The process $\gamma \gamma \to \varphi J/\psi$ is measured using a data sample of $825\mathrm{f}{\mathrm{b}}^{-1}$ collected with the Belle detector. A narrow peak of ${8.8}_{-3.2}^{+4.2}$ events, with a significance of 3.2 standard deviations including systematic uncertainty, is observed. The mass and natural width of the structure [named $X(4350)$] are measured to be $[{4350.6}_{-5.1}^{+4.6}(\mathrm{stat})\pm 0.7(\mathrm{syst})]\mathrm{MeV}/c^2$ and $[{13}_{-9}^{+18}(\mathrm{stat})\pm 4(\mathrm{syst})]\mathrm{MeV}$, respectively. The product of its two-photon decay width and branching fraction to $\varphi J/\psi$ is $[{6.7}_{-2.4}^{+3.2}(\mathrm{stat})\pm 1.1(\mathrm{syst})]\mathrm{eV}$ for $J^P=0^{+}$, or $[{1.5}_{-0.6}^{+0.7}(\mathrm{stat})\pm 0.3(\mathrm{syst})]\mathrm{eV}$ for $J^P=2^{+}$. No signal for the $Y(4140)\to \varphi J/\psi$ structure reported by the CDF Collaboration in $B\to K^{+}\varphi J/\psi$ decays is observed, and limits of ${\Gamma }_{\gamma \gamma }\mathbf{(}Y(4140)\mathbf{)}\mathcal{B}\mathbf{(}Y(4140)\to \varphi J/\psi \mathbf{)}<41\mathrm{eV}$ for $J^P=0^{+}$ or $<6.0\mathrm{eV}$ for $J^P=2^{+}$ are determined at the 90% C.L. This disfavors the scenario in which the $Y(4140)$ is a $D_s^{*+}{D}_s^{*-}$ molecule.

• Received 11 December 2009

DOI:https://doi.org/10.1103/PhysRevLett.104.112004