Shock heating of A521

Shock heating of the merging galaxy cluster A521


A521 is an interacting galaxy cluster located at z=0.247, hosting a low frequency radio halo connected to an eastern radio relic. Previous Chandra observations hinted at the presence of an X-ray brightness edge at the position of the relic, which may be a shock front. We analyze a deep observation of A521 recently performed with XMM-Newton in order to probe the cluster structure up to the outermost regions covered by the radio emission. The cluster atmosphere exhibits various brightness and temperature anisotropies. In particular, two cluster cores appear to be separated by two cold fronts. We find two shock fronts, one that was suggested by Chandra and that is propagating to the east, and another to the southwestern cluster outskirt. The two main interacting clusters appear to be separated by a shock heated region, which exhibits a spatial correlation with the radio halo. The outer edge of the radio relic coincides spatially with a shock front, suggesting this shock is responsible for the generation of cosmic ray electrons in the relic. The propagation direction and Mach number of the shock front derived from the gas density jump, , are consistent with expectations from the radio spectral index, under the assumption of Fermi I acceleration mechanism.

Subject headings:
galaxies: clusters: general, — galaxies: clusters: individual( (catalog A521)), — galaxies: clusters: intracluster medium, — shock waves

1. Introduction

Collisions between massive galaxy clusters are the most energetic events in the present universe. Part of the kinetic energy released during these collisions is dissipated through supersonic shock fronts propagating in the intracluster medium (ICM) and turbulent motions. While heating the thermal component of the ICM, shocks and turbulence may also accelerate (or reaccelerate) relativistic particles (e.g., Cassano & Brunetti, 2005; Hoeft & Brüggen, 2007; Ryu et al., 2003; Pfrommer et al., 2006; Brunetti & Lazarian, 2007; Vazza et al., 2009). Radio observations probe these complex mechanisms through the detection of diffuse synchrotron emission from the ICM, in the form of giant radio halos, Mpc-scale radio emission in the cluster central regions, and radio relics, sharp-edged radio sources in the cluster periphery (e.g., Ferrari et al., 2008; Cassano, 2009; Venturi, 2011; Brunetti, 2011, for recent reviews)

Observable as sharp X-ray brightness and temperature discontinuities, few shock fronts have been detected so far, because they can only be visible in the brightest cluster regions and in particularly favorable projections (Markevitch et al., 2002, 2005; Russell et al., 2010; Finoguenov et al., 2010; Macario et al., 2011). Peripheral radio relics are believed to be shock fronts that propagated far outside the X-ray bright region, while still accelerating (or re-accelerating) electrons, which produce radio emission and quickly cool after the shock passes, resulting in a characteristic narrow feature (e.g., Ensslin et al., 1998; van Weeren et al., 2011). The physics of giant radio halos is probably more complex. Radio halos plausibly result from the (re)-acceleration and transport of relativistic particles in large turbulent regions of the ICM, although many aspects of the mechanisms generating radio-emitting electrons remain unclear (e.g., Brunetti, 2011, for a recent review). Sharp radio edges (and radio relics) are frequently observed at the border of giant radio halos, suggesting a possible link between merger shocks and the generation of turbulence in the ICM (e.g., Markevitch, 2010; Macario et al., 2011).

A521 is a moderately distant (z=0.247) and X-ray luminous (, Arnaud et al., 2000) 9 galaxy cluster, presenting several signatures of dynamical activity. As revealed already in ROSAT images, its X-ray and optical components appear spatially segregated, with an N–S bimodality of the X-ray emission (Arnaud et al., 2000), and a more complex galaxy number density distribution revealing two NW/SE and NE/SW major elongations (Ferrari et al., 2003). As further shown from Chandra data analysis, the ICM in A521 exhibits an irregular thermal structure with indications for gas heating at the interface between the two main gas components (Ferrari et al., 2006). A521 exhibits a giant radio halo that is the prototype of the class of ultra-steep spectrum radio halos (Brunetti et al., 2008). The halo being spatially connected to a radio relic, A521 provides us with an ideal test case to investigate the effects of shocks on the properties of thermal and nom-thermal components of the ICM and their connection with giant radio halos. In this merging cluster, a shock has been suggested by the presence of an X-ray brightness edge on the SE side of the cluster, coinciding with the edge of the radio relic (Giacintucci et al., 2008). The larger-scale cluster radio halo shows a very steep synchrotron spectrum supporting a picture where relativistic electrons are stochastically re–accelerated by the non-linear interaction with turbulence in the ICM (Brunetti et al., 2008).

The present article will focus on the analysis of a deep observation of A521 recently performed with XMM-Newton , with particular goal to probe the ICM structure up to the outermost regions covered by the cluster radio halo, and radio relic. After discussing data preparation and analyzes issues in Sections 2 and 3, we present the various X-ray brightness and temperature features revealed by this observation in Section 4. We comment on the interplay between thermal and non-thermal components of the ICM in Section 5. Unless otherwise noted, any energy distribution is normalized as a probability density function, while confidence ranges on individual parameter estimates are 68 . In the following, intra-cluster distances are computed as angular diameter distances, assuming a -CDM cosmology with , , . Given these assumptions, an angular separation of 1 arcmin corresponds to a projected intra-cluster distance of 232.5 kpc.

