Search for pair production of first-and second-generation scalar leptoquarks in pp collisions at sqrt ( s ) = 7 TeV "

Results are presented from a search for the pair production of firstand second-generation scalar leptoquarks in proton-proton collisions at sqrt(s)= 7 TeV. The data sample corresponds to an integrated luminosity of 5.0 inverse femtobarns, collected by the CMS detector at the LHC. The search signatures involve either two charged leptons of the same-flavour (electrons or muons) and at least two jets, or a single charged lepton (electron or muon), missing transverse energy, and at least two jets. If the branching fraction of the leptoquark decay into a charged lepton and a quark is assumed to be beta=1, leptoquark pair production is excluded at the 95% confidence level for masses below 830 GeV and 840 GeV for the first and second generations, respectively. For beta = 0.5, masses below 640 GeV and 650 GeV are excluded. These limits are the most stringent to date.

Search for pair production of first-and second-generation scalar leptoquarks in pp collisions at ffiffi ffi s p ¼ 7 TeV S. Chatrchyan et al. * (CMS Collaboration) (Received 24 July 2012;published 27 September 2012) Results are presented from a search for the pair production of first-and second-generation scalar leptoquarks in proton-proton collisions at ffiffi ffi s p ¼ 7 TeV. The data sample corresponds to an integrated luminosity of 5:0 fb À1 , collected by the CMS detector at the LHC. The search signatures involve either two charged leptons of the same flavor (electrons or muons) and at least two jets or a single charged lepton (electron or muon), missing transverse energy, and at least two jets. If the branching fraction of the leptoquark decay into a charged lepton and a quark is assumed to be ¼ 1, leptoquark pair production is excluded at the 95% confidence level for masses below 830 GeV and 840 GeV for the first and second generations, respectively. For ¼ 0:5, masses below 640 GeV and 650 GeV are excluded. These limits are the most stringent to date. DOI: 10.1103/PhysRevD.86.052013 PACS numbers: 14.80.Sv, 12.60.Ài, 13.85.Rm

I. INTRODUCTION
The structure of the standard model (SM) of particle physics suggests a fundamental relationship between quarks and leptons. There are many models beyond the SM that predict the existence of leptoquarks (LQ), hypothetical particles that carry both baryon number and lepton number and couple to both quarks and leptons. Among these scenarios are grand unified theories [1,2], composite models [3], extended technicolor models [4][5][6], and superstring-inspired models [7]. Leptoquarks are color triplets with fractional electric charge and can be either scalar or vector particles. A leptoquark couples to a lepton and a quark with a coupling strength , and it decays to a charged lepton and a quark with an unknown branching fraction or to a neutrino and a quark with branching fraction 1 À . To satisfy constraints from bounds on flavor-changing neutral currents and from rare pion and kaon decays [3,8], it is assumed that leptoquarks couple to quarks and leptons of a single generation. Leptoquarks are classified as first, second, or third generation, depending on the generation of leptons to which they couple. The dominant mechanisms for the production of leptoquark pairs at the Large Hadron Collider (LHC) are gluon-gluon (gg) fusion and quarkantiquark (q " q) annihilation, shown in Fig. 1. The dominant processes depend only on the strong coupling constant and have been calculated at next-to-leading order (NLO) [9]. The cross section for production via the unknown Yukawa coupling of a leptoquark to a lepton and a quark is typically smaller. This paper reports on a search for pair production of scalar leptoquarks. Several experiments have searched for pair-produced scalar leptoquarks, but none has obtained evidence for them [10][11][12][13][14][15]. This search uses a data sample corresponding to an integrated luminosity of 5:0 fb À1 recorded with the Compact Muon Solenoid (CMS) detector during the 2011 proton-proton run of the LHC at ffiffi ffi s p ¼ 7 TeV. The analysis performed in this paper considers the decay of leptoquark pairs into two charged leptons of the same flavor (either electrons or muons) and two quarks; or into a charged lepton, a neutrino, and two quarks. As a result, two distinct classes of events are selected: one with two high-transverse-momentum (p T ) electrons or muons