Search for high-mass new phenomena in the dilepton final state using proton–proton collisions at $\sqrt{s}=13$ TeV with the ATLAS detector

Contents lists available atScienceDirect

Physics Letters B

www.elsevier.com/locate/physletb

Search for high-mass new phenomena in the dilepton final state using proton–proton collisions at √

s = ^{13 TeV with the ATLAS detector}

.TheATLAS Collaboration^

a r t i c l e i n f o a b s t ra c t

Articlehistory:

Received14July2016

Receivedinrevisedform17August2016 Accepted24August2016

Availableonline30August2016 Editor:W.-D.Schlatter

Asearchisconductedforbothresonantandnon-resonanthigh-massnewphenomenaindielectronand dimuonfinalstates.Thesearchuses3.2fb⁻¹ofproton–protoncollisiondata,collectedat√

s=^13TeV bytheATLASexperimentattheLHCin2015.Thedileptoninvariantmassisusedasthediscriminating variable. NosignificantdeviationfromtheStandardModel predictionisobserved;thereforelimits are setonthesignalmodelparametersofinterestat95%credibilitylevel.Upperlimitsaresetonthecross- section timesbranchingratio for resonancesdecaying todileptons, and thelimits are converted into lowerlimitsontheresonancemass,rangingbetween2.74 TeV and3.36 TeV,dependingonthemodel.

Lowerlimitsontheqq contactinteractionscalearesetbetween16.7 TeVand25.2 TeV,alsodepending onthemodel.

©^{2016TheAuthor.PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense} (http://creativecommons.org/licenses/by/4.0/).FundedbySCOAP³.

1. Introduction

Thedilepton(ee or μμ)final-statesignaturehasexcellentsen- sitivity to a wide variety of new phenomena expected in theo- riesbeyondtheStandardModel(SM).Itbenefitsfromhighsignal selection efficiencies and relatively small, well-understood back- grounds.

Models with extended gauge groups often feature additional U(1) symmetries with corresponding heavy spin-1 Z^ bosons whose decays would manifest themselves as narrow resonances inthedileptonmassspectrum.GrandUnifiedTheories(GUT)have inspired modelsbased onthe E6 gauge group [1,2], which,fora particularchoiceofsymmetry-breakingpattern,includestwoneu- tralgaugebosonsthatmixwithanangleθE6.Thisyieldsaphysical statedefinedby Z^(θE6)=^Z_ψ^cosθE6+^Z_{χ sin}θE6,wherethegauge fields Z_ψ^ and Zχ are^ associated withtwo separate U(1) groups resultingfromthebreakingofthe E6 symmetry. All Z^ signalsin thismodelaredefinedbyspecific valuesofθE₆ rangingfrom−π

to π, and the sixcommonly motivated casesare investigatedin this search, namely Z_ψ^, Z^η , Z^_N, Z^_I, Z^_S, and Z^χ . The widths of thesestatesvaryfrom0.5%to1.2%oftheresonancemass,respec- tively.In addition tothe GUT-inspired E₆ models, the Sequential Standard Model(SSM)[2] provides acommon benchmarkmodel that includes a Z_SSM^ boson with couplings to fermions identical to those of the SM Z boson. This search is also sensitive to a seriesofmodelsthatpredictthepresenceofnarrowdileptonres-

 E-mailaddress:atlas.publications@cern.ch.

onances;howeverconstraintsarenotexplicitlyevaluatedonthese models. Theseinclude theRandall–Sundrum (RS)model [3]with a warped extra dimension giving rise to spin-2 graviton excita- tions,thequantumblackholemodel[4],theZ*model[5],andthe minimalwalkingtechnicolourmodel[6].

Some modelsofphysicsbeyondtheSMresultinnon-resonant deviations fromthe predicted SM dilepton massspectrum. Com- positenessmodelsmotivatedbytherepeatedpatternofquarkand lepton generations predict new interactions involving their con- stituents.Theseinteractionsmayberepresentedasacontactinter- action (CI)betweeninitial-statequarksandfinal-state leptons [7, 8].Othermodelsproducingnon-resonanteffects,butnotexplicitly evaluated here, are models with large extra dimensions [9] mo- tivated by the hierarchy problem. The following four-fermion CI Lagrangian [7,8]isusedtodescribeanewinteraction orcompos- itenessintheprocessqq→ ⁺⁻:

L= ^g2

²[ηLL(q_Lγμq_L) (Lγ^μL)+ηRR(q_Rγμq_R) (Rγ^μR) (1) +ηLR(q_Lγ_μqL) (Rγ^μR)+ηRL(q_Rγ_μqR) (Lγ^μL)] ,

where g isacouplingconstantsettobe√

4πbyconvention, is the CI scale, andq_L,R andL,R are left-handedand right-handed quark and leptonfields, respectively. Thesymbol γμ denotes the gamma matrices,andthe parameters ηi j,where i and j areLor R(leftorright),definethechiralstructureofthenewinteraction.

