Selection of nearby microlensing candidates for observation by the Space Interferometry Mission

I investigate the prospects for using the Space Interferometry Mission (SIM) to measure the masses of nearby stars from their gravitational deflection of light from more distant sources, as originally suggested by Paczynski and by Miralda-Escude. I derive an analytic expression for the total observing time T-tot required to measure the masses of a fixed number of stars to a given precision. I find that T-tot proportional to r(max)(-2), where r(max) is the physical depth to which candidates are searced, or T-tot proportional to mu(min)(2), where mu(min) is the minimum proper motion to which candidates are searched. I show that T-tot can be reduced by a factor of 4 if source availability is extended from V-s = 17 to V-s = 19. Increasing r(max) and V-s and decreasing mu(min) require a significantly more aggressive approach to finding candidates. A. search for candidates can begin by making use of the Luyten proper motion catalog together with the USNO-A2.0 all-sky astrometric catalog. However, a thorough search would require an all-sky proper-motion catalog such as USNO-B or Guide Star Catalog II, which are not yet available. The preliminary observations necessary to prepare for the mission will become more difficult the longer they are delayed because the candidate pairs are typically already within 1" and are getting closer.