New measurements of ${\rm Y}(1S)$ decays to charmonium final states

Using data collected by the CLEO III detector at CESR, we report on measurements of ${\rm Y}(1S)$ decays to charmonium final states. The data sample used for this analysis consists of $21.2\times {10}^6$ ${\rm Y}(1S)$ decays, representing about 35 times more data than previous CLEO ${\rm Y}(1S)$ data samples. We present substantially improved measurements of the branching fraction $\mathcal{B}({\rm Y}(1S)\to J/\psi +X)$ using $J/\psi \to {\mu }^{+}{\mu }^{-}$ and $J/\psi \to e^{+}{e}^{-}$ decays. The branching fractions for these two modes are averaged, thereby obtaining: $\mathcal{B}({\rm Y}(1S)\to J/\psi +X)=(6.4\pm 0.4(\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{t})\pm 0.6(\mathrm{s}\mathrm{y}\mathrm{s}\mathrm{t}))\times {10}^{-4}$. A greatly improved measurement of the $J/\psi$ momentum distribution is presented and indicates a spectrum which is much softer than predicted by the color-octet model and somewhat softer than the color-singlet model. First measurements of the $J/\psi$ polarization and production angle are also presented. In addition, we report on the first observation of ${\rm Y}(1S)\to \psi (2S)+X$ and evidence for ${\rm Y}(1S)\to {\chi }_{cJ}+X$. Their branching fractions are measured relative to $\mathcal{B}({\rm Y}(1S)\to J/\psi +X)$ and are found to be $\{[\mathcal{B}({\rm Y}(1S)\to \psi (2S)+X)]/[\mathcal{B}({\rm Y}(1S)\to J/\psi +X)]\}=0.41\pm 0.11(\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{t})\pm 0.08(\mathrm{s}\mathrm{y}\mathrm{s}\mathrm{t})$, $\{[\mathcal{B}({\rm Y}(1S)\to {\chi }_{c1}+X)]/[\mathcal{B}({\rm Y}(1S)\to J/\psi +X)]\}=0.35\pm 0.08(\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{t})\pm 0.06(\mathrm{s}\mathrm{y}\mathrm{s}\mathrm{t})$, $\{[\mathcal{B}({\rm Y}(1S)\to {\chi }_{c2}+X)]/[\mathcal{B}({\rm Y}(1S)\to J/\psi +X)]\}=0.52\pm 0.12(\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{t})\pm 0.09(\mathrm{s}\mathrm{y}\mathrm{s}\mathrm{t})$, and $\{[\mathcal{B}({\rm Y}(1S)\to {\chi }_{c0}+X)]/[\mathcal{B}({\rm Y}(1S)\to J/\psi +X)]\}<7.4$ at 90% confidence level. The resulting feed-down contributions to $J/\psi$ are $[24\pm 6(\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{t})\pm 5(\mathrm{s}\mathrm{y}\mathrm{s}\mathrm{t})]\%$ for $\psi (2S)$, $[11\pm 3(\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{t})\pm 2(\mathrm{s}\mathrm{y}\mathrm{s}\mathrm{t})]\%$ for ${\chi }_{c1}$, $[10\pm 2(\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{t})\pm 2(\mathrm{s}\mathrm{y}\mathrm{s}\mathrm{t})]\%$ for ${\chi }_{c2}$, and $<8.2\%$ at 90% confidence level for ${\chi }_{c0}$. These measurements (apart from ${\chi }_{c0}$) are about a factor of 2 larger than expected based on the color-octet model.

Figure 1

Feynman diagrams for production of charmonium in ${\rm Y}(1S)$ decays from (a) color-octet processes and (b) color singlet ${\rm Y}(1S)\to J/\psi c\overline{c}g+X$. For the color-octet processes, the $J/\psi$ is produced in a color octet and becomes a color singlet through emission of a soft gluon.

Figure 2

Dilepton invariant mass distributions for (a) $J/\psi \to {\mu }^{+}{\mu }^{-}$ and (b) $J/\psi \to e^{+}{e}^{-}$ candidates for data taken on the ${\rm Y}(1S)$ resonance (points) and data taken just below the ${\rm Y}(4S)$ resonance (shaded). The ${\rm Y}(4S)$ distributions are scaled to account for the different integrated luminosities and center-of-mass energies for the two data samples.

Figure 3

Invariant mass distributions for $J/\psi \to {\mu }^{+}{\mu }^{-}$ candidates in $x$ bins of size 0.2 for data taken on the ${\rm Y}(1S)$ resonance.

Figure 4

Invariant mass distributions for $J/\psi \to e^{+}{e}^{-}$ candidates in $x$ bins of size 0.2 for data taken on the ${\rm Y}(1S)$ resonance.

Figure 5

Invariant mass distribution for (a) $J/\psi \to {\mu }^{+}{\mu }^{-}$ and (b) $J/\psi \to e^{+}{e}^{-}$ in data taken below the ${\rm Y}(4S)$. The data are integrated over all momenta.

