Search for supersymmetry in events with soft leptons, low jet multiplicity, and missing transverse energy in proton-proton collisions at root s=8 TeV

Contents lists available atScienceDirect

Physics

Letters

B

www.elsevier.com/locate/physletb

Search

for

supersymmetry

in

events

with

soft

leptons,

low

jet

multiplicity,

and

missing

transverse

energy

in

proton–proton

collisions

at

√

s

=

8 TeV

.

CMS

Collaboration



CERN,Switzerland

a

r

t

i

c

l

e

i

n

f

o

a

b

s

t

r

a

c

t

Articlehistory: Received25December2015 Receivedinrevisedform7April2016 Accepted11May2016

Availableonline16May2016 Editor:M.Doser

Keywords: CMS Physics Supersymmetry

Results arepresentedfromasearchforsupersymmetricparticlesinscenarioswithsmallmasssplittings. Thedatasamplecorrespondsto19.7 fb−1ofproton–protoncollisionsrecordedbytheCMSexperiment at √s=8 TeV. The search targets top squark (t) pair productionin scenarios with mass differences

m₌m(t)−m(

χ

0

1) below the W-boson mass and with top-squark decays in the four-body mode

(t_→b

ν



χ

0

1), where the neutralino(

χ

10) isassumedto bethe lightest supersymmetricparticle(LSP).

Thesignatureincludesahightransversemomentum(pT)jetassociatedwithinitial-stateradiation,one

ortwolow-pTleptons,andsignificantmissingtransverseenergy.Theeventyieldsobservedindataare

consistentwith theexpectedbackground contributionsfromstandard model processes.Limitsare set onthecrosssectionfortop squarkpairproductionasafunctionofthet and LSPmasses.Assuminga 100%branchingfractionforthefour-bodydecaymode,top-squarkmassesbelow316 GeVareexcluded form=25 GeV at95% CL.Thedileptondataarealsointerpretedunder theassumptionofchargino– neutralinoproduction,withsubsequentdecaystosleptonsorsneutrinos.Assumingadifferencebetween thecommon

χ

1+/

χ

20massandtheLSPmassof20 GeVanda

τ

-enricheddecayscenario,massesinthe

rangem(

χ

1+)<307 GeV areexcludedat95%CL.

©2016TheAuthor.PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense (http://creativecommons.org/licenses/by/4.0/).FundedbySCOAP3_.

1. Introduction

The main objectives of the CERN LHC programme include searches for new physics, in particular supersymmetry (SUSY)

[1–5],oneofthemostpromisingextensionsofthestandardmodel (SM) of particle physics. Supersymmetric models can offer solu-tionstoseveralshortcomingsoftheSM,inparticularthoserelated tothemasshierarchyofelementaryparticles

[6,7]

andtothe pres-enceofdarkmatterintheuniverse.

Supersymmetrypredicts superpartners ofSM particles (sparti-cles)whosespinsdifferbyone-halfunit withrespecttotheir SM partners. In SUSY models with R-parity [8] conservation, sparti-clesarepair-producedandtheirdecaychainsendwiththelightest supersymmetricparticle(LSP).Inmanyofthesemodelsthe light-est neutralino (



χ

0

1) takes the role of the LSP and, beingneutral

andweakly interacting,would matchthe characteristics required ofadarkmattercandidate.TheLSPswouldremainundetectedand

 _{E-mailaddress:cms-publication-committee-chair@cern.ch.}

yieldacharacteristicsignatureofhighmissingtransverse momen-tum,themagnitudeofwhichisreferredtoasE_Tmiss.

In this paper we investigate the production of supersymmet-ricparticlesinascenarioinwhichthemasssplittingbetweenthe next-to-lightest SUSY particle(NLSP) andtheLSP is small, which isreferred toascompressedSUSY. Inthiscase,the eventswould escape classical search strategies because of the low transverse momenta (pT) of the decay products of the NLSP. Signal events

canstillbe distinguishedfromSM processesifahigh-pT jetfrom

initial-state radiation (ISR) leads to a boost of the sparticle pair systemandenhancestheamountof Emiss

T ,whiletheother decay

products typically remain soft. In the signal scenarios studied in thispaper,SUSYparticlescandecayleptonically,andthepresence oflow-pTleptonscanbeusedtodiscriminatefurtheragainst

oth-erwisedominantSMbackgrounds,suchasmultijetproductionand Z

+

jets eventswithinvisibleZ bosondecays.

SUSY models withlight top squarks (



t) arewell motivatedas they control the dominant correction to the Higgs boson mass andtherebypreserve“naturalness”

[6,7,9–14]

.SUSYscenarioswith masssplittingsof15–30 GeVbetweenthetopsquarkandtheLSP

http://dx.doi.org/10.1016/j.physletb.2016.05.033

0370-2693/©2016TheAuthor.PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense(http://creativecommons.org/licenses/by/4.0/).Fundedby SCOAP3_.

Fig. 1. Signalmodelsfortopsquarkpairproductionwithsubsequentfour-body de-cays(left),andchargino–neutralinopairproductionwithdecaysviasleptonsand sneutrinos(right).Antiparticlelabelsaresuppressed.TheISRjetusedinthe analy-sisisnotshowninthesediagrams.

are especiallyinteresting because they wouldlead, through



t–

χ



0 1

co-annihilation, to the observed cosmologicalabundance ofdark matter [15]. For mass differences below the W-boson mass, top squarkscouldundergoeitheratwo-bodydecay(suchas



t

→

c

χ



0

1)

or a four-body decay (



t

→

bff



χ

0

1, where ff representsa pair of

quarks or leptons), as shown in the left panel of Fig. 1, with branching fractionsand kinematicpropertiesthat depend on de-tailsofthemodel

[16,17]

.Thesearchstrategy basedonthe pres-enceofanISRjethasbeenusedtosearchforthetwo-bodydecay inamonojettopologybytheCMSCollaboration

[18]

,andforboth decaymodesbytheATLASCollaboration[19–21].Inthispaperwe assumethatotherSUSYparticlesaredecoupledandthatthe four-bodydecayproceedsexclusivelyviavirtualSMparticles.

Finalstates with a hard ISR jet, high Emiss

T , and one or more

charged leptons can also occur in the production of chargino– neutralinopairs incompressedSUSYmodels

[22–24]

.Amodelof pair-production of the lightest chargino (

χ



₁+) with the second-lightest neutralino (



χ

0

2) is shown in Fig. 1 (right). Decay chains

couldproceedviaintermediatesleptonsorsneutrinosandgiverise to final states with one or three charged leptons. In this model,



χ

₁+and

χ



0

2 areassumedtobealmostdegenerateandareassigned

a common mass m

(



χ

)

. In general, the same signature can arise fromtheproductionofheavyparticleswhosedecaychainscontain undetected,slightlylighterparticlesplusleptons.PreviousLHC re-sultsforthemodelofelectroweakproductiondescribedaboveand masssplittings belowm

(

Z

)

canbe found inRefs. [25–29], where thelast two referencesalso reportan alternative approachbased onthevector-bosonfusiontopology.

In thispaper we describe a search forpair production oftop squarks withsubsequent four-body decays via virtual top quarks andW bosonsineventswithahigh-pT jet, EmissT ,andoneortwo

softleptons,corresponding tosignaleventswithaleptonic decay ofat leastone of the virtual W bosons. The single-lepton topol-ogyoffers thesecond-highest branching fractionafter the purely hadronic mode. In this channel we consider only muons, which canbeefficientlyreconstructedandidentifiedwithtransverse mo-menta as low as 5 GeV. Forthe dilepton topology we requirea second lepton (electronor muon) of opposite charge. The single anddouble electron final states are not used because they have reduced sensitivity compared to the muon channels due to the higher pT thresholds required forelectrons. Inaddition, selected

eventsare requiredto have an energeticjet compatiblewiththe ISR signature,at mostoneadditional jetof moderatetohigh pT,

nohardleptons, anda significantamountof Emiss

T .The dominant

SM backgroundsto thissearch are pairproductionoftop quarks, W bosonorZ

/

γ

∗productioninassociationwithjets,anddiboson (VV)production.Theircontributions tothe signal region(SR)are estimatedby correctingthepredictionsfromsimulationusingthe eventyieldsobservedinseveralcontrolregions(CRs)indata.Data

are also usedto validate thisprocedure andtoderive systematic uncertainties.

