EUROPEAN ORGANISATION FOR NUCLEAR RESEARCH (CERN) CERN-PH-EP-2014-298 Submitted to: Eur. Phys. J. C 5 1 0 2 n Search for direct pair production of a chargin√o and a neutralino u decaying to the 125 GeV Higgs boson in s = 8 TeV pp J 5 collisions with the ATLAS detector ] x e - p e h The ATLAS Collaboration [ 2 v 0 1 1 7 Abstract 0 . 1 A search is presented for the direct pair production of a chargino and a neutralino pp → χ˜±χ˜0, where 0 the chargino decays to the lightest neutralino and the W boson, χ˜± → χ˜0(W± → (cid:96)±ν), wh1ile2the neu- 5 1 1 1 tralino decays to the lightest neutralino and the 125 GeV Higgs boson, χ˜0 → χ˜0(h → bb/γγ/(cid:96)±νqq). 2 1 : The final states considered for the search have large missing transverse momentum, an isolated elec- v i tron or muon, and one of the following: either two jets identified as originating from bottom quarks, X or two photons, or a second electron or muon with the same electric charge. The analysis is based ar on 20.3fb−1 of √s = 8 TeV proton–proton collision data delivered by the Large Hadron Collider and recorded with the ATLAS detector. Observations are consistent with the Standard Model expectations, and limits are set in the context of a simplified supersymmetric model. (cid:13)c 2015CERNforthebenefitoftheATLASCollaboration. ReproductionofthisarticleorpartsofitisallowedasspecifiedintheCC-BY-3.0license. Noname manuscript No. (will be inserted by the editor) Search for direct pair production of a chargino and a neutralino decaying to the 125 GeV Higgs boson in √ s = 8 TeV pp collisions with the ATLAS detector The ATLAS Collaboration 1CERN,1211Geneva23, Switzerland,E-mail:[email protected] thedateofreceiptandacceptanceshouldbeinsertedlater Abstract Asearchispresentedforthedirectpairpro- attheLargeHadronCollider(LHC)[17].Furthermore, duction of a chargino and a neutralino pp → χ˜±χ˜0, directpairproductionofcharginosandneutralinosmay 1 2 where the chargino decays to the lightest neutralino be the dominant production of supersymmetric parti- and the W boson, χ˜± → χ˜0(W± → (cid:96)±ν), while the cles if the superpartners of the gluon and quarks are 1 1 neutralino decays to the lightest neutralino and the heavier than a few TeV. 125 GeV Higgs boson, χ˜0 → χ˜0(h → bb/γγ/(cid:96)±νqq). In SUSY scenarios where the masses of the pseu- 2 1 The final states considered for the search have large doscalar Higgs boson and the superpartners of the lep- missing transverse momentum, an isolated electron or tons are larger than those of the produced chargino muon,andoneofthefollowing:eithertwojetsidentified andneutralino,thecharginodecaystothelightestneu- asoriginatingfrombottomquarks,ortwophotons,ora tralinoandtheW boson,whilethenext-to-lightestneu- second electron or muon with the same electric charge. tralino decays to the lightest neutralino and the SM- √ The analysis is based on 20.3fb−1 of s = 8 TeV like Higgs or Z boson. This paper focuses on SUSY proton–proton collision data delivered by the Large scenarios where the decay to the Higgs boson is the Hadron Collider and recorded with the ATLAS de- dominant one. This happens when the mass splitting tector. Observations are consistent with the Standard between the two lightest neutralinos is larger than the Modelexpectations,andlimitsaresetinthecontextof Higgs boson mass and the higgsinos are much heavier a simplified supersymmetric model. than the winos, causing the composition of the lightest charginoandnext-to-lightestneutralinotobewino-like and nearly mass degenerate. 1 Introduction A simplified SUSY model [18,19] is considered for theoptimisationofthesearchandtheinterpretationof Supersymmetry (SUSY) [1–9] proposes the existence results.Itdescribesthedirectproductionofχ˜± andχ˜0, 1 2 of new particles with spin differing by one half where the masses and the decay modes of the relevant unit from that of their Standard Model (SM) part- particles(χ˜±,χ˜0,χ˜0)aretheonlyfreeparameters.Itis 1 1 2 ners. In the Minimal Supersymmetric Standard Model assumed that the χ˜± and χ˜0 are pure wino states and 1 2 (MSSM) [10–14], charginos, χ˜± , and neutralinos, degenerate in mass, while the χ˜0 is a pure bino state. 