Search for the Decays $B_{(s)}^0\to e^{+}{\mu }^{-}$ and $B_{(s)}^0\to e^{+}{e}^{-}$ in CDF Run II

We report results from a search for the lepton flavor violating decays $B_s^0\to e^{+}{\mu }^{-}$ and $B^0\to e^{+}{\mu }^{-}$, and the flavor-changing neutral-current decays $B_s^0\to e^{+}{e}^{-}$ and $B^0\to e^{+}{e}^{-}$. The analysis uses data corresponding to $2{\mathrm{fb}}^{-1}$ of integrated luminosity of $p\overline{p}$ collisions at $\sqrt{s}=1.96\mathrm{TeV}$ collected with the upgraded Collider Detector (CDF II) at the Fermilab Tevatron. The observed number of $B^0$ and $B_s^0$ candidates is consistent with background expectations. The resulting Bayesian upper limits on the branching ratios at 90% credibility level are $\mathcal{B}(B_s^0\to e^{+}{\mu }^{-})<2.0\times {10}^{-7}$, $\mathcal{B}(B^0\to e^{+}{\mu }^{-})<6.4\times {10}^{-8}$, $\mathcal{B}(B_s^0\to e^{+}{e}^{-})<2.8\times {10}^{-7}$, and $\mathcal{B}(B^0\to e^{+}{e}^{-})<8.3\times {10}^{-8}$. From the limits on $\mathcal{B}(B_{(s)}^0\to e^{+}{\mu }^{-})$, the following lower bounds on the Pati-Salam leptoquark masses are also derived: $M_{\mathrm{LQ}}(B_s^0\to e^{+}{\mu }^{-})>47.8\mathrm{TeV}/c^2$, and $M_{\mathrm{LQ}}(B^0\to e^{+}{\mu }^{-})>59.3\mathrm{TeV}/c^2$, at 90% credibility level.

Figure 1

Invariant mass distribution of $e^{+}{\mu }^{-}$ pairs for events where one track passed the electron identification and the other track the muon identification. The $B_s^0$ ($B^0$) search window is indicated by the solid line (short dashed line). The sideband regions are indicated by the dashed lines.

Figure 2

Invariant mass distributions of $e^{+}{e}^{-}$ pairs for events where both tracks passed the electron identification. The $B_s^0$ ($B^0$) search window is indicated by the solid line (short dashed line). The sideband regions are indicated by dashed lines.

Figure 3

Leptoquark mass limit corresponding to the 90(95)% C.L. on $\mathcal{B}(B_s^0\to e^{+}{\mu }^{-})$. The error band is obtained by varying the values entering the theoretical calculation within their uncertainties. The uncertainties stemming from approximating ${\alpha }_s$ are not included.