We study with unprecedented detail the chemical composition and stellar parameters of the solar twin 18 Sco in a strictly differential sense relative to the Sun. Our study is mainly based on high resolution (R ~ 110 000) high S/N (800-1000) VLT UVES
spectra, which allow us to achieve a precision of about 0.005 dex in differential abundances. The effective temperature and surface gravity of 18 Sco are Teff = 5823+/-6 K and log g = 4.45+/-0.02 dex, i.e., 18 Sco is 46+/-6 K hotter than the Sun and log g is 0.01+/-0.02 dex higher. Its metallicity is [Fe/H] = 0.054+/-0.005 dex and its microturbulence velocity is +0.02+/-0.01 km/s higher than solar. Our precise stellar parameters and differential isochrone analysis show that 18 Sco has a mass of 1.04+/-0.02M_Sun and that it is ~1.6 Gyr younger than the Sun. We use precise HARPS radial velocities to search for planets, but none were detected. The chemical abundance pattern of 18 Sco displays a clear trend with condensation temperature, showing thus higher abundances of refractories in 18 Sco than in the Sun. Intriguingly, there are enhancements in the neutron-capture elements relative to the Sun. Despite the small element-to-element abundance differences among nearby n-capture elements (~0.02 dex), we successfully reproduce the r-process pattern in the solar system. This is independent evidence for the universality of the r-process. Our results have important implications for chemical tagging in our Galaxy and nucleosynthesis in general.
[Context]. The standard solar model fails to predict the very low lithium abundance in the Sun, which is much lower than the proto-solar nebula. This Li problem has been debated for decades, and it has been ascribed either to planet formation or to s
ecular stellar depletion. In order to test the evolution of Li, it is important to find solar twins in a range of ages. Also, the study of stars similar to the Sun is relevant in relation to the signature of terrestrial planet formation around the Sun. [Methods]. We acquired high-resolution (R=110,000), high S/N (~300) ESO/VLT UVES spectra of several solar twin candidates and the Sun (as reflected from the asteroid Juno). Among the solar twin candidates we identify HIP 114328 as a solar twin and perform a differential line-by-line abundance analysis of this star relative to the Sun. [Results]. HIP 114328 has stellar parameters Teff = 5785+/-10 K, log g = 4.38+/-0.03, [Fe/H] = -0.022+/-0.009, and a microturbulent velocity 0.05+/-0.03 km/s higher than solar. The differential analysis shows that this star is chemically very similar to the Sun. The refractory elements seem even slightly more depleted than in the Sun, meaning that HIP 114328 may be as likely to form terrestrial planets as the Sun. HIP 114328 is about 2 Gyr older than the Sun, and is thus the second oldest solar twin analyzed at high precision. It has a Li abundance of A(Li)NLTE <= 0.46, which is about 4 times lower than in the Sun (A(Li)NLTE = 1.07 dex), but close to the oldest solar twin known, HIP 102152. [Conclusions]. Based on the lower abundances of refractory elements when compared to other solar twins, HIP 114328 seems an excellent candidate to host rocky planets. The low Li abundance of this star is consistent with its old age and fits very well the emerging Li-age relation among solar twins of different ages.
Context: Recent works with improved model atmospheres, line formation, atomic and molecular data, and detailed treatment of blends, have resulted in a significant downward revision of the solar oxygen abundance. Aims: Considering the importance of
the Sun as an astrophysical standard and the current conflict of standard solar models using the new solar abundances with helioseismological observations we have performed a new study of the solar oxygen abundance based on the forbidden [OI] line at 5577.34 A, not previously considered. Methods: High-resolution (R > 500 000), high signal-to-noise (S/N > 1000) solar spectra of the [O I] 5577.34 A line have been analyzed employing both three-dimensional (3D) and a variety of 1D (spatially and temporally averaged 3D, Holweger & Muller, MARCS and Kurucz models with and without convective overshooting) model atmospheres. Results: The oxygen abundance obtained from the [OI] 5577.3 A forbidden line is almost insensitive to the input model atmosphere and has a mean value of A(O) = 8.71 +/- 0.02 (sigma from using the different model atmospheres). The total error (0.07 dex) is dominated by uncertainties in the log gf value (0.03 dex), apparent line variation (0.04 dex) and uncertainties in the continuum and line positions (0.05 dex). Conclusions: The here derived oxygen abundance is close to the 3D-based estimates from the two other [OI] lines at 6300 and 6363 A, the permitted OI lines and vibrational and rotational OH transitions in the infrared. Our study thus supports a low solar oxygen abundance (A(O) ~ 8.7), independent of the adopted model atmosphere.
