Measurement of the b-jet cross-section with associated vector boson production with the ATLAS experiment at the LHC HeatherM.GrayaonbehalfoftheATLASCollaboration CERN 2 Abstract. Ameasurementofthecross-sectionforvectorbosonproductioninassociationwithjetscontaining 1 b-hadronsispresentedusing35pb−1 ofdatafromtheLHCcollectedbytheATLASexperimentin2010.Such 0 processes are not only important testsof pQCD but also large, irreducible backgrounds tosearches such as a 2 low massHiggsboson decaying topairsofb-quarks whentheHiggsisproduced inassociationwithavector n boson.TheoreticalpredictionsoftheV+bproductionratehavelargeuncertaintiesandpreviousmeasurements a havereporteddiscrepancies. Cross-sectionsmeasuredusingintheelectronandmuonchannelswillbeshown. J Comparisonswillbemadetorecenttheoreticalpredictionsatthenext-to-leadingorderinα . S 4 2 Vectorbosonproductioninassociationwithjets(V+jets) taining additional jets are vetoed to control backgrounds. ] isoftenusedasatestinggroundforperturbativeQCDcal- TheZ+b-jetcross-sectionismeasuredatthejetlevel,such x e culations. Despite substantial progress in the understand- thateventswithtwojetsenterthedistributionstwice.The - ingandmodellingofinclusivejetproductioninvectorbo- W+b-jetcross-section,ismeasuredattheeventlevel,with p son events,less study hasbeenmade of heavy-flavourjet eachevententeringeachdistributiononce. e production.V+jetsisalsoanimportantbackgroundtomany h TheATLASdetectorconsistsofaninnertrackingsys- [ searchesattheLHC.TheseincludesearchesfortheHiggs boson produced in association with a vector boson and tem(|η|<2.5)surroundedbya2Tsuper-conductingsole- 1 noid,electromagneticandhadroniccalorimeters(|η|<4.9), with the Higgs decaying to a pair of b-quarks [1] or su- v and a muon spectrometer(|η| < 2.7) [7]. Eventsare col- persymmetric models with b-quarks in the final state[2]. 6 lectedusingsingleelectronormuonhigh p triggersand Significantprogresshasbeenmaderecentlyinincreasing T 7 requiredtocontainatleastonereconstructedprimaryver- the accuracy of theoretical calculations [3,4]. Challenges 9 tex.Leptonsarerequiredtohave p > 20GeVwithelec- 4 to theoretical calculations stem from the non-necessarily T tronswithin|η|< 2.47andmuonswithin|η|< 2.4.Events . negligibleb-quarkmassandtheinterplaybetweencontri- 1 containingaZbosonareselectedbyrequiringtwoisolated butionsfromb-quarkproductionininitialandfinalstates. 0 WediscussresultsofameasurementoftheWandZ+b- leptons of the same flavour but with opposite charge and 2 withtheinvariantmassoftheleptonstobeconsistentwith 1 jet cross-sections made by the ATLAS experiment using : 35 pb−1 of LHC data collected in 2010 [5,6]. Both mea- theZbosonmass:76<mll <106GeV.Eventscontaining v surements are challenging, but in different respects: the a W boson are selected by requiring exactly one isolated Xi Z+b-jet is statistically limited by its small cross-section, lepton with missing transverse energy (ETmiss > 25 GeV) r while the W+b-jet process has a larger cross-section, but and transverse mass, mW = 2pℓpν(1−cos(φell−φν)), a also largerbackgroundsresultingina lowsignaltoback- T q T T consistentwiththatofaWboson(mW >40GeV). groundratio. T TheV+b-jetcross-sectionsare measuredin a fiducial Jetsarereconstructedusingtheanti-k algorithmwith T phase space, bycorrectingthe resultsback to the particle a radius parameter R = 0.4 and required to have p > T levelwhileaccountingforalldetectoreffects.Thecutsap- 25 GeV. Jets with anangularseparation∆R < 0.5froma pliedtodefinethefiducialcross-section,closelymatching