Searches for supersymmetry using the MT2 variable in hadronic events produced in pp collisions at 8 TeV

Searches for supersymmetry (SUSY) are performed using a sample of hadronic events produced in 8 TeV pp collisions at the CERN LHC. The searches are based on the MT2 variable, which is a measure of the transverse momentum imbalance in an event. The data were collected with the CMS detector and correspond to an integrated luminosity of 19.5 inverse femtobarns. Two related searches are performed. The first is an inclusive search based on signal regions defined by the value of the MT2 variable, the hadronic energy in the event, the jet multiplicity, and the number of jets identified as originating from bottom quarks. The second is a search for a mass peak corresponding to a Higgs boson decaying to a bottom quark-antiquark pair, where the Higgs boson is produced as a decay product of a SUSY particle. For both searches, the principal backgrounds are evaluated with data control samples. No significant excess over the expected number of background events is observed, and exclusion limits on various SUSY models are derived.


Searches for supersymmetry using the M T2 variable in hadronic events produced in pp collisions at 8 TeV Abstract
Searches for supersymmetry (SUSY) are performed using a sample of hadronic events produced in 8 TeV pp collisions at the CERN LHC. The searches are based on the M T2 variable, which is a measure of the transverse momentum imbalance in an event. The data were collected with the CMS detector and correspond to an integrated luminosity of 19.5 fb −1 . Two related searches are performed. The first is an inclusive search based on signal regions defined by the value of the M T2 variable, the hadronic energy in the event, the jet multiplicity, and the number of jets identified as originating from bottom quarks. The second is a search for a mass peak corresponding to a Higgs boson decaying to a bottom quark-antiquark pair, where the Higgs boson is produced as a decay product of a SUSY particle. For both searches, the principal backgrounds are evaluated with data control samples.

Link to arXiv
This analysis has been documented in the arXiv:1502.04358 . It has been submitted to JHEP.

Plots from SUS-13-019
Figures from the paper pdf, png

Figure Formats
Caption pdf, png Figure 3: The ratio r(M T2 ), described in the text, as a function of M T2 for events satisfying the medium-H T and the (N j = 3 5, N b = 0) requirements of the inclusive-M T2 search. The solid circle points correspond to simple data yields, while the points with open circles correspond to data after the subtraction of the non-multijet backgrounds, as estimated from simulation. Two different functions, whose exponential components are

Figure Formats
Caption pdf, png  The lost-lepton background is estimated from data control samples, while the +jets is evaluated using simulation. The uncertainties in each plot are drawn as the shaded band and do not include the uncertainty in the shape of the lost-lepton background. The signal model consists of gluino pair production events with one of the two gluinos containing an h boson in its decay chain. For this model it is assumed = 750 GeV and = 350 GeV.
pdf, png Figure 16: Exclusion limits at 95&percnt; CL as a function of and for the cMSSM/mSUGRA model with tanβ = 30, A 0 = -2max(m 0 , m 1/2 ), and µ < 0. Here, is the average mass of the the first-generation squarks.
pdf, png Table 5: Summary of observed mass limits (at 95&percnt; CL) for different SUSY simplified models and for the cMSSM/mSUGRA model. The limits quoted are the observed limits using the signal cross section minus one standard deviation (σ theory ) of its uncertainty. For the simplified models, the limit on the mass of the parent particle is quoted for , while for the LSP the best limit on its mass is quoted. The best limit on the mass splitting between the parent particle mass and the LSP mass is also given. Finally, the absolute limits on the squark and gluino masses are quoted for the cMSSM/mSUGRA model.

Figure:
Extended maximum likelihood fit to the shower shape variable σ iηiη for barrel photons (left) and endcap photons (right). The green and pink histograms correspond to the signal and background templates respectively. The blue histogram shows the total sum. distribution: pdf, png ratio: pdf, png

Figure:
Left: Comparison between data and simulation for the Z(ll) and γ M T2 distributions with N j ≥ 2, N b = 0, H T > 450 GeV, and boson-p T > 80 GeV. The boson-p T is added to the E T miss vector. The data correspond to 19.5 fb -1 . Backgrounds have been subtracted using simulation. Right: The corresponding ratios Z(ll)/γ in data and simulation. distribution: pdf, png ratio: pdf, png

Additional Material for Results and Exclusion limits, Signal Efficiencies
We provide in electronic from (as a ROOT file) the signal efficiency for all used search regions separately. You access a specific bin following scheme: For inclusive M T2 analysis: The root files contain histograms (TH2F) for each search bin called: where [HT] is the H T region under consideration, i.e. lowHT, mediumHT, or highHT, • [topological] is the topological region under consideration, i.e. 2j0b, 2j1to2b, 3to5j0b, 3to5j1b, 3to5j2b, ge6j0b, 6j1b, 6j2b, or 3b,
• As an example: For the signal region with 3 ≤ N j ≤ 5, 2 b jets, medium H T , and M T2 = 270-380 GeV the histograms name is mediumHT_3to5j2b_MT2-270to280.

PhysicsResultsSUS13019 < CMSPublic < TWiki
Additional Material for Results and Exclusion limits, Signal Efficiencies topological regions used are (3 ≤ N j ≤ 5, N b = 0), (3 ≤ N j ≤ 5, N b = 1), (N j ≥ 6, N b = 0), (N j ≥ 6, N b = 1). root Signal efficiency for gluino pair production, with . The topological regions used are (3 ≤ N j ≤ 5, N b = 1), (3 ≤ N j ≤ 5, N b = 2), (N j ≥ 6, N b = 1), (N j ≥ 6, N b = 2), (N j ≥ 3, N b ≥ 3). root Signal efficiency for gluino pair production, with . The topological regions used are (N j ≥ 6, N b = 1), (N j ≥ 6, N b = 2), (N j ≥ 3, N b ≥ 3). root Signal efficiency for gluino pair production with , , and , . In this scenario, neutralino and chargino are degenerate, with a mass . pdf, png Figure: The ratios of the data yields divided by the data-driven background estimates for the three H T signal regions as well as the topological regions in the jet − b-jet multiplicity. The M T2 signal regions are summed up for each of the H T and topological regions. The uncertainties on the ratio is only the statistical uncertainty of the data. The uncertainties of the data-driven background estimates are shown in the shaded uncertainty band. The data correspond to an integrated luminosity of 19.5 fb −1 .
We evaluate the compatibility of the data and the standard model prediction by computing the pull value for every signal bin. The pull is defined via , where N obs is the observed number of events with σ obs its statistical uncertainty and N bkg is the background estimate with a total uncertainty σ bkg . A study using simulated pseudo-data shows that the observed pull distribution in data is compatible with the prediction with a probability of 11&percnt;. pdf, png  analysis after each stage of the event selection, gluino pair production with , with a gluino mass of 550 GeV and an LSP mass of 500 GeV. The baseline selection requires at least two jets with p T > 100 GeV and for the jet/E T miss quality criteria requirements. The uncertainties are statistical. pdf, png Table: Predicted number of signal events in the inclusive M T2 analysis after each stage of the event selection, gluino pair production with , with a gluino mass of 1275 GeV and an LSP mass of 100 GeV. The baseline selection requires at least two jets with p T > 100 GeV and for the jet/E T miss quality criteria requirements. The uncertainties are statistical.