Measurement of the top quark mass in the $t\bar{t}\rightarrow$ dilepton channel from $\sqrt{s}=8$ TeV ATLAS data

(1)

Contents lists available atScienceDirect

Physics Letters B

www.elsevier.com/locate/physletb

Measurement of the top quark mass in the t t ¯ → dilepton channel from

√ s = 8 TeV ATLAS data

.TheATLAS Collaboration

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

Articlehistory:

Received8June2016

Receivedinrevisedform21July2016 Accepted8August2016

Availableonline24August2016 Editor:W.-D.Schlatter

Thetopquarkmassismeasuredinthet¯t→dilepton channel(lepton=^e,μ)usingATLASdatarecorded intheyear2012attheLHC.Thedataweretakenataproton–protoncentre-of-massenergyof√_s

=^{8 TeV} and correspond toan integratedluminosity of about20.2 fb⁻¹.Exploitingthe template method,and usingthedistributionofinvariantmassesoflepton–b-jet pairs,thetop quarkmassismeasuredtobe mtop=¹⁷².99±⁰.41 (stat)±⁰.74 (syst) GeV,withatotaluncertaintyof0.84 GeV.Finally,a combination withpreviousATLASmtopmeasurementsfrom√

s=^{7 TeV data}ⁱⁿ^the^t¯t→dilepton andtt¯→^lepton+ jets channelsresultsinmtop=¹⁷².84±⁰.34 (stat)±⁰.61 (syst) GeV,withatotaluncertaintyof0.70 GeV.

©²⁰¹⁶^TheÂuthor.^Published^byÊlsevier^B.V.^Thisîsânôpenâccessârticleûnder^the^CC^BY^license (http://creativecommons.org/licenses/by/4.0/).FundedbySCOAP³.

1. Introduction

Themassofthetopquark (mtop) isanimportantparameterof theStandardModel (SM)ofparticlephysics.Precisemeasurements ofm_top provide crucial informationfor globalfits ofelectroweak parameters[1–3]whichhelpassesstheinternalconsistencyofthe SM and to probe its extensions. In addition, the value of mtop affectsthe stability ofthe SM Higgspotential, whichhas cosmo- logicalimplications[4–6].Manymeasurementsofmtop havebeen performedby the Tevatronand LHC Collaborations. Combining a selection of those, the first Tevatron+LHC mtop result is mtop= 173.34±⁰.27 (stat)±⁰.71 (syst) GeV,withatotal uncertaintyof 0.76 GeV [7].Meanwhile, a number ofnewresults havebecome available [8–13],some ofwhich aremore precisethan theabove combination. The latest ATLAS results in the t¯t→^lepton+^jets and tt¯→dilepton decay channels, both with electrons (e) and muons(μ)inthefinalstate[14],aremtop=¹⁷².33±⁰.75 (stat)± 1.02 (syst) GeV andmtop=¹⁷³.79±⁰.54 (stat)±¹.30 (syst) GeV, respectively.

ThisLetterpresentsanewmeasurementofm_topobtainedinthe tt¯→dilepton decay channel using 2012data takenat a proton–

proton (pp) centre-of-mass energy of √

s=^{8 TeV,} ^with ^an ^inte- gratedluminosityofabout20.2 fb⁻¹.Theanalysisexploitsthede- caytt¯→^W⁺^W⁻^b_b¯→ ⁺⁻νν¯b_b,¯ _which_is_realised_when_both _W bosonsdecayintoachargedleptonanditscorrespondingneutrino.

In theanalysis, thett decay¯ channelsee, eμ and μμ (including

τ_→e, μ) are combinedandreferred to asthe dilepton channel.

Single-top-quark eventswiththe samelepton ﬁnal states are in-

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

cluded in the signal. Given the larger data sample compared to Ref.[14],theeventselectionwasoptimisedtoachievethesmallest total uncertainty. The measurement is based on the implemen- tation of the template method described in Ref. [14], which is calibrated using signal Monte Carlo (MC) samples. Consequently, thetopquarkmassmeasuredinthiswaycorrespondstothemass deﬁnitionusedintheMCprogram.

