Titolo della tesi: Searches for Heavy Neutral Leptons in B Meson Decays with the CMS Experiment
The Standard Model(SM) of particle physics is the most accurate scientific theory ever conceived. It describes the fundamental constituents of the Universe and their interactions with astounding precision, as its predictions have been confirmed by countless experiments. However, the model is flawed with theoretical drawbacks and evidence of unpredicted phenomena. Evidence for neutrino flavor oscillations are among the most surprising hints to physics beyond the SM, as they provide compelling evidence for the neutral leptons to be mass-given particles; the SM does foresee neutrino dynamics, yet it does not account for neutrino masses. Several SM extensions accounting for neutrino masses in multiple scenarios have been developed: the nuMSM provides a minimal extension to the model through the addition of right-handed heavy neutral leptons.
This thesis reports the first search for Majorana heavy neutral leptons in B meson decays with the CMS experiment. It is made possible by the 2018 B-Parking dataset, a sample of 10^10 B\bar{B} events collected with novel trigger and data-taking strategies.
The analysis targets heavy neutral leptons with masses below the mass of the B meson, mB = 5.27 GeV, and spans a range of neutrino lifetimes ctau = [0.0001,1] m. The targeted signature is formed by three low-energy objects, two leptons and a pion. One of the lepton and the pion originate from the heavy neutrino possibly displaced vertex: the analysis configures as a search for a peak in the invariant mass spectrum of the heavy neutral lepton decay products.
The main challenge of the analysis lies in the reconstruction and identification of low-energy displaced objects, as CMS is optimized for the reconstruction of high-energy events originating in the interaction region. Novel identification strategies and an event categorization depending on displacement-related variables are used to enhance the analysis sensitivity. Parametric deep-learning discriminators are used, in the analysis categories, in order to ensure optimal signal efficiency over the range of investigated neutrino masses while rejecting the background.
The expected results for the heavy neutrino coupling |V|^2 to the SM neutrinos are extracted for several flavor coupling scenarios thanks to the combination of two channels containing two muons or a muon and an electron in the final state. The analysis is currently blinded. The analysis reaches a sensitivity, expressed as the best 95% exclusion limit on the total coupling |V|^2, which ranges between 5 10^{-5} for a mass hypothesis of 2 GeV to 1.66 10^{-4} for mass 3 GeV.