1,2 1 χ˜0 , are the mass-ordered eigenstates formed from The prompt decays χ˜± → W±χ˜0 and χ˜0 → hχ˜0 are 1,2,3,4 1 1 2 1 the linear superposition of the SUSY partners of the assumed to have 100% branching fractions. The Higgs Higgs and electroweak gauge bosons (higgsinos, winos boson mass is set to 125 GeV, which is consistent with and bino). In R-parity-conserving models, SUSY par- themeasuredvalue[20],anditsbranchingfractionsare ticles are pair-produced in colliders and the lightest assumed to be the same as in the SM. The latter as- SUSYparticle(LSP)isstable.InmanymodelstheLSP sumption is motivated by those SUSY models in which is assumed to be a bino-like χ˜0, which is weakly in- themassofthepseudoscalarHiggsbosonismuchlarger 1 teracting. Naturalness arguments [15,16] suggest that than the Z boson mass. the lightest of the charginos and neutralinos may have The search presented in this paper targets leptonic masses at the electroweak scale, and may be accessible decays of the W boson and three Higgs boson decay 2 (a) Onelepton andtwob-quarkschannel (b) Onelepton andtwophotonschannel (c) Same-signdilepton channel Fig. 1 Diagrams for the direct pair production of χ˜±χ˜0 and the three decay modes studied in this paper. For the same-sign 1 2 dileptonchannel(c),onlythedominantdecaymodeis shown. modes as illustrated in Fig. 1. The Higgs boson decays 2 The ATLAS detector into a pair of b-quarks, or a pair of photons, or a pair of W bosons where at least one of the bosons decays ATLAS is a multipurpose particle physics experi- leptonically. The final states therefore contain missing ment [33]. It consists of detectors forming a forward- transverse momentum from neutrinos and neutralinos, backward symmetric cylindrical geometry.1 The inner one lepton ((cid:96) = e or µ), and one of the following: two detector (ID) covers |η|<2.5 and consists of a silicon b-quarks ((cid:96)bb), or two photons ((cid:96)γγ), or an additional pixel detector, a semiconductor microstrip tracker, and leptonwiththesameelectriccharge((cid:96)±(cid:96)±).TheHiggs a transition radiation tracker. The ID is surrounded bosoncandidatecanbefullyreconstructedwiththe(cid:96)bb by a thin superconducting solenoid providing a 2T and (cid:96)γγ signatures. The (cid:96)±(cid:96)± signature does not allow axial magnetic field. A high-granularity lead/liquid- for such reconstruction and it is considered because of argon (LAr) sampling calorimeter measures the en- its small SM background. Its main signal contribution ergyandthepositionofelectromagneticshowerswithin is due to h → WW, with smaller contributions from |η|<3.2. Sampling calorimeters with LAr are also used h → ZZ and h → ττ when some of the visible decay to measure hadronic showers in the endcap (1.5< products are missed during the event reconstruction. |η|<3.2) and forward (3.1<|η|<4.9) regions, while √ The analysis is based on 20.3fb−1 of s = 8 TeV a steel/scintillator tile calorimeter measures hadronic showers in the central region (|η|<1.7). The muon proton–proton collision data delivered by the LHC and spectrometer(MS)surroundsthecalorimetersandcon- recorded with the ATLAS detector. Previous searches sistsofthreelargesuperconductingair-coretoroidmag- forcharginosandneutralinosattheLHChavebeenre- nets, each with eight coils, precision tracking cham- ported by the ATLAS [21–23] and CMS [24,25] collab- bers(|η|<2.7),andfasttriggerchambers(|η|<2.4).A orations. Similar searches were conducted at the Teva- three-level trigger system selects events to be recorded tron [26,27] and LEP [28–32]. for permanent storage. The results of this paper are combined with those of the ATLAS search using the three-lepton and miss- ing transverse momentum final state, performed with the same dataset [21]. The three-lepton selections may 3 Monte Carlo simulation containuptotwohadronicallydecayingτ leptons,pro- vidingsensitivitytotheh→ττ/WW/ZZ Higgsboson The event generators, the accuracy of theoretical cross decaymodes.Thestatisticalcombinationoftheresults sections,theunderlying-eventparametertunes,andthe is facilitated by the fact that all event selections were parton distribution function (PDF) sets used for simu- constructed not to overlap. lating the SM background processes are summarised in Table 1. This paper is organised in the following way: the ATLASdetectorisbrieflydescribedinSect.2,followed 1ATLAS uses a right-handed coordinate system with its ori- byadescriptionoftheMonteCarlosimulationinSect.3. gin at the nominal interaction point (IP) in the centre of In Sect. 4 the common aspects of the event reconstruc- the detector and the z-axis along the beam line. The x-axis tionareillustrated;Sects.5,6,and7describethechannel- points from the IP to the centre of the LHC ring, and the y-axis points upward. Cylindrical coordinates (r,φ) are used specific features; Sect. 8 discusses the systematic un- in the transverse plane, φ being the azimuthal angle around certainties;theresultsandconclusionsarepresentedin thez-axis.Thepseudorapidityisdefinedintermsofthepolar Sects. 