162
5 Direct Searches for New Physics
analysis. The resolved search is optimised for m VLQ < 1 TeV and requires six smallR jets within |η| < 4.5, three of which have to be b-tagged. In this kinematic regime,
the difficulty lies in the suppression of the abundant QCD multijet background and a
method for its estimation, able to model the trigger turn-on. These challenges are met
by a cascade of selection steps and a background extrapolation from regions with
loosened b tagging requirements. The boosted analysis selects a dijet topology with
two high- p T large-R jets, where one is t-tagged and the other either Z - or H -tagged.
The Z and H taggers use τ 21 < 0.6 and two b-tagged subjets. Jets have to have
a pruned mass in the range 65–105 GeV to be Z tagged and 105–135 GeV to be
H tagged. Similar requirements are imposed on t-tagged jets, with τ 32 < 0.57, one
b-tagged subjet and a soft drop mass in the range 105–220 GeV. The overlap in the
Z /H and t taggers results in an ambiguity when selecting jets, but increases the signal
efficiency in this analysis where a confusion between the Z /H and t tagged jets does
not matter. The main backgrounds are tt and multijet production, where the latter
is estimated from control regions, obtained by reversing the tagging requirements.
Upper cross section limits are reported for B(T → Ht) = 1, B(T → Zt) = 1 and
B(T → Ht) = B(T → Zt) = 0.5, the two different production modes with either
an associated b or t, and different fractional widths. The limits are based on whichever
of the two selections, resolved or boosted, achieves the best estimated expected
sensitivity. In general, for m VLQ 1 TeV the resolved analysis is better and for masses
above 1 TeV the boosted analysis is better. The limits on the product of T production
cross section and B(T → Ht) are between 0.028 and 0.12 pb for T masses of 1.8 TeV
and 1 TeV, respectively, thus improving the previous limits obtained in the +jets final
state by nearly an order of magnitude. Comparable sensitivity is obtained for T → Zt
and a mixture of T → Ht and T → Zt.
The diversity of Z boson decays allows for searches for T → Zt also in leptonic
final states, including charged leptons and neutrinos. Neutrinos from the Z → νν
decay result in large p
miss
T , balanced by a boosted top quark. This particular topology,
which does not have a counterpart in the SM, has been searched for by ATLAS using
36.1 fb
−1 of 13 TeV data [908]. Large-R jets are identified as t jets based on the
trimmed jet mass and τ 32 . A loose working point is chosen with a t-tagging efficiency
of 80% above p T of 400 GeV, for high signal efficiency. Contrary to the all-hadronic
analysis, tighter requirements on the t tagging are not needed because the background
from SM multijet production can be reduced efficiently by a selection based on p
miss
T .
This is achieved by p
miss
T
> 200 GeV, and requiring the azimuthal angle between
p
miss
T
and the t-tagged jet to be large. In order to ensure well-reconstructed p
miss
T ,
the p T of the t-tagged jet and p
miss
T
have to be balanced, and no small-R jets should
be present close in azimuthal angle to p
miss
T . This selection reduces the multijet
background to a negligible level of <1% of the total background in the signal region,
quite remarkable for an all-hadronic analysis. In addition to the presence of one ttagged large-R jet with p T > 250 GeV, one b-tagged small-R track-jet is required,
which reduces the background from V +jets. The obtained limits on the product
of production cross section and B(T → Zt) are between 0.045 and 0.15 pb for
m VLQ = 1.8 and 1 TeV, respectively, comparable to the sensitivity of the all-hadronic
search by CMS [907]. Searches in the dilepton final state, targeting T → Z ((()t,
5 Direct Searches for New Physics
analysis. The resolved search is optimised for m VLQ < 1 TeV and requires six smallR jets within |η| < 4.5, three of which have to be b-tagged. In this kinematic regime,
the difficulty lies in the suppression of the abundant QCD multijet background and a
method for its estimation, able to model the trigger turn-on. These challenges are met
by a cascade of selection steps and a background extrapolation from regions with
loosened b tagging requirements. The boosted analysis selects a dijet topology with
two high- p T large-R jets, where one is t-tagged and the other either Z - or H -tagged.
The Z and H taggers use τ 21 < 0.6 and two b-tagged subjets. Jets have to have
a pruned mass in the range 65–105 GeV to be Z tagged and 105–135 GeV to be
H tagged. Similar requirements are imposed on t-tagged jets, with τ 32 < 0.57, one
b-tagged subjet and a soft drop mass in the range 105–220 GeV. The overlap in the
Z /H and t taggers results in an ambiguity when selecting jets, but increases the signal
efficiency in this analysis where a confusion between the Z /H and t tagged jets does
not matter. The main backgrounds are tt and multijet production, where the latter
is estimated from control regions, obtained by reversing the tagging requirements.
Upper cross section limits are reported for B(T → Ht) = 1, B(T → Zt) = 1 and
B(T → Ht) = B(T → Zt) = 0.5, the two different production modes with either
an associated b or t, and different fractional widths. The limits are based on whichever
of the two selections, resolved or boosted, achieves the best estimated expected
sensitivity. In general, for m VLQ 1 TeV the resolved analysis is better and for masses
above 1 TeV the boosted analysis is better. The limits on the product of T production
cross section and B(T → Ht) are between 0.028 and 0.12 pb for T masses of 1.8 TeV
and 1 TeV, respectively, thus improving the previous limits obtained in the +jets final
state by nearly an order of magnitude. Comparable sensitivity is obtained for T → Zt
and a mixture of T → Ht and T → Zt.
The diversity of Z boson decays allows for searches for T → Zt also in leptonic
final states, including charged leptons and neutrinos. Neutrinos from the Z → νν
decay result in large p
miss
T , balanced by a boosted top quark. This particular topology,
which does not have a counterpart in the SM, has been searched for by ATLAS using
36.1 fb
−1 of 13 TeV data [908]. Large-R jets are identified as t jets based on the
trimmed jet mass and τ 32 . A loose working point is chosen with a t-tagging efficiency
of 80% above p T of 400 GeV, for high signal efficiency. Contrary to the all-hadronic
analysis, tighter requirements on the t tagging are not needed because the background
from SM multijet production can be reduced efficiently by a selection based on p
miss
T .
This is achieved by p
miss
T
> 200 GeV, and requiring the azimuthal angle between
p
miss
T
and the t-tagged jet to be large. In order to ensure well-reconstructed p
miss
T ,
the p T of the t-tagged jet and p
miss
T
have to be balanced, and no small-R jets should
be present close in azimuthal angle to p
miss
T . This selection reduces the multijet
background to a negligible level of <1% of the total background in the signal region,
quite remarkable for an all-hadronic analysis. In addition to the presence of one ttagged large-R jet with p T > 250 GeV, one b-tagged small-R track-jet is required,
which reduces the background from V +jets. The obtained limits on the product
of production cross section and B(T → Zt) are between 0.045 and 0.15 pb for
m VLQ = 1.8 and 1 TeV, respectively, comparable to the sensitivity of the all-hadronic
search by CMS [907]. Searches in the dilepton final state, targeting T → Z ((()t,