Theresultsofthedileptonsearcharealsointerpretedinterms ofthemodelof



χ

₁+–

χ



₂0pairproductiondiscussedabove.Forsmall



χ

₁+

− 

χ

0

1 masssplittings, the leptons inthe final state wouldbe

softandthereforewithinthesignalregionofthedileptonsearch.

2. Detectordescriptionandeventreconstruction

The CMSdetectorhasbeendescribed indetail inRef.[30].Its central feature isa superconducting solenoidthat provides a ho-mogeneous field of 3.8 T in a volume containing a silicon pixel andstriptracker,aleadtungstatecrystalelectromagnetic calorime-ter, and a brass and scintillator hadron calorimeter. Muons are measured ingas-ionizationchambersembedded inthesteel flux-return yokesurroundingthe solenoid. The acceptanceofthe sili-contrackerandthemuonsystems extendstopseudorapiditiesof

|

η

|

<

2

.

5 and

<

2

.

4,respectively.Thebarrelandendcap calorime-terscovertherange

|

η

|

<

3

.

0 andarecomplementedbyextensive forward calorimetry. Eventsare selected forfurther analysisby a two-tier triggersystem that uses custom hardware processors to makeafastinitialselection,followedbyamoredetailedselection executedonadedicatedprocessorfarm.

The measurement of jets and Emiss

T is based on candidates

reconstructed by the particle-flow (PF) algorithm [31,32], which identifies leptons, photons, and charged and neutral hadrons by combining information from all subdetectors. The PF candidates are clusteredintojetsby usingtheanti-kT algorithm [33] witha

distance parameterof 0.5. Jetsare required tohave pT

>

30 GeV

and

|

η

| <

4

.

5,andtopassloosequalitycriteria[34] basedonthe energy fractions associated with electromagnetically or hadroni-cally interacting chargedor neutralparticles.The negative vector sumof thetransverse momenta ofthePF candidatesdefines the value of Emiss_T and the corresponding direction. Jet energies and

Emiss_T are corrected for shifts in the energy scale, contributions from additional, simultaneous proton–proton collisions (pileup), andresidualdifferencesbetweendataandsimulation

[35,36]

.Jets originatingfromb quarksareidentified(“tagged”)usingthe com-bined secondaryvertexalgorithm [37,38]ata workingpoint cor-responding to an efficiencyof about70% and a misidentification probability forlight-quarkjetsofabout1%. Hadronicdecays of

τ

leptonsareidentifiedusingthehadrons-plus-stripsalgorithm[39]. Muons and electrons are required to have pT above 5 and

7 GeV, respectively. Inthe single-muonsearch, thelepton accep-tance isrestricted to

|

η

| <

2

.

1,while inthe dileptonsearch, this limit is tightened to 1.5for both electrons andmuons. Standard looseidentificationrequirements[40,41]areappliedtoreducethe background fromnonprompt (NPR)leptons produced in semilep-tonic hadron decays and from jets showing a lepton signature. Furtherbackgroundreductionisachievedbyrequiringtheleptons tobeisolated.Theabsoluteisolation Iabs iscomputedbysumming

the transversemomenta ofPFcandidates,except thatof the lep-ton,inaconeofsize



R

<

0

.

3 aroundtheleptondirection,where



R

≡



(φ)

2

_{+ (}

_η

₎

2 _and

_φ

_{istheazimuthalanglemeasuredin}

radians. The energy inthe isolation cone is correctedfor the ef-fects of pileup. The relative isolation Irel is obtained by dividing

Iabs by the pT of thelepton. Thedetails of theisolation

require-mentsdifferbetweenthesingle-leptonanddileptontopologiesdue todifferencesinthedominantbackgroundsandthepurities.They aredescribedinSections4and

5

.

3. Samplesandeventpreselection

Thedatasamplecomprisesproton–protoncollisionsrecordedin 2012at acentre-of-mass energyof 8 TeVandcorresponds toan

integratedluminosityof19

.

7 fb−1.Thesearchuseseventspassing oneofseveralonline E_Tmiss selections.Thesetriggersevolvedover thedata-takingperiodandrequiredeitherEmiss_T

>

120 GeV,where

Emiss_T isreconstructedfromtheenergydepositedinthe calorime-ters,or Emiss_T

>

95 GeV andajetwith pT

>

80 GeV and

|

η

|

<

2

.

6,

where both objects are reconstructed using the PF algorithm. In thesecond partofthedata-taking period,thethresholdon Emiss_T

was raised from95 to 105 GeV. Control samples were collected basedonasingle-muontriggerwithapTthresholdof24 GeV.

SimulatedMonteCarlo(MC)samplesofSMbackgroundevents areproducedbyusingseveralgenerators.Singleandpair produc-tion of top quarks are simulated by using the powheg 1.0 [42]

program.Simulationsofmultijetanddibosoneventsaredonewith pythia6.4 [43]. The generation of all other relevant samples, in particular Z

/

γ

∗ processes, W

+

jets events, andtt production in association with a W, Z, or Higgs boson, is performedwith the MadGraph _{5.1 [44] generator. Alternative samples of tt and} di-boson events are also produced using MadGraph to investigate possiblesystematicdifferences,whicharefoundtobeinsignificant inthecontextoftheanalysesdescribed inthispaper.Allsamples generated with MadGraph or powheg are passed to pythia 6.4 withtheZ2*tune [45] forhadronizationandshowering. The de-tectorresponseissimulatedwiththe Geant4[46]program.Finally, alleventsarereconstructedwiththesamealgorithmsastheones usedfordata.Pileupeventsareincludedinthesimulationandall samplesare reweighted tomatch the distribution ofthe average numberoftheseeventsindata.

The signal simulation for



t pair production is done on a grid inthe



t–

χ



0

1 massplane withm

(



t

)

rangingfrom100–400 GeVin

stepsof25 GeV,and



m

≡

m

(



t

)

−

m

(



χ

₁0

)

rangingfrom10–80 GeV instepsof10 GeV.Theproductionoftop-squarkpairswithupto twoadditionaljetsandthefour-bodydecaysofthetopsquarksare generatedwith MadGraph. Thedecaysare forcedtoproceedonly throughvirtual SM particles.Chargino–neutralinopairproduction isalsomodelledwith MadGraph,whiletheirdecaysaregenerated with pythia.Weassumeabino-likeLSPandwino-like

χ



0

2 and

χ



1+

inorder to allow a direct comparison withRef. [25]. A rangein thecommongauginomassof100–400 GeViscoveredwithsteps of20 GeV, maintaining afixed massdifference of 20 GeV above the



χ

₁0. As forthe background samples,the generation steps for bothsignal models are followedby hadronization andshowering in pythia.For the signal samples, the modelling of the detector responseisperformedwiththeCMSfastsimulationprogram

[47]

. Differencesintheefficienciesoftheleptonselectionandtheb-jet identificationbetweenthefastandthedetailed Geant4simulation arecorrectedby usingscalefactors.Deficienciesinthemodelling ofISRinthesimulation

[48]

arecorrectedbyapplyingaweightas afunctionofthe pToftherecoilingsystem.

Theeffectsofresidualdifferencesbetweendataandsimulation aretakenintoaccountintheanalysis.The systematicuncertainty relatedtopossiblevariationsinthejetenergyscale[35]is evalu-atedbyacoherentchangeofalljetenergies,whichisalso propa-gatedto Emiss_T .Thejetenergyresolutioninsimulationisfoundto beslightlybetterthanindata[35].Tocompensateforthiseffect, the energies of simulated jetsare smeared and a corresponding systematicuncertaintyisassigned. Simulationiscorrectedfor dif-ferencesintheefficienciesofthereconstructionofleptons,andof theidentificationofleptons[40,41]andb jets[37,38]withrespect to the values measured in data. The corresponding uncertainties arepropagatedtothefinalresults.