XMM-Newton Centre coordinates MOS1 effective MOS2 effective PN effective
obs. IDs exposure time (ks) exposure time (ks) exposure time (ks)
0603890101 (S) 04h54m22.00s -1016’30.” 15.7 (50.8 %) 15.7 (68.9 %) 12.8 (36.1 %)
0603890101 (U) 04h54m22.00s -1016’30.” 64.4 (77.3 %) 64.5 (74.6 %) 57.6 (52.5 %)
Note. The fraction of the useful exposure time after solar-flare filtering is shown in brackets.
Table 1Effective exposure time of each XMM-Newton -EPIC observation.
Figure 1.— EPIC XMM-Newton exposure of A521.
Figure 2.— Background spectrum observable in the A521 outskirts. Light blue: particle background. Cyan blue: Cosmic X-ray Background emission. Blue and Violet: TAE emission (kT1 = 0.099 keV, kT2 = 0.248 keV, see Kuntz & Snowden (2000), and details in Sect. 3.1). Green: Residual soft proton emission. Red and black: overall fit and data set.

2. Observations and data preparation

The EPIC-XMM-Newton  data set is a dual observation of A521, performed with focal aim point and central EPIC-MOS CCDs located in the SE cluster outskirts (see Figure 1). In order to remove the contribution of soft proton flares, we filtered the histogram of the photon arrival times through a temporal wavelet analysis. A summary of the ‘good’ exposure time remaining on each of the three EPIC cameras is provided in Table 1. The average good exposure time is about 75 ks.

In order to perform imaging and spatially resolved spectroscopy, we binned photons in sky coordinates and energy , matching the angular and spectral resolution of each focal instrument. To map the surface brightness of extended sources, these photon counts may have to be normalized for spatial and spectral variations of the telescope effective area and detector exposure times. We thus associate an ‘effective exposure’ array, , to the photon event cube. Expressible e.g. in , is computed as a linear combination of CCD exposure times, , related to individual observations , with local corrections for useful CCD areas, , Reflexion Grating Spectrometer (RGS) transmissions10, , and mirror vignetting factors 11.


3. Data analysis

3.1. Background noise modeling

The cluster emissivity must be separated from an additive, spatially extended and mostly stationary background noise including false photon detections due to charged particle-induced and out-of-time events, but also the Cosmic X-ray Background (CXB), and some Galactic foreground components.

The XMM EPIC background is dominated by the particle component, which is modeled from observations performed in the Filter Wheel Closed (FWC) position during revolutions 230 to 2027 as for the EPIC-MOS cameras, and 355 to 1905 as for the EPIC-PN camera. Following an approach proposed in e.g. Kuntz & Snowden (2008) or Leccardi & Molendi (2008) this model sums a quiescent continuum to a set of florescence emission lines convolved with the energy response of each detector. It is completed with a residual emission associated with soft protons, presently only detectable in the case of the EPIC-PN camera and modelled as a power-low spectrum normalized to 1.4 cts in the 0.5–1. band. To account for two different spectral shapes in the soft and hard bands, the quiescent continuum is modelled as the product of a power law with an inverted error function increasing in the soft band. We set the emission line energies to the values reported in Leccardi & Molendi (2008), while the soft proton residual is modelled using an additional power law. Presumably due to differences in the collecting areas of the imaging and readout detector regions, the EPIC-MOS quiescent continuum exhibit a small emissivity gradient along the RAWY CCD coordinate, which has been measured and taken into account in the model. Because the fluorescence lines exhibit a more complex spatial variation (Lumb et al., 2002; Kuntz & Snowden, 2008), we modelled the emissivity distribution of the most prominent lines 12 from the wavelet filtering of a set of FWC event images in narrow energy intervals around each line.

Secondary background components include the Cosmic X-ray background and Galactic foregrounds. Being associated with real photon detections, these components are corrected for the effective exposure. The Cosmic X-ray background is modelled with an absorbed power law of index (see, e.g., Lumb et al., 2002), while the Galactic foregrounds are modelled by the sum of two absorbed thermal components accounting for the Galactic Ôtransabsorption emissionÕ (TAE; kT1 = 0.099 keV and kT2 = 0.248 keV, see Kuntz & Snowden, 2000). We estimate emissivities of each of these components from a “joint-fit” of all background components in a region of the field of view located beyond the boundary of X-ray emission in the SE cluster outskirt, but covered by the central MOS CCDs (see also Figure 2). This estimates yields 13.4, 28.0 and 29.5 cts. in the 0.5–1. band as for the two transabsorption and CXB components, respectively (/d.o.f. = 1.23). Our background model, , eventually includes a contribution for the EPIC-PN out-of-time count rate, which is estimated in each energy band as 6.3 % of all photon counts registered along the CCD columns.

