Inevitably, in this process of infall along filaments, there will sometimes be more dramatic collisions between massive clusters of comparable sizes, such as illustrated in Figure 3. This is the well-known “Bullet cluster” where the broght bow-shaped structure is a pressure- matched discontinuity. The brighter regions correspond to the cool, dense remnants of the original clusters. Beyond these are the fainter, but hotter regions heated by shocks formed in the > 1057 Joules encounter. In this spectacular collision, the cluster galaxies and dark matter have low enough cross sections that they passed right through each other, while the diffuse plasma bears the brunt of energy dissipation.
 |
Figure 3. X-ray emission from the Bullet Cluster, courtesy M. Markevitch. |
Such shocks and the accompanying post-shock turbulence can generate a rich variety of phenomena in the ICM. A spectacular example is seen in Abell 2256, (Figure 4), the ongoing merger of two or three massive clusters. Temperature maps of the X-ray emitting ICM show several pressure-matched discontinuities, different in shape but similar in physical properties, to the “bullet” shown above. Coincident with the X-ray emission is a benign looking white contour, enclosing a region of diffuse, low brightness radio synchrotron emission, which we call a “radio halo.” It reveals that the smooth-looking X-ray emission is deceptive, because the ICM must be highly turbulent in order to accelerate the cosmic ray electrons responsible for the radio emission. At observed frequencies around 1 GHz, these ∼1 GeV electrons will radiate away their energy in ∼ 108 years, which means that they cannot move far from their energization sites. CR electron acceleration must therefore be happening throughout the cluster — and that puts the burden on otherwise invisible turbulence in the ICM.
 |
Figure 4. Abell 2256, in the 2-10 keV X-ray band (purple), near infrared band (orange) and radio synchrotron (green, [6]). The white contour indicates the extent of the diffuse radio “halo”. The white patches with dark spots indicate a few of the cluster galaxies that host a radio structure bent back by relative motion through the ICM. |
In Figure 4, the large filamentary radio structure in the upper right is assumed to be associated with shocks generated in the cluster collision, although in this case, the X-ray shock structure has yet to reveal its presence. Such synchrotron features are often found on the periphery of cluster X-ray emission and are termed “radio relics” for historical reasons. Individual galaxies in the clusters are themselves also sources of radio emission, and their swept back appearance indicates relative motion through the ICM. In some cases, sharp deflections in these radio tails, or curious features such as the ring-like structure in the left of Figure 4 can point to otherwise invisible flows in the ICM Such indications of “weather” in the environment surrounding radio galaxies are spotted occasionally in cluster mapping projects; an emerging source of such critical probes are the sharp-eyed “citizen scientists” participating in projects such as Radio Galaxy Zoo.
The observed and inferred shocks and turbulence in the ICM, and features such as radio halos and relics appear almost exclusively in clusters that have recently experienced major mergers. But what is the mechanism by which they transfer energy to a small and highly energetic population of CRs? The answer lies in the small-scales of the turbulence, where the repeated scattering of electrons by self-excited and other MHD waves slowly restores their radiated away energy. Simultaneously, the magnetic fields required for synchrotron radiation are seeded by processes currently unknown, and then amplified largely through stretching, as they are sheared in the turbulent ICM. Approaching such a complex set of interconnected processes requires numerical simulations, such as illustrated in Figure 5.
 |
Figure 5. Numerical MHD simulations of merging clusters showing widespread amplification of the magnetic field. (Courtesy of F. Vazza, University of Bologna, see https://vimeo.com/289054145). |
Such simulations, with increasing support from both radio and X-ray observations, indicate that the turbulence and its consequences are relatively more important in the outer, lower density regions, as opposed to cluster cores where the thermal energy is highly dominant. Extending such simulations to predict the distribution of synchrotron radiation requires understanding the microphysics of cosmic ray acceleration. In particular, we need to know how the turbulent energy is dissipated on the smallest scales, far below the resolution of cluster-wide numerical simulations. Processes such as the relative roles of Alfvenic, compressive, solenoidal turbulence and CR self-excited modes are under intense study. Interested readers are commended to the review article by Gianfranco Brunetti and Tom Jones [7].
Much of the ICM turbulence is likely generated downstream from shocks and gas perturbations that are produced during the merger process. In numerical simulations, a dominant pair of opposite moving shocks often forms. When these reach the peripheries of the merging cluster they become separated from the rest of the cluster emission and become more accessible to observations. One such shock, dubbed the “Toothbrush” because of its morphology, is shown in Figure 6. Here we see the radio emission generated immediately behind the shock, and a profile plot of the corresponding X-ray brightness.
 |
Figure 6. Radio emission from the Toothbrush Relic and its approximate correspondence with the X-ray brightness profile. See [8] and references therein. |
The X-ray and radio emission each give us a way to estimate the shock Mach number, a key parameter in determining both what has taken place during the merger and how the ICM will be affected. In the X-ray, the Rankine-Hugoniot jump conditions can be applied pre- and post- shock to deprojected values of either the ICM density and temperature, or preferably both. In the case shown here, the X-ray shock is quite modest, trans-sonic, with a Mach number of ∼ 1.3. Projection and smearing effects often lead to a factor of 2 uncertainty in the actual amount of compression in such shocks.
The radio emission provides an alternative insight into the shock structure. If the enhanced radio emissivity post-shock arises from diffusive shock (or Fermi type I) acceleration, then the logarithmic slope of the synchrotron radiation spectrum will reach a steady-state value that depends on the Mach number. Observations of the Toothbrush reveal that instead of the low X- ray derived Mach number, the synchrotron derived value is M ∼ 3.5. To first order, it's remarkable that these simple first-order theories with very different underlying physics lead to similarly low Mach numbers. At second order, the apparent discrepancy between X-ray- and radio- derived values have led us to look deeper into our models.
One additional consideration is whether all of the radiating CR electrons are generated at the shock, or whether some actually come from a pre-existing population left behind by cluster radio galaxies. This latter possibility has the extra advantage of avoiding the low efficiency that low Mach number shocks have for accelerating CR electrons out of the thermal pool. Before electrons can be efficiently accelerated at the shock, e.g., they must have gyro-radii large enough to carry them back and forth across the shock front. Whether mechanisms such as shock drift acceleration can be effective, or whether pre-existing seed populations are required, is yet to be determined. Pre-existing electrons will change the radio spectral signatures, so better radio- and X-ray- observational diagnostics and a range of numerical experiments are the order of the day.