Direct imaging constraints on planet populations detected by microlensing

Context. Results from gravitational microlensing suggested the existence of a large population of free-floating planetary mass objects. The main conclusion from this work was partly based on constraints from a direct imaging survey. This survey determined upper limits for the frequency of stars that harbor giant exoplanets at large orbital separations. Aims. We want to verify to what extent upper limits from direct imaging do indeed constrain the microlensing results. Methods. We examine the current derivation of the upper limits used in the microlensing study and re-analyze the data from the corresponding imaging survey. We focus on the mass and semi-major axis ranges that are most relevant in context of the microlensing results. We also consider new results from a recent M-dwarf imaging survey as these objects are typically the host stars for planets detected by microlensing. Results. We find that the upper limits currently applied in context of the microlensing results are probably underestimated. This means that a larger fraction of stars than assumed may harbor gas giant planets at larger orbital separations. Also, the way the upper limit is currently used to estimate the fraction of free-floating objects is not strictly correct. If the planetary surface density of giant planets around M-dwarfs is described as df(Planet) proportional to a(beta)da, we find that beta less than or similar to 0.5-0.6 is consistent with results from different observational studies probing semi-major axes between similar to 0.03-30 AU. Conclusions. Having a higher upper limit on the fraction of stars that may have gas giant planets at orbital separations probed by the microlensing data implies that more of the planets detected in the microlensing study are potentially bound to stars rather than free-floating. The current observational data are consistent with a rising planetary surface density for giant exoplanets around M-dwarfs out to similar to 30 AU. Future direct imaging surveys will show out to what semi-major axis the above mentioned range of beta is valid and what fraction of the planetary mass objects detected by microlensing are indeed bound.