Combination of searches for $\mathit{WW}, \mathit{WZ}$, and \mathit{ZZ} resonances in $\mathit{pp}$ collisions at $\sqrt{s}=8$ TeV with the ATLAS detector

Contents lists available atScienceDirect

Physics Letters B

www.elsevier.com/locate/physletb

Combination of searches for W W , W Z , and Z Z resonances in pp collisions at √

s = ^{8 TeV with the ATLAS detector}

.ATLASCollaboration^

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

Articlehistory:

Received16December2015

Receivedinrevisedform9February2016 Accepted9February2016

Availableonline11February2016 Editor:W.-D.Schlatter

TheATLASexperimentattheCERNLargeHadronColliderhasperformedsearchesfornew,heavybosons decayingto W W ,W Z and Z Z finalstatesinmultipledecaychannelsusing20.3 fb⁻¹ ofpp collision dataat√

s=^{8 TeV.Inthecurrentstudy,theresultsofthesesearchesarecombinedtoprovideamore} stringenttest ofmodelspredictingheavy resonanceswithcouplingsto vectorbosons.Directsearches forachargeddibosonresonancedecayingtoW Z intheν^^(=μ,e),q^q,¯ νq^{q and}¯ ^{fullyhadronic} finalstatesare combinedandupperlimits ontherateofproductiontimesbranchingratiototheW Z bosonsarecomparedwithpredictionsofanextendedgaugemodelwithaheavyW^boson.Inaddition, directsearchesforaneutraldiboson resonancedecayingtoW W and Z Z intheq^q,¯ νq^q,¯ ^andfully hadronicfinalstatesarecombinedandupperlimitsontherate ofproductiontimesbranchingratioto the W W and Z Z bosons arecompared withpredictionsfor aheavy,spin-2 gravitoninan extended Randall–Sundrummodel wheretheStandardModel fieldsareallowedtopropagateinthe bulkofthe extradimension.

©^{2016CERNforthebenefitoftheATLAS}Collaboration.PublishedbyElsevierB.V.Thisisanopen accessarticleundertheCCBYlicense(http://creativecommons.org/licenses/by/4.0/).FundedbySCOAP³.

1. Introduction

Thenaturalnessargumentassociatedwiththesmallmassofthe recentlydiscoveredHiggs boson[1–4] suggeststhat theStandard Model(SM)isconceivablytobeextendedbyatheorythatincludes additionalparticles and interactions atthe TeV scale. Many such extensionsoftheSM,suchasextendedgaugemodels[5–7],mod- elsofwarpedextradimensions[8–10],technicolour[11–14], and more generic composite Higgs models [15,16], predict the exis- tenceofmassiveresonancesdecayingtopairsofW and Z bosons.

Intheextendedgaugemodel(EGM)[5]anew,chargedvector boson(W^)couplestotheSMparticles.Thecouplingbetweenthe W^andtheSMfermionsisthesameasthecouplingbetweenthe W bosonandtheSMfermions.TheW^W Z couplinghasthesame structureasthe W W Z couplingintheSM,butisscaledbyafac- torc× (^mW/m_W)²,where c isa scalingconstant,mW isthe W boson mass,andm_W is the W^ boson mass.The scaling of the coupling allows the width of the W^ boson to increase approx- imately linearly withmW^ at mW^ ^mW and to remain narrow forc∼^1.Forc=^{1 andm}W^>0.5 TeV the W^ widthisapproxi- mately3.6%ofitsmassandthebranchingratioofthe W^→^{W Z} rangesfrom1.6%to1.2%dependingonm_W.Productioncrosssec-

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

tions in pp collisionsat √

s=^{8 TeV forthe W bosonaswell as} theW^widthandbranchingratiosofW^→^{W Z foraselectionof} W^ bosonmassesintheEGMwithscalefactorc=^{1 aregivenin} Table 1.

SearchesforaW^bosondecayingtoνhavesetstrongbounds on the mass of the W^ when assuming the sequential standard model (SSM) [17,18], which differs from the EGM in that the W^W Z coupling is set to zero. Forc∼^{1 the effect ofthis cou-} plingontheproductioncrosssectionoftheW^bosonattheLHC isverysmall,thustheproductioncrosssectionoftheW^bosonin theSSMandtheEGMis verysimilar.Moreover, duetothesmall branching ratiooftheW^→^{W Z in theEGMwiththe scalefac-} tor c∼^{1, the branching ratios of the W boson to fermions are} approximatelythesameasintheSSM.Nevertheless,modelswith narrowvectorresonanceswithsuppressedfermioniccouplingsre- main viable extensions to theSM, andthus the EGM provides a usefulandsimplebenchmarkinsearches fornarrowvectorreso- nancesdecayingto W Z .

