Measurement of the cross section of high transverse momentum $Z \rightarrow b\bar{b}$ production in proton--proton collisions at $\sqrt{s}=8$ TeV with the ATLAS Detector

Contents lists available atScienceDirect

Physics Letters B

www.elsevier.com/locate/physletb

Measurement of the cross section of high transverse momentum Z → ^b b production ¯ in proton–proton 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:

Received28April2014

Receivedinrevisedform8September2014 Accepted8September2014

Availableonline16September2014 Editor:H.Weerts

Keywords:

LHC

Boostedbb topologies¯

ThisLetterreportstheobservationofahightransversemomentumZ→^bb signal¯ inproton–protoncol- lisions at√

s=^{8 TeV and the} measurementof its productioncross section. The data analysed were collectedin2012 withthe ATLASdetector attheLHC andcorrespond toan integratedluminosity of 19.5 fb⁻¹.TheZ→^bb decay¯ isreconstructedfromapairofb-taggedjets,clusteredwiththeanti-kt jet algorithmwithR=⁰.4,thathavelowangularseparationandformadijetwithpT>200 GeV.Thesignal yieldisextractedfromafittothedijetinvariantmassdistribution,withthedominant,multi-jetback- groundmassshapeestimatedbyemployingafullydata-driventechniquethatreducesthedependence oftheanalysisonsimulation.Thefiducialcrosssectionisdeterminedtobe

σ_Z^fid_→bb¯=2.02±0.20(stat.) ±0.25(syst.)±0.06(lumi.)pb=2.02±0.33 pb, ingoodagreementwithnext-to-leading-ordertheoreticalpredictions.

PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense (http://creativecommons.org/licenses/by/3.0/).FundedbySCOAP³.

1. Introduction

High transverse momentum (pT), hadronically decaying, elec- troweak-scale bosons have already been used in searches at the LHC [1–5], and are expected to play an increasingly sig- nificant role as the LHC moves to higher centre-of-mass ener- gies in 2015. Therefore it is important to study them directly.

This Letter presents the observation of a high-p_T Z →^bb sig-¯ nal in a fully hadronic final state, and a measurement of its production cross section. The measurement is compared to the next-to-leading-order (NLO) matrix element plus parton-shower predictions of POWHEG [6–9] and aMC@NLO [10], where the parton-shower,hadronisation andunderlying-event modellingare provided by Pythia-8.165 [11] and Herwig++ [12], respectively.

This first measurement of a high-p_T electroweak-scale boson in anall-hadronicfinalstate attheLHCdemonstratesthevalidityof boththeanalysistechniquesused andofthestate-of-the-art NLO plusparton-showerparticle-levelpredictionsforelectroweak-scale bosons decaying to bb.¯ It is therefore of great relevance for the search for the H→^bb signal¯ in the (most sensitive) highHiggs boson pT range [13], aswell as for searches for TeV-scale reso- nances decaying to bbb¯ b via¯ Z Z , Z H or H H [14,15]. A high-p_T Z→^bb signal¯ can also provide a usefulbenchmark for validat-

 E-mailaddress:Atlas.Publications@cern.ch.

ingtheperformanceoftheATLASdetector(forexample,theb-jet energy scale¹); and for testing and optimising analysis methods relevant for physics studies involving high-pT jets that contain b-hadrons(b-jets).

TheanalysisdescribedinthisLetterisdesignedtoselectb_{b de-}¯ caysofZ -bosonswithp_T>200 GeV,inproton–protoncollisionsat

√s=^{8 TeV.Thehigh-p}T requirementhelpstoenhancethesignal relativetobb production¯ inmulti-jetevents(predominantlygluon splitting to bb in¯ this high-pT regime), which is the dominant sourceofbackgroundandhasamoresteeply-falling p_T spectrum.

In orderto minimisethe dependenceon simulation,the analysis employsafullydata-driventechniqueforthedeterminationofthe invariantmassspectrumofthemulti-jetbackground.Thisisespe- ciallyimportantgiventhatMonteCarlo(MC)generators havenot beentestedthoroughlyin thisregionofthebb production¯ phase space.

