Searching for Supersymmetry with Tri-Leptons and Missing Transverse Energy Using the CMS Experiment at the Large Hadron ...

MISSING IMAGE

Material Information

Title:
Searching for Supersymmetry with Tri-Leptons and Missing Transverse Energy Using the CMS Experiment at the Large Hadron Collider
Physical Description:
1 online resource (167 p.)
Language:
english
Creator:
Skhirtladze, Nikoloz
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Physics
Committee Chair:
KORYTOV,ANDREY
Committee Co-Chair:
MITSELMAKHER,GUENAKH
Committee Members:
MATCHEV,KONSTANTIN TZVETANOV
RAY,HEATHER
RAKOV,VLADIMIR ALEK SANDROVICH

Subjects

Subjects / Keywords:
cmscharginoneutralino
Physics -- Dissertations, Academic -- UF
Genre:
Physics thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
A new era in collider physics hasarrived with the launch of the Large Hadron Collider (LHC) located at theborder of Switzerland and France. Proton beams of center-of-mass energy of 8 TeV provides a great opportunity to exploit new physics beyond the Standard Model, the theory which has proved to be very successful at low energies over the past several decades. At the energy scale available at the LHC, it is possible to further our understanding of the nature of physics and attempt to answer the long-awaited question of whether new physics exists at TeV scale.The work presented here is the effort to search for supersymmetry via the direct electroweak production of chargino and neutralino using the tri-lepton plus missing energy final state at the Compact Muon Solenoid (CMS) experiment. The 19.5 fb-1 data produced by the LHC during the 2012 run year at the center-of-mass energy of 8 TeV allows us to probe multiple supersymmetry scenarios in the context of R-parity conserving Minimal Supersymmetry Standard Model (MSSM). Supersymmetry is a leading candidate for beyond the Standard Model of the particle physics.  Here, we target supersymmetry scenarios where the colored superpartners are assumed to be heavy (and decoupled). We consider the possibility of light sleptons, giving rise to pp?W*? ? ~1± ? ~20? l+l-l±?l?~10 ? ~10, where the decays to leptons are mediated by intermediate on-shell sleptons. We alsoconsider the possibility that the sleptons are heavy, leaving only three-body decays through Standard Model W’s and Z’s. An extensive overview of the data-driven methods used to model the behavior of background processes is given as well. This analysis focuses on this type of direct electroweak supersymmetry production, and we set further constraints on the masses of chargino and neutralino particles.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Nikoloz Skhirtladze.
Thesis:
Thesis (Ph.D.)--University of Florida, 2013.
Local:
Adviser: KORYTOV,ANDREY.
Local:
Co-adviser: MITSELMAKHER,GUENAKH.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0046250:00001


This item is only available as the following downloads:


Full Text

PAGE 1

SEARCHINGFORSUPERSYMMETRYWITHTRI-LEPTONSANDMISSINGTRANSVERSEENERGYUSINGTHECMSEXPERIMENTATTHELARGEHADRONCOLLIDERByNIKOLOZSKHIRTLADZEADISSERTATIONPRESENTEDTOTHEGRADUATESCHOOLOFTHEUNIVERSITYOFFLORIDAINPARTIALFULFILLMENTOFTHEREQUIREMENTSFORTHEDEGREEOFDOCTOROFPHILOSOPHYUNIVERSITYOFFLORIDA2013

PAGE 2

c2013NikolozSkhirtladze 2

PAGE 3

ACKNOWLEDGMENTS IwouldliketothankmyadvisorAndreyKorytovandco-advisorGuenakhMitselmakherforallthesupportandencouragementovertheseyearsofmygraduatestudies.Iwouldliketothankmycommitteemembers:HeatherRay,KonstantinMatchev,andVladimirRakovfortheirtimeandinterestinoverseeingmyresearchactivities.ItwasaprivilegetomovetoCERNformyfurtherresearchandtohavetheopportunitytoworkandshareideaswithsomanyinspirationalpeople.IwouldparticularlyliketothankDidarDobur,RonnyRemington,MingshuiChen,andLesyaShchutskaforthequalitydiscussionsthathelpedmetounderstandtheconceptoftheexperimentalhighenergyphysicsandtheirinuenceovermyacademicdevelopment.IwouldliketoextendaspecialthankstoDidarDoburforpushingmetoachievemyfullpotentialandforherongoingsupportthroughoutmyresearch.Finally,Iwouldliketodrawattentiontothecontinuoussupportofmyfamilyandfriendswithoutwhichthisstudywouldnothavebeenpossible. 3

PAGE 4

TABLEOFCONTENTS page ACKNOWLEDGMENTS .................................. 3 LISTOFTABLES ...................................... 7 LISTOFFIGURES ..................................... 9 ABSTRACT ......................................... 13 CHAPTER 1INTRODUCTION ................................... 15 2THESTANDARDMODELOFPARTICLEPHYSICS ............... 17 2.1OverviewoftheStandardModel ....................... 17 2.1.1Fermions ................................ 17 2.1.2GaugeBosons ............................. 18 2.1.3HiggsBoson .............................. 19 2.2BeyondtheStandardModel .......................... 20 3SUPERSYMMETRY ................................. 22 3.1MotivationofSupersymmetry ......................... 22 3.2TheMinimalSupersymmetryStandardModel ................ 23 3.3SimpliedSupersymmetryModels ...................... 26 4THELARGEHADRONCOLLIDER ......................... 28 4.1DescriptionoftheLargeHadronAccelerator ................. 28 4.2PerformanceoftheLHCforthe2010-2012runperiod ........... 30 5THEDESCRIPTIONOFTHECOMPACTMUONSOLENOIDDETECTOR .. 32 5.1OverviewoftheCMSDetector ........................ 32 5.2CMSCoordinateSystem ........................... 33 5.3CMSSolenoid ................................. 33 5.4CMSTrackingSystem ............................. 34 5.4.1CMSSiliconPixel ............................ 35 5.4.2CMSSiliconStrip ............................ 36 5.5CMSElectromagneticCalorimeter ...................... 37 5.6CMSHadrinocCalorimeter .......................... 40 5.7CMSMuonSystem ............................... 42 5.7.1DriftTubeChambers .......................... 43 5.7.2CathodeStripChambers ........................ 44 5.7.3ResistivePlateChambers ....................... 45 5.8CMSTriggerandDataAquisitionSystem .................. 46 4

PAGE 5

6OBJECTRECONSTRUCTIONANDIDENTIFICATION .............. 50 6.1MuonReconstructionandIdentication ................... 50 6.2ElectronReconstructionandIdentication .................. 51 6.3JetReconstructionAlgorithm ......................... 52 6.4Particle-FlowAlgorithm ............................ 53 6.5Jetandb-TaggedJetReconstructions .................... 53 6.6TauReconstructionandIdentication ..................... 55 6.7MissingTransverseEnergyReconstruction ................. 57 7EVENTSIMULATION ................................ 59 7.1OverviewofEventSimulation ......................... 59 7.1.1MonteCarloSimulation ........................ 59 7.1.2PartonDistributionFunctionsandtheHardProcess ........ 60 7.1.3PartonShowersandHadronization .................. 60 7.1.4EventGenerators ............................ 61 7.2DetectorSimulation .............................. 62 7.3FullandFastSimulationsoftheCMSdetector ............... 62 8SEARCHFORSUPERSYMMETRYINEVENTSWITHTRI-LEPTON,MISSINGENERGYANDB-TAGGEDJETVETO ....................... 64 8.1Introduction ................................... 64 8.2ObjectSelectionCriteriafortheAnalysis ................... 67 8.2.1MuonSelection ............................. 67 8.2.2ElectronSelection ........................... 68 8.2.3SelectionofTauLepton ........................ 69 8.2.4Jetandb-TaggedJetSelections ................... 69 8.3DataSamples .................................. 70 8.4TriggerSelectionandEfciencies ....................... 70 8.5EventSelectionandDenitionofAnalysisVariables ............ 73 8.5.1EventSelection ............................. 73 8.5.1.13-leptonSelection ...................... 73 8.5.1.2LowInvariantMassVeto .................. 75 8.5.1.3b-TaggedJetVeto ...................... 75 8.5.1.4LargeMissingEnergy .................... 77 8.5.2VariableDenitions ........................... 77 8.5.2.1DenitionofDileptonInvariantMass ............ 79 8.5.2.2DenitionofTransverseMass ................ 81 8.5.3DenitionofBaselineandSearchRegions .............. 83 8.6BackgroundEstimationMethods ....................... 85 8.6.1DeterminationofBackgroundduetoNon-PromptLightLeptons .. 87 8.6.2DeterminationofBackgroundduetoNon-PromptElectrons .... 90 8.6.3DeterminationofBackgroundduetoNon-PromptTauLeptons ... 94 8.6.4DeterminationofBackgroundduetoWZprocess .......... 101 5

PAGE 6

8.6.4.1TheRecoilMethod ...................... 103 8.6.4.2BinnedGaussianMethod .................. 107 8.6.4.3ApplyingtheRecoilCalibration ............... 108 8.6.4.4LeptonEnergyScaleandResolution ........... 109 8.6.4.5CorrectionstotheNormalization .............. 112 8.6.4.6SystematicUncertaintiesAssociatedforthisMethod ... 112 8.6.5DeterminationofBackgroundduetoPhotonConversion ...... 113 8.6.6BackgroundduetoRareSMprocesses ............... 115 9FINALRESULTS ................................... 117 10SIGNALACCEPTANCE ............................... 126 11SIGNALACCEPTANCEUNCERTAINTIES .................... 134 11.1TheoreticalUncertainty ............................ 134 11.2LuminosityUncertainty ............................. 134 11.3TriggerUncertainty ............................... 134 11.4LeptonReconstructionEfcienciesandAssociatedUncertainties ..... 134 11.5MissingTransverseEnergyandTransverseMassAcceptancesandAssociatedUncertainties .................................. 136 11.6b-TaggedJetUncertainty ........................... 137 12INTERPRETATIONOFRESULTS ......................... 142 13CONCLUSIONS ................................... 150 APPENDIX ACOMPARISONOFTHERESULTSBETWEENATLASANDCMS ....... 151 BANEXCESSINTHEON-ZREGIONOFTHEOSSFSELECTION ....... 155 REFERENCES ....................................... 163 BIOGRAPHICALSKETCH ................................ 167 6

PAGE 7

LISTOFTABLES Table page 3-1SummaryofsupermultipletsintheMSSM .................... 25 6-1Branchingfractionofthedecayofthetaulepton ................. 57 8-1CMS-recordeddatausedintheanalysis ...................... 70 8-2MonteCarlosamplesusedintheanalysis ..................... 70 8-3Asetoftriggerswhichareusedintheanalysis .................. 73 8-4DenitionofelectronIDsattheHLTlevel ..................... 73 8-5Overalltriggerefciencies .............................. 73 8-6DenitionofsearchregionsforOSSFeventselection .............. 84 8-7DenitionofsearchregionsforNoOSSFeventselection ............. 84 8-8DenitionofsearchregionsforSSTaueventselection .............. 84 8-9ClosuretestinthebaselineselectionfordifferentOSSFchannels ....... 90 8-10Closuretestforthepredictionofthebackgroundduetonon-promptelectronsarising ......................................... 95 8-11ClassicationofSSTaueventsinthedifferentMCsamples ........... 97 8-12Summaryofnon-promptleptonoriginsinsimulatedttevents .......... 97 8-13Summaryofnon-prompttauoriginsinZ+jetscontrolsample ......... 98 8-14Electronenergyscaleandresolutioncorrections ................. 111 8-15ExpectedcontributionstothedifferentrareSMprocessesinthebaselineselection 116 9-1ThesummaryoftheobservedandpredictedbackgroundyieldsfortheOSSFselection ....................................... 122 9-2ThesummaryoftheobservedandpredictedbackgroundyieldsfortheNoOSSFselection ....................................... 124 9-3ThesummaryoftheobservedandpredictedbackgroundyieldsfortheSSTauselection ....................................... 125 A-1DenitionofthesearchregionsusedintheATLASanalysis ........... 154 A-2AsummaryofpredictedandobservedyieldsinallsixsearchregionsusedbyATLAS ......................................... 154 7

PAGE 8

B-1Thesummaryof26eventsfortheOSSFselectionintheon-ZandhighMT(>160GeV)regionwithEmissT>50GeV ...................... 156 8

PAGE 9

LISTOFFIGURES Figure page 2-1Thebuildingblocksofmatterandtheparticlesresponsibleforthefundamentalinteractions ...................................... 20 4-1ThefullviewoftheLHCacceleratorcomplex ................... 29 4-2Evolutionoftheintegratedluminosityforthe2010-2012runperiod ....... 31 5-1ThefullviewoftheCMSdetector .......................... 32 5-2SuperconductingSolenoidduringtheassemblyoftheCMSdetector ...... 34 5-3FourlayersoftheCMSsolenoid .......................... 35 5-4r-zviewoftheCMStracker ............................. 36 5-5Simulationrepresentinganelectromagneticshower ............... 37 5-6ThefullviewoftheCMSelectromagneticcalorimeter .............. 39 5-7TheviewoftheCMShadroniccalorimeter ..................... 42 5-8TheviewoftheCMSmuonsystem ......................... 43 5-9Schematicdesignofadrifttubecell ........................ 44 5-10LayoutoftheCMSbarrelmuonDTchambers ................... 45 5-11Schematicdesignofacathodestripchamber ................... 46 5-12ArchitectureoftheCMSL1system ......................... 48 8-1Feynmandiagramsforthechargino-neutralinoproductionandtheirdecaymodes 66 8-2Dileptontriggerefciencies ............................. 74 8-3b-taggedjetvetoefciencyinthesignalsamples ................. 76 8-4Thebackgroundexpectationforb-taggedjetmultiplicityusingsimulation .... 77 8-5DistributionofEmissTforOSSFeventsinthesignalsamples ........... 78 8-6DistributionofEmissTforthe3-leptonselectionsintheSMMCprocesses .... 79 8-7DistributionofM``forthe3-leptonselectionsintheSMMCprocesses ..... 81 8-8DistributionofMTforthe3-leptonselectionsintheSMMCprocesses ..... 82 8-9DistributionofMTusingOSSFeventsinthesignalsamples ........... 83 8-10DistributionoftheobservedyieldsintheMTandM``space ........... 85 9

PAGE 10

8-11Tight-to-LooseobtainedusingdataandQCDMCsamples ........... 89 8-12ClosuretestobtainedinttMCsample ....................... 91 8-13Isolationdistributionofelectronswhichdescendfromdifferentmotherjetsinthecontrolsample .................................. 93 8-14Tight-to-LooseobtainedinthecontrolsampleusingdataandQCDMCsamples 94 8-15Originsofelectrons ................................. 95 8-16DistributionofTightelectronsinthecontrolsampleusedtoestimatethesystematicuncertaintyofthemethod .............................. 96 8-17Tight-to-Looseofthetauleptonfordifferentsourcesinthecontrolsample ... 98 8-18LoosetauleptonpTdistributionindifferentMCsamples ............ 99 8-19Tight-to-LooseforofthetauleptonasafunctionofpTand .......... 100 8-20Tight-to-Looseforthetauleptonasafunctionofanumberofvertices ..... 101 8-21Tight-to-LooseforthetauleptonasafunctionofpTinbinsofandthenumberofvertices ....................................... 102 8-22Closuretestfornon-prompttauleptoninthettMCsample ........... 103 8-23Closuretestfornon-prompttauleptonintheZ+jetsMCsample ........ 104 8-24Gaussianttotherecoilcomponents ....................... 106 8-25ResponseandresolutionoftherecoilcomponentsindataandZ+jetMCsamples ........................................ 108 8-26ResponseandresolutionoftherecoilcomponentsintheZ+jetsandWZMCsamples ........................................ 109 8-27InvariantmassdistributionforZcandidatesindataandMC ........... 111 8-28Feynmandiagramforthephotonconversion .................... 115 8-29MCanddatacomparisonforinternalandexternalconversions ......... 115 9-1DistributionofMTinthebaselineselection .................... 117 9-2DistributionofMTinthebaselineselection .................... 118 9-3PredictedandobservedyieldsfortheOSSFselectionasafunctionofEmissT .. 119 9-4PredictedandobservedyieldsfortheNoOSSFselectionasafunctionofEmissT 120 9-5PredictedandobservedyieldsfortheSSTauselectionasafunctionofEmissT .. 121 10

PAGE 11

10-1Signalacceptanceefciencyfortheavor-democraticchargino-neutralinoproductionwithx~`=0.95 .............................. 126 10-2Signalacceptanceefciencyfortheavor-democraticchargino-neutralinoproductionwithx~`=0.5 ............................... 127 10-3Signalacceptanceefciencyfortheavor-democraticchargino-neutralinoproductionwithx~`=0.05 .............................. 128 10-4Signalacceptanceefciencyforthetau-enrichedchargino-neutralinoproductionwithx~`=0.95 ..................................... 129 10-5Signalacceptanceefciencyforthetau-enrichedchargino-neutralinoproductionwithx~`=0.5 ..................................... 130 10-6Signalacceptanceefciencyforthetau-enrichedchargino-neutralinoproductionwithx~`=0.05 ..................................... 131 10-7Signalacceptanceefciencyforthetau-dominatedchargino-neutralinoproductionwithx~`=0.5 ..................................... 132 10-8Signalacceptanceefciencyfortheavor-democraticchargino-neutralinoproductionwithWZ+LSPinthenalstate .................... 133 11-1Fittothedielectroninvariantmassdistributions .................. 136 11-2Reconstructionefciencyforelectrons ....................... 137 11-3Reconstructionefciencyformuons ........................ 138 11-4DistributionofM``fordifferentb-taggedjetrequirements ............. 141 12-1Amapofthemostsensitivesearchregionsfortheavor-democraticscenario 144 12-2Amapofthemostsensitivesearchregionsforthetau-enrichedscenario ... 145 12-3Amapofthemostsensitivesearchregionsforthetau-dominatedandavor-democraticscenarios ................................ 146 12-4ConstraintsoncharginoandLSPmassesfortheavor-democraticscenario .. 147 12-5ConstraintsoncharginoandLSPmassesforthetau-enrichedscenario .... 148 12-6ConstraintsoncharginoandLSPmassesforthetau-dominatedandavor-democraticscenarios ................................ 149 A-1Ratioofthechargino-neutralinoproductioncross-sectionsusedbyATLASandCMS ....................................... 152 A-2ComparisonoftheexclusionlimitsbetweenCMSandATLASfortheavor-democraticscenariowiththexparameterof0.5 ................. 153 11

PAGE 12

A-3ComparisonoftheexclusionlimitsbetweenCMSandATLASfortheavor-democraticscenariodecayingviaWandZgaugebosons ........... 153 B-1ComparisonofEmissTshapebetweenthesimulationandthedata-correctedsimulationmethod .................................. 157 B-2DistributionsofEmissTandMTfortheOSSFselectionintheon-Zregion .... 158 B-3Distributionoftheleptonvariables ......................... 159 B-4ElectronmultiplicityandasumofchargesforOSSFselectionintheon-Zregion 160 B-5JetmultiplicityandHTdistributionfortheOSSFselectionintheon-Zregion .. 160 B-6b-taggedjetmultiplicityfortheOSSFselectionintheon-Zregion ........ 161 B-7DileptonmassdistributionfortheOSSFselectionintheon-Zregion ...... 161 12

PAGE 13

AbstractofDissertationPresentedtotheGraduateSchooloftheUniversityofFloridainPartialFulllmentoftheRequirementsfortheDegreeofDoctorofPhilosophySEARCHINGFORSUPERSYMMETRYWITHTRI-LEPTONSANDMISSINGTRANSVERSEENERGYUSINGTHECMSEXPERIMENTATTHELARGEHADRONCOLLIDERByNikolozSkhirtladzeDecember2013Chair:AndreyKorytovMajor:PhysicsAneweraincolliderphysicshasarrivedwiththelaunchoftheLargeHadronCollider(LHC)locatedattheborderofSwitzerlandandFrance.Protonbeamsofcenter-of-massenergyof8TeVprovidesagreatopportunitytoexploitnewphysicsbeyondtheStandardModel,thetheorywhichhasprovedtobeverysuccessfulatlowenergiesoverthepastseveraldecades.AttheenergyscaleavailableattheLHC,itispossibletofurtherourunderstandingofthenatureofphysicsandattempttoanswerthelong-awaitedquestionofwhethernewphysicsexistsatTeVscale.Theworkpresentedhereistheefforttosearchforsupersymmetryviathedirectelectroweakproductionofcharginoandneutralinousingthetri-leptonplusmissingenergynalstateattheCompactMuonSolenoid(CMS)experiment.The19.5fb)]TJ /F9 7.97 Tf 6.58 0 Td[(1dataproducedbytheLHCduringthe2012runyearatp s=8TeVallowsustoprobemultiplesupersymmetryscenariosinthecontextofR-parityconservingMinimalSupersymmetryStandardModel(MSSM).SupersymmetryisaleadingcandidateforbeyondtheStandardModeloftheparticlephysics.Here,wetargetsupersymmetryscenarioswherethecoloredsuperpartnersareassumedtobeheavy(anddecoupled).Weconsiderthepossibilityoflightsleptons,givingrisetopp!W!~1~02!`+`)]TJ /F7 11.955 Tf 7.08 -4.34 Td[(``~01~01,wherethedecaystoleptonsaremediatedbyintermediateon-shellsleptons.Wealsoconsiderthepossibilitythatthe 13

PAGE 14

sleptonsareheavy,leavingonlythree-bodydecaysthroughStandardModelW'sandZ's.Anextensiveoverviewofthedata-drivenmethodsusedtomodelthebehaviorofbackgroundprocessesisgivenaswell.Thisanalysisfocusesonthistypeofdirectelectroweaksupersymmetryproduction,andwesetfurtherconstraintsonthemassesofcharginoandneutralinoparticles. 14

PAGE 15

CHAPTER1INTRODUCTIONAtheorycalledtheStandardModelofparticlephysics,thatexplainselementaryparticlesandtheirinteractions,hasbeendevelopedoverthepastseveraldecades.ThistheoryprovedtoberobustandsuccessfulatthelowenergyregimeandrepeatedlyrewardedwithNobel-prizes.However,thereareseveralkeyquestionsthatareleftunansweredbytheStandardModel(section 2.2 );therefore,lookingfortheanswersbeyondtheStandardModelseemsaverynaturalthingtodo.OneparticularcaseofphenomenabeyondtheStandardModelistheexistenceofsupersymmetry,thetheorywhichsetsaconnectionbetweenfermionandbosonparticles.Thistheoryiswellmotivatedandremainsfavoredinthephysicscommunityduetoitsbeautifulformalismandpossiblesolutionstothependingquestions.ItisexpectedthatsupersymmetrywillemergeasonebeginstoprobetheTeVenergyscale.TheLargeHadronCollider(LHC)atCERNwasbuilttoexploretheTeVenergyregime.Thisenergyisobtainedbycollidingtwoprotonbeams,producinganunprecedentedcenter-of-massenergyof8TeVin2012runyear.TheCompactMuonSolenoid(CMS)detectorisoneofthemultipurposedetectorsoperatingattheLHC,anditallowsustopreciselymeasurethepropertiesoftheparticlesproducedfromtheprotoncollisions.Therearemanymodelsofsupersymmetry,manifestingindifferentwaysinthenalstate.Inthisdissertation,Iamsearchingthesupersymmetrysignalinanelectroweakproduction,theproductionofchargino-neutralinosupersymmetricparticles.TheStandardModelbackgroundprocessescanbesignicantlyreducedbyspecicallydesignedselections,leavingarathercleansignaturetolookforapresenceofthesupersymmetrysignatureviachargino-neutralinoproduction.Thisworkisorganizedinthefollowingway:Chapter2brieydescribesthetheoreticalfoundationoftheparticlephysics;Chapter3introducesthemotivationandtheoverviewofthesupersymmetrytheory;Chapters4and5giveanoverviewof 15

PAGE 16

theLargeHadronCollider(LHC)andtheCompactMuonSolenoid(CMS),respectively;Chapter6describesthereconstructionalgorithmsfortheobjectswhicharetheprimaryinterestinthisanalysis.Chapter7describeseventsimulation;Chapters8-10arededicatedtothesearchesofsupersymmeteryviachargino-neutralinoproduction,includinganextensiveoverviewofmultiplebackgroundestimationmethods;Chapter11isdedicatedtotheinterpretationoftheresults;Chapter12presentsasummaryofthework. 16

PAGE 17

CHAPTER2THESTANDARDMODELOFPARTICLEPHYSICS 2.1OverviewoftheStandardModelTheStandardModel(SM)ofparticlephysicsisoneofthegreatestachievementsofhumanintelligence.Itcombinesthetheoriesthatdescribethestrongforce,theweakforce,theelectromagnetic(EM)force,andtheirrespectiveparticles.Thisisprimarilydonebyemployingtheoryofgroupsymmetriesunitedbytherequirementoflocalgaugeinvariance.ManyexperimentalobservationswhichhavebeencarriedoutoverthepastseveraldecadesindicatethattheSMisindeedaverysuccessfultheory;and,itisworthnotingthatalltheseexperimentsexploitedlowenergyregimes(O(100GeV)).However,ithastoevolvetodescribephysicsprocessesathigherenergy.TheStandardModeldictatesthatallmatteriscomposedofparticlescalledfermionsandtheforcesbetweenfermionsaremediatedbybosons.AlargeamountofvaluableworkhasbeendoneassociatedwiththeStandardModeloverthepastseveraldecadesandthereforemanypapersandexcellentintroductorybooksareavailable,forexample[ 1 2 ].Here,IdonotattempttogiveadetaileddescriptionoftheStandardModel,butratherabriefoverviewispresented. 2.1.1FermionsMatteriscomposedofspin-1/2fermions.Fermionscanbeorganizedintotwoelementarycategories:quarksandleptons.Botharepoint-likeparticlesthathavenosubstructure;theyaredenedbytheirmassandquantumnumbers(likespin,charge,weakisospin,hypercharge,andcolor).Therealsoexistsananti-fermionforeveryfermion;ananti-particlecanbeunderstoodastheregularparticlewiththesamemassbutoppositequantumnumbers.Leptonsandquarkshavedifferentinteractionnatures:leptonsparticipateinallinteractionsexceptthatwhichisstrong,whereasquarksparticipateinallinteractions. 17

PAGE 18

Fermionsaregenerallyorganizedintothreegenerationswithvaryingmassesbutwithothersimilarproperties,asshowninFigure 2-1A .Ineachquarkgeneration,thereisaquarkwithanelectricchargeof+2/3(inunitsoftheelectroncharge)andaquarkwithachargeof-1/3.Ineachleptongeneration,thereisaverylightneutralleptoncalledaneutrino,andaleptonwith1electriccharge.Inadditiontoeachgeneration,thereexistsacorrespondinggenerationcomposedofantiparticles.Therstgenerationisstable,andallordinarybaryonicmatterismadeoftheparticlesfromtherstgeneration.Thesecondandthethirdgenerationparticlesusuallyliveveryshorthalflivesandcansubsequentlydecayintoparticlesoftherstgeneration.AsummaryofallfermionsareshowninFigure 2-1A .Furthermore,quarkscarrythechargeassociatedwithstronginteractions,calledcolor,andquarksbehaviorisdictatedbythephenomenonofcolorconnement[ 3 ].Thiscolorconnementensuresthattheydonotexistfreeinnature,butareinsteadgroupedwithotherquark(s)insuchawaytocreatecolorlesscompositeparticlescalledhadrons.Asimpleexamplewouldbeprotonsandneutrons.Therearethreedistincttypesofcolorchargeandconsequentlytwodifferentwaystomakeacolorneutralcombination:Baryonsarecomposedofthreequarksorantiquarks(protons(uud)andneutrons(udd)areexamplesofbaryons);Mesonsarecomposedofaquarkandanantiquark,andthecolorcongurationissuchthattheycancel.Itisworthnotingthatbothfermionsfeatureapropertywhichisrelatedtotherightorleft-handedrepresentationofthePoincaregroup[ 1 ].Allthesethreegenerationsdescribedaboveareleft-handed;thereisalsorighthandedrepresentation,butduetoalackofexperimentalevidencethatthereexistsright-handedneutrinoinnature,right-handedfermionsareasingletofSU(2)group. 2.1.2GaugeBosonsForcesamongthefermionsaregovernedviatheexchangeofvirtualgaugebosons,andtheyaredescribedbythesymmetrytransformationsoftheSU(3)CSU(2)L 18

PAGE 19

U(1))]TJ /F1 11.955 Tf 8.64 1.79 Td[(gaugegroups.Forcesworkoverdifferentrangesandhavedifferentstrengthsandcharacteristics.However,theyareallorganizedbyanalogousprocesses.Thephotonistheelectromagneticforcecarrier,itismasslesswithnoelectriccharge.Therangeoftheelectromagneticforceisinniteandfollowstheinversesquarelaw.TheintermediatevectorbosonsWandZ0arethecarriersoftheweakinteraction,whichislimitedatsufcientlyshortdistances(anorderof10)]TJ /F9 7.97 Tf 6.59 0 Td[(16cm)duetothehighmassofthevectorbosons(MW'81GeV,MZ'92GeV).Gluons,likephotons,aremasslessandcoupleonlytothequarks;However,thereisafundamentaldifferencebetweenphotonsandgluons,gluonscarrycolorandanti-colorwhichallowthemtohaveself-coupling.Atsufcientlyshortdistancesthestrongforcebehavesanalogouslytoelectromagnetism;however,afteralimitingdistanceontheorderof10)]TJ /F9 7.97 Tf 6.59 -.01 Td[(16cm,thestrongforceremainsataconstantstrength,independentofthedistancebetweenthequarks.AsummaryofallgaugebosonsareshowninFigure 2-1A 2.1.3HiggsBosonAhypothesizedscalar(spin-0)particlecalledtheHiggsbosonrepresentsthemysteriouspieceoftheStandardModel.Itwaspredictedalmost40yearsagobutithasbeenatremendouschallengetoobservethisparticleatthecolliders.However,duetotheoutstandingperformanceoftheLHC,theCMSandtheATLAScommunitieshaveindependentlyannouncedobservationofthebosonatthemassrangeof125GeV[ 15 16 ],whichhasthepropertiesconsistentwiththeStandardModelHiggsboson.AccordingtotheStandardModeldescription,theSMHiggseldisacomplexscalarHwithclassicalpotential[ 1 ] V=m2HjHj2+jHj4.(2)TheHiggseldacquiresanon-zerovacuumexpectationvalueattheminimumofthepotential,giving=p )]TJ /F4 11.955 Tf 9.3 0 Td[(m2H=2witharequirementthatm2H<0and>0.TheHiggsbosonisaconsequenceofelectroweaksymmetrybreaking,whichisbelievedtoberesponsibleforgeneratingthemassesoftheWandZ0gaugebosons, 19

