148 Pages
English

Development of a cryogenic silicon detector system and study of strange particle production in deep inelastic scattering [Elektronische Ressource] / by Rita De Masi

Gain access to the library to view online
Learn more

Description

DevelopmentofacryogenicsilicondetectorsystemandstudyofstrangeparticleproductionindeepinelasticscatteringDissertationbyRitaDeMasiPhysik DepartmentE18TechnischeUniversitat¨ Munchen¨July2004Fakultat¨ fur¨ PhysikderTechnischenUniversitat¨Munchen¨PhysikDepartmentE18DevelopmentofacryogenicsilicondetectorsystemandstudyofstrangeparticleproductionindeepinelasticscatteringRitaDeMasiVollstandiger¨ Abdruck der von der Fakultat¨ fur¨ Physik der Tech nischen Universitat¨ Munchen¨ zur Erlangung des akademischenGradeseinesDoktorsderNaturwissenschaften(Dr. rer. nat.)genehmigtenDissertation.Vorsitzender: Univ. Prof. Dr. A.J.BurasPrufer¨ derDissertation:1. Univ. Prof. Dr. St. Paul2. Univ. Prof. Dr. F.vonFeilitzschDie Dissertation wurde am 25/10/04 bei der Technischen Univer-sitat¨ Munchen¨ eingreicht und durch die Fakultat¨ fur¨ Physik am9/11/04angenommen.ContentsIntroduction 11 TheCOMPASSexperiment 51.1 Thephysicsaims . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51.1.1 Physicswiththemuonbeam . . . . . . . . . . . . . . . . . . 51.1.2withthehadronbeam . . . . . . . . . . . . . . . . . 101.2 TheCOMPASSdetector . . . . . . . . . . . . . . . . . . . . . . . . . 111.2.1 Thepolarisedbeams . . . . . . . . . . . . . . . . . . . . . . . 121.2.2 Thespectrometer . . . . . . . . . . . . . . . . . . . . . . . . . 131.2.3 ProcessingofCOMPASSdata . . . . . . . . . . . . . . . . . . 192 ThebasicprinciplesofsilicondetectorsandtheLazaruseffect 232.

Subjects

Informations

Published by
Published 01 January 2004
Reads 37
Language English
Document size 18 MB

Developmentsystemandofastudyofcryogenicstrangesiliconparticledetector
scatteringinelasticdeepinproduction

Dissertation

Rita

by

De

Masi

Physik-Department

echnischeT

at¨Universit

July

2004

E18

¨Munchen

Fakult¨atf¨urPhysikderTechnischenUniversit¨at
unchen¨ME18DepartmentPhysik

Developmentofacryogenicsilicondetectorsystem
andstudyinofdeepstrangeinelasticparticlescatteringproduction

MasiDeRita

Vollst¨andigerAbdruckdervonderFakult¨atf¨urPhysikderTech-
nischenUniversit¨atM¨unchenzurErlangungdesakademischen
einesGrades

DoktorsderNaturwissenschaften(Dr.rer.nat.)

Dissertation.genehmigten

Vorsitzender:Univ.-Prof.Dr.A.J.Buras
Dissertation:derufer¨Pr1.Univ.-Prof.Dr.St.Paul
2.Univ.-Prof.Dr.F.vonFeilitzsch

DieDissertationwurdeam25/10/04beiderTechnischenUniver-
sit¨atM¨uncheneingreichtunddurchdieFakult¨atf¨urPhysikam
angenommen.9/11/04

Contents

Introduction

experimentASSCOMPThe11.1Thephysicsaims.............................
1.1.1Physicswiththemuonbeam..................
1.1.2Physicswiththehadronbeam.................
1.2TheCOMPASSdetector.........................
1.2.1Thepolarisedbeams.......................
1.2.2Thespectrometer.........................
1.2.3ProcessingofCOMPASSdata..................

2ThebasicprinciplesofsilicondetectorsandtheLazaruseffect
2.1Siliconmicrostripdetectors.......................
2.1.1Basicfeatures...........................
2.1.2Particledetection.........................
2.2Radiationdamage.............................
2.2.1Defectsinsilicon.........................
2.2.2Effectsoftheradiationonthepropertiesofthedetector...
2.2.3Annealing.............................
2.3Lazaruseffect...............................
2.3.1Experimentalevidences.....................
2.3.2TheLazaruseffectmodel....................

3TheimplementationofsilicondetectorinCOMPASS
3.1ThesilicondetectorinCOMPASS....................
3.2Thedetectordesign............................
3.2.1Processingcharacteristics....................

I

1

5551011121319

2324242729293133343435

39394040

4

5

3.2.2Geometry.............................41
3.2.3Radiationhardness........................41
3.3Theelectronics...............................43
3.3.1TheAPV25............................44
3.3.2TheL-board............................46
3.3.3Therepeatercard.........................47
3.3.4TheADCcard...........................47
3.3.5TheGeSiCA............................48
3.4Thecryostat................................48
3.5Thegroundingscheme..........................51
3.6Anoverallview..............................52

Preparationandinstallationofthefirstcryogenicsiliconstationin
55ASSCOMP4.1Laboratorysetup.............................56
4.2CharacterisationoftheAPV25chipatcryogenictemperatures...57
4.2.1Coldmeasurementsetup....................57
4.2.2APV25-S0characterisation....................58
4.3Testofthecomponents..........................66
4.3.1Testoftheglue..........................68
4.3.2Testoftheconnector.......................69
4.4Temperaturedistribution.........................69
4.4.1TemperaturedistributionontheL-board:measurements..70
4.4.2Temperaturedistributiononthewafer:simulation......73
4.5AcoldstationinCOMPASSexperiment................74
4.5.1Installationintheexperiment..................74
4.5.2The2003datataking.......................75
4.5.3Detectorperformances......................76
4.6Thecoolingcontrolsystem........................76
4.6.1Themechanicaldesign......................77
4.6.2Thecontroloftheux......................80
4.6.3Theprincipleofoperation....................81

−−ThesearchfortheΞpentaquarkinCOMPASS83
5.1Introduction................................83
5.1.1Theoreticalmodels........................84

II

5.1.2Experimentalresults.......................86
5.2TheanalysisoftheCOMPASSdata...................89
5.2.1DataSampleandluminosity..................89
5.2.2Eventtopology..........................90
5.2.3Preliminaryselectionofevents.................92
005.2.4ΛandΛreconstruction....................93
5.2.5ΞandΞreconstruction....................98
−−0∗0∗−−−−
5.2.6Ξ,Ξ,ΞandΞselection................100
5.2.7MonteCarlosimulations.....................109
05.2.8Ξreconstructionefficiencyandproductioncrosssection.112
∗5.2.9AnupperlimitfortheΞproductioncrosssection.....116
−−05.2.10Ξπselection...........................118
∗5.2.11Discussionoftheresults.....................120

OutlookandConclusions

Bibliography

Acknowledgements

OwnContributions

III

125

127

133

135

FiguresofList

1.1ThePhoton-GluonFusion........................
01.2TheDforopencharmanalysis.....................
1.3TheCOMPASSsetup...........................
1.4TheM2beamline.............................
1.5TheCOMPASSspectrometerin2003..................
1.6TheDAQarchitecture..........................
1.7Thedataprocessing............................

2.1TheBethe-Blochformula.........................
2.2Defectclustersinthesiliconlattice...................
2.3Displacementdamagecrosssection...................
2.4CCEvs.temperature...........................

3.1Detectorsarrangementupstreamofthetarget.............
3.2Crosssectionofthedetector.......................
3.3Readoutstripsonthedetectors.....................
3.4Layoutofthen-sideofthesilicondetector...............
3.5LayoutoftheAPV25readoutchip...................
3.6APV25outputframe...........................
3.7L-boardslayout..............................
3.8Schematicdrawingofthecryostatforthetargetregion.......
3.9Asiliconmoduleinthecryostat.....................
3.10Thegroundingschemeforthesilicon..................
3.11Viewofthetargetregion.........................
3.12Siliconreadoutchain...........................

671212141821

25303235

404142434445465050515252

4.1SchematicviewofLNux.......................56
24.2PCBforthecoldmeasurements.....................57

V

4.3Set-upforcoldmeasurements......................59
4.4APV25-S0outputdataframe......................59
4.5APV25-S0pulseshapedependenceat130K..............60
4.6APV25-S0pulseshapedependenceat130K..............61
4.7APV25-S0noisedependence.......................62
4.8APV25-S0pulseshape..........................63
4.9Dependenceofsignalamplitudefromthetemperature........63
4.10APV25-S0gainatdifferenttemperatures................64
4.11Noisevs.capacitanceatdifferenttemperatures............64
4.12APV25-S0pipelinepedestal.......................65
4.13r.m.s.pipelinepedestal..........................65
4.14Pipelinegainuniformityat130K.....................66
4.15APV25-S0pulseshapeafter50thermalcycles.............67
4.16APV25-S0noiseafter50thermalcycles.................67
4.17ArrangementofthewaferontheL-board...............68
4.18Epoxyconnector..............................70
4.19PCBforthermaldistributionstests...................71
4.20Temperaturevs.powerdissipatedfordifferentLNuxes.....72
24.21Temperaturevs.positiononthewaferfordifferentLNuxes...72
24.22Temperaturedistributiononthesiliconwafer.............73
4.23Thecoldsetupintheexperimentalhall................75
4.24Temperaturebehaviourduring2003run................76
4.25Flowscheme................................77
4.26Distributionbox..............................78
4.27Coolingsetup...............................81

5.1Thebaryonicantidecuplet10......................85
5.2TheNA49pentaquarksignal......................88
5.3Theeventtopology............................91
5.4Theprimaryvertexdistribution.....................93
5.5Thecosαdistribution...........................95
05.6TheKpeak................................96
−5.7ThepπinvariantmassspectraforthreedifferentintervalsofZ097
Λ005.8TheΛandΛ...............................98
05.9TheKpeak................................99

VI

+005.10TheΛπandΛπinvariantmassspectra.............100
−5.11Ξπinvariantmassdistributions...................102
−−005.12ΞandΞpeaks.............................102
∗∗5.13ThexdistributionforΞπinthepγCMS..............103
∗−F5.14Ξπinvariantmassdistributionindifferentintervalsofx....104
−−F+5.15Ξπinvariantmassdistributionindifferentintervalsofx....105
−F+5.16Ξπ/Ξπdistributionindifferentintervalsofx.........106
−−−F5.17Ξπinvariantmassdistributionindifferentintervalsofx.....107
−F05.18Ξindifferentintervalsofxbackgroundsubtracted........108
∗F05.19ΞVs.x..................................109
∗F05.20TheΛπinvariantmassspectrumfromdataandMonteCarlo..111
−+5.21TheΞπinvariantmassspectrumfromdataandMonteCarlo..111
−+5.22TheΞπinvariantmassspectrumfromdataandMonteCarlo
−backgroundsubtracted..........................112
+5.23TheΞπinvariantmassspectraindifferentxintervals......113
−F+5.24ThebackgroundsubtractedΞπinvariantmassspectraindiffer-
−entintervalsofx.............................114
F05.25TheΞefficiencyasfunctionofxintervals..............115
∗F05.26TheproducedΞasfunctionofx...................116
∗F5.27TheΞπInvariantmassspectraindifferentxintervals......117
−−F05.28Ξπinvariantmassdistribution....................119
∗5.29TheNA49pentaquarksignal......................122
5.30Ξπinvariantmassdistributions...................123
−−

VII

Introduction

Theandmainunderstandgoalofhowhightheyenergycombinephysicsinisortoderfindtotheformultimatematterinthecomponentsuniverseofasmatterseen
thattoday.matterAlriseadytheGrdiscontinuouseekandphilosophernameditsDemocritosconstituents(about400atoms.a.C.)Fromthehypothesisedendof
the19thcenturyseveralexperimentalobservationshaveconfirmedthediscon-
mentaltinuousstrmodelsucturehaveofthebeenmatter.developed:Sincenowthenwemanyknowthatscientificatomstheoriesareandcomposedexperi-of
onselectr)areons,notprotonspoint-like,andbutneutrtheyons.arePrinotonsturnandcomposedneutronsofpointlike(collectivelyfermions,callednucle-with
fractionalelectriccharge,calledquarks.Physicistsdiscoveredtheexistenceofsix
tiontypeofcomposedquarks,ofcalleduanddavours,quarks,u,d,buts,c,bquarksandt.combineNucleonstoarformeinalsofirstseveralapproxima-other
particleswithshortlifetimecalledhadrons.
Quarksaresujecttoelectromagnetic,weakandstrongforces.TheQuantum
Electrodynamics(QED)describessuccesfullytheinteractionbetweenelectrically
chargedparticles.Thestrongforceisresponsiblefortheformationofprotonand
neutronsanditsresidualeffectsaccountforthebindingofnucleonsintheatomic
nucleus.InanalogywithQED,theQuantumChromodynamics(QCD)describes
thestronginteractionsbetweenquarks.ThebasicideaofQCDisthatquarks
carryastrongcolourcharge,analoguetotheelectriccharge,whichcantakeon
threedifferentvalues,labelledred,blueandgreen.Quarksaresubjecttothecolour
forcefieldandtheyinteractviatheexchangeoffieldquantacalledgluons,which
carryalsocolour.ThegluonsinQCDplayananalogousroleasthephotonsin
QED.assumingPrtheesentlystrnoongfreeforcequarkstobehavesignificantbeenatobserved:largethisdistancespeculiarityandisnegligibleexplainedat
grshortedientofdistances.QCD.ThisThismeansbehaviourthatisaifresultquarksofartheeclosetoself-couplingeachofother(gluons,1afm)keytheyin-
tualbehaveattralikectionfreebecomesparticlesstrbut,ongeras.soonTheefasfecttheirofrthiselativeisthatdistancequarksincrremaineases,theconfinedmu-
inthehadronvolume,whichiscolourless,andonlythiscanbeobservedasfree
[CM84].[Per82],particle

1

Allthehadronswhichareobserveduntilnowareeitherformedofthreequarks
(baryons)oraquark−antiquarkpair(mesons).HoweverQCDdoesnotexclude
theexistenceofstateswithdifferentquarkcontent,providedthattheyarecolour-
less,i.e.multipleofthreecoloursorcolouranti-colourpairs.Thus,inprinciple
hybridstateswhichhaveexplicitlyexcitedgluoniccomponentsarepossible,as
wellasstatescomposedoftwoquark-antiquarkpairs(tetraquark,qqqq),ofthree
quarksandaquark-antiquarkpair(pentaquark,qqqqq)andsoon.Theyarecollec-
[jaf04].exoticscalledtivelyTheactualtheoreticalunderstandingofexoticstatesingeneralandofpen-
taquarksinparticularisnonethelessquitepoor.Severalmodelshavebeende-
veloped,intheframeworkofQCD,whichpredicttheexistenceofpentaquarks
andtheircharacteristics(spin,parity,lifetime,etc.).Atthispointexperimental
resultsareneededtoconfirmthevalidityofthepredictions.+From−−theexperi-
mentalside,evidencefortwopossiblepentaquarkstates,ΘandΞ,havebeen
giveninthelastyear.Thesestatesaresupposedtohaveminimumquarkcontent
uuddsandddssuforΘ+andΞ−−,respectively.ThestudiesofΘ+pentaquark
candidateswereperformedinphoto-andhadro-production,whiletheΞ−−was
studiedonlyinhadro-production.Howevertheclaimedpentaquarksignalshave
arathermoderatesignificance[LEP03],[na403]and,moreover,severalexperi-
mentsgavecontradictingresults.Thisclearlyasksformoreexperimentswith
higherstatisticstoclarifythematter.ThestudyoftheΞ−−pentaquarkcandidate
withintheCOMPASSexperimentistheanalysispartofthiswork.
COMPASS[COM96]isafixedtargetexperimentpresentlyrunningatCERN-
SuperProtoSyncrotron(SPS)usinga160GeVmuonbeam.Itsprincipalcharacter-
isticsarehighcountrateandhighprecision.Bothofthesefeaturesareneededin
ordertostudyexoticstates.OneoftheproposedgoalofCOMPASSisthestudy
ofexoticsusingahadronbeam;thistaskwillbeperformedinthenextyears.
However,thedataacquiredwithmuonbeamofferthepossibilityofperforming
systematicstudiesofpentaquarksinphoto-production.
ThehighprecisionoftheCOMPASSspectrometerisachievedwithanaccurate
trackingofincomingandoutcomingparticleswithrespecttothetarget.Aprecise
reconstructionofthebeamparticlesbeforetheyinteractinthetargetisobtained
byusingsiliconmicrostripdetectors.SinceCOMPASSuseshighintensitybeams
(∼108particle/s)thesilicondetectorsabsorbehighdoseand,hence,sufferra-
diationdamage.Tomaintainanoptimaldetectionefficiencyforlongperiodsof
datataking,detectorsmustbeoperatedatcryogenictemperatures,exploitingthe
[RD39].effectLazarusso-calledTheworkreportedinthisthesiscanbedividedintwoparts.Inthefirstpartthe
design,installationandoperationofasilicondetectorsystemoperatedatcryo-

2

withgenicemphasistemperaturonestheisprsearchesented.forΞIn−−thesecondpentaquarkpartthecandidate.dataTheanalysisfirstischapterdescribedof
thisthesisdescribesthephysicsgoalsandtheexperimentalsetupofCOMPASS.
AnwherintretheoductioneffecttooftheradiationphysicsofdamagesiliconandthedetectorsLazarisusgiveneffectinthearesecondoutlined.chapterThe,
thirdchapterisdedicatedtothespecificaccomplishmentofasiliconmicrostrip
rdetectorealisationinofaCOMPsiliconASS,withdetectoraspecialsystememphasisoperatingatonthecryogenicreadoutelectrtemperaturonics.esisrThee-
portedinthefourthchapter;theworkalsoincludedthedesign,development
andtestingofthecryogenicinfrastructures.Thefifthchapter,afteranoverview
ofthepresentstatus,bothfromatheoreticalandexperimentalpointsofview,de-
scribestheanalysisperformedwiththeCOMPASSdatatosearchfortheexotic
pentaquarkcandidateΞ−−.Theresultsarediscussedwithrespecttothegeneral
experimentalscenarioandfinallyconclusionsaregiven.

3

1Chapter

experimentASSCOMPThe

TheCOmmonMuonandProtonApparatusforStructureandSpectroscopy
(COMPASS(NA58))isafixedtargetexperimentattheCERN1Super-Proton-
Syncrotron(SPS).TheCOMPASSexperimentistheresultofthemergingoftwo
differentexperiments,namelyHMCandCHEOPSwhichhadsimilarexperimen-
talneeds,buthaddifferentphysicsobjectives.HMCwasdesignedtoinvestigate
thenucleonstructureusingamuonbeam;CHEOPShaditsmainfocusonhadron
spectroscopyusinghadronbeam.COMPASSisthusabletoaddressawiderange
ofphysicsgoalsachievablewithbothmuonandhadronprobes.Theexperimen-
talapparatuswasdesignedtobesuitableforbothconfigurations.
COMPASSbegantotakedatain2001and,sincethen,ithasbeenrunningwith
themuonbeam.ThefirstrunwithhadronbeamswilltakeplaceinSeptember
2004andwillcontinuein2006,aftertheSPSbreakplannedfor2005.
Thefirstpartofthischapterfocusesonthephysicsgoalsoftheexperiment:Sec.
1.1.1describesthegoalsofthemuonbeamprogram;Sec.1.1.2describesthe
topicstobeinvestigatedwithhadronbeams.Thesecondpartillustratesand
explainstheCOMPASSdetector,includingadescriptionofthecomponentwhich
havetobemodifiedinordertochangefromoneconfigurationtotheother.

aimsphysicsThe1.1

1.1.1Physicswiththemuonbeam

ThemaingoaloftheCOMPASSmuonprogramistounderstandthespincom-
positionofthenucleon[COM96].Thespinofthenucleonisthetotalspinand
1EuropeanCentreforNuclearResearch

5

Figure1.1:ThePhoton-GluonFusionprocess:thevirtualphotoninmuonscatter-
inginteractswiththegluonviaanintermediatequark.Theintermediatequark
providesthecolourandelectricchargeconservation.

angularmomentumofitsconstituents,thatis:

21=21ΔΣ+ΔG+Lzq+Lzg(1.1)
11wherquarkethespinleft-handcontribution,sideΔofGtheistheequationgluonisspinthespincontribution,2oftheLzqandnucleon,Lzg2arΔeΣtheisorthe-
bitalangularmomentumofquarksandgluons,respectively.Variousexperiments
([EMC88],[SMC99],[HER98])havemeasuredΔΣ=0.27±0.13.Theotherterms
arepresentlystillunknown.
WithSemiInclusiveDeepInelasticScattering(SIDIS)measurements,COMPASS
canprovideinformationonthespincontributionsofindividualquarkavours
(Δq)andonthegluonpolarisation(ΔG),aswellasonthetotalorbitalangular
momentum.Additionaltopicofinterestisthemeasurementofthetransverse
spincontributionwhichisaccessiblewithatransverselypolarisedtarget.All
thesemeasurementsarediscussedinthefollowing.

ΔG/G

ThemeasurementofthegluonpolarisationΔGiscrucialtounderstandthepuz-
zleofthenucleonspin.InCOMPASS,ΔGcanbeaccessedviaPhoton-Gluon
Fusion(PGF)process,whichisshowninFig.1.1.
Threedifferentchannelscanbestudied,namelytheopencharmchannel,thehigh
pTpairchannelandthehighpTsinglechannel.

6

Figure1.2:TheK−π+invariantmassspectrumwhereaclearpeakforD0isvisi-
ble.

OpenchamrDuetoitsgelarmass,theesenceprofthecquarkinthequarkseaquarksofainnucleonleadingisorofderlowatprCOMPobabilityASS.enerMorgieseovergoes,theprpredominantlyoductionofviaheavyPGF,
leadingtotwocharmtracks(c.f.Fig.1.1).Therefore,theobservationof
charmedparticlesisaagforthePGFprocess.Thelightestcharmedpar-
ticlesareD0mesons,whicharecleanlyidentifiedinCOMPASSinthede-
caychainD0→K−π+.Fig.1.2showstheK−π+invariantmassspectrum
0grwheroundeaisclearachievedpeakbyforrDequiringisvisible.thattheAD0substantialcomefrromDeduction∗+decayof,thethatback-is:

D∗+→D0π+→K−π+π+(1.2)
Itispossibletocorrelatethenumberofcharmeventstothegluonpolarisa-
as:[Mal96])(seetion

expNc↑c↓−Nc↑c↑ΔσγN→ccXΔσ(sˆ)⊗ΔG(η,sˆ)
A=N↑↓+N↑↑∝σγN→ccX=σ(sˆ)⊗G(η,sˆ)(1.3)
ccccwhereAexpisthespin-dependentasymmetry,Nccisthenumberofcharm
eventsfortargetspinparallelorantiparalleltothespinofthebeammuon,
ΔσγN→ccXandσγN→ccXarethehelicity-dependentandhelicity-averaged
crosssections,respectively;theselasttwotermscanbeexpressedasacon-
volutionofthephoton-gluoncrosssectionΔσ(sˆ)andσ(sˆ)withthegluon
distributionΔG(η,sˆ)andG(η,sˆ).sˆistheinvariantmasssquaredofthe
photon-gluonsystemandηisthegluonmomentumfraction.

7

•HighpTpairInprincipletheasymmetryAexpcanbemeasuredforquarks
ofanyavour.AsignatureofPGFistheproductionoftwojetsemittedin
oppositedirection(inthecentreofmasssystem).Thischannelhashigher
statisticscomparedtothecharmcase,butalsoahigherbackground.Byse-
lectingonlythehadronswithhightransversemomentum(pT>1GeV/c),
asensiblereductionofthebackgroundisachieved.HowevertheComp-
tonscatteringoffaquark(γq→gq)hasthesameorderinαslikethePGF:
thetwoprocessescannotbedistinguishedexperimentally,andMonteCarlo
simulationsareneededtoestimatethecontributionoftheComptonscatter-
ingtotheasymmetry.Fortheasymmetry:

∗AγLLd→hhX=−0.065±0.036(stat)±0.005(syst)(1.4)
wasmeasured[OC04].TheMonteCarlosimulationtoextractΔG/Gisstill
development.underThepresentstatusoftheΔG/Gmeasurementinboththeopencharmand
highpTchannelscanbefoundin[OC04].
•HighpTsingleparticleRecentlythepossibilityofmeasuringΔGviasingle
particlescreatedwithhighmomentumwasraised.Thismeasurementis
interesting,duetoitslowtheoreticalbackgroundandtothehighstatistics.
Onthecontrary,theexperimentalbackgroundisdifficulttoestimate.More
theoreticalstudiedareneededtoextractΔGfromthemeasuredasymmetry.
Thefeasibilityofthemeasurementisatpresentunderinvestigation[SHP].

Spinfunctionsdistribution

SIDISmeasurementscanaccessthespindistributionfunctions(Δq)ofdifferent
avours.Asymmetrieswithvariousleadinghadronsinthefinalstatearethe
meanstomeasurethefractionalcontributionofthedifferentquarkavoursto
thenucleonspin.TheresultsofthemeasurementoftheΛ0polarisation,which
allowtheextractionoftheΔsspindistribution,canbefoundin[Lam04].

ransversityT

Acompletedescriptionofthenucleonstructureatleadingorderisgivenbythe
threeindependentPartonDistributionFunctions(PDF)q(x),Δq(x)andΔTq(x).
Thelatter,calledthetransversespindistribution,describesthequarkdistribution
inatransverselypolarisednucleonasprobedbyalongitudinallypolarisedbeam
anditisΔTq=q↑→−q↓→.

8

Inanalogywithg1(x)onecandefinethestructurefunctionh1(x)as:
h1(x)=1åeq2[ΔTq(x)+ΔTq(x)].(1.5)
2q

hit1(xcan)becannotaccessedbebymeasurthoseedinprDISocessprwhichocessesshowduetotwoitshelicityoddips.chiralityOnenaturofe,thosebut
processisthepolarisedSemi-InclusiveDeepInelasticScattering(p-SIDIS)on
atransversallypolarisedtarget,whereh1(x)canbeextractedbytheazimuthal
dependenceoftheleadinghadron(Collinsangle).Thestruckquarkavourcan
beidentifiedbymeasuringtheinvolvedhadron.
ThepresentstatusoftheCOMPASStransversityanalysiscanbefoundin[h1].