Thefirst step inthe eventselection is designedto matchthe onlinerequirementsandtoserveasacommonbasisforthe anal-ysisinbothchannels. Itisguidedbythegeneralcharacteristicsof signalevents.TheleadingjetofeacheventisconsideredasanISR

jet candidate.It isrequiredtopasstighter jetidentification crite-ria andto fulfil pT

>

110 GeV and

|

η

| <

2

.

4. Since jets resulting

from



t decays are soft, andno jetsare expectedfrom



χ

0 2 or

χ



1+

decays,atmost oneadditional jet with pT

>

60 GeV isaccepted.

Atleastoneidentifiedmuonwith pT

>

5 GeV and

|

η

| <

2

.

1 must

be present.Finally,a requirementof Emiss

T

>

200 GeV isimposed.

Byusingacontrol samplecollected withthesingle-muontrigger, thesignaltriggersarefoundtobefullyefficientafterthese prese-lectioncriteriaareapplied.

4. Searchinthesingle-leptonchannel

Thesingle-leptontopologyisselectedbyrequiringatleastone muon within the acceptance described in the previous sections. Events are rejected if an electron, a

τ

lepton, or an additional muon with pT

>

20 GeV is present. Toavoidstrong variations of

the muon selection efficiencywith pT, a combined isolation

cri-terion, Iabs

<

5 GeV or Irel

<

0

.

2, is used, equivalent to a

transi-tion froman absolutetoa relative isolation requirementat pT

=

25 GeV. Theimpact parameters of themuon withrespect to the primary collision vertex in the transverse plane, dxy, and

longi-tudinal direction, dz, are required to be smaller than 0.02 and

0.5 cm,respectively.Theprimaryvertexischosenastheonewith the highestsum of pT2 of its associated tracks.Furthermore,

re-quirements are imposed on Emiss

T and on the scalar sumof the

transverse momenta of all jets, HT. Since thesetwo observables

are correlated, a simultaneous selection is applied by using the combined variable CT

≡

min

(

Emiss_T

,

HT

−

100 GeV

)

. To match the

preselection, CT

>

200 GeV isrequired.BackgroundfromSM dijet

andmultijetproductionissuppressed by requiringtheazimuthal anglebetweenthe momentum vectorsofthe twoleading jetsto be smaller than 2.5 rad for all events with a second hard jet of pT

>

60 GeV. According to simulation, the remaining sample

is dominated by W

+

jets and, to a lesser extent, by tt produc-tion witha single promptlepton inthe finalstate. Therefore,we use the transverse mass mT [49] computed from the transverse

components of the muon momentum and the E_Tmiss vector asa discriminant.

Distributions of the muon pT andof mT at this stage of the

selection are presentedin

Fig. 2

.Theyshow good agreement be-tween data and simulation.The variation of the signal shapes is illustrated withtwo extremecasesof the masssplitting (10 and 80 GeV).

To maintain sensitivity over a large range of



m values, sev-eralSRsaredefinedaslistedin

Table 1

.Sincesignalleptons have low pT, we impose an upper limit of pT

<

30 GeV in all these

selections. Because the muon pT spectrum ofthe signal changes

rapidly with



m, the full range of muon pT is subdivided into

threebinsinthecalculationofthefinalresults:5–12,12–20,and 20–30 GeV.

The signalregion labelledasSRSL1isdesignedforlow values of



m, wheretheb jetsproducedinthe



t decaysrarelypassthe selection thresholds.Avetoon b-taggedjetsstrongly reducesthe contribution from tt events. In addition, only events with nega-tively chargedmuons( Q

= −

1) areaccepted, usingthefact that theremainingW

+

jets backgroundshowssignificantlymore posi-tivelythannegativelychargedmuons[50]whilethesignalis sym-metricinthe muoncharge. Theacceptanceformuonsisreduced tothecentralregion,

|

η

| <

1

.

5,andtherequirementonCTis

tight-ened to 300 GeV. For signal points at low



m, mT is typically

small, mainly dueto the soft lepton pT spectrum. With

increas-ing



m, theaveragemT increasesandeventuallythe distribution

extendstovaluesabovem

(

W

)

.Tocoverthefullrangeof



m

Fig. 2. Distributionsof(top)muon pTand (bottom)mT afterthepreselectionof

thesingle-muonanalysis.Foreachplot,thevariableshownhasbeenexcludedfrom theselection.Dataareindicatedbycircles.Theuncorrectedbackgroundpredictions fromsimulationarerepresentedasfilled,stackedhistograms,andtheshapesfor twosignalpointswithm(t)=225 GeV andmasssplittingsofm=10 and80 GeV asdashedredandsolidbluelines,respectively.Theerrorbarsandthedark,shaded bandsindicatethestatisticaluncertaintiesofdataandsimulation,respectively.The lowerpanelsshowtheratioofdatatothesumoftheSMbackgrounds.(For inter-pretationofthereferencestocolourinthisfigurelegend,thereaderisreferredto thewebversionofthisarticle.)

definedbymT

<

60 GeV,60

<

mT

<

88 GeV,andmT

>

88 GeV,

re-spectively.

The second signal region (SRSL2) targets signals with higher mass splitting, where some of the b jets enter the acceptance. Therefore, the b jet veto in the region 30

<

pT

<

60 GeV is

re-versed, andatleast one such jetis required.Events with one or more b-taggedjetswith pT

>

60 GeV arestill rejectedto reduce

the tt background. In addition, the pT threshold of the ISR jet

candidateis raised to 325 GeV.This second SR receives a strong contributionfromtt events.

4.1. Backgroundestimation

The following fourbackground contributionsare estimatedby using data: W

+

jets and tt production,which are the dominant components forthe single-muon search;

(

Z

→

νν

)

+

jets, which is relevant for a signal region at high mT as explained below;

andmultijetproduction.Forthe firstthree ofthesebackgrounds, data/simulation scale factors are determined in suitable CRs and applied to the simulated yields in the SR. The contribution of multijeteventsisestimatedbyusingdataonly.Rarebackgrounds (other Z

/

γ

∗ processes,anddibosonandsingle top quark produc-tion)arepredictedbyusingsimulation.

Simulation provides only an imperfect description of the pT

spectrum for the main background samples (W

+

jets, tt). Since the extrapolations fromcontrol tosignal regions involvethe lep-ton pT spectrum,the pT distributions ofW bosons(for W

+

jets

events)andtopquarks(fortt events)arecorrectedbasedon mea-surementsindatasamplesdominatedbytt,Z

+

jets,andW

+

jets eventsbeforederivingthescalefactors.

Fortheestimationofthett background,asinglecontrolregion (CRSL

(

tt

)

)is used:eventsare requiredtopassthebasicselection defined above andmust include atleast two b-tagged jets, with one of them inthe pT region above 60 GeV. ThisCR hasan

es-timated purity of 80% in tt events. The observed event count in CRSL

(

tt

)

iscorrectedforotherbackgroundcontributionsand com-paredtotheyieldestimatedfromtt simulation.Theresultingscale factorof1.05isthenusedtomodifythepredictionsofthett sim-ulationinallSRs.

The W

+

jets yields from simulation are normalized in con-trol regions associated to each of the four signal (sub-)regions SRSL1a–c (CRSL1a–c) and SRSL2 (CRSL2). Control and signal re-gions differonlyby themuon pT range:intheCRsa muonwith

pT

>

30 GeV isrequired.ThecontrolregionsCRSL1a–chavean

es-timated purity of 80% in W

+

jets events. For region CRSL2 this numberisabout50%,theremainderbeingdominatedbytt events. Again,scalefactorsarederivedaftersubtractingnon-W

+

jets back-groundsfromtheobservedyieldsintheCR.Thett yieldsusedin thesubtractionarecorrectedbythescalefactordeterminedas de-scribed in the previous paragraph. The scalefactors for W

+

jets simulationvaryfrom0.88–1.18inthefoursignalregionsSRSL1a–c andSRSL2.

