Discovery of a shock front in the merging cluster of galaxies A2163
aa r X i v : . [ a s t r o - ph . H E ] J a n MNRAS , 1– ?? (2016) Preprint January 29, 2021 Compiled using MNRAS L A TEX style file v3.0
Discovery of a shock front in the merging cluster of galaxies A2163
N. Mhlahlo, ⋆ L. Guennou, L. Feretti, School of Physics, University of the Witwatersrand, Private Bag 3, 2050-Johannesburg, South Africa University of KwaZulu-Natal, King George V Ave, Durban, 4041, South Africa Istituto di Radioastronomia INAF, Via P. Gobetti 101, 40129 Bologna, Italy.
Accepted XXX. Received YYY; in original form ZZZ
ABSTRACT
ACO2163 is one of the hottest (mean kT = − . z = . XMM-Newton and
Chandra obser-vations hinted at the presence of a shock front that is associated with the gas ‘bullet’ crossingthe main cluster in the west-ward direction, and which heated the intra-cluster medium, lead-ing to adiabatic compression of the gas behind the ’bullet’. The goal of this paper is to reporton the detection of this shock front as revealed by the temperature discontinuity in the X-rayXMM-Newton image, and the edge in the Very Large Array (VLA) radio image. We alsoreport on the detection of a relic source in the north-eastern region of the radio halo in theKAT-7 data, confirming the presence of an extended relic in this cluster.The brightness edge in the X-rays corresponds to a shock front with a Mach number M = . ± .
3, at a distance of 0.2 Mpc from the cluster centre. An estimate from the luminosityjump gives M = . ± .
4. We consider a simple explanation for the electrons at the shockfront, and for the observed discrepancy between the average spectral index of the radio haloemission and that predicted by the M = . Key words:
Galaxies: clusters: intracluster medium; Physical Data and Processes: accelera-tion of particles; shock waves
Radio studies of supernovae remnants have indicated that inthese sources, a fraction of the shock energy can be convertedinto the acceleration of the electrons to relativistic speeds (e.g.Blandford & Eichler 1987). It is generally accepted that this pro-cess also takes place during cluster mergers, and is responsible forthe production of the large-scale diffuse synchrotron radio emis-sion observed in galaxy clusters’ peripheries in the form of relicsources (see e.g. Giacintucci et al. 2008; Finoguenov et al. 2010;Macario et al. 2011; Akamatsu & Kawahara 2013; Bourdin et al.2013; Ogrean & Brüggen 2013). The other large-scale diffusesources that are observed at cluster centres in approximately 60clusters so far, known as radio halos, are thought to result fromthe acceleration of the electrons by magnetohydrodynamic turbu-lence in the presence of magnetic fields, which also occurs duringcluster mergers (e.g. Buote 2001; Feretti et al. 2002; Feretti 2005;Feretti et al. 2012, and references therein).However, the spatial coincidence of shocks with the edges ofradio halos, which has been observed in a few clusters (e.g. ⋆ E-mail: [email protected]
Markevitch et al. 2005), has raised questions about the role thatshock acceleration processes can play in the formation of radiohalos (see also the review by Markevitch 2010). Out of all theclusters where detections of merger shock fronts have been re-ported, roughly seven have their radio halo edges coincident withthe shock location, suggesting that shocks could be responsible for(re)acceleration of relativistic electrons at the halo edge locationsin the presence of magnetic fields, leading to the production of atleast some of the observed radio emission at those sites.A number of models have been used to explain the electronsat the shock front. Direct acceleration of relativistic electronsby the shock (e.g. Blandford & Eichler 1987) has been cited asthe main mechanism responsible for the emission at cluster out-skirts, and at least for some of the halo emission, in a fewclusters (Markevitch et al. 2005; Macario et al. 2011). Compres-sion by the shock of fossil electrons in the intra-cluster medium(ICM) (e.g. Ensslin et al. 1998; Enßlin & Gopal-Krishna 2001;Markevitch et al. 2005; Macario et al. 2011) has been consideredto be another alternative. The compression of pre-existing elec-trons at the shock front is expected to significantly increase theirsynchrotron emission at the observing frequency and to produceradio emission in front of the bow shock (Enßlin & Gopal-Krishna © 2016 The Authors hock front in ACO21632001). Another process for radio emission production, which isexpected to be more efficient than the alternatives, is shock re-acceleration of fossil (pre-existing) relativistic plasma which couldbe the remnant of a radio galaxy, by Diffusive Shock Accel-eration (DSA) (Blandford & Eichler 1987; Shimwell et al. 2015;Kang & Ryu 2016). This model has been used to explain theobserved discrepancy between the X-ray Mach number and theDSA predictions for some of the clusters (e.g. Kang & Ryu 2016;Ogrean et al. 2013).There is no strong agreement about which mechanism best de-scribes the electrons at the shock front. However, it is generallyagreed that the sharp radio edges which are observed at the bordersof some of the giant radio halos suggest a possible connection be-tween merger shocks and the generation