9 and 10. angleθ asη=−lntan(θ/2). 3 Table1 Simulatedsamplesusedforbackgroundestimates.“Tune”referstothechoiceofparametersusedfortheunderlying- eventgeneration. Process Generator Crosssection Tune PDFset Singletop,t-channel AcerMC[34]+Pythia6[35] NNLO+NNLL[36] AUET2B[37] CTEQ6L1[38] Singletop,s-channel Powheg[39,40]+Pythia6 NNLO+NNLL[41] Perugia2011C[42] CT10[43] tW Powheg+Pythia6 NNLO+NNLL[44] Perugia2011C CT10 t¯t Powheg+Pythia6 NNLO+NNLL[45–50] Perugia2011C CT10 t¯tW,t¯tZ MadGraph[51]+Pythia6 NLO AUET2B CTEQ6L1 W,Z ((cid:96)bbchannel) Sherpa[52] NLO – CT10 W,Z ((cid:96)±(cid:96)± channel) Alpgen[53]+Pythia6 NLO Perugia2011C CTEQ6L1 WW,WZ,ZZ Sherpa NLO – CT10 Wγ Wγγ Alpgen+Pythia6 NLO AUET2B CTEQ6L1 Zγ,Zγγ Sherpa NLO – CT10 Wh,Zh Pythia8[54] NNLO(QCD)+NLO(EW)[55] AU2[56] CTEQ6L1 t¯th Pythia8 NLO(QCD)[55] AU2 CTEQ6L1 The SUSY signal samples are produced with Her- tainty of 2.8% derived following the methodology de- wig++ [57] using the CTEQ6L1 PDF set. Signal cross tailed in Ref. [65]. sections are calculated at next-to-leading order (NLO) Verticescompatiblewiththeproton-protoninterac- in the strong coupling constant using Prospino2 [58]. tionsarereconstructedusingtracksfromtheID.Events These agree with the NLO calculations matched to re- are analysed if the primary vertex has five or more summation at next-to-leading-logarithmic (NLL) accu- tracks,eachwithtransversemomentump >400 MeV, T racy within ∼2% [59,60]. For each cross section, the unlessstatedotherwise.Theprimaryvertexofanevent nominalvalueanditsuncertaintyaretakenrespectively is identified as the vertex with the largest (cid:80)p2 of the T fromthecentreandthespreadofthecross-sectionpre- associated tracks. dictions using different PDF sets and their associated Electron candidates are reconstructed from cali- uncertainties,aswellasfromvariationsoffactorisation bratedclusteredenergydepositsintheelectromagnetic and renormalisation scales, as described in Ref. [61]. calorimeter and a matched ID track, which in turn de- The propagation of particles through the ATLAS termine the p and η of the candidates respectively. detector is modelled with GEANT4 [62] using the full T Electrons must satisfy “medium” cut-based identifica- ATLAS detector simulation [63] for all Monte Carlo tion criteria, following Ref. [66], and are required to (MC) simulated samples, except for tt¯production and have p >10 GeV and |η|<2.47. T the SUSY signal samples in which the Higgs boson de- Muon candidates are reconstructed by combining cays to two b-quarks, for which a fast simulation based tracks in the ID and tracks or segments in the MS [67] on a parametric response of the electromagnetic and andarerequiredtohavep >10 GeVand|η|<2.5.To hadronic calorimeters is used [64]. The effect of mul- T suppress cosmic-ray muon background, events are re- tiple proton–proton collisions in the same or nearby jectediftheycontainamuonhavingtransverseimpact beam bunch crossings (in-time or out-of-time pile-up) parameter with respect to the primary vertex |d | > is incorporated into the simulation by overlaying ad- 0 ditionalminimum-biaseventsgeneratedwithPythia6 0.2mm or longitudinal impact parameter with respect to the primary vertex |z |>1mm. ontohard-scatterevents.Simulatedeventsareweighted 0 so that the distribution of the average number of in- Photon candidates are reconstructed from clusters teractionsperbunchcrossingmatchesthatobservedin of energy deposits in the electromagnetic calorimeter. data,butareotherwisereconstructedinthesameman- Clusterswithoutmatchingtracksaswellasthosematch- ner as data. ing one or two tracks consistent with a photon con- version are considered. The shape of the cluster must matchthatexpectedforanelectromagneticshower,us- ing criteria tuned for robustness under the pile-up con- 4 Event reconstruction ditions of 2012 [68]. The cluster energy is calibrated separately for converted and unconverted photon can- The data sample considered in this analysis was col- didates using simulation. In addition, η-dependent cor- lected with a combination of single-lepton, dilepton, rection