2008ApJ...684.1093R

We reanalyze the M31 microlensing event WeCAPP-GL1/POINT-AGAPE-S3 taking into account that stars are not pointlike but extended. We show that the finite size of stars can dramatically change the self-lensing event rate and (less dramatically) also the halo lensing event rate, if events are as bright as WeCAPP-GL1. The brightness of the brightest events mostly depends on the source sizes and fluxes and on the distance distribution of sources and lenses and therefore can be used as a sensitive discriminator between halo-lensing and self-lensing events, provided the stellar population mix of source stars is known well enough. Using a realistic model for the 3D light distribution, stellar population, and extinction of M31, we show that an event like WeCAPP-GL1 is very unlikely to be caused by self-lensing. In the entire WeCAPP-field (17.2'x17.2' centered on the bulge) we expect only one self-lensing event every 49 years with the approximate parameters of WeCAPP-GL1 (FWHM timescale between 1 and 3 days and a flux excess of 19.0 mag or larger in R). On the other hand, if we assume only 20% of the dark halos of M31 and the Milky Way consist of 1 M_☉MACHOs, an event like WeCAPP-GL1 would occur every 10 years. Furthermore, if one uses position, FWHM timescale, flux excess, and color of WeCAPP-GL1, self-lensing is even 13 times less likely than lensing by a MACHO, if MACHOs contribute 20% to the total halo mass and have masses in the range of 0.1-4 M_☉. We also demonstrate that (1) the brightness distribution of events in general is a good discriminator between self- and halo lensing; (2) the timescale distribution is a good discriminator if the MACHO mass is larger than 1 M_☉. Future surveys of M31 like the PAndromeda key project of Pan-STARRS 1 should be able to provide many more such events within the next 4 years.