3.2. Spectroscopic and surface brightness measurements

To estimate average ICM temperatures, , and metal abundances, , along the line of sight and for a given location of the field of view , we add a source emission spectrum to the background model, and fit the spectral shape of the resulting function, , to the photon energy distribution registered in the energy band (0.3–12 keV):


In this modeling, the source emission spectrum assumes a redshifted and absorbed emission modelled from the Astrophysical Plasma Emission Code (APEC, Smith et al., 2001), with the element abundances of Grevesse & Sauval (1998) and neutral hydrogen absorption cross sections of Balucinska-Church & McCammon (1992). The value has been fixed to , from measurements obtained near A521 in the Leiden/Argentine/Bonn Survey of Galactic HI (Kalberla et al., 2005). It is altered by the mirror effective areas, filter transmissions and detector quantum efficiency, and convolved by a local energy response matrix M(k,l,e,e’) computed from response matrixes files (RMF) tabulated in detector coordinates in the XMM-Newton -EPIC calibration data base. 13

To compute images and radial profiles of the intra-cluster gas distribution, we estimate a cluster surface brightness map, , from photon counts registered in a given energy band, and corrected for effective exposure and additive background. Assuming an average ICM energy distribution, , we define as a function of an effective exposure map, :


All parameters of are practically determined from spectral fitting of the main cluster emission spectrum: , , , while is estimated in a “soft” energy band ([.5–2.5] ) in order to lower the dependence of on .

3.3. Surface brightness, ICM density and temperature profiles

In the following, surface brightness and temperature profiles have been extracted within cluster sectors oriented approximately along the surface brightness gradients. We derived the radial surface brightness by averaging the surface brightness of Equation (3) in each profile annulus composed of N pixels , as follows:


The background contribution being estimated within a much larger area of the field of view than any sector annulus used to derive , we neglected any systematic uncertainty related to its modeling and estimated the variance on from a weighted mean of the local Poisson fluctuations in :


Projected temperatures and associated confidence interval have been computed within each annulus by fitting a uniform emission model to the data set. To do so, we averaged the emission models of Equation (2) associated with each pixel of the annulus, and estimated the model parameters , via a minimization.

These brightness and temperature profiles have been used to model the underlying density and temperature of the ICM, assuming spherical symmetry of the cluster atmosphere in the vicinity of the features of interest. This was undertaken by projecting and fitting parametric distributions of the three-dimensional (3D) emission measure , , and temperature, , to the observed profiles. In this modeling, projected brightness profiles are convolved with the XMM-Newton focus Point Spread Function (PSF), while projected temperatures are computed assuming the ‘spectroscopic-like’ weighting scheme proposed in Mazzotta et al. (2004). In Section 4.3, the ICM emission measure 14 and temperature profiles across two shock fronts have been modelled by step-like distributions with a common jump radius :

Figure 3.— EPIC XMM-Newton observation of A521. Top panels: Photon rate image in the .5-2.5 keV band. Photon counts in this image have been corrected for spatially variable effective area, background flux and wavelet detected point-like sources. Bottom-left panel: Anisotropic details in the ICM emissivity map. These details have been enhanced from subtraction of a wavelet denoised map to the photon rate (further details are provided in Sect. 3.4.1) Bottom-right panel: ICM temperature map obtained from wavelet spectral-imaging. Prominent brightness jumps are indicated by dashed lines on the photon rate image and temperature map.

3.4. Imaging and spectral-imaging


An image of the cluster is presented on the top panels of Figure 3. To obtain this image, , we corrected the EPIC-XMM-Newton raw photon image for spatially variable effective area and background flux, following Equation (3). The point-like sources have also been modelled by means of an isotropic undecimated B3-spline wavelet analysis (see e.g. Starck et al., 2007), and subtracted from the image.

A map of anisotropic details in the ICM structure is shown on the bottom-left panel of Figure 3. To create this image, we subtracted a wavelet filtered map of the photon rate, , from the photon rate itself, then smoothed the residual image with a gaussian function of typical width . The wavelet filtering has been performed by means of a soft 3 thresholding of B3-spline wavelet coefficients, the significance thresholds being directly computed from the raw —Poisson distributed— photon map, following the multiscale variance stabilization scheme introduced in Zhang et al. (2008).

Figure 4.— Projected gas brightness measured across two cluster sectors intercepting the brightness jumps and . The projection of a step-like gas density distribution (Equation 6) convolved with the XMM-Newton PSF is superimposed as a dotted line, assuming density jump amplitudes of and as for and , respectively.
Figure 5.— Top panels: Projected gas brightness and temperature profiles measured across two cluster sectors intercepting the brightness jumps and , as shown on Figure 7. Bottom panels: ICM density and temperature profiles modelled as step-like 3D distributions matching the projected profiles (see also Equations  6 and 7). Dispersions on these profiles have been estimated from random realizations of the data set and corresponding models, each profile envelope delimiting 68 % of the realizations with closest distance from the original data set. The projection of these distributions is reported as a dotted line on the projected profiles.


In order to map the ICM temperature in A521, we used the EPIC-XMM-Newton data set and applied the spectral-imaging algorithm detailed in Bourdin et al. (2004) and Bourdin & Mazzotta (2008, , hereafter B08). Following this algorithm, a set of temperature arrays with associated fluctuations are first computed on various analysis scales , then convolved by complementary high-pass and low-pass analysis filters in order to derive wavelet coefficients. The wavelet coefficients are subsequently thresholded according to a given confidence level in order to restore a de-noised temperature map. Here, the signal analysis have been performed over 6 dyadic scales within an angular resolution range of = [1.7 – 110] arcsec. This was undertaken by averaging the emission modelled by Equation (2) within overlapping meta-pixels , and computing the and arrays by means of a likelihood maximization. The resulting ICM temperature map shown in Figure 3 was then obtained from a B2-spline wavelet analysis (see B08 for details) with coefficients thresholded to the 1 confidence level.

