Contents lists available atScienceDirect
Physics Letters B
www.elsevier.com/locate/physletb
Search for R-parity-violating supersymmetric particles in multi-jet final states produced in p–p collisions at √
s = 13 TeV using the ATLAS detector at the LHC
.TheATLAS Collaboration
a r t i c l e i n f o a b s t ra c t
Articlehistory:
Received11April2018
Receivedinrevisedform26July2018 Accepted15August2018
Availableonline17August2018 Editor:M.Doser
Results ofasearchfor gluinopairproductionwithsubsequentR-parity-violatingdecaystoquarks are presented.Thissearchuses36.1 fb−1ofdatacollectedbytheATLASdetectorinproton–protoncollisions withacentre-of-massenergyof√s
=13 TeVattheLHC.Theanalysisisperformedusingrequirements onthenumberofjetsandthenumberofjetstaggedascontainingab-hadronaswellasatopological observable formedbythescalarsum ofmassesoflarge-radiusjetsintheevent.Nosignificantexcess abovetheexpectedStandardModelbackgroundisobserved.Limitsaresetontheproductionofgluinos inmodelswiththeR-parity-violatingdecaysofeitherthegluinoitself(directdecay)ortheneutralino produced inthe R-parity-conservinggluino decay(cascadedecay).Inthegluino cascadedecaymodel, gluinomassesbelow1850 GeVareexcludedfor1000 GeVneutralinomass.Forthegluinodirectdecay model,the95%confidencelevelupperlimitonthecrosssectiontimesbranchingratiovariesbetween 0.80fb atm˜g=900 GeVand0.011 fb atm˜g=1800 GeV.
©2018TheAuthor.PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense (http://creativecommons.org/licenses/by/4.0/).FundedbySCOAP3.
1. Introduction
Supersymmetry(SUSY) [1–6] isa theoreticalextension ofthe Standard Model (SM) which fundamentally relates fermions and bosons. It is an alluring theoretical possibility given its poten- tialto solve thehierarchy problem[7–10]. ThisLetter presentsa searchforsupersymmetricgluinopairproductionwithsubsequent R-parity-violating(RPV) [11–16] decaysintoquarksineventswith manyjets using 36.1 fb−1 of p–p collision data at√
s=13 TeV collectedbytheATLASdetectorin2015and2016.Intheminimal supersymmetricextensionoftheStandardModel,theRPVcompo- nentofagenericsuperpotentialcanbewrittenas [15,17]:
WRPV=1
2λi jkLiLjE¯k+λi jkLiQjD¯k+1
2λi jkU¯iD¯jD¯k+κiLiH2,(1) wherei,j,k=1,2,3 aregenerationindices.Thegenerationindices are omittedinthe discussions that followifthe statement being made is not specific to any generation. The first three terms in Eq. (1) areoftenreferredtoasthetrilinearcouplings,whereasthe lasttermisreferredtoasbilinear.TheLiandQirepresentthelep- tonandquark SU(2)Ldoubletsuperfields, whereas H2 represents theHiggs superfield. The E¯j, D¯j, andU¯j are thechargedlepton,
E-mailaddress:atlas.publications@cern.ch.