GeneralisedPartonDistributions(GPD)

InrnucleonecentstrtimesucturetheasGPDtheypr[GPD02]ovideaareaprcompleteomisingdescriptiontoolforofthethenucleonunderstandingbyuni-of
fyingtheconceptsofpartondistributionsandhadronicformfactors.Exclusive
measurementssuchasHardExclusiveMesonProduction(HEMP)andDeeply
VirtualComptonScattering(DVCS)canexperimentallyaccesstheGPDs.
AnimportantremarkaboutDVCSmeasurementsisthattheycanaccesstheor-
bitalmomenta,giving,aspreviouslydescribed,acrucialcontributionintheun-
spin.nucleontheofderstandingTheHEMPandDVCSmeasurementsarenotpresentintheoriginalCOMPASS
proposalof1996.Nonetheless,feasibilitystudies[DVC]haveshownthatthe
COMPASSapparatusishighlysuitedtoperformthem,withadditionallyalarge
domainin(xBj,Q2).Preliminarymeasurementshavebeenalreadycarriedout,
andacompleteprogramwilltakeplaceduringthesecondphaseofCOMPASS
2005.after

0ΛransverseTpolarisation

verseAnotherΛ0topicpolarisation.whichItwaswasnotobservedincludedin[Les75]thethatCOMPΛ0sprASSproducedoposalinnonisthepolarisedtrans-
processeshaveapreferredspinorientation,i.e.apolarisation.Thepreferredaxis
isnormaltotheproductionplaneoftheΛ0.Thisphenomenonindicatestherole
ofparticlespinintheproductionmechanism.COMPASShasmeasuredthetrans-
verseΛ0polarisationinphoto-production;theresultsofthismeasurementcanbe
ie03].[Winfound

9

1.1.2Physicswiththehadronbeam

Thephysicsprogramtobecoveredbyusinghadronbeamscomprisesthreemain
aims:thestudyofdoublycharmedbaryons;thespectroscopyofexoticstatesand
glueballs;themeasurementofmesonpolarisabilitiesviaPrimakoffscattering.A
completedescriptionofthehadronphysicsprogramcanbefoundintheproposal
oftheCOMPASSexperiment[COM96].

baryonscharmedDoubly

Prcomposedesently,oflittleaisheavyknowcc-diquarkaboutdoublysurroundedcharmedbyalightbaryons.quarkTheywitharemassesmostinlikelythe
region3.6-3.8GeVasfirstobservedbytheSELEXexperiment[Moi02].Doubly
highcharmedintensitybaryonsbeamsareinexpectedordertotohaveachieveaveryhighlowcrossstatistics;forsection.thisrCOMPeasonitASSisusesex-
pectedtogiveasignificantcontributiontothespectroscopyofdoublycharmed
baryons,aswellastothecharacterisationoftheirdecaymechanism.

glueballsandstatesExotic

AnimportantpropertyoftheQuantumChromoDynamic(QCD)theoryisthat
thecharge.exchangeAsabosonsconsequence,ofthethecolourgluonsforcance,theinteractgluons,betweenhaveeachthemselvesotheraandcolourthere
isthennotheorglueballs,eticalpureimpedimentgluonicstates,forthemandtohybridsform,statesboundwithstates.bothOnequarkexpectsandtovalencefind
gluons.Thesestatesareexpectedtohaveashortlifetime.Thecharacteristicsthat
distinguishthemfromotherhadronicresonancesisthattheycanhavequantum
numberswhichareforbiddentoanyqqorqqqstate.Forthisreasontheyare
.exoticscalled

Also,anothersetofhadronicstatescanhaveexoticnumbers,namelythesocalled
TetraquarksandPentaquarks.Thesestatesarecomposedofmorethanthreeva-
lencequarksorantiquarks,howeverhavebaryonicnumberB=0or1,likethe
usualhadronicstates.Sincethestudyofpentaquarksisoneofthetopicofthis
thesis,amoredetaileddescriptionofthetheoreticalbackgroundwillbegivenin
chapters.subsequentsthe

10

scatteringfPrimakof

Thepionpolarisabilitiesareimportantobservablesfortheunderstandingof
hadronstructureinthelowenergyrangeofQCD.Thebestwaytoaccessthem
isviaγπComptonscattering.Thedifficultyofhavingapiontargetcanbeover-
comewithPrimakoffscattering.Insteadofaphotonbeamonapiontarget,the
used:isinteractionfollowing

π+Z→π+γ+Z(1.6)
TheincidentpionscattersonavirtualphotonoftheCoulombfieldofanucleus
withatomicnumberZ.Arealphotonisproducedandcanbedetected,together
withthescatteredpion,inthefinalstate.ThePrimakoffscatteringcanbeseenas
aComptonscatteringininversekinematics.
Thesemeasurementscanaccesstheelectricandmagneticpolarisabilitiesofthe
pionand,forthefirsttime,ofthekaon.AcompletereviewonthePrimakoff
measurementsinCOMPASScanbefoundin[pol02]and[Kuh01].

detectorASSCOMPThe1.2

Inordertopursueitsvariousphysicsgoals,theCOMPASSexperimenthasto
achievehighstatisticswithhighprecision.Therealisationofthesetworequire-
mentsischallengingforthedetectorsandtheDataAcQuisitionsystem(DAQ).
Asanexample,becauseofthehighintensitybeam(∼108particles/s)bothafast
readoutelectronicsandradiationharddetectorsarerequired.
Fig.1.3showsthesetupforthemuonrunasintheproposal.
Thebeamhitsasolidstatetarget.Thespectrometeriscomposedofdifferent
detectorsplacedalongthebeam,whichenablesthereconstructionofthetracks
andthemomentaoftheinteractionproductsandparticleidentification.
Thespectrometeris≈60mlonganddividedintotwostagespositionedbehind
eachother.Thefirststageisdesignedtodetectparticlesemittedatlowmomenta
(5-50GeV)andatlargeanglesanditisthereforecalledLargeAngleSpectrometer.
Detectorswithhighinteractionlength(calorimetersandmuonwall)haveahole
inthemiddletoallowthehighmomentumparticlestopassundisturbed.These
particlesaredetectedintheSmallAngleSpectrometerthatdetectsparticleswith
momentumfrom30to100GeV.
Inthesubsequentsectionsthedifferentcomponentscomprisingthespectrometer
described.ear

11

Figure1.3:TheCOMPASSsetupforthemuonrun.Thebeamgoesfromleftto
right.Thepositionofthetargetandofthedetectorsisshown.Theverticallines
detectors.trackingesenteprr

beamspolarisedThe1.2.1

COMPASSisinstalledintheexperimentalhallEHN2oftheCERNNorthArea.
Thehadronexperimentbeams.isThelineservedcanbyprtheovideM2alsobeamanlineelectrwhichoncanbeamprwhichovidecanbothbemuonusedandfor
purposes.test

Figure1.4:TheM2beamline

TheextractionlineisshowninFig.1.4.A400GeV/cprimaryprotonbeamis
intensityextractedonfromthethetargetSPSandvariesisdirbetweenected101towar2andds1the013prprimaryotonstarpergetSPST6.Thecycle.prFrotonom
theT6targetasecondarybeamisderived.Inthecaseofthemuonbeam,the

12

tertiarymuonsarisefrompionandkaondecays.ABerylliumabsorberstops
thehadronsinthebeam.Inthecaseahadronbeamiswished,theabsorber
aremovedandthesecondaryparticlesaredirectlytransportedtotheCOMPASS
target;inthiscaseCherenkovdetector(CEDAR)areinstalleinthebeamline
toperformparticleidentification.Thebeamisfocusedviaasetofdipoleand
quadrupolemagnets.Thepolarisationofthemuonisobtainedbyselectinga
certainenergyrangeviathebendingmagnets.Duetothespillstructureofthe
protonbeamtotheSPS,theuxofmuonsisnotcontinuousbutconcentratedin
4.8s(extractionattheSPS)followedby12nswhennobeamisdelivered,fora
totalof16.8scycle.Thecharacteristicsofthebeamforbothmuonandhadron
programsareshowninTab.1.1.

muonprogramhadronprogram
particlesµ+π,k,p
100-30060-160(GeV/c)gyenerintensity(particles/spill)2∙108108
beamsizeonthetarget(RMSincm)∼0.80.3−0.5
Table1.1:Characteristicsofthebeamforthemuonandthehadronprograms.

ForacompletedescriptionoftheM2beamlineandespeciallytheCOMPASS
[M2].seemodifications,dedicated

spectrometerThe1.2.2

Duewithtoonlytimeaandpartiallymanpowerequippedconstraints,spectrometerfor.theFig.time1.5being,showstheCOMPsetupASSisasratunning2003.
ThemaindifferenceistheabsenceoftheRICHII(RingImagingCHerenkov)
thedetectorfollowingandthethedifelectrferonicentpartscalorimeteroftheinCOMPtheASSsecondspectrstageometerofthewillspectrbeometerdiscussed..In
Theinformationconcerningthedetectorswereobtainedfrom[Wie03],[Fau04],
[Kuh02],[Leb02],[Ilg03],[Ehl02],[Tak03]and[COM].

targetsThe

COMPASSusesbothmuonandhadronbeamsthatcanaddressdifferentphysical
problems.Inordertopassfromoneprogramtotheother,thespectrometersetup
mustbeslightlymodified.Themostimportantdifferenceliesinthetarget.

13

eFigur

right.

1.5:

The

tectors.

The

wto

ASSCOMP

entferdif

ometerspectr

stages

ear

in

marked

14

2003:

the

togheter

beam

with

the

passes

omfr

position

of

left

the

to

de-

•Thetargetforthemuonprogramismadeoftwocylindricalrodsof6LiD
of1.5cmradiusand60cmlenghtseparatedby10cm.Thetwocellsare
polarisedviaDynamicNuclearPolarisation(DNP)atatemperatureof0.5
Kandinamagneticfieldof2.5T.Accordingtothephysicalproblemtobe
investigated,thecellscanbelongitudinallyortransverselypolarisedwith
respecttothebeamdirection.Thetwocellshaveoppositedirectionsofpo-
larisationwithrespecttoeachother,toavoidsystematicerrorsintheofine
reconstruction;forthesamereasonthepolarisationofthecellsisinverted
every8hours.Becauseoftechnicalproblems,theCOMPASSsolenoidmag-
nethasnotbeencompletedintime.ThisfactforcedtheSMCtargetmagnet,
whichhasaloweracceptancewithrespecttotheCOMPASSmagnet(±70
mradinsteadof±160mrad),tobereactivated.Apolarisationof55%was
achievedinthe2003datarun.
•Adifferenttargetwillbeusedinthehadronprogram:thePrimakoffand
charmprogramsrequireathin(2-3mm)solidhigh-Ztarget.Aprecision
vertexreconstructionisobtainedwith3ormoresiliconstationsinstalled
downstream.ForthePrimakoffprogramanadditionalvetobox(abarrel
ofscintillatorsplacedaroundthetarget)allowsunwantedevents,where
hadronicfragmentsareproduced,tobevetoed.Forthediffractiveandcen-
tralproductionprogramsaliquidhydrogentargetisused.Arecoildetector
madeoflayersofscintillatorswillbeinstalledaroundthetarget.Thedetec-
torisneededtoidentifytherecoilproton.

magnetsThe

COMPASSusesconventionaldipolemagnetstoreconstructtheparticlemo-
menta.Trackingdetectorsplacedupanddownstreamofthemagnets,permit
thereconstructionofthedeectedtracks.Byknowingthepropertiesofthemag-
neticfielditispossibletoextractthemomentumoftheparticles.Thefirstmagnet
SM1hasacentralgapof110×153×172cm3;forthehadronprogramtheheight
ofthegapwillbereducedfrom172a82cm.SM1hasaintegratedfieldon1Tm
at2500A.ThesecondmagnetSM2hasagapof400×200×100cm3andamax-
imumintegratedfieldof5.2Tm.Duringthe2003runitwasoperatedat4.4Tm
A.4000at

trackingThe

canCOMPbeASSdividedusesdifintoferthrenteetrackingclassesdetectorsdependingalongontheirtheentirsize:eVSAspectrT(Vometerery.STheymall

15

AreaTracker);SAT(SmallAreaTracker)andLAT(LargeAreaTracker).Abrief
reviewofthetrackersisgiven,whileasummaryofdetectorssizeandspatial
resolutionisgiveninTab.1.2.
•VSAT:Fortheregionupstreamofthetargetandfortheareainproximityof
thescatteredbeam,wheretheparticledensityishigh,detectorswithhigh
spatialresolutionandsmallsizeareused.Therearetwodifferenttypes
ofScintillatingFibresStations(SCIFI-JandSCIFI-G),thatadditionallyhave
excellenttimeresolution(≈400ps)andareusedtoassignthecorrecttime
totheevent.TheSilicondetectors(SI)areusedinthemuonsetuponlyup-
streamofthetargetforbeamreconstruction.Instead,forthehadronsetup,
moredetectorsareforeseendownstreamoftheinteractionpointtoimprove
thevertexreconstruction.Sincethesilicondetectorsformoneofthemain
topicsofthisthesis,theircharacteristicsandusageinCOMPASSaremore
extensivelydiscussedinthechapters2and3.
•SAT:TheSATshavealargeractiveareathantheVSAT.TheSATsareMi-
cromegas(MM)[MM]andGEM[Sau97]detectors:theyarebothgaseousde-
tectorswithinnovativesystemsforthechargeamplificationstage(ametal-
licmicromeshandaperforatedcopper-cladpolymerfoil,respectively).The
centralpartofthedetectors,wherethebeampassesthrough,isusuallydis-
activatedtoavoiddischargesinthegasduetothehighintensity.
•LAT:Themostouterareawithrespecttothebeamdirection,wherethein-
tensityislowandhighresolutionisnotrequired,iscoveredbytheLAT.
ConventionalDriftChambers(DCandW45),MultiWireProportionalCham-
bers(MWPC)andStrawchambersareused.
DetectorActivearea(mm2)Spatialresolution(µm)
SCIFI-J52.25×52.25120
SCIFI-G123×123410
SI50×707
MM380×38070
GEM300×30046
DC1400×1250240
W452400×50002000
MWPC1500×1200500
STRAW3250×2770250
Table1.2:Spatialresolutionforthevarioustrackingdetectors.
16

Usuallythreedetectors,onepertype,aremountedclosetoeachother,centred
alongthebeamdirection.Thisnestedconfigurationisparticularlyefficient:a
largeareaiscoveredtomaximisethetrackingefficiency;differentspatialreso-
lutionsinregionswithdifferentintensityminimisestheoccupancy,reducingthe
event.eachforratedata

(PID)identificationparticleThe

Inordertodistinguishbetweenpions,protonsandkaons,COMPASSusesRICH
detectors.ARICHdetectormeasuresthevelocityofparticlesviatheirCherenkov
emissionangleattheirpassagethoughthematerialradiator.Forthepartialsetup,
onlytheRICH-1detectorwasinstalled.Itspurposeistoseparateπ,pandK
withmomentaupto60GeV/c.ThephotonsaredetectedviaMWPCswithCsI
photocatodes.Theenergiesofallparticles,exceptthemuonsandneutrinos,aremeasuredby
thecalorimeters,wheretheyareabsorbedanddepositalltheirenergy.Duetothe
highdensityofthematerialinthecalorimeter,theparticlecreatesashowerthat
allowstoreconstructthepositionoftheincidentparticle.Calorimetersarealso
theonlydetectorsinCOMPASSwhicharesensitivetoneutralparticles.Each
stageisequippedwithelectronicandhadroniccalorimeters,installeddown-
streamoftheRICH.TheelectroniccalorimetersECAL1andECAL2aremade
outofleadglassblocksfromtheformerexperimentGAMS.In2003ECAL-2was
installed,andECAL-1willbeinstalledlaternextyear.Thehadroniccalorimeters
HCAL1andHCAL2haveasimilarstructure,consistingofsandwichofscintilla-
torsandironplates.TheinformationfromHCAL1andHCAL2arealsousedin
theformationofthetrigger.
Thehighpenetrationcapabilityofhighenergymuonsisusedtoidentifythem
inthemuonwalldetectorsMW1andMW2.Aparticleisidentifiedasamuonif
detectedinbothlayersoftrackingdetectors(Iaroccianddrifttubes,respectively)
upstreamanddownstreamablockofiron(≈1mthick).

riggerTThe

Thetriggerinitiatesthedataacquisition.Atriggerisidentifiedviathegeomet-
ricalpropertiesofthescatteringmuontrackandoftheenergydepositedbythe
producedhadronsinHCAL1andHCAL2.Themuontrackisreconstructedwith
dedicatedscintillatorhodoscopesplacedallalongtheexperiment.Adifferent
triggercalibrationallowsquasirealphotonevents(Q2<1GeV2)andinclu-
sivedeepinelasticscatteringevents(Q2>1GeV2)tobedistinguished.Forthe

17

Figure1.6:TheDAQarchitecture:thedatafromthedetectorsarefirstcollected
fromthedetectorfrontends,thentransmittedtotheDAQcomputers,wherethey
arecombinedinoneeventblockandtransferredtothecentraldatarecording.

hadronprogram,additionalinformationfromtheelectroniccalorimeterswillbe
.triggertheinused

(DAQ)systemAcQuisitionDetectorThe

TheCOMPASSdetectorcontainsapproximately190000detectorschannels.The
eventtriggersizerateisgoesonfraverageom450kHztokByte,thethermaximumeforethedesigndatafrrateequencygoesfrofom100200kHz.to4000The
SPSMByte/s.spillisInused.ordertoDuringcopethewith4.8sofsuchhighextraction,datatherate,datathearerpeculiareadoutstranducturbufeoffered;the
thecycledataforartheedatatransferredacquisition,intheobtainingfollowinga12s.smallerThispeakmethoddataowuses.Fig.the1.6wholeshowsSPS
e.chitecturarDAQthe

atTheeachsignalstriggerfr.omThetherdifeadoutferentmodulesdetectors(CAareTCHreadandoutfrGeSiCAom)thecollectdetectorthefrdatafrontendsom
ers,severalcalledfrRontendseadOandutBuftransferfer(ROBthem).viaTheopticalROBshostconnectionsPCIcartodsthecalledDAQspillbufferscomput-

18

wherethedataarebuffered.Fromthespillbuffersthedataaretransferredvia
variousGigabitEthernetdetectorstoaretheEventcombinedBuilderinoneeventcomputers,blockwherandethetransferredinformationstothefromCERNsthe
centraldatarecordingsystem.AcompleteviewoftheCOMPASSDAQcanbe
[Sch03].infoundTheDAQcomputersalsoperformtheonlinefiltering.Itconsistsofarejection
oftriggeredeventswhichdonothaveacertaintopology.Theinformationfrom
thedetectorsupstreamofthetargetareanalysed:ifadefinitenumberofplanes
intershowaestingsignaltopologieswithcohercanentbetimetaggedandevolution,eventsthewitheventaislargeaccepted.pile-upInofthistracksway
fromofftimeeventsorsecondaryinteractionscanbesuppressed,reducingthe
numberofeventstorecordtotape.Duringthe2003run,theonlinefilterwas
usedonlytotag,whilein2004itisactivelyusedtofiltertheevents.Acomplete
descriptionoftheonlinefilteranditsperformancescanbefoundin[Nag05].

(DCS)SystemControlDataThe

Duringoperation,theDCSmonitorsandcontrolstheCOMPASSsetupparame-
tersrameterssuchasliketheprhighessure,andlowtemperaturvoltagee,etc.systems,TheDCSgasusessupplies,acommerslowlycialvaryingSCADApa-2,
PVSS.called

1.2.3ProcessingofCOMPASSdata

Asillustratedintheprevioussection,thedatafromthedetectorsarecollectedvia
thetargetDAQ.areThethencollectedinformationsinonecorruniqueespondingeventtotheblocksameandprimarytransferredtointeractiontheincentralthe
datarecordingweretheyarestoredontapeonastoragefacilitycalledCASTOR
(CernAdvancedStorageManager)[CAS].
Atthislevel,aneventisacollectionofrawinformationsfromthedetectors
onandthehasadata,typicaltheysizehaveofto∼be40prKbytes.ocessedinBefororederatorphysicseconstranalysisuctthecanbetopologyperformedofthe
event(tracks,vertices,momenta,particleidentification)fromthedetectorsinfor-
mations.ThistaskisperformedbytheCOmpassReconstructionandAnaLysis
(CORAL)program.Afterthereconstructionandprocessing,thesizeofoneevent
becomesabout4Kbytes,thatmeansareductionto10%respecttotheoriginal
eventsize.ThestandardoutputofCORALisaDataSummaryTape(DST)file
2SupervisoryControlsandDataAcquisitionsystem

19

thethatpackagecontainsPHtheysicsfullyArnalysiseconstrSuctedoftwareTevents.ools(ThePHASTphysics).analysisisdonewith

CORAL

CORALisaC++basedsoftwarepackagethatreconstructstheeventtopology
fromthedetectorsrawdata.Thisprocesscanbedividedinfoursteps:decoding,
vertexing.tracking,clusterisation,

•Decoding:Auniquehardwar
ThisASS.COMPinchannel

eidentifierisassignedtoeachdetectors•channelinCOMPASS.Thisidentifierisrelatedtothefrontendmodules
usedforthereadoutofthechannel.Thecorrespondencebetweenidentifier
ofthechannelandthenameofthedetectorstowhomthechannelbelong
isstoredinthemappingfiles.Asfirststepofthedataprocessing,CORAL
extractsthehitchannelsfromtherawdataandlocatesthemwiththehelp
files.mappingtheof•Clusterisation:Whenaparticlecrossesadetector,iteventuallycreates
signalsonmanyneighbouringchannels.CORALcombinesthesesignals,
eventuallyamplitude-weighted,inacluster,whosecentrerepresentsthe
impactpointoftheparticle,andfindsitsspacialcoordinatesintheCOM-
PASSreferencesystem.
•Tracking:Atthispointthehitshavetobecombinedtoreconstructthe
tracks.Thisisdoneintwosteps:atfirsttheexperimentisdividedinfive
zonesinwhichthetracksareassumedtobestraight(nomagneticfields,
littlemultiplescattering)andstraightlinesarereconstructed;thenthetrack
segmentsfromtwoneighbouringzonesarefittogethertakinginaccount
thebendingduetomagneticfieldsorthedeviationduetothemultiplescat-
tering.•Vertexing:Aslaststep,CORALanalysesthetrackstofindiftheycan
originatefromacommonvertex.Thiswouldsuggestthattheparticles,
whichthetracksareassociatedwith,wereproducedinthesameinterac-
tion.Tracksareassumedtooriginatefromonevertexifthisisconsistent
withtheirgeometryandkinematics.Thevertexpositionistheaverageof
thepointsofclosestapproachofthetracks.

PHAST

COMPPHASTASS[PHA]physicsisasoftwaranalysis.epackagePHASTcanbasedroneadC++theandrawdataROOTaswelldevelopedastheforDSTthe

20

Figure1.7:Aschematicviewofthedataprocessingow.Foradetaileddescrip-
text.seetion

files.Ithasaspecificoutputformat,theminiDataSummaryTape(mDST).A
mDSTcontainstwodifferenttypesofinformations:thesetupinformations,like
magneticfieldmaps,materialmaps,detectorgeometricalinformations,storedin
thePaSetupobject;theeventinformations,likereconstructedparticlesandver-
tices,whichstormeansedinathefurtherPaEventreductionobject.rTheespecttypicaltothesizeDSTofaeventmDSTsize.eventPHASTis1.2addition-KBytes,
isallyrreducedejectstothe∼1%eventsinwithmDSTnonecomparreconstredtotheuctedrawvertex.data.Overall,thedatavolume

PHASTcontainsalsoasectionwithspecificfunctionsthatcanbemodifiedby
theuserinordertoreadthemDSTdataandaccomplishhis/herownphysical
goals.eventsThewiththeuserdesircanedaccessthecharacteristicsreconstranductedreconstreventsuctinthephysicalmDSTobservafiles,bleselectlike
invariantmasses,kinematicvariables,etc..TheoutputcanbeeitheraROOTfile,
toberbeeprfurtherocessedanalysedwithPHASTwithifthemorestandardinformationsROOTframe,aboutortheaeventsmDSTarfile,erthatequircaned.
Fig.1.7thedifferentstepsofthedataprocessing.

PHASTwasusedinthisworkfortheanalysisofstrangeparticlesdescribedin
5.Chap.

21

MontesimulationsCarlo

AbetterunderstandingofthedetectorperformancesisachievedwithMonte
Carlosimulation.ThedetectorsimulationisdoneinCOMpassGEANT
(COMGEANT).ItisaninterfacetotheGEANT3.21programwhichisacurrently
frozenversionoftheCERNsimulationpackageGEANT.Itwasdevelopedfor
theWA89experimentandhasbeenupgradedtobeusedalsoforCOMPASS.

Theeventsaregeneratedbyprogramscalledeventgenerators(PHYTIA,LEPTO,
etc.).COMGEANTcalculatestheoutputofthespectrometertoacertainevent.
TheCOMGEANToutputfilecanbeprocessedwithCORALandanalysedwith
PHASTlikeanormaldatafile.Theadvantageofsimulateddataisthatthey
containalsotheoriginalinformationsoftheevents,likewhichparticleswere
generated,theirmomenta,position,etc..MonteCarlodataareusedtoreconstruct
interestingquantities,likeefficiency,resolution,etc..Forexample,theratioofthe
numberofΞparticlesreconstructedintheanalysisandthenumberofΞinthe
eventgeneratorsfileistheefficiencyofthesystem(detector+analysis)forthe
Ξ.ThismethodwillbeusedinChap.5forthecalculationoftheΞ0∗andΞ−
.ficiencyef

22

2Chapter

ThedetectorsbasicandprinciplestheLazarusofsiliconeffect

clesSiliconinhighdetectorsenergybelongphysicstothemostexperiments.preciseTheytrackinghaveandetectorsexcellentforspacecharrgedesolutionparti-
(fewmicrometers)andneighbourtracksseparation,theirsignalcollectiontimeis
short(∼10ns)andforathicknessof300µmithasonly0.3%radiationlength,
thatisanextremelysmallvaluefortrackingdetectors.Theyhavealreadycon-
tributedsignificantlytothestudyoftheτlepton[PER75],ofcharmandbeauty
quarksand,lastbutnotleast,tothediscoveryofthetopquark[Erl95].

Theradiationcausesanappreciabledeteriorationoftheperformanceofsilicon
signaldetectors.toInnoisetheratioplanningafteronephaseyearofofCOMPdataASStakingawassignificantestimated,degradationwhichofisthenot
tolerablefortheenvisagedoperationlifetime.

Theradiationtoleranceofsilicondevicesandoftheirread-outelectronicsisthere-
foreofbiginterestandledtointenseeffortforthedevelopmentofradiationhard
techniques.Oneofthose,discoveredatCERNbytheRD39collaboration,con-
cernstherecoveryofirradiatedsilicondetectorswhenoperatedatlowtempera-
ture(around130K).ThiseffectwascalledLazaruseffectanditisexploitedinthe
experiment.ASSCOMP

Inthischapterthebasicoperationalprinciplesofsilicondetectorswillbeillus-
trated.Additionally,areviewofthemechanismofradiationdamageandofthe
Lazaruseffectwillbegiven.