selectedleptonareremoved.Jetsarerequiredtohave|η|< theexperimentalacceptance,arelistedinTable1.Thelep- 2.1toensuregoodb-taggingperformancewiththefulljet tonmomentumincludestheenergyfromphotonsradiated containedwithin the InnerDetector acceptance.Jets con- within a cone of radius 0.1. The neutrino pT is used as tainingb-quarksareidentifiedusingthedecaylengthsig- thetransversemissingenergy.Particleleveljetsarerecon- nificancebetweentheprimaryandasecondaryvertex[8]. structedusingtheanti-kT algorithmusingallstable parti- Jets with a significance greater than 5.85 are tagged as cles(τ>10ps).Ab-jetisdefinedasajetwithab-hadron b-jets. In a simulated tt¯sample such a cut has a 50% b- with pT > 5GeV withanangularseparationfromthejet jetidentificationefficiencyandamis-tagrateof10%and ∆R<0.3. 0.5%forcandlightjetsrespectively[8].Theb-taggingef- The Z+b-jetcross-section is measured inclusively for ficiencywasmeasuredindatausingab-jetenrichedsam- eventscontainingatleastoneb-jet.DuetothelargerW+b- pleobtainedbyselectingjetscontainingmuons,whichare jet cross-section a differential measurement is made for producedpredominantlyfromb-hadrondecay.Thedistri- eventscontainingeitheroneortwob-jets,buteventscon- bution of the muon transverse momentum relative to the jetaxis,prel,isthenusedtodiscriminatebetweenb,c-and T a e-mail:[email protected] lightjets,beforeandafterb-tagging. EPJWebofConferences Table1.Definitionofthephasespaceforthefiducialcrosssec- VV tioRneqfourirtehmeeWnt+b-jetandZ+b-jWetm+ebasurements[5Z,6+].b Events / 5 GeEvents / 5 Ge110022 ATLAS ∫ L dt = 36 pb-1 DtZWZt((a(→→t→aµτ µτ2+)νµ0 )+-1 )+ 0j+e j( tejsetsst s= 7 TeV) 1100 Diboson Leptonp pℓ >20GeV Single top T T Leptonη |ηℓ|<2.5 11 Dileptonmass - 76−106GeV Neutrinop pν >25GeV - T T Wtransversemass mT >40GeV - 1100--11 Jet p pj >25GeV T T Jetrapidity |yj|<2.1 6600 8800 110000 112200 114400 116600 118800 220000 222200 224400 mmµµ++µµ-- [[GGeeVV]] Jetmultiplicity n≤2 - b-jetmultiplicity n =1orn =2 n ≥1 b b b Fig.1.Di-leptonmassdistributionforeventswithatleastoneb- Jet-leptonseparation ∆R(ℓ,jet)>0.5 - taggedjetwith p >25GeVand|y| <2.1forthemuonchannel T intheZ+b-jetanalysis[5].Thecontributionestimatedfromthe simulatedMCsamplesofthesignalandvariousbackgroundsis Table2.Numberofsignalandbackgroundeventsin35pb−1 of shown. The multi-jet background, estimated with a data-driven data [5,6]. The W+b numbers show the ALPGEN predictions method,isnotshown. normalised to the inclusive NNLO W cross-section, while the Z+jetyieldhasbeennormalisedtothefitresult.TheQCDback- groundinbothcaseshasbeenestimatedfromdata.Thetopback- eV120 G ATLAS Data 2010, s= 7 TeV gdraotau-nddrivinenthmeeWth+odb,-jwethailnealfyosrisZw+basiteswtiamsaetesdtimuasitnedg afropmartisaimlly- nts/ 8 100 Electron + 1 or 2 jets WW++bc e W+light ulation. Thesmaller singletopand dibosons backgrounds were Ev 80 multi-jet estimatedfromsimulation.Uncertaintiesarenotindicated. tt 60 Single top Other EW Z+≥1b-jet W+1b-jet W+2-jet ∫ 40 L dt = 35 pb-1 V+b 64 43 45 V+c 60 192 81 20 V+l 0.0 68 37 Top 6 19.1 77 0 50 100 150 Missing E [GeV] SingleTop 0.3 31 41 T QCD 1.0 18 15.8 Fig.2.MissingETdistributionintheW+b-jetelectronchannel Diboson 0.5 5.8 4.6 in the combined 1- and 2- jet bin after applying the b-tagging requirement [6]. All backgrounds, but QCD multi-jet, are nor- malisedtotheirMCprediction. Theestimatednumberofsignalandbackgroundevents in 35 pb−1 of data is shown in Table. 