An Analysis of Soft X-ray Structures at Kiloparsec Distances from the Active Nucleus of Centaurus A Galaxy
Dominika Ł. Król, Volodymyr Marchenko, Michał Ostrowski, Łukasz Stawarz
DDraft version September 17, 2020
Typeset using L A TEX twocolumn style in AASTeX63
An Analysis of Soft X-ray Structuresat Kiloparsec Distances from the Active Nucleus of Centaurus A Galaxy
Dominika (cid:32)L. Kr´ol, Volodymyr Marchenko, Micha(cid:32)l Ostrowski, and (cid:32)Lukasz Stawarz Astronomical Observatory, Jagiellonian University, ul. Orla 171, 30-244 Krak´ow, Poland
ABSTRACTHere we re-analyze the archival
Chandra data for the central parts of the Centaurus A radio galaxy,aiming for a systematic investigation of the X-ray emission associated with the inner radio lobesin the system, and their immediate surroundings. In particular, we focus on four distinct featurescharacterized by the soft excess with respect to the adjacent fields. Those include the two regionslocated at kpc distances from the nucleus to the West and East, the extended bow-shock structure tothe South, and a fragment of a thin arc North from the center. The selected North, West, and Southfeatures coincide with the edges of the radio lobes, while the East structure is seemingly displaced fromthe radio-emitting plasma. Our X-ray spectral analysis reveals (i) a power-law emission componentwith photon index Γ ∼ ∼ . − ) and relatively cold (temperature ∼ . Keywords: radiation mechanisms: non-thermal — ISM: jets and outflows — galaxies: active — galax-ies: jets — galaxies: individual (Centaurus A) — X-rays: galaxies INTRODUCTIONThe famous radio galaxy Centaurus A, hosted byNGC 5128, is located at the distance of 3 . ± .
35 Mpc(Rejkuba 2004; Ferrarese et al. 2007). NGC 5128 is an el-liptical with a prominent dust lane and other morpholog-ical features indicating multiple merger events which oc-curred about 200–700 million years ago (see the reviewsby Israel 1998; Morganti 2010, and references therein).It harbors a central black hole with the estimated massof (0 . − . × M (cid:12) (Marconi et al. 2006; Neumayeret al. 2007; Cappellari et al. 2009). Cen A is a strongsource of a multi-frequency emission detected on vari-ous scales. The large scale structure of the system hasbeen resolved in the radio domain, from very low fre-quencies up to mm wavelengths, to have an angular sizeof 8 ◦ × ◦ , equivalent to the projected linear dimensionsof 500 kpc ×
250 kpc (Hardcastle et al. 2009; Feain et al.
Corresponding author: Dominika (cid:32)L. Kr´[email protected] γ -rays by the Large AreaTelescope (LAT) onboard the Fermi satellite (Abdo etal. 2010a; Sun et al. 2016), and selectively mapped inX-rays with
Suzaku (Stawarz et al. 2013).On smaller scales, in particular at the distance ofseveral–to–tens of kpc North from the Cen A nucleus, adiffuse and low-surface brightness radio structure calledthe“northern middle lobe” is seen (Morganti et al. 1999).Around and within this structure, a complex net of op-tical filaments of ionized gas, clouds of atomic gas withanomalous velocities, young stars, and large-scale X-ray filaments composed of discrete knots, have been ob-served, all suggestive of a complex interaction betweenthe evolving large-scale radio jet with the interstellarmedium (Morganti et al. 1999; Oosterloo & Morganti2005; Kraft et al. 2009; Crockett et al. 2012; Neff et al.2015; Salom´e et al. 2016).Finally, on yet a smaller scale of a few/several arcmin,the inner structure of Cen A radio galaxy consists of sev-eral components clearly visible in radio and X-rays whenimaged with the arcsec resolution, including the bright a r X i v : . [ a s t r o - ph . H E ] S e p Kr´ol et al.
Figure 1.
