Gas Ejection from Spiral Galaxy Disks
Abstract
We present the results of three proposed mechanisms for ejection of gas from a spiral arm into the halo. The mechanisms were modelled using magnetohydrodynamics (MHD) as a theoretical template. Each mechanism was run through simulations using a Fortran code: ZEUS-3D, an MHD equation solver. The first mechanism modelled the gas dynamics with a modified Hartmann flow which describes the fluid flow between two parallel plates. We initialized the problem based on observation of lagging halos; that is, that the rotational velocity falls to a zero at some height above the plane of the disk. When adopting a density profile which takes into account the various warm and cold H$_I$ and H$_{II}$ molecular clouds, the system evolves very strangely and does not reproduce the steady velocity gradient observed in edge-on galaxies. This density profile, adopted from Martos and Cox (1998), was used in the remaining models. However, when treating a system with a uniform density profile, a stable simulation can result. Next we considered supernova (SN) blasts as a possible mechanism for gas ejection. While a single SN was shown to be insufficient to promote vertical gas structures from the disk, multiple SN explosions proved to be enough to promote gas ejection from the disk. In these simulations, gas ejected to a height of 0.5 kpc at a velocity of 130 km s$^{-1}$ from 500 supernovae, extending to an approximate maximum height of 1 kpc at a velocity of $6.7 \times 10^3$ km s$^{-1}$ from 1500 supernovae after 0.15 Myr, the approximate time of propagation of a supernova shock wave. Finally, we simulated gas flowing into the spiral arm at such a speed to promote a jump in the disk gas, termed a hydraulic jump. The height of the jump was found to be slightly less than a kiloparsec with a flow velocity of 41 km s$^{-1}$ into the halo after 167 Myr.
The latter models proved to be effective mechanisms through which gas is ejected from the disk whereas the Hartmann flow (or toy model) mechanism remains unclear as the heliocentric velocity profile becomes unstable when run through a time-dependent simulation. Though the cause of this instability is unclear, pressure fluctuations in the system are suspected to play a role.