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

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 the populations. Yet, for many elemental abundance ratios, it remains difficult to make this distinction at low metallicities. Aluminium abundances, on the other hand, have been empirically found to be an effective discriminator, allowing for the separation of accreted and in-situ populations in the Milky Way halo even at low metallicities and low orbital energies. 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 [Al/Fe] ability to separate the in-situ from accreted stars and test its limitations. We derived aluminium abundances from the Al I 3944 and 3961 {\rm 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. We confirm that the low-$α$ population systematically has lower [Al/Fe] compared to high-$α$ stars. Aluminium abundances, when carefully measured and NLTE effects taken into account, are effective tracers of the chemical history of halo stars. They provide an independent constraint on origin, complementing $α$-element abundances trends, and help us to disentangle subpopulations within the accreted halo, especially in the metal-poor regime.