D. Gonzalez-Diaz, KEK, 19-01-2017 I. A contemporary recap II. - - PowerPoint PPT Presentation

• D. Gonzalez-Diaz, KEK, 19-01-2017

• I. A contemporary recap
• II. Historical introduction

III.Technological pillars / few things not to forget

• IV. Faster
• V. Bigger
• VI. Better

• II. Cherenkov light + solid converters
• III. Multi-sampling

V

• I. parallel plate geometries

Limited by fluctuations in the multiplication and primary ionization statistics Limited by primary statistics and electron diffusion Limited by ionization mean free path and drift velocity

• D. Gonzalez-Diaz et al. arXiv:1606.08172v2
• R. de Oliveira arXiv:1503.05330v1

[?]

History of timing with gaseous detectors(I)

• E. K. Zavoisky and G. E. Smolkin, At. Energ. (USSR), 4(1956)46

trigger

chambers

readout

track σ=100 ps tails!

• Long dead-time (time to reload the chamber capacitance after a spark was ~1-10ms) .
• Read-out was performed by optical means .

1956

1 mm

• Time resolution limited due to a relatively big gas gap .
• Discharge takes energy only from the locally affected area, limited by surface resistivity .

1970

History of timing with gaseous detectors(II)

0.1 mm

• High efficiency enforces operation at around 12bar .
• Time resolution consistently below 100ps .

1978

History of timing with gaseous detectors(III)

1.5 mm

• Time resolution limited due to a relatively big gas gap .
• Avoids the necessity of using soviet technology (Bakelite plates used instead) .
• In subsequent works, the authors introduced C2H2F4 and SF6 and a new operation mode

(limited proportionality / saturated avalanche mode). 1981

History of timing with gaseous detectors(IV)

3 mm

• Time resolution limited due to a relatively big gas gap .
• Time resolution and efficiency can be improved with addition of more gaps. Each new gas

gap behaves like a detector replica (i.e., a parallel current generator) improving avalanche statistics, and resistive plates can be simply left floating . 1996

History of timing with gaseous detectors(V)

0.3 mm

2000

• Time resolution good due to small gap .
• Efficiency good due to large number of gaps .
• Standard materials, standard gases, standard pressure, standard 1GHz electronics .

History of timing with gaseous detectors (VI)

+V

• V

charged particle

readout pads/strips

• J. Wang et al. NIM A, 621(2010)151 [Tsinghua University]

>2004

• Small improvements since 2000 in order to stablish reliable production techniques .
• 50-90ps achieved on large 2m-scale areas .
• Technology stabilized.

History of timing with gaseous detectors (VIII)

I. The characteristics of the induced signal are only mildly affected by the resistive material (through its dielectric constant and thickness). II. The transition avalanche-streamer-filamentary discharge-spark is quenched down to the energy available in a small local area, provided the flow of current through the resistive electrode is limited. In practice big avalanches end as streamers, and stable operation of the amplification electronics is possible.

• III. The use of multiple gas gaps (acting in practice as parallel current generators) allows to

keep the high efficiency characteristic of large gaps and the high time resolution characteristic of narrow ones. The associated reduction in the induced charge fluctuations improves both.

• IV. The use of electronegative gases helps at stabilizing le operation up to very high gains,

even in the presence of strong Space-Charge effects.

• V. The maximum operating rate is limited by charge build-up and the characteristic time for

the released charge to abandon the system, by conduction through the resistive electrodes.

• VI. Scalability is relatively easy due to use of common materials.

) / log( A s

1GeV 10GeV 100GeV 1000GeV 10000GeV E 2000 2005 2010

FOPI ALICE EEE HADES HARP CBM R&D MPD R&D

year

STAR (barrel, MTD)

2015

PHENIX BGO LEPS2

• technology introduced (100ps)

NIM A 443(2000)201

• simulations demonstrate operation

under deep space-charge conditions NIM A 517(2004)54

• ALICE module (50ps)

NIMA 533 (2004)93

• warm RPCs

NIM A 527(2004)471 NIM A 555(2005)72

• Chinese glass RPC

NIM A 621(2010)151

• Electrostatic compensation

NIM A 648 (2011) 52

• 24-gap module (20ps)

NIM A 594(2008) 39

• Ceramics RPC

NIM A818 (2016) 45–50

RPC wall hadron blind RICH inner MDCs (I-II)

