1: %\documentclass[12pt, preprint]{aastex}
2: \documentclass{emulateapj}
3: \usepackage{times}
4: \newcommand{\etal}{{et al}\/.}
5: \newcommand{\hh}{^{\rm h}}
6: \newcommand{\mm}{^{\rm m}}
7: \begin{document}
8: \slugcomment{Draft of \today}
9: \shorttitle{AGN feedback in NGC\,6764}
10: \shortauthors{J.H. Croston \etal}
11: \title{{\it Chandra} evidence for AGN feedback in the spiral galaxy NGC\,6764}
12: \author{J.H. Croston\altaffilmark{1}, M.J. Hardcastle\altaffilmark{1},
13: P. Kharb\altaffilmark{2}, R.P. Kraft\altaffilmark{3}, A. Hota\altaffilmark{4}}
14: \altaffiltext{1}{School of Physics, Astronomy and Mathematics, University of
15: Hertfordshire, College Lane, Hatfield AL10 9AB, UK}
16: \altaffiltext{2}{Department of Physics, Purdue University, 525 Northwestern Avenue, West Lafayette, IN 47907, USA}
17: \altaffiltext{3}{Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, MA~02138, USA}
18: \altaffiltext{4}{Institute of Astronomy and Astrophysics, Academia Sinica, P.O. Box 23-141, Taipei 10617, Taiwan, R.O.C.}
19:
20: \begin{abstract}
21: We report the {\it Chandra} detection of X-ray emission spatially
22: coincident with the kpc-scale radio bubbles in the nearby ($D_{L}
23: \sim 31$ Mpc) AGN-starburst galaxy NGC\,6764. The X-ray emission
24: originates in hot gas ($kT \sim 0.75$ keV), which may either be
25: contained within the radio bubbles, or in a shell of hot gas
26: surrounding them. We consider three models for the origin of the hot
27: gas: (1) a starburst-driven galactic wind, (2) shocked gas associated
28: with the expanding radio bubbles, and (3) gas heated and entrained
29: into the bubbles by jet/ISM interactions in the inner AGN outflow. We
30: rule out a galactic wind based on significant differences from known
31: galactic wind systems. The tight correspondence between the brightest
32: X-ray emission and the radio emission in the inner outflow from the
33: Seyfert nucleus, as well as a correlation between X-ray and radio
34: spectral features suggestive of shocks and particle acceleration,
35: lead us to favour the third model; however, we cannot firmly rule out
36: a model in which the bubbles are driving large-scale shocks into the
37: galaxy ISM. In either AGN-driven heating scenario, the total energy
38: stored in the hot gas is high, $\sim 10^{56}$ ergs, comparable to the
39: energetic impact of low-power radio galaxies such as Centaurus A, and
40: will have a dramatic impact on the galaxy and its surroundings.
41:
42: \end{abstract}
43: \keywords{galaxies: active -- X-rays: galaxies}
44:
45: \maketitle
46:
47: \section{Introduction}
48: \label{intro}
49:
50: Galaxy feedback processes are now thought to be an important
51: ingredient in galaxy formation models (e.g. Croton et al. 2006; Bower
52: et al. 2006), potentially solving the long-standing problem of
53: explaining the deficit of galaxies at the low and high mass ends of
54: the galaxy luminosity function relative to model predictions.
55: Disentangling the feedback contributions from AGN outbursts and from
56: star formation to the energetics of gas in galaxies, galaxy groups and
57: clusters is a key problem in galaxy evolution, as it is clear that
58: both can inject large quantities of energy into their surroundings.
59: While there is growing evidence that AGN outbursts are the dominant
60: feedback process operating at the high-mass end of the galaxy
61: luminosity function, it might be expected that at lower masses, and
62: particularly in late-type galaxies, the energy input from star
63: formation processes will dominate.
64:
65: Recent work has shown that kpc-scale radio bubbles connected to an
66: active nucleus can be found in many Seyfert galaxies (e.g. Gallimore
67: et al. 2006; Kharb et al. 2006). Although it has been argued that
68: these bubbles may be powered by stellar winds (e.g. Baum et al. 1993),
69: the fact that no convincing examples exist in galaxies without an AGN
70: strongly suggests that they are inflated by the active nucleus (e.g.
71: Hota \& Saikia 2006, hereafter HS06). Many of the known Seyfert radio
72: bubbles are predicted to be overpressured with respect to their
73: surroundings (e.g. Capetti et al. 1999), and therefore may be shocking
74: the surrounding medium, as has been observed for kpc-scale radio lobes
75: in massive elliptical galaxies (e.g. Centaurus A: Kraft et al. 2003;
76: NGC\,3801: Croston et al. 2007).
77:
78: In order to search for evidence for galaxy feedback associated with
79: kpc-scale radio bubbles in spiral galaxies, we carried out a {\it
80: Chandra} observation of the nearby Seyfert 2/LINER galaxy NGC\,6764 ,
81: which is a barred spiral galaxy (type SB(s)bc) with non-thermal radio
82: bubbles perpendicular to the plane of the galaxy disk that extend for $\sim
83: 1$ kpc (Fig.~\ref{ropt}). Here we report on the results of this
84: observation.
85:
86: Throughout this paper, we adopt a cosmology with $H_{0} = 70$ km
87: s$^{-1}$ Mpc$^{-1}$, $\Omega_{M} = 0.3$ and $\Omega_{\Lambda} = 0.7$.
88: We adopt a luminosity distance for NGC\,6764 of 31.3 Mpc, obtained by
89: correcting the heliocentric velocity of 2146 km s$^{-1}$ to the CMB
90: frame of reference. This gives an angular scale of $1^{\prime\prime} =
91: 0.15$ kpc at the distance of NGC\,6764.
92:
93: \section{Data analysis}
94:
95: We observed NGC\,6764 with {\it Chandra} (ACIS-S) for 20 ks on 2008
96: January 20th. The observation was taken in VFAINT mode to minimize
97: the background level. The data were reprocessed from the level 1
98: events file with {\sc ciao} 3.4 and CALDB 3.4.2, including VFAINT
99: cleaning. The latest gain files were applied and the 0.5-pixel
100: randomization removed using standard techniques detailed in the {\sc
101: ciao} on-line documentation\footnote{http://asc.harvard.edu/ciao/}. An
102: inspection of the lightcurve for our observation using the {\it
103: analyze\_ltcrv} script showed that there were no periods of high
104: background level, and so no additional GTI filtering was applied.
105:
106: We produced a 0.5 -- 5 keV filtered image to examine the X-ray
107: emission associated with the galaxy, presented in Fig.~\ref{data}. The
108: most prominent feature in the image is the region of extended
109: emission, which is located exactly coincident with the bubbles of
110: radio emission shown in the right hand panel. In addition, there is a
111: region of brighter X-ray emission in the centre of the emitting region
112: on scales of $\sim 4^{\prime\prime}$, coincident with the AGN and
113: extending to the west, while becoming broader. This bright region
114: coincides with a region of higher surface brightness radio emission,
115: as shown in the higher resolution 5-GHz radio map in Fig.~11 of HS06.
