1: %\documentclass[12pt,preprint]{aastex}
2: \documentclass{emulateapj}
3: \usepackage{times}
4: %\usepackage{epsf}
5: %\usepackage{apjfonts}
6: %\usepackage{epsfig}
7: %\usepackage{amsmath}
8: %\usepackage{amssymb}
9: %\usepackage{amsfonts}
10: %\usepackage{natbib} % This is required to use the apj.bst
11: %\citestyle{aa} % and be able to use the AASTEX 5.x
12: % citations (\citep, \citet, etc.) - bph
13: %\bibliographystyle{apj}
14:
15:
16: \newcommand{\source}{\hbox{Cen\,A}}
17: \newcommand{\ntof}{\hbox{NGC\,315}}
18: \newcommand{\chandra}{\textit{Chandra}}
19:
20: \begin{document}
21:
22: \slugcomment{Dec 13th 2007: Accepted for publication in ApJ Letters}
23:
24: \shorttitle{Cen A X-ray knottiness}
25: \shortauthors{D.M.~Worrall et al.}
26:
27: \title{Where Centaurus A gets its X-ray knottiness}
28:
29: \author{
30: D.M.~Worrall\altaffilmark{1,2},
31: M.~Birkinshaw\altaffilmark{1,2},
32: R.P.~Kraft\altaffilmark{2},
33: G.R.~Sivakoff\altaffilmark{3},
34: A.~Jord\'{a}n\altaffilmark{2},
35: M.J.~Hardcastle\altaffilmark{4},
36: N.J.~Brassington\altaffilmark{2},
37: J.H.~Croston\altaffilmark{4},
38: D.A.~Evans\altaffilmark{2},
39: W.R.~Forman\altaffilmark{2},
40: W.E.~Harris\altaffilmark{5},
41: C.~Jones\altaffilmark{2},
42: A.M.~Juett\altaffilmark{6},
43: S.S.~Murray\altaffilmark{2},
44: P.E.J.~Nulsen\altaffilmark{2},
45: S.~Raychaudhury\altaffilmark{7,2},
46: C.L.~Sarazin\altaffilmark{6},
47: K.A.~Woodley\altaffilmark{5}
48: }
49: \altaffiltext{1}{Department of Physics,
50: University of Bristol,
51: Tyndall Avenue,
52: Bristol\ \ BS8 1TL, UK}
53: \altaffiltext{2}{Harvard-Smithsonian Center for Astrophysics,
54: 60 Garden Street,
55: Cambridge, MA\ \ 02138}
56: \altaffiltext{3}{
57: Department of Astronomy,
58: The Ohio State University,
59: 4055 McPherson Laboratory,
60: 140 W.~18th Avenue,
61: Columbus, OH\ \ 43210}
62: \altaffiltext{4}{School of Physics, Astronomy \& Mathematics,
63: University of Hertfordshire, College Lane, Hatfield\ \ AL10 9AB, UK}
64: \altaffiltext{5}{
65: Department of Physics and Astronomy,
66: McMaster University,
67: Hamilton, ON\ \ L8S 4M1, Canada}
68: \altaffiltext{6}{
69: Department of Astronomy,
70: University of Virginia,
71: P. O. Box 400325,
72: Charlottesville, VA\ \ 22904
73: }
74: \altaffiltext{7}{
75: School of Physics and Astronomy,
76: University of Birmingham,
77: Edgbaston, Birmingham\ \ B15 2TT, UK
78: }
79:
80: %\email{d.worrall@bristol.ac.uk}
81: %\email{mark.birkinshaw@bristol.ac.uk}
82: %\email{rkraft@cfa.harvard.edu}
83: %\email{sivakoff@astronomy.ohio-state.edu}
84: %\email{ajordan@cfa.harvard.edu}
85: %\email{mjh@star.herts.ac.uk}
86: %\email{nbrassington@cfa.harvard.edu}
87: %\email{J.H.Croston@herts.ac.uk}
88: %\email{devans@cfa.harvard.edu}
89: %\email{wrf@cfa.harvard.edu}
90: %\email{harris@physics.mcmaster.ca}
91: %\email{cjf@cfa.harvard.edu}
92: %\email{amj3r@virginia.edu}
93: %\email{ssm@cfa.harvard.edu}
94: %\email{pnulsen@cfa.harvard.edu}
95: %\email{somak@star.sr.bham.ac.uk}
96: %\email{cls7i@virginia.edu}
97: %\email{woodleka@physics.mcmaster.ca}
98:
99:
100:
101: \begin{abstract}
102:
103: We report an X-ray spectral study of the transverse structure of the
104: Centaurus~A jet using new data from the \chandra\ \source\ Very Large
105: Project. We find that the spectrum steepens with increasing distance
106: from the jet axis, and that this steepening can be attributed to a
107: change in the average spectrum of the knotty emission. Such a trend
108: is unexpected if the knots are predominantly a surface feature
109: residing in a shear layer between faster and slower flows. We suggest
110: that the spectral steepening of the knot emission as a function of
111: distance from the jet axis is due to knot migration, implying a
112: component of transverse motion of knots within the flow.
