HAL Id: in2p3-00024894
http://hal.in2p3.fr/in2p3-00024894
Submitted on 17 Oct 2005
HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires
Construction, assembly and tests of the ATLAS electromagnetic barrel calorimeter
B. Aubert, B. Beaugiraud, J. Colas, P. Delebecque, L. Di Ciaccio, M. El Kacimi, P. Ghez, Christian Girard, M. Gouanère, D. Goujdami, et al.
To cite this version:
B. Aubert, B. Beaugiraud, J. Colas, P. Delebecque, L. Di Ciaccio, et al.. Construction, assembly and tests of the ATLAS electromagnetic barrel calorimeter. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, Elsevier, 2006, 558, pp.388-418. �10.1016/j.nima.2005.11.212�. �in2p3-00024894�
CERN PH EP=2005 034
27 July2005
Constrution, assembly and tests of
the ATLAS eletromagneti barrel
alorimeter
ATLAS Eletromagneti Barrel Liquid Argon Calorimeter
Group
Abstrat
The onstrutionand assembly of the twohalf barrels of the ATLAS entral
eletromagneti alorimeterand theirinsertion intothe barrelryostat are de-
sribed. The results of the qualiationtests of the alorimeter before installa-
tion inthe LHCATLAS pitare given.
(submitted to Nul. Instrum. Methods A)
|||||||||||||||||||{
) See next pages for the list of authors
L. Di Ciaio, M. El Kaimi a;b;
, P. Ghez, C. Girard
M. Gouanere, D. Goujdami a;b;
, A. Jeremie, S. Jezequel,
R. Lafaye, N. Massol, P. Perrodo, H. Przysiezniak, G. Sauvage 1
J. Thion, I. Wingerter-Seez, R. Zitoun, Y. Zolnierowski
Laboratoire de Physique dePartiules (LAPP), IN2P3-CNRS,
F-74941 Anney-le-VieuxCedex, Frane b;
.
R. Alforque, H. Chen, J. Farrell, H. Gordon,
R. Grandinetti, R. Hakenburg, A. Homann, J. Kierstead,
J. Koehler, F. Lanni, D. Lissauer, H. Ma,
D. Makowieki, T. Muller, S. Norton, V. Radeka,
D. Rahm, M. Rehak, S. Rajagopalan, S. Resia,
K. Sexton, J. Sonderiker , I. Stumer, H. Takai
Brookhaven National Laboratory (BNL) Upton, Ny 11973-5000, USA.
A. Belymam, D. Benhekroun, C. Driouihi,
A. Hoummada, M. Hakimi
Faulte desSienes An Chok,Casablana, Moroo b;d
.
M. Knee, R. Stroynowski, B. Wakeland
Southern Methodist University, Dallas, Texas 75275-0175, USA.
V. Datskov, V. Drobin
Joint Institute for Nulear Researh, Dubna.
M. Aleksa, J. Bremer, T. Carli, M. Chalifour e
,
J.L. Chevalley, F. Djama f
, L. Ema, C. Fabre,
P. Fassnaht, F. Gianotti, A. Gonide, J.B. Hansen g
,
L. Hervas, T. Hott, C. Laaste, C.P. Marin,
P. Pailler, A. Pleskath h
, D. Sauvagey, G. Vandoni,
V. Vuillemin, H. Wilkens
European Laboratory for Partile Physis (CERN), CH-1211 Geneva 23,
Switzerland.
D. Dzahini, A. Ferrari, J. Fulahier, M.L. Gallin-Martel,
J.Y. Hostahy, G Laborie, F Ledroit-Guillon, P. Martin,
J.F. Muraz, F. Ohlsson-Malek, S. Saboumazrag, S. Viret
Laboratoire de PhysiqueSubatomique et de Cosmologie (LPSC),UniversiteJoseph
Fourier, IN2P3-CNRS, F-38026 Grenoble, Frane b;d
.
R. Othegraven, C. Zeitnitz
Johannes Gutenberg-Universitat Mainz, Institut f ur Physik,D-55099 Mainz,
Germany.
D. Ban, L. Carminati, D. Cavalli, M. Citterio, G. Costa,
M. Delmastro, M. Fanti, L. Mandelli, M. Mazzanti, F. Tartarelli
Dipartimento di Fisia dell'Universita di Milano and INFN
I-20133 Milano, Italy.