116:
117: We extracted spectra from the Seyfert nucleus, the bright central
118: region, the two lobes (both separately and together to improve the
119: signal-to-noise ratio) and from a large region encompassing the galaxy
120: disk, using the {\it specextract} script, which also builds the
121: appropriate response files. Local background regions adjacent to or
122: surrounding the source extraction regions were used. Spectra were
123: grouped to 20 counts per bin after background subtraction prior to
124: spectral fitting, which was carried out using {\sc xspec}. We assumed
125: a fixed Galactic absorption of $N_H = 6.0 \times 10^{20}$ cm$^{-2}$ in
126: our spectral fitting (Dickey \& Lockman 1990), except where otherwise
127: specified.
128: \begin{figure}[!t]
129: \begin{center}
130: \plotone{f1.eps}
131: \caption{1.4-GHz radio contours of NGC\,6764 overlaid on a DSS-2
132: optical image of the host galaxy. The radio map is made from VLA
133: archival data as published in Hota \& Saikia (2006). Contour levels
134: are $(1.5 \times 10^{-4}) \times 1, 2, 4,...,512$ Jy beam$^{-1}$.}
135: \label{ropt}
136: \end{center}
137: \end{figure}
138:
139: \begin{figure*}
140: \begin{center}
141: \caption{Left: 0.5 - 5.0 keV image of the {\it Chandra} data, smoothed
142: with a Gaussian of 0.6$^{\prime\prime}$ FWHM, showing emission
143: associated with the radio bubbles; right: X-ray emission in colour
144: smoothed with a Gaussian of 1.8$^{\prime\prime}$ FWHM, with radio
145: contours overlaid. Contour levels are as for Fig.~\ref{ropt}.}
146: \label{data}
147: \vskip 10pt
148: \plottwo{f2a.eps}{f2b.eps}
149: \end{center}
150: \end{figure*}
151:
152:
153: \section{Results}
154:
155:
156: \subsection{Radio-related X-ray emission}
157:
158: We measure a total of $\sim 1150$ net 0.5 -- 5 keV ACIS-S counts from
159: the regions coincident with the radio bubbles (excluding the bright
160: central region). In what follows we will refer to these X-ray emitting
161: regions as ``X-ray bubbles'' for simplicity but without any implied
162: geometry for the emitting fluid. We fitted a series of models to the
163: bubble spectra to distinguish between different origins for the X-ray
164: emission. We initially considered the spectrum for the entirety of the
165: bubble regions, using a local background region (source and background
166: extraction regions are shown in Fig.~\ref{data}). We fitted both
167: single power law and single {\it mekal} model (with Galactic
168: absorption, as described above, and free abundance for the thermal
169: model). The power-law model gave an unacceptable fitting statistic
170: ($\chi^{2} = 181$ for 25 d.o.f.), and allowing the value of Galactic
171: column density to vary as a proxy for any absorption intrinsic to the
172: galaxy did not allow an acceptable fit to be obtained. We obtained a
173: good fit with the {\it mekal} model, with best-fitting values of $kT =
174: 0.75\pm0.05$ keV and $Z = 0.13^{+0.05}_{-0.03}$ Z$_{\sun}$ and
175: $\chi^{2} = 25.5$ for 24 d.o.f. This model fit is shown in
176: Fig.~\ref{spec}. Allowing the total absorbing column to vary to
177: account for any intrinsic absorption does not significantly change the
178: best-fitting temperature or abundance values (although the total
179: absorbing column density was left free, it did not fall below the
180: assumed Galactic column). We note that the low abundance values may be
181: due to fitting a single-temperature model to a multi-temperature
182: plasma (e.g. Buote et al. 2000), and so are likely to be unreliable.
183: Given the very poor fitting statistics obtained for a single power-law
184: model, we can firmly rule out a non-thermal origin for the X-ray
185: emission associated with the radio bubbles. We added a power-law
186: component to the spectrum to investigate whether some contribution
187: from non-thermal emission is present; however, this did not
188: significantly affect the fit statistic or the parameters of the {\it
189: mekal} component. While we cannot rule out some contribution from
190: non-thermal emission, the X-ray emission predominantly originates from
191: gas of $kT \sim 0.75$ keV. We measure a total, unabsorbed 0.5 -- 2.0
192: keV flux from the X-ray bubbles of $1.2 \times 10^{-13}$ erg cm$^{-2}$
193: s$^{-1}$, corresponding to a luminosity of $\sim 1.4 \times 10^{40}$
194: erg s$^{-1}$.
195:
196: To investigate whether there is any spatial variation in the
197: properties of the X-ray-emitting gas, we examined spectra of the
198: regions corresponding to the northern and southern bubbles separately.
199: The properties of the two bubbles appear to be fairly similar both to
200: each other and to the results for the joint spectrum discussed above.
201: In neither case is a power-law model an adequate fit (either with
202: fixed or free $N_{{\rm H}}$), and in both cases good fits could be
203: obtained with a single {\it mekal} model. For the northern bubble we
204: obtained a best-fitting temperature of $kT = 0.81^{+0.08}_{-0.07}$ keV
205: and $Z = 0.14^{+0.09}_{-0.05}$ Z$_{\sun}$ with $\chi^{2} = 10.3$ for 9
206: d.o.f., and for the southern bubble we obtained a best-fitting
207: temperature of $kT = 0.64^{+0.06}_{-0.05}$ keV and $Z =
208: 0.12^{+0.06}_{-0.04}$ Z$_{\sun}$ with $\chi^{2} = 15.9$ for 13 d.o.f.
209: Therefore there is a small, but statistically significant difference
210: in the temperatures of the two bubbles, with the southern bubble
211: slightly cooler than the northern one. The origin of this small
212: difference in unclear, but could be related to differences in the
213: distribution of material being heated in the north and south.
214:
215: Finally, we also examined the spectrum of the brighter central region
216: of X-ray emission (Fig.~\ref{inner}; $2 - 4^{\prime\prime}$ scales) in
217: the same way. We found that for this emission neither a {\it mekal}
218: nor a power-law model gave an acceptable fit if Galactic absorption
219: was assumed; however, an acceptable fit could be obtained for a {\it
220: mekal} model if the value of $N_{\rm H}$ was permitted to vary (again
221: we varied the overall column density, but it did not drop below the
222: assumed Galactic value). We obtained a good fit with $N_{H} =
223: (2.0^{+1.5}_{-0.7}) \times 10^{21}$ cm$^{-2}$, $kT =
224: 0.93^{+0.12}_{-0.18}$ keV, and abundance fixed at the value measured
225: for the outer regions, with $\chi^{2} = 16.7$ for 12 d.o.f. If the
226: abundance is fixed at solar abundance, we find a best fit with $N_{\rm
227: H} = (7.7^{+1.3}_{-1.0}) \times 10^{21}$ cm$^{-2}$ and $kT =
228: 0.65^{+0.22}_{-0.11}$ keV, with $\chi^{2} = 18.3$ for 12 d.o.f. A
229: power-law model with free $N_{\rm H}$ did not give an acceptable fit,
230: and so we conclude that the brighter central X-ray emission also has a
231: thermal origin. The temperature of this gas is consistent with that in
232: the regions associated with the radio bubbles.
