed, High-reliability 0.5-um-emitter InP-based Heterojunction Bipolar Transistors’, NTT Technical Review 7(12), December 2009 (fetch-tool summary).

Ques: Is the end of life of an HBT a 15 or a 20 per cent loss of current gain?

Both are used. Lin and Wang call a 20 per cent drop of the gain the usual maxim for failure, while Kashio et al. use a 15 per cent drop of $β / β_{0}$ for their InP HBTs. The choice changes the reported life little when the gain falls quickly at the end, as Wuerfl et al. describe with a burn-in, a stable phase and a rapid end-of-life decline, but it matters for a gradual mode. With the gain equation of Kashio et al. and assumed values, doubling the periphery recombination current lowers the gain from 156 to 128, an 18 per cent drop (computed), which fails one criterion and passes the other. The criterion must therefore be stated with every MTTF, and the circuit’s own gain margin is the more relevant limit (my reading). Reference: B. Lin and L. Wang, ‘InGaP emitters make HBTs more reliable’, Compound Semiconductor, May 2001 (fetch-tool summary); N. Kashio, K. Kurishima, Y. K. Fukai and S. Yamahata, ‘High-speed, High-reliability 0.5-um-emitter InP-based Heterojunction Bipolar Transistors’, NTT Technical Review 7(12), December 2009 (fetch-tool summary); J. Wuerfl, P. Kurpas, F. Brunner, M. Mai, M. Rudolph and M. Weyers, ‘Degradation properties of MOVPE-grown GaInP/GaAs HBTs under combined temperature and current stressing’, Microelectronics Reliability 41(8), 1103-1108 (2001) (abstract only, FBH page).

Ques: Does hydrogen shift the threshold of a GaAs FET up or down?

The sources report both. Northrop Grumman finds a positive shift of often 300 to 500 $mV$ in Pt-gate GaAs FETs after high hydrogen exposure at 200 $^{\circ} C$ , with $g_{m}$ and linear gain falling. Kayali says that atomic hydrogen neutralises the Si donors, which lowers the current, but also that in pHEMTs the drain current can increase, which he attributes to a change of the built-in potential at the gate. Blanchard et al. cite a negative threshold shift in InP HEMTs and add a 0.9 $V$ rise of the breakdown voltage after a forming-gas anneal. The sign depends on whether the gate barrier or the donors dominate. For Pt-gate GaAs pHEMTs the positive shift is the more likely, since it is the measured result of a maker for such devices, and the sign must be checked for each technology. Reference: Northrop Grumman, ‘Effects of Hydrogen on Hermetically Packaged GaAs MMICs’, application note, revision April 2008 (fetch-tool summary); S. Kayali, ‘Basic Failure Modes and Mechanisms’, chapter 4 (pp. 89-108) of the JPL GaAs MMIC Reliability Assurance Guideline (parts.jpl.nasa.gov; title of the guideline and year not shown in the text read, references to 1996) (fetch-tool summary); R. R. Blanchard, J. A. del Alamo and A. Cornet Calveras, ‘Hydrogen-Induced Changes in the Breakdown Voltage of InP HEMTs’, IEEE Transactions on Device and Materials Reliability 5(2) (2005) (MIT copy, fetch-tool summary).

Ques: How much hydrogen may a hermetic package with GaAs FETs contain?

Northrop Grumman says that the hydrogen must be kept below 100 ppm and preferably below 10 ppm for 10 years at typical commercial conditions, and that getters reach below 10 ppm while a high-temperature bake only reaches 200 to 1000 ppm. Their plot of life against residual hydrogen covers 1 to 10,000 ppm and about $1 \times 1 0^{3}$ to $1 \times 1 0^{7}$ $h$ at a 60 $^{\circ} C$ baseplate. The activation energy of about 0.6 to 0.7 $eV$ means that a 200 $^{\circ} C$ hydrogen test is accelerated by 485 to 1360 times relative to 60 $^{\circ} C$ use (computed), so a 100 $h$ test covers only about 5.5 to 15.5 years of use. The hydrogen comes from the package base materials or the plated metal finishes (Analog Devices). Their HBTs also contain Pt, but show no hydrogen degradation, so the limit applies to FETs. Reference: Northrop Grumman, ‘Effects of Hydrogen on Hermetically Packaged GaAs MMICs’, application note, revision April 2008 (fetch-tool summary); Analog Devices, ‘Hydrogen Effects on Exotic Material Devices’, Version 1.1, 19 August 2019.

Ques: How long do GaAs and InP HBTs last at a junction of 125 degrees?

