COSMIC BULLETS AS PARTICLE ACCELERATORS AND RADIO-SOURCES

We have simulated in two dimensions the dynamical evolution of dense gas clouds (''cosmic bullets'') moving supersonically through a uniform low-density medium. The diffusive shock acceleration of relativistic protons (cosmic rays) and their dynamical feedback on the background flow are included by the two-fluid model for this process. The acceleration of relativistic electrons is approximated by a test-particle model, and a passive magnetic field is followed by a simple advection scheme. Strong bow shocks, with Mach numbers similar to that of a bullet's motion, are the most important particle accelerators in the flow, while tail shocks and shocks inside the bullets do not play generally significant roles in this regard. Analogous to properties seen in previous time-dependent diffusive shock acceleration studies, the evolution of cosmic-ray-modified bullet shocks and the particle acceleration around them are dependent mainly upon initial conditions, the ratio of the age of the bullet to the particle acceleration timescale within the shocks, and the ratio of the bullet size to the particle diffusion length. For our simulation parameters, greater than or similar to 10% of the initial bullet kinetic energy is converted to a combination of internal energy of gas and cosmic-ray protons by the time the bullets begin to be disrupted. The fraction of that energy transfer going to the cosmic rays can in some (but not all) cases exceed that transferred to the gas. Characteristically, the cosmic rays gain several percent of the available kinetic energy. Bullet destruction on timescales only a little larger than the ram pressure bullet crushing time begins in response to Kelvin-Helmholtz and especially to Rayleigh-Taylor instabilities along the forward bullet surface. For dense bullets this happens before the bullet is stopped by ram pressure. Development of strongly sheared flows may greatly amplify an embedded magnetic field. That could alter or impede the final bullet destruction. According to our simple model for synchrotron emission from relativistic electrons accelerated and transported within the flows, that emission increases rapidly as the bullet begins to fragment, when it is strongly dominated by field enhancement in sheared flows. Synchrotron emission from the acceleration region within the bow shock is, by contrast, much weaker. This means that attempts to identify bright ''knot'' radio features with bow shocks may be misleading and that a more sophisticated understanding of the dynamical state of the material, the magnetic field, and the nature of acceleration processes other than first-order shock acceleration is necessary to interpret these important laboratories pro