Figure 6.— Tentative interpretation of the ICM thermal and entropy structure observed in the central region of A521. Left: Early stage of a two-cluster merger: the cluster boundaries start to collide and develop two shock fronts propagating within the densest regions of each cluster. In the meanwhile, the cluster develop two cold fronts while pushing the higher entropy gas away from their interacting region. Right: The shock fronts have now propagated to the most external regions of the interacting clusters, but could not penetrate the two cool cores. A shocked gas region with high entropy remains at the interface between the two cold fronts.
Figure 7.— Photon rate image of A521 extracted in the .5-2.5 keV band. The image has been re-binned to a 6.8 arcsec angular resolution in order to enhance the brightness jumps and . The two annular sectors show the two cluster regions where temperature and brightness profiles of Figure 5 have been extracted.
Sector 1 Sector 2
Detector Region 1 Region 2 Region 3 Region 4 Region 1 Region 2 Region 3
Note. The fraction of the total counts is shown in brackets.
Table 2Estimated cluster photon counts within the .3–5. keV band, in the regions shown on Figure 7

4. ICM thermodynamics

4.1. Intra-cluster gas brightness and thermal structure

The X-ray photon image of Figure 3 reveals us the complex morphology of the intra-cluster gas in A521. On large scales, a northern subcluster with comet shape is apparently falling on the main component. The photon image also reveals the strongly irregular morphology of the surface brightness, presenting various edges indicated with dashed lines. Some of these brightness jumps have been enhanced in the bottom-left image of anisotropic details. They are also noticeable on the surface brightness profiles of Figure  4 and 5. At the interface between the two main interacting cluster components, we observe two bow-shaped brightness jumps, and , joining each other to form a low brightness cross-shaped feature. A third brightness jump with higher curvature radius, , is crossing the southern cluster component from SE to NW, while a fourth one, , is visible at the South-east cluster outskirts.

The ICM temperature map of Figure 3 is strongly irregular, and presents various noticeable features. The northern sub-cluster is clearly cool (). The interacting region separating this cool core from the main cluster to the South appears hotter () and strongly disturbed. The cross-shaped brightness depression observable on the photon image seems to coincide with a hot cross (), in particular, along the brightness jump . The southern part of the main cluster is cooler () than the interacting region, in particular, to the South of the brightness jump .

Bringing together the brightness and temperature maps of Figure 3, we observe that the two brightness jumps and are associated with temperature increases as the brightness decreases, while the brightness jumps and are associated with a temperature decrement. The two jumps and are thus likely to be cold fronts separating the densest parts of the two sub-clusters from their interacting region, while and are probably shock fronts propagating outwards from the colliding clusters.

4.2. Cold fronts and shock heating at the interface between two interacting sub-clusters

One of the most striking features seen in our data is a cross-shaped brightness depletion separating the two colliding sub-clusters. This feature also corresponds to a temperature and entropy enhancement, in particular just outside the two cold fronts, and . What we see is probably shocked gas with high entropy being squeezed by the converging cool core remnants and flowing around the densest part of the two interacting clusters, without penetrating the two cold fronts. The projected layer of shocked gas would thus exhibit maximal temperature and entropy near the two cold fronts, where it is tangentially intercepted by the line of sight. This shocked gas layer might also partly overlay in projection the main sub-cluster from its boundary delineated by cold front , to the southern brightness jump, . A possible interpretation for the origin of this hot gas flow is illustrated on Figure 6. Originally located at the cluster boundary (if there is one), the high entropy gas may have been shock heated between the two clusters starting to interact. It would now expand over the cluster atmosphere, following shock fronts presently propagating outside the cluster cores. One of these shock fronts might be observed to the South of the main cluster as the brightness and temperature jump, .

Figure 8.— Left panel: Anisotropic details in the X-ray emissivity (same as Figure 3) overlaid with emissivity iso-contours in the 240 MHz radio band (Giant Metrewave Radio Telescope; Brunetti et al., 2008). Right panel: ICM temperature map overlaid with the same radio emissivity iso-contours as for the left figure.
Density estimates Temperature estimates
Shock front Jump amplitude Mach number Jump amplitude Mach number
Table 3Density, temperature jumps and Mach numbers estimated across the shock fronts and .

4.3. Shock fronts propagation to the cluster outskirts

The 2D gas brightness and temperature maps of Figure 3 suggest the brightness jumps and to be shock fronts propagating outwards the cluster center. Located at various distances from the cluster center, these two shock fronts might have been developed during two successive cluster collisions. In order to analyze ICM thermodynamics across these jumps, we extracted the brightness and temperature profiles shown on Figure 5, corresponding to the two sectors of Figure 7. An estimation of the cluster photon counts in each sector region is provided in Table 2. The brightness jumps and exhibit the typical shape of a projected spherical density jump, convolved with the XMM-Newton PSF. We model the underlying gas density and temperature profiles as two step-like functions with common jump location, following equations (6) and (7). A discussion about the validity of the assumption of the ICM spherical symmetry in the vicinity of the shock fronts is provided in the Appendix. The gas density and temperature distributions corresponding to this model are reported under the projected profiles on Figure 5. The 3D density and temperature jumps associated with these distributions are reported in Table 3, with confidence intervals estimated from the 68 % percentiles of a parameter sample matching several random realizations of the data set.

