The properties of cooling flows in X-ray luminous clusters of galaxies

We discuss the X-ray properties of the cooling flows in a sample of 30 highly X-ray luminous clusters of galaxies, observed using the ASCA and ROSAT satellites. We demonstrate the need for multiphase models to consistently explain the spectral and imaging X-ray data for the clusters. The mass deposition rates of the cooling flows, independently determined from the ASCA spectra and ROSAT images, exhibit good agreement and exceed 1000 M_⊙ yr⁻¹ in the largest systems. We confirm the presence of intrinsic X-ray absorption in the clusters using a variety of spectral models. The measured equivalent hydrogen column densities of absorbing material are sensitive to the spectral models used in the analysis, varying from average values of a few 10²⁰ atom cm⁻² for a simple isothermal emission model, to a few 10²¹ atom cm⁻² using our preferred cooling flow models, assuming in each case that the absorber lies in a uniform foreground screen. The true intrinsic column densities are likely to be even higher if the absorbing medium is distributed throughout the clusters. We summarize the constraints on the nature of the X-ray absorber from observations in other wavebands. Much of the X-ray absorption may be caused by dust.     

The impact of cooling flows on the TX–LBol relation for the most luminous clusters

We examine the effects of cooling flows on the T _X– L _Bol relation for a sample of the most X-ray luminous ( L _Bol > 10⁴⁵ erg s⁻¹) clusters of galaxies known. Using high-quality ASCA X-ray spectra and ROSAT images we explicitly account for the effects of cooling flows on the X-ray properties of the clusters and show that this reduces the previously noted dispersion in the T _X– L _Bol relationship. More importantly, the slope of the relationship is flattened from L _Bol ∝  T ³_X to approximately L _Bol ∝  T ²_X, in agreement with recent theoretical models which include the effects of shocks and pre-heating on the X-ray gas. We find no evidence for evolution in the T _X– L _Bol relation within z  ∼ 0.3. Our results demonstrate that the effects of cooling flows must be accounted for before cosmological parameters can be determined from X-ray observations of clusters. The results presented here should provide a reliable basis for modelling the T _X– L _Bol relation at high X-ray luminosities.     

The 6.4-keV fluorescent iron line from cluster cooling flows

The fate of the cooling gas in the central regions of rich clusters of galaxies is not well understood. In one plausible scenario clouds of atomic or molecular gas are formed. However the mass of the cold gas, inferred from measurements of low-energy X-ray absorption, is hardly consistent with the absence of powerful CO or 21-cm emission lines from the cooling flow region. Among the factors which may affect the detectability of the cold clouds are their optical depth, shape and covering fraction. Thus, alternative methods to determine the mass in cold clouds, which are less sensitive to these parameters, are important.   For the inner region of the cooling flow (e.g. within a radius of ∼50–100 kpc) the Thomson optical depth of the hot gas in a massive cooling flow can be as large as ∼ 0.01. Assuming that the cooling time in the inner region is few times shorter than the lifetime of the cluster, the Thomson depth of the accumulated cold gas can be accordingly higher (if most of the gas remains in the form of clouds). The illumination of the cold clouds by the X-ray emission of the hot gas should lead to the appearance of a 6.4-keV iron fluorescent line, with an equivalent width proportional to τ_T. The equivalent width only weakly depends on the detailed properties of the clouds, e.g. on the column density of individual clouds, as long as the column density is less than a few 10²³ cm⁻². Another effect also associated exclusively with the cold gas is a flux in the Compton shoulder of bright X-ray emission lines. It also scales linearly with the Thomson optical depth of the cold gas. With the new generation of X-ray telescopes, combining large effective area and high spectral resolution, the mass of the cold gas in cooling flows (and its distribution) can be measured.     

ASCA and ROSAT observations of nearby cluster cooling flows

We present a detailed analysis of the X-ray properties of the cooling flows in a sample of nearby, X-ray-bright clusters of galaxies using high-quality ASCA spectra and ROSAT X-ray images. We demonstrate the need for multiphase models to consistently explain the spectral and imaging X-ray data for the clusters. The mass deposition rates of the cooling flows, independently determined from the ASCA spectra and ROSAT images, exhibit reasonable agreement. We confirm the presence of intrinsic X-ray absorption in the clusters using a variety of spectral models. We also report detections of 100-μm infrared emission, spatially coincident with the cooling flows, in several of the systems studied. The observed infrared fluxes and flux limits are in good agreement with the predicted values owing to reprocessed X-ray emission from the cooling flows. We present precise measurements of the abundances of iron, magnesium, silicon and sulphur in the central regions of the Virgo and Centaurus clusters. Our results firmly favour models in which a high mass fraction (70–80 per cent) of the iron in the X-ray gas in these regions originates from Type Ia supernovae. Finally, we present a series of methods which may be used to estimate the ages of cooling flows from X-ray data. The results for the present sample of clusters indicate ages of between 2.5 and 7 Gyr. If the ages of cooling flows are primarily set by subcluster merger events, then our results suggest that in the largest clusters, mergers with subclusters with masses of ∼30 per cent of the final cluster mass are likely to disrupt cooling flows.     

