4.1 Measurements of Jet Substructure Observables
97
and systematic uncertainties of high p T V H and H +jet production in the H → bb
channel. After a high p T photon selection, Z jets are identified using two b tags from
associated track jets. The dominant background originates from γ +jets production
with g → bb splittings. After the signal region selection, this background yields
about a factor of 20 more events than signal events. It can be estimated in data from
control regions with either one b tagged jet, or a less restrictive photon selection. It
should be noted that for this background estimation to work, simulated contributions
from tt+γ and W +γ production need to be subtracted from data, as the shape of the
mass distribution differs from that of γ +jets. The resulting distribution in the soft drop
jet mass of γ +jet events, where the jet has two b tags, shows a peak from Z → bb
on a falling distribution. After subtracting the background, the observed significance
over the background-only hypothesis is 2.7 standard deviations for Z → bb, with
an expected significance of 2.7. The unfolded soft drop jet mass distribution at the
particle level is shown in Fig. 4.3 (right). It peaks at around 95 GeV and is described
well by the Sherpa event generator. The trimmed jet mass distribution peaks at lower
values, around 85 GeV, and shows a higher significance of 3.9 standard deviations,
because of a narrower jet mass distribution at the particle level.
4.1.3 Jet Mass of Top Quark Jets 1
The jet mass of top quark jets, with the full top quark decay merged into a single
large-R jet, has been measured in +jets tt events, where stands for an electron or
muon, recorded by CMS at 8 [438] and 13 TeV [557]. The 8 TeV measurement is the
first jet mass distribution unfolded at the particle level probing three prong decays.
Large-R jets are reconstructed with the CA algorithm using a distance parameter
of R = 1.2. The larger value of R in this measurement compared to the default
R = 0.8 applied for top tagging applications in CMS is due to an optimisation of
statistical precision versus the width of the jet mass distribution at the particle level
and the JMR. The number of fully merged top quarks grows with increasing R,
but so does the width of the jet mass distribution and the susceptibility to pileup
and the underlying event. The leading jet p T is required to be above 400 GeV to
ensure the hadronic top quark decay to be fully captured within the large-R jet. No
substructure selection is applied on the high- p T large-R jet in order not to bias the jet
mass measurement. The shape of the particle-level differential tt cross section as a
function of the leading jet mass is well described by the simulations. The simulations
predict a larger cross section than observed in the measurement, consistent with the
tt cross section measurements by ATLAS and CMS at high p T (see Sect. 4.2.3).
The experimental systematic uncertainties are dominated by the uncertainties on the
jet mass and energy scale, but are smaller than the uncertainties due to the signal
modelling, coming from the choice of the top quark mass, the parton showering
1 The text in this subsection is based on Refs. [26, 557], written by the author. It has been adjusted
to fit this book.
97
and systematic uncertainties of high p T V H and H +jet production in the H → bb
channel. After a high p T photon selection, Z jets are identified using two b tags from
associated track jets. The dominant background originates from γ +jets production
with g → bb splittings. After the signal region selection, this background yields
about a factor of 20 more events than signal events. It can be estimated in data from
control regions with either one b tagged jet, or a less restrictive photon selection. It
should be noted that for this background estimation to work, simulated contributions
from tt+γ and W +γ production need to be subtracted from data, as the shape of the
mass distribution differs from that of γ +jets. The resulting distribution in the soft drop
jet mass of γ +jet events, where the jet has two b tags, shows a peak from Z → bb
on a falling distribution. After subtracting the background, the observed significance
over the background-only hypothesis is 2.7 standard deviations for Z → bb, with
an expected significance of 2.7. The unfolded soft drop jet mass distribution at the
particle level is shown in Fig. 4.3 (right). It peaks at around 95 GeV and is described
well by the Sherpa event generator. The trimmed jet mass distribution peaks at lower
values, around 85 GeV, and shows a higher significance of 3.9 standard deviations,
because of a narrower jet mass distribution at the particle level.
4.1.3 Jet Mass of Top Quark Jets 1
The jet mass of top quark jets, with the full top quark decay merged into a single
large-R jet, has been measured in +jets tt events, where stands for an electron or
muon, recorded by CMS at 8 [438] and 13 TeV [557]. The 8 TeV measurement is the
first jet mass distribution unfolded at the particle level probing three prong decays.
Large-R jets are reconstructed with the CA algorithm using a distance parameter
of R = 1.2. The larger value of R in this measurement compared to the default
R = 0.8 applied for top tagging applications in CMS is due to an optimisation of
statistical precision versus the width of the jet mass distribution at the particle level
and the JMR. The number of fully merged top quarks grows with increasing R,
but so does the width of the jet mass distribution and the susceptibility to pileup
and the underlying event. The leading jet p T is required to be above 400 GeV to
ensure the hadronic top quark decay to be fully captured within the large-R jet. No
substructure selection is applied on the high- p T large-R jet in order not to bias the jet
mass measurement. The shape of the particle-level differential tt cross section as a
function of the leading jet mass is well described by the simulations. The simulations
predict a larger cross section than observed in the measurement, consistent with the
tt cross section measurements by ATLAS and CMS at high p T (see Sect. 4.2.3).
The experimental systematic uncertainties are dominated by the uncertainties on the
jet mass and energy scale, but are smaller than the uncertainties due to the signal
modelling, coming from the choice of the top quark mass, the parton showering
1 The text in this subsection is based on Refs. [26, 557], written by the author. It has been adjusted
to fit this book.