The direction of the temperature jumps is consistent with the shock front interpretation. The cold front hypothesis would instead imply a temperature increase across the jumps (), which is excluded by the data. Assuming two shocks propagating outwards in the main cluster, one should be able to estimate the shock Mach numbers from the Hugoniot-Rankine density, temperature or pressure jump conditions across the fronts. Such Mach number values are reported on Table 3. The Mach numbers independently estimated from the density and temperature jumps are consistent with each other, though estimates from the temperature jumps have larger uncertainties. We will hereafter use Mach number estimates for both shocks and , from their density jumps: and .

5. Non-thermal ICM emission

A521 hosts a radio relic in its southeastern peripheral region, and a rare low-frequency giant radio halo. In order to investigate the interplay between thermal and non-thermal components of the ICM emission, the 240 MHz radio image obtained from observations performed at the Giant Metrewave Radio Telescope (GMRT) has been superimposed on the X-ray photon and ICM temperature maps of Figure 8.

5.1. The A521 radio halo

The A521 radio halo has been discovered from low-frequency observations at the GMRT (240, 325, 610 MHz, Brunetti et al., 2008, see also Figure 8) and then studied in detail through a deep follow-up Very Large Array observation at 1.4 GHz (Dallacasa et al., 2009). Its very steep spectrum, with spectral index between 325 and 1400 MHz, suggests magneto-hydrodynamic turbulence to be responsible for the in-situ re-acceleration of the relativistic electrons (Brunetti et al., 2008). The radio halo is covering the cluster central region, exhibiting an E-W elongation and reaching the radio relic to the Southeast. When excluding the relic region, the halo appears spatially correlated with the cluster X-ray emission. There is an even better correlation between the radio brightness and the hottest regions of the ICM, –in particular, the radio brightness exhibits a quick drop across the shock.

The complex thermodynamics of the ICM in the cluster center hint at the possible origin of the turbulence that may re-accelerate non-thermal particles in the halo. The two cold fronts and may have developed K-H instabilities at large angles from the main cluster collision axis. As suggested by the spatial correlation between shock-heated regions and the radio emission, turbulence may alternatively have been generated behind the two shocks and , now propagating to the cluster outskirts. In addition, the merger disturbance has likely generated turbulence within the two subcluster core remnants.

Figure 9.— Projected galaxy density distribution derived from photometric observations performed at the CFH telescope (dark color indicate higher densities, Dressler algorithm, see Ferrari et al., 2003). Pink iso-contours: radio emissivity in the 610 Mhz band (Giacintucci et al., 2008). Black iso-contours: X-ray emissivity in the .5-2.5 band (Curvelet de-noising of the XMM-Newton image).

5.2. Shock front propagation and the radio relic

A521 has been known to host a SE radio relic observed at various frequencies (Ferrari et al., 2006; Giacintucci et al., 2006, 2008, hereafter, G08). As shown in G08, the integrated synchrotron radiation in the relic exhibit a power-law spectrum with spectral index in the frequency range 235-4890 Mhz, with evidences of steepening of the radio spectrum with increasing distance from the eastern edge. As further noted in G08, the outer edge of the radio relic coincides with the X-ray edge , which we have shown in this work to be a shock front propagating to the cluster outskirts. As observed in several peripherical radio sources of galaxy clusters (see e.g. Brüggen et al. (2012) for a recent review), these facts support the shock electron (re)-acceleration to be at least partly responsible for the radio emission from the relic. Assuming diffuse shock acceleration for the origin of the emitting electrons, in the test particle approach the slope of the injection spectrum of cosmic rays is related to the shock Mach number, M, by (Blandford & Eichler, 1987) . This leads to a spectrum of electrons in the downstream region with slope (implying a synchrotron spectral index ) taking into account radiative losses and assuming stationary conditions. G08 thus predicted a shock propagation with Mach number, , from their measurement of . The steepening of the radio spectrum with increasing distance from the eastern edge further allowed them to predict a shock propagation to the cluster outskirts. The propagation direction and Mach number of the shock front , (cf. section 4.3), are fully consistent with this hypothesis.

It is further worth noticing that the X-ray edge corresponding to the shock front seems to extend in North–South direction more than the radio relic. A first interpretation for this limited extent of the relic might be that the shock would re-accelerate pre-existing relativistic electrons in the ICM. In this case the radio relic could reflect the spatial and energy distribution of the pre-existing electrons across the shock front. In line with this hypothesis, recent analyzes (Kang & Jones, 2007; Kang & Ryu, 2011a, b) suggest that the presence of pre-existing particles in addition to the thermal pool can significantly increase the average efficiency of the particle acceleration and the expected synchrotron emission at weak shocks (M 3). Differences in extension between the shock and the radio relic might alternatively indicate some changes in the efficiency of electron acceleration changes along the shock front, possibly due to local variations of the Mach number (e.g., Hoeft et al., 2008). In this respect the radio relic in A521 is located at the extremity of the NW/SE major galaxy alignment evidenced in Ferrari et al. (2003, see also Figure 9), where indeed recent accretion of subcluster material may have produced inhomogeneities in the ICM.