The ATLAS andCMS Collaborations haveset exclusionbounds on the productionand decayof the EGM W^ boson. In searches usingtheν^^ (≡^e, μ) channel,theATLAS [19] andCMS[20]

Collaborations have excluded, at the 95% confidence level (CL), EGM (c=^{1) W bosons decaying to W Z for W masses below} 1.52 TeVand1.55 TeV,respectively. InadditiontheATLAS Collab- oration has excluded EGM (c=^{1) W bosons for masses below}

http://dx.doi.org/10.1016/j.physletb.2016.02.015

0370-2693/©^{2016CERNforthebenefitoftheATLAS}Collaboration.PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense (http://creativecommons.org/licenses/by/4.0/).FundedbySCOAP³.

1.59 TeVusing the q^q¯ ^{[21] channel, andbelow 1.49 TeV using} theνqq¯[22]channel.Thesehavealsobeenexcludedwithmasses between1.3and1.5 TeVandbelow1.7 TeVbytheATLAS[23]and CMS[24]Collaborations,respectively,usingthefullyhadronicfinal state.

Diboson resonances are also predicted in an extension ofthe original Randall–Sundrum (RS) [8–10] model with a warped ex- tradimension.Inthisextension totheRSmodel[25–27],theSM fields are allowed to propagate in the bulk of the extra dimen- sion,avoiding constraintson the original RSmodelfrom flavour- changing neutral currents and from electroweak precision mea- surements.Thisso-calledbulk-RS modelischaracterised bya di- mensionlesscouplingconstantk/ ¯M_Pl∼^{1,wherek isthecurvature} of the warped extra dimension, and M¯_Pl=^MPl/√

8π is the re- ducedPlanckmass.InthismodelaKaluza–Klein excitationofthe spin-2graviton,G^∗,candecaytopairsofW orZ bosons.Forbulk RS models with k/ ¯M_Pl=^{1 and for G∗ masses between 0.5 and} 2.5 TeV, the branching ratio of G^∗ to W W ranges from 34% to 16% and the branching ratioto Z Z ranges from 18% to 8%. The G^∗ widthrangesfrom3.7%to6.2%dependingontheG^∗ mass.Ta- ble 1 lists widths, branching ratio to W W and Z Z for G^∗, and productioncross sectionsin pp collisions at 8 TeVin thesebulk RSmodels.

The ATLAS Collaboration has excluded, at the 95% CL, bulk G^∗ →^{Z Z with masses below 740 GeV, using the} qq chan-¯ nel[21],aswell asbulk G^∗→^{W W withmassesbelow760 GeV,} usingtheνq^{q channel}¯ ^assumingk/ ¯M_Pl=^{1[22].TheCMSCollab-} orationhasalsoexcluded atthe95% CLtheG^∗ oftheoriginal RS model,decayingtoW W and Z Z withmassesbelow1.2 TeVusing thefullyhadronicfinalstate[24]andhassetlimitsontheproduc- tionanddecayofgenericdibosonresonancesusingacombination ofqq,¯ νqq and¯ fullyhadronicfinalstates[28].

To improve the sensitivity to new diboson resonances, this article presents a combination of four statistically independent searchesfordibosonresonancespreviouslypublishedbytheATLAS Collaboration [19,21–23]. The searches are combined while con- sideringthe correlations betweensystematicuncertainties inthe differentchannels.Thefirstsearch,sensitivetochargedresonances decaying to W Z , uses the ν^^ [19] final state. The second search,sensitivetochargedresonancesdecayingtoW Z and neu- tralresonancesdecayingto Z Z ,usestheqq final¯ state[21].The thirdsearch,sensitivetochargedresonancesdecayingtoW Z and neutralresonancesdecayingtoW W ,usestheνq^{q final}¯ ^state[22].