factors determined from Z → e+e− events are and diphoton triggers. After applying beam, detector, applied to the cluster energy, as described in Ref. [68]. anddata-qualityrequirements,thedatasetcorresponds The photon candidates must have p > 20 GeV and T toanintegratedluminosityof20.3fb−1,withanuncer- |η|<2.37, excluding the transition region 1.37<|η|< 4 1.56 between the central and endcap electromagnetic Table2 Summaryoftheoverlapremovalprocedure.Poten- calorimeters. The tighter η requirement on photons, as tial ambiguities are resolved by removing nearby objects in the indicated order, from top to bottom. Different ∆R sepa- compared to electrons, reflects the poorer photon reso- rationrequirementsareusedinthe threechannels. lutioninthetransitionregionandfor2.37≤|η|<2.47. Jetsarereconstructedwiththeanti-k algorithm[69] Candidates ∆R threshold Candidateremoved t with a radius parameter of 0.4 using three-dimensional (cid:96)bb (cid:96)γγ (cid:96)±(cid:96)± clusters of energy in the calorimeter [70] as input. The e–e 0.1 — 0.05 lowest-pT e clusters are calibrated, weighting differently the energy e–γ — 0.4 — e deposits arising from the electromagnetic and hadronic jet–γ — 0.4 — jet jet–e 0.2 0.2 0.2 jet componentsoftheshowers.Thefinaljetenergycalibra- τ–eorτ–µ — — 0.2 τ tion corrects the calorimeter response to the particle- µ–γ — 0.4 — µ level jet energy [71,72]; the correction factors are ob- e–jetorµ–jet 0.4 0.4 0.4 eorµ tained from simulation and then refined and validated e–µ 0.1 — 0.1 both µ–µ 0.05 — 0.05 both usingdata.Correctionsforin-timeandout-of-timepile- jet–τ — — 0.2 jet up are also applied, as described in Ref. [73]. Events containing jets failing to meet the quality criteria de- scribedinRef.[71]arerejectedtosuppressnon-collision In the same-sign channel, e+e− and µ+µ− pairs with backgroundandeventswithlargenoiseinthecalorime- m <12 GeVarealsoremoved.Theremaininglep- (cid:96)+(cid:96)− ters. tons and photons are referred to as “preselected” ob- JetswithpT >20 GeVareconsideredinthecentral jects. pseudorapidity (|η|<2.4) region, and jet pT >30 GeV Isolation criteria are applied to improve the pu- is required in the forward (2.4 < |η| < 4.5) region. For rity of reconstructed objects. The criteria are based central jets, the pT threshold is lower since it is possi- on the scalar sum of the transverse energies ET of the ble to suppress pile-up using information from the ID, calorimeter cell clusters within a radius ∆R of the ob- the “jet vertex fraction” (JVF). This is defined as the ject (Econe∆R), and on the scalar sum of the p of the T T pT-weightedfractionoftrackswithinthejetthatorigi- trackswithin∆R andassociatedwiththeprimaryver- nate from the primary vertex of the event, and is −1 if tex(pcone∆R).Thecontributionduetotheobjectitself T there are no tracks within the jet. Central jets can also isnotincludedineithersum.Thevaluesusedintheiso- be tagged as originating from bottom quarks (referred lationcriteriadependonthechannel;theyarespecified toasb-jets)usingtheMV1multivariateb-taggingalgo- in Sects. 5, 6 and 7. rithmbasedonquantitiesrelatedtoimpactparameters Themissingtransversemomentum,pmiss(withmag- T oftracksandreconstructedsecondaryvertices[74].The nitude Emiss), is the negative vector sum of the trans- T efficiencyoftheb-taggingalgorithmdependsontheop- verse momenta of all preselected electrons, muons, and erating point chosen for each channel, and is reported photons, as well as jets and calorimeter energy clus- in Sects. 5 and 7. ters with |η|<4.9 not associated with these objects. Hadronically decaying τ leptons are reconstructed Clustersthatareassociatedwithelectrons,photonsand as1-or3-pronghadronicjetswithin|η|<2.47,andare jetsarecalibratedtothescaleofthecorrespondingob- required to have pT >20 GeV after being calibrated to jects [76,77]. the τ energy scale [75]. Final states with hadronically The efficiencies for electrons, muons, and photons decaying τ leptons are not considered here; however, to satisfy the reconstruction and identification criteria identified τ leptons are used in the overlap removal are measured in control samples, and corrections are procedure described below, as well as to ensure that applied to the simulated samples to reproduce the ef- thesame-signleptonchanneldoesnotoverlapwiththe ficiencies in data. Similar corrections are also applied three-leptonsearch[21]thatisincludedinthecombined