2010A&A...523A..91G

The homogeneous spectroscopic determination of the stellar parameters is a mandatory step for transit detections from space. Knowledge of which population the planet hosting stars belong to places constraints on the formation and evolution of exoplanetary systems. We used the FLAMES/GIRAFFE multi-fiber instrument at ESO to spectroscopically observe samples of stars in three CoRoT/Exoplanet fields, namely the LRa01, LRc01, and SRc01 fields, and characterize their stellar populations. We present accurate atmospheric parameters, T_eff, log g, [M/H], and [α/Fe] derived for 1227 stars in these fields using the MATISSE algorithm. The latter is based on the spectral synthesis methodology and automatically provides stellar parameters for large samples of observed spectra. We trained and applied this algorithm to FLAMES observations covering the MgI b spectral range. It was calibrated on reference stars and tested on spectroscopic samples from other studies in the literature. The barycentric radial velocities and an estimate of the Vsini values were measured using cross-correlation techniques. We corrected our samples in the LRc01 and LRa01 CoRoT fields for selection effects to characterize their FGK dwarf stars population, and compiled the first unbiased reference sample for the in-depth study of planet metallicity relationship in these CoRoT fields. We conclude that the FGK dwarf population in these fields mainly exhibit solar metallicity. We show that for transiting planet finding missions, the probability of finding planets as a function of metallicity could explain the number of planets found in the LRa01 and LRc01 CoRoT fields. This study demonstrates the potential of multi-fiber observations combined with an automated classifier such as MATISSE for massive star spectral classification.