Target Selection for the SDSS-III MARVELS Survey

Abstract

We present the target selection process for the Multi-object APO Radial Velocity Exoplanets Large-area Survey (MARVELS), which is part of the Sloan Digital Sky Survey (SDSS) III. MARVELS is a medium-resolution ($R \sim 11000$) multi-fiber spectrograph capable of obtaining radial velocities for 60 objects at a time in order to find brown dwarfs and giant planets. The survey was configured to target dwarf stars with effective temperatures approximately between $4500$ and $6250 \, \mbox{K}$. For the first 2 years MARVELS relied on low-resolution spectroscopic pre-observations to estimate the effective temperature and $\log(g)$ for candidate stars and then selected suitable dwarf stars from this pool. Ultimately, the pre-observation spectra proved ineffective at filtering out giant stars; many giants were incorrectly classified as dwarfs, resulting in a giant contamination rate of $\sim$30\% for the first phase of the MARVELS survey. Thereafter, the survey instead applied a reduced proper motion cut to eliminate giants and used the Infrared Flux Method to estimate effective temperatures, using only extant photometric and proper-motion catalog information. The target selection method introduced here may be useful for other surveys that need to rely on extant catalog data for selection of specific stellar populations.