Search for the Standard Model Higgs boson decay to μ+μ− with the ATLAS detector

(1)

Article

Reference

Search for the Standard Model Higgs boson decay to μ+μ− with the ATLAS detector

ATLAS Collaboration

ALEXANDRE, Gauthier (Collab.), et al.

Abstract

A search is reported for Higgs boson decay to μ+μ− using data with an integrated luminosity of 24.8 fb−1 collected with the ATLAS detector in pp collisions at s=7and8 TeV at the CERN Large Hadron Collider. The observed dimuon invariant mass distribution is consistent with the Standard Model background-only hypothesis in the 120–150 GeV search range. For a Higgs boson with a mass of 125.5 GeV, the observed (expected) upper limit at the 95% confidence level is 7.0 (7.2) times the Standard Model expectation. This corresponds to an upper limit on the branching ratio BR(H→μ+μ−) of 1.5×10−3 .

ATLAS Collaboration, ALEXANDRE, Gauthier (Collab.), et al . Search for the Standard Model Higgs boson decay to μ+μ− with the ATLAS detector. Physics Letters. B , 2014, vol. 738, p.

68-86

DOI : 10.1016/j.physletb.2014.09.008

Available at:

http://archive-ouverte.unige.ch/unige:55783

Disclaimer: layout of this document may differ from the published version.

1 / 1

(2)

Contents lists available atScienceDirect

Physics Letters B

www.elsevier.com/locate/physletb

Search for the Standard Model Higgs boson decay to μ

⁺

μ

⁻

with the ATLAS detector

.ATLASCollaboration

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

Articlehistory:

Received1July2014

Receivedinrevisedform2September2014 Accepted3September2014

Availableonline8September2014 Editor: H.Weerts

A search is reported for Higgs boson decay to μ⁺μ⁻ using data with an integrated luminosity of 24.8 fb⁻¹collectedwith theATLASdetector inpp collisions at√_s

=⁷ and 8 TeV atthe CERNLarge Hadron Collider. The observed dimuon invariant mass distribution is consistent with the Standard Model background-only hypothesis inthe120–150 GeV searchrange.For aHiggsbosonwithamass of 125.5 GeV, theobserved (expected) upper limitatthe 95% conﬁdencelevel is 7.0 (7.2)times the StandardModelexpectation.ThiscorrespondstoanupperlimitonthebranchingratioBR(H→μ⁺μ⁻) of1.5×¹⁰⁻³^.

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

1. Introduction

The Standard Model (SM) describes a wide range of particle physics phenomena toa highdegree ofprecision. Inthe SM,the Brout–Englert–Higgs(BEH)mechanism[1–3]spontaneouslybreaks the electroweak(EW) gauge symmetry and generatesmasses for the W and Z gauge bosons as well asfor the chargedfermions via Yukawacouplings[4–6]. Insearchesforthe Higgsbosonpre- dictedbytheBEHmechanism,theATLAS andCMSCollaborations havediscoveredanewparticle,viadecaysintogaugebosons[7,8], withamassofapproximately125.5 GeVandmeasuredproperties consistentwiththosepredictedbytheSM[9–12].

Higgsbosondecaystob_b,¯ τ⁺τ⁻ andμ⁺μ⁻ can bemeasured at the LHC with their SM branching ratios proportional to the squares of the fermion masses. The SM branching ratio for the H→μ⁺μ⁻ decayis21.9×¹⁰⁻⁵ ^for^a^Higgs^boson^mass^(mH)of 125GeV[13,14].TheH→μ⁺μ⁻decayhasacleanﬁnalstatesig- naturethatallowsameasurementoftheHiggsbosoncouplingto second-generationfermions.Thedominantirreduciblebackground isthe Z/γ^∗→μ⁺μ⁻ process,whichhasanapproximatelythree orders of magnitudehigher production ratecompared to that of theexpectedsignal.

