The Hills Mechanism and the Galactic Center S-stars

Abstract

Our Galactic Center contains young stars, including the few million year old clockwise disk of O and B stars between 0.05 pc to 0.5 pc from the Galactic Center, and the S-star cluster of early B-type stars at a galactocentric distance of ∼0.01 pc. Recent observations suggest the S-stars are remnants of tidally disrupted binaries from the clockwise disk. In particular, Koposov et al. 2020 recently discovered a hypervelocity star that was ejected from the Galactic Center five million years ago with a velocity vector that is consistent with this disk. We perform a detailed study of this scenario. First, we quantify the plausible range of binary semi-major axes in the clockwise disk. We find that the dynamical evaporation of such binaries is dominated by other disk stars rather than more numerous old, isotropic stellar population. For the expected range of binary semi-major axes in the clockwise disk, binary tidal disruptions (also known as the Hills mechanism) can reproduce the observed S-star semi-major axis distribution. Reproducing the observed thermal eccentricity distribution of the S-stars requires an additional relaxation process. The flight time of the recently discovered hypervelocity star from the Galactic Center and the most recent constraints on the S-stars' ages both suggest this process has to be effective within 10 7 years. We consider three possibilities: (i) scalar resonant relaxation from the surrounding isotropic star cluster (ii) torques from the clockwise disk, and (iii) an intermediate mass black hole. We conclude that only the latter would be fast enough to reproduce the observed S-star eccentricity distribution. Finally, we show that the primary star from an unequal mass binary would be deposited at larger semi-major axes than the secondary, possibly explaining the dearth of O stars among the S-stars.

The former companion of the hyper-velocity star S5-HVS1

“…In the following, we draw the distribution of pericenter distance rp by considering two extreme cases of empty and full loss cones. The reality should be somewhere in between, especially considering that binary may come from the nuclear star cluster within the central 0.5 pc from SgrA* (e.g., in the eccentric disk instability scenario, Madigan et al 2009;Generozov & Madigan 2020), and in this case the loss cone has a larger size and may be closer to the empty side. We also note that even when ∆J J lc , the phase space deeply within the loss-cone can still be populated by rare strong scatterings (Weissbein & Sari 2017).…”

A stream of hypervelocity stars from the Galactic Center

“…S5-HVS1's velocity vector is also consistent with a disk origin. As discussed in Madigan et al (2009) and Generozov & Madigan (2020), this disk could excite binaries to tidal disruption via a gravitational instability early in its evolutionary history. In particular, if this structure starts as an apsidally-aligned, eccentric disk some of its member stars and binaries can be excited to extreme eccentricities and tidally disrupted.…”

“…These simulations are described in detail in Generozov & Madigan (2020). We initialize disks of 300 equal mass particles in an eccentric, apsidally-aligned configuration.…”

“…The semimajor axis is drawn from a log-uniform distribution. The minimum and maximum binary semimajor axis are functions of the binary mass and eccentricity, and are set by the prescriptions in § 4 of Generozov & Madigan (2020) (the semimajor axis is typically between a few×10 −2 and a few au). The binaries' centers of mass are on eccentric orbits with a semimajor axis of 0.05 pc (approximately the inner edge of the clockwise disk).…”

“…This can be understood at least qualitatively in the Hills model. For nearly parabolic disruptions the primary and secondary are equally likely to be left bound to the SMBH, but the primary would be deposited at larger semimajor axis (Kobayashi et al 2012;Generozov & Madigan 2020). Assuming that the masses of the observed S-stars are drawn from the m −1.7 mass function of the disk (Lu et al 2013) with a truncation at low mass due to observability and a truncation at higher masses due to 1 The distribution of binary properties in these simulated encounters is similar to the input distribution for Figure 7 of Generozov & Madigan (2020).…”