2. ATLASdetector

TheATLASexperiment[15]attheLHCisamulti-purposeparti- cledetectorwithaforward–backwardsymmetriccylindricalgeom- etryandanear4π coverageinsolidangle.¹Itconsistsofaninner tracking detector surroundedby a thinsuperconducting solenoid providing a 2 T axialmagnetic ﬁeld, electromagneticandhadron calorimeters,andamuon spectrometer.The innertrackingdetec- torcoversthepseudorapidityrange|η_|<2.5.Itconsistsofsilicon pixel,siliconmicrostrip,andtransitionradiationtrackingdetectors.

Lead/liquid-argon(LAr)samplingcalorimetersprovideelectromag- netic(EM)energymeasurementswithhighgranularity.A hadronic (steel/scintillator-tile) calorimetercovers thecentral pseudorapid- ityrange(|η_|<1.7).The end-capandforwardregionsareinstru- mented withLArcalorimeters for EMand hadronicenergymea-

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

Thex-axispointsfromtheIPtothecentreoftheLHCring,andthe y-axispoints upwards.Cylindricalcoordinates (r,φ)areusedinthetransverseplane,φ being theazimuthalanglearoundthe z-axis.Thepseudorapidityisdeﬁnedintermsof the polarangleθ asη= −ln tan(θ/2).Angulardistance ismeasuredinunitsof R≡

(η)²+ (φ)². http://dx.doi.org/10.1016/j.physletb.2016.08.042

0370-2693/©²⁰¹⁶^TheÂuthor.^Published^byÊlsevier^B.V.^Thisîsânôpenâccessârticleûnder^the^CC^BY^license(http://creativecommons.org/licenses/by/4.0/).Fundedby SCOAP³.

(2)

The ATLAS Collaboration / Physics Letters B 761 (2016) 350–371 351

surementsupto|η|=⁴.9.Themuon spectrometersurroundsthe calorimetersandisbasedonthreelargeair–coretoroidsupercon- ductingmagnetswitheightcoilseach.Itsbendingpowerisinthe rangefrom2.0to7.5 T m.It includesasystemofprecisiontrack- ingchambersandfastdetectorsfortriggering.A three-leveltrigger system is used to select events. The ﬁrst-level trigger is imple- mentedinhardwareandusesasubsetofthedetectorinformation toreduce theaccepted eventratetoatmost75 kHz.This isfol- lowedby two software-based trigger levels that together reduce the acceptedrate to 400 Hz on average depending on the data- takingconditionsduring2012.

3. DataandMCsamples

Thisanalysisisbasedon pp collision datarecordedin2012at

√s=^{8 TeV.}^The^integrated^data^luminosity^amounts^to²⁰.2 fb⁻¹ withan uncertaintyof 1.9% determined withthe procedures de- scribedinRef.[16].

Themodelling oftt and¯ single-top-quark signal eventsandof mostbackgroundprocessesreliesonMCsimulations.Forthesim- ulationofsignaleventsthe Powheg-Box program[17–19]isused.

The simulation of the top quark pair [20] and single-top-quark productioninthe W t-channel[21] usesmatrixelementsatnext- to-leading order (NLO) in the strong coupling constant αS, with theNLO CT10[22]partondistributionfunction (PDF)andthepa- rameter h_damp= ∞^. ^The ^hdamp parameter sets the resummation scale,whichcontrolsthetransitionfromthematrixelementtothe partonshower (PS) simulation.Giventhat theeventselection de- scribedbelowrequiresleptonic decayproductsoftwo W bosons, single-top-quarkeventsinthe s-channelandt-channelare found nottocontributetothesample.

The Pythia (v6.425)program[23] withtheP2011C[24]setof tunedparameters(tune)andthecorrespondingCTEQ6L1PDF [25]

are employed to provide the parton shower, hadronisation and underlying-eventmodelling. Theuncertainties duetoQCD initial- and ﬁnal-state radiation (ISR/FSR) modelling are estimated with samples generated with the Powheg-Box program interfaced to the Pythia program for which the parameters of the generation arevariedtospantherangescompatiblewiththeresultsofmea- surementsoft¯t productioninassociationwithjets[26–28].