6. Discussion and conclusions

A521 is a complex cluster system where optical analyzes have revealed at least three galaxy groups to the SE, and four groups to the NW including the cluster BCG group coinciding with the X-ray peak (Ferrari et al., 2003). The X-ray morphology of the BCG group suggests an infall along a NNW–SSE direction (projected onto the sky plane), which is slightly offset with respects to the major NW–SE galaxy alignment (Ferrari et al., 2006). The cluster atmosphere exhibits various brightness and temperature edges associated with cold fronts and shock fronts, that our XMM-Newton data revealed.

The main two interacting gas components in the central region of this system are separated by a region of gas with lower density, higher temperature and entropy. We interpret this feature a flow of high-entropy gas being squeezed by two converging subcluster cores that are delimited by cold fronts. We suggest this high entropy gas to have been heated by shocks formerly developed when the two gas components started to interact. One of these shocks is currently observed to the South of the main component, with Mach number . The hot gas region separating the two interacting components appears spatially correlated with the cluster radio halo. The development of turbulence in the hot gas flowing between the two cool cores may be responsible for high energy electron re-acceleration, yielding the radio halo emission. Merger shock propagation and/or cold fronts may have contributed to the development of these instabilities.

A shock front is observed at the Southeast cluster outskirt. An X-ray brightness edge there has been hinted at by Chandra data (G08), though the statistical significance was marginal. The orientation of this shock front and its large distance from the cluster center suggest that it is associated with a cluster collision that has occurred prior to the current two-component interaction. Our Mach number for this shock, , is consistent with that expected from the spectrum of the radio relic in G08 under the assumption of Fermi I acceleration mechanism. As observed in X-ray follow-ups of other radio relics –A3667, Finoguenov et al. (2010); RXCJ1314.4-2515, Mazzotta et al. (2011)–, its detection supports the shock electron (re)-acceleration to be at least partly responsible for the radio emission from the relic. The detection of a polarization of the relic would be an additional support for this process, complementary to the extension of its synchrotron spectrum to very high radio frequencies, and to evidences for spectral steepening downstream to the shock (G08). Delimited by the shock front, the radio relic seems however to subtend only a fraction of the shock front. Differences in the spatial extent of a radio relic and its companion shock front have also been observed in the colliding cluster RXCJ1314.4–2515 (Mazzotta et al., 2011), where a radio relic seems to be confined to a small section of the shock front presumably distorted by a nonuniform gas flow. These differences may thus reflect variations of the efficiency of particle acceleration across the shock that could be driven by local variations of the Mach number and shock velocity. Deeper X-ray or SZ observations may enable us to investigate this hypothesis, though the present XMM-Newton image does not seem to evidence any strong variation in the amplitude of the surface brightness edge, and thus in the shock Mach number. An alternative hypothesis is that the radio relic would reveal us local inhomogeneities in the properties of the pre-existing relativistic electrons, that would be re-accelerated by the shock passage. The observed connection between the radio halo and the relic may suggests that pre-existing relativistic electrons have first been accelerated by turbulent gas motions responsible for the radio halo emission, then re-accelerated at the shock front.

We thank the reviewer for her/his constructive comments and suggestions aiming at improving the manuscript. H.B. thanks the Harvard-Smithsonian Centre for Astrophysics, where this work has been initiated, for its hospitality. We thank Chiara Ferrari for providing us a map of the projected galaxy density distribution in A521, derived from photometric observations performed at the CFH telescope. This work is based on observations obtained with XMM-Newton, an ESA science mission funded by ESA Member States and the USA (NASA). H.B and P.M acknowledge support by grants NASA grant NNX09AP45G and NNX09AP36G grant ASI-INAF I/088/06/0 and ASI-INAF I/009/10/0. S.G. acknowledges the support of NASA through Einstein Postdoctoral Fellowship PF0-110071 awarded by the Chandra X-ray Center, which is operated by the Smithsonian Astrophysical Observatory. G.B. acknowledges partial support from PRIN-INAF2009
Figure 10.— ICM geometry across the shocks. Left: Cluster volume cut along the line of sight. Right: Cluster volume cut within the sky plane.
Shock ICM asphericity ICM slope Shock curvature Density jump Temperature jump Mach number
() () () (derived from )
Table 4ICM asphericity parameters across the shock fronts S1 and S2.

The ICM density and temperature distributions intercepting the shocks and has been modelled in Section 3.3 as two step-like functions, assuming the shock center of curvature and the ICM centroid coincide to coincide with each other.

