Search for neutral resonances decaying into a Z boson and a pair of b jets or τ leptons

(1)

Contents lists available atScienceDirect

Physics Letters B

www.elsevier.com/locate/physletb

Search for neutral resonances decaying into a Z boson and a pair of b jets or τ leptons

.The CMS Collaboration

CERN,Switzerland

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

Articlehistory:

Received9March2016

Receivedinrevisedform15May2016 Accepted27May2016

Availableonline31May2016 Editor: M.Doser

Keywords:

CMS Physics Higgs 2HDM BSM b-Tagging Tau Lepton

AsearchisperformedforanewresonancedecayingintoalighterresonanceandaZboson.Twochannels arestudied,targetingthedecayofthelighterresonanceintoeitherapairofoppositelychargedτleptons orabb pair. TheZbosonis identiﬁedviaitsdecays toelectrons ormuons. The searchexploitsdata collectedbytheCMSexperiment atacentre-of-massenergyof8TeV,correspondingtoan integrated luminosityof19.8fb⁻¹.Nosigniﬁcantdeviationsareobservedfromthestandardmodelexpectationand limitsaresetonproductioncrosssectionsandparametersoftwo-Higgs-doubletmodels.

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

1. Introduction

Theobservationofanewparticlewithamassofapproximately 125GeV was reportedby theATLASandCMSexperiments atthe CERNLHC inthe WW,ZZand

γ γ

ﬁnal states [1–3].Evidence of the decay of the particle to pairs of fermions (

τ τ

and bb) has also been reported in Refs. [4–6]. The measurements of branching fractions, production rates,spin andparity are all consistent with the predictions for the standard model (SM) Higgs boson [7,8],whereinasingledoubletofHiggsﬁeldsispresent.However, additionalHiggsbosons are expectedinsimpleextensions ofthe SM scalarsector, such asmodels withtwo Higgs-bosondoublets (2HDMs) [9]. These models predict ﬁve physical Higgs particles thatariseasaconsequenceoftheelectroweaksymmetry-breaking mechanism:twoneutralCP-evenscalars(h,H),oneneutralCP-odd pseudoscalar(A),andtwochargedscalars(H^±).

Animportant motivationfor 2HDMsis that such modelspro- vide away toaccommodate the asymmetry betweenmatterand antimatterobservedintheuniverse[9,10].AnextensionoftheSM scalarsectorwithtwo Higgsbosondoublets wouldalsonaturally arisein supersymmetry[11,12],which requiresa scalarstructure morecomplexthanasingle doublet.Axionmodels[13]providea

E-mailaddress:[email protected].

stronginteractionthatdoesnotviolateCPsymmetryandgiverise toaneffectivelow-energytheorywithtwoHiggsdoublets.Finally, ithasrecentlybeennoted[14] thatcertainrealisationsof2HDMs can accommodate the muon g-2 anomaly [15] without violating presenttheoreticalandexperimentalconstraints.

In the most general case, 14 parameters describe the scalar sector of a 2HDM[9].Only sixfree parameters remain once the experimental observationsareincludedby imposingtheso-called Z2 symmetry to suppressﬂavour changing neutralcurrents, and by ﬁxing both thevalues of themass of the recentlydiscovered SM-like Higgsboson(125GeV)[16] andtheelectroweakvacuum expectationvalue (246GeV).The compatibilityofaSM-likeHiggs boson with 2HDMs is possible in the so-called alignment limit.

Thealignmentlimitisreachedwhencos(β−

α

)→^0,^where^tanβ is the ratioof thevacuum expectationvaluesand

α

is themix- ingangleofthetwoHiggsdoublets. Insucharegime,one ofthe CP-evenscalars, h or H,is identiﬁed withthe SM-like Higgs boson. Arecenttheoreticalstudy[10] hasshownthat,in thislimit, a large mass splitting (>100 GeV) betweenthe A andH bosons would favour the electroweak phase transition that wouldbe at theoriginofbaryogenesisintheearlyuniverse,satisfyingthereby thecurrentlyobservedmatter–antimatterasymmetry.Inthiscon- text, themost frequentdecaymode ofthe pseudoscalarA boson would be A→^ZH. ^Since ^the ^analysis ^strategy ^presented ⁱⁿ ^this paperisindependentoftheassumedmodelandparityoftheres- http://dx.doi.org/10.1016/j.physletb.2016.05.087

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

(2)

onance,theresultscanalsobeinterpretedinthereversedtopology H→^ZA,^where^theêxpected^2HDM^mass^hierarchyîsînvertedând themassofAisexpectedtobelight[17].Forbothtopologies,the lighterscalarresonance(AorH)isnotidentifiedwiththeSM-like Higgsboson.

Thispaperdescribestheﬁrst CMSsearch foranewresonance decayingintoalighterresonanceandaZboson.Twosearchesare performed,targetingthedecayofthelighterresonanceintoeither apairofoppositely charged

τ

leptonsorabb pair.Inbothcases, the Z boson is identiﬁed via its decay into a pair of oppositely chargedelectrons ormuons(lightleptons),labelledinthetextby thesymbol.Thechoiceofbb and

τ τ

ﬁnalstatesismotivatedby thelargebranchingfractionspredictedinmostofthe2HDMphase space[18].Forthe

τ τ

channel,thefollowing

τ τ

ﬁnalstatesare considered:e

μ

,e

τ

h,

μτ

h,and

τ

h

τ

h,where

τ

hindicatesthedecays

τ

→hadrons+

ν

τ . Given its sensitivity to the 2HDMparameter spaceregion where cos(β−

α

)≈^0,^the ^search ^presented ⁱⁿ ^this paperiscomplementarytootherrelatedsearchesperformedinthe sameﬁnalstatebytheATLASandCMScollaborations[19,20].

