202
5 Direct Searches for New Physics
Fig. 5.29 Distribution in the
photon subjet energy fraction
f γ for large-R jets with
p T > 200 GeV, three
reconstructed subjets and
τ 31 < 0.4. Taken from
[1188]
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
1
γ
f
0
0.05
0.1
0.15
0.2
0.25
Fraction of jets
Data
QCD simulation
1TeV
g
~
150 GeV, M
1
0
χ
∼
M
1TeV
g
~
200 GeV, M
1
0
χ
∼
M
1TeV
g
~
300 GeV, M
1
0
χ
∼
M
(13 TeV)
-1
35.9 fb
CMS
1194].
7 Current experimental constraints on non-SM branching fractions of the H
boson result in upper bounds of 19 at 95% CL [1195, 1196], which can be taken as
the upper bound for B(H → aa). Since the a boson mixes with the H boson, the
hierarchy of predicted branching fractions into SM particles is similar, with a → bb,
a → τ τ and a → μμ the largest, if kinematically accessible [1197]. In parameter
regions resulting in small fermionic couplings, the loop induced decays a → γ γ
and a → gg can be dominant [1198]. Searches at the LHC for the decays H → aa
are facilitated by constraints from the fact that the a bosons have the same mass
and their combination gives m H . The experimental challenges include soft decay
products of the a bosons, as well as merged a decays for small m a . The Jacobian
peak of the H → aa decay implies that most of the a bosons have p T of about
60 GeV or smaller, which is shared between the a decay products, making their
reconstruction challenging. For masses m a 15 GeV, these are separated in by
0.5 and smaller, resulting in insufficient selection efficiencies when using isolation
requirements or considering resolved decays only. Searches for H → aa have been
conducted by ATLAS and CMS using 8 TeV data, analysing the channels 4τ [1199,
1200], 4μ [1201], 2μ2τ [1200, 1202] and 2μ2b [1200]. A combination of 8 TeV
searches by CMS excludes B(H → aa) between 4 and 17%, depending on the specific model and the choice of parameters, for m a between 5 and 65 GeV [1200].
At 13 TeV, leptonic final states have been analysed by ATLAS in search for 4μ
decays [1203], and by CMS in search for 2μ2τ decays [1204, 1205], where collimated signatures are reconstructed with dedicated lepton reconstruction techniques.
An ATLAS analysis targets loop-induced decays in the 2γ 2g channel [1206]. The
analysis is carried out in the VBF production mode since it has a higher cross section
than VH production, but allows to suppress SM backgrounds through the presence
of forward jets. In order to reduce the contamination of pileup jets in the forward
region, forward jet vertex tagging is used [445]. This significantly improves the VBF
7 Note that this signature also arises in other models of new physics, for example two-Higgs-doublet
models with an additional pseudoscalar [1076].
5 Direct Searches for New Physics
Fig. 5.29 Distribution in the
photon subjet energy fraction
f γ for large-R jets with
p T > 200 GeV, three
reconstructed subjets and
τ 31 < 0.4. Taken from
[1188]
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
1
γ
f
0
0.05
0.1
0.15
0.2
0.25
Fraction of jets
Data
QCD simulation
1TeV
g
~
150 GeV, M
1
0
χ
∼
M
1TeV
g
~
200 GeV, M
1
0
χ
∼
M
1TeV
g
~
300 GeV, M
1
0
χ
∼
M
(13 TeV)
-1
35.9 fb
CMS
1194].
7 Current experimental constraints on non-SM branching fractions of the H
boson result in upper bounds of 19 at 95% CL [1195, 1196], which can be taken as
the upper bound for B(H → aa). Since the a boson mixes with the H boson, the
hierarchy of predicted branching fractions into SM particles is similar, with a → bb,
a → τ τ and a → μμ the largest, if kinematically accessible [1197]. In parameter
regions resulting in small fermionic couplings, the loop induced decays a → γ γ
and a → gg can be dominant [1198]. Searches at the LHC for the decays H → aa
are facilitated by constraints from the fact that the a bosons have the same mass
and their combination gives m H . The experimental challenges include soft decay
products of the a bosons, as well as merged a decays for small m a . The Jacobian
peak of the H → aa decay implies that most of the a bosons have p T of about
60 GeV or smaller, which is shared between the a decay products, making their
reconstruction challenging. For masses m a 15 GeV, these are separated in by
0.5 and smaller, resulting in insufficient selection efficiencies when using isolation
requirements or considering resolved decays only. Searches for H → aa have been
conducted by ATLAS and CMS using 8 TeV data, analysing the channels 4τ [1199,
1200], 4μ [1201], 2μ2τ [1200, 1202] and 2μ2b [1200]. A combination of 8 TeV
searches by CMS excludes B(H → aa) between 4 and 17%, depending on the specific model and the choice of parameters, for m a between 5 and 65 GeV [1200].
At 13 TeV, leptonic final states have been analysed by ATLAS in search for 4μ
decays [1203], and by CMS in search for 2μ2τ decays [1204, 1205], where collimated signatures are reconstructed with dedicated lepton reconstruction techniques.
An ATLAS analysis targets loop-induced decays in the 2γ 2g channel [1206]. The
analysis is carried out in the VBF production mode since it has a higher cross section
than VH production, but allows to suppress SM backgrounds through the presence
of forward jets. In order to reduce the contamination of pileup jets in the forward
region, forward jet vertex tagging is used [445]. This significantly improves the VBF
7 Note that this signature also arises in other models of new physics, for example two-Higgs-doublet
models with an additional pseudoscalar [1076].