23

detectorsmicrostripSilicon2.1

featuresBasic2.1.1

lossEnergy

Whenachargedparticlecrossesasemiconductormediumitloosespartofitsen-
ergy,mainlyviaelectromagneticinteractionswiththeatomicelectrons.Iftheen-
ergygainedintheinteractionissufficient,theelectronwillmovefromthevalence
bandtotheconductionbandleavingbehindahole,ormissingelectron.Theatom
inthelatticeisionisedandanelectron-holepairiscreated.TheenergyEneeded
tocreateanelectron-holepairinsiliconis3.6eVat300Kand3.8eVat77gK,where
athirdofitisusedbytheelectrontoovertaketheenergygapandtherestisab-
sorbedbylatticeexcitations.Thisthresholdvalueisreallylowifcomparedforex-
amplewiththeionisationenergyinagas(O(10eV)),makingthesiliconahighly
suitablematerialforparticledetection.Itisimportanttonoticethateventhough
thethermalenergyatroomtemperature(kT=8.617×10−5eV/K∙300K∼26
tomeVcr,eatewheremanykistheelectron-holeBoltzmannpairs.Aconstant)detailedisonly∼description7%ofEofg,theitisprhighocesscanenoughbe
.[Bic72]infoundperTheunitBethe-BlochlengthdE/dformulaxasaprovidesfunctiontheofeneraveragegyorenergymomentum:lossofthecrossingparticle

2222
−dE=2πNAre2mec2ρZz2ln2meγβ2cWmax−2β2−δ.(2.1)
βAxdI

whernumberer,eIisandthemearmeanetheexcitationclassicalelectrpotential,onZ,radiusAandandρarmass,etheNAisatomicAvogadrnumberos,
themassnumberandthedensityoftheabsorbingmaterialrespectively,zisthe
chargeoftheincidentparticleinunitsofe,βcitsvelocityandWmaxthemaximum
energytransferinasinglecollision.δistheso-calleddensityeffectcorrection
(seeforexample[Leo94]).Thisformulahasaminimumforparticleswithβ∼
0.96whicharethencalledMinimumIonisingParticles(MIP).Fromthispointon
thedE/dxchangeslowlywithincreasingtheenergy,thentheparticlesarestill
consideredMIPs.InFig.2.1theBethe-Blochformulaforprotons,pionsand
muonsindifferentmaterialsisdepicted.
Themeanenergydepositedaccordingtotheequation2.1coincideswiththemost
probabledepositedenergyonlyforthickdetectors.Forthindetectors(d10cm
fortransferincidentisstrpiononglyofr∼1educed20GeV)andthethepraverageobabilityenerofgylossinteractionsishigherwiththanhightheenermostgy

24

Figure2.1:Theenergylossindifferentmaterialsasfunctionofthemomentumof
theincidentparticle,describedbytheBethe-Blochformula

probableone.Inthiscasetheprobabilitydistributionoftheenergylossiswell
describedbyaLandaucurve,whichisasymmetricwithalongtailtowardsthe
highvalues,beinghoweverrareeventswithhighenergyloss.Fromthisdistri-
butiononefindsthatthemostprobableenergylossofaMIPinasilicondetector
of300µmthicknessis∼86keV,producing∼24000electron-holepairs.

junctionp-n

theyOncehavethecrtobeossingextractedparticleinhasordercrtoeatedadeterminecertainthenumberpassageofofaelectrparticle.on-holeApairs,solu-
tionWhenistothermalbringtoequilibriumcontacttwoisbulksestablished,ofp-aandn-certaindopednumbersiliconoffreechar(diode-principle).gecarri-
erscharfrgeomthecarriersp-andandan-rpotentialegionrdifferecombine.enceisThebuiltjunctionup(rbuilt-inegionisvoltage).thenItisdepletedpossi-of
blebiastoincrvoltage.easeThethesizedepthoftheofthedepletionspacercharegiongerisegiongivenby:applyinganexternalreverse

W=2(V+Vbi)/Neffe=2ρµ(V+Vbi)
25

(2.2)

whereWisthethicknessofthedepletedregion,Vtheexternalbiasvoltage,Vbi
isthebuilt-involtage,Nefftheeffectivedopingconcentration,etheelectronic
charge,thedielectricconstant,ρtheresistivityandµthechargecarriermobility
[PDG00a].Anasymmetricdopantconcentrationallowsthedepletionregiontobemoreex-
tendedononesideofthejunction.Atypicalconfigurationisasmall(fewµm)
highlydopedp-region,socalledp+,onan-dopedbulk(somehundredsofµm).
Usuallyn-dopedmaterialisusedasbulk,forthemobilityofitsfreechargecarri-
ers(electrons)ishigherthaninthepmaterial,allowingabiggerdepletionregion
withthesamebiaspotential(Eq.2.2).
Theconductivepandnregionstogetherwiththedepletedvolumeformacapac-
capacitance:thewithitor

whereAistheareaofthediode.

currentLeakage

C=A/W

(2.3)

Evenwhenthep-ndiodeisoperatedinreversebias,asmallleakagecurrentis
present,duetothethermalcreationofelectron-holepairs.Inordertoidentifythe
passageofaparticlethroughthedetector,thesignalgeneratedhastobehigher
thantheleakagecurrent,beingthelatterthenoiseofthedetector.Theleakage
currentIleakdependslinearlyonthewidthWofthedepletionregion:
Ileak∝W∝√V(2.4)
Equation2.4showsthatwhenthedetectorisfullydepletedtheleakagecurrent
saturates.

Changeofsiliconpropertieswiththetemperature

Reducingtheoperationtemperatureofasilicondetectorreducesthethermalen-
ergyoftheatomsinthelattice.Thisaffectsmainlytwoquantitieswhichare
significantforusingthesiliconasaparticledetector[Rug99]:

•leakageperature:curraentdecrIleakeasedueoftotheitsorigin,temperaturithaseisarstrongeectedinadependencereductionontheofthetem-
26

quencenumbertheofnoiseelectrofon-holethedetectorpairscrdecreatedeasesbywiththermalthetemperaturexcitation.e.Asconse-

mobilityoftheeefrgecharcarriersiteasesincrwitheasingdecre.temperaturThisfacthastwoimportantconsequences:asfirsttheresistivityofthesemi-
conductorincreases,permittingtofullydepletethedetectorwithasmaller
biasvoltagewithrespecttoroomtemperature(Eq.2.2);thenthesaturation
valueforthefreechargecarriersvelocityisreachedatalowervoltage.The
observableeffectisthatthesignalbecomesfasterandthechangeismore
holes.theforsignificant

Thismeansthatalsoanondamagedsilicondetectorcanbenefitfrombeingop-
e.temperaturlowaterated

detectionParticle2.1.2

operationofPrinciple

Asilicondetectorworkslikeap-ndiodeoperatedunderreversebias.Whena
particlecrossesthedepletedvolume,itionisestheatomsofthelattice,creating
electron-holepairs.Theseareseparatedbytheelectricfieldappliedanddrift
throughtheelectrodes,whereasignalisobserved.Thereisnochargemultiplica-
volume.silicontheinsidetionAsilicondetectorwidelyusedinhighenergyphysicsisthemicrostripdetector.
Itisconstitutedofan-typehighresistivitysubstrate(ρn>1KΩcm),onwhich
highlydopedp+stripsareimplanted,formingajunctionmatrix.Atypicaldis-
tancebetweenthestripsis∼50µm.Onthebacksidethereisanhomogeneous
implantationofhighlydopedn+silicon.Theelectronscreatedbythecrossing
particlearecollectedonthestripsclosetotheimpactpointonthedetector,pro-
vidingaonedimensionalinformationwithatypicalresolutionofsomeµm.The
detectoristhencoupledtotheread-outelectronicseitherdirectly(DCcoupling)
orwithasiliconoxidelayerinbetween(ACorcapacitivecoupling).
Itispossibletoachieveadoublesidedread-outsegmentingthen+sideaswell
asthep+one.Inthiscasetheholeswillbecollectedonthen+stripsarranged
orthogonallytothep+ones.Thedoublesideread-outprovidesaninformation
onthetwo-dimensionalpositionoftheimpactpoint,minimisingtheamountof
materialtheparticlehastocross.

27

resolutionSpatial

Asalreadymentionedthesilicondetectorscanhaveareallygoodspaceresolu-
tion.Thereasonsarethesmallsizeofthechargeclouds(aGaussiandistribution
withσ∼10µm),andtheprocessingtechnologywhichpermitstosegmentthe
readoutstripsdownto∼10µmpattern[Pei92].
Ifthechargecloudiscompletelycollectedononestrip,thecentreofthe√strip
givesthepositionmeasurement,whiletheresolutionisgivenbyσx=δx/12,
whereδxisthestrippitchandσxisthesecondmomentofaconstantdistribution
ofwidthδx.Ifinsteadthechargecloudisdistributedonmorethanonestrip,
thecentreofthecloudcanbeestimatedwithmuchhigherprecision,fitting,for
example,thechargedistributiononthestripwithaGaussianfunction.

resolutionimeT

Themobilityofchargecarriersinthesiliconis1350and480cm2/Vsforelectrons
andholes,respectively.Thismeansthatthecollectingtimeofthesignalisofthe
orderoffewns,foratypicalthicknessof300µm.Itispossibletodecreasethe
collectingtimeincreasingthebiasvoltageoverthefulldepletionvalue(overdeple-
tion).Thelimitofthisoperationisthediodebreakdown.

Thesignalshapeofthedetectorisaconvolutionofthecollectingtimeofthe
detectorandofthecharacteristictimeoftheread-outelectronics,beingusually
thelatterthedominantcontribution.Inthenextchaptersamethodforimproving
thetimeresolutionoftheCOMPASSsilicondetectorwillbedescribed.Forthe
momentletsjustmentionthatvaluesdownto4nscouldbeachieved.

ficiencyEfCollectionCharge

Thechargecollectionefficiency(CCE)istheratiobetweenthechargecreatedinthe
siliconbulkbyacrossingparticleandthechargecollected.Thisratiodepends
onthedepthofthedepletionregion,sinceinthenotdepletedregionthecharges
crfullyeatedbydepletedacrossingdetectortheparticleCCEcanisr∼100ecombine%.Itiswiththeworthwhilefreetochargementioncarriers.hereForthata
theexposureofdetectortoradiationreducesitsCCE,accordingtoamechanism
whichwillbedescribedinthefollowingsection.

28

Noise

Athecorrdetectorectitselfestimationandofthethenoisenoiseofintraoduceddetectorbymusttheread-outconsidertheelectrintrinsiconics.Thenoisefirstof
termisconstitutedbytheleakagecurrent,asexplainedinSec.2.1.1.Thesecond
termisrelatedtothecapacitanceofthedetectorasitisseenbythepreamplifier.
ThisquantitycanbeexpressedasEquivalentNoiseCharge(ENC)topermitan
easiercomparisonwiththesignal,thatitisusuallyexpressedintermofnumber
ofelectrons.TheENCfromthereadoutelectronicscanbewrittenas:

ENC=a+bC(2.5)
whereCistheloadcapacitanceandaandbareparametersofthepreamplifier.

damageRadiation2.2

Radiationdamageinsilicondetectorscanbedividedintosurfaceandbulkdam-
age.Inthefollowingonlythebulkdamagewillbediscussed,beingthemost
relevantcontributionforminimumionisingparticledetectors.

siliconinDefects2.2.1

mechanismDamage

Thebulkdamageproducedinsilicondetectorbyhighenergyparticlesiscaused
primarilybyelasticcollisionofthelatterwithlatticeatoms.Iftheenergytrans-
ferredtothesiliconatomishighenoughtodisplaceitoutofthelattice(primary
knockonatom,PKA)aFrenkelpairiscreated,thatisaninterstitialatomandaleft
overvacancy.TheenergyneededtocreateaFrenkelpairis∼25eV.Theenergy
oftherecoilatom˜canofcoursebemuchhigher(forexample,amuonof160GeV
cantransferupto2GeVtoaPKA)andthePKAcaneitherioniseotheratoms
orremovethemfromtheirpositioninthelattice,creatingacascadeofdislocated
atoms.Thenon-ionisingprocessesarepredominantwhentheatomloosesitslast
5-10keV.Then,attheendofeachheavilyrecoiledatompathaclusterofdefects
iscreated,asdisplayedinFig.2.2.
Thedisplacementofseveralsiliconatomsfromtheirpositioninthelatticeisnot
theendofthedamageprocess.Interstitialatomsandvacanciesareverymobile
inthesiliconlattice.Thereforeabout90%ofthevacanciesproducedannihilate
29

eFigur

ters

of

2.2:

Monte

defects.

Carlo

simulation

of

the

30

track

of

a

PKA

with

formation

of

clus-

thiswithprintocesserstisitialO(10siliconsminutes).andnoTherdamageemainingremains10%(seemigrateSec.thr2.2.3).oughThethetimelatticescaleandof
combineeitherwitheachotherorwiththeimpurityinthesiliconcreatingawide
varietyofdefects.Thesearecalledpointdefectsandtheyare,togetherwiththe
clusterdefects,theoriginofthebulkdamage.Theirelectricalpropertiesresults
inthemacroscopicdeteriorationofthedetectorproperties.Foramoredetailed
[Mol99].seedefectsofclassification

hypothesisscalingNIELThe

Inordertocomparethedefectscreatedinthesiliconbydifferentparticles,the
NonIonisingEnergyLoss(NIEL)hypothesiscanbeassumed.Itsbasicassumption
isthattheamountofdefectsinducedinthebulkisproportionaltotheamountof
energyreleasedbytheparticlevianonionisingcollisionswiththelatticeatoms.
Usuallytheuenceofneutronof1MeVischosenasunity.Toscaleacertain
uenceΦmeastoitsΦ1eqMeVonecanusetherelation:

Φ1eqMeV=kΦmeas
kiscalledhardnessfactoranditisdefinedas

(2.6)

Emaxk=1∙EminEdmaxEpφ(Ep)D(Ep)(2.7)
Dn(1MeV)EmindEpφ(Ep)
whereφ(Ep)istheenergyspectrumoftheparticlep,Episitskineticenergyand
D(Ep)isthedisplacementdamagecrosssectionwhombehaviourasfunctionofEis
showninFig.2.3.Forneutronsof1MeV,thedisplacementdamagecrosssection
is1andDn(1MeV)=95MeV∙mb.

2.2.2Effectsoftheradiationonthepropertiesofthedetector

Thefinalbulkdamageofasilicondetectorisproducedbythermallystablepoint
defectsandclusters.Duetothesedefects,newenergeticlevelsforelectronsand
holesarecreatedinsidetheforbiddengapbetweenthevalenceandtheconduc-
tionenergyband.Concerningthepropertiesofthedetector,thedefectsassoci-
atedtothedeeplevelsarethemostperniciousones,sincetheyhaveenergiesin
themiddleoftheforbiddengap.
ofMacrtheoscopicleakageefcurrfectsentofandtheofdefectstheefonfectivetheprdopingopertiesoftheconcentrationdetectorwitharethetheuence.change

31

Figure2.3:Displacementdamagecrosssectionfordifferentparticlesasfunction
gyenerof

leakagecurrentAsmentionedinSec.2.1.1,theleakageentcurristheori-ginofthenoiseofthedetector.Ifintermediateenergylevelsarecreated
intheforbiddengap,moreelectron-holepairsaregeneratedduetother-
malenergy.Itwasexperimentallyobservedthattheleakagecurrentinan
inverselybiaseddetectorincreaseswiththeuenceas

ΔI=I(φ)−I(φ=0)=αVφ(2.8)
whereαisthedamagefactorwhichis∼6∙10−17Acm−1forneutronsof1
MeVandVisthesensitivevolumeofthedetector.

•effectivedopingconcentrationTheeffectivedopingconcentrationNeffin
anonirradiatedsemiconductorisdefinedas

Neff=|ND−NA|(2.9)
whereNDisthedonorconcentrationandNAtheacceptorconcentrationin
thespacechargeregion.Thedefectscreatedbytheirradiationwithlevels
intheforbiddenenergygapcancaptureandemitelectronandholes,act-
ingasdopants.Theycanbeclassifiedasdonorsoracceptorsifneutralor
negative,respectivelywhenoccupiedwithanelectron.Thedefects,then,
changethedopingconcentration.Sincethenumberofdefectsisafunction
oftheuenceφ,inirradiateddetectorsNeffchangeswithφ.Forabetter

32

understandingofthephenomenonitisusefultoconsiderthecaseofade-
spacetectorcharwithgen−rtypeegionisbulk.Apositiven−typeandthesiliconfreebulkcharisgedopedcarrierswitharedonors;electrons.the
Whenirradiated,donorsandacceptorsdefectsarecreated.Thelevelsare
filledwiththefreeelectrons;donorsbecomeneutralandacceptorsbecome
removalnegative,prdecrocess,easingduetothethetotalcharpassificationgeoftheofbulk.donorOndopantstopofthis,whenthereactingdonor
withdefects,reducefurtherthetotalcharge.TheNeffdecreasesand,after
thedetectorhasbeenenoughexposed,then−bulkmaterialwillchangeto
p−type.Afterthetypeinversion,Neffandthedepletionvoltage(Eq.2.2)in-
creaseuntilthediodebreaksdown.SinceW∝1/√Neffandρ∝1/Neff(Eq.
2.2),thedepthofthespacechargeregionWandtheresistivityρdecrease
uence.theeasingincrwithThevariationofNeffwiththeφcanbeexpressedas

Neff=|NDe−cφ−NA−βφ|(2.10)
wherecandβparametrisethechangeofdonorsandacceptorsconcentra-
tion,respectively.Foran−typematerialwithstandarddopingconcentra-
tion,thetypeinversiontakesplaceforuences∼1014cm−2.

Fromthedetectionpointofview,anincreaseoftheleakagecurrentcorresponds
toanincreaseofthenoise,whichreducestheperformancesofthedetector.Fur-
thermore,increasingtheeffectivedopingconcentrationmeansanhigherbias
voltagehastobeappliedtothedetectorinordertoreachthefulldepletion.For
veryhighdamageddetector,thevoltageneededforacompletedepletioncanbe
higherthanthebreakdownvoltageofthep-njunction.Asdescribedinsection
2.1.2,ifadetectorisnotfullydepleteditschargecollectionefficiency(CCE)is
smallerthan100%.
Tosummarise,theradiationdamagesthelatticestructureofthesilicon,creating
intermediateenergylevelsbetweenthevalenceandconductivebands.Theselev-
elsmodifythemacroscopicpropertiesofthesilicon,degradingitsperformances
asparticledetector.Whentheamountofradiationthedetectorundergoesistoo
high,thedegradationissuchtomakethedetectornolongerusable.

2.2.3Annealing

aAssiliconmentioneddetectorintheprchangeseviouswithsection,theuence.theeffectiveAfterthedopingirradiationconcentrationtheNefNfeffstillin

33

changes,duetothereorganisationofthedefects.Thedefectsmigratethrough
thesiliconlatticeuntiltheyrecombinewiththeircounterpartofformnewdefects
byassosiationwithotherdefectsorimpurities.Thisphenomenoniscalledauto-
annealing.Rightaftertheirradiation,Neffdecreases(beneficialannealing)andafter
fewdaysNeffstartstoincreaseagain(inverseannealing).Theinverseannealing
canlastforyearsuntilNeffreachesasaturationvalue.Keepingthedetectorat
higherorlowertemperaturewithrespecttotheroomtemperaturecanspeedup
orslowdowntheinverseannealing,butitdoesnotchangethesaturationvalue.
Foradetailedreviewaboutsilicondetectorssee[Pei92]and[Mol99].

fectefLazarus2.3

Asdiscussedinthepreviouschapter,theradiationdamageposesanharshlimit
onthelifetimeofsilicondetectors.Differentwaysofimprovingtheovercom-
ingtheproblemareunderstudy.Betweenthem,itisworthwhiletomentionthe
implementationofmultipleguardringsstructuresonthesiliconwafer,which
ensuresbetterperformancesathighbiasvoltages[Abt96];and,asdefectengi-
neering,animprovedradiationhardnessofthesiliconbulkwhenenrichedwith
[ROSE].concentrationhighatoxygenAnalternativeapproachtotheproblemwasenvisagedbytheRD39collabora-
tiontheselatter[RD39w].wereInsteadelectricallyofavoidingpassifiedthebyrformationeducingtheofthermalradiationenerinducedgy.Theefdefects,fect
wasatemperaturrevivales.ofThishighlyphenomenondamagedwassiliconnameddetectorsLazaruswheneffectafteroperatedtheatbiblicalcryogenictale.
PrTheLazaroton-SynchrusefotrfectonwasandalrwilleadybeusedexploitedintheinTtheOTEMNA50experimentexperiment[T[NA5]OTEM]atatCERNthe
LargeHadronCollideratCERN.

evidencesExperimental2.3.1

ThefirstexperimentalevidenceoftheLazaruseffectwasobservedin1998
[Laz98].InFig.2.4thecorrespondingplotisshown:theCCErecoveryisclearly14
andobservedoperatedintheatthertemperatureverseebiasscan.voltageForVthe=100detectorV,thesignalirradiatedbyamplitudeφ=r1∙1eaches0
∼its90%atmaximum130Kat.TThem=r13ecovery0K,iswithalsoaCCEevincedwhichathigherpassesfruencesom∼and20dif%atfer1ent80Kbiasto
34

Figure2.4:Thevariationofthechargecollectionefficiency(inpercentage)as
functionoftemperaturefordetectorsirradiatedwithdifferentuences

voltages.ThevalueofTmdoesnotdependontheuence,whereasthemaximum
oftheCCEdecreaseswithincreasinguences.
Afterthisfirstevidence,furtherstudieswereperformedforafullcharacterisation
ofbythethetemperaturphenomenon.eatItwaswhichfoundthethatdetectortherwasecoverystoredoftheafterCCEtheisnotirradiation.inuencedAd-
ditionally,nosignificantdifferenceswerefoundwhenthesilicondetectorsare
irradiatedwhileoperatedatcryogenictemperatures[Cas01].
IncludingthebenefitsofalowtemperatureoperationdescribedinSec.2.1.1,one
canconcludethatalongerlifetimeforsilicondetectorsoperatedinhighradia-
tionenvironmentcanbeachieveddecreasingtheoperationtemperaturedown
to130K.Inthiswaycomplicationsconcerningwafergrowthanddesignare
avoided,butthechallengebecometheinfrastructuresneededforcryogenicop-
eration.

2.3.2TheLazaruseffectmodel

TherearetwopossiblereasonsforthedegradationoftheCCEinirradiatedde-
tectors:

gecharlossduetothetrappingofcarriersirradiation,asmentionedinSec.2.2.2.

35

bydeeplevelseatedcrduringthe

•partialdepletionofthedetector.

Contributionofchargeloss(CCEt)

Theintermediateenergylevelscreatedbytheradiationactastrappingcentres
fortheelectronsandholescreatedbyanionisingparticle.Theelectronsand
prholesocess.canIfbeτtrapisdetrapped:smallerletthanτtrapthebethecollectiontimetimeconstantoftheofcharthegetcoll,thetrapping-detrappingchargecan
bethetotalcollectedcharingecrspiteeatofedthebythetrapsandionisingthefinalparticleissignaliscollectednotafinafected.longerIfτtraptimeortcollis
partiallycollectioneflost,ficiencyandasafactorrconsequenceelatedtothetheCCEcarrieristrapping,deteriorated.CCEt,Theisrdefinedelativeascharge

CCEt=1−K[1−exp(−tcoll/τtrap)](2.11)
whereK=1/3forMIPs[Laz].ThetimedependenceofEq.2.11arisesonlyfrom
thedependenceofτtrapfromthetemperature,whereastcollisconstantduetothe
saturationofthedriftvelocity(Sec.2.1.1)

Contributionofthedepletiondepth(CCEg)

ForMIPdetection,theCCEtermoriginatingfromageometricalfactor(CCEg)is

2WCCEg=d2(2.12)
whereWisthespacechargeregiondepthanddthedetectorthickness.Wde-
pendsontheeffectivedopingconcentrationNeff(Eq.2.2),thenstudyingthede-
pendencyofNefffromthetemperatureispossibletoextrapolatethebehaviourof
.CCEg

CCEemperatureTVs.

ThetotalCCEcanbewrittenas

CCE=CCEt∙CCEg(2.13)
ItcanbedemonstratedthatCCEgismoresensitivetothetemperaturethanCCEt.
36

TheexperimentalresultsontheCCErecoverycanbethenexplainedasfollowing:
theture;efthefectiveratioWdoping/dthenincrconcentrationeasesNandeffdecrhenceeasesthetotalwhenCCEdecrincreasingeases.theWhentempera-the
ingdetectorfurtheristhefullydepletedtemperaturWe./dFor=T1<andTL,thewherCCEeTgLisstaystheconstanttemperaturwhenewherdecreeas-the
aturCCEehasintervalathemaximum,trappingtheCCEtconstantbecomesτtdecrtheneasestheandleadingcorrterm:espondininglthisythetemperCCE-
eases.decr

AmorecompleteexplanationoftheLazaruseffectisgivenin[Laz].

37

3Chapter

ThedetectorinimplementationCOMPASSofsilicon

Inthischapterthedifferentcomponentscomprisingthedetectorareexplained.
Atdepicted.firsttheThesiliconlastpartdetectorofthedesign,chapterthendescribesthetheelementsofcryostat,thertheead-outliquidnitrchainogenare
distributorandanoverallviewofthecompleteapparatus.

3.1ThesilicondetectorinCOMPASS

Untilthe2003run,silicondetectorswereusedonlyasbeamtrackingdevicesin
theregionupstreamofthetarget.Thiswasduebothtothefactthatinthemuon
programthespacebetweenthetargetmagnetandthefirstspectrometermagnet
wassmallthusleavinglittlespacefortrackingdevices,andthatsuchahighspace
resolutionasprovidedbysilicondetectorswasnotasmandatoryforthisphysics
programasforthehadronprogram.

Duringinstalledtheupstr2003eamrun,ofthrtheeetarsiliconget,inorstationsdertoandprtwoovidescintillatingexcellentspacefibreandstationstimerweres-e
olutionforthereconstructedbeamtracks.TheFig.3.1showsthearrangementof
differentstationsupstreamofthetarget.

Asrate(∼1mentioned08inparticles/s;chapterup1,toCOMP100ASSkHz),haswhichahighimposesbeamtheintensityusageandoffasthighrtriggereadout
electronics.Thisisespeciallytrueinthecaseofsilicondetectors,sinceeachde-
tectorisexposedtothefullintensitybeam.

39

Figure3.1:Detectorarrangementupstreamofthetarget:SFarescintillatingfibre
stations,SMCisthetargetmagnet,SIarethesiliconstations.SI01andSI02are
mountedonacommonsupport,whileSI03isonaseparatebench.

designdetectorThe3.2

ThesilicondetectorsusedinCOMPASSwereoriginallydesignedfortheHERA-B
experiment.TheyweredesignedbytheHalbleiter-Labor(HLL)ofMax-Planck-
Institutf¨urPhysik(M¨unchen).ThewafersusedinCOMPASSwerepartofa
batchproducedbySINTEF(Oslo,Norway)[Ber97].Fromthegeometricalpoint
ofview,theywereidenticaltotheoriginalHLLdesign,butpresentedsomeslight
technologicaldifferenceswithrespecttoit.ForinformationabouttheHERA-B
[Abt99].seeperformancesdetector

characteristicsProcessing3.2.1

Thesistivitydetectorswafers.areThepractiocessedvearoneaais2580×µ7mcm2.thick,Thenr-typeead-outsubstrate,pitchis2-354.k6Ωµmcmrande-
51.7µmforthen-andp-sidesrespectively.Onbothsides,intermediatestrips
arepresentinordertoimprovethespatialresolutionbychargedivisionreadout.
ThedetectorcrosssectionisshowninFig.3.2.
Thep-sideisdividedinto1024strips,inclinedat2.5◦withrespecttothewafer
edge;then-sidepresents1280strips,perpendiculartothep-sidestrips.Therea-
sonofthispeculiararrangementwillbeclarifiedinthenextsection.
nThe-typereadoutsiliconstripsbulk,onwhilethepthe-siden-sidearenaturally(ohmic-side)rinsulatedequiresfrommoreeachcomplicatedotherbyde-the
signsforelectricalisolationbetweenthestrips.InthecaseoftheHERA-Bsilicon
detectors,thep−stopstechnologywasused.Itconsists+ofhighlyp+dopedcom-
pensationimplantationsiliconarrangedbetweenthenreadoutstrips.

40

Figure3.2:Crosssectionofthesilicondetector.DetailsinSec.3.2.1

Geometry3.2.2

Sincemultiplescatteringcontributionsdominatethespatialresolutioninmi-
crostripsilicondetectors,areductionoftheamountofmaterialusedisconve-
nient.Asaconsequencedouble-sidedsilicondetectorsarepreferable,thatpro-
videinformationregardingtwospatialcoordinateswiththesameamountofma-
terialrequiredforasingle-sidedreadout.

Placingtwodouble-sideddetectorsclosetogether,allowsthemeasurementof4
independentprojectionsprovidingthestripsarenotparallel,sothatambiguities
inmulti-hiteventscanbedisentangledandghosthits(noiseabovethesignal
threshold)canberejected.Asmentionedinthepreviouschapter,thereisa2.5◦
stereoanglebetweenthestripsandthedetectoredge.Puttingtwodetectorsback-
to-back(aso-calledstation)willresult5◦anglebetweenstripsoncorresponding
sidesofthedetectors,asshowninFig.3.3.

Thisdesignenablesthelargestpossibleoverlapofthetwodetectorsand,atthe
sametime,theuseofthesamewaferdesignforbothdetectorsinonestation.
IntheCOMPASScoordinatessystem,onedetectorwillgivethexandyprojec-
tions(inclinedof0◦and90◦)andtheinclinedcoordinates(withanangleof−5◦
and85◦,respectively)arecalledtheuandvprojections.

hardnessRadiation3.2.3

Duetothehighuences(thedetectorhadtostandintheHERA-Benvironment
upto3∙1014/cm2peryear),adedicatedradiation-harddesignforthedetector

41

Figure3.3:Readoutstripsonthetwodouble-sideddetectorsmountedinone
stationtracking

wasdeveloped.Inthefollowingthedifferentcomponentsofthedesignareillus-
namely:trated,

•multiguard-ringstructure
•capacitivecouplingofthereadoutstrips
•polysiliconresistorsforstripbiasing

structureguard-ringMulti

Forafulldepletionofasiliconbulkafterradiationdamage,higherbiasvoltages
thannominal(∼90V)areexpected(upto300-500V).Inordertobeableto
operatedetectorsuptosuchconditions,thesensitiveareaofthedetectorissur-
roundedbya16punch-throughguard-ringstructureonbothsides.Thisstruc-
tureshieldsthesensitiveareafromsurfaceandedgeleakagecurrentsand,in
addition,providesacontrolledandgradualdropofpotentialfromthedetectors
rimtowardsthepotentialoftheundepletedsubstrate.
InFig.3.4acornerofthen-sidelayoutisshown.
Thedetectorwastestedforstableoperationatbiasvoltagewellabove500V
[Abt96].