2. TheCMSdetector

Thecentralfeature oftheCMSapparatusisasuperconducting solenoid of6 m internal diameter, providing a magnetic ﬁeld of 3.8 T.Locatedinconcentriclayers withinthesolenoidvolumeare a silicon pixel andstrip tracker, a lead tungstate crystalelectro- magnetic calorimeter(ECAL), anda brass and scintillator hadron calorimeter (HCAL), each composed of one barrel and two end- capsections.Theselayersprovidecoverageuptoapseudorapidity

|

η

|=².5.Extensiveforwardcalorimetrycomplementsareprovided bytheendcapdetectorsfor|

η

_|<5.2.Combiningtheenergymea- surementin the ECAL with the measurement in the tracker,the momentumresolutionforelectronswithp_T≈⁴⁵^{GeV from}^Z→^ee decaysrangesfrom1.7% fornonshowering electronsinthe barrel regionto4.5%forshoweringelectronsintheendcaps[21].Muons are measured in gas-ionisation detectors embedded in the steel ﬂux-return yoke outside the solenoid. They cover the pseudorapidity range |

η

_|<2.4, with detection planes made using three technologies: drift tubes, cathode strip chambers, and resistive platechambers.Matchingmuonstotracksmeasuredinthesilicon tracker results in a relative transverse momentum resolution for muonswith20<p_T<100GeV of1.3–2.0%inthebarrelandbet- terthan6% intheendcaps[22].The ﬁrstleveloftheCMStrigger systemusesinformationfromthecalorimetersandmuondetectors toselectthe mostinteresting events.Ahigh-level triggerproces- sorfarmdecreasestheeventratefromapproximately100 kHzto 600 Hzbeforedatastorage.AmoredetaileddescriptionoftheCMS detector,together witha deﬁnitionof thecoordinate systemand kinematicvariables,canbefoundinRef.[23].

3. Dataandsimulatedsamples

ThedatausedforthissearchwerecollectedbytheCMSexper- imentat √

s=⁸^TeV, ând^correspond ^to â ^total întegrated^lumi- nosityof19.8fb⁻¹.Theaveragenumberofinteractionsperbunch crossing(pileup)inthedatawas21[24].Eventswereselectedus- ingdielectronanddimuontriggers[21,22].ThesetriggershavepT thresholdsof17and8GeV for theleadingandsubleadinglepton respectively,andrequirerelativelyloosereconstructionandidenti- ficationcriteria.

Themain SMbackgroundprocesses givingrise toprompt lep- tonsareW/Z +^jets, ^tt+ ^jets,^tW, ^and^diboson^production^(WW, ZZ, and WZ). The SM background contribution from ZZ is gen- erated at next-to-leading order (NLO) with powheg 1.0 [25] for the

τ τ

channelandusingtheleading-order(LO) MadGraph 5.1

Monte Carlo(MC) program [26],matched to pythia 6.4 [27] for the parton showering and hadronization, for the bb channel.

Single top quark events are generatedat NLO using powheg 1.0.

Simulated eventsfor other samplesare obtainedusing the Mad- Graph 5.1 MC matched to pythia 6.4. The pythia parameters affecting thedescription oftheunderlying eventare setto those of the Z2*tune [28]. All generators used forprocesses including

τ

leptons in the ﬁnal state are interfaced with tauola 2.4 [29]

forthe simulationofthe

τ

decays.Thedetectorresponse issim- ulated usinga detaileddescriptionof CMS, basedon the Geant4 toolkit[30].Thesimulatedsamplesaccountforcontributionsfrom pileupcollisionsthatreﬂectthedistributionsobservedindata.The trigger andreconstruction eﬃciencyinthe simulationis rescaled byasmuchas2%inordertomatchthatmeasuredinthedata[24].

Twobenchmark2HDMprocessesareconsideredassignal:H→ ZA and A→^ZH,^where^the^lightest^boson(pseudoscalarorscalar, according to the process) can decayto

τ τ

or bb,and the Z decays to . The MadGraph 5.1program,interfacedto pythia 6.4 and tauola 2.4,wasusedtogeneratesignalsamplescorresponding to differentvaluesofAandH masses(mA andmH,respectively).

The same propertiesof the SM Higgs boson are assigned to the lightest scalarboson,h, andits massm_hisﬁxed at125GeV. The identiﬁcation of the observed Higgs boson, together with all its measuredproperties,withthescalarh constrainsthephenomeno- logically reliable parameterspace regions to not depart fromthe SM-likeconditioncos(β−

α

)≈^0.^Thiscorrespondstotheso-called alignmentlimit[31].Consideringtheparameterspacestillallowed bydirectsearches[12],thechosenvaluesforcos(β−

α

)andtanβ are0.01and1.5,respectively,andtype-IIYukawacouplingsareas- sumedforthebenchmarkprocesses.

The massesofthe chargedHiggsbosons (m_H±) are keptequal to the highest mass involved in the signal process (mH or mA) topreserve thedegeneracym²_H_±≈^m²H/A [17],denotingwithmH/A the massofthe scalarH or the massofthe pseudoscalarA. The value ofthem12 parameter,thesoftZ2 symmetrybreakingmass, was set to m²₁₂=^m²_H±tanβ/(1+^tan²β), accordingto the mini- malsupersymmetricstandardmodel(MSSM)parametrisation[11].

