5.7 Supersymmetry
201
associated hits in the pixel detector.
6 For each jet with a photon candidate, the jet
is reclustered with the k T algorithm using the photon and all other jet constituents
as input. Reversing the clustering history and looking at the last clustering step,
the less massive of the two pseudojets is the first subjet. In the second step the
more massive pseudojet is declustered, resulting in the second and third subjets.
Signal jets are required to have three subjets with p T > 10 GeV and τ 31 < 0.4. The
photon subjet energy fraction f γ is defined as the fraction of the photon p T relative
to the p T of the subjet containing the photon. This is a measure of the additional
activity in proximity to the photon and serves as a way to discriminate between
signal photons and photons from the decay of hadrons, which are abundant in SM
multijet production. The distribution of f γ is shown in Fig. 5.29 for data, multijet
background and signal jets. The distributions for signal jets show a more pronounced
peak at f γ ≈ 1 for increasing ˜
χ
0
1 mass, because of a larger relative photon p T with
respect to the jet axis, resulting in more isolated photons. Signal jets are classified
into tight and loose photon jets by f γ > 0.9 and f γ < 0.9, respectively. Six signal
regions are defined for events with three or four and more large-R jets, where at
least one jet has to be identified as tight photon jet. In each of these signal regions,
the H T distribution is measured and serves as a discriminator between signal and
background. The background is obtained from data by weighting events with at
most one loose photon jet with the misidentification probability per jet, obtained
as a function of p T and η. The largest uncertainty in this search originates from
the measurement of the signal efficiency, which is performed in a control region
enriched with dileptonic tt decays. Fully-merged t decays of the kind t → eνb with
an additional gluon from ISR, reconstructed in the same jet, are utilised as a proxy
for the ggγ signal jets. The pixel veto is reversed to allow for an efficient selection of
electrons instead of photons. Differences between samples where the hadronisation
is simulated with Pythia and Herwig result in uncertainties in the signal efficiency
between 24 and 85% for p T between 200 and 500 GeV. This analysis would benefit
from a better understanding of the differences in the parton shower and hadronisation
in these simulations. The analysis excludes ˜
g masses up to 1.7 TeV for a ˜
χ
0
1 mass of
200 GeV. This is the first result exploiting jet substructure methods to identify single
jets originating from two gluons and a photon.
In non-minimal SUSY scenarios [1189, 1190], the extended scalar sector contains
additional CP-even and CP-odd scalar particles. These are the result of the symmetry breaking of a scalar superfield, which acquires a vacuum expectation value and
dynamically generates the interactions of the Higgs doublet superfields. The lightest
CP-odd scalar, the axion a, is considerably lighter than the other scalars for a wide
range of the parameter space [1191, 1192]. For axion masses m a < m H /2, the decay
H → aa is possible and is an intriguing channel to discover these new states [1193,
6 Photon conversions γ → e + e − in the tracking detector may lead to tracks pointing in the direction
of the photon candidate, such that a higher selection efficiency is obtained by a pixel veto, instead
of discarding all candidates with an associated track.
201
associated hits in the pixel detector.
6 For each jet with a photon candidate, the jet
is reclustered with the k T algorithm using the photon and all other jet constituents
as input. Reversing the clustering history and looking at the last clustering step,
the less massive of the two pseudojets is the first subjet. In the second step the
more massive pseudojet is declustered, resulting in the second and third subjets.
Signal jets are required to have three subjets with p T > 10 GeV and τ 31 < 0.4. The
photon subjet energy fraction f γ is defined as the fraction of the photon p T relative
to the p T of the subjet containing the photon. This is a measure of the additional
activity in proximity to the photon and serves as a way to discriminate between
signal photons and photons from the decay of hadrons, which are abundant in SM
multijet production. The distribution of f γ is shown in Fig. 5.29 for data, multijet
background and signal jets. The distributions for signal jets show a more pronounced
peak at f γ ≈ 1 for increasing ˜
χ
0
1 mass, because of a larger relative photon p T with
respect to the jet axis, resulting in more isolated photons. Signal jets are classified
into tight and loose photon jets by f γ > 0.9 and f γ < 0.9, respectively. Six signal
regions are defined for events with three or four and more large-R jets, where at
least one jet has to be identified as tight photon jet. In each of these signal regions,
the H T distribution is measured and serves as a discriminator between signal and
background. The background is obtained from data by weighting events with at
most one loose photon jet with the misidentification probability per jet, obtained
as a function of p T and η. The largest uncertainty in this search originates from
the measurement of the signal efficiency, which is performed in a control region
enriched with dileptonic tt decays. Fully-merged t decays of the kind t → eνb with
an additional gluon from ISR, reconstructed in the same jet, are utilised as a proxy
for the ggγ signal jets. The pixel veto is reversed to allow for an efficient selection of
electrons instead of photons. Differences between samples where the hadronisation
is simulated with Pythia and Herwig result in uncertainties in the signal efficiency
between 24 and 85% for p T between 200 and 500 GeV. This analysis would benefit
from a better understanding of the differences in the parton shower and hadronisation
in these simulations. The analysis excludes ˜
g masses up to 1.7 TeV for a ˜
χ
0
1 mass of
200 GeV. This is the first result exploiting jet substructure methods to identify single
jets originating from two gluons and a photon.
In non-minimal SUSY scenarios [1189, 1190], the extended scalar sector contains
additional CP-even and CP-odd scalar particles. These are the result of the symmetry breaking of a scalar superfield, which acquires a vacuum expectation value and
dynamically generates the interactions of the Higgs doublet superfields. The lightest
CP-odd scalar, the axion a, is considerably lighter than the other scalars for a wide
range of the parameter space [1191, 1192]. For axion masses m a < m H /2, the decay
H → aa is possible and is an intriguing channel to discover these new states [1193,
6 Photon conversions γ → e + e − in the tracking detector may lead to tracks pointing in the direction
of the photon candidate, such that a higher selection efficiency is obtained by a pixel veto, instead
of discarding all candidates with an associated track.