and at least two high-p T jets (''jj) and the other with one high-p T electron or muon, large missing transverse energy (E miss T ), and at least two high-p T jets ('jj). The CMS detector, described in detail elsewhere [16], uses a cylindrical coordinate system with the z axis along the counterclockwise beam axis. The detector consists of an inner tracking system and electromagnetic (ECAL) and hadron (HCAL) calorimeters surrounded by a 3.8 T solenoid. The inner tracking system consists of a silicon pixel Published by the American Physical Society under the terms of the Creative Commons Attribution 3.0 License. Further distribution of this work must maintain attribution to the author(s) and the published article's title, journal citation, and DOI. and strip tracker, providing the required granularity and precision for the reconstruction of vertices of charged particles in the range 0 2 in azimuth and jj < 2:5, where the pseudorapidity is defined as ¼ À ln½tanð=2Þ, and is the polar angle measured with respect to the z axis. The crystal ECAL and the brass/scintillator sampling HCAL are used to measure with high resolution the energies of photons, electrons, and hadrons for jj < 3:0. The three muon systems surrounding the solenoid cover a region jj < 2:4 and are composed of drift tubes in the barrel region (jj < 1:2), of cathode strip chambers in the endcaps (0:9 < jj < 2:4), and of resistive plate chambers in both the barrel region and the end caps (jj < 1:6). Events are recorded based on a trigger decision using information from either the calorimeter or muon systems. The final trigger decision is based on the information from all subsystems, which is passed on to the high-level trigger (HLT), consisting of a farm of computers running a version of the reconstruction software optimized for fast processing.
The ''jj and 'jj analyses are performed separately and the results are combined as a function of the branching fraction and the leptoquark mass M LQ for first and second generations independently. The analysis in all four decay channels searches for leptoquarks in an excess of events characteristic of the decay of heavy objects. Various triggers are used to collect events depending on the decay channel and the data taking periods as described in Sec. II. An initial selection isolates events with high-p T final-state particles (two or more isolated leptons and two or more jets; or one isolated lepton, two or more jets, and large E miss T indicative of the emission of a neutrino). Kinematic variables are then identified to further separate a possible leptoquark signal from the expected backgrounds, and optimized thresholds on the values of these variables are derived to maximize the sensitivity to the possible presence of a signal in each decay mode. The variables used in the optimization are the invariant mass of jet-lepton pairs (M 'j ), the scalar sum (S T ) of the p T of each of the final-state objects, and either the invariant mass of the dilepton pair (M '' ) in the ''jj channels or E miss T in the 'jj channels.
Major sources of SM background are Z= Ã þ jets, W þ jets processes, and t " t. Smaller contributions arise from single-top production, diboson processes, and QCD multijet processes. The major backgrounds are determined either from control samples in data or from Monte Carlo (MC) simulated samples normalized to data in selected control regions.
After final selection, the data are well described by the SM background predictions, and upper limits on the leptoquark pair-production cross section are set using a CL S modified frequentist approach [17,18]. Using Poisson statistics, 95% confidence level (CL) upper limits are obtained on the leptoquark pair-production cross section times branching fraction as a function of leptoquark mass (M LQ ). This is compared with the NLO predictions [9] to determine lower limits on M LQ for ¼ 1 and ¼ 0:5. The ''jj and 'jj channels are combined to further maximize the exclusion in and M LQ , especially for the case $ 0:5, where combining the two channels increases the sensitivity of the search.
The data and the initial event selection are detailed in Sec. II of this paper, followed by a description of the signal modeling and background estimates in Secs. III and IV, respectively. Section V contains the final event selection, and Sec. VI describes the systematic uncertainties. The results of the search are presented in Sec. VII and summarized in Sec. VIII.

