Simulations of the formation of stellar discs in the Galactic centre via cloud-cloud collisions

TL;DR

This study uses numerical simulations to explore how cloud-cloud collisions near the Galactic center can produce the observed young stellar discs, matching their structure, warping, and stellar distribution.

Contribution

It demonstrates that collisions of massive gaseous clouds can form stellar discs with properties consistent with observations, providing a plausible formation scenario.

Findings

Simulations produce nearly circular gaseous discs and eccentric filaments matching observations.

Discs are warped by 30-60 degrees, consistent with recent data.

Radial stellar mass distribution follows a steep R^-2 profile.

Abstract

Young massive stars in the central parsec of our Galaxy are best explained by star formation within at least one, and possibly two, massive self-gravitating gaseous discs. With help of numerical simulations, we here consider whether the observed population of young stars could have originated from a large angle collision of two massive gaseous clouds at R approx. 1 parsec from Sgr A*. In all the simulations performed, the post-collision gas flow forms an inner, nearly circular gaseous disc and one or two eccentric outer filaments, consistent with the observations. Furthermore, the radial stellar mass distribution is always very steep, Sigma proportional to R^-2, again consistent with the observations. All of our simulations produce discs that are warped by between 30 to 60 degrees, in accordance with the most recent observations. The 3D velocity structure of the stellar distribution is sensitive to initial conditions (e.g., the impact parameter of the clouds) and gas cooling details. For example, the runs in which the inner disc is fed intermittently with material possessing fluctuating angular momentum result in multiple stellar discs with different orbital orientations, contradicting the observed data. In all the cases the amount of gas accreted by our inner boundary condition is large, enough to allow Sgr A* to radiate near its Eddington limit over approx. 10^5 years. This suggests that a refined model would have physically larger clouds (or a cloud and a disc such as the circumnuclear disc) colliding at a distance of a few parsecs rather than 1 parsec as in our simulations.