Search for ${W}' \rightarrow tb$ decays in the hadronic final state using $\mathit{pp}$ collisions at $\sqrt{s}=13$ TeV with the ATLAS detector

Contents lists available atScienceDirect

Physics Letters B

www.elsevier.com/locate/physletb

Search for W

→ ^{tb decays in the hadronic final state using pp} collisions at √

s = 13 TeV with the ATLAS detector

.TheATLASCollaboration^

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

Articlehistory:

Received25January2018

Receivedinrevisedform13March2018 Accepted13March2018

Availableonline4April2018 Editor:W.-D.Schlatter

AsearchforW^-bosonproductionintheW^→^tb¯→^qq¯^bb decay¯ channelispresentedusing36.1 fb⁻¹ of13 TeVproton–protoncollisiondatacollectedbytheATLASdetectorattheLargeHadronColliderin 2015and2016.Thesearchisinterpretedintermsofbothaleft-handedandaright-handedchiralW^ bosonwithinthemassrange1–5TeV.Identificationofthehadronicallydecayingtopquarkisperformed using jetsubstructuretaggingtechniques basedonashower deconstruction algorithm. Nosignificant deviationfromtheStandardModelpredictionisobservedandtheresultsareexpressedasupperlimits onthe W^→^tb production¯ cross-sectiontimesbranching ratioas afunctionof the W^-boson mass.

TheselimitsexcludeW^bosonswithright-handedcouplingswithmassesbelow3.0 TeVandW^bosons withleft-handedcouplingswithmassesbelow2.9 TeV,atthe95%confidencelevel.

©^{2018TheAuthor.PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense} (http://creativecommons.org/licenses/by/4.0/).FundedbySCOAP³.

1. Introduction

Severaltheories beyondthe Standard Model (SM) involve en- hancedsymmetriesthatpredictnewgauge bosons,usuallycalled W^or Z^bosons.TheW^bosonisthemediatorofanewcharged vector current that can be massive enough to decay into a top quark anda b-quark (as in Fig. 1). Many models such as those withextradimensions[1],strongdynamics[2–5],compositeHiggs [6], or the Little Higgs [7,8] predict new vector charged-current interactions,some withpreferential couplings toquarks orthird- generation particles [6,9–12]. Due to the large mass of the top quark,itsinteractions decouplefromtherestofthephenomenol- ogyin many theories beyond the SM. An effective Lagrangian is used to capture the relevant phenomenology of the Sequential StandardModel(SSM)[13] W^→^tb signal¯ [14,15],whichhasthe samecouplingstrengthtofermionsastheSMW bosonbuthigher mass.

Searchesfora W^ bosondecayingintot_b,¯¹ _{classifiedaseither} leptonic or hadronic according to the decay products of the W bosonoriginatingfromthetopquark,wereperformedattheTeva- tron[16,17] andtheLargeHadronCollider(LHC)infinalstatesthat includeleptons[18–21] orthatarefullyhadronic[22].Thespecific searchfora W^ bosondecayingintot_{b allows}¯ _foraright-handed W^ boson(W^_R) in models in which the right-handedneutrino’s mass isassumed to be much higher than that of the W^ boson

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

1 Forsimplicity, the notation “tb”¯ is used todenote the finalstate for both W^+→tb and¯ W^−→ ¯tb decays.

Fig. 1. FeynmandiagramforW^-bosonproductionwithdecayintotb and¯ ahadron- icallydecayingtopquark.

(mνR >mW^), which the leptonic decay mode cannot access. In such a model,the branchingratio fora W^_R bosondecaying into t_{b is}¯ O(10%)higherrelativetothatforaW^_L bosondecayinginto tb since¯ aW^_L bosoncanalsodecaytoaleptonandneutrino.Lim- its on a SSMleft-handed W^ boson (W^_L) decaying into alepton andaneutrinohavebeensetpreviously[23,24].Previoussearches intheall-hadronicfinal stateexclude W_R^ bosonswithmassesup to2 TeV,setatthe95% confidencelevel(CL)using20.3fb⁻¹ ofpp collisiondataatacentre-of-massenergy(√

s)of8 TeV [22].Are- centsearchbytheCMSCollaborationinthelepton+^{jetsfinalstate} excludes W^_R-bosonmasses up to 3.6TeV using 35.9fb⁻¹ of pp collisiondatacollectedat√

s=13 TeV [18].

