This page covers the degradation of semiconductor lasers: edge-emitting Fabry-Perot, DFB and high-power ridge-waveguide lasers, and oxide-confined vertical-cavity surface-emitting lasers (VCSELs). It covers gradual wear-out and its acceleration by current and temperature, the lognormal life statistics and random failures used in qualification, catastrophic optical damage at the facet, dark-line defects, and electrostatic and electrical overstress. The device physics is on the laser device pages of this vault, the ESD models on the Electrical Stress page, the displacement-damage models on the Radiation page and the life statistics on the Failure Rate page.

1. Why It Happens

Trigger: Drive current and junction temperature (ESA; IEEE 802.3dh deck); high optical power density at the output facet of edge emitters (Boschker et al.; Zhao et al.); ESD and switching transients during handling and test (ESA).
Mechanism, gradual wear-out: Non-radiative defects grow in the active region, which raises the threshold current and lowers the output; in edge emitters this is tracked as a threshold rise (Vilokkinen et al.).
Mechanism, dark-line defects: In oxide VCSELs, climb dislocations start at the edge of the mirror oxide and grow as <100> dark-line defects, which can take years to appear; GaAs/AlGaAs is more prone than InP-based or InGaAs/GaAs lasers because of mid-gap recombination centres and self-propagating dislocations (ESA).
Mechanism, catastrophic optical damage: Light absorbed at the facet, enhanced by surface recombination and strain-induced band-gap shrinkage, heats the facet until thermal runaway melts it; facet oxidation drives the absorption (Zhao et al.); oxides and defects at a passivation interface lower the facet stability (Boschker et al.).
What sets the rate: The junction temperature, with 0.4 to 1.2 $eV$ by source and device (Vilokkinen et al.; ESA; IEEE 802.3dh deck); the current density, with exponents from 1 to about 10 (ESA; IEEE 802.3dh deck); the facet power density (Boschker et al.).
Failure criterion: A 50 per cent rise of the threshold current (Vilokkinen et al.); a cumulative failure fraction such as 1 ppm or 1 per cent at the end of the mission (IEEE 802.3dh deck); a sudden loss of output at COD (Boschker et al.).

2. Where to Look

Location: The edge of the oxide aperture in VCSELs (ESA); the front facet and its coating or passivation interface in edge emitters (Boschker et al.; Zhao et al.); the active region for gradual degradation (Vilokkinen et al.).
Signature: A slow threshold rise at constant output power (Vilokkinen et al.); an abrupt failure for random ‘maverick’ failures and COD (ESA; Boschker et al.); dark lines in electroluminescence images for dislocation growth (ESA; background knowledge); a higher reverse leakage after ESD or EOS (ESA).
Confirm with: Life tests at several currents and temperatures with lognormal analysis (IEEE 802.3dh deck; Vilokkinen et al.); step-stress of power to find the COD limit (Boschker et al.); reverse-bias leakage after handling (ESA); constructional analysis (ESA).
Look-alikes: Self-heating, which makes a current exponent look large in a model without the junction temperature (IEEE 802.3dh deck); ESD damage that appears months after burn-in and looks like a random failure (ESA).

3. Prevention

Design: Choose the operating current and aperture for the mission temperature profile, since a smaller aperture raises current density and thermal resistance (ESA; IEEE 802.3dh deck); keep the facet power density below the demonstrated limit, about 42 to 58 $MW/c m^{2}$ for the passivated lasers of Boschker et al.; consider 980 $nm$ InGaAs VCSELs, which showed no wear-out in the TRUMPF tests (IEEE 802.3dh deck).
Material and process: Passivate facets on clean surfaces, since UHV-cleaved ZnSe-passivated facets were three times more stable than hydrogen-cleaned ones (Boschker et al.); use graded oxide layers or inactivate the oxide edge against dislocations (ESA).
Operation: Protect against ESD and switching transients in assembly and test (ESA); keep the case cool, since each 10 $K$ matters at 1.2 $eV$ (IEEE 802.3dh deck).
Screening and test: Burn-in to remove infant failures, up-screening of out-of-family parts and a reverse-leakage test (ESA; Vilokkinen et al.); qualification with enough device hours to bound the random failure rate (Vilokkinen et al.).

