Impact of selection criteria on the structural parameters of the Galactic thin and thick discs

TL;DR

This study systematically evaluates how five different classification schemes for thin versus thick disc stars affect the inferred vertical and radial structure of the Milky Way's discs. Using APOGEE DR17 red giants and astroNN ages, the authors fit a simple two-exponential disc model to vertical density profiles across radial bins, comparing chemical, dynamical, kinematic, and age-based selections. Abundance- and age-based methods yield the cleanest disc separations, while kinematic and dynamical methods suffer from contamination due to mixing of well-overlapped populations; all methods consistently show a thinner disc with a longer scale length than the thicker disc, and a flaring thin disc with radius. The results underscore that the method used to separate Galactic components can significantly influence the derived disc parameters, with implications for interpreting the Galaxy’s formation history and for comparing with external galaxies. The work also demonstrates how combining multiple selection criteria could improve classification and encourages leveraging Gaia with upcoming spectroscopic surveys for more robust chemodynamical modelling.

Abstract

Context: The Milky Way contains a thick and a thin disc that differ in chemical, kinematic, structural, and spatial properties. There is significant overlap in the distributions of these properties, especially so at higher metallicities. Distinguishing between these major structural components is crucial for understanding the formation and evolution of the Galaxy. Multiple selection methods exist to classify stars as thin or thick disc stars, each with its own advantages and limitations. Aims: We investigate how different classification methods for categorising stars into the thick and thin disc populations influence the determination of structural properties of the two discs. Methods: We apply five different selection methods. Two methods use cuts in the [$α$/Fe]-[Fe/H] and [Mg/Mn]-[Al/Fe] planes; one uses a dynamical separation in $J_φ$ -$J_Z$ space; one uses an age-based cut; and the last one uses a kinematic likelihood method. For each method, we derive relative density profiles of each component as functions of height above the Galactic plane and Galactocentric radius, and fit these to a simple two-exponential disc model. We use red giant stars from APOGEE DR17 and stellar ages from astroNN. Results: Methods based on abundance or age data produce the cleanest separations, while kinematic and dynamical methods suffer higher contamination due to difficulties in separating well-mixed populations. The thin disc scale heights show a clear flaring as they increase with radius, while the thick disc stays approximately constant at around 1 kpc over most radii for all methods. All methods find the thin disc to have a longer scale length than the thick disc, with the difference being greatest for the chemical selection methods. A scale length of the thick disc of 2.0 kpc leads to one of between 2.3 and 3.0 kpc for the thin disc.

Figures (20)

• Figure 1: Selection of stars for this study. The parts of the sample that fall outside our radial limits, shown with red lines at $R_{\mathrm{gal}} = 4$ kpc and $R_{\mathrm{gal}} = 14$ kpc, are shown with transparency. The Galactic centre is marked with a red cross. Top: X-Y view of the distribution of selected stars. Bottom: $R_{\mathrm{gal}}$-Z view of the distribution of selected stars.
• Figure 2: How the accreted population was split from the thin- and thick discs using the [Mg/Mn]-[Al/Fe] plane. Left: How the [Mg/Mn]-[Al/Fe] plane was divided into different regions, marked with red lines. Region I is the thick disc, region II is the thin disc, and region III is the halo or accreted stars. Right: Distribution of stars from the selection. The panels show [Mg/Fe]-[Fe/H] in the top row and [Mg/Mn]-[Al/Fe] in the bottom row. Left columns: Region I, the thick disc; Right columns: Region II, the thin disc.
• Figure 3: How the [Mg/Fe]-[Fe/H] plane was split into the thin disc, thick disc, and halo. Left: Region I is the thick disc, Region II is the thin disc, and Region III is accreted and halo stars, separated by red lines. Right: Distribution of stars from the selection. The panels show [Mg/Fe]-[Fe/H] in the top row and [Mg/Mn]-[Al/Fe] in the bottom row. Left columns: Region I, the thick disc; Right columns: Region II, the thin disc.
• Figure 4: Distribution of stars from the kinematic selection. Top row: [Mg/Fe]-[Fe/H]. Bottom row: [Mg/Mn]-[Al/Fe]. Left: Region I, the thick disc; Right: Region II, the thin disc.
• Figure 5: How the $J_\phi$-$J_Z$ plane was split into the thin disc, thick disc, and halo. Left: Relative density of the chemically defined high- and low-$\alpha$ populations in the $J_\phi$-$J_Z$ plane. The colour of each pixel is logarithmic relative density as defined by $\log(N_{thin}) - \log(N_{thick})$, where red means thin disc dominated and blue means thick disc dominated. The disc components are separated by a black line following the area where $\log(N_{thin}) - \log(N_{thick}) = 0$. Region I is the thick disc, Region II is the thin disc, and the dashed line at $J_\phi = 0$ delineates Region III, which is the halo and accreted stars. Right: Distribution of stars from the selection. The panels show [Mg/Fe]-[Fe/H] in the top row and [Mg/Mn]-[Al/Fe] in the bottom row. Left columns: Region I, the thick disc; Right columns: Region II, the thin disc.