Table of Contents
Fetching ...

The origin of B-type runaway stars based on kinematics

Yanjun Guo, Chao Liu, ZhiCun Liu, Chunyan Li, Qida Li, Kun Chen, Zhanwen Han, XueFei Chen

TL;DR

This work investigates the origins of 39 B-type runaway stars from LAMOST through orbital traceback with GALPY, combining spectral typing, NLTE-informed atmospheric parameters from SLAM, and mass-age estimation via a PARSEC-trained random forest. The trajectory analysis reveals 29 stars fully confined to the Galactic disk and 10 that disk-pass yet trace back to the disk, with two potential cluster-origin cases. A notable dichotomy in the kinematic plane, $V_{ m Sp}$ versus $v\sin i$, suggests two distinct origins: low $V_{ m Sp}$ with high $v\sin i$ aligning with BSS and high $V_{ m Sp}$ with low $v\sin i$ aligning with DES. While these findings support a mixed origin scenario for B-type runaways, the conclusions are limited by the lack of high-resolution abundance measurements, underscoring the need for future spectroscopic follow-up.

Abstract

Runaway stars depart their birthplaces with high peculiar velocities. Two mechanisms are commonly invoked to explain their origin, the binary supernova scenario (BSS) and the dynamical ejection scenario (DES). Investigating the kinematic properties of runaway stars is key to understanding their origins.We intend to investigate the origins of 39 B-type runaway stars from LAMOST using orbital traceback analysis. From the catalog of LAMOST, we selected 39 B-type runaway stars and determined their spectral subtypes from key absorption lines. We then derived atmospheric parameters for each star using the Stellar Label Machine (SLAM), which is trained on TLUSTY synthetic spectra computed under the non-local thermodynamic equilibrium (NLTE) assumption. Using the derived atmospheric parameters as input, we estimated stellar masses and ages with a machine learning model trained on PARSEC evolutionary tracks. We finally performed orbital traceback with GALPY to analyze their origins. Through orbital traceback, we find that 29 stars have trajectories entirely within the Galactic disk, whereas 10 are disk-passing yet still trace back to the disk. Two stars have trajectories that intersect those of known clusters. Their orbits show similar morphologies in both the $X-Y$ and $R-Z$ planes, and their [M/H] values are comparable, suggesting possible cluster origins. However, definitive confirmation will require additional evidence. In addition, the $V_{\rm Sp} - v\sin{i}$ plane shows that runaway stars with low peculiar space velocities but high $v\sin{i}$ remain on the Galactic disk, whereas those with high peculiar space velocities but low $v\sin{i}$ pass through the disk, possibly reflecting two distinct origins.

The origin of B-type runaway stars based on kinematics

TL;DR

This work investigates the origins of 39 B-type runaway stars from LAMOST through orbital traceback with GALPY, combining spectral typing, NLTE-informed atmospheric parameters from SLAM, and mass-age estimation via a PARSEC-trained random forest. The trajectory analysis reveals 29 stars fully confined to the Galactic disk and 10 that disk-pass yet trace back to the disk, with two potential cluster-origin cases. A notable dichotomy in the kinematic plane, versus , suggests two distinct origins: low with high aligning with BSS and high with low aligning with DES. While these findings support a mixed origin scenario for B-type runaways, the conclusions are limited by the lack of high-resolution abundance measurements, underscoring the need for future spectroscopic follow-up.

Abstract

Runaway stars depart their birthplaces with high peculiar velocities. Two mechanisms are commonly invoked to explain their origin, the binary supernova scenario (BSS) and the dynamical ejection scenario (DES). Investigating the kinematic properties of runaway stars is key to understanding their origins.We intend to investigate the origins of 39 B-type runaway stars from LAMOST using orbital traceback analysis. From the catalog of LAMOST, we selected 39 B-type runaway stars and determined their spectral subtypes from key absorption lines. We then derived atmospheric parameters for each star using the Stellar Label Machine (SLAM), which is trained on TLUSTY synthetic spectra computed under the non-local thermodynamic equilibrium (NLTE) assumption. Using the derived atmospheric parameters as input, we estimated stellar masses and ages with a machine learning model trained on PARSEC evolutionary tracks. We finally performed orbital traceback with GALPY to analyze their origins. Through orbital traceback, we find that 29 stars have trajectories entirely within the Galactic disk, whereas 10 are disk-passing yet still trace back to the disk. Two stars have trajectories that intersect those of known clusters. Their orbits show similar morphologies in both the and planes, and their [M/H] values are comparable, suggesting possible cluster origins. However, definitive confirmation will require additional evidence. In addition, the plane shows that runaway stars with low peculiar space velocities but high remain on the Galactic disk, whereas those with high peculiar space velocities but low pass through the disk, possibly reflecting two distinct origins.

Paper Structure

This paper contains 14 sections, 9 figures.

Figures (9)

  • Figure 1: Gaia DR3 parallaxes as a function of $G$ magnitude. The error bars represent the parallax uncertainties.
  • Figure 2: Sample spectra spanning B3, B5, B7, and B9 subtypes, arranged from top to bottom. Key spectral lines, including He I 4026, 4121, 4471, 4713, H$_{\delta}$ 4101, H$_{\gamma}$ 4340, and Mg II 4481, are labeled to highlight their variations across different subtypes. The RA and Dec of each target are listed below the corresponding spectrum.
  • Figure 3: Training grid and predicted sample in the $\log g$–$T_\mathrm{eff}$ plane. The background shows the PARSEC 1.2S main-sequence grid color-coded by stellar mass. The black crosses and triangles represent our 45 B-type runaway stars, with colors corresponding to their predicted masses using the same color scale. Crosses denote stars lying within the parameter range of the training grid (high certainty), while triangles indicate extrapolated sources outside the grid (low certainty).
  • Figure 4: Orbital trajectories for two example stars with coordinates RA: 57.671321, Dec: 57.178426 (top panels) and RA: 41.448800, Dec: 54.544701 (bottom panels), shown in the Galactic $X-Y$ plane (left panels) and the Galactic $R-Z$ plane (right panels). The color scale indicates travel time (Myr), with the red star marking each star's birthplace and the black “x” marking its current position. The red-shaded contours denote the uncertainty regions of the birthplaces.
  • Figure 5: Orbital trajectories for two example stars with coordinates RA: 19.599800, Dec: 43.974800 (top panels) and RA: 140.390003, Dec: 50.098992 (bottom panels), shown in the Galactic $X-Y$ plane (left panels) and the Galactic $R-Z$ plane (right panels). The color scale indicates travel time (Myr), with the red star marking each star's birthplace and the black “x” marking its current position. The red-shaded contours denote the uncertainty regions of the inferred birthplaces.
  • ...and 4 more figures