• uter MDCs (III-IV)

magnet high angle TOF SHOWER detector

8 m2

• 0.27 mm × 4 gaps
• minimum for good efficiency
• Aluminum and glass, 2mm-thick electrodes
• minimize amount of glass for maximum

rate capability

• try to keep good mechanics
• Heat-tolerant materials

Glass Aluminium spring-loaded pressure plate fully shielded HV & readout at the center

wall gas box of one sector column (1-3) row (1-31) layer(1-2) 1116 tRPC individual detectors 6 sectors x 2layers x 3columns x 31 cells

variable cell overlap for providing full angular coverage variable cell width for matching occupancy

detailed info in D.Gonzalez-Diaz 2006 JINST TH 003 and D.Belver et al. NIM A, 602(2009)687

S5 S4 S3 S2 S1 S6

• verall resolution 77ps

σT [ps]

• verall cross-talk 0.4%

deterioration of resolution for a coincident track (around 100ps)

• A. Blanco et al., NIM A, doi.10.1016/2010.08.068

distance (in rows) between primary and secondary hits 13cm length 55cm length

• G. Kornakov et al., 2014 JINST 9 C11015

Normal t=0 t=∞ Resistive t Rgap V Cgap E g d σ, εr V E g E? E? Rgap Cgap Rplate Cplate E? E?

g V E = g V E = g V C C C E

gap plate plate

+ = g V E =

transient! (in practice, few sec)

V=V

• V=0

d g V=V

• V=0

Cglass Cglass V=0 Cgas V=V

• (‘AC limit’)

1/Ctot =1/Cglass+1/Cgas+...

+- +- +- +- +- +- +- +- +- +- +- +- +- +- +- +- +- +- +- +- +- +- +- +- +- +- +- +- +- +- +- +- +- +- +- +-

Eglass Eglass Egas Egas Eglass Egas Eglass = Ctot/Cgas Vo/g Eglass = Ctot/Cglass Vo/d Rtot =Rglass+Rgas+... Egas Egas Egas Eglass Eglass Eglass Eglass = Vo/3g Eglass ~0

V=V

• V=0

Rglass Rgas Rglass V

• 2/3V
• (‘DC limit’)

ΔV

• =1/3V
• RPC capacitance loaded

glass capacitance un-loaded, gap capacitance fully loaded

V=V

• V=0
• - - -

++++ +

• ++++++
• - - - - -

V

• 2/3V
• 1/3V
• +

V=V

• V=0

+ +

• - -
• - -

V

• 2/3V
• +ΔV1

1/3V

• -ΔV2

V=V

• V=0

V

• <V>=2/3V
• <V>=1/3V
• ...
• E. Cerron Zeballos et al., NIM A, 374(1996)132

*Indeed, it has not been possible to quantify the effect of the fluctuations on the plates’ potentials. But

no practical difference has been observed between leaving them floating or fixing the potential…

Avalanche field (Ez=100 kV/cm, 0.3 mm gap)

• C. Lippmann, W. Riegler, NIM A 517(2004)54

!!

Semi-quantitative derivation of the maximum attainable gain before streamers appear:

2

4 ) ( r e q x E

• x

e av

πε

α

≅

• x

e c av

E r e q x E

c c

≅ ≅

2

4 ) ( πε

α th

v D 3 λ =

• th

e

E mv q v λ ≅ v x D r 4

2 = 2 th

• e

k

mv v D E q = = ε ]) cm [ ln( ]) eV [ ln( 16

c k c c

x x + + ≈ ε α

characteristic energy: in equilibrium

20 ≈ g α

Experimentally observed limit for ~ cm gap PPCs: The Raether limit, 1964

]) cm [ ] eV [ ln( 16 g g

k

ε α + ≈

αg=25- 33 in tRPCs!

) (

2 , 2 , th avalanche v noise v T

V V dt dV = + = σ σ σ

d T

v ) ( 28 . 1 η α σ − = Exact solution for a single electron avalanche!