Wear-out and Life Statistics

Name	Formula	Related Formulae	Variables/Constants
Acceleration of laser wear-out by current and temperature	$A F = (\frac{I _{a cc}}{I _{u se}})^{n} exp [\frac{E _{a}}{k _{B}} (\frac{1}{T _{J, u se}} - \frac{1}{T _{J, a cc}})]$	$T_{J} = T_{A} + R_{t h} (I V - P_{o}), T T F_{x} = C J^{- n} e^{E_{a} / (k_{B} T_{J})}$	$A F : a cce l er a t i o n f a c t or$ $I_{a cc} an d I_{u se} : s t r ess an d u se c u r r e n t s$ $n : c u r r e n t e x p o n e n t$ $E_{a} : a c t i v a t i o n e n er g y$ $k_{B} : B o l t z mann co n s t an t (8.617 \times 1 0^{- 5} eV / K)$ $T_{J, u se} an d T_{J, a cc} : ab so l u t e j u n c t i o n t e m p er a t u r es$ $T_{A} : ambi e n t (or s u b s t r a t e) t e m p er a t u r e$ $R_{t h} : t h er ma l r es i s t an ce$ $V : v o l t a g e$ $P_{o} : o pt i c a l o u tp u t$ $T T F_{x} : t im e t o x p er ce n t f ai l e d$ $C : co n s t an t$ $J : c u r r e n t d e n s i t y; E S A (W i tt i g) : n o f ab o u t 2 f or V C S E L s an d 1 f or e d g e e mi tt er s, E_{a} o f ab o u t 0.7 eV, r e l iabi l i t y w or se n s ab o u t a s I^{2}, a s ma l l er o x i d e a p er t u r e m os tl y w or se n s r e l iabi l i t y t h r o ug h hi g h er t h er ma l r es i s t an ce an d c u r r e n t d e n s i t y, an d ab o u t 50 p er ce n t f ai l u r es n ee d m or e t han a y e a r a t A F o f ab o u t 100; I E E E 802.3 d h d ec k : E_{a} = 1.18 eV an d n = 1.64 (T R U M P F), 1.15 eV an d 8.21 (B r o a d co m), 1.15 eV an d 9.8 (I I - V I), t h e hi g h n p r o bab l y ab sor bin g se l f - h e a t in g, J_{O P} = 18.2 kA/c m^{2} a t 7 mA f or a 7.0 μ m a p er t u r e, an d T R U M P F 850 nm s t r ess l i v es (50 p er ce n t f ai l e d) o f 853 h a t 140^{\circ} C an d 10 mA, 561 h a t 155^{\circ} C an d 8 mA an d 307 h a t 170^{\circ} C an d 6 mA$ $(co m p u t e d : 7 mA in 7 μ m i s 18.2 kA/c m^{2}; w i t h T R U M P F^{'} s E_{a} an d n o n s u b s t r a t e t e m p er a t u r es, A F = 1550 f r o m 140^{\circ} C an d 10 mA t o 70^{\circ} C an d 7 mA, t_{50} = 1.3 \times 1 0^{6} h an d t_{1} = 4.1 \times 1 0^{5} h f or σ = 0.5; t h e s am e m o d e l p r e d i c t s 385 an d 209 h f or t h e o t h er tw o s t r esses a g ain s t 561 an d 307 h m e a s u r e d, s in ce l o w er c u r r e n t s a l so r u n coo l er; E S A^{'} s 0.7 eV an d n = 2 w o u l d g i v e A F = 113 o n l y)$
Lognormal failure percentiles	$F (t) = Φ (\frac{l n t - μ}{σ}), t_{x} = t_{50} e^{z_{x} σ}$	$λ_{F I T} (t) = 1 0^{9} \frac{f ( t )}{1 - F ( t )}$	$F (t) : c u m u l a t i v e f r a c t i o n f ai l e d$ $Φ : s t an d a r d n or ma l d i s t r ib u t i o n$ $μ : m e an o f l n t$ $σ : l o g n or ma l s ha p e$ $t_{x} : t im e t o f r a c t i o n x$ $t_{50} : m e d ian l i f e$ $z_{x} : s t an d a r d n or ma l q u an t i l e (- 2.326 f or 1 p er ce n t, - 4.753 f or 1 pp m)$ $λ_{F I T} : ha z a r d r a t e in f ai l u r es p er 1 \times 1 0^{9} d e v i ce h o u r s, f (t) : d e n s i t y; I E E E 802.3 d h d ec k : σ = 0.5 (T R U M P F), 0.8 (B r o a d co m) an d 0.52 (I I - V I), a 15 y e a r a u t o m o t i v e mi ss i o n o f 32000 h w i t h 1 p er ce n t o f t h e t im e a t 105^{\circ} C, 1 pp m a f t er ab o u t 1.1 y e a r s a t 105^{\circ} C an d 28.8 y e a r s a t 70^{\circ} C ambi e n t (T R U M P F), T T F_{1} o f ab o u t 