Improved Measurement of the Electroweak Penguin Process B->Xsll

TL;DR

This paper presents an improved measurement of the inclusive B -> X_s l^+ l^- branching fraction using a semi-inclusive reconstruction that covers 18 hadronic X_s states with M_{X_s} < 2.0 GeV/c^2. By leveraging beam-energy constrained mass and energy difference variables and a likelihood-based candidate selection, the analysis suppresses backgrounds and extracts the signal yield via an extended unbinned maximum likelihood fit, carefully treating peaking backgrounds from charmonium and hadronic sources. The combined branching fraction for M_{ll} > 0.2 GeV/c^2 is Br(B -> X_s l^+ l^-) = (4.11 ± 0.83 (stat) +0.85 -0.81 (syst)) × 10^-6, with a significance of 5.4σ; differential distributions in M_{X_s} and q^2 are also provided. The result is consistent with SM predictions and prior measurements, and demonstrates the viability of semi-inclusive techniques to probe FCNC processes and constrain potential new physics in electroweak penguin loops.

Abstract

We present an improved measurement of the branching fraction for the electroweak penguin process B->Xsll, where l is an electron or a muon and Xs is a hadronic system containing an s-quark. The measurement is based on a sample of 152*10^6 Upsilon(4S) -> BB events collected with the Belle detector at the KEKB energy asymmetric e+e- collider. The Xs hadronic system is reconstructed from one K^{+-} or K^{0}_{S} and up to four pions, where at most one pion can be neutral. Averaging over both lepton flavors, the inclusive branching fraction is measured to be Br(B->Xsll)=(4.11+-0.83(stat)+0.85-0.81(syst))*10^{-6} for M_ll > 0.2GeV/c2.

Figures (5)

• Figure 1: Likelihood ratio for (a) signal and (b) background events. Points represent data and solid histograms represent MC. Samples used are described in the text.
• Figure 2: $M_{\rm bc}$ distributions for MC charmonium events for the (a) dielectron and (b) dimuon channels, and for $B \rightarrow X_s h^+h^-$ candidates in data for the (c) dielectron and (d) dimuon channels. The normalization corresponds to the expected charmonium and hadronic peaking background in $140~{\rm~fb}^{-1}$ of data.
• Figure 3: $M_{\rm bc}$ distributions of (a) $B\to X_s e^+e^-$, (b) $B\to X_s\mu^+\mu^-$, (c) $B\to X_s\ell^+\ell^-$ ($\ell = e, \mu$), and (d) $B\to X_s e^\pm\mu^\mp$ candidates in data. The solid lines represent the result of the fits, and dashed lines represent the sum of all background components under the signal peaks, respectively.
• Figure 4: Distributions for the signal yield as a function of (a) hadronic mass $M_{X_s}$ and (b) $q^2 \equiv M_{\ell^+\ell^-}^2$ for $B\to X_s\ell^+\ell^-$ signal. The points with error bars are the data (electron and muon channels combined) while the histogram is the MC signal normalized to the data.
• Figure 5: Differential branching fraction as a function of (a) hadronic mass $M_{X_s}$ and (b) $q^2 \equiv M_{\ell^+\ell^-}^2$ for $B\to X_s\ell^+\ell^-$ signal. The points and error bars represent the data (electron and muon channels combined) while the histograms represent the MC signal normalized to the data statistics. The outer (inner) error bars represent the total (statistical) errors.