of turbulence in the ICM,and that acceleration by turbulence is responsible for the cluster-scale radio emission (e.g. Markevitch 2010; Macario et al. 2011).The evolution of cluster-scale radio emission in clusters has beeninvestigated by Donnert et al. (2013) who combined an idealizedmodel for cluster mergers, with a numerical model for the injec-tion, cooling and re-acceleration of cosmic-ray electrons. The sim-ulations have shown that re-acceleration of cosmic-ray electronscan potentially reproduce key observables of radio halos such asthe transient nature of radio halos and their connection to merg-ers and merger-driven turbulence. Furthermore, non-thermal ef-fects connected to cosmological shock waves and AGN feedbackhave also been studied with cosmological simulations. Vazza et al.(2013) modelled the injection and evolution of cosmic rays, as wellas their effects on the thermal plasma, and further investigated theinjection of turbulent motions into the ICM from both the accretionof matter and AGN feedback in order to constrain the energeticsand mechanisms of feedback models in clusters. Also, the simula-tions of Vazza et al. (2011) have shown that merger shock heating isthe leading source of entropy production in clusters and is respon-sible for generating most of the entropy of the large-scale structuresin the Universe.Though there is a relatively large number of merging clusters (>60),shock fronts are rare (at a fraction of ∼ Ω m = .
32 and Ω Λ = .
68 and H = . − Mpc − is as-sumed. ACO2163 is a moderately distant (z=0.203; Struble & Rood 1999),rich cluster, and is one of the hottest (mean kT=12-15.5 keV,(Elbaz et al. 1995; Markevitch et al. 1996; Maurogordato et al.2008)) and extremely X-ray overluminous compared to its mass( L X [2-10 keV] = × erg s − , (Arnaud et al. 1992)). This clus-ter has been a subject of extensive studies at multiple wave-lengths (see Elbaz et al. 1995; Markevitch et al. 1996; Squires et al.1997; Govoni et al. 2004; Feretti et al. 2001; Nord et al. 2009;Bourdin et al. 2011, and references therein). Evidence for a recentmerger that involves two or more components has been revealedby X-ray morphological studies of Elbaz et al. (1995) based onROSAT data, the spectroscopic analysis of Markevitch et al. (1996)based on ASCA data, and that of Bourdin et al. (2011) based onXMM-Newton and Chandra data. In Figure 1 we show a smoothedX-ray image of the cluster, and we observe hints indicating thepresence of a major merger, with the main axis of the cluster alongthe north-east, south-west direction.Dynamical studies based on joint weak gravitational lensing and X-ray observations (Squires et al. 1997) have shown that ACO2163is in a disturbed dynamical state, showing irregular distributionof mass and galaxies. This has been supported by weak-lensingstudies of the dark matter distribution in ACO2163 which havesuggested that a multiple merger is taking place in this cluster(Soucail 2012). Further evidence for a merger has come from X-ray temperature maps which have shown strong temperature varia-tions and complex X-ray thermal structure in their central regions(Govoni et al. 2004; Bourdin et al. 2011). Temperature maps havealso revealed a cold front in the south-west direction from the clus-ter centre associated with ∼ ∼
11 keV gas (e.g. Owers et al. 2009).The merger scenario in ACO2163 is also supported by the presenceof a radio halo in this cluster. The detection of a radio halo was firstreported by Herbig & Birkinshaw (1994). Feretti et al. (2001) re-observed the cluster at 1.4 GHz with the VLA, and the detectionof one of the largest radio halos with a total extent of ∼ ′ wasreported (see also Feretti et al. 2004, for low frequency observa-tions). Diffuse emission at the N-E side of the halo (Feretti et al.2004) was also detected and this feature was identified with a relicsource.Recent studies on ACO2163 done by Bourdin et al. (2011), whichrevealed striking similarities between ACO2163 and the ‘Bullet’cluster, have suggested the presence of a shock front in this cluster.Evidence for a gas ’bullet’ or cool-core separated from its galax-ies and crossing the ACO2163 atmosphere along the east-west di-rection was presented. From evidence of pressure excess, partic-ularly in the innermost regions of the main cluster ACO2163-A,Bourdin et al. (2011) infered adiabatic compression of the gas inthe ICM behind the westward moving gas ‘bullet’ as a result ofpresumed shock heating. This has suggested that the main-clusterin ACO2163 has accreted a subcluster along the east-west direc-tion at a supersonic velocity. Though the merger event might haveshocked the main-cluster atmosphere, however, there was no evi-dence for a shock front preceding this ‘bullet’, contrary to what isobserved in the ‘Bullet’ cluster.More recently, Thölken et al. (2018) detected three shock frontsin the spectral analysis of data from Suzaku XIS observations ofA2163, one in the NE direction and two in the SW direction. Theirinner shock in the SW direction is located at 3.3’ from the clustercentre, which is further out compared to the location of the shockreported here, which is at ∼ ′ from the centre.2 X-RAYOBSERVATIONS,ANALYSISANDRESULTS Shock front in ACO2163 Figure 1.
