Searches for high mass resonances with the CMS detector

New heavy resonances are predicted by many extensions of the standard model of particle physics. Recent results for high mass resonance searches with the Compact Muon Solenoid detector, in the diphoton, dilepton, dijet and ttbar channels, are discussed. Limits for numerous benchmark models are presented.


Diphoton Resonances
Kaluza-Klein (KK) gravitons predicted by RS warped extra dimensions may manifest themselves as high mass resonances in the diphoton invariant mass spectrum. The diphoton channel has the advantage that the branching ratio for spin-2 gravitons is twice that to leptons. Two isolated photons are selected with E T > 70 GeV and |η| < 1.44. The resulting diphoton invariant mass distribution with 2.2 fb −1 of data is shown in Fig. 1 [4]. The expected background arising from irreducible SM diphoton production is estimated using simulation, scaled by a next-to-leading order mass dependent K factor. Instrumental backgrounds, arising from γ+jet and dijet processes, in which the jets are misidentified as photons, are estimated using a datadriven fake rate method. Observing no excess in the diphoton invariant mass distribution above SM expectations, upper limits are set on the production cross section for RS a e-mail: toyoko.orimoto@cern.ch gravitons, using the CL S technique [5,6]. The limits on the cross section are translated into lower limits on the model parameters (Fig. 2), where M 1 is the mass of the first graviton excitation, andk is a dimensionless parameter which quantifies the strength of the graviton coupling to SM fields. We exclude at the 95% confidence level (CL) resonant graviton production in the RS1 model with values of M 1 < 0.86 − 1.84 TeV, depending onk.
[TeV] tron and dimuon invariant mass distributions from 1.1 fb −1 of data can be seen in Fig. 3 [7]. The expected background from irreducible SM Drell-Yan production is estimated using simulation, normalized to the Z 0 peak in data. Backgrounds with prompt leptons (tt, tW, dibosons) are crosschecked using a data-driven method counting e-µ pairs. Backgrounds with jets misidentified as leptons (W+jet, dijets) are estimated using a fake rate measured from a jetenriched data sample. Lastly, cosmic muon backgrounds are rejected by imposing topological criteria. The dilepton analysis is a shape-based search, making no assumptions on the absolute background rate; this is achieved by normalizing the results to the Z 0 peak. We set limits, using a Bayesian technique, on the ratio (R σ ) of the cross section for Z (or G KK ) production to the cross section for SM Z 0 production (Fig. 4). The limits on R σ can be interpreted as lower limits on Z (G KK ) mass. We exclude at 95% CL a Z with SM-like couplings (Z SSM ) with mass < 1940 GeV, the superstring-inspired Z ψ < 1620 GeV, and RS G KK < 1450 (1780) GeV fork =0.05 (0.10).
We require the two leading jets have |η| <2.5 and |∆η| <1.3 and dijet invariant mass > 838 GeV. To recover radiation lost through FSR and to improve the dijet mass resolution, we combine particle flow [13] jets with the anti-k T algorithm (R = 0.5) into "wide jets". QCD multijets comprise the main background, following a smoothly falling dijet mass distribution predicted by the SM. Fig. 5 shows the dijet invariant mass spectrum with 1.0 fb −1 , where the expected background from QCD multijets is described with a functional fit [14]. The systematic uncertainties from the   Fig. 4. Upper limits on the production ratio R σ of cross section times branching fraction into lepton pairs as a function of resonance mass, for Z SSM , Z ψ , and G KK . The limits are shown from the combined dilepton (ee + µ + µ − ) result. Shaded yellow and red bands correspond to the 68% and 95% quantiles for the expected limits. The predicted cross section ratios are shown in bands, with widths indicating theoretical uncertainties.  jet energy scale and resolution are 2% and 10%, respectively. The 95% CL upper limits on the product of the cross section and branching ratio and acceptance, computed using a Bayesian approach, are shown in Fig. 6

tt Resonances: Semileptonic Decay
New bosons with enhanced coupling to the top quark appear in many SM extensions, such as those predicting axigluons and KK gluons [15]. We present a search for heavy tt resonances in the semileptonic (qqb)(µνb) final state, focusing on highly boosted top pairs with decay products narrowly collimated along the direction of the top. Backgrounds arise from SM tt, W/Z+jets, single top, and QCD multijets.
For high mass tt, the decay products of the hadronicdecaying top can have small opening angles in the detector frame. Thus, instead of requiring four jets, we require two particle flow jets with p T > 50 GeV and |η| < 2.4, with the leading jet p T > 250 GeV; jets are reconstructed with the anti-k T algorithm (R =0.5). The high top p T also results in low ∆R = (∆φ) 2 + (∆η) 2 between the µ and b, making it difficult to require the muon be well-isolated. To suppress QCD multijet backgrounds, we thus apply a two-dimensional requirement, ∆R > 0.5 or p T,rel > 25 GeV, where p T,rel is the magnitude of the p µ component orthogonal to the jet axis. In addition, muons are required to have p T > 35 GeV and |η| < 2.1. Events with additional muons or electrons (from tt and Z 0 decays) are vetoed. Lastly, H T,lep , the scalar sum of the muon p T and missing transverse energy (MET), is required to be > 150 GeV. Fig. 7 shows the resulting 95% CL upper limits on the cross section for a benchmark topcolor Z [16], computing using a Bayesian method. The largest uncertainties come from the jet energy resolution (10−20%) and the jet energy scale (2 − 3%). With 1.1 fb −1 , we exclude a topcolor Z of width 3% in the mass regions 805 < m Z < 935 GeV and 960 < m Z < 1060 GeV.

tt Resonances: All-Hadronic Decay
The motivations for studying the fully hadronic decay of tt are similar to those for the semileptonic search. Likewise, the all-hadronic search exploits the highly boosted nature of the top quarks from high mass resonances. Moreover,  Fig. 8. The 95% CL upper limit on a product of the production cross section of Z and the branching fraction for its decay into tt pairs, as a function of assumed Z mass, for a combination of "1+2" and "1+1'" channels. Three theoretical models are examined in the dashed lines: a Kaluza-Klein gluon model and a topcolor Z model (updated to √ s = 7 TeV via private communication) [17] with widths 3% and 1.2%. the all-hadronic decay benefits from a higher branching ratio than the semileptonic decay.
In this analysis, each event is divided into hemispheres, such that each hemisphere contains the final products of each top. Then, the top decays are classified into categories, depending on the how boosted the top is: (1) "high boost" tops are those in which all three jets are merged into one top jet and (2) "moderate boost" tops are those in which only two out of three of the jets are merged. We conduct the search in two categories: "type 1+1", which have two highly boosted top jets, or "type 1+2", which are threejet events. Jets are reconstructed using particle flow and Cambridge-Aachen clustering algorithms. The dominant background comes from QCD multijets, which is estimated with a data-driven top-tagging mistag rate. The small continuum tt contribution is estimated with simulation. The limits are evaluated with a counting experiment, using a Bayesian procedure. Fig. 8 depicts the 95% CL upper limits on the product of the cross section of Z and the branching ratio for its decay into tt pairs [18]. With 886 pb −1 , we exclude the KK gluon masses between 1.0 − 1.5 TeV.

Conclusions
We present searches for high mass resonances with the CMS detector in the diphoton, dilepton, dijet, and tt channels. Observing no excess above standard model predictions, we set limits on a variety of benchmark models, including those predicting gravitons and Z .