Search for new physics in events with two soft oppositely charged leptons and missing transverse momentum in proton–proton collisions at √s=13TeV

Contents lists available atScienceDirect

Physics

Letters

B

www.elsevier.com/locate/physletb

Search

for

new

physics

in

events

with

two

soft

oppositely

charged

leptons

and

missing

transverse

momentum

in

proton–proton

collisions

at

√

s

=

13

TeV

.

The

CMS

Collaboration



CERN,Switzerland

a

r

t

i

c

l

e

i

n

f

o

a

b

s

t

r

a

c

t

Articlehistory: Received 5 January 2018

Received in revised form 1 May 2018 Accepted 15 May 2018

Available online 25 May 2018 Editor: M. Doser Keywords: CMS SUSY Compressed Leptons Missing energy

A searchis presentedfornewphysics inevents withtwolow-momentum,oppositelycharged leptons (electronsormuons)andmissingtransversemomentuminproton-protoncollisionsatacentre-of-mass energy of13 TeV. The data collectedusingthe CMS detector atthe LHCcorrespond toan integrated luminosityof35.9 fb−1.Theobservedeventyieldsareconsistentwiththeexpectationsfromthestandard model.Theresultsareinterpretedintermsofpairproductionofcharginosandneutralinos(

χ

₁±and

χ

0

2) with nearlydegeneratemasses,asexpectedinnaturalsupersymmetrymodelswithlighthiggsinos, as wellasintermsofthepairproductionoftopsquarks(t),whenthelightestneutralinoandthetopsquark havesimilarmasses.At95%confidencelevel,wino-like

χ

₁±/

χ

0

2 massesareexcludedupto230 GeV for amass differenceof20 GeV relativetothelightest neutralino.Inthe higgsino-like model,massesare excluded upto168 GeV forthe samemassdifference.Fort pairproduction,top squarkmassesupto 450 GeV areexcludedforamassdifferenceof40 GeV relativetothelightestneutralino.

©2018TheAuthor(s).PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense (http://creativecommons.org/licenses/by/4.0/).FundedbySCOAP3_.

1. Introduction

Supersymmetry(SUSY) [1–5] isa widelyconsidered extension ofthe standard model(SM) ofparticle physics, asit canprovide solutionsto severalopen questionsintheSM, inparticularthose relatedtothehierarchyproblem [6–8] andthenatureofdark mat-ter.SUSYpredictssuperpartnersofSMparticleswhosespinsdiffer byone-halfunitwithrespecttotheirSMpartners.InR-parity con-servingmodels [9],SUSYparticlesarepair-producedandtheir de-caychainsendinthestable,lightestSUSYparticle(LSP),whichin manymodelscorrespondstothelightestneutralino(

χ



0

1).Astable

LSPwouldescapeundetected,yieldingacharacteristicsignatureof alargemagnitudeofmissingtransversemomentum(pmiss_T )in col-lisionsattheCERNLHC.Asastable,neutralandweaklyinteracting particle,theneutralino matchesthepropertiesrequiredofa dark mattercandidate [10].

The absence ofSUSY signals in previous experiments,as well asattheLHC,canbe interpreted asan indication that SUSY par-ticleshave verylarge mass,leading to theexpectationthat SUSY eventshavelarge visibleenergyandmomentum.As aresult, the manysearchesthatyieldthemoststringentlimitson themasses

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

of the SUSY particles are based on events with large pmiss_T and energetic final-stateobjects such asleptons andjets. Another in-terpretation for the absence of a SUSY signal is that the SUSY particlesareinapartoftheparameterspacethatisnoteasily ac-cessible.Onesuch scenario,wherepreviously mentionedsearches wouldnotbesensitive,iswherethemassspectrumiscompressed, i.e. the mass splitting between the produced SUSY particles and theLSPissmall.WhenthemasssplittingsbetweenSUSYparticles aresmall, thevisibleenergyintheevent,andalsopotentiallythe

pmiss

T , is relatively low, which motivates searches in events with

low-momentumobjects.

Compressed mass spectra arise in several SUSY models, in-cluding natural SUSY, i.e. SUSY models that solve the hierarchy problem with little fine tuning. It has been pointed out in sev-eral studies, for example in Refs. [6–8,11–15], that naturalness imposes constraints onthemassesof higgsinos,top squarks,and gluinos. Natural SUSY is generally considered to require at least onecolouredSUSYparticleofmassbelowapproximatelyone TeV. Further,itisoftenassumedthatthisparticleisthetopsquark(



t). More recently,however, thehypothesis ofnaturalSUSY requiring a light top squark hasbeen disputed asarising from oversimpli-fiedassumptions[16–18].Irrespectiveofthetopsquark,higgsinos remainacomplementarywindowtonaturalSUSYastheyare gen-erally expectedto be light.As pointedout in Refs. [19–22], light

https://doi.org/10.1016/j.physletb.2018.05.062

0370-2693/©2018 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Funded by SCOAP3_.

higgsinosare likelyto haveacompressed massspectrum, poten-tiallyleading tosignatures withsoftleptons andmoderate pmiss_T . Thusfar,themostsensitivesearchesinthismodelhavebeen car-riedout by experiments atLEP [23,24] and ATLAS [25]. The LEP experimentsexcluded

χ



₁±massesupto103.5 GeV foramass split-tingbetweenthe

χ



₁±and

χ



0

1 ofatleast3 GeV.

Thesearch described inthis letteris designedfor neutralinos andcharginos,whicharecollectivelyreferredtoas “electroweaki-nos”,inamodelwheretheseelectroweakinosformacompressed mass spectrum [19,21,22,26]. Two models are considered where the electroweakinosare either pure wino/bino-like orwhere the lightestelectroweakinosareofmostlyhiggsinonature.Thesearch hasdiscoverypotential alsowhen alight topsquarkandthe LSP arenearly degenerate inmass andthetop squarkdecays tofour fermions.Amoredetaileddiscussionofsuchmodelscanbefound inRef. [27].Thenear-degeneracyinmassofthetopsquarkandthe LSP is typical ofthe so-called“co-annihilation region”, in which theLSPisthesolesourceofdarkmatter [28].

In the models considered in this analysis, the visible decay products inthe SUSY signal have low momentum, which can be distinguishedfromSM processeswhenajet withlargetransverse momentum(pT) frominitial-stateradiation(ISR) leads toa large

boostoftheSUSYparticlepair.Thisboostalsoenhancesthe pmiss_T

intheevent.Asimilarsearchhaspreviouslybeenreportedbythe ATLASCollaboration [25].Forthesignalstudiedinthisletter,SUSY particles can decayleptonically, and the presenceof low-pT

lep-tonscanbe usedto discriminateagainst otherwisedominantSM backgrounds,such asmultijetproductionthrough quantum chro-modynamics(QCD)andZ

+

jets eventswithinvisibleZ boson de-cays.

Thecurrentstrategyissimilarto thatintheprevious publica-tionbasedon8 TeV data [29],withthemaindifference beingthe deploymentofanewtriggerselectionthatimprovesthesensitivity ofthesearchineventswithtwomuonsandlow pmiss_T .Inaddition, theselectionhasfurtherbeenoptimizedforelectroweakinoswith a compressedmass spectrum. At least one jet is requiredin the finalstate;inthe caseofthesignal,thisjet mustarise fromISR, whichprovidesthefinal-stateparticleswithaboostinthe trans-verseplane,andtherebythepotentialformoderateorlarge pmiss

T

intheevent. Unlikethe8 TeV analysis,thereisnoupperlimiton thenumberofjetsintheevent.

2. CMSdetector

The central feature of the CMS apparatus is a superconduct-ingsolenoidof 6 minternal diameter,providing amagnetic field of3.8 T.Withinthesolenoidvolumeare asilicon pixelandstrip tracker,aleadtungstatecrystalelectromagneticcalorimeter(ECAL), andabrassandscintillatorhadroncalorimeter(HCAL),each com-posedofa barreland twoendcap sections.Forwardcalorimeters extendthepseudorapidity(

η

)coverageprovidedbythebarreland endcapdetectors.Muonsare detectedin gas-ionizationchambers embeddedinthesteelflux-returnyokeoutsidethesolenoid.

Events ofinterest are selected usinga two-tieredtrigger sys-tem [30].The firstlevel(L1),composed ofcustom hardware pro-cessors,uses informationfromthe calorimetersandmuon detec-tors to select eventsat a rate of around 100 kHz within a time intervaloflessthan4 μs.Thesecondlevel,knownasthehigh-level trigger (HLT),consists of a farm of processors running a version of the full event reconstruction software optimized forfast pro-cessing,andreduces theevent rateto around 1 kHzbefore data storage.

