|
|
Mtt FlaME Resonance Search (319 pb-1) |
|
Univ. of Florida: Valentin Necula, Jacobo Konigsberg, Roberto Rossin
We performed a blind search for non-SM resonant production of
top anti-top pair in the lepton+jets channel. Each event is analyzed assuming
SM ttbar production using a likelihood that incorporates the differential production
cross section and the probabilty to produce the observed jet momenta and event Pt.
The ttbar invariant mass (Mtt) is reconstructed at each step (for all parton level
momenta considered and for all final state combinations) and the mean Mtt (over
all steps, each weighted by its likelihood) is chosen
as *the Mtt* for that event.
The resulting Mtt distribution is then tested for possible resonant production (Xo)
using a Bayesian approach that yields the posterior probability that the data contains
resonant events produced with a given mass (MXo) and a given cross section (sigma).
We compare the 95% C.L. points for these posterior probabilities for the data with those expected
from SM sources only and determine how likely it is for the SM to fluctuate to the data for
each mass point. A very unlikely fluctuation could then indicate the presence of New Physics
in the sample.
The posterior for the data also yields the most probable value for the cross section
for each MXo tested. We then compare [KS test] the data to the SM Mtt shape (null hypothesis) and
to the Mtt shape containing the most likely cross section for a resonance for the most likely MXo.
A very unlikely SM shape and a very likely SM+resonance shape will indicate the possible
presence of New Physics in the sample.
We tested for vector resonances with masses between 450 and 900 GeV and widths equal
to 1.2% of their mass. We selected events with 4 or more jets and assumed a top mass = 175 GeV. The dataset corresponds to an integrated luminosity of 319 pb-1.
The data shows an intriguing peak at 500 GeV. This is also the mass value most inconsistent
with SM fluctuations (a 2-sigma effect). If such resonance existed, the most probable
value for the cross section times branching-ratio is 2 pb. The KS probability of the data
coming from just SM is 15.2 % and for the data to come from SM and 2 pb from a 500 GeV resonance
is 70.2 %.
Expected 95% C.L. upper limit on the cross section for resonance production assuming SM only.
Mtt distribution from data (left) and vs expected SM background
in the >400 GeV search region (right)
95% CL upper limits for the cross section for resonance production: data vs SM expectations (left) and posterior from data for a 500 GeV resonance contribution (right)
KS tests: Data vs SM [left] and Data vs SM+Xo(500) [right]
95% C.L. upper limits on the cross section for resonance production: data vs SM expectation vs Leptophobic Topcolor model