The value ofthecomplex couplingsλ6 andλ7 in thisparametri- sation are set to zero, in order to avoid tree-level CP violation.

The productioncross sections, used for the normalisation of the signal samples, are computed using the SusHi 1.4 program [32], which provides next-to-next-to-leading-order (NNLO) predictions.

The branching fractionfor the heavy andlight Higgs bosons are obtainedusing the 2hdmc 1.6program [33], followingtheguide- linesinRefs.[34,35].

The signal benchmark where the light boson decays into

τ τ

is simulated for values of mH/A andmA/H varying in the ranges 200–1000 and 15–900 GeV, respectively, with the constraint mH/A>mA/H+^mZ.Forthebb analysisthelowerboundforthe invariant massmA/H goesdownto10GeV. TheregionwheremH issmallerthanm_hisnotpertinentinthismodel.

4. Eventreconstructionandselection

Event reconstruction is based on the particle-ﬂow algorithm [36,37], which exploits information from all the CMS subdetec- tors to identify andreconstruct individual particles in theevent:

muons,electrons,photons,chargedandneutralhadrons.Suchpar- ticles are algorithmically combined to form the jets, the

τ

h candidates, the missing transverse momentum p^miss_T , deﬁnedas the projectionontheplaneperpendiculartothebeamsofthenegative vector sumoftheparticles momentaandits magnitude, denoted as E^miss_T . To minimisethe contributions from pileup interactions, charged tracks are requiredto originate fromthe primary vertex (reconstructed using the deterministic annealing algorithm [38]),

(3)

whichistheonecharacterisedbythelargestp²_Tsumofitsassoci- atedtracks.

Electronsare identified by combininginformation fromtracks and ECAL clusters, including energy depositions from final-state radiation[21].Muonsareidentifiedthroughacombinedfittopo- sition measurements from both the inner tracker and the muon detectors[22].The

τ

hobjectsareidentiﬁedandreconstructedus- ing the “hadron-plus-strips” algorithm [39], which uses charged hadrons and photons to reconstruct the main hadronic decay modesofthe

τ

lepton: onechargedhadron, one chargedhadron and photons, and three charged hadrons. Electrons and muons can be misidentified as hadronic taus if produced in jets or if close-byactivityfrompile-uporbremsstrahlungispresent.These misidentifications are suppressed using dedicated criteria based ontheconsistencybetweenthe measurementsinthetracker,the calorimeters,andthemuondetector[39].Torejectnonpromptor misidentified leptons, requirements are imposed on the isolation criteria,basedonthesumofdepositedenergies.Theabsolutelep- ton isolation I_abs is definedby the scalar sum of the p_T of the charged particles from the primary vertex, neutral hadrons, and photonsinan isolationcone ofsize R=√

(

η

)²+ (φ)²=⁰.4 (R=⁰.3 for electrons),centred around the leptondirection. To reduce theeffectfrompileup, theenergydepositreleased inthe isolation cone by charged particles not associated with the primary vertex is subtracted from the neutral particles pT scalar sum. For electrons and muons the relative isolation, deﬁned as Irel=^Iabs/pT,isused.

Jetsare clusteredusingtheanti-k_T algorithm[40],withadis- tance parameter of 0.5, as implemented in the Fastjet software package [41]. Charged particles not associated with the primary vertexare excluded by means of the charged-hadronsubtraction technique [42]. Theremaining energy originatingfrompileupin- teractions, includingthe neutralcomponents,is subtracted based onthe medianenergydensityin the detectorcomputedthrough theeffectivejet areatechnique [43].The identiﬁcationofb quark initiated jetsisachieved throughthe combined secondaryvertex (CSV) algorithm [44], which exploits observables related to the longlifetimeofBhadrons.

4.1.SelectionforZ→

Inselecting bb and

τ τ

events, theleptons fromZ boson decayare requiredto be well within the CMStrigger anddetec- tor acceptance of pT>20 GeV and |

η

_|<2.5 for electrons, and pT>20 GeV, |

η

_|<2.4 for muons. Muon momentum-scale [22]

and electron energy corrections [21] are applied to recover the globalshiftofthescaleobservedbetweendataandsimulation.The requirementontherelativeisolationfortheleptonsissetto I_rel<

0.15forelectronsandIrel<0.2formuonsinselectingbb events.

For the leptons from the Z boson, in the case of

τ τ

events, the required relative isolation is I_rel< 0.3. The presence of two reconstructed same-ﬂavour, oppositely charged lepton candidates forming a pair with invariant mass in the range of 76–106 GeV isrequired to suppress contamination of non-resonant Drell–Yan + jets and tt processes. In events where multiple Z candidates arepresent,theleptonpairwiththeinvariantmassclosesttothe nominalZbosonmass[45]ischosen.

4.2.Eventselectionforbb

For the bb search, the jets are selected to be in the kine- maticregionpT>30GeV and|

η

_|<2.4.AtleasttwoCSVb-tagged jetsare requiredto bepresentintheevent, toreduce thecontri- bution of Z+ light-parton jets(originating from gluons oru, d,

or s quarks) events. The threshold on the b tagging discrimina- torcorresponds toa btaggingeﬃciencygreater than65% andto amisidentiﬁcationprobabilityforlight-partonjetsof1%[44].The twob-taggedjetswithhighestvaluesoftheCSVdiscriminantare consideredascandidatedecayproductsofthenewlightresonance.

