A Contact Binary Misclassified as an Ellipsoidal Variable: Complications for Detached Black Hole Searches

Abstract Identifying sources exhibiting ellipsoidal variability in large photometric surveys is becoming a promising method to search for candidate detached black holes (BHs) in binaries. This technique aims to exploit the orbital-phase-dependent modulation in optical photometry caused by the BH distorting the shape of the luminous star to constrain the mass ratio of the binary. Without understanding if, or how much, contamination is present in the candidate BH samples produced by this new technique it is hard to leverage them for BH discovery. Here, we follow up one of the best candidates identified from Gaia Data Release 3, Gaia DR3 4042390512917208960, with a radial velocity (RV) campaign. Combined photometric and RV modeling, along with spectral disentangling, suggests that the true mass ratio (the mass of the unseen object divided by the mass of the luminous star) is an order of magnitude smaller than that inferred assuming the modulations arise from ellipsoidal variability. We therefore infer that this system is likely a contact binary, or on the boundary of both stars nearly filling their Roche lobes; however, further observations are required to confidently detect the secondary. We find that the well-known problem of discriminating between ellipsoidal and contact binary light curves results in a larger contamination from contact binaries than previously suggested. Until ellipsoidal variables can be reliably distinguished from contact binaries, samples of BH candidates selected based on ellipsoidal variability are likely to be highly contaminated by contact binaries or similar systems.