Thisanalysissearchesfora W^ bosondecaying intot_{b with}¯ _a massin therangeof 1–5TeV,inthe invariant massspectrum of the top quark and bottom quark (m_tb) reconstructed in the fully hadronicchannel.Thisincludesa W^_R bosonthat isnot kinemati- callyallowedtodecayintoaleptonandneutrinoanda W_L^ boson that candecayinto quarksorleptons. Thelarge W^ massresults https://doi.org/10.1016/j.physletb.2018.03.036

0370-2693/©^{2018TheAuthor.PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense}(http://creativecommons.org/licenses/by/4.0/).Fundedby SCOAP³.

ina top quark and ab-quark that have hightransverse momen- tum (pT).² The decay products of the top quark become more collimated asthe top-quark pT increases, andtheir showerspar- tiallyoverlap[25].Thishigh-pTtopology isreferred toasboosted.

Theboostedtop-quarkdecayis reconstructedasa single jet.The shower deconstruction(SD) algorithm[26,27] is employed to se- lect, or tag, jets from boosted top-quark decays. A signal would be reconstructedasalocalised excessin them_tb distribution ris- ingabovethesmoothlyfallingbackgroundoriginatingmostlyfrom jetscreatedbythestronginteractiondescribedbyquantumchro- modynamics (QCD). This analysis represents an improvement on theprevious ATLAS analysisinthis channel [22] dueto a higher centre-of-mass energy, higher integrated luminosity, and better top-taggingtechniques,understanding ofsystematicuncertainties, andstatisticaltreatment.

2. ATLASdetector

The ATLAS detector [28] at the LHC covers almost the entire solid angle around the collision point. Charged particles in the pseudorapidity range |η_|<2.5 are reconstructed with the inner detector (ID), which consists of several layers of semiconductor detectors (pixel and strip) and a straw-tube transition-radiation tracker,the latter covering |η_|<2.0. The high-granularity silicon pixel detector provides four measurements per track; the clos- est layer to the interaction point is known as the insertable B- layer (IBL) [29]. The IBL was added in 2014 and provides high- resolution hits at small radius to improve the tracking perfor- mance. The ID isimmersed in a 2 T magneticfield provided by a superconducting solenoid. The solenoid is surrounded by elec- tromagneticandhadroniccalorimeters,anda muonspectrometer incorporatingthreelarge superconductingtoroid magnetsystems.

Thecalorimetersystemcoversthepseudorapidityrange|η_|<4.9.

Electromagneticcalorimetry isperformedwithbarrelandendcap high-granularitylead/liquid-argon(LAr)electromagneticcalorime- ters,within the region |η|<3.2. There is an additional thinLAr presamplercovering|η|<1.8,tocorrectforenergylossinmate- rialupstreamofthecalorimeters.For|η|<2.5,theLArcalorime- ters are divided into three layers in depth. Hadronic calorimetry is performed with a steel/scintillator-tile calorimeter, segmented intothree barrelstructureswithin |η_|<1.7,andtwo copper/LAr hadronicendcapcalorimeters,which covertheregion1.5<|η_|<

3.2. The forward solid angle up to |η_|=4.9 is covered by cop- per/LAr and tungsten/LAr calorimeter modules, which are opti- misedforenergymeasurementsofelectrons/photonsandhadrons, respectively.Themuonspectrometer(MS)comprisesseparatetrig- gerandhigh-precisiontrackingchambersthatmeasurethedeflec- tion ofmuons in a magnetic field generated by superconducting air-coretoroids.TheATLASdetectorusesatieredtriggersystemto selectinteresting events.Thefirstlevelisimplementedincustom electronicsandreducesthe eventratefromtheLHCcrossing fre- quencyof40MHztoadesignvalueof100kHz.Thesecondlevelis implementedinsoftwarerunningona general-purposeprocessor farmwhichprocessestheeventsandreducestherateofrecorded eventsto∼^1kHz[30].