Eachfactorisappliedtoallthreemuon pT binsofasignal

re-gion.Systematicuncertaintiesareassignedrelatedtothestatistical uncertainties of the factors (6–30%), andto the shape of the pT

spectrum asdescribed later inthissection.Thedefinitions ofthe single-leptonsignalandcontrolregionsaresummarizedin

Table 1

, andtheexpectedcompositionsoftheeventsinthecontrolregions areshownin

Table 2

.Forthebenchmarksignalmodels,thecontrol regionswouldtypicallyreceiveacontributionfromsignaleventsat thelevelofafewpercent.Thiseffectistakenintoaccountinthe statisticalanalysisoftheresults.

After applying the signal selection, withthe exception of the requirementonmuon pT,themuon pT spectraoftt andW

+

jets

eventsaresimilar.Therefore,thecorrectionprocedureleadstoan anti-correlationoftheestimatesforthetwocategoriesanda rela-tiveuncertaintyinthesumofthetwocontributionsthatissmaller than the uncertainty in a single component. Forthis reason, the analysisisrobustagainstvariationsintherelativeyieldsoftt and W

+

jets events: a validation based on the direct estimation of thesumofbothbackgroundcomponentsfromthecontrolregions CRSL1a–c and CRSL2 yields almost identical results in terms of the totalbackground.Theanti-correlationbetweenthe two back-grounds is takeninto account in the computation of the results describedinSection6.

Table 1

Definitionofsignalandcontrolregionsforthesingle-muonsearch.Forjets,theattributes“soft”and“hard”refertothepTranges

30–60 GeVand>60 GeV,respectively.Forthecalculationofthefinalresults,eachsignalregion(SRSL1a–c,SRSL2)issubdivided intothreebinsaccordingtopT(μ):5–12,12–20,and20–30 GeV.

Variable SRSL1a–c, CRSL1a–c SRSL2, CRSL2 CRSL(tt)

Emiss

T (GeV) >300 >300 >200

HT(GeV) >400 – >300

pT(ISR jet)(GeV) >110 >325 >110

Number of hard jets ≤2 ≤2 ≤2

φ(hard jets)(rad) <2.5 <2.5 <2.5

Number of b jets 0 ≥1 soft (≥1 soft and≥1 hard)

0 hard or (≥2 hard) pT(μ)(GeV) 5–30 (SR),>30 (CR) 5–30 (SR),>30 (CR) >5 |η(μ)| <1.5 <2.4 <2.4 dxy(μ)(cm) <0.02 <0.02 <0.02 dz(μ)(cm) <0.5 <0.5 <0.5 Q(μ) −1 any any

Lepton rejection no e,τ, or additionalμwith pT>20 GeV

mT(GeV) <60 (a), 60–88 (b),>88 (c) – –

Table 2

Contributionstothecontrolregionsofthesingle-muonanalysisasdeterminedfromsimulationbeforeapplicationofscalefactors, togetherwiththeobservedeventcounts.Alluncertaintiesarestatistical.

Background CRSL(tt) CRSL1a CRSL1b CRSL1c CRSL2

W+jets 67.9±3.6 323.3±6.4 141.9±4.3 30.3±2.0 36.5±2.3 tt 471.0±9.6 19.5±2.2 9.9±1.5 6.1±1.2 37.5±3.5 Z/γ∗+jets 2.1±0.5 16.1±1.0 0.8±0.2 0.3±0.1 0.7±0.2 VV 3.8±0.6 13.7±1.3 8.0±1.1 2.5±0.5 1.1±0.4 Single top quark 58.6±12.6 4.6±1.4 3.3±1.2 1.1±0.7 3.5±1.2 Total SM 603.4±16.2 377.1±7.1 165.0±4.8 40.3±2.5 79.4±4.3

Data 628 347 172 46 75

Theextrapolationofthecorrection factorsfromcontrol to sig-nalregionshasbeenvalidatedbycomparingcorrectedyieldsfrom simulationtodatainsidebandregions.Eachofthesesidebandsis definedby one ofthefollowing changes withrespectto the sig-nalselection:(a)aloweringofthe Emiss_T

/

HTrequirementto200

<

CT

<

300 GeV,(b)a changeinthemuon chargerequirement, and

(c)theconditionofexactlyoneb-taggedjetwithpT

>

60 GeV.The

predictionsinthesidebandsarecompatiblewiththeobservations, andtheresultsareusedtoassignsystematicuncertaintiesonthe extrapolationof the scale factors to the SRs. These uncertainties are20%fortheestimateofthett backgroundand10–30%forthe estimateoftheW

+

jets backgroundwherethehighestuncertainty appliestoregionSRL1c.

At high values of mT, only a few W

+

jets events pass the

SRSL1c selection. In this signal region, Z

+

jets production, with the Z boson decaying to neutrinos, plus a nonprompt muon re-latedto oneof thejets, constitutesa non-negligiblecontribution. This contribution is estimated from simulation, together with a correctionderivedfromadatasampleofeventswithtwoormore muons,selectedbythesingle-muontrigger.Inthiscontrolsample, Z

+

jets eventswithZ bosondecaystomuonpairsareused.By us-ingtightermuonselectioncriteriaandrestrictingthemassofthe dimuonsystemto bewithin 15 GeV ofm

(

Z

)

,a high-purity sam-pleisobtained. The eventsare usedto mimicZ

→

νν

decaysby removingthe two daughtermuonsandaddingtheir momentato the Emiss_T vector. Thecorrection is applied asthe productof two factors:Rμμ,theinclusivedata-to-simulationratio,andRμμμ/μμ,

theratiooftheprobabilitiestoobserveathird,softmuon.Thefirst factorcorrects the crosssection inthe

μμ

channel for a signal-like region. Its measured value is 0

.

80

±

0

.

03. The double ratio

Rμμμ/μμ is determined in a looser selection to be 1

.

26

±

0

.

27,

yieldingatotal correctionfactorof1

.

01

±

0

.

22.The uncertainties quoted above are statistical. Systematic uncertainties due to the evolutionwithEmiss

T andHT,orduetodifferencesinthemuon

ef-ficiencyoracceptancebetweendataandsimulation,arenegligible withrespecttothestatisticaluncertainty.

Thecontributionfrommultijeteventsisestimatedbyinverting therequirementsonmuon isolation,themuonimpactparameter, andthevetoonleadingjetsinback-to-backconfiguration. Assum-ingsmallcorrelationsamongthethreevariablesmentionedabove, the yield ofmultijet eventscan be estimated fromthe yield ob-tainedwiththefullyinvertedselectioncombinedwiththeproduct of three reduction factors (onefor each variable). The estimated contributions to SRSL1 and SRSL2 are below 0.1 events and are thereforeneglected.

A summary of the expected contributions of different back-groundprocessestotheSRsisshownin

Table 3

togetherwiththe yieldsoftwobenchmarksignalpoints.

4.2. Backgroundsystematicuncertainties

Inadditiontothesystematicuncertaintiesestimatedinthe pre-vious subsections,the followingsystematiceffects andassociated uncertaintieshavebeenevaluated.

The full difference in the background estimates induced by the correction of the pT spectrum of simulated tt and W

+

jets

eventsis assignedasasystematic uncertainty.The impactofthe reweighting applied to tt eventsis onlysignificant for thesignal region SRSL2, where the contribution of this background is the highest.

Changes inthe polarization ofthe W boson can have an im-pactontheresults sincethey changethebalance betweenmuon

pT and EmissT . To quantify this effect, the polarization fractions

fλ=+1, fλ=−1, and fλ=0, associated withhelicity

+

1,

−

1, and 0

amplitudeshavebeenmodifiedfollowingthreedifferentscenarios: a 10%variationof f₋1

−

f+1 forbothW+andW−,a5%variation

of f₋1, f+1,and a 10% variation ofthe longitudinal polarization

Table 3

Estimatedbackgroundcontributionsforthesignalregionsofthesingle-muonanalysis.Thescalefactors deter-minedinthecontrolregionsareapplied.Forthesignalsamples,m(t)andm(χ0

1)areshowninparentheses.All

uncertaintiesarestatistical.