E. Auge, S. BaÆoni, J. Bonis, W. Bonivento i
,
C. Bourdarios, C. De la Taille, L. Fayard, D. Fournier,
G. Guilhem, P. Imbert, L. Ionomidou-Fayard,G. Le Meur,
M. Menik, J-M. Noppe, G. Parrour, P. Puzo,
D. Rousseau, A-C. Shaer, N. Seguin-Moreau, L. Serin,
G. Unal, J-J. Veillet, F. Wiek, D. Zerwas
Laboratoire de l'Aelerateur Lineaire, Universite deParis-Sud, IN2P3-CNRS,
F-91898 Orsay Cedex, Frane.
F. Astesan, W. Bertoli, B. Canton, F. Fleuret,
D. Imbault, D. Laour, B. Laforge, Ph. Shwemling 1
Laboratoire de PhysiqueNuleaireet de Hautes Energies,Universite
Paris VIet VII, IN2P3-CNRS, F-Paris,Frane.
E.M. Abouelouafa, A. Ben Mansour, R. Cherkaoui, Y. El Mouahhidi,
H. Ghazlane, A. Idrissi
Faultedes Sienes and CentreNational de l'Energie, desSienes et Tehniques
Nuleaires, Rabat, Moroo b;d
.
F. Lobkowizy, P. Slattery
University of Rohester, River Campus, B&L Building, Rohester, New York
14627-0171, USA.
J. Belorgey, N. Besson, M. Boonekamp, D. Durand,
J. Ernwein, B. Mansoulie, F. Molinie, J.P. Meyer,
P. Perrin, J. Shwindling, J.P. Taguet, H. Zaone
DAPNIA, CEA Salay, F-91191 Gif-sur-Yvette Cedex, Frane.
B. Lund-Jensen, S. Rydstrom, Y. Tayalati
RoyalInstitute of Tehnology, Stokholm, Sweden b
.
B. Bothev, G. Finohiaro, J. Homan, R.L. MCarthy,
M. Rijssenbeek, J. Steens, M. Zdrazil
State University of New York,Stony Brook, NewYork 11794, USA.
H.M. Braun
University of Wuppertal, Wuppertal, Germany.
(a) visitorfrom LPHEA,FSSM-Marrakeh (Moroo).
(b) partlysupportedbythe"Calorimetrieeletromagnetique a argonliquide
d'ATLAS"GDRI betweenIN2P3/CNRS,theuniversitiesJosephFourier of
Grenoble,of Mediterranee ofAix-Marseille II, and of Savoie, theMoroan
CNRSTand KTHSweden.
()partly supported bythe PAI MA/02/38.
(d)partly supported bythePAI MA/01/05.
(e)also CEA-Salay.
(f) on leave from CPPM,Marseille.
(g)now at NielsBohr Institute,Copenhagen.
(h)on leave from IFVE, Russia.
(i)nowat INFN-Cagliari.
y Deeased.
The eletromagnetibarrel alorimeterof the ATLAS experiment atCERN's
futureproton-protonollider,theLHC,isalead-liquidargonsamplingalorime-
terwithaordionshaped absorbersandeletrodes.Calorimetrywillbearu-
ial tool for the understanding of proton-proton ollisions at the LHC, sine
manyphysisproesseswillmanifestthemselvesthroughnalstateswithele-
tronsorphotons:Wand Zprodution,H !,H !4e,H !WW.Preise
measurements of the properties of the eletrons and photons is of utmost
importane, and dynami range, resolution and uniformity are the main pa-
rameterstobeoptimized.Fromphysissimulations[1℄,ithasbeendetermined
that the eletromagneti alorimetershould meet the followingrequirements:
ForaHiggsboson deayingtotwophotons ortofoureletrons,inthemass
rangefrom90to180 GeV/
2
,ATLAS shouldmeasure the Higgsmass with
1% preision using the alorimeter system alone. This translates into the
requirements of a sampling term of less than 10%=
p
E, assoiated with a
onstant term better than 1%.
The alorimetermust beable to separate the two photons from a 0
deay
up totransverse energies of 60GeV.
The dynami range has to over 30MeV up to 1TeV, i.e. fromthe typial
noiselevelinone singleelluptothe singleellenergy depositionexpeted
inthe ase of the deay of a heavy Z 0
orW 0
with masses up to 6TeV = 2
.