The X-ray image of Figure 3 seems however to show that the shock fronts and S2 are less curved than the closest cluster brightness isophotes. To investigate the systematic uncertainties inherent to our spherical symmetry approximation, we alternatively tried to model the shock front and the ICM density as two spherical distributions with distinct centers. Assuming these two centers to be located in the plane of the sky, the ICM emission measure is now expressed per volume unit, as:


where and refers to the norm of each radius vector in the shock and ICM frames, respectively. Introducing and , the projection of these radius vectors onto the sky plane, a surface brightness profile intercepting the shock is obtained from integration of Equation (1) along the line of sight:


where is related to as a function of , the distance separating the shock from the center of the ICM distribution and , the angle separating the projected radius vector to the shock propagation axis (, see also Fig. 10). In addition to an ICM density slope, , and the shock curvature radius, density and temperature jumps, , and , respectively, the ICM emission measure thus depends on an asphericity parameter: .

We tried to invert and its parameters from a minimization of the distance separating (Equation 2) from the X-ray surface brightness profiles extracted across each shock front (see Figure 5). Some of the searched parameters being degenerated with one another, we first performed this inversion by fixing the asphericity parameter to 0 and 0.5, corresponding to shocks located at distances of and from the cluster center, respectively. We subsequently left all parameters free to vary and report the results of our measurements in Table 4, the confidence interval on each parameter being estimated from the 68 % percentiles of a parameter sample matching several random realizations of the data set. As expected, the shock curvature radius, density and temperature jumps obtained when fixing the asphericity to 0 are consistent with their estimates derived from the spherical model of section 3.3. A marginal difference in the amplitude of the density jump is still noticeable, since Equation (1) yields as for shock , while Equation (6) yields . This difference is probably related to the lack of any variation of the ICM density slope at the shock crossing, following Equation (1). Fixing the aspericity to 0.5 instead of 0 also marginally affect the density jump, essentially due to the degeneracy between the ICM asphericity and density slope. This degeneracy is noticeable in the case of , the shock front observed with the highest statistics. Leaving the ICM asphericity free to vary yields estimates of 0.45 and 0.75 in the case of and , respectively, consistent with the shock curvature radii observed on the X-ray image of Figure 3. The shock density, temperature jump and Mach numbers derived from these various assumptions are in any case consistent with one another, and with their estimates obtained from the sperical model of section 3.3. Given the limited statistics available, it is difficult to break the degeneracy between the ICM asphericity, ICM density slope and shock curvature radius in the vicinity of the shocks. For simplicity purposes, we consequently adopted the spherical model of section 3.3 in order to derive the amplitudes of the density jumps and Mach numbers of the two shocks and .


  1. affiliation: Dipartimento di Fisica, Università degli Studi di Roma ‘Tor Vergata’, via della Ricerca Scientifica, 1, I-00133 Roma, Italy;
  2. affiliation: Dipartimento di Fisica, Università degli Studi di Roma ‘Tor Vergata’, via della Ricerca Scientifica, 1, I-00133 Roma, Italy;
  3. affiliation: Harvard Smithsonian Centre for Astrophysics, 60 Garden Street, Cambridge MA, 02138, USA
  4. affiliation: NASA Goddard Space Flight Center, Code 662, Greenbelt, MD 20771, USA
  5. affiliation: Joint Space-Science Institute, University of Maryland, College Park, MD, 20742-2421, USA
  6. affiliation: Joint Space-Science Institute, University of Maryland, College Park, MD, 20742-2421, USA
  7. affiliation: Astronomy Department, University of Maryland, College Park, MD 20742, USA
  8. affiliation: INAF - Istituto di Radioastronomia, via Gobetti 101, I-40129 Bologna, Italy
  9. ; X-ray luminosity in the 0.1-2.4 keV band has been corrected for luminosity distance assuming km s Mpc, and .
  10. EPIC-MOS detectors share a common optical path with the RGS
  11. Information about these instrumental effects have been obtained from the XMM-Newton -EPIC Current Calibration Files (CCFs)
  12. Namely the Al, Si and Cu, Ni complexes as for the EPIC-MOS and EPIC-PN cameras, respectively.
  13. EPIC response matrixes are computed from canned RMFs corresponding to the observation period provided by the XMM-Newton Science Operation Centre.
  14. More precisely, the ICM emission measure per volume unit.


