Digital Hadronic Calorimeters

Imad Laktineh

Institut de Physique Nucléaire de Lyon

Outline

• Some History

• Why digital hadronic calorimeters are good for ILC

• Gaseous detectors for the DHCAL

• DHCAL present Activities

• Towards the technological prototype

Some historyThe use of gaseous detector as a sensitive medium in calorimeters is not a new idea. Calorimeters based on gaseous detectors using proportional, saturated avalancheand streamer modes were developed since the seventies of the last century. In some cases digital readout (counting the number of tracks the shower) were used.

Example : PEP4 Electromagnetic Calorimeter Gaseous (argon-Ethyl-bromide) Geiger cells+ Lead Resolutions obtained with gaseous calorimeters were shwon

to be equivalent to those obtained with scintillators calorimeters.

PEP4 experiment scheme

PEP4 ECAL scheme

Linearity of gaseous calorimeter

Some examples of gaseous HCAL

: The four LEP experiments at CERN (ALEPH, DELPHI, L3, OPAL) had their hadronic calorimeters or part of it made with gaseous detectors:

ALEPH : Iron + Streamer tubes 85%/√E

DELPHI :Iron + Streamer tubes (only barrel) 21%+112%/ √ E L3 : Uranium +Wire chambers

OPAL : Iron + Streamer tubes 120%/√E

ATLAS combined beam test :σ/E ≈ (0.52/√E + 0.03/E) + 1.6/E (LAr + Tile)Expected jet energy resolutionσ(jet)/E ≈ 0.6/√ E(GeV)

CMS-brass-scintillator readout by wavelength shifter in the barrel and endcaps-Fe-quartz fibers (Cerenkov) in the very forward directionATLAS-Fe-scintillator tile / WLS fibre readout in the barrel-Cu/Liquid Argon in the endcap-W/Liquid Argon in the very forward region

At LHC with very high energy jets the non linearity of the standard gaseous calorimeters would have been a problem

Why Digital HCAL For ILC?

Jet energy resolution is a key feature of the future Linear Colliders experiments

Zqq (70%), Wqq’(68%), H(120 GeV)qq,WW,ZZ (85%)

Improving jet energy resolution is of important interest for bosons identification

60%/E 30%/E

Motivation

Charged tracks resolution ∆p/p ~ few10-5 Photon(s) energy resolution ∆E/E ~ 12% Neutral hadrons energy resolution ∆E/E ~ 45%

Ejet = Echarged tracks + E + Eh0fraction 65% 26% 9%

2 jet = ch. + 2 + 2 h0 + confusion

= (0.15)2 EJet + confusion

PFA: Particle Flow Algorithms : High granularity Topological separation Reduce confusion Improve on jet energy resolution

Motivations

Ed

ep

Nh

its

E pion (GeV) E pion (GeV)

Digital-1bitAnalog

Gaussian

E (GeV) Number of hits

/mean ~22% /mean ~19%

++ 5GeV 5GeV

Simulation

Magill

Comparison analog%digital

Sandwich

Hadron Calorimeter

incident

lead 1m×1m×8mmsci 1m×1m×2mm

×100 Layers

・・・

1m×1m

1mYamada et al

Analog

lead 1m×1m×8mm

sci 1m×1m×2mm

×100Layers

measure all energy

each layer

Measure Energy

Digital

lead 1m×1m×8mm

sci 1cm×1cm×2mm

×10000

×100Layers

measure number

of all hits

NhitYamada et al

Histogram of π（ 4GeV 、 1000event ）

analog digital

Mean ： 113.9MeVSigma ： 34.43MeV

Mean ： 35.59hitsSigma ： 9.027hits

Yamada et al

Threshold Dependence(Nhit)

electron π

Yamada et al

Measure Energy , Nhit（ Incident Energy Dependence ）

analog digital

Yamada et al

Energy Resolution（ Incident Energy Dependence ）

electron π

Yamada et al

Cell Size Dependence（ Measure Energy , Nhit ）

electron π

Yamada et al

Cell Size Dependence（ Energy Resolution ）

electron π

Yamada et al

Matsunaga et al

Matsunaga et al

Why Gaseous HCAL For ILC?

0.1

1

10

100

1000

0 20 40 60 80 100

z position ( cm )

Hit

tim

e (

ns

)pi-4GeV

The green cells are neutron hits. (by G4Track Information)

E(neutron) = 50MeV

Z

The hits are due to neutron scattering.

Takeshita et al

Single pion simulation

Neutrons

Neutrons

Scintillator Gas(Xenon)neutron 50MeV

The box size: 1m x 1m x 1m

Neutrons incident at random positions to the pure scintillator and the gas. Takeshita et al

Homogeneity

The detector is segmentedEffect to be measured

Segmentation is provided by the readout. No effect observed

Why Gaseous Digital HCAL For ILC?

