Prospects of measuring a metallicity trend and spread in globular clusters from low-resolution spectroscopy

Abstract

Context. The metallicity spread, or the metallicity trend along the evolutionary sequence of a globular cluster, is a rich source of information to help understand the cluster physics (e.g., multiple populations) and stellar physics (e.g., atomic diffusion). Low-resolution integral-field-unit spectroscopy in the optical with the MUSE spectrograph is an attractive prospect if it can provide these diagnostics because it allows us to efficiently extract spectra of a large fraction of the cluster stars with only a few telescope pointings. Aims. We investigate the possibilities of full-spectrum fitting to derive stellar parameters and chemical abundances at low spectral resolution (R ∼ 2000). Methods. We reanalysed 1584 MUSE spectra of 1061 stars above the turn-off of NGC 6397 using FERRE and employing two different synthetic libraries. Results. We derive the equivalent iron abundance [Fe/H]e for fixed values of [α/Fe] (solar or enhanced). We find that (i) the interpolation schema and grid mesh are not critical for the precision, metallicity spread, and trend; (ii) with the two considered grids, [Fe/H]e increases by ∼0.2 dex along the sub-giant branch, starting from the turn-off of the main sequence; (iii) restricting the wavelength range to the optical decreases the precision significantly; and (iv) the precision obtained with the synthetic libraries is lower than the precision obtained previously with empirical libraries. Conclusions. Full-spectrum fitting provides reproducible results that are robust to the choice of the reference grid of synthetic spectra and to the details of the analysis. The [Fe/H]e increase along the sub-giant branch is in stark contrast with the nearly constant iron abundance previously found with empirical libraries. The precision of the measurements (0.05 dex on [Fe/H]e) is currently not sufficient to assess the intrinsic chemical abundance spreads, but this may change with deeper observations. Improvements of the synthetic spectra are still needed to deliver the full possibilities of full-spectrum fitting.