Single-avalanche response mesurement method for MPGD … · Single-avalanche response mesurement method for MPGD detectors András László [email protected] Wigner RCP,

Single-avalanche response mesurement methodfor MPGD detectors

András László

[email protected] RCP, Budapest, Hungary

joint work with Gergo Hamar, Gábor Kiss, Dezso Varga

ISSP2015, Erice, Italy

Single-avalanche response mesurement method for MPGD detectors – p. 1

Gas based particle detectors

Gas based particle detectors are widely used in HEP experiments.

Relatively low price, very low material budget in the way of particles.

Signal amplification based on electron multiplication by avalanche process.

High field necessary for avalanche process obtained in various ways:

Classical solution: Multi-Wire Proportional Chamber, high field near wire.

Newer solutions: Micro Pattern Gas Detectors (such as GEMs),high field by microstructures.


Intrinsic limits to energy deposit resolution

When particles traverse the sensitive volume, they leave trace of ionization.

Before electronic detection, large signal amplification (∼ 104) is needed.

This amplification done by electron multiplication avalanche process in high electric field.

Response to ionization energy deposit depends on the properties of avalanche process.

Measurement of ionization energy deposit (dE/dx) used for particle identification⇒ good response resolution to dE/dx is important.


Idealized large avalanches: system does not have memory, any intermediate electron inthe avalanche generates the same response.⇒ Final population has exponential (geometric) distribution.⇒ σ/mean = 1 for the response distribution.⇒ Intrinsic limit to dE/dx resolution after amplification process.

Lot is known on the behavior on classic MWPC avalanches (NIM 75 161, NIM 89 155):available space and time for avalanche development is large.⇒ Large, close-to-idelized avalanches with large number of generations.⇒ Approximately exponential single-avalanche response.⇒ σ/mean ≈ 1.

For MPGD detectors? Micromegas: T.Zerguerras et al. NIM A608 397. Deviation fromexponential due to limited space and time for avalanche development, σ/mean < 1.

What about GEMs? This talk is about that. Motivated also by ALICE TPC upgrade.


Exp. setup for multiplication response measurement

working gasvolume

quartz window

drift region

gas envelope

1.5mm

5.0mm

8.0mm

8.5mm

4.0m

m

4.0m

m

1.0m

m

drift cathode wires (30um)

field wires (100um) / sense wires (20um)

drifting photoelectrons

pulsed UV LED (defocused)

(gold plated GEM, no multiplication)photoelectron source

amplifier GEM under study(effective gain=1−100)

readout MWPC (high gain stage)MWPC cathode pad


Defocused pulsed UV LED used to extract photoelectrons (PEs)from surface of gold plated classic GEM.

These PEs were transferred to the reverse side,providing a source of laminarly drifting PEs.

Electron multiplication takes place in a subsequent classic GEM (effective gain 1-100).

Multiplied electrons enter into an MWPC amplifier region to enhance their signal.

Electronic FEE amplifier detects the MWPC-amplified signal from sense wires.


Front-End Electronics and LED electronics from the Leopard project(NIM A694 16, POS TIPP2014 056).

MWPC design based on CCC concept (NIM A648 163, NIM A698 11).

Due to fast recording, statistics of 10M pulses were recorded for each setting.(Negligible statistical errors, systematics dominated measurement.)

Various working gases: Ar(80)CO2(20), CH4, Ne(90)CO2(10), Ne(90)CO2(10)N2(5).


Amplitude distribution from direct measurement for typical setting.

Measured signal distribution (10M pulses)

ADC-2000 -1000 0 1000 2000 3000 4000 5000 6000 7000 8000

Ent

ries

100

101

102

103

104

105

106

0 PE signal

1 or more PE signal

mean ADC = 735.506

DataPedestal


Single-avalanche response reconstruction

Signal formation:

Photoelectron generation with Poisson distribution Pν , here ν is the per pulse PE yield.

Multiplication of each PE by GEM avalanche process, with distribution p each.

Multiplication of each GEM-multiplied PE by MWPC avalanche process, with distributioneγ each (γ: MWPC gain). Note: eγ is approximately exponential with slope γ.

An additive noise is mixed to the signal at the FEE amplifier input, with distribution g.Note: g is approximately Gaussian, and can be measured with zero-signal events.

GEM response distribution to an incoming single-photoelectron:

p.

GEM+MWPC response distribution to an incoming single-photoelectron:

f =∞∑

k=0

e(k-fold convolution)γ p(k).

GEM+MWPC+FEE response distribution to an incoming LED pulse:

h = g (convolution)∞∑

n=0

f(n-fold convolution) Pν(n).


Using convolution theorem, one has in Fourier space:

H = G exp (−ν (1− F ))

where F , G and H is the Fourier transform of f , g and h.Here, F decays to zero because of Riemann-Lebesgue lemma.

⇒ H/G∣

∣

∣

high freq.limit= e−ν , i.e. the PE yield ν can be determined.

