2008ApJ...679..379S

The predicted central densities of dark matter halos in ΛCDM models exceed those observed in some galaxies. Weinberg & Katz argue that angular momentum transfer from a rotating bar in the baryonic disk can lower the halo density, but they also contend that N-body simulations of this process will not reveal the true continuum result unless many more than the usual numbers of particles are employed. Adopting their simplified model of a rotating rigid bar in a live halo, I have been unable to find any evidence to support their contention. I find that both the angular momentum transferred and the halo density change are independent of the number of particles over the range usually employed, up to that advocated by these authors. I show that my results do not depend on any numerical parameters and that field methods perform equally with grid methods. I also identify the reasons that the required particle number suggested by Weinberg & Katz is excessive. I further show that when countervailing compression by baryonic settling is ignored, moderate bars can reduce the mean density of the inner halo by 20%-30%. Long, massive, skinny bars can reduce the mean inner density by a factor ∼10. The largest density reductions are achieved at the expense of removing most of the angular momentum likely to reside in the baryonic component. Compression of the halo by baryonic settling must reduce, and may even overwhelm, the density reduction achievable by bar friction.