Search for the Elusive Higgs Boson Using Jet Structure at LHC

Abstract

We consider the production of a light non-standard model Higgs boson of order $100~\GEV$ with an associated $W$ boson at CERN Large Hadron Collider. We focus on an interesting scenario that, the Higgs boson decays predominately into two light scalars $\chi$ with mass of few GeV which sequently decay into four gluons, i.e. $h\to 2\chi \to 4g$. Since $\chi$ is much lighter than the Higgs boson, it will be highly boosted and its decay products, the two gluons, will move close to each other, resulting in a single jet for $\chi$ decay in the detector. By using electromagnetic calorimeter-based and jet substructure analyses, we show in two cases of different $\chi$ masses that it is quite promising to extract the signal of Higgs boson out of large QCD background. Comment: 20 pages, 7 figures

Full text

Page 1

arXiv:1006.1151v1 [hep-ph] 6 Jun 2010
Preprint typeset in JHEP style - HYPER VERSION
KEK-TH-1369
IPMU 10-0091
Search for the Elusive Higgs Boson Using Jet
Structure at LHC
Chuan-Ren Chen
Institute for the Physics and Mathematics of the Universe,
The University of Tokyo, Chiba, 277-8583, Japan
Mihoko M. Nojiri
Institute for the Physics and Mathematics of the Universe,
The University of Tokyo, Chiba, 277-8583, Japan
Theory Group, KEK,1-1 Oho, Tsukuba, Ibaraki 305-0801, Japan
The Graduate University for Advanced Studies (Sokendai),
1-1 Oho, Tsukuba, Ibaraki 305-0801, Japan
Warintorn Sreethawong
Theory Group, KEK,1-1 Oho, Tsukuba, Ibaraki 305-0801, Japan
The Graduate University for Advanced Studies (Sokendai),
1-1 Oho, Tsukuba, Ibaraki 305-0801, Japan
Abstract:
We consider the production of a light non-standard model Higgs boson of order
100 GeV with an associated W boson at CERN Large Hadron Collider. We focus
on an interesting scenario that, the Higgs boson decays predominately into two light
scalars η with mass of few GeV which sequently decay into four gluons, i.e. h →
2η → 4g. Since η is much lighter than the Higgs boson, it will be highly boosted
and its decay products, the two gluons, will move close to each other, resulting in
a single jet for η decay in the detector. By using electromagnetic calorimeter-based
and jet substructure analyses, we show in two cases of different η masses that it is
quite promising to extract the signal of Higgs boson out of large QCD background.
Keywords: Collider Phenomenology, Higgs Boson Decay, Jet Structure.

Page 2

Contents
1. Introduction1
2. Higgs Decay h → 2η → 4g at the LHC
2.1 A Reference Model
2.2 Signal and Background Generations
3
3
4
3. Narrow Jet and mη= 4 GeV Case5
4. Jet Substructure and mη= 8 GeV Case10
5. Discussions and Conclusions 14
A. Appendix 15
1. Introduction
Even though the Standard Model (SM) successfully explains almost all of the data
in collider experiments, the Higgs boson, which is associated with the electroweak
symmetry breaking mechanism in the SM, has not yet been found. The current
limits on the Higgs boson mass, mh, come from the direct searches at LEP II and
Tevatron, where a lower bound of 114 GeV has been obtained and a window of
162 GeV < mh< 166 GeV has been excluded, respectively, at 95% confidence level
(C.L.) [1, 2]. Combined with the direct search limit, the global fits to electroweak
precision data prefer a light Higgs boson, mh ? 191 GeV at 95% C.L. [3]. The
Large Hadron Collider (LHC) at CERN in-operation is expected to discover the
SM Higgs boson with mass up to 1 TeV, if such a particle exists. At the LHC,
the Higgs boson is predominately produced in the gluon-gluon fusion process via a
top quark loop. Other important production processes are vector-boson fusion and
production of the Higgs boson with an associated W/Z boson (associated hW/hZ
production). The search strategy for the Higgs boson depends on its mass. When
the Higgs boson is heavy (mh? 135 GeV), the channels of h → WW(∗)→ ℓ+
and h → ZZ → ℓ+
GeV), h → γγ is the main search channel, while the dominant decay channel h → b¯b
suffers from a very large QCD background.
1ℓ−
2ν1¯ ν2
1ℓ−
1ℓ+
2ℓ−
2are very promising. For a light Higgs boson (mh? 135
– 1 –