Left: Background-subtracted XMM-Newton X-ray image of the cluster ACO2163 in the 0.5 - 8.0 keV energy band, representing ∼ FWHM =
10 arcsec. To the south-west is the brightness edge that we argue is a shock, and the twolines near the bright emission region show the approximate extent of the edge. The linear scale bar has units counts s − . The radio contours of the radio haloin VLA at 1.4 GHz with a resolution of 15 ′′ are overlayed on the X-ray image. The σ noise level in the image plane is 0.03 mJy/beam. Contours start at 0.09mJy/beam and then scale by a factor of √
2. The radio image is from (Feretti et al. 2001). Right: Sectors showing cluster regions where spectra were extractedboth toward the south-western direction and around the edge (top) and in the perpendicular direction (bottom).
RG (J1615-061)
600 kpc
RG (J1615-061)
600 kpc
The XMM-Newton observations of ACO2163 were done in August2000. We retrieved the data of this cluster from the XMM-NewtonScience Archive (obsID 0112230601, PI M. Turner, public data).The raw time of observation was 16.3 ks. The data were obtainedwith the THIN1 filter with the EPN and EMOS 1 and 2 camerasand then analysed using the SAS (Science Analysis System devel-oped by the XMM-Newton team) tool from the Heasarc packageto do the main part of the reduction. After removing the flares, weobtained a reduced image of ∼ phabs and mekal models. The mekal model is an emission spectrum from hot diffuse gas based onthe model calculations of Mewe et al. (1985, 1986) with Fe L cal-culations by Liedahl et al. (1995). The model includes line emis-sions from several elements, with the plasma temperature (in keV),H density (cm − ) and Metal abundance as free parameters, to fitthe different spectra taken from MOS1, MOS2 and PN, extractedin the 0.5-8.0 kev band range from several rectangular bins in the same sector across the cluster and the brightness edge.To this end, we used the sectors shown in Figure 1 - right panels- to extract spectra and obtain the temperature. We also extracteda background which we then subtracted from our spectra and tookcare of masking possible point sources in the background and clus-ter that would bias our measurements. The choice of our bins wasto ensure that there is enough signal to provide a good spectral fit.We divided the brightness region into seven bins. Three of the sevenbins were chosen to have the size of 163 ′′ x 40 ′′ . However, the binsaround the edge (four bins) were made to be smaller (163 ′′ x 13 ′′ - see Figure 1 - top-right panel). Higher resolution spatial binningis ideal for minimizing contamination due to bright emission justinside the edge and closely probes the shock (Russell et al. 2010). We thus obtained a temperature profile along the main axis of thecluster to determine the gas temperature jump across the edge. Asa check, we also extracted spectra in the area shown in the bottom-right panel of Fig. 1, using the same method described above, toobtain the temperature variation behaviour in the northwest - south-east perpendicular direction. The temperatures and the profiles areshown in Table 1 and in Fig 2 - left panel, respectively, and the er-rors are approximately 7% -20%.We see a sudden drop in temperature (Fig. 2) at the location of thebrightness edge shown in Fig. 1, where we argue there is a shock,which occurs at ∼
200 kpc ( ∼ ′′ ) from the cluster centre. The3.2 LuminosityProfileacrosstheEdge Shock front in ACO2163 −400 −200 0 200 400 Distance [Kpc] T e m p e r a t u r e ( K e V ) −300 −200 −100 0 100 200 300 Distance [Kpc] T e m p e r a t u r e ( K e V ) −400 −200 0 200 400 Distance [kpc] L u m i n o s i t y ( e r g / s ) Figure 2.
Left panel: Temperature profiles along the main axis (red line) and in the perpendicular direction (green line). The vertical line represents the locationwhere the shock is expected. The null distance represents the center of the cluster. The temperature obtained in the inner regions of the cluster is similar to theones found in the literature (see e.g. Markevitch & Vikhlinin 2001; Bourdin et al. 2011). Middle panel: Temperature profile around the region of the shock.The shaded areas show 68% confidence intervals. The solid line is the best-fit model (see text). Right panel: Plots of the luminosity obtained along the mainaxis (red dots) and in the perpendicular sector (green dots). The location of the luminosity jump is shown by the vertical line.
