Viscosity of hot gas in galaxy clusters
September 01, 2019
Clusters of galaxies are the most massive virialized objects in our Universe. More than 80% of their mass is provided by Dark Matter, while their number density depends on Dark Energy, which affects the expansion of the Universe and the growth of clusters. These links to the Dark Side of the Universe makes clusters an important tool for Cosmology. At the same time, most of the normal (baryonic) matter in clusters is in the form of a hot (100 million degrees) and very tenuous (one particle per 1000 cubic centimeters) plasma permeated with a weak magnetic field. It is hardly possible to produce and study such a plasma in the lab, but clusters offer us this opportunity. In particular, we do not know the thermal conductivity or viscosity of such plasma, and the level of uncertainty can be as high as ten orders of magnitude.
Such uncertainty implies that numerical modeling of clusters’ properties and evolution may not reach the level of accuracy required for identifying subtle effects related to, e.g., non-trivial properties of the Dark Matter or Dark Energy. In numerical simulations, scientists often assume that both the thermal conduction and the viscosity are small. And indeed, plasma physics theory offers a possibility that particles are scattered by the microscopic fluctuations of the magnetic field so often that the viscosity and conductivity become very low. However, another extreme possibility is that only particle collisions are important. In this case, particles would be able to move freely over large distances in the hot and low-density plasma of clusters, transporting momentum between adjacent regions and therefore making the plasma very viscous. Which of the two extremes is closer to reality? Answering this question was the primary goal of very long observations of the Coma cluster with NASA’s Chandra observatory. The results of these observations were published in Nature Astronomy in June 2019.
The gas in the Coma cluster is very hot – almost a hundred million degrees and has a very low density, especially in regions away from the cluster core that Chandra was observing for more than ten days (Fig.1). These are exactly the conditions where the mean free path of particles is very long (tens of kiloparsecs) and viscosity and conduction are expected to smear out any fluctuations of gas velocity, temperature or density on sufficiently small spatial scales. The higher the viscosity the larger are the affected scales. Therefore, by identifying the scales where fluctuations are suppressed, we can infer the viscosity.
With the superb angular resolution of Chandra, measuring the gas density fluctuations that lead to the brightness variations in X-ray images is a straightforward exercise (see Fig.2). The fluctuations found in the Coma cluster fall exactly in the range of scales, which should be severely affected by viscosity. The presence of fluctuations at these small scales falsifies the assumption that the plasma viscosity is high and that it is set by particle collisions. The opposite assumption of strongly suppressed transport coefficients seems to be consistent with observations. Therefore, numerical simulations that ignore the viscosity maybe not far from reality, even although the exact recipe for modeling weakly collisional cluster plasma is yet to be determined. Importantly, these results demonstrate that observations of turbulence in clusters are giving rise to a new branch of astrophysics that can sharpen theoretical views on such plasmas.