  1. Arnaud, M., Maurogordato, S., Slezak, E., & Rho, J. 2000, A&A, 355, 461
  2. Balucinska-Church, M., & McCammon, D. 1992, ApJ, 400, 699
  3. Blandford, R., & Eichler, D. 1987, Phys. Rep., 154, 1
  4. Bourdin, H., & Mazzotta, P. 2008, A&A, 479, 307
  5. Bourdin, H., Sauvageot, J., Slezak, E., Bijaoui, A., & Teyssier, R. 2004, A&A, 414, 429
  6. Brüggen, M., Bykov, A., Ryu, D., & Röttgering, H. 2012, Space Sci. Rev., 166, 187
  7. Brunetti, G. 2011, Mem. Soc. Astron. Italiana, 82, 515
  8. Brunetti, G., Giacintucci, S., Cassano, R., Lane, W., Dallacasa, D., Venturi, T., Kassim, N. E., Setti, G., Cotton, W. D., & Markevitch, M. 2008, Nature, 455, 944
  9. Brunetti, G., & Lazarian, A. 2007, MNRAS, 378, 245
  10. Cassano, R. 2009, in Astronomical Society of the Pacific Conference Series, Vol. 407, The Low-Frequency Radio Universe, ed. D. J. Saikia, D. A. Green, Y. Gupta, & T. Venturi, 223
  11. Cassano, R., & Brunetti, G. 2005, MNRAS, 357, 1313
  12. Dallacasa, D., Brunetti, G., Giacintucci, S., Cassano, R., Venturi, T., Macario, G., Kassim, N. E., Lane, W., & Setti, G. 2009, ApJ, 699, 1288
  13. Ensslin, T. A., Biermann, P. L., Klein, U., & Kohle, S. 1998, A&A, 332, 395
  14. Ferrari, C., Arnaud, M., Ettori, S., Maurogordato, S., & Rho, J. 2006, A&A, 446, 417
  15. Ferrari, C., Govoni, F., Schindler, S., Bykov, A. M., & Rephaeli, Y. 2008, Space Sci. Rev., 134, 93
  16. Ferrari, C., Maurogordato, S., Cappi, A., & Benoist, C. 2003, A&A, 399, 813
  17. Finoguenov, A., Sarazin, C. L., Nakazawa, K., Wik, D. R., & Clarke, T. E. 2010, ApJ, 715, 1143
  18. Giacintucci, S., Venturi, T., Bardelli, S., Brunetti, G., Cassano, R., & Dallacasa, D. 2006, NewA, 11, 437
  19. Giacintucci, S., Venturi, T., Macario, G., Dallacasa, D., Brunetti, G., Markevitch, M., Cassano, R., Bardelli, S., & Athreya, R. 2008, A&A, 486, 347
  20. Grevesse, N., & Sauval, A. J. 1998, Space Science Reviews, 85, 161
  21. Hoeft, M., & Brüggen, M. 2007, MNRAS, 375, 77
  22. Hoeft, M., Brüggen, M., Yepes, G., Gottlöber, S., & Schwope, A. 2008, MNRAS, 391, 1511
  23. Kalberla, P. M. W., Burton, W. B., Hartmann, D., Arnal, E. M., Bajaja, E., Morras, R., & Pöppel, W. G. L. 2005, A&A, 440, 775
  24. Kang, H., & Jones, T. W. 2007, Astroparticle Physics, 28, 232
  25. Kang, H., & Ryu, D. 2011a, Mem. Soc. Astron. Italiana, 82, 648
  26. —. 2011b, ApJ, 734, 18
  27. Kuntz, K. D., & Snowden, S. L. 2000, ApJ, 543, 195
  28. —. 2008, A&A, 478, 575
  29. Leccardi, A., & Molendi, S. 2008, A&A, 486, 359
  30. Lumb, D. H., Warwick, R. S., Page, M., & De Luca, A. 2002, A&A, 389, 93
  31. Macario, G., Markevitch, M., Giacintucci, S., Brunetti, G., Venturi, T., & Murray, S. S. 2011, ApJ, 728, 82
  32. Markevitch, M. 2010, ArXiv:1010.3660
  33. Markevitch, M., Gonzalez, A. H., David, L., Vikhlinin, A., Murray, S., Forman, W., Jones, C., & Tucker, W. 2002, ApJ, 567, L27
  34. Markevitch, M., Govoni, F., Brunetti, G., & Jerius, D. 2005, ApJ, 627, 733
  35. Mazzotta, P., Bourdin, H., Giacintucci, S., Markevitch, M., & Venturi, T. 2011, Mem. Soc. Astron. Italiana, 82, 495
  36. Mazzotta, P., Rasia, E., Moscardini, L., & Tormen, G. 2004, MNRAS, 354, 10
  37. Pfrommer, C., Springel, V., Enßlin, T. A., & Jubelgas, M. 2006, MNRAS, 367, 113
  38. Russell, H. R., Sanders, J. S., Fabian, A. C., Baum, S. A., Donahue, M., Edge, A. C., McNamara, B. R., & O’Dea, C. P. 2010, MNRAS, 406, 1721
  39. Ryu, D., Kang, H., Hallman, E., & Jones, T. W. 2003, ApJ, 593, 599
  40. Smith, R. K., Brickhouse, N. S., Liedahl, D. A., & Raymond, J. C. 2001, ApJ, 556, L91
  41. Starck, J.-L., Fadili, J., & Murtagh, F. 2007, IEEE Transactions on Image Processing, 16, 297
  42. van Weeren, R. J., Brüggen, M., Röttgering, H. J. A., & Hoeft, M. 2011, Journal of Astrophysics and Astronomy, 32, 505
  43. Vazza, F., Brunetti, G., & Gheller, C. 2009, MNRAS, 395, 1333
  44. Venturi, T. 2011, Mem. Soc. Astron. Italiana, 82, 499
  45. Zhang, B., Fadili, J. M., & Starck, J. 2008, IEEE Transactions on Image Processing, 17, 1093
Comments 0
Request Comment
You are adding the first comment!
How to quickly get a good reply:
  • Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
  • Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
  • Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
The feedback must be of minumum 40 characters
Add comment
Loading ...
This is a comment super asjknd jkasnjk adsnkj
The feedback must be of minumum 40 characters
The feedback must be of minumum 40 characters

You are asking your first question!
How to quickly get a good answer:
  • Keep your question short and to the point
  • Check for grammar or spelling errors.
  • Phrase it like a question
Test description