AmoredetaileddescriptionoftheCMSdetector,togetherwith adefinitionofthecoordinatesystemusedandtherelevant kine-maticvariables,canbefoundinRef. [31].

3. Dataandsimulatedsamples

Thedatausedinthissearchcorrespondtoanintegrated lumi-nosityof35.9 fb−1 ofproton–proton(pp)collisionsata centre-of-massenergyof 13 TeV,recordedin2016using theCMSdetector. The data are selected usingtwo triggers: an inclusive pmiss

T

trig-ger,whichisusedforsignalregions(SRs)withanofflinepmiss

T cut

>

200 GeV andanadditionaltriggerwhichrequirestwomuonsto lowertheoffline pmiss_T cutto 125 GeV.Boththemuon pT andthe

muonpairpThaveatriggeronlinecutofpT

>

3 GeV.Theinclusive

pmiss_T triggerscorrespondtoan integratedluminosity of35.9 fb−1, whereastheeventsrecordedwiththedimuon

+

pmiss_T trigger cor-respondto33.2 fb−1.

Simulated signal and major background processes, such as tt, W

+

jets, and Z

+

jets are generated with the MadGraph5_ amc@nlo 2.2.2[32,33] event generatoratleading order(LO) pre-cision in perturbative QCD usingthe MLM merging scheme [34]. Additional partons are modelled in these samples. The diboson processesWW, ZZ,andW

γ

aregenerated withthe MadGraph5_ amc@nlo 2.2.2eventgeneratoratnext-to-leadingorder(NLO) pre-cisionusingtheFxFxmergingscheme [33],whiletheWZprocess isgenerated atNLO with powheg v2.0[35–39]. Rarebackground processes(e.g. ttW,ttZ,WWW,ZZZ,WZZ,andWWZ)arealso gen-erated at NLO precision with MadGraph5_amc@nlo 2.2.2 (2.3.2.2 for ttZ) [32,33].The rarebackgroundfromsingle top quarks pro-ducedinassociationwithaW bosonisgeneratedatNLOprecision with powheg v1.0 [40]. The NNPDF3.0 [41] LO and NLO parton distribution functions (PDF) are used for the simulated samples generated atLO andNLO. Showering, hadronization andthe un-derlying eventdescriptionare carriedout using the pythia 8.212 package [42] withthe CUETP8M1 underlying event tune [43,44]. AdetailedsimulationoftheCMSdetectorisbasedonthe Geant4 [45] package.Afastdetectorsimulation [46] isused forthelarge numberofsignalsamples,correspondingtodifferentSUSYparticle masses.Thetrigger,leptonidentification,andb taggingefficiencies arecorrectedinthesimulationthroughapplicationofscalefactors measured indedicateddatasamples [47].Corrections fortheuse ofthefastdetectorsimulationarealsoapplied.

For the signal, we consider the neutralino–chargino (

χ



0 2–

χ



1±)

pair production where the mass degenerate

χ



0

2 and



χ

1± are

as-sumedtodecaytotheLSPviavirtualZ andW bosons.Thedecays ofelectroweakinosare carriedout using pythia,assuming a con-stant matrix element. The SM branching fractions are assumed forthe decays ofthe virtual Z and W bosons.The simulation of the

χ



0

2 (

χ



1±) decay takes into account the Breit–Wigner shape

of the Z (W) boson mass. The production cross sections corre-spondtothoseofpurewinoproduction [48–50] computedatNLO plus next-to-leading-logarithmic (NLL) precision. A second mass scan simulates a simplified model of



t-pair production,in which a heavy chargino mediates the decay of the



t into leptons and



χ

0

1, namely



t

→

b

χ



1±

→

bW∗

χ



10. The mass of the

χ



1± is set to

(

m_t

+

mχ_0

1

)/

2,andthemassdifferencebetween



t and



χ

0

1 issetto

be lessthan 80 GeV,thus bjetsareexpectedtohavea pT below

25 GeV.Fig.1showsdiagramsforthesetwosimplifiedmodels.We denotetheupperdiagraminFig.1asTChiandthelowerdiagram asT2tt.Themassesaregivenwiththemodelname,i.e. TChi150/20 (T2tt150/20)denotesa

χ



0

2-



χ

1±(



t pair)production,wherethe

pro-ducedparticleshavea massof150 GeV anda massdifferenceto theLSPof20GeV.

Weinterprettheresultsofthissearchintwovariations ofthe electroweakinomodel.Whilethemodeldescribedaboveusespure winocross sectionswiththe

χ



0

2 and

χ



1± mass degenerate,these

additional models resemble a scenario where the electroweaki-nos are of higgsino nature. The first ofthese higgsinosimplified

Fig. 1. Production

and decay of an electroweakino pair (upper) and of a

chargino-mediated t pair (lower).

models features associated

χ



0

2 and

χ



1± production and as such

corresponds to the same diagram as the one shown in Fig. 1

(upper). The second higgsino model considers associated

χ



0 2-

χ



10

production. In both cases, the mass of the chargino is given as

m_χ± 1

= (

mχ 0 2

+

mχ 0 1

)/

2, andthe

χ



0

2 decaysvia an off-shellZ

bo-son, and if applicable, the

χ



₁± decays via an off-shell W boson. Thesimplifiedmodels donotincludeanyspin correlationsinthe decays.Inthesimplifiedhiggsinomodel,thiscanleadtoa differ-entM

()

distributionthatwedonotaccountfor.

In addition to the electroweakino models, we interpret the results in a phenomenological minimal supersymmetric model (pMSSM) [51], in which the higgsino (

μ

), bino (M1), and wino

(M2) mass parameters are varied. There is only a small

depen-dencyontan

β

,whichissetto10.Allother massparametersare assumedtobedecoupled.Toreducetheparameterspacetoa two-dimensionalgrid,M2 issetto2M1.Thisconventionisinspiredby

electroweakinomassunificationatthegrandunifiedtheoryscale. Sincethefocusisonelectroweakproductiononly,thegluinomass parameter M3 isassumedtobe decoupled.Alltrilinearcouplings

arediscarded.Inthismodel,thehiggsinomassparameter

μ

is var-iedbetween100and200 GeV, while M1 variesbetween300 GeV

and1 TeV. Events for this“higgsino pMSSM” are generated with MadGraph5_amc@nlo [52]. TheNLO crosssectionsare computed usingProspino2 [53].Severaladditionalpackages[54–58] areused tocalculatemassspectraandparticledecays.

4. Objectreconstruction

Theanalysismakesuseoftheparticle-flow(PF)algorithm [59], whichreconstructsandidentifieseach individualparticlethrough an optimized combination of information from the various ele-ments of the CMS detector.The difficulties inreconstructing the event of interest, because of the presence of the large average numberofinteractionsper bunchcrossing (pileup),aremitigated by a primary vertex selection and other methods described be-low. The reconstructed vertexwith the largest value of summed physics-objectp2_Tistakentobetheprimary pp interactionvertex. Thephysicsobjectsarethejets,clusteredusingthejetfinding al-gorithm [60,61] withthetracks assignedto the vertexasinputs, andtheassociated pmiss_T , takenasthenegative vector pT sumof

thosejets.

Theleadingandsubleadingmuon(electron)arerequiredto sat-isfy pT

>

5GeV,

|

η

|

<

2

.

4 (2

.

5).A requirementof pT

<

30GeV on

the leptons is alsoapplied; thisthresholdis identified asthe pT

value below which the current analysis is more sensitive in the compressed regionscompared toother CMSanalyses.To increase the sensitivityinthecompressed massregime, thelower thresh-old on the pT of the subleading muon is set to 3.5 GeV in the

high-pmiss_T regionsofthe



t search.

Muons are required to satisfy standard identification crite-ria [62],andtobe isolated withina conein

η

–

φ

spaceofradius

R

=



(

η

)

2

_{+ ( φ)}

2

₌

₀

_.

_{3: the p}

T sumof other charged

par-ticle tracks within the cone, Isoabs, is required to be less than

5 GeV.In addition,thequantity Isorel,whichis theratioofIsoabs

andthe pT ofthe muon,isrequiredtobelessthan0.5.

Contam-ination frompileup within the isolation coneis subtracted using techniques that utilize charged particle depositswithin the cone itself [62].