Calibration

The calibration is an important issue in case of high granular analog scintillator HCAL :

•SiPM output dependence on temperature and voltage •Scintillator and fiber evolution with time (forward regions)•Electronics readout system

In the case of gaseous digital HCAL only electronics readout system needs to be calibrated and it is very simple for digital or semi-digital readout.

Analog

HCAL 85%/√E 21%+112%/ √ E 120%/√E

Digital

DHCAL For ILC

ILD with GRPC

DHCAL For ILC

SiD with GRPC,GEM,MICROMEGAS

Energy resolution of K0-L in the ILD DHCAL Barrelwith 1-bit readout using PANDORA

Gaseous Detectors For DHCAL

• GRPC

• GEM

• MICROMEGAS

Gaseous detectors

ln M

Voltage

Attachment

Collection

Multiplication

StreamerBreakdown

IONIZATION CHAMBER

PROPORTIONAL COUNTER

Saturation

n1

n2

Sauli@tipp09: http://kds.kek.jp/conferenceTimeTable.py?confId=2376

Gaseous detectors

M (x) nn0

e x

Sauli@tipp09: http://kds.kek.jp/conferenceTimeTable.py?confId=2376

QMAX ≈ 107 e SPARK

Gain:

= N (Townsend coefficient)

Raether limit

lE x

Ions

Electrons

fast

slow

GRPC Glass Resistive Plate Chamber1981: Santoninco,Cardarelli

Avalanche modeHV=6-8 KV/mmTFE+IB+SF6Gain about 106

Streamer modeHV>8 KV/mmArgon+IB+SF6Gain about 108

glass

glass

glass

No spark : V = R I with R very high

GEM 1996: F. Sauli: Gas Electron Multiplier (GEM)

500 V on the 2 copper sides of the kapton foil of 50 µ E 100 KV/cminto the holes

Gain 102-103 / foilGas mixture dependent

1996 : MICRO MEsh GASeous detectorY. Giomataris, Ph. Rebourgeard, J.-P. Robert, G. Charpak,

or pads

MICROMEGAS

300-400 VBetween the mesh and the pickup padsD= 50-120 µ In the Multiplication space E 60-100 KV/cm

Gain = 104-105

Gas mixture dependent

GRPC GEM MICROMEGAS

Cost low high high

Efficiency high high high

Multiplicity 1.1-1.8 1.05-1.1 1.07-1.15

Charge(pc) 0.1-10 .002-.5 .002-.5 Rise time few ns few ns few ns

Sparks absent moderate frequent

Detection rate < 100Hz/cm2 high high

Thickness < few mm <few mm < few mm

Fabrication simple complex complex

DIGITAL HADRONIC CALORIMETERS DEVELOPMENT in CALICE

µMEGAS-based DHCAL development

The detector development is realized essentially by the LAPP group

• Setup– 6 x 16 cm2 mesh– Source collimated above 1

pad– Readout mesh signals

• Collection efficiency– Plateau for field ratios

larger than 50

• Gas gain (assuming 230 primary e-)– Gain doubles every 20 V– Decreases with pressure– Max around 2.104

• Energy resolution at 5.9 keV– 17 % FWHM

Slope yields -2 fC/mbarA

DC

Cou

nts

Pressure (mbar)

55Fe X-ray source test

Beam test setup• Trigger: 3 scintillators• 3 Micromegas 6x16 pads• 1 Micromegas 12x32 pads• Steel absorber option

• GASSIPLEX readout– VME modules (ADC) +

CENTAURE

• Characterisation of the prototypes– efficiency and multiplicity– response uniformity over area– X-talk studies

• Aug & Nov 2008 at CERN PS and SPS

H2 line at SPS-CERN(4th August-15th August 2008)

• Select event with one and only one hit in each chamber– Insure all charge collected on

one pad– Hit if ADC > 27 counts

• Landau distribution MPV is at 45 fC– Shows variations of 10% RMS

over all pad platinum events

MPV ~ 45 fC

en

trie

s

all triggers

ADC counts

Chamber 0Chamber 1Chamber 2Chamber 3

2.7 fC

Efficiency and multiplicity• Require one and only one « safe » hit in

three chambers– ADC > 51 counts (5.3 fC)– Hit position in fourth chamber

extrapolated– Count hit (ADC > 27 counts) in 3x3

pad area

• Efficiency between 92 and 99 %• Multiplicity between 1.07 and 1.15

Efficiency map of one chamber

gold events

Hit multiplicity distributions of 2

chambers

Micromegas with Digital readout

• PCB with DIRAC1 64 channels ASIC – Digital link to DAQ (possibility to chain detectors)– 3 thresholds – Synchronous architecture

• First operational Bulk Micromegas with embedded readout electronics !