II. DATA SET AND OBJECT RECONSTRUCTION
For the first-generation eejj analysis, events are required to pass a double-electron trigger or a double-photon trigger, with an electron or photon p T > 33 GeV. For the first-generation ejj analysis, events are required to pass either a single-electron trigger or a trigger based on the requirement of one electron with p T threshold between 17 and 30 GeV, missing transverse energy threshold between 15 and 20 GeV, and two jets with p T threshold between 25 and 30 GeV. The trigger thresholds vary according to the run period. For the second-generation leptoquark analyses, events are required to pass a single-muon trigger without isolation requirements and with a p T threshold of 40 GeV. For the eejj channel, the trigger efficiency is greater than 99%. For the ejj channel, the electron trigger efficiency is measured to be 95%. For the jj and jj channels, the single-muon trigger efficiency is measured to be 92% per muon.
Electron candidates [19] are required to have an electromagnetic cluster with p T > 40 GeV and pseudorapidity jj < 2:5ð2:2Þ for the eejj (ejj) analysis, excluding the transition region between the barrel and the end cap detectors, 1:44 < jj < 1:57. The ejj analysis requires lower electron jj to reduce the QCD multijet background, with negligible reduction of the signal acceptance. Electron candidates are required to have an electromagnetic cluster in the ECAL that is spatially matched to a reconstructed track in the central tracking system in both and the azimuthal angle , and to have a shower shape consistent with that of an electromagnetic shower. Electron candidates are further required to be isolated from additional energy deposits in the calorimeter and from reconstructed tracks beyond the matched track in the central tracking system. In addition, to reject electrons coming from photon conversion in the tracker material, the track associated with the reconstructed electron is required to have hits in all inner tracker layers.
Muons are reconstructed as tracks in the muon system that are matched to the tracks reconstructed in the inner tracking system [20]. Muons are required to have p T > 40 GeV and to be reconstructed in the HLT fiducial volume, i.e. with jj < 2:1. In addition, muons must be isolated by requiring that the tracker-only relative isolation be less than 0.1. Here, the relative isolation is defined as a sum of the transverse momenta of all tracks in the tracker in a cone of ÁR ¼ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi ðÁÞ 2 þ ðÁÞ 2 p ¼ 0:3 around the muon track (excluding the muon track), divided by the muon p T . To have a precise measurement of the transverse impact parameter of the muon track relative to the beam spot, only muons with tracks containing more than 10 hits in the silicon tracker and at least one hit in the pixel detector are considered. To reject muons from cosmic rays, the transverse impact parameter with respect to the primary vertex is required to be less than 2 mm.
Jets and E miss T are reconstructed using a particle-flow algorithm [21], which identifies and measures stable particles by combining information from all CMS subdetectors. The E miss T calculation uses calorimeter estimates improved by high precision inner tracking information as well as corrections based on particle-level information in the event. Jets are reconstructed using the anti-k T [22] algorithm with a distance parameter of R ¼ 0:5. The jet energy is calibrated using p T balance of dijet and þ jet events [23]. In the eejj and jj (ejj and jj) channels, jets are required to have p T > 30ð40Þ GeV, and jj < 2:4. Furthermore, jets are required to have a spatial separation from electron or muon candidates of ÁR > 0:3.
The initial selection of eejj or jj events requires two electrons or two muons and at least two jets satisfying the conditions described above. The two leptons and the two highest-p T jets are selected as the decay products from a pair of leptoquarks. The invariant mass of the two electrons (muons) is required to be M '' > 60ð50Þ GeV. To reduce the combinatorial background, events with a scalar transverse energy S '' For the ejj (jj) initial selection, events are required to contain one electron (muon) satisfying the conditions described above and at least two jets with p T > 40 GeV and E miss T > 55 GeV. The jet p T threshold is higher than that in the dilepton channels to account for jet p T thresholds in triggers used in the ejj channel. A veto on the presence of extra muons (electrons) is also applied. The angle in the transverse plane between the leading p T jet and the E miss T vector is required to be ÁðE miss T ; j 1 Þ > 0:5 to reject events with misreconstructed E miss T . To reduce the contribution from QCD multijet events, the lepton and the E miss T are required to be separated by ÁðE miss T ; 'Þ > 0:8. In addition, events are rejected if the scalar transverse energy GeV. The initial selection criteria are summarized in Table I.

