Search for gluinos in events with two same-sign leptons, jets and missing transverse momentum with the ATLAS detector in pp collisions at sqrt(s) = 7 TeV

A search is presented for gluinos decaying via the supersymmetric partner of the top quark using events with two same-sign leptons, jets and missing transverse momentum. The analysis is performed with 2.05 fb^-1 of integrated luminosity from pp collisions at sqrt(s) = 7 TeV collected by the ATLAS detector at the LHC. No excess beyond the Standard Model expectation is observed and exclusion limits are derived for simplified models where the gluino decays via the supersymmetric partner of the top quark and in the MSUGRA/CMSSM framework. In those scenarios, gluino masses below 550 GeV are excluded at 95% CL within the parameter space considered, significantly extending the coverage with respect to existing limits. Depending on the model parameters, gluino masses up to 750 GeV can also be excluded at 95% CL.

In SUSY models, the gluino is a strongly-interacting Majorana fermion. Pair-produced gluinos therefore have equal probability to produce a pair of leptons that have the same charge (same-sign, SS) and the opposite charge from their decays. The supersymmetric partner of the top quark (stop) has two mass eigenstates witht 1 being the lightest. Top quarks andt 1 squarks can be produced in the decay of the gluino viagg → ttt 1t * 1 , ttt * 1t * 1 ,ttt 1t1 [10][11][12][13][14]. Thet 1 squark can further decay to the lightest chargino (χ ± 1 ) or lightest neutralino (χ 0 1 ) viat 1 → bχ ± 1 or t 1 → tχ 0 1 producing isolated leptons in the semi-leptonic top quark decay or in the leptonicχ ± 1 decay and enriching the signal events with two or more leptons, jets and missing transverse momentum (E miss T ) from the undetected neutralinos. This analysis considers events with a pair of isolated SS leptons, multiple high-p T jets and large E miss T . The requirement of SS leptons in the event suppresses the contribution from SM processes and thus enhances the potential signal significance even for final state topologies with relatively soft jet kinematics. This article presents the first search for gluino-mediated stop production using the SS signature although other searches with SS leptons have been performed [15][16][17][18]. The analysis presented here complements the results of the ATLAS search based on single lepton plus b-jets [19], enhancing the sensitivity in the experimentally difficult region near the kinematic limit for the production of two top quarks and a neutralino, and for the region with low stop masses.
ATLAS is a general-purpose detector [20] at the LHC. This search uses pp collision data recorded from March to August 2011 at a center-of-mass energy of 7 TeV. The data set corresponds to a total integrated luminosity of 2.05 ± 0.08 fb −1 [21,22] after the application of data quality requirements. Events are selected using single lepton and dilepton triggers that have constant efficiency as a function of lepton p T above the offline p T cuts used in the analysis.
Jets are reconstructed from three-dimensional calorimeter energy clusters by using the anti-k t jet algorithm [23,24] with a radius parameter of 0.4. The measured jet energy is corrected for inhomogeneities in, and for the non-compensating nature of, the calorimeter by using p T -and η-dependent correction factors [25]. Only jet candidates with p T > 20 GeV within |η| < 2.8 are retained. Events with any jet that fails the jet quality criteria designed to remove noise and non-collision backgrounds [25] are rejected.
Electron candidates are required to have p T > 20 GeV, |η| < 2.47 and satisfy the 'tight' selection criteria defined in Ref. [26]. They are also required to be isolated: the scalar sum of p T of tracks within a cone in the η −φ plane of radius ∆R = 0.2 around the candidate excluding its own track, Σ pT , must be less than 10% of the electron p T . Muon candidates are required to have p T > 20 GeV, |η| < 2.4 and are identified by matching an extrapolated inner detector track and one or more track segments in the muon spectrometer [27]. They must have longitudinal and transverse impact parameter within 1 mm and 0.2 mm of the primary vertex [28], respectively, and Σ pT < 1.8 GeV.
The calculation of E miss T [29] is based on the vectorial sum of the p T of the reconstructed jets (with p T > 20 GeV and |η| < 4.5), leptons and the calorimeter energy clusters not belonging to reconstructed objects.
During part of the data-taking period, a localized electronics failure in the electromagnetic calorimeter created a dead region (∆η × ∆φ ≃ 1.4 × 0.2). For jets in this region, a correction to their energy is made using the energy depositions in the neighbouring cells, and is propagated to E miss T . If this correction projected onto the direction of the E miss T is larger than 10 GeV or 10% of the E miss T , the event is discarded [30]. Events with reconstructed electrons in the calorimeter dead region are also rejected.