The E^miss_T signiﬁcance[46,47],representinga

χ

²differencebe- tweenthe observedresultfor E^miss_T andthe E^miss_T =0 hypothesis, is used to suppress background events originating from tt processes.Thisvariableprovidesanevent-by-eventassessmentofthe likelihood that the observedmissing transverse energyis consistent with zero giventhe reconstructed content of the eventand known measurement resolutions. This variable is a stronger dis- criminantagainsttt backgroundthanE^miss_T aloneandalsoprovides smaller systematic uncertainties. The distribution of the tt com- ponentmotivatesthe requirementonthe E^miss_T signiﬁcancetobe smallerthan 10.

4.3. Eventselectionfor

τ τ

To increase thesignal sensitivityin thehigh

τ τ

mass region, the

τ τ

event selection includes the requirement of a trans- versely boosted Z boson (pT>20 GeV), together with a large (>1.5 rad) azimuthal anglebetween the Z boson ﬂight direction and p^miss_T ,particularlyeffectivein suppressingtheZ+ jetsbackground.Inadditiontothetwolightleptonsrequiredtoreconstruct theZboson,twoadditionaloppositelychargedanddifferent-ﬂavor leptons (e,

μ

, and

τ

h) are used to reconstructthe Aor H boson candidate. The requirements on the pseudorapidity forlight lep- tonsarethesameasfortheZdecayleptons,withthe pTthresh- old lowered to 10 GeV. The

τ

h candidates are required to have pT>20GeV and|

η

|<2.3.Therelativeisolationforelectronsand muons, andtheabsoluteisolationfor

τ

leptons arerequiredtobe smallerthan0.3and2GeV,respectively.SincetheZ+ ^jets^backgroundis characterisedby a softer lepton transversemomentum spectrum than thesignal one, this backgroundis reducedby se- lecting events with high LT, where LT indicates the scalar sum ofthe visible p_T ofthe decayproducts froma

τ τ

pair. Both the isolation requirements andthe value ofthe LT thresholdare de- terminedasa resultofanoptimisationprocedurethatmaximises the expectedsigniﬁcance of thesearched signal. The optimal requirementonthe LT quantity isfound byscanning thethreshold between20and200GeV,atintervalsof20GeV.

Jetsare requiredtohave p_T>30GeV and |

η

|<4.7.Toreduce thelargett background,alleventswithatleastonejetwithpT>

20GeV and|

η

_|<2.4,reconstructed asa jetoriginatingfroma b quark according to the output of the CSV discriminator used for tagging,arevetoed.

To calculate the

τ τ

invariant mass, the secondary-vertex ﬁt algorithm(svﬁt)[48]isused,alikelihood-basedmethodthatcom- binesthereconstructed p^miss_T anditsresolutionwiththemomen- tumofthevisible

τ

decayproducts toobtainanestimatorofthe massoftheparentparticle.

5. Modellingofthebackground 5.1. Thebb channel

The relevant sources of background for the bb ﬁnal state originate fromZ+ ^jets^processes,^{tt and} ^tWproduction,diboson production,andvectorbosonproductioninassociationwithaSM Higgsboson.ThecontributionsofZ+ ^jets^andtt backgroundsare measuredby meansofadata-basedmethod,thedibosonandtW backgroundsare normalisedto theCMSmeasurements.Forthese backgrounds, the shapes are taken from MC, while the normalisations are extracted fromdata. The vector boson production in

(4)

associationwithaSMHiggsbosonisnormalisedtothetheoretical prediction.

The comparisonof data andpredictions after the selection of events for the bb final state shows the importance of an ac- curate theoretical calculation of the Z + ^jets^production ^rate. În particular, in the 400–700 GeV range of the mbb distribution, thedataisfound toexceedtheLOpredictionbyup totwo stan- darddeviations,depending ontheconsideredmass.Thisexcessis nolongersignificant whenNLO QCDcorrections, asimplemented in amc@nlo [49],are included in the modellingof the Z + ^jets process.For thisreason, theLOpredictions arecorrected usinga reweightingtechnique,inordertoaccountforNLOQCDeffects.To thisend,itbecomesnecessarytoapplythereweightingaccording totheparton(orhadron)flavourofthejetsinthegeneratedevent.

The ratio NLO/LO of the light- and heavy-flavour components of themjj distributioniseach fittedwitha third-orderpolynomial and a separate reweighting of the shape of the light and heavy flavour components ofmjj isapplied, resulting inbetter agree- mentwiththedata.

Todetermine the Z+ ^jets^andtt normalisation, a data-based methodisexploited.Data-derivedcorrectionfactorsforsimulation areobtainedafteranadditionalcategorisationoftheZ+^jet^back- groundevents,basedontheﬂavour(bjetornot)andmultiplicity (exactly twojetsor threeormore jets)ofthe reconstructedjets.

Thesecategoriesare sensitive to NLO effectsrelatedto the mod- ellingofextra jets[50].Scale factors(SFs)are introducedforthe tt backgroundandthelightandheavyflavourcomponentsofZ+ jetsbackground.Theseareleft free tofloatin atwo-dimensional fitofthedistributionspredictedbythesimulationtothedata.The distributions used as input are the product of the CSV discrimi- nantsofthetwoselectedjets,andtheinvariantmassofthelepton pair from the Z boson decay in the range 60<m<120 GeV.