Liquid argon alorimetryhas been hosen for ATLAS beause of its intrinsi
linear behaviour, stability of the response over time and radiation tolerane.
The aordion geometry has been hosen beause it allows avery high gran-
ularity and a very good hermetiity, sine the high voltage and signal ables
runonlyatthefrontandbakfaesofthedetetor.Inaddition,theaordion
geometry minimizes indutanes in the signal paths, allowing the use of the
fastshapingneededforoperationwith25nsbunhintervalsbetweenollisions
atLHC.
The rst studiesof liquidargon alorimetryfor LHC date bak to 1990[2{5℄.
A \2 meter" long prototype was built by the R&D ollaboration RD-3, to
evaluate the performanes of liquidargon alorimetryfor a LHCexperiment.
After the hoie of the liquid argon tehnology by the ATLAS ollaboration
for its eletromagnetialorimetry in1995, the geometryof the detetor and
1
Correspondingauthors: sauvagelapp.in2p3.fr,shwemlilpnhep.in2p3.fr.
reassessed in view of assembly line prodution, using the experiene gained
withthe 1992prototype.The assembly lineprodutiononept wasvalidated
bytheonstrutionin1998andextensivetestingofaseondprototype,alled
Module 0,usingnal orloseto nalseries elementsforitsonstrution.The
fabriation of some of the elements of the alorimeter, like the absorbers,
started in the beginning of 2000, and the staking of the modules took plae
between 2001 and mid-2003.During the staking, four modules were studied
inbeam tests.
The paper is organized as follows. In Setion2, a general desription of the
barrel alorimeter is given and the onstrution of the main elements ne-
essary to assemble a barrel module is desribed. The staking of the barrel
modules is the subjet of setion3. The old eletronis and the ablingof a
module are desribed insetion4. The dierent monitors assoiated with the
alorimeter are detailed in setion5. A desription of the barrel ryostat and
assoiated feedthroughs is given in setion6. The assembly of 16 modules in
a halfbarrel is desribed in setion7. The insertion of the two half barrels in
the barrel ryostat and the qualiation tests performed at room and liquid
argon temperature onthe entire alorimeter are the subjet of setion8.The
mainresults of the alibrationof the 4modules testedwith aneletron beam
are summarized insetion9. Finally, the onlusions are given insetion10.
2 Calorimeter desription
TheATLASeletromagnetibarrelalorimeter,shown inFig.1anddesribed
indetailin[6℄,ismadeoftwohalfbarrels,enteredaroundthez-axis(ATLAS
beam axis). One half barrel overs z > 0 (pseudorapidity > 0) and the
otheronez < 0( < 0),fromjj = 0tojj = 1:475;thelengthofeahhalf
barrel is 3.2m, the inner and outer diameters are 2.8m and 4mrespetively.
The alorimeter is loated behind the superonduting solenoid, in the same
ryostat and weighs 114ton. Sine the material in front of the alorimeter
amounts to about 1.5X
0
on average, the barrel alorimeter is omplemented
with a liquid argon presampler detetor plaed in front of its inner surfae,
over the full range.
A half barrel is made of 1024 aordion shaped absorbers, interleaved with
readouteletrodes.Theeletrodesarekeptinthemiddleofthegapbyspaers.
The driftgap on eahside of the eletrode is 2.1mm, whihorresponds to a
total drifttime of about 450nsfor an operatingvoltage of 2000V.
One assembled, there is no disontinuity along the azimuthal angle; but
for ease of onstrution, a half barrel is divided into 16 modules. The total
0
22X
0
to30X
0
between jj=0and jj=0:8,andfrom24X
0
to33X
0
between
jj = 0:8 and jj = 1:3.
A module (Fig.2) has three ompartments in depth (front, middle, bak).
The readout granularity of the dierent ompartments is shown in Table1
for < 1:3. In total there are 3424 readout ells per module, inluding the
presampler ells.
Table 1
Readout segmentation and depth(in X
0
) ofthe ompartments ofa barrelmodule
and its assoiatedpresamplersetors for < 1:3.