DIRAC

Sparks protections mask for bulk laying 8x8 pads with bulk

Bulk from R. De Oliveira & O. Pizzirusso

Micromegas with digital readout

• Tested during August 2008 test beam• Minimum threshold of 19 fC• Only one detector available

– No efficiency/multiplicity measurements yet

– Prepare more prototype for next beam test

Beam Profilewhen moving the X-Y table

Micromegas with digital readout

• PCB with 4 HARDROC1– 64 channels ASIC, detector active area 8x32 cm2

sparks protections

HARDROC1

bulk

Some problems withthe mesh HV but analysisIs going on

2008 beam test setup is being simulated to Compare

with real data

30 cm iron absorber in front

Simulation

Test beam

1 m² prototype: 6 Active Sensor Units9 216 channels - 96 x 96 cm² active area - 3 DIF + interDIF boards

Expected in the second half of 2009

DHCAL with Micromegas

• Full 1 m² geometry implemented– Readout: from 0.5 x 0.5 cm2

to 4x4 cm² pads– 3 mm drift gap– Gas mix. Argon/Isob. 95/5– 1.9 cm thick absorber

between layers– different absorber materials– 40 or 80 layers (~4, 5 or

~9 λ)– Thickness of active layer: 3.2

mm

– Ideal Micromegas, digitization not yet fully implemented

DHCAL 40 planes

100 GeV Pions

• Sum up hits and energy in all cells• Apply a single threshold (1 MIP)• Look at distributions RMS

E(GeV)

Nb

of H

its

1 MIP thresholdPions

Nb of Hits

Nb

of e

vent

s

Nb of Hits

Nb

of e

vent

s

Analogresponseat 10 GeV

Digitalresponseat 10 GeV

1 MIP threshold

Energy resolutionPion energy

E (GeV)

σE/E

Worse resolution at High energy Need more than 1 threshold ?

GEM-based DHCAL development

The detector development is realized essentially by the UTA group

Feb. 20, 2009 63

GEM-based Digital Calorimeter Concept

Use Double GEM layers

{~6.5mm

-2100V

∆V ~400V

∆V ~400V

0V

GEM DHCAL Status & Plans J. Yu

1cmx1cm readout pads

GEM 30cmx30cm Foil Mounting Jig

Anode Board & Preamp for 30cm x 30cm Chamber

Preamps configured to read 96 pads in the center

Use 32 channel FNAL preamps

30cmx30cm D-GEM Detector Signal

Signal from Cs137 Source

30cm x 30cm GEM Chamber for KAERI Beam Exposure

UTA GEM Chamber in KAERI Electron Beam

4-pad area (2cm x 2cm) exposed to scanning beams for ~2000 sec.

•e- beam: 1010 particles in 30ps pulse ~every 43s

•Scans 4cmx60cm area every 2 seconds

G10 boards in the exposed area discolorized.But no damage to the GEM foils

GEM Beam Test Detector Setup

3 Slice test finger counters

Slice test 19x19cm2

counter

30x30cm2 GEM chamber

Slice test 19x19cm2

counter

• Beam Trigger – 5Fold scintillation counters

– Three 1cmx1cm finger counters, 5cm apart, upstream – Two 19cmx19cm counters envelop the chamber active

area, separated by about 3m’s, downstream – Coincidence of all 5 counters defines a beam spot less

than or equal to 1cmx1cm -->Size of one readout pad

• GEM Chamber self trigger – Use negative chamber output – Threshold set at -30mV

• Beam constrained chamber trigger formed of 5F*GEM: 6Fold

– Allows to look at data from neighboring pads while triggering on the pad centered at the beam

– Had to use this since there were no independent means of ensuring the beam containment in one pad

Signal shape Efficiency

GEM with Analog KPix Chip

M. Breidenbach/R. Herbst SLAC

1cmx1cm cells

KPiX (64-ch, developed for ILC Si-Ecal) was modified to accommodate

smaller GEM signals (>~20fC)

10cm x 10cm Pad Board Test Chamber

KPiX chip uses ILC-clock synchronization. No external trigger system. This is not appropriate to measure efficiency in test beams or cosmic ray data taking.