Formtop hypothesis testing, thet¯t and single-top-quark event samplesaregeneratedforﬁvevaluesofm_topintherange167.5 to 177.5 GeV instepsof2.5 GeV.Foreachm_top value,theMC samples are normalised according to the best available cross-section calculations, which for mtop=¹⁷².5 GeV are σ_t_t_¯ =²⁵³⁺₋¹³₁₅^pb [29–34]fort¯t productionand σW t=²².4±¹.5 pb[35]forsingle- top-quarkproduction inthe W t-channel.The PDF + αS-induced uncertainties in these cross-sections are calculated using the PDF4LHCprescription[36]withtheMSTW200868% CL NNLOPDF [37,38],CT10 NNLO PDF [22,39] andNNPDF2.3 5f FFN PDF [40], and are added in quadrature with the uncertainties due to the choicesofthefactorisationandrenormalisationscales.

The simulation of W^± or Z boson production in association withjetsisperformedwiththe Alpgen (v2.13)program[41]inter- facedtothe Pythia6 programusingtheCTEQ6L1PDFandthecor- respondingAUET2tune[42].Dibosonproductionprocesses(W W , W Z and Z Z ) aresimulatedusingthe Alpgen programinterfaced tothe Herwig (v6.520)program[43] withtheAUET2tuneandto the Jimmy (v4.31)program[44].Allsamplesare simulatedtaking intoaccount theeffects ofmultiplesoft pp interactions(pile-up) registered in the 2012 data. These interactions are modelled by overlaying simulated hits from events with exactly one inelastic (signal)collisionperbunchcrossingwithhitsfromminimum-bias eventsthat are produced with the Pythia (v8.160) program [45]

usingtheA2Mtune[46]andtheMSTW2008 LO PDF.Forthisanal- ysis,theobservedvaluesofthepile-up-relatedquantitiesμ,the meannumberofinteractionsperbunchcrossing,andnvtx,theav- eragenumberofverticesperevent,areμ=²⁰.7 andn_vtx=⁹.2.

Finally,thesamples undergoasimulation oftheATLAS detector[47]basedon Geant4[48],andarethenprocessedthroughthe same reconstruction software asthe data. A number of samples usedtoassesssystematicuncertainties areproducedwithafaster versionofthesimulationwhich,inadditiontothefullsimulation ofthe tracking,uses smearing functionsandinterpolates particle behaviourandcalorimeterresponse,basedonresolutionfunctions measuredinfull-simulation studies,toapproximatetheresultsof thefullsimulation.

4. Dataselectionandeventreconstruction

Triggers basedonisolated single electrons ormuonswithen- ergy or momentumthresholds of 24 GeV areused. The detector objects resulting fromthe top quark pair decayare electron and muon candidates,jetsandmissingtransverse momentum(E^miss_T ).

Inthefollowing,thetermleptonisusedforchargedleptons (ex- cluding τ leptons)exclusively.

Electron candidates [49] are required to have a transverse energy of ET>25 GeV, a pseudorapidity of the corresponding EM cluster of |ηcluster|<2.47, withthe transition region 1.37<

|ηcluster|<1.52 between the barrel and the end-cap calorimeter excluded. The muon candidates [50] are required to have transverse momentum pT>25 GeV and|η_|<2.5. Toreduce thecon- tamination by leptons from heavy-ﬂavour decays inside jets or fromphotonconversions,referred toasnon-prompt (NP)leptons, strictisolationcriteriaareappliedtotheamountofactivityinthe vicinityoftheleptoncandidate[49,50].

Jetsarebuiltfromtopologicalclustersofcalorimetercells[51]

with the anti-kt jet clustering algorithm [52] using a radius parameter of R=⁰.4. Jets are reconstructed using the local clus- terweighting (LCW) andglobalsequentialcalibration (GSC) algo- rithms[53–55]andrequiredtosatisfy pT>25 GeV and|η_|<2.5.

MuonsreconstructedwithinaR=⁰.4 conearoundtheaxisofa jetwith p_T>25 GeV arenot consideredascharged-leptoncandi- dates.Inaddition,jetswithinaR=⁰.2 conearoundanelectron candidate are removed and finally electrons within a R=⁰.4 cone around anyofthe remaining jetsare discarded.The identi- fication ofjets containing b-hadrons, b-tagging,isused forevent reconstruction and background suppression. In the following, ir- respective of their origin, jets tagged by the b-tagging algorithm are referred to as b-tagged jets, whereas those not tagged are referred to as untagged jets. Similarly, whether they are tagged or not, jets originating from bottom quarks are referred to as b-jetsandthosefrom(u,d,c,s)-quarksorgluonsaslightjets.The working point of the neural-network-based MV1 b-tagging algo- rithm [56] corresponds to an average b-tagging efficiency of70%

