Optical links for the CMS Tracker at CERN

J. Troska (on behalf of Karl Gill)

CERN EP Division

jan.troska@cern.ch

Outline

1. Introduction Project

LHC/CMS/Tracker/Optical Links Environment

2: Radiation damage testing at CERN Lasers fibres/connectors photodiodes System considerations

3: Summary

Copies of slides available: http://gill.home.cern.ch/gill/talks.html

jan.troska@cern.ch

LHC at CERN

CERN, Geneva

Large Hadron Collider

27 km circumference7 TeV proton beams

CMS

jan.troska@cern.ch

CMS at CERN/LHC

**

jan.troska@cern.ch

CMS Tracker development

Barrel layer prototype

Forward disk prototype

Layers of silicon microstrip detectors

~10 million detector channels

jan.troska@cern.ch

CMS Tracker analogue optical link

Transmitter - 1310nm InGaAsP EEL Fibres and connectors - Ge-doped SM fibre Receivers - InGaAs 12-way p-i-n plus analogue electronics - rad-hard deep sub-micron

jan.troska@cern.ch

Tracker radiation environment

Charged hadron fluence (/cm2 over ~10yrs)

high collision ratehigh energylarge number of tracks

radiation damage

lifetime >10 years temperature -10C

jan.troska@cern.ch

CMS Tracker readout and control links final system:

PLL Delay

MUX 2:1

Timing

APVamplifierspipelines

128:1 MUX

Detector Hybridprocessingbuffering

TTCRx

ADC

Rx Hybrid

FED

TTCRx

FEC

CCUCCU

CCU CCU

DCU

Control

processingbuffering

Front-End Back-End

TTC

DAQ

Analogue Readout50000 links @ 40MS/s

Digital Control2000 links @40MHz

Parts under test for radiation

damage

Tx Hybrid

jan.troska@cern.ch

Environments Optoelectronics already employed in variety of harsh

radiation environments e.g. civil nuclear and space applications

Totaldose(Gy)

Dose rate (Gy/hr)

1E-21E-2 1E+0 1E+2 1E+4 1E+6

1E+0

1E+2

1E+4

1E+6

1E+8Space (p,e)

Nuclear(, also n)CMS TK

CMS cavern

jan.troska@cern.ch

COTS issues

Extensive use of commercial off-the-shelf components (COTS) in CMS optical links

Benefit from latest industrial developments cheaper “reliable” tested devices

However COTS not made for CMS environment no guarantees of long-life inside CMS

validation testing of COTS mandatory before integration into CMS

jan.troska@cern.ch

COTS components for CMS Tracker links

Some examples:

single fibre and MU connector

1-way InGaAsP edge-emitting lasers on Si-submount with ceramic lid

12-way optical ribbon and MT-connector

96-way cable

jan.troska@cern.ch

Validation procedures for rad-resistance E.g. lasers and photodiodes

irradiation

n irradiation

irradiation

annealingageing

(in-system) lab tests

Highlighted:1999 Market survey

validation tests

(in-system) lab tests

Feedback test results into system specifications radiation damage effects can then be mitigated

jan.troska@cern.ch

Neutron tests at UC Louvain-La-Neuve

deut

eron

s

neutrons

Recent validation tests of laser diodes

~20MeV neutronsflux ~ 5x1010n/cm2/s

fluence ~ 5x1014n/cm2

(Similar to CMS Tracker10 year exposure)

neutrons

Samples stackedinside cold box (-10C)

jan.troska@cern.ch

Radiation test system

Test setup for in-situ measurements

In-situ measurements give powerful capability to extrapolate damage effects to other radiation environments e.g. CMS Tracker, if damage factors of different particles known

Similar test setup for p-i-n and fibre studies

MUX + DMM

I/O register

DAC

I generator

set V

Vout

Vin

photodetector

laserundertest

Mac + Labviewoptical fibre

current

Iout

signal

DataloggerUnit

Control roomIrradiationsource

jan.troska@cern.ch

neutron damage - Italtel/NEC lasersLight output vs current characteristics

before and after neutron irradiation

Fluence = 5x1014n/cm2

temperature =23C

600

500

400

300

200

100

0

optic

al p

ower

, L (

µW

)5040302010

current, I (mA)

LD1 pre-irrad LD2 pre-irrad LD1 post-irrad LD2 post-irrad

Damage effects consistent with build-up ofnon-radiative recombination centres in/around

active laser volume, causing a decrease ininjected charge carrier lifetimes

jan.troska@cern.ch

neutron damage - Italtel/NEC LDDamage vs fluence and time

annealing vs time

50

40

30

20

10

thre

shol

d cu

rren

t (m

A)

2.82.42.01.61.20.80.40.0

irradiation time (hrs)

4.03.02.01.00.0neutron fluence (1014n/cm2)