In this Letter a search for the H→μ⁺μ⁻ decay of the SM Higgs boson is presented.This search for the presence of a nar- row H→μ⁺μ⁻ resonance, with a signal width determined by experimentalresolution,isperformedbyﬁttingtheinvariantmass distribution inthe 110–160 GeV region.This rangeallows deter- mining background shape and normalisation and setting a limit

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

on the dimuon decay of the SM Higgs boson with a mass of 125.5GeV.Section2givesadescriptionoftheexperimentalsetup and summarises the data sample andMonte Carlo (MC) simulation samplesusedtomodelthesignal process andto developan analytical model for the background processes. Sections3 and 4 describe the event selectionsand categorisation. Analytical models used to describe invariant mass distributions for signal and background processes are discussed inSection 5, andsystematic uncertaintiesaredetailedinSection6.Theresultsarepresentedin Section7.

2. Experimentalsetup,dataandsimulatedsamples

Thissearch isperformedonthedatasamplerecordedin2011 and 2012 by the ATLAS detector in pp collisions at the LHC at

√s=⁷ ^{and 8 TeV,} respectively. ATLAS [15] is a general-purpose particle detectorwithacylindrical geometryandconsistsofsev- eral subdetectors surroundingthe interaction point and covering almost the full solid angle.¹ The trajectories and momenta of charged particles are measured within the pseudorapidity range of |η_|<2.5 by multi-layer silicon pixel and microstrip detec- tors as well as a transition radiation tracker. The tracking system is immersed in a 2 T magnetic ﬁeld produced by a superconducting solenoid, and is surrounded by a high-granularity

1 ATLAS usesaright-handedcoordinatesystemwithitsoriginatthe nominal interactionpoint(IP)inthecentreofthedetectorandthe z-axisalongthebeam pipe.Thex-axispointsfromtheIPtothecentreoftheLHCring,andthe y-axis points upward.Cylindricalcoordinates(r,φ)areusedinthetransverseplane,φ beingtheazimuthalanglearoundthebeampipe.Thepseudorapidityisdeﬁnedin termsofthepolarangleθasη= −^{ln tan}(θ/2).

http://dx.doi.org/10.1016/j.physletb.2014.09.008

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

(3)

liquid-argon (LAr) electromagnetic samplingcalorimeter covering

|η_|<3.2.Anironandscintillatortilehadroniccalorimeterprovides coverageintherange|η_|<1.7.TheLArcalorimetersalsoprovide measurements ofelectromagneticandhadronicenergydepositsin theregion1.5<|η_|<4.9.The muonspectrometer surroundsthe calorimetersand consists ofthree large air-core superconducting magnetswithatoroidal ﬁeldof0.5 Tinthebarrelregionand1 T intheforwardregions, anda systemofprecision trackingcham- bers; it provides coverage for |η_|<2.7. The muon spectrometer alsoincludesfastdetectorsfortriggeringcovering|η|<2.4.

Events were recorded with a trigger [16] requiring at least onemuoncandidatewithtransversemomentum pT>18 GeV for 7 TeVdataandp_T>24 GeV for8 TeVdata.Thistriggerisapprox- imately70%efficientformuonswith|η|<1.05 andapproximately 90%efficientformuonswith|η_|>1.05.Thedifferencesintheef- ficiencyaremostly duetothedifferentgeometricalacceptanceof themuon triggersysteminthese regions.For8 TeV data,events were alsorecorded with a dimuontrigger requiring atleast two muon candidates with transverse momenta of pT>18 GeV and p_T>8 GeV.Afterapplyingdataqualityrequirements,thetotalin- tegratedluminosity oftheselected datasample is4.5±⁰.1 fb⁻¹ for7 TeVdataand20.3±⁰.6 fb⁻¹ for8 TeVdata.Theassociated systematicuncertaintiesaresummarisedinSection6.

At the LHC, SM Higgs boson production is dominated by the gluonfusion(ggF)processwiththenexttwomostimportantcon- tributionsarising fromthevectorbosonfusion (VBF)process and productioninassociation withvector bosons(VH).Thesignal MC samples are generated in 5 GeV steps of the Higgs boson mass from120 GeV to 150 GeV. The ggF and VBF samples are gener- atedatnext-to-leading-order (NLO) inQCD withPowheg[17,18]

withthepartonshoweringmodelledbyPythia8[19]forthe8TeV samples and Pythia [20] for the 7 TeV samples. The CT10 [21]

parton distribution functions (PDFs) are used in Powheg, with the ATLAS Underlying Event Tune, AU2 [22]. The ggF Higgs boson pT spectruminPowhegistunedtoagreewiththeprediction fromHRES[23,24];thisprocedureshiftstheHiggsboson p_T spec- trumtoslightlysmallervalues.TheVH samplesaregeneratedwith Pythia8,usingAU2andtheCTEQ6[25]PDFs.