The first observable is sensitive to the contribution from non-b jets,whereasthesecondoneissensitivetothecontributionofthe tt productionprocess.Thefitisperformedsimultaneouslyinfour differentcategories:electrons, muons, exactlytwo jets,andmore than two jets. The SF forthe tt is foundto be very close tothe unity,whilefortheZ+ ^jets^process^the^SFs^depart^fromûnity^by asmuchas1.3forthelightflavourcomponent.

The overall yields from diboson and tW processes are nor- malisedtotheCMSmeasurements[51–54].Theassociatedproduc- tionofaZbosontogetherwiththeHiggs-likescalarboson(Zh)is alsoaccountedforasbackground,andnormalisedtotheexpected theoreticalcrosssection[55].

5.2. The

τ τ

channel

Methods basedon both data andsimulationare used to esti- matetheresidualbackgroundaftereventselection.Normalisations andmassdistributionsintheZZ,Zh,aswellasfortheminorfully leptonicWWZ,WZZ,ZZZandttZ backgroundsareestimatedfrom simulation. The Z + ^jets ^and^WZ + ^jetscontributions are mea- suredbymeansofadata-basedmethod.

Production of Z + ^jets ând ^WZ + ^jets constitutes the main source of background when at least one lepton is misidentified. Misidentified light leptons arise from semileptonic decays of heavy-flavour quarks,decays in flight of hadrons, and photon conversions, while jetsoriginating fromquarks orgluons can be misidentified as

τ

h. Backgrounds with at leastone misidentified lepton are estimated from control samples in data starting from the estimation of the lepton misidentification probabilities. The lepton misidentification probability is defined as the probability that a genuine jet, satisfying loose lepton identification criteria (whichrefertotheso-called“loose”lepton),alsopassestheidenti- ficationcriteriarequiredforaleptoncandidateinthesignalregion

(so-called“tight”lepton).Thisprobabilityismeasuredforeachlep- ton ﬂavour using a datasample where a Zcandidateis selected, andanadditionalsinglelepton(electron, muon,or

τ

h)passesthe loose identification requirements. Counting the fraction of such loose leptons that also pass the tight lepton identification criteria inthe p_T bins ofthe reconstructedjet closest, in R, to the loose lepton,yields themisidentification probability f as a func- tionofpT.Thecontributionfromgenuineleptonsarisingfromthe WZ andZZproduction aresubtracted. Once themisidentification probabilitiesarecomputed,threecontrolregions(C R)aredefined withaZcandidateandtwo opposite-signleptons,asfollows:the C R₀₀ wherein both leptons pass loose identification criteria but not the tightones; C R10 region,whereinone lepton passestight identification requirements,the other onlyloose criteria, andthe looseleptonisthe

τ

hwithlower p_T inthe

τ

h

τ

hchannel,thelight lepton in the

τ

_h channels, andthe electron in the e

μ

channel;

the C R01 region,whichis similartoC R10 butthelooseleptonis the

τ

hwithhigherp_Tinthe

τ

h

τ

hchannel,the

τ

hinthe

τ

hchan- nels,andthemuoninthee

μ

channel.TheestimatedN_misIDofthe background withatleastone misidentiﬁed lepton froma pairof closest-jetp_Tbinsisgivenby:

N_misID

=

^N10

f₁

1

−

^f1

+

^N01

f₂

1

−

^f2

−

^N00

f₁f₂

(

1

−

^f1

)(

1

−

^f2

) ,

(1) where N00, N01, and N10 denotethe numberofevents fromthe C R₀₀, C R₀₁, and C R₁₀ control regions, respectively, with closest jetsintheconsideredpTbins,and f1 and f2 indicatethemisiden- tiﬁcationprobabilitiesassociatedwiththetwodifferentﬂavor(ex- cept for the

τ

h

τ

h final state) loose leptons in the p_T bins. The expressioninEq.(1)takesintoaccountboththebackgroundwith twomisidentifiedleptons(mostlyfromZ+^jets)ând^that^fromônly onemisidentifiedlepton(primarilyfromWZ+^jets).

Thecontaminationfromgenuineleptonsinthecontrolregions from the SM Zh, WWZ, WZZ,ZZZ, ttZ,and ZZprocesses is estimated fromsimulation, and subtracted from N₀₀, N₀₁, and N₁₀. Thetotalbackgroundinthesignalregionisobtainedbysumming thecontributionsfromallpairsofpT bins.

6. Systematicuncertainties

Thesystematicuncertaintiesarereportedinthefollowingpara- graphsandsummarisedinTable 1.

TheuncertaintyontheintegratedluminosityrecordedbyCMS isestimatedtobe2.6%[56].

The systematic uncertainties associated with the lepton eﬃ- ciencySFs,usedtocorrectthesimulationandderivedfromstudies atthe Zpeak usingthe tag-and-probe (T&P)method[22,21], are approximately1%formuonsand2%forelectrons, andaffectboth signalandbackgroundprocessesinthesameway.Also,theuncertainties onthedoublemuon anddoubleelectrontrigger eﬃcien- ciesareevaluatedtobe1%fromsimilarstudiesattheZpeak[24].

The uncertainty on the jet energy scale is derived from the method of Ref. [57] and the parameters describing the shape of theenergydistributionarevariedbyone standarddeviation(SD).

The effectis estimatedseparately on the backgroundandon the signal, resulting ina 3–5% variation,depending onthe p_T and

η

of the jets. The uncertainty on signal and backgroundyields in- ducedby theimperfectknowledge ofthejet energyresolutionis estimatedtobe3%[57].