Cooling flows induced by Compton cooling resulting from luminous quasars in clusters of galaxies

We have studied the effects of Compton cooling on cooling flow by performing numerical hydrodynamic calculations of the time evolution of hot gas in clusters of galaxies with luminous quasars. We assumed various temperatures for the hot intracluster gas. We have shown that the Compton cooling resulting from very luminous quasars is effective in inducing cooling flow, before radiative cooling flow is realized. The mass flux resulting from the cooling flow increases with time, as the Compton-cooled region expands. However, the mass flux of the Compton cooling flow is not large, less than 1 M⊙ yr⁻¹ in our model, because the Compton-cooled region is limited to an inner galactic region around a quasar. Even after the quasar active phase has ceased, the cooling flow will continue for at least 10⁹ yr. The accreted mass is enough to explain X-ray absorption lines in high-redshift quasars, if the Compton-cooled gas is compressed by high-pressure intracluster gas.     

X-ray colour maps of the cores of galaxy clusters

We present an analysis of X-ray colour maps of the cores of clusters of galaxies, formed from the ratios of counts in different X-ray bands. Our technique groups pixels lying between contours in an adaptively smoothed image of a cluster. We select the contour levels to minimize the uncertainties in the colour ratios, whilst preserving the structure of the object. We extend the work of Allen & Fabian by investigating the spatial distributions of cooling gas and absorbing material in cluster cores. Their sample is almost doubled: we analyse archive ROSAT Position Sensitive Proportional Counter (PSPC) data for 33 clusters from the sample of the 55 brightest X-ray clusters in the sky. Many of our clusters contain strong cooling flows. We present colour maps of a sample of the clusters, in addition to adaptively smoothed images in different bands. Most of the cooling flow clusters display little substructure, unlike several of the non-cooling-flow clusters.
We fitted an isothermal plasma model with galactic absorption and constant metallicity to the mid-over-high energy colours in our clusters. Those clusters with known strong cooling flows have inner contours which fit a significantly lower temperature than the outer contours. Clusters in the sample without strong cooling flows show no significant temperature variation. The inclusion of a metallicity gradient alone was not sufficient to explain the observations. A cooling flow component plus a constant temperature phase did account for the colour profiles in clusters with known strong cooling flow components. We also had to increase the levels of absorbing material to fit the low-over-high colours at the cluster centres. Our results provide more evidence that cooling flows accumulate absorbing material. No evidence for increased absorption was found for the non-cooling-flow clusters.     

The missing soft X-ray luminosity in cluster cooling flows

The gas temperature in the cores of many clusters of galaxies drops inward by about a factor of 3 or more within the central 100-kpc radius. The radiative cooling time drops over the same region from 5 or more Gyr down to below a few 10⁸ yr. Although this indicates that cooling flows are taking place, XMM-Newton spectra show no evidence for strong mass cooling rates of gas below  1–2 keV  . The soft X-ray luminosity expected from steady cooling flows is missing. Here we outline and test the energetics of a cold mixing model in which gas below  1–2 keV  falls from the flow and is rapidly cooled by mixing with cold gas. The missing X-ray luminosity can emerge in the ultraviolet, optical and infrared bands, where strong emission nebulosities are commonly seen. We explore further the requirements for any heat sources that balance the radiative cooling in cluster cores.     

Chandra temperature and metallicity maps of the Perseus cluster core

We present temperature and metallicity maps of the Perseus cluster core obtained with the Chandra X-ray Observatory. We find an overall temperature rise from  ∼3.0 keV  in the core to  ∼5.5 keV  at 120 kpc and a metallicity profile that rises slowly from  ∼0.5  solar to  ∼0.6  solar inside 60 kpc, but drops to  ∼0.4  solar at 120 kpc. Spatially resolved spectroscopy in small cells shows that the temperature distribution in the Perseus cluster is not symmetrical. There is a wealth of structure in the temperature map on scales of  ∼10  arcsec (5.2 kpc) showingswirliness and a temperature rise that coincides with a sudden surface brightness drop in the X-ray image. We obtain a metallicity map of the Perseus cluster core and find that the spectra extracted from the two central X-ray holes as well as the western X-ray hole are best-fit by gas with higher temperature and higher metallicity than is found in the surroundings of the holes. A spectral deprojection analysis suggests, however, that this is due to a projection effect; for the northern X-ray hole we find tight limits on the presence of an isothermal component in the X-ray hole, ruling out volume-filling X-ray gas with temperatures below 11 keV at 3σ.     