1.0

0.9

0.8

0.7

0.6

relative effiency, E/E

(0) I

thr LD1 Eff

LD1

Ithr LD2

EffLD2

1.0

0.9

0.8

0.7

0.6

0.5

frac

tion

rem

aini

ng d

amag

e

0.1 1 10annealing time (hrs)

Ithr

LD1

Ithr

LD2

eff LD1 eff LD2

Roughly linearincrease in damage

with fluence

Same annealing dynamics for threshold and

efficiency therefore indicate same underlying

physical damage mechanism

jan.troska@cern.ch

Different vendors compared

Ithr and Eff changes vs neutron fluence

4

2

0

86420neutron fluence, (1014n/cm2)

4

2

0

4

2

0

OptobahnIthr(0)~3.3mA

NortelIthr(0)~10.5mA

NEC (Italtel)Ithr(0)~12.2mA

Test B

Test B

Test B 1.0

0.8

0.6

86420

neutron fluence, (1014

n/cm2)

1.0

0.8

0.6

1.0

0.8

0.6

OptobahnE(0)~12µW/mA

NortelE(0)~8.5µW/mA

NEC (Italtel)E(0)~60µW/mA

Test B

Test B

Test B

similar effects in all 1310nm InGaAsP lasers

jan.troska@cern.ch

Radiation source comparison - Italtel/NEC laser For comparison with other sources:Normalize to fluence = 5x1014/cm2

in 96 hours irradiation

40

30

20

10

0

thre

shol

d in

crea

se,

I thr (

mA

)

543210

fluence, (1014/cm2 )

0.8MeV n0

~6MeV n0

330MeV +

24GeV p+

~20MeV n0

Damage factor ratios:0.8MeV n = 1

~6MeV n = 3.1~20MeV n = 4.9

330MeV pi = 11.524GeV p = 9.4

80

60

40

20

0

dam

age

(% 1

0yrs

ful

l lum

inos

ity)

1086420

LHC operating time (years)

LHC luminosity profile:

year 1: 10%year 2, 33%year 3, 66%

years 4-10, 100%

total damage annual components

Extend to 10 years, taking into account LHC luminosity profile

possible to estimate damage to laser threshold in CMS Tracker:

in worst case, at low radiiIthr = 14mA for Italtel/NEC lasers

jan.troska@cern.ch

Tests of fibre attenuation

Gamma damage (CMS-TK COTS single-mode fibres) at 1310nm

jan.troska@cern.ch

Tests of photodiodes - leakage

leakage current (InGaAs, 6MeV neutrons)

10-8

10-7

10-6

10-5

leak

age

curr

ent,

I leak

(A

)

0.1 1 10neutron fluence, (10

14n/cm

2)

all devices, -5V

Epitaxx (Italtel) Lucent Alcatel Epitaxx Nortel Fermionics

similar damage over many (similar) devices

jan.troska@cern.ch

Photodiodes - response

Photocurrent (InGaAs, 6MeV neutrons)

1.0

0.5

0.01050

neutron fluence, (1014

n/cm2)

1.0

0.5

0.0

1.0

0.5

0.0Fermionics

Epitaxx (FI)

Epitaxx (BI) /Italtel

Tests A and B

Test B

Test C

Significant differences in damage

depends mainly if front or back-illuminated

• front-illuminated better

jan.troska@cern.ch

Photodiode Single-event-upset Bit-error-rate for 80Mbit/s transmission with 59MeV protons in InGaAs p-i-n (D=80m) 10-90 angle, 1-100W optical power flux ~106/cm2/s (similar to that inside CMS Tracker)

10-12

10-10

10-8

10-6

10-4

10-2

100

Bit

-err

or-r

ate

-30 -25 -20 -15 -10

Optical power amplitude at PD (dBm)

p+ (beam off)

p+ (59MeV, 90°)

p+ (59MeV, 45°+)

p+ (59MeV, 30°+)

p+ (59MeV, 10°+) Ionization dominates for angles close

to 90 nuclear recoil dominates for smaller

angles BER inside CMS Tracker similar to

rate due to nuclear recoils should operate at ~100W opt. power

45° 10°90°beam

jan.troska@cern.ch

System considerations

Build in radiation damage compensation into optical link system

Lasers provide adjustable d.c. bias to track threshold increase provide variable gain to compensate for efficiency loss (and other gain

factors) Fibres and connectors

No significant damage Photodiodes

provide leakage current sink provide adjustable gain use sufficient optical power (~100W) to avoid significant SEU effects

jan.troska@cern.ch

Summary

CMS Tracker Optical links project 50000 analogue + 1000 digital links harsh radiation environment, 10 years at -10C COTS components

Extensive validation testing of radiation hardness necessary ionization damage (total dose) displacement damage (total fluence) and annealing reliability SEU

