First look at limit setting on GMSB models in the final state with 2 fb-1

1
First look at limit setting on GMSB models in the
final state with 2 fb-1

• R. Culbertson, M. Goncharov, Eunsin Lee,
• A. Pronko, D. Toback
• Fermilab, Texas AM University

Search for Heavy, Long-Lived Neutralinos
that decay via
Eunsin Lee
2
Outline

• Analysis Introduction
• Progress
• GMSB Signal Acceptance
• 1-DIM Optimization with MetSig
• Expected Cross Section Limits
• Conclusion and Plan

3
Dominant Signal Process (GMSB)

• Looking for with low
  lifetime
• - Both Neutralinos decay in the detector --gt Two
  Photons
• (Optimal for low
  lifetimes)

R. Culbertson, D. Kim, M. Kim, S. Lee, and
D.Toback Search for Anomalous Production of
di-photon MET at CDF and Limits on GMSB Models
(CDF/PRD) Phys.Rev.D71, 031104 (2005),
hep-ex/0410053
D0 Collaboration Search for supersymmetry in
di-photon final states at sqrt(s)1.96 TeV
(D0/PLB) Phys.Lett.B659, 856 (2008),
hep-ex/0710.3946
4
Exclusion Region were trying to cover
Delayed Photon Analysis M.Goncharov, V.Krutelyov,
E.Lee, D.Toback and P.Wagner Phys. Rev. Lett 99,
121801 (2007)
Were trying to cover here
5
Background Sources
Collision (SM) Background
Non-Collision Background Cosmic and Beam Halo
Triggers, Data Sets and Final Objects with
Preselection requirements
Search for Heavy, Long-Lived Neutralinos
that decay via
Eunsin Lee
6
Progress

• - Sashas MET Resolution Model (Search for
  Anomalous Production of Events in 2
  fb-1 of Data) blessed (02/21/08)
• CDF NOTE 9184
• - Produced GMSB signal (6000 events) Acceptance
  
• Example point (
  )
• - 1-Dim Optimization (MetSig cut) to set limit

Search for Heavy, Long-Lived Neutralinos
that decay via
Eunsin Lee
7

• Background Sources
• - QCD events(
  ) with fake
  due to energy mismeasurement
• - EWK process with real
• (mostly
  ,
• W(e)pho 44, W(e)jet 17, W(tau)pho
  14, W(tau)jet 11 )
• - Non-Collision Backgrounds (PMT spikes,
  cosmic rays, beam-halo interaction)
• To reduce these backgrounds, we require
  events to pass the global event selection, photon
  ID , non-collision background rejection, and
  cleanup cuts.
• CDF note 9184 (R. Culbertson, A. Pronko, M.
  Goncharov)

Search for Heavy, Long-Lived Neutralinos
that decay via
Eunsin Lee
8
Data Sample and Event Quality

• - Ntuples and Good Runs
• Data Stntuples cdfpstn cdipa(d,h,i,j) ,
  cdfpstn bhelb(d,h,i,j)
• Goodrun list Up to and including period 11
• V17 with good CAL,
  CES, COT (allow COT
• degraded period but
  require Silicon)
• Code Release cdfsoft 6.1.4, Stntuple dev_243
• - Data Sample and Event quality
• Diphoton trigger
• Nvx12 ? 1
• Highest ?PT vertex
• For any vertex Zvxlt60 cm
• Any 2 vertices DZvxlt90 cm
• Vertex re-assignment to minimize MET if Nvxgt1
• 2 leading photons pass tight ID cuts

9
Central Photon ID cuts
cuts Tight cuts Loose cuts
Calorimeter fiduciality central central
Photon ET gt13 GeV (7 GeV for pre-selection) gt13 GeV (7 GeV for pre-selection)
CES fiduciality XCESlt21.0 cm 9.0 cmltZCESlt230.0 cm XCESlt21.0 cm 9.0 cmltZCESlt230.0 cm
Average CES ?2 lt20 lt20
Had/Em lt0.0550.00045ET lt0.125
Corrected CalISO lt2.00.02(ET-20) or lt0.1ET if ETlt20.0 GeV lt3.00.02(ET-20) or lt0.15ET if ETlt20.0 GeV
TrkISO lt2.00.005ET lt5.0
N3D N3D0,1 N3D0,1
Trk PT (if N3D1) lt1.00.005ET lt0.25ET
2nd CES cluster (wire or strip) lt0.14ET if ETlt18 GeV lt2.40.01ET if ETgt18 GeV No cut
Phoenix rejection No photons matched to phoenix track No photons matched to phoenix track
10
GMSB MC Simulation