Capacitivecoupling

trDueonicstothehadlartogebeleakagecapacitivelycurrentcoupledinducedtobytheradiationdetector.Fordamage,thisthercapacitiveeadoutelec-cou-

42

Figure3.4:Layoutofthen-sideofthesilicondetector(corner)

pling,thealuminiumreadoutlinesareseparatedfromtheimplantedstripsbya
220nmthicklayerofsilicondioxide(SiO2)and90nmofsiliconnitride(Si3N4),
3.2.Fig.inshownas

stripforresistorsPolysiliconbiasing

Usuallythebiasvoltagefordepletingthedetectorisprovidedbypunch-through
contactstothereadoutstrips.Forahighlyirradiateddetector,thistechniqueis
notsufficientastraps,whicharecreatedinthepunch-throughchannel,leadto
anincreaseofthenoise.Therefore,polysiliconresistorsof∼1MΩwereimple-
[Ber97].mented

electronicsThe3.3

Thesignalinducedinthesiliconbyacrossingparticle(Chap.2)hastobecol-
lected,acquisitionamplifiedsystem.andAdigitised,specialisedrbeforeeadoutbeingchaintransferr(Fig.ed3.12)tothehasbeenCOMPASSdevelopeddata
rforeadout.SiliconOpticaldetectors,dataemployingtransmissionisfeaturesused,likesincedatathertwoeductionsidesofonealgorithmsdetectorandarfaste
readoutwiththesamereadoutchain,althoughtheyareondifferentpotentials
duesystemtotheandabiasthorvoltageoughapplieddescriptiontotheofthedetectordata.rDetailseductiononthealgorithmsCOMPinASSthereadoutADC
aregivenin[Gru01].

43

APV25The3.3.1

Asread-outchiptheAPV25waschosen.ItisananaloguepipelineASIC1in-
tendedfortheread-outofsiliconmicrostripdetectorsinthetrackeroftheCMS
CERN.atexperimentDuetodifferentassemblytimes,twoversionsofthechip,namedS0andS1),were
usedintheCOMPASSsilicondetectors.Sincethereareonlysmalldifferences
betweenthetwoversions,inthefollowingonlythedescriptionofonewillbe
given,namelytheS1version,asitwastheversionusedwhenthedetectorswere
es.temperaturcryogenicatoperatedTheAPV25isa128channelanaloguepipelinewith192columnsofanaloguestor-
age.Siliconstripsignalsareamplifiedintoashapedpulsesofcharacteristictime
τCR−RC=50nsandmagnitude100mV/25,000electrons.Thesearesampledat
arateof40MHz(thatiseach25ns)andstoredintheringbuffer.Usefuldataare
markedafteraprogrammablelatency,andheldinthepipelinebufferuntilthey
canbereadout.A32deepFIFOholdstheaddressesofpipelinecolumnsholding
markeddata.WhendataisreadfromthepipelineitisprocessedusingaFIR2
filterbeforebeingreadoutthroughananaloguemultiplexer.Theoutputsignal
isadifferentialcurrent.InFig.3.5theschematicviewoftheAPV25isdepicted.

Figure3.5:LayoutoftheAPV25readoutchip

TheoutputfromtheAPV25chipisintherangeof±4mA.Whenthereisno
datatoreadout,theoutputisatthelogic0level,withsynchronisationpulses
12FAinitepplicationImpulseSRpecificesponseIntegratedCircuit

44

every70clock-cycles.Whenaneventistriggered,adatasetisthenoutputandis
composedoffourparts.Thefirstthreearedigital:aheader(threebitsatlogic1
level);anaddress(eightbits)definingthecolumnaddressusedtostorethesignal;
128andanchiperrorchannelsbit.rTheeadfourthoutatparttheisthememoryanaloguelocationdataset,indicatedthatbyisthetheaddroutputess.ofThethe
digitalanaloguesignal,dcvalueandthe(baseline)signalisgainprontopogrammableofthetobebaselinebetweenis1themA/mip.twolevelsofthe
TheFig.3.6showsatypicaloutputframe.

Figure3.6:OutputframeoftheAPV25operatinginmulti-mode

Thechiphasthreemodesofoperation:

•TheDeconvolutionmodeisusedinnormaloperationwhendataratesare
thesufFIRficientlyfilterhighremovessuchthattheefthefectseffectofsofpile-uppile-uptoareconfinesignificant.thesignalsInthistoonemode,25
intervalns

•Thesecondmode,Peak,isusedwhenpile-upisnotsignificantandalarger
signaltonoiseratioisrequired.InthismodetheFIRfilteractssimplyasan
amplifier

•Thethirdmode,Multi,canbeusedwhencalibratingapulseshape,andin
thismodethreemultipleconsecutivepipelinecolumnscanbetriggeredand
readoutwithoutFIRfiltering.

45

FortheCOMPASSsiliconreadout,theMultimodewaschosen,inordertoim-
provethetimeresolution.Foramoredetaileddescriptionofthemethodusedto
extractthetimeresolutionwiththeCOMPASSsilicondetectorsee[Wie03].

L-boardThe3.3.2

3TheThewafersupportingisgluedstructurbetweeneofthethemwithdetectoraiscomposedmono-componentoftwosilicon6-layerbasedPCBsnon-.
conductiveglue,inordertohaveaccesstothereadoutpadsonbothsides.The
boardsalsoprovidepadsforthebiasvoltageofthewafer.Thebiasvoltageon
eachsideofthedetector,isfixedtothereferencevoltageonthecorresponding
L-boards.Thatmeansthetworeadoutsidesareondifferentgroundpotentials.
theForrwafereducingonthetheL-boarpickupds.noiseTherthereadouteadoutpadschipsonthearesiliconplacedarasecloseconnectedasviapossiblewirtoe
onbondingthetodetectortheandinputchips,padsaonpitchthechips.adaptorAsthe(aluminiumpatternsforwirestheonpadsa30ar0eµdifmferthickent
glasssupport)isplacedbetweenthem,toallowstraightwirebonding.
boarDuedtoonthethedifnfer-sideenthostsnumber10ofAPV25,channelswhileonthetheonetwoonsidesthepof-sidethehostsdetector8,(seetheFig.L-
3.7).

Figure3.7:LayoutofthetwoL-boards

Asmentionedintheprevioussection,theoutputoftheAPV25isadifferential
3PrintedCircuitBoard

46

current.OntheL-boards,thedatafromeachchiparereadoutthrougha4.7kΩ
resistor,inordertoconvertthecurrentoutputinavoltageoutput.
TheAPV25chipsareverysensitivetonoiseonthepowerandgroundinglines,
leadingtooscillationsofthebaselineintheanalogueoutput.Inaddition,due
tothedouble-sidedoperation,thereisacross-talkbetweenthereadoutsonthe
twoprojectionsviathesiliconwafer.Thiscouplingallowsthetransmissionof
thenoisethroughthetwosidesofthereadout.Toachieveastableoperation
atlownoise,thepowerlineswereequippedwithcapacitancesdirectlyunder-
neaththechipsandacrosstheL-boardsforstabilisationandfrequencyfiltering.
Rightbelowthepadsforthechips,atcableconnectorsprovidedatatransmis-
sionandpowerlines,aconfigurationbusandclockandtriggerinformation.
TheL-boardsalsoprovidethecoolingfacilitiesneededforcryogenicoperations:
Copper-Nickelcapillaries(douter=1.6mm;dinner=1.3mm)aresolderedon
dedicatedpadsonthebacksideofthePCBsandareushedwithliquidnitrogen
(LN2),whichisusedascoolant.Thepadsareconnectedviametalizedholesto
thepadswherethewaferisglued.Toassurebettercryogenicoperations,thecop-
perlayersinsidethePCBsarearrangedinsuchwaythatthewaferisthermally
decoupledfromthereadoutchipsastheyareanheatsource.EachPCBalsohosts
thereadoutfortwoPT100thermometersfortemperaturecontrolonthedetector.

cardrepeaterThe3.3.3

Afirstsignalpreamplificationisdoneintherepeatercards.Onerepeaterboard
foreachsideisused.Apartfromamplification,theboardsprovidepowerforthe
aturAPVechips,monitoringdistributionontheofL-boarclockds.andInCOMPtriggerASS,lines,powerandansupplyinterfacefortheforrtemperepeater-
boardsisprovidedbyoatingC.A.E.N.lowvoltagepowermodules,whichallow
monitoringmonitoring,ofthesethecurrcurrentsentsprthatovedthetobeAPVsanconsume.importantTtoologetherforwithcheckingthetemperaturchipse
statusandforremotedebuggingpurposes.

cardADCThe3.3.4

FordigitisationoftheAPVsignals,anADCcardforSiliconreadoutwasdevel-
oped(sgadc).Onecardprocessesinformationreceivedfromtherepeaterboard
fromonesilicondetectorside.Digitisationisperformedwith10bitprecision.
TheAPVheaderinformationismergedwithadditionaleventinformationinto
anewheaderanddataissentoutviaopticalfibres.TheimplementedHotlink
protocolpermitstransferratesofupto40MByte/sonopticalfibres.

47

Modesofoperation.Thesgadcprovidestwomodesofoperation.Inlatch-
allmode,theamplitudesofallchannelsandallthreesamplesaretransmittedto
thereadoutchain.Anamountof9kBdatapereventanddetectorisproduced,
limitingreadouttoarelativelylowtriggerrateofabout1kHz.Toincreaseperfor-
mance,dedicateddatareductionmethodshavebeenimplementedforthesgadc.

Zerosuppression&commonmodecorrection.Insparsemode,aXilinxFPGA4
chiponthecardperformspedestalsubtractionforeachchannel.5Inaddition,
acorrectionforcollectivechangesofallpedestals—thecommonmode—duetoa
baselineshiftofindividualAPVchipsisdone.Onlydatafromchannelswhichex-
ceedagiventhresholdaretransmitted.Sinceahighfractionofchannelscontain
low-amplitudenoise,thedatarateisreducedconsiderablybyzerosuppression.
ThresholdsandpedestalsareloadedintotheADCviasoftwareusingtheI2C
protocolandareobtainedfromdedicatedpedestalrunsinlatch-allmode.For
thefuture,mechanismstoobtainpedestalsduringtheexecutionofdatarunsare
ed.consider

GeSiCAThe3.3.5GeSiCA6isa9UVMEcontrolmodule,whichmanagesthecommunicationbe-
tweenADCsandthedataacquisitionsystem.Itsequentiallymultiplexesdata
ofuptofourADCsconnectedviaHotlinkandsendsthemviaa160MB/sopti-
calSlinkconnection[Bij]totheCOMPASSdataacquisition.Moreover,GeSiCA
receivestheexperiment-wideclockandtriggerthroughaTCS7receiveranddis-
tributesthemtotheADC(see[Gru01]).Forconfiguration,theAPVsI2Cbuscan
beaccessedandconfigurationregistersofGeSiCAcanbesetviatheVMEbus
andaVMEfront-endmachinerunningLinux.

cryostatThe3.4

Asmentionedinsection3.2,twosilicondetectorsaremountedbacktobackto
formastation.Asafirststep,twodetectorsaremountedonaglassfibrestar-
shapedstructure,calledshuricane.Theshuricaneisthenmountedinaholding
structurecalledcryostat.Thestructurehastohavethefollowingrequisites:
4Field-programmablegatearray
5Inanoisespectrum,thebaselinediffersslightlyfromchanneltochannel.Thezerovaluesof
theindividualchannelsarecalledpedestals.
6GEMandSiliconcontrolandacquisition
7Triggerandcontrolsystem

48

•lighttightness:silicondetectorscandetectvisiblephotons.Inanormally
lightenedenvironment,thenumberofthesephotonsismuchhigherthan
thenumberofdetectedbeammuons.Foravoidingtodetect,then,visible
photonsthecryostathastobelighttight.
•vacuumtightness:cryogenicoperationofthedetectorsimplytheiroptimal
thermalinsulationwhichrespecttotheouterenvironment.Inordertomin-
imisetheamountofliquidnitrogenusedforcooling.Theabsenceofwater
isalsoanissue,toavoidanyformationoficeonthewafer,whichwould
leadtoanincreaseofmaterialalongthebeampath.Boththeserequire-
mentsarefulfilledwhileoperatingthedetectorsinvacuum.
•lowmass:tominimisethemultiplescatteringcontributiontothespatial
resolution,theradiationlengthofthedetectorsandtheirhostingstructure
hastobeassmallaspossibleinthebeamthroughdirection.
•Faradaycage:ametallicholdingstructurecanhelptoreducethenoisein-
ducedbytheelectromagneticradiationintheexperimentalenvironment,
apparata.otherbygenerated

Fig.3.8isaschematicviewofthededicatedcryostat.Itsmainpartisastainless
steelframewheretheshuricaneispositionedwithaprecisionof∼100µm.The
cryostathosts3KFangesonthebottom,whereavacuumpump,avacuum
gaugeandaspillvalveforvacuumcompensationareinstalled.Thetopand
thetwosideshost6anges,twoangeseach,usedforcryogenicandelectric
connections.Ontheframe,perpendiculartothebeamdirections,twoplatesclosethecryostat.
Inordertominimisethematerialinthisdirection,a10×7cm2windowwas
openedoneachplate.Thewindowiscoveredwitha100µmaluminizedMylar
foillyingonanetofKearwirestoavoidtheimplosionofthefoilduringthe
pumping.Thealuminizationguaranteesbothlighttightnessandelectromagnetic
shield.Detailsontheplatestructurecanbefoundin[Fuc04].
Thecryostatisoperatedwithaninnerpressure∼10−6mbar.Theelectricalcon-
nectionsbetweentheL-boardandrepeatercardsaredoneviaalong,curvedPCB,
whichisgluedintheangewithtwo-componentepoxyglue,halfremaninigin
thecryostatandhalfoutside.Oneofthetopsideangeshostsaconnectorforthe
liquidnitrogeninlet:atthebottomoftheconnectortheLN2uxissplitandthe
coolingofthetwodetectorsisdoneinparallel.ThetwoLN2outletarelocated
onthesides.Inordertoavoidanyelectricalcoupling,thecapillariesaresolded
ontothedetectorstotheinletandoutletsviaceramicconnectors.Fig.3.9shows
amoduleinstalledinthecryostat.

49

eFigur

side

3.8:

view;

Schematic

c)

ontfr

drawing

view

with

eFigur

3.9:

of

the

window

A

cryostat:

installed;

silicon

module

50

onta)fr

d)

in

top

the

view

.view

of

cryostat

open

cryostat;

b)

Figure3.10:Thegroundingschemeforthesilicondetector.Siisthedetector,C
isthecryostat,Rtherepeatercards,AtheADCs,PSthepowersupplies,ΔVthe
appliedbiasvoltagebetweenthetwosidesofthedetector.

schemegroundingThe3.5

Thebiasvoltageisappliedbetweenthep-andn-sideofthesilicondetector.Be-
causeofthecapacitivecouplingbetweenreadoutstripsandsiliconimplants(see
Sec.3.2.3),thereadoutelectronicscanbeinprincipleonanypotential.Any-
sidehow,wertheeputL-boaratd,thethersameepeaterpotentialcardofandthetheADCimplant,rineadingorderouttotheavoidsameprdetectoroblems
in3.10.caseTheofshortsiliconcutdetectorofa(stripSi)isreprcapacitor.esentedTheasgradiodoundingeandΔschemeVisistheshownappliedinbiasFig.
thevoltage.onesforThethepowerADCsarsupplieseofconnectedthertoepeaterthegrcardsoundareviaa1intrinsically0MΩroatesistoring,.whileSince
outtheADCsGeSiCAiswhicharopticallyeatadifconnectedferentrtoefertheenceADCs,voltage.itisThepdecoupled-sideofandtheitcandetectorreadis
alsoconnectedtothecryostat,inordertodefineareferenceground.Respectto
it,then-sideisathighpotential.Sincethedifferentcomponentsofthereadout
chainandtheirconnectingcablescannotbegrounded,theyactlikeantennasand
onpicktheupsignal,electraonic10nFnoiseoncapacitortheandpotentiala200lines.µFToonewerminimiseetheinstalledeffectacrofossthisthenoisebias
voltagetofilterthehighfrequencynoise.

51

3.6

Figure3.11:Viewofthetargetregion

viewoverallAn

Fig.3.11isapictureofthesiliconstationsasinstalledduringthe2003run.The
cryostatsaremountedonanopticalbench.Thebenchhasaprecisionmechanical
structurethatpermitsittobealignedwithsomereferencedirectiontoaprecision
of∼100µm.
ThecompletereadoutchainisdepictedinFig.3.12.

Achip.particleThecrsignalossingisthefirstdetectoramplifiedcreates(raepeatersignalcard),whichtheniscollectedtransferedontothethereadoutADC

Figure3.12:Siliconreadoutchain(nottoscale)

52

them

to

the

general

the

educingr

ession,suppr

ozer

performs

also

ADC

The

digitised.

is

it

ewher

the

collects

transferredtotheGeSiCA.TheGeSiCA

earwhichdata

amountof

ASSCOMP

data

acquisition

transfers

and

merges

dataoffourADC(onecompletestation),

system.

53

4Chapter

theofinstallationandPreparationinstationsiliconcryogenicfirstASSCOMP

Asmentionedinchapter2,therecoveryofthechargecollectionefficiency(CCE)
ofdamagedsilicondetectorsoperatedatcryogenictemperatureshasamaximum
for∼130K,thesocalledLazarustemperature.Inordertofulfilthistemperature
requirement,thesupportingboardsforthedetectorswereequippedwithtiny
capillaries(innerdiameter1.3cm;outerdiameter1.6cm)whereacontinuous
two-phaseowofnitrogenwasachieved,inanattempttoensureanoptimal
thermalcontactbetweenthemandthedetector(seeChap.3).Theheattransfer
proceedsfromthedetectortothecapillarywhereliquidnitrogenisushed.Since
theboilingpointofnitrogenisat77K(atatmosphericpressure),thecapillaries
willcontainamixtureofliquidandgas(2-phaseux),thesmallertheamountof
gasthehigherthecoolingcapabilityofthemixture.

Avariationofnitrogenuxresultsinachangeofthetemperatureofthedetector.

Inthischapterallthestepswhichbroughtthesiliconstationatcryogenictem-
peratureintooperationarelisted.Firstthelaboratorysetup,thenthetestsofthe
differentcomponents,andfinallytheinstallationofthesiliconstationintheex-
perimentalareaaredescribed.Thefinalsectionisdedicatedtothedescriptionof
adistributionboxforliquidnitrogenthatwillbeusedduringtherunsinthenext
years.

55

4.1

Figure4.1:SchematicviewofLN2uxforthelaboratorysetup

setupLaboratory

Withinthescopeofthisthesis,alaboratorysetupwasrealisedfrom2000to2002.
Liquidnitrogenwassuppliedbyaself-pressuriseddewar.Thedewarhasaca-
pacityof120l.Aninternalpressurebuildingcircuitpermitapre-setpressureto
bemaintainedwhichcanbesetbytheuser.Nitrogenistransferedtothecryostat
viaasuper-insulatedtransferline.Thetransferlinehasadoubleaxialstructure:
theinnerpipe,whereliquidnitrogenisushing,issurroundedbyasecondcham-
ber.Theyareconnectedinawaysuchthatasmallamountofnitrogenisushing
backintothesecondchamber;thethirdchamberisevacuatedtominimisethe
thermalexchangewiththeexterior.Thispeculiararrangementwasdesignedto
minimisethelossinliquidheliumtransfer;incaseofliquidnitrogenthetransfer
occurswithpracticallynolosses.Foramoredetaileddescriptionofthedewar
andthetransferlinesee[wes].

Inthefollowingmeasurements,manualowmeterswereusedtomeasureand
controltheliquidnitrogenuxdownstreamofthedetector.Thegaswasguided
totheowmeterswithaPVCtube.Inordertoavoidanydis-functionoftheow
metersduetofrost,theywereplaced∼2mawayfromtheoutletofthecryostat,
allowingforthegastowarmupbeforepassingthroughthedevice.

InFig.4.1,aschematicviewoftheliquidnitrogenpathisdepicted.

56

4.2CharacterisationoftheAPV25chipatcryogenic
temperatures

Inordertooptimisetheperformancesofthesilicondetectorreadout,thechiphas
tobemountedataminimumdistancefromthesiliconmicrostrips.Therefore,in
ordertooperatethesilicondetectoratcryogenictemperatures,itismandatory
thatthereadoutelectronicsshouldworkinthesamerangeoftemperatures.The
firststepwasthentocheckthefunctionalityofthereadoutchipatcryogenic
es.temperaturInthissection,resultsoftestsperformedontheAPV25-S0atcryogenictempera-
turesarepresented:inparticular,thefeaturesofdifferentsamplesintherangeof
temperaturebetween77Kandroomtemperature;thechipfeaturesstabilitywith
temperatureoscillationsintheLazarusregion(130±20K)andtheresistance
ofthechiptothermalcycles.TheS0versionoftheAPV25chipwasused.

measurementCold4.2.1setup

Inordertoperformmeasurements,theAPV25-S0hasbeenmountedonadedi-
cated4-layerprintedcircuitboardPCBasshowninFig.4.2.

Figure4.2:4-layerPCBusedforthecoldmeasurements

ThePCBcanhost,initslowerside,uptotwosiliconsensors,oneperside,with
theirreadoutandthetemperaturemonitoringsystem.Thedesignhasbeenmade
withtheperspectiveoftestingsilicontrackingdetectorsandtheirelectronicread-
outatcryogenictemperature.However,inthefollowingmeasurements,onlyone
sideoftheboardwasusedtocharacterisetheAPV25-S0chip.Theboardispar-
tiallyinsertedintoadedicatedcryostatthatworksinvacuumconditions.The
designofthePCBhasbeenoptimisedinordertoachievetwomainrequisites:

57

•homogeneouscoolingintheregionofthedetectorandelectronicreadout
•thermaldecouplingbetweenthebottominthecryostatsvacuumandthe
topatatmospherepressure.

Inparticular,thesecondrequisiteisneededtoavoidthatatthetopofthePCB,
whermationeallonthetheelectricalcryogenicconnectionscharacteristicareofsuchplaced,adesignwatercanbecondenses.foundinFurther[gra01]inforor-
[RD39w].inInmm,orderexternaltocooldiameterdownthe1.6PCB,mm)awasused,nickel-coppertinmicrsolderedo-pipeonthe(innerdedicateddiameterpad1.3
inwhichauxofliquidnitrogenwasdirected.Thetemperaturecanbetuned
thuscontrollingtheuxrateofliquidnitrogen.Thechiphasbeengluedwith
2-componentsconductiveglue,toensuretheelectricalconnectiontothebiaspad
onthebottom.TosimulateaMinimumIonisingParticle(MIP)injectiononone
chipchannel,aRCnetworkconnectedtoanexternaltunablepowersupplywas
used.Thisarrangementpermitssimulationfrom0.5uptoseveralMIPs,insteps
of0.5.Moreover,fiveotherchannelsoftheAPV25-S0wereconnectedtothe
groundviaaseriesofdifferentcapacitancesand50Ωresistorstomeasurethe
noiseasfunctionofcapacityatdifferenttemperatures.
Tomeasurethechippropertiesasafunctionoftemperature,avacuumtightcryo-
stathasbeenused.ThededicatedmeasurementsetupisshowninFig.4.3.From
thefigure)figurtoeithostisinstrpossibletoumentationdistinguishandfivetwo(onsetstheofbottom)anges:tofourhost(onPCBs.thetopTheofthePCB
usedforthesemeasurementsispluggedintothesecondangefromleft.APVC
transparenttubeemergesfromthePCB,thisistheliquidnitrogenoutletcon-
nectedtoaowmeterthatmeasuresandtunestheuxrate.Ontheotheranges
fromtheleftthereisaventilationgauge,apressuregauge,aliquidnitrogeninlet
lineandthevacuumturbopumpgauge.Theworkingpressureinsidethecryo-
stathasbeenmeasuredas1×10−6mbar.Themaximumliquidnitrogenux
allowedbytheowmeterwas250N2gas/hour.Thetemperaturewasmoni-
toredviaPlatinumthermometersPT100connectedtoaKeithleymultimeter.The
wholesystemallowstemperaturecontrolwithin1K.

characterisationAPV25-S04.2.2

Fig.4.4showstheoutputoftheAPV25wherea1-MIPequivalentsignalisin-
jectedintooneinputchannel.Afullcharacterisationofthechipcanbefoundin
[Ray01].Theamplitudeoftheconnectedchanneliscorrelatedwiththenumber

58

eFigur

gen,

c)

4.3:

Experimental

ogennitr

outlet,

eFigur

set-up:

d)ange

4.4:

a)

to

oughbeam-thr

host

APV25-S0

the

chip

output

59

ange,

d.boar

data

frame

b)

at

inlet

130

for

K

liquid

o-nitr

ofinjectedelectrons.Measurementswereperformedtodeterminetheanalogue
pulseshapeandlinearity,noiseasfunctionofinputcapacitance,temperature
behaviouranduniformityofpipeline.Itisworthpointingoutthatallmeasure-
mentsatlowtemperaturehavebeenperformedwithoutchangingtheparame-
terssettingoftheAPV25-S0withtherespecttoroomtemperature,inorderto
verifytheeffectsofthetemperatureontheCMOScircuit.

linearityandshapepulseAnalogue

Fig.4.5showstheamplifierpulseshapemeasuredatroomtemperatureand130
K,inpeak(broadcurve)anddeconvolution(narrowcurve)modes(Sec.3.3.1).
ApplyingaprogressivedelayΔTbetweenthetimeofthechargeinjectionandof
thereadout(depictedonthex-axis)theshapeoftheAPV25outputsignalwas
ucted.econstrr

Figure4.5:APV25-S0pulseshapedependenceonsignalamplitudeinpeakand
deconvolutionmodeat130K(dashedblackcurves)andatroomtemperature(con-
tinousgreencurves)for1MinimumIonisingParticle(MIP)equivalentsignal

Thecurvesat130Kshowasmallerriseandfallingtime,whilethepositionof
thepeakisbasicallythesame.Moreover,incaseofdeconvolutionmode,onecan
noteanundershooting.Finallyaslightlyhigherpeakvalueispresentinboth
workingmodeswhenthechipisoperatedat130K.Infactatlowtemperature,
thebehaviourofaCMOScircuitisexpectedtodecreaseitscharacteristictime
τ,whichinotherwordsmeansthatthecircuitbecomesfaster.Thisconsistently
explainsallthethreemainfeaturesatlowtemperature.Abasicevaluationshows
thattherisingtimetat130Kisabout15%smallerthanatroomtemperature.
60

Figure4.6:APV25-S0pulseshapedependenceonsignalamplitudeinpeakmode
at130KfordifferentMIPsignals

Fig.4.6illustratesthepulseshapedependencewiththesignalamplitudeinpeak
modeat130K.Theinputsignalvariesbetween1and5MIPsin1MIPsteps.The
saturationofthemaximumofthepulseshapebeginsathighervaluesofMIPthan
atroomtemperature.Thisisduetothefactthatwhenloweringthetemperature
thepositionoftheAPV25-S0sbaselineshiftstolowervalueswithrespecttothe
chip.theofrangedynamic

Noise

Fig.4.7showstheAPV25-S0noisedependenceasfunctionoftheinputcapaci-
tanceatroomtemperatureandat130K.Itisclearlyseenthatatlowtemperature,
thenoisebecomeshigher.Thisbehaviourcanbeexplainedstartingfromthe
measurementsdescribedintheprevioussection.Atlowtemperaturethecharac-
teristictimeτofthechipdecreases;sincethenoiseofthechipsamplifiergoesas
1/τ[NBW],anincreaseofthenoiseisexpectedatlowtemperatures.However,
theobservedvariationissmallanddoesnotinhibitthefunctionalityofthechip.

emperatureT.behaviour

Inordertocheckthebehaviourofthechipscharacteristicversustemperature
variationaroundtheLazarustemperature,thenoiseandthepulseshapeatdif-
ferentMIPsignalsintherange110-150Kin10Kstepweremeasured.Fig.4.8

61

Figure4.7:APV25-S0noisedependenceoninputcapacitanceatroomtempera-
eturK.130and

showspulseshapesfor1MIPsignalatdifferenttemperatures.Asshowninthe
plot,nosignificantchangesinamplitudeandtimeareobserved.
InFig.4.9thedependenceofthemaximumamplitudeof1MIPsignalfromthe
temperatureisshown.Ifoneestimatesthenoiseofthermaloscillationsasthe
variationofADCcountsperKatthemaximumamplitude,onefinds0.125ADC
channels/K.InTab.4.1,thevaluesoftypicalsourcesofnoisearelisted.Thenoiseduetotem-
peraturevariationsisnegligiblecomparedwithothersourcesofnoise.Thisresult
isverypromisinginviewofausageoftheAPV25-S0atcryogenictemperature.

channels)(ADCNoiseceSourNotIntrinsicbondedADCchannelnoise∼0.51
Bondedchannel∼3.5
Fromtemperaturevariation∼0.125/K
Table4.1:DifferentsourcesofnoiseinADCchannelunits.