New version (KPiX v7) includes an external trigger

Right under the sourceAway from the source

KPiX4-GEM Source Response Extraction

• A method based on simulation developed to overcome the triggering complication with weak source Jacob’s Method

• Simulate KPiX4 readout of GEM signal using GEM pulse signal, actual pedestal distributions and previously measured Landau response curves– Since the source signal is random, KPiX integrates charge

partially• Let the Landau MPV and width as well as the normalization of

the ped gaussian float till the source data is well described by the simulation

Feb. 20, 2009 GEM DHCAL Status & PlansJ. Yu

76

Extracted Response – GEM-KPix

MPV=1.9fC (normally ~20fC)

DataSimulationResponse

Random sampling of signal & inadequate gas supply

Charge (fC)

Base steel plate, t=2 mm

Readout Board330x500 mm2

330 mm

1000 mm

UTA’s 33cmx100cm DHCAL Unit Chamber

GEM DHCAL Future Plans

Late 2009 – Mid 2011– 33cmx100cm thin GEM unit chambers

• Complete production of 15 33cmx100cm unit chambers

– Construct five 100cmx100cm GEM DHCAL planes• Using DCAL or KPiX readout chips

– Beam test GEM DHCAL planes in the CALICE beam test stack together with RPC

– TGEMs and RETGEMs• Construction and characterization of a prototype

chamber using an analog readout chip• Beam test of TGEM prototype chamber

UTA’s 100cmx100cm Digital Hadron Calorimeter Plane

Feb. 20, 2009 79GEM DHCAL Status & PlansJ. Yu

RPC-based DHCAL development

Efforts in the U.S

Efforts in Europe&Asia

GRPC DHCAL Activities in U.S

The detector development is realized essentially by the ANL group

I Resistive Plate Chambers

Pursued two design options

Two-glass design

One-glass design

Resistive paint

Resistive paint

Mylar

1.2mm gas gap

Mylar Aluminum foil

1.1mm glass

1.1mm glass

-HV

Signal padsG10 board

Resistive paint

Signal pads

Mylar

Aluminum foil

1.1mm glass1.2mm gas gap -HV

G10 board

Vertical Slice Test

Test of whole system with

20 x 20 cm2 GRPC 9 2-glass designs 1 1-glass design Only use RPC0 – RPC5 in analysis of e+, π+

Only use RPC0 – RPC3 for rate dependence Absorber For cosmic rays, muon, pions, electrons: Steel (16 mm) + Copper (4 mm) Rate capability measurement (120 GeV protons): 16 mm PVC with whole cut out in centerTest beamPrimary beam (120 GeV protons) with beam blocker for muonsPrimary beam without beam blocker for rate measurementsSecondary beam for positrons and pions at 1,2,4,8, and 16 GeV/c

Electronic Readout System

Attempt to be as similar as possible to what’s needed for the PS

Components

DCAL ASIC ANL/FNAL Pad-boards ANL Front-end boards ANL Data concentrators ANL Data collectors Boston Timing and trigger module FNAL

Prototyping and commissioning

Used 2nd round of DCAL prototypes All other components: 1st prototypes → all worked very well

A few events… μ Calibration Runs120 GeV protons with 1 m Fe beam block no μ momentum selection

One of many perfect μ event

μ event with double hits in x

μ at an angle or multiple scattering

μ event with δ ray or π punch through

A few events…e+ Run 1 - 16 GeV secondary beam Čerenkov signal required

8 GeV e+ event

8 GeV e+ event with satellites 8 GeV e+ event

8 GeV e+ event

A few events…π+ Run 1 – 16 GeV secondary beam Veto on Čerenkov signal

8 GeV π+ event (typical)

8 GeV π+ event (early shower)8 GeV π+ event (early shower)

8 GeV π+ event (early shower)

Efficiency = ___________________ All triggers

Events with hits in RPC

At high rate efficiency drops and then levels out

Fits to exponential + constant appear adequate

Time constant for efficiency drop shorter at higher rate (as expected)

Efficiency drops for rates ≥ 100 Hz/cm2

In agreement with previous measurements with sources

Simulation Strategy

• Generate muons (at some energy) with GEANT4 (with same x-y distribution and slope as in the data)

• Get x,y,z of each energy deposit (point) in the active gaps• Generate charge from measured charge distribution for each point (according to our own measurements)

• Introduce charge offset Q0 for flexibility• Introduce dcut to filter close-by points (choose one randomly)

(RPCs do not generate close-by avalanches) • Noise hits can be safely ignored

• Distribute charge according to exponential distribution with slope a• Apply threshold T to flag pads above threshold (hits)• Adjust a, T, dcut and Q0 to reproduce measured hit distributions

• Generate positrons at 8 GeV with GEANT4 (with same x-y distribution and slope as in the data)

• Introduce material upstream to reproduce measured shapes etc… • Re-adjust dcut if necessary (Muon data not very sensitive to dcut)

• Generate predictions for other beam energies• Generate pions at any beam energy

Lateral shower profile

Without material in beam

First layer too narrow

Subsequent layers OK

Data

Simulation

Lateral shower profile

With lots of material (1/4 X0) in beam

Looks good everywhere

Data

Simulation

1 m3 – Physics Prototype

Description 40 layers each ~ 1 x 1 m2

Each layer with 3 RPCs, each 32 x 96 cm2

Readout of 1 x 1 cm2 pads with one threshold (1-bit) ~400,000 readout channels using the same electronics as the one or the VST Layers to be inserted into the existing Calice AHCAL structure

