Ageing of VCSELs in non-hermetic 85/85 condition Ageing of VCSELs in non-hermetic 85/85 condition S. Hou, 2014/09/18 Academia Sinica.

Ageing of VCSELs in non-hermetic 85/85 condition

S. Hou, 2014/09/18Academia Sinica

2

Sample preparation− VCSELs of various manufacturers

die/wire bonds by FOCI

− TOSAsTrueLight 10 Gbps on PCB

3

DC light power measurements− V-I-L scan by a LabView setup

VCSELs covered by a large (10×10 mm2) GaAs PIN, Mechanical alignment is requiredNI 6024E PCMCIA to an XP notebook

− Power meter measurement

4

DC bias to VCSELS− DC bias by Agilent E3631A, Keithley 2304A

current kept at ~ 5 mA/ch

− Bias at 1.65 V, 1.75 V, 2.0 V, to VCSELs

channel current measured, bundled.

5

85/85 chamber − Temperature stable ~ 0.1 oC− Humidity stable ~ 0.2% RH− Cooled to 30/55 before opening, to prevent condensation

Burn periods conducted30/50 : 118 hrs85/85 : 12 hrs85/85 : 94 hrs85/85 : 275 hrs85/85 : 316 hrs85/85 : 363 hrs85/85 : 534 hrs

6

Reference samples

− Spare from 85/85 test,to examine systematics in DAQ

1F454.25 Gbps

1F5910 Gbps

1F59

1F5810 Gbps

1F58

− TOSA of Truelight, 3 types

7

Reference samples− FINISAR 2092-001 5 Gbps− a board has two 4x1 arrays

VCSELs are not centered at PIN light collection deviates

8

TOSA in 85/85 tests

A board with12 TOSAof Truelight1F45 4.25 Gb1F58 10 Gb1F59 10 Gb

No obvious lossafter ~1200 hr biased in 85/85

1F45

1F45

1F58

1F59

9

Bare-die VCSLEs of FOCI Lightpeak in 85/85One board having 10 VCSELs of Lightpeak (A-brand) Biased in 85/85 for 1600 hrsMinor degradation within systematic uncertainties

L-I shapes differ due to angle to PIN photo detector

ch closer to PIN center near PIN edge

10

FINISAR 4x1 arrays in 85/85

V850-2902-001, 5 Gbps

1600 hrs in 85/85 − One board, two arrays,

total 8 channels

− One channel showing large degradation, reason not clear

− others are consistent with small degradation

11

Oxide-confined VCSELs, degradation in 85/85

− Three types of Truelight, not expected hermeticTSD-000, 2.5 Gbps, 10 lost 6, after ~400 hrs TSD-008, 4.25 Gbps, 10 lost 4, after ~1100 hrs TSD-051, 10 Gbps, 10 lost 2, after ~1100 hrs

− Degradation in ligth level,then sudden loss of light, drawing current still

ch closer to PIN center near PIN edgeNot sure what cause degradation,, For channels near PIN centerPIN geometrical mattersNot observed in other brands TSD-051

12

Summary

− Environment tests conducted to VCSELs85 oC, 85% RH, up to 1600 hrs

− TOSAs having VCSELs sealed is show negligible degradation

− Bare Die VCSELs of LightPeak A-brand, Finisar arrayshow negligible degradation

− Degradation of Oxide-confined VCSELSfollowed by sudden loss of light

− Two brands of VCSELS supposed for non-hermetic Chip-on-board Failed in room condition within a week

collaborate with the manufacturer to find the cause 1. compare sealed samples with open ones humidity damage2. use current source power supply surge in voltage power supply 3. checking on signs of ESD damage