Background SRSL1a SRSL1b SRSL1c SRSL2 W+jets 116.8±8.8 73.2±7.6 8.8±2.1 16.0±4.9 tt 7.4±1.3 4.1±1.0 1.2±0.5 13.8±1.8 Z→νν+jets 1.1±0.4 1.2±0.4 1.5±0.5 0.3±0.2 Z/γ∗→  +jets 4.4±0.5 0.2±0.1 0.2±0.1 0.5±0.2 VV 4.6±0.7 1.8±0.5 0.7±0.3 0.5±0.2 Single top quark 0.1±0.1 0.6±0.4 <0.3 1.0±0.7 Total SM 134.5±8.9 81.3±7.8 12.3±2.3 32.1±5.3 tt signal(250,230) 32.5±2.8 6.2±1.2 4.7±1.0 7.1±1.3 tt signal(300,250) 11.0±1.0 4.2±0.6 5.1±0.7 10.7±1.0

Table 4

Relativesystematicuncertaintiesinthebackgroundpredictionsinthesignalregions ofthesingle-muonsearch.Thelabelsrefertosourcesofsystematicuncertainties discussedinSections3and4.2.

Systematic effect Uncertainty (%)

SRSL1a SRSL1b SRSL1c SRSL2

Pileup 0.5 0.8 3.0 0.5

W pTreweighting 7.1 8.8 8.1 3.7

tt pTreweighting 0.8 0.5 0.1 5.4

Jet energy scale 2.4 3.2 2.1 6.0 Jet energy resolution 1.1 4.4 7.3 3.4

b tagging 0.1 0.1 0.5 1.3 W polarization 2.9 2.8 3.9 0.8 Muon efficiency 3.5 3.5 3.5 3.5 W+jets validation 8.8 18.1 21.7 10.2 tt validation 1.0 0.9 1.7 8.4 Other backgrounds 3.8 2.4 9.8 3.6 Total uncertainty 13.1 21.5 27.0 17.1

The uncertainties basedon thecomparisonof dataand simu-lationinthe validationregions described inthe previous subsec-tion are propagated to the final estimate. An uncertaintyof 50% is assigned to the cross sectionsof all non-leadingbackgrounds, includingZ

→

νν,

andpropagatedthroughthefullestimation pro-cedure.An overviewofallsystematicuncertainties relatedtothe backgroundpredictionispresented inTable 4.The dominant un-certainties are related to the limited statistical precision of the validationprocedure andtotheuncertainties inthe shapeofthe muon pTspectruminW

+

jets events.

5. Searchinthedileptonchannel

Theanalysisinthedileptonchannelalsostarts fromthe com-monbaselineselectiondescribedinSection3.Inthistopology,less background is expected, and thus the selection requirements on

Emiss

T andthe pT oftheISRjet candidatearesettobeabove 200

and150 GeV,respectively,justabovethetriggerthresholds.To in-creasesensitivity, we select eventsin two signal regions defined by the pT of the leading lepton: 5–15and 15–25 GeV. The

sec-ond lepton is required to have pT

<

15 GeV. We requireexactly

twoidentifiedleptons ofoppositesign, withatleastoneofthem amuon.Finally,eventswithaninvariantmassofthedileptonpair

m

()

<

5 GeV arerejectedtoremove aregionthat isdifficultto simulateandtoavoidanypotential J

/ψ

background.Becausethe relativefractionofreconstructedleptons not arisingfromthe de-cayofaW or Z boson(“nonprompt”leptons)ishighercompared tothesingle-leptonchannel,theisolation andidentification crite-riaon theleptons are stricter.On top ofthe muon identification used for the single-leptontopology, stricter requirements on the numberoftrackerhits,thequalityofthetrackfit,andthematch

tosignalsinthemuondetectorareapplied.Thisselectionis sim-ilar to the soft muon identification used for b-quark physics in CMS[54].Forelectrons,thedefinitionsfortheH

→

ZZ

→

4



[55]

analysisareusedtogetherwithastrongerrejectionofphoton con-versions.Forbothflavours,theleptons arerequiredtobeisolated (Iabs

<

5 GeV andIrel

<

0

.

5) andtohaveimpactparametervalues

dxy anddz smaller than 0.01 cm. As in the region SRSL1 of the

single-muonanalysis,b-taggedjetsarevetoedtosuppresstt back-grounds.Toremovepotentialmultijetbackgrounds,a selectionon

Emiss_T

/

HT

>

2

/

3 isapplied.

Afterthisselection,one ofthemainbackgroundsisZ

/

γ

∗ pro-ductionof

τ

pairs,withboth

τ

leptonsdecayingleptonically. Un-der theassumptionthat thedirectionofthereconstructed lepton isparalleltothe

τ

direction,whichistruetogoodapproximation, theinvariantmassofthe

τ

pair,mττ ,canbereconstructedby set-ting its transverse momentum equal to the hadronic recoil (the missing transverse momentum without the leptons). All events withmττ

<

160 GeV arerejected.

Thedefinitionsofthedileptonsignal(SRDL)andcontrol(CRDL) regionsaresummarizedin

Table 5

.

5.1. Backgroundprediction

Four different backgroundcategories are predictedfrom data: dileptonic tt events (tt

(

2

)

,



:

e

μτ

), which constitutethe largest background;dibosonproductionsuch asWW orWZ (the second-largestbackground);andZ

/

γ

∗ productionof

τ

pairswithleptonic

τ

decays.Backgrounds with one nonprompt lepton, i.e. W

+

jets andsemileptonictt events(tt

(

1

)

),arethefourthcategory.Halfof the background eventscontain atleast one

τ

lepton that decays leptonically. The negligible (

≈

1%) contribution of rare processes (ttV, ttH, tW, andW±W±) is predictedby using simulation.For eachofthefourcategories,aCRenrichedinsuchprocessesis de-fined in data,fromwhich we derive correction factors tocorrect yieldsfromsimulation.

Inall CRs,therequirementsonjetsarethesameasintheSR. SeveralCRsuseeventswithhigherlepton pTcomparedtotheSR.

Intheseregions,theleadingleptonhastobe amuon,andevents areselectedbyusingthesingle-muontriggerdescribedbefore.The relativeleptonisolationhastobesmallerthan0.12andthemuon identification criteriaare tightened. Apart fromthe Z

/

γ

∗ control region,theEmiss_T requirementisloweredto125 GeVandtheEmiss_T

selection ofthesignalregion isinsteadappliedto LT,thesumof

Emiss

T and the pT of the leading lepton. The presentselection is

LT

>

225 GeV to take into account that for the default selection

LT is alsoupto 25 GeVhigherthan EmissT .Inthisway,theevent

yields inthe CRs can be increasedwhile maintaining kinematics similartotheSReveninthepresenceofahigher-pT lepton.

To achieve a clean control sample of dileptonic tt events (CRDL

(

tt

(

2

))

), werequireexactlyone b-taggedjet.Thisjet must

Table 5

Definitionofsignalandcontrolregionsforthedileptonsearch.FortheCRs,onlychangeswithrespecttotheSRareshown.Dashesindicatethatnoselectionisapplied.For thelowerlimitsonleptonpTintheSR,thevalueusedforelectronsisshowninparentheses.TheSRissubdividedintotwobinsaccordingtothepToftheleadinglepton:

5–15and15–25 GeV. Variable SRDL CRDL tt(2) NPR1 NPR2 VV Z τ τ Q(1)Q(2) −1 +1 +1 12 μμ,μe, eμ μμ,μe μμ,μe μμ,μe μμ pT(1)(GeV) 5(7)–25 >25 >25 >25 >125 pT(2)(GeV) 5(7)–15 >15 >15 >10 |η|() <1.5 <2.1 dxy, dz()(cm) <0.01 <0.02,<0.5

pT(ISR jet) (GeV) >150

pT(jet3) (GeV) <60

Number of b jets 0 1 0 (loose id.)

