Blind chemical tagging with DBSCAN: prospects for spectroscopic surveys

Abstract: Chemical tagging has great promise as a technique to unveil our Galaxy's history. Grouping stars based on their similar chemistry can establish details of the star formation and merger history of the Milky Way. With precise measurements of stellar chemistry, chemical tagging may be able to group together stars born from the same gas cloud, regardless of their current positions and kinematics. Successfully tagging these birth clusters requires high quality chemical space information and a good cluster-finding algorithm. To test the feasibility of chemical tagging on data from current and upcoming spectroscopic surveys, we construct a realistic set of synthetic clusters, creating both observed spectra and derived chemical abundances for each star. We use Density-Based Spatial Clustering of Applications with Noise (DBSCAN) to group stars based on their spectra or abundances; these groups are matched to input clusters and are found to be highly homogeneous and complete. The percentage of clusters with more than 10 members recovered is 40% when tagging on abundances with uncertainties achievable with current techniques. Based on our fiducial model for the Milky Way, we predict recovering over 600 clusters with at least 10 observed members and 70% membership homogeneity in a sample similar to the APOGEE survey. Tagging larger surveys like the GALAH survey and the future Milky Way Mapper in SDSS V could recover tens of thousands of clusters at high homogeneity. Access to so many unique co-eval clusters will transform how we understand the star formation history and chemical evolution of our Galaxy.