2. TheATLASdetector

ATLAS is a multi-purpose particle physics experiment [17] at the LHC. The detector layout² consists of inner tracking devices

1 TheuseoftheZ→b_{b peak}¯ _{toconstraintheb-jetenergyscaleattheTevatron} experimentswasdemonstratedinRef.[16].

2 ATLASuses aright-handedcoordinatesystemwith itsoriginat thenominal interactionpoint(IP)inthecentreofthedetectorandthez-axisalongthebeam http://dx.doi.org/10.1016/j.physletb.2014.09.020

0370-2693/PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense(http://creativecommons.org/licenses/by/3.0/).FundedbySCOAP³.

surrounded by a superconducting solenoid, electromagnetic and hadroniccalorimeters,andamuonspectrometer. Theinner track- ingsystemprovidescharged-particle trackinginthepseudorapid- ity region |η|<2.5 and vertex reconstruction. It consists of a silicon pixel detector, a silicon microstrip tracker, and a straw- tube transition radiation tracker. The system is surrounded by a solenoid that produces a 2 T axial magnetic field. The central calorimetersystemconsistsofaliquid-argonelectromagnetic(EM) samplingcalorimeterwithhighgranularityandaniron/scintillator- tile calorimeter providing hadronic energy measurements in the central pseudorapidityrange(|η|<1.7). Theendcap andforward regions are instrumented withliquid-argon calorimeters forboth electromagnetic and hadronic energy measurements up to |η_|=

4.9. The muon spectrometer isoperated ina magnetic field pro- vided by air-core superconducting toroids and includes tracking chambersforprecisemuonmomentummeasurementsupto|η_|=

2.7 and trigger chambers covering the range |η|<2.4. A three- leveltrigger systemis used toselect interesting events[18].The Level-1 trigger reduces the event rate to below 75 kHz using hardware-based triggeralgorithms acting ona subset of detector information.Twosoftware-basedtriggerlevels, referredto collec- tivelyastheHigh-LevelTrigger(HLT),furtherreducetheeventrate toabout400 Hzusinginformationfromtheentiredetector.

3. Data,simulatedsamples,andeventreconstruction

Thedatasample usedinthisanalysis, afterrequiringthatcer- tain quality criteriabe satisfied,corresponds to an integrated lu- minosityof L=¹⁹.5±⁰.5 fb⁻¹, and was recorded by ATLAS in 2012.Theuncertaintyontheintegratedluminosityisderived,fol- lowingthe same methodologyas that detailedin Ref. [19], from acalibration oftheluminosity scaleusing beam-separationscans performedinNovember2012.

MC event samples simulated with the GEANT4-based [20]

ATLAS detectorsimulation[21]areusedtomodeltheZ→^b_{b sig-}¯ nal and the small t¯t, Z→^c¯^{c and W} →^qq¯^{ background contri-} butions. In addition, multi-jet MC event samples are used for studyingthetriggermodellinginsimulation.Theeffectofmultiple proton–protoninteractionsinthesamebunchcrossing(pile-up)is includedinthesimulation.

The Z→^b_{b signal}¯ _{is modelled} using Sherpa-1.4.3 [22], with the CT10[23] NLO partondistribution function (PDF) set. An al- ternative Z→^b_{b model}¯ _wasgeneratedwith Pythia-8.165[11]and the CTEQ6L1 [24] leading-order (LO) PDF set and is used to de- terminethesystematicuncertaintyassociatedwith Z→^bb mod-¯ elling.The Z→^c¯c backgroundisalsogeneratedwith Sherpa-1.4.3 and the CT10 PDF set. The t¯t background is simulated with MC@NLO-4.06[25] interfacedto Herwig-6.520 [26] forthe frag- mentationandhadronisationprocesses,including Jimmy-4.31[27]

fortheunderlying-event description.The topquark mass isfixed at 172.5 GeV, and the PDF set CT10 is used. The W →^qq¯^ and multi-jet MC samples are generated using Pythia-8.165 withthe CT10PDFset.

