Chandra Measurements of Non-thermal-like X-rayEmission from Massive, Merging, Radio-halo Clusters

Evan Million1 & S. W. Allen1emillion@stanford.edu; KIPAC (Stanford/SLAC)

AbstractWe report the discovery of spatially-extended, non-thermal-like emission components in Chandra X-ray spectra for five of a sample of seven massive, merging galaxy clusters with powerful radio halos.The emission components can be fitted by power-law models with mean photon indices in the range1.5 < Γ < 2.0. A control sample of regular, dynamically relaxed clusters, without radio halos but withcomparable mean thermal temperatures and luminosities, shows no compelling evidence for similarcomponents. Detailed X-ray spectral mapping reveals the complex thermodynamic states of the ra-dio halo clusters. Our deepest observations, of the Bullet Cluster 1E 0657-56, demonstrate a spatialcorrelation between the strongest power-law X-ray emission, highest thermal pressure, and brightest1.34GHz radio halo emission in this cluster. We confirm the presence of a shock front in the 1E 0657-56 and report the discovery of a new, large-scale shock front in Abell 2219. We explore possibleorigins for the power-law X-ray components. These include inverse Compton scattering of cosmicmicrowave background photons by relativistic electrons in the clusters; bremsstrahlung from supra-thermal electrons energized by Coulomb collisions with an energetic, non-thermal proton population;and synchrotron emission associated with ultra-relativistic electrons. Interestingly, we show that thepower-law signatures may also be due to complex temperature and/or metallicity structure in clustersparticularly in the presence of metallicity gradients. In this case, an important distinguishing charac-teristic between the radio halo clusters and control sample of predominantly cool-core clusters is therelatively low central X-ray surface brightness of the former. Our results have implications for previousdiscussions of soft excess X-ray emission from clusters and highlight the importance of further deepX-ray and radio mapping, coupled with new hard X-ray, γ-ray and TeV observations, for improving ourunderstanding of the non-thermal particle populations in these systems.

The SampleOur targets are drawn from the sample of X-ray luminous clusters with large, extended radio ha-los discussed by Feretti & Giovannini (2008). These clusters have high quality X-ray data availableon the Chandra archive. All of the clusters show clear evidence for recent, major merger activitywith no cooling cores. Clusters with radio halos are expected to produce non-thermal X-ray emis-sion via Inverse Compton (IC) scattering of the Cosmic Microwave Background (CMB). However,other processes can also contribute to a non-thermal-like X-ray emission signature including shocks(see Markevitch & Vikhlinin 2007 and references within), a range of non-thermal bremsstrahlung pro-cesses (eg.Wolfe & Melia 2008) and synchrotron emission by ultra-relativistic electrons and positrons(eg.Inoue et al.2005). Our targets include 1E 0657-56 (the ‘Bullet Cluster’), Abell 665, 2163, 2219,2255, 2319, and 2744. As a control we have also observed five comparably X-ray luminous clustersthat are known not to contain radio halos: Abell 478, 576, 1795, 2029, and 2204 (Ferretti & Giovannini2008). Most are highly relaxed and contain central cooling cores.

100 300 500

45.0 40.0 35.0 6:58:30.0 25.0 20.0 15.0

54:30.0

55:00.0

30.0

56:00.0

30.0

-55:57:00.0

30.0

58:00.0

30.0

2 4 6 8 10 12 14 16

Shock

0.002 0.004 0.006 0.008 0.01 0.012 0.014

Fig. 1 Surface brightness (left), temperature map (middle), and pressure map (right) of the ‘Bullet Cluster. Regions have

∼ 3, 000 net counts which results in a statistical uncertainty in temperature of ∼ 10%. Radio halo contours are from Liang

et al.2000. The radio halo morphology corresponds well with that of the pressure map.

Spectral Regions and AnalysisIndividual regions for the spectral analysis were determined using the contour binning technique ofSanders (2006). Regions follow contours of constant surface brightness until a desired signal to noisethreshold is met. Regions are small enough to be reasonably approximated by an optically-thin, sin-gle temperature plasma model (MEKAL). For our standard analysis, each independent spectral regioncontains 104 counts and is modelled by a single temperature model with the absorption fixed at theGalactic value (model A) determined by the LAB HI survey (Kalberla et al.2005). A series of moresophisticated spectral models were examined and compared to those obtained with model A. Theseinclude (model B) allowing the line-of-sight absorption to vary as a single free parameter and the in-clusion of a power-law (the photon index is fixed at Γ = 2.0) or a second temperature component withfree normalizations in each independent region.