Number of jets ≥1 1 or 2

|φ(1,ISR jet)|(rad) – >1

Emiss T (GeV) >200 >125 >125 >125 – EmissT /HT >2/3 – – – – LT(GeV) – >225 >225 >225 – LT/HT – >2/3 >2/3 >2/3 pT(μμ)(GeV) – >200 pT(μμ)/HT – >2/3 m()(GeV) >5 >50 >10 m(τ τ)(GeV) >160 – <160

not be the leading jet to ensure a distribution in pT of the tt

system similar to that in the SR. We require one muon with

pT

>

25 GeV and a subleading lepton with pT

>

15 GeV.

Back-groundsother than tt

(

2

)

are subtracted fromdatabefore calcu-latingtheratiobetweendataandthe predictionfromsimulation for tt

(

2

)

in the CR. This ratio is used to rescale the simulated tt

(

2

)

yieldsintheSR.

FortheCRenrichedinnonpromptleptons(CRDL(NPR)),weuse the union of two samples. The first sample (CRDL(NPR1)) corre-spondstotheSRwiththeexceptionthattheleptonsarerequired tohavethesamecharge.Itwascheckedthatintheselected kine-matic region, the origins for NPR leptons, mainly heavy quarks, occuratasimilarfractionasintheSR.Inaddition,thekinematics ofthesenonpromptleptonsisverysimilarinsignalandcontrol re-gions.Forthesecondsample(CRDL(NPR2)),same-signeventswith aleadinglepton pTabove25 GeVareused,andtheCRselectionof

Emiss_T

>

125 GeV,LT

>

125 GeV,andLT

/

HT

>

2

/

3 isapplied.Under

theseconditionstheoriginsandkinematicsofthenonprompt lep-tonsaresimilarbetweensignalandcontrolregions,sincetheNPR contributioninthesignalregion ismostly relatedto the sublead-ing lepton.Again the datayield inthe combinedCRis corrected forother backgrounds, such as dibosonevents, by using simula-tion.TheratioofthecorrectedyieldtothesimulatedNPRyieldin theCRisusedtorescalethesimulatedNPRyieldintheSR.

ForthepredictionofZ

/

γ

∗events,twoseparateCRsaredefined. Thefirstone isusedto correctforanyeffectsonmττ (CRDL

(

Z

)

). Forthis purpose, a clean sample of Z

/

γ

∗ eventswith decays to a pair of muons is used. The invariant mass of the muon pair hasto be higher than 10 GeV andthe Emiss

T selection is applied

tothe pT ofthe muon pair.Three binsare definedasa function

of this momentum: 200–300, 300–400, and

>

400 GeV. We use thereconstructedmuon pair pT tomeasurethe resolutionofthe

hadronicrecoilalong andperpendicularto thedirectiongivenby themuon pairboth indataandsimulation. Theresulting scaling factorsoftherecoilresolutionareappliedtothesimulationto re-compute the efficiencyof themττ selection in the SR. A second controlregion(CRDL

(

τ τ

)

)isusedtomeasureindatathe probabil-ityofZ

/

γ

∗

→

τ τ

eventsleadingtotwosoftleptonsandveryhigh

Emiss

T .Todoso,weusetheSRDLselectionwiththerequirementon

mττ invertedto

<

160 GeV.Aftersubtractingotherbackgroundsin thisregionbyusingsimulation,theobservedyieldismultipliedby

thecorrectedmττ efficiencytopredictthenumberofZ

/

γ

∗ events intheSR.

For the diboson control region (CRDL

(

VV

)

) one muon with

pT

>

25 GeV is required. The pT of the second lepton has to be

>

15 GeV.Tofurtherenhancethedibosonfractionandreducethe otherwise dominanttt background,atmosttwo jetsareallowed, eventswithajetpassing alooserworkingpointoftheb tagging algorithmarerejected,andtheazimuthalanglebetweenthe lead-ingleptonandtheleadingjethastobe

>

1 rad.Finally,werequire

m tobeabove50 GeV.Contributionsoftt

(

2

)

,NPR,andZ

/

γ

∗ to CRDL

(

VV

)

are estimatedwith methods similar to those used for the SR. Backgrounds dueto rare processes are subtracted by us-ing the simulation.Afterthiscorrection, the ratioofthe number of data to simulated dibosonevents is built andused to rescale thesimulatedVV yieldintheSR.

Thebackgroundcontributionfrommultijeteventsisnegligible in our final selection. Apart from the fact that we require high

Emiss

T andtwoleptons,wealsoselect ETmiss

/

HT

>

2

/

3 torejectany

residualmultijetevents.Toevaluate theefficacyofthisselection, a testwas performedby invertingthisrequirementtohavea re-gionthatshouldhavesignificantmultijetbackgroundiftherewere any.Forthisregion, datayields were comparedwiththe simula-tionresultsofallconsideredbackgroundcategories(whichdonot includemultijetevents)andwere foundto beinagreement. Fur-thertestswereperformedbyusingtheelectron–electronchannel andby relaxingtheupperlimitsondxyanddzto0.05 cm,which

showednoindicationforacontaminationbymultijetevents.These tests confirmed that, as expected, we can assume that multijet backgroundisnegligibleinourfinalselection thatemploysmuch tighterrequirementsagainstmultijetevents.

The event yields for data andsimulation in the different CRs thatarethebasisforthescalefactorsappliedintheSRsareshown in Table 6. The predicted event yields per background for each searchbinarepresentedin

Table 7

.Theimpactofapotential sig-nalcontaminationisfoundtobe onlyrelevantforcontrolregions CRDL(NPR)andCRDL

(

τ τ

)

,withaneffectofafew percentonthe totalbackgroundpredictioninthesignalregions,andistakeninto accountinthestatisticalanalysisoftheresults.

Totestthepredictionmethods,wedefineseveralvalidation re-gionsthatareenrichedinspecificbackgroundsbutexpectedtobe free of signal. The first region is equivalent to the signal region, except foran inversion ofthe veto on b-tagged jets. This region

Table 6

Contributionstothecontrolregionsofthedileptonanalysisasexpectedfromsimulationbeforeapplicationofscale factors,togetherwiththeobservedeventcounts.Alluncertaintiesarestatistical.

Background CRDL(tt(2)) CRDL(NPR) CRDL(VV) CRDL(τ τ) tt(2) 119.1±2.4 0.27±0.11 30.3±1.2 0.15±0.08 tt(1) 1.09±0.29 4.7±0.6 0.30±0.14 0.11±0.11 W+jets <0.4 3.4±1.3 <0.4 0.7±0.7 Z/γ∗+jets 0.4±0.4 <0.30 4.9±1.3 2.8±0.9 VV 2.4±0.6 0.62±0.11 45.9±1.8 0.13±0.09 Rare backgrounds 14.9±2.7 1.0±0.5 6.4±1.7 <0.21 Total SM background 138.0±3.7 10.0±1.5 87.8±3.0 3.9±1.1 Data 119 11 8 5 Table 7

Estimatedbackgroundcontributionsforthetwosignalregionsofthedileptonsearch.Thescalefactorsdeterminedin thecontrolregionsareapplied.Forthesignalsamples,m(t)andm(χ0

1)areshowninparentheses.Alluncertainties

arestatistical.

Background pT(1): 5–15 GeV pT(1): 15–25 GeV Inclusive

tt(2) 0.75±0.19 2.08±0.37 2.8±0.4 tt(1), W+jets 0.60±0.33 1.3±0.7 1.9±0.8 Z/γ∗+jets <0.30 0.5±0.5 0.5±0.5 VV 0.74±0.27 1.6±0.5 2.4±0.5 Rare backgrounds 0.03±0.01 0.08±0.04 0.11±0.04 Total SM background 2.1±0.5 5.6±1.0 7.7±1.1 tt signal(225,145) 4.2±1.2 6.3±1.6 10.4±2.0 tt signal(300,250) 4.0±0.6 3.8±0.6 7.8±0.8

is usedto test the predictionof low-pT leptons in tt

(

2

)

events.