Differentchiralstructuresareinvestigatedhere,withtheleft–right (right–left) model obtainedby setting ηLR= ±^{1 (}ηRL= ±^{1) and} allotherparameterstozero.Likewise,theleft–leftandright–right models are obtained by setting the corresponding parameters to http://dx.doi.org/10.1016/j.physletb.2016.08.055

0370-2693/©^{2016TheAuthor.PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense}(http://creativecommons.org/licenses/by/4.0/).Fundedby SCOAP³.

±^{1,andtheotherstozero.Thesignof} ηi j determineswhetherthe interferencebetweentheSM Drell–Yan(DY) qq→^Z/γ^∗_{→ }⁺⁻ processandtheCIprocessisconstructive(ηi j= −^1)ordestructive (ηi j= +^1).

Themostsensitive previous searches fora Z^ decayingto the dileptonfinalstatewerecarriedoutbytheATLASandCMSCollab- orations[10,11].Using20 fb⁻¹of pp collisiondataat√

s=^{8 TeV,} ATLAS seta lower limit at95% credibility level(CL) onthe Z_SSM^ pole mass of 2.90 TeV for the combined ee and μμ channels.

Similarlimitswere setby CMS.Themoststringentconstraintson CI searches are also provided by the CMS and ATLAS Collabora- tions[11,12].The strongestlower limitsonthe qq CIscale are >21.6 TeV and >17.2 TeV at 95% CL for constructive and destructiveinterference,respectively,inthecaseofleft–leftinter- actionsandgivenauniformpositivepriorin1/ ².Previousdilep- tonsearchesatATLAShavealsosetlowerlimitsontheresonance massinothermodelssuchas:anRSgravitonupto2.68 TeV,quan- tumblackholesat3.65 TeV,theZ*bosonat2.85 TeV,andminimal walkingtechnicolourupto2.27 TeV[10].Similarlowerlimitswere setbyCMSwhereequivalentsearcheswereperformed[11].

Inthisletter,asearchforresonantandnon-resonantnewphe- nomenaispresentedusingtheobservedee and μμmassspectra extractedfrompp collisionswithintheATLASdetectorattheLarge HadronCollider(LHC)operatingat√

s=^13TeV.The pp collision datacorrespondtoanintegratedluminosityof3.2fb⁻¹.Theanal- ysisandinterpretationofthesespectrarelyprimarilyonsimulated samplesofsignalandbackgroundprocesses.The Z masspeakre- gionisusedtonormalisethebackgroundcontributionandperform cross-checksofthe simulatedsamples.The interpretationis then performedtakingintoaccounttheexpectedshapeofdifferentsig- nalsinthedileptonmassdistribution.

2. ATLASdetector

The ATLAS experiment[13,14] at the LHC is a multi-purpose particle detectorwith a forward–backward symmetric cylindrical geometry and near 4π coverage in solid angle.¹ It consists of an inner tracking detector surrounded by a thin superconduct- ingsolenoidprovidinga2 Taxial magneticfield,electromagnetic and hadronic calorimeters, and a muon spectrometer. The inner trackingdetector(ID)coversthepseudorapidityrange|η_|<2.5.It consistsofsiliconpixel,siliconmicrostrip,andtransition–radiation tracking detectors. Lead/liquid-argon (LAr) sampling calorimeters provide electromagnetic (EM) energy measurements with high granularity. A hadronic (steel/scintillator-tile) calorimeter covers thecentral pseudorapidity range(|η_|<1.7). The endcapandfor- wardregions areinstrumented withLArcalorimetersforEM and hadronic energy measurements up to |η|=⁴.9. The total thick- nessoftheEMcalorimeterismorethantwentyradiationlengths.

