Constraining the parameters of globular cluster NGC 1904 from its variable star population

TL;DR

This study analyzes variable stars in globular cluster NGC 1904 using multi-year observations to refine stellar parameters, derive cluster properties, and confirm variable star classifications, including the discovery of a new RR Lyrae star.

Contribution

It provides new high-precision light curves, refines periods of known variables, and estimates the cluster's metallicity, distance, and age using Fourier analysis and isochrone fitting.

Findings

Confirmed 10 out of 13 previously classified variables

Discovered a new RR Lyrae variable in NGC 1904

Estimated cluster metallicity and distance with improved precision

Abstract

We present the analysis of 11 nights of V and I time-series observations of the globular cluster NGC 1904 (M 79). Using this we searched for variable stars in this cluster and attempted to refine the periods of known variables, making use of a time baseline spanning almost 8 years. We use our data to derive the metallicity and distance of NGC 1904. We used difference imaging to reduce our data to obtain high-precision light curves of variable stars. We then estimated the cluster parameters by performing a Fourier decomposition of the light curves of RR Lyrae stars for which a good period estimate was possible. We also derive an estimate for the age of the cluster by fitting theoretical isochrones to our colour-magnitude diagram (CMD). Out of 13 stars previously classified as variables, we confirm that 10 are bona fide variables. We cannot detect variability in one other within the precision of our data, while there are two which are saturated in our data frames, but we do not find sufficient evidence in the literature to confirm their variability. We also detect a new RR Lyrae variable, giving a total number of confirmed variable stars in NGC 1904 of 11. Using the Fourier parameters, we find a cluster metallicity [Fe/H]_ZW=-1.63 +- 0.14, or [Fe/H]_UVES=-1.57 \pm 0.18, and a distance of 13.3 +- 0.4 kpc (using RR0 variables) or 12.9 kpc (using the one RR1 variable in our sample for which Fourier decomposition was possible).