PAGE 20

leavingthephotonmassless.Thenon-zerovacuumexpectationvalueisresponsibleforgeneratingmassforeveryparticlethatcouplestotheHiggsbosonviaYukawacouplings.ExperimentalmeasurementsoftheweakinteractionsconstrainHiggsmassintheorderof100GeV,whichisconsistentwiththeobservedmass[ 15 16 ]. A BFigure2-1. Thebuildingblocksofmatterandtheparticlesresponsibleforthefundamentalinteractions.A)showsasummaryofallelementaryparticlesintheSM.B)showstherelationshipbetweenfermionsandbosons. 2.2BeyondtheStandardModelTheStandardModel(SM)provedtobeaverysuccessfultheoryofparticlephysicswhichdescribesallthefundamentalparticlesandtheirinteractionswell;however,itisbelievedthattheSMisamanifestationofamoregeneralormorefundamentaltheoryatlowenergy.Therearemanyreasonswhythephysicscommunityhasneverattemptedtoconsideritasthecompletetheory;tomentionafew: Gravity-Asoftoday,gravitationalinteractionisnotconsideredwithintheStandardModelframework.ItsincorporationtotheStandardModelsetsseriouschallenges:rstly,thegravitationalcouplingconstantGisnotdimensionlesslikethecouplingsoftheotherfundamentalforces;andquantumgravityisthusanon-renormalizabletheory;secondly,itisimpossibletotestgravitationalinteractionatthecurrentenergiesavailableatthecolliders. UnicationofCouplings-InthecontextoftheGrandUniedTheory(GUT),thethreegaugeinteractionsoftheStandardModeltendtomergeathighenergies, 20

PAGE 21

muchthesamewaythatelectromagneticandweakinteractionsarecombinedintheStandardModelattheelectroweakscale.TheGUTremainsthefavoredscenariobythephysicscommunity.However,extrapolationoftheStandardModelgaugeinteractions,whicharemeasuredatrelativelylowenergy,doesnotallowthemtomergeatsomehighervalues. DarkMatter-TheWilkinsonMicrowaveAnisotropyProbe(WMAP)indicatesthattheuniverseconsistsof4%ofbaryonicmatter[ 4 ].Theremainingpartisbelievedtobeassociatedwithdarkmatteranddarkenergy[ 4 5 ].TheStandardModelisincapableofdescribingdarkmatteranddarkenergyphenomena. FineTuningProblem-WiththediscoveryofthenewbosonparticleattheLHC,moreeffortsareneededtofurtherourunderstandingaboutthenatureoftheStandardModel.Unfortunately,theStandardModelisincapableoffullyformulatingtheHiggsphenomenology.Inparticular,theSMpredictsthattheHiggsbosonmassissensitivetothequantumcorrections,thusforcingthemassesofallknownSMparticlestoremainsensitiveaswell. 21

PAGE 22

CHAPTER3SUPERSYMMETRY 3.1MotivationofSupersymmetryAllthesependingfundamentalquestionsencouragedmanyphysiciststosearchfortheanswersbeyondtheSM.Overthepastfewdecades,manymodelsweredevelopedasanextensionoftheSM,theleadingcandidatebeingsupersymmetry,thesymmetrybetweenfermionandbosonstates.ItisworthnotingthatthemostcompellingtheoreticaljusticationforthistheoryisrelatedtoanaturalcancellationofthedivergentquantumcorrectionstotheHiggsmass.ThishasbeenanoutstandingissuesincetheHiggsbosonwaspredicted,andthisisuniquetothescalarparticleswithnon-zerovacuumexpectationvalue(VEV).TheHiggsmassissensitivetotheenergyscalecut-off,UV;inotherwords,thisenergyscalecanbeunderstoodasthescaleatwhichnewphysicscantakeplace.TheSMHiggsbosonmassincludingthequantumcorrectionforfermionscanbewrittenas[ 6 7 ] M2H=M20+2UV+(ln),(3)whereM0isthebaremassofHiggsbosonandtheothertermsrepresentdifferentorderquantumcorrections,andwhereisaconstantfactordependingonthecoupling.Theheavierthefermion,thestrongercouplingithastotheHiggsbosonduetotheYukawacoupling;intheSM,theheaviestmassscaleissetbythetop-quark.Onesimpleconclusionfromtheaboveformulaisobvious:theHiggsmassisverysensitivetotheenergyscalecut-off.Ontheotherhand,thisindirectlyleadsallmassesintheSM,whichhavebeenexperimentallymeasured,tobesensitivetothatenergyscaleaswell.However,ifsupersymmetryreallyexistsinnature,thentherewillalsobeacontributionfromthebosonsuperpartnersofthefermions.Thesameformulastillholdsinthiscase,butwithanexceptionthatthesecondtermenterswithaminussign: M2H=M20)]TJ /F7 11.955 Tf 11.95 0 Td[(2UV+(ln),(3) 22

PAGE 23

Thisminussignisresponsibleforthemiraculouscancellationofthedivergentquantumcorrections.Itmustbementionedthatthiscancellationeventakesplaceatanyenergyscale.Thatisanotherreasonwhysupersymmetryisconsideredasaverypromisingmodel.Furthermore,inthecontextofsupersymmetry,unicationoftheStandardModelgaugecouplingsisfeasibleduetosuperpartnersparticipatingintheloop[ 6 ].Supersymmetryshouldbeunderstoodasanextensionoftheusualspace-timesymmetries.ToaccommodateaSMparticle(fermionorboson)anditscorrespondingsuper-partnerintoasingleentity,weconsideranotionofasupermultipletwhichgroupsbothstates.IncontrasttotheLorentzandtranslationtransformations,thegeneratorsofthesupersymmetrytransformationareaspinone-halfMajoranaspinoranddonotcommutewiththeLorentztransformations.Theyactinthisfashion: QjBoson>=jFermion>,QjFermion>=jBoson>.(3)Thesupersymmetryalgebrasatisesthefollowingcriteria: fQ,Qyg=P(3) fQ,Qyg=fQy,Qyg=0(3) [P,Q]=[P,Qy]=0,(3)wherePisthefour-momentumgeneratorofthespace-timetranslation.ItisworthnotingthatQandQycommutewiththegeneratorsofthegaugetransformation,leadingtotheveryimportantcharacteristicthatparticlesinthesamesupermultipletshouldbeinthesamerepresentationofthegaugegroup;inotherwords,theymusthavethesameelectriccharge,weakisospinandcolordegreeoffreedom.Massesofparticlesandtheirsuperpartnersmustalsoequalbeforethesymmetrybreaking. 3.2TheMinimalSupersymmetryStandardModelTheMinimalSupersymmetryStandardModel(MSSM)isconsideredasthesimplestextensionofthestandardmodel.Thismodelcontainsthesmallestnumber 23

PAGE 24

ofsupersymmetricparticlesandassumeslocalsymmetryasintheSM.Inordertobephenomenologicallyviable,itshouldbeconsistentwiththeknownfeaturesoftheStandardModel.AsageneralruleinSUSY,allSMparticleshavecorrespondingsuperpartners.Todistinguishonefromanother,ageneralnotationwasdened:thenamesofthesuperpartnersoftheSMfermionsareprependedbys,whereasthenamesofthesuperpartnersoftheSMscalarandvectorbosonsareappendedbyino.Forinstance,thesuperpartnerofanelectronisreferredtoasselectronandtheHiggssuperpartneriscalledhiggsino.Astheleft-andright-handedcomponentsofthefermionsintheSMaredecoupledfromeachotherduetouniquetransformationproperties,theirbosonsuperpartnersaretreatedseparatelyaswell.Therearetwobasicsupermultiplets:theSMfermionstogetherwiththeirbosonsuperpartnersarearrangedintoachiralsupermultiplet,whilegaugebosonsandtheirfermionsuperpartnersformagaugesupermultiplet.AsummaryofallMSSMparticlesarelistedinTable 3-1 .Thecolumnwithspin-0andspin-1/2belongtothechiralsupermultipletandthecolumnwithspin-1/2andspin-1representthegaugesupermultiplet.ThegauginosectorintheMSSMresemblestheStandardModelinasensethattheSMgaugebosonmasseigenstatesareformedafterthesymmetrybreakingandtheyareamixofthegaugebosoneigenstatesoftheelectroweakinteractionSU(2)LU(1)Y.Similarly,intheMSSM,theneutralhiggsinos(H0uandH0d)mixwiththeneutralgauginos(B0andW0)toformfourmasseigenstatescalledneutralinos,denoted~0i(withindexi=1,2,3,4).Thechargedhiggsinos(HuandHd)combinewiththechargedwinos(W)toformtwomasseigenstatescalledcharginos,denoted~i(withindexi=1,2).Byconvention,thesubscriptorderinghappensinascendingorder;inotherwords,~00isthelightestneutranilo(LSP)and~04istheheaviestneutralino,andthesameappliestochargino. 24

PAGE 25

Table3-1. SummaryofsupermultipletsintheMSSM Supermultiplet spin-0spin-1/2spin-1 squark,quark (~uL,~dL)(uL,dL)~uRuR~dRdR-slepton,lepton (~L,~`L)(L,`L)~`R`R-Higgs,Higgsinos (H+u,H0u)(~H+u,~H0u)(H+d,H0d)(~H+d,~H0d)-gluinogluon -~ggwinos,Wbosons -~W,~W0W,W0binos,Bbosons -~B0B0 TheHiggssectorintheMSSMtheoryrequiresspecialattention.UnlikeintheStandardModel,thepresenceoftwoHiggsdoubletsSU(2)areassumedintheMSSM.Thisisanecessaryrequirementinordertoensurethatelectro-weakinteractionisanomaly-free[ 6 7 ].Furthermore,oneHiggsdoubletisresponsibleforthemassesofquarkswiththecharge+2/3,andthusitislabeledHu,whereasHdisresponsibleforthemassesforquarkswiththecharge-1/3[ 6 7 ].Eachdoubletconsistsofchargedandneutralparticles.AllconservedquantitieswhicharevitalintheStandardModel,likeBaryonandLeptonnumbers,shouldberespectedinthecontextoftheMSSMaswell.ABaryonnumberisproportionaltothedifferenceofthequarkandanti-quarknumberswhilealeptonnumberisdenedasthedifferencebetwenleptonandanti-leptons(L=n`)]TJ /F4 11.955 Tf 12.14 0 Td[(n`).However,inthesimplestformsoftheMSSM,thereisnoconstraintonthesequantities,andthereforeitcanleadtoB-andL-violatingcouplings(forexample,p!e0).Onesolutionistoimposethesymmetry,calledR-parityormattersymmetry,whichisdeinfedas PR=()]TJ /F5 11.955 Tf 9.3 0 Td[(1)3(B)]TJ /F6 7.97 Tf 6.58 0 Td[(L)+2S(3) 25

PAGE 26

ThisistheexactsymmetryanddictatesthattheproductofPRforallinitial-stateparticlesshouldequaltotheproductofPRforthenal-stateparticles.TheSMparticleshavePR=+1,whilesupersymmetricparticleshave-1.Thisfurtherleadstotheveryimportantphenomenologicalandexperimentalconsequences: ThelightestsupersymmetricparticleisstablebecauseitcannotfurtherdecayintotheStandardModelparticlesduetothePRsymmetry;itisoftenreferredtoasLSP.InthedecaychaintheheavierSUSYparticleswilldecaytolightersparticles/particlesandyieldLSPinthenalstate.IftheLSPisaweaklyinteractingneutralparticle,thenitcanbeconsideredasaveryattractivedarkmattercandidate; Atthecolliders,atleastonepairofsupersymmetricparticlescanbeproducedandthereforeatleasttwoLSPsareexpectedinthenalstate(ingeneral,thisisverymodeldependent);and IfLSPisweaklyinteractingparticle,thenitcanpassthroughthedetectorwithoutinteracting,leavingnoindicationofitspassing.ThisisabigadvantageintheSUSYsearchesbecauseLSPwillenhancemissingenergyinthenalstatewhichisagoodhandletosegregatesignalfromthebackground.SincenohintontheexistenceofSUSYparticleshavebeenobservedyetinanycolliderexperiment,itisnaturaltothinkthatsupersymmetryisabrokensymmetry.Thereareseveralmodelswhichaddressthespontaneousbreakingmechanismsofsupersymmetry:supergravity,gauge,andanomalymediation.Moreonthismattercanbefoundin[ 6 ]. 3.3SimpliedSupersymmetryModelsTheMSSMcontainsatotalof124parameters[ 6 ],where9parametersareassociatedwiththegaugesector,5parametersareassociatedwiththeHiggssector,and110parametersareassociatedwiththesoftSUSYbreakingcomponentsoftheLagrangian.Somanyparametersmakethetheoryextremelydifculttoworkwith,and,especially,tomakeagenericstatementabouttheresultofanyparticularsearch.Asignicantreductioncanbeachievableoncetheunderlyingmechanismofsupersymmetrybreakingisunderstood.Overthepastfewdecades,severalmodelshavebeendenedasthebenchmarkpointsforthesupersymmetrysearchesatthe 26

PAGE 27

colliders,incorporatingspecicbreakingmechanismandcertainassumptionsontheproductionanddecayofsparticles,forinstancemSUGRA.Thismodelallowsonly5parametersinsteadof124underthecertainassumptions,but,rstly,itintroducestheoryprejudice,secondly,theCMSandATLASexperimentsexcludedawiderageofmSUGRAphacespace.Therefore,asetofsimpliedmodelspectra(SMS)weredesigned[ 8 ].EachSMSmodelconsistsofasmallnumberofkinematicallyaccessiblesparticlesandaxedsequenceofsparticleproductionandsubsequentdecay.Forinstance,thisworkfocusesonthechargino-neutralinoproduction,giving3-leptonandLSPinthenalstateviasleptonorW=Zgaugebosons.Here,squarksandgluinoareassumedtobeheavyandthereforetheyarenotconsideredingenericLagrangian.Thesemodelsoffertheadvantageofestablishingthesensitivityofanyspecicsearch,andtheyprovideastraightforwardinterpretationwithoutbeingfullymodel-dependent. 27

PAGE 28

CHAPTER4THELARGEHADRONCOLLIDER 4.1DescriptionoftheLargeHadronAcceleratorThelargehadroncollider(LHC)isinstalledundergroundattheborderbetweenFranceandSwitzerland.Itscircumferenceisapproximately27km.Thetunnelwasoriginallybuiltforthelargeelectron-positroncollider(LEP)during1984-1989,witheightstraightsectionsandeightarcs.Itliesbetween45mand170mbelowthesurfaceonaplaneinclinedat1.4%slopingtowardstheGenevalake.Tosavecosts,LHCinheritedthetunnelfromtheLEP[ 9 ].However,twomajorlimitationsimposedsomeconstraintsonthedesignoftheLHC:thelackofavailablespaceinthetunnel,whichhasacylindricalshapewiththediameterof3.7m,andthetargetedhighluminosity.AsolutionwasfoundindesigningtheLHCasaproton-protoncollider,unliketheotherpredecessorproton-antiprotonhadronicaccelerators;itisabigtechnologicalchallengetorunproton-antiprotonbeamsathighluminosities.LHCcanalsooperatewithheavyionbeams(Pb),whichusuallytakesplaceattheendofanoperationalyear.MoredetailsoftheheavyionoperationcanbefoundinReference[ 9 ].Themaximumenergyperprotonbeamcanreachupto7teraelectronvolt(TeV)whichleadstoamaximumthecenter-of-massenergy(ECM)of14TeV.Fortheheavyion,perbeamenergycanreachonly2.8TeVwithEhiCMof5.6TeV.Tokeepthepathof7TeVbunchesofprotonsalongthering,LHCemployssuperconductingdipolemagnetsprovidingamagneticeldupto8.3T.Thishugeamountofmagneticeldisachievedbycoolingthemagnetsdownto1.9K.Thesemagnetsareofatwin-boredesign,allowingtheLHCtocirculatetwoprotonbeamssimultaneously.Intotal,thereare1232ofthesesuperconductingdipolemagnets,spreadalongtheentirecircumferenceofthetunnel.Thetwocounter-rotatingprotonbeamsrunalongtworings.LHCisthelargesthadronacceleratoreverbuiltanditalsoprovidesthehighestcenter-of-massenergy.LHCdesignluminosityis1034cm)]TJ /F9 7.97 Tf 6.58 0 Td[(2s)]TJ /F9 7.97 Tf 6.59 0 Td[(1,andthatisachievedbyapproximately2800injected 28

PAGE 29

Figure4-1. ThefullviewoftheLHCacceleratorcomplex. protonbuncheswiththenominalbunchseparationof25ns.Todate,thisisthehighestluminosityeverachievedatthecolliders.Toaccelerateprotonbeamsuptoitsdesignedenergy,arequiredelectromagneticeldisproducedinfourregionsalongtheringbyradiofrequency(RF)cavities.Thecross-sectionofthebeamalongthering,excepttheregionsnearbytheinteractionpoints,areapproximately200m.Thissuppressionisobtainedbythe858superconductingquadrupolesspreadalongthering.Neartheinteractionpoints,thecrosssectionisfurtherreducedbyaspecialtripletofquadrupoles.Thisisrequiredtomaintainstablebeamsafterthecollisionsattheinteractionpoints.Protonsineachbeamareinitiallyobtainedfromhydrogenatoms.Thehydrogengasisionizedwithanelectriceldafterinjectingitintotheduoplasmatron,leadingtotheseparationofprotonsandelectronsofthehydrogengas.Theprotonsthengothroughachainofdifferentsub-partsoftheLHCbeforeendingupinthemainLHCring.Figure 4-1 showsacompleteschemeoftheprotoninjectionchain.Theparticles 29

PAGE 30

leavingtheduoplasmatronareacceleratedto750KeVandareformedintoasegmentedbeambyaradiofrequencyquadrupole.Forthenextstep,thesegmentedbeammovestothesecondLinac,Linac2,amulti-chamberresonancecavity,wheretheprotonsacquireenergyupto50MeV.TheythenmovetotheProtonSynchrotronBooster(PSB)whichacceleratestheprotonsupto1.4GeV.AfterthattheyareinjectedtotheProtonSynchrotronwithinamicrosecond.Duringthisstep,theprotonsareorganizedintobuncheswithuniformspacingappropriatetotheLHCoperationalscheme.Bunchesareacceleratedupto450GeVandaresenttotheLHCmainring.AdetaileddescriptionoftheLHCmachinecanbefoundin[ 9 ].TherearefourdistinctdetectorsoperatingattheLHCring.Theyare atoroidalLHCapparatus(ATLAS); thecompactmuonsolenoid(CMS); alargeioncolliderexperiment(ALICE)and; thelargehadroncolliderbeauty(LHCb).TheATLASandCMSdetectorscanoperateatthedesignedluminosityoftheLHC.Theyarecompetingmulti-purposedetectorsandhavedifferentdesignsintermsofmagneticeldconguration,electromagneticcalorimetryandmuonsystems.ThedesignoftheATLASdetectorcanbefoundelsewhere[ 10 ].ALICEisspecicallydesignedtostudyheavyioncollisionsatthepeakluminosityof1027cm)]TJ /F9 7.97 Tf 6.59 0 Td[(2s)]TJ /F9 7.97 Tf 6.59 0 Td[(1withabeamenergyof2.8TeV.Itsdesigncanbefoundelsewhere[ 11 ].LHCbisdesignedtostudyheavyavorphysicsandcanonlyoperateatlowluminosityof1032cm)]TJ /F9 7.97 Tf 6.58 0 Td[(2s)]TJ /F9 7.97 Tf 6.59 0 Td[(1.Itsdesigncanbefoundelsewhere[ 12 ]. 4.2PerformanceoftheLHCforthe2010-2012runperiodIttookabout14years(from1994to2008)toconverttheideaofanLHCintothecompleteandoperationalacceleratorcomplex[ 9 ].Therstcollisiontookplacein2008;however,asucceedingdisasterdelayedthedatatakingfortwomoreyears[ 14 ].Withasuccessfulrestartin2010,theLHCperformedgreatlyanddeliveredvaluabledatauntil 30

PAGE 31

2013,whentheupgradeoftheLHCwasscheduled.Forthersttwoyears,2010and2011,theLHCwasrunningwiththecenterofmassenergyof7TeV,halfofitsdesignedcapacity.Withthegradualincreaseoftheinstantaneousluminosityfrom1030cm)]TJ /F9 7.97 Tf 6.58 0 Td[(2s)]TJ /F9 7.97 Tf 6.59 0 Td[(1to21033cm)]TJ /F9 7.97 Tf 6.58 0 Td[(2s)]TJ /F9 7.97 Tf 6.59 0 Td[(1,LHCwasabletodeliverintegratedluminosityofapproximately6fb)]TJ /F9 7.97 Tf 6.59 0 Td[(1.The2012operationalyearwasmoreaggressiveandthereforeagreatsuccessforthehighenergyphysicscommunity.Thecollisionswithacenter-of-massenergyof8TeVandapproximately23fb)]TJ /F9 7.97 Tf 6.59 0 Td[(1deliveredintegratedluminosityallowedCMSandATLASphysicscommunitiestoindependentlyreporttheobservationofabosonresonancewithamassconsistentwith125GeV[ 15 16 ].TheworkinthisthesisisbasedonthedatarecordedbytheCMSdetectoroverthe2012runyearwhichcorrespondstothecertiedintegratedluminosityof19.5fb)]TJ /F9 7.97 Tf 6.59 0 Td[(1.Fig. 4-2 showstheevolutionoftheluminositydeliveredbytheLHC.Itmustbenotedthattheproton-protonrunwasextendedin2012toaccumulatemoredatafortheHiggsbosonsearches. Figure4-2. Evolutionoftheintegratedluminosityforthreeconsecutiverunyears. 31

PAGE 32

CHAPTER5THEDESCRIPTIONOFTHECOMPACTMUONSOLENOIDDETECTOR 5.1OverviewoftheCMSDetectorTheCompactMuonSolenoiddetector(CMS)isoneofthefourdetectorsoperatingattheLHC.Itislocated100metersundergroundclosetotheFrenchvillageofCessy.ItisamultipurposedetectorwhichcanoperateatthedesignedluminosityoftheLHC,1034cm)]TJ /F9 7.97 Tf 6.58 0 Td[(2s)]TJ /F9 7.97 Tf 6.59 0 Td[(11.ThedesignoftheCMSdetectorallowsittoparticipateintheproton-protoncollisionsaswellasintheheavy-ioncollisions.CMShasacylindricalshapewithalengthof21mandadiameterof15m.ThetotalweightofthewholeCMSdetectoris12500tones[ 13 ].Figure 5-1 showsthewholeviewoftheCMSdetector. Figure5-1. ThefullviewoftheCMSdetector[ 13 ]. 1Duringthe2010-2012operation,instantaneousluminositydeliveredbytheLHCincreasedfrom1032to71033cm)]TJ /F9 7.97 Tf 6.59 0 Td[(2s)]TJ /F9 7.97 Tf 6.59 0 Td[(1. 32

PAGE 33

5.2CMSCoordinateSystemAcoordinatesystemoftheCMSdetectoriswell-suitedforacylindricalapparatusoperatinginahadroncolliderenvironment.Thepositivex-axispointstowardsthecenteroftheLHCring,whereasthepositivey-axispointstowardthesky.Thez-axisliesalongthebeampipewiththepositivedirectiontowardtheJuraMountains.Theazimuthalangleismeasuredbetweenthex-axisandthey-axisandiscountedfromthepositivex-axis.Thezenithangleismeasuredbetweenthez-axisandthey-axisandiscountedfromthepositivez-axis.However,inthehadronicacceleratorsangleisnotuseddirectlyfordescribingparticle'strajectory.Instead,thepseudo-rapidityispreferred,whichisdenedas=)]TJ /F5 11.955 Tf 11.29 0 Td[(lntan 2 (5)Thedifferenceinpseudo-rapidity(())betweentwoparticlesemergingfromacollisionrepresentsalongitudinallyinvariantquantity. 5.3CMSSolenoidOneofthemainpurposesofanydetectorinexperimentalhighenergyphysicsistomeasurethemomentumofachargedparticle.Todoso,theparticleshouldbebentwithinanactivematerialofadetector.MagneticeldprovidesthebendingpowerduetotheLorentzforce.Sinceenergeticparticles(pTTeV)areexpectedatthecollisions,ahighmagneticeldisrequiredtoprovidemomentummeasurements.Therefore,CMSusessuperconductingsolenoidgeneratinga4Tmagneticeld.Itspans12.5minlengthand6mindiameter,andithasenoughspacetoaccommodatethewholetrackingsystem,electromagneticandhadroniccalorimetersaltogether.Thetotalweightofthesolenoidis120T.Figure 5-2 showsthesolenoidwithmuonsystemsaroundit.Thesolenoidconsistsofvemodules,andeachmoduleismadeoffourwendinglayers,showninFig. 5-3 ,towithstandthedeformationduetotheenormous 33

PAGE 34

magneticeldgeneratedfromthenominal19kAcurrent.EachlayerismadeofNbTicablesembeddedintoapurealuminumstabilizerandaluminumalloy.Toachievethemaximummagneticeld,theentiresolenoidiscooleddownto4.5K,thepointwhenthesuperconductivityturnson.Thesolenoidalsoservesasanadditionalouterabsorberforthehadroniccalorimeter.MoredetailsoftheCMSsolenoidcanbefoundelsewhere[ 13 ]. Figure5-2. SuperconductingSolenoidduringtheassemblyoftheCMSdetector.Thesoleonoidissurroundedbythemuonsystem.PhotocourtesyoftheCMScollaboration. 5.4CMSTrackingSystemTheCMStrackingsystem,whichisinstalledclosesttothebeampipe,isthekeysub-detectorfortheidenticationofthetracksandmomentummeasurementsofthechargedparticlesastheytraversethedetector.InordertocopewiththewiderangeofCMSphysicsprograms,thetrackingsystemshouldbeabletoidentifylonglivedparticles(b/c-quarks,taulepton),impactparameter,andprimaryverticeswithprecision,aswellasprovidemomentummeasurements.Thehostileenvironmentneartheinteractionpointandtherequirementforprecisemeasurementsputconstraintsonthedesignofthetrackingsystem.Thus,theCMScommunitydesignedthefully 34

PAGE 35

Figure5-3. Fourlayersofthesolenoid.Thewholesolenoidiscoveredbythe50mmmainaluminummandrel.ThepipesoutsideofthemainmandrelensurethecoolingfortheCMSsolenoid[ 13 ].PhotocourtesyoftheCMScollaboration. siliconbasedtrackerwithatotalsensitiveareaof200m2whichspansto5.8minlengthand2.5mindiametertoprovidehighperformance.Itiscomposedoftwoparts:siliconpixelsandsiliconstrips.TheCMStrackingsystemnormallyoperatesatlowtemperature,approximately-10C.Lowtemperatureisanecessaryrequirementtoavoidthermalrunawayforthesiliconsensors[ 13 ].Figure 5-4 showsther-zprojectionoftheCMStrackingsystem. 5.4.1CMSSiliconPixelThesiliconpixelsub-detectoriscomposedofthreebarrellayerslocatedbetween4.4cmand10.2cmfromtheinteractionpoint,andtwoendcapdisksoneithersideofthebarrellocatedatz=34.5andz=46.4cm.Thepixeldetectorcoversthepseudo-rapidityjj<2.5.Thepixelcellsizeof100(r)]TJ /F7 11.955 Tf 12.99 0 Td[()150(z)mallowsittoachievehighgranularity.Highgranularityisanecessaryrequirementtomaintaintheoccupancyatalowrate,allowingittobefunctionalattheenormousradiationuxfromthecollisions,andtoprovideagoodresolution.Thepixeldetectoroccupiesatotalareaofabout1m2andhas66millionpixels. 35

PAGE 36

Figure5-4. r-zviewoftheCMStracker[ 13 ]. 5.4.2CMSSiliconStripThesiliconstriptrackerfollowsthepixelpartandislocatedbetween10cmand116cmintheradialdirection.Itconsistsofthreedifferentsub-systems.Thetrackerinnerbarrel(TIB)residesbetween10cmand55cm,complementedby3trackerinnerdisks(TID)oneitherside.Bothofthemhave320mthicksiliconmicro-strips,whichprovide4hitsinr)]TJ /F7 11.955 Tf 12.53 0 Td[(spaceofthetrajectorymeasurements.InTIB,stripsareparalleltothebeamaxiswithapitchsizeof80monthersttwolayersand120monthelasttwolayers.Theyhaveresolutionsofapproximately23mand35mrespectively.IntheTID,thepitchsizevariesbetween100mand141m.TheTIB/TIDareenclosedbytheoutersilicontracker(TOB)inbarrelandtrackerendcaps(TEC)intheendcapregionswhichspansupto116cmintheradialdirectionandupto2.6minthezdirection.TheTOBconsistsof6layerswithsiliconstripsthicknessof500m.Thepitchsizesvarybetween122mand183mamongdifferentlayers,thusprovidingresolutionsof35mand53m,respectively.TheTECdetectorspansfrom124cmto282cminthezdirectionandfrom22.5cmto113.5cminradialdirection.Itconsistsof9disks,eachwith7ringswithinthesiliconstripdetectors.Thethicknessofthesilicon 36