ThepredictedSMHiggsbosoncross-sectionsandbranchingra- tios are compiled in Refs. [13,14,26]. The cross section for the ggF processis calculatedat next-to-next-to-leading-order(NNLO) [27–32] in QCD and NLO electroweak (EW) corrections are applied[33,34],assumingthattheQCDandEWcorrectionsfactorise.

Thecross section fortheVBF process iscalculated withfull NLO QCDandEWcorrections [35–37]andapproximateNNLOQCDcor- rections[38].ThecrosssectionfortheVH processiscalculatedat NNLO[39,40] inQCD, andNLO EWradiative corrections[41] are applied.Thebranchingratiosfor H→μ⁺μ⁻ decaysasafunction ofm_H arecalculatedusingHdecay[42].

ThefollowingMC eventgeneratorsare usedto simulateback- ground processes: Alpgen [43] + ^Herwig ^[44] ^for ^W +^jets, MC@NLO[45]+^Herwig^for^tt,¯ t W andtb,AcerMC[46]+^Pythia for tqb, Powheg + ^Pythia⁸ ^for ^Z/γ^∗ and q^q¯→^{W W ,} ^gg2WW [47]+^Herwig^for^gg→^{W W ,}^Powheg^[48]+^Pythia^for^{W Z and} Z Z andMadGraph[49,50]+^Pythia^for^Wγ^∗.W W ,W Z ,Z Z and Wγ^∗productionarereferredtoasdibosonproductionlaterinthe text.Inaddition,contributionsfromW bosonproductioninasso- ciationwithoneormorejets(W+^jets),^whereôneôf^the^jetsîs misidentifiedasamuon,areestimatedusingadatacontrolregion containingsame-signdimuoneventsasdescribedinRef.[9].

ThesignalandbackgroundMC samplesareprocessedthrough theATLASdetectorsimulation[51]basedonGeant4[52],followed bythe samereconstruction algorithms that areused forcollision data.The effects arising from multiple collisions in the same or neighbouring bunch crossings (pile-up) are included in the MC

simulation, matching the pile-up conditions of the selected data sample.

3. Eventselection

Events are required to contain at least one reconstructed pp collisionvertexcandidatewithatleastthreeassociatedtrackseach withpT>0.4 GeV.Thevertexwiththelargestsumofp²_T oftracks isconsidered tobethe primaryvertex. Muon candidates[53]are reconstructed by matchingtracks inthe inner detectorto tracks reconstructed in themuon spectrometer. Inaddition to stringent track quality requirements imposed for muon identiﬁcation, the muon tracksmust be consistent withhavingoriginatedfromthe primary vertex. All selected muon candidates are required to be within|η|<2.5.Muoncandidatesmustpasstrackandcalorimeter isolation requirements that scalewith the pT of themuon track.

TheisolationiscalculatedasthescalarsumofthepTofadditional tracks or the E_T of calorimeter energy deposits within cone of radius R=

(φ)²+(η²)=⁰.3 aroundthe muon track,nor- malisedbypμ

T.For8 TeVdata,eachmuonwithpμ

T >20(15)GeV isrequiredtohavea normalisedtrackisolationsmallerthan 0.12 (0.08)andanormalisedcalorimetricisolationlessthan0.30(0.18).

For7 TeVdata,equivalentvaluesforpμ

T >15 GeV are0.15and0.2 for trackandcalorimetric isolation, respectively. Due to different pile-up conditions, different isolation criteria are used for7 TeV and8 TeVdata.

Jets are reconstructed from clusters of calorimeter cells using the anti-kt algorithm [54,55] witha radius parameter of0.4.

The selected jets must satisfy ET >25 GeV for |η_|<2.4 and E_T>30 GeV for 2.4≤ |η|<4.5. Muon candidates overlapping with the selected jets within a cone of radius R=⁰.4 are re- moved from the analysis. In the pseudorapidity range |η_|<2.5, jetsoriginatingfromb-quarksareidentifiedusingab-taggingalgo- rithm[56,57]withanefficiencyofapproximately80%,determined fromtt MC¯ events,andwithamisidentificationrateforselecting light-quark orgluon jets ofless than 1%. The missingtransverse momentum[58],E_T^miss,isthemagnitudeofthevectorsumofthe pT of muons, electrons, photons, jetsand clustersof calorimeter cellswith|η_|<4.9 notassociatedwiththeseobjects.