• W. Riegler et al., Nucl. Inst. Meth. A 500(2003)144

MC simulation consider only 1 gap V I(t)=qδ(t) simple circuit calculation long transient behavior, 'equilibration time' average voltage drop in stationary conditions

DC

V ∆

voltage fluctuations

ΔVglass/R

• D. Gonzalez-Diaz et al., Nucl. Phys. B (Proc. Suppl) 158(2006)111
• D. Gonzalez-Diaz et al., Nucl. Instr. Meth. A 602(2009)713

‘cell model’

N q rms E E rms

q gap

• Egap

2 1

2 2

        + = −

τφ A N =

) 1 ( ) ( d a g V V g V E

th

• th

gap

φρ + − + =

) 1 ln( ~ d a d a teq φρ φρ τ +

the DC limit must satisfy ) ( 1 ) ( 1 q d V g R I V g E

• gap

φρ − = − = average field fluctuations around the average field transient time for the typical values A>1mm2 the importance is small. for float glass the equilibration time typically

• bserved is ~2-3s.

Campbell theorem for shot noise no dependence on the area A influenced (cell) by the shot!

( )

th gap

V V a q − ~

if we further assume that

Campbell theorem

simple analytical estimate MC rmsEgap/(Eo-Egap) teq/ τ Φ [Hz/cm2] Φ [Hz/cm2] DC model

• D. Gonzalez-Diaz et al., Nucl. Phys. B (Proc. Suppl) 158(2006)111

From M. Morales, C. Pecharromán, PoS(RPC2012)024

(dry) Next generation experiments

Q/A = q x ϕ x Δt = 100 mC/cm2 electrical ageing (material) avalanche ageing (material+gas) 2pC 5 years 300 Hz/cm2

• S. Gramacho, L. Lopes et al.,

NIM A 602(2009)775 Deposits observed for ~100mC/cm2 but no degradation of performance!

+V

• V

I(t)? Cg Cm Lo Lm x z y D~1m D~1m

• D. Gonzalez-Diaz, H. Chen, Y. Wang,
• Nucl. Instr. Meth. A 648, 1(2011)

) , ( ˆ ˆ ) , ( ) , ( ˆ ˆ ) , (

2 2 2 2 2 2 2 2

t y V dt d C L t y V dy d t y I dt d L C t y I dy d = =

standard RPC ‘compensated’ RPC

center end center end

• D. Gonzalez-Diaz et al arXiv:1606.08172v2

intrinsic time resolution average ohmic drop at the insulator time resolution at high particle flux

From C. Pecharroman (CSIC-Madrid), "Understanding the ageing process in RPC’s from an ion

conductivity approach'. Talk at the X Workshop on Resistive Plate Chambers and related detectors.

good insulators bad insulators semi-conductors conductors not enough streamer quenching too low rate capability

The problem is to find a 'bad insulator' developed by industry!

1 2 3 4 5 6 7 8 9 10 0.1 1 10 100

Roughness(nm) Measured point

Ra_d Rq_d Ry_d Ra_c Rq_c Ry_c

~1C/cm2

continuous line: doped glass dashed line: common glass

• J. Wang et al. NIM A 621(2010)151

good surface uniformity chinese glass low resistivity float glass x100 new insulators are needed ! ...

) ( ) (

gap gap T

E S K E = σ

) (

1 ) (

ref gap

E E gap

e E

− −

+ =

ϑ

ε ε ) (φ σ T ) (φ ε

for small particle fluxes:

d q K d q KT

T

φρ ε ε φρ σ σ

ε

− = + =

approximate linear behavior

• I. Deppner et al.,doi:10.1016/j.nima.2010.09.165

from theoretical considerations: phenomenological: yields the following scaling

Φ[kHz/cm2]

) 1 ( ) ( d a g V V g V E

th

• th

gap

φρ + − + =

σT[ps]

ε

rate capability [kHz/cm2] ρ0d0/ρd

Typical sizes of interconnects, cross-talk and signal attenuation inside the counter limit the performances for counters at the 1-2m scale unless proper precautions are taken. Virtually all next-generation RPCs are based, to be based, or considering the multi-strip technology (EEE, STAR-MTD/RHIC, CBM/SIS300, MPD/NICA, BGO, LEPS2)

arXiv:1606.08172v2

BGO-EGG counter

use shielding vias fancier: electrostatic compensation! x10!

• m
• m

L L C C =

• D. Gonzalez-Diaz, H. Chen, Y. Wang,
• Nucl. Instr. Meth. A 648, 1(2011)

compensated under-compensated

• ver-compensated

dielectric losses in time-domain in frequency-domain What makes the trick?

!