PAGE 37

stripsvariesbetween300mand520mandthethepitchsizefrom97mto184m,leadingtotheresolutionsofapproximately29mand55m,respectively.Theyprovide9measurementsinr)]TJ /F7 11.955 Tf 11.96 0 Td[(space. 5.5CMSElectromagneticCalorimeterTheelectromagneticcalorimeter(ECal)isusedtomeasuretheenergyofincidentphotonsandelectronsbyabsorbingthem.Theunderlyingprocessesfortheabsorptionaredifferentbetweenelectronsandphotons.Atenergiesabove100MeV,electronsradiatephotonsaroundthenucleusofthematter(bremsstrahlung),whilephotonscreateapairofe)]TJ /F4 11.955 Tf 7.09 -4.34 Td[(e+.Thus,theseprocessesleadtoaso-calledelectromagneticshower.Theshowercontinuesuntiltheenergyofthesecondaryparticlesintheshowerreachesthecriticalenergylevelwherebytheirenergydissipatesbyexcitingorionizingtheatomsofthedetectormediumandnomorefurthershowercandevelop.Figure 5-5 showsapictorialrepresentationofanelectromagneticshower. Figure5-5. Simulationshowinganelectromagneticshowerproducedinthecrystalbyanincidentparticle(photonorelectron). InthemajorityofphysicssearchesattheCMS,thepreciseenergymeasurementofphotonsandelectronsiscrucial.TheCMSelectromagneticcalorimeterismadeoflead-tungstate(PbWO4)crystals.Itiscompletelyhermeticinthedirectionanditisahomogeneoussub-detector2ThedesignoftheCMSelectromagneticcalorimeter 2Inhighenergyexperimentphysics,calorimetersareusuallydividedintotwocategories:homogeneousandsampling.Homogeneouscalorimetersaremadeof 37

PAGE 38

wasmotivatedbyseveralaspects:rstly,therequirementofthehighresolutionfortheenergymeasurements;secondly,smallMoliere3radius(2.2cm)ofthePbWO4crystals,leadingtotheefcientdetectionoftheelectromagneticshower;andthirdlytheabilitytowithstandhighlevelsofradiationoverthefulllifetimeoftheCMSoperation.Thecalorimeterisdividedintotwopartsinregion,barreljj<1.47andendcap1.47
PAGE 39

Figure5-6. ThefullviewoftheCMSelectromagneticcalorimeter[ 13 ].Differentcolorsillustratedifferentsubparts:greenandblueillustratethebarrelregion,greyillustratestheendcapregion,andmagentaillustratesthepreshowersub-detector. Whereas,vacuumphototriodes(VPT)areusedintheendcappart.EachVPTis25mmindiameter,withanactiveareaofapproximately280mm2.CMSemploysapreshowersub-detectormountedinfrontofECalintherangeof1.653and2.6.Itisasamplingdetectorwithtwolayers,andithasthicknessof20cm.Theabsorbersaremadeofleadandtheactivemediumisdesignedusingthesiliconstripssensors.Eachsensor,withthenominalthicknessof320m,isdividedinto32strips,eachwith1.9mmpitch.Thus,thepreshowerdetectorprovidesenoughspatialresolutiontodiscriminatetwophotonscomingfrom0decay.ThisisaverysignicantbackgroundforHiggssearches.Thisbackgroundisnotrelevantforthiswork.Fortheenergiesbelow500GeV,energyresolutionoftheECaldetectorcanbecharacterizedinthefollowingway[ 13 ]: E2=S p E2+N E2+C2 (5)Intheformula,S,NandCstandforthestochastic,thenoise,andtheconstantterms,respectively.Thestochastictermdependsontheuctuationsintheshower 39

PAGE 40

event-by-event.Thenoisetermisrelatedtotheelectronicnoiseandunderlyingevents.Whereastheconstanttermtakesintoaccountcrystalrelatedeffectssuchasleakageofenergyfromthebackofthecrystal,calibrationofthecrystalsandthelightcollectionfromtheshower.MoredetailsoftheCMSECalsystemcanbefoundelsewhere[ 13 ]. 5.6CMSHadrinocCalorimeterTheCMShadroniccalorimeter(HCal)islocatedjustaftertheECalbetween1.77mand2.95mintheradialdirection.ThepurposeofHCalistomeasuretheenergyoftheincidenthadrons,andtoallowthemeasurementofthemissingtransverseenergy(duethepresenceofaneutrinoorLSP).HCalisasamplingcalorimeterunlikeECal.TheprincipleoftheHCaloperationresemblesECal'soperation.Theincidenthadronsproduceashowerduetothestronginteractionwiththeabsorbers,yieldingthelightwhichisdetectedbytheactivemediumcalledscintillators.Theshowerischaracterizedbythenuclearinteractionlength0andiscomposedofsecondaryhadronslike,0,Ketc.Duetothepresenceof0(immediatelydecayingintotwophotons)particlesintheshower,thehadronicshowercontainstwoparts:electromagneticandhadronic(,etc).Themomentumtransferamongthesecondaryparticlesinthehadronicshowerisanorderof0.35GeV/c,resultinginamorepronouncedlateralshowerthaninECal,andthusbothshowersdevelopdifferently.Theshapeoftheshowerisusedtodistinguishbetweenthetwoparts.Furthermore,lateralshowerinuencesenergyresolutionoftheHCal.CMShadoniccalorimeterconsistsoffourparts:Barreljj<1.3(HB),endcapjj<3.0(HE),forward3.0
PAGE 41

interactionlengthhasthedependenceanditchangesfrom5.820at900to10.60at=1.3.TheCMSECaladdsanadditional1.10.ThefrontandbacksidesofeachwedgearecoveredbytheactivematerialtodetectthehadronicshowersproducedoutsideofHCal.Theabsorberplateshavedifferentthicknesseswithinawedge,therstandthelastonearemadeofstainlesssteelwith40mmand75mmrespectively.Betweenthem,thersteightbrassplatesare50.5mmthick,followedbysix56.5mmthickbrassplates.Eachwedgeissegmentedintofourazimuthalanglesectors.Theactivemediumissplitinto16sectorswhichleadstothegranularityof0.0870.087in)]TJ /F7 11.955 Tf 11.96 0 Td[(space.TheHEpartcoverstheregionofbetween1.3and3.0BothHBandHEarelocatedwithinthesolenoid.Brassplatesare79mmthickwithagapof9mmbetweentheplatesforthescintillatorinstallation.TogetherwithECal,HEcanprovideabout100radiationlength.ThegranularityofHEisalso0.0870.087in)]TJ /F7 11.955 Tf 12.92 0 Td[(between1.31.6.Theouterhadroncalorimeterliesoutsideofthemagnetcoilandcoversarangeofjj<1.3.Itprovidesadditionallayersofscintillationtocapturetheleakageoflongorlate-startinghadronshowers.Itusesthemagnetcoilasanadditionalabsorberwhichcanprovidedependenceradiationlengthequalto1.4 sin(). (5)TheHOisdividedinto5ringsinof300-sectors,followingtheshapeofthemuonbarrelsystem.Theforwardcalorimetersarelocatedatadistanceofjzj=11.2mfromthenominalinteractionpointandhaveacoverageof3.0
PAGE 42

mmseparation.Thesignalischanneledtoconventionalphotomultipliertubesbehindthickshielding.EachHFdetectorissegmentedinto13ringsof18wedges,leadingtogranularityof0.170.17in)]TJ /F7 11.955 Tf 12.42 0 Td[(.MoredetailsoftheCMSHCalsystemcanbefoundelsewhere[ 13 ]. Figure5-7. CMShadroniccalorimeterwithitsallpartsexceptforwardhadroniccalorimeter[ 13 ]. 5.7CMSMuonSystemInthedesignoftheCMSdetector,theroleofmuonsystemhadalotofimportancebecausemanyinterestingphysicsphenomenayieldmuonsinthenalstateandmuons'distinctresponsetothedetector.Nottomentionthatmuonidenticationisverycleancomparedtoelectronsorevenhadronicjets.TheCMSmuonsystemislocatedoutsidethesolenoidandcoversthejjregionupto2.4.TheCMSmuonsystemcomprisesthreedistinctgaseousdetectors.Inthebarrelregion(jj<1.2),drifttubes(DT)areemployed,whileintheendcapregion(1.2
PAGE 43

traversedit.Asaresult,themuondetectorresolutiondegradeswithincreasingmuonmomentum,unlikethecalorimeter.Itisworthnotingthatmuonsystemresolutionalonecanbeasbest10%.Thislowerboundisdrivenbythemultiplescatteringofthemuoncandidatesinthereturnyoke.But,withtheuseofthetrackersystem,theresolutionhasasignicantimprovement:formuonswithpTupto200GeV,theresolutioncanbeasgoodasfewpercent,whileformuonswithpTabove200GeV,theresolutiondegradesandstaysbetween15to45percentdependingthepseudo-rapidity[ 13 ]. A BFigure5-8. TheviewoftheCMSmuonsystem.A)showslocationoftheCMSmuonsystemswithrespecttothewholeCMSdetector.B)showscongurationofDT,RPCandCSCmuonsub-systems[ 13 ]. 5.7.1DriftTubeChambersInthebarrel,fourcylindricalstationsaroundthesolenoidformthewholedrifttube(DT)sub-systemofCMS.Inthisregion,themuonbackgroundislowandthemagneticeldisuniform;therefore,DTwith380nanoseconddrifttimeisperfectlysuitedforthebarrelregion.Oneachstation,the3innercylindershave60driftchamberseachandtheoutercylinderhas70chambers,resultinginatotalof172000sensitivewires.Eachchamberislledwith85/15%mixtureofArandCO2gasandiskeptattheatmosphericpressure.Allchambersaremountedtotheironyokewhichalsoservesasahadronabsorber.2.4mlongwiresinsideofeachchambermeasuresr)]TJ /F7 11.955 Tf 12.6 0 Td[(projection.Eachchamberis2.4mlongandismadeof3or2superlayers(SL).Thewiresinthe2outer 43

PAGE 44

SLsareparalleltothebeam,thereforetheymeasurether)]TJ /F7 11.955 Tf 12.54 0 Td[(coordinates,whereasthewiresintheouterSLareperpendiculartothebeamandmeasurethezcoordinatesalongthebeam.Thewireshave50mdiameterandaremadeofgold-platedstainlesssteel.Thesinglewireresolutionisabout250m,butthetotalresolutionof100mperchamberisachievedbytakingintoaccount8trackpointswhicharemeasuredinthetwoSLs. Figure5-9. Sketchofacellshowingdriftlinesandisochrones.Theplatesatthetopandbottomofthecellareatgroundpotential.Thevoltagesappliedtotheelectrodesare+3600Vforwires,+1800Vforstrips,and1200Vforcathodes[ 13 ]. 5.7.2CathodeStripChambersTheCSCsystemislocatedoneithersideofthecoil,withfourstationsforeachside.Therststation,whichisclosesttothecoil,containsthreerings,whiletheotherstationscontainonlytworings.Chambershaveatrapezoidalshapeandcovereither100or200indirectionandaremountedtotheironreturnyokes.Inthisregion,themagneticeldisnon-uniformandthemuonrateishigherthaninthebarrelregion.Eachchamberismadeof6anodewireplanesinterleavedamong7cathodeplanes,andislledwithAr,CO2andCF4gasmixture.Wiresarelocatedperpendiculartothebeam(alongdirection)andmeasurer)]TJ /F7 11.955 Tf 12.28 0 Td[(coordinates,whilestripswhicharemilledtotheplanerunlengthwiseprovidingthezcoordinates.ThedesignoftheCSCsub-detectorwasspecicallychosenasitdoesnotrequireauniformmagneticeldandcanoperate 44

PAGE 45

Figure5-10. LayoutoftheCMSbarrelmuonDTchambersinoneofthe5wheels[ 13 ]. atahighrate.CSCcanalsoparticipateintriggering.Theofinespatialresolutionisabout75mforthersttworingsontherststations,anddropsto150mfortherestchambers. 5.7.3ResistivePlateChambersCMSResistivePlateChambersaremadeoftwoparallelplates,eachdesignedwithphenolicresinwithabulkresistivityof1010-1011.Thus,theRPCchamberwasdesignedasthedouble-gapchamber,referredtoasupanddown,andwithstripsinbetween,tomaintainthefasttimeresponserequirement.ThegapislledwithagasmixtureofC2H2F4,C4H10,andSF6,andithasthicknessof2mm.Itisworthnotingthatthegapwidthinuencestimeresponseperformance,whereastheresistivity 45

PAGE 46

Figure5-11. AviewofaCSCchamber,whichconsistsof6anodewireplanesinterleavedamong7cathodeplane.Wiresrunalongthedirection,whilestripsrunalongtherdirection[ 13 ]. determinestheratecapability.ARPCchamberoperatesunderhighvoltageinordertoallowavalanchemode.AtotalofsixlayersaremountedinthebarrelpartoftheCMSdetector,withtwolayersinstrumentedontherstandsecondstations,whilethelasttwostationshaveoneRPCeach.OnlyfourRPCsarelocatedintheendcappartoftheCMSdetector,andtheycoverupto<1.6.However,itwasinitiallyplannedthatRPCswouldcoverthewholerangeoftheCSC,butthisplanwillbeimplementedlaterduetothebudgetissues. 5.8CMSTriggerandDataAquisitionSystemAtpeakluminosity,thecollisionsattheLHCwillhappenatarateof40MHz.Mostcollisionswillnotbeofprimaryinterestand,moreover,thelimitationsofadiskstorageallowsonlyasmallfractionofallcollisionstobestoredforlaterofineanalysis.Therefore,thetriggersystemanddataacquisition(DAQ)systemweredesignedtoselectinterestingeventsfortheanalysesconductedatCMS;forinstance,allevents 46

PAGE 47

whichcontain3leptonsaregreatlyimportantforthisanalysis.Atthedesignluminosity,theprotonbunchescollideevery25nanosecondswhichchallengesthedesignandtheimplementationoftriggerandDAQsystems.TheCMSdesignedatwo-leveltriggersystem:Level1(L1)andHighleveltrigger(HLT).L1isbasedfullyonthecustom-made,programmablehardwaredesignedtoreducetheeventratefrom40MHztofewhundredsofKHz.Adecisionwhethertoacceptaneventornotshouldbemadein3.2s(approximately28continuousbunchcrossingsatthedesignluminosity),duringwhichtimethedatacanbestoredintheelectronics.Duetothetimeconstraint,L1triggerprocessesinformationonlyfromcoarselysegmentedcalorimetersandthemuonsystemwhichsimpliesdecision-making.TheECALandHCALdetectorsaresegmentedintotriggertowers,eachconsistingofablockofECALcrystalsandacorrespondingsetofHCALtowers.Theenergypatternsinthetriggertowersareprocessedtoidentifycandidateelectrons,photons,andjets.Theyarecalledtriggerprimitives.Triggerprimitivesarealsoformedinallthreemuonsystemsseparately:DT,RPCandCSC.ThetracksegmentsinthesuperlayersoftheDTchambersformtriggerprimitives.TheyaresenttotheDTTrackFindertomatchsegmentsfromdifferentchambersandtoassignkinematicandqualityparameterstothetrack.Similarly,tracksegmentsareconstructedintheCSCfromthecathodeandanodereadoutineachchamberusingpatterntemplates.Three-dimensionaltracksareconstructedfromthesesegmentsusingtheCSCTrackFinder.IntheRPC,hitpatternsareformedusingthespatialandtemporalcoincidenceofhitsinseveralRPClayers,andareassignedtotheproperbunchcrossingduetoitsfasttimingcapability.TheGlobalMuonTriggerreceivesuptofourmuoncandidatesfromeachoftheDT,CSC,RPC-barrel,andRPC-endcapdetectors.Lookuptablesaredesignedandusedtocombinecandidatesassociatedwithacommonmuonandtoassignaqualitycriteriatothecandidates.ThefourhighestqualitycandidatemuonsareforwardedtotheGlobalTrigger. 47

PAGE 48

TheGlobalTriggercarriesoutthenaldecisiontoacceptanevent.L1ranksallobjectsinaneventaccordingtothequalityandkinematicvaluesandthenitforwardsthemtotheHLTsystem.TheL1systemisfullymountedtotheCMSdetector,whereastheHLTsystemislocatedonthesurface,approximately100mawayfromthedetector.Thedataowsbetweenthetwosystemsusingaswitchingnetworkthatoperatesat100GB/s.AschematicrepresentationofthetriggersystemisshowninFigure 5-12 Figure5-12. ArchitectureoftheCMSL1system[ 13 ]. TheHLTsystemanalyzeseventsbasedfullyonthesoftwareanditsdecisionhastobemadeinloosertimewindowof100ms.Ittakesintoaccountthefulldetectorinformation,thereforeitcanperformamoreaccurateeventevaluationthanL1.Itconsistsofaprocessorfarmwithabout1000nodes,9200cores,and18TBofmemory.AttheHLTlevel,itispossibletoapplymorestringentkinematicandqualityrequirementsfortheobjectselectionthanatL1,resultinginfurtherreductionoftheeventratestofewhundredsofHzwhichwillbenallytapedforlaterofineanalysis.Theaveragesizeofatapedeventisapproximately1.5MB.ThegoaloftheCMSdataacquisitionsystemistoprovideasmoothtransferofdatafromtheL1tothestoragedevice.EventbuilderwithinDAQmonitorsthedataowto 48

PAGE 49

avoidunwantedevents.Forinstance,ifanysub-detectorfailsduringthedata-taking,thenDAQwillnoticeandcanmarkthedataasunwanted. 49

PAGE 50

CHAPTER6OBJECTRECONSTRUCTIONANDIDENTIFICATION 6.1MuonReconstructionandIdenticationMuonreconstructionattheCMSdetectorisarelativelyeasytaskduetoitsrelativelylonglifetime(10)]TJ /F9 7.97 Tf 6.58 0 Td[(6second)andrelativelyhighmass(105GeV),whichallowsittotraversethewholeCMSdetectorwithasmallenergyloss.Theenergylosstakesplacewhenmuonpassesthroughthedetectormaterial,mostlytheelectromagneticandhadroniccalorimetries,anddepositsenergywhichisconsistentwiththeenergyoftheminimumionizingparticle(MIP).CMSdevelopedtwoalgorithmstodealwiththemuonreconstructionandidentication,thesearebasedontheinformationsolelyfromthemuonsystemandthetracker:GlobalmuonreconstructionandTrackermuonreconstruction[ 18 ].Globalmuonreconstructionperformsthereconstructionbyidentifyingthemuoncandidatesegments1inthemuonsystem(DTorCSC).Segmentsshouldbepresentatleastintwomuonstationsinaconsistentway.Thealgorithmthenpropagatesthetrackstowardsthetrackertondthecorrespondingtrackertracksprovidedthatthetracksweregeneratedbyarealmuon.ThetbetweenthetracksfromthetwosystemsiscarriedoutusingtheKalman-ltertechnique[ 17 ].GlobalmuonreconstructionimprovestheperformanceofthemuonreconstructionformuonswithpT>200GeV.Thetrackermuonreconstructionalgorithmstartsareconstructionprocedurefromthetrackersystem,whereallpreselectedtrackswithpT>0.5GeVandp>2.5GeVwithinthetrackerpseudo-rapidityacceptanceareconsideredasmuoncandidates.Thealgorithmthenpropagatesthemtowardsthemuonsystem(DTandCSC)bytakingintoconsiderationmagneticeldconguration,expectedenergyloss,andmultiple 1Amuonsegmentisformedwithinamuonchamberusingalineartofthehitsindifferentlayers.FortheDTsubsystem,8-12layersareused,whereas4outof6layersareusedforCSC. 50

PAGE 51

scattering.Ifanytrackertrackisconsistentwithonemuonsegment,itisthenqualiedasthetrackermuon.Thematchingbetweentheextrapolatedtrackertrackandthemuonsegmentisperformedinthelocalcoordinatesofamuonchamber,andthematchingcriteriavariestoincreasethereconstructionefciency:thematchingissuccessfuliftheclosesttrackisselectedwithin3cmorifthevalueofthepullofthelocalcoordinates(thedifferenceofthepositionsofthetrackertrackandmuonsegmentdividedbytheiruncertainties)islessthan4.Thetrackermuonalgorithmismoreefcientforlowenergymuons.Oncethebasicreconstructionprocedurehasbeenperformed,tofurtherapplystringentselectionisanecessarysteptoincreasethepurityofreconstructedmuoncandidates.TheCMSmuonworkinggrouphasdevelopedseveralmuonidenticationselections[ 19 ]whichallowscientistseasyaccesstothemandtohavecommonmuonselectioncriteriaamongdifferentsearches.Currently,therearethreemuonselectionsavailable:Loose,SoftandTightmuonselections.Inthiswork,Loosemuonselectionisused,andadetaileddescriptionispresentedinsection 8.2.1 6.2ElectronReconstructionandIdenticationElectronreconstructionisgenerallyamoredifculttaskthanmuonforseveralreasons:electroncanradiatephotonswhilepassingthroughthematerial,thereforespecialcareisneededtoaccountfortheenergyloss;photonscancreateanelectron-positronpairinthetrackermaterialandeitherisfalselyidentiedasarealelectron;orjetscaneasilyfakesignalelectronsignatureasbotharecapturedbythecalorimetries.TheCMSEGammagroupdesignedtwocomplementaryalgorithmstoaddressthesechallenges[ 20 ]:trackerdrivenandECaldriven.Bothalgorithmsarebasedonthecombinedinformationfromthetrackerandcalorimetries.ThetrackerdrivenalgorithmismoresuitableforlowpTelectronsorforelectronswhichoriginatedfromthejetactivities.ReconstructionalgorithmowstartsbyselectingtracksinthetrackerandpropagatingthetrackstotheECalsystem.Whereas,the 51

PAGE 52

otheralgorithmismoresuitableforelectronsoriginatedfromthedecayofWorZbosonswhichnaturallycomewithouthadronicactivityaroundthem.TheECal-seededalgorithmbeginsareconstructionfromthedepositedenergy(ET>4GeV)intheECalsupercluster2.Itthenpropagatesthetrackbacktothetrackersystembyconsideringthepresenceofmagneticeldtondthecorrespondingtrackseeds(pairsortripletsofhits)intheinnertrackerlayers.ThenaltrajectoriesarereconstructedusingaGaussianSumFilter(GSF)[ 21 ].Reconstructedelectronsfurtherrequirestringentselectioncriteriaasthepurityofreconstructedelectronsintherststageislow.Severalpowerfulvariablesareemployedintheselectionswhichprimarilyrejectmis-identiedelectronsfromjetsorphotonconversions.TheEGammagroupdesignedseveralstandardizedworkingpoints(WP)[ 22 ]toaddressdifferentneedsforthedifferentanalyses:Veto,Loose,MediumandTight.Forthiswork,LooseWPisselected,andadetaileddescriptionispresentedinsection 8.2.2 6.3JetReconstructionAlgorithmManystandardmodelparticlesareproducedattheLHC.However,jetproduction,asprayofhadronsoriginatingfromafragmentationofquarkorgluon,dominateduetoitshugeproductioncross-section.Itsefcientreconstructionisakeytaskforfuturediscoveries.Overthepastfewdecadesmanyalgorithmshavebeendevelopedtoessentiallygroupthesprayofhadronsintoajet.Sincemanypile-upeventsareexpectedinthecollisionsleadingtosoftjets,thealgorithmshouldberobust,efcient,andinsensitivetosoftjetswhilemergingthecomponentsofahardjet.Therefore,CMS 2SuperclusterisagroupofoneormoreassociatedclustersofenergydepositsintheECalsystem.ItwasprimarilydesignedtocollectenergyofanelectronwhichradiatesphotoninthematerialofthetrackerwhichisinstalledjustbeforetheECalsystem.Radiationisspreadinthedirectionduetothemagneticeldwhilethedirectionstaysconstant.Consequently,superclusterhasnarrowwidthinthedirectionandcoversawiderrangeinthedirection. 52

PAGE 53

employsanti)]TJ /F4 11.955 Tf 12.67 0 Td[(kTalgorithmforjetreconstructionduetoitsrobustnessandefcientperformance.Allavailablealgorithmsarebasedontheseformulas[ 24 ]:dij=min(k2nTi,k2nTj)42ij R2 (6)diB=k2nTi. (6)Where,4ijisthedistancesbetweeniandjparticlesinandspace,4ij=(i)]TJ /F7 11.955 Tf 12.51 0 Td[(j)2+(i)]TJ /F7 11.955 Tf 12.51 0 Td[(j)2andkTirepresentsthetransversemomentum.Differentvaluesofthenvalueclassifydifferentalgorithms[ 24 ],theanti)]TJ /F4 11.955 Tf 12.44 0 Td[(kTalgorithmisobtainedbysettingn=-1.Havinggroupedhadronswiththeanti)]TJ /F4 11.955 Tf 10.8 0 Td[(kTalgorithm,ashapeofthejetcreatedbyahardhadronisnotaffectedbythenearbysofthadronswithin4
PAGE 54

trackerallowsittodetermineindividuallythemomentumofthechargedhadronsinsideofthejetwithhighprecision,buttheneutralpartofthejettraversingthetrackercanonlybereconstructedinthecalorimeters.Thenon-linearitywithrespecttotheincidentparticleenergyisaresultofthedifferentresponseoftheHCaltothehadronicandelectromagneticcomponentsofahadronshower.Moreover,thepresenceofpile-upinteractionswillcontributetoanadditionalunwantedenergytothereconstructedjets.Toaccountforsucheffects,themeasuredfour-momentumofajetisscaledbyacorrectionfactordependingonthetransversemomentumandpseudo-rapidityofthejet[ 26 ].Bottomquark-jetexhibituniquecharacteristicsthatdistinguishesthemfromjetsoriginatedfromlight-avor(u,dors)quarks,gluons,andtoalesserextentc-quarks.Thediscriminatingpropertiesofheavyavorhadronsinclude:theirrelativelylargemass,highkineticenergies,longlifetime(c400m),thelargenumberofchargedparticlesproducedintheirdecayandtheirlargebranchingratiostoleptons.Thus,thereareseveralalgorithmsforb-jetreconstructions.However,inthisanalysistheonlyonealgorithmisused,combinedsecondaryvertex(CSV).Thedescriptionofotheralgorithmscanbefound[ 27 ].CSValgorithmusesapresenceofasecondaryvertexinthejetwhichisthedistinguishingfeatureofb-jets.Asinthecaseofprimaryvertexreconstruction,secondaryverticesarereconstructedwiththeadaptivevertextter.Tracksenteringthevertextmustsatisfymorestringentrequirementsthanintheprimaryvertexreconstructiontomaintainthereasonablepurity.Inaddition,tracksarerequiredtoliewithinR<0.3ofthejetaxis.Riscalculatedbyformingaconein)]TJ /F7 11.955 Tf 12.08 0 Td[(spaceofradiusR=p 2+2=0.3centeredaroundthejetcandidate.Secondaryverticesarerequiredtohavelessthan65%oftheirassociatedtracksincommonwiththeprimaryvertextoavoidambiguity.Intheinstanceswhereasecondaryvertexisnotreconstructed,highqualitytrackscanstillbecombinedtoformapseudo-vertex,ifneitherisconstructed,thediscriminationismadeusingonlythetrackinformation[ 27 ].TheCSValgorithmisconstructedbyajointlikelihoodusingasetofninevariables. 54

PAGE 55

Thesecondaryvertexcategory(i.e.vertex,pseudo-vertex,orno-vertex); Thetransverseightdistancesignicance; Avertexmassconstructedfromthetotaltracksassociatedwiththevertex; Thenumberoftracksatthevertex; Thefractionofthejetenergycarriedbytracksfromthevertex; Thepseudorapidityofvertextrackswithrespecttothejetaxis; Thetransverseimpactparametersignicanceofthersttrackthatraisesthevertexmassabovethec-quarkthresholdof1.5GeV(withtracksorderedindecreasingtransverseimpactparametersignicance); Thenumberoftracksinthejet;and Thethree-dimensionalimpactparametersignicanceofthetracks.Foreachjetavor(b,c,orlight/gluon),weconstructalikelihoodfunctionandthenab-jetdiscriminantusingaweightedsumoflikelihoodratios[ 27 ]. 6.6TauReconstructionandIdenticationThetauleptonplaysaveryimportantroleinstandardmodelphysics,aswellasinnewphysicssearches.Itdecaysapproximately17.82%(17.34%)intotheelectron(muon)lepton,so,inthiscase,itsreconstructionisessentiallybasedonthelightleptonreconstruction;but,mostofthetime,itdecayshadronically.Itisalmostimpossibletodirectlydetectduetoitsveryshortlifetime(2.910)]TJ /F9 7.97 Tf 6.58 0 Td[(13s)andseveraldecaymodes(Table 6-1 ),therefore,itspresencecanonlybeinferredbythedecayproducts.Ontheotherhand,itsdecayproductresemblesthenatureofhadronicjetsoriginatingfromthequarks,whichareproducedinenormousamountsinthecollisions.Thus,itisaverychallengingtasktoidentifyhwithareasonablepuritywhilemaintaininghighefciency. 55