Table 1.
Results from the fit of the temperature profile across the edge and in the perpendicular direction. The errors are within the 68% confidence limit.Shock Profile Perpendicular ProfileTemperature (keV) Distance (kpc) Chi-squared Temperature (keV) Distance (kpc) Chi-squared13.3 ± ± ± ± ± ± ± ± ± ± ± ± ± ± temperature drops by more than 6 keV between before and after theedge. However, in the case of the perpendicular profile, we observetemperatures between 10.283 keV and 13.124 keV with no suddendrop in temperature on the galaxy cluster edges. This implies thatthe behaviour we observe in the area of the edge is specific andis found only on the brightness edge of this galaxy cluster. Thus,the temperature profile confirms that the brightness edge is a shockfront. The temperature jump of the right sign across the brightnessedge leads to an approximate value of the ratio of the post-shock topre-shock temperatures of 2.5 ± γ = /
3, we determinethe Mach number: M = (cid:18) ( γ + ) ( T / T − ) γ ( γ − ) (cid:19) (1)to be 2 . ± .
3. This value of the Mach number is a lower limitsince for the plane shock the maximum of temperature should bejust behind the shock (see Fig 2 - left panel).The drop in temperature near the centre of the cluster in both pro-files (Fig 2 - left panel) is due to the presence of the cool corereported by Bourdin et al. (2011). The cool core is located at ap-proximately ∼
50 - 100 kpc from the centre, in the SW direction.In Fig. 2-middle panel, we show a simple best-fit shock model onour temperature profile. To model the temperature profile we useda simple step function T ( r ) = (cid:26) T r ≤ r j , T r > r j . (cid:27) where r j is the location of the jump discontinuity. From the fit weobtained the step amplitude (difference between temperature values before and after the discontinuity caused by the shock) which is7.0 ± ± χ / d . o . f = .
2. This jump leads tothe Mach number M = . ± We looked at the luminosity profile in the region around the pre-sumed shock to check for a sudden decrease in luminosity. To thisend, we used bin sizes (165 ′′ x 29 ′′ ) which provided enough signalto extract spectra and obtain the luminosity. We chose the smallestpossible window providing converging X-ray parameter estimatesto have as many measurements as possible along the direction ofthe presumed shock.To obtain the luminosity, we used the mekal model using tempera-ture, H density and abundance as free parameters. Since we knowthe redshift of Abell 2163, we set its value to 0.203 and we let theother 3 parameters free. The luminosity profile is shown in Fig. 2-right panel (solid line).The brightness edge is clearly visible in the luminosity profile asa sudden drop in luminosity by a factor of ∼ ≈ . × ergs / s . The jump in luminosity occurs at adistance of ∼
200 kpc from the cluster centre, which is the samelocation where we observe a jump in temperature. We also lookedat the luminosity profile in the sector that is perpendicular to theSW direction where the shock is located (dashed line). We observeno luminosity jump in this direction. We thus conclude that the lu-minosity jump is associated to the shock.Since the X-ray luminosity L x ∝ n T − / V for the XMM band we4 RADIOANALYSISANDRESULTS Shock front in ACO2163 Table 2.
