vacuum photodetectors: present and future

EE Photodetectors for SLHC FNAL 20/1108 R M Brown - RAL 1

STFC

RALVacuum photodetectors: present and future

• What do we expect at SLHC?

• VPT properties

• Anticipated losses in VPT performance at LHC- Photocathode fatigue

- Faceplate darkening

• Possible future approaches

• A word of caution

• Summary


STFC

RAL What do we expect at SLHC?

KoutchoukIoP Liverpool

Jun 2007

Inst

anta

neou

s lu

min

osity

(cm

-2s-1

)

Year

SLHCLHC∫L ~ 440 fb-1 ∫L ~ 3040 fb-1

Total ∫ L ~ 3500 fb-1

(107 s/year, 50% efficiency)(cf ECAL TDR: ∫L ~ 500 fb-1)

~ x7


STFC

RAL Dose versus in EE (LHC)


STFC

RAL Neutron fluence (>100keV) in EE


STFC

RAL Fluence and Dose for 500 fb-1

Neutron + charged hadron fluence (cm-2) for E>100 keV

500 fb-1

500 fb-1

Absorbed dose (Gy)

Immediately behind crystals

500 fb-1

500 fb-1


STFC

RAL CMS Vacuum phototriodes

= 26.5 mm

MESH ANODE

Vacuum Phototriode (VPT):Single stage photomultiplier tube with fine mesh grid anode

• Favourable B-field orientation in EE (VPT Axis: 8.5o < || < 25.5o wrt to B )

• Active area of ~ 280 mm2/crystal

• ‘ Bialkali’ (CsK2Sb) photocathode+ dynode coating

• Gain 8 -10 at B = 4 T

• Q.E. ~ 20% at 420 nm

Type PMT188National Research Institute Electron,

St Petersburg


STFC

RAL Gain in a strong magnetic field

0

2

4

6

8

10

12

0 200 400 600 800 1000

Dynode Voltage

Gai

n

V(A)=1000V

V(A)=800V

Gain ~ 10 achieved with:

- High bias voltages: V(A)/V(D) ~ 1000/800

- CsK2Sb coating on dynode

secondary emission coefficient ~20

0 0.5 1.0 1.5 2.0Magnetic field (Tesla)

600

500

400

300

200

100

0

Ano

de r

espo

nse

(arb

itrar

y un

its)

B-field immunity requires a very fine anode mesh Anode pitch =10 m

Primary electrons should pass through the anode and strike the dynode, but secondary electrons should be captured by the anode Anode transparency = 50%

Tilt angle =15O


STFC

RAL Response vs tilt angle

p

a

t

p

a

t

p =10 m(a2/p2) = 0.5

a′

p′

a′

p′

a′

p′

Effect of tilt on anode transparency

Angle scan at 4.0 T (UVa)

Angle (deg)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

-90 -60 -30 0 30 60 90

VPT angle (deg.)

Rel

. A

node

Res

pons

e

measured anoderesponse vs tilt angle

Grid transmissionfactor (t=2.75um)

Angle scan at 1.8 T (RAL)Schematic of anode grid


STFC

RAL Phototube ageing

Photocurrent (nA) L =1034 cm-2s-1

2.9 8.0

2.5 2.5

2.0 0.6

1.6 0.1

0

10

20

30

40

50

60

70

80

90

100

0 5 10 15 20 25 30 35

Time [ Days]

Me

an

[%

of

firs

t re

ad

ing

]

0

4

8

12

16

20

DC

Ga

in

RIE #72 @ B = 0 Tesla, HT = 800/600 V

RIE #72 @ B = 0 Tesla, DC Gain

IK(0) = 200 nA

0

20

40

60

80

100

0 2 4 6 8 10 12

Time[days]

Me

an

[% o

f fi

rst

rea

din

g]

RIE #50, B = 1.8 T, DC LED @ 200 nA

RIE #50, B = 0 T, DC LED @ 200 nA

IK(0) = 200 nA

30 days at IK(0) = 200 nA

~ 650 fb-1 at = 2.9

~ 2000 fb-1 at = 2.5

The fall in anode response is dominated by degradation of the photocathode

The gain remains ~ constant

B = 0

B = 1.8 T

B = 0

~ 10 years ago, ageing tests were made at RAL and at Brunel on 1 inch VPTs from several manufacturers, at B=0 and B=1.8T. Most tubes showed similar behaviour. These plots are for RIE tubes.

