A search is performed for the pair production of a new heavy resonance that decays into high jet-multiplicity final states with missing transverse energy using 4.98 ± 0.11 fb^{-1} of data collected by the CMS experiment in 2011 at the CERN Large Hadron Collider. The razor variables are used to study events with at least six jets, where at least one jet is identified as coming from the decay of a b-quark. These events are characterised by large M_{R} and small R^{2} and are in a different kinematic region than previously studied. This analysis is sensitive to SUSY and SUSY-like signatures with the production of a high number of jets, such as top squark or t′ pair production.
The analysis uses the Razor kinematic variables to discriminate between signal and background. One of the variable, M_{R}, is a measurement of the typical energy scale of the process and it is related to the masses of the SUSY particles for signal events. The other variable, R = M_{R}^{T}/M_{R}, quantifies the unbalance the event, as a result of the presence of two missing particles. The numerator of R, MRT, is smaller than the value M_{Δ} at which M_{R} peaks. This results in a value of R smaller than one. While signal events have R ~ 0.3, events from background with no real missing transverse energy have R ~ 0.
p-value of the observed yield in bins of M_{R} and R^{2} with respect to the prediction from the background model. The background model is obtained from a ML fit in the green region, then extrapolated to the signal region.
Projection of the 2D fit result on M_{R} (left) and R^{2} (right) for the BJet box using the full 2011 CMS MultiJet dataset. The blue histogram is the total SM prediction as obtained from a single large pseudo-experiment based on the 2D fit, while the histogram uncertainties show the 68% range from the covariance matrix. The breakdown of the different background components is also shown. The fit is performed in the R^{2}-M_{R} sideband and projected into the full region to allow comparison between the prediction from the fit and the data. The ratio shown is the full SM prediction divided by the data distribution. No significant trend is seen.
We interpret the absence of an excess in the M_{R} vs R^{2} tail in terms of an cross-section limit on the pair production of a heavy top partner t', which then decays to a top quark and a missing particle. The cross-section limit is translates into a mass limit under the assumption that the t' is a scalar (e.g. a stop) or a fermion (e.g. a KK partner of the top in UED)
Expected and observed 95% CL limits in the (mt′, mLSP) SMS T2tt plane from the razor multijet analysis as derived from the BJet box. The black (red) lines show the observed (expected)limit for the scalar (left) and the spin-1/2 fermion (right) t'.The theoretical uncertainties arising from 0 1/2 scale and parton distribution function uncertainties are also shown by the narrow lines. The colour scale shows the model independent cross-section excluded in this SMS. The solid grey region indicates model points where the analysis was found to have dependence on initial state radiation modelling in the simulation of signal events above a pre-defined tolerance; no interpretation is presented for these model points.
Expected and observed limits for pair-produced top partners with the mass of the invisible particle fixed to 50 GeV. In each case, the relevant theoretical cross-sections and uncertainties are shown.
Signal MC efficiencies,relative to the total number of events per SMS point, for the LeptonBVeto (left), BVeto (center), and BJet (right) boxes as a function of the SMS T2tt model parameters. It can be seen that the signal efficiency in the two data control regions is very low.
Average values of M_{R} (left) and R^{2} (right) for the BJet box as a function of the SMS T2tt parameters.
We interpret the absence of an excess in the MR vs R2 tail in terms of an cross-section limit on the pair production of two gluinos, each decaying to a pair of top quarks and a LSP.
Expected and observed limits for pair-produced gluinos with the mass of the invisible particle fixed to 50 GeV. In each case, the relevant theoretical cross-sections and uncertainties are shown. No model points where the analysis was seen to have dependence on initial state radiation modelling in the simulation of signal events above a pre-defined tolerance were found.
Signal MC efficiencies,relative to the total number of events per SMS point, for the LeptonBVeto (top left), BVeto (top right), and BJet (bottom) boxes as a function of the SMS T1tttt model parameters. It can be seen that the signal efficiency in the two data control regions is very low.
Average values of M_{R} (left) and R^{2} (right) for the BJet box as a function of the SMS T1tttt parameters.
Expected limit for the pair-production of gluinos (left) and the top-partners t' (right). The black lines correspond to the expected limit in the case of a SUSY-like QCD cross section, as well as for a factor-3 enhanced and a factor-3 reduced cross section. The red lines in the right plot show the corresponding expected limits for a sermonic partner of the top.