100 y e a r s a t 70^{\circ} C s u b s t r a t e (B r o a d co m), T T F_{1} ab o v e 3.7 y e a r s a t 7 mA an d 105^{\circ} C (I I - V I), an d r an d o m f ai l u r es a ss u m e d b e l o w 10 F I T; E S A : F ini s a r u ses σ = 0.7 (0.6 in an o t h er b u l l e t)$ $(co m p u t e d : t_{1} / t_{50} = 0.31 f or σ = 0.5 an d 0.16 f or 0.8, an d t_{1 pp m} / t_{50} = 0.093 an d 0.022, so a w i d er d i s t r ib u t i o n n ee d s a f o u r t im es l o n g er m e d ian f or t h e s am e pp m l i f e; t h e 1.1 t o 28.8 y e a r r a t i o o f 26 co m p a r es w i t h an A r r h e ni u s f a c t or o f 40 b e tw ee n 105 an d 70^{\circ} C a t 1.18 eV)$
Gradual degradation of edge-emitting lasers	$\frac{I _{t h} ( t ) - I _{t h, 0}}{I _{t h, 0}} = r t, t_{E O L} = \frac{0.5}{r}$	$M L_{2} = M L_{1} exp [\frac{E _{a}}{k _{B}} (\frac{1}{T _{2}} - \frac{1}{T _{1}})]$	$I_{t h} : t h r es h o l d c u r r e n t$ $I_{t h, 0} : ini t ia l v a l u e$ $r : r e l a t i v e d r i f t r a t e f r o m a l in e a r f i t$ $t_{E O L} : e n d o f l i f e a t a 50 p er ce n t r i se$ $M L_{1} an d M L_{2} : m e d ian l i v es a t ab so l u t e t e m p er a t u r es T_{1} an d T_{2}$ $E_{a} : a c t i v a t i o n e n er g y$ $k_{B} : B o l t z mann co n s t an t (8.617 \times 1 0^{- 5} eV / K); V i l o k k in e n e t a l . (1.3 μ m A l G a I n A s l a ser s, A P C a t 85^{\circ} C, 13 l a ser s a t 7 mW f or o v er 110000 d e v i ce h o u r s an d 11 a t 10 mW f or o v er 100000) : e n d o f l i f e i s a 50 p er ce n t r i se o f I_{t h}, f ai l u r e t im es a r e l o g n or ma l an d r e a d a t 50 p er ce n t, E_{a} = 0.4 eV i s t h e T e l cor d ia d e f a u l t, n o t m e a s u r e d, m e d ian l i v es o f 1.79 \times 1 0^{6} h (204 y e a r s) an d 1.58 \times 1 0^{6} h (180 y e a r s) a t 85^{\circ} C, 2764 an d 1312 y e a r s a t 25 an d 40^{\circ} C f or t h e 2.5 G bi t / s d es i g n, b u r n - in a t 80 mA an d 100^{\circ} C, w e a r - o u t 86 t o 148 F I T a t 40^{\circ} C f or 5 t o 20 y e a r d es i g n l i v es$ $(co m p u t e d : a t 0.4 eV t h e f a c t or f r o m 85^{\circ} C i s 6.44 t o 40^{\circ} C an d 13.6 t o 25^{\circ} C, g i v in g 1314 an d 2769 y e a r s f r o m 204 y e a r s, ma t c hin g t h e p a p er)$
Random failure rate from a test without failures	$λ = 1 0^{9} \frac{g ( N , C L )}{t _{t o t}} FIT$	$g (0, 60%) = 0.92, g (0, 90%) = 2.30$	$λ : r an d o m f ai l u r e r a t e$ $g : co n f i d e n ce co n s t an t (ha l f t h e c hi - s q u a r e q u an t i l e w i t h 2 N + 2 d e g r ees o f f r ee d o m)$ $N : r an d o m f ai l u r es$ $C L : co n f i d e n ce l e v e l$ $t_{t o t} : t o t a l d e v i ce h o u r s; V i l o k k in e n e t a l . : r an d o m f ai l u r e r a t es a t 40^{\circ} C o f 1145 an d 3616 F I T a t 60 p er ce n t an d 2862 an d 6963 F I T a t 90 p er ce n t co n f i d e n ce, w i t h 0 r an d o m f ai l u r es in t h e 2.5 G bi t / s g r o u p an d 1 in t h e 10 G bi t / s g r o u p$ $(co m p u t e d : 110000 h w i t h o u t f ai l u r e a t 85^{\circ} C g i v es 8300 F I T a t 60 p er ce n t an d 20 900 F I T a t 90 p er ce n t, or 1290 an d 3250 F I T a t 40^{\circ} C w i t h 0.4 eV, c l ose t o t h e p a p e r^{'} s 1145 an d 2862; t h e r an d o m r a t e i s ab o u t t e n t im es t h e w e a r - o u t r a t e o f 109 F I T)$

