X-ray analysis of a complete sample of galaxy clusters

Abstract

Modern cosmology tries to trace the history and predict the future of our Universe. Great achievements have been made by the discovery of the accelerated expansion, the claim for Dark Matter and Dark Energy, the detection and interpretation of the cosmic microwave background (CMB), or the study of the visible large scale structure of our Universe. <br /> Galaxy clusters can be described as the most massive and gravitationally bound systems that evolved out of collapsed overdensities in the early Universe, and therefore witness its history. Observationally one can study galaxy clusters at almost any wavelength, but the X-ray regime takes a key role due to the visibility of the hot plasma between the galaxies, which is the most massive visible component. The aim of this work is to analyze a complete sample of galaxy clusters in detail and constrain cosmological parameters, like the matter density, OmegaM, or the amplitude of initial density fluctuations, sigma8. The purely flux limited sample (HIFLUGCS) consists of the 64 X-ray brightest galaxy clusters, which are excellent targets to study the systematic effects, which can bias results and lead to wrong conclusions.<br /> With current X-ray observatories like Chandra and XMM-Newton, galaxy clusters can be analyzed in detail, e.g., by measuring the plasma temperature and the surface brightness to constrain the total gravitating mass. This quantity is of extraordinary importance for cosmological studies. Unfortunately, the calibration of X-ray instruments is challenging, because of the absence of absolute calibration targets. <br /> In a cross calibration study of the instruments onboard Chandra and XMM-Newton using the HIFLUGCS galaxy clusters sample, I find that systematic differences exist, which cause temperature measurements to deviate significantly and systematically: Chandra ACIS gives higher temperatures than any XMM-Newton EPIC detector (MOS1, MOS2, PN), and the difference increases with increasing temperature (23% between ACIS and EPIC-PN in the full energy band at 10 keV plasma temperature). Even the three EPIC detectors do not agree with each other. In the hard energy band the differences are not significant. Systematics like the different angular resolutions or possible multitemperature structure of the gas do not explain the observed differences. Tests such as a comparison of the soft band absorption by the Milky Way ISM (free-NH), or the consistency between soft and hard band temperatures of the same instrument seem to theoretically be able to select the best calibrated instrument, but due to large uncertainties (abundance distribution of heavy elements, multitemperature structure) no clear conclusion can be made.<br /> In a second part I derive the total (hydrostatic) and gas masses of all HIFLUGCS clusters individually from the X-ray data.<br /> The cosmological analysis of the HIFLUGCS masses using Chandra data (HICOSMO) involves a likelihood estimation of a halo mass function with a Markov Chain Monte Carlo algorithm. The result is OmegaM = 0.168 +0.021 -0.019 and sigma8 = 0.898 +0.051 -0.048, assuming a flat LambdaCDM Universe. Since the sample consists of local clusters, no tight constraints can be made for the Dark Energy. The gas mass fraction of each galaxy cluster is also compared with simulations, which constrain OmegaM = 0.246 +0.007 -0.007. Since the halo mass function results deviate from current CMB anisotropy results, several tests to understand the systematics involved are performed. Subsamples containing only the high redshift clusters (z > 0.05) or relaxed objects seem to show more agreement with WMAP9 results, especially when adding a hydrostatic bias of 10-30%, which increases all hydrostatic masses due to nonthermal pressure in the gas. Also galaxy groups, which are not represented in the high redshift sample, seem to influence results, as shown by the decrease of the predicted number of groups in the halo mass function, which also shifts results toward the CMB constraints. The more dominant influence of baryonic physics in galaxy groups cannot be proven to solve the discrepancy, since a mass function including feedback, gas heating and radiative cooling has almost no influence on the results. Since also the fgas test suggests an insufficient modeling of low mass systems by the simulations, this aspect needs to be studied in more detail in future hydrodynamic simulations. <br /> The final chapter looks in more detail at one galaxy cluster (Z8338) experiencing an interesting interaction of the cluster ICM with the gas of an infalling galaxy, which leads to one of the longest X-ray tails ever observed. In this case, the tail, caused by stripping of galactic gas, is offset from the galaxy, which has never been observed so clearly before, but predicted by simulations. This scenario can show how the cluster environment is enriched with heavy elements and helps to understand the evolution of galaxy clusters.<br /> Future prospects include the study of the larger flux limited sample (eHIFLUGCS) to further quantify the effect of galaxy groups and reduce the uncertainties by the 50% increase in statistical power. Instrumental uncertainties can be solved, e.g., by incorporating the emission line ratio temperatures in the analysis and increase the number of objects for the free-NH test.