Disentangling Milky Way halo populations at low metallicities using [Al/Fe]^★

¹ Lund Observatory, Department of Geology, Sölvegatan 12, 223 62 Lund, Sweden
² Dipartimento di Fisica e Astronomia, Universitá degli Studi di Firenze, Via G. Sansone 1, 50019 Sesto Fiorentino, Italy
³ Observational Astrophysics, Department of Physics and Astronomy, Uppsala University, Box 516, 751 20 Uppsala, Sweden

^★★ Corresponding author: heitor.ernandes@geol.lu.se

Received: 8 September 2025
Accepted: 26 September 2025

Abstract

Context. Differentiating between in situ and accreted populations in the Milky Way halo is a challenging task. Various kinematic spaces are often used to identify distinct accreted populations from the in situ Milky Way halo. However, this approach has limitations, especially at low orbital energies. To overcome this ambiguity, elemental abundances are typically used to distinguish between populations. However, for many elemental abundance ratios (e.g. [Mg/Fe]), it remains difficult to make this distinction at low metallicities ([Fe/H]~−1.5). Aluminium abundances, on the other hand, are empirically effective discriminators and separate accreted from in situ populations in the Milky Way halo, even at low metallicities and low orbital energies.

Aims. We aim to test the discriminating power of [Al/Fe] using a well-studied sample of high-velocity stars in the solar vicinity with high-quality spectra. With these stars, we explore the ability of [Al/Fe] to separate in situ from accreted stars and test its limitations.

Methods. We derived aluminium abundances from the Al I 3944 and 3961 Å lines for 45 stars observed in two ESO programmes, along with 11 stars with archival spectra. Aluminium abundances were determined using 1D LTE and 1D NLTE spectral synthesis and line-profile fitting.

Results. We confirm that the halo low-α population (associated with accreted stars) systematically exhibits lower [Al/Fe] compared to halo high-α (in situ) stars. However, at low [Fe/H] ≈ −1.4, we identify three stars, previously classified as in situ or thick-disc stars, as having low [Al/Fe] < −0.3, indicating that they might actually be accreted.

Conclusions. Aluminium abundances, when carefully measured and with NLTE effects taken into account, are effective tracers of the chemical history of halo stars within the [Fe/H] limits analysed in the present work (−1.5 ≲[Fe/H]≲ −0.5). They provide an independent constraint on origin, complementing α-element abundance trends, and help disentangle subpopulations within the accreted halo, especially in the metal-poor regime.