InFig.4.10thegaincorrespondingatdifferenttemperaturesisshown.Thesat-
urationforinputsignalsbiggerthan3MIPsisduetothelimitationimposedby
range.dynamicaltheThenoiseasfunctionoftheinputcapacitanceatdifferenttemperaturesisshown
inFig.4.11.Asexpected,thenoiseincreaseswhentemperaturedecreases.

62

eFigur

at

4.8:

ferdifent

APV25-S0

pulse

es.temperatur

eFigur

4.9:

shape

Dependence

of

in

peak

signal

mode

for

amplitude

63

1MIP

omfr

equivalent

the

input

e.temperatur

signal

eFigur

eFigur

4.10:

4.11:

APV25-S0

Noise

vs.

gain

at

capacitance

64

entferdif

at

estemperatur

entferdif

es.temperatur

testsPipeline

rTheespondingAPV25pipelinepedestalcellvalues.capacitanceThehasuniformitytobeofuniformpipelinetoatavoid130Kvariationwasfortestedcorto-
verifythatnopeculiaritiesoccuratcryogenictemperatures.Fig.4.12showsthe
aftervalueofconvertingpedestalstheversusrmsvaluepipelinetolocation,EquivalentwhileNoiseFig.4.13Charge.showsThethefigursameerclearlyesults
demonstratesthatevenatlowtemperaturethenoiseassociatedtopipelinedis-
negligible.istribution

Figure4.12:APV25-S0pipelinepedestalforatypicalchannelat130K.

Figure4.13:r.m.s.pipelinepedestalsforall128channels.

65

TheshowninamplitudeFig.on4.14.theThe192widthofcapacitancethepipelinedistributionforisaas1MIPsmallassignalwasat130measurKedas
forroomtemperature,indicatingaclosematchingofcapacitancebetweencells
[5].

Figure4.14:Pipelinegainuniformityat130K.

thermalatoleranceTcycles

InordertousetheAPV25-S0inarealenvironment,itismandatorytomake
surethatthebehaviourofthechipstaysstableevenafterseveralthermalcycles.
Inthecourseofthetestpreviouslydescribed,somechipsunderwentapproxi-
mately50thermalcycles.Additionally,systematictestsononesamplechipwere
performed.Theminimumtemperatureattainedbythechiphasbeenoftheorder
of100K;afterperformingthemeasurementsofthechipcharacteristics,suchas
outputdataframe,pulseshape,noise,thechiphasbeenquicklywarmedupto
roomtemperature.Themeasurementswerethenrepeated50times.Theresults
inFig.4.15andFig.4.16clearlyshowthatnoappreciablechangeoccurredbefore
cycles.thermaltheafterand

4.3Testofthecomponents

Asforthereadoutchip,theusabilityoftheassemblingcomponentsatcryogenic
checked.wasestemperatur

66

eFigur

4.15:

eFigur

APV25-S0

4.16:

pulse

APV25-S0

shape

noise

ebefor

ebefor

67

and

and

after

after

thermal

thermal

cycles.

cycles.

Figure4.17:TheArrangementofthewaferontheL-board

4.3.1Testoftheglue

InFig.4.17,acrosssectionofthedetectorisdepicted.Thecoolingproceeds
throughthesidesofthewaferwhicharegluedtotheL-boardwithasiliconbased
glue.Thegluehadtobeelectricallynon-conductiveandatthesametimepossess
adetectorcertainbyelasticitytemperaturinoredertovariations.absorbtheThemostmechanicaleligiblestrglue,essescrNEEeated001ontweiss,hewholeused
alreadyfortheHERA-Bsilicondetectors[Abt98],wastestedatcryogenictem-
peraturconditions.es,toA5check×7cmwhether2pieceitsofglasscharacteristicsof300µwhermethicknessmaintainedwasinthesesandwichedworkingbe-
ThetweenglasstwowasL-boargluedds,roneprtwooducingsidestheonlyassemblywithsmallusedforspotstheofglueCOMPandASSonthedetectors.two
remainingsideswithacontinuousline.Thesamplewassqueezedinliquidni-
trogenfor3hours.Thesideswiththespotsofgluewereungluedinsomepoints,
whilethetworemainingsideswerestillstuckwithcontinuouslinesofglue.The
glasssamplewaswasthenthengluedsqueezedwithainliquidcontinuousnitrlineogenonforall10thefourminutes,sidesthenoftheextractedboard.andThe
keptafteratrwhichoomnotemchangesperatureappearfor5edinminutes.thesetup.Thecyclewasrepeatedforfivetimes

Thechipsweregluedwithabi-componentepoxy-basedconductiveglue.This
gluewasalsotestedforliquidnitrogentemperature;itnevershowedanydegra-
cycles.thermalmanyafterdation

68

4.3.2Testoftheconnector

InlightedChapterthat3thethetwolayoutL-boarofdsthebetweenL-boardswhichwasthedescribed,waferisitwasgluedhaveparticularlytwodifhigh-fer-
entgroundpotentials.Itwasthenpreferabletousetwodifferentcapillariesfor
thetwosidesofthedetector,inordertohaveacompleteelectricaldecouplingof
thetwoL-boards.Toavoidanymechanicalcomplication,auniqueuxcooling
ofonedetectorwasusedaswellasaninsulatorconnectorbetweenthecapillaries
onthetwosidesofthedetector.Theconnectorwasdesignedwitha180◦angle
tobefittedinthededicatedholeintheL-boards(seeFig.3.7).
Concerningthematerial,thefirstchoicewastouseStesalit,afibreglassmate-
rial.OnbothsidesofasmalltubeofStesalit(∼1.5cmlength,∼1cmexternal
diameter,1.7mminnerdiameter)twocapillarieswerefittedandgluedwitha
twousedforcomponentthemeasurepoxyementsglue.Thedescribedpieceinwassectionthen4.2.insertedTheintocryostatthewassameevacuatedcryostat
at∼10−6mbarandliquidnitrogenushedinthecapillaries.Withthevacuum
gaugemountedonthecryostatitwaspossibletocheckeventualleaksdueto
ofcracks∼60of0thel/hofStesalitgas,orthat,moris∼epr1l/hobablyof,ofliquidtheinglue.theThetube,piecebut,aftercouldfewstandathermalux
cycles,thevacuuminsidethecryostatwaslost.
TheuseoftheStesalittuberequiresthreedifferentmaterialstobeconnected
tobasedtheglue.sameSincepoint:thesetheStesalitmaterialsitself,havethediffercoppenter-nickelexpansioncapillarycoefficientands,thetheepoxyther-
malstressatthispointwashigh.Theideaofmodellingaconnectorwithepoxy
gluenectorandhasgluingapprtheoximatelycapillariesthesametoitwithdimensionthesameoftheglueStesalitwasone.successful.TheepoxyThiscon-con-
nectorunderwentthesametestsoftheStesalit,showingitsreliabilityaftermany
thermalcyclesandaneasierhandlingrespecttotheStesalitone.
InFig.4.18theepoxyconnectorisshown.

T4.4distributionemperature

Asalreadymentioned,inordertoexploittheLazaruseffect,thesilicondetectors
havetobeoperatedatcryogenictemperatures,preferably130K,sinceforthis
valueoftemperaturetherecoveryoftheCCEismaximum.Thesilicondetec-
torsusedinCOMPASShaveanactiveareaof5×7cm2,andthechallengewasto
ensureuniformtemperatureonthewafer.Afirstsetofmeasurementswasper-
formedinordertohaveadistributionofthetemperatureontheL-board,while

69

Figure4.18:Theepoxyconnector

thetemperaturedistributiononthesiliconwaferistheresultsofsimulations.

4.4.1TemperaturedistributionontheL-board:measurements

Thereadoutchipsdissipateanon-negligibleamountofpower(∼5W)andcon-
stitutethemainsourceofheatontheboard.Thisisthereasonwhythecooling
capillarieshavetopassunderneaththewaferonthesidewherethechipsare
prperaturesent,einorderdistributiontostoponthetheheatL-boarowd,toathePCBprdetector.ototypeInorwasdertobuild.measurTheethedistribu-tem-
tionofthecopperlayersinitsinnerreproducesthedistributionoftheL-boards,
ofbut,theinsteadchips,1ofkΩusingSMDtwo1resistorsL-shapedwerPCBs,eused.asingleFiverpieceesistorswasweredesigned.placedatInsteadeach
chipsurfaceofposition,thetochip.simulateThertheesistorsfactwerthatetheconnectedpowertoisandissipatedexternalovertunablethepowerwhole
supply,sothatavariationofthevoltageappliedcorrespondedtoavariationof
thepowerdissipated.Copper-Nickelcapillariesweretin-solderedondedicated
padslocatedonthefoursidesofthewaferposition.Allaroundthewaferposi-
tionandacrossthechips,23Pt100thermometerswereplaced.Thethermometers
werereadoutviaaNI4351NationalInstrumentreadoutcard,withaLabview
tion,graphicalinordertointerface.haveAthepiecesameofglassthermal300loadµmasthicktherwasealglueddetectoron.Thethegoalwaferwasposi-to

1SurfaceMountableDevice

70

Figure4.19:ThePCBdesignedforthermaldistributiontests

measurethetemperaturedistributionalongthewaferposition.InFig.4.19the
depicted.isdboardedicated

Theboardwastheninsertedintothecryostatdescribedinsection4.2andthe
cryostatwasevacuated.Theworkingpressureinsidethecryostatwas∼10−6
mbar.Byvaryingthevoltageappliedtotheresistorsandthenitrogenuxin
thecapillaries,acompletescanoftemperatureversuspowerdissipatedandux
achieved.becould

InFig.4.20thetemperaturebehaviourregisteredbyonethermometerasafunc-
tionofthepowerdissipatedfordifferentliquidnitrogenuxesisshown.The
∼curve130Kforeven250ifl/htheshowspowerthatisdissipatedpossiblevaries.toachieveaconstanttemperatureof

InFig.4.21,thetemperaturedistributiononthetwosidesofthewaferposition
≥for1504Wl/h,powerananduniformdifferenttemperaturLN2euxesonisbothshown.sidesofOnethecouldwaferseecouldthatbeforaachieved.ux
l/h(Referring∼250tol/hFig.of4.20,gasonemeasurcaned)thenitisassertpossiblethattowithaachieveliquidanitruniformogenuxoftemperatur∼0.e4
of∼imately130Kconstantonbothevensidesforaofthevariationwaferoftheposition.powerThedissipatedtemperaturofe±3staysWarapproundox-
.W5

71

Figure4.20:Thetemperaturebehaviourregisteredbyonethermometerasfunc-
tionofthepowerdissipatedfordifferentLN2uxes

Figure4.21:Temperaturedistributiononthetwosidesofthewaferpositionfor
4WpoweranddifferentLN2uxes

72

Figure4.22:Temperaturedistributiononthesiliconwafer(fromsimulation)

4.4.2Temperaturedistributiononthewafer:simulation

Oncethepossibilitytoachieveanuniformtemperatureof∼130Konbothsides
ofthewaferontheL-boardwasinvestigated,thenextaimwastochecktheuni-
formityofthetemperatureonthewafer.Thiswasperformeddoneviasimula-
tions.Thethermalconductivityofsiliconcrystalsat300K(roomtemperature)
is156W/Km(foriron80W/Km,forair0.024W/Km).Withdecreasingtem-
perature,thethermalconductivityofsiliconreachesvaluesbetween950and420
W/Kminthetemperaturerange100-150K(see[GLA64]),implyingthatifone
coolsthesystemdownto130Konbothsidesofthewafer,anuniformtempera-
tureof130Kwillbeachievedovertheentiresurface.

Totestthishypothesis,aC++simulationpackagewasdeveloped,tocalculate
thetemperaturedistributionofasiliconwaferwithfixedtemperaturealongtwo
sides.Thewaferwassimulatedasablackbody,suchthatthepowerdissi-
patedviairradiationfollowstheStephan-Boltzmannlawanditcanbewrittenas
R=σΔT4wereσistheStephan-BoltzmannconstantandΔTisthetemperature
gradientbetweenthewaferandthecryostat(thelatterbeingatroomtempera-
ture).Thisistheonlypowerdissipatedinthesetup.Thethermalconductivityas
functionofthetemperaturewasalsotakeninaccountandinthiswaythetime
evolutionofthetemperatureonthewafercouldalsobeevaluated.InFig.4.22the
temperaturedistributiononthesiliconwaferatthermalequilibriumisshown.

frTheomplotsbothshownsurfacesinwhileFig.4.22oncewertheedetectorobtainedisinstalledsupposinginthatthethecryostat,waferoneirradiatessurface

73

willseeanotherwaferwhichisatthesametemperatureof130Kandthusthe
powerirradiatedisreducedofafactortwo.Moreover,theapproximationthatthe
siliconwaferisablackbodyoverestimatesthepowerdissipated.Onecanthen
statethatthevaluesobtainedviasimulationsareupperlimitsonthetemperature
gradientonthewafer.Nevertheless,thevalueobtainedof7Konthesilicon
acceptable.edconsiderwassurface

4.5AcoldstationinCOMPASSexperiment

Afterensuringthereliabilityoftheindividualcomponentsforcryogenicopera-
tion,thecompletedetectorwasassembled.Thedetectorwasinstalledinthecryo-
statandoperatedatcryogenictemperature.Themeasurementwereperformed
intheMunichlaboratoryusingmeasurementsetupdescribedinsection4.1.The
mainfunctionalitiesofthedetectorweretestedusingaStrontiumsourceanda
.setup,triggerdedicatedOncethedetectorwasevaluatedtobecorrectlyfunctioning,acompletestation
equippedwithcryogenicfacilitieswasassembledandinstalledintheCOMPASS
experiment.Inthissection,theinstallationprocesswillbedescribed,followed
bythecharacterisationofthesilicondetectors.

Installation4.5.1experimentthein

Inordertoensureastablecooling,adedicatedsystemforcontrollingandtuning
wasinstalleddevelopedduringintheMunich.2003beamDuetotimeexternalandsotheconstraints,laboratorythesetupsystemwascouldused.notbeA
descriptionofthesystemwillbegiveninthenextsection.Electronicmassow
meterswereusedinsteadintheexperimentofthemanualones.Theowmeters
werlowedethecontruxolledtobeandrcontreadolledoutviaandamonitorLabviewedprwithogramahighrprunningecisiononaandPC.Thiswithoutal-
accesstheexperimentalarea.Thenitrogendewarwasconnectedtothemainni-
trogenvalveintheexperimentalarea;unfortunately,itwasnotpossibletorefill
itautomatically,butthisprocedurewasperformedmanuallytwiceaweek.The
cryostatwasevacuatedtillapressureof∼10−5mbar.Toavoidwatercondensa-
tiononthebeamthroughwindows,forcedventilationontheouterofthecryostat
ganised.orwasThecompletesetupisshowninFig.4.23.

74

4.5.2

Figure4.23:Thecoldsetupintheexperimentalhall

takingdata2003The

ExploitingatemporaryfailureoftheSPS,thelaboratoryset-upwasinstalledand
testedintheareaattheendofJuly.Unfortunately,thisset-updidnotguaranteea
longtermreliability.Firstthecapacityofthedewaronlyprovidedsufficientliq-
uidnitrogenforapproximatelythreedays,afterthatithadtobemanuallyrefilled
andthisoperationrequiredaccesstotheexperimentalarea.Additionally,itwas
observedthatthecapillarytubeswereregularlyobstructedwithice,preventing
anyfurthercooling.Theicemostlikelyoriginatedfrommicroscopicwatercrys-
talsintermingledwiththeliquidnitrogen.Thisblockagecausedaslowdecrease
inthesupplythroughthecapillariesandaconsequentincreaseinthetempera-
tureofthedetectorafterapproximatelyonedayofstableconditions.Fig.4.24
showsthisbehaviour.Therisingslopeinthesecondpictureisexplainedbythe
formationoficeinthecapillaries.

Inordertoavoidlossofbeamtimeduringthedatataking,itwasdecidedto
operatethedetectoratroomtemperatureuntilSeptember,whenabeamwitha
peculiar25nsstructure(notusefulforphysicsdatataking)wouldbedelivered
bytheSPS,allowingacompletetestofthedetectorfeaturesatcryogenictemper-
atureswithintheCOMPASSenvironment.

75

Figure4.24:Temperaturebehaviourduring2003beamtime

performancesDetector4.5.3

Thedatatakenduringthe25nsbeamtimewereanalysedinordertocomparethe
detectorperformancesatroomandatcryogenictemperatures.Asexplainedin
Sec.2.1.2,thesizeofthechargecloudisdirectlyrelatedtothespacialresolution.
Nochangewasfoundinthesetwoparameterswhilethedetectorwasoperatedat
130K.Asignificantimprovementwasfoundinsteadinthetimeresolution:from
atypicalvalueof≈2.3nsatroomtemperatureto≈1.2ns,animprovementby
almostafactortwo.Thevariationismostlikelyduetothechangeofthereadout
chippreamplifierparameters.Theobservationsareconsistentwiththemeasure-
mentsdescribedinSec.4.2.Acompletedescriptionoftheanalysisperformed
[Fuc04].infoundbecan

systemcontrolcoolingThe4.6

Operatingmoresiliconstationsatcryogenictemperaturewithastand-alonecool-
ingsystem,inanexperimentalenvironmentwheretheaccesstotheapparatusis
notalwayspossibleandspecialsafetyconditionsarerequired,presentedsev-
eralproblemsforwhichfindingasolutionhasbeenachallengingtask.Acom-
pletesystemwasdevelopedwhichallowsindependentcoolingofseveralsta-
tions,withaselfregulatingtemperaturesystem.Itconsistsofthemechanical
infrastructurestohandleliquidnitrogenandthetemperaturecontrolelectronics
(hardware+software).Asmentioned,thissetupwasnotusedduringthe2003
run,butitwillbeinstalledinautumn2004.Inthissectionanextensivedescrip-
tionofthecomponentsisgiven.

76

Figure4.25:Theowscheme(theobjectsarenotonscale)

designmechanicalThe4.6.1

ThedetectordesigninfortheaCOMPliquidASSnitrogenexperimenttransferislinedescribedandindistributionthissection.boxtoIttheisasiliconfixed
installationaimedtotransferliquidnitrogenbetweenthemaingauge,locatedon
asideoftheexperimentalarea,andthesilicondetectors.Themaingaugeisde-
liveringLN2at3.2bar.Theremotelycontrolledcryovalveinstalledonthegauge
isconnectedviaabayonettoa30mlongvacuuminsulatedexibletransferline
tothedistributionbox.Thenitrogenuxistheredividedinordertoprovide
separatecoolingtothedifferentsiliconstations.Liquidnitrogenissuppliedto
thedetectorsisdonevia2mvacuuminsulatedtransferlineswithbayonetcon-
nectionsoneachend.Eachlinehasanestimateduxof2l/hourofLN2.The
nitrogenowsthroughthecapillariesandcoolsdownthedetectorasdescribed
inSec.3.3.Thenitrogenexhaustiscollectedandevacuatedoutsideoftheexperi-
mentalhall.Thenitrogenowfromthemaingaugetothedetectorisdepictedin
4.25.Fig.

Itinisordertoworthwhileavoidtothementionformationthatofaicewaterinsidefilterthewillbecapillariesinstalledthatatthecomprmainomisedgauge,the
correctoperationofcryogenicsilicondetectorsduring2003beamtime(seeSec.
4.5.2).

AllthedimensionsofthecomponentshavebeenoptimisedaccordingtotheGer-
mansecuritycodeforunderpressuredevicesAD-Merkbl¨atter.Inthefollowinga
moredetailedviewoftheindividualcomponentsisgiven.

77

boxdistributionThe

Figure4.26:Thedistributionbox

Thedistributionboxisavacuumcryostatwhichdividesthemainsupplyofliq-
uidnentsnitrareogentheandvessel,redirtheectsitshieldtotheandtheindividualdistributionsiliconsystem.stations.Fig.Its4.26mainshowscompo-a
box.distributiontheofsketch

•Thevesselisastainlesssteelcylinderclosedatthebottomend,withadi-
ameterof2m,1.5mheightand5mmthickwalls.Ithostthedistribution
systemanditisclosedonthetopwithaangewhereallserviceconnections
aremounted.Aserviceangeforpumpingisinstalledonthesideofthe
vessel.ThevesselisevacuateddowntoP≈10−2mbar,toensureagood
thermaldecouplingbetweenthedistributionsystemandtheouterenviron-
ment.Thevesselisadditionallyequippedwithanvalvethatopensinofan
internalpressure0.2barhigherthanatmospheric.Thissystemismeantto
preventpressurebuildingupinsidethevesselinthecaseofnitrogenleak
system.distributiontheomfr

•Acoppershieldisinstalledbetweenthevesselandthedistributionsystem
toimprovethethermaldecouplingofthelatter.Theshieldiscooledto
≈100Kviaadedicatedcoolingline.Theamountofenergytransferredvia
radiationbetweentwobodieswithtemperaturesT1andT2is∝T14−T24.
Withouttheshield,theenergytransferredtothenitrogeninthedistribution

78

systemwouldbe∝3004K−774K.Thepresenceoftheshieldreducesthis
quantityto∝1004K−774K.Inthiswayalowergas/liquidratiointhe
tubescanbeachieved,increasingthecoolingpropertiesofthemixture.

Fig.4.26showsthedistributionsystem.Theogennitrcomesomfrthemaingaugethroughthetransferline,thatisconnectedwithabayonettothedis-
tributionbox.Themaintubeisthendividedinseventubeswithsmaller
diameters.Eachofthesetubesisconnectedtoacryogenicvalve,inorderto
openandclosethesupplyofLN2foreachsiliconstationindividually.One
ofthetubesissolderedalongthewallsofthecoppershielding,whilethe
othersixaredirectedtothetopangewheretheyareconnectedviabayo-
nettothetransferlinesofthestations.Tominimisethethermallossdueto
radiation,thetubesarewrappedinseverallayersofsuperinsulatingmate-
rials:aMylarfoilcoatedwithaluminium.Themaintubeisequippedwith
anoverpressurevalve(P=4bar),toavoidpressureincreasesthatwould
beoverthesecurityvalues.ThesystemwasbuiltintheMunichworkshop,
readaptinganexistingstructureusedintheTritronexperiment.Thefunc-
tionalityofthesystemhasbeensuccessfullytestedinthelaboratory.

linestransferThe

Asalreadymentioned,thenitrogensupplyisprovidedviaamaingaugeinstalled
intheexperimentalhall.Thedistributionboxislocated∼30mfromthegauge,
closetothesilicondetectorplatform.Thetransferofliquidnitrogenisdonevia
atransferline.Thetransferlineisexible,sinceitmustfollowacomplicated
pathfromthegaugetothedistributionbox.Thelineiscomposedoftwocon-
centricexiblestainlesssteeltubes,betweenwhichisavacuum.Thenitrogen
owsthroughtheinnertubeandthevacuumensuresgoodthermalinsulation.
Custom-madeTeonspacersareinsertedevery20cmthroughoutthewholetube
lengthtoavoidanycontactbetweenthetwotubeswhichwouldreducethether-
malinsulation.Theinnertubeiswrappedinseverallayersofsuperinsulating
materials.ThetransferlinewasmanufacturedattheCERNcryolab.Othertrans-
ferlinesareusedfromthedistributionboxtothedetector.Themainlinestructure
isrepeatedonsmallerscale.

Thegasformationfractioninthetransferlineswasestimatedtobesmallfor
therequiredcoolinguxof≈2l/heachstation.Inthiswaythecoolingofthe
optimised.wasdetectors

79

insulationElectrical

AspowermentionedandgrinoundingSec.3.3.2,lines.theInrordereadouttochipsavoidartheeverypickupsensitivenoisetofromthethenoisefrcoolingom
system(transferlinesanddistributionbox),ithastobeelectricallydecoupled
fromthecryostat.Forthispurpose,aconnectingstructureofTeonwasde-
signed,thatwasinstalledonthebayonetwhichconnectsthetransferlinewiththe
cryostat.Teonisusedsinceitisanexcellentinsulatorandadditionallyhighly
suitableforcryogenictemperaturesenvironment.

4.6.2Thecontroloftheux

Thetemperatureofthesiliconwaferdependsupontheamountofnitrogenthat
owsinthecapillaries.Therefore,afastandreliableregulationofthenitrogen
uxisofcapitalimportancetoachieveastabletemperature.Forthisreason,the
uxatdevelopedtheoutlettemperaturoftheecontrcryostats.olisThisbasedchoiceontheallowscontrolofindependentthegaseoustuningnitrofogenthe
twodetectorsinonestation,minimisingthenumberofinletsinthecryostat.In
thefollowingtheindividualcomponentsoftheuxcontrolleraredescribed.

ControllerFlowThe

TheBronkhorstEL-FLOWmassowcontrollerwasusedtomeasureandtunethe
nitrogengas.Viaananaloginput(0-5V)itispossibletocontrolgasowsfrom0
inuptheto900rangel/h.0-5TheV.Formeasurmoreementdetailedoftheefinformationfectiveuxseegoes[Bro].viaanalogoutputalso

EmbeddedLocalMonitorBoard(ELMB)

TheactiveuxcontrolisperformedviaELMB[ELMa].TheELMBisamulti
purposefrontendI/OsystemwithaCANinterface,developedfortheATLAS
experimentatCERNandalreadyusedinCOMPASSforseveralmonitoringtasks
[ELMb].TheELMBreadsthethermometeronthesiliconwafer(seeSec.3.3.2)
and,dependingonthetemperaturevalue,adjuststheuxattheowmeter,ac-
cordingtoanalgorithmdevelopedspecificallyforthispurpose.Furthermorethe
ELMBreadsseveralparametersofthesystem(e.g.pressureinthecryostat)and
comunicatesthemtotheCOMPASSDetectorControlSystem(DCS)thatmonitors
andstoresthemeasuredvaluesintoadatabase.Foramoredetaileddescription
oftheusageoftheELMBinthesiliconcoolinginfrastructure,see[See04].

80

Figure4.27:Schematicalviewofthecoolingsetup

boxsafetyThe

Topreventthedamageofthesystemthatmaybecausedbythefailureofsome
components,asafetydevicewasdeveloped.TheSafetyBoxconsistsofaseriesof
surlogicale,ux)cirandcuitsthatcomparesmonitorsthemwithseveralsomevaluesprofedefinedthesystemvalues.(e.g.Incaseoftemperaturanomalies,e,pres-
ofthethesafetydetectorboxandswitchesclosesoffthethevoltagecryogenicofvalvetherthateadoutprovideselectrtheonics,liquidthebiasnitrogenvoltageto
thedefectivestation.Thesafetyboxisdesignedtooperateasstandalonecom-
ponent.

operationofprincipleThe4.6.3

Fig.4.27showsaschematicviewofthecoolingcontrolsystem.Thetemperature
ofthesilicondetectorisreadoutviaPT100bytheELMB.Ifthetemperatureis
differentthenrequired,theELMBsendsacommandtotheowcontroller,to
varytheuxaccordingly.Thesafetyboxswitchesoffthestationinthecaseof
failureofacomponentofthesystem.