Finally,the fourth search, sensitive tocharged resonances decay- ing to W Z andtoneutralresonances decayingtoeither W W or Z Z ,usesthefullyhadronicfinal state [23].Due tothelarge mo- menta ofthebosons fromthe resonancedecay, theresonancein thischannel is reconstructed withtwo large-radius jets, andthe fullyhadronicchannelishereafterreferredtoasthe J J channel.

TosearchforachargeddibosonresonancedecayingtoW Z the

ν^^, qq,¯ νqq,¯ and J J channelsare combined.The result of this combination is interpreted using the EGM W^ model with c=^{1 asabenchmark.}

TosearchforneutraldibosonresonancesdecayingtoW W and Z Z theq^q,¯ νq^q,¯ ^and J J channels arecombined,andtheresult isinterpretedusingthebulkG^∗,assumingk/ ¯M_Pl=^1,asabench- mark.

The ATLAS Collaborationhasperformed additionalsearches in whichnew dibosonresonancescould manifest themselvesasex- cessesoverthebackgroundexpectation. Intheanalysispresented inRef.[29]the^^,νν,qq and¯ qq¯νν finalstateshavebeen explored inthecontext ofthesearch fora new,heavy Higgsbo- son.Also, in thecontext ofsearchesfor darkmattera final state ofahadronically decayingbosonandmissingtransversemomen- tum[30],andafinalstateofaleptonicallydecaying Z bosonand

missingtransversemomentumhavebeenexplored[31].Thesead- ditional searches are not included in this combination. They are not expected to contribute significantly to the sensitivity of the combined search duetothe lower branching ratioincaseof the leptonic channels, andthe useofonly narrowjetsincaseof the qq¯ννfinalstate.

2. ATLASdetectoranddatasample

The ATLAS detector isdescribed in detail in Ref. [32]. It cov- ers nearly the entire solid angle¹ around the interaction point and has an approximatelycylindrical geometry. It consistsof an inner tracking detector (ID) placed within a 2 T axial magnetic fieldsurroundedbyelectromagneticandhadroniccalorimetersand followedbyamuonspectrometer(MS)withamagneticfieldpro- videdbyasystemofsuperconductingtoroids.

Theresultspresentedinthisarticleusethedatasetcollectedin 2012byATLASfromtheLHC pp collisionsat√

s=^{8 TeV,usinga} single-lepton(electronormuon)trigger[33] witha p_T threshold of24 GeV,orasinglelarge-radiusjettriggerwithapTthresholdof 360 GeV. Theintegratedluminosityofthisdatasetafterrequiring data quality criteriato ensure that all detectorcomponents have beenoperational duringdata takingis 20.3 fb⁻¹.Theuncertainty on theintegratedluminosityis ±².8%.It isderived followingthe methodologydetailedinRef.[34].

3. Signalandbackgroundsamples

Theacceptanceandthereconstructedmassspectrafornarrow resonancesareestimatedwithsignal samplesgeneratedwithres- onance masses between 200 and 2500 GeV, in 100 GeV steps.

The bulk G^∗ signal events are produced by CalcHEP 3.4 [35]

withk/ ¯MPl=1.0,andthe W^ signal samplesare generatedwith Pythia 8.170 [36], setting the coupling scale factor c =^{1. The} factorisation and renormalisation scales are set to the gener- ated resonance mass. The hadronisation and fragmentation are modelled with Pythia 8 in both cases, and the CTEQ6L1 [37]

(MSTW2008LO[38]) partondistributionfunctions(PDFs)areused for the G^∗ (W^) signal. The leading-order cross sections and branching ratios for the W^ andbulk G^∗ signal samples for se- lectedmasspointsandassumedvaluesofthecouplingparameters areprovidedinTable 1.

The backgroundsinthedifferentdecaychannelsare modelled with simulated event samples. The W+^{jets and Z} +^{jets back-} groundsaregeneratedusing Sherpa 1.4.1[39]withCT10PDFs[40].

A separatesampleisgeneratedusing Alpgen 2.14[41]toestimate systematiceffects,usingCTEQ6L1PDFsand Pythia 6[36]forfrag- mentationandhadronisation.