Compartment X
0
Presampler 0:025 2=64 0.08to 0.15
Front 0:025=8 2=64 2.5to 4.5
Middle 0:025 2=256 16.5 to 19
Bak 0:050 2=256 1.4 to 7
2.1 Module omponents
The main omponents of amodule are listed below:
Absorbers: the absorbers (Fig.2and3) are made of lead plates (thikness
1.53mm for < 0:8, 1.13mm for > 0:8), glued between two 0.2mm
thik stainless steel sheets by resin-impregnatedglass ber fabri. The two
lead thiknesses have been hosen to ensure a depth of at least 22X
0 2
for
the alorimeter. The derease inlead thikness after =0:8 limitsthe de-
rease of the sampling fration at high . The ber fabriompensates for
the dierene inthikness of the two types of lead plates so that the nom-
inal thikness of an absorber is 2.2mm. The stainless steel sheets provide
mehanial strength tothe absorber.
G10preision bars: at the inner and outer edges, eah absorber is enased
into the groove of a preision mahined G10 berglass-epoxy omposite
bar. The purpose of these bars (Fig.4) is to position eah absorber with
respet to its neighbours, and also to provide spae for the onnetors of
theeletrodes.Theultimategeometrialaurayoftheassembledylinders
is a diret onsequene of the preise mahining of the azimuthal ontat
surfaesof the bars.
2
Inludingthe materialinfront,theminimaldepthis 24X
0 .
toa half barrel.Eah ring (Fig.5) is made of 16ring-piees orresponding
to the 16 modules. Similarly, 8 omposite inner rings help to dene the
inner geometry of a half barrel. Eah inner ring (Fig.6)is also made of 16
ring-piees.
Eletrodes:thereadouteletrodes(Fig.7)onsistofthreeondutiveopper
layers separated by insulatingpolyimidesheets.The twoouter layersare at
the high voltage potential; the inner one is used for readingout the signal
through apaitive oupling. Eah gap between two absorbers is equipped
withtwo eletrodes: typeA ( < 0:8) and typeB ( > 0:8).
Spaers: the spaers keep the eletrodes entered between two onseutive
absorbers. Eah spaer (Fig.8)onsists of strips of honeyomb laidon the
at setionsof the aordion absorber.
2.2 Presampler
The presampler is needed to orret for the energy lost upstream of the
alorimeter, espeially at low energy. For this purpose a separate liquid ar-
gon presampler, providing shower sampling in a thin liquid argon layer (11
mmindepth), is plaedinfront of the eletromagnetialorimeter insidethe
ommon barrel ryostat. It is made of 64 idential azimuthal setors (i.e. 32
fora halfbarrel).Eah setoris3.1mlong and0.28mwidethusovering the
halfbarrellengthandgivingaoverageinandof1.52and0.2respetively.
It is omposed of eight dierent sized modules with a length inreasing with
to obtainthe same overage of 0.2for eah module exept the one atthe
end of the barrel whose overage is redued to 0.12. The eight modules of
a setor are housed in an external envelope whih onsists of a 0.4mm thin
glass-epoxy shell. Detailsof the ontrutionan be found inref.[7℄.
The modules are made of interleaved athode and anode eletrodes glued
between FR4 3
plates. The eletrode spaing slightly varies with presampler
module type from 1.9 and 2.0mm. The athodes are (270 30)m thik,
double-sided printed iruit boards while the (330 30
40
)m thik anodes have
3 ondutive layers separated by nominally 150 and 130m FR4 layers. The
requiredgranularity( = 0:025, = 2=64)for eah moduleisobtained
byputtingtogether theappropriatenumberofanodesinthe(orz)diretion
andbysubdividing(i.e.ething)eahanodeintotwohalvesinthe-diretion.
A +2kV high voltage potential is applied to the outer layers of the anodes
and the signal is readout through apaitive oupling to the entral layer
at ground potential. Eah anode has four 1M surfae-mounted resistors to
feed the high voltage. Quality assurane tests of the anodes prior to module
3
Fiberglassepoxy omposite.
urrentatthisvoltageandaRC-measurementtoverifytheonnetivityofthe
dierent layers and the soldered resistors. The anode prodution and quality
is doumented separately[8℄. Fig.9shows a presamplermodule.
A mother board whih is a ve layers printed iruit is mounted on eah
module. The mother boards, whih ome in eight dierent sizes to math
the modules, ollet signals from the readout ells. In addition the mother
boards are equipped with a set of aurate surfae-mounted resistors (0.1%
and 25ppm/
Æ
C) for the injetion of alibration pulses.
