146
5 Direct Searches for New Physics
Searches for tt resonances in the boosted dilepton channel have been performed
by CMS using 8 [806] and 13 TeV [808] data. In these analyses, the presence of two
neutrinos complicates the reconstruction of the tt system. Studies have shown that S T ,
defined as the scalar sum of p T of jets, leptons and p
miss
T , provides higher sensitivity
in this channel than m tt . The large mass of the resonance leads to a collimated system
from each top quark decay, consisting of a lepton and a b jet. To account for the
overlap between the lepton and the b jet, two-dimensional isolation criteria are used,
based on p
rel
T and R((, j), where the latter is the angular distance between the lepton
and the nearest small-R jet. The distribution in R sum = R(( 1 , j) + R(( 2 , j)
is found to have high discrimination power between signal and background, where
1 and 2 denote the p T -leading and p T -subleading leptons in the event. Lepton-jet
pairs from resonance decays populate the low-R sum region. Events are separated
into signal and control regions, defined by R sum < 2 and R sum > 2, respectively.
The S T distribution in the signal region is used to search for a signal in ee, eμ and μμ
final states. The expected signal distributions are broader compared to distributions
in m tt in the +jets channel, resulting in a lower sensitivity for high mass resonances.
This is exacerbated by the smaller branching fraction in this channel. However, for
resonances with masses of 1 TeV or smaller, this channel achieves similar sensitivity
as the +jets channel.
The best sensitivity is achieved by a combination of analyses in different decay
channels of the tt system. This allows for the most precise determination of the
SM backgrounds, as well as for better constraints on experimental and theoretical
uncertainties. Such combinations have been performed by CMS considering the allhadronic and +jets final states [805, 807], as well as all three tt decay channels
considered in searches [806, 808]. In these combinations, the dilepton and +jets
channels contribute in approximately the same amount to limits for masses below
1 TeV. Above 1.5 TeV, the sensitivity from the all-hadronic and +jets channels is
about the same.
The off-shell component in the production of high mass resonances is strongly
enhanced by the available parton luminosity at lower m tt . This effect becomes more
apparent for wider resonances. So far, LHC searches for tt resonances have focussed
on narrow signals with relative widths /M between 1 and 30%. For masses up to
about 3 TeV, these show pronounced peaks in the m tt distribution, even though signals with /M > 10% have already a large off-shell component with a similar shape
as SM tt production. For masses of 5 TeV and higher, only signals with /M ≈ 1%
show a resonant structure. The searches presented here have only limited sensitivity to
broad, high mass resonances. For example, the ATLAS analysis presented in Fig. 5.8
has remarkable sensitivity to narrow resonances with /M ≈ 1%, but it is blind to
non-resonant enhancements of the tt production cross section at high m tt . In general,
recent searches at 13 TeV are already limited by systematic uncertainties for wide,
high-mass resonances. However, recently these resonances or general enhancements
in the high energy tails of tt production have received considerable theoretical attention. Examples are five-dimensional BSM models with Kaluza-Klein excitations that
result in an increase of the continuum production of tt [818], or low-mass axion-like
5 Direct Searches for New Physics
Searches for tt resonances in the boosted dilepton channel have been performed
by CMS using 8 [806] and 13 TeV [808] data. In these analyses, the presence of two
neutrinos complicates the reconstruction of the tt system. Studies have shown that S T ,
defined as the scalar sum of p T of jets, leptons and p
miss
T , provides higher sensitivity
in this channel than m tt . The large mass of the resonance leads to a collimated system
from each top quark decay, consisting of a lepton and a b jet. To account for the
overlap between the lepton and the b jet, two-dimensional isolation criteria are used,
based on p
rel
T and R((, j), where the latter is the angular distance between the lepton
and the nearest small-R jet. The distribution in R sum = R(( 1 , j) + R(( 2 , j)
is found to have high discrimination power between signal and background, where
1 and 2 denote the p T -leading and p T -subleading leptons in the event. Lepton-jet
pairs from resonance decays populate the low-R sum region. Events are separated
into signal and control regions, defined by R sum < 2 and R sum > 2, respectively.
The S T distribution in the signal region is used to search for a signal in ee, eμ and μμ
final states. The expected signal distributions are broader compared to distributions
in m tt in the +jets channel, resulting in a lower sensitivity for high mass resonances.
This is exacerbated by the smaller branching fraction in this channel. However, for
resonances with masses of 1 TeV or smaller, this channel achieves similar sensitivity
as the +jets channel.
The best sensitivity is achieved by a combination of analyses in different decay
channels of the tt system. This allows for the most precise determination of the
SM backgrounds, as well as for better constraints on experimental and theoretical
uncertainties. Such combinations have been performed by CMS considering the allhadronic and +jets final states [805, 807], as well as all three tt decay channels
considered in searches [806, 808]. In these combinations, the dilepton and +jets
channels contribute in approximately the same amount to limits for masses below
1 TeV. Above 1.5 TeV, the sensitivity from the all-hadronic and +jets channels is
about the same.
The off-shell component in the production of high mass resonances is strongly
enhanced by the available parton luminosity at lower m tt . This effect becomes more
apparent for wider resonances. So far, LHC searches for tt resonances have focussed
on narrow signals with relative widths /M between 1 and 30%. For masses up to
about 3 TeV, these show pronounced peaks in the m tt distribution, even though signals with /M > 10% have already a large off-shell component with a similar shape
as SM tt production. For masses of 5 TeV and higher, only signals with /M ≈ 1%
show a resonant structure. The searches presented here have only limited sensitivity to
broad, high mass resonances. For example, the ATLAS analysis presented in Fig. 5.8
has remarkable sensitivity to narrow resonances with /M ≈ 1%, but it is blind to
non-resonant enhancements of the tt production cross section at high m tt . In general,
recent searches at 13 TeV are already limited by systematic uncertainties for wide,
high-mass resonances. However, recently these resonances or general enhancements
in the high energy tails of tt production have received considerable theoretical attention. Examples are five-dimensional BSM models with Kaluza-Klein excitations that
result in an increase of the continuum production of tt [818], or low-mass axion-like