III. SIGNAL AND BACKGROUND MODELING
The MC samples for the signal processes are generated for a range of leptoquark mass hypotheses between 250 and 900 GeV, with a renormalization and factorization scale % M LQ . The MC generation uses the PYTHIA generator [24] (version 6.422) and CTEQ6L1 parton distribution functions (PDF) [25]. The MC samples used to estimate the contribution from SM background processes are t " t þ jets events, generated with MADGRAPH [26]; single-top events (s, t, and tW channels), generated with POWHEG [27]; Z= Ã þ jets events and W þ jets events, generated with SHERPA [28]; VV events, where V represents either a W or a Z boson, generated with PYTHIA; QCD muon-enriched multijet events, generated with PYTHIA in bins of transverse momentum of the hard-scattering process from 15 GeV to the kinematic limit. The simulation of the CMS detector is based on GEANT4 [29] and includes multiple collisions in a single bunch crossing corresponding to the luminosity profile of the LHC during the data taking periods of interest.

IV. BACKGROUND ESTIMATE
The main processes that can mimic the signature of a leptoquark signal in the ''jj channels are Z= Ã þ jets, t " t, VV þ jets, W þ jets, and QCD multijets. The Z= Ã þ jets background is determined by comparing events from data and MC samples in two different regions: in the region of low (L) dilepton invariant mass around the Z boson mass (70 < M '' < 100 GeV for electrons and 80 < M '' < 100 GeV for muons) and in the region of high (H) mass M '' > 100 GeV. The low mass scaling factor R Z ¼ N L =N MC L is measured to be 1:27 AE 0:02 for the eejj channel and 1:29 AE 0:02 for the jj channel, where N L and N MC L are the number of data and MC events, respectively, in the Z mass window. The number of Z= Ã þ jets events above 100 GeV is then estimated as is the number of MC events with M '' > 100 GeV. The estimated number of Z= Ã þ jets events is obtained with the selection criteria optimized for different leptoquark mass hypotheses, and it is used in the limit setting procedure.
The number and kinematic distributions for the t " t events with two leptons of the same flavor are estimated from the number of data events that contain one electron and one muon. This type of background is expected to produce the ee final state or the final state with half the probability of the e final state. In the data the number of ee or events is estimated to be where and e are the muon and electron reconstruction and identification efficiencies and trig are the HLT efficiencies to select ee, , and e events.  Events / 20 GeV No QCD multijet MC events pass the jj final selection. A cross-check made using a data control sample containing same-sign muons confirms that the QCD multijet background is negligible in this channel.
For the first-generation leptoquark analyses the multijet background contribution is estimated from a data control sample as follows. The probability that an electron candidate passing loose electron requirements additionally passes all electron requirements is measured as a function of p T and in a data sample with one and only one electron candidate, two or more jets, and low E miss T . This sample is dominated by QCD multijet events and is similar in terms of jet activity to the eejj and ejj analysis samples. A correction for a small contamination of genuine electrons passing all electron requirements is derived from MC simulations. The QCD multijet background in the final eejj (ejj) selection is predicted by applying twice (once) the above probability to a sample with two electron candidates (one electron candidate and large E miss T ), and two or more jets, which satisfy all the requirements of the signal selections. The resulting estimate is $1% ( $ 8%) of the total background for the selections corresponding to the region of leptoquark masses where the exclusion limits are placed.
Contributions to the ''jj background from VV þ jets processes and single-top production are small, and they are estimated using MC simulation.