Theuncertainties affectingbtaggingeﬃcienciesare pT-depen- dent,andvaryfrom3%to12%(forpT>30GeV)[44].Theimpact oftheseuncertaintiesonthenormalisationofsignalis5%forback- groundand4–6%forsignalinthebb analysis,andabout1%in the

τ τ

analysis.Theuncertaintyinthemistaggingrateisfound tohaveanegligibleimpact.

(5)

Table 1

Summaryofsystematicuncertaintiesforbothτ τ andbb ﬁnalstates.

Source Uncertainty [%]

H→^ZA→ ^bb ^H→^ZA→ τ τ

Luminosity 2.6 2.6

Lepton identiﬁcation/isolation/scale 1–2 1–2

Lepton trigger eﬃciency 1 1

Jet energy scale 3–5 3–5

Jet energy resolution 3 3

b-tagging and mistag eﬃciency 4–6 1

Signal modelling (PDF, scale) 5–6 5–6

Background norm. (Z Z ) 11 11

Background norm. (Z+jets and tt) <8 –

Background norm. (tW, WW, WZ and Zh) 8–23 –

Z+jet background modelling 5–30 –

Signal eﬃciency extrapolation 3–50 –

Tau identiﬁcation/isolation – 6

Tau energy scale – 3

Reducible background estimate – 40

Thesystematicuncertaintyonthesignal isevaluated byvary- ing the set of parton distribution functions (PDFs) according to thePDF4LHCprescriptions[58–60]andthefactorisationandrenor- malisationscalesby varyingtheir valuesby afactoronehalf and two.Aneffectof5–6% isestimatedfortheentiremassrangefor both

τ τ

andbb ﬁnalstates.This uncertaintyisestimatedby propagatingthesevariationsthroughthesignalsimulationandre- constructionsequenceandthusaccountsforuncertaintiesrelated tobothsignalcrosssectionandacceptance.

Finally,an 11%uncertaintyisassignedtotheZZnormalisation fromthecrosssectionmeasuredbyCMS[51].

Forthe bb final state, the uncertainty on the SFs used for normalisationofZ+^jetsândtt backgrounds isderivedfromthe statisticaluncertainty resulting fromthe fit used to derive these SFsanditisestimatedtobe<8%.Anadditionalsystematicuncer- taintyassociatedwiththem_bbspectrumcorrection,describedin Sec.5,rangesfrom5%form_bbbelow700GeV to30%formasses attheTeVscale.An uncertaintyof8% isassignedtothenormal- isationof theWWprocess, corresponding to theuncertainties in thecrosssection measured by CMS[52].Asimilar uncertaintyis assignedalso tothe WZ process, whichshares the same sources ofuncertainties inthecross section measurement.Forthe minor tWbackground,theuncertaintyisestimatedas23%,alsobasedon themeasured cross section [54].A 7% uncertainty is assignedto theZhprocess,reflecting theuncertaintyon thetheoreticalcross section [55]. Given the small cross section for this SM process compared to other background processes,its contribution to the background normalisation uncertainty has been calculated to be lessthan 1% andis thus considered negligible.In order tointer- polate smoothly thesignal efficiencyacross the parameter space, additionalmasspointsforthebb finalstateareprocessedusing a parametric simulation [61], tuned for delivering a realistic ap- proximationofthe CMSresponse inthe reconstructionofphysics objects usedin this search. Forthis reason, an additional source ofuncertaintyisintroducedfortheSFappliedtothesesamplesto reproducetheefficiencymeasuredwiththefullsimulation.Thisis measuredforthedifferentsignal pointsinthem_H–m_A plane and itiscloseto3%inmostofthephasespace,butrisesto50%atthe boundariesofthesensitivityregion.

In the

τ τ

ﬁnal state, the uncertainty of 6% [39] in the

τ

h identiﬁcationeﬃciency, whichhas been determined usinga T&P method,hasbeentakenintoaccount.The

τ

h energyscale uncer- taintyiswithin3%[39]andonlyaffectstheshapesofthe

τ τ

mass distributions. The systematicuncertainties estimated fore,

μ

,

τ

h andjetenergyscalesarepropagatedto ^p^miss_T ând^to^the^mass^distributions.Thepropagationto^p^miss_T învolvesâ^sumôf^theênergies

of each objectﬁrst and a consequent subtractionof such contri- butionsoncethenominalenergyscales(orresolutions)arevaried up anddown byone SD(fore,

μ

,

τ

h,andjets).Oneofthe main systematic uncertainties is related to the nonprompt background estimation. Thisuncertainty hasbeen evaluated using simulation bycomparingthedirectestimateofthebackgroundswiththatob- tained using the procedure adopted in the analysis, but applied to simulated events.The discrepancybetween the two estimates neverexceeds40%.Thisvalueisthusconsideredastheuncertainty ontheestimatesofthereduciblebackgroundyieldforallchannels andall LTthresholds.

7. Results

The analysissearchesfornewresonancedecays bycomparing data to simulation in the two-dimensional plane defined by the four-body (m_bb ormτ τ )andtwo-body(m_bb ormττ )invariant masses. The numerical valuesforthe upper limitsorthe signifi- canceofalocalexcessareobtainedusingtheasymptoticmethod describedinRef.[62].TheCLSmethod[63,64]isusedtodetermine the95% confidencelevel(CL)upperlimitsonthe excludedsignal crosssection.Forbothfinalstates,thelimitsinthelower-righttri- angleofthemassplane,whichcorrespondstotheprocessA→^ZH, are obtainedby mirroring the results obtained in the upper-left triangle,sincethe signal efficienciesforH→^{ZA and}Â→^{ZH are} equalforthesamemassesoftheheavyandlightHiggsbosonsin thetwoprocesses.