Corrections are applied to simulated MC samples in order to account for differences betweendata and MC simulation for the triggerandidentificationefficiencyandforthemuonmomentum scaleandresolution.Thetriggerandreconstructionefficiencycor- rectionsare measuredusing Z→μ⁺μ⁻ eventsandarefound to bewithin2%ofunity.Themuonmomentumcorrectionsaredeter- minedbycomparingthereconstructedinvariantmassdistribution of Z→μ⁺μ⁻ events in data with that fromsimulated events;

thesecorrectionsarewithin0.1%ofunity.

H→μ⁺μ⁻ candidateeventsare selectedbyrequiringexactly twooppositelychargedmuonswithtransversemomentum pμ1

T >

25 GeV and pμ2

T >15 GeV for theleading andsubleadingmuon, respectively.Selectedeventsmustcontainatleastonemuoniden- tiﬁed by the trigger system within a cone of radius R=⁰.15 centredonthereconstructedmuoncandidate.Thedominantback- groundin this search is Z/γ^∗_→μ⁺μ⁻ production, followedby smallerbackgroundsfromsingleandpairproductionoftopquarks and diboson processes.To suppress backgrounds from top quark pairproductionanddibosonprocesses,eventsarerequiredtohave E^miss_T <80 GeV. The dimuon invariant mass distribution mμ⁺μ⁻ and the dimuon transverse momentum pμ⁺μ⁻

T fordata and MC

eventspassing allthe selectionrequirementsso farare shownin Fig. 1.The numberof expectedsignal events formH =^{125 GeV,} the expected background contributions, and the number of observed dataeventsinthemμ⁺μ⁻ regionfrom122.5to127.5 GeV

(4)

Fig. 1. Thedistributionofthedimuoninvariantmass(top)anddimuonmomentum (bottom)for7 TeVand8 TeVdatawithalltheselectionrequirementsdescribedin Section3.TheexpectedsignalisshownformH=^{125 GeV.}

are shown in Table 1. The MC background yields are given to illustratetheexpectedbackgroundcomposition.Theselectioneﬃ- ciencytimesacceptanceforsignaleventswithmH=125 GeV after allselectioncriteriadescribedthusfarisapproximately55%.

The expected background processes produce smooth mμ⁺μ⁻ distributionsinthesearchwindow,allowing thetotalbackground normalisationandshape ineach categoryto be derivedfromﬁt- ting the data asdescribed in Section 5. The mμ⁺μ⁻ distribution isexamined inthe range110–160 GeV. Thisrangeis largerthan the 120–150 GeV search window in order to account for signal resolutioneffectsandtoallowsuﬃcientsidebandsforbackground normalisation.

4. Eventcategorisation

Toincrease sensitivityto the Higgs bosonsignal, the selected eventsareseparatedintosevenmutuallyexclusivecategorieswith differentsignal-to-backgroundratiosbasedontheirmuonpseudo- rapidity(ημ), pμ⁺μ⁻

T ,andVBF dijetsignature.Eventsproduced in theVBF process are characterisedby two forwardjetswith little hadronicactivitybetweenthem.The VBF categoryisthusdeﬁned byrequiringtheeventstohaveatleasttwojetswithaninvariant massgreaterthan500GeV, |ηjet₁−ηjet₂|>3 and ηjet₁×ηjet₂<0.

Ineventswithmorethan twojets,those withthehighest pT are used in the selection. Events with at least one jet identiﬁed as originatingfromab-quarkareexcludedfromtheVBFcategory.

Theeventsthatarenot selectedfortheVBF category areclas- siﬁedusing pμ⁺μ⁻

T .Signaleventshaveonaveragelargervaluesof pμ⁺μ⁻

T thantheZ/γ^∗ backgroundevents.Therefore,theremaining

Table 1

NumberofexpectedsignaleventsformH=^{125 GeV,}^number^of^expected^back- groundeventspredictedbytheMCsimulation,andnumberofobserveddataevents withinawindowof|mH−mμ⁺μ⁻|≤2.5 GeV afterallselectioncriteriaareapplied.