Purpose

Validate DHCAL concept Gain experience running large RPC system Measure hadronic showers in great detail Validate hadronic shower models

Status Started construction in fall 2008 DCAL chips ok but pacjaging is on critical path GRPCs : Adequate glass, resistive painting, cassetes : ongoing Completion is expected by end of 2009

RPCs and cassettes

RPC design

2 – glass RPCs 1 – glass RPCs (developed by Argonne)

Prototypes

Number of RPCs

Number of glass plates

Glass thickness [mm]

Size [cm]

Status Tests Problems

~15 2 1.1 20 x 20 built 2 years None

1 1 1.1 20 x 20 built 2 years None

1+3 2 1.2 32 x 96 built 1 month High pad multiplicity

3 1 1.1 20 x 20 built 2 months None

2 2 0.85/1.2 32 x 96 being built

Not on critical path

Comment I: Glass thickness

Pad multiplicity of 32 x 96 cm2 too large: due to glass of 1.2 mm (and track extrapolation) Difficulty to obtain 0.85 mm glass in the U.S. Vendor from Europe identified, provided 10 samples

Comment II: 1 – glass RPCs

Advantages: pad multiplicity ~1, thinner, simpler, surface resistivity not critical, better rate capability, compression with electric field Disadvantage: can’t be assembled without final electronics, recent design (less tested) Some layers for the physics prototype will be equipped with 1 – glass RPCs

Comment III: Resistive paint

LICRON paint (we all used for years) not available anymore New LICRON product difficult to apply (backup solution) Explored two alternatives

Artist paint (currently preferred solution)Floor paint (possible solution)

Black-Green Artists PaintX 1 = 20% HumidX 4 = 20% HumidX 5 = 60% Humid

0

0.5

1

1.5

2

2.5

3

0 2 4 6

TIME ( Test Number )

Res

ista

nce

,

ME

G

Left side

Right side

Center X

Center Y

GRPC DHCAL Activities in Europe & Asia

Bologna-CERN,IPNL,LAL,LLR,Louvain-la-Neuve,Protvino,Tsinghua

RPC in avalanche mode(TFE/Iso/SF6 = 93/5/2)1.2 mm,8.4 kV - working point,2.2 mV thr

Protvino

The challenge for ILC DHCALHow to have a detector of few thousands m² fully equipped with low consumption semi-digital readout and still very compact !!!!?Embedded Daisy-chained electronics can be the solution

VFE ASIC

DataADC

I/OBuffer

FE-FPGA

BOOT CONFIG

Data FormatZero SuppressProtocol/SerDes

FPGA Config/Clock Extract

Clock

Bunch/Train Timing

Config Data

Clk

SlabSlab

FEFPGA

PHY

Data

Clock+Config+Control

VFEASIC

Conf/Clock

VFEASIC

VFEASIC

VFEASIC

RamFull

Analog output

HARDROC

• 64 channels, 16mm²• Digital/analog output.• 2bit: 2-3 thresholds• low consumption, power pulsing (< 10 µW/ch)

Electronics

30 fC

10 fC

Piedestal

Dac

un

it

chanel

•Digital memory able to store up to 128 evts.• Large gain range • Xtalk <2% Adequate for GRPC* (threshold>10 fc)

Mini DHCALMini DHCAL project projectAim: Validate the new electronics/acquisition scheme for theDHCAL

(GRPC/ µMEGAS)•8-layer, 800 µ thick PCB buried and blind vias x-talk <0.3 %•4 hardroc chips •Readout FPGA USB•8×32 pads detector

FPGAasics

Acquisition modes : different modes are allowed:

a) Triggerless (ILC mode)

b) External trigger : cosmic rays & test beam

Data output: digital and analogue

First Daisy chain measurement

LLR

before

after

Gain correction

2.5 fc

Injected charge = 100 fc

8x32 pads prototypes

Example of a recorded mip

5-GRPC cosmic rays test bench

1st GRPC

2d GRPC

3d GRPC

4th GRPC

5thGRPC

Hit pads

Pads associated to the 4 asics

second threshold

First threshold

Threshold 100 fc, gas mixture :TFE(93%,Isobutane 5%, SF6 2%)

Resistive painting Graphite(400 k)

Resistive painting Licron(20 M/)

X-talk =1.3@7.4 kv

X-talk =1.7@7.4 kv

GRPC

GRPC

Test Beam @PS-CERN

Goals :

• Validate the semi-digital electronics readout system in beam conditions

• Evaluate the performance of different kinds of GRPCs

PS T10 17-24 july: (260 kEvents)PS T10 17-24 july: (260 kEvents)

• Use of Eudet telescopeEudet telescope with 4 graphite GRPC from protvino. Track reconstruction (accuracy <100μm of one telescope 100μm of one telescope armarm) to evaluate efficiency at interpad and edges

• Hight voltage scanHight voltage scan monitoring efficiency/multiplicity.