7.1. Thebb channel

Forthe bb final state, resultsare obtainedusinga counting approach,which canbe reinterpreted inother theoretical models withthesamefinal state.Resultsarereportedinbinsofmbband m_bbmasses,intherangefrom10GeV to1TeV form_bb,andfrom 140GeV to1TeV form_bb.Todefinethepropergranularityofthe binning,a studyisperformedusingsignal benchmarkpoints and evaluatingthewidthofthem_bbandm_bbpeaksintheconsidered mass range.Theaverage reconstructed width, definedasone SD, form_bbandm_bbisfoundtobeapproximately15%oftheconsid- eredmass.Thebinwidthshavebeenchosentobe±^{1.5 SD}âround eachconsideredmasspoint.

Theefficiency,definedasthefractionofgeneratedsignalevents reconstructed after the final selection, is calculated withthe full CMS simulationand reconstruction softwareat 13representative signalpoints inthemH–mA massplane.Thesignalefficienciesfor the rest ofthe plane are obtained by interpolating theratio be- tweenthefullsimulationandtheparametricsimulation(typically

(6)

Fig. 1. Observed95%CLupperlimitsonσH/A→^ZA/H→^bb asafunctionofmAand mH.

0.9),calculated ineach of the13 signal masspoints, andscaling theeﬃcienciescalculatedusingtheparametricsimulationby this interpolatedratio.Theresultingsignaleﬃciencyrangesfrom8%at (mA,mH)= (¹⁰⁰,300)GeV to13%at(300,600)GeV.

Fig. 1 showsthe observedupper limitsonthe product ofthe crosssection(

σ

)andbranchingfraction(B⁾^for^thebb ﬁnalstate in them_H–m_A plane. The achieved sensitivityprovides an exclu- sionlimitat95% CLofapproximately10 fbforalargefractionof thetwo-dimensionalmass plane.Inparticular,the observedlimit ranges fromjustabove 1 fbformH close to 1TeV to 100 fb for m_H<300GeV.Thevalidityoftheseresultsisapplicableto mod- elsallowing theexistenceofboth AandHbosons witha natural widthsmallerthan15%oftheirmasses.

Two moderate excesses are observed for the bb channel in the regions around (m_bb,m_bb)= (⁹⁵,285) GeV and (575,660)GeV.Accordingtotheproceduredescribed atRef.[65], theyhavelocalsigniﬁcancesof2.6and2.85 SDrespectively,which become globally 1.6 and 1.9 SD, once accounting for the look- elsewhere effect [66]. The low-mass excess is more compatible with the signal hypothesis, both in terms of yield and width.

The reconstructed invariant mass distributions for the bb and

bb systems, in the regions around this excess,are reported in Fig. 2andcomparedwiththeexpectationsfrombackgroundpro- cesses. A 2HDM type-IIbenchmark signal at m_H=²⁷⁰ ^{GeV and} mA=¹⁰⁴^GeV,^normalised ^to ^theNNLO SusHi prediction, is also superimposed.

7.2. The

τ τ

channel

Inthecontextofthe

τ τ

analysis,asearchbasedonthemττ distribution is performed. For every considered pair of H and A mass values, the search is performed in eight

τ τ

^svfit binned mass distributions, each corresponding to one of the eight considered final states. Variable bin widths are adopted in order to account for themass resolution. A simultaneous likelihoodfit to the observed distributions is performed with the expected distributions from the background-only and signal plus background hypotheses.The normalisation of thesignal distributionis a free parameterinthefit.Nosignificantdeviationsareobservedindata fromthe SM expectation. The svfit massdistributions ofthe

τ τ

pairintheeightdifferentﬁnalstatesareshowninFig. 3.Thecho- sensignalcorresponds tomH=³⁵⁰^{GeV and}^mA=⁹⁰^GeV,^which is the one closest to the centreof the bin in which the highest

Fig. 2. (Top)Thembbspectrumforeventsselectedinthe222<mbb<350GeV re- gionfordataandsimulationandtherelativeratio.(Bottom)Thembbspectrumfor eventsselectedinsidethe72<mbb<114GeV regionfordataandsimulationand therelativeratio.ThesignalcorrespondingtomH=²⁷⁰^{GeV and}^mA=¹⁰⁴^GeV, normalisedtotheNNLO SusHi crosssection,issuperimposedfortanβ=¹.5 and cos(β−α)=⁰.01 inthe2HDMtype-IIscenario.Theoverallsystematicuncertainties inthesimulationarereportedasahatchedband.

excess isobservedinthe bb channel.The shownshapescorre- spondto LT >40GeV fore

μ

,LT>60GeV for e

τ

hand

μτ

h,and LT>80GeV for

τ

h

τ

h.

Fig. 4 showsthe limit on

σ

B ^for ^the

τ τ

ﬁnal state in the mH–mA plane.Signalcrosssectionsofabout5–10 fbareexcluded in mostofthe mH–mA plane (500<mH/A<1000 GeV and90<

m_A_/_H<400GeV).