233:
234: \begin{figure}
235: \begin{center}
236: \caption{ACIS-S counts spectrum for the entire bubble region in the
237: energy range 0.4 -- 7.0 keV, with best-fitting {\it mekal} model, as
238: described in the text, overplotted.}
239: \label{spec}
240: \plotone{f3.eps}
241: \end{center}
242: \end{figure}
243:
244: \subsection{The Seyfert nucleus}
245: \label{sey}
246: We extracted a spectrum from a 1.5$^{\prime\prime}$ radius circle
247: centred on the Seyfert nucleus, which we assumed to be located at the
248: peak of the X-ray emission. This position is in agreement with the
249: position of the Seyfert nucleus determined from optical observations
250: (Clements 1981) to within the accuracy of {\it Chandra}'s astrometry.
251: We found that single power-law models were unacceptable whether the
252: overall absorption was fixed at the Galactic value or left as a free
253: parameter as a proxy for a contribution from intrinsic absorption. A
254: single {\it mekal} model was also unacceptable with either Galactic or
255: free absorption. We therefore fitted a model consisting of a {\it
256: mekal} component plus an intrinsically absorbed power law. Since the
257: number of spectral bins is fairly low, the power-law index was fixed
258: at $\Gamma = 1.5$, a typical value for nuclear emission (a steeper
259: value, e.g. $\Gamma = 2.0$, does not significantly alter the results).
260: We obtained a good fit for this model, with $\chi^{2} = 10.1$ for 7
261: d.o.f., $kT = 0.92^{+0.16}_{-0.05}$ keV, and $N_{H} =
262: (2.9^{+3.6}_{-1.8}) \times 10^{21}$ cm$^{-2}$. For this best-fitting
263: model, we obtained an unabsorbed 2.0 -- 10.0 keV flux for the
264: power-law component of $(3.4^{+1.0}_{-1.2}) \times 10^{-14}$ ergs
265: cm$^{-2}$ s$^{-1}$, which corresponds to a 1-keV flux density of
266: $4.0^{+1.1}_{-1.4}$ nJy and implies a 2.0 - 10.0 keV unabsorbed
267: nuclear luminosity of $(4.0^{+1.2}_{-1.4}) \times 10^{39}$ ergs
268: s$^{-1}$. The uncertainities here take account of the large
269: uncertainty in the absorbing column. The unabsorbed 0.5 - 2.0 keV flux
270: of the thermal component is $(3.0^{+0.6}_{-0.5}) \times 10^{-14}$ ergs
271: cm$^{-2}$ s$^{-1}$, corresponding to a luminosity of
272: $(3.5^{+0.8}_{-0.6}) \times 10^{39}$ ergs s$^{-1}$. The total
273: unabsorbed nuclear 2.0 -- 10.0 keV luminosity is $\sim 4.4 \times
274: 10^{39}$ ergs s$^{-1}$.
275:
276: It is common for type 2 Seyfert nuclei to have strong intrinsic
277: absorption, artificially lowering the inferred X-ray luminosity [e.g.
278: Awaki et al. (2000), Terashima \& Wilson (2001), Matt et al. (2004),
279: Matsumoto et al. (2004)]. A more conservative upper limit can be
280: obtained by assuming a typical value of the intrinsic absorption for
281: Seyfert 2s where detailed X-ray analysis is available. We therefore
282: also fitted a model with a {\it mekal} plus heavily absorbed power
283: law, with $N_{H} = 3 \times 10^{23}$ cm$^{-2}$ [the value measured for
284: the Seyfert 2 M51 by Terashima \& Wilson (2001)] in order to determine
285: a conservative upper limit on the intrinsic nuclear luminosity. Fixing
286: the power-law normalisation at the upper limit allowed by the
287: best-fitting model, we found that the data are consistent with a
288: heavily obscured power-law component with a 2.0-10.0 keV luminosity as
289: high as $\sim 2 \times 10^{41}$ erg s$^{-1}$. Hence we cannot rule out
290: the presence of an intrinsically bright nucleus that is heavily
291: obscured.
292:
293: \section{Energetics of the X-ray-emitting gas and radio bubbles}
294: \label{energetics}
295: As shown in Fig.~\ref{data}, the X-ray emission appears to trace the
296: 1.4-GHz radio structure in NGC\,6764 very closely. There is evidence
297: for a surface-brightness deficit in both the X-ray and radio emission
298: in the centre of the northern bubble, and to a lesser extent in the
299: centre of the southern bubble; however, the X-ray structure is not
300: strongly edge-brightened as seen in shocked shells surrounding the
301: kpc-scale lobes of radio galaxies (e.g. Kraft et al. 2003; Croston et
302: al. 2007). We cannot distinguish between a model where the hot gas is
303: inside the radio bubbles and one where it is in shell structures
304: surrounding them based on the X-ray morphology alone.
305:
306: To investigate the energetics of the bubbles we therefore considered
307: two possible scenarios: (1) the X-ray-emitting gas fills the radio
308: bubbles, (2) the X-ray-emitting gas is located in shells surrounding
309: the radio bubbles. Based on the outer extent of the X-ray emission
310: from the southern bubble (which has a more spherical appearance), we
311: assumed an outer radius for the emitting region of
312: $4.2^{\prime\prime}$, which corresponds to 0.7 kpc. For scenario (1)
313: we assumed a sphere of this radius, and for scenario (2) we assumed a
314: spherical shell with thickness 0.1 kpc. Assuming a constant density
315: and a filling factor of unity in the emitting region, these geometries
316: lead to inferred electron densities of 0.19 cm$^{-3}$ and 0.30
317: cm$^{-3}$, for scenarios 1 and 2, respectively. For the measured
318: temperature of the southern bubble X-ray emission, these correspond to
319: pressures of $(4 - 7) \times 10^{-12}$ dyne cm$^{-2}$. For the
320: northern bubble, we assumed an outer radius of $4.0^{\prime\prime}$,
321: and the same shell thickness for scenario 2, which led to inferred
322: electron densities of 0.17 cm$^{-3}$ and 0.27 cm$^{-3}$ for scenarios
323: 1 and 2, respectively, and to pressures consistent with those of the
324: southern bubble. For lower filling factors, the gas densities (and
325: pressures as inferred below) will increase. The total gas mass is
326: $\sim 10^{7}$ M$_{\sun}$, a small fraction of the ISM mass likely to
327: be contained in the galaxy's bulge.
