Search for supersymmetry in events with two leptons including a tau

Searches for new physics in events with hadronic jets, missing transverse energy, and two leptons of which at least one is a hadronically decaying tau are presented. The result is based on a data sample corresponding to an integrated luminosity of 1 $\fbi$ at a center-of-mass energy of 7 TeV collected by the CMS experiment at the LHC. No significant excess with respect to the standard model predictions is found.


Introduction
This article summarizes searches for physics beyond the Standard Model (BSM), analyzing Ldt ≈ 1fb −1 of data recorded by the Compact Muon Solenoid Experiment (CMS). Proton-proton collisions where provided by the LHC at center of mass energies of √ s = 7TeV in 2011 . CMS conducted several searches for BSM in finale states characterized by large missing transverse energy (E miss T ) and hadronic activity. On the one hand, the high E miss T signature occurs in models with weakly interacting particles that escape detection, which are favored by cosmological measurements. On the other hand, hadronic activity accours naturally in colored particle interactions dominating BSM cross sections in proton-proton collisions. In this article we focus on finale states containing a combination of two leptons, at least one of which is required to be a hadronically decaying tau (eτ h , µτ h , or τ h τ h ). The charges of these two leptons can be either of same sign [1] or opposite sign [2], leading to different selection and background estimation strategies. The study concerning same and opposite sign finale are described in sections 3 and 4, respectively.
In both cases hadronic jets and E miss T are reconstructed using the particle flow technique [3] and jets are clustered using the anti-kt algorithm [4]. The amount of hadronic activity in the event is measured by the quantity H T = p jet T and the requirement of two or more jets per event. Hadronically decaying tau leptons (τ h ) are identified using the HPS algorithm [5]. To suppress multi-jet QCD backgrounds all leptons are required to be isolated.

The CMS Detector
The central feature of the Compact Muon Solenoid (CMS) apparatus is a superconducting solenoid, of 6 m internal diameter, providing a field of 3.8 T. Within the solenoids volume are the silicon pixel and strip tracker, the crystal electromagnetic calorimeter (ECAL) and the brass/scintillator hadron calorimeter (HCAL). Muons are measured in gasionization detectors embedded in the steel return yoke. In a e-mail: edelhoff@cern.ch addition to the barrel and endcap detectors, CMS has extensive forward calorimetry. A much more detailed description of CMS can be found elsewhere [6].

Same Sign Search
Finale states with two leptons of the same charge are rare in the Standard Model (SM). Thus, the main backgrounds for this search are quark or gluon jets misidentified as a τ h (e.g. in W+jets events) and events where the charge of one of the leptons is misidentified (e.g. in dileptonic tt events). Both backgrounds are estimated directly from data as described in sections 3.2 and 3.3. The influence of light leptons not produced in the hard scattering (e.g. from heavy flavor decays) is small compared to that of misidentified τ h due to the abundance of hadronic jets in the region of interest. Contributions from rare same sign SM processes such as diboson production, double W strahlung, or double parton scattering are small and estimated using simulation.

Event Selection
The same sign search region is selected requioring H T > 350 GeV and E miss T > 80 GeV. For the requiroment of two or more jets and in calculating H T jets with transverse momenta p T > 40 GeV are considered. Leptons are required to be in |η| < 2.4 and have transverse momenta of p e T > 10 GeV, p µ T > 5 GeV, p τ h T > 15 GeV, to ensure efficient trigger selection.

Estimating Misidentified Jet Contribution
The HPS τ h identification algorithm distinguishes hadronic jets created in the decay of a τ lepton from those created in the hadronisation of a quark or gluon by means of isolation and reconstructed particle content. Nonetheless, the selection of hadronically decaying τ leptons always includes a remaining contamination by misidentified quark or gluon jets.
In order to estimate this contamination we employ the tight-to-loose method (TL). First, we define a loose τ h selection by relaxing the isolation requirement, in addition to arXiv:1201.5008v1 [hep-ex] 24 Jan 2012 EPJ Web of Conferences Table 1. Predicted backgrounds and observed event yields in the search region (H T > 350 GeV and E miss T > 80 GeV). Statistical and systematic errors have been added in quadrature. The upper limit is set using the CL S method.
the tight τ h selection used in the analysis. Second, we measure the fraction f T L of loose candidates, which pass the tight criteria in a sample containing predominately quark and gluon jets. Finally, we extrapolate the expected number of misidentified tight τ h candidates from the number of observed loose candidates in the signal region.
The tight-to-loose ratio f T L mainly depends on the transverse momentum and pseudorapidity of a given τ h candidate and is measured in bins of those variables. The difference in H T of the region where f T L is measured and the search regions are minimised in the definition of the loose selection. The results of this procedure have been shown to be in good agreement with background simulation.