The muon spectrometer (MS) surrounds the calorimeters and is basedon three large superconducting air-core toroids witheight coilseach.Thefieldintegralofthetoroidsrangesbetween2.0and 6.0T·^{mformostofthedetector.Itincludesasystemofprecision} tracking chambers and fast detectors for triggering. A dedicated trigger system is used to select events. The first-level trigger is implementedinhardwareandusesthecalorimeterandmuonde- tectorstoreduce the acceptedeventratefrom40MHz to below

1 ATLASCollaboration usesaright-handedcoordinatesystemwithitsoriginat thenominalinteractionpoint(IP)inthecentreofthedetectorandthez-axisalong thebeampipe.Thex-axispointsfromtheIPtothecentreoftheLHCring,and they-axispointsupwards.Cylindricalcoordinates(r,φ)areusedinthetransverse plane,φbeingtheazimuthalanglearoundthez-axis.Thepseudorapidityisdefined intermsofthepolarangleθasη= −^{ln tan}(θ/2).Angulardistanceismeasuredin units ofR≡

(η)²+ (φ)^2.

100kHz.Thisisfollowedbyasoftware-basedtriggerthatreduces theacceptedeventrateto1kHzonaverage.

3. DataandMonteCarlosamples

Thedatasampleusedinthisanalysiswas collectedduringthe 2015LHC run with pp collisions at√

s=^{13 TeV.After selecting} periodswithstablebeamsandrequiringthatrelevantdetectorsys- temsarefunctional,thedatasetusedfortheanalysiscorresponds to3.2fb⁻¹ ofintegratedluminosity.Eventqualityisalsochecked to remove those events which contain noise bursts or coherent noiseinthecalorimeters.

Modellingofthevariousbackgroundsourcesreliesprimarilyon MonteCarlo(MC)simulation.Thedominantbackgroundcontribu- tionarisesfromtheDYprocess[15].Otherbackgroundsourcesare top-quark [16] anddiboson (W W , W Z , Z Z ) [17] production. In thecaseofthedielectronchannel,multi-jetandW+jets processes also contribute due to the misidentification of jets as electrons.

A data-drivenmethod,describedinSection 5,isusedtoestimate thesebackgroundcontributions.Themulti-jetandW+jets contri- butioninthedimuonchannelisnegligible.

DY events are simulated using Powheg-box v2 [18] at next- to-leading order(NLO) in QuantumChromodynamics (QCD), and interfacedtothe Pythia 8.186[19]partonshowermodel.TheCT10 parton distribution function (PDF) set [20] is used in the ma- trixelementcalculation.The AZNLO[21] setoftunedparameters (“tune”)isused,withtheCTEQ6L1PDFset[22],forthemodelling ofnon-perturbativeeffects.TheEvtGenv1.2.0program[23]isused forpropertiesofthebottomandcharmhadron decays. Photos++

version3.52[24]isusedforQuantumElectrodynamic(QED)emis- sionsfromelectroweakverticesandchargedleptons.Eventyields are corrected with a mass-dependent rescaling to next-to-next- to-leading order(NNLO)in theQCDcoupling constant,computed with VRAP 0.9 [25] andthe CT14NNLO PDF set [26]. The NNLO QCD corrections are a factor of ∼ ^{0.98 at m}=^{3 TeV. Mass-} dependent electroweak (EW) corrections are computed at NLO with Mcsanc 1.20 [27]. The NLO EW corrections are a factor of

∼ ^{0.86 atm}=^{3 TeV. Those include} photon-induced contribu- tions (γ γ →  ^{via t- and u-channel processes) computed with} theMRST2004QEDPDFset[28].

Dibosonprocesseswithfourchargedleptons,threechargedlep- tonsandoneneutrino, ortwochargedleptons andtwo neutrinos are simulated using the Sherpa 2.1.1 generator [29]. Matrix ele- ments contain all diagrams with four electroweak vertices. They are calculated forup to one (4, 2+²ν) or no additional par- tons(3+¹ν)atNLO.Diboson processeswithoneofthebosons decayinghadronicallyandtheotherleptonicallyaresimulatedus- ing the Sherpa 2.1.1 generator. Theyare calculatedfor up toone ( Z Z ) or no (W W , W Z ) additional partons at NLO. All are cal- culated withup tothree additionalpartonsatleading-order (LO) using the Comix [30] and OpenLoops [31] matrix element gen- erators and merged with the Sherpa parton shower [32] using the ME+^PS@NLO prescription [33]. The CT10 PDF set is used in conjunction with dedicated parton shower tuning developed by the Sherpa authors.The Sherpa dibosonsamplecross-sectionwas scaled down to account for its use of αQED=¹/129 rather than 1/132corresponding totheuse ofcurrentPDGparameters asin- puttotheGμ scheme.