Accumulated knowledge of radiation damage effects compensation built into optical link system confidence of capacity to operate for 10 years inside CMS Tracker

http://cern.ch/cms-tk-opto/

jan.troska@cern.ch

CMS at CERN/LHC

**

jan.troska@cern.ch

Radiation environment in CMS

** high collision ratehigh energylarge number of tracks

radiation damage

require lifetime >10 years operating temperature -10C

fluences up to 2x1014cm-2

(dominated by charged pions)

total dose up to 150kGy

jan.troska@cern.ch

Single-mode optical fiber

Optical attenuator

Bit Error Rate testerOutput signal

Input signal

1310nm laser transmitter

Optical power-meter

Optical receiver circuit

PIN diode

Experimental setup for particle-induced BER

Incident beam

jan.troska@cern.ch

Annealing (current)

damage anneals faster at higher forward bias

0.90

0.80

0.70

0.60

0.50

rem

aini

ng d

amag

e,

I thr(

t)/

I thr(

)

1 10 100annealing time, t (hrs)

dc bias 0 60mA 80mA 100mA

“recombination enhanced annealing”

jan.troska@cern.ch

Reliability irradiated device lifetime > 10 years?? Ageing test at 80C

No additional degradation in irradiated lasers

acc. Factor ~400 relative to -10C operation

lifetime >>10years

60

50

40

30

20

10lase

r th

resh

old

curr

ent,

I thr (

mA

)

40003000200010000time in oven - 1st batch (hrs)

40003000200010000

time in oven - 2nd batch (hrs)

unirrad (batch 1) (10LD) neutron irrad (batch 1) (10 LD) neutron irrad (batch 2) (20 LD)

jan.troska@cern.ch

Fibre attenuation vs fluence

damage actually most likely due to gamma background

‘Neutron’ damage (CMS TK)

jan.troska@cern.ch

Damage picture

Defects introduced into band-gap by displacement damage non-radiative recombination at defects

competes with laser recombination

undoped InGaAsP MQW structure

stsp

nrdefect levels

n-type InP

p-type InP

Ev

Ec

jan.troska@cern.ch

Fibre annealing damage recovers after irradiation (e.g. gamma)

Damage therefore has dose-rate dependence

jan.troska@cern.ch

InGaAs p-i-n characteristics

Output current vs incident power

InGaAs p-i-n -5V

before/after 2x1014/cm2

200

150

100

50

0250200150100500

incident power, P in (µW)

pre-irr post-irr

Increase in Ileak

decrease in Iphoto

jan.troska@cern.ch

leakage current (InGaAs, different particles, 20C)

10-8

10-7

10-6

10-5

10-4

I leak

(A

)

1013

1014

1015

fluence (cm-2

)

330MeV 24GeV p ~6MeV n

all at -5V

higher energy , p more damaging than n

Different particles (leakage)

jan.troska@cern.ch

Different particles (response)

different particles:

1.0

0.8

0.6

0.4

0.2I PC /

I PC(0

) @

100

µW

1013

1014

1015

fluence (cm-2

)

330MeV 24Gev p ~6MeV n

all at -5V

higher energy , p more damaging than n

jan.troska@cern.ch

InGaAs p-i-n annealing After pion irradiation (room T, -5V)

1.0

0.9

0.8

rela

tive

I leak

ann

eal [

I leak

(t)/

I leak

()]

0.1 1 10 100Annealing time (hrs )

1.00

0.98

0.96

relative Iphoto anneal [Iphoto (t)/Iphoto ()]

after 330MeV leakage photocurrentall at -5V

Leakage anneals a little No annealing of response

jan.troska@cern.ch

InGaAs p-i-n reliability irradiated device lifetime > 10 years?? Ageing test at 80C

No additional degradation in irradiated p-i-n’s

lifetime >>10years

1.04

1.02

1.00

0.98

0.96

norm

aliz

ed p

hoto

curr

ent,

I pc(

t)/I p

c(0)

40003000200010000

time in oven - 1st group (hrs)

40003000200010000

time in oven - 2nd group (hrs)

unirrad control (group 1) irrad (group 1) irrad (group 2)

jan.troska@cern.ch

PD SEU

photodiodes sensitive to SEU

strong dependence upon particle type and angle

jan.troska@cern.ch

CMS/LHC at CERN

**

jan.troska@cern.ch

CMS pit excavations (to -80m)

jan.troska@cern.ch

CMS experimental cavern excavations

jan.troska@cern.ch

CMS experiment assembly at surface

jan.troska@cern.ch

LHC Radiation environments: experiments e.g CMS neutrons

(courtesy M. Huhtinen)

jan.troska@cern.ch

LHC Radiation environments: experiments e.g CMS photons

(courtesy M. Huhtinen)

jan.troska@cern.ch

LHC Radiation environments: Trackers (charged hadrons) (courtesy M. Huhtinen)

jan.troska@cern.ch

LHC Radiation environments: Trackers (neutrons) (courtesy M. Huhtinen)