Testing blend scenarios for extrasolar transiting planet candidates. I. OGLE-TR-33: A false positive

We report high-resolution spectroscopic follow-up observations of the faint transiting planet candidate OGLE-TR-33 (V 14: 7), located in the direction of the Galactic center. Small changes in the radial velocity of the star were detected that initially suggested the presence of a large planet or brown dwarf in orbit. However, further analysis revealed spectral line asymmetries that change in phase with the 1.95 day period, casting doubt on those measurements. These asymmetries make it more likely that the transit-like events in the light curve are the result of contamination from the light of an eclipsing binary along the same line of sight (referred to as a blend). We performed detailed simulations in which we generated synthetic light curves resulting from such blend scenarios and fitted them to the measured light curve. Guided by these fits and the inferred properties of the stars, we uncovered a second set of lines in our spectra that correspond to the primary of the eclipsing binary and explain the asymmetries. Using all the constraints from spectroscopy, we were then able to construct a model that satisfies all the observations and to characterize the three stars based on model isochrones. OGLE-TR-33 is fully consistent with being a hierarchical triple system composed of a slightly evolved F6 star (the brighter object) near the end of its main-sequence phase and an eclipsing binary with a K7-M0 star orbiting an F4 star. The application to OGLE-TR-33 of the formalism developed to fit light curves of transit candidates illustrates the power of such simulations for predicting additional properties of the blend and for guiding further observations that may serve to confirm that scenario, thereby ruling out a planet. Tests such as this can be very important for validating faint candidates.