The reported values span more than an order of magnitude. Lin and Wang project above $7.4 \times 1 0^{6}$ $h$ for InGaP/GaAs HBTs at 75 $kA/c m^{2}$ , from about 3200 $h$ at a 330 $^{\circ} C$ junction with 0.8 $eV$ , which the acceleration of 2770 reproduces as $8.8 \times 1 0^{6}$ $h$ (computed). Wuerfl et al. report typically $3 \times 1 0^{7}$ $h$ at $1 \times 1 0^{5}$ $A/c m^{2}$ for GaInP/GaAs HBTs, and Kashio et al. above $1 \times 1 0^{8}$ $h$ for InP HBTs with 1.5 $eV$ . The differences come from the activation energies, the current densities and the gain criteria (20 or 15 per cent), not only from the devices. All are far above a 10 year mission, $8.8 \times 1 0^{4}$ $h$ , so at 125 $^{\circ} C$ the HBT itself is rarely the limit; the lowest value, $7.4 \times 1 0^{6}$ $h$ , is the safe one for design. Reference: B. Lin and L. Wang, ‘InGaP emitters make HBTs more reliable’, Compound Semiconductor, May 2001 (fetch-tool summary); J. Wuerfl, P. Kurpas, F. Brunner, M. Mai, M. Rudolph and M. Weyers, ‘Degradation properties of MOVPE-grown GaInP/GaAs HBTs under combined temperature and current stressing’, Microelectronics Reliability 41(8), 1103-1108 (2001) (abstract only, FBH page); N. Kashio, K. Kurishima, Y. K. Fukai and S. Yamahata, ‘High-speed, High-reliability 0.5-um-emitter InP-based Heterojunction Bipolar Transistors’, NTT Technical Review 7(12), December 2009 (fetch-tool summary).

Ques: Why does a small change of the ohmic-contact activation energy matter so much?

Because Kayali gives a range of 0.5 to 1.8 $eV$ for the degradation of AuGe/Ni contacts, and the factor between two temperatures depends exponentially on it. Between 100 and 150 $^{\circ} C$ the life changes by 6.3 for 0.5 $eV$ and by 746 for 1.8 $eV$ (computed). This is why he says that the contacts have a reasonable life below 100 $^{\circ} C$ but deteriorate rapidly above 150 $^{\circ} C$ : with a high activation energy the cliff is steep. A life test above 150 $^{\circ} C$ therefore cannot be extrapolated to use without knowing the energy, and a contact process should be qualified at two or three temperatures. The contact resistance of 0.1 to 0.5 $Ω mm$ is set by alloying above 400 $^{\circ} C$ . Reference: S. Kayali, ‘Basic Failure Modes and Mechanisms’, chapter 4 (pp. 89-108) of the JPL GaAs MMIC Reliability Assurance Guideline (parts.jpl.nasa.gov; title of the guideline and year not shown in the text read, references to 1996) (fetch-tool summary).

Ques: How much does gate sinking shift the threshold of a MESFET?

It depends on how far the gate moves into the channel. For a uniformly doped channel the pinch-off voltage is $q N_{D} a^{2} / (2 ε_{s})$ , so a front that moves by $Δ a$ shifts it by about $q N_{D} a Δ a / ε_{s}$ . With an assumed doping of $2 \times 1 0^{17}$ $c m^{- 3}$ and a 100 $nm$ channel, the pinch-off is 1.40 $V$ and 5 $nm$ of sinking shifts it by 0.14 $V$ (computed). The front grows as the square root of $D t$ in a diffusion-limited reaction, as for the diffusion profiles of the Diffusion Furnace page, so the shift slows down with time but does not stop. Kayali says that Au diffusion reduces the channel depth and that Pt or Pd barriers slow it, with grain boundaries as fast paths, but gives no activation energy, so a projection needs its own Arrhenius data. Reference: S. Kayali, ‘Basic Failure Modes and Mechanisms’, chapter 4 (pp. 89-108) of the JPL GaAs MMIC Reliability Assurance Guideline (parts.jpl.nasa.gov; title of the guideline and year not shown in the text read, references to 1996) (fetch-tool summary).

References

H. Livingston (BAE Systems), ‘A Survey of Heterojunction Bipolar Transistor (HBT) Device Reliability’, updated 3 June 2003 (fetch-tool summary).
B. Lin and L. Wang, ‘InGaP emitters make HBTs more reliable’, Compound Semiconductor, May 2001 (fetch-tool summary).
N. Kashio, K. Kurishima, Y. K. Fukai and S. Yamahata, ‘High-speed, High-reliability 0.5-um-emitter InP-based Heterojunction Bipolar Transistors’, NTT Technical Review 7(12), December 2009 (fetch-tool summary).
J. Wuerfl, P. Kurpas, F. Brunner, M. Mai, M. Rudolph and M. Weyers, ‘Degradation properties of MOVPE-grown GaInP/GaAs HBTs under combined temperature and current stressing’, Microelectronics Reliability 41(8), 1103-1108 (2001) (abstract only, FBH page).
Northrop Grumman, ‘Effects of Hydrogen on Hermetically Packaged GaAs MMICs’, application note, revision April 2008 (fetch-tool summary).
S. Kayali, ‘Basic Failure Modes and Mechanisms’, chapter 4 (pp. 89-108) of the JPL GaAs MMIC Reliability Assurance Guideline (parts.jpl.nasa.gov; title of the guideline and year not shown in the text read, references to 1996) (fetch-tool summary).
R. R. Blanchard, J. A. del Alamo and A. Cornet Calveras, ‘Hydrogen-Induced Changes in the Breakdown Voltage of InP HEMTs’, IEEE Transactions on Device and Materials Reliability 5(2) (2005) (MIT copy, fetch-tool summary).
Analog Devices, ‘Hydrogen Effects on Exotic Material Devices’, Version 1.1, 19 August 2019. Available: https://www-cce.cldnet.analog.com/media/en/technical-documentation/application-notes/hydrogen-sensitivity-adi-v3.pdf.
Diffusion Furnace, Metal Semiconductor Field Effect Transistor (MESFET), Conventional Semiconductor’s Material Values, Thermal Stress pages of this vault.

Quartz 5

Explorer

GaAs and InP HEMT, MESFET and HBT