Jets are reconstructed using the anti-kt jet clustering algo- rithm [28], withradius parameter R=⁰.4. The inputsto there- constructionalgorithmaretopologicalcalorimetercellclusters[29]

calibratedattheEMenergyscale.Theeffectsofpile-uponjeten- ergies are accounted forby a jet-area-based correction [30]. Jets are then calibrated to the hadronic energy scale using pT- and

η-dependentcalibrationfactorsbasedon MCsimulationsandthe

pipe.Thex-axispointsfromtheIPtothecentreoftheLHCring,andthe y-axis pointsupwards.Cylindricalcoordinates(r,φ)areusedinthetransverseplane,φis theazimuthal anglearound thebeam pipe.Thepseudorapidity, η,is definedin termsofthepolarangleθasη= −^ln[^tan(θ/2)]^.

combinationof severalinsitu techniques appliedto data[29].To remove jetswith a significant contribution from pile-up interac- tions, itisrequiredthat atleast50% ofthesummedscalar pT of tracksmatchedtoajetbelongstotracksoriginatingfromthepri- maryvertex.³

TheflavourofjetsintheATLASsimulationisdefinedbymatch- ing jetstohadronswithp_T>5 GeV.A jetislabelledasab-jetif ab-hadronisfoundwithinR=

(η)²+ (φ)²=⁰.3 ofthejet axis; otherwise, ifa c-hadron is found within thesame distance the jet is labelled as a c-jet; andif neitheris the case then the jetislabelledasalight(quark)jet.Thelifetimeandother proper- tiesofb-hadronsareusedtoidentify(b-tag)b-jetswith|η_|<2.5, by exploitingthepropertiesandtopologyoftheirdecayproducts, such as theimpact parameter oftracks (definedas atrack’s dis- tance ofclosestapproach tothe primaryvertexin thetransverse plane), thepresence ofdisplaced vertices,andthereconstruction ofc-hadronandb-hadrondecays.Theb-taggingalgorithmusedin thisanalysis[31]combinestheaboveinformationusingmultivari- atetechniquesandisconfiguredtoachieveanefficiencyof70%for taggingb-jetsinaMC sampleoftt events,¯ whilerejecting80%of c-jetsandmorethan99%oflightjetsinthesamesample.

4. Eventselection

The events of interest in this analysis were triggered by a combination ofsixjet-based triggers. The mostefficient ofthese triggers (accepting about 70% of the signal events) requires two jets identified asb-jets by a dedicated HLT b-tagging algorithm, withtransverse energies(ET) above35 GeV,anda jetwith ET>

145 GeV thatmayormaynotbeoneoftheb-taggedjets.Thetrig- gerefficiencyforthe Sherpa signaleventspassing thefull offline eventselectionis88.1%.

The event selection requires that there be at least three but no morethan fivejetswith|η|<2.5 and pT>30 GeV, andthat exactlytwoofthembeb-tagged.Theb-taggedjetsmusteachhave pT>40 GeV. The angulardistance, R, betweenthem must be smallerthan1.2andthetransversemomentumofthedijetsystem theyform, p^dijet_T ,mustbegreaterthan200 GeV.

The final step of the event selection uses two variables with significant discrimination betweensignal and background to de- fine twosets ofevents,one signal-enrichedandtheothersignal- depleted,referredtohereafteras“SignalRegion”and“ControlRe- gion”. The two variables, which are combined with an artificial neuralnetwork(ANN)intoasinglediscriminant,SN N,are:(1) the dijet pseudorapidity, ηdijet; and(2) the pseudorapidity difference,

η,betweenthedijetandthe balancingjet,wherethebalancing jet is chosen to be the one that, when added to the dijet, gives thethree-jetsystemwiththesmallesttransversemomentum.The correlation of these two variables with the dijet invariant mass, mdijet,isverysmall,allowingtheANNtobetrainedusingselected dataeventsoutsidethemasswindow[⁸⁰,110]^{GeV asbackground} and Z→^bb MC¯ eventsassignal.Fig. 1depictsthedistributionsof

ηdijet, η andSN N inthesignal MCsample andinthedata.The data shownhere include all events with 60<mdijet<160 GeV, and are representative ofthe background asthe signal contribu- tionisestimatedtobeonlyabout1%.TheSignalRegionisdefined by SN N>0.58 and the Control Regionby SN N<0.45. The dis- criminatingpowerof ηdijetandηstemsfromthefactthatsignal productionproceedspredominantlyviaaquark–gluonhardscatter, asopposedtothedominantmulti-jetbackgroundwhichislargely initiatedbygluon–gluonscattering.Duetothedifferencesbetween

3 Amongstallreconstructedproton–protoncollisionsinabunchcrossing,thepri- maryvertexisdefinedasthevertexwiththehighestsummedtrackp²_T.