The
Chandra
ACIS intensity map of the inner parts of Cen A, within the photon energy range 0 . − . . − . σ Gaussian. nucleus, the jet extending to the North-East up to fourkpc from the core, and the counter-lobe pronounced tothe South (e.g., Hardcastle et al. 2007; Croston et al.2009). The X-ray emission spectrum of the nucleus inthe energy range 3 −
10 keV is well fitted by a heavilyabsorbed power-law model plus a neutral and narrowfluorescence iron line (Evans et al. 2004); the jet X-ray emission continuum, contributed by multiple brightknots and a diffuse component, is best described as anunabsorbed steep power-law (see Kataoka et al. 2006;Snios et al. 2019). Another X-ray feature within the in-ner parts of the Cen A system, is a ring-like structureextending to several kpc in the direction perpendicularto the jet, as reported by Karovska et al. (2002).The Cen A core has been detected in soft and high-energy γ -rays by all the instruments onboard the Comp-ton Gamma Ray Observatory (Steinle et al. 1998, andreferences therein), as well as by the Fermi -LAT (Abdoet al. 2010b). The radio galaxy has also been detectedin the very high energy γ -ray range ( >
100 GeV) by theH.E.S.S. observatory (Aharonian et al. 2009). The mostrecent analysis of the broad-band γ -ray continuum of thesource reveals a spectral hardening above the photon en-ergies of a few GeV (Sahakyan et al. 2013; Abdalla etal. 2018); the extension of the H.E.S.S. source reportedrecently by Sanchez et al. (2018), seems to point out thekpc-scale jet as the most likely origin of this observed“excess” γ -ray emission (see, e.g., Tanada et al. 2019).In this paper we re-analyze the archival Chandra datafor the central parts of the Cen A radio galaxy, focusingon the spectral analysis for the diffuse features associ-ated with the inner radio lobes, and characterized by theexcess soft X-ray emission with respect to the adjacentfields. The data acquisition, analysis, and modeling isdescribed in §
2; the interpretation and the discussion of the obtained results are given in § § CHANDRA
DATAWe have reviewed all the available
Chandra
AdvancedCCD Imaging Spectrometer (ACIS) data for the in-nermost region of the Cen A system, and selected theexposures for which the ACIS readout streaks are re-stricted (as much as possible) to the plane perpendicu-lar to the jet axis, in order to avoid any overlaps withthe lobes and their immediate surroundings. The Ob-sIDs of thus selected observations are: 316, 962, 2978,3965, 8489 and 8490. For these, the analysis was car-ried out with the software package
CIAO 4.10 (Fruscioneet al. 2006) and the calibration database
CALDB 4.7.9 .Before the analysis, the data was reprocessed using the chandra repro script recommended in the
CIAO analy-sis threads. Next the data was merged and binned witha factor of 1.0, which corresponds to the original
Chan-dra pixel size of 0 . (cid:48)(cid:48) . The images of the selected ob-servations, merged and smoothed with 3 σ Gaussian, areshown in the Figure 1 for the energy ranges 0 . − . . − . X-ray hardness analysis
The
Chandra map of Cen A radio galaxy in the energyrange 0 . − . . − . oft X-ray Structures in Cen A Galaxy Figure 2.
The smoothed (Gaussian of σ = 5 px) hard-ness ratio map for the inner parts of the Cen A radio galaxy,with the radio continuum intensity contours superimposed(black), starting from the 0.2 Jy/beam level, and increasingby a factor of √
2. The regions selected for the spectral anal-ysis — East, West, South and North — are denoted withsolid red contours, and the corresponding background re-gions by dashed red contours. Also marked are the pointsources (mostly foreground XRBs) removed from the spec-tral extraction regions. ray range, although both structures can also be notedon the hard X-ray
Chandra map. In addition to those,in the soft image, a fragment of a thin but distinct arclocated to the North from the nucleus is clearly visibleas well; its position and orientation are both consistentwith the extension of the edge of the main radio lobe.Figure 2 presents the hardness ratio map of the ana-lyzed system, based on the exposure-corrected images,and smoothed with the Gaussian of σ = 5 px. The hard-ness ratio here is defined as the ratio of the 2 . − . . − . Chandra ’s ACIS (whichis of the order of 150 eV). For comparison, in the fig-ure we also display the contours of the lobes’ continuumradio emission at 21 cm, taken from the NRAO VeryLarge Array (VLA) archive (Condon et al. 1996). Asshown in the figure, while the jet, the Southern bow-shocks, and the Northern arc, are all still prominenton the hardness map, the soft diffuse X-ray hour-glass The National Radio Astronomy Observatory is a facility of theNational Science Foundation operated under cooperative agree-ment by Associated Universities, Inc. structure “collapses” to two distinct but symmetric fea-tures with relatively sharp and well defined boundaries,located into the East and into the West of the core.Moreover while the aforementioned North, West, andSouth regions do overlap well with the lobes’ edges onthe radio map, the East feature appears seemingly dis-placed from the radio-emitting plasma of the main lobe.We note that the two peculiar structures East andWest are inclined at ∼
35 deg with respect to the jetaxis, and are approximately cone-like shaped; the core-vertex angular separation for both reads as ∼ . Spectral modelling
For the spectral analysis of the selected regions, firstwe extracted the corresponding spectra from each Ob-sID considered in our study, using the specextract toolfrom the
CIAO software package. Next we combinedthe spectra for each region using the combine spectra script. Because of the extended wings of the pointspread function (PSF) of the extremely bright nucleus,which in addition is subjected to a severe pile-up inthe instrument, we restricted the analysis to the pho-ton energy range up to 3 keV (see in this context also
Kr´ol et al. C oun t s / s e c / k e V Energy (keV) S i g m a −202 C oun t s / s e c / k e V Energy (keV) S i g m a −4−2024 C oun t s / s e c / k e V Energy (keV) S i g m a −4−202 C oun t s / s e c / k e V Energy (keV) S i g m a −2−1012 Figure 3.
