Atomic diffusion and mixing in old stars VII. Abundances of Mg, Ti, and Fe in M30

Abstract

We attempt to constrain the efficiency of additional transport or mixing processes that reduce the effect of atomic diffusion in stellar atmospheres. We apply spectral synthesis methods to spectra observed with the GIRAFFE spectrograph on the VLT to estimate abundances of Mg, Ti, Fe, and Ba in stars in the metal-poor globular cluster M30. To the abundances we fit trends of abundances predicted by stellar evolution models assuming different efficiencies of additional transport or mixing processes. The fitting procedure attempts to take into account the effects of parameter-dependent systematic errors in the derived abundances. We find that the parameter $T_0$, which describes the efficiency of additional transport or mixing processes, can almost certainly be constrained to the narrow range $\log_{10}{\left( T_0 / \left[ \mathrm{K} \right] \right)}$ between $6.09$ and $6.2$. This corresponds to decreased abundances for stars at the main sequence turn-off point compared to the red giant branch by $0.2\,\mathrm{dex}$ for Mg, $0.1\,\mathrm{dex}$ for Fe, and $0.07\,\mathrm{dex}$ for Ti. We also find that while our estimates do have non-negligible systematic errors stemming from the continuum placement and the assumed microturbulence, our method can take them into account. Our results partly amend the results of an earlier paper in this article series, that tentatively used a value of $\log_{10}{\left( T_0 / \left[ \mathrm{K} \right] \right)} = 6.0$ when modelling the Spite plateau of lithium. To more easily distinguish physical effects from systematic errors, we recommend that studies of this kind focus on elements for which the expected surface abundances as functions of effective temperature have a distinct structure and cover a wide range.