Onlystatisticaluncertaintiesaregiven.Thetheoreticalsystematicuncertaintyon thedominantZ/γ^∗backgroundisabout4%.TheMCbackgroundyieldsaregiven toillustratetheexpectedbackgroundcomposition.

7 TeV 8 TeV

Signal (125 GeV) 5.6±⁰.1 32.7±⁰.2

Z/γ^∗ 3110±⁴⁰ ^{16 660}±²⁷⁰

W Z/Z Z/Wγ 2.2±0.2 12.3±0.7

t¯t 75.6±1.8 509.2±2.7

W W 23.2±0.5 123.3±1.6

Single top 7.2±⁰.9 54.5±⁰.6

W+^jets ³.2±¹.5 38±⁴

Total Bkg. 3220±⁴⁰ ^{17 390}±²⁷⁰

Observed 3344 17 745

Table 2

Expectedsignalyields(NS)formH=125 GeV andtheratioNS/√

NBusingthesim- ulatedMCbackgroundyields(NB)withinawindowof|^mH−^mμ⁺μ⁻|≤².5 GeV for eachoftheeventcategoriesunderstudy.Inaddition,thefullwidthathalfmaxi- mum(FWHM)ofthesignalm_μ+μ⁻distribution,modelledasdescribedinSection5, isgiven.Alsoshownareχ²/ndof ofthestandalonefitstothemμ⁺μ⁻distribution ineachcategoryusingmodelsdescribedinSection5.The7 TeV VBFcategorydoes notincludeasufficientnumberofeventstocomputeameaningfulχ²value.The goodnessoffitforthebackgrounddescriptioninthiscategorywasverifiedusing MCsimulationtoensurethatthebackgroundmodelprovidesaconsistentdescrip- tionofthe7 TeVand8 TeVVBFcategories.

√s [TeV]

Category NS √N_S

NB FWHM

[GeV]

χ²/ndof

8 non-cen. low p^μ_T⁺^μ⁻ 6.1 0.07 6.6 49.8/48

8 cen. low p^μ_T⁺^μ⁻ 2.6 0.06 5.5 52.8/48

8 non-cen. medium p^μ_T⁺^μ⁻ 10.4 0.15 6.6 45.1/48

8 cen. medium p^μ_T⁺^μ⁻ 4.7 0.13 5.6 36.7/48

8 non-cen. high p^μ_T⁺^μ⁻ 5.5 0.13 7.2 26.7/48

8 cen. high p^μ_T⁺^μ⁻ 2.6 0.10 6.0 32.3/48

8 VBF 0.8 0.09 7.0 18.6/19

7 non-cen. low p^μ_T⁺^μ⁻ 1.0 0.03 6.8 42.0/48

7 cen. low p^μ_T⁺^μ⁻ 0.5 0.03 5.3 43.5/48

7 non-cen. medium p^μ_T⁺^μ⁻ 1.8 0.06 6.9 41.2/48

7 cen. medium p^μ_T⁺^μ⁻ 0.8 0.05 5.5 34.4/48

7 non-cen. high p^μ_T⁺^μ⁻ 0.9 0.05 7.5 60.0/48

7 cen. high p^μ_T⁺^μ⁻ 0.5 0.05 5.9 56.2/48

eventsareseparatedintothree pμ⁺μ⁻

T categories:low(<15 GeV), medium (15–50 GeV) and high (>50 GeV). To further improve the search sensitivity, each of these threecategories is also sub- divided intoa central category with|η_μ₁_|<1 and |η_μ₂_|<1 and anon-central categorycontainingall remainingevents.Thisvalue forthe ημ boundaryhasbeenchosen by scanninga rangeofημ values and selecting a value with the highest signal sensitivity.

Themuonmomentummeasurementforthecentralmuonsismore precise,producinganarrowermμ⁺μ⁻ distributionforsignalevents inthecentralcategoryandthusresultinginahigheroverallsignal sensitivity. Table 2showsthe signaleventyields, NS/√

NB ratios, approximate signal width andresults ofthe ﬁts to the data, de- scribedinSection5,forallanalysiscategories.

5. Signalandbackgroundmodels

Analytical models are used to describe the mμ⁺μ⁻ distributions for signal and background processes. The simulated samples detailedinSection2areusedtodevelopbackgroundmodels