The W+^{jets and Z}+jets productioncrosssectionsarescaled tonext-to-next-to-leading-order(NNLO)calculations[42].Thetop quarkpair,s-channelsingle-topquarkandW t processesaremod- elled by the MC@NLO 4.03 generator [43,44] with CT10 PDFs, interfaced to Herwig [45] for fragmentation and hadronisation and Jimmy [46] for modelling of the underlying event. The top quark pairsampleisscaledtotheproductioncrosssectioncalcu- latedat NNLOin QCDincluding resummationofnext-to-next-to- leading logarithmic soft gluon terms withTop++2.0 [47–52]. The

1 ATLASusesaright-handedcoordinatesystemwithitsoriginatthenominalin- teraction point(IP)inthecentreofthedetector andthe z-axisalongthebeam pipe.Thex-axispointsfromtheIPtothecentreoftheLHCring,andthe y-axis pointsupward.Cylindricalcoordinates(r,φ)areusedinthetransverseplane,φbe- ingtheazimuthalanglearoundthebeampipe.Thepseudorapidityisdefinedin termsofthepolarangleθasη= −^{ln tan}(θ/2),andthedistancein(φ,η)spaceas R≡

( φ)²+ ( η)².

Table 1

Leading-ordercrosssections,widths,andbranchingratiosfortheW^bosonintheEGMwithscalefactorc=^{1 andfortheG∗inthebulkRSmodelwithk}/MPl=^{1 inpp} collisionsat√

s=8 TeV foravarietyofmasspoints.

m [TeV]

W

[GeV]

σ(W^) [fb]

BR(W^→W Z) [%]

G_RS [GeV]

σ(G^∗) [fb]

BR(G^∗→W W) [%]

BR(G^∗→Z Z) [%]

0.5 18.0 2.00×10⁵ 1.6 18.4 3.11×10³ 34 18

1.0 36.0 1.17×^{104 1.3 55}.4 5.60×^{101 19 10}

1.5 54.0 1.44×^{103 1.3 89}.5 3.14×^{100 17 8}

2.0 73.3 2.42×^{102 1.2 122}.5 2.90×^{10−1 16 8}

2.5 90.7 5.31×10¹ 1.2 155.0 3.20×10⁻² 16 8

t-channel single-top events are generated by AcerMC [53] with CTEQ6L1PDFsand Pythia 6forhadronisation.Thedibosonevents areproduced with the Herwig generator andCTEQ6L1PDFs, ex- ceptfortheν^^channelwhichuses POWHEG[54,55]interfaced to Pythia 6.Thedibosonproductioncrosssectionsarenormalised tonext-to-leading-order predictions[56].Additionaldibosonsam- plesfortheνq^{q channel}¯ ^{areproducedwith}the Sherpa generator.

QCD multijetsamples are simulated with Pythia 6, Herwig, and POWHEGinterfacedto Pythia 6.

GeneratedeventsareprocessedwiththeATLASdetectorsimu- lationprogram[57]basedontheGEANT4package[58].Signaland background samples simulated or interfacedwith Pythia use an ATLASspecifictuneof Pythia[59].Effectsfromadditionalinelastic pp interactions (pile-up)occurringinthesameandneighbouring bunch crossingsare takeninto account by overlaying minimum- biaseventssimulatedby Pythia 8.

4. Objectreconstructionandselection

Thesearch channelsincludedinthecombinationpresentedin thisarticleusereconstructedelectrons, muons,jetsandthemea- surementofthemissingtransversemomentum.

Electron candidates are selected from energy clusters in the electromagneticcalorimeterwithin|η|<2.47,excluding thetran- sition region between the barrel and the endcap calorimeters (1.37<|η|<1.52), that match a track reconstructed in the ID.

Electronssatisfying‘tight’identificationcriteriaareusedtorecon- struct W→^eν candidates,while Z→^{ee are} reconstructed from electrons thatsatisfy ‘medium’ identificationcriteria. Thesecrite- riaare described inRef. [60]. Muon candidatesare reconstructed within the range |η_|<2.5 by combining tracks with compati- ble momentum in the ID and the MS [61]. Only leptons with pT>25 GeV areconsidered.

Backgroundsduetomisidentifiedleptonsandnon-promptlep- tonsaresuppressedbyrequiringleptonstobeisolatedfromother activity in the event and also to be consistent with originating fromtheprimaryvertexoftheevent.² Upperboundsoncalorime- terandtrack isolation discriminants are used to ensure that the leptonsareisolated.