2 ATLASusesaright-handedcoordinatesystemwithitsoriginatthenominalin- teractionpoint(IP)inthecentreofthedetectorandthez-axisalongthebeampipe.

Thex-axispointsfromtheIPtothecentreoftheLHCring,andthey-axispoints upwards.Cylindricalcoordinates(r,φ)areusedinthe transverseplane,φ being theazimuthalanglearoundthez-axis.Thepseudorapidityisdefinedintermsof thepolarangleθasη= −ln tan(θ/2).Angularseparationismeasuredinunitsof

R≡

(η)²+ (φ)^2,whereηandφaretheseparationsinηand φ.Mo- mentuminthetransverseplaneisdenotedbypT.

3. Dataandsimulationsamples

This analysis uses data from proton–proton (pp) collisions at

√s=13 TeV collectedwiththeATLAS detectorin2015and2016 that satisfy a numberof criteriatoensure that theATLAS detec- tor wasin goodoperatingcondition. Theamountofdata usedin thisanalysiscorrespondstoanintegratedluminosityof36.1 fb⁻¹. The average number of pp interactionsdelivered per LHC bunch crossingwas23.7.

MonteCarlo(MC)eventgeneratorswere usedtosimulatesig- nalandbackgroundevents.Signal eventsweregeneratedatlead- ingorder(LO)inQCDby MadGraph5_aMC@NLO v2.2.3[31],using a chiral W^-boson model in which the coupling strength of the W^bosonto theright- andleft-handedfermionsarethesameas thoseoftheSM W bosontoleft-handedfermions.The W_L^ boson can decay into all left-handed fermions, but the W^_R boson can decayonlyintoright-handedquarksastheright-handedneutrino is assumed to be more massive than the W^_R boson. MadGraph was used to simulatethetop-quark andW -boson decays,taking spincorrelationsintoaccount. Pythia v8.186[32] wasusedforthe modelling of the parton shower, fragmentation andthe underly- ing event.The NNPDF23LO partondistributionsfunction(PDF)set [33] and theA14setoftuned parameters[34] wereusedforthe eventgeneration.Allsimulatedsampleswererescaledtonext-to- leading-order (NLO) calculations using NLO/LO K -factors ranging from1.3to1.4,dependingonthemassandhandednessoftheW^ boson,calculatedwith Ztop [15].Thewidthofthe MadGraph sim- ulated W^ bosonissettotheNLO Ztop width calculation,O^(3%) of its mass. Signal samples with gauge-boson massesbetween 1 and3 TeVwere generated in250 GeVsteps, andbetween3and 5 TeVin500 GeVsteps.

The dominantSM backgroundprocess ismulti-jet production.

Inordertoreducethedependenceonthemodellingofthesimu- lation adata-driven methodis implementedasdescribed inSec- tion 5. Correctionsin thismethod areestimated usingQCD dijet simulationproduced atLOby Pythia v8.186.Uncertaintiesinthis methodareobtainedusingsimulatedQCDdijeteventsproducedat LOby Herwig++ v2.7.1[35] and Sherpa v2.1.1[36],andatNLO by Powheg-Box v2[37,38] witheither Pythia8or Herwig+Jimmy [39]

for the parton shower, fragmentation and the underlying event simulation (referred to as Powheg+Pythia and Powheg+Herwig, respectively). Vector bosons (W/Z ) produced inassociation with jetsareincludedinthedata-drivenapproach.Theseprocessesare expected to contribute lessthan 1% of themulti-jet background.

ThisW/Z+^{jetspredictionischeckedusingeventssimulatedwith} the Sherpa v2.2.1[36] generatorandtheCT10PDFset[40].

Top-quarkpairproductionisanimportantbackgroundwithan inclusive cross-section of σ_t¯t=832⁺₋⁴⁶₅₁ pb for a top-quark mass of 172.5 GeV as obtained from calculations accurate to next- to-next-to-leading order and next-to-next-to-leading logarithms (NNLO+^NNLL)inQCDwith Top++2.0[41–47].Simulatedtop-quark pairprocesseswereproducedusingtheNLO Powheg-Box v2gen- erator withthe CT10PDF. The partonshower, fragmentationand theunderlyingeventwereaddedusing Pythia v6.42[48] withthe Perugia2012setoftunedparameters[49].Toincreasethenumber ofsimulatedeventsathighmass,sampleswere producedbinned int¯t mass.InterferenceandbackgroundcontributionsfromtheSM s-channelsingle-topprocessarefoundtobenegligibleandarenot consideredfurtherinthisanalysis.

