Co-evolution of the Milky Way high- and low-α sequences with chemical evolution models

Observational data have revealed a clear dichotomy in the α/Fe versus Fe/H diagram of the Milky Way thick and thin disc stars. Many recent studies have shown evidence of a co-evolution phase between the high- and low-α disc sequences as well as the presence of very old low-α stars. We aim to revise the parallel chemical evolution model that assumes two parallel histories of star formation for the two discs, by considering a pre-enriched delayed second infall episode in our revised scenario. By means of our chemical evolution models, we aim to explore the effects of a phase of co-evolution and the presence of old low-α stars, as recently observed. We consider a new version of the parallel scenario for the Milky Way thick and thin disc formation, which consists of two distinct infall episodes of slightly pre-enriched gas. The gas is considered to be extragalactic but possibly contaminated by the chemically enriched gas of a massive dwarf galaxy as Gaia-Enceladus, which merged with the Milky Way at least 10 Gyrs ago. Moreover, we test in our model observationally derived star formation histories of kinematically selected thick and thin discs, suggesting that the star formation is triggered by the passages of the Sagittarius galaxy. Our models can explain well the α/Fe versus Fe/H diagram from APOGEE DR17. The high-α sequence should have formed fast (in fewer than one billion years), but the low-α one on a much longer timescale in the solar vicinity (several gigayears). A hiatus arises as a period of low star formation between the formation of the high- and low-α sequences. We are able to predict the existence of low-α stars older than 11 Gyrs, as found in the considered observational sample. Concerning the observationally motivated histories of star formation, we find that the star formation rate for the thin disc can also reproduce the data well with the inclusion of bursts; on the other hand, a prolonged star formation history for the thick disc is not compatible with its observed stellar age distribution of a very old population. Our revised chemical evolution model with a pre-enriched and delayed (roughly 1 Gyr) second infall episode explains not only the abundance patterns of high- and low-α stars but also stellar age distributions for the selected observational sample. We predict a short co-evolution period in between the two phases and we can explain the observed old low-α stars, but still further data for precise stellar ages would be needed to put more stringent constraints on their physical nature.