328:
329: \begin{figure}
330: \caption{Surface brightness profiles for the northern (l) and southern
331: (r) bubbles obtained in a slice towards the bubble edge. Centre of
332: the profile is the bubble centre in each case (not the galaxy
333: nucleus).}
334: \label{profs}
335: \plottwo{f4a.eps}{f4b.eps}
336: \end{figure}
337: \begin{figure*}
338: %\begin{center}
339: \caption{Left: Close-up of the X-ray emission from the inner
340: jet/outflow region from the 0.5 - 5.0 keV image smoothed with a
341: Gaussian of 0.6$^{\prime\prime}$ FWHM, with 5-GHz contours overlaid
342: from a map made from archival data as published in Hota \& Saikia
343: (2006). Contour levels are $(8.0 \times 10^{-5}) \times 1, 2, 4,
344: ..., 512$ Jy beam$^{-1}$; right: hardness ratio map, as described in
345: the text. White is hard and red is soft. There is a statistically
346: significant increase in hardness to the western edge associated with
347: a region of flatter spectrum radio emission.}
348: \label{inner}
349: \vskip 10pt
350: \plotone{f5.eps}
351: %\end{center}
352: \end{figure*}
353:
354: We determined the equipartition internal pressures of the two bubbles
355: using measurements of the 1.4-GHz flux density for each lobe to
356: normalize the synchrotron spectrum. We assumed a filling factor of
357: unity, a broken power-law electron distribution with initial electron
358: energy index, $\delta$, of 2.1, $\gamma_{{\rm min}} = 10$ and
359: $\gamma_{{\rm max}} = 10^{5}$, and a break at $\gamma_{{\rm break}} =
360: 10^{3}$, and modelled the lobes as spheres of radius 4.0 and
361: 4.2$^{\prime\prime}$, respectively, for the northern and southern
362: bubbles. The resulting equipartition internal pressures were $\sim 1.2
363: \times 10^{-13}$ dyne cm$^{-2}$ for both bubbles, which are a factor
364: of $\sim 30 - 50$ times lower than the pressure in the hot gas. Hence
365: if the hot gas is external to the radio bubbles, an additional source
366: of pressure in the bubbles is required to balance that in the hot gas.
367: However, it is known that external pressures from hot gas surrounding
368: most low-power radio-galaxy lobes are significantly higher than
369: internal equipartition pressures (e.g. Morganti et al. 1988; Croston
370: et al. 2003, 2008), and so this is not on its own a strong argument
371: against the gas being outside the radio bubbles. In particular, the
372: external pressures of the shocked shells in NGC\,3801 and Centaurus A
373: (Croston et al. 2007; Kraft et al. 2003) are factors of $\sim 70$ and
374: $\sim 15$ times higher, respectively, than the internal equipartition
375: radio-lobe pressures.
376:
377: Based on the two scenarios for the geometry of the X-ray-emitting
378: material discussed above, we determined the total thermal energy
379: stored in the hot gas to be $(3 - 5) \times 10^{55}$ ergs, which is
380: comparable to the energy stored in the shocked shells surrounding the
381: lobes of the radio galaxy NGC\,3801 ($\sim 8 \times 10^{55}$ ergs).
382:
383: We also determined the physical properties of the central region of
384: brighter, thermal X-ray emission. As shown in Fig.~\ref{inner}, this
385: region corresponds closely to a brighter region of radio emission. The
386: narrowest part of the brighter region coincides with the position of
387: the Seyfert nucleus, leading HS06 to interpret this emission as a
388: radio jet oriented towards the western edge of the large-scale radio
389: bubbles. The X-ray and radio emission from this brighter region
390: corresponds very closely, and it has the appearance of a poorly
391: collimated outflow. If this is indeed an outflow, then it is likely to
392: be supplying material to the large-scale bubbles. We therefore also
393: investigated the energetics of this inner region. Assuming a spherical
394: region of radius 1.85$^{\prime\prime}$, we find an electron density
395: between 0.76 -- 0.80 cm$^{-3}$ corresponding to a gas pressure of
396: $(8.3 - 11) \times 10^{-12}$ dyne cm$^{-2}$ (the ranges given here
397: take into account a range in metallicity from [0.15 -- 1]Z$_{\sun}$).
398: Hence the inner gas is overpressured by as much as a factor two
399: relative to the outer X-ray bubbles, and should be expanding and
400: possibly heating the outer gas through weak shocks. It is also
401: possible that the radio bubble is providing pressure support for this
402: higher pressure gas, or that the gas is being compressed by the
403: expanding radio bubble. We discuss the implications of this expanding
404: hot gas in the centre of the galaxy, and its relation to the larger
405: X-ray bubbles in the sections that follow.
406:
407: \section{Origin of the X-ray emission: galactic wind or jet-driven shocks?}
408:
409: The two most likely explanations for the origin of the X-ray emission
410: in NGC\,6764 are: (I) hot gas in a galactic wind, similar to the X-ray
411: emission seen in typical nearby starburst galaxies, or (II) emission
412: from material that has been heated by the expanding radio plasma.
413: However, there are some clear problems with both scenarios, as neither
414: can fully explain the very tight correspondence between radio and
415: X-ray morphology. A third possibility, suggested by this
416: correspondence, is that both the radio and X-ray emission are produced
417: by the same particles. However, the steep radio spectral index of the
418: bubbles is not consistent with a thermal origin for the radio
419: emission, and the X-ray spectra of the bubbles are clearly
420: inconsistent with a non-thermal origin. The observed flux level is
421: also much higher than would be predicted by an inverse-Compton model,
422: so that this model can be firmly ruled out. Therefore, we can find no
423: obvious explanation in which the X-ray emission is produced by the
424: same particles as the radio emission. Below we consider possibilities
425: (I) and (II) in more detail. We consider two scenarios for Model II: a
426: model in which the expanding bubbles are driving large-scale shocks
427: into the ISM (Model IIa), and one in which jet/ISM interactions in the
428: inner parts of the outflow are heating the ISM and causing it to be
429: entrained into the bubbles (Model IIb).
430:
431: \subsection{Model I: hot gas from a galactic wind}
432:
433: As NGC\,6764 is a galaxy with known starburst activity, the presence
434: of a galactic wind (e.g. Veilleux et al. 2005) producing the observed
435: X-ray emission must be considered. The morphology of the X-ray
436: emission from NGC\,6764 is similar to some of the known examples of
437: X-ray-detected galactic winds in nearby galaxies (e.g. Strickland et
438: al. 2004b). For example, the X-ray structures in the starburst
439: galaxies M82 and NGC\,1482 have similar opening angles to the bubbles
440: in NGC\,6764. In both of these cases the X-ray-emitting gas is
441: coincident with H$\alpha$ filaments extending out of the galaxy disk,
442: which is also the case for NGC\,6764 (see Fig. 12 of HS06). However,
443: radio bubbles coincident with the X-ray-emitting gas are not present
444: in either M82 or NGC\,1482. We note also that the X-ray morphology of
445: these two systems is not consistent with a uniformly filled sphere or
446: truncated cone, as it is slightly brighter towards the edges rather
447: than in the centre.
448:
449: NGC\,6764 shows strong evidence for circumnuclear starburst activity
450: (e.g. Eckart et al. 1991, 1996; Schinnerer et al. 2000). There is also
451: evidence that the nuclear starburst is producing an outflow: e.g. Leon
452: et al. (2007) estimated, based on CO observations, that $\sim 4 \times
453: 10^{6}$ M$_{\sun}$ of molecular gas is outflowing at speeds of $\sim
454: 25$ km s$^{-1}$. However, Eckart et al. (1991) estimated a total star formation
455: rate of $\sim 4$ M$_{\sun}$ yr$^{-1}$, with a supernova rate of $\sim
456: 1.2 \times 10^{-2}$ yr$^{-1}$, which is significantly lower than the
457: SFR of $13 - 33$ M$_{\sun}$ yr$^{-1}$ for M82 estimated by F\"{o}rster
458: Schreiber et al. (2003). The kinetic energies of the outflowing
459: material detected in HI absorption (HS06) and CO emission (Leon et al.