tothetriggerefficiencies,aswellastothejetb-tagging result. efficiency and misidentification probability. Potential ambiguities between candidate leptons, photons and jets are resolved by removing one or both (cid:112) objectsiftheyareseparatedby∆R≡ (∆φ)2+(∆η)2 5 One lepton and two b-jets channel below a threshold. This process eliminates duplicate objects reconstructed from a single particle, and sup- 5.1 Event selection presses leptons and photons contained inside hadronic jets. The thresholds and the order in which overlap- The events considered in the one lepton and two b- ping objects are removed are summarised in Table 2. jets channel are recorded with a combination of single- 5 Table 3 Selection requirements for the signal, control and validation regions of the one lepton and two b-jets channel. The numberofleptons,jets,andb-jetsislabelledwith nlepton,njet,and nb-jetrespectively. SR(cid:96)bb-1 SR(cid:96)bb-2 CR(cid:96)bb-T CR(cid:96)bb-W VR(cid:96)bb-1 VR(cid:96)bb-2 n 1 1 1 1 1 1 lepton njet 2–3 2–3 2–3 2 2–3 2–3 n 2 2 2 1 2 2 b-jet Emiss [GeV] >100 >100 >100 >100 >100 >100 T mCT [GeV] >160 >160 100–160 >160 100–160 >160 mW [GeV] 100–130 >130 >100 >40 40–100 40–100 T lepton triggers with a p threshold of 24 GeV. To en- suppresses the W +jets background. The two SRs are T surethattheeventistriggeredwithaconstanthighef- distinguished by requiring 100 < mW < 130 GeV for T ficiency, the offline event selection requires exactly one SR(cid:96)bb-1 and mW >130 GeV for SR(cid:96)bb-2. The first sig- T signal lepton (e or µ) with p > 25 GeV. The signal nal region provides sensitivity to signal models with a T electrons must satisfy the “tight” identification criteria mass splitting between χ˜0 and χ˜0 similar to the Higgs 1 2 of Ref. [66], as well as |d |/σ < 5, where σ is the boson mass, while the second one targets larger mass 0 d0 d0 error on d , and |z sinθ| < 0.4mm. The signal muons splittings. 0 0 must satisfy |η| < 2.4, |d0|/σd0 < 3, and |z0sinθ| < In each SR, events are classified into five bins of 0.4mm. The signal electrons (muons) are required to the invariant mass m of the two b-jets as 45–75–105– bb satisfy the isolation criteria ETcone0.3/pT < 0.18 (0.12) 135–165–195GeV. In the SRs, about 70% of the signal and pcTone0.3/pT <0.16 (0.12). events due to h → b¯b populate the central bin of 105– Events with two or three jets are selected, and the 135GeV. The other four bins (sidebands) are used to jets can be either central (|η| < 2.4) or forward (2.4 < constrain the background normalisation, as described |η|<4.9).CentraljetshavepT >25 GeV,andforward below. jets have p > 30 GeV. For central jets with p < T T 50 GeV, the JVF must be > 0.5. Events must contain exactly two b-jets and these must be the highest-p T centraljets.Thechosenoperatingpointoftheb-tagging 5.2 Background estimation algorithm identifies b-jets in simulated tt¯events with an efficiency of 70%; it misidentifies charm jets 20% Thecontributionsfromthett¯andW+jetsbackground of the time and light-flavour (including gluon-induced) sources are estimated from simulation, and normalised jets less than 1% of the time. to data in dedicated control regions defined in the fol- AftertherequirementofEmiss >100GeV,thedom- lowingparagraphs.Thecontributionfrommulti-jetpro- T inant background contributions in the (cid:96)bb channel are duction, where the signal lepton is a misidentified jet tt¯, W +jets, and single-top Wt production. Their con- or comes from a heavy-flavour hadron decay or photon tributionsaresuppressedusingthekinematicselections conversion,isestimatedusingthe“matrixmethod”de- described below, which define the two signal regions scribed in Ref. [22], and is found to be less than 3% (SR) SR(cid:96)bb-1 and SR(cid:96)bb-2 summarised in Table 3. of the total background in all regions and is thus ne- The contransverse mass m [78,79] is defined as glected. The remaining sources of background (single CT top, Z + jets, WW, WZ, ZZ, Zh and Wh produc- (cid:113) m = (Eb1 +Eb2)2−|pb1 −pb2|2, (1) tion) are estimated from simulation. CT T T T T Two control regions (CR), CR(cid:96)bb-T and CR(cid:96)bb-W, where Ebi and pbi are the transverse energy and mo- are designed to constrain the normalisations of the tt¯ T T mentum of the i-th b-jet. The SM tt¯background has andW+jetsbackgroundsrespectively.Theacceptance anupperendpointatm ofapproximatelym ,andis for tt¯events is increased in CR(cid:96)bb-T by modifying the CT t