5.4 Excited Third Generation Quarks
177
The first search for the resonant production of t ¯
t+2 jets at the LHC has been
carried out by CMS using 19.5 fb
−1 of 8 TeV data [982]. The analysis is carried
out in the +jets final state, and has been optimised for t
∗ masses below 1 TeV
with resolved t decays. Events are selected with one isolated lepton and at least
six jets with p T > 30 GeV, one of which has to be b tagged. The reconstructed t
∗
distribution is used as sensitive observable. It is obtained by all possible permutations
of assignments of the six p T -leading jets to the expected t
∗ decay partons. A kinematic
fit is performed in order to improve the resolution in the reconstructed t
∗ mass. The
fit has seven kinematic constraints from the W , t and t
∗ masses, and the conservation
of p T in the collision. The parameter estimation for this complex system is difficult
and computationally expensive, such that a sampling of the allowed parameter space
is performed before the actual numerical minimisation. However, only in about 11%
of the simulated signal events a correct assignment is found. In most of the events
with a wrong assignment, one of the actual jets from the t
∗ decay has failed the jet
p T requirement. However, it is found that lowering the jet p T threshold increases
the background without a gain in sensitivity. Despite this shortcoming, the t
∗ mass
reconstruction results in a peaking distribution at the correct value for signal events,
while the expected background is a smoothly falling distribution. The background is
estimated by a parametric function with three free parameters, fit to the distribution
in the reconstructed b
∗ mass. Lower mass limits of 800 GeV for a t
∗ with spin-3/2
are reported. Spin-1/2 particles can only be excluded between 470 and 510 GeV,
but their existence at lower masses can be excluded by differential tt cross section
measurements [627, 983] and tt production with additional jet activity [984, 985].
The CMS search for t
∗ ¯
t
∗ in the +jets final state has been updated using 35.9 fb
−1
of 13 TeV data [986]. Small adjustments to the analysis have been made, for example
two b-tagged jets instead of one are required in order to suppress contributions from
non-tt backgrounds, and to reduce the complexity of finding the correct jet-parton
assignment. In addition, this assignment is now based on an estimator using the
sum of squares of differences between measured and expected W , t and t
∗ masses,
without performing a kinematic fit. While this results in a worse t
∗ mass resolution
for correct assignments, the overall sensitivity is nearly unchanged since only in
11% of the events all six jets are correctly assigned. No excess over the parametrised
background is observed, resulting in lower mass limits of 1.2 TeV for a t
∗ with spin3/2.
In future searches for t+g resonances, a significant improvement over previous
results can only be achieved by the use of jet substructure information for the t
identification. This will enable to probe higher masses through an increase in signal
efficiency at high top quark p T , and it will reduce the complexity of the t
∗ mass
reconstruction. In addition, q/g discrimination applied on the two gluon jets could
help to improve the signal-to-background ratio. With larger datasets it will become
interesting to include the t
∗
→ tγ decay, resulting in tt gγ final states. While this
decay only has a branching fraction of about 3%, the presence of a photon can help
to reduce the SM backgrounds considerably [969]. Last but not least, the single
production channels t
∗ t and ¯
t
∗ t should be included [968] and t
∗ searches in the
all-hadronic and dilepton final states should be envisioned.
177
The first search for the resonant production of t ¯
t+2 jets at the LHC has been
carried out by CMS using 19.5 fb
−1 of 8 TeV data [982]. The analysis is carried
out in the +jets final state, and has been optimised for t
∗ masses below 1 TeV
with resolved t decays. Events are selected with one isolated lepton and at least
six jets with p T > 30 GeV, one of which has to be b tagged. The reconstructed t
∗
distribution is used as sensitive observable. It is obtained by all possible permutations
of assignments of the six p T -leading jets to the expected t
∗ decay partons. A kinematic
fit is performed in order to improve the resolution in the reconstructed t
∗ mass. The
fit has seven kinematic constraints from the W , t and t
∗ masses, and the conservation
of p T in the collision. The parameter estimation for this complex system is difficult
and computationally expensive, such that a sampling of the allowed parameter space
is performed before the actual numerical minimisation. However, only in about 11%
of the simulated signal events a correct assignment is found. In most of the events
with a wrong assignment, one of the actual jets from the t
∗ decay has failed the jet
p T requirement. However, it is found that lowering the jet p T threshold increases
the background without a gain in sensitivity. Despite this shortcoming, the t
∗ mass
reconstruction results in a peaking distribution at the correct value for signal events,
while the expected background is a smoothly falling distribution. The background is
estimated by a parametric function with three free parameters, fit to the distribution
in the reconstructed b
∗ mass. Lower mass limits of 800 GeV for a t
∗ with spin-3/2
are reported. Spin-1/2 particles can only be excluded between 470 and 510 GeV,
but their existence at lower masses can be excluded by differential tt cross section
measurements [627, 983] and tt production with additional jet activity [984, 985].
The CMS search for t
∗ ¯
t
∗ in the +jets final state has been updated using 35.9 fb
−1
of 13 TeV data [986]. Small adjustments to the analysis have been made, for example
two b-tagged jets instead of one are required in order to suppress contributions from
non-tt backgrounds, and to reduce the complexity of finding the correct jet-parton
assignment. In addition, this assignment is now based on an estimator using the
sum of squares of differences between measured and expected W , t and t
∗ masses,
without performing a kinematic fit. While this results in a worse t
∗ mass resolution
for correct assignments, the overall sensitivity is nearly unchanged since only in
11% of the events all six jets are correctly assigned. No excess over the parametrised
background is observed, resulting in lower mass limits of 1.2 TeV for a t
∗ with spin3/2.
In future searches for t+g resonances, a significant improvement over previous
results can only be achieved by the use of jet substructure information for the t
identification. This will enable to probe higher masses through an increase in signal
efficiency at high top quark p T , and it will reduce the complexity of the t
∗ mass
reconstruction. In addition, q/g discrimination applied on the two gluon jets could
help to improve the signal-to-background ratio. With larger datasets it will become
interesting to include the t
∗
→ tγ decay, resulting in tt gγ final states. While this
decay only has a branching fraction of about 3%, the presence of a photon can help
to reduce the SM backgrounds considerably [969]. Last but not least, the single
production channels t
∗ t and ¯
t
∗ t should be included [968] and t
∗ searches in the
all-hadronic and dilepton final states should be envisioned.