VPT Photo-currents at LHC


STFC

RAL Causes of photocathode degradation

- Positive ion bombardment- Cs desorption- Oxidation due to faulty metal in glass seals- Electrolysis of the glass of the window- Other

Not a problem with well-constructed tubes

Bias with cathode at 0V (a/c couple anode)

0

10

20

30

40

50

60

70

80

90

100

0 10 20 30 40Time [ Days]

Re

lativ

e a

no

de

re

spo

nse

.

R=R0 {0.225 exp(-t/0.65) + 0.775 exp(-t/280)}

I0(K) = 200 nA

B = 0

Photocathode lifetime is often expressed as the ‘charge lifetime’, in Coulombs/cm2

Measurements on an RIE tube show a behaviour that is well described by the sum of two exponential terms – indicating two distinct effects.

These are sometimes termed ‘conditioning’ and ‘ageing’.

Assuming linear scaling to a typical EE/LHC photocurrent of 2 nA:

C1 ~0.25 and 1 ~ 65 days

C2 ~0.75 and 2 ~ 3x104 days

1 is consistent with a simple estimate of the time to sweep up the residual gas in the tube (positive ion bombardment)

It is tempting to attribute 2 to Cs desorption


STFC

RAL Positive ion bombardment

e-e-

I+ I+

V(A)1000V

V(D)800V

V(K)0V

e-e-

I+ I+

V(A)1000V

V(D)800V

V(K)0V

The secondary emission coefficient of the dynode ~20.

Positive ion production is dominated by secondary electrons - both in the anode-dynode gap and the anode-cathode gap

Ions strike the cathode with <E(I+)> = 900

eV

Ions strike the dynode with <E(I+)> = 100 eV

Not to scale!!!

Photocathode damage caused by positive ion bombardment increases with ion energy.

Pre-condition by operating the tubes at low bias for ~100d at Ik = 2nA ??(tests are planned but note importance of gain)

(Note: in principle one could precondition the tubes as diodes with V(K) = V(A) =0 and V(D) = -200V, using the dynode as the photocathode. In this configuration all the ions would be swept on to the dynode, which appears to be less sensitive to damage.

However, without internal gain, this would take a prohibitively long time at practical levels of illumination.)

Note also that positive ion damage self-anneals to some extent when a tube is ‘rested’ for several months – so a pre-conditioning strategy would need repeating.


STFC

RAL

0

20

40

60

80

100

300 350 400 450 500 550 600 650 700 750Wavelength (nm)

Tra

ns

mis

sio

n (

%)

Emission spectrumCrystal transmission

Radiation-darkening of faceplates (n & )

Neutron fluence is 7x1014 n/cm2 (Reactor)

Accompanying -dose ~100 kGyRelative loss, weighted by emission, = 23.5%

For comparison, expected exposures at LHC (500 pb-1) at = 3 are:7x1014 n/cm2 and 50 kGy

(Super radiation hard vacuum phototriodes for the CMS endcap ECAL, NIM A535, 2004, 511-516 Yu.I.Gusev, et al.)

