Planetary Microlensing Perturbations: True Planets or Binary Sources?

A planetary microlensing event is characterized by a short-lived perturbation to the standard Paczy\'nski curve. Planetary perturbations typically last from a few hours to a day, and have maximum amplitudes, $\dmax$, of $5-20$ of the standard curve. There exist a subset of binary-source events that can reproduce these main features, and thus masquerade as planetary events. These events require a binary source with a small flux ratio, $\epsil \sim 10^{-2}-10^{-4}$, and a small impact parameter for the fainter source, $\beta_2 \lsim \epsil / \dmax$. The detection probability of events of this type is $\sim \beta_2$, and can be as high as $\sim 30$; this is comparable to planetary detection rates. Thus a sample of planetary-like perturbations could be seriously contaminated by binary-source events, and there exists the possibility that completely meaningless physical parameters would be derived for any given perturbation. Here I derive analytic expressions for a binary-source event in the extreme flux ratio limit, and use these to demonstrate the basic degeneracy between binary source and planet perturbations. I describe how the degeneracy can be broken by dense and accurate sampling of the perturbation, optical/infrared photometry, or spectroscopic measurements.