• Record first pions showerspions showers with miniDHCAL configuartion (use of iron slabs).

• Test of new GRPC prototypesnew GRPC prototypes over different angles (GRPC licronlicron, GRPC staguardstaguard).

PS T9 28 July 4 august (80 kEvents)PS T9 28 July 4 august (80 kEvents)

• Record more pions showers, with different slab different slab configurationsconfigurations.

• Test again new GRPC prototypes (GRPC licron, GRPC staguard).

PS T9 7-12 November (65 kEvents)PS T9 7-12 November (65 kEvents)• First test of multigap GRPC multigap GRPC (could reduce charge spread)

• Flux and detector ratesdetector rates testing (20-6000 triggers/spill)

• First attempts : replacing isobutane by COCO22.

All evtsAll evts

dt >50 msdt >50 ms dt >50 msdt >50 ms

Previous spillPrevious spill Current spillCurrent spill

NoiseNoise0.045 Hz/cell0.045 Hz/cell

Efficiency map for 3+1 GRPCs @7.4KV

Dead channels

Daq’s Thresholds: lower 120 fC 120 fC / higher 450fC 450fC Plateau: 7.2 to 8 kV 7.2 to 8 kV

-> Efficiency between 80 and 98%80 and 98% Lower multiplicity will be the best for us.

-> Best ratio Best ratio multiplicity/efficiency: around 7.4 kVaround 7.4 kV Until now the licron detectorlicron detector seems to be the best candidate:

-> it got the lower multiplicity and shows acceptable efficiency performances.

Multiplicity moving as expectedmoving as expected => lowering as threshold increase. Efficiency decreasing a bit (about 12%). ASIC’s dynamic range to short to have à kind of MIP spectraMIP spectra.

Next ASIC version will have a larger dynamic rangelarger dynamic range, so we can make this measurement again.

Eudet Telescope

Moveable table

photogrammetric spotsused for alignment

I.Laktineh-IPNL 112

Spots positioning on both EuDet telescope and DHCAL Setup

Using EuTelEuTel, we can evaluate efficiencyefficiency on the detector edgesedges, and between two pads.

Black (Trigger): spatial prediction of hits in GRPC, from EuTel.

Red : matched digital hits (EuTel + GRPC)

Efficiency:Efficiency: Red/Black

Angle scan:Angle scan:

Efficiency quite constantquite constant, even forlarge angles

Fvolution of performances with particle flux :Fvolution of performances with particle flux : Correlation with particle particle

flux flux (obtained with scintillators ), and chamber’s efficiency.

It gives us some preliminary results about GRPC running in ILC beam conditionsILC beam conditions.

96%

The firsts tests using COCO22 to replace isobutane are quite promising.

(Gas mix with C02 will be intensively testedintensively tested next test beam)

Hadronic shower are mostly uncontainedmostly uncontainedin MiniDHCAL (0.5 i)but these profiles gives a first ideafirst idea of shower development,and energy deposition.

Hadronic showersHadronic showers

BeamGRPCsGRPCs

GRPC 1GRPC 1

GRPC 2GRPC 2

GRPC3GRPC3

GRPC4GRPC4

Muon contamination area

Beam(pions)

Blue: 1st threshold

Red: 2d threshold

TILC09 April 17-21 Tsukuba

kieffer@ipnl.in2p3.fr 118

Distribution of total number of hits in mini DHCAL for test beam datatest beam data and geant4 simulated datageant4 simulated data.

PreliminaryPreliminary

Muon contamination area

Preparation for the technological prototype

Technology drivers :• Closed chamber design – no external gas-tight box• Reduce the dead zones: spacers, frame• uniform resistive coating • Low cost• ScalableComponents Borosilicate glass

Anode: 0.7 mmCathode: 1.1 mm

Resistive layer (~ 20μ)Graphite, Licron’ (polymer), Statguard’ (oxides of Fe, Ti)

Insulation layers – mylar:175 μ cathode side (HV ~7.5 kV)50 μ anode side (0 V)

GRPC activities

1M2 for the technological prototype

• Two types of chamber:– ‘Standard’ chamber

• Frame in G10, thickness 1.2 mm, width 3 mm• ‘Channelled’ gas distribution – ‘2 fishing lines’ (PMMA)

– ‘Capillary’ chamber• Capillary tube frame 1.2X.8 mm• Frame used to distribute gas (0.3 mm holes drilled in capillary

walls)• Advantage: reduction of dead zones

• Support between glass planes:– Ceramic balls diam. 1.2 +/- 0.02 mm– Distance between balls optimized (ANSYS): 100 mm (max. deformation 44 μ – 81 balls / m2)

