Measurement of the Inclusive Branching-Fraction Ratio \({{B}^{ - }} \to {{D}^{{*0}}}{{\pi }^{ - }}\) to \({{B}^{ - }} \to {{D}^{0}}{{\pi }^{ - }}\) at Belle and Belle II

Abstract

We report the status of an ongoing measurement of the branching-fraction ratio \(\mathcal{B}\)(B^– → \({{D}^{{*0}}}{{\pi }^{ - }}){\text{/}}\mathcal{B}({{B}^{ - }} \to {{D}^{0}}{{\pi }^{ - }})\) using \(711{\kern 1pt} \,{\text{f}}{{{\text{b}}}^{{ - 1}}}\) of Belle data and \(362\,{\kern 1pt} {\text{f}}{{{\text{b}}}^{{ - 1}}}\) of Belle II data collected at the \(\Upsilon (4S)\) resonance. The signal yields are extracted with an inclusive recoil-mass method: one B meson (the tag-B) is fully reconstructed in hadronic modes, and only the charged pion from the signal B is reconstructed. A deep neural-network classifier is used for background suppression. The invariant mass of the remaining unreconstructed particles, referred to as the recoil mass, is calculated using the known initial \({{e}^{ + }}{{e}^{ - }}\) state and the reconstructed tag-B and charged pion, and is fit to extract the \(D\pi \) and \(D{\text{*}}\pi \) yields. The fit uses separate signal shapes for correctly and incorrectly reconstructed tag-\(B\) candidates. The branching-fraction ratio remains blinded while the analysis procedure is finalized; we obtain a statistical uncertainty of \(1.7\% \). For comparison, the world-average PDG relative uncertainties are 2.2% for \({{B}^{ - }} \to {{D}^{0}}{{\pi }^{ - }}\) and \(2.9\% \) for \({{B}^{ - }} \to {{D}^{{*0}}}{{\pi }^{ - }}\). The \({{B}^{ - }} \to {{D}^{{(*)0}}}{{\pi }^{ - }}\) fits also determine tag-mode-dependent corrections to the \({{B}_{{{\text{tag}}}}}\) reconstruction efficiency and good-tag fractions. These corrections are used in the inclusive measurement of the branching fraction of \({{B}^{ - }} \to X(3872){{K}^{ - }}\).