5.8 Leptoquarks
207
to exclude LQ masses below 900 GeV for any combination of branching fractions
fulfilling B(LQ → tμ) + B(LQ → tτ ) = 1 or B(LQ → tμ) + B(LQ → bν) = 1.
The excluded mass range for LQs coupling to top quarks goes beyond 1 TeV in
many cases, suggesting that searches considering boosted top quarks and jet substructure can improve the sensitivity at the LHC. In particular, when combining signatures
from pair and single production, the sensitivity can be drastically improved by the use
of top tagging [1276, 1277]. The reason for the improvement lies in the mechanisms
of single production, which results in the same tt final state for LQs coupling
to top quarks as pair production, but with a low- p T t produced in association with
the + LQ(→ t) system from the production vertex. The corresponding signature can be captured by selecting events with an pair and a t-tagged large-R jet,
capturing events from single and pair production simultaneously. This strategy has
been followed by a CMS analysis using 137 fb
−1 of 13 TeV data [1278], which has
been optimised for the scenario of B(LQ → tτ ) = B(LQ → bν) = 0.5, resulting
in pp → tτ νb production. For single production, the b can be soft or very forward, outside of the detector acceptance. Events are selected with p
miss
T
> 200 GeV,
one τ h candidate, at least one b-tagged small-R jet and a reconstructed t candidate.
The analysis considers fully merged, partially merged and resolved t decays. Fully
merged t decays are reconstructed using the soft drop jet mass and τ 32 . Partially
merged decays are reconstructed with a W jet, identified using the pruned jet mass
and τ 21 , and a small-R jet with a combined mass close to m t . Resolved decays are
reconstructed using three small-R jets. Signal events are categorised by a successful
boosted or resolved t reconstruction and the b jet multiplicity. The distribution in
S T is measured in the resulting four signal regions, where S T is the sum of p
miss
T
and p T of the reconstructed t and τ h . The analysis excludes scalar LQ masses up
to 950 GeV, considering pair production only, which constitutes an improvement by
about 140 GeV compared to the previous CMS search for LQ → tτ , which obtained
a mass limit of 810 GeV for B(LQ → tτ ) = 0.5 [1269]. Taking into account single
and pair production, the limit improves to 980 and 1020 GeV for LQ-quark-lepton
couplings of 1.5 and 2.5, respectively.
Very recently, ATLAS has completed a search for cross-generation couplings in
the LQ → te and LQ → tμ channels using 139 fb
−1 of 13 TeV data [1279]. The
analysis aims at LQ masses above 1 TeV with decays of boosted t quarks. Events are
selected in e
+ e
− and μ
+
μ
− final states with dilepton masses greater than 120 GeV
and two trimmed large-R jets with p T > 200 GeV and m jet > 50 GeV. The analysis
treats the e
+ e
− and μ
+
μ
− channels separately, because simultaneous LQ couplings
to electrons and muons are tightly constrained by the absence of flavour-changing
neutral currents [1225]. The signal is enhanced over SM backgrounds with the use
of a BDT with 29 inputs in the electron channel and 32 inputs in the muon channel.
The inputs include the discriminating variables from a LQ reconstruction using the
recursive jigsaw technique [1280], kinematic variables derived from the leptons,
jets and p
miss
T , the large-R jet masses, and three jet substructure variables: the k T -
splitting scale
√
d 32 , τ 32 and the mass of the two pseudojets in the second-to-last
clustering step of the k T algorithm, Q W [486]. Interestingly, the jet substructure
variables are only included in the muon channel, because of an overlap removal of
207
to exclude LQ masses below 900 GeV for any combination of branching fractions
fulfilling B(LQ → tμ) + B(LQ → tτ ) = 1 or B(LQ → tμ) + B(LQ → bν) = 1.
The excluded mass range for LQs coupling to top quarks goes beyond 1 TeV in
many cases, suggesting that searches considering boosted top quarks and jet substructure can improve the sensitivity at the LHC. In particular, when combining signatures
from pair and single production, the sensitivity can be drastically improved by the use
of top tagging [1276, 1277]. The reason for the improvement lies in the mechanisms
of single production, which results in the same tt final state for LQs coupling
to top quarks as pair production, but with a low- p T t produced in association with
the + LQ(→ t) system from the production vertex. The corresponding signature can be captured by selecting events with an pair and a t-tagged large-R jet,
capturing events from single and pair production simultaneously. This strategy has
been followed by a CMS analysis using 137 fb
−1 of 13 TeV data [1278], which has
been optimised for the scenario of B(LQ → tτ ) = B(LQ → bν) = 0.5, resulting
in pp → tτ νb production. For single production, the b can be soft or very forward, outside of the detector acceptance. Events are selected with p
miss
T
> 200 GeV,
one τ h candidate, at least one b-tagged small-R jet and a reconstructed t candidate.
The analysis considers fully merged, partially merged and resolved t decays. Fully
merged t decays are reconstructed using the soft drop jet mass and τ 32 . Partially
merged decays are reconstructed with a W jet, identified using the pruned jet mass
and τ 21 , and a small-R jet with a combined mass close to m t . Resolved decays are
reconstructed using three small-R jets. Signal events are categorised by a successful
boosted or resolved t reconstruction and the b jet multiplicity. The distribution in
S T is measured in the resulting four signal regions, where S T is the sum of p
miss
T
and p T of the reconstructed t and τ h . The analysis excludes scalar LQ masses up
to 950 GeV, considering pair production only, which constitutes an improvement by
about 140 GeV compared to the previous CMS search for LQ → tτ , which obtained
a mass limit of 810 GeV for B(LQ → tτ ) = 0.5 [1269]. Taking into account single
and pair production, the limit improves to 980 and 1020 GeV for LQ-quark-lepton
couplings of 1.5 and 2.5, respectively.
Very recently, ATLAS has completed a search for cross-generation couplings in
the LQ → te and LQ → tμ channels using 139 fb
−1 of 13 TeV data [1279]. The
analysis aims at LQ masses above 1 TeV with decays of boosted t quarks. Events are
selected in e
+ e
− and μ
+
μ
− final states with dilepton masses greater than 120 GeV
and two trimmed large-R jets with p T > 200 GeV and m jet > 50 GeV. The analysis
treats the e
+ e
− and μ
+
μ
− channels separately, because simultaneous LQ couplings
to electrons and muons are tightly constrained by the absence of flavour-changing
neutral currents [1225]. The signal is enhanced over SM backgrounds with the use
of a BDT with 29 inputs in the electron channel and 32 inputs in the muon channel.
The inputs include the discriminating variables from a LQ reconstruction using the
recursive jigsaw technique [1280], kinematic variables derived from the leptons,
jets and p
miss
T , the large-R jet masses, and three jet substructure variables: the k T -
splitting scale
√
d 32 , τ 32 and the mass of the two pseudojets in the second-to-last
clustering step of the k T algorithm, Q W [486]. Interestingly, the jet substructure
variables are only included in the muon channel, because of an overlap removal of