down-type quark, and up-type quark SU(2)L singlet superfields, respectively.Thecouplingsforeachtermaregivenbyλ,λ,andλ, while κ isamassparameter.Inthebenchmarkmodelsconsidered inthissearch, thecouplingsofλ andλ are settozeroandonly thebaryon-number-violatingcouplingλi jk isnon-zero.Becauseof the structure of Eq. (1), scenarios in which only λi jk=0 are of- ten referred to as UDD scenarios. The diagrams shown in Fig. 1 representthe benchmark processesused inthe optimizationand design ofthesearch presented inthisLetter. Inthe gluinodirect decaymodel(Fig.1(a)),thegluinodirectlydecaysintothreequarks via the RPV UDDcouplingλ, leading tosix quarksattree level in thefinal state ofgluinopairproduction.In thegluinocascade decaymodel (Fig. 1(b)), thegluino decaysintotwo quarks anda neutralino, which,in turn, decays into three quarks via the RPV UDDcouplingλ,resultingintenquarksattreelevelinthefinal stateofgluinopairproduction.Eventsproducedintheseprocesses typically have a high multiplicity ofreconstructed jets. In signal modelsconsideredinthissearch,theproductionofthegluinopair isassumedtobeindependentofthevalueofλ.Decaybranching ratios ofall possible λ flavour combinations givenby thestruc- tureof Eq. (1) areassumedtobeequal,anddecaysofthegluino and neutralinoare implemented asprompt decaysvia modifying thedecaywidthsofgluinosandneutralinos.Inthisconfiguration, a significant portionof signalevents containatleast onebottom ortopquark. OthermodelsoftheRPVUDDscenario,suchasthe https://doi.org/10.1016/j.physletb.2018.08.021
0370-2693/©2018TheAuthor.PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense(http://creativecommons.org/licenses/by/4.0/).Fundedby SCOAP3.
Fig. 1.Diagramsforthebenchmarkprocessesconsideredforthisanalysis.Theblack linesrepresentStandardModelparticles,theredlinesrepresentSUSYpartners,the greyshadedcirclesrepresenteffectivevertices thatincludeoff-shellpropagators (e.g.heavysquarkscouplingtoaχ˜10 neutralinoandaquark),andthebluesolid circlesrepresenteffectiveRPVverticesallowedbythebaryon-number-violatingλ couplingswithoff-shellpropagators(e.g.heavysquarkscouplingtotwoquarks).
Quarkandantiquarkarenotdistinguishedinthediagrams.(Forinterpretationof thecoloursinthefigure(s),thereaderisreferredtothewebversionofthisarticle.)
Minimal Flavour Violation model [18,19], predict that the gluino decayspreferentiallyintofinalstateswiththird-generationquarks.
Thesetheoreticalargumentsmotivatetheintroductionofb-tagging requirementsintothesearch.
Thisanalysisis an updateto previous ATLAS searches forsig- nalsarisingfromRPVUDDscenarios [20,21] performedwithdata takenat√
s=8 TeV.The search strategy closelyfollowsthe one implementedinRef. [21],whichexcludesagluinowithmassupto 917 GeVinthegluinodirectdecaymodel,andagluinowithmass upto1000 GeVforaneutralinomassof500 GeVinthegluinocas- cadedecaymodel.Twootherpublications [22,23] fromtheATLAS Collaboration reported on the searches for signals from a differ- entgluinocascadedecaymodelwherethequarks/antiquarksfrom the gluino decay are top quark–anti-quark pairs and the quarks fromthe neutralino decays are u, d or s quarks. These searches probedeventswithatleastoneelectronormuon.Themoststrin- gentlower limit onthe gluinomass,fromRef. [22], is2100 GeV fora neutralino mass of1000 GeV. Ina recentpublication [24], theCMSCollaborationsetalowerlimitof1610 GeVonthegluino massinanRPVUDDscenariowherethegluinoexclusivelydecays into a final state of a top quark, a bottom quark and a strange quark,using√
s=13 TeV ppcollisiondata.
2. ATLASdetector
The ATLAS detector [25] covers almost the whole solid an- gle around the collision point with layers of tracking detectors, calorimeters and muon chambers. The inner detector, immersed in a magnetic field provided by a solenoid, has full coverage in φ and covers the pseudorapidity range |η|<2.5.1 It consists of asiliconpixeldetector,a siliconmicrostrip detectoranda transi- tion radiation straw-tube tracker. The innermost pixel layer, the insertable B-layer, was added between Run-1 and Run-2 of the LHC,ataradiusof33mmaroundanew,thinner,beampipe [26].
Inthepseudorapidityregion|η|<3.2,highgranularitylead/liquid- argon(LAr)electromagnetic(EM)samplingcalorimetersareused.