PAGE 56

CMSdevelopedtwoalgorithmstoreconstructandidentifyahadronicallydecayinghlepton:thehadronplusstrips(HPS)andthetauneuralclassier(TaNC)3.TheHPSalgorithmisbasedonthePFreconstructedparticlesandaredesignedinawaytocoveralldecaymodeswhicharelistedinTable 6-1 .Hadronictauidenticationisachievedbylookingattheintermediatedecayproducts,thetworesonants1withamassof1260GeVandmesonwithamassof770GeV.ThealgorithmstartsfromthejetswhicharereconstructedusingPFalgorithm,thealgorithmwhichcansegregatealldifferentparticlessuchaschargehadrons,neutralhadrons(0),muonsandelectrons.Electromagmeticclustersreferredtoasstripsareusedtoinferapresenceof0(withinthejet)throughthetwophotons.Thealgorithmalsotakesintoaccountthepossiblephotonconversioninthetrackermaterialwhichbroadensthesignalinthedirectionduetothemagneticeld.Astripwindowof4=0.05and4=0.2isdened,inwhichallenergydepositsaretakenintoaccounttodeterminethenalstripfour-momentumwhichmusthavepTabove1GeV.Itisthencombinedwiththechargedhadrontoproceedforthetauidentication.TheHPSalgorithmtakesintoaccountthefollowingtopologiestomaximizetheefciency: singlehadronforhandh0modes,whereinthelatterphotonsfrom0decayaresoftandthereforeastripdoesnotpasspTcut; onehadron+onestripfortheh0modewheretwophotonsfrom0decayaretooclosetoeachotherandtheyarereconstructedasonestrip.Aninvariantmasswhichisconstructedusingthehadronandthestripisrequiredtobeconsistentwiththemesonmasswindow(770400MeV); onehadron+twostripsfortheh0modewheretwophotonsfrom0decayareseparatedenoughtoreconstructthemseparately.Aninvariantmasswhichisconstructedusingthehadronandoneofthestripsisrequiredtobeconsistentwith 3TaNCalgorithmisnotusedanywhereinthisworksoitisnodiscussedhere.Itsdescriptioncanbefound[ 23 ] 56

PAGE 57

mesonwindow(770400MeV);andaninvariantmassconstructedfromthetwostripsshouldbeconsistentwith0hypothesis(14050MeV);inaddition, threehadronsforthehhhmode.Allhadronsarerequiredtobeconsistentwiththevertexandaninvariantmassshouldbeconsistentwiththe1meson.Ifmorethanonetautopologiesisreconstructedinanevent,thepreferredtopologyisconsideredtheonewithhighesttaupT.Tofurtherincreasethepurityofreconstructedtaus,tausarecleanedwithrespecttomuonorelectron;becausetheirenergydepositinthecalorimetrytogetherwiththeirtracksinthetrackercaneasilyfakethetausignature.Todoso,thetauleptonisrejectedifitsleadingchargehadrontrackmatchesthemuonortheelectronwhichpassestheLooseWP([ 19 22 ]).Thetauleptonisolationiscalculatedusingallenergydeposits(excludingallparticleswhichwereusedinthedeterminationoftauleptonmode)within4R=0.5fromthetauaxis.TheCMSTauworkinggroupdesignedthreeworkingpoints(whichareprimarilybasedonthepTthresholdoftheparticleswhichgointheisolationcalculation):Loose,MediumandTight.Inthiswork,theLoosetauselectionisused,andadetaileddescriptionispresentedinsection 8.2.3 Table6-1. Branchingfractionofthedecayofthetaulepton DecaymodeResonanceMass(Mev/c2)BranchingFraction !h11.6%!h077026%!h00a112009.5%!hhha112009.8%!hhh04.8% 6.7MissingTransverseEnergyReconstructionAtthehadroniccolliders,thetotaltransversemomentumofallthecomponentsinacollisionisalmostzero[ 28 ].Thus,thisinformationcanbeusedtoconcludeanyenergylosseventbases.Neutrinosandanyotherpossibleparticles(forexample,LSP)fromthenewphysicswhichcanonlyparticipateintheweakinteractiontraversethewholedetectorwithoutanyinteraction,therefore,thisleadstotheenergyimbalance.EmissTis 57

PAGE 58

denedasfollows:EmissT=)]TJ /F12 11.955 Tf 11.29 13.75 Td[(X~pT (6)wherethesumgoesthroughalltheparticlesinanevent.CMSusestheparticle-owalgorithm[ 25 ]toreconstructEmissT. 58

PAGE 59

CHAPTER7EVENTSIMULATION 7.1OverviewofEventSimulationInhighenergyphysicsexperiments,werelyontheknowledgeoftheoreticalexpectationsforthebackgroundaswellasforthesignal.Currenttheoreticalmodelsessentiallyinvolvecomplicatedintegralswhich,inmostcases,cannotbecalculateddirectlyduetotheircomplication.Therefore,simulationisinevitable.SimulationemployMonteCarlotechniquestoperformagreatnumberofsimulatedexperimentsusingrandomnumbergenerators. 7.1.1MonteCarloSimulationEventgenerationusingMonteCarloismodularanddevelopsinstages.Inmanycases,itisfeasibletousedifferentsoftwarepackagesforeachstep.Thegenerationstartswiththerelevantinformationaboutthecollisionstobesimulatedincludingtheparticlestobecollided(inthiscaseprotons)andthecenter-of-massenergyofthecollision(inthiscasep s=8TeV).Ingeneral,ausercanmanipulatethesesettings.Thestructureofthecollidingprotonsismodeledwiththeuseofpartondistributionfunctions(PDFs).ThisPDFdenestheprobabilitydensityforndingaparton(aquarkorgluon)withagivenlongitudinalmomentumfractionxatagivenvalueofmomentumtransferofthecollision.Thecollisionoccursbetweenindividualpartonswithineachproton,andisreferredtoasthehardprocess.Theparticlesofinterestfortheeventareproducedinthehardprocess(e.g.vectorbosons,heavyquarks,hypotheticalnewparticles,etc.).Oncetheparticlesofinterestareproduced,theirsubsequentdecaysarethengovernedbymatrixelements.ThesearecalculatedfromFeynmandiagramsinquantumeldtheory,wherethesquareofthematrixelementgivestheprobabilitydensityfortheprocess. 59

PAGE 60

7.1.2PartonDistributionFunctionsandtheHardProcessThedistributionofthemomentumfractionxofthepartonsinthecollidingprotonsdirectlyinuencesthecalculationoftheproductioncross-section.Inpractice,thePDFscontainingthisinformationareimpossibletocalculateperturbatively,andareinsteaddeterminedbytstodatafromtheprocessessuchasdeepinelasticscattering(DIS)andDrell-Yan(DY).GroupssuchasCTEQ[ 29 ]provideregularupdatesofthesePDFs.Theinteractionproducingthehardscatteroccursbetweentheindividualpartonsineachproton,withthePDFsdeterminingtheenergyavailablefortheprocess.Feynmandiagramsareusedtocalculatethematrixelementforeachprocess,andtheproductioncross-sectioncanbederivedfromthismatrixelement.Itisworthnotingthatacross-sectioncalculationusingtreeleveldiagramsresultsinleadingorder(LO)cross-sections,andsomeprogramsthatsimulateeventsincludeonlyLOcrosssectioncalculation;howeverthereareseveralseparatesoftwarepackagesavailablewhichcangobeyondLO. 7.1.3PartonShowersandHadronizationInadditiontothehardscatteringprocess,thecollisionincludesQCDradiationfromboththeincomingandoutgoingpartonsandisreferredtoasinitialandnalstateradiation(ISRandFSR),respectively.ThisradiationdependsprimarilyonthemomentumtransferscaleQ2,ratherthanthedetailsoftheparticularprocessbeingsimulated.Dependingontheanalysis,ISRandFSRcanhaveasignicantimpacttothenalstate.ItshouldbenotedthatforthisanalysisFSRisnotrelevantbecausetheintermediateandnalstatesinthedecaychaindonotinvolvecolorparticles;whereasISRtakesplacebutitsinuenceisnegligiblebecausetheanalysisstaysinclusiveinthejetmultiplicity.Inotherwords,wedonotapplyanyrequirementonthenumberofjetsinanevent.Duetocolorconnement,partonsproducedinthehardscattercannotexistontheirown.Theirkineticenergyistransferredtothecoloreld,whereitproduces 60

PAGE 61

additionalpartonsfromthevacuum.Thepartonsinthisshowerthenhadronize,formingcolor-neutralcombinations.Theresultingcollinearsprayofparticlesproducesareconstructedobjectthatwecallajet.Theprocessofhadronizationisnotwelldescribedtheoretically,andisinsteadmodeledphenomenologicallywiththeLundstringmodel[ 34 ].Inthismodel,thegluonsbindingtwoquarksaretreatedaseldlines.Gluonsareself-interacting,andthereforethecoloreldtheyproduceiscompressedintoanarrowtube(orstring).Bycontrast,theelectromagneticeldtendstospreadmuchmore,becausethephotonhasnoself-interactionterms.Whensufcientenergyisstoredinthecoloreldtoproducenewquark-antiquarkpairs,thestringbreaksup,andthenewlyproducedparticlesformbound-statehadronswiththeoriginalquarks.Duringthehadronizationthisprocesshappensrepeatedly,andtheproducedboundstatesmodelthekinematicsoftheoriginalparton.Theremainingpartonsfromtheprotonscannotbeignored;theyrepresenttheunderlyingeventandareincludedwhenconsideringhadronization.Theseinteractionsaretypicallysoftandthedescriptionoftheunderlyingeventreliesonnon-perturbativeorsemi-perturbativephenomenologicalmodels.Additionally,theeffectofpile-upmustalsobeconsidered,wheremultipleprotonsinasinglebunchcrossinginteract.Pile-upissimulatedingeneratedeventsbysuperimposingminimumbiaseventsontheeventsinthenominalsamples. 7.1.4EventGeneratorsAsmentionedbefore,differentMonteCarlosoftwarescanbeemployedinthesimulationprocesstoachieveasatisfactorylevelofperformance.Thisapproachismodeldependentandvariesfromanalysistoanalysis.TherearemanyMonteCarlosoftwaresavailableattheHEPmarket.However,onlytwoofthem,themostwidelyused,arediscussedhere: PYTHIA-Simulateseverystepintheevent,includingtheinitialinteraction,partonshower,underlyingeventandsubsequenthadronizationanddecaytonalstateparticles.Crosssectionsarecalculatedonlytoleadingorder(LO),andthese 61

PAGE 62

generatorsaremostsuitablefor2!2processes(thosewithtwoinitialandtwonalstateparticles).Moreinformationcanbefoundelsewhere[ 32 ]. MADGRAPH-Itisnotsufcienttoknowonlyleadingordercross-sectionsbuttoalsolookbeyondthenext-leadingorder.TheMADGRAPHgeneratorspecializesinmodelingthehardinteractionandiscapableofcalculatingnext-to-leadingorder(NLO)correctionstomatrixelements,aswellassimulating2!nprocesseswithmultiplenalstatepartons.Moreinformationcanbefoundelsewhere[ 33 ] 7.2DetectorSimulationOncemodelsofvariousprocessesaregenerated,wearelefttounderstandthedetectorresponsetotheparticles.ThiswouldimitatetherealCMSexperiment.ThesimulationoftheCMSdetectorisdonewithaprogramcalledGEANT4[ 30 ],atoolkitusedtomodeltheinteractionsofparticlesinthedetector.Thealgorithmincorporatesinformationonthematerials,magneticeldsandspecicgeometryoftheCMSdetector,andusesindependentpartstodeterminethedetectorresponsetothesimulatedparticles. 7.3FullandFastSimulationsoftheCMSdetectorAfterallthesestagesdescribedabove,thefullinformationaboutthephysicsprocessesandtheCMSdetectorresponseareknown.Inpractice,thisinformationishugeandtostoreitisabigchallenge.Inaddition,simulationrequiresalotofCPUtime.Thedetectorresponsecontributesthemost,whichincludesdetailedinformationfromeachsub-detector.Therefore,twoapproachesweredevelopedtofurtherdecreasethesimulationtimepereventandreducetheeventsize:fullandfastsimulations.Infullsimulation(asthenameimplies)allinformationfromeachsub-detectorresponseisaccessibleanddetectorresponseiscalculatedaspreciselyaspossible,includingelectromagneticandhadronicshowers,andthetracksinthetrackersystem.Whereas,fastsimulation[ 31 ]isasimpliedversionofthefullsimulation.Thesimplicationwasintroducedtoalmostallsub-detectors[ 31 ]: Thetrackersub-detector-Thetrackerismadeof30thinnestedcylindersrepresentingtheactivelayers; 62

PAGE 63

Theelectrommagneticcalorimetry-Representedasahomogeneousmedium,whichisagoodapproximationfortheactualcrystal.TheelectromagneticshowerissimpliedbyintroducingseveralthousandenergyspotswhicharedistributedlongitudinallyaccordingtoaGammafunction.Thedepositedenergyisintegratedover2X01radiationlength,andtheactualcrystalshapeistakenintoaccountatthelaststage,assumingnoenergyleakage; Thehadroniccalorimetry-Theresponseofchargedhadronsisderivedfromthefullsimulationandtabulatedaccordingtotheirenergyandpseudo-rapidity.Themeanandrmsaretakenfromtheresponsedistributiontosimulatetheresponseofthehadronsinfastsimulation;and Themuonsystems-Themuonsystemislessaffectedbythebackground,thereforetheonlysimplicationisrelatedtothemuonenergylossschemeandonlymultiplescatteringandtheenergylossbyionizationaretakenintoaccountforthefastsimulation.However,bothfullandfastsimulationsyieldidenticalresultsforthemuonsselectionsusedinthisanalysis. 1Ameanfreepathoverwhicharelativisticelectronloses63%ofitsinitialenergythroughBremsstrahlung. 63

PAGE 64

CHAPTER8SEARCHFORSUPERSYMMETRYINEVENTSWITHTRI-LEPTON,MISSINGENERGYANDB-TAGGEDJETVETO 8.1IntroductionAtthebeginningoftheoperationoftheLHC,asmallamountofdelivereddataallowedscientiststosearchforsupersymmetryinthestrongproductionwhichfeaturelargeamountsofhadronicactivityinthenalstate(e.g.,pp!~g~g,~g~q,~q~q[ 35 ]).Thesemodelswereveryattractiveastheproductioncross-sectionisrelativelyhighduetothestronglyinteractingparticlesintheinitialstate,especiallyifthesquarksandgluinos(thesuperpartnersofquarksandgluon,respectively)provetobelight.However,theproductioncross-sectionscalesinverselywiththemassesofthesecoloredsuperpartners.NosignicantdeviationfromtheStandardModel(SM)predictionsareobservedamongthesesearches,andmassesofgluinoandsquarkareprobedaround1TeV[ 35 ];ontheotherhand,squarksandgluinosbecomefurtherconstrainedtotakeonhighervaluesastheprotonPDFsaresteeplyfallingwithrespecttoacenter-of-massenergy.WithmorecollecteddatafromtheLHC,theelectroweak(EWK)supersymmetryproductionmechanisms(e.g.,chargino-neutralinoproduction,pp!~0~)canbeveryadvantageous.Here,weassumethatthecoloredsuperpartnersareheavyandthereforedecoupled,leavingonlyEWKproductionmechanisms.Similarsearcheswerecarriedoutinthepredecessorcolliders.TherststrictconstraintscamefromtheLEP2experiments[ 36 ],whichprobedthecharginoandneutralinomassesupto103GeV;theDexperimentthenprobedcharginomassesupto140GeV[ 37 ];and,recently,theCMSexperimentwith7TeVcenter-of-massenergyand5fb)]TJ /F9 7.97 Tf 6.58 0 Td[(1collecteddataprobedcharginoandneutralinomassesupto500GeV[ 38 ],stillleavingavastamountofviableparameterspacetobeprobedbyadedicatedsearch.Thisworkpresentsthedesign,results,andimplicationsofasearchforEWKsupersymmetrybasedontheexclusive3leptons,plusmissingtransverseenergy(EmissT) 64

PAGE 65

signatureusingthecollecteddataof19.5fb)]TJ /F9 7.97 Tf 6.59 0 Td[(1andthecenter-of-massenergyof8TeV[ 39 ].Thereareonlyafewdiagramsofinterestforthechargino-neutralinoproductionsuchasinFigure 8-1 .Inaddition,wedonotexpectmuchhadronicactivityinthenalstate,exceptthatwhichcanbeattributedtoinitialstateradiation(ISR).Thelargestbackgroundfromstrongproductionisexpectedtobefromtop-quarkevents(e.g.tt,tt+V).Wedemonstratethatthistypeofbackgroundcanbesignicantlyreducedbyemployingb-taggedjetveto.FromEWKproductionprocesses,theStandardModeloffersadiversecollectionofrareprocesseswhichcanmimicthissignature,albeitwithsmallrates.Byfar,thedominantbackgroundisexpectedtobeWZproduction.Infact,oursignaltopologyresemblestheSMWZproduction,howeverthesupersymmetricmasssplittingsgreatlyinuencethesignalacceptanceefciencies.AnoverlapwiththeSMWZproductionisafunctionofthemassscalesandsplittingsbetweenthesupersymmetricparticles.Thus,theleveltowhichtheWZbackgroundisirreduciblestronglydependsonthedetailsofthenewphysicsmodeloneconsiders.WeexploreavarietyofdifferentmanifestationsofEWKSUSYproduction.Figure 8-1 representsthegenerictopologytargetedbythissearchinthecontextofMinimalSupersymmetricStandardModel(MSSM).Theslepton(~`)massintheintermediatedecaychainisconstrainedbycharginoandLSPmasses,accordingtothisformula:m~`=m~01+x~`(m~1)]TJ /F4 11.955 Tf 11.95 0 Td[(m~01), (8)wherex~`cantakeanyvaluebetween0and1.Currently,weexplorethreedifferentx~`parameters:0.05,0.5,and0.95.Wealsoassumethatsleptonsdecay100%intothesameavorSMleptonandLSP.Thesignalacceptanceefciencyis,ingeneral,afunctionofx~`andthereforeweexpectthatdifferentx~`parametersmanifestdifferently.Theinterpretationoftheresultmayfurtherdependonwhetherthesleptonisright-handed~`Rorleft-handed~`L.Weconsidertwolimitingcases.Inonecase,in 65

PAGE 66

A B CFigure8-1. Feynmandiagramsforthechargino-neutralinoproductionandtheirdecaymodes.A)showschargino-neutralinoproductionwith~02decayingintoapairofslepton-lepton,and~1decayingintoapairofslepton-neutrino.B)showsthesameaspreviousdiagramexceptthat~1decaysintoapairoflepton-sneutrino.C)showsthecasewhensleptonisheavierthancharginoandneautralino,andthereforeadecayhappensviaW=Zbosons. whichonlyleft-handedsleptonsparticipateinthedecaychain(thismeansthattheright-handedsleptonistooheavytoconsideratthisenergyscaleandthereforeitisdecoupled),thediagrams 8-1A and 8-1B existandwiththeequalprobability.Itmustbenotedthat,inthiscase,theneutralino(~02)candecayintoapairofslepton-lepton(~02!~``)aswellasapairofsneutrino-neutrino(~02!~)withthesameprobability.Thelattercasewillhavezerosensitivitytothe3leptonnalstate,thusthispartwasnotgeneratedinthecurrentsignalsamples,andwewillhavetoapplyafactorof0.5tothesignalefcienciestoaccountforthisloss.Weassumethatthedecayofcharginoandneautralinotoallthreeavorshappenswiththesameprobability.Thisscenarioisreferredtoasavor-democratic.Inthesecondcase,inwhichonlytheright-handed 66

PAGE 67

sleptonparticipatesinthedecay,thedecayofcharginoisconstrainedbecausethereisnocouplingtoSU(2)Lgauginos.~`Rcanonlycoupletothehiggsinocomponentofchargino.Furthermore,thehiggsinocomponentofthecharginoisinuencedbyandthestrengthofthecouplingbetweenthehiggsinoandright-handedsleptonisinuencedbytheSMYukawacouplings.Asaresult,charginosprefertodecaytotheright-handedstauslepton,causingtheexistenceofthediagram 8-1A onlybecausethereisnoright-handedsneutrino.Then,dependingontheneutralinodecay,twoscenariosarefurtherfeasible:tau-enrichedandtau-dominated.Inthetau-enrichedcaseneutralinodecaysdemocraticallyintoapairofslepton-leptonofanyavor,whereasinthetau-dominatedscenariotheneutralinocanonlydecayintoapairofstau-tau.Itisworthnotingthatallthreescenariosavor-democratic,tau-enrichedandtau-dominatedremainvalidifthesleptonmassislighterthanchargino/neutralino.However,wedonotlimitourselftootherpossiblescenariossuchasshowninFigure 8-1C .Thisscenarioispreferredifthesleptonsprovetobetooheavytoconsider.Therefore,chargino/neutralinoyields3leptonsinthenalstateviaintermediateon-shellWandZgaugebosons,withbothbosonsdecayingintoleptons. 8.2ObjectSelectionCriteriafortheAnalysisInChapter6,objectreconstructionalgorithmsattheCMSdetectorwerebrieyintroduced.Inthissection,allobjectswhicharetheprimaryinterestfortheanalysisarecloselyrevisitedintermsofthekinematicandqualityrequirements.Inthisanalysisweusemuons,electrons,taulepton,andmissingtransverseenergy.Wevetoeventswithb-taggedjets.AlltheseobjectselectionsareadoptedfromtheCMSphysicsobjectworkinggroups(POGs),thustheyarewellunderstoodandapprovedbytheCMScollaboration. 8.2.1MuonSelection pT>10GeV. jj<2.4. 67

PAGE 68

MuonsshouldsatisfyParticleFlowandGlobalmuonqualityrequirements.Theserequirementsguaranteehighpuritymuonsinthesample. Normalizedglobaltbetweenthetrackerandmuon-systemtracks(2=ndof)shouldbelessthan10-thisisanimportantrequirementtoensurehighpurityglobalmuons. Thereshouldbeatleastonestandalonehitinthetrackersystem. Muonshouldhaveconsistenthitsatleastintwomuonstations. Impactparameterwithrespecttotheprimaryvertexkd0,pvk<0.02. Muontrackshouldbeconsistentwiththeprimaryvertexwithinkd0,zk<0.5. Muonrelativeisolation,whichisoftenabbreviatedasRelIso,RelIso<0.15.TheRelIsoobservableiscalculatedbyrstformingaconein)]TJ /F7 11.955 Tf 12.71 0 Td[(spaceofradiusR=p 2+2=0.3centeredaroundtheleptoncandidate.Then,thesumofthetransversecomponents(excludingtheenergyoflepton)arecalculatedwithinthisconeusingallthetracksmeasuredinthetracker,ECaltransverseenergydepositsmeasuredintheECal,andHCaltransverseenergydepositsmeasuredintheHCal;andthesumofthetransversecomponentsisdividedbythetransversemomentumofthelepton. 8.2.2ElectronSelection pT>10GeV. jj<2.4,region[1.4441,1.566]isrejectedtoavoidpoorelectronreconstructionduetoagapbetweenthetwopartsofECal,endcapandbarrel. Electronsshouldpassseveralimportantidenticationselections.Theserequire-mentsareprimarilydesignedtorejectfakeelectronswhicharemis-identiedfromjetsand/orfromthephotonconversion:ii(barrel=endcap)<0.01/0.034(barrel=endcap)<0.15/0.14(barrel=endcap)<0.007/0.009H E(barrel=endcap)<0.12/0.1. Toensurethatphotonconversionisfurthersuppressed,werequirethatthereisatleastonetrackerhitinthetracker. Impactparameterwithrespecttotheprimaryvertexkd0,pvk<0.02. Electrontrackshouldbeconsistentwiththeprimaryvertexwithinkd0,zk<0.5. ElectronrelativeisolationRelIso<0.15,denedthesamewayasformuons. 68

PAGE 69

Electronsthathaveaselectedmuonwithin4R(e,)<0.1arerejected.Thisre-quirementhelpstorejectelectronswhichoriginatefromthemuonBremsstrahlungradiation!+()]TJ /F7 11.955 Tf 12.62 0 Td[(>ee). 8.2.3SelectionofTauLepton pT>20GeV. jj<2.3. reconstructionisperformedusingHPSalgorithmdescribedinsection 6.6 identicationselectionalsousesoptionstodiscriminateitsidenticationfrompoorlyreconstructedelectronsandmuons.AgainstElectronMVAissettoTRUE.AgainstMuonTightissettoTRUE. WeusetauleptonswithLooseisolationrequirement.Isolationofthetauleptonsisbooleanvariable:trueorfalseifthetauleptonisisolatedornot,respectively. Tauleptonsarecleanedwithrespecttotheselectedelectronand/ormuonwithin4R(=e,)<0.1-thisrequirementminimizestherateatwhichthelightleptoncanfakethetaulepton. 8.2.4Jetandb-TaggedJetSelections pT>30GeV. jj<2.5. PFJetsreconstructedwithParticleFlowalgorithmusinganti)]TJ /F4 11.955 Tf 12.95 0 Td[(kTalgorithmdescribedinsection 6.3 Jetsshouldsatisfyrecommendedidenticationagtoensurehighpurityjets. Jetsareremovediftheyoverlapleptonswithin4R(=e=,jet)<0.4.Thisselectiondeservesmoreattention.Ingeneral,byintroducingthiscutwearetryingtosuppressthosejetswhichyieldrealleptons.Atthisstage,therearetworeconstructedobjects,theleptonandthejet.Bothareoriginatedfromthesamesource,themotherjet.So,oneofthemshouldbekeptforfurthersteps.Theotheralternativewaywouldbetorejectleptoninsteadofjet. B-taggersatisesmediumworkingpointrecommendedbytheCMSb-taggedjet(bTAG)workinggroup. 69

PAGE 70

8.3DataSamplesThecertiedCMS-recordeddataof19.5fb)]TJ /F9 7.97 Tf 6.59 0 Td[(1wasusedforthisanalysis.ThisamountofdatawasrecordedatCMSduringthewhole2012runyearatthecenter-of-massenergyof8TeV.Table 8-1 showsthedatasamplesrecordedduringthedifferentperiodsof2012runyear. Table8-1. CMS-recordeddatausedintheanalysis. Reconstructionname RunrangeL(pb)]TJ /F9 7.97 Tf 6.59 0 Td[(1) Run2012A-13Jul2012-v1 190456-196531809Run2012A-recover-06Aug2012-v1 190782-19094982Run2012B-13Jul2012-v1 190456-1965314403Run2012C-24Aug2012-v1 198022-198523495Run2012C-PromptReco-v2 198524-2038536394Run2012D-PromptReco-v1 203854-2078987224 AllMonteCarlosamplesusedinthisanalysisaresummarizedinTable 8-2 togetherwiththeirassociatedcross-sections. Table8-2. MonteCarlosamplesusedintheanalysis. Process GeneratornameCrosssection(pb) tt MADGRAPH225Z+jets MADGRAPH3532.8ttW MADGRAPH0.232ttZ MADGRAPH0.208ttg MADGRAPH2.166ttWW MADGRAPH0.002ZZZ MADGRAPH0.005ZZ MADGRAPH0.177ZZW MADGRAPH0.019WWW MADGRAPH0.082WWZ MADGRAPH0.063 8.4TriggerSelectionandEfcienciesAsdiscussedinChapter5,triggerisdesignedforanyanalysisconductedattheCMS,andingeneralatanyhighenergyphysicsexperiment,tocollectonlinesignaleventssuitablefortheanalysis;thissearchisnotanexception.Therefore,weadoptedanassortmentofpuredileptontriggerswhichhavebeendevelopedandcommissioned 70

PAGE 71

onbehalfofseveralimportantanalysesonCMS,listedinTable 8-3 .Thesetriggersarebasedonlightleptonobjects(muonorelectron).ArateofoccurrenceofelectroncandidatesattheHLTlevelisenormousbecausejets(whichareproducedinavastmajorityofevents)canverylikelyfakearealelectronsignature;electronsthereforeemployidenticationandisolationrequirements( 8-4 )inadditiontothekinematicrequirementsonpTandtokeeptheratetoareasonablelevel.Ontheotherhand,HLTmuoncandidatessatisfykinematicrequirementsandnoisolationisusedbecauseitislesslikelytomis-identifymuonsfromjets.SincetheleptonreconstructioniswellunderstoodandapprovedbytheCMScollaboration,thesetriggershavehighefciencyforselectingonlineevents.However,itisveryimportanttomonitorthetriggerefciencyoverthedatatakingperiodtoeliminateanypossiblemistakeintheimplementationofthetriggers.Forinstance,atthebeginningofthe2012run,anunexpectedbugwasintroducedinthemuononlineselectionrequirement,whichcanbeseeninFigures 8-2A and 8-2C ;fortunately,thisbugwasxedsoon.Wemeasurethetriggerefcienciesdirectlyfromdatausingeventsrecordedinthehadronicprimarydataset.Thehadronicprimarydatasetcontainseventsthathavepassedasetoftriggersemployingjetrelatedvariableslikethenumberofjets,sumofalljetspTetc,thusthissamplecanbeconsideredasunbiasedtomeasureleptonefciencies.Eventswhichareselectedwithofinedileptonpassingtheanalysisleptonfullselectionformadenominator,andafractionoftheseeventspassingthedileptontriggersformsanumeratortodetermineefciencies.TheseefcienciesaresummarizedinTable 8-5 andshowninFigure 8-2 asafunctionofthesoftestleptonpT.Theefcienciesdonotreach100%(asonewouldexpectinanidealcase)anddependonpTor.Thereforetocorrectlypropagatetheseefcienciesintotheanalysis,weevaluateintegratedefcienciesfordifferentpTorranges,whicharesummarizedinTable 8-5 .ItisworthnotingthatthechoiceoftriggersgreatlyinuencesthepTthresholdsoftheofineleptons,aswellasisolationandidenticationrequirements.Thetriggers 71