Electronsfrompromptdecaysareselectedusingamultivariate discriminant basedon theenergy distributionin theshower and trackquality variables. Thelooseworkingpoint employed by the H

→

ZZ∗

→

4



analysis [63] isusedfor pT

<

10GeV,andatighter

oneforpT

>

10GeV.Thesamedefinitionofisolationandthesame

isolationcriteriaareappliedforelectronsasusedformuons. To suppress nonprompt leptons, requirements on the three-dimensionalimpactparameter [64] relative totheprimary vertex, IP3D, anditssignificance, SIP3D,are applied.Leptons arerequired

tohaveIP3D

<

0

.

01 cmandSIP3D

<

2 standarddeviations(s.d.).

Thecombinedefficiencyforreconstruction,selectionand isola-tion dependson the pT of thelepton. The efficiencies are inthe

range 70%(50%) formuons(electrons) at5 GeV,up to80% (60%) formuons (electrons)at30 GeV.

Jetsare clusteredusing theanti-kT algorithm [60] with a

dis-tance parameterof0.4 [65], asimplemented inthe FastJet pack-age [61]. The momentum of a jet, which is determined by the vectorial sum of all particle momenta in the jet, is found from simulation to be within 5 to 10% of the true momentum over the full pT spectrum and detector acceptance. An offset

correc-tionisappliedtojetenergiestotakeintoaccountthecontribution frompileup [66].Jetenergycorrectionsareobtainedfrom simula-tion, andconfirmedthrough in situmeasurements ofthe energy balance in dijet and photon

+

jet events [67]. Jets are selected with pT

>

25GeV and

|

η

|

<

2

.

4. In the following, the transverse

hadronic energy, HT, is definedas the scalar pT sumof the

se-lectedjets.

Jets arising from the hadronization of b quarks are identi-fiedthrough thecombinedsecondaryvertex(CSV)tagger [68,69], whichemploysbothsecondaryvertexandtrack-basedinformation. Inthisanalysis,alooseworkingpointcorrespondingtoab tagging efficiencyofabout80%isusedwithmisidentificationratesof10% and40%forlight-quarkorgluonjetsandforcquarkjets, respec-tively [68].

The



p_Tmiss is determined using the PF-reconstructed objects. A varietyofeventfiltersareappliedtoremovedetector- andbeam relatednoise [70].

5. Eventselection

The analysisrequirestwooppositely chargedleptons(N

=

2),

of either same (ee,

μμ

) ordifferent flavour (e

μ

), andmoderate

pmiss_T inthefinalstate,togetherwithatleastonejetintheevent. The main backgrounds arise fromevents in whichone of the leptons isnotprompt(mainly fromW

+

jets events),eventsfrom fully leptonic tt decays (tt

(

2

)

), and Drell–Yan (DY) processes withsubsequentdecays

γ

/

Z∗

→

τ τ

→ 

ν



ν



ν

τ

ν

τ .Smaller

WW andZZ∗,withZ∗

→ 

and Z

→

νν

(VV).Processessuch as ttW, ttZ, WWW, ZZZ, WZZ and WWZ as well as processes in-cluding the Higgs boson have very small contributions, and are groupedtogether as“Rare”.The following eventselection shown inTable1includesanumberofrequirementsdesignedtoreduce thesebackgrounds:

•

0

.

6

<

pmiss

T

/

HT

<

1

.

4:thiscriterioniseffectiveinrejectingSM

eventscompriseduniquelyofjetsproducedthroughthestrong interaction, referredtoasQCDmultijetevents,while remain-ing efficient foreventswithISR,asin thecaseofthe signal. Theboundsontheratiopmiss

T

/

HT isdeterminedfromastudy

of a control region (CR) atlow-pmiss

T andwith dimuonmass

closetothat oftheJ/

ψ

meson. Thisrequirementrejects such eventswhileleavingthesignalunaffected.

•

b jet eventveto: requiring events where no jet is tagged as originating from b quarks significantly reduces the tt back-ground in which b jets originate fromthe decay of the top quarks. This requirement is applied to all jets with pT

>

25GeV and usesthe b taggingselection criteriadescribed in Section4.Theefficiencyforapotentialsignalfrom



t decaysis notaffectedsignificantlysinceinthecompressed



t-LSPmodel, theb jetsareexpectedtohavesmallpTandarethereforenot

tagged.

•

M

(

τ τ

)

<

0 orM

(

τ τ

)

>

160GeV:thisrequirementonthe esti-mate of the ditau mass isdesigned to reject the large back-ground from Z

→

τ τ

decays, with the

τ

leptons decaying leptonically. Thequantity M

(

τ τ

)

[22] iscomputedasfollows: since the

τ

leptons from the decayof a Z boson have large

pTcomparedtotheirmass,thedirectionoftheoutgoing

lep-ton is approximately the same as that of the

τ

lepton (i.e.

R

(,

τ

)

≈

0).Themagnitudesoftheleptonmomentum vec-tors are then rescaled so that the lepton pair balances the hadronicrecoil.ForZ

→

τ τ

events,thisleads toafairlygood approximation oftheoriginal

τ

momenta.Theinvariantmass of the two

τ

leptons, M

(

τ τ

)

, is estimated by the invariant mass ofthetwoscaled leptons. Insomeevents,the estimate of the magnitudeof the

τ

momentum results in a negative value whentheflightdirectionisopposite tothedirectionof thelepton.Insuchcases,M

(

τ τ

)

issettoitsnegativevalue.

•

MT

(

i

,

pmissT

)

<

70GeV,fori

=

1

,

2:thetransversemassMTis

definedas MT

(,

pmissT

)

=



2p TpmissT



1

−

cos



φ



,

pmiss_T



,

and



1 and



2 aretheleadingandsubleadingleptons,

respec-tively. For the signal, the leading lepton is typically aligned withtheboostdirectionoftheLSP(

φ (,

pmiss_T

)

≈

0).This re-quirementiseffectiveinfurthersuppressingthett background for the electroweakino search, butnot for the



t search. It is thereforeonlyappliedintheelectroweakinosearch.

•

J

/ψ

, and

ϒ

veto:to suppressbackground contributions from J

/ψ

,low-mass

γ

∗,and

ϒ

decays,thedileptoninvariantmass

M

()

isrequiredtosatisfy M

()

>

4GeV andtoalsolie out-sidetherange9

<

M

()

<

10

.

5GeV.Thisvetoisonlyapplied tosameflavourleptonpairs.

•

pmiss

T

>

125GeV: to ensure high trigger efficiency, both the

pmiss_T andthemuon corrected pmiss_T ,whichiscomputedfrom the vectorial sumof the pmiss_T and the pT ofthe muons

se-lected intheevent,isrequiredtobelargerthan125 GeV.The region 125GeV

<

pmiss_T

<

200 GeV is only accessible by the dimuon trigger andtherefore only dimuon pairs are consid-ered.Theregion pmiss

T

>

200 GeV includesalsoelectrons.

•

Trigger acceptance:in theonlineselection, thelepton pairis

requiredtohaveasmallboostofpT

>

3GeV,togetherwithan

Table 1

Common selection requirements for the signal regions. The subleading lepton pT

threshold is reduced to 3.5 GeV for muons in the high-pmiss

T , t-like signal region.

Variable SR selection criteria

N 2 (μμ,μe, ee)

q(1)q(2) −1

pT(1), pT(2) [5,30]GeV pT(μ2)for high-pmissT t-like SR [3.5,30]GeV

|ημ| <2.4

|ηe| <2.5

IP3D <0.01 cm

SIP3D <2

Isorel(1,2) <0.5

Isoabs(1,2) <5 GeV

pT(jet) >25 GeV

|η|(jet) <2.4

Nb(pT>25 GeV, CSV) 0

M() [4,9]or[10.5,50]GeV (forμμand ee)

pT() >3 GeV

pmiss

T >125 GeV (forμμ)

>200 GeV (forμe, ee)

pmiss

T (muon corrected) >125 GeV (forμμ)

>200 GeV (forμe, ee)

pmiss

T /HT [0.6,1.4]

HT >100 GeV

M(τ τ) veto[0,160]GeV

MT(i,pmissT ),i=1,2 <70 GeV (electroweakino selection only)

upperbound on thedimuon invariant mass M

()

<

60GeV, to limit the trigger rate. To remain fully efficient after of-flinereconstruction,anupperboundof50 GeV onM

()

and a lower requirement on the dilepton transverse momentum

pT

()

>

3GeV areimposed.

•

HT

>

100GeV:thisrequirementsuppressesbackgroundswith

lowhadronicactivityintheevent.

