Péridier Library Abstract Archive

Abstract No. UT 381

Title: Self-Organized Large-Scale Coherence in Simulations of Galactic Star Formation
Author(s): David Chappell and John Scalo
Keywords: none
E-Mail: John Scalo (to request a full copy of this paper)
Preprint: not yet established
Release date: 05/23/97 11:51:54
Publication status: submitted to Astrophysical Journal
Comments: 36 pages, 8 figures

It is often assumed that galaxies cannot generate large-scale coherent star-forming activity without some organizing agent, such as spiral density waves, bars, large-scale instabilities, or external perturbations due to encounters with other galaxies. We present simulations of a simple model of star formation in which local spatial couplings lead to large-scale coherent, and even synchronized, patterns of star formation without any explicit propagation or any separate organizing agent. At a given location, star formation is assumed to occur when the gas velocity dispersion falls below a critical value dependent on the density. Young stars inject energy into the gas in their neighborhood, increasing the velocity dispersion and inhibiting the instability. A dissipation function continually "cools" the gas. The stability of this local inhibitory feedback model is examined both analytically and numerically. A large number of two-dimensional simulations are used to examine the effect of spatial couplings due to energy injection into neighboring regions. We find that several distinct types of behavior can be demarcated in a phase diagram whose parameter axes are the density (assumed constant in most models) and spatial coupling strength. These "phases" include, with decreasing density, a spatially homogeneous steady state, oscillatory "islands," traveling waves of star formation or global synchronization, and scattered "patches" of star formation activity. The coherence effects are explained in terms of the ability of the energy injected near a star formation site to introduce phase correlations in the subsequent cooling curves of neighboring regions. It is suggested that phases such as these, which depend mostly on the density, may occur in different ranges of galactocentric distance within individual galaxies, and that galaxies as a whole may evolve through different phases as the gas is gradually depleted by star formation, or because the transient time to settle into a given phase may be very large. In particular, the results suggest that galaxies may develop large-scale or global oscillations or bursts in their star formation rates during some stage of evolution, without the necessity for any organizing agent or even propagation. Such activity may explain the large range in present-to-past average star formation rate ratios at a given morphological type found in two recent studies of disk galaxies, and is consistent with the scattered low-level star formation seen in low surface brightness galaxies and the outer disks of higher surface brightness galaxies. The long "incubation time" (~ 10⁹ yr) required for synchronized global oscillations to develop suggests a possible connection with observed redshift-dependent galactic phenomena. The results are also applicable to the evolution of star formation in dwarf galaxies and in even smaller regions, such as large molecular complexes. However, true hydrodynamical spatial coupling, differential rotation, and other effects remain to be examined, and so the models must be interpreted as merely suggestive of the type of self-organized pattern formation phenomena that might occur in real galaxies. The possible generic relation to synchronization phenomena in models for excitable biological systems, especially systems of integrate-and-fire formal neurons, is discussed.