Estimating Misidentified Charge Contribution
Backgrounds due to charge misidentification arise from the relative abundance of SM processes with two leptons of opposite charge, at least one of which is an electron or hadronically decaying τ. The contribution of muons with misidentified charge is found to be negligible. The misreconstruction of Electron charge occurs due to energy loss in the tracking volume. Furthermore, τ h charge misidentification occurs in three prong τ h decays, when a track form the background is wrongly associated with the τ h object.
To estimate the impact of these effects, we compare the number of opposite-and same sign dilepton pairs near the Z resonance. The Drell-Yan (DY) signal is fitted in the dilepton invariant mass spectrum alongside backgrounds form misidentified leptons and other SM processes. We identify the probability f q of lepton charge misidentification as the ratio of dilepton pairs recontructed with the same sign to those reconstructed with opposite sign.
For electrons we measure f e q = 2 · 10 −4 (3 · 10 −3 ) in the ECAL barrel (endcap). Differences arrise due to differences in the amount of tracker material in front of the ECAL crystals.
Again results from this data driven background estimation are in agreement with background simulation.

Results
A summary of the predicted background and the observed yield in the search region is given in table 1. We do not observe evidence of an event yield in excess of the SM based predictions and set 95% CL upper limits (UL) on the number of observed BSM events. The hybrid frequentistbaysian CL S method [7] is applied, including nuisance parameters and the signal strength maximizing the ratio of the signal with background and background only likelihoods.

Opposite Sign Search
In contrast to the same sign search there are several SM processes, such as DY and tt decays, with finale states containing two leptons of oposite charge. The background from DY processes can be sufficiently suppressed by the choice of search region. However, contributions from tt decays remain. For the channels eτ h and µτ h those are estimated from data as described in section 4.2. Naturally backgrounds due to misidentified quark and gluon jets remain and are estimated as described in section 3.2. Furthermore, the fully hadronic finale state τ h τ h is treated differently than the other finale states: Here we define background enriched sideband regions for each considered background. In this instance f bkg T L is taken from simulation. Backgrounds are predicted extrapolating from the number of loose candidates in a background enriched sideband to the search region.

Event Selection
For the finale states containing a light lepton (eτ h and µτ h ) two search regions are defined (Fig. 1). For brevity we focus on the high E miss

Estimating Dileptonic tt Contribution
We use the dilepton transverse momentum (p T ( )) method to estimate the contribution of dileptonic tt events in the signal region of opposite sign eτ h and µτ h events. We estimate the contribution of tt events to the corresponding light lepton channels (ee, eµ, and µµ), following the idea [8] that the variable p T ( ) can be used to model E miss T = p T (νν) [9].
Here, we exploit the fact that in dileptonic tt decays the p T distributions of the leptons are related to those of the neutrinos via the common boosts from the intermediate top and W decays. This relation is governed by the well understood W polarization, which can be reliably accounted for.
Contamination by events which stem from Z decays is first reduced by a 76 < m < 106 GeV and a E miss T > 50 GeV requirement. The remaining contribution is then predicted and subtracted using the same procedure as in Ref. [10]. The bias of the p T ( ) distribution due to the E miss T requirement is measured and accounted for. Finally, lepton universality allows us to extrapolate from the light lepton channels to the τ h channels in question. In this, τ h reconstruction efficiency, acceptance, and branching ratios are taken from simulation.

Results
A summary of the observed event yields and the data driven background predictions in the search regions is given in table 2. We observe no excess of events over the SM predictions. Also these predictions are shown to be in agreement with SM expectations from simulation.
We procede to evaluate three benchmark scenarios, referred to as LM1, LM2 and LM13, of the minimal supersymmetric extension of the standard model (cMSSM) [11]. , and µ > 0 [12]. We place 95% confidence level upper limits on the cross section of those scenarios using again the CL S method ( Table 3). The combination of the three channels takes differences in search regions and correlations of the uncertainties into account. All three scenarios are ruled out by the presented results. Furthermore we publish [2] additional information to allow testing of specific BSM models against these results.

Conclusion
Searches for physics beyond the standard model with τ h final states using ≈ 1fb −1 of integrated luminosity are summarized. Dominant backgrounds are estimated from the data taking the challenges of the hadronic τ finale state into account. No deviation from the SM is found and 95% CL upper limits are computed.