Chandra spectra along with the best-fit models (and residuals) for the four regions selected for the analysis: East(upper left), West (upper right), South (lower left), and North (lower right). The models displayed consist of a mixture ofabsorbed thermal and non-thermal components, xsphabs*(xsapec+xspowerlaw) , for the West and East regions, and absorbedpower-law, xsphabs*xspowerlaw , for the North and West regions.
Croston et al. 2009). We note, that both the East andWest regions are located at similar distances from thecore, and so are their background regions, and hence anyphoton leakage from the central PSF should affect thespectral analysis of both regions in a similar/comparablemanner; the North and especially the South regions, onthe other hand, are expected to be subjected to a muchlesser extent by the pollution from the core emission.After the background extraction, the fitting was per-formed for the grouped data (with the minimum signal-to-noise ratio = 5 for the West, East and South region,and minimum signal-to-noice ratio = 7 for the North re-gion because of a much lower photon statistics) using the
Sherpa fitting application (Freeman et al. 2001). Theinitial values of the model free parameters were chosenbased on the preliminary fitting using the Monte-Carlomethod in
Sherpa . The fitted model for all the regions included a ther-mal component ( xsapec ) plus a power-law component( xspowerlaw ), moderated by the Galactic absorptionand the internal absorption ( xsphabs ). All the modelparameters were set free, except of the Galactic equiv-alent hydrogen column density in the direction of thesource, which was frozen at N H , Gal = 7 . × cm − following Kalberla et al. (2005). For the North andSouth regions, we applied a simple absorbed power-lawmodel, with no contribution from a thermal compo-nent; in fact for these regions the two-component model(power-law plus apec) did not provide any substantialimprovement over the one-component model (consistingof a single power-law emission). The spectra of the Eastand West regions, on the other hand, could not be fittedat all with a one-component model, consisting of eitheran absorbed single power-law emission, or an absorbed oft X-ray Structures in Cen A Galaxy Table 1.
Spectral fitting results
Region/Model Parameter Value 1 σ errors Units East kT 0.23 0.01 keV xsphabs*(xsapec+xspowerlaw) norm 0 . . − × apec Abundanc 0.09 0.03 —Γ 0.24 0.66 —ampl 2 . . − × ph/keV/cm at 1keV N H cm − Final fit statistic 125.33Degrees of freedom 92
West kT 0.19 0.01 keV xsphabs*(xsapec+xspowerlaw) norm 1 . . − × apec Abundanc 0.20 0.09 —Γ 1.75 0.28 —ampl 32 . . − × ph/keV/cm at 1 keV N H cm − Final fit statistic 208.34Degrees of freedom 128
South kT 0.52 0.03 keV xsphabs*(xsapec+xspowerlaw) norm 0 .
06 0 .
01 10 − × apec Abundanc 0 (unconstr.) —Γ 1.00 0.20 —ampl 15 . . − × ph/keV/cm at 1keV N H cm − Final fit statistic 179.68Degrees of freedom 118 xsphabs*xspowerlaw
Γ 2.14 0.13 —ampl 46 . . − × ph/keV/cm at 1keV N H . +0 . − — cm − Final fit statistic 104.49Degrees of freedom 121
North kT 0.16 0.05 keV xsphabs*(xsapec+xspowerlaw) norm 8 × − (unconstr.) 10 − × apec Abundanc 0 .
24 (unconstr.) —Γ 2.52 0.71 —ampl 4 . . − × ph/keV/cm at 1keV N H +0 . − — cm − Final fit statistic 43.69Degrees of freedom 59 xspowerlaw
Γ 2.02 0.43 —ampl 3 . .