The same type of HV, alibration and signal ables (see subsetion4.2.1) as
for the alorimeter are used. Two separate HV lines per pair of modules,
onneted to the two sides of the anodes, are used to prevent the omplete
lossof a module inase of short-iruits.
Two assembly sites were used, eah equipped with a station for testing se-
tors in liquid nitrogen. Eah setor was rst measured at room temperature
with a test benh where the response to a low frequeny, 10kHz, sinusoidal
signalinjeted through eah ofthe HV ableswasregistered (alledTBFtest
later).Toeliminatedeviationsdue todierent readout ampliersbeingused
for dierent signal hannels, the response was also measured when the same
sinusoidalsignalwasinjetedthrough the alibrationables.Thisallowedthe
veriation of the onnetivity of all ables and anodes, inluding the 4 re-
sistors on eah anode. The test also inluded registering the response to a
triangular test pulse injeted through the alibration ables. The same tests
were then repeated at liquid nitrogen temperature. Any failed onnetivity
was, if possible, repaired and the old test repeated. Prior to the old test, a
voltageof 1kV wasappliedtothe anodes, for atleast one night.Finally eah
setor wastested at 2kV inliquid nitrogenforat least 24hours.
A total of 66 setors, inluding 2 spares, have been produed and validated.
The validatedsetorswere storedin drynitrogengasinside aboxmade ofan
anodized extruded aluminium prole into whih the setor ould slide. This
limitedthe humidity until installationonthe barrel alorimeter.
2.3 Absorbers
2.3.1 Lead sheet prodution
Thikness variations of the lead plates used to build the absorbers indue
loalvariationsof the alorimeterresponse, whihontribute tothe onstant
term. Typially,aninrease of1% ofthe leadplate thikness leads toa 0.5%
derease of the measured signal.
inindustry.Thethikness toleranethat themanufaturerwas readytoguar-
antee, was 30m (rms of 17m). This would translate into a ontribution
of about 1% to the onstant term in the worst ase, when neighbouring ab-
sorbers had orrelated thikness utuations. To avoid suh ontributions to
the onstant term, two steps were taken:
AnX-raymeasurementsystemwasset up andusedduringtherollingoper-
ationtohelpthemanufaturerinreduingthedispersionofleadthiknesses
withinthe allowedtolerane interval.Withthis real timemonitoring ofthe
thikness, the operator ouldreat very fast todrifts observed inthe sheet
thikness. In addition,the measurement system provided the possibility to
rejet out-of-tolerane lead. In this way, the lead sheets delivered to the
ollaborationhad a thikness r.m.s variationof 6.6m.
An ultrasound measurement system, with an auray of a few m was
usedtoproduedetailedthikness maps (ona55m 2
grid)for eahlead
plate. Using the fat that a typial shower develops, on average, in ve
onseutive absorbers, the detailedthikness maps of the plates have been
used to optimize the arrangement of onseutive absorbers, by ensuring
that neighbouring absorbers have opposite thikness utuations. It was
established[9℄ that the onstant term ould be estimated from the r.m.s.
variation of the normalized thikness of the lead plates, averaged over 5
onseutive plates. From the thikness measurements, it an be estimated
that the residual ontribution tothe global onstant term arising fromthe
lead thikness utuations has been redued to0.19% (Fig.10).
2.3.2 Absorber fabriation
First, a at sandwih of stainless steel, resin-impregnated glass ber fabri
andleadplateswasassembledfromtheseparate parts.Thentheatsandwih
was bent to an aordion shape, using a bending mahine with the following
priniple: theupperand lowerbendingknivesof the mahine were positioned
with a preision of 0.1mm with respet to eah other and to the entral top
knife,whihwasxed. The otherknivesouldbefreed tomovelaterally.The
sandwih rst got pinhed between the two sets of knives. At this stage, the
kniveswere freed, and theupperkniveswere pushed down by thepress, while
simultaneously all the knives moved laterally towards the xed knife in the
enter.The driving forewasthefrition between the surfae ofthe sandwih
and the tips of the knives. After bending, the sandwih was ured at 125 Æ
C
inaheatingpress whih onsisted oftwosingle-pieesteel matriesmahined
with a 0.02mmtolerane with respet to the absorber shape at 125 Æ
C. This
heating press applieda programmablepressure on the nished sandwih, up
to 7 bars, using mehanial jaks, and followed a omputer-ontrolled uring
yle. Sine the nal shape of the absorber was dened by the parameters