⇒ the single-PE GEM+MWPC response f can be reconstructed in mathematical way.

Using an unfolding method (A.László JPCS 368 012043, arXiv:1404.2787):

f =∞∑

k=0

e(k-fold convolution)γ p(k)

can be solved for p.⇒ the single-PE GEM response p can be reconstructed in mathematical way.


Due to the additivity of mean µ and variance σ2 of independent distributions:

µp =µh − µg

ν γ( = GEM gain),

σp

µp

=

√

√

√

√νσ2h− σ2

g

(µh − µg)2−

(

σeγ

µeγ

)21

µp

− 1 ( = GEM intrinsic resolution)

is valid for the moments.⇒ the GEM gain µp and the GEM intrinsic resolution σp/µp can be recovered.

Note: it is seen that systematic errors caused by possible small non-exponentiality ofMWPC response is suppressed by 1 / GEM gain.


Measurements can be performed in two extremes:

small PE yield (ν ≈ 0.01− 0.1) — direct measurement possible without reconstruction,but low amplitude region not directly visible on raw spectra,

large PE yield (ν ≈ 1− 3) — the discussed Fourier method is needed, and can alsorecover the small amplitude region.

Consistency check of single-PE reconstruction

Response [ADC]-2000 -1000 0 1000 2000 3000 4000 5000 6000 7000

Den

sity

func

tion

[1/A

DC

]

10−6

10−5

10−4

10−3

10−2

10−1

Reconstructed single-PE distribution (ν ≈ 1.52383)

Raw distribution (ν ≈ 0.205659)

Good consistency seen.Single-avalanche response mesurement method for MPGD detectors – p. 12

Shape bias by MWPC response is small, especially for not too small GEM gain settings.

Simulation example:GEM and GEM + MWPC response simulation (GEMGain=50, MWPCGain=1000)

Response [number of electrons in MWPC scale]0 50000 100000 150000 200000 250000

Ent

ries

0

50000

100000

150000

200000

250000

300000

350000

400000

GEM extractedGEM extracted + MWPC

Can be corrected in an exact way by the discussed unfolding method, and shape correctionis very small for GEM gain ≥ 10.


Preliminary results

Compilation of scaled GEM multiplicity distributions

Multiplicity / gain0 1 2 3 4

Pro

babi

lity·

gain

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Ar(80)CO2(20), GEMgain=93.1859

Ar(80)CO2(20), GEMgain=47.4713

Ar(80)CO2(20), GEMgain=16.8815

CH4 , GEMgain=12.2961



Ne(90)CO2(10), GEMgain=12.0316

Ne(90)CO2(10), GEMgain=35.2746

Ne(90)CO2(10), GEMgain=53.3248

Ne(90)CO2(10)N2(5), GEMgain=8.68682

Ne(90)CO2(10)N2(5), GEMgain=17.0717

Ne(90)CO2(10)N2(5), GEMgain=26.2785

Preliminary results show that shape of GEM avalanche response distribution isapproximately independent of the effective gain, and slightly depends on the working gas(better for Neon based).


The same effect is seen on the σ/mean values: independent of the effective gain, andslightly depends on the working gas (better for Neon based).

Compilation of GEM response σ/mean values

GEM gain0 25 50 75 100 125

σ/m

ean

0

0.25

0.5

0.75

1

Ar(80)CO2(20): σ/mean ≈ 0.840478

CH4 : σ/mean ≈ 0.911974

Ne(90)CO2(10): σ/mean ≈ 0.743731

Ne(90)CO2(10)N2(5): σ/mean ≈ 0.784677

Ar(80)CO2(20) (νeff ≈ 1.66423)

Ar(80)CO2(20) (νeff ≈ 0.0218017)

CH4 (νeff ≈ 3.20924)

CH4 (νeff ≈ 0.013401)

Ne(90)CO2(10) (νeff ≈ 1.5916)

Ne(90)CO2(10) (νeff ≈ 0.0155221)

Ne(90)CO2(10)N2(5) (νeff ≈ 2.3037)

Ne(90)CO2(10)N2(5) (νeff ≈ 0.0236394)

(Systematic errors of large PE yield measurements are better – about 5%.)Single-avalanche response mesurement method for MPGD detectors – p. 15

Summary

An experimental setup and methodology was developedfor determination of the multiplication distribution of GEMelectron avalanches.

Apparatus and methodology was designed specially forthe coverage of wide ranges of effective gain (1-100) aswell as for measurement of the low multiplication regionof the avalanche response.

Experimentally measured avalanche multiplicationdistributions show deviation from exponential for GEMs:σ/mean down to 0.8 were seen.

Preliminary results reveal universal properties of theshape of the single avalanche multiplication distribution.

Best intrinsic resolution (σ/mean) seen for Neon basedworking gases around 0.8 ⇒ improved dE/dx resolution.


Documents

Single-avalanche response mesurement method for MPGD … · Single-avalanche response mesurement method for MPGD detectors András László [email protected] Wigner RCP,