Thegenerationofthesimulatedeventsamplesincludestheef- fectofmultiple pp interactionsperbunchcrossing,aswell asthe effect on the detector response due to interactions from bunch crossingsbefore orafter theone containing the hard interaction.

Forall MadGraph, Powheg, Pythia and Herwig samples,theEvt- Gen v1.2.0 program [50] was used for the bottom and charm

Geant4-basedATLAS detectorsimulation[51,52] andwererecon- structedwiththesamealgorithmsasthedataevents.

4. Eventreconstructionandshowerdeconstruction 4.1.Eventreconstruction

Thisanalysisrelies onthe reconstruction andidentificationof jetsinitiated by the top- and bottom-quark daughters ofthe W^ boson.Jetsare builtfromtopologicallyrelatedenergydepositions inthecalorimeterswiththeanti-k_t algorithm[53] usingtheFastJet package [54].Two radiusparameters are used forjet reconstruc- tion: asmall radius (small-R) of0.4and alarge radius (large-R) of 1.0. The momenta of both the small-R and large-R jets are corrected for energylosses in passive material and for the non- compensating response of the calorimeter [55]. Small-R jets are alsocorrectedfortheaverageadditionalenergyduetopile-upin- teractions[56].Energydepositionsfrompile-upareremovedfrom large-R jetsusing the trimming algorithm [57]: the constituents ofthe large-R jet are reclusteredusing the kt jet algorithm [58, 59] with R=⁰.2.Constituentjetscontributinglessthan5% ofthe large-R jet’s pT are removed. The remaining energy depositions areusedtocalculatethetrimmed-jetkinematicsandsubstructure properties.In order to improve on the angular resolution of the calorimeter,themassofalarge-R jetiscomputedusingacombi- nationofcalorimeterandtrackinginformation[60].

Small-R jetsareusedtoidentifythejetscompatiblewithorig- inatingfromab-quark createdeitherdirectlyfromtheW^ boson orfromthetop-quarkdecay.Onlysmall-R jetswith pT>25 GeV and|η_|<2.5 (in orderto be within the coverage of theID) are consideredinthisanalysis.Additional p_Trequirementsareapplied toenhancethesensitivityofthesearch (seeSection5).Toreduce thenumber ofsmall-R jets originatingfrompile-up interactions, alikelihood discriminant, basedon trackandvertex information, isusedtodeterminewhethertheprimary vertex³ istheoriginof thecharged-particletracksassociatedwithajetcandidateandre- jects jets originatingfrom pile-up interactions [61]. This is done only for small-R jets with p_T<60 GeV and |η|<2.4. Small-R jetswhichoriginatefromb-quarksareidentifiedusingamultivari- ateb-taggingalgorithm[62,63].Severalobservables,suchasthose basedonthelonglifetimeofb-hadronsandtheb- toc-hadronde- caytopology,areusedasalgorithminputstodiscriminatebetween b-jets,c-jetsandotherjets.Theb-taggingrequirementcorrespond- ing to an efficiency of 77% to identify b-jets with pT>20 GeV, as determined from a sample of simulated tt events,¯ is found to be optimal for the statistical significance of this search. This 77% working point (WP) provides rejection factors against light- flavour/gluon jets and c-jets of 134 and 6 respectively [63,64].

Jetsidentifiedthiswayarereferred toasb-taggedjets. Sincethe b-taggingfactorsaremeasured ina differentpT region,an uncer- taintyisassignedtotheextrapolationofthemeasurementtothe highp_Tregionofinterest.

Eventswithreconstructedelectrons[65] ormuons[66] areve- toed in order to ensure statistical independence of this analysis fromanalyses usingthe leptonicdecayofthe W bosonfromthe top quark [19]. Electrons and muons with transverse momenta above25 GeV andselected withcriteriasimilarto those usedin Ref. [67] areconsideredforthisveto.

