ETD system

Electronic theses and dissertations repository


Tesi etd-09252018-193748

Thesis type
Tesi di laurea magistrale
email address
Search for e+e- -> mu+mu-Z'(Z'->Invisible) in the first data of the BelleII experiment
Corso di studi
relatore Forti, Francesco
Parole chiave
  • Dark Sector
  • particle physics
  • BelleII
Data inizio appello
Riassunto analitico
The thesis work concerns the search for the process e+e− -> µ+µ−Z' (Z' -> Invisible)
with the first data of the BelleII experiment, collected in the data taking period known
as Phase-2 (February 2018 - July 2018). This work includes also the development of
the calibration method of the algorithm used by the Silicon Vertex Detector (SVD) of
BelleII to estimate the hit time of particles crossing each sensor. The SVD is a detector
equipped with double-sided silicon strip sensors used for the reconstruction of charged
particle tracks. Together with the Pixel Detector (PXD), it constitutes the vertex detector
of the experiment. The BelleII experiment is installed on the SuperKEKB accelerator,
which is an asymmetric electron-positron collider, located on KEK laboratory, Tsukuba,
The Standard Model (SM) is the theory that currently describes the known particles and
their interactions. Recently some tensions between the theoretical predictions of the SM
and the experimental measurements have been observed. For example the measurement of
the anomalous magnetic moment of the muon (g −2)µ deviates from the SM prediction by
3.4σ, while the lepton flavor universality test performed by the LHCb, BABAR and Belle
experiments shows a deviation from the SM prediction of 4.1σ. These deviations can be
explained by extensions of the SM that predict a collection of hypothetical hidden particles
that could interact with SM particles, although very weakly, through new gauge bosons
called dark photons.
The purpose of this thesis is to investigate the production of a light dark gauge boson
Z0 in association with a muon pair in electron-positron annihilation at the center of mass
energy of 10.58 GeV. Through this analysis is possible to limit the Lµ − Lτ model, which
introduces the boson Z' by extending the symmetry group of the SM, U(1)Y ⊗ SU(2)L ⊗
SU(3)C, with the abelian group U(1)_{Lµ−Lτ} . This new symmetry introduces an interaction
term of the Z' boson with muons and talons and it preserves the difference between the
µ-leptonic number and the τ-leptonic number in the processes.
The Lµ − Lτ model is theoretically very well motivated and could explain some of the
observed experimental effects.
The process analyzed is e+e− -> µ+µ−Z';(Z' -> Invisible), where the muon pair
is produced at the Υ(4S) energy peak and the Z' boson is radiatively emitted by one
of the two muons. The signature of the process consists of two muon tracks coming
from the interaction point plus missing mass. The signal yield is extracted by fitting
the distribution of the recoil mass against the muon pair with respect to the center of
mass momentum, which is expected to peak at the Z' mass M_Z' for signal events. The
considered background processes are e+e− ! µ+µ−(γ) where the photon is not detected,
e+e− -> τ +τ− γ; τ -> µν and e+e− -> e+e−µ+µ− where the two electrons are undetected.
The recoil mass distribution of the first process peaks at small recoil masses, the second
dominates 2 < Mrecoil < 7 GeV/c2 and the last one dominates for Mrecoil > 7 GeV/c2.
The MC samples produced for the analysis consist of 10 thousand signal events for eight
different mass hypotheses and 10 fb−1 background events for each considered background
process. In order to reject background, a selection in two steps, based on the kinematics
of the events and on the particle identification, has been performed. The first selection
requires events with two muon track, i.e. two tracks characterized by a probability to be
identified as muon higher than 90%, coming from the interaction point and in which the
direction of the recoil points to the barrel region of the ECL without photons inside a cone
of 15 degrees from the direction of the recoil. Including a trigger efficiency of ∼82%, for M_Z' in the range [1,8] GeV/c2, the total efficiency of the preselection is ∼35%.
A further selection based on possible discriminant variables between signal events and
background events has been performed. Many variables are under study yet and currently
only the transverse momentum of the dimuon candidate pT µµ has been considered for the
background rejection. The selection been optimized for each Z' mass considered.
The range of Z' mass considered in the analysis goes from 1 GeV/c2 to 8 GeV/c2,
since the poor mass resolution for M_Z' . 1 GeV/c2 (the resolution for M_Z' ∼ 500MeV is
around 150 MeV) and the small cross section estimated for MZ0 & 8 GeV/c2 through the
MC simulation limit the sensitivity to signal events for MZ0 out of the considered range.
This analysis allowed to estimate the sensitivity region to the parameters of the Lµ−Lτ
model at a CL = 90%, assuming Z' -> ν_{µ;τ}ν_{µ;τ} and Z0 -> χχ, where χ is a dark matter
particle lighter than the Z' boson.
The sensitivity has been estimated for the following integrated luminosities: 20 fb−1,
which is the luminosity that was expected for Phase-2, 2 ab−1 and 50 ab−1, which is the full
data set expected for the end of the BelleII experiment. With an integrated luminosity of
20 fb−1, BelleII is already competitive with the other experiments, while with an integrated
luminosity of 50 fb−1 a wide part of the region of parameters that can explain the (g − 2)µ
anomaly is excluded.
Finally, a very preliminary comparison between the 505 pb−1 of data actually collected
during the Phase-2 and the produced MC samples has been performed.
In addition, a of the calibration method of the algorithm used by SVD to estimate the
hit time of particles that cross each sensor is presented, which is very important to be able
to use precise timing information to reject different background sources, events which are
off-time with respect to the event time, and ghost hits. The method consists of correcting
the SVD hit time estimation using the timing information provided by the CDC, which
are currently the only ones available, as reference. The final resolution on the hit time
obtained is of ∼4 ns for the V/n-side of the sensors.