Invariant Mass Distribution of Jet Pairs Produced in Association with a W boson in ppbar Collisions at √s = 1.96 TeV

We report a study of the invariant mass distribution of jet pairs produced in association with a W boson using data collected with the CDF detector which correspond to an integrated luminosity of 4.3 fb⁻¹. The observed distribution has an excess in the 120-160 GeV/c² mass range which is not described by current theoretical predictions within the statistical and systematic uncertainties. In this letter we report studies of the properties of this excess.

The published paper, Phys.Rev.Lett.106:171801 (2011), is available here and from arxiv here. The analysis was presented by Viviana Cavaliere at a Joint Experimental-Theoretical Seminar at FNAL on April-6. A copy of that talk is here.

We have updated this result and present further studies of the excess using additional data collected through to November 2010 corresponding to an integrated luminosity of 7.3 fb^-1. The significance of the excess increases from 3.2 (for 4.3 fb^-1) to 4.1 standard deviations consistent with expectation from the increase in the data sample size. The event selection and analysis methodology remain unchanged from the original 4.3 fb^-1publication. The 7.3 fb^-1 analysis is available here.

Additional studies of these W+dijet event are presented in the thesis of Viviana Cavaliere and below we include plots of additional distributions not included in the PRL submission due to space restrictions.

• Fig-1 of paper.
  

  The dijet invariant mass distribution. The sum of electron and muon events is plotted. In the left plots we show the fits for known processes only (a) and with the addition of a hypothetical Gaussian component (c). On the right plots we show, by subtraction, only the resonant contribution to M_JJ including WW and WZ production (b) and the hypothesized narrow Gaussian contribution (d). In plot (b) and (d) data points differ because the normalization of the background changes between the two fits. The band in the subtracted plots represents the sum of all background shape systematic uncertainties described in the text. The distributions are shown with a 8 GeV/c² binning while the actual fit is performed using a 4 GeV/c² bin size.

• Fig-2 of paper.
  

  The dijet invariant mass distribution for the sum of electron and muon events is shown after subtraction of fitted background components with the exception of resonant contribution to M_JJ including WW and WZ production and the hypothesized narrow Gaussian contribution (a). With respect to Figure 1, the subtracted background components are chosen as the systematic combination that best fit data. The fit χ²/ndf is 62.0/81. (b) ΔRjj distribution for events with M_JJ < 115 and M_JJ > 175 GeV of the data compared to the background estimation that corresponds to the same systematic combination of (a). The uncertainty band corresponds to background statistical uncertainty.

• Missing ET (MET) distributions and fits: CEM electrons, CMUP muons and CMX muons.

  These are used to determine the normalization of the multijet QCD events and a preliminary estimate of the W+jets fraction. The MET distribution of the data is fitted to a sum of background templates. The samples with known cross section (Z+jets, diboson, ttbar and single top) are fixed to their NLO cross sections, while the normalization of the QCD and W+jets samples are allowed to float in the fit. The parameters of the fit are the total normalization and the QCD fraction.

• Kinematic distribution of events with 3 or more jets which enhances the top contribution to the sample.

• W+jet/QCD discriminants : Δφ in the transverse plane between the purely calorimetric MET and MET evaluated using tracks.
  Δφ defined as the angle in the transverse plane between MET and the closest jet with raw E_T > 5 GeV.

• Z+jets. Z+2 jet events are selected by requiring one tight lepton (same selection of W+jets) and one lepton with P_T > 10 GeV/c and the invariant mass of the two leptons to be in the range 81-110 GeV/c².

• ΔR distributions.
  The ΔR distribution is studied to investigate possible effects that could result in a mismodeling of the dijet invariant mass distribution. Two control regions are studied: the first defined by events with M_JJ < 115 and M_JJ > 175 GeV/c² and the second defined by events with pT_JJ < 40 GeV/c. We use these regions to derive a correction as a function of ΔR so that ΔR agrees with the Monte-Carlo and apply this reweight to the events in the excess region. We find that the statistical significance of the result is reduced by one sigma for reweightings from the M_JJ sidebands and increased by one sigma using reweights based on the pT_JJ < 40 GeV/c sample.

• Alternate background shape.
  A simple exponential background shape was fitted to the M_JJ > 110 GeV region and used as an alternate background model. The resulting Gaussian fit to the excess (mean = 139 ± 5 GeV/c², width=15± 5 GeV/c²), is in good agreement with the default fit.

g

• Variation with E_T threshold
  The number of excess events is determined as a function of the jet E_T threshold. The number reduces from 262 (E_T=30 GeV) to 37 (E_T=65 GeV). If the E_T threshold is reduced to 20 GeV and the requirement on pT_JJ removed (as is the case in CDF Higgs analyses) then the significance of the excess is reduced and becomes 1.1 sigma (see plots below).

• High pT_JJ sample
  

  The significance of the excess for the subsample of events selected with pT_JJ> 60 GeV is 2.2 σ. The M_JJ distribution is shown below.

• Top Backgrounds

  The plots below show the diboson, single top, and t-tbar backgrounds. Left: normalised to unit area, and Right: normalised to expected cross sections

• "Inclusive" Distributions without # jets=2 cut

  The M_JJ and background subtracted M_JJ distributions are shown below when the njets=2 cut is relaxed i.e. these distributions are with njets ≥ 2.

• Resonant production : distribution of the total mass of the event and Q = M_{l ν+ JJ} - M_JJ -M _W distribution. The p_Z of the neutrino is obtained by constraining the W mass to the PDG value and extracting the two solutions. The plots thus have two entries per event. The distributions are compatible with the background-only hypothesis.