Fortheselectedevents,asetofSRsare defined,basedon the dileptoninvariantmassandpmiss

T .Foreventswithleptonsofsame

flavourandoppositecharge,fourSRsaredefinedin M

()

ranges of 4–9, 10.5–20, 20–30, and 30–50 GeV. These SRs are intended for searches for



χ

0

2

→

Z∗

χ



10 events, where M

()

is related to

themass differencebetweenthetwo electroweakinos.Forevents with leptons of different flavour and opposite charge, three SRs are definedin the leading lepton pT ranges of 5–12, 12–20, and

20–30 GeV. Thedefinition ofthebins oftheSRs canbe found in Table2.

Toexploitthepotentialofthedimuonplus pmiss_T trigger,events areseparatedaccordingtothevalueofpmiss_T :intotalthreeranges areusedforthesignalregions,namely pmiss

T

∈

125–200,200–300,

and

>

300 GeV forthe



t search,andpmiss

T

∈

125–200,200–250,and

>

250 GeV for the electroweakino search. Since the low-pmiss_T re-gioncontainseventsaccessibleonlyviathedimuon

+

pmiss_T trigger, only

μμ

pairs are considered. The muons need to be of oppo-site charge. Conversely, in the high-pmiss_T regions, both electron and muon flavours are considered. The electroweakino SRs are populated by ee and

μμ

pairs,where theleptons are oppositely charged.Forthe



t SRs,e

μ

pairsarealsoconsidered.Forthelatter, the pT thresholdon the trailinglepton is reducedto 3.5 GeV for

muonsinthehigh-pmiss_T regiontogainsensitivityinthesearchfor



t signal.

TheacceptancetimesefficiencyforthesignalmodelTChi150/20 (T2tt350/330) in the electroweakino (stop) selection is between 3

×

10−5(3

×

10−5)and7

×

10−5(15

×

10−5).Theefficiencytimes acceptanceformuonsisabout2to5timeshigherthan for elec-trons in the electroweakino selection and about 1.5 to 3 times higherinthestopselection.

Table 2

Definition of bins in the two SRs. The lowest pmiss

T region includes only muon pairs, since it is only accessible by

the dimuon trigger.

Electroweakino search region t search region

pmiss T [GeV] M()[GeV] p miss T [GeV] p lepton T [GeV] [125,200] [4,9] [125,200] [5,12] [12,20] [20,30] [10.5,20] [20,30] [30,50] [200,250] [4,9] [200,300] [ 5,12] [12,20] [20,30] [10.5,20] [20,30] [30,50] >250 [4,9] >300 [5,12] [12,20] [20,30] [10.5,20] [20,30] [30,50] Table 3

Summary of changes in selection criteria relative to Table1for CRs and the VV validation region (VR).

DY CR tt (2) CR VV VR

No upper requirement on pT()

Isorel<0.1 as an or condition with the SR isolation

0 <M(τ τ)<160 GeV

IP3D<0.0175 cm, SIP3D<2.5 s.d.

pT(1)>20 GeV, or IP3D>0.01 cm, or SIP3D>2 s.d. MTas for electroweakino SR

No requirements on MT

At least one b-tagged jet with pT>40 GeV

pT(1)>20 GeV

|same flavour M()−M(Z)|>10 GeV

MT>90 GeV

6. Backgroundestimation

BackgroundswithtwopromptleptonsareestimatedusingCRs chosentobemostlyfreefromsignalbutwhenpossible,with sim-ilar kinematic characteristics as the events in the signal regions. Different CRsare employed foreach SM process that contributes significantly to the signal region, i.e. the tt dilepton background andthe DY

+

jets background.The normalization of the diboson backgroundiscrosscheckedinavalidationregion(VR).

Foreachbackground,thenumberofeventsineachSRis esti-matedusingthenumberofeventsobservedinthecorresponding CR,andatransferfactorthatisusedtodescribetheexpectedratio ofeventsintheSRandCRfortheprocess inquestion.The trans-ferfactorforaspecificprocess,Fprocess,isdeterminedfromMonte

Carlo(MC)simulationoftheprocessthroughtheratio

Fprocess

=

NSR_{MC process} NCR_{MC process}

.

Sincea CRtypically contains contributions fromother physics processes,they needto besubtracted fromtheobserved number ofeventsintheCR, NCR_data.Thesecontributions,N_{MC other}CR ,aresmall compared to the main process for which the CRis defined, and arethusestimatedusingMCsimulation.Theestimateofthe back-groundfromaspecificphysicsprocessintheSRisthengivenby

NSR_process

=



NCR_data

−

N_{MC other}CR



Fprocess

.

SystematicuncertaintiesinthevalueofFprocess areincludedwhen

determiningthe fulluncertaintyin NprocessSR .Thetotal background

inthe SRis givenasthe sumofthe backgrounds expectedfrom eachprocess.

The differentCRsare split into two pmiss_T bins: The low pmiss_T

binwithpmiss_T between125and200 GeV isusedtoconstrainthe SRswiththesamepmiss

T range,whilethehighpmissT binwithpmissT

>

200 GeV is used toconstrain all SRs with pmiss

T above 200 GeV.

The shapes for M

()

andthe lepton pT are taken directlyfrom

simulation. Asummary ofallCRs forpromptlepton backgrounds isgiveninTable3.Forthedibosonbackground,avalidationregion enriched inVV(mainly WW events)isadded.Thisregionisused toestablishhowwell thesimulationagreeswithdatainorderto validatetheuncertaintyassignedtothedibosonsimulation.About halfoftheeventsinthisregionstemfromVV.

6.1. TheDY

+

jetscontrolregion

The main difference between the CR for the DY

+

jets back-ground andthe SRlies in therequirement imposed on the Mττ

variable; the CR consistsofevents that are vetoed in the SR se-lection, namely those events with Mττ in the range 0–160 GeV. To increase the efficiency for leptons from

τ

decays, the im-pactparameterrequirementsarerelaxedtoIP3D

<

0

.

0175 cm and

SIP3D

<

2

.

5 s.d.The variation ofthescale factors appliedto

sim-ulation by changing thecuts onIP3D and SIP3D was found to be

negligible.Inaddition,the30GeV upperboundonthelepton pTis

removed,andtheregionwithleptonpT

<

20GeV,IP3D

<

0

.

01 cm,

andSIP3D

<

2 isalsoremovedtoreducethepresenceofpotential

signal.Thedistributionsinkinematicquantitiesoftheseevents, in-cludingthevariablesusedtodefine thesignal regions,M

()

and theleadinglepton pT,arewell describedinsimulation.Theevent

yieldsestimatedfromsimulationandtheobservedeventyieldsare listedinTable4.

6.2. Thett (2



)controlregion

To obtain a sample enriched in tt events, at least one jet is required tobe identified asoriginating fromb quarks. Toreduce potentialsignalcontamination,theleadingb-taggedjetisrequired to satisfy pT

>

40GeV.To increase the number of events in the

CR, whilestillavoidingpotentially largesignalcontamination, the upper bound onthe lepton pT is alsoremoved. The eventyields

estimatedfromsimulationandtheobservedeventyields arealso showninTable4.

Table 4

Data and simulation yields for the DY and tt (2) CRs, corresponding to integrated luminosities of 35.9 fb−1(high-pmiss

T region) and 33.2 fb− 1

(low-pmiss

T region). The SR scale

factors are derived by subtracting the other processes from the observed data count, and dividing this number by the expected event yields from simulation for the process in question. The uncertainties are statistical only.

pmiss

T DY CR tt (2) CR

125–200 GeV >200 GeV 125–200 GeV >200 GeV

DY+jets or tt 70.1±5.1 64.5±3.3 1053.7±9.4 535.7±7.1

All SM processes 82.6±5.5 75.2±3.6 1170.0±11.0 710.4±11.1

Data 84 75 1157 680

SR scale factor 1.02±0.13 0.99±0.13 0.99±0.03 0.94±0.05

6.3.Nonpromptbackground

The background from nonprompt or misidentified leptons is evaluated using a “tight-to-loose” method. Events where atleast one lepton fails the tight identification andisolation criteriabut passesalooserselectiondefinethe“applicationregion”.Eventsin thisregion areweighted bya transferfactorbasedonthe proba-bilitythatnonpromptleptons passingthelooserequirementsalso satisfy thetight ones. The resultingestimate is corrected forthe presenceofpromptleptonsintheapplicationregion.