PAGE 72

in 8-5 imposeonlineasymmetricpTcutsof8GeVand17GeV,thusleadingustoadoptofinepTthresholdsof10GeVand20GeV,respectively,inordertobeontheplateauoftheefciencycurve.OfineelectronisolationandidenticationthresholdsarealsotighterthanthecorrespondingHLTthresholdstoavoidanyinefciencyintheleptonselections.ItshouldbenotedthattheseoverallefcienciesarerelevantonlytoobtainsignalacceptanceefciencyfromthesimulationandtheMC-basedbackgroundprediction.Fortheotherbackgroundswhicharedirectlyestimatedfromdata,triggerefcienciesareautomaticallyincorporated.However,different3-leptonselectionsrequiredifferenttreatmentstocorrectlyaccountforthesescalefactors.Forinstance,forthosechannelswherewerequireexactly3lightleptons(muonorelectron),thetriggeroverallefciencyreachesroughly100%.Thisissimplydrivenbythefactthetwoleptonsaresufcienttopassthetriggerrequirementwhilethreereconstructedleptonsinaneventleadstothehigherefciency.Onecouldattempttocalculatetheseeffectiveefcienciesusingsimplecombinatorics,theprobabilitythatthedileptontriggerispassedinatri-leptoneventis: P(trigger)=pass3+3pass2(1)]TJ /F1 11.955 Tf 11.96 0 Td[(pass),(8)SubstitutingthenumbersfromtheTable 8-5 ,wecanverifythattheassumptionaboveholds.Ontheotherhand,otherchannelswithexclusivetwolightdileptonandonehadronictau,weuseexactlythesamenumbersasshownin 8-5 .Toevaluatethetriggerefciencyuncertainty,wedevelopedaprocedurewhichtakesintoaccountthemaximumdeviationalongthepTorbinswithrespecttotheaverageefciency.Wend5%deviationinthelastbiningure 8-2D ,andthereforeweassign5%systematicuncertaintytoalltriggerpaths. 72

PAGE 73

Table8-3. Asetoftriggerswhichareusedintheanalysis.IDsaredenedin 8-4 Name purpose HLT Ele17 CaloIdT CaloIsoVL Ele8 CaloIdT CaloIsoVL usedforselectingee+`channels,`=eor HLT Mu17 Ele8 CaloIdT CaloIsoVL usedforselectinge+`channels,`=e,or HLT Mu8 Ele17 CaloIdT CaloIsoVL HLT Mu17 Mu8 usedforselecting+`channels,`=or Table8-4. DenitionofelectronIDsattheHLTlevel. ID CutsinBarrel/Endcap CaloIdT H/E<0.15/0.1,ii<0.014/0.035CaloIsoVL ECaliso/ET<0.2/0.2,HCaliso/ET<0.2/0.2TrkIdVL 4<0.1/0.1,4<0.15/0.1TrkIsoVL trkiso/ET<0.2/0.2 8.5EventSelectionandDenitionofAnalysisVariables 8.5.1EventSelectionAfterimposingthetriggerrequirementswedemandtheeventpassthefollowingcriteria. 8.5.1.13-leptonSelectionTheofineselectionbeginswitharequirementthatthreeleptonsarereconstructedandpassthefullselectioncriteria,includingisolation.WerequirethatleptonseachhavepT>10GeV,withtheleadingleptonpT>20GeVinordertobeconsistentwiththetriggerrequirement.Wewillinparalleltrackthenumberofeventswhichareusedtopredictbackgroundyields.Namely,wewillselecteventswhichfeaturetwo(one)isolatedandone(two)non-isolatedleptons,failingtheisolationcut.Theseeventsarereferredtoassidebandevents.Atthisstage,wealsorequireaparticularavorandchargeconguration.Becauseoursignaltopologyinvolvesbothcharginoandneutralino Table8-5. Overalltriggerefciencies. TriggerName pT<20GeV pT>20GeV HLT Ele17 ID Ele8 ID 0.820.01stat 0.950.002stat HLT Mu8(17) Ele17(8) ID 0.860.01stat 0.930.002stat jj<1.0 jj>1.0 HLT Mu17 Mu8 0.900.003stat 0.810.004stat 73

PAGE 74

A B C DFigure8-2. DileptontriggerefcienciesasafunctionofsoftestleptonpTand.A)showstheefciencyofdouble-muontrigger.B)showstheefciencyofdouble-electrontrigger.C)showstheefciencyofelectron-muontrigger.D)showstheefcienciesforallthreetriggerpathsasafunctionofpseudo-rapidity. decays,weexpectthatatleastonepairofleptonswillbeoppositelychargedandofthesameavor.Therefore,weimposethisrequirementattheoutset.BecausethepTspectraoftheleptonswillbeinuencedbythemassscalesoftheSUSYparticles,thesignalacceptanceisexpectedtobemodeldependent.Thereshouldbeatleastoneopposite-signpairofleptons.WefurtherclassifyeventsbasedontheavorcombinationoftheleptonswhicharesensitivetodifferentSUSYscenarios. 74

PAGE 75

OSSF-Eventsthatcontainoneofthefollowing3-leptonevents:(),()e,(ee)eor(ee).Wenaturallyexpectthesecombinationsfromasignalmodelwhereneutralinodecaysdemocraticallyvialeft-handed(avor-democratic)orright-handedslepton(tau-enriched),yieldingoppositesignandsameavorleptonpairtoobeyleptonavorconservation.Whereascharginocandecaydemocraticallyvialeft-handedsleptonsorviaright-handedstau,asdescribedbefore.Inthelattercase,tausarerequiredtodecayleptonically. NoOSSF-Eventsthatcontainoneofthefollowing3-leptonevents(e)or(e)e.Thisselectionissensitivetotau-dominatedscenario,whereneutralinoandcharginoalldecaytoright-handedstau,followedbytauleptonsdecayingleptonically. SSTau-Eventsthatcontainoneofthefollowing3-leptonevents(),(e),()eor(e)e.Thesechannelswerealsodesignedtomaximizethesensitivityforthetau-dominatedscenario. 8.5.1.2LowInvariantMassVetoOncewehaveasampleofclean3leptons,wefurtherrequirethatallpossibleopposite-signdileptonpairsformaninvariantmassabove12GeV,whichhelpstosuppressleptonpairsfromheavy-avordecays.Thisrequirementalsohelpsmitigatetheformidablebackgroundfromlow-massDrell-Yan.Insignalsamples,thedileptoninvariantmassisingeneraldictatedbythemasssplittingsoftheSUSYparticles;butatagoodapproximation,signalacceptanceisnegligibleforlowmass(M``<12GeV). 8.5.1.3b-TaggedJetVetoAftertheinvariantmasscut,weimposeanaggressivevetoonb-taggedjets.Thisisdoneprimarilytocleanthesamplefromeventswithtop-quarkproduction,forexample,thereductionoftteventsismorethanafactorof2.OtherrareSMprocesseslikettVarealsoreducedapproximatelywiththesameorder.Figure 8-4 showsthebackgroundexpectationforb-taggedjetmultiplicity.Thesignalisnotexpectedtobeassociatedwithanystrongproductiontorstorder.However,jetsareexpectedfrominitialstateradiation(ISR)orpile-up,whichcaninprinciplefalselymimicb-taggedjets.But,theb-taggingalgorithm,whichisusedinthisanalysis,hasverylow(approximately1%)mis-tagrate. 75

PAGE 76

Therefore,thesignalacceptancewithrespecttotheb-vetowillbemodelindependenttoagoodapproximation,asshowninFigure 8-3 A B CFigure8-3. b-taggedjetvetoefciencyintheavor-democraticsignalsamplesusingtheOSSFselection.Anumeratorcontainsthetotalnumberofeventswith3leptonsandnob-taggedjet,whereasadenominatorcontainsthetotalnumberofeventswith3leptonsandb-taggedjetrequirementrelaxed.ThereisnocutappliedtotheM``andMTvariablesforboththenumeratoranddenominator.Forallthreesignalsamples,efcienciesareatandreachapproximately98%.Thesameistruefortheotherscenarios.A)correspondstox~`=0.05.B)correspondstox~`=0.5.C)correspondstox~`=0.95. 76

PAGE 77

Figure8-4. Thebackgroundexpectationforb-taggedjetmultiplicityusingsimulation. 8.5.1.4LargeMissingEnergyFinally,werequireEmissT>50GeV.Fromoursignalmodels,weexpectthreeinvisibleparticles(twoLSPsandneutrino,asshowninFigure 8-1 )whichescapedetection.Thiswillinevitablyleadtomissingenergysignals.ThescaleofEmissTinthesignalwillbeinuencedbythedifferenceinmassscalesofthesupersymmetryparticleswhichareproducedinthedecaychain.Thus,thesignalacceptanceisexpectedtobemodeldependent,asshowninFigure. 8-5 .ThiscutalsohelpsustogreatlyreducetheZ+jetsbackgroundprocess,whichhasahugecross-section,astheonlypossiblesourcesofEmissTintheprocessareneutrinofromthedecayofheavyavorjetsand/ortheenergymis-measurementofjetsor,toalesserextent,ofleptons.Figure 8-6 showsEmissTdistributionsfromtheexpectedSMbackgrounds,normalizedtoaluminosityof19.5fb)]TJ /F9 7.97 Tf 6.58 0 Td[(1. 8.5.2VariableDenitionsItwasfoundthatM``,MTandEmissTprovidethebesthandlesforsegregatingsignaleventsfromthedominantSMbackground,WZproduction.InWZ,M``willbeclusteredaroundtheZmassandMTisconnedtobebelowtheWmass,aslongasthetheyareproducedon-shell.Ontheotherhand,aninvariantmassdistributioninthesignalsamplespopulatesinthehightailsofthedistribution.However,allthree3-lepton 77

PAGE 78

A B CFigure8-5. DistributionofEmissTintheavor-democraticsignalsamplesusingtheOSSFselection.NorequirementisappliedtotheM``andMTvariables.EmissTisroughlydrivenbythemasssplittingbetweenLSPandslepton.TodemonstratehowEmissTisdistributedindifferentregionsofSUSYmassspace,weconsidertwocharginomassesof400GeVand800GeV,whereastheLSPmassesaretakenclosetozeroandclosetothediagonal.Asexpected,diagonalregionfeatureslowEmissT,whileinthebulkEmissThasusuallyhighervalues.Theothersignalsampleshavethesamefeature.A)correspondstox~`=0.05.B)correspondstox~`=0.5.C)correspondstox~`=0.95. 78

PAGE 79

A B CFigure8-6. DistributionofEmissTfortheSMMCprocessesfordifferent3-leptonselections.ThereisnorequirementonMTandM``.A)showstheOSSFselection.B)showstheNoOSSFselection.C)showstheSSTauselection. selectionscannotbetreatedwiththesamedagreeregardinganumericalvalueofZmass. 8.5.2.1DenitionofDileptonInvariantMassM``iscalculatedusingadileptonpairwithoppositesigncharges.Dependingonthe3-leptoneventselection,wefurthertakeintoaccounttheavorsofadileptonpairtocalculateM``: forOSSF,same-avorandopposite-signpair(oree)isused.SinceitisassumedthatbothleptonsoriginatedirectlyfromZbosondecay,thuswithitsrealmassof91GeV; 79

PAGE 80

forNoOSSF,opposite-avorandopposite-signpair(e)isused.ThischannelisexpectedfromthescenariowhenZbosondecaysintoapairofleptons,withbothleptonsfurtherdecayingleptonically.Astheneutrinostakeplaceintheleptondecay,theZmassvalueisshiftedandthereforeitistakenas60GeV; forSSTau,opposite-avorandopposite-signpair(ore)isused.Thischannelisalsodesignedtomaximizethesensitivityfortheabovescenario.However,weconstructdileptoninvariantmassusingonetauleptonandonelightlepton(e/),andthemassvalueistaken50GeV.DifferentSMprocessescontributetodifferentregionsofM``distribution.MostSMprocessestendtohavelowvaluesofM``,withmulti-bosonbackgroundslocalizingaroundtheZmassregion.Multi-bosonprocessisthedominantoneinthisanalysis,thereforetheZregionisbackgroundenriched,asindicatedinFigure 8-7 .TheprocedureofassigningadileptonpairtohypotheticalZbosoncanbeambiguous,asinsomecasestherearetwopairsofdileptonwhichcanbeusedinM``calculation.wechooseasimplemethodtosettletheambiguity.WhicheverpairformsaninvariantmassclosesttotheZmasswillbeused.Thisapproachhasamainbenetinthefactthatitistrivialtoimplementalgorithmicallyanditisaccurateforthevastmajorityofeventsthatentertheanalysis(expectedtobeWZ).However,forthesmallbackgroundwhichcanpossiblycomefromW,suchanalgorithmwilllikelymaketheincorrectassignment.Aslongastheprobabilityforamis-assignmenttooccuriswellmodeledinMC,wedonotexpecttoincuranybiasestoourbackgroundpredictionsorcalculatedsignalacceptance.Weperformasimpletestusingeandeeevents,pretendingthatallleptonsarethesameavor.WecalculatedthefractionofeventsinwhichtheepairisselectedastheZcandidateindataandinsimulation.ThereisagoodagreementbetweendataandMC,andtheprobabilitytoreconstructaZmassusingthewrongpairisapproximately10%. 80

PAGE 81

A B CFigure8-7. DistributionofM``fortheSMMCprocessesfordifferent3-leptonselections.EventareselectedwithEmissT>50GeVandMT>0GeV.A)showstheOSSFselection.B)showstheNoOSSFselection.C)showstheSSTauselection. 8.5.2.2DenitionofTransverseMassTheMTvariableisdenedusingthethirdlepton,whichdoesnotparticipateintheM``calculation,andmissingtransverseenergyaccordingtothisformula: MT=q 2p`TEmissT(1)]TJ /F5 11.955 Tf 11.95 0 Td[(cos(`)]TJ /F7 11.955 Tf 11.96 0 Td[(EmissT)),(8)TheMTdistributionisgreatlyaffectedbythepresenceofEmissT.AllbackgroundsyieldingnorealEmissTdominatethelowpartofthedistribution,whereasthehighertailsofthe 81

PAGE 82

MTdistributionarepredominantlydrivenbythebackgroundswithrealandhighEmissT.Forinstance,theWZtransversemassdistributionpeaksaround80GeVandthenfallsoffrapidly,asshowninFigure 8-8 .SignalsamplestendtoresultinhighMTdueto A B CFigure8-8. DistributionofMTfortheSMMCprocessesfordifferent3-leptonselections.EventareselectedwithEmissT>50GeVandM``>12GeV.A)showstheOSSFselection.B)showstheNoOSSFselection.C)showstheSSTauselection. highEmissT,butinprincipleitisindirectlydictatedbytheSUSYmasssplittings.Figure 8-9 showstheMTdistributionforOSSFeventsfordifferentmasssplittingsinthesignalsample.ThehighMTregionishenceacleanregionseparatedfromthebackgroundwhichoffersagoodhandletosearchfornewphysics. 82

PAGE 83

Figure8-9. DistributionsofMTusingOSSFeventsintheavor-democraticsignalsample.HerethesameSUSYmassphasespaceisconsideredasinFigure 8-5 .EventareselectedwithEmissT>50GeVandM``>12GeV.Forsimplicity,onlyx~`=0.5isshown. 8.5.3DenitionofBaselineandSearchRegionsHavingimposedtheeventselectionrequirementsabove,wearenowlefttodeneoursearchregions(SR).WedeneexclusivesearchregionsusingM``,MTandEmissTvariables.SinceweexpectfromthesignalmodelsthatthesevariablesareinuencedbytheSUSYmasssplittings,wedesignedsearchregionsinawaytocoverallpossiblecornersofthephasespace,maximizingthesensitivitytothedifferentSUSYmodels.ThenumberingschemeforthesearchregionsaresummarizedinTables 8-6 8-7 and 8-8 .Wealsodenearegioncontainingareasonableamountofeventsinthesamplewhichwillbeusedtotestthebackgroundestimationmethods.ThisregionshouldbeenrichedwiththeSMprocessesinordertoeliminateanybiasfromthesignalcontamination.Thereasonableregioncanbedenedusingthefollowingselections:3leptons,EmissTabove50GeV,andnob-taggedjetinanevent.NopriorMTorM``requirementsareappliedunlessmentionedotherwise.Thisregionwillbereferredtoasbaselinethroughoutthethesis.Scatterplotsfortheobservedeventsforthreedifferent3-leptonselectionscanbefoundinFigure 8-10 83

PAGE 84

Table8-6. Denitionofsearchregions(SR)fortheOSSFselection. MT(GeV) EmissT(GeV) M``<75(GeV) 75105(GeV) 50-100 SR1 SR16 SR31 0-120 100-150 SR2 SR17 SR32 150-200 SR3 SR18 SR33 200-inf SR4 SR19 SR34 50-100 SR6 SR21 SR36 120-160 100-150 SR7 SR22 SR37 150-200 SR8 SR23 SR38 200-inf SR9 SR24 SR39 50-100 SR11 SR26 SR41 160-inf 100-150 SR12 SR27 SR42 150-200 SR13 SR28 SR43 200-inf SR14 SR29 SR44 Table8-7. DenitionofsearchregionsfortheNoOSSFselection. MT(GeV) EmissT(GeV) M``<100(GeV) M``>100(GeV) 50-100 SR46 SR76 0-120 100-150 SR47 SR77 150-200 SR48 SR78 200-inf SR49 SR79 50-100 SR51 SR81 120-160 100-150 SR52 SR82 150-200 SR53 SR83 200-inf SR54 SR84 50-100 SR56 SR86 160-inf 100-150 SR57 SR87 150-200 SR58 SR88 200-inf SR59 SR89 Table8-8. DenitionofsearchregionsfortheSSTauselection. MT(GeV) EmissT(GeV) M``<100(GeV) M``>100(GeV) 50-100 SR91 SR121 0-120 100-150 SR92 SR122 150-200 SR93 SR123 200-inf SR94 SR124 50-100 SR96 SR126 120-160 100-150 SR97 SR127 150-200 SR98 SR128 200-inf SR99 SR129 50-100 SR101 SR131 160-inf 100-150 SR102 SR132 150-200 SR103 SR133 200-inf SR104 SR134 84

PAGE 85

A B CFigure8-10. DistributionoftheobservedyieldsintheMT-M``spaceafterEmissTabove50GeV.A)showstheNoOSSFselection.B)showstheSSTauselection.Thesetwochannelsaresensitivetotau-dominatedSUSYscenario.B)containsOSSFeventstoexploreavor-democraticandtau-enrichedscenarios.Pinkdashedlinesaredrawntoseparatethesearchregions. 8.6BackgroundEstimationMethodsThissectionaddressesthebackgroundestimationmethodsforthissearch.InHEP,thekeyaspectofallkindsofsearchesistounderstandthebackgroundcompositionforthesearchesandthentoseehowcompatibleitcanbewiththeobservedsignalyields,whichwouldindicateanypossiblepresenceofnewphysics.Inthisanalysis,weemployseveralbackgroundestimationmethods,withmostofthembeingdata-driven(DD)techniques.Data-drivenmethodensuresahighaccuracyofthemethodbecause 85

PAGE 86

itallowsustobesimulationindependent-asthemostsensitivesearchregionstolookfornewphysicsaredenedatthetailsoftheMTandEmissTvariables,MCbackgroundpredictionsbecomelessreliablebecause,inmostcases,thesampleisstatisticallylimited(generatedeventsisnotenough)ortheNLOproductioncross-sectionmaynotbeaccurateatthetails.WearelookingfornewphysicsbeyondtheStandardModelwiththesignatureof3leptons,missingtransverseenergyandnob-taggedjet.NottoomanyStandardModelprocessesthatareknowncancontributetothistopology.Afterapplyingthesearchregionselectioncriteriatoreducethenumberofbackgroundevents,therearestillirreducibleeventcandidatesleftinthesignalsample;althoughsmall,itisvitaltoknowallsourcesofthebackground.ThebackgroundsassociatedwiththisanalysisareWZ,tt,Z+jetsandotherrareSMprocesseswhichhavenotbeenmeasuredyetattheacceleratorsduetotheirsmallproductioncross-sections.However,theyaretheoreticallycalculatedwithintheSMandtherearegoodreasonstotakethemintoconsideration.Theyaresimulatedusingthefullsimulationprocedure.Thebackgroundcanbecategorizedbasedonhowitistreated: Backgroundduetofake(non-prompt)leptons(e,and).Aleptonisconsideredfakeornon-promptifitdoesnotoriginatefromthedecayoftheWandZbosonsortheSUSYparticles.Itisworthnotingthatfakeisageneralizationhere.Inotherwords,afakeleptoncanbetherealleptonwhichdescendsfromajet(e.g.,fromthedecayofKorparticles)oritcanbereallyfakeonewhichwasmis-reconstructedfromajet.IfaleptondescendsfromthedecayoftheWandZbosonsortheSUSYparticles,wethenconsideritasthepromptlepton.Dependingontheeventselection,weemploydifferentDDmethods.TheyarepresentedinSection 8.6.1 8.6.2 and 8.6.3 ; Backgroundduetotheprompt3-leptondescribedinSection 8.6.6 .HerewefurtherdistinguishWZproduction,whichisthemajorbackgroundforthemostofthesearchregions,fromtheotherrareSMprocesses.TheWZbackgroundiscloselyexaminedusingdata-driventechniquesinSection 8.6.4 ;and Backgroundduetophotonconversion.Thedata-drivenmethodtoestimatethistypeofbackgroundispresentedinSection 8.6.5 86

PAGE 87

8.6.1DeterminationofBackgroundduetoNon-PromptLightLeptonsAsMonteCarlo-basedstudiessuggest(Figure 8-7 8-8 and 8-6 ),itisexpectedthat,dependingonthe3-leptonselection,thebackgroundcomponentduetothenon-promptleptoncanhaveaconsiderablecontribution;although,allvariablesandtheirthresholdsassociatedwiththisanalysisdramaticallyreducethiscontribution.Forinstance,theminimumthresholdonthetransversemissingenergyof50GeVsignicantlyreducesZ+jetsevents(thisprocessismorerelevantfortheOSSFchannel)wherenogenuineEmissTisexpected,butjetenergymis-measurementandpile-upresultinEmissTuctuationsandcanarticiallypopulateeventsintothesearchregions.Anotherprocesstt(thisprocesscancontributetoallthree3-leptonselections)isusuallyproducedwithhighEmissT,butwithtwogenuineb-quarks.Therefore,theb-taggedjetvetorequirementfurtherreducesit.However,asmallfractionoftheseeventscanstillpasstheb-vetorequirementduetotheinefciencyoftheb-taggingalgorithm1orthekinematiccutsonpTorofb-quarksbeingoutoftheacceptance.Themethodforestimatingnon-promptbackgroundisessentiallybasedonthefactthatthenon-promptlightlepton(e/),whichoriginatedfromthejetactivity,comeswithinthehadronicenvironmentaroundit,thustendingtohavelooseisolation.Wecalculatetheprobabilityofsuchnon-isolatedleptons(referredtoasLooseleptons)topasstheanalysisisolationcutof0.15(referredtoasTight).ThisTight-to-Loose(TL)ratioisthenappliedtotheevents,inwhichtwoleptonspassthefullleptonselections,includingisolation,andthethirdleptonfailstheisolationcut.Asmentionedbefore,theseeventsarereferredtoassidebandevents.Forinstance,thecaseofZ+jetsprocess,twoisolatedleptonscomefromtheZbosondecayandthethirdleptonmustbefromnon-promptsources,predominantlyfromlightavorjets;whereasintt,twoisolated 1Theb-taggingalgorithmwhichisemployedthroughoutthisthesishasapproximately70%efciency 87

PAGE 88

leptonsdescendfromtwo(t!b+W)Wbosonsdecaysandthethirdleptonoriginatesfromnon-promptsources,mostlyfromtheheavyavorjets.Itmustbementionedthatweonlyconsidersidebandeventswithonlyonefakelepton.ThismethodalsopredictsthecontributionfromotherSMprocesses,suchasWW+jets,ZZ+jets(whereoneZbosondecaysleptonicallyandtheotherZbosondecaysintohadronicallyorintotwoneutrinos.)Inprinciple,onemustalsotakeintoaccounttheprocessesyieldingtwoorthreefakeleptonssuchasW+jetsorQCD,butthiscontributionisnegligibleand,therefore,omittedfromtheanalysis.Noconstraintisimposedontheisolationofthenon-isolatedleptoninthesidebandeventsfromthetrigger(itisassumedthattwoisolatedofineleptonssatisfytriggerrequirements),therefore,isolationspansfrom0toinnity.WecanattempttomodelTLdistributionsformuonsandelectronsbystudyingQCDeventswhichfeaturebbproduction.InQCDmulti-jetdata,hereafterreferredtoascontrolsample(CS),wemeasureTLwhichisparameterizedinleptontransversemomentum.CSisdenedas: atleastoneelectron/muonwithpT(`)>10GeVthatpassesthefullleptonIDrequirementsexcepttheimpactparameterandisolationrequirements; weapplyaZ-vetointhecaseoftwoleptons:rejecteventswhereadileptonpairformsaninvariantmassthatiswithinawindowof10GeVaroundMZ; MT<15GeVtosuppresstheW=Zcontamination; EmissT<15GeVtofurthersuppresstheW=Zcontamination; atleast2jetswithjj<2.4andpT>30GeV; exactly1jetthatpassestheb-jetidentication.Thesameb-jetdiscriminatorisusedasforthesignalselection; inordertopreventbiasesinthemeasurementofnon-promptleptonisolationefciencies,weselectprobeleptonsawayfromtheb-taggedjet,(b)]TJ /F4 11.955 Tf -372.7 -14.44 Td[(jet,probelepton)>2; 88

PAGE 89

onlyleptonswithpT40GeVareconsideredfortheTL.Thisisdoneinordertoreducecontaminationfrompromptleptonsources(e.g.tt);themotivationonthischoiceistheassumptionthatbeyondpT=40GeVTLremainsat. A BFigure8-11. TLobtainedinthecontrolsampleusingdataandQCDMC.A)showsTLformuons.B)showsTLforelectrons.Theerrorbarsincludeonlystatisticaluncertainty. Figure 8-11 showsTight-to-LooseforelectronsandmuonsmeasuredindataandinQCDMCcontrolsamples.Wendthattheprobabilityfornon-promptleptonstopasstheisolationcriteriaofRelIso<0.15isaround2-3%formuonand6-7%forelectron.OnewouldwonderaboutthedisagreementbetweendataandMCinFigure 8-11 ;however,itisworthnotingthataperfectagreementbetweenthetwoisnotexpected.TLmeasuredinMCcouldbeaffectedbythesimulationeffectsorTLmeasuredindatacanbeinuencedbythecontaminationofthepromptleptonstosomeextent.Aslongasthemethodremainsdata-driven,wearemostlyinterestedinlookingatthedata.Itisveryimportanttovalidatethemethod.Todoso,wetestthemethodologyinthesimulation.WeconsidertwoMCprocessesttandZ+jetstoobtainobservedsignaleventsaswellassidebandevents.WealsocomputeTLinQCDMCandthenapplyittothesidebandevents.TheclosuretestobtainedinttsampleisshowninFigure 8-12 89

PAGE 90

TheseguresdemonstratethatwecanpredicttheshapeofEmissT,MTandM``usingthismethod.Thishasagreatimportanceasoursearchregionsareformedbyemployingthesevariables.Table 8-9 showsthepredictedandobservedyieldsforbothprocessesinthebaselineregion. Table8-9. ClosuretestinthebaselineselectionfordifferentOSSFchannels. Sample (+)]TJ /F5 11.955 Tf 7.08 -4.34 Td[()(+)]TJ /F5 11.955 Tf 7.08 -4.34 Td[()e(e+e)]TJ /F5 11.955 Tf 7.09 -4.34 Td[()(e+e)]TJ /F5 11.955 Tf 7.09 -4.34 Td[()eTotal tt Predicted51.981.6956.481.7348.761.6110.790.74168.013.00 Observed51725117191 Z+jets Predicted3.20.432.10.32.80.41.10.29.00.7 Observed542213 AscanbeseeninTable 8-9 ,thepredictionandobservednumberofeventsagreewell.Weusethemaximumrelativediscrepancybetweenthepredictedandobservedyieldsinttasoneofthesourcesforthesystematicuncertainty.Weassigna30%systematicuncertaintyonthepredictedbackgroundmotivatedbytheleveloftheclosuretestofthemethod. 8.6.2DeterminationofBackgroundduetoNon-PromptElectronsAsdiscussedpreviouslyinSection 8.4 ,thissearchusesdileptontriggerstoselecteventsonlinewithatleasttwoleptons.Atthetriggerlevel,muonshavenoisolationcutapplied,whereastheelectronpartemploysisolationwhichhasalooservaluethantheisolationcut(0.15)usedinthisanalysis.TheOSSFandNoOSSFselections,wherewerequirethreelightleptons,arenotaffectedbythetriggerconstraintasweonlyconsidersidebandeventswithtwoisolatedandonenon-isolatedleptons.However,specialcareisneededforthechannelswherewerequireanexclusivesame-signlightdilepton(withatleastoneelectron)pairandtaulepton(SSTau).Thepreviousdata-drivenmethod,whichisbasedontheisolationmodelingforelectronsisnolongerapplicableduetothetriggerconstraintandisasubjectofmodication.Themaindifferenceisintroducedbytheisolationrequirement(0.15
PAGE 91

