A beautiful paper by CLEO-c: J/psi → γX
CLEO-c have released a beautiful paper today on the search for narrow invisible states in decays of the J/ψ meson (arXiv:1003.0417):
Search for the Decay J/ψ → γ+Invisible
I have always liked this analysis and I am very happy to see this update.
The idea is simple: a light gauge boson or fundamental scalar particle might appear in the decays of the J/ψ mesons, which consists essentially of a charm and anti-charm quark in a bound state. If the decay also involves a photon, then the energy of that photon (Eγ) depends linearly on the mass of the new particle, X, and on the J/ψ mass, which is very well known. A plot of Eγ from J/ψ decays will display a peak above a smooth background – the smoking gun of this beyond-the-standard-model decay. Thus one could discover a new state without reconstructing it explicitly.
But how do you reconstruct the J/ψ if you don’t reconstruct its decay products? You have to infer the J/ψ final state through a kinematic tag. In the case of the CLEO-c analysis, they look for the decay ψ(2s) → π+π–, produced on top of the resonance e+e– → ψ(2s). By virtue of the initial state, the 4-momentum of the final state is known a priori. The two “tagging” pions are well measured, so the 4-momentum of the J/ψ is determined without any knowledge of its decay. Here is the result of this kinematical tagging:
I find this peak amazing – it has a perfect Breit-Wigner shape indicating excellent kinematic resolution. Very impressive – and a great reminder of the power of analysis techniques in an e+e– machine.
In order to search for invisible particles, you need to reject events with any activity beyond what is expected in the signal: the two tagging pions, and the photon. Sometimes a pion hitting a calorimeter can spit out debris back into the tracking volume, much like an asteroid hitting the surface of the moon. This debris can be reconstructed erroneously as a photon, which would spoil this analysis. The authors take care to set aside such fake photons on the basis of the shape of the energy deposit in the calorimeter, and by avoiding photon candidates that are simply too close to the tagging pions.
The final sample contains 73 events with Eγ > 1.25 GeV. Here is the spectrum:
The top plot is for the signal. The simulation, represented by the solid line, significantly underestimates the number of such events. According to the simulation, these events are caused by decays J/ψ to a neutron + anti-neutron pair, followed by the anti-neutron annihilating in the calorimeter and producing a fake photon. The authors are able to verify this hypothesis by preferentially selecting the kind of broad showers expected from the anti-neutron – this is the bottom plot. It is not a big surprise that the simulation does not get the level of this obscure background right – in fact, I am impressed that the shape of the spectrum is well reproduced by the simulation.
No narrow peak is seen in the spectrum, so there is no signal for J/ψ → γ X. Limits are computed as a function of the mass of the invisible object X. The result is determined mainly by statistics; the systematic uncertainties associated with the background description, etc., have a minor impact, which the authors took into account in a simple and pessimistic way. Overall, the upper limit on the branching ratio for J/ψ → γX is about 6.3×10-6. Here is the plot:
These limits constrain models with extended Higgs sectors. By combining with published limits coming from studies of Υ decays, the authors infer a bound on the mixing angle between the invisible scalar X and the standard model Higgs boson: cos²θ < 0.033. Thus, any such mixing must be small – a strong constraint on such models. Theses results also constrain models with a light gauge boson U.
This analysis is powerful and elegant, based almost entirely on in situ determinations of efficiencies and backgrounds. I hope theorists who think about light neutral bosons will take these new results into account.
Entry filed under: Particle Physics.