Sequential 4th Family Quarks at ATLAS

1
Sequential 4th Family Quarks at ATLAS
September 21, 2007 UCL ATLAS Physics

• V. E. Özcan
• University College London
• In collaboration with
• G. Unel S. Sultansoy

2
Sequential 4th Family

• SM itself does not make an argument on the number
  of generations
• Why 3 generations then?
• 1973 KM point out 6 flavors in 3 generations
  would accommodate CP violation in the SM
• 1979 studies on abundances of light elements
  start to put constraints on of light neutrinos
• 1989 SLC LEP experiments establish 3 light
  neutrinos (with mass lt mZ/2).
• So we naturally assume that 3 is the number.
• On the discovery of the muon, I. I. Rabi Who
  ordered that?
• Heavier quarks leptons expected in many
  theories t in Little Higgs models, iso-singlet
  iso-triplet fermions in E6 GUT, some models of
  dynamical symmetry breaking, etc.
• For this study, we look for a sequential 4th
  family a full new generation of fermions within
  the SM, much like the first three.

3
Signal

• Search for 4th family quarks as predicted under
  the assumption of Flavor Democracy.
• All flavors have comparable Yukawa couplings to
  start with, so the mass matrix is democratic.
  However this is slightly broken.
• Prediction of this model 4 families with
  quasi-degenerate 4th generation quarks, ie.
  m(u4)-m(d4)few GeV
• No fundamental reason to assume the 4x4 CKM
  follows the same trend of the 3x3 version
• ATLAS TDR 4th family mixes predominantly with
  the 3rd family.
• New study 4th family mixes predominantly with
  1st or 2nd.
• Final state
• pp gt q4q4 gt Wjjj Wlnj ( 2 hard u,d,s,c jets
  2 Ws)

4
Event Generation

• 12k signal events with CompHep for 3 different
  choices of mass.
• (Later dropped 250 GeV due to recent upper limit
  from CDF.)

• A total of 250k BG events generated with
  Madgraph
• WWjj, WZjj, WWbb (tt), WWbbj (ttj)

5
Reconstruction Selection

• Leptonic W from missing ET e/m
• Hadronic W from 3rd 4th highest-PT jets
• Combine W candidates with two hardest jets.
• Do both combinations and choose the min
  DmjWm1q4-m2q4
• All 4 jets used have to be non b-tagged.

6
Example Distributions
The tail due to cases where W has high PT and
ends up being a single jet. gt Analysis can be
improved.
7
Reco. mq4

• Tricky part Doing the fits

8
Fits

• Finding the right fit function is difficult.
• PT cuts on the hard jets effect the lower end of
  the BG mq4 distribution.
• Even for cases which initially looked promising,
  problems were encountered when we went to Toy MC
  studies.
• We want a fit that can run with as minimum human
  interaction as possible.
• Finally settled with
• For signal, a Breit-Wigner 3 parameters
• For BG, a reverted Crystal Ball function (a
  Gaussian core and a power-law tail added together
  so that the function is not only contentious, but
  also smooth.) 5 parameters

9
Results

• S Integrate background function within 2s of
  the signal peak
• B Integrate signal function within 2s of the
  signal peak

• The fit BG functions for the two masses are in
  agreement with each other (within statistics).
  Then, one can use these to generalize results to
  different q4 masses
• Compute the x-sections
• Estimate BG around the new peaks by integrating.
• Estimate cut efficiency by interpolating

10
5s Reach
1 fb-1 mq4 lt 650 GeV 30 fb-1 mq4 lt 850 GeV
11
Conclusion

• You can see the draft paper
• ATL-COM-PHYS-2007-044
• All comments will be highly appreciated!