460: 2007) are $8.5 \times 10^{52}$ ergs and $2.4 \times 10^{53}$ ergs,
461: respectively, which are both several orders of magnitude below the
462: thermal energy of the X-ray emitting gas. Leon et al. estimated the
463: typical energies available from supernovae and stellar winds given its
464: inferred star formation rate, finding values in the range $10^{54} -
465: 10^{55}$ ergs. It is therefore unclear that sufficient energy is
466: available from starburst activity to power the observed X-ray
467: emission.
468:
469: The median 0.3 -- 2.0 keV X-ray luminosity for galactic winds for the
470: starburst galaxies in the sample of Strickland et al. (2004a, 2004b)
471: is $2.5 \times 10^{39}$ erg s$^{-1}$, which is an order of magnitude
472: lower than the X-ray luminosity of the emission from NGC\,6764 in the
473: same energy range. The highest luminosity wind in their sample has a
474: 0.3 -- 2.0 keV X-ray luminosity of $L_{X} = 8.3 \times 10^{39}$ erg
475: s$^{-1}$, a factor of 2 lower than that of NGC\,6764. While NGC\,6764
476: is therefore more luminous than other known galactic winds, its
477: luminosity is not so extreme as to preclude a similar origin. However,
478: the X-ray emission from the bubbles of NGC\,6764 (excluding the bright
479: central region) originates from a smaller volume (e.g. the height in
480: kpc above the disk to which the emission extends is a factor of $\sim
481: 3$ times smaller in NGC\,6764 than in M82), hence the volume
482: emissivity is considerably higher.
483:
484: The diffuse $H\alpha$ emission from NGC\,6764 has a total luminosity
485: of $(6 - 8) \times 10^{40}$ erg s$^{-1}$ (Zurita et al. 2000), giving
486: a value of $L_{X}/L_{H\alpha} \sim 3 - 4$. In contrast, the ratio of
487: $L_{X}/L_{H\alpha}$ for the starburst wind in M82 is $\sim 19$ (using
488: the $H\alpha$ luminosity measured by McCarthy et al. (1987). We note
489: also that the ``mean'' temperatures for the galactic winds in the
490: Strickland et al. sample are typically significantly lower than that
491: of the gas in NGC\,6764, although these result from two-temperature
492: fits where the hotter component has temperatures similar to our
493: measured value for NGC\,6764. We find no evidence for a cooler
494: component of hot gas in NGC\,6764 (in two-temperature fits, the second
495: component tends to higher temperatures of $\sim 1$ keV with no
496: significant improvement in the fitting statistic).
497:
498: Another important difference between the X-ray emission in NGC\,6764
499: and galactic winds such as M82 is the sharp decrease in surface
500: brightness at its northern and southern boundaries. Rasmussen et al.
501: (2004) compare the inferred density slopes at the outer boundaries of
502: three starburst winds, M82, NGC\,253 and the dwarf galaxy NGC,1800,
503: finding slopes of $\alpha \sim 1$, where $n(r) \propto r^{-\alpha}$.
504: In constrast, NGC\,6764 shows an initially flat surface brightness
505: profile (see Fig.~\ref{profs}), dropping steeply at distances between
506: 2 and 5$^{\prime\prime}$ for both the northern and southern bubbles,
507: with a slope that corresponds to $\alpha \sim 2.5$. This is a steeper
508: decrease than would be expected for a freely expanding wind (e.g.
509: Chevalier \& Clegg 1985).
510:
511: The presence of hot, overpressured gas in the inner regions of the
512: galaxy is consistent with the presence of a galactic wind; however,
513: its one-sided structure and East-West orientation may be difficult to
514: reconcile with a galactic wind model. We are not aware of any other
515: galactic wind systems in which a central hot-gas outflow with this
516: type of structure has been observed. An alternative is a photoionized
517: wind that is driven directly by the AGN. Such a model cannot be ruled
518: out on energetic grounds alone: our upper limit on the possible hidden
519: nuclear X-ray luminosity is higher than the energy required to power
520: the bubbles for realistic timescales of bubble inflation (e.g. $10^{6}
521: - 10^{7}$ years). However, as for the starburst-driven wind model, the
522: AGN wind model does not appear consistent with the steep decrease in
523: surface brightness of the X-ray bubbles in NGC\,6764. The origin of
524: the radio emission is also unclear in this model, and the E-W
525: orientation of the brighter central outflow may be problematic for an
526: AGN wind explanation.
527:
528: We conclude that there are important differences between the
529: properties of the X-ray emission and starburst activity of NGC\,6764
530: compared to those of well-studied examples of starburst-driven galactic
531: winds. Most crucially, the close correspondence between extended
532: non-thermal radio emission and X-ray emission is not seen in any of
533: the starburst wind systems. HS06 argue that the lack of such radio
534: bubbles in any galaxies that do not possess an AGN indicates that the
535: bubbles are likely to be powered by the AGN radio jet, although their
536: dynamics and evolution may be affected by the presence of a galactic
537: wind.
538:
539: \subsection{Model IIa: shock heating by the expanding radio bubbles}
540:
541: The properties of the hot gas associated with the radio bubbles in
542: NGC\,6764 are similar to those of the shocked gas associated with the
543: kpc-scale radio lobes of Centaurus A and NGC\,3801. The temperatures
544: we measure are similar to those seen in NGC\,3801, and the gas
545: densities and pressures are comparable to those measured in the
546: shocked shells of gas surrounding Centaurus A [assuming the thermal
547: interpretation of Kraft et al. (2003)]. However, there is no strong
548: evidence for edge-brightening of the X-ray emission -- in fact both
549: the radio and X-ray emission appear to have similar structures, with a
550: slight surface brightness deficit towards the centre of the bubbles,
551: particularly in the north. It would be possible to contrive a shock
552: geometry in which there was more shocked material in front of or
553: behind the bubbles relative to that at the sides, thus counteracting
554: the effect of edge-brightening; however, the smoothness of the X-ray
555: emission argues against a model in which the shocked material is very
556: inhomogeneous.