Catastrophic and Electrical Damage

Name	Formula	Related Formulae	Variables/Constants
Facet power density and catastrophic optical damage	$S_{f a ce t} = \frac{P _{o}}{w d _{e f f}}$	$P_{C O D} = S_{C O D} w d_{e f f}$	$S_{f a ce t} : o pt i c a l p o w er d e n s i t y a t t h e f r o n t f a ce t$ $P_{o} : o u tp u t p o w er$ $w : l a t er a l m o d e w i d t h (r i d g e w i d t h)$ $d_{e f f} : e f f ec t i v e v er t i c a l m o d e s i z e (a c t i v e - l a y er t hi c k n ess d i v i d e d b y t h e co n f in e m e n t f a c t or)$ $P_{C O D} : p o w er a t c a t a s t r o p hi c o pt i c a l d ama g e$ $S_{C O D} : cr i t i c a l f a ce t p o w er d e n s i t y; B osc hk er e t a l . (980 an d 1064 nm r i d g e l a ser s, 5 μ m r i d g e, 3.9 mm, Z n S e f a ce t p a ss i v a t i o n 24.5 nm) : ma x im u m f a ce t p o w er d e n s i t i es o f 58 MW/c m^{2} (1064 nm D Q W) an d 42 MW/c m^{2} (980 nm an d 1064 nm T Q W), n o C O M D u p t o 8 A (4.1 kA/c m^{2}), 980 nm l a ser s ab o v e 1.75 W f or m or e t han 2000 h, m or e t han t h r ee t im es hi g h er f a ce t s t abi l i t y f or U H V - c l e a v e d t han f or h y d r o g e n - c l e an e d 1064 nm T Q W f a ce t s, an d o x y g e n a t t h e Z n S e / (A l, G a) A s in t er f a ce a f t er h y d r o g e n c l e anin g; Z ha o e t a l . (s im u l a t i o n) : f a ce t ab sor pt i o n, s t r ain - in d u ce d ban d - g a p s h r ink a g e an d s u r f a ce r eco mbina t i o n h e a t t h e f a ce t, C O M D b e tw ee n 2 an d 2.05 A an d r u na w a y a t 2.5 A, w i t h f a ce t o x i d a t i o n a s t h e main c a u se$ $(co m p u t e d : 2 W a t 42 MW/c m^{2} o v er a 5 μ m r i d g e im pl i es d_{e f f} o f ab o u t 0.95 μ m)$
Electrostatic and electrical overstress of VCSELs	$V_{H B M, f ai l} \approx 100 to 200 V$	$I_{R} (V_{R}) screen after handling$	$V_{H B M, f ai l} : h u man - b o d y - m o d e l E S D v o l t a g e t ha t c an d ama g e a V C S E L, I_{R} (V_{R}) : r e v er se l e ak a g e a t a r e v er se v o l t a g e, u se d t o d e t ec t E S D an d E O S d ama g e; E S A (W i tt i g) : V C S E L s a r e se n s i t i v e t o l o n g E S D p u l ses, H B M d ama g e c an occ u r a t ab o u t 100 t o 200 V an d t h e d e v i ce ma y f ai l m o n t h s a f t er b u r n - in, E O S f r o m a s l i ttl e a s 5 V s w i t c hin g t r an s i e n t s o f t es t e q u i p m e n t c an c a u se d ama g e, t h e m os t co mm o n w e a r - o u t so u r ce i s < 100 > d a r k - l in e d e f ec t s f r o m c l imb d i s l oc a t i o n s a t t h e e d g e o f t h e mi r r or o x i d e, w hi c h c an t ak e y e a r s t o a pp e a r, G a A s / A l G a A s i s l ess r e l iab l e t han I n G a A s P / I n P or I n G a A s / G a A s, an d V C S E L s a r e r a d ia t i o n t o l er an t b ec a u se o f t h e i r s ma l l a c t i v e a r e a, hi g h c u r r e n t d e n s i t y an d s h or t c a r r i er l i f e t im e, w i t h d i s pl a ce m e n t d ama g e ann e a l in g a t hi g h c u r r e n t, so^{'} o f f^{'} i s t h e w or s t c a se f or r a d ia t i o n an d^{'} o n^{'} f or h u mi d i t y$ $(co m p u t e d : n o n e; t h e E S D m o d e l s a r e o n t h e E l ec t r i c a l S t r ess p a g e)$

Important Questions

Ques: Is the activation energy of VCSEL wear-out 0.7 or about 1.2 eV?

The ESA guide quotes about 0.7 $eV$ with a current exponent of about 2 for VCSELs, the values Finisar used, while the 2022 IEEE 802.3dh deck reports 1.18 $eV$ and $n = 1.64$ for TRUMPF, and 1.15 $eV$ with $n = 8.21$ and 9.8 for Broadcom and II-VI. The newer values come from fits to recent 850 $nm$ devices with junction temperatures, and the deck notes that the high current exponents probably absorb self-heating. For a modern oxide VCSEL, about 1.15 to 1.2 $eV$ is the more likely value (my reading). The choice matters greatly: from 140 $^{\circ} C$ and 10 $mA$ to 70 $^{\circ} C$ and 7 $mA$ , TRUMPF’s parameters give an acceleration of 1550, while 0.7 $eV$ and $n = 2$ give 113 (computed), so the older values predict a much shorter life. Reference: S. Wittig (ESA ESTEC TEC-QTC), ‘Deploying VCSELs for space applications: reliability and procurement guide’, Intra-satellite fibre optic workshop, 10-11 December 2015 (slides; acceleration equation partly garbled, fetch-tool summary); R. Perez-Aranda (KDPOF), J. Pankert and R. King (TRUMPF Photonic Components), ‘VCSELs: Performance measurements and reliability analysis’, IEEE 802.3dh ad hoc meeting, 5 October 2022, with data of Broadcom (Giovane, 2021) and II-VI (Hoser, 2022) (slides, fetch-tool summary).