65 10 − × ph/keV/cm at 1keVFinal fit statistic 8.82Degrees of freedom 15 single-temperature plasma, and hence below we do notdiscuss those attempts.The final results obtained with the Levenberg-Marquardt optimalization method, using the chi-squared statistic ( chi2datavar ) with variance calcu-lated from the data, are presented in Figure 3, and thecorresponding best-fit values of the model free parame-ters are summarized in Table 1. In Figure 4, we provide also the confidence contour plots of the main model pa-rameters for the two-component model applied to theEast and West regions. The results of the spectral fit-ting for all four analyzed regions can be summarized asfollows: • In the East region, which is seemingly displacedfrom the radio-emitting plasma of the main lobe,we clearly see a thermal component with the best-
Kr´ol et al. fit temperature of kT ∼ . ∼ .
1, moderated by a hydrogen col-umn density ∼ . × cm − much in excessof the Galactic value. The power-law componentseen in the spectrum is characterized by a veryflat (although not well constrained) slope and alow amplitude. • In the West region, which does overlap with theedge of the Southern radio lobe, we see the samethermal component as in the East region ( kT ∼ . ∼ .
2, hydro-gen column density ∼ . × cm − ). How-ever, in addition we also detected a power-lawemission characterized by a relatively steep slope(Γ (cid:39) . ± .
28) and a high amplitude. • In the South region, which is located at larger dis-tances from the core and which corresponds to theSouthernmost edge of the radio counter-lobe, wedo not see any particularly pronounced thermalemission component, or a hydrogen column den-sity in excess of the Galactic value. The single ab-sorbed power-law component (though with hardlyconstrained absorbing column density, consistentwith zero) dominating the radiative output of theregion, is characterized by a relatively steep slope(Γ (cid:39) . ± . • In the case of the North region, which is also lo-cated at a larger distance from the core (whencompared to the West and East regions), andwhich overlaps with the edge of the main radiolobe, despite a very low photon statistics, an ac-ceptable fit could be obtained assuming a singlepower-law model, yielding a steep slope of the con-tinuum (Γ (cid:39) . ± . Chandra ’s PSF, and a significant instrumental pile-upaffecting predominantly low-energy segments of the tar-gets’ spectra. Hence, we believe that the low-amplitudeand flat-spectrum power-law component seen within theEast region, is due to this effect, i.e. represents only thevery broad PSF wings of the Cen A nucleus. As such,it should be also present in the West region, located ata comparable distance from the core. We have there-fore repeated the spectral fitting for the West regionadding an additional flat power-law component, withthe model parameters fixed at the values within 1 σ er-rors of the best-fit power-law component emerging fromthe spectral analysis of the East region. The remain-ing model parameters obtained in this way turned outbasically the same as the ones reported in Table 1, asexpected keeping in mind a much lower amplitude ofthe flat-spectrum power-law component to be comparedwith the steep-spectrum one.On the other hand, the steep-spectrum power-lawcomponent with the best-fit photon index Γ ∼ . (cid:39) . +0 . − . (model IV therein), is due to the syn-chrotron emission of very high-energy electrons ener-gized at the front of the bow shock, induced in the am-bient medium by the expanding radio counter-lobe.We also note that the field partly overlapping with ourWest region was also analyzed by Croston et al. (2009,“Region 3” therein); the best fit obtained by these au-thors assumed in this case a thermal model, and yieldeda relatively high plasma temperature of (cid:39) .
95 keV. Thereason for the discrepancy between our fitting resultsand those presented by Croston et al. for the West fea-ture, is the differences in the source and backgroundextraction regions, as well as in the fitting procedure;as a result, in our spectral modelling instead of a hotplasma we see a prominent power-law component in ad-dition to the relatively cold thermal gas emission. Webelieve that our estimates are however robust, becausealmost exactly same thermal gas parameters emerge forthe East region, where we do not see any physical power-law component (above the low-level emission related tothe PSF wings of the bright nucleus).The two issues should be clarified at this point re-garding the thermal component fitting. One is that theemerging low abundance values may not necessarily cor-respond to a real low metallicity, but instead may onlysignal an unresolved multi-temperature gas, for which oft X-ray Structures in Cen A Galaxy gamma A bundan c kT [keV] n H [ / c m ] gamma −2 −1 0 1 2 A bundan c kT [keV] n H [ / c m ] Figure 4.
Confidence contours of 1, 2, and 3 σ for the chi2datavar statistic on the two-thaw-parameters planes, for the Westregion (upper panels), and the East region (lower panel). the best-fit temperature of ∼ . χ values. Moreover, the best-fitline positions, ∼ ∼ .
