Extension of the MCP-PMT lifetime

K. Matsuoka (KMI, Nagoya Univ.)S. Hirose, T. Iijima, K. Inami, Y. Kato, K. Kobayashi,

Y. Maeda, R. Omori, K. Suzuki (Nagoya Univ.)

RICH2016Bled, Slovenia

Sep. 6, 2016

Photon sensors in novel RICH detectors

Requirements for the photon sensors are

Not only a spatial resolution to reconstruct Cherenkov

images

Very good time resolution for single photons

<50 ps for Time-Of-Propagation (TOP)

Large photocoverage

High efficiency

Work under a high background

Work in a high B-field

2

Only an MCP-PMT could meet every requirement.

nC

1cosTOP

Quartzp

K

C

MCP-PMT (Micro Channel Plate PMT)

Square shape multi-anode MCP-PMT with a large photocoverage

Developed for the Belle II TOP counter at Nagoya in collaboration with

HAMAMATSU Photonics K.K.

3

Micro channel

e–~

1 k

V /

40

0 m

m

23 mm

10 mm

40

0 m

m

13°

(Cross-section)

Photocathode (NaKSbCs)

MCP x 2

Photon

e–

5.275 mm

4 x 4 anodes

The best time resolution of photon sensors

10 mV

1 ns

Fast signal

KT0117 ch0 2480 V1.9 x 106 gainLimited by the oscilloscope.The actual signal is faster.

Major problem of the MCP-PMT4

Aging of the photocathode

In the electron multiplication, gas/ion is desorbed

from the MCP.

The photocathode is deteriorated by the gas/ion

and the QE is depressed.

The amount of QE depression depends on the accumulated

output charge.

Define the lifetime of the MCP-PMT as an accumulated

output charge 𝑄𝜏 at which QE(𝑄𝜏)/QEinital = 0.8 at 400 nm.

Estimated accumulated output charge for Belle II TOP

dominantly due to beam bkgd.:

~3.7 C/cm2 at 50 ab–1 with 5 x 105 gain

Photocathode

Ion

Gas

We have researched to achieve the lifetime longer than

the estimated accumulated output charge.

How to extend the lifetime

Three steps of approach

1. Block the gas/ion from reaching the photocathode

… Conventional MCP-PMT [NIM A629 (2011) 111]

2. Suppress outgassing from the MCP

… ALD (Atomic Layer Deposition) MCP-PMT (2013~)

3. Reduce residual gas on the MCP

… Life-extended ALD MCP-PMT (2015~)

5

Step 1 Step 2

Ion

Gas

Ceramic

blockAl layer

ALD coating on

the MCP surface

Gas

MCP

Step 3

Gas

MCP

++

Evaluated the lifetime of each type of MCP-PMT

Test of the lifetime

Monitor the QE as a function of the accumulated

output charge of the MCP-PMT.

LED is used to load the output charge, which is measured

by a CAMAC ADC.

QE is monitored as the hit rate by the laser single photons.

6

Reference PMT

MCP-PMTs

Pulse laser (400 nm)

LED (100 kHz)

The QE depression curve is represented by QE 𝑄

QEinital

= 1 − 0.2 𝑄/𝑄𝜏2

Longer lifetime with ALD and much longer with life-extended ALD

Result of the lifetime test7

Life-extended ALD (YH0205)

ALD (KT0074)

Conventional (XM0267)

1.0 5.6 29.4

Result of typical sample of each type

at 4

00 n

m

QE spectrum after the lifetime test

Measured by Xe lamp + monochromator

8

Consistent with the in-situ QE measurement by the laser

at 400 nm.

The QE drops more significantly at longer wavelengths

as the work function of the photocathode increases.

Conventional (XM0267) ALD (KT0074) Life-extended ALD (YH0205)

Lifetime estimation halfway through the test

QE drop of 4 life-extended ALD MCP-PMT samples at 4.0-5.5 C/cm2

was little. Stopped the test to keep them as spares for Belle II TOP.

Estimate the lifetime of those samples by comparing the QE spectrum

with another sample of which lifetime was measured to be 11.2 C/cm2.

1 − 0.2 Τ𝑄 𝑄𝜏2 ≥ 1 − 0.2 Τ3.3 11.2 2

𝑄𝜏 ≥ Τ𝑄 3.3 × 11.2 C/cm2

9

Halfway sample XM0240

Stopped at these

output charge

Lifetime = 11.2 C/cm2

≥ 13.6-18.7 C/cm2

Summary of the measured lifetime10

The lifetime varies broadly sample-by-sample.

Need to measure many samples to evaluate the lifetime.

Succeeded in extending the lifetime significantly.

0

5

10

15

20

25

30

35

Life

tim

e (

C/c

m2)

Sample

12 conventional

0.3-1.7 C/cm2

Average: 1.1 C/cm2

8 ALD

2.5-26.1 C/cm2

Average: 10.4 C/cm2

8 life-extended ALD

>13.6 C/cm2

Estimation of

the lower bound

Belle II TOP

Performance of the ALD MCP-PMT

Due to the emissive coating of the ALD, the ALD MCP

has a larger secondary electron yield (SE yield) than

the conventional MCP.

