1. Semi-Digital Hadronic CALorimeter
Introduction
I.Laktineh
CIEMAT, Gent, IPNL, LAL, LAPP, LLN, LLR, LPC, Protvino, Tsinghua, Tunis

2. OUTLINE
Physics motivation
Technology Motivation
GRPC detector

3. Motivations
For future colliders, jet energy resolution will be a determinant factor of
understanding high energy physics.
For instance:
Higgs production (e+ e-  H n n )
WW scattering measurement in absence of Higgs
60%/E
30%/E
30%/E

4. Motivations 2 jet = s2 ch. + s2  + s2 h0 + s2thr+ s2confusion
Ejet = Echarged tracks E Eh0
fraction % % %
Charged tracks resolution ∆p/p2 ~ few10-5
Photon(s) energy resolution ∆E/sqrt(E) ~ 16%
Neutral hadrons energy resolution ∆E/Sqrt(E) ~ 50%
2 jet = s2 ch. + s2  + s2 h0 + s2thr+ s2confusion
= (0.17)2 EJet + s2thr + s2confusion
PFA: Particle Flow Algorithms :
High granularity  Topological separation Reduce confusion
 Improve on jet energy resolution
Scheme : Increase granularity by going digital
J-C. Brient (LLR) 4

5. Motivations Detector choice: Gaseous detectors are excellentHV
Signal
Graphite
Resistive plates
Gas
Pick-up pads
Detector choice:
Gaseous detectors are excellent
candidates. They are homogenous,
cost-effective, and allow high
transverse and longitudinal
Granularity.
They can be very thin while still very efficient
3 detectors are proposed for
digital calorimetry :GEM, MICROMEGAS and (G)RPC.
GRPC was chosen for our prototype.
Two HCAL prototypes using GRPC are followed within CALICE
one is a physical prototype and ours which is a technological
prototype

6. Motivations Electronics readout And granularity choice 1 cm2 pad
The size of avalanche at the anode level
of our GRPC is about 1-2 mm2.
The best granularity is then 1mm2.
This however implies for ILD DHCAL option :
50 millions  5000 millions
The 1cm2 is a good compromise. Using PFA technique, Mark Thomson
found the performance of such calorimeter ( with binary readout) to
Be very similar to an analog HCAL(3X3 cm2 segmentation) for jets of
100 GeV energy. The semi-digital (3 thresholds) readout should improve
the resolution.
1 cm2 pad
Avalanches

7. Motivations Electronics readout and granularity choice
Why not larger pads? :
1- Granularity is very helpful to identify muons within Jets
2- Fine granularity is better for energy resolution
Single particle
Jet
KEK
(Matsunaga et al)

8. Motivations Electronics readout and granularity choice GRPC
At high energy the shower core is
very dense (up to 50 pc/cm2)
simple binary readout will suffer
saturation effect
semi-digital readout (2-bit)
can improve the energy
resolution.
1 cm2 pad
GRPC

9. Motivations The Semi-digital GRPC-based HCAL was proposed and accepted
as one of the two HCAL possible options in the ILD Letter Of Intent
A genuine mechanical structure was also proposed
It is self-supporting
Has negligible dead zones
Eliminates projective cracks
Minimizes barrel / endcap separation
(services leaving from the outer radius)
Barrel
Module
Module 9

10. Motivations
We intend to validate the SDHCAL concept by building a prototype which is
as close as possible to the proposed SDHCAL for ILD to understand key
issues of integration and operation :Technological prototype
 Self-supporting mechanics
 Minimized dead zone
 Minimized thickness
 One-side services
 Power pulsed electronics
gas
Beam
Beam
The prototype will be made of
40 units. Each unit is made of :
2 cm absorber
+ 0.6 cm sensitive medium
1 cm2 transversal granularity
This is about 5 λI
and channels
The modular structure we propose makes it possible to increase the number of units up to 48

11. Motivations Technological prototype vs Physical prototype
With respect to the physics prototype developed by U.S groups
our efforts to build a technological prototype led us to develop:
1- Large detector (1m2) with almost no dead zones :
(rather than assembling 3 chambers of 33X100 cm2 and using
fishing line as spacers)
2- Large and thin embedded electronics board
rather putting side by side electronics boards read by each sde
4- One-side services : readout, gas outlets..
rather having two-side access
5- Self-supporting mechanical structure
rather using the old and simple mechanical structure of AHCAL
6- Power-pulsed electronics
which is not considered in the US-DHCAL
7- New generation of DAQ system
rather using the old one (rate limitation)
gas
Beam
Beam

12. GRPC DETECTOR

13. Detector The GRPC choice was motivated also by:
1- High efficiency and stability
2- Low cost
3- Large detector can be easily built and well suited for ILD
4- Can be home-made
in addition to
1- Well known performance (BELLE, OPERA)
2- Expertise with thin GRPCs developed by IHEP group
The GRPC will be used in the avalanche mode
(2-4 pC/mip and 100 Hz/cm2) rather than in the streamer mode
( pC/mip and few Hz/cm2)
The GRPC to be used in the SDHCAL is very thin (gas gap 1.2 mm) with
a gas mixture made of TFE(93 % ,Isobutane/CO2 (5%) and SF6(2%)
at H.V = 7.4 kV.
10 primary electrons are expected  low probability to have no signal
gas
Beam
Beam

14. Summary of RPC features
From Ammosov LCWS04
All these are confirmed by new TB with our detector (R.Kieffer talk)

15. Detector More on the GRPC choice
Although the groups involved in this project (Protvino group)
had a good knowledge of GRPC, R&D was however necessary due to the
need to build :
large, thin and one-side service GRPC with almost no dead zone.
The R&D items that were developed were:
1- Spacers
2- Resistive coatings
3- Gas distribution system
4- Aging studies
5- High Voltage connections
gas
Beam
Beam

16. Backup slides

17. + 5GeV Digital-1bit Analogue Gaussian Simulation /mean ~22%
E (GeV)
Gaussian
Landau Tails
+ path length
Number of hits
/mean ~22%
/mean ~19%
+ 5GeV
Simulation
Analogue
Digital-1bit

18. RPC in avalanche mode Typical Q and m distributions
1.2 mm, 2% SF6, 8.4 kV - working point, 2.2 mV thr
Mean 2.8 pC
RMS 1.6 pC
Mean 1.47
RMS 0.58
Q ~ 107 e
2 adj pads
From Ammosov LCWS04

19. RPC in avalanche mode 1.2 mm gap RPC eff, <m> vs HV
- 2% and 5% of SF6
Thresholds  mV
 mV
 mV
2.2 mV is best threshold
eff >99%
low <m> ~ 1.4
For 2.2 mV Knee kV kV
V kV kV

20. Typical Q and M distributions, 200 V above knee
RPC in streamer mode
Typical Q and M distributions, 200 V above knee
1.2 mm gap, TFE/Ar/IB=80/10/10
RMS/Q=0.6
FWHM=20%
No ways to suppress multi streamer tail
From Ammosov LCWS04

21. for different thresholds
RPC in streamer mode
Eff, M and Q vs HV
for 1.2 and 1.6 mm gaps
Ar10 mix
for different thresholds
best choice - thr = 300 mV
From Ammosov LCWS04

22. Comparison of avalanche and streamer modes
Rate capability
streamer ~2-3 Hz/cm2
avalanche ~100 Hz/cm2
It is hard to work in streamer
mode even for usual beam
conditions
Streamer is suitable only for
very low rates like e+e- FLC
From Ammosov LCWS04

23. Comparison of avalanche and streamer modes
As example, for 1.2 mm gap
From Ammosov LCWS04