Page 3

Recently, the associated Zh/Wh production combined with h → b¯b search, which
has been considered less promising, was recovered by focusing in the regime where
the Higgs boson is highly boosted and employing a jet substructure technique, and
it was reported that the significance about 4σ can be achieved for 30 fb−1of an
integrated luminosity at the LHC [4, 5]. The idea is that when a heavy particle is
highly boosted, its decay products will be emitted almost collinearly so that they
are likely to be merged into a single jet rather than appearing as separate jets. Such
jet should have internal structures corresponding to decay products, including their
soft/collinear radiations, and the jet mass should peak around the mass scale of
the parent particle. Applications of jet substructure method for particle searches
and reconstructions also include searches for strongly interacting W bosons [6] and
supersymmetric particles [7].
Furthermore, in spite of the great agreement between the SM predictions and
experimental results, it is widely believed that the SM is only an effective theory at
the weak scale due to, for instance, the so-called ”hierarchy problem” and lack of
a dark matter candidate in the SM. Many new models beyond the SM have been
proposed over past decades, and in most of them the Higgs sector is extended. It
should be emphasized that the current bound on the Higgs boson mass cannot be
directly applied to such models especially when properties of the Higgs sector are
significantly modified. For example, the LEP lower bound is based on studies of the
associated Zh production with h → b¯b. Therefore, the limit can be relaxed if the
production cross section, or equivalently an effective coupling gZZh, is suppressed.
Another interesting possibility is that the Higgs boson is hidden, i.e. it has non-SM
decay modes and escapes from detection even though it might have been produced at
LEP and Tevatron. Particularly, if the decay into not-yet-discovered light particles
dominates, the direct search strategies have to be revised to capture the nature of
new decay topologies. Examples of scenario are the extensions of Minimal Supersym-
metric Standard Model (MSSM) including an additional singlet [8, 9] or CP-violating
interactions [10], the little Higgs models [11] and supersymmetric little Higgs models
[12, 13]. In most of these models, the Higgs boson dominantly decays into a pair of
new light scalars η which subsequently decay into two or more visible SM particles.
The phenomenology of such non-SM decays has been studied both at the Tevatron
[14, 15] and at the LHC [9, 15, 16]. The most commonly studied cascade decays are
h → 2η → 4b/4τ/2b2τ.
In this paper, we study a scenario in which a light Higgs boson (? 135 GeV)
decays predominately into a pair of very light pseudoscalars, η, ( mη< 2mb) each of
which then decays into two gluons, i.e. h → 2η → 4g. A model which predicts such
a decay channel will be briefly described later in Section 2.1. However, we stress
that we do not restrict ourselves to a particular model since we treat the branching
ratios of h → 2η and η → 2g as free parameters. This channel will encounter the
large QCD background, we therefore focus on the associated Wh production process
– 2 –