A steel/scintillator tile calorimeterprovides hadronic calorimetry coverage over |η|<1.7.The end-cap andforward regions, span- ning 1.5<|η|<4.9, are instrumented with LAr calorimetry for boththeEMandhadronicmeasurements.Themuonspectrometer
1 ATLASusesaright-handedcoordinatesystemwithitsoriginatthe nominal interactionpointinthecentreofthedetectorandthez-axisalongthebeamdi- rection.Thex-axispointstowardthecentreoftheLHCring,andthey-axispoints upward.Cylindricalcoordinates(r,φ)areusedinthetransverseplane,φbeingthe azimuthalanglearoundthebeampipe.Thepseudorapidityηisdefinedintermsof thepolarangleθbyη≡ −ln[tan(θ/2)].
surroundsthesecalorimeters,andcomprisesasystemofprecision trackingchambersandfast-responsedetectorsfortriggering,with threelargetoroidalmagnets,eachconsistingofeightcoils,provid- ingthemagneticfieldforthemuondetectors.A two-leveltrigger systemisusedtoselectevents [27].Thefirst-leveltriggerisimple- mentedinhardwareandusesasubsetofthedetectorinformation.
Thisisfollowedbythesoftware-basedhigh-leveltrigger,reducing theeventratetoabout1kHz.
3. Simulationsamples
Signalsampleswereproducedcoveringawiderangeofgluino and neutralino masses. In the gluino direct decay model, the gluino mass(m˜g) was varied from 900 GeV to1800 GeV. In the case of the cascade decays, for each gluino mass (1000 GeV to 2100 GeV), separate samples were generated withmultiple neu- tralino masses (mχ˜10) ranging from 50 GeV to 1.65 TeV. In each case, mχ˜10 <mg˜. In the gluino cascade decay model, the two quarks produced from the gluino decay were restricted to be first orsecond generationquarks.All threegenerations ofquarks were allowed to be in the final state of the lightest supersym- metric particle decay. Signal samples were generated atleading- order (LO) accuracy withup to two additionalpartons using the MadGraph5_aMC@NLOv2.3.3eventgenerator [28] interfacedwith PYTHIA8.186 [29] forthepartonshower, fragmentationandun- derlying event. The A14 set of tuned parameters [30] was used together withtheNNPDF2.3LO partondistribution function(PDF) set [31].TheEvtGenv1.2.0programwasusedtodescribetheprop- erties of the b- and c-hadron decays in the signal samples. The signalproductioncrosssectionswerecalculatedatnext-to-leading order(NLO)inthestrongcouplingconstant,addingtheresumma- tion ofsoftgluon emission atnext-to-leading-logarithm accuracy (NLO + NLL) [32–36]. The nominal cross section and its uncer- tainty were taken from Ref. [37]. Cross sections were evaluated assuming masses of 450 TeVfor the light-flavour squarks in the caseofgluinopairproduction.Inthesimulation,thetotal widths ofgluinosandneutralinosweresettobe1 GeV,effectivelymaking theirdecaysprompt.
While a data-driven method was used to estimate the back- ground, simulated events were used to establish, test and vali- date the methodology of the analysis. Multijet events constitute the dominant background in the search region, with small con- tributionsfromtop-quarkpairproduction(tt).¯ Contributionsfrom
γ + jets, W + jets, Z + jets, single-top-quark, and diboson backgroundprocessesarefoundtobenegligiblefromstudiesper- formedwithsimulatedevents.Themultijetbackgroundwasstud- ied with three different leading order Monte Carlo samples.The PYTHIA 8.186 event generator was used together withthe A14 tuneandtheNNPDF2.3LOpartondistributionfunctions,whilethe Herwig++ 2.7.1eventgeneratorwasusedtogetherwiththeUEEE5 tune [38] and CTEQ6L1 PDF sets [39]. The Sherpaevent genera- tor [40] wasalsousedtogeneratemultijeteventsforthestudyof backgroundestimation. Matrixelements were calculated withup to threepartons atLO, were showered withSherpaaswell, and weremergedusingtheME+PS@LOprescription [41].TheCT10PDF set [42] was used in conjunction with dedicated parton shower tuning developed by the Sherpa authors. For the generation of fullyhadronicdecaysoftt¯events,thePowheg-Box v2eventgener- ator [43] wasusedwiththeCT10PDFsetandwasinterfacedwith PYTHIA6.428 [44].TheEvtGenv1.2.0program [45] wasalsoused to describe theproperties ofthe b- and c-hadron decaysfor the backgroundsamplesexceptthosegeneratedwithSherpa[46].