ThenextvalidationregionisidenticaltoCRDL

(

tt

(

2

))

,exceptthat the pT of thesubleading leptonis requiredto be below15 GeV.

This region provides a further test of the prediction of the soft-lepton rate. Another validation region is the same as CRDL

(

VV

)

, apartfromthefactthatall selectionsusedtoenrichtheregionin dibosonevents are inverted. In addition,a validation region that hasa composition inbackgrounds similar to the signal region is defined. For this, one muon with pT above 25 GeV is required,

while the second lepton must be soft (pT

<

15 GeV). All

valida-tion regions show reasonable agreement betweenprediction and observation.

5.2. Backgroundsystematicuncertainties

In addition to the common uncertainties from object recon-struction and simulation as described in Sections 2 and 3, the followingsystematicuncertainties thatarespecific tothe individ-ualbackgroundpredictionsareconsidered.

Inestimatingthe tt

(

2

)

background,thepolarization oftheW bosonresultingfromthetopquarkdecayisvaried.Inaddition,the spin correlation between thetwo top quarksis changed by 20%, since this might affect how often both leptons are soft [56]. As inthesingle-leptonchannel,thedifferenceduetothereweighting ofthetopquark pTspectrumintt simulationistakenasafurther

uncertainty.However,itseffectissmallduetothebackground pre-dictionmethodused.

Foraconservativeassessmentoftheuncertaintyrelatedtothe estimateofNPRbackgrounds,thefractionsofleptons fromb and c hadrons are varied by 50% and100%, respectively. The relative fractionoftt toW

+

jets eventsisalteredby rescalingboth con-tributions by

±

50%.Thelargest ofthesevariations isusedasthe uncertainty. Furthermore, the pT and

|

η

|

distributions in the CR

arevaried toreflectpotentialresidualdifferencesinthe kinemat-icsbetweensignalandcontrolregions.Moreover,thepolarization oftheW bosonisvariedtoestimatetheuncertaintydueto polar-izationmodelling.

Table 8

Relativesystematicuncertaintiesinthebackgroundpredictionsinthesignalregions ofthedileptonsearch.

Systematic effect Uncertainty (%)

pT(1): 5–15 GeV pT(1): 15–25 GeV

Statistical uncertainty 21.9 18.3

Jet energy scale 1.0 2.8

b tagging 1.5 1.4 Electron efficiency 1.3 1.1 Muon efficiency 6.0 4.5 tt background 5.1 5.4 NPR background 10.1 5.6 Z/γ∗background <0.1 2.3 VV background 8.0 2.6 Rare backgrounds 3.7 3.3 Total uncertainty 26.9 21.1

The cross sectionsfor WW [57,58] andalso WZ and ZZ[59]

production have been measured at the LHC, and both the total anddifferentialcrosssectionsshowreasonableagreementbetween data and simulation.To estimate the uncertainties related to VV production,thepolarizationofthevectorbosonsisalteredby10%, aswellasthefractionofthedibosonpairmomentumthata sin-gle boson carries. Inaddition, the crosssection corresponding to eventswithlowm

(

γ

∗

)

between5and12 GeVisvariedby 100% to account for anypotential shape mismodelling of the dilepton mass.

IntheestimationoftheZ

/

γ

∗ background,theeffectofthe re-coilresolutioncorrectionisusedtoderiveanuncertaintyduetoa potential mismodellingoftheresolution.Thecrosssectionofrare processes isvaried by

±

50% throughouttheanalysis (alsoin the CRs),andtheeffectispropagatedtotheeventyieldsintheSR.

A summary of all uncertainties can be found in Table 8. The dominating uncertaintystems from thelimited numberof simu-latedeventswithnonpromptleptonsintheSRs.

Table 9

Summaryofobservedandexpectedbackgroundyieldsinthesignalregionsofthesingle-leptonanddileptonsearches.Theuncertaintiesinthebackgroundyieldsinclude statisticalandsystematiccontributions.TransversemomentaareshowninunitsofGeV.

Single muon Dilepton

pT(μ) SRSL1a SRSL1b SRSL1c SRSL2 pT(1) SRDL 5–12 exp. 41.1±6.3 29.7±7.2 4.3±1.5 11.3±2.9 5–15 exp. 2.1±0.6 obs. 42 17 3 16 obs. 2 12–20 exp. 44.2±6.8 25.1±6.2 3.1±1.2 8.5±2.4 15–25 exp. 5.6±1.2 obs. 39 14 4 16 obs. 4 20–30 exp. 49.2±7.5 26.5±6.5 5.0±1.8 12.2±3.0 obs. 40 28 5 9

All exp. 134.5±19.8 81.3±19.1 12.3±4.0 32.1±7.7 All exp. 7.7±1.4

obs. 121 59 12 41 obs. 6

6. Resultsandinterpretation

Theobservationsandbackgroundpredictionsforthesignal re-gions ofthe single-lepton anddileptonsearches are summarized inTable 9. Observedandpredictedyields are in goodagreement andgivenoindicationofthepresenceofsignal.

Themodified-frequentistCLSmethod[60–62]withaone-sided

profile likelihood ratio test statistic is used to define 95% con-fidence level (CL) upper limits on the production cross section asa function of thesparticle masses. Statistical uncertainties re-latedto the observed number of events in CRs are modelled as Poisson distributions. All other uncertainties are assumed to be multiplicativeandaremodelledwithlog-normaldistributions.The impactofapotentialsignalcontaminationinthecontrolregionsis takeninto accountinthe calculationofthelimitsforeach signal point.

Systematicuncertainties inthesignal yieldsrelatedtothe de-terminationoftheintegratedluminosity[63](2.6%),pileup(

≈

2%), energy scales [35,36] (up to 7%), object identification efficien-cies[40,41](upto10%), anduncertaintiesintheparton distribu-tion functions[64–68] (upto 6%) and the modellingof ISR [48]

(

≈

20%) have been evaluated. Correlations between the system-atic uncertainties in different signal regions are taken into ac-count, where applicable. All systematic uncertainties are treated asnuisance parameters inthe calculation of the limits, withthe exceptionofthetheoreticaluncertaintyontheinclusiveSUSY pro-ductioncross section. The latter is shownin theform of an up-anddownwardvariationoftheobservedmasslimits.

Thelimitsobtainedfortopsquarkpairproductioninthe single-muon and the dilepton searches are shown in Fig. 3 top and bottom, respectively, under the assumption of a 100% branching fraction of the four-body decay. By using the



t pair production crosssection calculatedatnext-to-leadingorder(NLO)

+

next-to-leadinglogarithm(NLL)precision

[69–73]

,thecrosssection limits can be converted into excluded regions in the



t–

χ



₁0 mass plane. Uncertaintiesinthesecrosssectionsaredeterminedasdetailedin Ref.[74].At



m

=

25 GeV,thedileptonsearch excludes



t masses below316 GeV. Here and in the following all quoted values for masslimitsconservativelyrefer to the

−

1σ variationof the pre-dictedcrosssection.Thesingle-muonsearchshowsasmallerreach inm

(



t

)

(

≈

250 GeV)buthasahighersensitivityatthelowest con-sideredmasssplitting of10 GeV,where valuesup to

≈

210 GeV areexcluded.Intheintermediate



m region(

≈

20–70 GeV),these resultsconsiderablyextendexistinglimits

[19–21]

.Theyare com-plementary to the results of searches in the monojet topology

[18,19,21].