In this particular configuration... increasing the coupling

within ±0.2mm cross-talk increases a factor 2-3!

t1 t2 FEE-DBOs (amplification, discrimination, Q-ToT algorithm) FEE-MBOs (power regulation and distribution, sensing, signal distribution)

• D. Belver et al. IEEE TNS 57(2010)2848

low-noise customized LV system based

• n switching

power supplies

• A. Gil et al. IEEE TNS 56(2009)382

charge threshold ~ 30fC jitter@100fC ~ 15ps gain ~ 100 BW ~ 2GHz Q-Width algorithm built-in

TDC resolution at the level of 25 LSB (HPTDC) or ~10ps (FPGA-based)

) ( ) ( )) ( ) ( ( ) ( ~ ) (

max max max max max

E v g E v E E N N K E

drift drift

• T

best T

∝ − = η α σ σ

• Accuracy on the gap definition better than ~0.05gap x sqrt(Ngap) ~10μm–30μm, since

the field variations are directly depending on it. So... make very thin capacitors! problems:

Increase number of gaps and field as much as possible

drift

• T

v N ) ( 1 ~ η α σ −

No streamers.

C g ≤ α

(~30 for 0.3mm gaps and electronegative gas)

• Low efficiency, due to the low ionization probability (use multi-gap!).
• Theoretical arguments apart, obtaining resolutions below 50ps is technically difficult

(FEE-BW, noise, interconnects, transmission characteristics, TDC resolution...).

• Detector stability?. Respecting the analogous to the Raether criteria for thin gaps

does not guarantee stability, due to unavoidable fluctuations in avalanche multiplication and primary ionization that can trigger streamers.

160um gas gap

• N. Tomida, C.-Y. Hsieh, et al JINST 7 (2012) P12005
• S. An et al., Nucl. Instr. Meth. A, 594(2008)39

D~1m D~1m +V

• V

α / 1

R R e--I+ I(t)? Ne-

+V

• V

Imain

• Ineigh
• Imain

+

Ineigh

+

currents produced by the 'mirror charges' x z y Ne-,v and summing over all gaps! D~1m D~1m

+V

• V

I(t)? Cg Cm Lo Lm x z y D~1m D~1m

I(t)? TDC ADC

• D. Gonzalez-Diaz

doi:10.1016/j.nima.2010.09.067

E=100kV/cm σT~1ns σT~100ps computing time <50ms per impinging particle

Zc = 21Ω (simulation) pad readout Jingbo Wang et al., NIM A 621, 1-3(2010)151

CBM-prototype (multi-pad)

6cm

sim (BW~0.5GHz)

injection at center

HADES (shielded)

Zc = 9.0Ω (measurement) Zc = 10.5Ω (simulation) 15cm-long, 2.2cm-wide (short cell, high granularity region) 20mV 100ns

15-55cm

sim (BW=2GHz) sim (BW=2GHz)

• D. Belver et al., NIM A

602, 3(2009)687 injection at center

40cm

Zc=9.3Ω

Fast neutron detector-I

sim (BW=1.25GHz) sim (BW=1.5GHz) sim (BW=1.25GHz)

• D. Yakorev et al. doi:10.1016

/j.nima.2011.05.031 injection at center

200cm

Zc = 13.0Ω

sim (BW=1.25GHz) sim (BW=1.5GHz) sim (BW=1.25GHz)

Fast neutron detector-II

injection at center

200cm

Zc = 14.1Ω

sim (BW=1.5GHz) sim (BW=1.25GHz)

very slightly modified (additional material required -> less than 10%)

Fast neutron detector-II (optimized)

injection at center

• Seventeen successful years of multi-gap timing RPCs development. This is still by far

the best technology for sub-100ps timing with gaseous detectors, although new approaches are being tried.

• Large module sizes (~1m x 1m, 2m x 2m) achieved.
• Large system sizes implemented (~100m2).
• Time resolution at the scale of 50-100ps for minimum ionizing particles (and 95-100%

efficiency) in any modern detector system.

• Demonstrated technological limit: σt=20ps, 10ps possibly achievable.
• Rate capability from ~0.5-1kHz/cm2 (standard float glass), to 5-10kHz/cm2 (warm glass),

30-50 kHz/cm2 (chinese glass), 100 kHz/cm2 (ceramics).

• Demonstrated maximum detector size keeping a 1-2GHz bandwidth: 2m.
• Good performance demonstrated also for relativistic heavy ions, neutrons and

annihilation photons (with accuracies around 100ps in either case).