Mechanical deformation of the detector

Deformation with HV and ceramics balls

Gas pressure reduces this

Gas distribution, ‘standard’ chambersGas distribution, ‘standard’ chambers

Simulation – gas circulation in standard chamber

Gas distribution,‘capillary’ chambers

Capillary 1.2 x 0.8

Anode glass

Cathode glass

PEEK capillary

PEEK joint

Gas distribution holes ∅= 300 microns

Simulation – gas circulation in capillary chamber

M2 GRPC Status4 chambers of 1M2 were built up to now in Lyon • All with the standard gas distribution system

• 2 with Licron (aerosol ,ρs ~ 30 MΩ/□)

• 1 with Statguard ( liquid, 500 MΩ/□ !!!!!)• 1 to be coated with Statguard (silk screen printing)

Three kinds of problems were encountered and solved:• Gas tightness

• High voltage connection

• Resisitivity control

Construction steps

– Clean glass and cover with resistive coating– Glue micro-balls, frame, gas spacers and capillary

tubes to cathode glass on gluing table– Add glue to upper surfaces of balls and gas spacers– Turn table to vertical position– Introduce anode glass– Turn table to horizontal position– Deposit glue lines between glass and frame to

make gas-tight– Glue 6mm gas connectors to capillaries and solder

HV connectons– Transfer to honeycomb support

Gas tightness

• First chambers inflated under gas pressure!

• Glue failure caused balls to become detached from upper glass

• Subequent failure of glue around perimeter → gas leaks

• Over-pressure in chamber not excessive (Δpexit ~2.5 mbar ≡ 250g / ball max.)

deltaP capillaire

0,00

0,20

0,40

0,60

0,80

1,00

1,20

1,40

1,60

1,80

2,00

0 0,5 1 1,5 2 2,5 3 3,5 4 4,5

Débit (l / hr)

Del

ta P

(m

bar

)

Glue test

• Usual glue – two-component epoxy AY103 + HY951: 2.7g/cm2• Dow Corning RTV Silicone 3140: 5.0g/cm2• Araldite epoxy 2011 / 2012: 108 g/cm2

HV connections

• Recurring problem – loss of HV connection on Licron chambers

• Apparent thinning of Licron layer near the copper strip glued to the glass

• Occurred using: After short time (few days to a couple of weeks)– Copper Scotch with conductive adhesive– Copper strips glued with silver-loaded varnish

• Solutions found:– Graphite Scotch– Epotek EE129 conductive epoxy

Both solutions seem to work up to now

Statguard resistivity (1)

• Commercial product used for ESD protection of floor surfaces

• Potential to silk-screen print onto glass• Relatively inexpensive• Good surface finish• Small chamber in Nov. 08 test beam performed

reasonably well (efficiency, multiplicity) Vincent talk

• 1M2 Statguard chamber in same test beam had static build-up problem → few HARDROCs damaged due to charge breakdown

This is due most probably to the very high Statguard resistivity (500MΩ/□ )

Statguard resistivity (2)

• Resistivity not easily controllable:– Varies from 10 MΩ/□ to >500 MΩ/□ for no

apparent reason– Same glass cleaning procedure– Same method of deposition (roller)– Same number of layers and approximate layer

thickness• Recent tests indicate roller may be to blame• Consistent results (~25 MΩ/□ for 1 coat) with

paint brush or skimmer• Silk-screen printing method has been investigated

Silk-screen printing method

• Silk-screen printing method provides a uniform thickness.• Suitable for coating of large surface detectors • Different screen configurations were tested using Statguard to obtain the needed resistivity• other coatings will be tested (colloidal graphite)

Resistivity evolution with time after silk-screen painting

• Resistivity depends on the layer thickness (up to some extent) • Using the screen structure allows to determine the thickness (less fibers/cm more paintingthicker layerless resistivity)

Statguard painting

GRPC activitiesMGRPCBologna-CERN(C.Williams)

•5 glass plates of 400 µ each4 gaps of 250 µ using fishing line as spacersand Licron as resistive coating

•32X8 cm2 MGRPC was built and tested with the SDHCAL electronics

•1M2 multigap GRPC is beingbuilt and will be tested with the same 1M2 SDHCAL electronics

Cathode -10 kV

Anode 0 V

(-2 kV)

(-4 kV)

(-6 kV)

(-8 kV)

Signal electrode

Signal electrode

10 11 12 13 14

1000

100

10

1

Applied Voltage [kV]

2 mm gap RPC with C2F4H2 gas mixture

2 mm gap RPC with C2F5H gas mixture

10 gap (250 micron) double stack MRPCwith C2F4H2 gas mixture

knee of eff. plateau

knee of eff. plateau

knee of eff. plateau

Charge produced by through-going charged particle [pC]

Semiconductive glass

210mm*70mm*0.7mm

~1010.cm

Rate: 28 k Hz/cm2

Semiconductive ceramics

80mm*50mm*1mm

106 ~109.cm

GRPC activitiesHigh rate GRPCTsinghua University (Y.Wang)

Few small chambers will be tested with the SDHCALIn the next TB at cern.