85 keV, correspond tothe well-known neon/iron L-shell and silicon blends, re-spectively (see, e.g., Peterson, & Fabian 2006; B¨ohringer& Werner 2010). We note that previously, Evans etal. (2004) reported on the detection of the neutral sil-icon K α line in the nuclear spectrum of Cen A, withthe source extracted region corresponding effectively to (cid:46) DISCUSSIONOne of the two main results emerging from the anal-ysis presented in the previous section, is the detectionof the power-law emission component with a relativelysteep photon index Γ ∼ γ ∼ enabled by a balance between the accelera- Kr´ol et al. tion and radiative cooling timescales (see the discussionin Croston et al. 2009), but regards also the potentialfor the formation of a “shock-type” energy spectrum ofultra-relativistic electrons ∝ E − s with s ∼
2, assumingthe observed synchrotron X-ray photons are producedin the strong cooling regime, i.e. that s = 2(Γ − V (cid:39) .
43 kpc and (cid:39) .
23 kpc ,respectively, assuming the structures are cone-shaped,with the main axes on the plane of the sky. Assumingfurther that each region is filled uniformly with fully ion-ized hydrogen, we derive the gas density n , total mass M = m p nV , and the internal energy ε int = pV forthe corresponding thermal pressure p = nkT , basedon the normalization parameter in the xsapec model,10 − n V / πD cgs (see Table 1), where D = 3 .
85 Mpcis the distance to the source. For the West region, thederived parameters are n ∼ . − , M ∼ . × M (cid:12) , p = 1 . × − dyn cm − , and ε int ∼ × erg.For the East structures, we obtain n ∼ . − , M ∼ . × M (cid:12) , p = 1 . × − dyn cm − , and ε int ∼ × erg.Note that by assuming a clumpy distribution of theX-ray emitting gas in the analyzed regions, the de-rived gas density, and hence also the pressure, wouldincrease. Such an increase would however be problem-atic, since even with the filling factor of the order ofunity, the thermal gas present in the analyzed regionsappears much denser than, and over-pressured with re-spect to, the diffuse ISM at the corresponding distancefrom the nucleus; for this “unperturbed” ISM, followingKraft et al. (2003) hereafter we adopt n ISM ∼ .
01 cm − , kT ISM ∼ .
35 keV, and p ISM ∼ − dyn cm − . Atthe same time, the thermal gas we see in the Eastand West regions turns out to be in a pressure balancewith the non-thermal plasma present around the edgesof the lobes, for which Croston et al. (2009) estimated p shell ∼ − dyn cm − . This could suggest that whatwe see is simply a result of a shock compression of theISM by the expanding radio lobes. However, the corre-sponding high density contrast n/n ISM ∼
50, togetherwith the temperature ratio
T /T
ISM (cid:46)
1, impose a gen-eral problem for any interpretation involving adiabati-cally shocked ISM, in which case one would expect thedensity jump ∼ T /T
ISM (cid:29)
1. An unusual density increase along with a rapid tem-perature drop down to the pre-shock value, on theother hand, could possibly be encountered in a radiativeshock, i.e. when the gas cooling in the near downstreamis sufficiently fast that a relatively narrow radiative re-laxation layer is formed. For such, in the specific case ofan “isothermal shock” with the upstream plasma bulkvelocity u − , the gas temperature in the far downstreamsettles at T ∼ T ISM , and the density contrast reaches n ∼ n ISM (cid:18) u − c s (cid:19) ∼ . (cid:16) u − km s − (cid:17) cm − , (1)consistently with the gas temperature and density de-rived above for the analyzed East and West regions, aslong as u − ∼ km s − .However, the problem with this scenario is that theradiative cooling of the ISM gas immediately behind theshock (related to the free-free and line emission), is notsufficiently short. In fact, assuming a strong shock with u − (cid:29) c s , the compression ratio in the near downstreamshould be simply n + ≈ n ISM ∼ .
04 cm − , and the gastemperature kT + = 316 m p u − ∼ (cid:16) u − km s − (cid:17) keV . (2)For such, with u − ∼ km s − the thermal coolingtimescale τ cool = 52 nkT + n Λ ∼
300 Myr , (3)where Λ stands for the radiative cooling function (Pe-terson, & Fabian 2006), and we assumed approximatelyone-third solar abundance. This would be then ordersof magnitude longer than the dynamical timescale in-volved, τ dyn (cid:39) du − ∼ , (4)where d (cid:39) . oft X-ray Structures in Cen A Galaxy τ ei ∼ . m p ( kT e ) / e n √ m e (cid:46) .
