5.1 Diboson Resonances
129
uses a categorisation into loose and tight V and H tags, defined by a selection on
τ 21 and the double-b tagging discriminator. In addition, the soft drop jet masses
have to be within 65 < m jet < 105 GeV for a V tag and 105 < m jet < 135 GeV for
an H tag. Jets tagged as V bosons are further categorised in W and Z jets, with
65 < m jet < 85 GeV and 85 < m jet < 105 GeV, respectively. Even though the W
and Z jet mass peaks cannot be fully resolved, this classification allows a partial
discrimination between W H and Z H resonances. The signal regions corresponding
to these two final states can be readily combined in a statistical analysis of the results,
as these correspond to independent data. In ATLAS, the W H and Z H signal regions
have an overlap of approximately 60% due to the overlapping jet mass selections,
such that these can not be combined easily. In both analyses, the multijet background
is estimated by one-dimensional, monotonically decreasing functions, similar to the
parametrisations used in V V resonances searches, given in (5.1). In addition, ATLAS
constrains the normalisation of the multijet background from control regions with no
b-tagged track-jets associated with the H jets. The two analyses achieve very similar
sensitivities, despite the very different choices made in the design of the analyses.
Upwards fluctuations in the excluded upper limits at 95% CL seen in one experiment
correspond to downward fluctuations in the other, such that overall no significant
excess or deficiency relative to the expected limits are observed. Recently, ATLAS
has updated the VH resonance search in the all-hadronic final state using 139 fb
−1 of
13 TeV data [721]. The analysis relies on the same strategy as the previous analysis,
but the H → bb identification has been improved by the use of VR track-jets, which
result in higher efficiency at high p T at constant background rejection. The larger
data set and improved H tagging results in an improvement in the expected upper
cross section limits of factors between 3 and 5, which translates into an extended
mass reach by about 500 GeV.
Searches in dilepton+jets final states include the channels ν J , J and νν J ,
similar to V V resonance searches, but with the fragmentation products of a bb pair
merged into the large-R jet. In analogy to V V resonance searches, the dilepton+jets
VH channels extend the mass range to lower values thanks to the leptonic V decay.
ATLAS has considered resolved and merged H decays into bb in an analysis of
36.1 fb
−1 of 13 TeV data [722], which succeeds an analysis of 8 TeV data [723] and
supersedes an early result based on 13 TeV data [724]. For events which satisfy both
the resolved and merged selections, priority is given to the resolved category with
two identified small-R b jets over the merged category with a large-R H jet. This
provides higher sensitivity to resonances with a mass near 1 TeV. This cascading
selection results in sensitivity for resonance masses as low as 200 GeV up to 5 TeV. In
a similar analysis by CMS [725] on the same amount of data, only merged final states
are considered, probing resonance masses above 800 GeV. Unlike in the analysis by
ATLAS, the background estimation is based on data from control regions with the
α method, instead of using simulated SM backgrounds. The sensitivities of the two
analyses are very similar and the versatility of the final states allows to constrain
several BSM models. Besides placing limits on the HVT model, also the parameter
space of two-Higgs doublet models and an extended two-Higgs doublet model with a
129
uses a categorisation into loose and tight V and H tags, defined by a selection on
τ 21 and the double-b tagging discriminator. In addition, the soft drop jet masses
have to be within 65 < m jet < 105 GeV for a V tag and 105 < m jet < 135 GeV for
an H tag. Jets tagged as V bosons are further categorised in W and Z jets, with
65 < m jet < 85 GeV and 85 < m jet < 105 GeV, respectively. Even though the W
and Z jet mass peaks cannot be fully resolved, this classification allows a partial
discrimination between W H and Z H resonances. The signal regions corresponding
to these two final states can be readily combined in a statistical analysis of the results,
as these correspond to independent data. In ATLAS, the W H and Z H signal regions
have an overlap of approximately 60% due to the overlapping jet mass selections,
such that these can not be combined easily. In both analyses, the multijet background
is estimated by one-dimensional, monotonically decreasing functions, similar to the
parametrisations used in V V resonances searches, given in (5.1). In addition, ATLAS
constrains the normalisation of the multijet background from control regions with no
b-tagged track-jets associated with the H jets. The two analyses achieve very similar
sensitivities, despite the very different choices made in the design of the analyses.
Upwards fluctuations in the excluded upper limits at 95% CL seen in one experiment
correspond to downward fluctuations in the other, such that overall no significant
excess or deficiency relative to the expected limits are observed. Recently, ATLAS
has updated the VH resonance search in the all-hadronic final state using 139 fb
−1 of
13 TeV data [721]. The analysis relies on the same strategy as the previous analysis,
but the H → bb identification has been improved by the use of VR track-jets, which
result in higher efficiency at high p T at constant background rejection. The larger
data set and improved H tagging results in an improvement in the expected upper
cross section limits of factors between 3 and 5, which translates into an extended
mass reach by about 500 GeV.
Searches in dilepton+jets final states include the channels ν J , J and νν J ,
similar to V V resonance searches, but with the fragmentation products of a bb pair
merged into the large-R jet. In analogy to V V resonance searches, the dilepton+jets
VH channels extend the mass range to lower values thanks to the leptonic V decay.
ATLAS has considered resolved and merged H decays into bb in an analysis of
36.1 fb
−1 of 13 TeV data [722], which succeeds an analysis of 8 TeV data [723] and
supersedes an early result based on 13 TeV data [724]. For events which satisfy both
the resolved and merged selections, priority is given to the resolved category with
two identified small-R b jets over the merged category with a large-R H jet. This
provides higher sensitivity to resonances with a mass near 1 TeV. This cascading
selection results in sensitivity for resonance masses as low as 200 GeV up to 5 TeV. In
a similar analysis by CMS [725] on the same amount of data, only merged final states
are considered, probing resonance masses above 800 GeV. Unlike in the analysis by
ATLAS, the background estimation is based on data from control regions with the
α method, instead of using simulated SM backgrounds. The sensitivities of the two
analyses are very similar and the versatility of the final states allows to constrain
several BSM models. Besides placing limits on the HVT model, also the parameter
space of two-Higgs doublet models and an extended two-Higgs doublet model with a