Jonathan Wolfgang Grumke ; Jochen Dingfelder ; Peter Lewis

2021
Physikalisches Institut Bonn

Abstract: Using the FEI algorithm with simulated data for hadronic tagging in the reconstruction of $B^0\to\pi^-\ell^+\nu_\ell$ results in a signal yield of $626\pm 39$ when applying the same selections as Sibidanov et al., who used the FR algorithm and obtained a signal yield of $463\pm 28$. However, this increase in efficiency does not result in an increase of significance when performing a fit to the $m^2_{\textrm{miss}}$ distribution. Sibidanov et al. obtained a significance of $16.7$ from a fit to experimental data, while an Asimov fit to simulated data reconstructed with the FEI yields a significance of $16.0$. By relaxing the event selection I was able to further increase the signal efficiency and obtain a signal yield of $754\pm44$ events. This also results in slight gain in significance of $17.2$. That the gain in efficiency by using the FEI does not result in a higher significance can be attributed to the worse resolution in $m^2_{\textrm{miss}}$ by the FEI compared to the FR. Additional decay channels that are reconstructed on the tag-side by the FEI are responsible for the gain in efficiency. Some of these new channels involve the reconstruction of $\pi^0$ and lead to a poor resolution in the reconstruction of the $B_{\textrm{tag}}$ four-momentum. This again leads to a poor resolution of $m^2_{\textrm{miss}}$. In order to still benefit from these lower quality $B_{\textrm{tag}}$ modes one could individually optimize the FEI signal probability selection for each decay channel reconstructed on the tag-side. One could choose a loose selection in high quality channels and therefore achieve a good $m^2_{\textrm{miss}}$ resolution, while retaining a high signal efficiency. In low quality channels one could choose a tighter selection on the FEI signal probability to still use these additional channels, which the FR does not reconstruct, while only keeping the candidates with a high kinematic resolution on the tag-side.

Note: Presented on 25 08 2021
Note: MSc