Context. To understand the formation and composition of planetary systems, it is essential to have insights into the chemical composition of their host stars. In particular, C/O elemental ratios are useful for constraining the density and bulk composition of terrestrial planets.
Aims: We study the carbon abundances with a twofold objective. On the one hand, we want to evaluate the behaviour of carbon in the context of Galactic chemical evolution. On the other hand, we focus on the possible dependence of carbon abundances on the presence of planets and on the impact of various factors (such as different oxygen lines) on the determination of C/O elemental ratios.
Methods: We derived chemical abundances of carbon from two atomic lines for 757 FGK stars in the HARPS-GTO sample, observed with high-resolution (R ~ 115 000) and high-quality spectra. The abundances were derived using a standard Local Thermodynamic Equilibrium analysis with automatically measured Equivalent Widths injected into the code MOOG and a grid of Kurucz ATLAS9 atmospheres. Oxygen abundances, derived using different lines, were taken from previous papers in this series and updated with the new stellar parameters.
Results: We find that thick- and thin-disk stars are chemically disjunct for [C/Fe] across the full metallicity range that they have in common. Moreover, the population of high-α metal-rich stars also presents higher and clearly separated [C/Fe] ratios than thin-disk stars up to [Fe/H] ~ 0.2 dex. The [C/O] ratios present a general flat trend as a function of [O/H] but becomes negative at [O/H] ≳ 0dex. This trend is more clear when considering stars of similar metallicity. We find tentative evidence that stars with low-mass planets at lower metallicities have higher [C/Fe] ratios than stars without planets at the same metallicity, in the same way as has previously been found for α elements. Finally, the elemental C/O ratios for the vast majority of our stars are below 0.8 when using the oxygen line at 6158 Å, however, the forbidden oxygen line at 6300 Å provides systematically higher C/O values (going above 1.2 in a few cases) which also show a dependence on Teff. Moreover, by using different atmosphere models the C/O ratios can have a non-negligible difference for cool stars. Therefore, C/O ratios should be scaled to a common solar reference in order to correctly evaluate its behaviour. We find no significant differences in the distribution of C/O ratios for the different populations of planet hosts, except when comparing the stars without detected planets with the stars hosting Jupiter-type planets. However, we note that this difference might be caused by the different metallicity distributions of both populations.
Conclusions: The derivation of homogeneous abundances from high-resolution spectra in samples that are modest in size is of great utility in constraining models of Galactic chemical evolution. The combination of these high-quality data with the long-term study of planetary presence in our sample is crucial for achieving an accurate understanding of the impact of stellar chemical composition on planetary formation mechanisms.

The table with parameters and abundances is only available at the CDS via anonymous ftp to cdsarc.u-strasbg.fr (ftp://130.79.128.5) or via http://cdsarc.u-strasbg.fr/viz-bin/cat/J/A+A/655/A99

Based on observations collected at the La Silla Observatory, ESO (Chile), with the HARPS spectrograph at the 3.6 m ESO telescope (ESO runs ID 72.C-0488, 082.C-0212, and 085.C-0063).