01 Myr , (5)assuming electron temperature kT e ∼ . c s = (cid:112) kT ISM / m p ∼
230 km s − . In the framework ofthis scenario, the velocity of the ejection, u ej , could beestimated by measuring on the Chandra maps the half-opening angles of the East and West cones, θ , which givethe Mach numbers M = 1 / sin θ . For both analyzed re-gions we obtain roughly M (cid:39) −
5, i.e. the ejection ve-locity within the range u ej = M c s ∼ − ,
000 km s − .This leads to the elapse time since the ejection event τ ej ∼ d/u ej ∼ − numberdensity (cf. equation 3). Hence, the ejection episodepostulated here could indeed be considered as accompa-nying/coinciding with the onset of the currently ongoingjet activity in the system.Moreover, the derived ejection velocities are relativelyhigh, but on the other hand consistent with the veloc-ities of nuclear outflows detected in several AGN. Inthe particular case of Cen A, we note in this contextthat, based on the Suzaku observations of the Cen A nu-cleus, Tombesi et al. (2014) claimed the detection of theFe XXV He α and Fe XXVI Ly α absorption lines with theequivalent widths of the order of 10 eV, correspondingto the ionized hot absorber with the outflow velocities ≤ ,
500 km s − . What is more, in the recent Herschel data for the central 500 pc of Cen A, Israel et al. (2017)detected an outflow of cold, neutral and ionized gas,roughly along the axis of the radio jet, with a mass ofseveral million solar masses, and the projected velocityof 60 km s − . We do not necessarily identify those cur-rently observed nuclear or circum-nuclear outflows with the ejection episode postulated here to explain the Eastand West features at kpc distances from the core. Thepoint is, rather, that one can indeed expect formationof massive gaseous outflows and plasma ejections in thesystem, with high and very high velocities.A possible complication to the above-drafted sce-nario could be, however, again the very high pressurecontrast between the analyzed regions and the ISM.On the other hand, as the required propagation veloc-ity is most likely super-Alfvenic, for the Alfven speed v A = B ISM / (cid:112) πm p n ISM <
200 km s − with n ISM ∼ .
01 cm − and the anticipated B ISM < µ G, one mayspeculate that the magnetic draping effect, expected aslong as the coherence scale of the ISM magnetic fieldis large enough (see in this context Moss & Shukurov1996), effectively increases the total ISM pressure aheadof the ejection by the piled-up magnetic field (see Lyu-tikov 2006), so that the pressure balance is maintained. SUMMARY AND FINAL REMARKSIn this paper we re-analyze the archival
Chandra datafor the central parts of the Centaurus A radio galaxy,aiming for a systematic investigation of the X-ray emis-sion associated with the inner radio lobes, and theirimmediate surroundings. After inspection of the X-rayhardness maps of the system, we focus on four distinctfeatures characterized by the soft excess with respect tothe adjacent fields. Those include the two regions lo-cated at kpc distances from the nucleus to the West andEast, the extended bow-shock structure to the South,and a fragment of a thin arc North from the center. Theselected North, West, and South features coincide withthe edges of the radio lobes, while the East structure isseemingly displaced from the radio-emitting plasma.We perform the spectral analysis for the selected re-gions, assuming a combination of the absorbed power-law and thermal emission components. We found outthat for the North and South features, a simple power-law model consistent with no thermal contribution andno intrinsic absorption, provided satisfactory fits tothe data. The spectra of the East and West regions,on the other hand, could not be fitted at all witha one-component model, consisting of either an ab-sorbed single power-law emission, or an absorbed single-temperature plasma; for those, a two-component modelwas indeed required.One of the two main results emerging from our spec-tral analysis, is the detection of the power-law emissioncomponent with a relatively steep photon index Γ ∼ Kr´ol et al. lap with the side edges of the main radio lobe and thecounter-lobe, respectively. This emission component canbe naturally explained as representing the synchrotroncontinuum of very high-energy electrons energised at thefront of the lobes’ termination shock. Hence, we con-clude that the efficiency of the electron acceleration atthe termination shock