Search for the Radiative Decays $B\to \rho \gamma$ and $B^0\to \omega \gamma$

A search of the exclusive radiative decays $B\to \rho (770)\gamma$ and $B^0\to \omega (782)\gamma$ is performed on a sample of about $84\times {10}^6$ $B\overline{B}$ events collected by the BABAR detector at the SLAC PEP-II asymmetric-energy $e^{+}{e}^{-}$ storage ring. No significant signal is seen in any of the channels. We set upper limits on the branching fractions $\mathcal{B}$ of $\mathcal{B}(B^0\to {\rho }^0\gamma )<1.2\times {10}^{-6}$, $\mathcal{B}(B^{+}\to {\rho }^{+}\gamma )<2.1\times {10}^{-6}$, and $\mathcal{B}(B^0\to \omega \gamma )<1.0\times {10}^{-6}$ at 90% confidence level (C.L.). Using the assumption that $\Gamma (B\to \rho \gamma )=\Gamma (B^{+}\to {\rho }^{+}\gamma )=2\times \Gamma (B^0\to {\rho }^0\gamma )$, we find the combined limit $\mathcal{B}(B\to \rho \gamma )<1.9\times {10}^{-6}$, corresponding to $\mathcal{B}(B\to \rho \gamma )/\mathcal{B}(B\to K^{*}\gamma )<0.047$ at 90% C.L.

Figure 1

(a) Pion efficiency and kaon misidentification of the charged pion selection for $0.4\mathrm{GeV}/c<p_{\mathrm{lab}}<4.0\mathrm{GeV}/c$, where $p_{\mathrm{lab}}$ is the track momentum in the laboratory frame. (b) $B^0\to {\rho }^0\gamma$ neural-network output for MC-simulated events with comparison to data control samples. Events with neural-network output greater than 0.9 are selected, as indicated.

Figure 2

$\Delta E^{*}$ vs $m_{\mathrm{ES}}$ fit regions for (a) $B^0\to {\rho }^0\gamma$, (b) $B^{+}\to {\rho }^{+}\gamma$, and (c) $B^0\to \omega \gamma$ candidates. The boxes indicate the regions where signal events would appear: $-0.2<\Delta E^{*}<0.1\mathrm{GeV}$ and $5.27<m_{\mathrm{ES}}<5.29\mathrm{GeV}/c^2$. Assuming $\mathcal{B}(B^0\to {\rho }^0\gamma )=\frac{1}{2}\mathcal{B}(B^{+}\to {\rho }^{+}\gamma )=\mathcal{B}(B^0\to \omega \gamma )={10}^{-6}$, we expect 9.9, 12.1, and 3.4 signal events in these regions, respectively.