The properties of ACO2163 and the details of KAT-7 observations.Cluster Name z RA (J2000) DEC (J2000) Diameter (arcmin)(h m s) (min s arcs) (at 5R500: X-ray)ACO2163 0.20 16 15 34.1 -06 07 26 37.81Cluster Name Freq. BW Observation Date Observing time Antennas FWHM, p.a. rms(MHz) (MHz) (hours) ( ′′ x ′′ ), degree (mJy beam − )ACO2163 1826.6875 234765 05-Oct-2012 4.48 7 226 × are using (Basu et al. 2010), hence we can re-write this relation as L postx L prex = (cid:16) n post n pre (cid:17) (cid:16) T post T pre (cid:17) − / × . n post n pre = (cid:16) L postx L prex (cid:17) / / h(cid:16) T post T pre (cid:17) − / × ( . ) / i (3)where n is the mean gas density and V is the gas volume, ’pre’ and’post’ refer to pre-shock and post-shock regions, respectively. Us-ing the previous values for the temperature jump (2.5 ± ± ρ ρ , which is 2.2 ± M = ρ ρ ( γ + ) − ( γ − ) ρ ρ ! / (4)and was found to be 1.9 ± ACO2163 was observed in 2012 with the decommissioned seven-dish KAT-7 array, an engineering testbed for the MeerKAT whichis South Africa’s pathfinder for the Square Kilometer Array (SKA).The KAT-7 telescope has been used to detect HI in nearby galax-ies (Carignan et al. 2013), to study extended radio haloes in galaxyclusters (Riseley et al. 2015; Scaife et al. 2015) and to study time-variability of radio sources (Armstrong et al. 2013), and we refer tothese papers for the technical specifications of the telescope.ACO2163 is one of the eight galaxy clusters in our sample that wereobserved at 1.83 GHz using the KAT-7 array. The sources were se-lected based on their extent, i.e. they had a diameter that is morethan 20 ′ , from a list of clusters that are observable in the South-ern Hermisphere. The data were obtained with a total observingbandwidth of 235 MHz divided in 601 channels. The properties ofACO2163 and the observation information are summarized in thecompound Table 2.The data were reduced and calibrated using the NRAO CommonAstronomy Software Applications (CASA) (McMullin et al. 2006),which is the standard data reduction package that is used for the re-duction of the KAT-7 data.Before the calibration, bad data and RFI were removed within CASA, and the 601 channels were averaged to 9, each having awidth of 25 MHz in order to reduce the size of the data set. Theinterval width for time averaging was 25s, and 64 channels wereaveraged to output each of the 9 channels. The choice for thesevalues was to ensure that time/frequency smearing are minimized.Flux calibrator source PKS 1934-638, and the phase calibrator PKS1621-115, were used to calibrate the ACO2163 data in flux and inphase, respectively. The flux density calibrator source was also usedto do bandpass calibration.After a number of self-cal procedures which were applied in orderto reduce residual phase variations, deconvolution was achieved us-ing the CLEAN algorithm in CASA. The images were produced us-ing a Briggs robustness parameter of 0.5. This value (Briggs=0.5)has the advantage of minimizing noise while giving a better res-olution with respect to natural weighting, allowing for good pointsource detection in the resultant images. Point sources in the back-ground were identified and subtracted in the uv data using the taskUVSUB before the search for diffuse emission was done. Onlysources above the confusion limit of 1 mJy/beam were subtracted. nterms=2 , which is the number of terms in the Taylor polynomial,was used to model the frequency dependence of the sky emission.A multi-scale CLEANing algorithm in CLEAN was used in orderto detect diffuse, extended structures on both smaller and large spa-tial scales. In order to pick up the diffuse emission, the final back-ground source-subtracted image was produced by means of naturalweighting. We have detected the ACO2163 radio halo and the extended north-eastern relic source in the KAT-7 data (Figure 3). The two sourcesseem to be connected by a wide bridge. The radio halo has a to-tal size of ∼ ′ and a total flux density S . GHz = ± . ∼
103 mJy obtained by scaling the 1.4 GHz flux tothat at 1.83 GHz using the spectral index across the 20 cm bandwhich is about 1.6 (Feretti et al., 2001). The discrepancy could bedue to some negative residuals, as suggested by the negative bowlin the bottom left in Figure 3, where there is extended emission de-tected from the VLA.Since the KAT-7 observations are of poor resolution, our discussionof the flux results for the halo are limited, as well as those of theextended north-eastern relic source which will follow later.In addition to the KAT-7 data, we obtained ready-to-use radio5.1 SpectralIndexAnalysis Shock front in ACO2163
Figure 3.
KAT-7 radio contours of ACO2163 (white) after subtraction of background sources in the field of the cluster are overlayed on the XMM-Newtonimage. The radio contours of the halo in NVSS are overlayed on the image in magenta. The restoring beam of the radio image is 226 ′′ × ′′ (PA=153 deg),and the noise level in the image plane is 1.1 mJy/beam. The contours start at 3.3 mJy/beam and then scale by a factor of √
2. The restoring beam is in thebottom-left hand corner. NVSS contours start at 1.6 mJy/beam and also scale by a factor of √
2. The yellow line indicates the location of the shock.
Figure 4.
