Simulation of a Space-based Microlensing Survey for Terrestrial Extrasolar Planets

We show that a space-based gravitational microlensing survey for terrestrial extrasolar planets is feasible in the near future and could provide a nearly complete picture of the properties of planetary systems in our Galaxy. We present simulations of such a survey using a 1-2 m aperture space telescope with a ~2 deg2 field of view to continuously monitor ~108 Galactic bulge main-sequence stars. The microlensing techniques allow the discovery of low-mass planets with high signal-to-noise ratio, and the space missions that we have studied are sensitive to planets with masses as low as that of Mars. By targeting main-sequence source stars, which can only be resolved from space, the space-based microlensing survey is able to detect enough light from the lens stars to determine the spectral type of one-third of the lens stars with detected planets, including virtually all of the F, G, and K stars, which comprise one-quarter of the event sample. This enables the determination of the planetary masses and separations in physical units as well as the abundance of planets as a function of stellar type and distance from the Galactic center. We show that a space-based microlensing planet search program has its highest sensitivity to planets at orbital separations of 0.7-10 AU, but it will also have significant sensitivity at larger separations and will be able to detect free-floating planets in significant numbers. This complements the planned terrestrial planet transit missions, which are sensitive to terrestrial planets at separations of ?1 AU. Such a mission should also detect ~50,000 giant planets via transits, and it is, therefore, the only proposed planet detection method that is sensitive to planets at all orbital radii.