Figure 1 and 2 show the Chandra X-ray image, the temperature, and pressure map for the radio haloclusters 1E 0657-56 (the ‘Bullet Cluster’) and Abell 2219. The ‘Bullet Cluster’ shows significant sub-structure in its surface brightness and temperature maps. We confirm the presence of a large scaleshock (Markevitch 2006). Our analysis indicates a Mach number M = 2.5+0.5

−0.4, corresponding to avelocity v = 4000+800

−600 km s−1.

Statistical Evidence for Non-thermal-like ComponentsUsing the F-test to compare the results for the simplest single-temperature models, A and B, we seethat for four out of six radio halo clusters, allowing the absorbing column to fit freely leads to signifi-cant improvement in the fit. In all four of these cases the best-fit column density is significantly lowerthan the Galactic value determined from HI studies. This suggests that model B does not provide acomplete description of the spectra. Conversely, none of the control sample of non-radio halo clustersshow comparably significant improvement when comparing models B and A.

The introduction of a power-law component with fixed photon index Γ = 2.0 leads to larger improve-ments in the fits for the same radio halo clusters. Again the control sample shows no comparablysignificant improvement in the fits obtained with model C over model A. Similar conclusions on thepresence or otherwise of non-thermal-like emission are also drawn from the fits with model D, wherea second MEKAL component, rather than a power-law model, is used.

2 4 6 8 10 12

35.0 30.0 25.0 16:40:20.0 15.0 10.0 05.0

45:00.0

44:00.0

43:00.0

42:00.0

41:00.0

46:40:00.0

10 12 14 16 18 20 22 24 26 0.002 0.004 0.006 0.008 0.01 0.012

Fig. 2 Surface brightness (upper), temperature map (middle), and pressure map (right) of Abell 2219. Details as for Fig.

1. We have discovered a large scale shock (Mach number of 1.3-1.9, corresponding to a velocity of ∼2500 km s−1) in the

cluster. The central temperature is also remarkably hot (kT = 22+7

−4keV) which may also be associated with recent/ongoing

shock activity.

Physical InterpretationsIt has been common in the literature to interpret limits on non-thermal X-ray emission from clusters interms of simple IC models. However, the flat spectrum and low implied magnetic fields appear at oddswith the simplest IC models. Strong theoretical challenges also exist for non-thermal bremsstrahlungprocesses and ultra-relativistic electrons emitting X-ray synchrotron emission.

We have generated and analyzed simulated X-ray data sets with characteristics similar to the Chandra

observations of Abell 2319. We incorporate a broad range of temperature and metallicity distributions(metallicity gradients) and fit with the same basic models. If both metallicity gradients and temperaturevariations are allowed, then a significant ‘non-thermal-like’ signal, of order the one detected here, canbe obtained if ∼ 50% of the flux originates from a lower metallicity component. The photon indexand flux level are consistent with that observed here. In this case, the physical difference betweenthe radio halo sample and the control sample becomes the presence of large surface brightness, coolcores. Fig. 3a-b shows a 3 dimensional representation of the surface brightness for the control clusterAbell 2029, and the radio halo cluster Abell 2319. For Abell 2029, most of the flux is from the bright,cool core whereas for Abell 2319, ∼ 50% of the flux is from projection effects. This may also explainpreviously reported claims of soft X-ray excesses in galaxy clusters though a true non-thermal originis still a possibility.

Fig. 3. 3D representation of the X-ray surface brightness for the control cluster Abell 2029 (middle) and the radio halo

cluster Abell 2319 (right).

REFERENCES:Feretti L., Giovannini G., 2008, in Plionis M., Lopez-Cruz O.,

Hughes D., eds, A Pan-Chromatic View of Clusters of Galaxies

and the Large-Scale Structure. Lecture Notes in Physics, Vol. 740.

Springer-Verlag, Berlin, p. 143

Inoue S., Aharonian F.A., Sugiyama N., 2005, ApJ, 628, 9

Kalberla P.M. et al., 2005, A&A, 440, 775

Liang H., Hunstead, R.W., Birkinshaw M., & Andreani P., 2000, ApJ,

544, 686

Markevitch M., 2006, ESA SP-604: The X-ray Universe 2005, 723

Markevitch M., Vikhlinin A., 2007, PhR, 443, 1

Sanders J.S., 2006, MNRAS, 371, 829

This work has recently been accepted to MNRAS (astro-ph:0811.0834).This work was supported by the U.S. Department of Energy and NASA.