Page 4

where W boson decays leptonically in order to suppress the background. Since the
mass splitting between Higgs boson and η particle is large, two pseudoscalars will be
highly boosted. An opening angle between two gluons from each η decay, which is
also limited by the small mass of η, will be so small that they are closely aligned and
typically present in a single jet in the detector. The collider signature is essentially
two central non-b jets, an isolated charged lepton and missing momentum. We will
show that the signal for a relatively light η can be feasibly discovered at the LHC by
imposing cuts on electromagnetic-calorimeter-based (E-cal-based) variables, together
with some kinematic cuts. For the case of heavier η, the angular separation between
daughter gluons from a pseudoscalar decay is larger, so the E-cal-based analysis
fails. However, by employing the jet substructure technique previously mentioned,
we demonstrate that such a scenario still can be discovered. For illustration, we
adopt mh= 120 GeV, 100% decay branching ratio for h → 2η and η → 2g, and two
masses for η: mη= 4 GeV (light) and mη= 8 GeV (heavy).
The paper is organized as follows. In the next section, we firstly give a brief
review on the model of elusive Higgs boson.
generations are described. In Section 3, we present a search strategy using narrow
jet at the LHC for the case mη= 4 GeV. Jet substructure method and its application
to the case mη= 8 GeV are presented in Section 4. Lastly, Section 5 is devoted to
discussions and conclusions.
Afterward, signal and background
2. Higgs Decay h → 2η → 4g at the LHC
2.1 A Reference Model
For models beyond the SM, it is not unusual that a light singlet scalar appears as
a result in the extended Higgs sector. If there exists a coupling between the Higgs
boson and the light scalar, the Higgs may predominately decay into a pair of such
light scalars. The dominant decay channel of the light scalar then depends on its mass
and couplings to SM fermions. Here we briefly review a supersymmetric extension
of the simplest Little Higgs model [12], in which the light scalar will mainly decay
into a gluon pair when its mass is lighter than twice the bottom quark mass, which
is the scenario we focus on in this paper. We only introduce the phenomenology in
Higgs sector and we refer readers to the original paper for the details of the model.
The model is a supersymmetric extension of the simplest Little Higgs model [12].
The gauge group is SU(3)C× SU(3)W × U(1)X, where the SU(3)W × U(1)X is
broken by two vector-like sets of Higgs superfields, Φu,dand Hu,d. The misalignment
between the vacuum expectation values (VEVs) of these two Higgs fields leads to
electroweak symmetry breaking. There exists an approximate SU(3)1×SU(3)2global
symmetry due to the absence of the cross terms between Φ and H. The VEV of Φ
at the order of 10 TeV breaks the SU(3)1 and the gauge group down to the SM
– 3 –

Page 5

SU(3)C× SU(2)W× U(1)Y, generating five massive gauge bosons. The VEV of H,
which is generated radiatively with an order of few hundreds GeV, breaks another
set of SU(3) × U(1) → SU(2) × U(1) and produces five Goldstone bosons, three
of which are eaten by W and Z bosons after electroweak symmetry breaking, and
the two remnants are identified as Higgs boson h and light pseudoscalar η. The
coupling between h and η is generated due to the fact that h lives partly in the third
component of Hu,d, where η mainly stays. The decay widths of the Higgs boson into
a pair of η’s and the SM fermions f are given as [12]
Γh→ηη≈
1
64π
?
?
1 −v2
ew
v2
H
?
?−1m3
hv2
v4
H
ew
,(2.1)
Γh→f¯f≈1 −v2
ew
v2
H
ΓSM
h→f¯f, (2.2)
where vew = 174 GeV is the electroweak scale related to the Higgs VEV, vH ≈
300 ∼ 500 GeV and ΓSM
pair. After electroweak symmetry breaking, η acquires mass naturally of order few
to O(10) GeV. The coupling between the singlet η to the SM fermion f is induced by
the mixing of the SM fermion and its heavy partner f′, and is suppressed by a factor
of m2
considered. As a result, η will predominately decay into two gluons via bottom and
top quark loops when η → b¯b is kinematically forbidden.
In summary, we review in this section a reference model in which the light Higgs
boson mainly decays into two singlet light scalars which sequently decay into four
gluons, i.e. h → 2η → 4g. In our numerical study, we take 100% branching ratio for
both h → 2η and η → 2g for mh= 120 GeV and mη< 2mb, since we only focus
on this non-SM Higgs decay scenario and will explore it in a model independent
approach. For any particular model which predicts such a scenario, our result can
be easily applied by rescaling the production cross section of the Higgs boson and
branching ratios.
h→f¯fis the decay width of the SM Higgs boson into a fermion
f/m2
f′. Therefore, only the couplings of bottom and top quarks to η should be
2.2 Signal and Background Generations
For signal event generation, we employed the Herwig 6.5 Monte Carlo generator [17]
at the LHC with√s = 14 TeV center-of-mass energy. We generated the heavier
neutral MSSM Higgs boson in association with a W boson production and forced
the Higgs boson to decay into two pseudoscalars each of which subsequently decays
into two gluons. All backgrounds were generated by Alpgen 2.13 [18] with showering
and hadronization by Herwig. In our analysis, we particularly forced W → eν
for both signal and background processes, therefore the final statistics should be
twice our result when W → µν is also considered. The leading order cross sections
incorporating branching fraction for signal and background processes are listed in
Table 1. Note that the signal cross section is normalized to the SM value. We
– 4 –