A B CFigure8-12. ClosuretestobtainedinttMCsample.Eventsareselectedwithnob-taggedjetand3-leptonwheretwoofthempassingthefullleptonselectionsincludingisolationandthethirdleptonhasinvertedisolation.Tight-to-LooseisalsoobtainedinQCDMCcontrolsample.A)showstheclosuretestasafunctionofM``.B)showstheclosuretestasafunctionofEmissT.C)showstheclosuretestasafunctionofMT. 91

PAGE 92

toasTightandLoose,respectively,andweobtainthenumberofelectronspereventintheseregions.Tight-to-Loose(TL)isessentiallyusedthesamewayasitwasdescribedintheprevioussubsection.Forthecontrolsample,weselecttheQCDmulti-jetdata,usingasetoftriggersrequiringonelooselyisolatedelectron(withthesameisolationrequirementasindileptontriggers)andonejetwithpT>30GeV.CSisdenedas: exactlyoneelectronwithpT(`)>10GeVintheeventthatpassesthefullleptonIDrequirementsexcepttheimpactparameterandisolationrequirement; electronwithimpactparameterinvertedd0>0.05mmforbothlooseandtightelectrons.Thischoiceismadetosuppressthepromptelectroncontamination,promptelectronisolationcanfailduetoaphotonwhichisemittedbytheelectron. MT<40GeVtosuppresstheW=Zcontamination. EmissT<20GeVtofurthersuppresstheW=Zcontamination.Thisrequirementislooserthaninthepreviousmethod.Thiswasprimarilydoneinordertoallowmoreeventsinthecontrolsample. atleast2jetswithjj<2.4andpT>30GeV. exactly1jetthatpassestheb-jetidentication.Thesameb-jetdiscriminatorisusedasforthesignalselection. inordertopreventbiasesinthemeasurementofnon-promptleptonisolationefciencies,weselectprobeleptonsawayfromtheb-taggedjet,(b)]TJ /F4 11.955 Tf -372.7 -14.44 Td[(jet,probelepton)>2; onlyleptonswithpT40GeVareconsideredfortheTight-to-Loose(TL)templates.Hereagain,itisimpliedthatforpT>40GeVTLratioremainsatasafunctionofelectronpT.Asexpected,controlsampleisdominatedbytheheavyavorjets,asshowninFigure 8-13 .Tight-to-LoosemeasuredindataandMCcontrolsamplesisshowninFigure 8-14 .Figure 8-14C showsthattheTLratioisstablewithinthestatisticalerrorwithrespecttopTthresholdsofthejet,whichisclosesttotheleptonofinterest;inagoodapproximation,thisleptonindeeddescendsfromthisjet.WederiveTight-to-Loose 92

PAGE 93

Figure8-13. Isolationdistributionofelectronswhichdescendfromdifferentmotherjetsinthecontrolsample. inthecontrolsampleindataandapplyittothesidebandeventsthatweselectedforeachsearchregion.Again,weproceedtovalidatethemethodbydoingaclosuretestinthesimulation,demonstratingthelevelofagreementbetweenobservedandpredictedyieldsinttandZ+jetssamples.Inordertogainstatistics,weuseabaselineselection,exceptEmissTrequirementisrelaxed,butdroptherequirementinanevent.FortheBaselineselection,thesidebandoftheZ+jetssamplecontainsasubstantialfraction(30%)ofpromptelectrons(Figure 8-15 )whichfailtherelativeisolationrequirementduetothebremsstrahlungphotonsthatisemittedbytheelectronsofinterest.Therefore,intheZ+jetssample,wevetoeventswith75
PAGE 94

A B CFigure8-14. TLobtainedinthecontrolsampleusingdataandQCDMC.A)showsTLforelectronsinMC.B)showsTLforelectronsindata.C)showsTLforelectronsindataandMCfordifferentclosest-jetpTthresholds.Theerrorbarsincludeonlystatisticaluncertainty. distribution.Wendanexcessof24eventsinaselectedTightsampleof95eventsandderiveasystematicuncertaintyof26%. 8.6.3DeterminationofBackgroundduetoNon-PromptTauLeptonsIntheprevioustwosections,themethodologiesfordealingwithlightnon-promptleptons(electronandmuon)weredescribed.Thissectionaddressesadata-drivenmethoddesignedforabackgroundemergingfromthenon-promptlepton.InaSSTau 94

PAGE 95

A B CFigure8-15. Originsofelectrons.Solidareashowsthenumberofeventswithcorrectchargereconstruction,whereasdashedareashowsthenumberofeventswithchargemisidentiedelectrons.Compositionoftheselectedelectronsinbothsignalandsidebandregionsasafunctionofrelativeisolationareshown.A)isobtainedusingthettMCsample.B)isobtainedusingtheZ+jetsMCsample.Orangelinedividesthehistogramsintotightandlooseregions.C)showstheinvariantmassdistributionoflooseelectronsinthesidebandselectionusingtheZ+jetsMCsample.Orangelinesindicatevetoedregions. Table8-10. Closuretestforthepredictionofthebackgroundduetoelectronsarisingfromnon-promptsources.Predictedandobservedeventyieldscorrespondtothebaselineselection,denedwithtwoselectedisolatedleptonsandzerob-taggedjets.Therelativecontributionofeachsourceofelectronstotheselectedsidebandisindicated.Thedeviationofthepredictedfromtheobservednumberofnon-promptelectronsintheTightregionisconsistentwiththecontaminationofthesidebandwithpromptelectrons. Samplefromb-jetsfromlightjetsprompteepred)]TJ /F4 11.955 Tf 11.96 0 Td[(obs obs tt87%12%1%Predicted99.08.52.1%Observed97 Z+jets20%74%6%Predicted278.719.35.6%Observed264 95

PAGE 96

Figure8-16. DistributionofTightelectronsinthecontrolsampleusedtoestimatethesystematicuncertaintyofthemethod.Reddashedlineshowsthetwithaconstantperformedintherange0.01
PAGE 97

single-fakeanddouble-fakebackgroundtopologiesmustbeconsideredinthenalbackgroundestimation.Theoverlapbetweenthetwotopologiesisappropriatelytakencareofbysubtractingthenalestimatedyields. Table8-11. ClassicationofSSTaueventsoriginatedfromdifferentsources.Forthisexercise,thereisnorequirementappliedtob-taggedjet,EmissTorM``.Itshouldbenotedthatbothlightleptons'pTareabove10GeVandthepTofthetauisabove20GeVforthepurposeofincreasingstatisticsofthesample.pandfindicateiftheleptonispromptorfake,respectively. Sample3-p2-p,1-f()2-p,1-f(e/)1-p(e/),2-f tt010210400WZ+jets5372810150 Table8-12. Summaryofnon-promptleptonoriginsinsimulatedttevents. SourceNNe= b82325c364s800lightavor(ud)13422lightavor(g)30unmatched5239 Itrarelyhappensthatallthreesignalleptonsarenon-prompt;forexample,thisscenariocantakeplaceinQCDortt(W!qq0)processes,butanalysisvariableslikemissingtransverseenergyandb-taggedjetvetoreducethiscontributiontothenegligiblelevel.Consequently,itisnotconsideredelsewhereintheanalysis.Themethodtoestimatethefakebackgroundisbasedondeningtheprobabilityofanon-isolatedleptontopassthesignalisolationrequirement.ThismethodresemblestheTight-to-Loose(TL)methodandtherefore,itwillalsobereferredtoasTL.However,thetwomethodsarecompletelyuncorrelated.WemeasureTLindatausingaspecicallydesignedcontrolsampledenedas: eventswithopposite-signandsame-avorlightdileptonpair(eor),withaninvariantmassconsistentwiththeZbosonmassof9015GeVwindow; eventswithanadditionalthirdlightleptonarevetoedinordertoeliminateanybias;and 97

PAGE 98

eventswhichcontainmissingtransverseenergyabove30GeVarerejectedtogetridofthemostbackgroundprocessesexceptZ+jets.Thechoiceofthecontrolsampleisacriticalpartofthemethod.ItshouldreectthesimilaritybetweenCSandsignalregionsintermsofkinematicandlepton-originpointofview.HavingseenTables 8-12 8-13 andFigure 8-18 ,thechoiceofCSisfairlyvalidatedastheZ+jetsprocessismostlyproducedinassociationwithlightavorjets,whichconsequentlyrepresentsthemajorsourceoffake.Figure 8-17 showstheTLfordifferentsourcesinZ+jetsCS. Table8-13. Summaryofnon-promptoriginsinZ+jetsCS. SourceNnumeratorNdenominator b3298033c84017451s283032212lightavor(ud)12127163318lightavor(g)99951064unmatched311120022Total20236292100 Figure8-17. TLofthetauleptonfordifferentsourcesobtainedintheZ+JetMCsampleasafunctionoftauleptonpT.Reconstructed(inthenumeratorandthedenominator)arematchedtoMCtruthparticleswithin4R(reco,gen)<0.5. Astheeventsareselectedwiththedileptontriggers,thereisnolimitationimposedonthecandidateselectionintermsofonlinekinematicandqualityrequirements.TLismeasuredasafunctionofpT,andnumberofreconstructedverticesandtheycan 98

PAGE 99

Figure8-18. pTdistributionofLoosetauleptonsintheZ+Jetandttsamples.Distributionsarenormalizedtotheoverallarea.NosignicantdifferencesinthepTshapearepresentbetweenthetwosamples. befoundinFigure 8-19 and 8-20 .Asstatedpreviously,thismethodisfullydata-drivenbutwealsocheckTLinsimulationasanadditionalcrosscheck,andwendagoodagreementbetweenthedataandMonteCarlo.However,thereisnoexpectancyofperfectagreementineachbin,forinstance,mis-modelingofjetscouldgreatlyinuenceTLinMCasreconstructedpredominantlyoriginatesfromjetsinthecontrolsample.Figures 8-19 and 8-20 indicatethatTLhasandpile-updependencies.GiventhehighstatisticsinCS,weparametrizeTLratiointhreevariablestominimizethedependence:pT,anddifferentrangesofpile-up.TheycanbefoundinFigure 8-21 Tonallyvalidatethemethod,weapplyTL,whichwasobtainedinMCCS,tottandZ+jetsMCsamples.Theclosuretestshowsgoodagreementbetweenobservedandpredictedevents.ThemainsourceofsystematicuncertaintyforthemethodisdrivenbythedifferencebetweenCSandavarietyofbackgroundprocesseswhichweaimtoestimateindata.IntheSSTaueventsectionwithEmissTabove50GeV,Z+Jetprocessbecomesnegligibleanditisnotexpectedtoconsiderablycontributetoanysignalregions,leavingttandperhapsasmallfractionofW+Jetsprocessesasthedominantnon-promptbackground.Weassigna30%systematicuncertaintytotheTL 99

PAGE 100

A B C DFigure8-19. TLdistributionsasafunctionofpTandoftaulepton.A)showsTLasafunctionofpTineventswithand.B)showsTLasafunctionofpTineventswitheeand.C)showsTLasafunctionofineventswithand.D)showsTLasafunctionofineventswitheeand.Theerrorbarsincludeonlystatisticaluncertainty. methodbasedonthemaximumdeviationbetweentheobserved(396)andtheexpected(55123)yieldsinFigure 8-22 .HavingobtainedTLandvalidatedthemethod,wethenproceedtoestimatethistypeofbackgroundindatabyapplyingTLtothesidebandeventsinallsignalregions,whichareselectedas: twosame-signisolatedlightleptonspassingallrequirementsandatleastoneoppositesignnon-isolatedtoestimatesingle-fakecontribution;and 100

PAGE 101

Figure8-20. TLasafunctionofvertices.BothchannelsfromFigure 8-19 arecombined. oneisolatedandonenon-isolated(Reliso>0.15)lightleptonhavingthesamechargeandatleastoneoppositesignnon-isolatedtoestimatedouble-fakecontribution.WeuseTLfromthetwopreviousmethodsforlightleptons. 8.6.4DeterminationofBackgroundduetoWZprocessThelargestbackgroundcontributioninthisanalysiscomesfromtheStandardModelWZprocess,wherebothbosonsdecayleptonically.WeestimatethisbackgroundfromtheWZsimulation,whereseveraleffectscaninuencetheacceptanceefciency,resultinginoverestimatingorunderestimatingtheprediction.Theseeffectsariseduetomis-modelingofthedetector,incompleteknowledgeofthecross-sectionbeyondnext-to-leadingorder(NLO),andpile-up.Tominimizetheseeffects,weemployadata-drivenmethod,whichhasbeenusedinseveralCMSanalyses(e.g.,[ 40 ]).Inparticular,thismethodwillallowustobetterevaluateandcompensateforthefollowingsourcesofsystematicerror: 1. CalibrationofthehadronicrecoiloftheWZsystem.PoorlymodeledhadronicrecoilhasanimplicationintheEmissTacceptanceaswellastheMTdistribution. 2. Calibrationoftheleptonmomentum/energyscale.Thisdirectlyinuencestheleptonacceptanceaswellasotherobservables:EmissT,MT,M``. 101

PAGE 102

A B C D E FFigure8-21. TLisparameterizedinpT,ofandthenumberofvertices.andeechannelsarecombinedtoincreasestatisticsasbothhaveverysimilarTLasseenfromFigure 8-19 .A)correspondstojj<1andverticesbetween1and9.B)correspondstojj>1andverticesbetween1and9.C)correspondstojj<1andverticesbetween9and17.C)correspondstojj>1andverticesbetween9and17.E)correspondstojj<1andverticesbetween18andabove.F)correspondstojj>1andverticesbetween18andabove.Theerrorbarsincludeonlystatisticaluncertainty. 102

PAGE 103

A B CFigure8-22. Closuretestfornon-prompttauleptoninthettMCsamplefordifferentobservables.Eventsareselectedwithsame-signlightdileptons(passingallleptonselectionsincludingisolation)+taulepton.A)showstheclosuretestasafunctionofEmissT.B)showstheclosuretestasafunctionofM``.C)showstheclosuretestasafunctionofvertices. 3. Underestimationoftheeventyieldnormalization.WehaveincompleteknowledgeoftheWZcross-sectionbeyondNLO.Thisisoneofthemainsourcesofthelargesystematicuncertaintythatistypicallyincurredfromthistypeofbackground.AcomprehensivedescriptionoftheentiremethodcanbefoundinReference[ 41 ].Hereabriefoverviewofthemethodrelevantforthisanalysisispresented. 8.6.4.1TheRecoilMethodThedistributionsofEmissTandMTintheWZsimulationcanbeinuencedifthedetectoreffectsorthemultipleinteractionsaremis-modeledinthesimulation,orif 103

PAGE 104

A B CFigure8-23. Closuretestfornon-prompttauleptonintheZ+jetsMCsamplefordifferentobservables.Eventsareselectedwithsame-signlightdileptons(passingallleptonselectionsincludingisolation)+taulepton.Inaddition,EmissTisrequiredtobeabove30GeVinordertobeorthogonaltothecontrolsample.A)showstheclosuretestasafunctionofEmissT.B)showstheclosuretestasafunctionofM``.C)showstheclosuretestasafunctionofvertices. 104

PAGE 105

thedetectorisnotperfectlycalibrated.Tocorrectfortheseeffects,wewilllookatthedifferencesofZ+jetsprocessbetweenthedataandthesimulation.ThisprocessisparticularlyusefulbecausethetransversemomentumoftheZbosoniscorrelatedwiththehadronicactivityduetotherecoileffect.Sincethehalf-lifeoftheZbosonisverysmall(about310)]TJ /F9 7.97 Tf 6.59 0 Td[(25second),itsdirectionshouldbedeterminedbythedecayproduct.WeonlyconsiderthoseeventswheretheZbosondecaysintoadileptonpair.Inthiscase,itwillbeverystraightforwardtomeasuretheZbosondirectionusingthedileptonsystem.TheZrecoilmethodtakesintoaccounttheinformationaboutthecalorimeterresponse,resolution,andtheunderlyingeventsfromtheZeventstorecalculatetheEmissTspectrumfortheWZMCprocess.WecancalculateEmissTintheWZprocessfromthetransverseenergyofthethreeleptonsandthehadronicrecoiltotheWandZbosonseventbases.Theenergyoftheleptonsisusuallywellmeasured,andanymis-measurementintheleptonenergyscalecanbecorrectedusingtheZmassshape.Contrary,natureoftherecoiltothegaugebosonsisusuallylessunderstoodandthereforeitiscriticaltolookatthedataaswell.ItisworthnotingthatatranslationfromtheZ+jetprocesstoWZisvalid,consideringthatthedominants-channelintheWZproductionhasaverysimilarmechanismtotheZproduction.ThedifferencesbetweenbothsampleswillbeinvestigatedinMCandtheywillbeusedasanadditionalsystematicuncertaintyonthepredictedEmissT.Thetransverserecoil(~uT)forWZeventscanbeformulatedas: ~uT=)]TJ /F7 11.955 Tf 18.25 4.51 Td[(~EmissT)]TJ /F7 11.955 Tf 12.53 3.16 Td[(~P1T)]TJ /F7 11.955 Tf 12.53 3.16 Td[(~P2T)]TJ /F7 11.955 Tf 12.53 3.16 Td[(~P3T,(8)where~EmissTdenotesthemissingtransverseenergyvectorand~PiTthetransversemomentaofleptonith.Similarly,wecandene~uTforZ+jetseventsas ~uT=)]TJ /F7 11.955 Tf 18.24 4.5 Td[(~EmissT)]TJ /F7 11.955 Tf 12.54 3.15 Td[(~P1T)]TJ /F7 11.955 Tf 12.53 3.15 Td[(~P2T,(8) 105

PAGE 106

where~P1Tand~P2Tarethetransversemomentumoftherstandsecondleptons,respectively.Therecoilvectorcannowbesplitintotwocomponents:parallel(u1)andperpendicular(u2)tothebosondirection.FortheWZsample,thedirectionoftheWZsystemistakenasthebosondirection.Theu1componentisthecalorimeterresponsebalancingthebosonandismainlysensitivetocalorimeterresponseandresolution,whiletheu2componentispredominantlydeterminedbytheunderlyingeventsandthemultipleinteractions.Wecalculatetherecoilinbothdataandsimulationasafunctionofthebosontransversemomentum,andthedifferencesarethentransferredtotherecoilinthesimulatedWZsample.ThesampleissplitinbinsofZbosonpTandineachbin,wecalibratetheresponseandresolutionoftherecoilcomponentsbyperformingthedouble-Gaussianttobothcomponents.Thedouble-GaussiantwasperformedbecauseasingleGaussiantshowedapoorperformance,ascanbeseeninFigure 8-24 .Thebosontransversemomentumbinningischoseninsuchawaythateverybincontainsasufcientamountofeventstoallowforapreciset. A BFigure8-24. Gaussianttotherecoilu1componentfor3GeV
PAGE 107

8.6.4.2BinnedGaussianMethodAfterthecalibrationoftherecoil,theresponseandresolutionofthetwocomponents,u1andu2,havetobedetermined.u1isafunctionofthebosonpTbecausethehadronicactivityisproducedintheoppositedirectionoftheZboson,therefore,theslopeoftheu1responsewillyieldthecalorimeterresponse.Ontheotherhand,theu2componentisexpectedtobeindependentofthevectorbosonpTandcenteredaroundzero,sinceitismainlydeterminedbytheunderlyingeventsandmultipleinteractions(pile-up),whichfeaturenogenuinemissingtransverseenergyonaverage.However,softradiationscanslightlyinuencetheu2component.ThedescriptionoftheresponseandresolutionforthedifferentrecoilcomponentsisextractedindataandMCinthefollowingway: 1. Therecoilcomponentsu1andu2arebinnedasafunctionofpToftheZboson;and 2. Ineverybinadouble-Gaussiandistributionisttedtothedistributionoftherecoilcomponents.ThemeanvalueofthettedGaussiandistributiongivestherecoilresponse,whilethewidthofthedistributionrepresentstherecoilresolution.Thetwowidthsofthedouble-Gaussianarealsocalculatedseparately.Figure 8-25 showsthedifferencesinhadronicrecoilscaleandresolutionbetweendataandsimulation.Thescaleandtheresolutionarewell-modeledinthesimulationafterpile-upreweighting.ThesesmalldifferenceswillbeappliedascorrectionfactorstotherecoilintheWZsimulation.OnemustkeepinmindthatthereisasmalldifferencebetweentheZ+jetsandWZprocesses,whichisalsoseeninFigure 8-26 .Thisisduetoacombinationofeffects:themassspectrumisdifferentbetweentheZandtheWZsamples,andtherearealsot-andu-channelsintheWZproductionwhichmayhavedifferenthadronicrecoilnature.Thus,thesedifferencesbetweenthetwoprocessesarepropagatedasasystematicuncertainty. 107

PAGE 108

A B C DFigure8-25. ResponseandresolutionoftherecoilcomponentsindataandZ+jetMCafterpile-upreweighting.A)showsu1componentresponse.B)showsu1componentresolution.C)showsu2componentresponse.D)showsu2componentresolution. 8.6.4.3ApplyingtheRecoilCalibrationTheratiosofdatatosimulationfortheresponseandresolutionarepresentedasafunctionofbosonpT,whicharethenappliedtosimulatedWZeventsinthefollowingway: 1. ForeacheventintheWZMC,lookuptheu1andu2responseandresolutionfromthecorrectedcurves; 2. RandomlysampleGaussianPDFsdenedwiththesevaluestodeterminenewrecoilcomponentsfortheevent.Inotherwords,wedetermineui=Gaus(i(pWZT),i(pWZT);and 108

PAGE 109

A B C DFigure8-26. ResponseandresolutionoftherecoilcomponentsinZ+jetsandWZMCafterpile-upre-weighting.A)showsu1componentresponse.B)showsu1componentresolution.C)showsu2componentresponse.D)showsu2componentresolution.Thesmalldifferencesbetweenbothsamplesareusedasasystematicuncertainty. 3. Combinethenewuicomponentstoformacorrectedrecoilvector~u.Addtheleptonvector(~`)backintodeterminethecorrectedEmissTdistribution. 8.6.4.4LeptonEnergyScaleandResolutionAftercalibratingthehadronicrecoilinthesimulation,wecannowevaluatetheleptonenergyscaleandresolutionfromdata.ThiswillallowustocorrecttheleptonmomentaandgetamorepreciseestimateoftheEmissT,MTandM``intheWZevents.LeptonsaredirectlyusedtocalculateMTandM``;thus,possibledifferencesinleptonenergyscaleandresolutionbetweendataandMCwillleadtomis-modelingofthesevariables. 109

PAGE 110

Inthecaseofthemuons,thenominalMCmomentumscalesufcientlydescribedthedata.Forelectrons,theenergyscaleandresolutionaredeterminedfromtheZeventsindata.AstheelectronenergymeasurementsdependstronglyontheECalmeasurements,wecalibratethemsimultaneouslyforthedifferentleptonpseudo-rapiditybins.Weconsiderdielectronsystem,whereelectronspasstheanalysisfullelectronselection,andcalculatetheZmassspectrumwhichwillallowustoderiveascalefactorandresolution.Weperformthisprocedureforsixdifferentbins,whichgivestwenty-onepossiblecombinationsforthedielectronsystem:6combinationsrelatetodielectronpairbeinginthesameregionand15combinationsrelatetothedielectronpairbeingindifferentregions.IneverycombinationtheZmassspectrumisttedbytheGaussiandistribution.Inthist,theMCmasstemplateisscaledwiththeenergyscalefactorsindata:massscale=masssim e1e2.Wheree1ande2aretheenergyscalefactorsoftheelectronsfromthetwobins,andmassscaleandmasssimaretherescaledandtheMCmassvariables,respectively.TheextraGaussiansmearingusedfortheMCmasstemplateisthequadraticsumoftheextraresolutionsforeachofthetwoelectrons:mass=e1e2.Wethencarryoutasimultaneoustoverallthecategoriestoextractthescalefactorsandresolutions.TheimprovementoftheZmassdistributionafterperformingthistcanbeseeninFigure 8-27 .Theresultingelectronenergyscalefactorsandresolutionsindifferent-binsarelistedinTable 8-14 .AstheCMStrackersystemhasnon-uniformmaterialbudgetdistribution[ 13 ],theelectronscalefactorsare-dependent,asseenfromTable 8-14 .Theelectronenergyscalecanbeupto2%differentbetweendataandMC.Forelectrons,theresolutionindataisconsiderablyworsethaninMC(upto2GeV).ThisisduetotheECalcrystaltransparencydegradationwiththeincreasingluminosity[ 40 ].Formuons,thedifferencesaremuchsmaller(maximum0.5GeV).TheuncertaintiesontheelectronenergyscalesandresolutionswillbeusedtogetherwiththeuncertaintiesrelatedtotherecoilmethodtocalculatethesystematicuncertaintiesontheWZ 110

PAGE 111

A BFigure8-27. InvariantmassdistributionforZcandidatesindataandMC.A)showsthedistributionwithoutanycorrections.B)showsthedistributionafterapplyingtheenergyscalecorrectionstodataandthenecessaryextrasmearingtotheMC. distributions.Thestatisticaluncertaintyofthetwillbecombinedwithasystematicuncertainty.Forthemuonmomentumscale,theCMSrecommendationsarefollowed:aat0.2%systematicuncertaintyonthemomentumscaleandaat0.6%uncertaintyontheresolutionareused. Table8-14. Electronenergyscaleandresolutioncorrections. -Region Scale(Data!MC) Resolution[GeV](MC!Data) [0,0.4] 1.00400.0001 0.3900.012 [0.4,0.8] 1.00430.0001 0.4090.013 [0.8,1.2] 1.00170.0001 0.6930.011 [1.2,1.4442] 1.00250.0001 1.2640.012 [1.566,2.0] 1.00270.0001 1.1080.013 [2.0,2.4] 1.00900.0001 1.1300.014 FortheEmissT,thecorrectionsfromtheTable 8-14 areappliedtotheleptonswhenconstructingthetemplate.TheenergyscalefactorsareappliedtotheelectronsinZMCandadditionalsmearingisappliedtotheZMCbeforettingfortherecoil 111

PAGE 112

components,asdescribedpreviously.Forthemuonsnocorrectionsareapplied,onlytheCMS-recommendeduncertaintiesarepropagatedthrough.TheleptontransversemomentaintheWZarealsocorrectedandsmearedtofurtherachievebetterMTandM``distributions. 8.6.4.5CorrectionstotheNormalizationThecorrectionstothehadronicrecoilandleptonenergyscale/resolutioncancausesimulatedeventstomigratealongthesearchregionswhicharedenedusingMT,M``,andEmissTvariables(Section 8.5.3 ).However,wedonotexpectthatthemigrationisuniformacrossthesearchregions.Forexample,theoff-Z,low-MTregionsaremainlysensitivetotheuncertaintiesontheleptonscaleandresolution,whiletheon-Z,high-MTregionsaremostlyaffectedbytheuncertaintiesfromtherecoilmethod.Theoff-Z,high-MTregionissensitivetobotheffects.TheothersourceofwhytheWZyieldscanbeoffarisesifthecross-sectionorintegratedluminosityareunderestimatedoroverestimated.Inthiscase,wecouldquasi-uniformlybiasourWZpredictionsinallsearchregions.Tominimizethiseffect,weimplementthefollowingprocedure.WeaddtheWZ-enrichedcontrolregion(denedas75
PAGE 113

binswillbepropagatedthrough.Thedouble-Gaussiantuncertaintiesareusedtoassignanuncertaintyontherecoil.ThetuncertaintiesindataandtheMCsamplescanthenbeusedtoassignanuncertaintytotherecoilpredictions.Thisisdonebyrecalculatingtherecoildistributionandtheotherquantities,assumingthattheresolutionsandmeansareuctuatedupordownby1standarddeviation.Anotherexperimentaluncertaintyisthe(predominantlytop)backgroundtotheZsampleindata.TherelativeamountofbackgroundincreaseswithlargerZtransversemomentum.Toremovethebackground,weusetheMCrecoilshapesofthebackgroundsampleswithoatingnormalizationinthet.ThesmalldifferencesbetweentherecoildistributionsintheWZMCandtheZMCalsohavetobetakenintoaccount,toallowforourlimitedquantitativeunderstandingofthem.InthiscasethemeanandresolutionofthehadronicrecoilcomponentsaretakenfromtheZdatasampleandusedtorecalculatethemissingtransverseenergy.TheuncertaintiesontheleptonenergyscaleandresolutionwillalsobeusedtoassignsystematicuncertaintiesontheWZdistributions.Asanuncertainty,thestatisticaluncertaintyofthetisusedandhalfofthecorrectionfactorisaddedinquadraturetoaccountforsystematiceffectslikethelimitedbinning.ThenaluncertaintyontheWZyieldsiscalculatedbyvaryingtheenergyscaleandresolutionseparatelydownandupby1sigmaandthendetermininghowtheextractedyieldsareinuencedinthecontrolregion. 8.6.5DeterminationofBackgroundduetoPhotonConversionAnothertypeofbackgroundtotheexclusivethreeleptonsearchescanarisefromthephotonconversion,althoughthisbackgroundisverysmall.Therearetwodistincttypesofphotonconversion:externalandinternal.Inthecaseonexternalphotonconversion,arealphotoniscreatedandwithinthedetectormaterialitcreatesanoppositesignandsameavordileptonpair.Anon-shellphotonalmostexclusivelydecaysintoeepairasthedecayintodimuonpairissuppressedbyafactorofme/m 113