• Use Pythia to generate the GMSB signal and
  cdfSim(Gen6) to simulate the detector
• The EMTiming system is simulated (CDF note 7982)
• - http//hepr8.physics.tamu.edu/elee/EmtimeSimul
  .html
• Generate Neutralino mass100 GeV and lifetime0
  ns with the following parameters fixed on the
  minimal GMSB Snowmass Slope SPS 8 with a
  neutralino NLSP

Search for Heavy, Long-Lived Neutralinos
that decay via
Eunsin Lee
11
Event Reduction for data and GMSB signal
Cuts Data GMSB MC signal (6000 events) Efficiency (Cumulative) Efficiency (Relative)
Pre-selection (trigger, good run, 2 EM objects) 54,802 5815 96.9 96.9
( photon ID, gt7 GeV ) 54,802 1017 17.0 17.5
PMT spikes 54,796 N/A N/A N/A
Vertex 51,440 969 16.2 95.3
Vertex swap gt13 GeV 35,673 945 15.8 97.5
Phoenix 32,899 921 15.4 97.5
Beam Halo 32,890 921 15.4 100
EMTiming 32,865 921 15.4 100
Trackless muon stub 32,720 916 15.3 99.5
MetSig gt 5.0 34 763 12.7 83.3
? Signal Efficiencies in Pre-selection under Study
Estimated from data
Search for Heavy, Long-Lived Neutralinos
that decay via
Eunsin Lee
12
Met and MetSig Distributions for Data and GMSB MC
Signal
DATA
GMSBMC signal (Not scaled yet)
Under Study
Internal integration limit on 1-dim integrals.
Not a big deal. Can be removed by increasing an
integration limit.
Conventional MetSig MET/?ET Sashas MetSig
log10(1-P(MET)) More info on Sashas MET model
Sashas Blessing Talk on Exotic
Meeting(02/21/08)
13
MetSig and MET Distributions for Data and GMSB
MC signal
14
95 C.L. Expected Corss Section Limit for MetSig
cut
MetSig Acceptance Expected Bkg Observed x_sec(exp) (fb) x_sec(obs) (fb) x_sec(prod) (fb)
3.0 0.131 0.013 106.6 14.5 120 131 173 162
4.0 0.129 0.013 63.6 8.9 52 93 60 162
5.0 0.127 0.013 48.6 7.5 34 81 42 162
6.0 0.125 0.012 38.0 8.0 25 79 39 162
8.0 0.119 0.012 23.5 5.5 13 66 28 162
9.0 0.106 0.011 18.0 4.5 12 63 39 162
10.0 0.090 0.009 16.0 4.0 11 68 46 162
12.0 0.067 0.007 8.1 2.8 6 69 54 162

• Acceptance of events passed the cut / total
  of events(6000)
• Acceptance Error 10
• Luminosity 2 fb(-1)
• Luminosity error 6
• x_sec(prod) NLO production cross section
  (including k-factor)

15
MetSig and x_sec limit plots

• MetSig cut kills most of QCD(Fake MET)
  background
• Most sensitive at MetSig gt 9.0

16
Conclusion and Future Plan

• - Considering a number of cuts for
  multidimensional optimization
• Et(pho1), Ht, M(di-pho), MET, Njet25, JetEt,
  IsoTrack,
• EWK rejection cuts, etc.
• - Produce GMSB signal Acceptance for higher
  mass points
• (110 GeV - 140 GeV ) with more events
  (120,000)
• - Finishing up reproducing delayed photon PRL
  acceptance
• - Finished up a preliminary gMettrack
  optimization(P. Geffert) but
• its clear we need to get rid of more QCD
  (DelPhi(met,jet)) cut
• in previous analysis)
• - Plan to add MetSig of Sashas Met Model to
  new analysis

Search for Heavy, Long-Lived Neutralinos
that decay via
Eunsin Lee