In the 'jj channel, the main backgrounds come from three sources: processes that lead to the production of a genuine W boson such as W þ jets, t " t, singletop production, diboson processes (WW, WZ); instrumental background, mostly caused by the misidentification of jets as leptons in multijet processes, thus creating misidentified electrons or muons and misreconstructed E miss T in the final state; and Z boson production, such as Z= Ã þ jets and ZZ processes. The contribution from the principal backgrounds, W þ jets and t " t, is estimated with MC simulation normalized to data at the initial selection level in the region 50 < M T < 110 GeV, where M T is the transverse mass calculated from the lepton p T and E miss T . The region 50 < M T < 110 GeV is used to determine both the W þ jets and the t " t normalization factors using two mutually exclusive selections (less than four jets or at least four jets with p T > 40 GeV and jj < 2:4) that separately enhance the samples with W þ jets and with t " t events. The results of these two selections are used to form a system of equations: where N i , N i;W , N i;t" t , N i;O , and N i;QCD are the number of events in data, W þ jets MC simulation, t " t MC simulation, other MC-simulated backgrounds, and QCD multijet events obtained from data, passing selection i. In the ejj analysis, the solution of the system yields the following normalization factors: R t" t ¼0:82AE0:04ðstatÞAE0:02ðsystÞ and R W ¼1:21AE0:03ðstatÞAE0:02ðsystÞ, where the systematic errors reflect the uncertainties on the QCD multijet background estimate. In the jj analysis, the normalization factors are R t" t ¼ 0:84 AE 0:03ðstatÞ and R W ¼ 1:24 AE 0:02ðstatÞ.
The contribution from QCD multijet processes after all selection criteria are applied to the jj sample is estimated to be negligible. No QCD MC events survive the full selection criteria optimized for any leptoquark mass hypothesis.   However, as multijet processes are difficult to accurately model by MC simulation, several cross-checks are made with data control samples to ensure that the QCD multijet background in the jj analysis is negligible. The method used to determine the QCD multijet background in the ejj analysis is similar to the one used for the eejj channel.

V. EVENT SELECTION OPTIMIZATION
After the initial selection, the sensitivity of the search is optimized by maximizing the Gaussian signal significance S= ffiffiffiffiffiffiffiffiffiffiffiffiffi S þ B p in all channels. Optimized thresholds on the following variables are applied for each leptoquark mass hypothesis in the ''jj channels: M '' , S '' T , and M min 'j . The invariant mass of the dilepton pair, M '' , is used to remove the majority of the contribution from the Z= Ã þ jets background. The variable M min 'j is defined as the smaller lepton-jet invariant mass for the assignment of jets and leptons to leptoquarks, which minimizes the LQ-LQ invariant mass difference.
Thresholds on the following variables are optimized for each leptoquark mass hypothesis in the 'jj channels: E miss T , S ' T , and M 'j . A minimum threshold on E miss T is used, primarily to reduce the dominant W þ jets background. The variable M 'j is defined as the invariant mass of the lepton-jet combination, which minimizes the LQ-LQ transverse mass difference. In addition, a lower threshold is applied on the transverse mass of the lepton (electron or muon) and E miss T in the event, M T > 120 GeV. The resulting optimized thresholds are summarized in Tables II and III for the ''jj and the 'jj channels, respectively.
After the initial selection criteria are applied, the yields in data are found to be consistent with SM predictions. Distributions of variables used in the final selection for the eejj, ejj, jj, and jj analyses are shown in Figs. 2-5. The number of events selected in data and estimated backgrounds are then compared at different stages of selection. This information is shown in Tables IV, V, VI, and VII for the initial selection and for the final selection for each channel separately.

VI. SYSTEMATIC UNCERTAINTIES
The uncertainty on the integrated luminosity is taken as 2.2% [30].
The statistical uncertainties on the values of R Z (R W ) after initial selection requirements are used as an estimate of the uncertainty on the normalization of the Z= Ã þ jets (W þ jets) backgrounds. The uncertainty on the shape of the Z= Ã þ jets and W þ jets distributions is calculated to be 15% (10%) and 20% (11%) for first (second) generation, respectively, by comparing the predictions of MADGRAPH samples produced with factorization or renormalization scales and matrix element-parton shower matching threshold varied up and down by a factor of 2.   The uncertainty on the estimate of the t " t background in the eejj and jj channels is derived from the statistical uncertainty of the ejj data sample and the ratio of electron and muon reconstruction uncertainties, which is calculated to be 2% and 3%, respectively. In addition, an uncertainty of 7% is assigned based on the estimated contamination from sources other than t " t in the data sample containing one electron and one muon. A 5.5% (5%) uncertainty on the normalization of the estimated t " t background in the ejj (jj) channel is given by the statistical uncertainty on the value of R t" t after the initial selection requirements. A 10% (10%) uncertainty on the shape of the t " t background distribution in the ejj (jj) channel is estimated by comparing the predictions of MADGRAPH samples produced with factorization or renormalization scales and matrix element-parton shower matching thresholds varied up and down by a factor of 2.