Detailsoftheleptonisolationcriteriaaregiveninthepublica- tionsfortheν^^[19],q^q¯ ^[21],andνq^q¯ ^{[22]channels.}

Jetsareformedbycombiningtopologicalclustersreconstructed inthecalorimetersystem[62],whicharecalibratedinenergywith the local calibration weighting scheme [63] and are considered massless. The measured energies are corrected forlosses inpas- sivematerial, thenon-compensatingresponse ofthecalorimeters andpile-up[64].

Hadronicallydecayingvectorbosons withlow pT (^{450 GeV)} are reconstructed using a pair of jets. The jetsare formed with theanti-kt algorithm[65]witharadiusparameter R=⁰.4.These

2 Theprimaryvertexoftheeventisdefinedasthereconstructedprimaryvertex withhighest

p²_Twherethesumisoverthetracksassociatedwiththisvertex.

jets are hereafter referred to as small-R jets. Only small-R jets with|η|<2.8(2.1)and pT>30 GeV areconsideredfortheνq^q¯ (qq)¯ channel. Forsmall-R jets with pT<50 GeV it is required that thesummedscalar pT ofthe tracksmatchedto theprimary vertex accountsfor at least 50% of the scalar summed p_T of all tracks matchedto the jet. Jetscontaining hadronsfrom b-quarks areidentifiedusingamultivariateb-taggingalgorithmasdescribed inRef.[66].

Hadronicallydecayingvectorbosonswithhigh p_T (^{400 GeV)} canbereconstructedasasinglejetwithalargeradiusparameter, orlarge-R jet, dueto the collimated nature oftheir decay prod- ucts. These large-R jets,hereafter denotedby J , are firstformed withtheCambridge–Aachen(C/A)algorithm[67,68]witharadius parameter R=¹.2.Afterthejet formationa setofcriteriaisap- pliedtoidentifythejetasoriginatingfromahadronicallydecaying boson(bosontagging).A groomingalgorithmisappliedtothejets to reduce the effectofpile-up andunderlyingevent activityand to identifya pairofsubjets associatedwiththe quarksemerging from the vector boson decay. The grooming algorithm, a variant ofthemass-dropfilteringtechnique [69],isdescribed indetailin Ref.[23].The groomingprocedureprovidesasmalldegreeofdis- criminatingpowerbetweenjetsfromhadronicallydecayingbosons andthoseoriginatingfrombackgroundprocesses.

Jet discrimination is further improved by imposing additional requirementsonthelarge-R jetproperties.First,inallofthechan- nels usinglarge-R jets, a requirement on thesubjetmomentum- balance found at the stopping point of the grooming algorithm,

√y>0.45,³ is applied to the jet. Second, jets are required to havethegroomedjetmasswithinaselectionwindow.Duetothe differentbackgroundsaffectingeachofthesearchchannels,differ- entmasswindowsareusedforeachchannel.Inthesinglelepton anddileptonchannels, masswindowsof65<mJ<105 GeV and 70<mJ<110 GeV,where mJ representsthe jetmass, are used forselectingW and Z bosons.Inthefullyhadronicchannel,mass windows of 69.4<m_J <95.4 GeV and 79.8<m_J<105.8 GeV, which are ±13 GeV around the expected W or Z reconstructed mass peak, are used for selecting W or Z boson candidates re- spectively.

The high-p_T jetsinbackgroundeventsare expectedtohavea larger charged-particle track multiplicity than the jets emerging from boson decays. This is due to the higher energy scale in- volved in the fragmentation process of background jetsand also duetothelargercolour chargeofgluonsincomparisontoquarks.

Hence,toimprovethesensitivityofthesearchinthefullyhadronic channel, a requirement on thecharged-particle trackmultiplicity matched to the large-R jet prior to the grooming, ntrk<30, is used to discriminatebetween jetsoriginatingfrom boson decays

3 √_y≡min(pT j₁,pT j₂) ^R_m^(j1,j2)

0 ,wherem0isthemassofthegroomedjetatthe stoppingpointofthesplittingstageofthegroomingalgorithm,pT j₁ andpT j₂ are thetransversemomentaofthesubjetsatthestoppingpointofthesplittingstageof thegroomingalgorithmand R(j₁,j₂)isthedistancein(φ,η)spacebetweenthese subjets.