Is the Galactic Bulge Devoid of ExoPlanets?

Is the Galactic bulge devoid of planets?

Authors:

Penny et al

Abstract:

Using a sample of 31 microlensing exoplanet hosts, we investigate whether or not the distances to these systems conform to the Galactic distribution of planets expected from a model. We derive the expected distribution of distances from a simulated microlensing survey, correcting (roughly) for the dominant selection effects that affect the detection sensitivity to planets as a function of their distance, and compare with the observed distribution using Anderson-Darling (AD) hypothesis testing. We find that the AD test rejects the hypothesis that the observed sample is drawn from our model distribution of distances with p-value 5.0×10−4. Interestingly, we find that an AD test can not reject (p=0.18) the hypothesis that the observed sample is drawn from a model distribution where only disk stars host planets (i.e., a model where there are no bulge planets), though AD tests of the relative proper motion distributions indicate tension with this extreme hypothesis. Allowing the relative abundance of planets in the bulge to that in the disk, fb, to be a free parameter, we find that AD tests do not reject models where fb less than 0.54 for p greater than 0.01. We find that allowing for a dependence of planet occurrence rate on host mass and metallicity does not significantly change our results, and if we fix these dependencies to their respective trends as inferred from RV surveys, the limit only weakens to fb less than 0.7. We suspect that some of the distance estimates for nearby (Dl less than 2 kpc) lenses could be significantly in error. If this is indeed the case, removal of two of the potentially problematic hosts we identify would allow for fb less than 0.96. While we cannot conclusively infer fb given current observations, more homogeneous samples from on-going surveys that are easier to model will allow the measurement of fb using similar methods to those developed in this work. (abridged)