PAGE 114

=10)]TJ /F9 7.97 Tf 6.59 0 Td[(4,whereasinternalconversioncanproducealmostasmanydimuonpairsasdielectronbecausephotonisvirtual.Internalconversionsmaynotbeproperlydescribedinthesimulationbecauseoflow-energythresholdsforemittedleptonsinthegenerator.So,itiscriticaltoemployadata-drivenmethodtoestimatethebackgroundduetothephotonconversion.ExternalconversionisdominatedbytheZevents,inwhichoneofthetwoleptonsradiatesphotonwhichthendecaysintodielectronpairasshowninFigure 8-28 .Tocontributetotheexclusivethreeleptonsignature,thephotondecayshouldbeasymmetric,yieldingonesoftlepton,whichescapesthedetection,andanotherharderleptonwhichcarriesthemostenergyofthemotherphoton.Inaddition,aninvariantmassofthesethreereconstructedobjectsnaturallyfallsintotheZmasswindowof9115GeV.Itispossibletoselectafairamountofeventsduetoexternalphotonconversionbyrequiringeventswith``ewhere`iseithereor.Eventsselectedwith``(where`iseithereor)aresensitivetotheinternalconversion.Figure 8-29 showsdataandMCcomparisonforbothconversions.ItmustbenotedthattheZMCsample,whichwasusedinFigure 8-29 ,wasproducedprivatelyandthesecomparisonsserveonlyasavisualdemonstration.Thedata-drivenmethodtopredictbackgroundduetophotonconversionreliesontheassumptionthattherateofarealphotonisproportionaltotherateofavirtualphoton.Thisassumptionisadvocatedbythefactthatthevirtualphotonmassspectrumisstronglypeakedinthelowmassregion,thereforethekinematicdistributionsaresimilarbetweenthetwophotons[ 42 ].Theconversionfactorisdenedastheratiooftheprobabilityforaphotontoproduceavalidleptoncandidateviaasymmetricconversiontotheprobabilityforarealphotontopassalloftheselectioncriteria[ 43 ].ThisratioismeasuredinthecontrolsampleindatawhichisselectedwitheventsatlowEmissT.Wecountthenumberof``eventsinthedenominatorwitharequirementthattheinvariantmassofthreeobjectsisconsistentwithaZmasswindowof9115GeV;whereasthenumberof```eventsdenotethenumerator.Again,theinvariantmassof 114

PAGE 115

threeobjectsshouldbeconsistentwithZmasswindowof9115GeV.Themeasuredconversionfactorsformuonandelectronare0.5%0.1%(stat)0.1%(sys)and3.1%0.3%(stat)1.5%(sys),respectively. Figure8-28. AFeynmandiagramforthephotonconversion.Thedifferentlengthsoftheleptonsfromphotonconversionindicatetheasymmetricdecay. A BFigure8-29. MCanddatacomparisonforinternalandexternalconversions.A)showstheinvariantmassdistributionofthreereconstructedleptons.Thesechannelsaresensitivetointernalaswellasexternalconversions.B)showstheinvariantmassdistributionforthoseeventswhichpredominantlyarisefrominternalconversion. 8.6.6BackgroundduetoRareSMprocessesWhilethedominantbackgroundstothissearchareestimatedbywell-designeddata-drivenmethods,afewrare,butirreducible,SMprocessesareexpectedto 115

PAGE 116

contributeasmallnumberofeventstooursearchregionsaswell.TheseprocessesincludeZZ,WWW,WWZ,WZZ,ZZZ,WW,ttW,ttZ,andtt.Atthetimeofwriting,noneoftheserareprocesseshavebeenestablishedormeasuredattheLHC.WethereforehavenorecoursebuttorelyonsimulationandtheNLOcross-sectionsthatwehaveatourdisposal.Fortunately,theeventswhichareexpectedtoenteroursignalregionsfromtheseprocesseswillcontainthreepromptleptonsandgenuinemissingenergy.Thus,thesimulationcanprovideuswithanestimationoftheexperimentalacceptancewithsufcientaccuracy.Table 8-15 showstheexpectedyieldsforthedifferenttopologiescomingfromrareSMprocessesforeachchannel,normalizedtotheluminosityof19.5fb)]TJ /F9 7.97 Tf 6.59 0 Td[(1.Thestatisticaluncertaintiesareoforder1%;however,weapplyasystematicuncertaintyof50%tocoverourincompleteknowledgeofthecross-sectionsbeyondLO.Thebackgroundfromtheseprocessesisexpectedtobeextremelysmall,andalmostnegligible.Thus,thenalresultwillbecompletelyinsensitivetothesystematicuncertaintyassumedhere. Table8-15. ExpectedcontributionstothedifferentrareSMprocessesinthebaselineselection.Yieldsarenormalizedtotheluminosityof19.5fb)]TJ /F9 7.97 Tf 6.58 0 Td[(1. Sample3lOSSFon-Z3lOSSFoff-Z3lNoOSSFSSTau ZZ18.563.220.280.83ttW(W)0.401.290.630.25ttZ3.160.730.180.24tt0.351.100.180.33WWW0.451.510.610.35WWZ3.240.290.130.15WW0.160.600.310.08WZZ0.980.040.010.02 ZZZ0.060.0010.0010.002 Total27.378.792.322.25 116

PAGE 117

CHAPTER9FINALRESULTSOncebackgroundestimationmethodologiesaredened,wearenowlefttotesthowthesemethodsmodelthebackgroundprocessesinthebaselineselection.Thebaselineselectionisdominatedbythebackgroundprocesses,sowearenotvulnerabletoanypossiblesignalcontamination.Figures 9-1 and 9-2 showthecomparisonbetweenobservedandpredictedyieldsinthebaselineselection.Itmustbenotedthatforthenalpredictions,allchannelsarecombined.Weseenosignicantdiscrepanciesinany3-leptonselection,whichisgood!Figures 9-3 9-4 and 9-5 showpredictedandobservedyieldsinallsearchregions.AnexcesswhichispresentedfortheOSSFselectionusingon-Zregion(75
PAGE 118

A B CFigure9-2. TransverseMassdistributionsforOSSFselectioninthebaselineselection.A)showstheoff-Z(lowmass)region.B)showstheon-Zregion.C)showstheoff-Z(highmass)region. 118

PAGE 119

CMSPreliminaryp s=8TeV,Lint=19.5fb)]TJ /F9 7.97 Tf 6.59 0 Td[(1Figure9-3. PredictedandobservedyieldsfortheOSSFselectionasafunctionofEmissT. 119

PAGE 120

CMSPreliminaryp s=8TeV,Lint=19.5fb)]TJ /F9 7.97 Tf 6.59 0 Td[(1Figure9-4. PredictedandobservedyieldsfortheNoOSSFselectionasafunctionofEmissT. 120

PAGE 121

CMSPreliminaryp s=8TeV,Lint=19.5fb)]TJ /F9 7.97 Tf 6.59 0 Td[(1Figure9-5. PredictedandobservedyieldsfortheSSTauselectionasafunctionofEmissT. 121

PAGE 122

Table9-1. ThesummaryoftheobservedyieldsandpredictedbackgroundsfortheOSSFselection.Thetotalerrorincludesbothstatisticalandsystematicaluncertainties. EmissT(GeV) WZNon-PromptRareSMZTotalbkgObserved MT>160GeV,M``<75GeV 50-100 1.80.423.110.880.390.370.146.21.212 100-150 1.30.272.20.730.80.460.020.014.30.913 150-200 0.310.130.860.340.220.110.040.011.40.382 200-inf 0.360.140.360.160.150.07000.870.230 MT>160GeV,75GeV160GeV,M``>105GeV 50-100 0.930.230.430.21.020.810.30.122.70.871 100-150 0.450.170.40.191.330.950.190.072.40.983 150-200 0.180.070.20.120.230.130.020.010.630.190 200-inf 0.110.050.060.050.180.10.030.010.380.120 120GeV
PAGE 123

EmissT(GeV) WZNon-PromptRareSMZTotalbkgObserved 50-100 1.40.350.520.220.540.250.250.12.70.494 100-150 0.130.070.740.30.120.070.020.0110.322 150-200 0.050.040.20.110.110.07000.360.140 200-inf 0.020.03000.180.16000.20.160 MT<120GeV,M``<75GeV 50-100 54.99.259.117.99.33.419.17.5142.421.7138 100-150 6.91.313.141.410.590.050.0221.54.316 150-200 1.70.4420.680.330.150.020.0140.835 200-inf 10.650.560.240.210.10.110.041.90.72 MT<120GeV,75GeV105GeV 50-100 23.84.415.84.94.51.720.7946.16.849 100-150 4.50.884.71.51.220.520.080.0310.51.810 150-200 1.20.280.970.370.380.170.070.032.60.494 200-inf 0.80.260.260.130.210.10.010.011.30.34 123

PAGE 124

Table9-2. ThesummaryoftheobservedyieldsandpredictedbackgroundsfortheNoOSSFselection.Thetotalerrorincludesbothstatisticalandsystematicaluncertainties. EmissT(GeV) WZNon-PromptRareSMZTotalbkgObserved MT>160GeV,M``<100GeV 50-100 0.370.12.60.880.520.250.160.063.70.922 100-150 0.230.061.20.450.560.320.150.062.10.553 150-200 0.060.050.540.220.20.110.030.010.830.250 200-inf 0.090.080.170.090.130.070.0100.40.141 MT>160GeV,M``>100GeV 50-100 0.050.030.10.090.230.130.030.010.410.160 100-150 0.050.030.240.140.110.070.060.020.460.160 150-200 0.010.010.060.070.010.010.0100.090.070 200-inf 0.040.02000.050.040.020.010.110.040 120GeV100GeV 50-100 0.110.040.030.030.050.040.060.020.250.071 100-150 0.010.030.120.090.030.020.020.010.180.10 150-200 000.030.030.010.01000.040.030 200-inf 00000.010.01000.010.010 MT<120GeV,M``<100GeV 50-100 3.40.3726.182.281.11.40.5533.28.129 100-150 0.580.116.520.580.310.10.047.82.15 150-200 0.110.041.10.40.170.10.020.011.40.421 200-inf 0.070.050.20.10.220.190.0100.50.220 MT<120GeV,M``>100GeV 50-100 0.090.031.10.40.290.160.210.081.70.441 100-150 0.010.010.180.090.110.070.040.010.340.120 150-200 000.030.030.050.04000.080.050 200-inf 000.030.0300000.030.030 124

PAGE 125

Table9-3. ThesummaryoftheobservedyieldsandpredictedbackgroundsfortheSSTauselection.Thetotalerrorincludesbothstatisticalandsystematicaluncertainties. EmissT(GeV) WZNon-PromptRareSMZTotalbkgObserved MT>160GeV,M``<100GeV 50-100 1.30.671.040.490.630.250030.862 100-150 0.640.331.140.410.560.28002.30.61 150-200 0.130.070.20.150.170.09000.50.190 200-inf 0.060.040.320.180.080.05000.460.192 MT>160GeV,M``>100GeV 50-100 0.070.050.380.20.020.02000.470.21 100-150 0.050.040.350.170.030.01000.430.181 150-200 000.190.110.020.03000.210.110 200-inf 000.030.040.020.02000.050.050 120GeV100GeV 50-100 0.070.050.170.150.050.04000.290.161 100-150 0.030.020.040.0500000.070.060 150-200 000.010.0100000.010.010 200-inf 0.010.010000000.010.010 MT<120GeV,M``<100GeV 50-100 157.533.110.23.171.20051.312.746 100-150 1.80.933.131.21.090.830061.71 150-200 0.840.430.980.450.150.090020.630 200-inf 0.450.230.330.250.060.03000.840.340 MT<120GeV,M``>100GeV 50-100 0.390.211.850.910.110.06002.40.943 100-150 0.050.040.390.20.050.04000.490.210 150-200 0.020.020.080.080.020.01000.120.080 200-inf 0.010.010.090.060.010.02000.110.070 125

PAGE 126

CHAPTER10SIGNALACCEPTANCESignalacceptanceefciencyisobtainedfromthesimulation.AsetofsimpliedmodelsaregeneratedwiththePYTHIAgeneratorandthenpropagatedthroughtheGEANT4detectorsimulationusingthefastsimulationapproach.Wealsoapplytriggerscalefactorswhichwereevaluatedinsecton 8.4 .SignalacceptanceefcienciesfordifferentscenarioscanbefoundbeinFigures 10-1 10-2 10-3 10-4 10-5 10-6 10-7 and 10-8 .Itmustbementionedthatonlythosesearchregionswhichareusedintheupperlimitcalculationonthesignalproductioncross-sectionareshown. Figure10-1. Flavor-democraticchargino-neutralinoproductionwithx~`=0.95. 126

PAGE 127

Figure10-2. Flavor-democraticchargino-neutralinoproductionwithx~`=0.5. 127

PAGE 128

Figure10-3. Flavor-democraticchargino-neutralinoproductionwithx~`=0.05. 128

PAGE 129

Figure10-4. Tau-enrichedchargino-neutralinoproductionwithx~`=0.95. 129

PAGE 130

Figure10-5. Tau-enrichedchargino-neutralinoproductionwithx~`=0.5. 130

PAGE 131

Figure10-6. Tau-enrichedchargino-neutralinoproductionwithx~`=0.05. 131

PAGE 132

Figure10-7. Tau-dominatedchargino-neutralinoproductionwithx~`=0.5. 132

PAGE 133

Figure10-8. Flavor-democraticchargino-neutralinoproductionwithWZ+LSPinthenalstate. 133

PAGE 134

CHAPTER11SIGNALACCEPTANCEUNCERTAINTIES 11.1TheoreticalUncertaintyThemaincontributionstotheoreticaluncertaintiesareassociatedwiththescaleoftheQCDcouplingandtheproton'spartondistributionfunction.Theseuncertaintiescanaffectthecross-sectioncalculationandtheexpectedsignalacceptance.BecausewewillconstrainmodelsofnewphysicsintermsofBR,wearenotconcernedwiththecross-sectionuncertainties.But,theuncertaintiesleadingtochangesinthesignalacceptancehavetobeevaluated.TheproceduretocarryoutthisstudyistouseanensembleofdifferenttheoryPDFstogeneratethesignalmodels.Thishasbeendoneonprevioussupersymmetryanalyses,whichwerebasedonstrongproduction[ 35 ],andfoundtorangefrom0-2%.Wethereforetake2%asaconservativeupperbound. 11.2LuminosityUncertaintyTheofcialluminositymeasurementonCMSquotesanuncertaintyof2.5%[ 44 ].Weconformtothisstandard. 11.3TriggerUncertaintyAnassortmentofdileptontriggers,thesupposedHiggstriggers,areusedinthisanalysis.Thetriggerefciencyanditsassociateduncertaintymeasurementsweredescribedinsection 8.4 .Weassign5%uncertaintytothetriggerperformance. 11.4LeptonReconstructionEfcienciesandAssociatedUncertaintiesTheidenticationandisolationefciencyiscalculatedindatausingZ!`+`)]TJ /F1 11.955 Tf -429.9 -28.25 Td[(eventsasafunctionofpTandusingthetagandprobemethod.ThismethodiswidelyusedinmanyanalysesattheCMStoevaluatepromptleptonreconstructionefciency.Onceitismeasuredindata,thenextstepistodirectlycompareittotheefcienciesobtainedintheZ+jetsMCsampleandthencalculatethescale-factors.Thisindicatesthelevelofprecisionofthesimulation.ThereconstructionefciencyisusuallymeasuredbytheCMSEGammaandMuonWorkingGroupsfortheirrecommendedlepton 134

PAGE 135

selections,andarepubliclyavailable.Here,weperformanindependentcross-check.Thescalefactorisshowntobeconsistentwith1.Dedicatedtriggersareusedforbothelectrons1andmuons2toselectunbiasedonlineevents.Toaccountforthedifferentpre-scales3,thedatawasrescaledusingthemeasuredZyieldsfromanunprescaledtrigger.Thetaggedleptonisawell-identiedandisolatedlepton(passingthefullleptonselections)anditismatchedtothetriggerHLTleptoncandidate.Theprobeisanoppositesignandsameavorlepton,whichmakesadileptoninvariantmassbetween60and120GeVwiththetaglepton.Theprobepassestheisolationcuts,whilemeasuringtheIDefciencyandviceversa.ThecorrelationbetweenisolationandIDistakenfromMCbycalculatingscalefactorsforeverycutandapplyingthattotheMC.Theassumptionbehindthisisthatthecorrelationsarequitewell-modeledinthesimulation.ForthelowpTbinsattothemassspectrumwasstillneeded,asonecanseeinFigure 11-1 .ThettingisdoneusingaGaussian-smearedsimulationshapeforthesignalandanexponentialfallingcurveforthebackground.TheGaussian-smearedMCwaschoseninsteadofananalyticalfunction(likeCrystal-Ball,convolutedBreit-Wigner)becausesomeoftheeffectsindifferentpTbinscannotbewell-modeledbythisanalyticalshape. 1HLT Ele17 CaloIdVT CaloIsoVT TrkIdT TrkIsoVT Ele8 Mass50-theIDandisolationrequirementisappliedonlytooneelectronleg,whilerelaxedfortheotherleg.2HLT Mu17 TkMu83AsdiscussedinChapter5,theCMSHighLevelTrigger(HLT)systemreceiveseventsfromtheL1systematarateof100KHz;itthenreduceseventsfurtherdownto300Hzbyapplyingmorestringentselections.SomesetsofHLTtriggershavequitelooseselectionsdesignedfordifferentpurposes,resultinginaveryhighrate.Therefore,tokeeptherateattheacceptablelevel,theCMSHLTsystemrandomlythrowsawaysomeevents.Thiseffectisreferredtoaspre-scale. 135

PAGE 136

A BFigure11-1. Fittothedielectroninvariantmassdistribution.ElectronpTrangeliesbetween10GeVto15GeV.A)showsthedistributionforelectronspassingtheID.B)showsthedistributionforelectronsfailingtheID. TheidenticationandisolationefcienciesfordataandMCareshowninFigure 11-2 andFigure 11-3 forelectronsandmuons,respectively.WeobservegoodagreementefcienciesbetweendataandMCwithin2%.Duetopossiblebackgroundcontamination,imperfectchoiceofsignalandbackgroundshapes,andpossiblemisunderstandingofthecorrelationbetweentheidenticationandisolationvariables,a3%systematicuncertaintyperleptonisassignedtotheefciencymeasurement. 11.5MissingTransverseEnergyandTransverseMassAcceptancesandAssociatedUncertaintiesAsitwasmentionedinSection 8.1 ,thesignalproductiondoesnotinvolvehadronicactivitiesexceptthosewhichcomefromtheinitialstateradiation(ISR).Thesignalisexpectedtofeaturethreeinvisibleparticles,whichwillyieldasignicantamountofgenuinemissingenergy,andthreeleptonsinthenalstate.AleptonmomentumiswellmeasuredatCMSanditsuncertaintyisusuallysmall,leadingtoasmalluncertaintyontheEmissTandMTobservables.However,thelargestsourcesofuncertaintyaffectingtheEmissTandMTobservablestypicallycomefromincompleteknowledgeofthejetenergyscaleandresolution.But,duetotheEWKproduction,thetrueEmissTacceptanceisnot 136

PAGE 137

A B C DFigure11-2. Reconstructionefciencyforelectrons.A)showsidenticationefciencyasafunctionofelectronpseudo-rapidity.B)showsidenticationefciencyasafunctionofelectronpT.C)showsisolationefciencyasafunctionofelectronpseudo-rapidity.D)showsisolationefciencyasafunctionofelectronpT. expectedtobeinuencedbyinstrumentaleffectstotherstorder,butratherbythemassscalesofthesignalmodel. 11.6b-TaggedJetUncertaintyToreducethebackgroundassociatedwiththetopquark,werejecteventswheretheb-taggedjetmultiplicityisnon-zero.Thealgorithmusedtoidentifyb-taggedjetswaspresentedinChapter 6 .Oncetheb-taggingalgorithmispositiveforagivenjet,itispracticallyimpossibletotellwhethertheb-taggedjetisgenuineornot.Therefore, 137

PAGE 138

A B C DFigure11-3. Reconstructionefciencyformuons.A)showsidenticationefciencyasafunctionofmuonpseudo-rapidity.B)showsidenticationefciencyasafunctionofmuonpT.C)showsisolationefciencyasafunctionofmuonpseudo-rapidity.D)showsisolationefciencyasafunctionofmuonpT. therearetwoimportantuncertaintieswhenb-taggingisusedinananalysis.Onerelatestotheefciencyoftagjetsarisingfromb-quarkproductionandtheotherrelatestotheprobabilityoffalselytaggingalight-avorjetasab-jet.Becausethesignalproductiondoesnotinvolveanygenuineb-taggedjet,theformerpracticallyhasnoimplicationinthesignalacceptance.But,thelattercanaffectthesignalacceptance.ThesignalmodelsusedforouranalysisaregeneratedwithPYTHIAandthenpropagatedthroughtheCMSFastSimulation,whichwasdescribedinChapter 7 138

PAGE 139

Becauseinoursignalsamples,weassumethatthestronglyinteractingsupersymmetricparticlesareheavy,thes-channelisthedominantproduction(andt-channelissuppressedduetoheavysquarks).Therefore,weexpectthatthejetactivityinoursignaleventswillcomefromISR.However,PYTHIAgeneratordoesnotpreciselymodelISRasitcanonlygenerate2!2scatteringprocesses.WealsohavealittlecondenceintheISRjetmomentaspectra.Ontheotherhand,MADGRAPH,whichisamatrixelementgenerator,provideabetterdescriptionofISRjetsandcansimulateupto2!4scatteringprocesses[ 33 ].TheWZsimulationusedinthisanalysisisproducedwiththeMADGRAPHgenerator.WecantakeadvantagefromthefactthattheSMWZproductionresemblesoursignalproductionandevaluatethesystematicuncertaintyoftheb-taggedjetvetoacceptanceonourPYTHIA-generatedsignalsamples.Wedevelopedthefollowingapproach: 1. FirstweverifythattheWZsimulationaccuratelymodelstheb-vetoacceptanceindata.Todothis,werelyontheeventsthatarepopulatedinaspecicallydenedcontrolregiondenedas:8160GeV.Inthisregion,weexpectaround526WZeventsfromthesimulation,whereastheotherMonteCarloprocessespredictaround23events.Inthisregionweobserve529events.Wethenrelaxtheb-vetorequirementtoseehowmanyWZeventsarerecovered.Here,weobserve573events.Fromthesimulationsweget537eventsforWZand44eventsfortheotherprocesses.Oppositesign-sameavormassM``distributionsfortheseeventsareshowninFigure 11-4 .TomakesurethatdataisalmostfullypopulatedbyWZevents,wesubtracttt,whichisestimatedfromthesimulation,andrareMonteCarloyieldsfromdata.Thisallowsustocalculateadata-drivenestimateoftheb-vetoacceptanceAWZ,datab-veto,AWZ,datab-veto=506=530=0.960.01stat. (11)WhenweperformthesametestintheWZsimulation,weobtaintheb-vetoacceptanceAWZ,MCb-veto,AWZ,MCb-veto=0.980.001stat. (11)Wetake(AWZ,MCb-veto)]TJ /F4 11.955 Tf 11.96 0 Td[(AWZ,datab-veto)=AWZ,datab-veto=2.1% (11) 139

PAGE 140

asanuncertaintythatcomesfrommodelingtheperformanceoftheb-taggingalgorithm. 2. WenowcomparetheWZb-vetoacceptanceinsimulationtoapointintheSMS-TChiSlepSlepmodelspacewhereM(~1)=100GeVandM(~01)=0GeV.ThisisthemodelpointthatmostcloselyresemblesWZproductionwithrespecttop ^s.Forthissignalpointweobtainab-vetoacceptanceASMSb-veto,ASMSb-veto=0.990.002. (11)Wetake(ASMSb-veto)]TJ /F4 11.955 Tf 11.95 0 Td[(AWZ,MCb-veto)=AWZ,MCb-veto=1% (11)asanadditionaluncertaintythatcomesfromtheinadequatemodelingoftheISRinPYTHIA.Weaddthesetwosourcesofuncertaintylinearlytoobtainatotaluncertaintyontheb-vetoacceptanceintheSMSsignalmodels,andweassigntheconservativeuncertaintyof5%. 140

PAGE 141

A B CFigure11-4. OSSFdileptonM``massdistributionsfordifferentb-taggedjetrequirements.A)showseventswithNb)]TJ /F6 7.97 Tf 6.59 0 Td[(jet0.B)showseventswithNb)]TJ /F6 7.97 Tf 6.59 0 Td[(jet=0.C)showseventswithNb)]TJ /F6 7.97 Tf 6.58 0 Td[(jet1. 141

PAGE 142

CHAPTER12INTERPRETATIONOFRESULTSGoodagreementisobservedinallsearchregionswithintheevaluateduncertainties,asshowninTables 9-1 9-2 and 9-3 .Havingobservednosignicantsignsofanexcessbeyondourbackgroundexpectation,wearenowlefttoevaluatetheconstraintsonnewphysicsimpliedbythissearch.Upperlimitsat95%condencelevel(CL)onthesignalproductioncross-sectionarecalculatedwiththeLandSsoftware[ 45 ]usingthemodiedfrequentistCLsmethod[ 46 ],whichisdenedastheratiooflikelihoodsforthetwohypothesesofinterest(inthiscasefortheexclusion):signal+backgroundandnull(orbackgroundonly)hypotheses.Aswehavedesignedexclusivesearchregions,withdifferentsearchregionshavingvarioussensitivitiestodifferentpartsofthesupersymmetryphasespace,weperformamultibinnedttomaximizethesensitivitytoagivenmodel.Todoso,weconstructthelikelihoodfunction:L(dataj,)=Poisson(datajs()+b()) (12)where,datarepresentstheobservation,isthesignalstrength,whiles()andb()representsignalandbackgroundyields,whicharethefunctionoftheirassociateduncertainties(referredtoasthenuisanceparameter).BothbackgroundaswellassignalyieldsaretreatedaccordingtothePoissonprobabilityineachsearchregion(bin),andL(dataj,)isdenedasY(si+bi)ni ni!e)]TJ /F9 7.97 Tf 6.59 0 Td[((si+bi). (12)Here,ni,siandbirepresentobserved,signalandpredictedStandardModelbackgroundyieldsinbinith.Weonlyemploythosesearchregionstocalculatea95%CLupperlimitwhichhavereasonablesignalacceptance.ThisisprimarilydonetoreducethehugeamountofCPUtimeneededtogenerateupperlimits;thosesearchregionswhichwereusedinthelimitcalculationcanbefoundinFigures 10-1 10-2 10-3 10-4 142

PAGE 143

10-5 10-6 10-7 and 10-8 .Weseparatelyincorporatealltypesofbackground,whicharerelevanttothisanalysis,togetherwiththeirassociateduncertainties;thus,theyareconsidereduncorrelatedwithrespecttoeachother.ItisworthnotingthatthetotaluncertaintiesincludestatisticalandsystematicuncertaintiesasindicatedinTables 9-1 9-2 and 9-3 ,therefore,eachbackgroundhasonlyonenuisanceparameteranditis100%correlatedacrossallsearchregions.Thesameisalsotrueforthesignal.Itisalsoimportanttodeterminethosesearchregionsgivingthebestsensitivityfordifferentmasspoints,whichwillessentiallydrivetheexclusionlimit.Wecarryoutthefollowingprocedure:rst,wegeneratethemodel-independentexpectedupperlimitontheyieldsfornewphysicsforeachexclusivesearchregionusingtheexpectedbackgroundsandtheirassociateduncertainties(statisticalandsystematicuncertaintiescombined).ThelimitisobtainedusingthesameLandSsoftware.Then,weclassifywhichsearchregionsexcludedagivensupersymmetrymasspoint1,andtakeonlythatsearchregionwhichyieldsthesmallestupperlimit.Inthecaseofnoexclusion,weonlytakethatsearchregionwhichresultsinthesmallestvalueofaratiooftheupperlimittothesignalyieldnormalizedtotheluminosityof19.5fb)]TJ /F9 7.97 Tf 6.58 0 Td[(1andthecross-section.Figures 12-1 12-2 and 12-3 showamapofthemostsensitivesearchregionsforall3-leptonselections.Figures 12-4 12-5 and 12-6 showconstraintsonthechargino-neutralinomassesfordifferentsupersymmetryscenariosusingthenext-to-leadingorder(NLO)cross-section[ 47 48 ]. 1Thelogicoftheexclusionisverystraightforward:iftheexpectedsignalyield(foragivenmasspoint)normalizedtotheluminosityof19.5fb)]TJ /F9 7.97 Tf 6.58 0 Td[(1isgreaterthanthegeneratedupperlimit,thenthismasspointisconsideredexcludedat95%CL 143

PAGE 144

A B CFigure12-1. Amapofthemostsensitivesearchregionsfortheavor-democraticscenario.Asexpected,inthebulkSR-44(theSRwithhighEmissT,highMTandhighM``cuts)dominates.Nearthediagonal,wherethesmallmasssplittingbetweenLSPandchargino/neutralinooccurs,searchregionswithmoderateEmissTbecomesensitive.A)showsthesignalsamplewithx~`=0.95.B)showsthesignalsamplewithx~`=0.5.C)showsthesignalsampleswithx~`=0.05. 144

PAGE 145

A B CFigure12-2. Amapofthemostsensitivesearchregionsforthetau-enrichedscenario.Asexpected,inthebulkSR-44(theSRwithhighEmissT,highMTandhighM``cuts)dominates.Nearthediagonal,wherethesmallmasssplittingbetweenLSPandchargino/neutralinooccurs,searchregionswithmoderateEmissTbecomesensitive.A)showsthesignalsamplewithx~`=0.95.B)showsthesignalsamplewithx~`=0.5.C)showsthesignalsampleswithx~`=0.05.. 145

