186
5 Direct Searches for New Physics
the s + χ ¯
χ signal is produced through a heavy mediator, qq → Z
→ sχ ¯
χ [1080].
For s with masses larger than 160 and 180 GeV, the decays s → W W and s → Z Z
become relevant, respectively [137]. The analysis considers the fully hadronic decay
chain s → V V → qq
() qq
() , where the four quarks are reconstructed in a single
large-R jet. This is one of the few analyses at the LHC targeting four-prong decays
in single jets. If the reconstruction of the fully merged decay fails, a partially merged
reconstruction is attempted, where a large-R jet and one or two accompanying smallR jets are combined to form the s → V V decay. Discrimination between signal
and background jets is improved by the use of track-assisted reclustered jets [430],
see Sect. 3.2. The improvement in the resolution of N -subjettiness ratios used for
V V → 4q tagging results in a stronger expected discovery significance by a factor
of up to 2.5. Besides the gain in resolution, another advantage of using track-assisted
reclustered jets is the possibility to propagate experimental uncertainties consistently.
This results in a realistic estimation of the signal efficiency without a dedicated measurement, which is not possible for this channel because of a lack of SM processes
with sufficiently large cross sections. In the fully merged category, signal jets are
required to have p T > 300 GeV, a trimmed mass in the range 100–400 GeV and
satisfy the combination of N -subjettiness ratios τ 42 < 0.3 and τ 43 < 0.6. In the partially merged category, the large-R jet is combined with two small-R jets if its mass
is between 60 and 100 GeV and with one small-R jet if it has a mass larger than
100 GeV. The considered small-R jets have to be within R < 2.5 of the large-R
jet. This recovers some of the signal events, but with the drawback of larger backgrounds in the partially merged category. The two distributions in the reconstructed
s mass, measured in the fully merged and partially merged categories, are used to
derive upper cross section limits on the signal model. The limits range between
between 0.32 and 0.03 pb for s with mass between 160 and 360 GeV, respectively,
and m χ = 200 GeV and m Z = 1 TeV.
Dark matter searches targeting models with mediators coupling to third generation
quarks have also profited from developments in jet substructure reconstruction. In an
ATLAS search for χ ¯
χ produced in association with b, bb or tt, using 36.1 fb
−1 of
13 TeV data [1081], the jet mass of reclustered jets is used to enhance the sensitivity.
Two signal regions in the jets+ p
miss
T
final state have been optimised for tt +χ ¯
χ
production. Mediators with 100 < m φ < 350 GeV result in a moderately boosted
tt topology, which is identified by requiring the jet masses of the two p T -leading
R = 1.2 reclustered jets to be >140 and 80 GeV. For lower masses, m φ < 100 GeV,
the on average smaller boost results in semi-merged topologies, which are selected
by requiring the jet masses to be greater than 80 GeV for both p T -leading R =
0.8 reclustered jets. Other topological selection criteria are applied to suppress SM
backgrounds, resulting in a negligible contribution from multijet production. Two
further signal regions in the jets+ p
miss
T
final state are optimised for bb + χ ¯
χ and
b + χ ¯
χ production, and a fifth signal region considers dilepton tt events. Mediator
masses between 10 and 50 GeV for scalar mediators coupling to top quarks are
excluded for m χ = 1 GeV. While the sensitivity of the search is driven by the dilepton
category for m φ < 150 GeV, for larger mediator masses the jets+ p
miss
T
final state is
better. The dilepton final state has been updated using the full 13 TeV data set with
5 Direct Searches for New Physics
the s + χ ¯
χ signal is produced through a heavy mediator, qq → Z
→ sχ ¯
χ [1080].
For s with masses larger than 160 and 180 GeV, the decays s → W W and s → Z Z
become relevant, respectively [137]. The analysis considers the fully hadronic decay
chain s → V V → qq
() , where the four quarks are reconstructed in a single
large-R jet. This is one of the few analyses at the LHC targeting four-prong decays
in single jets. If the reconstruction of the fully merged decay fails, a partially merged
reconstruction is attempted, where a large-R jet and one or two accompanying smallR jets are combined to form the s → V V decay. Discrimination between signal
and background jets is improved by the use of track-assisted reclustered jets [430],
see Sect. 3.2. The improvement in the resolution of N -subjettiness ratios used for
V V → 4q tagging results in a stronger expected discovery significance by a factor
of up to 2.5. Besides the gain in resolution, another advantage of using track-assisted
reclustered jets is the possibility to propagate experimental uncertainties consistently.
This results in a realistic estimation of the signal efficiency without a dedicated measurement, which is not possible for this channel because of a lack of SM processes
with sufficiently large cross sections. In the fully merged category, signal jets are
required to have p T > 300 GeV, a trimmed mass in the range 100–400 GeV and
satisfy the combination of N -subjettiness ratios τ 42 < 0.3 and τ 43 < 0.6. In the partially merged category, the large-R jet is combined with two small-R jets if its mass
is between 60 and 100 GeV and with one small-R jet if it has a mass larger than
100 GeV. The considered small-R jets have to be within R < 2.5 of the large-R
jet. This recovers some of the signal events, but with the drawback of larger backgrounds in the partially merged category. The two distributions in the reconstructed
s mass, measured in the fully merged and partially merged categories, are used to
derive upper cross section limits on the signal model. The limits range between
between 0.32 and 0.03 pb for s with mass between 160 and 360 GeV, respectively,
and m χ = 200 GeV and m Z = 1 TeV.
Dark matter searches targeting models with mediators coupling to third generation
quarks have also profited from developments in jet substructure reconstruction. In an
ATLAS search for χ ¯
χ produced in association with b, bb or tt, using 36.1 fb
−1 of
13 TeV data [1081], the jet mass of reclustered jets is used to enhance the sensitivity.
Two signal regions in the jets+ p
miss
T
final state have been optimised for tt +χ ¯
χ
production. Mediators with 100 < m φ < 350 GeV result in a moderately boosted
tt topology, which is identified by requiring the jet masses of the two p T -leading
R = 1.2 reclustered jets to be >140 and 80 GeV. For lower masses, m φ < 100 GeV,
the on average smaller boost results in semi-merged topologies, which are selected
by requiring the jet masses to be greater than 80 GeV for both p T -leading R =
0.8 reclustered jets. Other topological selection criteria are applied to suppress SM
backgrounds, resulting in a negligible contribution from multijet production. Two
further signal regions in the jets+ p
miss
T
final state are optimised for bb + χ ¯
χ and
b + χ ¯
χ production, and a fifth signal region considers dilepton tt events. Mediator
masses between 10 and 50 GeV for scalar mediators coupling to top quarks are
excluded for m χ = 1 GeV. While the sensitivity of the search is driven by the dilepton
category for m φ < 150 GeV, for larger mediator masses the jets+ p
miss
T
final state is
better. The dilepton final state has been updated using the full 13 TeV data set with