forb-jetsinsimulatedt¯t eventsandrejectionfactorsof5forjets containing a c-hadron and 137 for jets containing only lighter- ﬂavour hadrons.Tomatchtheb-tagging performance inthedata, p_T- and η-dependent scale factors [56], obtained fromdijet and tt¯→dilepton events, are applied to MC jets depending on their trueﬂavour.The reconstruction ofthe E^miss_T isbased onthevec- torsumofenergydepositsinthecalorimeters,projectedontothe transverseplane. Muonsare includedinthe E^miss_T usingtheir re- constructedmomentuminthetrackingdetectors[57].

The contribution of events wrongly reconstructed as tt¯→ dilepton events due to the presence of objects misidentiﬁed as leptons (fakeleptons),isestimatedfromdata[58].Thetechnique employedusesfake-leptonandreal-leptoneﬃcienciesthatdepend on η and p_T,measured inabackground-enhanced controlregion

(3)

Table 1

Theobservednumbersofeventsindataafterthepre-selectionandthefinalselec- tion.Inaddition,theexpectednumbersofsignaleventsformtop=172.5 GeV and backgroundeventscorrespondingtotheintegrateddataluminosityaregiven.Two significantdigitsareusedfortheuncertaintiesofthepredictednumbersofevents explainedinthetext.Thelowerrowsreportthematchingperformanceevaluated formtop=¹⁷².5 GeV,usingonesignificantdigitforthestatisticaluncertainties.

Selection Pre-selection Final selection

Data 36 359 9426

tt signal¯ 34 300±²⁷⁰⁰ ⁹⁶⁷⁰±⁷⁷⁰

Single-top-quark signal 1690±¹¹⁰ ³⁶³±²³

Fake leptons 240±240 31±31

Z+jets 212±83 20.6±8.5

W W/W Z/Z Z 57±21 10.2±3.8

Signal+^background ^{36 600}±²⁸⁰⁰ ^{10 100}±⁷⁷⁰

Expected background fraction 0.01±⁰.01 0.01±⁰.00 Data/(Signal+background) 0.99±⁰.07 0.93±⁰.07 Matching eﬃciency [%] 78.4±0.2 95.3±0.4 Selection purity [%] 51.6±0.1 69.8±0.3 Unmatched events [%] 34.2±0.1 26.7±0.1 Wrongly matched events [%] 14.2±⁰.1 3.4±⁰.0

withlow E^miss_T andfromeventswithdileptonmassesaroundthe Z peak[59].

TheselectionfromRef.[14]isappliedasapre-selectionasfol- lows:

1. Events arerequired tohave asignal fromthe single-electron orsingle-muontriggerandatleastoneprimaryvertexwithat leastﬁveassociatedtracks.

2. Exactly two oppositely chargedleptons are required,with at leastoneofthemmatchingthereconstructedobjectthatﬁred thecorrespondingtrigger.

3. In the same-lepton-ﬂavour channels, ee and μμ, E^miss_T >

60 GeV is required. In addition, the invariant mass of the leptonpairmustsatisfym>15 GeV,andmustnotbecom- patiblewiththe Z masswithin10 GeV.

4. In the eμ channel the scalar sumof p_T ofthe two selected leptonsandalljetsisrequiredtobelargerthan130 GeV.

5. The presence of at least two jets with pT >25 GeV and

|η_|<2.5 isrequired,andatleastone ofthesejetshasto be b-tagged.

The observed numbers of events in the data after this pre- selection,togetherwiththeexpectednumbersofsignal andback- groundeventscorrespondingtotheintegrateddataluminosity,are giveninTable 1.Assumingatopquarkmassofm_top=¹⁷².5 GeV, the predicted number of events is consistent with the one observed in the data within uncertainties. For all predictions, the uncertaintiesareestimatedasthesuminquadratureofthestatis- ticaluncertainty, a 1.9% uncertainty inthe integratedluminosity, and a number of additional components. For the signal, these are a 5.4% uncertainty in the t¯t cross-section, or a 6.0% uncertainty in the single-top-quark cross-section, as given in Sect. 3.

Finally, global 4.1%, 2.2% and 2.8% uncertainties are added, corresponding to the envelopes of the results from the eigenvector variations of the jet energyscale (JES), the relative b-to-light-jet energy scale (bJES) and the b-tagging scale factors, respectively.