efficiently suppressed by requiring m >160 GeV. requirementonm to100<m <160 GeV.Theac- CT CT CT The transverse mass mW, describing W candidates ceptanceofW+jetseventsisincreasedinCR(cid:96)bb-Wby T in background events, is defined as requiring mW >40 GeV and exactly two jets, of which T onlyoneisb-tagged.Thesetwocontrolregionsaresum- (cid:113) mW = 2E(cid:96)Emiss−2p(cid:96) ·pmiss, (2) marised in Table 3. The control regions are defined to T T T T T be similar to the signal regions in order to reduce sys- whereE(cid:96) andp(cid:96) arethetransverseenergyandmomen- tematicuncertaintiesontheextrapolationtothesignal T T tumofthelepton.RequiringmW >100 GeVefficiently regions; at the same time they are dominated by the T 6 eV ATLAS Data W+jets eV104 ATLAS Data W+jets G G s / 30 110023 s = 8 TeV, 20.3 fb 1 Ttmto(t∼χa±l∼χ S0,∼χM0) = (2SO5it0nh,g0el)re G teoVp s / 30 103 s = 8 TeV, 20.3 fb 1 Ttmto(t∼χa±l∼χ S0,∼χM0) = (2SO5it0nh,g0el)re G teoVp nt 1 2 1 nt102 1 2 1 e e v v E 10 E 10 1 1 10 1 10 1 M 2 M 2 S S a / 1 a / 1 at at D 0 D 0 100 150 200 250 300 100 150 200 250 300 m [GeV] m [GeV] CT CT (a) mCT inCR(cid:96)bb-T,SR(cid:96)bb-1andSR(cid:96)bb-2,centralmbb bin (b) mCT inCR(cid:96)bb-T,SR(cid:96)bb-1andSR(cid:96)bb-2,mbb sidebands V V Ge103 ATLAS Data W+jets Ge ATLAS Data W+jets 5 s = 8 TeV, 20.3 fb 1 Total SM Single top 5 103 s = 8 TeV, 20.3 fb 1 Total SM Single top nts / 1102 tmt(∼χ1±∼χ20,∼χ10) = (2O5t0h,0e)r GeV nts / 1102 tmt(∼χ1±∼χ20,∼χ10) = (2O5t0h,0e)r GeV e e Ev 10 Ev 10 1 1 10 1 10 1 M 2 M 2 S S a / 1 a / 1 at at D 0 D 0 40 60 80 100 120 140 160 180 40 60 80 100 120 140 160 180 mW [GeV] mW [GeV] T T (c) mW inVR(cid:96)bb-2,SR(cid:96)bb-1andSR(cid:96)bb-2,centralm bin (d) mW inVR(cid:96)bb-2,SR(cid:96)bb-1andSR(cid:96)bb-2,m sidebands T bb T bb s V103 nt ATLAS Data W+jets e ATLAS Data W+jets e104 G Ev103 s = 8 TeV, 20.3 fb 1 Ttmto(t∼χa±l∼χ S0,∼χM0) = (2SO5it0nh,g0el)re G teoVp s / 30 102 s = 8 TeV, 20.3 fb 1 Ttmto(t∼χa±l∼χ S0,∼χM0) = (2SO5it0nh,g0el)re G teoVp 1 2 1 nt 1 2 1 e 10 102 Ev 1 10 1 10 1 M1.5 M 2 S S a / 1 a / 1 at at D0.5 D 0 0 1 2 60 80 100 120 140 160 180 Number of b jets mbb [GeV] (e) Numberofb-jetsinSR(cid:96)bb-1andSR(cid:96)bb-2withouttheb-jet (f) m inSR(cid:96)bb-1andSR(cid:96)bb-2 bb multiplicityrequirement,centralm bin bb Fig. 2 Distributions of contransverse mass mCT, transverse mass of the W-candidate mWT, number of b-jets, and invariant mass of the b-jets m for the one lepton and two b-jets channel in the indicated regions. The stacked background histograms bb are obtained from the background-only fit. The hashed areas represent the total uncertainties on the background estimates after the fit. The rightmost bins in (a)–(d) include overflow. The distributions of a signal hypothesis are also shown without stacking on the background histograms. The vertical arrows indicate the boundaries of the signal regions. The lower panels show theratioofthedatatothe SMbackgroundprediction. 7 Table 4 Event yields and SM expectation in the one lepton and two b-jets channel obtained with the background-only fit. “Other” includes Z + jets, WW, WZ, ZZ, Zh and Wh processes. The errors shown include statistical and systematic uncertainties. SR(cid:96)bb-1 SR(cid:96)bb-2 SR(cid:96)bb-1 SR(cid:96)bb-2 CR(cid:96)bb-T CR(cid:96)bb-W VR(cid:96)bb-1 VR(cid:96)bb-2 105<m <135GeV m sidebands bb bb Observed events 4 3 14 10 651 1547 885 235 SMexpectation 6.0±1.3 2.8±0.8 13.1±2.4 8.8±1.7 642±25 1560±40 880±90 245±17 t¯t 3.8±1.2 1.4±0.7 8.0±2.4 3.1±1.4 607±25 680±60 680±90 141±18 W +jets 0.6±0.3 0.2±0.1 2.7±0.5 1.7±0.3 11± 2 690±60 99±12 62± 8 Singletop 1.3±0.4 0.7±0.4 1.9±0.6 2.5±1.1 20± 4 111±14 80±10 27± 4 Other 0.3±0.1 0.5±0.1 0.5±0.1 1.5±0.2 4± 1 76± 8 16± 2 15± 1 targeted background processes and the expected con- Figure 2 shows the data distributions of m , mW, CT T tamination by signal is small. n andm comparedtotheSMexpectationsinvar- b-jet bb Asinthesignalregions,thecontrolregionsarebin- iousregions. Thedata agreewellwiththe SMexpecta- ned in m (m in the case of CR(cid:96)bb-W). A “back- tions in all distributions. bb bj ground-only” likelihood fit is performed, in which the predictionsofthesimulatedbackgroundprocesseswith- out any signal hypothesis are fit simultaneously to the 6 One lepton and two photons channel data yields in eight m sideband bins of the SRs and bb thetenmbbbinsoftheCRs.Thisfit,aswellasthelimit- 6.1 Event Selection setting procedure, is performed using the HistFitter package described in Ref. [80]. The two free parame- Events recorded with diphoton or single-lepton trig- ters of the fit, namely the normalisations of the tt¯and gers are used in the one lepton and two photons chan- W +jets background components, are constrained by nel. For the diphoton trigger, the transverse momen- the number of events observed in the control regions tumthresholdsattriggerlevelforthehighest-p (lead- T andsignalregionsidebands,wherethenumberofevents ing) and second highest-p (sub-leading) photons are T is described by a Poisson probability density function. 