The heavy elements mixture and the stellar cluster age

S. Degl'Innocenti,P.G. Prada Moroni,B. Ricci
DOI: https://doi.org/10.48550/arXiv.astro-ph/0504611
2005-07-08
Abstract:Recent analyses of solar spectroscopic data (see Asplund, Grevesse & Sauval 2004 and references therein) suggest a significative variation of the heavy elements abundance. The change of heavy mixture might affect the determination of the clusters age for two different reasons: the change of theoretical isochrones at fixed metallicity and the variation of the inferred cluster metallicity from the observed [Fe/H]. The first point is analyzed discussing the effects of updating the metal distribution on theoretical evolutionary tracks and isochrones for metallicities suitable for the galactic globular and open clusters and for the bulk population of the Large Magellanic Clouds. The maximum variation of the estimated age ($\approx$0.5 Gyr), although not negligible, is still within the present uncertainty. The second point is addressed by comparing present theoretical predictions with the very precise observational data of the Hyades cluster from the Hypparcos satellite.
Astrophysics
What problem does this paper attempt to address?
The main problem that this paper attempts to solve is to explore the influence of the change in the mixing ratio of heavy elements on the age determination of star clusters. Specifically, the author analyzes the influence in two aspects: 1. **Changes in theoretical isochrones at a fixed metallicity**: Whether the updated mixing ratio of heavy elements will affect the isochrones in the stellar evolution model, thereby affecting the estimation of the age of star clusters. 2. **Changes in the metallicity of star clusters inferred from the observed [Fe/H]**: Due to the adjustment of the solar heavy - element abundances, whether the metallicity Z of star clusters inferred from the observed iron abundance [Fe/H] will change, and then affect the age estimate. ### Detailed Explanation #### 1. Changes in Theoretical Isochrones - **Background**: Recent analyses of solar spectral data have shown that the abundances of heavy elements have changed significantly (Asplund et al., 2004). These changes may affect theoretical isochrones, especially in the case of a fixed metallicity. - **Research Method**: The author calculated the stellar evolution tracks and isochrones under different mixing ratios of heavy elements (such as Asplund et al., 2004 and Grevesse & Noels, 1993), and compared their effects on stars in Galactic globular clusters, open clusters, and the Large Magellanic Cloud. - **Result**: For globular clusters, the maximum age difference is about 0.5 Gyr. Although it is not negligible, it is still within the current uncertainty range of age estimates. #### 2. Changes in Metallicity Inferred from Observed [Fe/H] - **Background**: The new solar heavy - element mixing ratio reduces the inferred metallicity Z of star clusters. For example, for nearby open clusters, the new (Z/X)⊙ value drops from 0.0230 to 0.0165, resulting in a significant decrease in the estimated metallicity. - **Research Method**: The author used very accurate Hipparcos satellite observation data to analyze the color - magnitude diagram (CMD) of the Hyades cluster and compared it with theoretical isochrones. - **Result**: The reduced metallicity causes the main sequence (MS) to shift blueward by about 0.05 magnitudes, resulting in inconsistency with the observed data. In order to achieve a reasonable match, the metallicity needs to be increased to at least Z = 0.02. ### Summary The core problem of this paper is to evaluate the influence of the change in the mixing ratio of heavy elements on the age estimate of star clusters. Through the comparison of theoretical models and observational data, the author finds that this change does have a certain influence on the age estimate, but the degree of its influence is still within the current uncertainty range. However, for some star clusters (such as Hyades), this change may lead to significant deviations, and further adjustment of model parameters is required to better match the observational data.