front, does not vary dramaticallyover the inner lobes’ extension.The other main finding following from our spectralanalysis, is the presence of a relatively cold (tempera-ture (cid:39) . ∼ . − )gas within the two regions located at kpc distances tothe West and East from the nucleus, which appears over-pressured (by one order of magnitude) with respect tothe surrounding diffuse ISM. We argue that the scenarioin which this gas represents the ISM shocked by the ex-panding radio lobes, is not self-consistent, because ofthe required effectively “isothermal compression”. In-stead, we propose that the presence of such a cold and dense gas could possibly be related to a massive nuclearoutflow from the central regions of the galaxy.ACKNOWLEDGMENTSD.K., V.M., and (cid:32)L.S. were supported by Polish NSCgrant 2016/22/E/ST9/00061. The authors thank theanonymous referee for her/his critical comments andsuggestions, which helped to improve the paper substan-tially. The authors are also grateful to Arti Goyal, foruseful discussions on the low-frequency radio maps ofCen A. Facilities:
Chandra (ACIS)
Software:
CIAO (Fruscione et al. 2006), Sherpa(Freeman et al. 2001)REFERENCES
Abdalla, H., Abramowski, A., et al. 2018, A&A, 619, A71.Abdo, A. A., Ackermann, M., Ajello, M., et al. 2010a,Science, 328, 725Abdo, A. A., Ackermann, M., Ajello, M., et al. 2010b, ApJ,719, 1433.Aharonian, F., Akhperjanian, A. G., Anton, G., et al. 2009,ApJ, 695, L40.B¨ohringer, H. & Werner, N. 2010, A&A Rv, 18, 127Cappellari, M., Neumayer, N., Reunanen, J., et al. 2009,MNRAS, 394, 660Condon, J. J., Helou, G., Sanders, D. B., et al. 1996, ApJS,103, 81Crockett, R. M., Shabala, S. S., Kaviraj, S., et al. 2012,MNRAS, 421, 1603Croston, J. H., Kraft, R. P., Hardcastle, M. J., et al. 2009,MNRAS, 395, 1999.Evans, D. A., Kraft, R. P., Worrall, D. M., et al. 2004, ApJ,612, 786.Feain, I. J., Cornwell, T. J., Ekers, R. D., et al. 2011, ApJ,740, 17Ferrarese, L., Mould, J. R., Stetson, P. B., et al. 2007, ApJ,654, 186Freeman, P., Doe, S., & Siemiginowska, A. 2001,Proc. SPIE, 4477, 76Fruscione, A., McDowell, J. C., Allen, G. E., et al. 2006,Proc. SPIE, 6270, 62701VHardcastle, M. J., Kraft, R. P., Sivakoff, G. R., et al. 2007,ApJ, 670, L81. Hardcastle, M. J., Cheung, C. C., Feain, I. J., & Stawarz,(cid:32)L. 2009, MNRAS, 393, 1041Hardcastle, M. J., Lenc, E., Birkinshaw, M., et al. 2016,MNRAS, 455, 3526Israel, F. P., G¨usten, R., Meijerink, R., et al. 2017, A&A,599, A53Israel, F. P. 1998, A&A Rv, 8, 237Kalberla, P. M. W., Burton, W. B., Hartmann, D., et al.2005, A&A, 440, 775Karovska, M., Fabbiano, G., Nicastro, F., et al. 2002, ApJ,577, 114.Kataoka, J., Stawarz, (cid:32)L., Aharonian, F., et al. 2006, ApJ,641, 158.Kraft, R. P., V´azquez, S. E., Forman, W. R., et al. 2003,ApJ, 592, 129Kraft, R. P., Nulsen, P. E. J., Birkinshaw, M., et al. 2007,ApJ, 665, 1129Kraft, R. P., Forman, W. R., Hardcastle, M. J., et al. 2009,ApJ, 698, 2036Lyutikov, M. 2006, MNRAS, 373, 73Marconi, A., Pastorini, G., Pacini, F., et al. 2006, A&A,448, 921McKinley, B., Briggs, F., Gaensler, B. M., et al. 2013,MNRAS, 436, 1286Mingo, B., Hardcastle, M. J., Croston, J. H., et al. 2011,ApJ, 731, 21Morganti, R., Killeen, N. E. B., Ekers, R. D., et al. 1999,MNRAS, 307, 750Morganti, R. 2010, PASA, 27, 463 oft X-ray Structures in Cen A Galaxy Moss, D., & Shukurov, A. 1996, MNRAS, 279, 229Neff, S. G., Eilek, J. A., & Owen, F. N. 2015, ApJ, 802, 88Neumayer, N., Cappellari, M., Reunanen, J., et al. 2007,ApJ, 671, 1329Oosterloo, T. A., & Morganti, R. 2005, A&A, 429, 469Peterson, J. R., & Fabian, A. C. 2006, PhR, 427, 1Rejkuba, M. 2004, A&A, 413, 903Sahakyan, N., Yang, R., Aharonian, F. A., et al. 2013, ApJ,770, L6.Salom´e, Q., Salom´e, P., Combes, F., et al. 2016, A&A, 586,A45Sanchez, D., Holler, M., Taylor, A. et al. for the HESSCollaboration 2018, talk during TeVPA2018 (Berlin)Snios, B., Wykes, S., Nulsen, P. E. J., et al. 2019, ApJ, 871,248.Stawarz, (cid:32)L., Szostek, A., Cheung, C. C., et al. 2014, ApJ,794, 164 Stawarz, (cid:32)L., Tanaka, Y. T., Madejski, G., et al. 2013, ApJ,766, 48Steinle, H., Bennett, K., Bloemen, H., et al. 1998, A&A,330, 97.Sun, X.-n., Yang, R.-z., Mckinley, B., & Aharonian, F.2016, A&A, 595, A29Tanada, K., Kataoka, J., & Inoue, Y. 2019, ApJ, 878, 139Tombesi, F., Tazaki, F., Mushotzky, R. F., et al. 2014,MNRAS, 443, 2154Quillen, A. C., Brookes, M. H., Keene, J., et al. 2006, ApJ,645, 1092Werner, N., de Plaa, J., Kaastra, J. S., et al. 2006, A&A,449, 475 Kr´ol et al.
