5.4 Excited Third Generation Quarks
175
signal and background contributions are fit to the data in the signal and control regions
simultaneously. The combination of the +jets and dilepton final states results in a
lower mass limit of 1.5 TeV for vector-like b
∗ couplings.
An analysis by CMS using 8 TeV data combines the all-hadronic, +jets and
dilepton channels in the search for b
∗
→ W t [980]. While the +jets and dilepton
channels target resolved topologies with exactly three and exactly one small-R jet,
respectively, the all-hadronic selection is optimised for fully merged final states.
This selection is based on two large-R jets with p T > 425 GeV, where one is W
and the other one is t-tagged. The W tagging uses the pruned jet mass and τ 21 ;
t tagging is achieved with the CMSTT, τ 32 and a subjet b tag.
The dominant
backgrounds in the leptonic final states are estimated from simulation and validated
in control regions. In the all-hadronic final state, the dominant multijet background
is obtained by weighting events prior to the t tagging selection with the t-tagging
misidentification rate. This misidentification rate is calculated as a function of jet
p T and η in a sample dominated by multijet events, obtained by inverting the W
selection criteria. Because the t tagging algorithm has a misidentification rate of about
0.1% [243], the background obtained from the weighted pre-t-tagged sample can be
considered statistically independent from the sample after the t tagging selection
has been applied. The sensitivity to signals for b
∗ masses below 1 TeV is driven
by the +jets and dilepton channels, but for higher masses the all-hadronic analysis
provides better sensitivity. Similar to the 8 TeV analysis by ATLAS, lower mass limits
of 1.5 TeV for vector-like b
∗ couplings are obtained.
A very recent analysis by CMS uses the full 13 TeV data with 137 fb
−1 to search
for b
∗
→ W t in the all-hadronic final state, targeting b
∗ masses >1.2 TeV [981].
The analysis uses two large-R jets with p T > 400 GeV and an angular separation
in azimuth greater than π/2 to ensure a back-to-back topology of the two jets. The
signal region is defined through a W - and a t-tagged jet. Both taggers use soft drop jet
mass, where t jets in the signal region are required to have 105 < m jet < 220 GeV
and W jets 65 < m jet < 105 GeV. In addition, τ 32 and subjet b tagging is used for the
definition of t jets, and τ 21 for W jets. The analysis is performed in two dimensions,
where the distribution in the plane (m t , m W t ) is analysed for a potential signal. Note
that a region in m t between 65 and 285 GeV is analysed, such that the signal region
with 105 < m jet < 220 GeV is enclosed by a low- and a high-mass sideband. In this
analysis, m t is the soft drop mass of the t jet and m W t is the mass of the dijet system.
This allows for the use of a novel method to construct the multijet background
template, which relies on a parametrisation of the pass-fail ratio as a function of m t .
The number of multijet events passing the t tagging requirement, n p , in a given
interval in the (m t , m W t ) plane is calculated as n p = n F · f (m t , m W t ), where n F is
the number of events failing the t tagging requirement and f (m t , m W t ) is the twodimensional pass-fail ratio. This ratio is obtained from data, with an initial estimate
obtained from simulation in order to reduce the complexity of the function. It is found
that a a surface parametrised by the product of a second-order polynomial in m t and a
first-order polynomial in m W t is sufficient to describe the data in the sideband regions.
The advantage of this method is that it interpolates the pass-fail ratio into the signal
region from the enclosing sidebands, such that the analysis can be fully tested and
175
signal and background contributions are fit to the data in the signal and control regions
simultaneously. The combination of the +jets and dilepton final states results in a
lower mass limit of 1.5 TeV for vector-like b
∗ couplings.
An analysis by CMS using 8 TeV data combines the all-hadronic, +jets and
dilepton channels in the search for b
∗
→ W t [980]. While the +jets and dilepton
channels target resolved topologies with exactly three and exactly one small-R jet,
respectively, the all-hadronic selection is optimised for fully merged final states.
This selection is based on two large-R jets with p T > 425 GeV, where one is W
and the other one is t-tagged. The W tagging uses the pruned jet mass and τ 21 ;
t tagging is achieved with the CMSTT, τ 32 and a subjet b tag.
The dominant
backgrounds in the leptonic final states are estimated from simulation and validated
in control regions. In the all-hadronic final state, the dominant multijet background
is obtained by weighting events prior to the t tagging selection with the t-tagging
misidentification rate. This misidentification rate is calculated as a function of jet
p T and η in a sample dominated by multijet events, obtained by inverting the W
selection criteria. Because the t tagging algorithm has a misidentification rate of about
0.1% [243], the background obtained from the weighted pre-t-tagged sample can be
considered statistically independent from the sample after the t tagging selection
has been applied. The sensitivity to signals for b
∗ masses below 1 TeV is driven
by the +jets and dilepton channels, but for higher masses the all-hadronic analysis
provides better sensitivity. Similar to the 8 TeV analysis by ATLAS, lower mass limits
of 1.5 TeV for vector-like b
∗ couplings are obtained.
A very recent analysis by CMS uses the full 13 TeV data with 137 fb
−1 to search
for b
∗
→ W t in the all-hadronic final state, targeting b
∗ masses >1.2 TeV [981].
The analysis uses two large-R jets with p T > 400 GeV and an angular separation
in azimuth greater than π/2 to ensure a back-to-back topology of the two jets. The
signal region is defined through a W - and a t-tagged jet. Both taggers use soft drop jet
mass, where t jets in the signal region are required to have 105 < m jet < 220 GeV
and W jets 65 < m jet < 105 GeV. In addition, τ 32 and subjet b tagging is used for the
definition of t jets, and τ 21 for W jets. The analysis is performed in two dimensions,
where the distribution in the plane (m t , m W t ) is analysed for a potential signal. Note
that a region in m t between 65 and 285 GeV is analysed, such that the signal region
with 105 < m jet < 220 GeV is enclosed by a low- and a high-mass sideband. In this
analysis, m t is the soft drop mass of the t jet and m W t is the mass of the dijet system.
This allows for the use of a novel method to construct the multijet background
template, which relies on a parametrisation of the pass-fail ratio as a function of m t .
The number of multijet events passing the t tagging requirement, n p , in a given
interval in the (m t , m W t ) plane is calculated as n p = n F · f (m t , m W t ), where n F is
the number of events failing the t tagging requirement and f (m t , m W t ) is the twodimensional pass-fail ratio. This ratio is obtained from data, with an initial estimate
obtained from simulation in order to reduce the complexity of the function. It is found
that a a surface parametrised by the product of a second-order polynomial in m t and a
first-order polynomial in m W t is sufficient to describe the data in the sideband regions.
The advantage of this method is that it interpolates the pass-fail ratio into the signal
region from the enclosing sidebands, such that the analysis can be fully tested and