Fig. 1. Thedistributionsof:(a)thedijetpseudorapidity,|ηdijet|;(b)thepseudorapid- itydifference,|η|,betweenthedijetandthebalancingjet;and(c)theneuralnet- workdiscriminantSN N,inthe Z→^bb signal¯ (redsquares)andinthedata(black circles),includingalleventswith60<mdijet<160 GeV.Thedataisdominatedby themulti-jetbackground.Thetwodashedlinesin (c)indicatetheSN Nvaluesdefin- ingtheSignal(S^{N N}>0.58)andControl(S^{N N}<0.45)Regions.(Forinterpretation ofthereferencestocolourinthisfigurelegend,thereaderisreferredtotheweb versionofthisarticle.)

thegluon andquark PDFs, the Z+^{jet system tends to be more} boostedalongthebeamaxis;hencethe Z -bosonisproducedwith higherηandsmaller ηseparation fromitsrecoil, compared to thebackground.

Since SN N is minimally correlated withm_dijet theControl Re- gion can be used as an unbiased model of the background in theSignal Region. Fig. 2showsthe normalisedratio ofthem_dijet distributions in the Signal and Control Regions, excluding the Z masswindow.A first-orderpolynomialfittothisdistributiongives a good χ² probability, 0.18, anda gradient consistent withzero, (−¹.37±¹.10)×^{10−4 GeV−1.In addition,thevalidity ofassum-}

Fig. 2. Thenormalisedratioofdijetmassdistributionsinthe SignalandControl Regions, excludingthe signalmasswindow,fittedwithafirst-order polynomial.

Thedashedlineindicatesunity.

ing minimal correlation is supported by a test, performed with eventsfrom a Pythia 8 multi-jetMC sample satisfying theabove analysisrequirements,givingaratiooftheabovedistributionscon- sistent with being flat.The impact ofpossible differences in the backgroundmdijetshapebetweentheSignalandControlRegionsis consideredasoneofthesystematicuncertaintiesonthemeasure- ment.

Thetotalnumberofdataeventssatisfyingthefullanalysisse- lection is 236 172 in the Signal Region and 474 810 inthe Con- trol Region. The signal-to-background ratio in a 30 GeV window around the Z -bosonmassis expectedtobe about6% (2%) inthe Signal (Control) Region. The tt events¯ are estimatedto represent about0.5%ofthetotalbackgroundinboththeSignalandControl Regions,andthe Z→^cc and¯ W→^qq¯^ backgroundsare approxi- mately8%and6%ofthesignal,respectively.

5. Cross-sectiondefinition

Thefiducialcrosssectionofresonant Z -bosonproduction,with Z decaying to b_b,¯ σ_Z^fid_→bb¯, isdefinedasfollows.Particle-level jets inMC Z→^bb events¯ arereconstructedfromstableparticles(par- ticleswithlifetimeinexcessof10 ps,excluding muonsandneu- trinos)usingtheanti-kt algorithmwithradiusparameter R=⁰.4.

There must be two particle-level b-jets in the event that satisfy the following fiducial conditions: pT>40 GeV, |η_|<2.5 for the individual jets; and R(jet1,jet2)<1.2, p^dijet_T >200 GeV, 60<

m_dijet<160 GeV forthedijetsystem.

The cross section is extracted from the measured yield of Z→^bb events¯ inthedata,N_Z→bb¯,as

σ_Z^fid_→bb¯= ^NZ→bb¯ L·C_Z→bb¯

,

where C_Z→bb¯ is the efficiency correction factor to correct the detector-level Z →^bb yield¯ to the particle level. The value of C_Z→bb¯ inthe Sherpa MCsignalisfoundtobe16.2%,whichcanbe factorisedintotheproductof:triggerefficiency(88.1%),b-tagging andkinematicselectionefficiency(52.7%),andtheefficiencyofthe SN NrequirementthatdefinestheSignalRegion(35.0%).Theuncer- taintiesonCZ→^bb¯ arediscussedinSection7.

6. Signalextraction

The signal yield is extracted by fitting simultaneously the m_dijet distributionsoftheSignal andControlRegionsintherange [⁶⁰,160]^{GeV withabinned,extended}maximum-likelihood(EML) fit,usingabinwidthof1 GeV.