557:
558: We first considered the case where the X-ray emission is produced from
559: shocked shells of gas surrounding the radio bubbles. Assuming a strong
560: adiabatic shock, the expected density contrast between shocked and
561: unshocked gas should be a factor of 4. Therefore a shocked gas model
562: requires a surrounding medium with $n_{e} \sim 0.08$ cm$^{-3}$. As X-ray surface brightness depends
563: primarily on the gas density, we can use an upper limit on the X-ray
564: surface brightness from regions surrounding the X-ray emission
565: associated with the radio bubbles to investigate whether a hot gas
566: medium could be being shocked. We considered a rectangular region
567: covering the eastern half of the southern radio bubble and a matched
568: background region next to the X-ray bright region. We found a surface
569: brightness of $\sim 9.4\pm0.1$ net ACIS-S 0.5 -- 5 keV counts
570: arcsec$^{-2}$ in the source region. If hot gas with a density a factor
571: of 4 lower than that in the bright regions were present, we would
572: expect it to have a surface brightnes a factor $\sim 16$ times lower,
573: e.g. $S_{X} \sim 0.59\pm0.01$ counts arcsec$^{-2}$ (if the unshocked
574: material is cooler this prediction decreases somewhat, e.g. to 0.5
575: counts arcsec$^{-2}$ for $kT = 300$ eV). We measure an upper limit of
576: $\sim 0.1$ counts arcsec$^{-2}$ from the region external to the
577: southern radio bubble, well below this prediction. It therefore seems
578: unlikely that hot ionized plasma can be the medium being shocked.
579:
580: The densities of molecular gas in the centre of NGC\,6764 are much
581: higher than our prediction for the density of the material being
582: shocked (e.g. $n_{e} \sim 3 \times 10^{3} - 3 \times 10^{4}$
583: cm$^{-3}$; Eckart et al. 1991). A possible candidate for material
584: being shocked is the warm, diffuse interstellar medium, which has
585: typical electron densities of $\sim 0.2$ cm$^{-3}$ (e.g. Walterbos
586: 1998). If the bubbles extend perpendicular to the plane of the galaxy
587: disk, then they are likely to be probing regions of lower than average
588: ISM density. In addition, a weaker shock could result in a density
589: contrast lower than 4, and so this candidate for shocked material
590: cannot be ruled out on grounds of density. As seen in Fig.~12 of HS06,
591: the radio bubbles of NGC\,6764 are coincident with an extended region
592: of H$\alpha$-emitting gas (Zurita et al. 2000), with brighter
593: filaments associated with the radio bubbles, indicating that the radio
594: bubbles are likely to be embedded in diffuse gas.
595:
596: The most likely large-scale shock scenario is one in which a strong
597: shock is propagating into a multiphase medium. We can use the
598: postshock temperature to estimate the shock speed, assuming that the
599: pre-shocked gas pressure is negligible. The inferred expansion speed
600: is $\sim 740$ km s$^{-1}$, which is comparable to the radio-lobe
601: expansion speeds measured from the shock properties in NGC\,3801
602: ($\sim 850$ km s$^{-1}$), and would imply a total kinetic energy in
603: the shocked material of $\sim 3 \times 10^{55}$ ergs, similar to, but
604: slightly lower than the thermal energy stored in the shells. If the
605: shells are expanding at such a high speed, then the age of the X-ray
606: features must be low ($\sim 10^{6}$ years, assuming a sound speed
607: appropriate for a hot gas ISM), implying that either the features are
608: short-lived or the gas supply is being replenished. If the latter is
609: the case, then an outflow of gas from the galaxy is likely to be
610: required to supply the ionized gas halo, even if it does not supply
611: the energy to power the X-ray emission.
612:
613: \subsection{Model IIb: Heating by jet/ISM interactions in the central
614: outflow region}
615:
616: The X-ray bright, overpressured region in the central few arcsec of
617: the galaxy suggests a second model for jet-related heating of the
618: X-ray-emitting gas. If this region is an outflow that is supplying
619: material for the large-scale bubbles, then it is clear from the good
620: radio and X-ray correspondence that it contains both hot gas and a
621: population of radio-emitting, non-thermal particles, presumably
622: supplied by a radio jet. The central X-ray/radio morphology is hard
623: to explain in Model IIa, although the X-ray emission could be caused
624: by expansion of the bubbles near to the nucleus. The most plausible
625: explanation for the close correspondence, assuming that the radio
626: plasma traces a nuclear outflow, is that the ISM is being heated and
627: compressed as it interacts with the radio-emitting outflow. The
628: interactions may also be responsible for decollimating the flow. If
629: the radio plasma and heated gas components can become well mixed in
630: the inner outflow, then it seems likely that they will remain mixed on
631: larger scales. Hence an alternative jet-heating scenario is one in
632: which the inner jet is disrupted by interactions with the clumpy ISM,
633: which also heat ISM gas and cause it to be mixed in with the jet
634: fluid. This is similar to the ``frustrated jet'' model discussed by
635: Gallimore et al. (2006), which predicts soft X-ray emission associated
636: with shocks and entrained material. While the energy is clearly
637: available from the jet to enable such heating on small scales to
638: occur, it is unclear how the details of such a heating and mixing
639: process can work in practice.
640:
641: In this scenario the bubbles are filled with hot, thermal gas. In this
642: case we would expect significant depolarization of the radio
643: synchrotron emission. Using the radio data discussed above, we found
644: that the integrated fractional polarization of the bubbles is $<2 \%$
645: and $<5 \%$ at 1.4-GHz and 5-GHz, respectively. This is consistent
646: with the level of expected internal Faraday depolarization given the
647: gas density and equipartition magnetic field strength in the bubbles
648: (see Cioffi \& Jones 1980). While the level of polarization is
649: consistent with Model IIb, it is also consisent with an external
650: Faraday screen having the density and path length expected for shocked
651: gas surrounding the bubbles in Model IIa (for reasonable values of the
652: magnetic field strength and plausible assumptions about the cell size
653: in the Faraday screen) using the analysis of Burn (1966), and so we
654: cannot rule out Model IIa on this basis.
655:
656: The right-hand panel of Fig.~\ref{inner} shows a hardness ratio map
657: obtained by dividing Gaussian smoothed images in the 1.0 -- 5.0 keV
658: and 0.4 -- 1.0 keV energy ranges. The region of harder X-ray emission
659: along the western edge of the bubbles is statistically significant (as
660: verified by comparing count ratios with uncertainties for the western
661: and eastern edges of the bubbles), and corresponds to a region of
662: flatter radio spectral index as shown in HS06. Gallimore et al. (1996)
663: propose a similar explanation for component C of the parsec-scale
664: jet in NGC\,1068 parsec-scale jet, which has a similar spectral index
665: to the regions we consider here and is associated with H$_{2}$0 maser
666: emission. These results suggest that the East-West outflow may be
667: shock heating the ISM in this region, leading to an increase in X-ray
668: temperature, and a flattening of the radio spectrum due to localized
669: particle acceleration. The radio and X-ray spectral structure
670: therefore offer support for a model in which shock heating plays at
671: least some role in heating the X-ray emitting gas. The edge-brightened
672: radio and X-ray structure of the bubbles suggests that shock heating
673: and particle acceleration may also be occurring at the edges of the
674: bubbles (although there is no flattening of the radio spectrum, or
675: hardening of the X-ray spectrum around the eastern edges). Magnetic
676: field compression offers an alternative explanation for the radio
677: structure, but does not explain the close correspondence with the
678: thermal X-ray emission. The surface brightness profiles of the bubbles
679: (Fig.~\ref{profs}) are too steep to be consistent with a uniformly
680: filled sphere of gas, supporting a model in which some compression and
681: possibly shocks are occurring at the edges of the bubbles.