The effect of additional p–p interactions per bunch crossing (“pile-up”)asafunctionoftheinstantaneousluminositywastaken
intoaccountbyoverlayingsimulatedminimum-biaseventsaccord- ingtotheobserveddistributionofthenumberofpile-upinterac- tionsindata.AllMonteCarlosimulatedbackgroundsampleswere passedthrougha fullGeant4simulation [47] oftheATLASdetec- tor [48]. Thesignal sampleswere passed through a fast detector simulation [49] based on a parameterization of the performance of the ATLAS electromagnetic and hadronic calorimeters and on Geant4 elsewhere. The compatibility of the signal selection effi- ciencybetweenthefastsimulationsample andthefullsimulation sample was validated ata numberof signal points in thegluino directdecaymodelandgluinocascadedecaymodelconsideredin thisLetter.
4. Eventselection
The data were recorded in2015 and2016, with the LHC op- erating at a centre-of-mass energy of √
s=13 TeV. All detector elements are requiredto be operational.The integrated luminos- ity is measured to be 3.2 fb−1 and 32.9 fb−1,for the 2015 and 2016datasets,respectively.Theuncertaintyinthecombined2015 and2016 integratedluminosity is 2.1%. Itis derived, following a methodologysimilar to that detailedin Ref. [50], froma calibra- tionoftheluminosityscaleusingx–y beam-separationscans.
The eventsused in thissearch are selected using an HT trig- ger,seeded froma first-level jet trigger with an ET threshold of 100 GeV,whichrequiresthescalarsumofjettransverseenergies atthehigh leveltriggerto be greater than1.0 TeV. Thisrequire- mentisfoundtobefullyefficientforsignalregionsconsideredin thisLetter. Events are required to havea primary vertex withat leasttwoassociatedtrackswithtransversemomentum(pT)above 0.4 GeV. Theprimaryvertexassignedtothehard-scatteringcolli- sionistheonewiththehighest
trackp2T,wherethesumoftrack p2T is taken over all tracks associated with that vertex. To reject eventswithdetectornoiseornon-collisionbackgrounds,eventsare removediftheyfailbasicqualitycriteria [51,52].
Jetsarereconstructedfromthree-dimensionaltopologicalclus- ters of energy deposits in the calorimeter calibrated at the EM scale [53], usingthe anti-kt algorithm [54,55] with two different radius parameters of R=1.0 and R=0.4, hereafter referred to as large-R jets andsmall-R jets, respectively. The four-momenta ofthejetsare calculatedasthesumofthe four-momentaofthe clusters, whichare assumed tobe massless. Forthe large-R jets, theoriginalconstituentsarecalibratedusingthelocalcellweight- ingalgorithm [53,56] priortojet-findingandreclusteredusingthe longitudinally-invariant kt algorithm [57] with a radius parame- terof Rsub-jet=0.2, to forma collectionof sub-jets.A sub-jet is discardedifitcarrieslessthan5%ofthelarge-Rjet pToftheorig- inaljet.The constituentsintheremaining sub-jetsare thenused to recalculate the large-R jet four-momenta, and the jet energy andmass are further calibratedto particle levelusing correction factorsderivedfromsimulation [58].Theresulting“trimmed” [58, 59] large-R jetsarerequiredtohave pT>200 GeVand|η|<2.0.
The analysisdoesnot place anyrequirementon the vertexasso- ciationoftrackswithinajetnoronthetimingofthecalorimeter cellswithinajet,whichpreservesthesensitivityofthisanalysisto modelscontainingnon-promptjets.Thesmall-R jetsarecorrected for pile-up contributions and are then calibrated to the particle levelusingsimulatedeventsfollowedbyacorrection basedonin situmeasurements [53,60,61].