On the soft X-ray spectrum of cooling flows

Strong evidence for cooling flows has been found in low-resolution X-ray imaging and spectra of many clusters of galaxies. However, high-resolution X-ray spectra of several clusters from the Reflection Grating Spectrometer on XMM-Newton now show a soft X-ray spectrum inconsistent with a simple cooling flow. The main problem is a lack of the emission lines expected from gas cooling below 1–2 keV. Lines from gas at about 2–3 keV are observed, even in a high-temperature cluster such as A1835, indicating that gas is cooling down to about 2–3 keV, but is not found at lower temperatures. Here we discuss several solutions to the problem: heating, mixing, differential absorption and inhomogeneous metallicity. Continuous or sporadic heating creates further problems, including the targeting of the heat at the cooler gas and also the high total energy required. So far there is no clear observational evidence for widespread heating, or shocks, in cluster cores, except in radio lobes which occupy only part of the volume. Alternatively, if the metals in the intracluster medium are not uniformly spread but are clumped, then little line emission is expected from the gas cooling below 1 keV. The low-metallicity part cools without line emission, whereas the strengths of the soft X-ray lines from the metal-rich gas depend on the mass fraction of that gas and not on the abundance, since soft X-ray line emission dominates the cooling function below 2 keV.     

A statistically selected Chandra sample of 20 galaxy clusters – I. Temperature and cooling time profiles

We present an analysis of 20 galaxy clusters observed with the Chandra X-ray satellite, focusing on the temperature structure of the intracluster medium and the cooling time of the gas. Our sample is drawn from a flux-limited catalogue but excludes the Fornax, Coma and Centaurus clusters, owing to their large angular size compared to the Chandra field of view. We describe a quantitative measure of the impact of central cooling, and find that the sample comprises nine clusters possessing cool cores (CCs) and 11 without. The properties of these two types differ markedly, but there is a high degree of uniformity amongst the CC clusters, which obey a nearly universal radial scaling in temperature of the form   T ∝ r ^∼0.4  , within the core. This uniformity persists in the gas cooling time, which varies more strongly with radius in CC clusters  ( t _cool∝ r ^∼1.3)  , reaching   t _cool < 1 Gyr  in all cases, although surprisingly low central cooling times (<5 Gyr) are found in many of the non-CC systems. The scatter between the cooling time profiles of all the clusters is found to be remarkably small, implying a universal form for the cooling time of gas at a given physical radius in virialized systems, in agreement with recent previous work. Our results favour cluster merging as the primary factor in preventing the formation of CCs.     

The entropy and energy of intergalactic gas in galaxy clusters

Studies of the X-ray surface brightness profiles of clusters, coupled with theoretical considerations, suggest that the breaking of self-similarity in the hot gas results from an 'entropy floor', established by some heating process, which affects the structure of the intracluster gas strongly in lower-mass systems. By fitting analytical models for the radial variation in gas density and temperature to X-ray spectral images from the ROSAT PSPC and ASCA GIS, we have derived gas entropy profiles for 20 galaxy clusters and groups. We show that, when these profiles are scaled such that they should lie on top of one another in the case of self-similarity, the lowest-mass systems have higher-scaled entropy profiles than more massive systems. This appears to be due to a baseline entropy of depending on the extent to which shocks have been suppressed in low-mass systems. The extra entropy may be present in all systems, but is detectable only in poor clusters, where it is significant compared with the entropy generated by gravitational collapse. This excess entropy appears to be distributed uniformly with radius outside the central cooling regions.
We determine the energy associated with this entropy floor, by studying the net reduction in binding energy of the gas in low-mass systems, and find that it corresponds to a pre-heating temperature of 0.3 keV. Since the relationship between entropy and energy injection depends upon gas density, we are able to combine the excesses of 70140 keV cm² and 0.3 keV to derive the typical electron density of the gas into which the energy was injected. The resulting value of implies that the heating must have happened prior to cluster collapse but after a redshift z 710. The energy requirement is well matched to the energy from supernova explosions responsible for the metals which now pollute the intracluster gas.