Figure 6

Invariant mass distributions for (a) $J/\psi \to {\mu }^{+}{\mu }^{-}$ and (b) $J/\psi \to e^{+}{e}^{-}$ in data taken on the ${\rm Y}(4S)$ resonance. To reject $J/\psi$’s from $B$ decay, we require the momentum of the $J/\psi$ to be larger than $2.0\mathrm{G}\mathrm{e}\mathrm{V}/c$.

Figure 7

Efficiency for reconstructing $J/\psi$’s in ${\rm Y}(1S)$ decays as a function of (a) scaled $J/\psi$ momentum, and (b) $\cos {\theta }_{J/\psi }$. The circles are for $J/\psi \to {\mu }^{+}{\mu }^{-}$ and the triangles are for $J/\psi \to e^{+}{e}^{-}$.

Figure 8

Distributions in $x$ for $e^{+}{e}^{-}\to J/\psi +X$ using the combined data taken on and just below the ${\rm Y}(4S)$. The circles show the results obtained using $J/\psi \to {\mu }^{+}{\mu }^{-}$ and the triangles show the corresponding distribution obtained using $J/\psi \to e^{+}{e}^{-}$.

Figure 9

Differential cross sections in $x$ for data taken on the ${\rm Y}(1S)$ (points) and data taken both on and just below the ${\rm Y}(4S)$ (triangles). The latter have been scaled to account for the ${\rm Y}(1S)\to {\gamma }^{*}\to q\overline{q}$ contribution, as discussed in the text. The upper figure shows the results obtained using $J/\psi \to {\mu }^{+}{\mu }^{-}$ and the lower figure shows the corresponding distributions obtained using $J/\psi \to e^{+}{e}^{-}$.

Figure 10

Differential cross sections in $x$ for ${\rm Y}(1S)\to J/\psi +X$ obtained using $J/\psi \to {\mu }^{+}{\mu }^{-}$ (circles) and $J/\psi \to e^{+}{e}^{-}$ (triangles). We also show the theoretical expectations based on the color-octet (solid line) [15] and color-singlet (dashed line) [13] models.

Figure 11

The helicity angular distributions, $\cos {\theta }_{\mathrm{h}\mathrm{e}\mathrm{l}}$ (top panel) and production angle, $\cos {\theta }_{J/\psi }$ (bottom panel) of the $J/\psi$ in ${\rm Y}(1S)\to J/\psi +X$. In each case, the points are the ${\rm Y}(1S)$ data, the dashed histogram is a pythia simulation of ${\rm Y}(1S)\to J/\psi +X$, and the dotted line is a fit to the ${\rm Y}(1S)$ data as described in the text.

Figure 12

Sideband-subtracted distributions of (a) number of charged tracks, $N_{\mathrm{t}\mathrm{r}\mathrm{k}}$, (b) neutral energy, $E_{\mathrm{N}\mathrm{E}\mathrm{U}}$, (c) the second Fox-Wolfram moment, $R_2$, and (d) the recoil mass, $M_{\mathrm{R}\mathrm{E}\mathrm{C}\mathrm{O}\mathrm{I}\mathrm{L}}$, for ${\rm Y}(1S)$ data and pythia simulation. The points (triangles) correspond to $J/\psi \to {\mu }^{+}{\mu }^{-}$ ($J/\psi \to e^{+}{e}^{-}$), and the histograms are the corresponding distributions from the simulation.

Figure 13

Invariant mass difference $M(l^{+}{l}^{-}{\pi }^{+}{\pi }^{-})-M(l^{+}{l}^{-})$ for both $J/\psi \to {\mu }^{+}{\mu }^{-}$ and $J/\psi \to e^{+}{e}^{-}$ candidates with invariant mass in the range from $3.00–3.14\mathrm{G}\mathrm{e}\mathrm{V}/c^2$.

Figure 14

Event-level distributions for events rejected in the $\psi (2S)$ analysis. Distributions shown are (a) number of charged tracks, (b) neutral energy, (c) missing event momentum, and (d) cosine of the $\psi (2S)$ production angle. Solid points are rejected ${\rm Y}(1S)\to \psi (2S)+X$ candidate events and the histogram is the $e^{+}{e}^{-}\to \psi (2S)+X$ radiative return simulation.

Figure 15

Event-level distributions for events accepted in the $\psi (2S)$ analysis. Distributions shown are (a) number of charged tracks, (b) neutral energy, (c) missing event momentum, and (d) cosine of the $\psi (2S)$ production angle. Solid points are accepted ${\rm Y}(1S)\to \psi (2S)+X$ candidate events and the histogram is a ${\rm Y}(1S)\to \psi (2S)D\overline{D}$ MC simulation.

Figure 16

Difference of invariant masses, $M(l^{+}{l}^{-}\gamma )-M(l^{+}{l}^{-})$, for $M(l^{+}{l}^{-})$ in the range from 3.00 to $3.14\mathrm{G}\mathrm{e}\mathrm{V}/c^2$ for $l=\mu ,e$ combined. The solid histogram represents the ${\rm Y}(1S)$ data and the shaded histogram is the below-${\rm Y}(4S)$ data scaled by the ratio of luminosities. The arrows indicate the mass differences corresponding to the three ${\chi }_c$ states.