Ques: Can 850 nm VCSELs last a 15-year automotive mission?

Only with margin, according to the IEEE 802.3dh deck. Its mission is 32 000 $h$ over 15 years, with 8 per cent of the time at 100 $^{\circ} C$ and 1 per cent at 105 $^{\circ} C$ . With TRUMPF’s model, 1 ppm of wear-out failures is reached after about 1.1 years at 105 $^{\circ} C$ ambient but 28.8 years at 70 $^{\circ} C$ , and the cumulative mission equals 2084 $h$ at the maximum temperature. Broadcom finds a 1 per cent failure time of about 100 years at 70 $^{\circ} C$ substrate and II-VI more than 3.7 years at 105 $^{\circ} C$ . The deck concludes that wear-out is the limiting factor. The hot hours dominate: the 1.1 to 28.8 year ratio of 26 is close to an Arrhenius factor of 40 between 105 and 70 $^{\circ} C$ at 1.18 $eV$ (computed). Reference: R. Perez-Aranda (KDPOF), J. Pankert and R. King (TRUMPF Photonic Components), ‘VCSELs: Performance measurements and reliability analysis’, IEEE 802.3dh ad hoc meeting, 5 October 2022, with data of Broadcom (Giovane, 2021) and II-VI (Hoser, 2022) (slides, fetch-tool summary).

Ques: How is the life of an edge-emitting laser extrapolated from a test at 85 degrees Celsius?

Vilokkinen et al. ran 1.3 $μ m$ AlGaInAs lasers at constant output power at 85 $^{\circ} C$ , fitted a straight line to each threshold current against time and took end of life at a 50 per cent rise. The failure times are lognormal, and the median was read at 50 per cent: $1.79 \times 1 0^{6}$ $h$ , 204 years, for the 2.5 Gbit/s design (the abstract gives $1.57 \times 1 0^{6}$ $h$ for the other design, against $1.58 \times 1 0^{6}$ in the text). With the Telcordia default of 0.4 $eV$ , which they did not measure, the factor to 40 $^{\circ} C$ is 6.44 and to 25 $^{\circ} C$ 13.6, giving 1314 and 2769 years (computed), matching their 1312 and 2764 years. Only devices with a degradation trend entered the fit, 9 of 13 and 10 of 11. Reference: V. Vilokkinen, P. Savolainen and P. Sipila, ‘Reliability analysis of AlGaInAs lasers at 1.3 um’, Electronics Letters 40(23) (2004), doi:10.1049/el:20046314 (fetch-tool summary).

Ques: Why does the random failure rate dominate the laser qualification result?

Because a test without failures can only bound it, and the bound depends on the device hours, not on the wear-out physics. Vilokkinen et al. report wear-out rates of 86 to 148 FIT at 40 $^{\circ} C$ , but random rates of 1145 FIT at 60 per cent confidence and 2862 FIT at 90 per cent for the 2.5 Gbit/s design, which had no random failure. From $λ = 1 0^{9} g / t_{t o t}$ with $g = 0.92$ and 2.30 for zero failures, 110 000 device hours at 85 $^{\circ} C$ give 8300 and 20 900 FIT, or 1290 and 3250 FIT at 40 $^{\circ} C$ with 0.4 $eV$ (computed), close to the paper. The random bound is ten times the wear-out rate (computed), so more device hours, not longer tests, improve the qualified rate. Reference: V. Vilokkinen, P. Savolainen and P. Sipila, ‘Reliability analysis of AlGaInAs lasers at 1.3 um’, Electronics Letters 40(23) (2004), doi:10.1049/el:20046314 (fetch-tool summary); Failure Rate and the Bathtub Curve page of this vault.