Events in which the two highest-p T leptons (ℓ = e, µ) have the same charge and with at least four jets with p T > 50 GeV are selected. In addition, two signal regions are considered: SR1, which requires E miss T > 150 GeV; and SR2, which in addition requires m T > 100 GeV, where m T is the transverse mass of the E miss T and the highest-p T lepton defined as m 2 ). This final m T cut helps reducing the tt background. The signal regions are optimized based on several models where SS dileptons are produced in gluino decays. In signals such as the MSUGRA/CMSSM (minimal supergravity or constrained minimal supersymmetric standard model) [31,32], the directions of the lepton andχ 0 1 are strongly correlated as they originate from the decay of a common parent particle (usuallyχ ± 1 or the next-to-lightest neutralino,χ 0 2 ). This leads to a softer m T spectrum than that found in gluino-mediated stop signal models, where the lepton and theχ 0 1 originate from different parent particles and are thus uncorrelated.
Simulated Monte Carlo (MC) event samples are used to aid in the description of the background and to model the SUSY signal. Top-quark pair and single-top production are simulated with MC@NLO [33], fixing the top-quark mass at 172.5 GeV, and using the next-to-leading-order (NLO) parton density function (PDF) set CTEQ6.6 [34]. Samples of W +jets and Z+jets with both light-and heavy-flavor jets are generated with ALPGEN [35] and PDF set CTEQ6L1 [36]. The fragmentation and hadronization for the ALPGEN and MC@NLO samples are performed with HERWIG [37], using JIMMY [38] for the underlying event. Samples of ttZ, ttW and ttW W (referred to as tt+X) are generated with MADGRAPH [39] interfaced to PYTHIA [40]. The total LO cross section for these samples is 0.39 pb and is normalized to NLO using a K-factor of 1.3 [41]. Diboson samples are generated with HERWIG for W ± W ∓ , W Z and ZZ processes and with MADGRAPH for W ± W ± qq processes. The total NLO cross section for the diboson background is 71 pb [42,43]. SUSY signal processes are simulated for various models using HERWIG++ [37]  [46] are used in the production of the MC samples, which are processed through a detector simulation [47] based on GEANT4 [48]. Effects of multiple proton-proton interactions per bunch crossing are included in the simulation.
The SM backgrounds are evaluated using a combination of MC simulation and data-driven techniques. SM processes that generate events containing jets which are misidentified as leptons or where a lepton from a b-or c-hadron decay is selected are collectively referred to as "fake-lepton" background. It generally consists of semileptonic tt, single top, W +jets and strong light-and heavy-flavor jet production. The contribution from the "fake-lepton" background is estimated from data with a method similar to that described in Ref. [49,50] by loosening the lepton identification and isolation criteria. For electrons the 'medium' criteria are used instead of the 'tight' criteria [26], and for both electrons and muons the isolation criterion is relaxed. The method counts the number of observed events containing loose-loose, loosetight, tight-loose and tight-tight lepton pairs. The probability of loose real leptons passing the tight selection criteria is obtained using a Z → ℓ + ℓ − sample. The probability of loose fake leptons to pass the tight selection criteria is determined as a function of the lepton p T using multijet control samples obtained by requiring two SS leptons and low E miss T . Using these probabilities, relations are obtained for the observed event counts in the signal regions as functions of the numbers of events containing fake-fake, fake-real, real-fake and real-real lepton pairs. These can be solved simultaneously to estimate the number of background events [49,50]. The results of the estimations have been validated with data in control regions obtained by reversing the E miss T or jet multiplicity cuts used in the signal regions.
Background events from charge misidentification (dominated by electrons which have undergone hard bremsstrahlung with subsequent photon conversion) are estimated using a partially data-driven technique [16]. The probability of charge misidentification is calculated from MC and corrected by consideration of the number of events in data with SS electron pairs and invariant mass within 15 GeV of the Z-boson mass. This probability is applied to tt MC events producing e ± ℓ ∓ pairs to evaluate the number of SS events from incorrect charge assignment in each signal region. The probability of misidentifying the charge of a muon and the contributions in the signal regions from charge misidentification of Z/γ * +jets and other SM backgrounds are negligible.
Contributions from other SM background sources (diboson and tt + X) are evaluated using the MC samples described above. In these processes, real SS lepton pairs are produced and their contribution to the signal regions can be described with MC. In particular, the contribution from the experimentally unmeasured tt+X processes has been studied using several MC generators. The background from cosmic rays is evaluated with data using the method in Ref.
[16] and its contribution is negligible in the signal regions.