35 GeV and 25 GeV respectively. For these events, the The remaining nuisance parameters correspond to the event selection requires exactly one signal lepton (e or sources of systematic uncertainty described in Sect. 8. µ) and exactly two signal photons, with p thresh- T They are taken into account with their uncertainties, olds of 15 GeV for electrons, 10 GeV for muons, and andadjustedtomaximisethelikelihood.Theyieldses- 40(27)GeVforleading(sub-leading)photons.Inaddi- timated with the background-only fit are reported in tion, events recorded with single-lepton triggers, which Table 4, as well as the resulting predictions in SR(cid:96)bb-1 have transverse momentum thresholds at trigger level andSR(cid:96)bb-2for105<m <135GeV.WhileCR(cid:96)bb-T of 24 GeV, are used. For these events, the selection re- bb isdominatedbytt¯events,CR(cid:96)bb-Wispopulatedevenly quiresp thresholdsof25 GeVforelectronsandmuons, T by tt¯and W +jets events, which causes the normali- and 40 (20) GeV for leading (sub-leading) photons. sations of the tt¯and W +jets contributions to be neg- In this channel, a neural network algorithm, based atively correlated after the fit. As a result, the uncer- on the momenta of the tracks associated with each tainties on individual background sources do not add vertex and the direction of flight of the photons, is up quadratically to the uncertainty on the total SM expectation. The normalisation factors are found to be 1.03±0.15fortt¯and0.79±0.07forW+jets,wherethe Table5 Selectionrequirementsforthesignalandvalidation regionsoftheoneleptonandtwophotonschannel.Thenum- errors include statistical and systematic uncertainties. ber of leptons and photons is labelled with nlepton and nγ To validate the background modelling, two valida- respectively. tion regions (VR) are defined similarly to the SRs ex- cept for requiring 40 < mW < 100 GeV, and requir- SR(cid:96)γγ-1 SR(cid:96)γγ-2 VR(cid:96)γγ-1 VR(cid:96)γγ-2 T ing 100 < mCT < 160 GeV for VR(cid:96)bb-1 and mCT > nlepton 1 1 1 1 160 GeV for VR(cid:96)bb-2 as summarised in Table 3. The nγ 2 2 2 2 Emiss [GeV] >40 >40 <40 — yields in the VRs are shown in Table 4 after the back- T ∆φ(W,h) >2.25 >2.25 — <2.25 ground-onlyfit,whichdoesnotusethedataintheVRs mWγ1[GeV] >150 <150 T to constrain the background. The data event yields are and or — — found to be consistent with background expectations. mWTγ2[GeV] >80 <80 8 V 5 nts / 10 Ge102 AsT =L 8A TSeV, 20.3 fb 1 DDTHNoaaiogtttnagaa ls H SS SiigMdMgesb aSnMds (Scaled) vents / 0.4102 AsT =L 8A TSeV, 20.3 fb 1 DDTHNoaaiogtttnagaa ls H SS SiigMdMgesb aSnMds (Scaled) e ∼∼ ∼ E ∼∼ ∼ v10 m(χ±χ0,χ0)=(165,35) GeV 10 m(χ±χ0,χ0)=(165,35) GeV E 1 2 1 1 2 1 1 1 10 1 10 1 0 10 20 30 40 50 60 70 80 90 100 0 0.5 1 1.5 2 2.5 3 Emiss [GeV] ∆φ(W,h) T (a) Emiss inSR(cid:96)γγ-1andSR(cid:96)γγ-2withoutEmiss cut (b) ∆φ(W,h)inSR(cid:96)γγ-1andSR(cid:96)γγ-2without∆φ(W,h)cut T T V V Ge ATLAS Data Ge ATLAS Data 25 102 s = 8 TeV, 20.3 fb 1 DToattaal SSiMdebands (Scaled) 20 102 s = 8 TeV, 20.3 fb 1 DToattaal SSiMdebands (Scaled) s / Higgs SM s / Higgs SM nt Non Higgs SM nt Non Higgs SM e ∼∼ ∼ e ∼∼ ∼ v10 m(χ±χ0,χ0)=(165,35) GeV v10 m(χ±χ0,χ0)=(165,35) GeV E 1 2 1 E 1 2 1 1 1 10 1 10 1 50 100 150 200 250 300 350 400 0 20 40 60 80 100120140160180200220240 Wγ Wγ m 1 [GeV] m 2 [GeV] T T (c) mWγ1 inSR(cid:96)γγ-1andSR(cid:96)γγ-2withoutmWγi cuts (d) mWγ2 inSR(cid:96)γγ-1andSR(cid:96)γγ-2withoutmWγi cuts T T T T Fig. 3 DistributionsofmissingtransversemomentumEmiss,azimuthdifferencebetweentheW andHiggsbosoncandidates T ∆φ(W,h),transversemassoftheW andphotonsystemmWγ1 andmWγ2 intheoneleptonandtwophotonssignalregionsforthe T T Higgs-mass window (120<mγγ <130GeV). The vertical arrows indicate the boundaries of the signal regions. The filled and hashedareasrepresentthestackedhistogramsofthesimulation-basedbackgroundcrosscheckandthetotaluncertainties.The