Simulations of Spiral Galaxies

The outer parts of the rotation curves of external galaxies provide the best evidence for the existence of dark matter. In almost all of them the rotation curve is flat or slowly rising out to the last measured point. Very few galaxies show falling rotation curves, and the ones that do either fall less than Keplerian or have nearby companions that perturb the velocity field. The simplest interpretation of these results is that spiral galaxies possess massive dark haloes that extend to larger radii than the optical disks.

The inner part of the rotation curve is crucial in determining the nature of the dark matter. The shape of the dark spherical halo is determined by the central density and the core radius. Haloes built up by hierarchical merging in dark matter cosmogonies are cusped at the very centre. Simulations by Navarro, Frenk & White (1997) found a nearly universal halo profile which rises as r^{-1} to the centre. These characteristics seem to disagree with a number of observations. Moreover, the number of subhaloes around typical galaxies, as identified by satellite galaxies, is an order of magnitude smaller than predicted by cold dark matter (CDM). The observed rotation curves of dwarf and low surface brightness galaxies seem to indicate that their dark matter haloes have a constant density core instead of steep cusps. For high surface brightness galaxies, the situation is not clear and there is not consensus as to whether bright galaxies are dark matter dominated at their very centres or not.

In order to clarify the problem of the mass distribution in the inner parts of bright galaxies and constrain the galaxy formation scenario, we are carrying out a project to test whether the luminous mass in the inner parts of spiral galaxies can account for their observed gas kinematics or whether an additional dark matter component is required. For this purpose we are planning to model the galaxy dynamics of a significant sample of barred spiral galaxies by running a 3-D composite N-body/hydrocode on the luminous underlying matter distribution of the sample galaxies and then compare it to the observe gas dynamics.

Principal Investigator

Roger Fux



Facilities Used



Ken Freeman
Isabel Perez

Oak-Kyoung Park
Lawrence Livermore National Laboratory

RFCD Codes


Significant Achievements, Anticipated Outcomes and Future Work

Due to the starting nature of the project, no final results have yet been obtained. However, we have explored and set the optimal initial conditions for the simulations. We have concluded that in order to reach gently a stationary state, without wobbling, we have to set the galaxy bar growth time to roughly 1.5 times the bar rotational period. This clashes with other people's claim that a shorter time is enough to reach a stationary state. We are now ready to start all the simulations and soon after we will be able to publish the results.


Computational Techniques Used

We are running a serial 3-D composite N-body/hydrocode based on a code developed by the Geneva Observatory galactic dynamics group. At each time step, it computes gravitational forces on all particles using a particle mesh with fast Fourier transform method. Secondly, it evaluates pressure and artificial viscous forces on the gas particles, using the Lagrangian smooth particle hydrodynamics method to solve the equations of motions for a fluid. Finally, it moves all particles one time step further by integrating the equations of motion with an adaptive time-step algorithm. The maximum resolution used in our test simulations was a grid of 147456 cells and 500000 gas particles. Given the number of galaxies to simulate and the high resolution required it would be impossible to achieve this without the supercomputer capabilities.