On the origins of extreme velocity stars as revealed by large-scale Galactic surveys

We assemble a large sample of 12,784 high-velocity stars with total velocity $\it{V}_{\rm{GSR}}\ge {\rm 300}$ km s$^{-1}$, selected from RAVE DR5, SDSS DR12, LAMOST DR8, APOGEE DR16, GALAH DR2, and $Gaia$ EDR3. In this sample, 52 are marginally hypervelocity star (HVS) candidates that have $\it{V}_{\rm{GSR}}$ exceeding their local escape velocities within $2\sigma$ confidence levels, 40 of which are discovered for the first time. All candidates are metal-poor, late-type halo stars, significantly different from the previous identified HVSs, which are largely massive early-type stars, discovered by extreme radial velocity. This finding suggests that our newly identified HVS candidates are ejected by different mechanisms from the previous population. To investigate their origins, for 547 extreme velocity stars with ${V}_{\rm{GSR}}\ge0.8V_{\rm{esc}}$, we reconstruct their backward-integrated trajectories in the Galactic potential. According to the orbital analysis, no candidates are found to be definitely ejected from the Galactic-center (GC), while 8 metal-poor extreme velocity stars are found to have a closest distance to the GC within 1kpc. Intriguingly, 15 extreme velocity stars (including 2 HVS candidates) are found to have experienced close encounters with the Sagittarius dSph, suggesting that they originated from this dSph. This hypothesis is supported by an analysis of the [$\alpha$/Fe]--[Fe/H] diagram. From a preliminary analysis of all the 547 extreme velocity stars, we propose a general picture: Star ejection from Galactic subsystems such as dwarf galaxies and globular clusters can be an important channel to produce extreme velocity stars or even HVSs, particularly the metal-poor late-type halo population.