Ques: What limits the power of a high-power edge-emitting laser before catastrophic optical damage?

The optical power density at the front facet. Light absorbed at the facet is enhanced by surface recombination and by band-gap shrinkage under strain, heats the facet and can run away thermally; Zhao et al. find in simulation a COMD threshold between 2 and 2.05 $A$ for their structure and identify facet oxidation as the main cause. Boschker et al. passivated 980 and 1064 $nm$ ridge lasers with 24.5 $nm$ of ZnSe and reached facet power densities of 42 to 58 $MW/c m^{2}$ without COMD up to 8 $A$ ; 2 $W$ at 42 $MW/c m^{2}$ over a 5 $μ m$ ridge implies an effective mode height of about 0.95 $μ m$ (computed). UHV-cleaved facets were more than three times more stable than hydrogen-cleaned ones, which showed oxygen at the ZnSe interface. Reference: T. Zhao, Z. Li, T. Li, P. Lu, Y. Qu, B. Bo, G. Liu, X. Ma and Y. Wang, ‘Analysis of Facet Temperature Distribution of Semiconductor Lasers’, Proc. MEBE 2015, Atlantis Press, 115-118 (simulation; fetch-tool summary); J. E. Boschker, U. Spengler, P. Ressel, M. Schmidbauer, A. Mogilatenko and A. Knigge, ‘Stability of ZnSe-Passivated Laser Facets Cleaved in Air and in Ultra-High Vacuum’, IEEE Photonics Journal 14(3), 1531606 (2022) (fetch-tool summary).

Ques: What are dark-line defects in VCSELs, and why are GaAs devices more prone to them?

They are lines of non-radiative recombination along <100> directions, formed by climb dislocations that start at the edge of the mirror oxide, where the oxidation creates stress and cracking. The ESA guide names them as the most common wear-out source of oxide VCSELs and notes that they can take years to appear, so they show up as late failures rather than at burn-in. GaAs/AlGaAs devices are worse than InGaAsP/InP lasers at 1.3 and 1.55 $μ m$ or InGaAs/GaAs lasers at 920 to 980 $nm$ , because GaAs has mid-gap recombination centres from dangling bonds and supports self-propagating dislocations. Graded oxide layers, electrically inactivated oxide edges or InP materials reduce the risk, and a smaller aperture mostly worsens reliability. Reference: S. Wittig (ESA ESTEC TEC-QTC), ‘Deploying VCSELs for space applications: reliability and procurement guide’, Intra-satellite fibre optic workshop, 10-11 December 2015 (slides; acceleration equation partly garbled, fetch-tool summary).

Ques: How sensitive are VCSELs to electrostatic discharge?

Very sensitive, and the damage can be latent. The ESA guide states that VCSELs are sensitive to long pulses, that human-body-model damage can occur at about 100 to 200 $V$ , and that a damaged device may fail months after burn-in, when it looks like a random field failure. Electrical overstress with longer pulses, from as little as 5 $V$ switching transients of test equipment, can also cause damage. A reverse-bias leakage test can detect ESD and EOS events, so it is part of up-screening together with removing out-of-family parts. The small active area that makes VCSELs vulnerable to ESD also makes them tolerant of radiation, since displacement damage partly anneals at high current, so an unbiased VCSEL is the worst case for radiation. Reference: S. Wittig (ESA ESTEC TEC-QTC), ‘Deploying VCSELs for space applications: reliability and procurement guide’, Intra-satellite fibre optic workshop, 10-11 December 2015 (slides; acceleration equation partly garbled, fetch-tool summary); Electrical Stress page of this vault; Radiation page of this vault.

Quartz 5

Explorer

Semiconductor Lasers (Edge-Emitting and VCSEL)