For the generation of t¯t and single top quarks in the W t- channel and s-channel the Powheg-box v2 generator with the CT10 PDF set in the matrix element calculations is used. EW t-channelsingle-top-quarkeventsaregeneratedusingthe Powheg- boxv1generator. Thisgeneratorusesthefour-flavour schemefor theNLOmatrixelementcalculationstogetherwiththefixed four- flavourPDFsetCT10f4.Foralltop-quarkprocesses,top-quarkspin

correlationsare preserved(for t-channel, top quarks are decayed using MadSpin [34]). The parton shower, fragmentation, andthe underlying eventare simulatedusing Pythia 6.428 [35] withthe CTEQ6L1PDFsetandthePerugia2012tune(P2012)[36].Thetop- quarkmassissetto172.5 GeV.TheEvtGenv1.2.0programisused forpropertiesofthebottomandcharmhadrondecays.Thett and¯ single-top-quarkMCsamplesare normalisedtoacross-section as calculatedwiththe Top++2.0program [37], whichis accurateto NNLOinperturbativeQCD,includingresummationofnext-to-next- to-leadinglogarithmicsoftgluonterms.

Resonantandnon-resonantsignalprocessesareproducedatLO using Pythia 8.186 with the NNPDF23LO PDF set [38] and A14 tune [39] for eventgeneration, partonshowering andhadronisa- tion.InthecaseofZ^production,interferenceeffects(suchaswith DY production) are not included.However, forthe production of non-resonant signalevents,both theDY andCI eventsaregener- atedtogetherinthesamesampletoaccountforthesignificantin- terferenceeffectsbetweenthosetwoprocesses.Higher-orderQCD corrections are computed asfor the DY backgroundand applied to both the resonant andnon-resonant MC samples. EW correc- tionsarenotappliedtotheresonantMCsamplesduetothelarge modeldependence.However,thesecorrectionsare appliedtothe non-resonantMCsamplesastheyinvolveinterferencebetweenthe DYandCIprocesses.Moreover,includingtheEWcorrectionsleads toamoreconservativeestimatewhensettingexclusionlimits.The generatorsettingsandcorrectionsdescribedherearealsousedto computethesignalcross-sectionsandbranchingratios.

The detectorresponse is simulated with Geant 4 [40,41] and the events are processed with the same reconstruction software asused forthedata.Furthermore,thedistributionofthenumber ofadditionalsimulated pp collisionsinthesameorneighbouring beamcrossings(pile-up)isaccountedforbyoverlaying simulated minimum-biaseventsandre-weighting theMCto matchthedis- tributionobservedinthedata.

4. Eventselection

Electronsare reconstructedin thecentral regionofthe ATLAS detectorcoveredbythetrackingdetectors(|η|<2.47),bycombin- ingcalorimetricandtrackinginformationasdescribedinRef.[42].

The transition region between the central and forward regions of the calorimeters, in the range 1.37≤ |η|≤¹.52, exhibits de- graded energy resolution and is therefore excluded. A likelihood discriminantisbuilttosuppresselectroncandidatesresultingfrom hadronicjets, photonconversions, Dalitzdecays andsemileptonic heavy-flavour hadron decays. The likelihood discriminant utilises lateral andlongitudinal shower shape, tracking andcluster–track matching quantities.Several operating points are defined forthe likelihooddiscrimination,asdescribedinRef.[42].Inthisanalysis, theMedium workingpointisusedinthesearch,andtheVeryLoose andLoose workingpoints areusedinthedata-driven background estimationdescribedinSection5.Inadditiontothelikelihooddis- criminant,selectioncriteriabasedontrackqualityareapplied.The selectionefficiencysmoothly decreasesfrom96% to95% forelec- tronswithtransverseenergy(E_T) between500 GeV and1.5 TeV.

