Study of <em>B → π ℓ ν</em> and <em>B → ρ ℓ ν</em> decays and extraction of |<em>V_ub</em>| at Belle II

Abstract

This thesis presents a measurement of the Cabibbo-Kobayashi-Maskawa (CKM) matrix element magnitude |<em>V_ub</em>| from a simultaneous study of the semileptonic decays <em>B⁰ → π^- ℓ⁺ ν_ℓ</em> and <em>B⁺ → ρ⁰ ℓ⁺ ν_ℓ</em>, where ℓ = e, μ. This analysis utilizes a data sample corresponding to 364 fb^-1 of integrated luminosity recorded by the Belle II detector at the SuperKEKB electron-positron collider between 2019 and 2022. The SuperKEKB collider targets a center-of-mass energy corresponding to the mass of the ϒ(4S) resonance. This results in nearly all produced ϒ(4S) mesons producing two <em>B</em> mesons, which in turn decay into lighter particles. In this thesis, we focus on an untagged measurement approach, where both decay modes are reconstructed without identifying the partner <em>B</em> mesons. After performing some signal selection steps, machine learning methods in the form of Boosted Decision Trees are trained and applied to suppress backgrounds. We simultaneously measure the differential branching fractions of <em>B⁰ → π^- ℓ⁺ ν_ℓ</em> and <em>B⁺ → ρ⁰ ℓ⁺ ν_ℓ</em> as functions of <em>q²</em> (momentum transfer squared). This is achieved by introducing a novel fit method, in which yields are inherently corrected for finite detector resolution. From these measurements, we determine the total branching fractions to be <em>B</em>(<em>B⁰ → π^- ℓ⁺ ν_ℓ</em>) = (1.516 ± 0.042 (stat) ± 0.059 (syst)) × 10^-4 and <em>B</em>(<em>B⁺ → ρ⁰ ℓ⁺ ν_ℓ</em>) = (1.625 ± 0.079 (stat) ± 0.180 (syst)) × 10^-4. Fitting the measured partial branching fractions of <em>B⁰ → π^- ℓ⁺ ν_ℓ</em> with constraints from lattice QCD calculations, we obtain |<em>V_ub</em>| = (3.93 ± 0.09 (stat) ± 0.13 (syst) ± 0.19 (theo)) × 10^-3. When we include additional LCSR constraints, the result is (3.73 ± 0.07 (stat) ± 0.07 (syst) ± 0.16 (theo)) × 10^-3. For <em>B⁺ → ρ⁰ ℓ⁺ ν_ℓ</em> decays, using light-cone sum rule constraints, we find |<em>V_ub</em>| = (3.19 ± 0.12 (stat) ± 0.17 (syst) ± 0.26 (theo)) × 10^-3. While the results are constrained by theoretical uncertainties, the leading systematic uncertainties result from the sizes of the off-resonance data set and the available simulated samples.