81

5Chapter

ThesearchASSCOMP

fortheΞ−−

pentaquarkinThediscoveryofstatescomposedoffourquarksandanantiquark,collectively
calledpentaquarks,in2003openedaneweraforhadronspectroscopy.TheCOM-
PASSexperimentcanmakeavaluablecontributiontotheinvestigationofpen-
taquarks,troughthehighstatisticsandprecisionitoffers.
Inthischapter−−thesearchforanexoticpentaquarkwithcharge-2andstrangeness
+1,calledΞ,usingdatafromtheCOMPASSexperimentisdescribed.Inthe
firstsectionanoverviewofthepresentunderstandingofpentaquarkstatesis
given,bothfromatheoreticalandexperimentalpointofview.Inthesecond
sectiontheanalysismethodisdescribed;theobtainedresultsarepresentedand
theirinterpretationinrelationwithotherexperimentsisgiven.

Introduction5.1

Formanydecades,highenergyphysicsexperimentshavebeendevotedtothe
searchofbaryonsandmesonsthatarenotcomposedofthreequarksora
quark−antiquarkpair,respectively.Someofthesestatesareassumedtohave
chargeandquantumnumbersincombinationswhichareforbiddenfornormal
states,theyarecalledexotics.Thesearchforexoticshasbeenratherunsuccesful,
andinits1988reviewtheParticleDataGroupstated[PDG00b]:

Thegeneralprejudiceagainstbaryonsnotmadeofthreequarksandthe
lackofanyexperimentalactivityinthisareamakeitlikelythatitwillbe
another15yearsbeforetheissueisdecided.

83

Thesituationremainedunchangeduntil2003,whentheevidenceofaverynar-
rowbaryonicstatewasreported[LEP03].Thestate,calledΘ+,hasstrangeness
one,chargeoneandmass1540MeVanddecaysinΘ+→nK+.TheΘ+s
minimumquarkcontentisuudds,i.e.acombinationoffourquarksandananti-
quark;thereforethestatehasbeencalledpentaquark.Despiteofcomprisingmore
thanthreequarks,thisstateisstillbaryonic,sinceithasbaryonicnumberB=1.
Almostayearlater,anotherpentaquarkwasobserved:theΞ−−(ddssu)at1862
MeV.Sincethesefirstobservations,severalexperimentshavetriedtoconfirmthe
existenceofthepentaquarksleadingtocontroversialresults.Stilltoday,thereis
noconvincingexperimentalevidencethatthepentaquarksexist.
Fromatheoreticalpointofview,thesituationisnotmuchclearer.Stateswith
morethanthreequarksarenotforbiddeninQCD,providedthattheyarecolour-
less.Howeverintheconstituentquarkmodelsuchastatewouldbeexpected
tofallapartsorapidlythatitswidthwouldbebroad,contrarytotheexperi-
mentalresults.AssumingQCDisstillvalid,severalmodelshavebeendevel-
opedproposingalternativestotheconstituentquarkmodel.Furtherexperimen-
talmeasurementsareneededtodiscriminatebetweenthem,accordingontheir
predictionsofobservablesquantitieslikemass,widthandparityofpentaquarks.
Inthissectionareviewoftwotheoreticalmodelsandoftheexperimentalresults
onpentaquarksearchesisgiven.

modelsTheoretical5.1.1

Manydeveloped.differentTheymodelscanbewhichseparatedpredictintothethoseexistencethatarofeanexoticextensionbaryonsofhavethebeencon-
stituentquarkmodel(i.e.diquarkmodel)andthosewhichproposeacompletely
taldifrferesultsentonapprtheoachΘ+,(i.e.chiralconcerningsolitonitslightmodel,masssee(m∼below).1540TheMeV)recentanditsexperimen-narrow
thiswidth(sectionΓ10twoMeV),modelshaveareputoutlinedharshwhichconstraintsareoncurrtheentlytheortheeticalmostprpredictions.omisingtoIn
baryons.exoticunderstand

modelsolitonChiral

Intheconstituentquarkmodel,abaryonisformedbythreequarkswhichin-
teractwitheachotherviagluonexchange.Theinteractionsbetweenthecon-
stituentquarksandtheonesintheDiracseaareneglected.Thequantumnum-
bers(strangeness,beauty,etc.)ofabaryonaresimplythesumofthequantum
numbersofitsconstituentquarks.Inthechiralsolitonmodel,orSkyrmemodel,

84

Figure5.1:Thebaryonicantidecuplet10.Theparticlesinthecornershaveexotics
numbers.quantum

abaryoniscomposedofthreequarkswhichinteractwitheachotherandwith
thequarksintheDiracsea.Thestrengthoftheinteractionbetweenquarksbe-
comesweakerasthedistancebetweenthecorrespondentenergylevelsincreases.
Inmostpractice,uppertheenerthrgyeelevelsandconstituentnotwithquarksalltheinteractquarksonlyinthewithDiracthesea,quarkswhichintheare
niteinfiniteprinocess.numberIfone.Thisconsidersmeansonlythatthethequarkscalculationintheoflightphysicalsectoru,quantitiesdandiss,athefi-
classificationofstatesincolourmultipletsoftheconstituentquarkmodelisstill
valid,antidecupletbut1besides0appearsthewellintheknowchiral8andsoliton10modelmultiplets(Fig.of5.1).baryons,Theanparticlesadditionalwhich
containquantumannumbers.antiquark,InFig.but5.1notthetheexoticcorrespondingbaryonsarequarktheones(i.e.atddsthesu)thrhaveeecornersexotic
triangle.theofThemostimpressivepredictionofthechiralsolitonmodelhasbeentheexistence
ofwidthanΓexotic<15baryonMeVwith[Pol97],charthegeΘ++1,,thatwasstrangeness+1,experimentallymassm∼claimed1530inMeV2003andby
[LEP03].experimentLEPSthe

modelDiquark

Themainideaofthediquarkmodelisthattwoquarksformastronglybounded
compoundqqiftheyformacolourantitriplet,avourantisymmetric,spinsinglet
withevenparity.Consideringonlythelightquarks(u,dands),itispossibleto
demonstrate[jaf04]thathighlycorrelatedscalardiquarksarethecombinations
ud,suandds.Ifastateisdominatedbyoneoftheseconfigurations,itshould
belighterandmorestablethanstatesformedofothertypeofdiquarks.This

85

modelwouldalsoexplainwhynoexoticsmesonsareobserved(anycombination
ofgooddiquarkandantidiquarkqqqqwillhavethesamequantumnumbers
ofastandardmeson).Ifoneappliesthismodeltopentaquarksonecanfind
thatallpossiblestatesformedbytwodiquarksandoneantiquark(qq)(qq)qare
organisedinthefollowingavourmultiplets:

•anonetwithspin1/2andnegativeparity
•anoctetwithspin1/2andpositiveparity
•amodel)antidecupletwithspin3/2andpositiveparity(likeinthechiralsoliton

Again,exoticpentaquarksareatthethreecornersoftheantidecuplet(Fig.5.1).
Acompletereviewofthediquarkmodelcanbefoundin[jaf04].

resultsExperimental5.1.2

Theexperimentalobservationofpentaquarkstatesisprobablythemostimpor-
tantresultinhadronspectroscopyinthelastyears,butitisnotaneasytask.Pen-
taquarksharerareparticlesandtheirmassesarepredictedinarangewheremany
otherresonancesarepresent(m∼1500−2000GeV).Ifoneconsidersthestates
atthecorneroftheantidecuplet(Fig.5.1),theyaremanifestlyexotics,which
meanstheirquantumnumberscannotbelongtostatescomposedofonlythree
quarks.Theidentificationofsuchexoticstatesismucheasier,sincenootherres-
onancesareexpectedinthecorrespondinginvariantmassspectrum.Thissection
containsareviewofthepublishedresultsfortheobservedpentaquarks.Many
experimentsarereanalysingtheexistingdatainthequestforpentaquarksand
newresultsareexpectedsoon.Otherdedicatedexperimentsarebeingplanned
andwillbecarriedoutinthenearfuturetoimproveourknowledgepentaquarks.

+ΘofMeasurements

Thefirstexperimentalobservationofanarrowbaryonresonancewith
strangeness+S=+1,calledΘ+,hasbeenreportedbytheLEPScollaboration
[LEP03].Θcontainsoneexcessstrangeantiquark.Accordingtothequark
bemodeluudditsandcannotitisbeathrmanifestlyeequarkexotic.stateThebutΘ+itsresonanceminimumcorrquarkespondscontenttoahassharpto
peakat1.54±0.001GeVintheK−missingmassspectrumoftheγn→K+K−n
86

reaction.Thewidthoftheresonancehasbeenestimatedtobesmallerthan25
MeVandtheGaussiansignificanceofthepeakis4.6σ.Severalexperimental
groupshaveconsistentlyconfirmedthisobservation[DIA03],[CLA03],[SAP03],
[HER03].TheyobservedasharppeaksinthenK+orpKS0invariantmassspectra
atthemass∼1540MeVandwithawidthlimitedbytheexperimentalresolution.
However,manyotherexperimentshavenotobservedanysignalcorresponding
toaΘ+baryoninthesameinvariantmassspectra.Significantdifferencemay
comefromthemechanismproduction,whichcanbesubstantiallydifferentin
e+e−collisionsratherthaninphoto-production,forexample.Presently,more
measurementswithhigherstatisticsarerequired.Moreovermeasurementsof
theparityandwidthoftheΘ+willprovidetheonlyconstraintstodiscriminate
betweendifferenttheoreticalmodels(outlinedinSec.5.1.1).

MeasurementsofΞ3−/2−andΞ30/2

Presently,theonlyexperimentalobservationofastatewithchargeQ=−2and
strangenessS=−2,calledΞ3−/2−(minimumquarkcontentuussd),hasbeenper-
formedattheNA49experimentatCERN[na403].NA49usedaprotonbeamon
aprotontargetatenergy√s=17.2GeVwithxFintherange[−0.25,0.25].By
chain:decaytheanalysing

Ξ−−→Ξ−π−→Λ0π−π−→pπ−π−π−(5.1)

anarrowresonanceintheΞ−π−invariantmassspectrumwithamass1.862±
0.002GeVhasbeenobserved.Thewidthoftheresonanceisbelowthedetector
resolutionΓ<0.018Gev.AtthesamemasspositionasignalintheΞ−π+in-
variantmassspectrumhasalsobeenobserved.Thisisoriginatedbytheneutral
andnonexoticmemberoftheS=−2isospinquartetΞ30/2.Fig.5.2showsthe
Ξ−π−andΞ−π+andtheirchargeconjugateinvariantmassspectraasmeasured
NA49.by

Thesamedecaychannelhasbeenanalysedbyotherexperiments(see,forexam-
ple,[HER04]and[wa804])withdifferentbeams,targetsandenergies.Nonehas
confirmedtheobservationofNA49.ThesearchoftheΞ3−/2−andΞ30/2withinthe
COMPASSexperimentisoneofthemaintopicsofscienceanalysispresentedin
thisthesisandwillbeextensivelydescribedinSec.5.2.

87

Figurby

ethe

5.2:InvariantNA49

massexperiment.

spectraforΞπ−

88

,−

π−Ξ

+,

+Ξπ

and

+Ξπ

+

asobserved

Measurementsofexoticsnotintheantidecuplet10(ΘcandΞ21−)
Consideringmoreavoursthantheu,dandsquarks,otherheavierpentaquarks
with∗−charmorbeautycanbeformed.Morerecentlyanarrowsignalinthe
pDinvariantmassspectrumhasbeenobservedbytheH1experimentatDESY
[H104]withmassm=3099MeVandwidthΓ<12MeV.Thesignalisacan-
didatefortheΘc(3099)pentaquarkwithminimumquarkcontentuuddc.The
ZEUSexperimentdoesnotseeanysignalinthesamechannel.
Thepentaquarksintheantidecupletarepredictedtohavespin3/2.Anexotic
statewithspin1/2hasbeenobservedbyNA49inthechannelΞ1−→Ξ0∗π.The
candidatehasamassm=1855MeV[KA04].Thesamedecaychain2hasalsobeen
investigatedinthisthesisanditwillbepresentedlater.

5.2TheanalysisoftheCOMPASSdata

ThespectroscopyofexoticstatesisoneofthefuturegoalsoftheCOMPASS
hadronprogramfrom2006(seeSec.1.1.2).Thedataalreadyobtainedwith
themuonbeamalsoofferthepossibilitytoperformasearchforpentaquarks.
Inparticular,COMPASSisoptimisedforthequasi−realphoto-pr2oduction,thatis
whentheenergytransferredinthescatteringprocessissmall(q≤1MeV)
andthescatteringphotonisalmoston-shell.TheΘ+pentaquarkshavebeenob-
servedinthisproductionregime.Moreoverthehighresolutionandthehigh
statisticsoftheCOMPASSspectrometermakeithighlysuitablefortheanaly-
sisofrareparticles,likethepentaquarks.Currently,studyofthedecaychannel
Θ+→pK0→pπ+π−withtheCOMPASSdataarebeingperformed.Inthis
workthechannelΞ−−→Ξ−π−hasbeenanalysed.Asresult,nopositiveevi-
denceofasignalcorrespondingtosuchpentaquarkshasbeenfound;however,
anupperlimitfortheproductioncrosssectionhasbeendetermined.

luminosityandSampleData5.2.1

ThefollowinganalysishasbeenperformedontheCOMPASSdatafrom2002and
2003.PHASTversion7.006hasbeenused(seeSec.1.2.3).
TheluminosityListheproductofthebeamuxIandthenumberofscattering
centresinthetargetC.AbuiltinfunctioninPHASTprovidesthevalueofthe
integratedmuonuxI.Intotalforthe2002and2003dataI=6.4∙1013.The
numberofscatteringcentresCpercm2isgivenby:

89

NC=VAlåAiPi(5.2)
AiwherisetheNAisatomictheAmassvogadrofostheinumberelement,Visofthethetartargetgetvolume,andPliisthethetarnumbergetoflength,its
mols.TheCOMPASStargetismainlycomposedof6LiD,withsomecontami-
Tab.nation5.1,frCom=3.other488∙102materials.5cm−2.TtheThecompositionluminosityisofthenthetargetmaterialisgivenin

L=I∙C=6.4∙1013×3.488∙1025cm−2=2240∙1039cm−2=2240pb−1(5.3)
Element#ofmolsAtomicnumber
DH44.53700.101912
34HHee3.620710.911834
76Li42.84296
71.7950iL120.0080C190.0160F58.60.0057iN63.60.0136uCTable5.1:Compositionofthetargetmaterial.

topologyEvent5.2.2AsignatureoftheS=-2pentaquarkΞ−−isrelatedtothefollowingdecaychain
5.3):(Fig.

Ξ−−→Ξ−π−→Λ0π−π−→pπ+π−π−.(5.4)
Alsothefollowingdecaychain,originatingfromthewellknownΞ0∗stateiscon-
ed:sider

Ξ0∗→Ξ−π+→Λ0π−π+→pπ+π−π+.(5.5)
90

Itisusefultorememberthat,intheΞ−π+invariantmassspectrum,NA49has
observedasignalcorrespondingtotheneutralmemberoftheisospinmultiplet
5.1.2).Sec.(see

Figure5.3:Thedecaychainconsideredinthisanalysis:Zprim,ZΞ−,ZΛ0arethe
longitudinalcoordinatesoftheprimaryvertex,theΞ−andtheΛ0decayvertex,
.espectivelyr

Thereconstructionofsucheventsstartfromtheendofthedecaychain:

•eventswithΛ0candidatesareselected;
•inthissubsampleeventswithΞ−areidentified;
•Ξ−arecombinedwithapiontracktoreconstructΞ−−andΞ0∗;
•Ξ0∗arecombinedwithapiontracktoreconstructΞ0∗πstates.
Duringthe2002and2003dataruns,theRICHdetectorcouldnotperforman
accurateparticleidentification.Aclearp/πseparationcouldbeobtainedonlyin
alimitedregionoftheacceptance.Thismeansaclearparticleidentificationcould
beachievedatthecostofasignificantreductionofthestatistics.Thereforethe
dataformtheRICHdetectorarenotusedinthisanalysis.Onlymuonsareclearly
identifiedusingthemuonfilters(Sec.1.2.2).Theinvariantmassofaparticleis
reconstructedmakingassumptionsonthemassesofthedaughterparticles.
AninstructiveexampleisthecaseoftheΛ0decayinginthepπ−channel.For
eachpairofparticlesofoppositecharge,thepositiveoneisassumedtobeapro-
tonandthenegativeoneapion.Ifthetwoparticlesarereallyaprotonandapion
comingfromaΛ0decay,theirinvariantmassisGaussianlydistributedaround
thevalueoftheΛ0masswithdispersionthatistheconvolutionofthedetector
resolutionandthewidthoftheparticle.Iftheyarenotcoming−fromtheΛ0decay,
theywillcontributetothebackground.Altogether,thepπinvariantmassspec-
trum,willlooklikeaGaussianpeakcorrespondingtotheΛ0superimposedon
thebackground.ComparingthenumberofeventsundertheGaussianpeakand
91

thebackgroundinthesameregiongivesanideaofthequalityofthereconstruc-
tion.Asitwillbeshowninthefollowingsections,aclearreconstructionofΛ0,Ξ−
andΞ0∗particleshasbeenachievedevenwithoutusingtheRICHinformations.

Preliminary5.2.3eventsofselection

Thefirstselectiontobeappliedtotheeventsampleconcernstheprimaryinterac-
tionvertex,thatistheinteractionpointofamuonfromthebeamandanucleon
get.tartheinInCORAL(Sec.1.2.3)aprimaryvertexisidentifiedbycontaininganincoming
muontrack.Areconstructedeventcanhavezero,oneorevenmoreprimaryver-
tices.ToreconstructthekinematicsoftheΞbaryonsproduction,theinformations
ontheenergytransferredintheprimaryinteractionareneeded.Thereforeevents
withzeroprimaryvertexhavebeendiscarded.Incaseofeventswithmorethan
oneprimaryvertex,theonewithmoreassociatedtrackshasbeenselected.The
eventsampleafterthecutsontheprimaryvertexisreducedto∼90%ofthe
sample.eventoriginalThebeammuonscaninteractsnotonlyinthetarget,butalsowithanyother
materialsurroundingit(air,magnet,coolant).Theseeventshavebeendiscarded
requiringthepositionoftheprimaryvertextobeinsidethetargetvolume.This
cutfurtherreducesthesampleto∼65%.Fig.5.4showsthepositionofthe
reconstructedprimaryvertexandthecutsapplied.Itisusefultorememberthat
thetargetforthemuonprogramismadeoftwocylindricalrodsof6LiDof1.5
cmradiusseparatedby10cm.Thepeaksat∼−150and∼60cminthez1
distributioncorrespondtothewindowsofthetargetspectrometer.
Alsoifthescatteredmuonfromtheprimaryvertexisnotreconstructedthemea-
surementoftheenergytransferredinthescatteringprocessisnotpossible.There-
foreonlyeventswithatleastonescatteredmuonareselected,correspondingto
afurther∼20%reductioninthenumberofevents.
Toeachreconstructedparticlethefollowingadditionalcutsareapplied:

•theparticleisneitheridentifiedasamuonnorasanantimuon;
•thelastpointoftheparticlestrackislocatedatZ>360cm(inorderto
measureitschargewithSM1);
•themomentumoftheparticleis≤140GeV,toexcludeunidentifiedmuons.
the1Inbeam.theTheCOMPxandASSyrefercoorencedinatessystemareinthetheztranscoorversedinatesplanecoincidesrespectintodirtheectionbeam.andversuswith

92

Figure5.4:Theprimaryvertexdistribution.Thezcoordinate(leftplot)isinthe
areinlongitudinalthedirtransverseectiondirrespectection.toThetheredbeamandshadedthearxeaandinythecoorleftdinatesplotand(righttherplot)ed
continuouslineintherightplotshowtheappliedcuts.

5.2.4Λ0andΛ0reconstruction

ThenextstepintheanalysisistheΛ0reconstruction.TheΛ0decaysinpπ−with
a63.9%branchingratio.Thisdecaychannelisbettersuitedtothereconstruction
withrespecttotheotherpredominantchannel(nπ0,35.8%branchingratio)for
thedecayproductsarechargedparticlesandcanbedirectlytrackedinthespec-
trometer(whileneutralparticlesarenot).IfdefiningtheV0asavertexwithonly
twooutgoingparticleswithoppositecharge,thedecayvertexofthe0Λ0isaV0in
theselecteddecaychannel.ThefollowingcutsareappliedtotheV:

•V0isnotclassifiedasprimaryvertex,todiscardΛ0producedintheprimary
interaction;

topology;•ZΛ0≥Zprim,tosearchfortherightevent
•thenormalisedχ2oftheV0issmallerthan4.

OneachoutgoingtrackfromtheV0thefollowingcutsareapplied:

93

•lastwhichpointcrossofthethefirsttrackismagnetlocatedSM1atZand,>360thercm.eforeThistheircutcharselectsgecanthebeparticlesclearly
1.2.2);Sec.(seeedmeasur•theparticleisneitheridentifiedasmuonnorasantimuon;
•themomentumoftheparticleis<140GeV,torejectmuonsfromthebeam
notcorrectlyidentifiedinthemuonfilters(seeSec.1.2.2).

Assumingthetrackcorrespondstoacertainparticle,itispossibletoreconstruct
itsLorentzvectorcombiningthe3−momentumofthetrackasgivenbyCORAL
andthemass.ForthereconstructionoftheΛ0,theprotonmassisassignedtothe
trackofthepositiveparticleandthepionmasstothenegativeone.IncaseofΛ0
themassassignmentsareopposite.WithvµtheLorentzvectorofaparticle

vΛµ0=vpµ+vπµ−(5.6)
andonecanreconstructthemass,thetrackandthemomentumoftheΛ0.

0Λenrichment

TheV0sdonotcomeonlyfromΛ0decay.Theconversionofaphotoninae+e−
pairorthedecayofotherneutralparticlesintwoparticlesofoppositechargealso
formaV0.Sincenodirectparticleidentification(asitcouldbedoneintheRICH)
isusedinthisanalysis,theabove-mentionedprocessesmaybeindistinguishable
fromaΛ0decayandcontributetotheΛ0background.Furthermoreithasbeen
observedthatthecombinatorialbackgroundfortheΛ0variesstronglydepending
onthepositionoftheV0alongtheincomingmuonbeamdirection,ZΛ0.The
backgroundstudiesarepresentedinthissection.
e+e−background.Oneparametertodistinguishbetweenthepπ−fromΛ0decay
ande+e−fromphotonconversionistheangleαbetweentheΛ0momentumvec-
torinthelabsystemandthemomentumvectorofoneofthedaughterparticles
intheΛ0CMS.Fig.5.5showsthecosαdistribution(leftplot).
ThedecayproductsofanupolarisedΛ0areuniformlydistributedinthesolid
angle.Thee+ande−fromtheconversionofarealphotonhaveasmalltrans-
versemomentumandhencearemostlyemittedatsmallangles,thenthepeaks
inthedistributionforcosα±1−comesmorelikelyfromthem.Onecanesti-
matetheeffectsofthecutonthepπinvariantmass(Fig.5.5,right):theblack
histogramisthepπ−invariantmassdistributionwithoutanycutonthecosα;
theredhistogramhasthecut|cosα|<0.9(redareacosineplot,Fig.5.5,left);
94

Figure5.5:Right:cosαdistribution.Left:pπ−invariantmassdistribution.The
continuouscurvesarethefitwithaGaussianfunctionplusapolynomial.For
text.seeexplanation

thebluehistogramhasthecut|cosα|≥0.9(blueareaintheleftplot).0Theblue
histogramdoesnotshowanypeakcorrespondingtoreconstructedΛ.Therefore
theeventswith|cosα|≥0.9canbediscardedwithonlyaverysmalllossinthe
0Λsignal.K0background.TheK0decaysinπ+π−anditisapotentialsourceofback-
groundfortheΛ0reconstruction.Fig.5.6showstheπ+π−invariantmassdis-
tribution(leftplot),withaK0peakat∼0.5GeV.Thepeakhasbeenfitwitha
Gaussianfunctionplusapolynomialforthebackground.ThewidthoftheGaus-
sianisσ=8MeV.
TheeffectofexcludingtheregionoftheK0peakcanbeseeninthepπ−invariant
massofFig.5.6(rightplot):theblackhistogramisthedistributionwithoutany
cut;theredhistogramcontainsalleventsundertheK0peakwithin3σ(redarea
intheleftplot);thebluehistogramcontainsalleventsoutoftheK0peakwithin
3σ(blueareain0theleftplot).Intheredhistogramasmallpeakcorrespondingto
reconstructedΛisstillvisible.Thesignaltobackgroundratioofthislastsample
issmall,moreovertheeventshavenotbeendiscardedsinceaclearselectionof
Λ0fromΞ−decaycanbedone,aswillbeexplainedinSec.5.2.5.
ThelongitudinalpositionoftheΛ0decayvertex,ZΛ0.Ithasbeenobservedthat
thebackgroundintheΛ0reconstructionstronglydependsonthepositionofthe

95

Figure5.6:Right:π+π−invariantmassdistribution.AclearpeakforK0isvisi-
ble.Left:pπ−invariantmassdistribution.Thecontinuouscurvesarethefitwith
aGaussianfunctionplusapolynomial.Forexplanationseetext.

ZΛ0.Fig.5.7showsthepπ−invariantmassspectraforthreedifferentintervals
ofZΛ0.TheleftplotcontainsallΛ0decayinginthemostdownstream0cellofthe
intarthegetormostinthedownstr10cmeambetweencellofthethecells.target.TheThecentralleftplotplotcontainscontainsthetheΛΛ0decayingdecaying
downstreamofthetarget.ApeakattheΛ0massisvisibleinallthreesamples,
butthefirsttwosamplesaredominatedbythebackground.Twotwomostlikely
e:areasonsr

•thecombinatorialbackgroundismuchhigherinthetarget,duetothe
highernumberofsecondaryinteractionsinthetargetmaterial.
•theacceptanceofthetargetmagnet:iftheΛ0decayproductsareemittedat
largeangle,theycanbeabsorbedbythemagnetandnoΛ0reconstructionis
possible;theacceptancedecreaseswithincreasingthedistancebetweenZΛ0
andtheendcupofthemagnet.Adetailedstudycanbefoundin[Wie03]

ThesignaltobackgroundratioforthethreedistributionsinFig.5.7hasbeen
calculated.TheinvariantmassdistributionsarefittedwithaGaussianfunction
plusabackgroundparabolicfunction.ThesignalandthebackgroundforΛ0and
Λ0areestimatedfromthefitparametersandareshowninTab.5.2.Becauseof
theextremelysmallsignaltobackgroundratio,onlyΛ0decayingdownstreamof
thetargethavebeenusedforfurtheranalysis(ZΛ0>30cm).

96

Figure5.7:Thepπ−invariantmassspectraforthreedifferentintervalsofZΛ0:Λ0
decayinginthefirsttargetcell(leftplot),inthesecondtargetcell(centralplot),
outsideofthetarget(rightplot).