Radio contours of the halo in VLA at 1.4 GHz, with resolution of45 ′′ × ′′ , are overlayed on the XMM-Newton X-ray image. The σ noiselevel in the radio image plane is 0.03 mJy/beam. Contours are at -3, 3, 6,12, 24, 48, 96 mJy/beam times the noise level. The units in the colour barare mJy/beam. The radio image is from Feretti et al. (2001). halo and spectral index maps of Feretti et al. (2001, 2004) that weutilised to read out spectral and flux profiles. The radio maps wereobtained at 1.4 GHz and 0.3 GHz with the VLA by Feretti et al. (2001, 2004). The high and low resolution images of the radiohalo at 1.4 GHz are shown in Figures 1 and 4 as contours over-laid on the X-ray image, and show the detection of a giant radiohalo, with a total extent of ∼ ′ (2.3 Mpc) and an elongation inthe E-W direction. After the subtraction of point sources, the to-tal fluxes in the halo were found to be S . GHz = ± S . GHz = ±
10 mJy, which led to an average spectral index α . . = . ± . The spectral index map was obtained by comparing the 90 cm and20 cm images produced with the same beam and cellsize. We thinkthat the flux is not missed at 20 cm and therefore the spectrum isnot affected by the slightly different u-v spacings for the followingreasons: (i) diffuse emission at the 2 frequencies has similar extent,(ii) the shortest baselines at 90 cm are provided by very few inter-ferometers, (iii) the spectral index map has values consistent withthe total spectral index, and (iv) the 20 cm map is much more sen-sitive than the 90 cm one (see details in Feretti et al. (2004))As expected from electron re-acceleration models, regions influ-6 RADIOANALYSISANDRESULTS Shock front in ACO2163 - : : . - : : . : . : . : . : . R Figure 5.
This figure, obtained from Thölken et al. (2018), shows the po-sitions of the shock fronts they detected in A2163 (blue), and the positionof our newly detected shock front (red) near the approximate position ofthe cool core (magenta ellipse). The black diamond represents the X-rayemission peak. enced by the shock (and turbulence) should show spectral flatten-ing, indicating the presence of more energetic radiating particlesand/or a larger value of a local magnetic field strength (Feretti et al.2001). Feretti et al. (2004) identified two regions where they sawevidence of flat spectra in the spectral index map: the vertical regioncrossing the cluster centre (in the N-S direction) to a distance of ∼ ′′ from the cluster centre, and the western halo region whichwas observed to be much flatter than the eastern region. This isthe same side of the cluster where the new shock front is located,as well as the two other shocks which have been detected recently(Thölken et al. 2018), one at ∼
700 kpc and the other at ∼ ∼ M = . ∼ ′′ from the core location. The sudden decline in radio fluxcorresponds with the radio edge. The shock front reported in this paper is embedded in the halo andas a result we do not observe any relic source coincident with theshock. A diffuse source on the N-E side of the halo in ACO2163claimed to be a relic, however, was detected by Feretti et al. 2001(labelled D3 in their Figure 2). Another emission, D4, was seensouth of D3, and no connection between the two sources was es-tablished. Feretti et al. 2001 ruled out the possibility that thesetwo sources are linked with the strong eastern radio source, J1616-061 (hereafter source S), located at RA=16h16m22s and Dec=-06d06m34s (see also overlaid NVSS contours in Figure 3). Laterobservations by Feretti et al. 2004 confirmed D3 (which is la-belled ’R’ in their Figure 2) as a relic source, with a spectral index α . . = . ± . ∼ ′ (1.8 Mpc) fromthe cluster center in our KAT-7 image, we detect large-scale dif-fuse emission which encompasses both D3 and D4. The size of thissource is ∼ ′ ( ∼
993 kpc).The measured flux of the relic before subtraction of the point sourceS is 55 mJy. The subtracted clean component of S contributes a fluxdensity of only S . GHz =
33 mJy in KAT-7. So the resulting fluxof the relic source should be ∼
22 mJy after subtraction. However,7.2 TheNorthEasternRadioRelic Shock front in ACO2163 −400 −200 0 200 400 600 800 1000
Distance (kpc) Sp e c t r a l I n d e x | RG Shock −400 −200 0 200 400 600 800 1000
Distance [kpc] F l u D e n s i t y ( m J y / b e a m ) | RG Shock
Figure 6.
