184
5 Direct Searches for New Physics
R jets. For Z
masses above 100 GeV, no fully merged selection is applied and the Z
decay is reconstructed in the resolved topology only. Constraints on several simplified
models resulting in a mono-Z
signature are reported, with upper cross section limits
in the range of 0.06–27 pb for Z
masses between 80 and 500 GeV [1068]. ATLAS
and CMS exclude vector mediators with masses below 1.8 TeV for m χ up to 700 GeV
for couplings g SM = 0.25 and g χ = 1. These analyses are related to diboson searches
in νν J final states, optimised for resonant signals [711].
Mono-H searches at 13 TeV improved the sensitivity to dark matter models considerably, where the higher
√
s and larger dataset are crucial in the H → γ γ channel [1069–1071]. The H → τ τ channel has been considered for the first time in dark
matter searches in a CMS analysis of 35.9 fb
−1 of data [1071], resulting in better
sensitivity for m φ > 700 GeV relative to the H → γ γ channel, which is better at
smaller masses. In addition to the simplified dark matter model, the results of this
analysis are interpreted in a Z
-two-Higgs-doublet model with a Z
mediator, with
the decay Z
→ H A with A → χ ¯
χ . The combination of the two channels excludes
Z
masses between 550 to 1265 GeV for m A = 300 GeV and m χ = 100 GeV. The
complementary channel with H → bb has been analysed by ATLAS and CMS in
several publications. Both collaborations have gradually improved the reconstruction
algorithms and selection criteria in the development of these analyses. An example
of the improvements in the ATLAS mono-H (bb) searches using 3.2 fb
−1 [1072],
36.1 fb
−1 [1073] and 79.8 fb
−1 [1074] is shown in Fig. 5.22. The trimmed jet mass
of large-R jets is used as sensitive observable in the boosted signal region with
p
miss
T
> 500 GeV. Large-R jets are required to have two ghost-associated b-tagged
track-jets. With respect to the analysis with 3.2 fb
−1 of data, further selection requirements are implemented in the analysis with 36.1 fb
−1 to suppress the tt background.
These include a veto on events containing a reconstructed τ lepton, events with a
b-tagged small-R jet not overlapping with the large-R jet are discarded, and the p T -
sum of all small-R jets should be smaller than 0.57 times that sum added to p T of the
large-R jet. These criteria, together with an improved b-tagging algorithm, results
in a large reduction of the SM backgrounds, as apparent when comparing the two
distributions on top of Fig. 5.22. A further improvement is obtained by considering
VR track-jets for subjet b tagging, as done in the updated analysis with 79.8 fb
−1 of
data. The signal efficiency is improved for high p T , where the track-jets merge when
using a fixed-R algorithm. This leads to a large improvement in sensitivity for mediator >2.5 TeV, as shown in the bottom right of Fig. 5.22. In CMS, the mono-H (bb)
search with 2.3 fb
−1 [1070] uses anti-k T R = 0.8 jets with a pruned jet mass and
subjet b-tagging for H tagging in the merged signal region, whereas small-R jets are
used to define the resolved signal region. The later analysis with 35.9 fb
−1 [1075]
uses CA R = 1.5 jets to gain sensitivity for moderately boosted H bosons, making the resolved selection obsolete. A variant of the double-b tagger, trained on CA
R = 1.5 jets, is used to identify the merged H → bb decay. Further discrimination
is achieved with a selection on the mass-decorrelated N 2 , which is designed to be
uncorrelated to the soft drop jet mass. The PUPPI algorithm enables the use of these
substructure algorithms for CA R = 1.5 jets, where the contribution from pileup
would result in a much worse resolution without a dedicated pileup mitigation. The
5 Direct Searches for New Physics
R jets. For Z
masses above 100 GeV, no fully merged selection is applied and the Z
decay is reconstructed in the resolved topology only. Constraints on several simplified
models resulting in a mono-Z
signature are reported, with upper cross section limits
in the range of 0.06–27 pb for Z
masses between 80 and 500 GeV [1068]. ATLAS
and CMS exclude vector mediators with masses below 1.8 TeV for m χ up to 700 GeV
for couplings g SM = 0.25 and g χ = 1. These analyses are related to diboson searches
in νν J final states, optimised for resonant signals [711].
Mono-H searches at 13 TeV improved the sensitivity to dark matter models considerably, where the higher
√
s and larger dataset are crucial in the H → γ γ channel [1069–1071]. The H → τ τ channel has been considered for the first time in dark
matter searches in a CMS analysis of 35.9 fb
−1 of data [1071], resulting in better
sensitivity for m φ > 700 GeV relative to the H → γ γ channel, which is better at
smaller masses. In addition to the simplified dark matter model, the results of this
analysis are interpreted in a Z
-two-Higgs-doublet model with a Z
mediator, with
the decay Z
→ H A with A → χ ¯
χ . The combination of the two channels excludes
Z
masses between 550 to 1265 GeV for m A = 300 GeV and m χ = 100 GeV. The
complementary channel with H → bb has been analysed by ATLAS and CMS in
several publications. Both collaborations have gradually improved the reconstruction
algorithms and selection criteria in the development of these analyses. An example
of the improvements in the ATLAS mono-H (bb) searches using 3.2 fb
−1 [1072],
36.1 fb
−1 [1073] and 79.8 fb
−1 [1074] is shown in Fig. 5.22. The trimmed jet mass
of large-R jets is used as sensitive observable in the boosted signal region with
p
miss
T
> 500 GeV. Large-R jets are required to have two ghost-associated b-tagged
track-jets. With respect to the analysis with 3.2 fb
−1 of data, further selection requirements are implemented in the analysis with 36.1 fb
−1 to suppress the tt background.
These include a veto on events containing a reconstructed τ lepton, events with a
b-tagged small-R jet not overlapping with the large-R jet are discarded, and the p T -
sum of all small-R jets should be smaller than 0.57 times that sum added to p T of the
large-R jet. These criteria, together with an improved b-tagging algorithm, results
in a large reduction of the SM backgrounds, as apparent when comparing the two
distributions on top of Fig. 5.22. A further improvement is obtained by considering
VR track-jets for subjet b tagging, as done in the updated analysis with 79.8 fb
−1 of
data. The signal efficiency is improved for high p T , where the track-jets merge when
using a fixed-R algorithm. This leads to a large improvement in sensitivity for mediator >2.5 TeV, as shown in the bottom right of Fig. 5.22. In CMS, the mono-H (bb)
search with 2.3 fb
−1 [1070] uses anti-k T R = 0.8 jets with a pruned jet mass and
subjet b-tagging for H tagging in the merged signal region, whereas small-R jets are
used to define the resolved signal region. The later analysis with 35.9 fb
−1 [1075]
uses CA R = 1.5 jets to gain sensitivity for moderately boosted H bosons, making the resolved selection obsolete. A variant of the double-b tagger, trained on CA
R = 1.5 jets, is used to identify the merged H → bb decay. Further discrimination
is achieved with a selection on the mass-decorrelated N 2 , which is designed to be
uncorrelated to the soft drop jet mass. The PUPPI algorithm enables the use of these
substructure algorithms for CA R = 1.5 jets, where the contribution from pileup
would result in a much worse resolution without a dedicated pileup mitigation. The