Event Selections and Detector Parameters

Return to home page.

We are using ISAJET 7.64 for our event generator and PGS 4 for our detector simulator.

Detector parameters used for detector simulation:
  • 320               ! eta cells in calorimeter
  • 200               ! phi cells in calorimeter
  • 0.0314159    ! eta width of calorimeter cells |eta| < 5
  • 0.0314159    ! phi width of calorimeter cells
  • 0.01              ! electromagnetic calorimeter resolution const
  • 0.2                ! electromagnetic calorimeter resolution * sqrt(E)
  • 0.8                ! hadronic calolrimeter resolution * sqrt(E)
  • 0.2                ! MET resolution
  • 0.01              ! calorimeter cell edge crack fraction
  • 5.0                ! calorimeter trigger cluster finding seed threshold (GeV)
  • 1.0                ! calorimeter trigger cluster finding shoulder threshold (GeV)
  • 0.5                ! calorimeter kt cluster finder cone size (delta R)
  • 2.0                ! outer radius of tracker (m)
  • 4.0                ! magnetic field (T)
  • 0.000013      ! sagitta resolution (m)
  • 0.98              ! track finding efficiency
  • 1.00              ! minimum track pt (GeV/c)
  • 3.0                ! tracking eta coverage
  • 3.0                ! e/gamma eta coverage
  • 2.5                ! muon eta coverage
  • 3.0                ! tau eta coverage, Abram has 2.0
  • -1.0               ! AAPedit! E_iso / E_tau(hadronic) maximum (negative value to turn this cut off)
  • 0.174533      ! AAPedit! tau signal cone size (radians)
  • 0.523598      ! AAPedit! tau isolation cone size (radians)

    Basic event selection: (used for generating Mττ , Mjττ , Mjτ , and PTvis)
  • Number of τ's > 2
  • PTτ > 20 GeV
  • |ητ| < 2.5
  • Highest PT τ must have PT > 40 GeV
  • τ identification efficiency = 50%
  • jet → τ fake rate = 1%
  • At least 2 non-b jets satisfying: PT > 100 GeV
  • |ηjet| < 2.5
  • MET > 180 GeV
  • MET + ETjet1 + ETjet2 > 600 GeV

    Event selection for generating Meff:
  • At least 4 non-b jets satisfying: PT > 50 GeV
  • PTjet1 > 100 GeV
  • |ηjet| < 2.5
  • Zero isolated electrons and muons with PT > 15 GeV and |η| < 2.5
  • Sphericity > 0.2
  • MET > 100 GeV
  • MET > 0.2 * Meff

    Event selection for generating Meff(b):
  • At least jets satisfying: PT > 50 GeV
  • PTjet1 > 100 GeV
  • |ηjet| < 2.5
  • Require the highest PT jet to be tagged as a b jet
  • Zero isolated electrons and muons with PT > 15 GeV and |η| < 2.5
  • Sphericity > 0.2
  • MET > 100 GeV
  • MET > 0.2 * Meff
    ) shows that the dominant production mechanism as well as the majority of 3 τ candidate events come from gluinos and squarks. In addition, the number of events produced in each scenario is nearly constant. Since gluinos decay to a quark squark pair, so the dominant events start with squarks and quarks. Squarks can decay in two ways:
        sq -> q C1
        C1 -> stau ν
        stau -> τ N1 60% of the time.
    or
        sq->q N2
        N2->τ stau
        stau->τ N1 40% of the time.
    τs which come from the decay of a stau are have low ET because of the near degneracy between the stau and the N1, but τs which come from the decay of an N2 have high ET because M(N2) ~ 2*M(stau). We choose to look at events with two hard τs and one soft τ because even though the first chain happens more often, soft τs are hard to detect. We must require at least one soft τ because all the information about ΔM is encoded in the decay of the stau to a τ and N1. Therefore our restriction to 3 τs limits the accepted production methods to gluino gluino, squark squark, and gluino squark, which amounts to about 90% of all supersymmetry production. Considering that we want the second decay chain, 14% of all supersymmetry events should have the proper decays. In addition, requiring three taus limits standard model background.

    In the attached files, three representative mSUGRA points are used:

  • A1: μ>0, tan β = 40, A0=0, m0=210, m1/2=360     ΔM = 7.36 GeV
  • A2: μ>0, tan β = 40, A0=0, m0=215, m1/2=360     ΔM = 12.38 GeV
  • A3: μ>0, tan β = 40, A0=0, m0=225, m1/2=360     ΔM = 21.99 GeV

    The quantity NOS-NLS = X is frequently used as an analysis variable in this paper. The statistical error on this variable is not as simple as Sqrt(X) since NOS and NLS are data subsets of the total N. To analyze the error, we use the binomial distribution, approximating pOS = NOS/N and pLS = NLS/N.
    The expressing for the variances are:
    σ2OS = NOS*NLS/N
    σ2LS = NOS*NLS/N

    Propagation of errors gives us:
    σ2stat = σ2OS +σ2LS - 2*σ2OS,LS
    where the last term is the covariance. If we write out the form of the covariance, it is easy to show that:
    σ2OS,LS = -σ2OS
    Therefore;
    σ2stat = 4*NOS*NLS/N

    While the number of events produced nearly constant, we see that the ET spectrum of the τs is well separated.
  • ET Spectrum of 3 Leading τs (Total and Visible) ()

    This plots shows that cuts on the leading jet and on missing ET can remove ttbar background.
  • ET Spectrum of Leading Jet and Missing ET Spectrum ()

    Plotting missing ET against the leading jet shows that a cut on the missing E_T + leading jet ET can separate ttbar background from SUSY signal effectivly.
  • Missing ET vs. Leading Jet ET ()

    After the cuts: 2 τ with ET > 40 GeV, 1 τ with ET > 20 GeV, and missing ET + leading jet ET > 400 GeV, we find that the number of events expected grows linearly with ΔM in the region ΔM > 5 GeV.
  • Graph of Events vs. ΔM ()

    Using the linear fit above, we find the number of expected events as a linear funtion of ΔM. The uncertainty in ΔM (neglecting systematic error) grows as a square root of this function in the region ΔM > 5 GeV.
  • Uncertainty in ΔM vs. ΔM (12 fb-1) ()

    The percent uncertainty decreases as luminosity-1/2
  • Percent Uncertainty vs. Luminosity ()
    -->
    (last updated on )