682:
683: Sutherland \& Bicknell (2007) presented simulations of jet/ISM
684: interactions in a non-homogeneous medium that show some similarities
685: with our observations: the radio and X-ray morphologies of the
686: large-scale bubbles in NGC\,6764 are somewhat similar to their
687: ``energy-driven bubble'' phase, with the inner region having a
688: morphology that could be consistent with their initial ``flood and
689: channel'' phase, which they find persists in the inner regions of the
690: outflow during the energy-driven phase. In their simulations X-ray
691: emission from shocks driven out by the bubbles is also significant.
692: However, the simulations were for much more powerful jets and for
693: initial environmental conditions quite different to those of NGC\,6764
694: (including a static clumpy ISM), Nevertheless they may provide some
695: qualitative support for a jet/ISM interaction interpretation in
696: NGC\,6764.
697:
698: If it is the radio jet and its environmental impact that are the
699: primary explanation for the bubbles, then it is perhaps surprising
700: that the inner jet/outflow is oriented in the East-West direction, not
701: North-South. The can be explained if the jet direction can change on
702: short timescales, perhaps due to the interactions with dense
703: surrounding material in the inner regions. Kharb et al. (2006)
704: proposed such a model for the Seyfert 2 galaxy Markarian 6, which
705: possesses an inner jet nearly perpendicular to its large-scale radio
706: bubbles, concluding that the jet direction changes on timescales of
707: $10^{5} - 10^{6}$ years. The radio and X-ray outflow properties of
708: NGC\,6764 are consistent with the picture described by Gallimore et
709: al. (2006) in which jet-ISM interactions in the inner kiloparsec lead
710: to frustration of the AGN jet. We conclude that this model is somewhat
711: more plausible than a large-scale radio-bubble driven shock (Model
712: IIa); however, neither model can be ruled out.
713:
714: \subsection{Comparisons with other Seyfert radio-bubble systems}
715:
716: HS06 presented a list of Seyfert radio-bubble systems similar to
717: NGC\,6764. {\it Chandra} X-ray studies of three of these systems are
718: published in the literature, NGC\,1068, M51 and NGC\,3079.
719: Below we discuss the similarities and differences between NGC\,6764
720: and each of these previously studied radio-bubble systems.
721:
722: The Seyfert 2 galaxy NGC\,1068 has two asymmetric radio bubbles with a
723: total extent of $\sim 1$ kpc, which are connected to a radio-bright
724: nucleus. The {\it Chandra} study of Young et al. (2001) showed the
725: presence of bright extended X-ray emission associated with the
726: north-eastern radio bubble, with a lack of corresponding emission
727: associated with the south-western bubble (likely to due to high
728: absorption from the disk of the galaxy). The X-ray and radio
729: morphology of this system are consistent with similar behaviour to
730: NGC\,6764. Young et al. required more complex spectral models to
731: account for the X-ray emission associated with the north-east radio
732: bubble region; however, they were able to obtain spectra of much
733: higher signal-to-noise than ours for NGC\,6764, and so it is possible
734: that a single-temperature {\it mekal} model would also prove
735: inadequate if we were able to obtain similar quality data for
736: NGC\,6764. Young et al. suggest that compression of the ISM by the
737: radio ejecta is responsible for the bubble-related X-ray emission.
738:
739: Terashima \& Wilson (2001) presented a {\it Chandra} study of the
740: extended emission from the Seyfert 2 galaxy M51, which possesses a
741: highly asymmetry two-sided radio bubble structure. As for NGC\,6764,
742: they found extended X-ray emission corresponding closely to the radio
743: structure. The X-ray emission is spectrally similar to that seen in
744: NGC\,6764 ($kT \sim 0.5 - 0.6$ keV), and interestingly a region of
745: harder X-ray emission is seen at the position where the radio jet is
746: seen to teminate, similar to what is seen in NGC\,6764. Terashima \&
747: Wilson adopt a shock heating model for M51, finding bubble expansion
748: velocities very similar to those inferred above for NGC\,6764 ($\sim
749: 700$ km s$^{-1}$).
750:
751: NGC\,3079 is another composite starburst/AGN system, which possesses
752: two kpc-scale scale radio bubble structures . Although the radio
753: structures have been modelled as originating in a nuclear wind (e.g.
754: Duric \& Seaquist 1988), long-term VLBI monitoring shows emitting
755: regions whose structure and time variability is consistent with the
756: expectation for a jet interacting with a dense and clumpy surrounding
757: medium (Middelberg et al. 2007). Middelberg et al. also argue that
758: these interactions can explain the large-scale morphology of the radio
759: bubbles. Cecil et al. (2002) present {\it Chandra} observations of
760: NGC\,3079, which show filamentary X-ray structure coinciding with
761: H$\alpha$ filaments; however, there is also a region of X-ray emission
762: associated with the southern radio bubble, which is too faint for
763: spectral analysis. The optical and X-ray filaments to the north do not
764: coincide with the radio bubble, but Cecil et al. claim that there is
765: some evidence for shock heating of gas surrounding the VLBI jet. It is
766: clear that starburst winds are responsible for some of the X-ray
767: structure in NGC\,3079, but it remains possible that X-ray emission
768: heated by the expanding radio plasma is also present. The relationship
769: between the radio structure and the optical and X-ray filaments in
770: NGC\,3079 is unclear.
771:
772: Based on the comparisons above, it seems likely that the radio/X-ray
773: bubbles observed in NGC\,6764, NGC\,1068 and M51 are examples of the
774: same phenomenon, whereas the situation in NGC\,3079 may be more
775: complex. We conclude that the behaviour seen in NGC\,6764 may be
776: widespread in the Seyfert population, as such outbursts are likely to
777: be short-lived. We would expect similar X-ray properties for radio
778: bubbles in Seyfert 1 galaxies; however, due to the small size scales
779: of the lobes and the expected projection, bubble-related X-ray
780: emission is likely to be difficult to disentangle from the brighter
781: nuclear X-ray emission in these systems.
782:
783: \section{Conclusions}
784:
785: We have identified X-ray counterparts to the kpc-scale radio bubbles
786: in the Seyfert galaxy NGC\,6764, whose morphology traces very closely
787: the radio structures. The energy input from these hot gas features is
788: $\sim 10^{56}$ ergs, comparable to the energy input from shock heating
789: in the low-power radio galaxy NGC\,3801. The inferred rate of energy
790: input in NGC\,6764 is $\sim 10^{42}$ erg s$^{-1}$, and is expected to
791: have a significant impact on the galaxy and its surroundings. Based on
792: the presence of an inner outflow that appears to emanate from the
793: Seyfert nucleus, seen in both radio and X-ray emission, similarities
794: to X-ray features in previously studied AGN-jet systems, key
795: differences compared to starburst outflows, and the spatial
796: coincidence of a region of flatter radio spectrum and hard X-ray
797: spectrum, suggesting particle acceleration, we conclude that the
798: thermal emission is produced by heating of the ISM due to compression
799: and shocks caused by the AGN outflow. The morphology of the radio
800: outflow in NGC\,6764 leads us to favour a model in which jet/ISM
801: interactions in the inner part of the galaxy lead to heating of the
802: gas and its entrainment into the bubbles, as well as jet disruption;
803: however, we cannot rule out a model in which the bubbles are driving
804: large-scale shocks into the ISM. The existence of a population of
805: kpc-scale radio lobe sources in Seyfert galaxies (expected to be
806: short-lived) suggests that AGN-jet or AGN-jet+starburst powered galaxy
807: feedback as seen in NGC\,6764 could be important for the galaxy
808: population as a whole.