PAGE 146

A BFigure12-3. Amapofthemostsensitivesearchregions.A)correspondstothetau-dominatedscenario.B)correspondstotheavor-democraticscenariodecayingviaon-shellW=Zgaugebosons. 146

PAGE 147

A B CFigure12-4. Limitsonchargino(=neutralino)andLSPmassesfortheavor-democraticscenario.Theexcludedregionsarethosewithinthekinematicboundariesandtotheleftofthecurves.TheeffectsofthetheoreticaluncertaintiesintheNLOproductioncross-sectioncalculationareindicatedbythethinblackcurves.Theexpectedlimitsandtheir1standarddeviationareshownbythethickandthinreddashedcurves,respectively.A)showsthesignalsamplewithx~`=0.95.B)showsthesignalsamplewithx~`=0.5.C)showsthesignalsampleswithx~`=0.05. 147

PAGE 148

A B CFigure12-5. Limitsonchargino(=neutralino)andLSPmassesforthetau-enrichedscenario.A)showsthesignalsamplewithx~`=0.95.B)showsthesignalsamplewithx~`=0.5.C)showsthesignalsampleswithx~`=0.05. 148

PAGE 149

A BFigure12-6. Limitsonchargino(=neutralino)andLSPmasses.A)showsthetau-dominatedscenario.B)showstheavor-democraticscenariodecayingviaon-shellW=Zgaugebosons. 149

PAGE 150

CHAPTER13CONCLUSIONSAsearchfortheelectroweaksupersymmetryscenarioviacharginoandneutralinoproductionhasbeenpresented.Eventswithexclusive3-lepton,missingtransverseenergyandnob-taggedjetshavebeenexploitedfromtheCMS-recordeddataof19.5fb)]TJ /F9 7.97 Tf 6.59 0 Td[(1duringthewhole2012runwiththeproton-protoncollisionsatthecenter-of-massenergyof8TeV.Toincreasethereachofthesensitivity,exclusivesearchregionsweredenedusingthefollowingvariables:EmissT,MTandM``;themostpowerfulvariablestosegregatesignalfromtheSMWZproduction.NosignicantdeviationfromtheSMpredictionareobservedalongthesesearchregions,thusmassesofcharginoandneutralinowereprobedupto750GeVatthe95%condencelevel.Sofar,thesearethebestconstraintsofthecharginoandneutralino.InterpretationoftheresultsweredoneinthecontextoftheMSSMsimpliedmodels,themodelswhichassumethatthecolorsupersymmetryparticlesareheavyandthereforedecoupledatthisenergyscale,andthebranchingratioofthecharginoandneutralinotoleptonicdecaysareassumedtobe100%. 150

PAGE 151

APPENDIXACOMPARISONOFTHERESULTSBETWEENATLASANDCMSThepurposeofdesigningandconstructingthetwocompetitiveexperiments(CMSandATLAS)attheLHCwasmainlyfortworeasons:toeliminateanypossibleawinphysicssearchesandtomaintaincompetitionamongthescientists.Vastamountofcomponentsareinvolvedinanyphysicssearch,varyingfromdetectorhardwareaspectstotheMCsimulation.Thus,independentresultsfromtwoexperimentsareanexcellentwayofcross-checkingtheresults,eliminatinganypossiblehumanerror.However,todesignandtooptimizeanalysesarenotnecessarilydonethesamewaybetweenthetwoexperiments,andinmostcases,itisuptothescientiststopursuetheresearchonewayoranother.ATLASalsoperformedsearchesforsupersymmetryviachargino-neutralinoproductionusingtheATLAS-recordeddataof20.7fb)]TJ /F9 7.97 Tf 6.58 0 Td[(1at8TeV[ 49 ].Acomparisonbetweenthetwosearchesaresummarizedbelow: 1. leptonselectionsarealmostidenticalinbothsearches,thusnosignicantdifferenceinthesensitivityreachisexpectedduetotheleptonselections; 2. ATLASsearchemploysthesameb-taggingalgorithmasusedinthiswork,butwiththerequirementofLooseworkingpoint(WP)andpTabove20GeV,whereasthroughoutthisthesisb-taggedjetsareselectedwithMediumWPandpTabove30(requirementremainsthesame).LooseWPhasoverall85%efciencyand10%mis-tagrateforlight-avorjets.Incontrast,MediumWPhasoverall70%efciencyand1%mis-tagrateforlight-avorjets.LooseWPleadstoabetterrejectionpowerofthebackgroundassociatedwiththegenuineb-jets.OtherbackgroundslikeWZandZZarealsosuppressedduetohighermis-tagrate.Asdiscussedpreviously,signalisnotexpectedtobeproducedinassociationwithjets,however,jetsareexpectedfromISRorpile-up,whicharepredominantlysoftandlight-avor,andduetohighermis-tagrateofLooseWP,asmallsignallossisexpectedaswell; 3. thebiggestdifferencebetweenthetwoanalysescomesfromthedifferencesofthesearchregiondenition.Inthiswork,wedesignedsearchregionsusingthreepowerfulvariables:MT,EmissTandM``,andwefurthersplitthesampleinexclusivebinsofthesevariables(Section 8.5.3 ).ATLASsearchalsoemploysthesamevariablesbutwithadifferentselection.Table A-1 showsthedenitionofsearchregionsusedinATLASsearch.TheyrequirelowerthresholdsonEmissTandMT 151

PAGE 152

cut,andlowandhighoff-Zregionsarenotdecoupled,asitisdoneinthiswork.Therefore,onlysixexclusivesearchregionsweredesignedandusedinATLASinterpretation A-1 ; 4. thecross-sectionusedbyATLASgroupislowerbyaround10%,asshowninFigure A-1 .Thisdifferenceisduetoamixofthehigssinoandbinocomponentsinneutralino(~02),whereas,inthiswork,neutralinoissolelybasedonamixofbinocomponents.Forthecomparisonpurposes,IusedthesameATLAScross-sectionfortheexclusion. 5. ATLASusesrecorded-dataof20.7fb)]TJ /F9 7.97 Tf 6.58 0 Td[(1,6%higherthanthedatarecordedbytheCMSexperiment.But,thiscanonlyimprovethesensitivitybyanorderofp 1.06,whichisunnoticeablysmall. FigureA-1. Ratioofthechargino-neutralinoproductioncross-sectionsusedbyATLASandCMS. AsTable A-2 indicates,ATLASgroupobservesagoodagreementbetweenpredictedandobservedyieldsallsixsearchregions.However,duetoloosercutsonEmissTandMTvariables,theyobservehigherbackground.Consequently,thesensitivitydegrades,andthiscanbeseeninFigure A-2 and A-3 .TheobservedexclusionlimitintheCMSresult(Figure A-3A )isworseduetothestatisticalupwarductuationon-Z,asshowninFigure 9-3 .Itisworthnotingthatinthisworkweonlyconsiderdecayviaon-shellgaugebosons.Whereas,ATLASgoesbeyondthisandexploresthescenariowhengaugebosonscanbeoff-shellaswell,asshowninFigure A-3B 152

PAGE 153

A BFigureA-2. ExclusionlimitsonmassesofcharginoandLSPfortheavor-democraticscenariowiththexparameterof0.5.A)showstheCMSresult.B)showstheATLASresult. A BFigureA-3. ExclusionlimitsonmassesofcharginoandLSPfortheavor-democraticscenariodecayingviaWandZgaugebosons.A)showstheCMSresult.B)showstheATLASresult. 153

PAGE 154

TableA-1. DenitionofthesearchregionsusedintheATLASanalysis. TableA-2. AsummaryofpredictedandobservedyieldsinallsixsearchregionsusedbyATLAS.Noexcessofeventsarepresented. 154

PAGE 155

APPENDIXBANEXCESSINTHEON-ZREGIONOFTHEOSSFSELECTIONFigure 9-3 showsthatthereissomeexcessoftheobservedyieldsoverthebackgroundpredictionintheon-Zregion(75
PAGE 156

TableB-1. Thesummaryof26eventsfortheOSSFselectionintheon-ZandhighMT(>160GeV)regionwithEmissTabove50GeV.AllrecordedeventsatCMSarelabeledbyrun,lumiandevent:runandlumiindicateinwhichrunnumberandluminositysectionaneventislocated,whereaseacheventatCMShasauniquenumberreferredtoasevent. runlumievent M``[GeV]NjetsHT[GeV]MT[GeV]EmissT[GeV]pTlep1lep1pTlep1lep2pTlep1lep3 195774593934316168 82.3500.00203.80155.4475.33e+49.14)]TJ /F19 10.909 Tf 28.88 -3.96 Td[(28.05+19821213979608257 89.1200.00173.56129.7175.72e)]TJ /F19 10.909 Tf 29.13 -3.96 Td[(63.70e+30.16e+19848710611138391807 92.47144.79216.0397.61120.66e)]TJ /F19 10.909 Tf 29.13 -3.96 Td[(51.05e)]TJ /F19 10.909 Tf 29.13 -3.96 Td[(40.67e+199008195231322993 91.201251.53233.1063.47219.40e+67.15+22.50)]TJ /F19 10.909 Tf -666.55 -17.51 Td[(199318127143294304 90.742232.64280.90211.01138.30e+49.84e)]TJ /F19 10.909 Tf 29.13 -3.96 Td[(40.92e+20049111292772322 91.1400.00257.72115.32147.96)]TJ /F19 10.909 Tf 28.87 -3.96 Td[(83.93)]TJ /F19 10.909 Tf 28.88 -3.96 Td[(23.02+2004919460672856 87.362166.35207.7861.12179.94e)]TJ /F19 10.909 Tf 29.13 -3.95 Td[(49.24e+29.78e)]TJ /F19 10.909 Tf -666.3 -17.5 Td[(200991168246650765 87.2600.00198.2297.69102.10+36.71+27.87)]TJ /F19 10.909 Tf -666.55 -17.51 Td[(201707792956394826 89.772266.28166.91174.6872.74e+51.25e+40.14e)]TJ /F19 10.909 Tf -666.3 -17.51 Td[(202054227190202126 89.9800.00198.7176.47135.86e+68.21+38.81)]TJ /F19 10.909 Tf -666.55 -17.51 Td[(202237109147128677 88.3800.00174.0654.23145.43)]TJ /F19 10.909 Tf 25.84 -3.96 Td[(123.50)]TJ /F19 10.909 Tf 28.88 -3.96 Td[(18.97+2022377351040633794 89.374243.24161.3689.1381.79e+74.59e)]TJ /F19 10.909 Tf 29.13 -3.96 Td[(28.35e+202272221243872441 91.9500.00183.1250.67177.45+165.73+16.30)]TJ /F19 10.909 Tf -666.55 -17.51 Td[(204577391503037640 85.8400.00169.66109.3866.93+31.38)]TJ /F19 10.909 Tf 28.88 -3.96 Td[(10.50+204599258385658326 88.6000.00208.3657.84190.41+152.92+16.65)]TJ /F19 10.909 Tf -666.55 -17.51 Td[(205519288335310582 83.8800.00232.72155.3991.52+91.02)]TJ /F19 10.909 Tf 28.88 -3.96 Td[(29.34+205667121129423645 82.89144.89168.81126.4957.82e)]TJ /F19 10.909 Tf 29.13 -3.96 Td[(29.91)]TJ /F19 10.909 Tf 28.88 -3.96 Td[(15.13+205921439647028523 89.125452.49330.42300.89105.38e+24.43e)]TJ /F19 10.909 Tf 29.13 -3.96 Td[(90.83+2065961719779754 78.744230.73168.70102.08108.54e)]TJ /F19 10.909 Tf 29.13 -3.95 Td[(69.71e+35.70e+20694071100579630 77.162171.09186.0963.86138.71+87.96)]TJ /F19 10.909 Tf 28.88 -3.96 Td[(14.89+2070999565290494 88.6300.00209.03120.02106.12e+54.80)]TJ /F19 10.909 Tf 28.88 -3.96 Td[(44.92+2070999667181797 91.822251.85209.9458.41192.25e+76.05)]TJ /F19 10.909 Tf 28.88 -3.96 Td[(40.28+207214676987349680 97.723257.38264.85108.15162.16e+41.72e+25.94e)]TJ /F19 10.909 Tf -666.3 -17.51 Td[(2074549501287560439 88.14165.28175.3667.18115.49e+75.72+19.40)]TJ /F19 10.909 Tf -666.55 -17.51 Td[(207920597876208309 97.7400.00181.07104.9880.16e)]TJ /F19 10.909 Tf 29.13 -3.96 Td[(45.90e+35.29e)]TJ /F19 10.909 Tf -666.3 -17.51 Td[(208487596870431859 89.54131.12170.8173.9899.06e+55.20)]TJ /F19 10.909 Tf 28.88 -3.96 Td[(32.42+ 156

PAGE 157

Cross-checkfortheWZBackground Inthisanalysis,weemployedthedata-driventechniquestoobtainabetterdescriptionoftheWZpredictionfromthesimulation.Thesetechniqueswerediscussedinsection 8.6.4 .Tomakesurethatthereisnomistakeorabuginthesetechniques,werstcomparethepuresimulationwiththetwodata-drivenmethodsusedtocorrectEmissTandtheleptonresolution.AllthreeestimatesagreeverywellinallbinsofEmissT,asshowninFigure B-1 .ItmustbenotedthatRayleighmethodwasnotemployedinthisthesisbutitsdescriptioncanbefoundinReference[ 39 ]. FigureB-1. ComparisonofEmissTshapebetweenthesimulationandthedata-correctedsimulationmethods.Thereisgoodagreementbetweenthethreemethods. TherecouldbetwoissuesregardingtheWZbackground.Firstly,theEmissTresolutioncouldbeoffanditcouldinuencetheMTshapewithitskinematicendpointsstrongerthantheEmissTshape.Theeffectshouldbelargerclosetothekinematicendpoints,andthustheexcessshouldbelargerfortheregionwhereMTisintherangeof120and160GeVthanintheotherregion,whereMTisabove160GeV.Furthermore,theEmissTresolutioneffectwouldshowupinthetailoftheEmissTdistribution.Figure B-2 showsEmissTandMTdistributions.Figure B-2A showstheEmissTdistributioninthehighMTregion,whichillustratesthattheexcessisvisibleforallEmissTbinsabove50GeV,withapeakofaround70GeV.Figure B-2B showstheMTdistributionforallon-ZeventswithEmissTabove50GeV.TheexcessispresentforalltheMTbins 157

PAGE 158

abovethekinematicendpoint,butitintensieswithlargerMT.Figure B-2C showstheEmissTdistributionforOSSFeventsusingtheon-ZregionwithoutanycutontheMTvariable;EmissTiswell-modeled,whichcanalsobeinferredfromFigure B-1 .Secondly,theproblemcouldberelatedtoaleptonreconstruction.Therefore,itisimportanttolookintotheleptonvariablesseparately. A B CFigureB-2. EmissTandMTdistributionsfortheOSSFselectionintheon-Zregion.A)showstheEmissTdistributionforthehighMT(>160GeV)region.B)showstheMTdistributionforEmissTabove50GeV.C)showstheEmissTdistributionforEmissTabove50GeVwithoutanyrequirementontheMTvariable. Figures B-3A and B-3B showthepTdistributionofaleptonassociatedwiththeWbosonandtheanglebetweentheleptonandtheEmissTvariable,respectively.ThesearethetwocomponentsinadditiontoEmissTwhichenterintothetransversemasscalculation.TheleptonpTdistributionseemstoagreewiththeexpectedshape,althoughthestatisticsisnotgreathere.TheanglebetweentheleptonandtheEmissTvariableseemstobeslightlymoreback-to-backthanexpectedfromthesimulation.However,thenumberofeventsissmallhere,therefore,theresultisnotfullyconclusive.Asanextracheck,wealsoplottedtheangleforthethoseeventscomingfromtheintermediateMTregion(120
PAGE 159

A B CFigureB-3. Distributionoftheleptonvariablesfortheon-Zregion.A)showsthetransversemomentumoftheleptonscomingfromtheWbosonsforthehighMT(>160GeV)region.B)showstheangledistributionbetweentheWleptonandtheEmissTforthehighMT(>160GeV)region.C)showstheangledistributionbetweentheWleptonandEmissTfortheintermediateMT(120GeV
PAGE 160

A BFigureB-4. ElectronmultiplicityandasumofchargesforOSSFselectionintheon-Zregion.A)showselectronmultiplicity.B)showsthesumofthe3-leptonchargesfortheregionwithMT>160GeVandEmissT>50GeV. thejetmultiplicityforthese26events,andtheonlyobviousexcesscanbeseeninthe0jetbin.Figure B-5B showstheHTdistributionwhichpeaksat0.But,anotherinterestingcaseshowsupathighhadronicactivity(around250GeV),butitisagainnotfullyclearwhetherthiscouldbeduetothestatisticaluctuationorduesomeothersources.Thiscouldalsobeduetopile-upjets,butthisshouldonlyhaveasecondordereffectontheMTdistribution.ItmightalsopointtoamissingStandardModelbackground. A BFigureB-5. HadronicactivitiesfortheOSSFselectionintheon-Zregion.A)showstheJetmultiplicity.B)showstheHTdistributionfortheregionwithhighMT(>160GeV)andEmissT>50GeV. 160

PAGE 161

Cross-checkfortheBackgroundProducedinAssociationwithb-Quarks Someextrarareprocesses(liketbZ)oramis-modelingofttcouldbeconsideredaspossibleculprits.Tofurtherinvestigatethis,theb-vetorequirementisrelaxed.Thenumberofb-taggedjetsfortheon-ZandhighMTregionoftheOSSFselectionisshowninFigure B-6 .Asmallexcessispresentedinthe1b-tagbin,butthisexcessisevensmallerandlesssignicantthanintheb-vetoedregion.Therefore,backgroundsproducedwiththegenuineb-quarksdonotseemtobethesourceoftheexcess.TheZmassshapeisalsoinvestigated,butnothingconclusive. FigureB-6. Numberofb-taggedjetsfortheOSSFselectionintheon-ZandhighMT(>160GeV)regionwithEmissTabove50GeV.Theexcessislesspronouncedineventswithnon-zerob-taggedjets. FigureB-7. ThedileptonmassshapeforthehighMT(>160GeV)regionwithEmissTabove50GeV.Asmallshifttolowerdileptonmassesisobserved. 161

PAGE 162

Inconclusion,nosmokinggunfortheexcesswasfoundandtoourknowledge,itismerelyastatisticaluctuation.MagnitudeoftheExcess Thep-valueassociatedwiththeexcesswascalculatedusingthesimpliedZBimetric[ 51 ].TheZBivariableisfastandeasytocalculatethesignicanceofapossiblenewphysicssignal.Theobservedyieldfortheon-Z,MT>120GeVandEmissTabove50GeVregionoftheOSSFselectionis26events,whereastheStandardModelpredictioninthisregionis12.062.95events.ThebackgrounduncertaintyassumesfullcorrelationoftheuncertaintiesacrossthedifferentEmissTbins.TheZBicalculationgivesap-valueof0.021,whichcorrespondstoaZBi(similartosignicance)of2.05.ThesignicancewasalsocalculatedbythrowingtoyexperimentsaccordingtothePoissondistributionaroundthemeanvalueof12.06andtheerrorof2.95.TheuncertaintyonthebackgroundpredictionistakenasaGaussiannuisanceparameter.Thep-valueobtainedwiththismethodis0.012,whichyieldsasignicanceof2.25.Thus,theexcessyieldsaround2-2.3localsignicance.ConsideringseveralSMbackgroundcomponentsandfourchannelsintheOSSFselection,a2uctuationisnotunlikely. 162

PAGE 163

REFERENCES [1] M.PeskinandD.V.Schroeder,AnIntroductionToQuantumFieldTheory,WestviewPress,(1995). [2] F.HalzenandA.Martin,QuarksandLeptons:AnIntroductoryCourseinModernParticle,JohnWiley&Sons,Inc.,(1984). [3] D.Politzer,AsymptoticFreedom:AnApproachtoStrongInteractions, Phy.Rept.14(1974)129 [ INSPIRE ]. [4] G.Bertone,D.HooperandG.SilkParticledarkmatter:evidence,candidatesandconstraints, Phys.Rept.405(2005)279 [ hep-ph/0404175 ][ INSPIRE ]. [5] D.Baueretal.,DarkMatterintheComingDecade:ComplementaryPathstoDiscoveryandBeyond, FERMILAB-CONF-13-166-A-AE [6] S.Martin,ASupersymmetryPrimer,[ hep-ph/9709356 ] [7] H.BaerandX.Tata,WeakScaleSupersymmetry:FromSupereldstoScatteringEvents,CambridgeUniversityPress,CambridgeU.K.(2006). [8] D.Alvesetal.,SimpliedmodelsforLHCnewphysicssearches, J.Phys.G39(2012)105005 [ hep-ph/1105.2838 ][ INSPIRE ]. [9] LHCCollaboration,LHCMachine, JINST3(2008)S08001 [ INSPIRE ]. [10] ATLASCollaboration,TheATLASExperimentattheCERNLargeHadronCollider, JINST3(2008)S08003 [ INSPIRE ]. [11] ALICECollaboration,TheALICEexperimentattheCERNLHC, JINST3(2008)08,S08002 [ INSPIRE ]. [12] LHCbCollaboration,TheLHCbDetectorattheLHC, JINST3(2008)S08005 [ INSPIRE ]. [13] CMSCollaboration,TheCMSexperimentattheCERNLHC, JINST3(2008)S08004 [ INSPIRE ]. [14] Summaryoftheanalysisofthe19September2008incidentattheLHC, TechnicalReport,CERN,Geneva [15] CMSCollaboration,Observationofanewbosonatamassof125GeVwiththeCMSexperimentattheLHC, Phys.Lett.B716(2012)30 [ hep-ph/1207.7235 ][ INSPIRE ]. [16] ATLASCollaboration,ObservationofanewparticleinthesearchfortheStandardModelHiggsbosonwiththeATLASdetectorattheLHC, Phys.Lett.B716(2012)1 [ hep-ph/1207.7214 ][ INSPIRE ]. 163

PAGE 164

[17] R.Fruhwirth,ApplicationofKalmanlteringtotrackandvertextting, Nucl.Instrum.Meth.A262(1987)444 [ INSPIRE ]. [18] CMScollaboration,PerformanceofCMSmuonreconstructioninppcollisioneventsatp s=7TeV, JINST7(2012)P10002 [ hep-ph/1206.4071 ][ INSPIRE ]. [19] https://twiki.cern.ch/twiki/bin/view/CMSPublic/SWGuideMuon [20] CMScollaboration,Electronreconstructionandidenticationatp s=7TeV, CMS-PAS-EGM-10-004 [21] W.A.etal.,ReconstructionofElectronswiththeGaussian-SumFilterintheCMSTrackerattheLHC, J.Phys.G31(2005) [ arXiv:1206.4071 ]. [22] https://twiki.cern.ch/twiki/bin/view/CMS/EgammaCutBasedIdentication [23] CMScollaboration,Performanceoftau-leptonreconstructionandidenticationinCMS, JINST7(2011)P01001 [ arXiv:1109.6034 ]. [24] M.Cacciari,G.Salam,andG.Soyez,Theanti-kTjetclusteringalgorithm, JHEP04(2008)063 [ hep-ph/1109.6034 ]. [25] CMScollaboration,Particle-FlowEventReconstructioninCMSandPerformanceforJets,Taus,andMET, CMS-PAS-PFT-09-001 [26] CMScollaboration,JetPerformanceinppCollisionsat7TeV, CMS-PAS-JME-10-003 [27] CMScollaboration,Identicationofb-quarkjetswiththeCMSexperiment, JINST8(2013)P04013 [ arXiv:1211.4462 ]. [28] CMScollaboration,METperformancein8TeVdata, CMS-PAS-JME-12-002 [29] P.M.Nadolskyetal.,ImplicationsofCTEQglobalanalysisforcolliderobserv-ables, Phys.Rev.D78(2008)013004 [ hep-ph/0802.0007 ][ INSPIRE ]. [30] S.Agostinellietal.,Geant4:asimulationtoolkit, Nucl.Instrum.MethA506(2003)250 [ INSPIRE ]. [31] S.Abdullinetal.,TheFastSimulationoftheCMSDetectoratLHC, J.Phys.:Conf.Ser.331(2011)032049 [ INSPIRE ]. [32] T.Sjstrand,S.MrennaandP.Skands,AbriefintroductiontoPYTHIA8.1, Comput.Phys.Commun.178(2008)852 [ hep-ph/0710.3820 ]. [33] J.Alwalletal.,MadGraph5:goingbeyond, JHEP06(2011)128 [ hep-ph/1106.0522 ][ INSPIRE ]. [34] B.Andersson,G.Gustafson,G.IngelmanandT.Sjostrand,Partonfragmentationandstringdynamics, Phys.Rept.97(1983)31 [ INSPIRE ]. 164

PAGE 165

[35] CMScollaboration,Searchfornewphysicswithsame-signisolateddileptoneventswithjetsandmissingtransverseenergyattheLHC, Phys.Rev.Lett.109(2012)071803 [ hep-ph/1205.6615 ]. [36] ParticleDataGroup,ReviewofParticlePhysics(RPP), Phys.Rev.D86(2012 )[ INSPIRE ]. [37] DCollaboration,Searchforassociatedproductionofcharginosandneutralinosinthetrileptonnalstateusing2.3fb)]TJ /F9 7.97 Tf 6.59 0 Td[(1ofdata, Phys.Lett.B680(2009)34 [ hep-ph/0901.0646 ][ INSPIRE ]. [38] CMScollaboration,Searchforelectroweakproductionofcharginosandneutrali-nosusingleptonicnalstatesinppcollisionsatp s=7TeV, JHEP11(2012)147 [ hep-ph/1209.6620 ][ INSPIRE ]. [39] CMScollaboration,Searchforelectroweakproductionofcharginos,neutrali-nos,andsleptonsusingleptonicnalstatesinppcollisionsatp s=8TeV, CMS-PAS-SUS-13-006 [40] CMScollaboration,MeasurementoftheInclusiveWandZProductionCrossSectionsinppCollisionsatp s=7TeVwiththeCMSexperiment, JHEP10(2011)136 [ arXiv:1107.4789 ][ INSPIRE ]. [41] P.EveraertsandS.Nahn,WCrossSectionMeasurementintheElectronChannelinppCollisionsatp s=7TeV, CERN-THESIS-2011-025 [42] R.C.Gray,C.Kilic,M.Park,S.Somalwar,S.Thomas,BackgroundstoHiggsbosonsearchesfromW?!``(`)asymmetricinternalconversion,[ arXiv:1110.1368 ]. [43] CMSCollaborationSearchforanomalousproductionofmultileptoneventsinppcollisionsatp s=7TeV,JHEP06(2012)169. [44] CMSCollaboration,CMSLuminosityBasedonPixelClusterCounting-Summer2013Update, CMS-PAS-LUM-13-001 [45] M.Chen, LandS:Astatisticaltoolforcalculatinglimitsandsignicance [46] A.L.Read,Presentationofsearchresults:theCLstechnique, J.Phys.G28(2002)2693 [ INSPIRE ]. [47] W.Beenakkeretal.,TheProductionofCharginos,Neutralinos,andSleptonsatHadronColliders, Phys.Rev.Lett.83(1999)3780 [ hep-ph/9906298 ]. [48] W.Beenakkeretal.,Erratum:productionofcharginos,neutralinos,andsleptonsathadroncolliders, Phys.Rev.Lett.100(2008)029901 165

PAGE 166

[49] ATLASCollaboration,Searchfordirectproductionofcharginosandneutralinosineventswiththreeleptonsandmissingtransversemomentumin21fb)]TJ /F9 7.97 Tf 6.59 0 Td[(1ofppcollisionsatp s=8TeVwiththeATLASdetector, ATLAS-CONF-2013-035 [50] CMSCollaboration,Searchforelectroweakproductionofcharginos,neutrali-nos,andsleptonsusingleptonicnalstatesinppcollisionsatp s=8TeV, CMS-PAS-SUS-12-022 [51] R.D.Cousins,J.T.LinnemannandJ.Tucker,Evaluationofthreemethodsforcal-culatingstatisticalsignicancewhenincorporatingasystematicuncertaintyintoatestofthebackground-onlyhypothesisforaPoissonprocess, Nucl.Instrum.Meth.A595(2008)480 [ arXiv:0702156 ]. 166

PAGE 167

BIOGRAPHICALSKETCH NikolozSkhirtladzewasborninTbilisi,Georgia(thecountry).HegrewupinTbilisiandattended6thGymnasiumHighSchool.Duringhisteenageperiod,hewasalwaysamazedbyandadmiredalltheselargescalescienticprojectsandscienticrevolutionaryideas,butheneverwasinvolvedinscience.Infact,hisoneofthemaindreamswastobecomeaprofessionalrugbyplayerwhileplayingrugbywiththerugbyclubAcademia.But,hisdreamcametotheendwhenhewasinjuredduringthegameandhadtoabandonhisrugbycarrier.Then,hedecidedtorealizehishobby-dreamtobecomescientist.HewassuccessfullyenrolledtothephysicsdepartmentattheTbilisiStateUniversity.Hebroadenedhisphysicsandmathematicalskillsthere,andthenafterearningB.ScandM.Scdegrees,hedecidedtogoabroadforhisgraduatestudies.HeenrolledattheUniversityofFloridaforgraduateschoolinphysics,wherehejoinedtheexperimentalhigh-energyphysicsgroupandlatertheCMScollaborationunderthedirectionofProfessorAndreyKorytov.HeenjoysbeingapartoftheCMScollaboration,andtoparticipatinginandcontributingtocuttingedgeresearch. 167