00ΛΛSNS/NSNS/N
−30>ZΛ0>−1003×1051.5×1070.023.8×1053.6×1070.01
30>ZΛ0>−301.1×1061.6×1070.070.5×1063.1×1070.02
ZΛ0>302.5×1061.2×1062.091.3×1062.3×1060.5
Table5.2:Signal(S),noise(N)andsignaltonoiseratio(S/N)forΛ0andΛ0.The
firstlinecorrespondstoparticlewhichdecaysinthefirstcellofthetarget,the
secondlinetothesecondcell,thethirdlinetoparticleswhichdecaysoutsideof
get.tarthe

97

FinalresultsAcleanΛ0andΛ0sampleisselectedforfurtheranalysis.Fig.5.8
showsthepπ−andpπ+invariantmassspectrawithrespecttothenominalmass
oftheΛ0.Thecontinuous(black)curvesresultfromafitofaGaussiananda
parabolicbackground.Thedotted(red)linesshowthebackground.Thetotal
numberofΛ0andΛ0eventsforCOMPASS,thecentralvalueandthewidthof
thefitontheirmassdistributionareshowninTab.5.3.

Figure5.8:Thepπ−(left)andpπ+(right)invariantmassspectrarespecttothe
mΛ0PDG.ThelateralverticallinesshowthemasscutappliedintheΛ0π−recon-
uction.str

00ΛΛEvents1245800±2400638000±2200
mass(MeV)1115.390±0.0031115.440±0.005
width(MeV)2.278±0.0042.242±0.007
Table5.3:NumberofΛ0andΛ0inCOMPASSdatafrom2002and2003,thecen-
tralvalueandthewidthofthefitontheirmassdistribution

5.2.5Ξ−andΞ−reconstruction
ThenextstepistheΞ−reconstruction.TheΞ−decaysinΛ0π−witha99.89±
0.035%[PDG00a]branchingratio.Λ0candidateswithmass|mpπ−mΛ0|≤
3∙σ=7MeVareselectedforfurtheranalysis(seeSec.5.2.4).ForΛ0candi-
dateswithinthismasswindowthenominalmassmΛ0PDG[PDG00a]isassumedfor
furtheranalysis.Anadditionalnegativeparticlewiththefollowingcharacteris-
equested:ristics•theparticlehasnotbeenusedforthereconstructionoftheΛ0;
98

Figure5.9:Right:π+π−invariantmassdistribution.AclearpeakforK0isvis-
ible.Left:Λ0π−invariantmassdistribution.Thecontinuouscurvesarethefit
withaGaussianplusalinearfunctions.Forexplanationseetext.

•thetrackisnotconnectedtotheprimaryvertex;
•theclosestdistanceofapproach(CDA)betweentheparticlestrackandthe
Λ0trackis<0.8cm;
•Zprim≤ZΞ−≤ZΛ0

Thepionmassisthenattributedtothisnegativeparticle.IntermsofLorentz
vectors:

vΞµ−=vΛµ0+vπµ−(5.7)
andreconstructsthemass,thetrackandthemomentumoftheΞ−.Inthecaseof
aΞ−,apositivepioniscoupledwithaΛ0.
AninterestingpointconcernstheeffectoftheK0backgroundontheΞ−recon-
struction(Sec.5.2.4).Aspreviouslymentioned,despitebeingasourceofback-
groundfortheΛ0signal,nocutontheK0masshasbeenapplied.
TheleftplotinFig.5.9showstheπ+π−invariantmassdistributionwithapeak
at∼0.5Gev.TheeffectontheΛ0π−invariantmassofexcludingtheregionofthe
K0peakcanbeseeninFig.5.9(rightplot):theblackhistogramisthedistribution

99

withoutanycut;theredhistogramcontainsalleventsundertheK0peakwithin
3σ(redareaintheleftplot);thebluehistogramcontainsalleventsoutoftheK0
peakwithin3σ(blueareaintheleftplot).Inthered−histogramaclearpeakwith
negligiblebackgroundcorrespondingtoreconstructedΞisvisible.Mostlikely,
falseΛ0arerejectedintheanalysiswhenthematchwithanadditionalparticleis
required.NotrejectingtheK0backgroundintheΛ0reconstructionisjustifiedby
analysis.−Ξthe

Figure5.10:TheΛ0π−(left)andΛ0π+(right)invariantmassspectrarelativeto
themΞ−PDG.TheverticallinesshowthemasscutappliedintheΞ−πreconstruction.

Fig.5.10showstheΛ0π−invariantmassspectrawithrespecttotheΞ−nominal
mass.Theblack(continuous)curvesresultfromafitofaGaussianandalinear
background.Thered(dotted)linesshowthebackground.
InviewofthereconstructionoftheexoticΞ−−stateobservedbytheNA49exper-
iment(seeSec.5.1.2),itisinterestingtocomparethetotalnumberofΞ−andΞ−
eventsforNA49andCOMPASSasshowninTab.5.4.FortheCOMPASSdatathe
centralvalueandthewidthofthefitontheirmassdistributionarealsoshown.
InthepresentanalysisoftheCOMPASSdatathenumberofreconstructedΞ−
usedforthesearchoftheΞ−−pentaquarkismorethan10timesthenumberof
reconstructedΞ−inNA49.

5.2.6Ξ0∗,Ξ0∗,Ξ−−andΞ−−selection
Oneofthemoreinterestingpointofthisworkisthesearchoftheexoticbaryon
resonanceΞ−−→Ξ−π−andofitsnon−exoticpartnerintheantidecuplet(Sec.
5.1.1)Ξ30/2→Ξ−π+,inordertoconfirmtheresultsoftheNA49experiment
[na403].Toperformthisanalysis,Ξ−candidateswithmass|mΛπ−mΞ−|≤3∙σ=
10MeVareselected;σiscalculatedfromthemΞ−distribution(Fig.5.10).Tothese
100

Ξ−−Ξ5511640NA49COMPASS17930±21010620±180
mass(MeV)1.32154±4×10−21.32153±5×10−2
width(MeV)3.37±0.043.43±0.06
Table5.4:NumberofΞ−andΞ−inCOMPASSandNA49FortheCOMPASS
datathecentralvalueandthewidthofthefitontheirmassdistributionarealso
shown.

candidatesthemΞ−PDG[PDG00a]isassigned.Anadditionalparticlewiththefol-
equested:rischaracteristicslowing

•theparticlehasnotbeenusedforΛ0andΞ−reconstruction;
•thetrackisconnectedtotheprimaryvertex;sincethesearchedresonances
decayimmediatelyafterproduction,theirproductionanddecayvertices
coincide;tohave

Thepionmassisthenattributedtothisparticle.MatchingtheΞ−withapositive
pion,onereconstructsthespectrumcontainingΞ0∗candidates,withanegative
piontheΞ−−candidates.AnoppositematchingisdonefortheantiparticlesΞ0∗
andΞ−−.Fig.5.11showstheinvariantmassspectraofthesefourstates.No
significantexcessofeventsisdetectedinthemassregionaroundm=1862MeV
whereNA49hasobservedasignal,whileclearpeaksforΞ0∗(1530)anditsan-
tiparticleareseen.ZoomedspectraareshowninFig.5.2.6,showingthemass
regionaroundtheΞ0∗(1530).
Theblack(continuous)curvesresultfromafitofaGaussianandaparabolic
background.Thered(dotted)linesshowthebackground.Thetotalnumberof
reconstructedΞ0∗(1530)andΞ0∗(1530),thecentralvalueandthewidthofthe
fitontheirmassdistributionareshowninTab.5.5.Thephasespaceshown
withthetheNA49signaldoesnotincludetheΞ0∗region(seeFig.5.2),therefore
nocomparisonbetweenthenumberofΞ0∗reconstructedeventsinNA49and
possible.isASSCOMPItisconvenienttointroducethedimensionlessvariablex-FeynmanxFthatquan-
tifies,inthecentreofmassframe,thefractionofbeamsmomentumcontainedin
thelongitudinalmomentumcomponentofthedetectedparticle.xFisdefinedas
[PDG00a]:

101

Figure5.11:Ξ−π+(topleft),Ξ−π−(bottomleft),Ξ−π−(topright),Ξ−π+(bottom
spectra.massinvariantright)

Figure5.12:TheΞ−π+(left)andΞ−π−(right)invariantmassspectra.Thever-
ticallinesshowthemasscutappliedinthefurtherΞ0∗πreconstruction(see
5.2.10).Sec.

102

Ξ0∗Ξ0∗
Events1080±90780±80
mass(MeV)1531.5±0.61530.7±0.8
width(MeV)8.2±0.89.5±1.0
Table5.5:NumberofΞ0∗andΞ0∗inCOMPASS2002and2003data,thecentral
valueandthewidthofthefitontheirmassdistribution

xF2p√zcm(5.8)
swherepzcmisthelongitudinalmomentumoftheparticleintheCentreofMass
System(CMS)oftheprotonpandthevirtualphotonγ∗,ands=(pp+pγ∗)2is
thesquareofthetotalenergyavailablefortheprocess.ThexFisausefulparam-
etertounderstandtheproductionmechanismoftheparticlesinagivenreaction.
IntheCMS,xFcangofrom−1to1.xF1meansthattheparticleproducedin
theinteractiontakesmostofthemomentumofthebeamparticle;thisregimeis
calledcurrentfragmentationregion.xF−1meansthattheparticleisproduced
inthefragmentationofthetargetparticleandonespeaksofatargetfragmentation
region.ThexFdistributionfortheΞ−πstatesinthepγ∗CMSisshowninFig.
5.13.

Figure5.13:ThexFdistributionforΞ−πinthepγ∗CMS.

−−ThecheckΞifπtheΞ−invariant−isseenmasswithadistributionsparticulararekinematicsplottedfor(Fig.differ5.14).entTheintervalssameofisxF,doneto
103

fortheΞ−π+invariantmassdistributions(Fig.5.15).xFgoesfrom-0.1to0.7in
stepsof0.1.NovisibleexcessofeventscorrespondingtotheΞ−−inanyofthexF
observed.isintervals

Figure5.14:Ξ−π−invariantmassdistributionsindifferentintervalsofxF.Start-
ingfromthefirstplotintheupperleftcorner,xFgoesfrom-0.1to0.7instepsof
0.1.

IncaseoftheΞ0∗,itispossibletodeterminethenumberofparticlesineachxF
intervaland,withthehelpofthereconstructionefficiencyfromMonteCarlo,one
cancalculatetheproductioncrosssectionasfunctionofxF.SincetheΞ0∗liesat
∼1530MeVandthemassthresholdonthephasespaceis∼1460MeV,itis
difficulttoestimatethebackgroundcorrectly.IntheΞ−π−invariantmassdistri-
bution,nofurthersignalsareexpected,anditcanbeusedtoestimatetheΞ−π+
phasespaceandthusthebackgroundundertheΞ0∗signal.SincetheΞ−π+and
theΞ−π−distributionshaveadifferentnumberofevents,thelatterhastobe
scaledtobeagoodestimationofthephasespace.Themultiplicationfactorfhas
beencalculatedforeachxFintervalbydividingthetwohistogramsbinbybin
(Ξ−π+/Ξ−π−),asshowninFig.5.16,andtakingtheaveragevaluefexcluding
theΞ0∗peakregion.Theratioofthedistributionsisfittedwithaconstantexclud-
ingtheregionoftheΞ0∗peak.ThevaluesoffforeachxFintervalareshownin
5.6.ab.T

104

Figure5.15:Ξ−π+invariantmassdistributionsindifferentintervalsofxF.Start-
ingfromthefirstplotintheupperleftcorner,xFgoesfrom-0.1to0.7instepsof
0.1.

fxF0.990.0-0.1−1.240.10.0−1.290.20.1−1.380.30.2−1.490.40.3−1.750.50.4−1.950.60.5−2.200.70.6−Table5.6:MultiplicationfactorsffortheΞ−π−invariantmassdistributionin
differentintervalsofxF.

105

Figure5.16:Ξ−π−/Ξ−π+distribution
beenfittedwithaconstantfunction

earconstantthe

ab.Tinshown

5.6.

excluding

inexcluding

ferdifexcluding

106

entthe

intervals0∗Ξ

peak

r

ofegion.

.xFegion.

TheThe

ratiovalues

hasvalues

of

eFigurdistributions

5.17:distributions

Ξπ−

difindistributions

+

dif

(blackentferdif

entferdifindistributions

histograms

have

been

histograms)andΞπ−

(fr-0.1om

-0.1

ed(rintervalsofxF(from-0.1to

scaled

dingaccor

to

107

the

factors

histograms)stepsin0.7

of

Tab.

5.6.

invariantof

0.1).

massThe

edr

eFigur5.18:Ξstepsin0.7to

function.

0∗

invariantof

0.1).

The

massdistributionsdistributions

ear

108

indiffitted

entferwith

a

intervalsGaussian

ofx

F

plus

(fra

om-0.1constant

InFig.5.17theΞ−π+(blackhistograms)andthescaledΞ−π−(redhistograms)
invariantmassdistributionsindifferentintervalsofxFareshown.Thetwodis-
tributionareingoodagreement,exceptfortheΞ0∗peak.Bysubtractingthetwo
distributions(Ξ−π+−Ξ−π−)theplotsinFig.5.18areobtained.TheΞ0∗
peakhasbeenfittedwithaGaussianplusaconstantfunctionforeachinterval.
Theconstantfunctioniszeroforalltheplots,whichconfirmthevalidityofthe
multiplicationfactorf.ThenumberofΞ0∗(within3σ)isplottedasfunctionofxF
(Fig.5.19).ThesharpdecreaseforxF0isduetotheoverallacceptance(ofthe
spectrometerandthereconstruction)whichreducestheefficiencyforparticles
emittedatlargeangle,thuswithsmallpzcm(Eq.5.8).

Figure5.19:NumberofreconstructedΞ0∗asfunctionofxF.

Tosummarise,theanalysisoftheΞ−πchannelhasnotrevealedanyexcessof
eventsattheeffectivemassm=1862MeV,correspondingtothesignalobserved
bytheNA49.ThesameholdsfordifferentxFintervals.Thewellknownreso-
nanceΞ0∗isclearlyseenintheΞ−π+channel.Thesameisvalidfortheantipar-
ticles.

simulationsCarloMonte5.2.7

MonteCarlosimulationsareaverypowerfultooltounderstandtheperformance
ofthedetectors.Thesimulationshavetobecomparedwithrealdata,inorderto
besuretheycorrectlydescribethedetectorbehaviour.MonteCarlosimulations
0∗willhavebebeenexplainedusedininthisthenextanalysistosection).measurInethisthersectioneconstrtheΞuction−andeftheficiencyΞ0∗ofΞdistribu-(as
tionsfromrealdataandMonteCarloarecompared,toverifythevalidityofthe
109

simulations.TheMonteCarlodatasimulatethefollowingdecaychain:

Ξ0∗→Ξ−π+→Λ0π−π+→pπ+π−π+.(5.9)
TheleftplotinFig.5.20showstheΛ0π−invariantmassspectrumfromrealdata
andsimulations.TheMonteCarlohistogramhasbeenscaledtothedatainorder
tohavethesamearea,correspondingtothesamenumberofevents,underthe
twoGaussiansfromdataandsimulations.ThewidthsoftheGaussianarevery
similar,butthebackgroundissensitivelydifferent.
Ausefulquantitytocheckthevalidityofthesimulationsisthemomentumdis-
tributionoftheΞ−.
Underthepeak,bothΞ−andbackgroundeventsarepresentandtheycannotbe
distinguished,buttheirseparatecontributiontothemomentumdistributioncan
beestimatedusingthemethodofsidebandsubstraction:theinvariantmassdistri-
butionisfittedwithaGaussianplusalinearfunctionandthewidthofthedistri-
butioniscalculated;themomentumoftheparticleisplottedfortheeventswith
massat±3σaroundthemeanvalueoftheGaussian(grayshadowedareainthe
leftplotofFig.5.20)andfortheeventsintworegionsoftheinvariantmassdis-
tributionasidethepeakwhereonlybackgroundisexpected(yellowshadowed
areaintheleftplotofFig.5.20).Subtractingthesecondhistogramfromthefirst
oneeliminatesthecontributionofthebackgroundevents.Themomentumdistri-
butionsfromrealdataandsimulationsaredepictedintherightplotofFig.5.20,
andtheyshowasufficientagreementonthequalitativelevelaimedhere.
ThesamemethodhasbeenusedforthereconstructionoftheΞ0∗(Fig.5.21).
Inthiscase,thecombinatorialbackgroundisrelevantalsoincaseofsimulated
eventsandthesidebandsubtractionmethodhasbeenappliedtobothmomen-
tumspectra.SincethepeakoftheΞ0∗liesonthetopofthephasespace,onlyone
backgroundregionhasbeenused.Alsointhiscasethetwomomentumdistribu-
tionsareinsufficientagreement.InFig.5.21thepositionsoftheΞ0∗peakforreal
dataandtheMonteCarlocoincide,butthewidthsaresignificantlydifferent.This
isduetoainadequateestimationofthebackgroundfunctionaroundthepeakre-
gion.Abetterwaytocomparetheinvariantmassdistributionistousethesame
methoddescribedinSec.5.2.6,thatistousetheΞ−π−spectrumtoestimatethe
backgroundoftheΞ−π+invariantmassdistribution.Inthiscasethetwohis-
togramsinFig.5.22forrealdataandMonteCarloareobtained.TheMonteCarlo
histogramisscaledtothedatainordertohavethesamenumberofeventsunder
thepeak.Afullagreementbetweenthetwodistributionisachieved.
FromthewidthoftheΞ0∗distributionσ=9.4±1.3MeVitispossibletoextract
theresolutionoftheCOMPASSspectrometerforΞ0∗events.Thetotalmeasured

110

Figure5.20:Right:Λ0π−invariant
andMonteCarlo(redhistogram).

massumspectromfrdata(blackhistogram)ThecurvesarethefitofaGaussianplusa

linearfunction.Theshadowedareasshowtheregionsusedforthesideband

subtractionofthemomentumdistribution(leftplot).Forexplanationseetext.

eFigur5.21:Right:sameasFig.111

5.20,buteherforΞ+π−

Figure5.22:Ξ−π+invariantmassspectrumfromdata(blackhistogram)and
MonteCarlo(redhistogram)withsubtractedbackground.Thecurvesarethe
fitofaGaussianplusaconstantfunction.

FWHMΓtot2.3∙σistheconvolution
andofthedetectorresolutionΓd.

ofthenaturallinewidthoftheparticleΓpΓtot=Γp2+Γd2(5.10)
ThewidthoftheΞ0∗isΓp=9.1±0.5MeV[PDG00a]andΓtot=22±3MeV.
Then

Γd=Γ2tot−Γp2=20±2MeV.
andforthemassresolutionσ=8.7±0.9MeV.

(5.11)

5.2.8Ξ0∗reconstructionefficiencyandproductioncrosssection

Theconstrructedeconstructionparticlesefandficiencytheoftotalaparticlenumberisoftheparticleratioandbetweengoesthefrom0numberto1.ofItrise-
tionimportantefficiencytofordetermineΞ0∗intheprCOMPoductionASShascrossbeensection.estimated,InthisusingworkthetherMonteeconstrCarlouc-

112

Figure5.23:TheΞ0∗massspectraindifferentxF(from-0.1to0.7instepsof0.1)
intervalsoftheMonteCarlogeneratedevents.

simulations.Thesamplehasbeendividedin8intervalsofxF(from-0.1to0.7in
stepsof0.1)andtheefficiencyineachintervalofxFhasbeenestimated.

InMonteCarlosimulationseachtrackhasanidentifierthatspecifiesthetypeof
particle.ThereforeΞ0∗canbeunambiguouslyidentified.TheΞ0∗spectrafordif-
ferentintervalsofxFareshowninFig.5.23.ThespectraarefitusingaGaussian
functionandthenumberofeventsunderthepeakwithin3σiscountedineachxF
interval.TheΞ−π+invariantmassspectrumhasbeenreconstructed.Theback-
groundhasbeenestimatedwiththehelpoftheΞ−π−invariantmassdistribution
(seeSec.5.2.6).TheinvariantmassdistributionsareshowninFig.5.24.Thepeak
correspondingtotheΞ0∗isfittedwithaGaussianplusaconstantfunction(red
continuousline).TheΞ0∗underthepeakwithin3σarecounted.

Theratiobetweenthenumberofreconstructedeventsandgeneratedeventshas
beencalculatedineachintervalofxFandplottedinFig.5.25.Thesmallvalue
oftheefficiencyforxF∼0isduetotheacceptanceofthespectrometer,caus-
inganotoptimalreconstructionforparticlesemittedatbiganglesrespecttothe
beamdirection.Thepointshavebeenfittedwithapolynomialfunctionoforder
3.Thischoiceofthepolinomialforthefitfunctiondoesnothaveatheoretical

113

Figur-0.1

eto

5.24:0.7

in

TheΞsteps

π

of

+

invariant0.1)

for

massspectrauctedeconstrr

114

inMonte

ferdifCarlo

entintervalsdata.

ofx

F

om(fr

Figure5.25:TheΞ0∗efficiencyasfunctionofxF.

justification,butitisobviouslysufficienttodescribethexFdependencewithin
thestatisticalerrors.Consequently,forfurtheranalysisthemeasuredefficiency
forthedifferentvaluesofxFwasused,insteadoftheinterpolationfromthefit
function.ByknowingtheefficiencyoftheΞ0∗reconstruction,theproductioncrosssection
ofΞ0∗producedinCOMPASSin2002and2003asfunctionofxFcanbecalcu-
lated.ThenumberofreconstructedΞ0∗ineachxFinterval(nr,Fig.5.19)hastobe
dividedforthereconstructionefficiencyηandtheluminosityL.SincetheΞ0∗de-
cayswith100%BranchingRatio(BR)inthechannelΞ−π+,theΞ−with99.887%
BRinthechannelΛ0π−andtheΛ0with63.9%BRinthechannelpπ−[PDG00a],
thesefactorshavetobeincludedtoestimatetheproductionoftheΞ0∗without
channel.decaytheonconstraints

Finally

ddxσF=nr/(η∗1∗0.99887∗0.639∗L∗ΔxF)

(5.12)

Fig.5.26showsthecrosssectionofΞ0∗producedinCOMPASSinthe2002and

115

Figure5.26:ThenumberofΞ0∗producedinCOMPASSasfunctionofxF.

2003asfunctionofxF.ThepointsarefittedwiththefunctionC(1−xF)1.725where
[Ada99].constantaisCTheintegratedcrosssectionfortheproductionofΞ0∗,σ,withxFfrom−0.1to0.7
isobtainedbysummingthebins:

σdσ=ådxFΔxF=71pb.

5.2.9AnupperlimitfortheΞ−−productioncrosssection

(5.13)

In−−ordertodeterminetheupperlimitfortheproductioncrosssectionofthe
ΞinCOMPASS,thesamemethodusedfortheWA89datahasbeenapplied
[wa804].TheΞ−π−invariantmassdistributionhasbeendividedin8binsof
xFfrom-0.1to0.7(Fig.5.27).Basedontheclaimedexperimentalwidthofthe
Ξ−−,Γ<17MeVFWHM[na403],threeregionswithwidth20MeVaround
1.852,1.862and1.872GeVhavebeenselected(greenshadedarea).Thenumber
ofeventsineachregionisni(i=1,2,3).FromafittotheΞ−π−invariantmass
spectrumintheregion1.55−2.6GeV,excludingthesignalregion(continuous
116

Figure5.27:TheΞ−π−InvariantmassspectraindifferentxFintervals(from-0.1
to0.7instepsof0.1).

redline),theexpectedbackgroundpereachbinbi(i=1,2,3)hasbeencalcu-
lated.The3σlimitnmaxforthenumberofeventsinanhypoteticalpeakoverthe
backgroundineachxFintervalis:

(5.14)

√nmax=maxi=1,2,3(max(0,ni−bi)+3bi)(5.14)
andarelistedinthesecondcolumnofTab.5.7.
FromthesenumberstheupperlimitoftheΞ−−differentialproductioncrosssec-
tiondσ/dxF(weightedwiththeeventualbranchingratioBR(Ξ−−→Ξ−π−))has
isformulaThecalculated.been

−−−−nmax
BR(Ξ→Ξπ)∙dσ/dxF=LηBR(Λ0→pπ−)BR(Ξ−→Λ0π−)ΔxF(5.15)
whereL=2240pb−1istheluminosity,ηthereconstructionefficiency,BR(Λ0→
pπ−)=0.639andBR(Ξ−→Λ0π−)=0.999thedecaybranchingratios.Since
117

MonteCarlosimulationsforaresonancewiththecharacteristicsofΞ−−propa-
gatedintheCOMPASSspectrometerarepresentlynotavailable,theefficiency
measuredfortheΞ0∗hasbeenused(Sec.5.2.8).However,nosignificantdiffer-
enceisexpected,sincetheΞ0∗andtheΞ−−haveasimilardecaychannel.
TheresultsarelistedinthethirdcolumnofTab.5.7.
xFnmaxΔxF∙BR∙dσ/dxF[pb]
6.08170.0-0.1−2.31300.10.0−340.20.11.10−1.40430.30.2−1.18380.40.3−1.25200.50.4−0.50130.60.5−0.4890.70.6−Table5.7:numberofeventsnmaxin3σandthecorrespondinglimitsonthedif-
ferentialcrosssectionforΞ−−productionfordifferentxFintervals.

Thelimitfortheintegratedproductioncrosssectionhasbeencalculatedbysum-
mingquadratically−−the−contributions−ΔxF∙BR∙dσ/dxF.TheresultisBRσmax≤
7pbintheΞ(1860)→ΞπdecaychannelwithxFfrom-0.1to0.7.

5.2.10Ξ0∗πselection
ThislaststudyfollowedthepreliminaryanalysisdonebyNA49inthechannel
Ξ0∗πwhereasignalatamass1855MeVintheinvariantmassspectrumwas
[KA04].observedΞ0∗candidateswithmass|mΞ0∗−mΞ0PDG∗|≤2∙σ=0.017GeVareselectedfor
furtheranalysis;σiscalculatedfromthemΞ0∗distribution(Fig.5.2.6).
Anadditionalparticlefulfillingthefollowingrequirementsisthenrequested:

•theparticlehasnotbeenusedforΞ0∗,Λ0andΞ−reconstruction;
•thetrackisconnectedtotheprimaryvertex,sincethesearchedresonances
oduction;prafterimmediatelydecay

0∗Theitiveorpionmassnegativeispion.attributedThetosamethisparticle.matchingisTheΞdoneisforthenthematchedantiparticleswithΞ0a∗.pos-In
118

Figurtom

eright)

5.28:Ξ0∗π+
(topinvariant

mass

0Ξπ∗left),spectra.

+

(bottom119

left),Ξ0π∗

(topright),Ξ0π∗

(bot-

Fig.5.28theinvariantmassspectraofthesefourstatesarepresented.Nosignifi-
cantsignalatm=1855MeVisobserved.Applyingthesamemethoddiscussed
in0∗Sec.−5.2.9theupperlimitforthenumberofeventshasbeencalculated.Forthe
Ξcombination:π√nmax=maxi=1,2,3(max(0,ni−bi)+3bi)=56(5.16)
SincenoMonteCarlosimulationsarepresentyavailableforthisstudy,theupper
limitforthecrosssectionhasnotbeencalculated.

resultstheofDiscussion5.2.110∗ΞofReconstructionWiththestudyofΞ0∗,COMPASScancontributetoclarifyitsproductionmecha-
nismandpropertiesinγ∗Ninteractions.
0∗PASSAbout1data.000ΞUsing(Tab.Monte5.5)Carlohavebeensimulatedreconstrdata,uctedthefrcromossthesection2002andforΞ02003∗prCOM-oduc-
tionasfunctionofxF,dσ,intherange[-0.1;0.7]hasbeenmeasured(Fig.5.26)
andddxσ∝(1−xF)1.725.dxFTheintegratedcrosssectionforxFintherange[−0.1;0.7]
hasbeenFcalculatedanditisσ=71pb.
Itisinterestingto0∗comparetheseresultswithotherexperiments,inordertobetter
understandhowΞareproduced.
TherearenostudiesfortheproductioncrosssectionofΞ0∗inphoto-production
experiments,whileitisstudiedinhadro-production.Forcomparison,theresults
oftheWA89experimentarepresentedhere[Ada99].
WA89measuredtheσΞ0∗inΣ−Nreactionswith0<xF<1.TheddxσFbehaveslike
C(1−xF)1.265.ThedifferentialcrosssectioninWA89goeslessrapidlytozerofor
−xFparticle,→1thatthanincontainsCOMPalrASS.eadyaThisisstrangeduetoquark:theiffacttheΞthat0∗isWA89formedusesinΣtheascurrbeament
atedtofragmentationformtherΞegion0∗.(xThisF∼1),phenomenononlyoneiscalledadditionalleadingstrangeparticlequarkeffect.hastoCOMPbecrASS,e-
0∗formedinstead,outusesofthemuonsquarkasbeamsea,makingparticlestheandprtheocessinstrangecurrentquarksinthefragmentationΞhasrtoegionbe
obable.prless21One8µbcanforalso0<comparx<e1the.Fromintegratedthecrplotsossinsection:[wa804]WA89itwasmeasuredextractedσΣ−Nthe→Ξ0∗Xvalue=
F120

σΣ−N→Ξ0∗X∼170µbintheregion0<xF<0.7tobecomparedwiththeCOM-
PASSvalueσµN→Ξ0∗X=44pbforthesamexFregion;theydifferbyaboutsixor-
dersofmagnitude.Twopointscanclarifythisdifference:thefirstoneisthemen-
tionedleadingparticleeffect,whichpredictsahigherproductionrateofstrange
particlesinthecurrentfragmentationregionifthebeamparticlecontainsalready
astrangequark.Thesecondandmostsignificantoneisthetypeofinteraction.
Forthecrosssectionσ∝α2whereαisthecouplingconstantoftheinteraction.
Theγ∗Ninteractioniselectromagneticandα=1/137,whiletheΣ−Ninterac-
tionisstrongandα=αs∼1fortheconsideredenergy.Thismeansthatthecross
sectionσΞforaΣbeamis∼104higherthenforaphotonbeam.Theresultsfor
WA89andCOMPASSaresummarisedinTab.5.8.