Top panel: Radio contours of the halo in VLA at 1.4 GHz, with resolution of 45 ′′ × ′′ , are overlayed on the colour-scale spectral index imagebetween 0.3 GHz and 1.4 GHz. The image is obtained with a resolution of 60 ′′ × ′′ . The contours are at -3, 3, 6, 12, 24, 48, 96 mJy/beam times the noiselevel. The location of the shock is represented by a solid white line and that of the radio galaxy is also indicated. The spectral index image is from Feretti et al.(2004). Bottom panels: 1-dimensional slice, Left: of the spectral index image showing the spectral index distribution along the direction indicated by thedashed line segment that passes through the indicated location of the radio galaxy in the spectral index image. The origin of the distance scale starts at theapproximate cluster centre at RA=16h15m51.2s, -06d08m01.6s. The locations of the radio galaxy and the shock are indicated. Right: showing the flux densitydistribution in the low resolution VLA radio image along the same dashed line segment shown in the spectral index image. The vertical line indicates thelocation of the radio edge. after subtraction we measure a flux density of 42.6 mJy for the relic, ∼
50% more than what we expect. The reason for this discrepancycould be that the point source was picked up using rhobust = 0.5which gives high resolution. After subtraction in the uv data the fi-nal point source subtracted image was produced with weighting =natural. So the point source in the weighting = natural image wouldbe more extended than in the rhobust = 0.5 image, thus leavingsome residual emission after subtraction.Feretti et al. 2004 measured a value of S . GHz = . S . GHz =
22 mJy suggeststhat the relic source in ACO2163 is more extended due to the KAT- 7 high sensitivity, with possibly more flux density at 1.4 GHz thanwas previously observed (i.e. less flux was possibly picked up at1.4 GHz due to the high resolution of the long-baseline interferom-eter).Feretti et al. 2004 suggested the possibility that source S is a tailedradio galaxy, extended toward the south. If this is the case thenthat would mean our relic emission is also contaminated by the tailemission of this source (source D), which would also explain thehigh flux value that we measured for the relic. Though it is not pos-sible to determine the accurate flux value for this source in KAT-7,it is reassuring to notice that the location of the M = . The main goal of this paper is to present evidence for the existenceof a shock front in ACO2163, located at ∼ ′′ ( ∼
200 kpc) fromthe cluster center in the X-ray map, on the south-western regionof the cluster. We have derived the Mach number associated withthis shock from the direct measurement of a gas temperature jump.The temperature jumps from 4.1 keV to 10.3 keV by a factor of2.5 ± M = v s / c s = ± c s isa sound speed of ∼ − . The shock velocity v s = Mc s hasbeen determined to be ∼ − .A detection of a shock front in ACO2163 supports the results ofBourdin et al. (2011) which have revealed I) the ‘bullet’ nature ofthe supersonic infalling substructure in ACO2163, II) the westwardmotion of a stripped cool-core/‘bullet’ embedded in the hotter at-mosphere of ACO2163-A, and III) adiabatic compression of the gasbehind the ‘bullet’ due to shock heating. These observations pointto a shock front being present in the inner regions of ACO2163,which we argue we have detected. The results of Bourdin et al.(2011) also showed the similar nature of ACO2163 to that of the‘Bullet’ cluster where a shock front that was preceding a ’bullet’was detected (Markevitch et al. 2002).Given the location of the ‘bullet’ which is at a projected distanceof 290 kpc from the main cluster (see Bourdin et al. 2011), and as-suming that the subcluster moves at the shock velocity, this velocityimplies that the ‘bullet’ crossed the main cluster nearly 0.1 Gyr ago.Our value of the Mach number is a typical value for a shock frontin merging clusters ((see e.g. Markevitch et al. 2005, for ACO520),(Macario et al. 2011, for ACO754), or (Akamatsu & Kawahara2013, for CIZA2242, ACO3667, and ACO3376)), and since shockfronts are routinely observed in merging clusters, the detection ofa shock in ACO2163 supports the results (e.g. Bourdin et al. 2011)which show the merging nature of this cluster.Radio observations of a number of merging galaxy clusters pro-vide evidence for the presence of relativistic electrons and mag-netic fields in the intracluster medium through the detection of syn-chrotron radio emission in the form of Mpc-scale sources called ra-dio halos (e.g. Feretti et al. 2012; Brunetti & Jones 2014). The rela-tivistic electrons can be (re)accelerated by turbulence and shocks inthe ICM, and the process of shock re-acceleration can be efficientfor shocks with M > Fermi mechanismof relativistic electrons by the M ∼ α = ( p − ) / + . ∼ α = α in j + . α in j =( p − ) / ∼ p is the slope of the energy spectrum of theelectrons generated by the shock, and is related to the Mach number of the shock through p = M + M − α . . = . ± ∼ ∼ α ( int ) . . ∼ . ∼ ∼ α in j or α s + ∼ M radio = (cid:18) α s + α s − (cid:19) (6) ∼ .
2, where M radio is the Mach number from the radio observa-tions for a steady planar shock. This value of the Mach number isin excellent agreement with that derived from the X-ray observa-tions, in support of re-acceleration by DSA. This could mean thatthe electron population injected at the shock with M ∼ α ∼ . In this paper we have presented a combined X-ray and radio studyof the nearby galaxy cluster ACO2163.The following are the main findings from our study and analysis:1. The XMM-Newton observations have confirmed the presenceof a shock front south-west of the cluster centre in ACO2163. Wehave estimated the Mach number from the X-ray temperature ratiobefore and after the shock, and have found a value of M=2.2 ± . M = . α ∼ .
6) whichwas observed for two nearby frequencies in ACO2163, implecatesa Mach number M = .