-0.02

0

0.02

0.04

0.06

0.08

0.1

250 350 450 550 650 750 (nm)

Induce

d a

bso

rpti

on

Light loss (PWO) ~ 9%

Induced absorption vs for US-49C faceplate (1mm) exposed to 20kGy (60Co)( = 2.6 at LHC)

0

20

40

60

80

100

300 350 400 450 500 550 600 650 700 750Wavelength (nm)

Tra

ns

mis

sio

n

US-49C before irradiationUS-49C after irradiation

PWO emission properties


STFC

RAL‘Ruggedizing’ vacuum photodetectors (1)

Window transparency:

Fused silica (‘quartz’) is extremely radiation hard, but requires ‘graded seals’ increased cost, increased length, increased vulnerability to He ingress.

However, UV-transmitting and Ce-doped glasses with improved radiation resistance are now available

0

20

40

60

80

100

300 350 400 450 500 550 600 650 700 750Wavelength (nm)

Tra

ns

mis

sio

nC-96 before irradiationC-96 after irradiation

Neutron fluence 7x1014 n/cm2

Accompanying -dose ~100 kGy

Yu.I.Gusev et al. VCI Conf, Vienna, 2004

P.R.Hobson et al., Astroparticle, Particle and Space Physics, Detectors and Medical Physics Applications, Como, 2003

US-49A faceplate exposed to 1016 n/cm2 together with a dose of 1600 250 kGy

[Yu.I. Gusev et al., NIM A 581, 438, (2007)]

Unfolding neutron damage using extrapolated 60Co data (and ignoring s from induced activity in the glass):

T/T0 (neutron) < 15% for EE < 3.0 at SLHC


STFC

RAL‘Ruggedizing’ vacuum photodetectors (2)

Positive ion damage (‘conditioning’):

Improve the vacuum: for example, incorporate a getter– tests on a single device during R&D for CMS showed a marked improvement.

Operate at low bias voltage- incompatible with large internal gain use vacuum photodiodes?

Caesium desorption:

Popular high efficiency photocathodes for visible light almost all incorporate Cs.

However, alternatives are available:E.g. ‘High temperature bi-alkali’ (Na2KSb) (used in oil-well logging)- Q.E. ~16% at 400 nm.


STFC

RALEE replacement/upgrade??

PbWO4 LYSO ?

emiss ~ (380-460) nm, LY(LYSO) ~ 200 x LY(PWO)

What photo-detector?

Silicon devices:

Neutron damage high leakage currents amplifier noise?

‘Nuclear counter effect’ (for a simple photodiode, direct sensitivity to shower leakage particles >> sensitivity to scintillation light high energy tail on energy measurement)

APD?

Vacuum devices: - Good match to biakali photocathode - Internal gain not necessary

Vacuum photodiode?e-

I+

V(A)<100V

V(K)0V

Na2KSb

e-

I+

V(A)<100V

V(K)0V

Na2KSb


STFC

RAL EE Activation

0 100 200 300 400z(cm)

0

100

150

50

r(cm)

Estimated dose rate in Sv/h after 60 d at L = 5x1033cm-2s-1 and 1 d cooling. (CMS closed)

After 4 months cooiling the dose rates are ~2.5x lower

150

54

24

Occupational dose limit: 20 mSv/yr(Note: this is the legal limit, the normal CERN limit is6 mSv/yr – except for the (very few) ‘Class A’ workers)

Assume induced activity levels at SLHC ~10xLHC Time to Annual limit at = 3 is ~12 h

LHC (ECAL TDR)


STFC

RAL Layout of EE elements


STFC

RALPartial endcap upgrade for SLHC?

At VPTs:

Dose(=2.2)/Dose(=3.0) ~1/10 (neutron fluence ~1/3)

~25% (18/71) Supercrystals are at >2.2

Only replace detectors at small radius ?Very challenging because of complexity of EE construction and high radiation levels Needs detailed study


STFC

RAL Summary

EE photodetectors at small radius will be significantly degraded after 500 fb-

1

(Anode response ~ 50% at = 3.0)

Development of ‘ruggedized’ vacuum photo-detectors appears feasible(Interest from Brunel, RAL, UVa……)

A partial replacement of EE would be challenging

vacuum photodetectors: present and future

Documents