A systematic uncertainty of 50% (25%) on the QCD multijet background estimate for the eejj (ejj) channel is estimated from the difference between the number of observed data and the background prediction in a QCD multijet-enriched data sample with lower jet multiplicity.
PDF uncertainties on the theoretical cross section of leptoquark production and on the final selection acceptance have been calculated using the PDF4LHC [31] prescriptions, with PDF and s variations of the MSTW2008 [32], CTEQ6.6 [33], and NNPDF2.0 [34] PDF sets taken into account. Uncertainties on the cross section vary from 10% to 30% for leptoquarks in the mass range of 200-900 GeV, while the effect of the PDF uncertainties on signal acceptance varies from 1% to 3%. The PDF uncertainties are not considered for background sources with uncertainties determined from data. An uncertainty on the modeling of pileup interactions in the MC simulation is determined by varying the mean of the distribution of pileup interactions by 8%.
Energy and momentum scale uncertainties are estimated by assigning a 4% uncertainty on the jet energy scale, a 1% (3%) uncertainty on the electron energy scale for the barrel (end cap) region of ECAL, and a 1% uncertainty on the muon momentum scale. The effect of electron energy, muon momentum, and jet energy resolution on expected signal and backgrounds is assessed by smearing the electron energy by 1% and 3% in the barrel and end caps, respectively, by smearing the muon momentum by 4%, and by varying the jet energy resolution by an -dependent value in the range 5%-14%. In the 'jj analyses, the uncertainty on the energy and momentum scales and resolutions is propagated to the measurement of E miss T . The effect of these uncertainties is calculated for the (minor) background sources for which no data rescaling is applied. For the background sources for which data rescaling is applied, residual uncertainties are calculated (i.e. relative to the initial selection used to derive the rescaling factor).
Recent measurements of the muon reconstruction, identification, trigger, and isolation efficiencies using Z ! events show very good agreement between data and MC events [35]. An $1% discrepancy is observed in the data-to-MC comparison of the muon trigger efficiency. This discrepancy is taken as a systematic uncertainty per muon, assigned to both signal and estimated background. The electron trigger and reconstruction and identification uncertainties contribute 3% (4%) to the uncertainty in both signal and estimated background for the eejj (ejj) channel.
The systematic uncertainties and their effects on signal and background are summarized in Table VIII for all channels, corresponding to the final selection optimized for M LQ ¼ 600 GeV.

VII. RESULTS
The number of observed events in data passing the full selection criteria is consistent with the SM background prediction in all decay channels. An upper limit on the leptoquark pair-production cross section is therefore set using the CL S modified frequentist approach [17,18]. A log-normal probability function is used to integrate over the systematic uncertainties. Uncertainties of a statistical nature are described with À distributions with widths determined by the number of events simulated in MC samples or observed in data control regions.