Theprobabilityfornonpromptormisidentified leptonstopass the tight selection criteriais referred to as the misidentification probability,whichisdeterminedasafunctionoflepton pT and

η

.

This probability is measured using a dedicated data sample, the “measurementregion”(MR),whichisenriched inthebackground fromSMevents containingonly jetsproducedvia strong interac-tion, referred to as QCD multijet events. This method has been used in severalmultilepton analyses atCMS and is described in moredetailin Ref. [71]. The MRisdefinedthrough thepresence of one loose lepton, obtained by relaxing the isolation and im-pactparameterrequirements,andthroughajetwithpT

>

30GeV,

separatedfromtheleptonby

R

>

0

.

7.Formuons,eventsare se-lected through prescaled single-lepton triggers with no isolation requirements.Forelectrons, a mixtureofprescaledjet triggers is used.Themethodincludesacorrectionforthepresenceofprompt leptonsintheMR,mostlyduetoW andZ bosonproductionin as-sociationwithjets. Theprobabilityforpromptleptonstopassthe tight selection criteria is taken fromsimulation and is corrected withadata-to-simulationscalefactorextractedfromdataenriched inZ

→ 

decays.

In thisanalysis, the misidentification probability measured in QCD multijet events is applied to loosely identified leptons in eventsthataredominatedbyW

+

jets andtt production.The lat-tercanhavebothadifferentcompositionintermsoftheflavourof thejetsthatgiverisetothenonpromptleptons,aswellasdifferent kinematicproperties, potentially resulting in a differenteffective misidentification probability. These effects are studied by com-paring the misidentification probabilities measured in simulated events of these two processes in the kinematic regions probed bythisanalysis.Aclosuretest isthenperformedbyapplyingthe misidentificationprobabilitymeasuredintheQCDsimulated mul-tijet events to a sample of W

+

jets events.The yield of events passing the tight identification criteriais compared with the es-timate obtained by applying the misidentification probability to eventsintheapplicationregion. Themethodis foundto be con-sistentwithin alevel of

<

40%; thisvalue isusedasasystematic uncertaintyinthe estimate ofthe normalizationofthereducible background.

Tofurtherconstrainthe contributionofthe nonpromptlepton backgroundintheSR, adedicatedCRconsistingofsame-sign (SS) leptonsisdefined.Requiringthetwoleptoncandidatestohavethe samesignincreasessignificantly the probabilitythat atleast one ofthe two is a nonprompt ormisidentified lepton. The SS CRis definedusingthe



t selectioninthe pmiss

T

>

200GeV region,where

Fig. 2. Same-sign

CR for

t selection and pmiss

T >200 GeV. The distribution of the

leading lepton pT is used as input to the final signal extraction. A signal from

neutralino–chargino (χ0

2–χ1±) production is superimposed.

the opposite charge requirement of the two leptons is modified to same-sign.In theSSCR, theprediction ofthenonprompt lep-ton backgroundis derived from the “tight-to-loose” method and agreeswiththedata.Fig.2showstheleadinglepton pT

distribu-tionintheSSCR. Italsoshowsthenearabsenceofasignal.The distributionofthe leading lepton pT isusedasinputto thefinal

fit that performs the signal extraction, as its constraining power is significant, given the significant uncertainty on the measured misidentificationprobability.

7. Systematicuncertainties

Thissectionsummarizesthesystematicuncertaintiesinthe es-timateofthebackgroundfromthevariousSMprocesses.Foreach sourceofsystematicuncertainty,wepresentboththeeffectonthe correspondingspecificbackgroundandtheoveralleffectonthe to-talbackgroundpredictionsarelistedinTable5.

Theuncertaintyinthepredictednonpromptleptonbackground contains astatisticalcomponentduetothe statisticaluncertainty in theapplication region eventyield, it rangesfrom 10% to 50%. When applied in the SR, the uncertainty is 4% to 20%. Another source of statistical uncertainty arises from limited statistics in data and simulationin the DY

+

jets andtt (2



) CRs. The effect onthepredictedyields intheSR, obtainedusingthetransfer fac-tordescribed inSection6,isapproximately13%fortheDY

+

jets backgroundand3%forthett background.

Table 5

Relative uncertainties in the final total background predictions for each individual systematic source of uncertainty.

Systematic source of uncertainty Typical uncertainty (%)

VV background normalization 3–25

Nonprompt lepton background normalization 4–20

DY+jets background normalization 4–20

tt background normalization 2–8

Rare background normalization 1–3

Jet energy scale 2–12

b tagging 2–6 Pileup 1–5 Lepton selection 1–4 Integrated luminosity 2.5 Trigger 1–2 tt modelling <1

Forthett background,we haveconsidereda setofsystematic uncertainties arising from the modelling ofthe kinematic distri-butions in thesimulation ofthis process. The spin correlation of thetop quarks hasbeen varied by20%, based onthe ATLAS and CMS [72,73] measurements and a comparison between different generators (MadGraph5_amc@nlo versus powheg). The helicity amplitudesoftheW bosonintopquarkdecayshavebeenvaried by5%.Atopquark pTmodellinguncertaintyhasalsobeenderived

by reweighting the simulated tt events based on the number of ISRjets(NISR_jets), soastomakethejetmultiplicityagreewithdata. Thereweightingfactorsrangefrom0.92to0.51forNISR_jets between 1and6.Thesystematicuncertaintyinthesereweightingfactorsis takento beequal toone half ofthedeviation ofthefactor from unity.Thecombinedeffectofthissetoftt modellinguncertainties onthetotalnumberofpredictedtt backgroundeventsisfoundto beintherange3–5%.

FortheDY

+

jetsbackground,theuncertaintyintheresolution of the pT ofthe system recoiling against the two leptons is

ob-tained fromdata dominatedby Z

→

μμ

events. The uncertainty affectsthe DY estimate,which usesthe efficiencyofthe require-mentsonMττ fromsimulation.Theeffectontheestimatedyields ofDY

+

jetsisfoundtobenegligible(

<

1%).

As presented in Section 6, the method used to estimate the backgroundfromnonpromptandmisidentified leptons leads toa 40%uncertaintyonthenormalization. Intheglobalfitthis uncer-taintyisreducedto 25%.

A50%uncertaintyisassignedforthedibosonbackground nor-malization,whichischeckedinthededicatedregiondescribedin Section 6.In thisregion, whichis enriched in W W eventswith similar kinematicproperties astheevents inthe SR, the simula-tionisfoundtoagree,withinthegivenuncertainty,withthedata. Aconservative 100% uncertaintyisassigned tothe very small rarebackgroundsthataredominatedbythetW process.

Theexperimentaluncertaintiesrelatedtob tagging,trigger, lep-ton reconstruction,identification, andisolation criteriahavebeen propagatedandtheireffectonthefinalresultsrangesfrom2%up to12%.Thejetenergyscalecorrections(JEC)areappliedtomatch jetenergiesmeasuredindataandsimulation.TheJECareaffected byanintrinsicuncertainty,whichaffectsallsimulatedbackground, leadingtotypically2–12%uncertaintiesinthefinalpredictions.

Anuncertaintyof2.5%isassignedtotheintegratedluminosity measuredbyCMSforthe2016datatakingperiod [74].Thisaffects theestimateoftherareSMbackgroundsthatrelyonthemeasured dataluminosity.

Finally,theuncertaintyrelatedtopileuphasbeenestimatedby varying theminimum-biascrosssection by

±

5% andreweighting the pileup distributionaccordingly. The systematic uncertaintyis foundtobeintherange1–5%.

As thesignal yields arefromsimulation,additionalsystematic uncertainties are applied in two categories. One arises from the systematic uncertainty in the inclusiveNLO

+

NLL [48–50] cross section used for the normalization, determined by varying the renormalizationand factorizationscales andthePDF. The depen-dence on these QCD scales yields a total uncertainty of 3%. The other category arisesfrom the uncertaintyin the product of the signalacceptanceandefficiency.

ItisimportanttoproperlymodeltheISRthatleadstotheboost ofthe producedSUSY particlesinthe transverseplane. In partic-ular, forthe electroweakinobenchmark, themodelling oftheISR with MadGraph5_amc@nlo affectsthetotaltransversemomentum

pISR_T of the system ofSUSY particles, which can be improved by reweighting pISR_T inthesimulatedsignalevents.Thisreweightingis basedon pTstudiesofeventscontainingaZ boson [75],inwhich

the factors range between 1.18 at pISR_T of 125 GeV, and0.78 for

pISR_T

>

600GeV.Thedeviationfrom1.0istakenasthesystematic uncertaintyofthereweightingprocedure.Forthe



t benchmarkto improve themodelling ofthe multiplicity ofadditional jetsfrom ISR,theeventsarereweightedbasedonthe NISR_jets,usingthesame correctionsusedforthetopbackgroundasdescribedearlierinthis section.The typicaluncertaintiesonthefinalresultsfromtheISR modellingarefoundtobeintherange2–7%.

