4.1 Measurements of Jet Substructure Observables
103
distribution in z g has first been measured using CMS open data
2 on an inclusive jet
sample [589, 590]. The theoretical prediction at modified leading-logarithmic accuracy describes the data well, even though the data have not been corrected for detector
effects and the all-particle prediction is compared to z g calculated from reconstructed
tracks only. Other quantities, like the opening angle θ g = R 12 /R or angularities
e
(g)
β are not described by the analytical calculations [590], which can be attributed
to larger detector effects. In a later measurement, ATLAS has reported unfolded
distributions of ρ, z g and θ g (called r g in this publication) [591]. The analytical
calculations, including non-global logarithms to achieve full NLL accuracy [592],
describe the unfolded distributions in θ g very well. The θ g distribution for soft drop
grooming with β = 0 is very similar for quark and gluon jets, as expected, since θ g is
independent of α S to leading order. For β > 0, different shapes are observed, where
the gluon jet distribution tends towards a larger splitting. The splitting function z g
has also been measured by CMS in pp and Pb-Pb collisions [593]. The results from
pp and peripheral Pb-Pb collisions agree within 15%. The z g distribution in central
Pb-Pb collisions is steeper, indicating that the parton splitting process is modified
by the hot medium created in heavy ion collisions.
The full radiation pattern of a jet can be conceived as emissions in a plane in
ln 1/z and ln 1/θ , first described by Lund diagrams [468], where the ungroomed
versions of z g and θ g are used. The Lund jet plane [469], also called primary Lund
plane, is obtained by reversing the clustering history of a jet and following the
harder branch at each clustering step. Each emission corresponds to one entry in the
Lund jet plane, such that any jet can be represented by a number of entries in the
plane of ln 1/z and ln 1/θ . Different regions in this plane correspond to different
physical effects as shown in Fig. 4.6. While only primary emissions are followed
in the construction of the Lund jet plane and secondary emissions are discarded, its
measurement visualises the salient features of jet fragmentation and radiation patterns
inside jets. The soft drop splitting variable z g is identical to the value of z at the first
splitting that fulfils the soft drop condition (2.41), such that the variable pair (z g , θ g )
is also included in the Lund jet plane. A first measurement has been presented very
recently by ATLAS using a dijet selection on the full available 13 TeV data [470],
and is shown in Fig. 4.6 (right). As expected, hard-collinear (bottom right) and softwide angle (top left) radiation have the highest probability. The average density
of emissions is approximately uniform for hard-wide angle radiation (bottom left).
The measurement can entangle regions with high sensitivity to the choice of parton
shower and hadronisation modelling, and will provide important input in future
developments of these processes.
Instead of probing the substructure of a mixture of quark and gluon jets, it is
possible to probe primarily quark jets in identified tt events. ATLAS has exploited
this idea in a measurement of differential and integrated jet shapes in dileptonic and
lepton+jets final states of tt production at 7 TeV [594]. The two final states enable a
comparison of jet shapes between light quark and b jets. At low p T , it is observed that
2 In 2014, CMS has publicly released the 7 TeV pp collision data recorded in 2010 through the
CERN open data portal [588]. By now, also data from the years 2011 and 2012 have been released.
103
distribution in z g has first been measured using CMS open data
2 on an inclusive jet
sample [589, 590]. The theoretical prediction at modified leading-logarithmic accuracy describes the data well, even though the data have not been corrected for detector
effects and the all-particle prediction is compared to z g calculated from reconstructed
tracks only. Other quantities, like the opening angle θ g = R 12 /R or angularities
e
(g)
β are not described by the analytical calculations [590], which can be attributed
to larger detector effects. In a later measurement, ATLAS has reported unfolded
distributions of ρ, z g and θ g (called r g in this publication) [591]. The analytical
calculations, including non-global logarithms to achieve full NLL accuracy [592],
describe the unfolded distributions in θ g very well. The θ g distribution for soft drop
grooming with β = 0 is very similar for quark and gluon jets, as expected, since θ g is
independent of α S to leading order. For β > 0, different shapes are observed, where
the gluon jet distribution tends towards a larger splitting. The splitting function z g
has also been measured by CMS in pp and Pb-Pb collisions [593]. The results from
pp and peripheral Pb-Pb collisions agree within 15%. The z g distribution in central
Pb-Pb collisions is steeper, indicating that the parton splitting process is modified
by the hot medium created in heavy ion collisions.
The full radiation pattern of a jet can be conceived as emissions in a plane in
ln 1/z and ln 1/θ , first described by Lund diagrams [468], where the ungroomed
versions of z g and θ g are used. The Lund jet plane [469], also called primary Lund
plane, is obtained by reversing the clustering history of a jet and following the
harder branch at each clustering step. Each emission corresponds to one entry in the
Lund jet plane, such that any jet can be represented by a number of entries in the
plane of ln 1/z and ln 1/θ . Different regions in this plane correspond to different
physical effects as shown in Fig. 4.6. While only primary emissions are followed
in the construction of the Lund jet plane and secondary emissions are discarded, its
measurement visualises the salient features of jet fragmentation and radiation patterns
inside jets. The soft drop splitting variable z g is identical to the value of z at the first
splitting that fulfils the soft drop condition (2.41), such that the variable pair (z g , θ g )
is also included in the Lund jet plane. A first measurement has been presented very
recently by ATLAS using a dijet selection on the full available 13 TeV data [470],
and is shown in Fig. 4.6 (right). As expected, hard-collinear (bottom right) and softwide angle (top left) radiation have the highest probability. The average density
of emissions is approximately uniform for hard-wide angle radiation (bottom left).
The measurement can entangle regions with high sensitivity to the choice of parton
shower and hadronisation modelling, and will provide important input in future
developments of these processes.
Instead of probing the substructure of a mixture of quark and gluon jets, it is
possible to probe primarily quark jets in identified tt events. ATLAS has exploited
this idea in a measurement of differential and integrated jet shapes in dileptonic and
lepton+jets final states of tt production at 7 TeV [594]. The two final states enable a
comparison of jet shapes between light quark and b jets. At low p T , it is observed that
2 In 2014, CMS has publicly released the 7 TeV pp collision data recorded in 2010 through the
CERN open data portal [588]. By now, also data from the years 2011 and 2012 have been released.