The background uncertainties contain jet-multiplicity-dependent uncertainties of about 40% inthe normalisation of the Z + ^jets backgroundanda 100% uncertaintyinthenormalisationoffake- leptonbackground.

ThetwojetscarryingthehighestMV1weightaretakenasthe twob-jetsoriginatingfromthedecaysofthetwotopquarks,and thetwoleptonsaretakenastheleptonsfromtheleptonic W de-

cays.Fromthetwopossibleassignmentsofthetwopairs,thecom- bination leading to thelowest averageinvariant mass ofthe two lepton–b-jet pairs (mb) is retained.To estimate theperformance ofthisalgorithminMCsimulatedsamples,thereconstruction-level objectsarematchedtotheclosest generator-levelobjectbasedon a maximumallowed R, being0.1 for leptons and 0.3 for jets.

A matched objectis definedasa reconstruction-level objectthat falls within R of anygenerator-level object ofthat type, and a correct match means that this generator-level object is the one it originated from. Due to acceptance losses and reconstruction inefficiency, not all reconstruction-level objects can successfully be matchedto theirgenerator-level counterparts,resultinginun- matchedevents.Thematchingefficiencyisthefractionofcorrectly matched events among all the matched events, and the selection purity isthe fractionof correctlymatched eventsamongall events,regardless ofwhetherthey could be matched ornot. The corresponding numbers formtop=¹⁷².5 GeV arereported inTa- ble 1.

Startingfromthispre-selection,anoptimisationofthetotalun- certaintyinm_top isperformed.A phase-spacerestrictionbasedon the average pT of the two lepton–b-jet pairs (p_T_,_b) is used to obtain the smallest total uncertainty in mtop. The corresponding p_T_,_b distribution is shownin Fig. 1(a). The smallest uncertainty in mtop corresponds to p_T_,_b>120 GeV. The difference in shape between data and prediction is covered by the systematic uncertainty as detailed in Sect. 6. This restriction is found to also increase the fraction of correctlymatched eventsin the tt sam-¯ ple, andreduces the number ofunmatched orwrongly matched events.

Toperformthetemplateparameterisationdescribed inSect.5, an additionalselection criterion is applied, restrictingthe recon- structedmbvalue(m^reco_b )totherange30 GeV<m^reco_b <170 GeV.

Applyingbothrestrictions,thenumbersofpredictedandobserved events resulting fromthe ﬁnal selection are reportedin Table 1.

Using this optimisation, the matching eﬃciency and the sample purity are much improved as reported in the bottom rows of Table 1, while retaining about26% of the events. Using thisse- lection, and the objects assigned to the two lepton–b-jet pairs, the kinematicdistributions inthe dataare well described by the predictions, as shown in Fig. 1 for the transverse momenta of b-jets and leptons, and for the Rb of the two lepton–b-jet pairs.

5. Templateﬁtandresultsinthedata

Theimplementationofthetemplatemethodusedinthisanal- ysis isdescribed inRef. [14]. Forthis analysis, the templates are simulateddistributions ofm^reco_b ,constructedforanumberofdis- cretevaluesofm_top.Appropriatefunctionsarefittedtothesetem- plates, interpolatingbetweendifferent input mtop. Theremaining parameters ofthe functionsare fixed by asimultaneous fittoall templates,imposinglineardependencesoftheparametersonm_top. The resulting template fit functionhas mtop asthe only free pa- rameterandanunbinnedlikelihoodmaximisationgivesthevalue ofm_top that bestdescribesthedata.Statisticallyindependentsig- naltemplates,comprisingtt and¯ single-top-quarkevents,arecon- structedasafunctionofthetopquark massusedintheMCgen- erator. Withinthestatisticaluncertainties, the sumofaGaussian distributionandaLandaufunctiongivesagooddescriptionofthe shapeofthem^reco_b distributionasshowninFig. 2(a)forthreeval- uesofmtop.Withthissignalchoice,thebackgrounddistributionis independentofm_top,andaLandaufunctionisfittedtoit.Thesum ofthesignaltemplateatmtop=¹⁷².5 GeV andthebackgroundis compared to data in Fig. 2(b).It gives a good description of the dataexceptfordifferencesthatcanbeaccountedforbyadifferent