2. The dilepton in- variant mass in the muon channel for the Z+b-jet analy- enceinefficiencybetweenpromptandnon-promptmuons sisisshowninFig.1.ThedominantbackgroundisZ+jet topassthestandardselectioncriteria.Intheelectronchan- eventswith a lightor c-jet has beenmistagged as a b-jet. neltheQCDbackgroundisestimatedfromafittothemiss- Otherbackgrounds,includingtopandsingle-top,aresmall ingenergydistributionwithatemplateobtainedbyrevers- andthereforeestimatedfromsimulation.Thesmallback- ingcertainelectronidentificationcriteria(seeFig.2).The groundfrommulti-jetproduction,referredtoasQCD,was 50% uncertainty on the QCD backgroundestimate trans- estimatedfromdata,byfittinganexponentialdistribution latesintoa7%uncertaintyontheW+b-jetcross-section. in the dilepton mass in a QCD enriched sample obtained Amaximumlikelihoodfittothesecondaryvertexmass witharelaxedleptonselectioncriteria. distributionwasusedtodiscriminatebetweenb−,c−and In the W+b-jet measurement, particularly in the 2-jet light-jetsandtoextracttheflavourfractiononastatistical selection,thecontributionfromnonW+jetbackgroundsis basis.Theothernon-V+jetsbackgroundswereallowedto largewithaS/B0.1-0.2.ThereforeboththeQCDmultijet floatwithinuncertainties.Thetemplateshapesforeachjet andthetopbackgroundwereestimatedfromdata.Thetop flavourwereestimatedfromsimulationandthesystematic background was estimated in a control region with more uncertaintiesontheshapewerederivedbycomparingdata than 4 jets and extrapolated into the signal region using andMCinmulti-jetsamplesenrichedineitherlightorb- Monte Carlo (MC) simulation. The uncertainty on the b- andc-jets.ExamplesofthefitsareshowninFig.3and4. taggingefficiencycancelsinthisratiothereforeobtaining For the Z+b-jet measurement, the fit is performedfor all an estimate largely independent of the b-tagging uncer- b-jetsintheelectronandmuonchannelscombined,while tainty. The veto on additionaljets was used to reducethe fortheW+b-jetmeasurementitisperformedseparatelyby topbackground. leptonflavourandjetmultiplicity. Heavy-flavour jet production is the dominant process Figure3showsthesecondaryvertexmassdistribution contributingtotheQCDmulti-jetbackground.AstheQCD for W+1-jet events. The distribution is shown both with backgroundislargeandhasalargeuncertaintyaverytight theW+jetcontributionnormalisedtotheNNLOWcross- requirementonleptonisolationwasintroduced.Inthemuon section (left) and after the W+jet contribution has been channel the QCD background largely results from non- scaled by the fit results(right).The secondaryvertexdis- promptmuonsandwasextractedbyexploitingthe differ- tributionfortheZ+b-jetanalysisisshownwiththeflavour 2011HadronColliderPhysicsSymposium Events/0.5 GeV12468000000 AETleLcAtroSn + 1 Jets ∫ L dt DWWWQtSO=t iaCtn+++3htgDbcla5eigl r e2 phE 0TtbW1o-10p, s= 7 TeV Events / 0.5 GeV 123456789000000000 AETleLcAtroSn + 1 jet ∫ L dtDWWWmtSO t=iatun+++h tl3gbclateiigl5- re2j he E0ptttoW1bp0-1, s= 7 TeV ≥ν→ 1 b-jet) [pb] + lW 112050 EEMNA(A(PbbllLLYLuee--jjPPOToeeccGGttnHtt rr 5 oofoIEECrFAnnnoNNhN lmyCaa S++ n nhfM rd.JJao EIInmMMM .a uMMMnoYYdEn )PCSh)an. Data 20∫1L0d,t =s A=3T75L TpAebSV-1 0 2 Secondary V4ertex Mass [GeV]6 00 1 2 Sec3ondary V4ertex Ma5ss [GeV]6 σ( 5 Fig.3.Thesecondary vertexmassdistributionforb-taggedjets inW+b-jeteventscontaininganelectronandasinglejet[6].The distribution is shown with the b, c and light contributions nor- 0 1 jet 2 jet 1+2 jet malisedtotheinclusiveNNLOW+jetcross-section(left)andto thefitresults(right). Fig.5.Comparisonofthemeasuredandpredictedfiducialcross- sections of b-jet production in association with a W boson [6]. VV Thecross-sectionisshowninexclusivebinsfor1,2and1+2jets ee Jets / 0.5 GJets / 0.5 G45450000 dZZZbaa+++tcbclaikg gh2rt0o1u0n d(ss=7TeV) ataonnddNstLoeOptahproarsteeedlfiyrcotfimoornAlseLpotbPotGnaiEflnNeadvaounusdrisnP.gyTtthhheieam.5eflaasvuoreumrnenutmibsecromscphaermede ∫ L dt = 36 pb-1 Table3.Measuredandpredictedfiducialcross-sectionofinclu- 3300 siveb-jetproductioninassociationwithaZ-boson[5] 2200 ATLAS Cross-section[pb] Sherpa 3.29±0.04(statonly) 1100 ALPGEN 2.23±0.01(statonly) MCFM 3.88±0.58 00 00 11 22 33 44 55 66 77 88 99 1100 ATLAS 3.55+0.82(stat.)