Research Highlight
June 1, 2005
Modeling Formation of Galaxy Clusters
by Daisuke Nagai

Picture:
Modeling Formation of Galaxy Clusters
Clusters of galaxies can serve as a powerful cosmological probe of the nature of mysterious dark matter and dark energy. However, to derive tight cosmological constraints, it is critical to understand the observable properties and evolution of clusters in detail. KICP members Daisuke Nagai and Andrey Kravtsov are modelling the formation and evolution of galaxy clusters using supercomputer simulations designed to follow the evolution of dark matter, gas and stars starting from realistic cosmological initial conditions. These simulations provide new insights into the physical processes operating during cluster evolution and interpretation of the recent and upcoming X-ray, Sunyaev-Zeldovich, and optical observations of clusters.

Picture:
<b>Cosmological Cluster Simulations</b>: Clusters of galaxies consist of approximately 85% of dark matter, 10% of hot gas and 2-5% of stars. Detailed theoretical modeling of clusters is thus a complicated astrophysics problem involving a variety of physical phenomena from the nonlinear collapse and merging of dark matter to radiative cooling of gas, star formation, chemical enrichment of the intergalactic medium by supernovae and energy feedback. The numerical simulations of cluster formation performed at the KICP start from the well-defined cosmological initial conditions and use the technique called Adaptive Mesh Refinement to greatly increase the resolution in the high density regions within the clusters. The high resolution is required to follow formation and evolution of cluster galaxies and their impact on the hot intracluster medium. The panel illustrates the complexity of physical processes involved in a typical simulation. It shows distribution of dark matter, stars, and the properties of gas at the present epoch in one of the simulated clusters forming in the concordance Cold Dark Matter model with vacuum energy. In the hierachical structure formation scenario, clusters form and grow through continuous mergers and accretion of small clumps and diffuse matter from voids and filaments. The figure reveals a rich and complex structure of the gas density and temperature distributions, such as strong and highly aspherical accretion shocks surrounding the cluster and turbulent gas motions within the cluster. The cluster gas is also enriched with heavy elements ("metals"), as the metal-enriched gas is stripped off from galaxies when they orbit within the cluster.

Picture:
<b>Properties of Cluster Galaxies</b>: To assess the effects of galaxy formation on the properties of cluster gas, one has to be confident that galaxies themselves are modelled correctly. One way of ensuring this is through a variety of comparisons of simulated galaxies with their observed counterparts. Here we describe one example in which radial distribution of simulated galaxies around the cluster center is compared to observations. In <a target="_blank" href="http://www.journals.uchicago.edu/ApJ/journal/issues/ApJ/v618n2/61141/61141.html">a recent paper</a>, Nagai and Kravtsov showed that a process called "tidal stripping" operates with different efficiency on dark matter and stars associated with cluster galaxies. When the galaxies and their dark matter accrete onto the cluster, the dark matter, owing to its more extended distribution, is stripped more efficiently than the stars that are located near the center of the galaxy and are tightly bound by gravity. Consequently, the radial distribution of galaxies depends on whether the galaxies are selected using their dark matter or stellar mass. Selection based on the dark matter mass favors galaxies less affected by tidal mass, located preferentially on the outskirts of clusters. This results in a more extended, "flatter" radial distribution. Selection based on the stellar mass does not suffer from such "selection bias" and results in a more centrally concentrated distribution. This difference in selection is important when one compares simulations to the data. The figure shows that the surface density of galaxies in projected radial bins does not agree with the optical/IR observations of cluster galaxies (triangles and solid circles), if the simulated galaxies are selected using their total mass (blue dashed line). The radial distribution of galaxies selected using stellar masses (red dotted line), on the other hand, is in good agreement with the data. The stellar mass selection is similar to the selection using the K-band luminosity of galaxies used to construct the observed radial profile. It is therefore not too surprising that stellar mass selection works better. Nevertheless, this example illustrates the importance of star formation in realistic cluster simulations.

Picture:
<b>The Baryon Fractions in Clusters</b>: The observed clusters are massive and dense. This means that in the past, when the density everywhere in the universe was close to the average, the cluster matter was spread over a large (tens of megaparsecs) region of space. In fact the initial cluster volume is so large that the clusters are expected to contain a representative mix of dark matter and baryons, close to the cosmic average (in professional jargon, the universal baryon fraction). Kravtsov, Nagai and their collaborator, Alexey Vikhlinin of the Center for Astrophysics at Harvard, recently showed that gas cooling and galaxy formation modify the baryon fractions in clusters. The effects are large enough to have a significant impact on the interpretations of measurements and derived cosmological constraints. The same processes also modify the cluster gas fractions because a fraction of gas is converted into stars. The results of the simulations are compared to the recent measurements of cluster gas fractions from the Chandra X-ray observations. These comparisons show that the gas fractions in the simulations with cooling and star formation (blue solid circles) match the data (stars with error bars) remarkably well. A paper describing these results appeared in the June 2005 issue of the Astronomical Journal and at <a target="_blank" href="http://xxx.lanl.gov/abs/astro-ph/0501227">Astrophysics abstracts</a>.

Picture:
<b>Cosmology with Sunyaev-Zeldovich Effect</b>: At the KICP, theorists and experimentalists are working together to understand the relations between the observable cluster properties and their total mass. The figure illustrates that a tight relation exists between the <a target="_blank" href="http://astro.uchicago.edu/sza/primer.html">Sunyaev-Zeldovich Effect (SZE)</a> signal and cluster gas mass. The numerical simulations performed by Daisuke Nagai are used to study the effects of galaxy formation on the SZ scaling relations, and these theoretical predictions are tested using SZE+Chandra cluster observations for a sample of 36 clusters obtained by a group led by the KICP faculty John Carlstrom. These comparisons show that simulations neglecting galaxy formation (red triangles and dotted line) are inconsistent with the observed correlation. The simulations that include gas cooling and star formation (blue triangles and dashed line), on the other hand, are in good agreement with the data. This comparison highlights the importance of inclusion of galaxy formation in cluster simulations and shows that the current generation of simulations produce clusters with gross properties remarkably similar to their observed counterparts.

Related Links:
KICP Members: Andrey V. Kravtsov
KICP Students: Daisuke Nagai