Weaccountfordifferencesobservedinpmiss_T reconstruction ef-fectsinfull andfastsimulationusedforsignal. Theuncertainties varybetween3and5%.Theuncertaintiesrelatedtopotential dif-ferencesin b taggingbetweenthefullandfastsimulationandin theJECvaryintherange1–2%.

These uncertainties, together with those related to the pre-dicted backgrounds described in Section 6, are included as log-normaldistributednuisanceparametersinthelikelihoodapproach.

8. Results

The estimatedyields oftheSM backgroundprocesses andthe data observed inthe SRs are shownin Figs. 3and 4.No signifi-cant excess hasbeen observed.The estimates in theSR bins are extractedfromamaximumlikelihoodfitofthedatausingthe ex-pectedyieldsdescribedinSection6,namelytheDY

+

jets,tt (2



), andSSCRs.Log-normaldistributions fornuisanceparameters are usedtodescribethesystematicuncertaintiesofSection7.The un-certaintiesinthepredictedyieldsquotedinthefollowingarethose determinedfromthefit.

The predictedyieldsalong withthedataare alsosummarized in Tables 6and7 foreach bin ofthe SR.Thetotal uncertaintyin theyieldforeachSMprocessincludesthesystematicand statisti-cal uncertaintiesdescribed inSection7,addedinquadrature.The largest deviationfromthe SMexpectationisseen ina binofthe electroweakino search region.Thebinwith pmiss_T

∈ [

200

,

250

]

GeV andM

()

∈ [

10

.

5

,

20

]

GeV has3

.

5

±

0

.

9 expectedeventsbut0 ob-served. Thesmaller numberofeventsobservedinthisbindrives the observedexclusionto highervaluesthanexpected, ascan be seeninthenextsection.Overall,thereisgoodagreementbetween expectationandobservation.

9. Interpretation

The results are interpreted in terms of the simplified mod-els withcompressedmass spectrafor



χ

0

2



χ

1±

→

Z∗W±∗



χ

10



χ

10 and

for



t



t

→

b



χ

₁±b

χ



₁∓ with the subsequent decay



χ

₁±

→

W±∗



χ

0 1 as

discussed inSection 3. A binned likelihood fit of signal and the backgroundexpectationstothedataisperformed.Thisfittakesas input theyields intheSRs (12fortheelectroweakino interpreta-tion and9forthetop squarkinterpretation),together withthose

Fig. 3. Left:

electroweakino search regions in bins of

M()for 125 <pmiss

T <200 GeV (muon only channel) for 33.2 fb−

1_{; middle: 200}

<pmiss

T <250 GeV (muon and electron

channel) for 35.9 fb−1_{; right: p}miss

T >250 GeV (muon and electron channel) for 35.9 fb−

1_{. A signal from neutralino–chargino (χ}0

2–χ1±) production is superimposed. The gap

between 9 and 10.5 GeV corresponds to the ϒveto.

Fig. 4. Left:t search regions in bins of leading lepton pTfor 125 <pmissT <200 GeV (muon only channel) for 33.2 fb−

1_{; middle: 200}

<pmiss

T <300 GeV (muon and electron

channel) for 35.9 fb−1_{; right: p}miss

T >300 GeV (muon and electron channel) for 35.9 fb−

1_{. A signal from t pair production is superimposed.}

inthe two CRs (125

<

pmiss_T

<

200GeV and pmiss_T

>

200GeV) for thett andDY

+

jetsestimates,andthethreepT binsforsame-sign

leptons forthe pmiss_T

>

200GeV CR.These background-dominated bins also help to constrain the uncertainties in the background taken from simulation and the one predicted by the “tight-to-loose”method.

Upper limits on the cross sections in the benchmark models at95%confidencelevel(CL)areextracted.Weuseasymptotic for-mulae [76] to derive the results. To set limits, the CLs criterion,

as described in [77,78], is used. Figures 5 and 6 show the ob-servedandexpectedupperlimitsontheelectroweakinoand



t pair production cross sections for the benchmarks considered in this search.

Fortheelectroweakino simplified model,the productioncross sections are computed at NLO

+

NLL precision in the limit of a massdegeneratewino

χ



0

2 and



χ

1±,alightbino



χ

10,andassuming

allotherSUSYparticlestobeheavyanddecoupled [48–50].Masses

of

χ



₂0upto230 GeV fora

m

(

χ



₂0

,

χ



₁0

)

of20 GeV areexcluded.The existenceof



t massesupto450 GeV witha

m

(



t

,

χ



0

1

)

of40 GeV is

ruledoutforthisspecificmodel.

Theexpectedandobservedexclusioncontoursforthehiggsino pMSSM are shown in Fig. 7. The higgsino mass parameter

μ

is excluded up to 160 GeV, when the bino mass parameter M1 is

300 GeV and thewino massparameter M2 is 600 GeV.For larger

values of M1 and M2, the mass splitting

m

(

χ



₂0

,

χ



₁0

)

becomes

smallerandthesensitivityisreduced.ForM1

=

700GeV,

μ

is

ex-cludedupto100 GeV.

Fig.8showstheexpectedandobservedexclusioncontoursand upperlimits oncross sectionsat95% CL ina higgsinosimplified model. To calculate the cross sections in this model, a scan in

|

μ

|

, M1, M2 andtan

β

iscarriedout. Allparametersare required

to be real, M2 to be positive and tan

β

∈ [

1

,

100

]

. The

remain-ingSUSYparticlemassesaredecoupled,andalltrilinearcouplings arediscarded.Theparameterspaceisthenscannedtoachievethe

Table 6

The number of events observed in the data and the result of the fit of the backgrounds to the data in the electroweakino search regions. The uncertainty indicated is determined from the fit to the 33.2 and 35.9 fb−1_{integrated luminosities. Values for the}

M()ranges are in GeV. Rare background event yields are omitted when they do

not contribute to the SR bin.

125<pmiss T <200 GeV 4<M() <9 10.5<M() <20 20<M() <30 30<M() <50 tt(2) 0.23±0.16 1.9±0.52 2.80±0.65 3.60±0.75 DY+jets 0.83±0.63 3.7±1.5 4.9±1.5 1.60±0.99 VV 0.82±0.48 0.71±0.65 1.7±1.0 2.2±1.2 Nonprompt lepton 1.7±0.7 5.7±1.5 7.5±1.7 3.3±1.1 Rare — 0.46+0.64 −0.45 — 0.33+ 0.49 −0.32 Total SM prediction 3.5±1.0 12.0±2.3 17.0±2.4 11.0±2.0 Data 2 15 19 18 200<pmiss T <250 GeV 4<M() <9 10.5<M() <20 20<M() <30 30<M() <50 tt(2) 0.21±0.17 0.38±0.18 0.11+0.11 −0.10 — DY+jets 0.69±0.62 0.67±0.32 0.42±0.27 — VV 0.26+0.28 −0.25 0.29+ 0.32 −0.28 0.42±0.33 0.33±0.29 Nonprompt lepton 0.44±0.32 2.0±0.7 1.0±0.6 0.03+₋00..1402 Rare — 0.14+0.39 −0.13 — 0.17+ 0.37 −0.16 Total SM prediction 1.6±0.7 3.5±0.9 2.0±0.7 0.51+0.52 −0.50 Data 1 0 3 1 pmiss T >250 GeV 4<M() <9 10.5<M() <20 20<M() <30 30<M() <50 tt(2) — 0.19±0.14 0.091±0.091 0.27±0.14 DY+jets 0.24±0.19 0.24±0.17 0.17±0.16 0.014+0.019 −0.013 VV 0.43±0.35 0.29+0.29 −0.28 0.41±0.29 0.66±0.45 Nonprompt lepton 0.28₋+00..3327 0.77±0.44 0.38±0.30 0.23±0.18 Rare 0.45+0.57 −0.44 — 0.49+ 0.62 −0.48 0.04+ 0.28 −0.03 Total SM prediction 1.4±0.7 1.5±0.6 1.5±0.8 1.2±0.6 Data 2 1 2 0 Table 7