±0.12(syst) SSVV00 mmaassss [[GGeeVV]] −0.74 Fig.4.Thesecondary vertexmassdistributionforb-taggedjets quarksandcontributionwithamasslessbquarkintheini- in selected Z+b events. The b, c and light templates have been tialstate treatedusingaschemebasedonbquarksPDFs. normalised to the fit results. The small contribution from other TheW+b-jetcross-sectionissystematicallyfoundtohave backgroundsisalsoindicated[5]. a small excess over theoretical predictions, both in the 1 and2-jetexclusivebinsandthe1+2-jetexclusivebin.How- evertheexcessissmallandthemeasurementisconsistent contributionsnormalisedtothefitresults.Thereisnovisi- blecontributionfromZ+lightjetsbecausethebestfitvalue atthe1.5σlevel. wasconsistentwithzerowithinuncertainties. In conclusion,ATLASmeasurementsof vectorboson in association with b-jets using the first 35 pb−1 of data ThefiducialV+jetcross-sectionswereobtainedbyun- havebeenpresented.Despitelargeuncertainties,theZ+b- foldingthemeasuredb-fractionusingMCsimulationsam- jet cross-section is found to be consistent with NLO pre- ples. Alpgen was used for the W+b-jetmeasurementand dictions,whileasmallexcessatthe1.5σlevelisobserved SherpafortheZ+b-jetmeasurement.Themeasuredcross- intheW+b-jetmeasurement. sectionshaveuncertaintiesfromthebackgrounds,template shapeandtheunfoldingfactors. Uncertaintiesontheb-taggingefficiencyresultinasys- References tematicuncertaintyof10%(12%)fortheZ(W)+b-jetmea- surement. The unfolding factor is sensitive to the accu- 1. ATLAS Collaboration, ATLAS-CONF-2011-103, racy of the Monte Carlo modelling such as the b-jet p T https://cdsweb.cern.ch/record/1369826/ spectrum and uncertainty on the opening angle between 2. ATLAS Collaboration, submitted to PRL, the pairs of b-quarks.The b-jet p spectrum is important T arXiv:1112.3832 because the b-tagging efficiency depends on p and the T 3. Campbell,J.M.etal,Phys.Rev.D.79(2009)034023. openingangledistributiondeterminesthe probabilitythat 4. Caola,F.etal,arXiv:1107.3714,(2011) the pair of b-quarks will be reconstructed as a single jet. 5. ATLAS Collaboration, Phys. Lett. B 706 (2012) 295- Uncertaintiesonthejetenergyscaleresultsina4%(7%) 313,arXiv:1109.1403 uncertaintyintheZ(W)+b-jetmeasurement. 6. ATLAS Collaboration, Phys. Lett. B 707 (2012) 418- TheresultsfortheZ+b-jetmeasurementaresummaris- 437,arXiv:1109:1470 ed inTable 3.The Z+b-jetcross-sectionagreeswellwith 7. ATLASCollaboration,JINST01(2008),S08003 theNLOMCFMpredictionof3.88±0.58pb[9,10]. 8. ATLAS Collaboration, ATLAS-CONF-2011-089 TheresultsfortheW+b-jetmeasurementaresummaris- (2011),https://cdsweb.cern.ch/record/1356198 edinFig.5.ThemeasurementsarecomparedtoNLOpre- 9. Campbell, J.M. et al, Nucl.Phys.Proc.Suppl. 205-206 dictions obtained using the 5 flavour number scheme [3] (2010)10-15,arXiv:1007.3492 after includinga non-perturbativecorrectionto the cross- 10. Campbell, John M. et al. Phys.Rev. D69 (2004) sectionof0.93±0.07[6]. Thecorrectionincludesa con- 074021,hep-ph/0312024 tribution from a bb¯ pair in the final state with massive b-