The selection efficiency modelling isevaluated in the datausing a tag-and-probe method [43] up to ET of 500 GeV and the un- certaintiesduetothemodellingoftheshowershapevariablesare evaluatedasdescribedinSection6.Theelectronenergyscaleand resolution has been calibrated up to ET of 500 GeV using data takenat√

s=^{8TeV[44].Theenergyresolutionforhigh-E}T elec- tronsisapproximately1%.Tosuppressbackgroundfrommisiden- tifiedjetsaswellasfromlight- andheavy-flavourhadron decays insidejets, electrons arerequiredto satisfythe calorimeter-based and track-based isolation criteria with a fixed efficiency of 99%

over thefullrangeofelectron momentum.Thecalorimeter-based isolationreliesontheratioofthetotalenergydepositedinacone ofsizeR=⁰.2 centredattheelectroncluster barycentretothe electron E_T.Likewise, thetrack-basedisolation reliesontheratio ofthe scalarsumoftransverse momentaoftrackswithin a cone ofsizeR=^10GeV/p_Ttothetransversemomentum(p_T)ofthe electron track. The tracksare required to originate fromthe pri- maryvertex(definedasthevertexwiththehighestsumoftrack p²_T),havep_T>1GeV,|η|<2.5,andmeettrackqualitycriteria.

Candidatemuontracksare,atfirst,reconstructedindependently in the ID and the MS [45]. The two tracks are then used asin- put to a combined fit which takes into account the energy loss in the calorimeter and multiple-scattering effects. The ID track used forthe combinedfit isrequired tobe within the ID accep- tance,|η_|<2.5,andtohaveaminimumnumberofhitsineachID sub-system.Muoncandidates intheoverlapoftheMS barreland endcap region (1.01<|η_|<1.10) are rejecteddue to thepoten- tialfor pT mismeasurementresultingfromrelative barrel–endcap misalignment. Inordertoreduce the backgroundfromlight- and heavy-hadron decays insidejets, muonsare requiredto fulfilrel- ativetrack-basedisolation requirementswitha fixedefficiencyof 99%, asdefinedabove forelectron candidates.Theselectedmuon candidatesmustalsopassneartheprimaryinteractionpointinthe z coordinate to suppress cosmic-ray background. Since momen- tum resolution is a key ingredient of this analysis, muon tracks are required to have at least three hits in each of three preci- sion chambers in the MS andnot to traverse regions of the MS whichare poorlyaligned. Thisrequirementreducesthe muonre- construction efficiencyby about20%formuonswitha pT greater than 1.5 TeV. Finally, the q/p (charge divided by momentum) measurements performed independently in the ID and MS must agree within seven standard deviations,calculated fromthe sum inquadratureoftheIDandMSmomentumuncertainties.

To search for high-mass dilepton signatures of new physics, requirements are applied to the data and MC samples to select events with two high-ET electrons or high-pT muons, satisfying the criteria described above. In the dielectron channel, a two- electron triggerbased onthe Loose identificationcriteriawithan ET thresholdof 17 GeV for each electron is used.Events in the dimuon channel are requiredto pass at leastone oftwo single- muontriggerswithpT thresholdsof26 GeV and50 GeV,withthe former alsorequiringthe muonto be isolated.These triggersse- lect events from a simulatedsample of Zχ with^ a pole mass of 3 TeV with anefficiencyofabout87% and94% forthedielectron anddimuonchannels,respectively.Electron(muon)candidatesare required to have E_T (p_T) greater than 30 GeV and havea trans- verseimpact parameterconsistentwiththebeam-line.Events are requiredtohaveatleastone reconstructedprimaryvertexandat leastonepairofsame-flavourleptoncandidates.

Onlytheelectron(muon)pairwiththehighestscalarsumofET (p_T)isretainedineacheventandanopposite-chargerequirement is applied in the dimuon case. The opposite-charge requirement is not applied in the dielectronchannel due to higherchance of chargemisidentificationforhigh-ET electrons.

Energy (momentum) calibration and resolution smearing are appliedtoelectron(muon)candidatesinthesimulatedsamplesto matchthe performance observedindata [44,45].Event-levelcor- rectionsareappliedinthesimulatedsamplestomatchthetrigger, reconstructionandisolationefficiencies.

Representative values of the total acceptance times efficiency fora Z^χ bosonwithapolemassof3 TeV are 69%inthedielectron channeland46%inthedimuonchannel.