Systematic uncertainties are estimated in the signal regions for the background and the SUSY signal processes. The primary sources of systematic uncertainties in the background are the jet energy scale calibration (35%), the jet energy resolution (10%), uncertainties on lepton and jet reconstruction and identification (5%), MC modeling and theoretical cross section uncertainties (40%-70%). In particular, the theoretical uncertainties on the cross section of the tt + X processes are found to be between 35%-55% by varying factorization and renormalization scales and 25% due to PDF uncertainties. In addition, a 50% uncertainty is assigned on the K-factor used to obtain the NLO cross section [41]. In the fake-lepton background estimation systematic uncertainties are assigned to the probabilities for loose fake leptons to pass the tight selection. This accounts for potentially different compositions of the signal and control regions. These uncertainties vary in the 10%-80% range depending on the lepton p T and are evaluated using data samples with jets of different energies. The absolute uncertainty for each background source is given in Table I. Systematic uncertainties on the signal expectations are evaluated through variations of the factorization and renormalization scales between half and twice their default values, and by including the uncertainty on α s and on the PDF provided by CTEQ6. Uncertainties are calculated for individual SUSY processes. The total uncertainty varies in the 20%-40% range for the considered MC signals. Any correlations of the systematic uncertainties in signals and background are taken into account.  Figure 1 shows the distribution of the number of jets with p T > 50 GeV for events with 2 SS leptons, and the E miss T distribution for events with 2 SS leptons and at least four jets with p T > 50 GeV. The contributions from all the SM backgrounds are shown together with their total statistical and systematic uncertainties. For illustration, the distribution for a signal obtained with the decayg → ttχ 0 1 ingg pair-produced events with mg = 650 GeV and mχ0 1 = 150 GeV is also shown. The data are in agreement with the SM background expectation and once four jets of p T > 50 GeV are required no event is observed with E miss T > 150 GeV. Table I shows the number of expected events in the signal regions for each background source together with the observed number of events. The expectation from the Errors on data points are statistical, while the error band on the SM background represents the total uncertainty. The component labelled "Fake-lepton" is evaluated using data as described in the text. The component labelled "Z/γ * +jets" is estimated from MC. No estimation of the charge mis-ID is included in the distribution. The component labelled "Signal" corresponds to a signal obtained with the decayg → ttχ 0 1 via off mass-shellt (mt = 1.2 TeV) ingg pair-produced events with mg = 650 GeV and mχ0 1 = 150 GeV.
SM is estimated to be less than one event for each signal region with no events observed in data. Limits at 95% confidence level (CL) are derived on the visible cross section σ vis = σ ×ǫ×A where σ is the total production cross section for any new signal producing SS dileptons, A is the acceptance defined by the fraction of events passing geometric and kinematic cuts at particle level and ǫ is the detector reconstruction, identification and trigger efficiency. For the signal shown in Figure 1, the acceptance and efficiency are 1.5% and 55%, respectively. Limits are set using the CL s prescription, as described in Ref. [51].
The results are given in Table I. The results obtained in SR2 are interpreted in a simplified model where gluinos are only produced in pairs, the stop (mt = 1.2 TeV) is heavier than the gluino, and only the gluino three-body decayg → ttχ 0 1 via an offshell stop is allowed. Figure 2 shows the limit in the gluino-neutralino mass plane. For a gluino mass of 650 GeV, neutralino masses below 215 GeV are excluded at 95% CL. For a neutralino mass of 100 GeV, gluino masses below 715 GeV are similarly excluded. The −1σ uncertainty limit on the expected limit lies outside the range of the figure as a consequence of the low number of expected signal events and a total signal uncertainty that reaches close to 50%. The results can be generalised in terms of production cross section upper limits at 95% CL forgg pair production processes with the produced particles decaying into ttχ 0 1 final states, as also shown in Figure 2.  The results in SR2 are also interpreted considering gluino pair production followed by theg →t 1 t decay. Only stop decays viat 1 → bχ ± 1 are considered with mχ± 1 ≈ 2mχ0 1 and mχ0 1 = 60 GeV. Figure 3 shows the exclusion limit as a function of gluino and stop masses, where gluino masses below 660 GeV are excluded at 95% CL for stop masses below 460 GeV.
The results in SR1 are interpreted within the MSUGRA-CMSSM framework in terms of limits on the universal scalar and gaugino mass parameters m 0 and m 1/2 , as shown in Figure 4. These are presented for fixed values of the universal trilinear coupling parameter A 0 = 0, ratio of the vacuum expectation values of the two Higgs doublets tan β = 10, and Higgs mixing parameter µ > 0. In this model, values of m 1/2 below 300 GeV are excluded at 95% CL for m 0 values below 750 GeV, and m 1/2 values below 180 GeV are excluded over the entire m 0 region considered. These are equivalent to the exclusion of gluino masses below ∼550 GeV independent of the squark mass (and gluino masses below ∼750 GeV for squark masses below 1 TeV).
In summary, a search for SUSY with two SS leptons, jets and missing transverse momentum has been performed using 2.05 fb −1 of ATLAS data. With no events observed in the signal regions, limits have been derived in the context of models where top quarks are produced    in gluino decays and MSUGRA/CMSSM scenarios. In all these signal models, gluino masses below ∼550 GeV are excluded at 95% CL within the parameter space considered and gluino masses up to ∼750 GeV are excluded at 95% CL, depending on the model parameters. The results of this analysis are complementary to and extend the current exclusion limits on the gluino mass beyond those from other ATLAS searches [19, 56,57].
We thank CERN for the very successful operation of the LHC, as well as the support staff from our institutions without whom ATLAS could not be operated efficiently. We