2010ApJ...725..721G

This work presents a homogeneous derivation of atmospheric parameters and iron abundances for a sample of giant and subgiant stars which host giant planets, as well as a control sample of subgiant stars not known to host giant planets. The analysis is done using the same technique as for our previous analysis of a large sample of planet-hosting and control sample dwarf stars. A comparison between the distributions of [Fe/H] in planet-hosting main-sequence (MS) stars, subgiants, and giants within these samples finds that the MS stars and subgiants have the same mean metallicity of <[Fe/H] > ≃ +0.11 dex, while the giant sample is typically more metal poor, having an average metallicity of [Fe/H] = -0.06 dex. The fact that the subgiants have the same average metallicities as the dwarfs indicates that significant accretion of solid metal-rich material onto the planet-hosting stars has not taken place, as such material would be diluted in the evolution from dwarf to subgiant. The lower metallicity found for the planet-hosting giant stars in comparison with the planet-hosting dwarfs and subgiants is interpreted as being related to the underlying stellar mass, with giants having larger masses and thus, on average, larger-mass protoplanetary disks. In core accretion models of planet formation, larger disk masses can contain the critical amount of metals necessary to form giant planets even at lower metallicities.