2, in good agreement with our X-ray results.The case of the shock in ACO2163 is a unique example since, incombination with re-acceleration of pre-existing electrons whichis obvious, turbulent acceleration seems also important behind theshock (within the halo).The KAT-7 results seem to suggest a more extended radio emissionat the relic source location. This is due to the large beam and en-hanced sensitivity of this instrument. With the MeerKAT we willbe able to detect halo and relic sources with radio power approxi-mately ten times better when compared to existing arrays, and thiswill enable a more detailed study of the interaction between theshock front and the radio plasma.
NM would like to thank Thomas Reiprich, Sofia Tholken, Kaus-tuv Basu and Reionout van Weeren for the insightful discussionswhich helped improve the quality of this paper. This work is based on the research supported by the National Research Foundation ofSouth Africa (grant number 111735). The KAT-7 telescope is op-erated by the South African Radio Astronomy Observatory, whichis a facility of the National Research Foundation, an agency of theDepartment of Science and Innovation.
Data is available on request from the authors.
References
Akamatsu H., Kawahara H., 2013, PASJ, 65Arnaud M., Hughes J. P., Forman W., Jones C., Lachieze-Rey M., Ya-mashita K., Hatsukade I., 1992, ApJ, 390, 345Basu K., et al., 2010, A&A, 519, A29Blandford R., Eichler D., 1987, Phys. Rep., 154, 1Bourdin H., et al., 2011, A&A, 527, A21Bourdin H., Mazzotta P., Markevitch M., Giacintucci S., Brunetti G., 2013,ApJ, 764, 82Brunetti G., Jones T. W., 2014, International Journal of Modern Physics D,23, 1430007Buote D. A., 2001, ApJ, 553, L15Colafrancesco S., Marchegiani P., Paulo C. M., 2017, MNRAS, 471, 4747Donnert J., Dolag K., Brunetti G., Cassano R., 2013, MNRAS, 429, 3564Elbaz D., Arnaud M., Boehringer H., 1995, A&A, 293Enßlin T. A., Gopal-Krishna 2001, A&A, 366, 26Ensslin T. A., Biermann P. L., Klein U., Kohle S., 1998, A&A, 332, 395Feretti L., 2005, Advances in Space Research, 36, 729Feretti L., Fusco-Femiano R., Giovannini G., Govoni F., 2001, A&A,373, 106Feretti L., Gioia I. M., Giovannini G., eds, 2002, Merging Processes inGalaxy Clusters Astrophysics and Space Science Library Vol. 272,doi:10.1007/0-306-48096-4.Feretti L., Orrù E., Brunetti G., Giovannini G., Kassim N., Setti G., 2004,A&A, 423, 111Feretti L., Giovannini G., Govoni F., Murgia M., 2012, A&ARv, 20, 54Finoguenov A., Sarazin C. L., Nakazawa K., Wik D. R., Clarke T. E., 2010,ApJ, 715, 1143Gabici S., Blasi P., 2003, Astroparticle Physics, 19, 679Giacintucci S., et al., 2008, A&A, 486, 347Govoni F., Markevitch M., Vikhlinin A., van Speybroeck L., Feretti L., Gio-vannini G., 2004, ApJ, 605, 695Herbig T., Birkinshaw M., 1994, in American Astronomical Society Meet-ing Abstracts. p. 1403Kang H., Ryu D., 2015, ApJ, 809, 186Kang H., Ryu D., 2016, ApJ, 823, 13Landau L. D., Lifshitz E. M., 1959, Fluid mechanicsLiedahl D. A., Osterheld A. L., Goldstein W. H., 1995, ApJ, 438, L115Macario G., Markevitch M., Giacintucci S., Brunetti G., Venturi T., MurrayS. S., 2011, ApJ, 728, 82Markevitch M., 2010, preprint, ( arXiv:1010.3660 )Markevitch M., Vikhlinin A., 2001, ApJ, 563, 95Markevitch M., Mushotzky R., Inoue H., Yamashita K., Furuzawa A.,Tawara Y., 1996, ApJ, 456, 437Markevitch M., Gonzalez A. H., David L., Vikhlinin A., Murray S., FormanW., Jones C., Tucker W., 2002, ApJ, 567, L27Markevitch M., Govoni F., Brunetti G., Jerius D., 2005, ApJ, 627, 733Maurogordato S., et al., 2008, A&A, 481, 593Mewe R., Gronenschild E. H. B. M., van den Oord G. H. J., 1985, A&AS,62, 197Mewe R., Lemen J. R., van den Oord G. H. J., 1986, A&AS, 65, 511Nord M., et al., 2009, A&A, 506, 623Ogrean G. A., Brüggen M., 2013, MNRAS, 433, 1701