Inthecaseofchargino–neutralino pairproduction,theresults of the dilepton analysis are used. For the model involving de-caychainswithsleptons,theslepton massesaresetto

(

m

(



χ

0

1

)

+

m

(



χ

₁+

))/

2.Insteadofusingbranchingfractionsderivedfrom com-pleteSUSYmodels,twoextremedecayscenariosarestudiedin

or-Fig. 3. Crosssectionandmasslimitsat95%CLinthem(χ0

1)andm(t)massplane

forthe(top)single-muonand(bottom)dileptonsearches.Thecolourshading corre-spondstotheobservedlimitonthecrosssection.Thesolid(dashed)linesshowthe observed(expected)masslimits,withthethicklinesrepresentingthecentralvalue andthethinlinesthevariationsduetothetheoretical(experimental)uncertainties.

dertoillustratethedependenceonthefinalstate.Inthe “flavour-democratic”scenario,boththeneutralinoandthecharginowould decayviathesupersymmetricpartnersoftheleft-handedleptons (





L) andoftheneutrinos(



ν)

withequalbranchingfractionstoall

Fig. 4. Crosssection limits at 95%CLobtainedfrom the searchinthe dilepton channelasafunctionofthecommonχ1±/χ10mass.Theblacklineswithsymbols

correspondtotheobservedlimit,whilethesolidanddashedcolouredlines rep-resenttheexpectedlimitandthe±1σ bandscorrespondingtotheexperimental uncertainties,respectively.Theflavour-democratic(τ-enriched)casesofthemodel areindicatedbygreen(orange)linesandupward- (downward-)pointingtriangular symbols.Thesolidanddashedbluelineswithoutsymbolscorrespondtothe pre-dictedcrosssectionforchargino–neutralinoproductionanditsuncertainties.(For interpretationofthereferencestocolourinthisfigurelegend,thereaderisreferred tothewebversionofthisarticle.)

leptonflavours.Inthisscenario,thefractionofeventswithatleast two charged leptons is reduced by 50% due to the



χ

0

2

→

νν

χ



0 1

decaychannel.Inthe“τ-enriched”scenario,thedecayswould pro-ceedviathesupersymmetricpartnersoftheright-handedleptons. Inthiscase,thedecay



χ

0

2

→

νν



χ

10isnotpresent,andweassume

equalbranchingfractionsofthe

χ



0

2 intothethree chargedlepton

flavoursandexclusivedecaysofthecharginoto

τ

leptons.In

Fig. 4

, the95% CLcrosssection limitsarepresentedforamasssplitting of



m

≡

m

(

χ



)

−

m

(

χ



0

1

)

=

20 GeV.Comparingwiththe predicted

crosssection, calculatedatNLO

+

NLLprecision withthe Resum-mino_{[75–77]program,95%CLlimitsonm}

₍



χ

₎

_{of212and307 GeV} areobtainedfortheflavour-democratic and

τ

-enrichedscenarios, respectively.Inthesecompressedscenarios,thenewlimitsslightly improvecurrentresults[25,29]intheflavour-democraticscenario and exceed them by

≈

200 GeV for the

τ

-enriched scenario. As forthe lattercase, the dominantdecays leadto final states with opposite-signleptons.

7. Summary

Asearch forsupersymmetry withcompressed massspectrais performedineventswithsoftleptons,moderatetohighvaluesof

Emiss

T ,andone ortwo hardjets, compatiblewiththeemission of

initial-stateradiation.Thedatasamplecorrespondsto19

.

7 fb−1 of proton–protoncollisionsrecordedbytheCMSexperimentat

√

s

=

8 TeV.Twoeventcategoriesare considered: eventswithasingle, softmuonandeventsinwhichasecond,softelectronormuonis present.

The first target of this search is the pair production of top squarks witha masssplittingof atmost80 GeV withrespectto theLSP. At low masssplitting, leptonmomenta are low,andthe b jetsdo not enter the acceptance.At higher values of



m, the averageleptonmomentumincreasesandsoftb jetscanbe recon-structed.Therefore,signalregionsarefurtherdividedaccordingto thepToftheleadingleptonandthepresenceorabsenceofasoft

b-taggedjet.Inthesingle-leptonsearchthetransversemassofthe lepton-Emiss

T systemisusedasanadditionaldiscriminant.

The mainbackgrounds tothissearch are W

+

jets andtt pro-duction.Contributionstothesignalregionsfromtheseandseveral nonleadingbackgroundsourcesareestimatedbyusingdatain con-trolregionstonormalizethesimulatedyields.Theseestimatesare testedwithdatainvalidationregions.

Theobservationsinthesignalregions arecompatiblewiththe SMbackgroundpredictions.Intheabsenceofanyindicationof sig-nal,crosssectionlimitsaresetat95%CLinthe



t–

χ



0

1 massplane.

Theseresultsareusedtoextractmasslimitsbasedonareference crosssectionfortopsquarkpairproductionandassuminga 100% branching fraction for the four-body decay



t

→

bff

χ



₁0. The most stringent limit on the massof the top squark isobtained in the dileptonchannel:m

(



t

)

>

316 GeV at95%CLforamasssplittingof 25 GeV.Theseresultsextendexistinglimitsinthefour-bodydecay channel of the top squark[19–21] andcomplement theanalyses performedinthe



t

→

c



χ

0

1 channel[18,21].

The results obtainedin the dilepton channel are also used to setlimitsonmodelsofchargino–neutralinoproductionina com-pressedspectrumwithamassdifferencebetween



χ

0

2

/

χ



1+ and

χ



10

of20 GeV.Based onthe 95%CL upperlimiton thecrosssection in the caseof flavour-democratic leptonic decays of these parti-cles,alowerlimitonthecommon



χ

₁+

/



χ

0

2 massissetat212 GeV.

If chargino decays proceed exclusively via the

τ

channel, andin the absence ofthe

χ



0

2

→ 

νν

decay mode, thislimit increases to

307 GeV,wellaboveexistinglimits[25,29].

Acknowledgements

WecongratulateourcolleaguesintheCERNaccelerator depart-ments for the excellent performance of the LHC and thank the technical andadministrativestaffs atCERNand atother CMS in-stitutes for their contributions to the success of the CMS effort. Inaddition,wegratefullyacknowledgethecomputingcentresand personneloftheWorldwideLHCComputingGridfordeliveringso effectively thecomputinginfrastructure essentialto our analyses. Finally, we acknowledge the enduring support for the construc-tion andoperationofthe LHCandtheCMSdetectorprovided by thefollowingfundingagencies:BMWFWandFWF(Austria);FNRS and FWO (Belgium); CNPq, CAPES, FAPERJ, and FAPESP (Brazil); MES(Bulgaria);CERN;CAS,MOST,andNSFC(China);COLCIENCIAS (Colombia);MSESandCSF(Croatia);RPF(Cyprus);MoER,ERCIUT andERDF(Estonia); AcademyofFinland,MEC,andHIP (Finland); CEA andCNRS/IN2P3 (France); BMBF, DFG, and HGF (Germany); GSRT (Greece); OTKA and NIH (Hungary); DAE and DST (India); IPM (Iran); SFI (Ireland); INFN (Italy); MSIP and NRF (Republic of Korea); LAS (Lithuania); MOE andUM (Malaysia); CINVESTAV, CONACYT,SEP,andUASLP-FAI(Mexico);MBIE(NewZealand);PAEC (Pakistan);MSHEandNSC(Poland);FCT(Portugal);JINR(Dubna); MON,RosAtom,RAS,andRFBR(Russia);MESTD(Serbia);SEIDIand CPAN(Spain);SwissFundingAgencies(Switzerland);MST(Taipei); ThEPCenter,IPST, STARandNSTDA(Thailand);TUBITAK andTAEK (Turkey); NASUand SFFR(Ukraine); STFC (United Kingdom);and DOEandNSF(USA).

Individuals have received support from the Marie-Curie pro-gramme andtheEuropean Research Council andEPLANET (Euro-peanUnion);theLeventisFoundation;theAlfredP.Sloan Founda-tion; the Alexander von Humboldt Foundation; the Belgian Fed-eral Science Policy Office; the Fonds pour la Formation à la Recherchedansl’Industrieetdansl’Agriculture(FRIA-Belgium);the AgentschapvoorInnovatiedoorWetenschapenTechnologie (IWT-Belgium); the MinistryofEducation, Youth andSports(MEYS) of theCzechRepublic;theCouncilofScienceandIndustrialResearch, India; the HOMING PLUSprogramme of the Foundation for Pol-ish Science, cofinanced from European Union, Regional Develop-ment Fund; the OPUS programme of the National Science