3 Collisionverticesareformedfromtrackswith pT>400 MeV.Ifaneventcon- tainsmorethanonevertexcandidate,theonewiththehighest

p²_Tofitsassoci- atedtracksisselectedastheprimaryvertex.

The SD algorithm can be used to identify the jets compati- ble withthe hadronicdecayof a W/Z boson, Higgsboson, ora top quark aswell asto discriminate between quark- and gluon- initiated jets. In this analysis, an SD-algorithm-based tagger (SD tagger) is used to identify jets originating from the top quark.

TheSDtaggercalculateslikelihoodsthat agivenlarge-R jetorigi- natesfromahadronictop-quarkdecayorfromahigh-momentum light quarkorgluon. Theconstituentsof thetrimmedlarge-R jet areusedtobuild exclusivesubjets[54],andthefour-momentaof these subjets serve as inputsto the SD algorithm. Thesesubjets are usedassubstitutesforindividual quarksandgluonsoriginat- ingfromthehardscatter.Alikelihoodweightiscalculatedforeach possibleshowerhistorythat canleadtotheobservedsubjetcon- figuration.Thisstep isanalogoustorunninga partonshowerMC generator in reverse, where emission and decay probabilities at each vertex, colour connections, andkinematic requirements are considered. For each shower history, the assignedweight is pro- portionaltotheprobabilitythattheassumedinitialparticlegener- atesthefinalconfiguration,takingintoaccounttheSMamplitude fortheunderlyinghard processandtheSudakov formfactorsfor the parton shower. A variable called χSD is defined as the ratio of the sum of the signal-hypothesis weights to the sum of the background-hypothesisweights.Foraset{^pk_i}^{ofN observedsub-} jetfour-momenta,wherei∈ [¹,N]^,thevalueof χSD isgivenby:

χSD({^pk_i}) =



permP({p^k_i}|top-quark jet)



permP({^pk_i}|gluon/light-quark jet),

where P({^pk_i}|top-quark jet) isbuiltusingtheweightsforthehy- pothesis that a signal process leads to the observed subjet con- figuration{^pk_i}^{and P}({^pk_i}|gluon/light-quark jet)isbuiltusingthe weightsforthehypothesis thatabackgroundprocessleads tothe observedsubjetconfiguration.The 

perm notation representsthe sumover all the shower histories inwhich signal processeslead to the subjet configuration. The large-R jet is tagged as a top- quark jetif χSD islargerthan agivenvalue,whichisadjusted to achievethedesiredtaggingefficiency.Thereisaninternalmecha- nismintheSDalgorithmtosuppresspile-upcontributionstothe jets,throughtheapplicationofadditionalweightsinthelikelihood ratio,whichcontaintheprobabilitythatasubsetofthesubjetsdid notoriginatefromthehardinteractionbutfrompile-up[68].

TheSDalgorithmselectseventsthatarekinematicallycompat- ible witha hadronictop-quarkdecay.The followingrequirements are made to optimise the algorithm to achieve a balance be- tween goodtop-quarkjet signalselection efficiencyandrejection of gluon/light-quarkjet backgrounds: the large-R jet hasat least threesubjets; twoormoresubjets musthaveacombinedinvari- ant mass ina 60.3–100.3 GeV windowcentred on the W -boson mass;andatleastonemoresubjetcanbeaddedtoobtainatotal massina132–212GeVwindowcentredonthetop-quarkmass.

TheSDtaggerwasoptimisedforthisanalysissothatitismore efficientfortop-quarkjetsignalselectionandgluon/light-quarkjet backgroundrejectionforp_T>800 GeVcomparedtotheversionof the SD taggerfirst studied by the ATLAS Collaboration [25]. This is done bybuilding subjetsobtainedby using an exclusivekt al- gorithm [54].First, thekt algorithmwith R=¹.0 is runover the large-R jet constituents and the k_t reclusteringis stopped if the splitting scale [69] is larger than 15 GeV. Oncethe kt recluster- ing is stopped the reclusteredprotojets are used assubjets. The choiceofa15 GeV requirementisbasedontheexpecteddiscrim- inationbetweensignalandbackgroundevents.Thesixhighestp_T