148
5 Direct Searches for New Physics
an order of magnitude smaller compared to the ATLAS analysis. Consequently, the
obtained upper cross section limits are better by a factor of two. A combination with
the +jets channel results in an exclusion of a W
boson with right-handed couplings
below 2.15 TeV.
For analyses in the +jets channels at 13 TeV, ATLAS and CMS have changed the
lepton selections to non-isolated leptons. Both experiments have performed analyses
of 36 fb
−1 of data, where ATLAS uses mini-isolation for the lepton selection [826]
and CMS uses the two-dimensional R, p
rel
T selection [827]. The CMS analysis
is optimised for high mass signals, with higher p T thresholds on leptons and jets
compared to the ATLAS analysis. In turn, ATLAS achieves sensitivity for W
masses
starting at 0.5 TeV, whereas the CMS analysis starts at 1 TeV. Besides this, the two
analyses strategies are similar, and both analyses estimate the SM backgrounds from
simulation. The CMS analysis achieves higher sensitivity at high masses, because
of a categorisation of the signal regions based on the p T of the four-vector sum of
the two p T -leading small-R jets. The achieved sensitivities are comparable, where
ATLAS achieves better sensitivity for masses below 2 TeV, and CMS above this
value. The mass exclusion limits improve by nearly 1.5 TeV compared to the 8 TeV
analyses, to 3.6 TeV for a right-handed W
.
A search in the all-hadronic channel using 36.1 fb
−1 of 13 TeV data has been
performed by ATLAS [521]. The tb final state is reconstructed using one small-R
b-tagged jet with p T > 420 GeV, which is required to have a large angular separation to the t-tagged jet, also with p T > 420 GeV. The shower deconstruction (SD)
algorithm [257, 258] is used to identify t jets. The input to the SD algorithm are
subjets, which are obtained by reclustering the selected large-R jet with the k T algorithm. The clustering is stopped once the splitting scale is larger than 15 GeV, at
which point the remaining protojets are used as subjets. The six p T -leading subjets
are used as input for the SD tagger. Large-R jets are only considered further if they
have at least three subjets and two or more subjets have a combined invariant mass
between 60.3 and 100.3 GeV, and by adding one or more subjets a total jet mass
between 132 and 212 GeV is obtained. These jets are passed to the SD algorithm,
and its output is used to define two working points. The loose working point has a t
jet efficiency of 80% with a misidentification rate of 4–10%; the tight working point
has an efficiency of 50% with a misidentification rate of 1.3–3.3% for p T between
0.45 and 1.3 TeV, respectively. The lose and tight working points, together with the
requirement of the small-R jet being b-tagged or not, define six regions. Depending
on the presence of a b-tagged small-R jet, overlapping with the t-tagged jet, six more
regions can be defined. Out of these twelve regions, the two regions with the tight
t-tagging requirement and one b-tagged jet opposite to the t jet, define the signal
regions. A third signal region is obtained from loose t-tagged jets, but two b-tagged
jets in the event. The multijet background is estimated from the other eight regions,
where one region is used to validate the procedure. The multijet background makes
up more than 90% of the background in two signal regions, and 75% in the third,
with the remainder being tt. The measured m tb distributions are described by the
estimated background within the uncertainties over the full accessible range between
1 and 6 TeV. The sensitivity achieved by this analysis is nearly identical to the sensi-
5 Direct Searches for New Physics
an order of magnitude smaller compared to the ATLAS analysis. Consequently, the
obtained upper cross section limits are better by a factor of two. A combination with
the +jets channel results in an exclusion of a W
boson with right-handed couplings
below 2.15 TeV.
For analyses in the +jets channels at 13 TeV, ATLAS and CMS have changed the
lepton selections to non-isolated leptons. Both experiments have performed analyses
of 36 fb
−1 of data, where ATLAS uses mini-isolation for the lepton selection [826]
and CMS uses the two-dimensional R, p
rel
T selection [827]. The CMS analysis
is optimised for high mass signals, with higher p T thresholds on leptons and jets
compared to the ATLAS analysis. In turn, ATLAS achieves sensitivity for W
masses
starting at 0.5 TeV, whereas the CMS analysis starts at 1 TeV. Besides this, the two
analyses strategies are similar, and both analyses estimate the SM backgrounds from
simulation. The CMS analysis achieves higher sensitivity at high masses, because
of a categorisation of the signal regions based on the p T of the four-vector sum of
the two p T -leading small-R jets. The achieved sensitivities are comparable, where
ATLAS achieves better sensitivity for masses below 2 TeV, and CMS above this
value. The mass exclusion limits improve by nearly 1.5 TeV compared to the 8 TeV
analyses, to 3.6 TeV for a right-handed W
.
A search in the all-hadronic channel using 36.1 fb
−1 of 13 TeV data has been
performed by ATLAS [521]. The tb final state is reconstructed using one small-R
b-tagged jet with p T > 420 GeV, which is required to have a large angular separation to the t-tagged jet, also with p T > 420 GeV. The shower deconstruction (SD)
algorithm [257, 258] is used to identify t jets. The input to the SD algorithm are
subjets, which are obtained by reclustering the selected large-R jet with the k T algorithm. The clustering is stopped once the splitting scale is larger than 15 GeV, at
which point the remaining protojets are used as subjets. The six p T -leading subjets
are used as input for the SD tagger. Large-R jets are only considered further if they
have at least three subjets and two or more subjets have a combined invariant mass
between 60.3 and 100.3 GeV, and by adding one or more subjets a total jet mass
between 132 and 212 GeV is obtained. These jets are passed to the SD algorithm,
and its output is used to define two working points. The loose working point has a t
jet efficiency of 80% with a misidentification rate of 4–10%; the tight working point
has an efficiency of 50% with a misidentification rate of 1.3–3.3% for p T between
0.45 and 1.3 TeV, respectively. The lose and tight working points, together with the
requirement of the small-R jet being b-tagged or not, define six regions. Depending
on the presence of a b-tagged small-R jet, overlapping with the t-tagged jet, six more
regions can be defined. Out of these twelve regions, the two regions with the tight
t-tagging requirement and one b-tagged jet opposite to the t jet, define the signal
regions. A third signal region is obtained from loose t-tagged jets, but two b-tagged
jets in the event. The multijet background is estimated from the other eight regions,
where one region is used to validate the procedure. The multijet background makes
up more than 90% of the background in two signal regions, and 75% in the third,
with the remainder being tt. The measured m tb distributions are described by the
estimated background within the uncertainties over the full accessible range between
1 and 6 TeV. The sensitivity achieved by this analysis is nearly identical to the sensi-