We have 277 conventional, 231 ALD and 65 life-

extended ALD MCP-PMTs for Belle II TOP.

Systematically studied the performance of the (life-

extended) ALD MCP-PMTs compared with the

conventional ones.

11

HV (V)

SE y

ield

~200~85~65

~2

Rough drawing

ALD

Conventional

Setup of the laser test

Single photon irradiation to each anode one by one.

12

Pico-second

pulse laser

(l = 400 nm)

Slit

MCP-PMT

Slit

Light spot ≈ 1 mm f

ND filters

Reference PMT

MCP-PMT

Laser

Fiber

ADC

Dark boxMoving stage

Variable amp

+19.5~35 dB

–10 dB +33 dB

AmpATT Discrim-inator

Threshold:

–20 mV

TDC

Gain

Define the gain as the mean of the output charge

distribution.

13

HV Gain

2750 V (3.9 x 106)

2650 V (2.0 x 106)

2550 V (1.0 x 106)

KT0449_20140717 ch4

Conventional

ALD

Life-extended ALD

HV for 5 x 105 gainALD MCP-PMT

Lower HV for ALDs to have the same gain owing to

the higher SE yield.

Relative collection efficiency

Count the number of TDC hits.

Correct the laser intensity variation with the reference PMT.

14

Higher collection efficiency of ALDs by ~15%.The variation of the collection efficiency among PMTs of each type

seems to be independent of the HV for the same gain or SE yield.

Conventional

ALD

Life-extended ALD

TTS (Transit Time Spread)

Fit a double Gaussian to the TDC distribution after time-

walk correction.

Define the TTS as s of the primary Gaussian.

15

ALD

Conventional(same gain of 2 x 106)

Normalized to the number of

incident photonsJT0763_20140626 ch6 3460 VKT0162_20140612 ch6 2550 V

Higher collection efficiency is due to increase of the efficiency

for the recoil photo electrons.

Conventional

ALD

Life-extended ALD

RecoilPhotocathode

MCP1

MCP2

Performance under a high rate

After a micro-channel fires, electrons are depleted on

the channel wall.

The channel wall 𝑂(10−16 F) is recharged by the strip

current through a high resistance of the micro-channel

𝑂(1014 Ω).

Dead time of a single micro-channel

𝜏𝑑 = 𝑅𝐶 = 𝑄𝑜𝑢𝑡/𝐼𝑠 = 𝑂(10 ms) for the 2nd MCP

16

Under a high rate of background,

a large fraction of micro-channels

could be dead.(Belle II TOP: ~200 kHz/anode)

Drop of the gain and efficiency,

depending on the MCP resistance.

Strip current

𝐼𝑠~𝑂(10 pA)

Electrons

𝑄𝑜𝑢𝑡~𝑂(0.1 pC)

+++

+++

Test under a high background

Added an LED to the laser test bench to emulate

background single photons of high rate.

17

CLOCK

10 MHz

20 ns width

ATT

–9 ~ –5 dBChange the

light intensity

LED

Tape

to fade light Life

-exte

nd

ed

ALD

MC

P-P

MT

Bkgd. single

photons at

a high rate

LED

MCP-

PMT

Laser

Gain under a high background18

0

1

2

3

4

5

6

7

8

0.001 0.01 0.1 1 10

Ga

in (

x10

5)

Background rate (MHz/anode)5.275 x 5.275 cm2

Gain drops above ≳100 kHz/anode for high R MCPs.

R of 2nd MCP

Be

lle II TO

P

Efficiency under a high background19

0.5

0.6

0.7

0.8

0.9

1

1.1

0.001 0.01 0.1 1 10

Re

lativ

e e

ffic

ien

cy

Background rate (MHz/anode)

Efficiency drops above ~1 MHz/anode.

Be

lle II TO

P

TTS under a high background20

0

50

100

150

200

250

300

350

0.001 0.01 0.1 1 10

TTS (

ps)

Background rate (MHz/anode)

A rough guide

TTS deteriorates above ~200 kHz/anode.

Be

lle II TO

P

Summary

An MCP-PMT is a key photon sensor for novel RICH detectors.

A major concern is a use under a high background:

Lifetime of the photocathode

Lifetime has been extensively improved by three steps:

1. Conventional MCP 1.1 C/cm2 on average of 12 samples

2. ALD MCP 10.4 C/cm2 on average of 8 samples

3. Life-extended ALD MCP >13.6 C/cm2 for all 8 samples

Performance degradation

Drop of the gain and efficiency is little up to 1 MHz/anode for RMCP ≲ 174 MW. (A smaller RMCP is better.*)

TTS deteriorates above ~200 kHz/anode

Independent of RMCP.

Correspond to 3.7 C/cm2/50 ab–1 for 5 x 105 gain at Belle II TOP.

Current limit of background rate to use the MCP-PMT.

We succeeded in developing the MCP-PMT for Belle II TOP, but

further R&D is necessary for next next generation experiments.

21

* Cannot be much smaller to avoid thermal runaway.