APPENDIX A. EMISSION LINES IN THE WEST AND THE EAST REGIONSResiduals are clearly seen in the best-fit models xsphabs*(xsapec+xspowerlaw) presented in Figure 3 for the Eastand West regions, indicating the presence of line-like emission features. We have investigated this issue by introducingfirst an additional gaussian component to the model, xsphabs*(xsapec+xspowerlaw + xsgaussian) , and next al-lowing for two different gaussian features, xsphabs*(xsapec+xspowerlaw + xsgaussian1+ xsgaussian2) , each timewith the source frame line widths frozen at σ = 10 eV, and plasma abundances frozen at the best-fit values emergingfrom the basic xsphabs*(xsapec+xspowerlaw) fits for the two regions (see Table 1). The results of the modellingare presented in Figure 5, and summarized in Table 2. As shown, by introducing two additional gaussian spectralcomponents, one can indeed improve the quality of the fitting in terms of the reduced χ values, with no significantchange in the best-fit values of the other model parameters. The best-fit positions of the first line-like feature readsas ∼ α line, or the iron L-shell blend (the position of which dependshowever on the plasma temperature; see B¨ohringer & Werner 2010); the best-fit position of the other line, ∼ .
85 keV,allows for the identification with the Si XIII blend (e.g., Peterson, & Fabian 2006). C oun t s / s e c / k e V Energy (keV) S i g m a −4−2024 C oun t s / s e c / k e V Energy (keV) S i g m a −4−2024 C oun t s / s e c / k e V Energy (keV) S i g m a −202 C oun t s / s e c / k e V Energy (keV) S i g m a −4−2024 Figure 5.
Chandra spectra along with the best-fit models (and residuals) for the East region (left column) and Westregion (right column). The models displayed consist of a mixture of absorbed thermal and non-thermal components, xsphabs*(xsapec+xspowerlaw) , with the addition of either one or two gaussian features xsgaussian (upper and lower rows,respectively). oft X-ray Structures in Cen A Galaxy Table 2.
Spectral fitting results with additional gaussian components for the East and West regions.
Region/Model Parameter Value 1 σ errors Units East kT 0.22 0.01 keV xsphabs*(xsapec+xspowerlaw norm 0 . . − × apec+ xsgaussian) Γ 0.11 0.50 —ampl 1 . . − × ph/keV/cm at 1keVLine position 1.85 0.01 keVLine normalization 0.93 0.02 10 − × ph/cm /s N H cm − Final fit statistic 124.08Degrees of freedom 91
West kT 0.19 0.01 keV xsphabs*(xsapec+xspowerlaw norm 2 . . − × apec+ xsgaussian) Γ 1.4 0.34 —ampl 22 . . − × ph/keV/cm at 1keVLine position 1.83 0.02 keVLine normalization 1.4 0.3 10 − × ph/cm /s N H cm − Final fit statistic 190.5Degrees of freedom 127
East kT 0.22 0.01 keV xsphabs*(xsapec+xspowerlaw norm 0 . . − × apec+ xsgaussian1+ xsgaussian2) Γ 0.25 0.74 —ampl 2 . .
65 10 − × ph/keV/cm at 1keVLine 1 position 1.0 — keVLine 1 normalization 7.6 1.6 10 − × ph/cm /sLine 2 position 1.83 0.03 keVLine 2 normalization 0.42 0.21 10 − × ph/cm /s N H cm − Final fit statistic 99.55Degrees of freedom 89
West kT 0.19 0.01 keV xsphabs*(xsapec+xspowerlaw norm 0 . . − × apec+ xsgaussian1+ xsgaussian2) Γ 1.75 0.32 —ampl 30 . . − × ph/keV/cm at 1keVLine 1 position 1.02 — keVLine 1 normalization 17.1 3.4 10 − × ph/cm /sLine 2 position 1.83 0.02 keVLine 2 normalization 1.2 0.3 10 − × ph/cm /s N H cm −2