DAC output (Vth)

Trigger

25 µs

PWR ON

HR2

4.7 mm

4.3m

mElectronics readout

3 thresholds with “independent” gain correctionNew PCBs are under design for a second M2

FSB0 scurves: HR1 /HR2 before and after gain correction

9% 5% 6.7% 1.4%

HR1 before cor.

HR1 after cor.

HR2 before cor.

HR2 after cor.

Qinj=100fC

Pedestal substracted

PCB connection : Assemble 2 X 3 PCBs on 1M2 Support

hardrocHoles

Readout electronicsDIFs (×120) ASUs

DAQ PC

DCC

Clock& Control

LDA

ODR

LDA

DCC

×10×9

:×14

:

DAQ

DHCAL DIFDHCAL DCC

PCB DESIGN

50 cm3

3.3

cm

1536 pads on Bottom Layer

DIF connector

ASU to ASU connector

ASU to ASU connector

Power and Gnd ConnectorASU to ASU on X axis

GND Connector ASU to ASU on Y axis

Y

X

HR1

HR24

GND Connector ASU to ASU on Y axis

Buffers(Other signals)

Buffers(Clocks)

Buried and Blind Vias (Same as the last PCB with 4 HR)

30 holes for M1 screws were distributed on the PCB for fixation on the absorber

T7 T7

1

2

3

1

2

3

weld

Via

0 ΩResistance

4

4 Pist

2 3

Schematic view of possible connections between two pcb

Electrical issue

4 PCB with 8X8 pads with the same structure of those used in the4-Chip ones are built to study the connection and its effect on the signal.

ER

i

Retrour de i

ER

i

Retrour de i

ER

i

Retrour de i

ER

i

Retrour de i

For low frequency the signal is not affected by the discontinuity

For high frequency the signal is affected by the discontinuity(an inductance of 1nH/cm)

The measurements we realized confirm the prediction and give an estimateof the number of discontinuities that we may allow for one detector plan.

Electronics readout status for M2 detector

• 8 PCB of 50X33.3 cm2 were conceived and produced• 8-layer, class 6 (buried vias)• 6 were equipped with hardroc1 (plastic packaging) 144 ASICs• PCB are connected 2 by 2 using zero resistor

DIF Slab 1 Slab 2

Electronics reqdout status for M2 detector

Problems found and fixed :

• Slow control and data readout failure: “Clock signal arriving before data signal after few ASICs”

buffers added (2/24 asics) critical line were adapted to avoid reflections

• DIF firmware failures state machines “latched” external trigger system correctly implemented

Data taking with cosmics started last week with one PCB-doubletIf ok we equip one 1M2

Hardroc 42 has been hit on channel 1Charge injected is 3.2pC (1.6V on 2pF)

1m2 GRPC chambers were tested with the small electronics board (4HR1)

PCB-doublets (3072 ch each) are tested independently

PCB-doublet on 1m2 are being tested

Fully equipped large detector to be soon tested in cosmic rays bench and in test beam at cern in June09

X(cm)Y(cm)

COSMICS

1 m2 GRPC

3-plan trigger

Slab #1

Slab #2

Slab #3

DIF #1

DIF #2

DIF #3

GRPC

Preparation for the 1M3 technological prototype

The aim is to come as close as possible to what we would like to have for ILC.

Technological prototype :40 planes of 1M2 :16mm s.steel absorber4mm s.steel support6mm GRPC

Important points:Semi-digital readout, mechanical structure, gas systemDAQ, event building, data storage.

g sense

Mechanical structure for the 1 m3

1 CNC machine of aprox 4x1 m2 working table. Accuracy of aprox 0.03 mm/m, with temperature compensation.

This is the machine that can be used to produce the plates for the HCAL prototype.

CIEMAT

Preparation for the 1M3 technological prototype

• Pions with different energies were simulated to better understand the containment• Analyses to exploit the three thresholds have started by having an idea of the energy/particles going in one pad• Work has started to develop algorithms for energy reconstruction using the 3 thresholds • Digitization is worked out.

ECAL DHCAL

100 GeV pions

1 GeV Muon:1Mip ~ 500eV

Assume Eion = 20eV1Mip ~ 25 ionization

G0, mean gain

For muon of 1 GeV

Conclusion

Digital hadronic calorimeters can be a powerful tool for physics achievement in the future ILC. They provides the granularity needed while keeping homogeneity and efficiency high

Still many efforts are needed to build a technological1m3 prototype and smart ideas to exploit it….