The number of events observed in the data and the result of the fit of the backgrounds to the data in the t search regions. The uncertainty indicated is determined from the fit to the 33.2 and 35.9 fb−1_{integrated luminosities. Values for the p}

T(1)ranges are in GeV. Rare background event yields are omitted when they do not contribute to the

SR bin. 125<pmiss T <200 GeV 5<pT(1) <12 12<pT(1) <20 20<pT(1) <30 tt(2) 1.9±0.4 11.0±1.9 23.0±3.5 DY+jets 2.9±1.4 5.6±1.9 4.6±1.7 VV 0.8±0.7 4.9+₋46..38 9.4±5.4 Nonprompt lepton 8.5±1.9 15.0±2.6 15.0±2.6 Rare 0.10+₋00..1609 0.93+ 1.0 −0.92 1.8±1.7 Total SM prediction 14.0±2.3 37.0±6.8 54.0±6.5 Data 16 51 67 200<pmiss T <300 GeV 5<pT(1) <12 12<pT(1) <20 20<pT(1) <30 tt(2) 1.3±0.35 9.9±1.2 15±2.2 DY+jets 0.92±0.83 2.4±0.9 1.6±0.6 VV 2.5±1.4 7.1±4.0 12.0±6.2 Nonprompt lepton 18.0±3.2 20.0±3.4 15.0±2.7 Rare 0.52₋+00..5451 1.96±1.46 1.45±1.13 Total SM prediction 23.0±3.5 41.0±5.6 45.0±7.0 Data 23 40 44 pmiss T >300 GeV 5<pT(1) <12 12<pT(1) <20 20<pT(1) <30 tt(2) 0.39±0.25 1.6±0.5 1.6±0.4 DY+jets 0.33±0.26 0.28±0.18 0.19±0.07 VV 0.93±0.53 2.5±1.4 4.2±2.2 Nonprompt lepton 3.1±1.1 5.6±1.3 4.0±1.3 Rare — 0.15+0.18 −0.14 0.45+ 0.50 −0.44 Total SM prediction 4.7±1.3 10.0±1.9 10.0±2.5 Data 4 11 9

Fig. 5. The

observed 95% CL exclusion contours (black curves) assuming the NLO

+

NLL cross sections, with the variations corresponding to the uncertainty in the cross section for electroweakino. The dashed (red) curves present the 95% CL expected limits with the band covering 68% of the limits in the absence of signal. Results are based on a simplified model of χ0

2χ1±→Z∗W∗χ10χ10process with a pure wino

pro-duction cross section. (For interpretation of the colours in the figure(s), the reader is referred to the web version of this article.)

Fig. 6. The

observed 95% CL exclusion contours (black curves) assuming the NLO

+ NLL cross sections, with the variations corresponding to the uncertainty in the cross section for t. The dashed (red) curves present the 95% CL expected limits with the band covering 68% of the limits in the absence of signal. A simplified model of the t pair production, followed by the t→bχ1±and the subsequent χ1±→W∗χ10

decay is used for the t search. In this latter model, the mass of the χ1±is set to be

(mt+mχ0 1)/2.

maximum higgsinocontent for

χ



₂0,



χ

₁±, and

χ



₁0 [79]. For a

m

between15and20 GeV,theproductionmodelofpp

→ 

χ

0 2

χ



1±and

pp

→ 

χ

0

2

χ



10 isexcludedformassesupto



χ

20

∼

167GeV. 10. Summary

A search is presented for new physics in events with two low-momentum leptons of opposite charge and missing trans-verse momentum in data collected by the CMS experiment at a centre-of-mass energy of 13 TeV, corresponding to an integrated

Fig. 7. The

observed 95% CL exclusion contours (black curve) assuming the NLO cross

sections, with the variations corresponding to the uncertainty in the cross sections for the higgsino pMSSM, which has been introduced in the text. The dashed (red) curves present the band covering 68% of the limits in the absence of signal. The model considers all possible production processes.

Fig. 8. The

observed 95% CL exclusion contours (black curves) assuming the NLO

+

NLL cross sections, with the variations corresponding to the uncertainty in the cross sections for the higgsino simplified models. The dashed (red) curves present the expected limits with the associated band covering 68% of the limits in the absence of signal.

luminosity of up to 35.9 fb−1. The data are found to be con-sistent with standard model expectations. The results are inter-preted in the framework of supersymmetric simplified models targetingelectroweakinomass-degeneratespectraand



t-



χ

0

1

mass-degeneratebenchmarkmodels.Forthe



t chargino-mediateddecay into bW∗

χ



0

1, top squark masses of up to 450 GeV are excluded

ina simplified modelfor

m

(



t

,



χ

0

1

)

=

40GeV.The search further

probesthe

χ



0

2

χ



1±

→

Z∗W∗

χ



10

χ



10 processformassdifferences(

m)

between

χ



0 2 and

χ



0

1 oflessthan 20 GeV. Assumingwino

produc-tioncrosssections,

χ



0

2 massesupto230 GeV areexcludedfor

m

of20 GeV. The search is alsosensitive to higgsinoproduction; in a simplified higgsino model,



χ

0

2 masses up to 167 GeV are

ex-cluded for

m of 15 GeV, while in a higgsino pMSSM, limits in thehiggsino-binomassparameters

μ

–M1planeareextracted.

Acknowledgements

WecongratulateourcolleaguesintheCERNaccelerator depart-ments for the excellent performance of the LHC and thank the technicalandadministrativestaffs atCERN andatother CMS in-stitutes for their contributions to the success of the CMS effort. Inaddition,wegratefullyacknowledgethecomputingcentresand personneloftheWorldwideLHCComputingGridfordeliveringso effectivelythe computinginfrastructureessential to ouranalyses. Finally, we acknowledge the enduring support for the construc-tionandoperation oftheLHCandthe CMSdetectorprovidedby thefollowingfundingagencies:BMWFWandFWF(Austria);FNRS and FWO (Belgium); CNPq, CAPES, FAPERJ, and FAPESP (Brazil); MES (Bulgaria); CERN; CAS, MoST, and NSFC (China); COLCIEN-CIAS(Colombia);MSESandCSF(Croatia);RPF(Cyprus);SENESCYT (Ecuador); MoER, ERC IUT, and ERDF (Estonia); Academy of Fin-land,MEC,andHIP(Finland);CEAandCNRS/IN2P3(France);BMBF, DFG, and HGF (Germany); GSRT (Greece); OTKA and NIH (Hun-gary);DAEandDST(India);IPM(Iran);SFI(Ireland);INFN(Italy); MSIPandNRF(RepublicofKorea);LAS (Lithuania);MOE andUM (Malaysia); BUAP, CINVESTAV,CONACYT, LNS, SEP, and UASLP-FAI (Mexico); MBIE (New Zealand); PAEC (Pakistan); MSHE and NSC (Poland);FCT(Portugal);JINR(Dubna);MON,RosAtom,RAS,RFBR andRAEP(Russia);MESTD (Serbia);SEIDI,CPAN,PCTI andFEDER (Spain);SwissFundingAgencies(Switzerland);MST(Taipei); ThEP-Center, IPST, STAR, and NSTDA (Thailand); TUBITAK and TAEK (Turkey);NASUandSFFR(Ukraine); STFC(United Kingdom);DOE andNSF(USA).

Individuals have received support from the Marie-Curie pro-gramme and the European Research Council and Horizon 2020 Grant,contract No. 675440 (EuropeanUnion);theLeventis Foun-dation;the A. P. Sloan Foundation; theAlexander von Humboldt Foundation; the Belgian Federal Science Policy Office; the Fonds pourlaFormationàlaRecherchedansl’Industrieetdans l’Agricul-ture(FRIA-Belgium); the Agentschapvoor Innovatie door Weten-schap en Technologie (IWT-Belgium); the Ministry of Education, YouthandSports(MEYS)oftheCzechRepublic;theCouncilof Sci-enceandIndustrialResearch,India;theHOMINGPLUSprogramme of the Foundation for Polish Science, cofinanced from European Union,Regional DevelopmentFund,the MobilityPlusprogramme oftheMinistryofScienceandHigherEducation,theNational Sci-ence Center (Poland), contracts Harmonia 2014/14/M/ST2/00428, Opus2014/13/B/ST2/02543,2014/15/B/ST2/03998,and2015/19/B/ ST2/02861, Sonata-bis 2012/07/E/ST2/01406; the National Priori-tiesResearchProgram byQatar NationalResearch Fund;the Pro-grama Severo Ochoa del Principado de Asturias; the Thalis and Aristeia programmes cofinanced by EU-ESF andthe Greek NSRF; theRachadapisekSompotFundforPostdoctoralFellowship, Chula-longkornUniversityandtheChulalongkornAcademicintoIts 2nd CenturyProjectAdvancement Project(Thailand);theWelch Foun-dation,contractC-1845;andtheWestonHavensFoundation(USA).