contributionsfromnon-Higgsbackgroundsarescaledby10GeV/50GeV=0.2fromthemγγ sideband(100<mγγ <120GeV and130<mγγ <160GeV)intotheHiggs-masswindow.Therightmostbinsin(a),(c),and(d)includeoverflow.Scaleddata inthesidebandareshownassquares,whileeventsintheHiggs-masswindowareshownascircles.Thedistributionsofasignal hypothesisarealsoshownwithoutstackingonthebackground histograms. used to select the primary vertex, similarly to the AT- h → γγ candidate (∆φ(W,h) > 2.25). Only events LAS SM h→γγ analysis described in Ref. [81]. Signal with a diphoton invariant mass, m , between 100 and γγ muons must satisfy |d | < 1mm and |z | < 10mm. 160 GeV are considered. Events in the sideband, out- 0 0 The isolation criteria for both the electrons and muons sidetheHiggs-masswindowbetween120and130GeV, are Econe0.4/p < 0.2 and pcone0.2/p < 0.15. Signal are included to constrain the non-Higgs background as T T T T photons are required to satisfy Econe0.4 < 6 GeV and described in Sect. 6.2. T pcone0.2 <2.6 GeV. Selected events are split into two SRs with differ- T The two largest background contributions are due ent expected signal sensitivities based on two variables to multi-jet and Zγ production, with leptons or jets mWTγ1 and mWTγ2, which are defined as misreconstructed as photons. These background con- (cid:113) tributionsaresuppressedbyrequiringEmiss >40 GeV. mWγi = (mW)2+2EWEγi −2pW ·pγi, (3) T T T T T T T The p of the W → (cid:96)ν system, reconstructed as- T suming background events with neutrino p = pmiss, wheremW,EW andpW arethetransversemass,energy T T T T T is required to be back-to-back with the p of the and momentum of the W candidate, and Eγi and pγi T T T 9 V V 12 e ATLAS Data e ATLAS Data G 6 G 5 s = 8 TeV, 20.3 fb 1 5 10 s = 8 TeV, 20.3 fb 1 2. 5 Fit to Data 2. Fit to Data s / s / 8 nt 4 nt e e v v 6 E 3 E 4 2 1 2 0 0 100 110 120 130 140 150 160 100 110 120 130 140 150 160 m [GeV] m [GeV] γγ γγ (a) SR(cid:96)γγ-1 (b) SR(cid:96)γγ-2 25 V V e30 ATLAS Data e ATLAS Data G G 2.5 25 s = 8 TeV, 20.3 fb 1 Fit to Data 2.5 20 s = 8 TeV, 20.3 fb 1 Fit to Data nts / 20 nts / 15 e e v v E15 E 10 10 5 5 0 0 100 110 120 130 140 150 160 100 110 120 130 140 150 160 m [GeV] m [GeV] γγ γγ (c) VR(cid:96)γγ-1 (d) VR(cid:96)γγ-2 Fig.4 Resultsofthebackground-onlyfittothediphotoninvariantmass,mγγ,distributionintheoneleptonandtwophotons signal and validation regions. The contributions from SM Higgs boson production are constrained to the MC prediction and associatedsystematicuncertainties.Thebandshowsthesystematicuncertaintyonthefit.Thefitisperformedoneventswith 100 GeV < mγγ < 160 GeV, with events in SR(cid:96)γγ-1 or SR(cid:96)γγ-2 in the Higgs-mass window (120 GeV ≤ mγγ ≤ 130 GeV), indicatedbythearrows, excludedfromthefit. are the transverse energy and momentum of the i-th, Emiss.Thesignalandvalidationregionsaresummarised T p -ordered, photon. Including a photon in the trans- in Table 5. T versemasscalculationprovidesameanstoidentifylep- tonically decaying W bosons in the presence of a final- Distributions in the Higgs-mass window of the four state radiation photon. Events with mWγ1 > 150 GeV kinematic variables used to define the SRs are shown T and mWγ2 > 80 GeV are classified into SR(cid:96)γγ-1, and in Fig. 3. For illustration purposes, the observed yield T those with either mWγ1 <150 GeV or mWγ2 <80 GeV in the sideband region is shown for each distribution, T T into SR(cid:96)γγ-2. Most of the sensitivity to the signal is scaled into the corresponding Higgs-mass window by provided by SR(cid:96)γγ-1, while SR(cid:96)γγ-2 assists in con- the relative widths of the Higgs-mass window and the straining systematic uncertainties. sideband region, 10 GeV / 50 GeV = 0.2. Also shown, for each distribution, is a simulation-based cross-check Two overlapping validation regions are defined by of the background estimate. To reduce statistical un- inverting and modifying the Emiss and ∆φ(W,h) cri- certainties originating from the limited number of sim- T teria relative to those of the signal regions. The first ulatedevents,thenon-Higgscontributionsareobtained regionVR(cid:96)γγ-1requiresEmiss <40 GeVandhasnore- inthesidebandandscaledintotheHiggs-masswindow T quirementon∆φ(W,h),andthesecondregionVR(cid:96)γγ-2 by 0.2. The simulation-based prediction of the non- requires ∆φ(W,h) < 2.25 and has no requirement on Higgs background is estimated from the W/Z(γ,γγ)