Search for Light CP-odd Higgs in Radiative Decays of Upsilon(1S)

Abstract

We search for a non-SM-like CP-odd Higgs boson (a0_1) with m(a0_1)< 2m(b) in radiative decays of the Upsilon(1S), using 21.5M Upsilon(1S) mesons directly produced in e+e- annihilation. We investigate a0_1 --> tau+tau- and a0_1 --> mu+mu- decay channels. No significant signal is found. We obtain upper limits on the product of B(Upsilon(1S)-->gamma a0_1) and B(a0_1-->tau+tau-) or B(a0_1-->mu+mu-). Our tau+tau- results are almost two orders of magnitude more stringent than previous upper limits. Our data provide no evidence for a Higgs state with a mass of 214 MeV decaying to mu+mu-. Existence of such a state was previously proposed as an explanation for 3 Sigma+ --> p mu+mu- events, having mu+mu- masses just above the kinematic threshold, observed by the HyperCP experiment. Our results constrain NMSSM models.

Figures (5)

• Figure 1: Photon energy and dimuon mass distributions in $\gamma\tau^+\tau^-$ (a,c) and $\gamma\mu^+\mu^-$ (b,d) data, respectively. Bin size for the right column plots is given in the axes labels in parentheses. In the top row, the $\Upsilon$Υ$(1S)$ data (points with error bars) are compared to the estimated backgrounds (dashed and solid lines). In the bottom row, the $\Upsilon$Υ$(1S)$ data (solid line) are shown in fine binning comparable to the detector resolution (see bottom row of Fig. \ref{['fig:llmc']}). In (b) the $J/\psi$ ISR peak is shifted in the background estimate, since we scaled $\mu$ momenta down by the ratio of the beam energies when scaling the higher energy data to the $\Upsilon$Υ$(1S)$ distribution.
• Figure 2: Efficiency (a,b) and $a^0_1$ mass resolution (c,d) obtained from the fits to the $a^0_1\to\tau^+\tau^-$ (left column) and $a^0_1\to\mu^+\mu^-$ (right column) signal MC (points) and interpolated for the regions in between (solid line). In (d) relative dimuon mass resolution was multiplied by a factor of 10. See Appendix C of Ref. ManganoNason for explanation of improvement of the dimuon mass resolution near the kinematic threshold. The hollow point with the error bar in (d) represents the fit of the mass resolution to the $J/\psi\to\mu^+\mu^-$ ISR peak observed in the $\Upsilon$Υ$(1S)$ data.
• Figure 3: Upper limits on ${\cal B}(\Upsilon$Υ$(1S)\to\gamma a^0_1)$ (a) $\times{\cal B}(a^0_1\to \mu^+\mu^-)$ (b) $\times{\cal B}(a^0_1\to \tau^+\tau^-)$ as a function of the $a^0_1$ mass (90% C.L.). The color coding corresponds to the one used in Fig. \ref{['fig:nmssm']}. The dashed line indicates the region ($m_{a^0_1}>9.2$ GeV) where $a^0_1$ is likely to mix with $\eta_b$ and acquire a non-negligible width, thus invalidating our analysis method.
• Figure 4: Comparison of CLEO upper limits on ${\cal B}(\Upsilon$Υ$(1S)\to\gamma a^0_1)\times{\cal B}(a^0_1\to\tau^+\tau^-)$ (solid and dashed lines) to the NMSSM predictions by Dermisek, Gunion, McElrath (points) Dermisek. See the text for explanations.
• Figure 5: A fit of a peak at a dimuon mass of 214.3 MeV with fixed width at the expected mass resolution, on top of cubic polynomial to our $\gamma\mu^+\mu^-$ data obtained at the $\Upsilon$Υ$(1S)$ center-of-mass energy (a). The polynomial describing the continuum backgrounds was simultaneously constrained by the data collected at and near the $\Upsilon$Υ$(4S)$ resonance (b).