References

[1] J. Wess, B. Zumino, Supergauge transformations in four dimensions, Nucl. Phys. B 70 (1974) 39, https://doi.org/10.1016/0550-3213(74)90355-1.

[2] H.P. Nilles, Supersymmetry, supergravity and particle physics, Phys. Rep. 110 (1984) 1, https://doi.org/10.1016/0370-1573(84)90008-5.

[3] H.E. Haber, G.L. Kane, The search for supersymmetry: probing physics beyond the standard model, Phys. Rep. 117 (1985) 75, https://doi.org/10.1016/0370 -1573(85)90051-1.

[4] R. Barbieri, S. Ferrara, C.A. Savoy, Gauge models with spontaneously broken lo-cal supersymmetry, Phys. Lett. B 119 (1982) 343, https://doi.org/10.1016/0370 -2693(82)90685-2.

[5] S. Dawson, E. Eichten, C. Quigg, Search for supersymmetric particles in hadron–hadron collisions, Phys. Rev. D 31 (1985) 1581, https://doi.org/10.1103/ PhysRevD.31.1581.

[6] R. Barbieri, G. Giudice, Upper bounds on supersymmetric particle masses, Nucl. Phys. B 306 (1988) 63, https://doi.org/10.1016/0550-3213(88)90171-X. [7] E. Witten, Dynamical breaking of supersymmetry, Nucl. Phys. B 188 (1981) 513,

https://doi.org/10.1016/0550-3213(81)90006-7.

[8] S. Dimopoulos, H. Georgi, Softly broken supersymmetry and SU(5), Nucl. Phys. B 193 (1981) 150, https://doi.org/10.1016/0550-3213(81)90522-8.

[9] G.R. Farrar, P. Fayet, Phenomenology of the production, decay, and detection of new hadronic states associated with supersymmetry, Phys. Lett. B 76 (1978) 575, https://doi.org/10.1016/0370-2693(78)90858-4.

[10] Particle Data Group, C. Patrignani, et al., Review of particle physics, Chin. Phys. C 40 (2016) 100001, https://doi.org/10.1088/1674-1137/40/10/100001. [11] B. de Carlos, J. Casas, One-loop analysis of the electroweak breaking in

super-symmetric models and the fine-tuning problem, Phys. Lett. B 309 (1993) 320, https://doi.org/10.1016/0370-2693(93)90940-J, arXiv:hep -ph /9303291. [12] M. Dine, W. Fischler, M. Srednicki, Supersymmetric technicolor, Nucl. Phys. B

189 (1981) 575, https://doi.org/10.1016/0550-3213(81)90582-4.

[13] S. Dimopoulos, S. Raby, Supercolor, Nucl. Phys. B 192 (1981) 353, https://doi. org/10.1016/0550-3213(81)90430-2.

[14] N. Sakai, Naturalness in supersymmetric GUTS, Z. Phys. C 11 (1981) 153, https://doi.org/10.1007/BF01573998.

[15] R.K. Kaul, P. Majumdar, Cancellation of quadratically divergent mass corrections in globally supersymmetric spontaneously broken gauge theories, Nucl. Phys. B 199 (1982) 36, https://doi.org/10.1016/0550-3213(82)90565-X.

[16] J.A. Casas, J.M. Moreno, S. Robles, K. Rolbiecki, B. Zaldívar, What is a natural SUSY scenario?, J. High Energy Phys. 06 (2015) 070, https://doi.org/10.1007/ JHEP06(2015)070, arXiv:1407.6966.

[17] H. Baer, V. Barger, D. Mickelson, M. Padeffke-Kirkland, SUSY models under siege: LHC constraints and electroweak fine-tuning, Phys. Rev. D 89 (2014) 115019, https://doi.org/10.1103/PhysRevD.89.115019, arXiv:1404 .2277. [18] A. Mustafayev, X. Tata, Supersymmetry, naturalness, and light higgsinos, Indian

J. Phys. 88 (2014) 991, https://doi.org/10.1007/s12648-014-0504-8, arXiv:1404 . 1386.

[19] G.F. Giudice, T. Han, K. Wang, L.-T. Wang, Nearly degenerate gauginos and dark matter at the LHC, Phys. Rev. C 81 (2010) 115011, https://doi.org/10.1103/ PhysRevD.81.115011, arXiv:1004 .4902.

[20] H. Baer, A. Mustafayev, X. Tata, Monojet plus soft dilepton signal from light higgsino pair production at LHC14, Phys. Rev. D 90 (2014) 115007, https://doi. org/10.1103/PhysRevD.90.115007, arXiv:1409 .7058.

[21] C. Han, A. Kobakhidze, N. Liu, A. Saavedra, L. Wu, J.M. Yang, Probing light hig-gsinos in natural SUSY from monojet signals at the LHC, J. High Energy Phys. 02 (2014) 049, https://doi.org/10.1007/JHEP02(2014)049, arXiv:1310 .4274. [22] Z. Han, G.D. Kribs, A. Martin, A. Menon, Hunting quasidegenerate Higgsinos,

Phys. Rev. D 89 (2014) 075007, https://doi.org/10.1103/PhysRevD.89.075007, arXiv:1401.1235.

[23] A. Heister, et al., ALEPH, Search for charginos nearly mass degenerate with the lightest neutralino in e+_e−_{collisions at center-of-mass energies up to 209 GeV,} Phys. Lett. B 533 (2002) 223, https://doi.org/10.1016/S0370-2693(02)01584-8, arXiv:hep -ex /0203020.

[24] J. Abdallah, et al., DELPHI, Searches for supersymmetric particles in e+_e− col-lisions up to 208 GeV and interpretation of the results within the MSSM, Eur. Phys. J. C 31 (2003) 421, https://doi.org/10.1140/epjc/s2003-01355-5, arXiv: hep -ex /0311019.

[25]ATLAS Collaboration, Search for electroweak production of supersymmetric statesinscenarioswith compressedmassspectraat √s=13 TeVwith the ATLASdetector,arXiv:1712.08119,2017.

[26] P. Schwaller, J. Zurita, Compressed electroweakino spectra at the LHC, J. High Energy Phys. 03 (2014) 060, https://doi.org/10.1007/JHEP03(2014)060, arXiv: 1312 .7350.

[27] R. Gröber, M.M. Mühlleitner, E. Popenda, A. Wlotzka, Light stop decays: impli-cations for LHC searches, Eur. Phys. J. C 75 (2015) 420, https://doi.org/10.1140/ epjc/s10052-015-3626-z, arXiv:1408 .4662.

[28] C. Balázs, M. Carena, C.E.M. Wagner, Dark matter, light stops and electroweak baryogenesis, Phys. Rev. D 70 (2004) 015007, https://doi.org/10.1103/PhysRevD. 70.015007, arXiv:hep -ph /0403224.

[29] CMS Collaboration, Search for supersymmetry in events with soft leptons, low jet multiplicity, and missing transverse energy in proton–proton collisions at √

s=8 TeV, Phys. Lett. B 759 (2016) 9, https://doi.org/10.1016/j.physletb.2016. 05.033, arXiv:1512 .08002.

[30] CMS Collaboration, The CMS trigger system, JINST 12 (2017) P01020, https:// doi.org/10.1088/1748-0221/12/01/P01020, arXiv:1609 .02366.

[31] CMS Collaboration, The CMS experiment at the CERN LHC, JINST 3 (2008) S08004, https://doi.org/10.1088/1748-0221/3/08/S08004.

[32] J. Alwall, R. Frederix, S. Frixione, V. Hirschi, F. Maltoni, O. Mattelaer, H.S. Shao, T. Stelzer, P. Torrielli, M. Zaro, The automated computation of tree-level and next-to-leading order differential cross sections, and their matching to parton shower simulations, J. High Energy Phys. 07 (2014) 079, https:// doi.org/10.1007/JHEP07(2014)079, arXiv:1405 .0301.

[33] R. Frederix, S. Frixione, Merging meets matching in MC@NLO, J. High En-ergy Phys. 12 (2012) 061, https://doi.org/10.1007/JHEP12(2012)061, arXiv:1209 . 6215.