(7)

Fig. 3. svfit massdistributionsfordifferentfinalstatesoftheH→^ZA→ τ τprocess,wheretheZ bosondecaystoee (rightcolumn)andμμ(leftcolumn).Theexpected signalcorrespondingtomH=³⁵⁰^{GeV and}^mA=⁹⁰^GeV,^whose^cross^section^times^branching^fractionîs^normalised^to^theNNLO SusHi prediction,issuperimposedfor tanβ=¹.5 andcos(β−α)=⁰.01 inthe2HDMtype-IIscenario.Onlystatisticaluncertaintiesarereportedasahatchedband.

(8)

Fig. 4. Observed95%CLupperlimitsonσH/A→ZA/H→τ τ asafunctionofmAand mH.

7.3. Combinationinthecontextof2HDM

Observed and expected upper limits on the signal cross sec- tionmodiﬁer

μ

=

σ

95%/

σ

_tharealsoderivedandreportedinFig. 5, where

σ

th isthe theorycrosssection ofthe2HDMsignal bench- markusedinthisanalysis.Theresultsareobtainedfromthecom- binationofthebb and

τ τ

ﬁnalstates.Thissearchisnotable toexcludethehigh-massregionswheremA>300GeV andmH>

300GeV, dueto the drop in thesignal cross section, wherethe A/H→^{tt channel}ôpensûp ^for^mA/H>2mt, wheremt isthetop quarkmass[18].Furthermore,intheregionwherehighly-boosted topologiesstartcontributing(mH≈¹⁰^mA),thesensitivityislower relative to the rest of the plane, primarily caused by the ineffi- ciency in reconstructing signal decay products insuch a regime.

Still,asigniﬁcantportionofthebenchmarkmassesisexcludedfor a2HDMtype-IIscenario withtanβ=¹.5 and cos(β−

α

)=⁰.01, delimitedbythesolid contourinFig. 5.Theobserved95% CLex- clusion region is localised in the range mH=200–700 GeV and mA=20–270 GeV forthedecayH→^ZA,^and^similarlyⁱⁿ^the^range m_A=200–700 GeV andm_H=120–270 GeV fortheA→^{ZH decay.}

Thefeature observedintheexclusionlimit forthe regionaround (mA,mH)= (^75–100,200–300) GeV is the result of an interplay between the larger Z + ^jets ^background ^yields ^expected ⁱⁿ ^this regionandthe quicklyevolving signal crosssection.The effectis visibleintheexpectedlimitsandbecomesslightlybroaderinthe observed onesgiventhe concurrent presence inthe sameregion of a moderatedata excess. The region where |^mH−^mA|<mZ is kinematicallyinaccessible.

The limits on

μ

can be also visualised as a function of the 2HDMparameterstanβ andcos(β−

α

)foragivenpairofmAand mH,fromthe combinationof bb and

τ τ

ﬁnal states.Results are given in Fig. 6, where the exclusion limits on the parame- tersareshownform_H=³⁷⁸^{GeV and}^mA=¹⁸⁸^GeV,^a^mass^pair chosen tobe within theexclusionregion ofFig. 5.The area con- tainedwithin the solid lineshowsthe parameter spaceexcluded for the chosen mass pair, where tanβ lies between 0.5 and 2.3 andcos(β−

α

)between−⁰.7 and0.3.

8. Summary

Thepaperdescribesthe ﬁrstCMSsearch foranewresonance decaying into a lighter resonance and a Z boson. Two searches havebeenperformed,targetingthedecayofthelighterresonance

Fig. 5. Observedlimitsonthesignalstrengthμ=σ95%/σthforthe2HDMbench- mark,aftercombiningresultsfrombb andτ τ ﬁnalstates.Thecrosssections arenormalisedtothe NNLO SusHi prediction,for a2HDM type-IIscenariowith tanβ=¹.5 andcos(β−α)=⁰.01.Thedashedcontourshowstheregionexpected tobeexcluded.Thesolidcontourshowstheregionexcludedbythedata.

Fig. 6. Observedlimitsonthesignalstrengthμ=σ95%/σthforthe2HDMbench- markaftercombiningresultsfrombb andτ τ ﬁnalstates.Thecrosssectionsare normalisedtotheNNLO SusHi prediction.Limitsareshowninthe2HDMparame- terscos(β−α)andtanβforthesignalmassesofmH=³⁷⁸^{GeV and}^mA=¹⁸⁸^GeV.

Thedashedcontourshowstheregionexpectedtobeexcluded.Thesolidcontour showstheregionexcludedbythedata.

into either a pair of oppositely charged

τ

leptons or a bb pair.

TheZbosonisidentiﬁedviaitsdecaystoelectronsormuons.The searchisbasedondatacorrespondingtoanintegratedluminosity of19.8 fb⁻¹inproton–protoncollisions at√

s=⁸^TeV.^Deviations fromtheSM expectationsare observedwithaglobalsigniﬁcance oflessthan2SDandupperlimitsontheproductofcrosssection andbranchingfractionareset.The searchexcludes

σ

B ^as^low^as 5 fband1 fbforthebb and

τ τ

ﬁnal states,respectively, de- pendingonthelightandheavyresonancemassvalues.

Limitsarealsosetonthemassparametersofthetype-II2HDM modelthat predicts theprocessesH→^{ZA and}^A→^ZH,^where^H andAareCP-evenandCP-oddscalarbosons,respectively.Combin- ingthebb and

τ τ

ﬁnalstates,thespeciﬁcmodelcorrespond- ing tothe parameterchoice cos(β−

α

)=⁰.01 andtanβ=¹.5 is excluded form_H inthe range200–700 GeV andm_A in therange