The 95% CL upper limits on Â 2 or Â 2ð1 À Þ as a function of leptoquark mass are shown together with the NLO predictions for the scalar leptoquark TABLE VII. Individual background (BG) sources, expected signal, data, and total background event yields after the initial (first row) and final selections for the jj analysis. Other BG includes single top, Z þ jets, and VV þ jets. Only statistical uncertainties are reported.  . Top (bottom) frame: the expected and observed upper limits at 95% CL on the leptoquark pairproduction cross section times 2 ð2ð1 À ÞÞ as a function of the first-generation leptoquark mass obtained with the eejj (ejj) analysis. The expected limits and uncertainty bands represent the median expected limits and the 68% and 95% confidence intervals. The systematic uncertainties reported in Table VIII are included in the calculation. The left and middle shaded regions are excluded by the current ATLAS limit [15] and CMS limits [12,13] for ¼ 1ð0:5Þ in the eejj (ejj) channel. The right shaded regions are excluded by these results. The theory curves and their bands represent, respectively, the theoretical scalar leptoquark pair-production cross section, and the uncertainties attributable to the choice of PDF and renormalization/factorization scales.  FIG. 11 (color online). Top (bottom) frame: the expected and observed upper limits at 95% CL on the leptoquark pairproduction cross section times 2 ð2ð1 À ÞÞ as a function of the second-generation leptoquark mass obtained with the jj (jj) analysis. The expected limits and uncertainty bands represent the median expected limits and the 68% and 95% confidence intervals. The systematic uncertainties reported in Table VIII are included in the calculation. The left and middle shaded regions in the top frame are excluded by the current ATLAS limit [14] and CMS limit [11] for ¼ 1 in the jj channel only. The left shaded region in the bottom frame is excluded by the current ATLAS limit [14] for ¼ 0:5 in the jj channel only. The right shaded regions are excluded by these results. The theory curves and their bands represent, respectively, the theoretical scalar leptoquark pair-production cross section, and the uncertainties attributable to the choice of PDF and renormalization/factorization scales. second-generation scalar leptoquarks with masses less than 840 (620) GeV are excluded for ¼ 1ð0:5Þ. This is to be compared with median expected limits of 790 (640) GeV for first-generation scalar leptoquarks and 800 (610) GeV for second-generation scalar leptoquarks.
The observed and expected limits on the branching fraction as a function of leptoquark mass can be further improved using the combination of the ''jj and 'jj channels, as shown in Fig. 12. These combinations lead to the exclusion of first-and second-generation scalar leptoquarks with masses less than 640 and 650 GeV for ¼ 0:5, compared with median expected limits of 680 and 670 GeV.

VIII. SUMMARY
In summary, a search for pair production of first-and second-generation scalar leptoquarks has been performed in decay channels with either two charged leptons of the same flavor (electrons or muons) and at least two jets, or a single charged lepton (electron or muon), missing transverse energy, and at least two jets, using 7 TeV protonproton collisions data corresponding to an integrated luminosity of 5 fb À1 . The selection criteria have been optimized for each leptoquark signal mass hypothesis. The number of observed candidates for each hypothesis agree with the estimated number of background events. The CL S modified frequentist approach has been used to set limits on the leptoquark cross section times the branching fraction for the decay of a leptoquark pair. At 95% confidence level, the pair production of first-and secondgeneration leptoquarks is excluded with masses below 830 (640) GeV and 840 (650) GeV for ¼ 1ð0:5Þ. These are the most stringent limits to date.

ACKNOWLEDGMENTS
We extend our thanks to Michael Krämer for providing the tools for calculation of the leptoquark theoretical cross section and PDF uncertainty. We congratulate our colleagues in the CERN accelerator departments for the excellent performance of the LHC machine. We thank the technical and administrative staff at CERN and other CMS institutes, and acknowledge support from the following: . Top (bottom) frame: the expected and observed exclusion limits at 95% CL on the first-(second-) generation leptoquark hypothesis in the versus mass plane using the central value of signal cross section for the individual eejj and ejj (jj and jj) channels and their combination. The dark green and light yellow expected limit uncertainty bands represent the 68% and 95% confidence intervals. Solid lines represent the observed limits in each channel, and dashed lines represent the expected limits. The systematic uncertainties reported in Table VIII are included in the calculation. The shaded region is excluded by the current ATLAS limits [14,15]. Formation à la Recherche dans l'Industrie et dans l'Agriculture (FRIA-Belgium); the Agentschap voor Innovatie door Wetenschap en Technologie (IWT-Belgium); the Council of Science and Industrial Research, India; the Compagnia di San Paolo (Torino); and the HOMING PLUS programme of Foundation for Polish Science, cofinanced from European Union, Regional Development Fund.