809:
810: \acknowledgments
811:
812: We thank the referee for helpful comments. We gratefully acknowledge
813: support from the Royal Society (research fellowship for MJH). This
814: work was partially supported by NASA grant G08-9108X.
815:
816: \begin{thebibliography}{}
817: \bibitem[]{awa00} Awaki, H., Ueno, S., Taniguchi, Y., Weaver, K.A.,
818: 2000, ApJ, 542, 175
819: \bibitem[]{bau93} Baum, S.A., O'Dea, C.P., Dallacassa, D., de Bruyn,
820: A.G, Pedlar, A., 1993, ApJ, 419, 553
821: \bibitem[]{bow06} Bower, R.G., Benson, A.J., Malbon, R., Helly, J.C., Frenk, C.S., Baugh, C.M., Cole, S., Lacey, C.G., 2006, MNRAS, 370, 645
822: \bibitem[]{buo00} Buote, D., 2000, MNRAS, 311, 176
823: \bibitem[]{bur66} Burn, B.J., 1966, MNRAS, 133, 67
824: \bibitem[]{cap99} Capetti, A., Axon, D.J., Macchetto, F.D., Marconi,
825: A., Winge, C., ApJ, 516, 187
826: \bibitem[]{cec02} Cecil, G., Bland-Hawthorn, J., Veilleux, S., 2002,
827: ApJ, 576, 745
828: \bibitem[]{cc85} Chevalier, R.A., Clegg, A.W., 1985, Nature, 317. 44
829: \bibitem[]{cj80} Cioffi, D.F., Jones, T.W., 1980, AJ, 85, 368
830: \bibitem[]{cle81} Clements, I., 1981, MNRAS, 197, 829
831: \bibitem[]{cro06} Croton, D.J. et al. 2006, MNRAS, 365, 11
832: \bibitem[]{c07} Croston, J.H., Kraft, R.P., Hardcastle, M.J., 2007, ApJ, 660, 191
833: \bibitem[]{c08} Croston, J.H., Hardcastle, M.J., Birkinshaw, M.,
834: Worrall, D.M., Laing, R.A., 2008, MNRAS, in press (arXiv:0802.4297)
835: \bibitem[]{dl90} Dickey, J.M., Lockman, F.J., 1990, ARA\&A, 28, 215
836: \bibitem{ds88} Duric, N., Seaquist, E.R., 1988, ApJ, 326, 574
837: \bibitem[]{eck91} Eckart, A., Cameron, M., Jackson, J.M., Genzel, R., Harris, A.I., Wild, W., Zinnecker, H., 1991, ApJ, 372, 67
838: \bibitem[]{eck96} Eckart, A., Cameron, M., Boller, Th., Krabbe, A., Blietz, M., Nakai, N., Wagner, S.J., Sternberg. A., 1996, ApJ. 472, 599
839: \bibitem[]{fab03} Fabian, A.C., Sanders, J.S., Allen, S.W., Crawford, C.S., Iwasawa, K., Johnstone, R.M., Schmidt, R.W., Talor, G.B., 2003, MNRAS, 344, L43
840: \bibitem[]{for05} Forman, W. et al. 2005, ApJ, 635, 894
841: \bibitem[]{for03} F\"{o}rster Schreiber, N.M., Genzel, R., Lutz, D.,
842: Sternberg, A., 2003, ApJ, 599, 193
843: \bibitem[]{gal96} Gallimore, J.F., Baum, S.A., O'Dea, C.P., Pedlar,
844: A., 1996, ApJ, 458, 136
845: \bibitem[]{gal04} Gallimore, J.F., Baum, S.A., O'Dea, C.P., 2004, ApJ,
846: 613, 794
847: \bibitem[]{gal06} Gallimore, J.F., Axon, D.J., O'Dea, C.P., Baum, S.A., Pedlar, A., 2006, AJ, 132, 546
848: \bibitem[]{hot06} Hota, A., Saikia, D.J., 2006, MNRAS, 371, 945 [HS06]
849: \bibitem[]{kha06} Kharb, P., O'Dea, C.P., Baum, S.A., Colbert, E.J.M., Xu, C., 2006, ApJ, 652, 177
850: \bibitem[]{kra03} Kraft, R.P., V\'{a}zquez, S.E., Forman, W.R., Jones, C., Murray, S.S., Hardcastle, M.J., Worrall, D.M., Churazov, E., 2003, ApJ, 592, 129
851: \bibitem[]{kra07} Kraft, R.P., et al., 2007, ApJ, 665, 1129
852: \bibitem[]{leo07} Leon, S. et al., 2007, A\&A, 473, 747
853: \bibitem[]{mas04} Matsumoto, C., Nava, A., Maddox, L.A., Leighly,
854: K.M., Grupe, D., Awaki, H., Ueno, S., 2004, ApJ, 617, 930
855: \bibitem[]{mat04} Matt, G., Bianchi, S., Guainazzi, M., Molendi, S.,
856: 2004, A\&A, 414, 155
857: \bibitem[]{mid07} Middelberg, E., Agudo, I., Roy, A.L., Krichbaum,
858: T.P., 2007, MNRAS, 377, 731
859: \bibitem[]{mcc87} McCarthy, P.J., van Breugel, W., Heckman, T., 1987,
860: AJ, 93, 264
861: \bibitem[]{ras04} Rasmussen, J., Stevens, I.R., Ponman, T.J., 2004,
862: MNRAS, 354, 259
863: \bibitem[]{sch00} Schinnerer, E., Eckart, A., Boller, Th., 2000, ApJ, 545, 205
864: \bibitem[]{str1} Strickland, D.K., Heckman, T.M., Colbert, E.J.M., Hoopes, C.G., Weaver, K., 2004a, ApJSS, 151, 193
865: \bibitem[]{str2} Strickland, D.K., Heckman, T.M., Colbert, E.J.M., Hoopes, C.G., Weaver, K., 2004b, ApJ, 606, 829
866: \bibitem[]{sut} Sutherland, R.S., Bicknell, G.V., 2007, Ap\&SS, 311, 293
867: \bibitem[]{ter01} Terashima, Y., Wilson, A.S., 2001, ApJ, 560, 139
868: \bibitem[]{wal98} Walterbos, R.A.M., 1998, PASA, 15, 99
869: \bibitem[]{you01} Young, A.J., Wilson, A.S., Shopbell, P.L., 2001,
870: ApJ, 556, 6
871: \bibitem[]{zur00} Zurita, A., Rozas, M., Beckman, J.E., 2000, A\&A,
872: 363, 9
873:
874: \end{thebibliography}
875:
876:
877:
878: \end{document}
879: