2002ApJ...566..452G

Recently Sahu et al., using the Hubble Space Telescope to monitor stars in the direction of the old (∼12 Gyr) globular cluster M22, detected six events in which otherwise constant stars brightened by ∼50% during a time of ≲1 day. They tentatively interpret these unresolved events as being caused by microlensing of background bulge stars by free-floating planets in M22. Using simple analytic arguments, I show that if these spike events are caused by microlensing, the lensing objects are unlikely to be associated with M22 and unlikely to be part of a smoothly distributed Galactic population. Thus, either there happens to be a massive dark cluster of planets along our line of sight to M22 or the spike events are not caused by microlensing. The lensing planets cannot be bound to stars in the core of M22: if they were closer than ∼8 AU, the lensing influence of the parent star would have been detectable. Moreover, in the core of M22, all planets with separations ≳1 AU would have been ionized by random stellar encounters. Most unbound planets would have escaped the core via evaporation, which preferentially affects such low-mass objects. Bound or free-floating planets can exist in the outer halo of M22; however, for reasonable assumptions, the maximum optical depth to such a population falls short of the observed optical depth, τ∼3x10^–6, by a factor of 5-10. Therefore, if real, these events represent the detection of a significant free-floating Galactic planet population. The optical depth to these planets is comparable to and mutually exclusive from the optical depth to resolved events measured by microlensing survey collaborations toward the bulge and, thus, implies a similar additional mass of lensing objects. If the spatial and kinematic distributions of the two populations are the same, there are more than 10³ planets per bulge microlens. Such a population is difficult to reconcile with both theory and observations.