Theidentificationofjetscontainingb-hadronsisbasedonthe small-R jets with pT>50 GeV and |η|<2.5 and a multivari- ate tagging algorithm [62,63]. This algorithm is applied to a set of trackswith loose impact parameter constraintsin a region of interest around each jet axis to enable the reconstruction ofthe b-hadron decay vertex. The b-tagging requirements result in an
efficiency of70% for jetscontaining b-hadrons, asdetermined in a sample of simulated tt¯ events [63]. A small-R jet passing the b-taggingrequirementisreferredtoasab-taggedjet.
The analysis of data is primarily based on observables built fromlarge-R jets.Thesmall-R jetsareusedtoclassifyeventsand for categorization of the large-R jetsbased on the b-tagging in- formation.Specifically, eventsselectedin theanalysisare divided intoab-tagging samplewhereatleastoneb-taggedjetispresent in the event, and a b-veto sample where no b-tagged jet is presentinthe event.Events selectedwithouttakinginto account anyb-taggingrequirementarereferred toasinclusiveevents.
Large-R jetsare classifiedas either those that are matched to a b-tagged jetwithin R=1.0 (b-matchedjets), orthose that are notmatchedtoab-taggedjet.
5. Analysisstrategy
The analysis uses a kinematic observable, the total jet mass, MJ [64–66],astheprimarydiscriminatingvariabletoseparatesig- nalandbackground.Theobservable MJ isdefinedasthesumof themassesofthefourleadinglarge-Rjets.
MJ =
pT>200 GeV
|η|≤2.0 j=1−4
mjetj (2)
This observable provides significant sensitivity for gluinos with very highmass.Fig.2(a)presentsexamples ofthediscrimination that the MJ observable provides between the background (rep- resented here by Sherpa, PYTHIA 8.186 and Herwig++ multijet MonteCarlosimulation)andseveralsignalsamples,aswellasthe comparisonofthedatatothesimulatedmultijetbackground.
Another discriminating variable that is independent of MJ is necessary in order to define suitable control and validation re- gionswherethebackgroundestimationcanbestudiedandtested.
The signal is characterized by a higherrate of central-jet events ascomparedtotheprimary multijetbackground.Thisisexpected due to the difference in the production modes: predominantly s-channelforthesignal,whereasthebackgroundcanalsobepro- ducedthroughu- andt-channelprocesses.Fig.2(b)showsthedis- tributionofthepseudorapiditydifferencebetweenthetwoleading large-R jets,|η12|forseveralsignalandbackgroundMonteCarlo samples, as well as data. A high-|η12| requirement can be ap- pliedtoestablishacontrolregionoravalidationregionwherethe potentialsignalcontaminationneedstobesuppressed.
TheuseofMJinthisanalysisprovidesan opportunitytoem- ploythefullydata-drivenjetmasstemplatemethodtoestimatethe background contributionin signal regions.The jet mass template method is discussed in Ref. [66], and its first experimental im- plementation is described in Ref. [21]. In thismethod, single-jet masstemplatesareextractedfromsignal-depletedcontrolregions.
These jet masstemplates are createdin bins thatare defined by a number ofobservables, which include jet pT and |η|, and the b-matchingstatus.They providea probabilitydensityfunction that describestherelativeprobabilityforajetwithagivenpTand ηto haveacertainmass.Thismethodassumesthatjetmasstemplates onlydependontheseobservablesandarethesameinthecontrol regions andsignal regions. A sample wherethe background MJ distributionneedstobeestimated,suchasavalidationregionora signal region,isreferredtoasthekinematicsample. Theonlyin- formationusedisthejetpTand η,aswellasitsb-matchingstatus, which are inputsto the templates. Foreach jet inthe kinematic sample, its corresponding jet mass template is used to generate a random jet mass.An MJ distribution canbe constructed from