ExperimentReactionnσΞ0∗production
COMPASSγ∗N1.72544pbphoto-production
WA89Σ−N1.265170µbhadro-production
Table5.8:ComparisonbetweenCOMPASSandWA89experimentsintheΞ0∗
reconstruction.nistheexponentoftheC(1−xF)nwhichwasusedtofittheddxσF
distribution.

Tosummarise,itisnotpossibletoquantitavelycomparetheΞ0∗productioncross
sectioninCOMPASSandWA89duetotheirdifferentproductionmechanisms.
However,ithasbeenshownthattheyarequalitativelyconsistent.

ofReconstruction−−Ξ

Thesearchforamanifestlyexoticpentaquarkcandidate,Ξ−−,followstheresults
oftheNA49experimentwhichclaimedasignalinthedecaychannelΞ−−→
Ξ−π−→Λ0π−π−→pπ−π−π−ata−mass−m−=1+862MeVinppinteractions
[na403].TheinvariantmassspectraforΞπ,Ξπandtheirchargeconjugate
areshowninFig.5.29.
Theanalysisofthesamedecaychaininthe2002and2003COMPASSdatadoes
notshowanypeakstructureintheΞ−π−channel,andinanyofthechargecon-
jugatechannels,asitcanbeobservedinFig.5.30.Thesameanalysisindifferent
intervalsofxFintherange−0.1<xF<0.7didalsoshownon-confirmative
rlated,esults.BR(MorΞ−−eover→,Ξan−π−)upper∙σmaxlimit≤of7pb.theproductioncrosssectionhasbeencalcu-
NA49observedtheΞ−−−−candidateinhadro-production,howeverthereisnofun-
damentalreasonwhytheΞshouldnotbeobservedinphoto-production.Itis

121

eFigurserved

5.29:by

Invariantthe

NA49

massspectraexperiment.

forΞ122

π

,

−Ξ

π

+

,

Ξ

+

π

andΞ+

π

+

asob-

Figurright)

π−Ξ5.30:emassinvariant

+

(topmass

left),Ξspectra.

π−

(bottom123

left),Ξ−

π

(topright),Ξ−

π+

(bottom

worthwhiletoremarkthatalsotheWA89[wa804]andHERA-B[HER04]exper-
imentshaveinvestigatedtheΞ−−→Ξ−π−channelwithnegativeresults.The
maincharacteristicsoftheseexperimentsarelistedinTab.5.9.
ExperimentReaction√s(GeV)productionΞ−−observation
NA49pp17.2hadro-productionyes
COMPASSγ∗N∼18photo-productionno
WA89Σ−N∼26hadro-productionno
HERA-BpN41.6hadro-productionno
Table5.9:ComparisonbetweendifferentexperimentswhosearchedfortheΞ−−

candidate.pentaquark

Withthisinmind,onecanmakesomequantitativecomparisonbetweenthere-
sultsofNA49andCOMPASS.NA49observed∼36eventsoverthebackground
atm∼1862GeV[na403]intheΞ−π−channeland1640Ξ−,thenΞ−−/Ξ−=
36/1640=0.022.Inthepresentanalysis17930−−Ξ−−eventswerereconstructed.
AssumingthesamerelativeproductionrateΞ/Ξ,fromthepurecomparison
ofthenumberofevents,COMPASSshouldthenobserve17930∙0.022∼394Ξ−−
events(seeTab.5.4).ThestatisticalsignificanceoftheCOMPASSresultsisthata
peakat1862MeVwithmorethan45eventscanbeexcludedwith2σC.L..This
meansthatthenonobservationofanexpectedsignalof364eventscannotbedue
toexplainstatisticalthispeculiaructuationsbehaviourwellofabovethe95%claimedC.L..Ξ−−Otherrpentaquarkeasonshavecandidate.tofoundto
−−Tomentsconclude,didnottheconfirmstatustheoftheΞobservationsearofchisNA49prinesentlyphoto-andpuzzling.hadro-prSeveraloduction.experi-
IfsultsotherachievedhighinstatisticsthisanalysisexperimentswithwouldCOMPASSconfirmdatathemayNA49pointtoobservation,anexotictheprre-o-
ductiondynamicsofthishypothesisedpentaquark.

124

OutlookandConclusions

ThisworkwasdoneintheframeworkoftheCOMPASSexperiment.

Inthefirstpart,thedesignandtherealisationofasiliconmicrostripdetector
systemoperatedatcryogenictemperaturehasbeendiscussed.Silicondetectors
areusedinCOMPASSforthereconstructionofbeamparticleupstreamofthe
target.Topreventthedegradationofthedetectorsperformancesduetohigh
radiationdosestowhichtheyareexposed,theyareoperatedatcryogenictem-
perature.ThisfeatureiscalledLazaruseffectandconsistsintherecoveryofthe
ChargeCollectionEfficiency(CCE)ofheavilydamagedsilicondetectorswhen
operatedatcryogenictemperature,withamaximumoftheCCEatT∼130K.
TheCOMPASSsilicondetectorscanbeusedfortheentirelifetimeoftheexperi-
mentifoperatedatcryogenictemperatures.Inordertooperateasilicondetector
atcryogenictemperatures,acoolingsystemhasbeendesigned.Itconsistofatiny
(1.3mminnerdiameter)tubeinthermalcontactwiththedetector,inwhichliquid
nitrogenuxes.Thetemperatureatthedetectorcanbevariatedbychangingthe
nitrogenuxinthetube.Thetermalinsulationofthedetectorfromtheouteren-
vironmentisachievedoperatingitinsideanevacuatedcryostat.Thecryostathas
beendesignedwithasmallradiationlengthinthebeamdirection,tominimise
itsinterferencewiththebeamparticles.Coolingthesilicondetectorimpliesthat
itsmechanicalsupport,whichcontainsalsothefirstcomponentofthereadout
chain,iscooledaswell.Thegluesusedinthisareahaveprovedtomantaintheir
propertyatlowtemperatures.ThereadoutchipAPV25,whichislocatedcloseto
detector,hasnotshownanydramaticchangeofitsfeatures,inspiteofnotbeing
designedforlowtemperatureoperation.

inDuringtheCOMP2003ASStwosiliconexperiment.detectorsNeverthelessoperatedaatstable130Kweroperationecouldsuccesfullybeachievedinstalled
onlyforfewdays,duetodiscontinuousmethodofprovidingtheliquidnitrogen
uidwithnitraogen,self-prwhichessurisedcanprdewarovide.Athemorecoolantsophisticatedwithoutanydistributioninterruptionsystemtoforseveralliq-
detectors,hasbeendesignedandtestedandwillbeinstalledinthenearfuturein
ASS.COMP

125

Thesecondpartofthisworkhasbeendevotedtothesearchofthemanifestly
exoticpentaquarkcalledΞ−−withmassm=1862MeVusingthe2002and2003
COMPASSdata.AhintforaΞ−−signalhadbeengivenbytheNA49exper-
imentinthedecaychannelΞ−−→Ξ−π−→Λ0π−π−→pπ−π−π−.NA49
claimedtheΞ−−inhadro-production,insteadCOMPASSprovidesdatainphoto-
production.TheonlyotherpentaquarkcandidateΘ+hasbeenobservedinboth
hadro-andphoto-productionexperiments,withpreferencetothelatterone,and
thereisnotheoreticalreasonforadifferentmechanismproductionfortheΞ−−.
FurthermoretheproductionofthestrangeparticleΞ0∗hasbeenstudied.These-
lecteddecaychannelisΞ0∗→Ξ−π+→Λ0π−π+→pπ−π−π+.Sincethedecay
channelsforΞ0∗andΞ−−aresimilar,theirstudyhasbedonedoneinparallel.
Atfirst,areconstructionmethodforΛ0,Ξ−andΞ0∗hasbeendevelopped,lead-
ingtoagoodsignaloverbackgroundforallthreeparticles.Thentheproduction
characteristicsoftheΞ0∗hasbeenstudied.Forthisparticlethereconstructionef-
ficiencyhasbeenevaluatedusingMonteCarlosimulation.Theproductioncross
sectionasfunctionofxF(dσµN→Ξ0∗X/dxF)hasbeenmeasuredforxFintherange
[−0.1;0.7].AstrongenhancementinthenumberofproducedΞ0∗canbeob-
servedfornegativevalueofxF,whichpointstoadominantcontributionofΞ0∗
fromthetargetfragmentationregion.Theintegratedproductioncrosssectionσ
hasbeencalculatedσ=71pb−1.
ThesameanalysismethodhasbeenappliedtosearchfortheΞ−−.TheΞ−π−
invariantmassspectrumdoesnotexhibitanyvisibleexcessofeventsatm∼
1862MeVinanyintervalofxF.Itisworthwhiletomentionthatotherexper-
iments[wa804][HER04]haveanalysedtheΞ−π−invariantmassspectrumin
hadro-productionandnoneofthemcouldconfirmthefindingsofNA49.TheΞ0∗
efficiencyhasbeentakenasestimatetocalculatetheupperlimitoftheproduction
crosssectionforΞ−−.TheresultisBR(Ξ−−→Ξ−π−)σmax≤7pb.
InordertoconfirmorrefusetheexistenceofaΞ−−pentaquark,moremeasure-
mentsareneeded.Whileexperimentswithhighenergyhadronbeamshaveob-
tainedresultswhicharedifficulttoimprove,onemightenvisagetheuseoflow
energybeams(γorkaons)tosearchforΞ−−.Assuchexperimentsareplannedfor
Θ+,theiroutcomewillhavestronginuenceonthefuturemeasurementplans.
Incaseofapositiveandconclusiveobservation,itisalsoenvisageabletomea-
surethewidth,thespinandtheparityofsuchastate,inordertohaveabetter
understandingofexoticsproductionand,moreingeneral,oftheQCD.

126

Bibliography

[Abt96]

[Abt98]

[Abt99]

[Ada99]

[Ber97]

[Bic72]

[Bij]

o][Br

[CAS]

[Cas01]

[CLA03]

[CM84]

[COM]

I.Abtetal.IrradiationTestsofDouble-SidedSiliconStripDetectors
withaSpecialGuardRingStructure.IEEETrans.Nucl.Sci43,1113–
(1996).1118

I.Abtetal.GluingSiliconwithSilicone.Nucl.Instr.andMeth.inPhys.
Res.A411,191–196(1998).

I.Abtetal.Doublesidedmicrostripdetectorsforthehighradiation
environmentintheHERA-Bexperiment.MPI-PhE99-05(1999).

−∗rM.I.eactionsatAdamovich345etGeV/c.al.PrEur.Phys.oductionJ.Cof11Ξ,r271–278esonances(1999).inΣinduced

SINTEFElectronics&Cybernetics,R.W.Bernstein,andT.West-
gaard.TechnicalSpecificationsandTestSpecificationsforMPIHERA-
(1997).SiliconDouble-sidedB

H.Bichsel.Passageofchargedparticlesthroughmatter.Mcgraw-Hill,3nd
edition,1972.

http://hsi.web.cern.ch/HSI/s-link/.Bij.dervanE.

Bronkhorst:High-TechB.V.ThermalMassFlowMetersandCon-
trollersforGases:InstructionManual.

.elcome.html.web.cern.ch/castor/Whttp://castor

L.Casagrandeetal.Reviewonthedevelopmentofcryogenicsilicon
detectors.Nucl.Instrum.Meth.A461,150–154(2001).

S.Stepanyanetal.hep-ph/0307018(2003).

P.D.B.CollinsandA.D.Martin.HadronInteractions.Hilger,2nd
1984.edition,

http://wwwcompass.cern.ch/

127

[COM96]

[DIA03]

[DVC]

[Ehl02]

[ELMa]

[ELMb]

[EMC88]

[Erl95]

[Fau04]

[Fuc04]

[GLA64]

[GPD02]

[gra01]

u01][Gr

[h1]

TheCOMPASScollaboration.AproposalforaCommonMuonand
ProtonApparatusforStructureandSpectroscopy.CERNSPSLC96-
(1996).14

V.V.Barminetal.hep-ph/0304040(2003).

N.dHose.FeasibilityStudyofDeeplyVirtualComptonScattering
usingCOMPASSatCERN.Inpreparation.

J.Ehlers.PerformanceoftheTripleGEMDetectorsintheCOMPASSEx-
periment.Diplomathesis,Universit¨atHeidelberg,September2002.

http://atlas.web.cern.ch/Atlas/GROUPS/DAQTRIG/DCS/LMB/SB/index.html

http://pccosrv1.cern.ch/compass/detector/dcs/elmb.html

TheEuropeanMuonCollaboration.Ameasurementofthespinasym-
metryanddeterminationofthestructurefunctiong(1)indeepinelas-
ticmuon-protonscattering.Nucl.Phys.B206364(1988).

J.andErlertheandCDFP.topLangackerquark.candidates.ImplicationsPhys.ofRevhigh.D52pr,ecision441–450(1995).experiments

P.Fauland.TheCOMPASSExperimentandtheRICH-1Detector.Phd
thesis,Universit¨atBielefeld,March2004.

A.ExperimentFuchs..SetupDiplomaforalowthesis,TUtemperaturM¨eunchen,siliconJanuarydetectorfor2004.theCOMPASS

C.J.GlassbrennerandG.A.Slack.ThermalConductivityofSiliconand
Germaniumfrom3KtotheMeltingPoint.Phys.Rev.4A134,1058–1069
(1964).

MichelGarcon.Anintroductiontothegeneralizedpartondistribu-
(2002).hep-ph/0210068tions.

V.Granataetal.CryogenicTechnologyforTrackingdetectors.NIMA,
197-199461(2001).

B.Grube.ThetriggercontrolsystemandthecommonGEMandSilicon
readoutfortheCOMPASSexperiment.Diplomathesis,TUM¨unchen,
2001.December

P.Paganoetal.PreliminaryMeasurementsofTransversityinCOM-
eparation.prInASS.P

128

[H104]

[HER98]

[HER03][HER04]

[Ilg03]

[jaf04]

[KA04]

[Kuh01]

[Kuh02][Lam04]

[Laz]

[Laz98]

[Leb02]

[Leo94]

[LEP03]

[Les75]

A.Aktasetal.Evidenceforanarrowanti-charmedbaryonstate.hep-
(2004).ph/0403017

A.g1(p)withAirapetianapureteal.hydrMeasurogentarementget.ofthePhys.prLett.otonB442spin,str484–492ucture(1998).function

A.Airapetianetal.hep-ph/0312044(2003).
K.T.Knoepeetal.SearchforΘ+Ξ3−/2−pentaquarksinHERA-B.hep-
(2004).ph/0403020

Ch.Ilgner.FertigungundInbetriebnahmeeinerStrohdriftkammersta-
tionfrdasCOMPASS-Experiment.Phdthesis,Ludwig-Maximilians-
2003.chMarMnchen,Universitt

R.L.Jaffe.Exotica.hep-ph/0409065(2004).

M.Karliner.http://gate.hep.anl.gov/berger/Aspen04/Prog04/Karliner.pdf

atR.COMPKuhn.ASS.SimulationsDiplomaforthethesis,MeasurTUM¨ementunchen,oftheNovemberPolarizabilities2001.ofthePion

P.Pagano.PreliminarymeasurementofTransversityatCOMPASSexperi-
ment.Phdthesis,Trieste,November2002.
A.Ferreroetal.Λ0andΛ0polarizationatCOMPASS.XLIIInterna-
tionalWinterMeetingonNuclearPhysicsFebruary2004.

V.Ereminetal.Inpreparation.

V.Palmierietal.Evidenceforchargecollectionefficiencyrecovery
inheavilyirradiatedsilicondetectorsoperatedatcryogenictemper-
atures.Nucl.Instr.andMeth.inPhys.Res.A413,475–478(1998).

M.Leberig.DasCOMPASS-TriggersystemzurMessungdesGluonbeitrags
zumProtonspin.Phdthesis,JohannesGutenberg-Universit¨atMainz,
2002.

W.R.Leo.TechniquesforNuclearandParticlePhysicsExperiments.
1994.edition,2nd,Springer

T.Nakanoetal.Phys.Rev.Lett.91,012002(2003).

A.Lesnicketal.Observationofadifferencebetweenpolarizationand
analyzingpowerinΛ0productionwith6GeV/cpolarizedprotons.
Phys.Rev.Lett.35,770(1975).

129

[M2]

[Mal96]

[MM]

[Moi02]

[Mol99]

[na403]

[NA5]

[Nag05]

[NBW][OC04]

oup/NewM2/main.htmlhttp://sl.web.cern.ch/SL/eagr

G.Mallot.TheSpinStructureoftheNucleonfromtheSMCExperiments.
Habilitationsschrift,JohannesGutenberg-Universitt,August1996.

G.tivePuillgasetal.detectorwithDevelopmentmicrofomesh(InFrMICROMEGAS,ench).anovelDAPNIA-SED-00-01-Tpositionsensi-.

M.A.Moinesteretal.Firstobservationofdoublycharmedbaryons.
(2002).hep-ex/0212029

M.fectsMoll.andmacrRadiationoscopicprDamageopertiesin.PhDSiliconthesis,ParticleUniversitDetectors¨at-micrHamburg,oscopic1999.de-

C.Altetal.ObservattionofanExoticS=-2,Q=-2BaryonResonance
inProton-ProtonCollisionsatCERNSPS.Phys.Rev.Lett.92-042003,
(2003).441–450

http://na50.web.cern.ch/NA50/collaboration.NA50

T.Nagel.Diplomathesis,TUM¨unchen,2005.Inpreparation.

http://www-physics.lbl.gov/˜spieler/HeidelbergNotes/pdf/IIIElectronicNoise.pdf
C.Bernetetal.TheGluonPolarizationΔGGatCOMPASS.XXXIXth
RencontresdeMoriond:QCDandHighEnergyHadronicInteractions
2004.chMar

[PDG00a]D.E.Groometal.Reviewofparticlephysics.Eur.Phys.J.C15(2000).

[PDG00b]G.P.Yostetal.Reviewofparticlephysics.Phys.Lett.B204,1(2000).

[Pei92]A.Peisert.SiliconMicrostripDetectors.DELPHI92-143MVX2(1992).

[PER75]M.L.Perletal.Phys.Rev.Lett.35,1489(1975).

[Per82]D.H.Perkins.IntroductiontoHighEnergyPhysics.Addison-Wesley,
1982.edition,2nd

[PHA]http://ges.home.cern.ch/ges/phast/

[Pol97]V.PetrovD.DiakonovandM.V.Polyakov.Exoticanti-decupletof
baryons:Predictionfromchiralsolitons.Z.Phys.A359,305–314(1997).

[pol02]M.Moinesteretal.PionpolarizabilitiesatCERNCOMPASS.hep-
(2002).ex/0206038

130

[Ray01]M.Raymondetal.TheAPV250.25µmCMOSReadOutChipforthe
(2001)..rackerTCMS[RD39]RD39collaboration.RadiationHardPositionSensitiveCryogenicSil-
iconDetectors:TheLazaruseffect..PhysicaB.
[RD39w]RD39collaboration.http://www.cern.rd39/
[ROSE]RD48collaboration(ROSE).http://rd48.web.cern.ch/
[Rug99]G.Ruggiero.Resistenzaallaradiazionedirivelatorialsiliciooperantiatem-
peraturecriogeniche.Diplomathesis,UniversitadeglistudidiNapoli,
1999.[SAP03]J.Barthetal.hep-ph/0307083(2003).
[Sau97]F.Sauli.GEM:Anewconceptforelectronamplificationingasdetec-
tors.Nucl.Inst.MethodsA386,531–534(1997).
[Sch03]L.Schmittetal.TheDAQoftheCOMPASSexperiment.13thIEEE
ReakTimeConference,Montreal(2003).
[See04]D.Seeharsch.AMicrocontrollerbasedSteeringMechanismfortheCOM-
PASSSiliconDetectorCoolingSystem.Diplomathesis,TUM¨unchen,
2004.comunication.privateKuhn.R.[SHP][SMC99]G.Abriletal.Spindependentstructurefunctiong1:Recentexperi-
mentalresults.Nucl.Phys.Proc.Suppl.71,100–108(1999).
[Tak03]N.Takabayashi.Polarizedtargetforthemeasurementofthegluoncon-
tributiontothenucleonspinintheCOMPASSexperiment.Phdthesis,
2003.chMarNagoya,[TOTEM]TOTEMcollaboration.http://totem.web.cern.ch/Totem/
[wa804]M.I.Adamovichetal.SearchfortheexoticΞ−−Resonancein340
GeV/cΣ−-NucleusInteraction.hep-ph/0405042(2004).
.wessingtoncryogenics.co.ukhttp://www[wes][Wie03]M.Wiesmann.ASiliconMicrostripDetectorforCOMPASSandaFirst
MeasurementoftheTransversePolarizationofΛ0-Hyperonsfromquasi-real
Photo-production.Phdthesis,TUM¨unchen,2003.

131

Acknowledgements

DuringmyPh.D.manypersonssupportedmeandcollaborated,inmanydiffer-
entways,totherealisationofthisworkandIwanttothankallofthem.
IfirstwishtothankmysupervisorProf.StephanPaulforhisguidance.Histrust
andencouragement,theautonomyheleftmeandthediscussionswehadhave
providedmewithmotivationtopursuemygoals.
IwanttothankAntonelloEspositoandJanFriedrich.Undertheirscientificguid-
anceIwasintroducedtothemarvellousworldofcryogenicsandsilicondetectors
andofthephysicsanalysis,respectively.Fromthecriticalandconstructivecom-
mentsofJanandhisprecioussuggestionsmuchofthisworkisindebt.Antonello
hasalsotheresponsibilityofbringingmehereandIcanfinallythankhimforthat.
Thanksalottoallpastandpresentmembersofthesilicongroup:MichaelWies-
mann,RobertWagner,AnnemarieFuchs,MatthiasBeckerandalltheworking
studentwithwhomIworked,shouted,butespeciallyhavegreattimeinMunich
CERN.atandThanktoIgorKonorov,forallwhatIlearnedfromhimintheseyears,electroni-
callybutnotonly.AspecialthankfortheArbat.
ManymanythankstotherestoftheMunichCOMPASSgroupforthehelpwith
allesm¨ogliches.AspecialmentionisforBorisGrube,RolandKuhn,(without
themIwouldbestillscreamingatmycomputer),LarsSchmitt(foransweringto
allpossiblequestions),HeinzAngererandBernhardKetzer.
ThankstoIgorAltarev,UweTrinksandtheTritronH¨uttefortheirgreatideasand
thepatiencetheyhadwithmines.
ThanksalottomyofficemateDanieleTortorella,whostoodmeduringthemost
stressingtimeofthiswork.
ThankstoFrauFrankforthehelpwithalltheGermanpaperswhopassedinmy
years.fourthisinhandsIwanttothankalsoalltheE18fortheniceatmosphereinourgroup,forhaving
funtogetherandforlaughingatallmyjokes.

133

DuringmythesesCERNbecameawellfamiliarplace;Imetthereseveralex-
ceptionalpersonsfromwhomIhavelearnedmuch:Iwanttothankthewhole
COMPASScollaboration,JanMcGillandTapioNiinikoski.

AlandbigIhavesupportbenefitcamefrinomthesethehelpyearsandfromthemywarmthfriends:oftheashistoricnewcomerinfriendsaforIlefteign
inNapoliandthefantasticonesImethere.ThankstoRosannaandPeppefor
alwaysbeentherewhenIneededtobecherished.ThankstoLaura,forherdeci-
siveroleintechnical,jobandemotionalbusiness.ThankstoAprajitaforteaching
meenglish(andhowtowriteaPhDthesis)andwarmingmeinthecoldGerman
years.

Dedicoquestatesiallamiafamiglia:mamma,papa,Marco,TataezioAdriano,
pertuttolaffettochemihannodato,peravercredutoinmeeperaversempre
appoggiatolemiescelte,nonultimaquestaavventuratedesca.

Und,herzlich,Christophder,trotzeinesganzenOzeans(undmehr),mirHeit-
erkeitgegebenhatumdieseArbeitfertigzumachen.

134

ContributionsOwn

Ihavespentthebiggestpartoftimeduringmythesisinthedesign,develop-
mentCOMPandASSsiliconinstallationgroup,oftheIhavecryogenicbeeninvolsiliconvedinsystem.allDuephasestoofthethesmallsizedevelopmentofthe
uction.constrandOneofmymaintaskshasbeenthedesign,productionandtestingofthecryo-
genicinfrastructure.Thisincludedcontacttodifferentworkshopsandprivate
companies.SincetheCERNsecuritypolicyrequireseverycryogenicandunder
pressurconcernedetheinstallationpreparationtorespectofthercertainequiredsecurityrdocumentationequirements,toaobtainpartoftheTmy¨UVworkap-
provalforthecryogenicinfrastructure.Partofthisworkhasbeencompletedby
.BeckerMatthiasIhavealsocontributedtotheassemblyandtestingofthefirstdetectorprototypes.
Thiselectronicsincludedtoconnectedlearntothebondingdetectorortechnologies,theAPV25todebugandtothepartparticipateofthetorseveraleadout
CERN.atbeamstestTsemblyogetherofwithsiliconMichaeldetectorsWandiesmannhaveIhavecontacteddeveloppedandinstrtheuctedpraocedurprivateeforthecompanyas-
forstartingamassproductionofthedetectors.
theirSincethelocationsiliconinthedetectorsCOMPASShadtobeexperiment,installedIandparticipatedeinstalledseveraltimeseveryinyearthisfrpromo-
e.cedurIIorwasganisedinvolvedinthealsoininstallationtheofplanningtheanddetectorsschedulingandofcryogenicthesiliiconnfraprstructoject.ureInin2003the
ICOMPsupervisedASSseveralexperiment.workIparticipatestudentstowhotheCOMPcollaboratedASSinshifts,thesiliconpartlyasprshiftoject.leader.
Startingfromtheendof2003IstartedtheanalysisoftheCOMPASSdata.Inthe
frameworkofthePHASTanalysispackageIimplementedC++basedroutines
forandtheDr.JandedicatedFriedrichphysicsImeasuranalysis.edtheUndercrossthesectionsupervisionfortheofprProf.oductionStephanofΞ0∗Paulin

135

in

section

osscr

the

for

limit

upper

an

oductionmuon-pr

calculated

and

−−

some

international

conferences.

136

ASSCOMPthetoworkmyofesultsrtheesentedprhaveI.Ξcandidatetaquarkcollaboration

and

pen-

the

of