113:
114: \end{abstract}
115:
116: \keywords{galaxies: active ---
117: galaxies: individual (\objectname{Centaurus A, NGC 5128}) ---
118: galaxies: jets
119: --- X-rays: galaxies}
120:
121: \section{Introduction}
122: \label{sec:intro}
123:
124: In relatively few extragalactic radio jets can the kinematics of the
125: flow on kpc scales be studied with any degree of certainty. Most
126: arguments are indirect, relying either on beaming statistics for
127: samples \citep[e.g.,][]{hard03a} or, for powerful jets oriented along
128: the line of sight, on models relying on an assumed inverse-Compton
129: origin for the X-ray emission \citep[e.g.,][]{schwartz06}. For the
130: few low-power radio galaxies with heavily studied, straight, radio
131: jets and counterjets, kinematic models have been constructed based on
132: the jet-counterjet asymmetry \citep[e.g.,][]{canvin05}. These models
133: have supported jet deceleration through mass entrainment, and ongoing
134: work has already shown consistency between the density and pressure
135: model inferred from the jet and the properties of the external
136: X-ray-emitting gas for 3C\,31 \citep{laing02}.
137:
138: In models of mass entrainment, the outer parts (``sheath'') are
139: decelerated before the inner (``spine''), consistent with a range of
140: observational evidence at radio frequencies \citep{laing96}. Applied
141: to more central regions, the consequence that emission from a slower
142: sheath becomes relatively more important in jets at larger angle to
143: the line of sight then resolves difficulties in models that unify BL
144: Lac objects with low power radio galaxies
145: \citep[e.g.,][]{chiaberge00}.
146:
147: \chandra\ has not only allowed confirmation that X-ray synchrotron
148: emission in low-power jets is common \citep{worrall01}, but has also
149: resolved transverse structure in the nearest and brightest sources. A
150: particularly interesting example is \ntof\ \citep{worrall07}. Here
151: the diffuse emission contains a knotty structure in the radio and
152: X-ray that appears to describe an oscillatory filament. Although the
153: structure could be the result of a chance superposition of
154: non-axisymmetric knots, the level of coherence led us to suggest that
155: the knots might be predominantly a surface feature residing in the
156: shear layer between the spine and sheath. If this latter
157: interpretation is correct, we might expect the X-ray spectra of the
158: knots to be similar across the transverse width of the jet.
159:
160: In the case of \ntof\ the distinct knotty emission is only about 10\%
161: of the total emission in X-rays and radio along the $\sim 2.5$ kpc of
162: projected jet length over which it is detected, and with a source
163: distance of $\sim 70$~Mpc the observations did not allow us to measure
164: separate X-ray spectra for the knots and diffuse emission. At
165: 3.7~Mpc\footnote{The average of five distance indicators, see \S6 in
166: \citet{ferrarese07}} (1~arcmin is $\sim 1.1$~kpc projected length),
167: Centaurus~A is a much closer example of a low-power radio galaxy whose
168: inner jet shows bright resolved X-ray knots and diffuse emission over
169: a similar projected length scale to \ntof\ \citep[e.g.][]{kraft02}.
170: Here we report an X-ray spectral study of the transverse structure of
171: the jet, and in particular we investigate whether the \source\ knots
172: are likely to occupy a shear layer between faster inner and slower
173: outer flows.
174:
175: \section{Observations \& analysis}
176: \label{sec:obs}
177:
178: \begin{figure*}
179: \centering
180: %emulateapj%
181: \includegraphics[width=1.9\columnwidth,clip=true]{f1.eps}
182: %aastex
183: %\epsscale{0.67}
184: %\plotone{jetimage-regions0.8-3keV-cmyk3.ps}
185: \caption{0.8-3-keV unsmoothed rotated image of the six
186: \chandra\ exposures comprising the \source-VLP,
187: with $0.492 \times 0.492$ arcsec$^2$ pixels. }
188: \label{fig:jet}
189: \end{figure*}
190:
191: \begin{figure}
192: \centering
193: %emulateapj%
194: \includegraphics[width=1.0\columnwidth,clip=true]{f2.eps}
195: %aastex
196: %\epsscale{0.67}
197: %\plotone{jetimage-regions0.8-3keV-cmyk3.ps}
198: \caption{Image from Fig.~\ref{fig:jet} with regions superposed, and
199: rotated to the sky orientation. The regions, called the spine, sheath and
200: background, working
201: outwards from the jet axis, are pie slices from a base position
202: of R.A.~$=13^{\rm h} 25^{\rm m} 26\fs98$, Dec.~$=-43\arcdeg01\arcmin14\farcs06$
203: (not the core) running from position angle (pa) 51\fdg2--57\fdg8 for the spine,
204: pa 49$^\circ$--51\fdg2 and 57\fdg8--60$^\circ$ for the sheath, and
205: pa 31$^\circ$--47$^\circ$ and 63$^\circ$--79$^\circ$ for the
206: background with the exclusion regions shown. From the core the inner
207: spine and sheath regions are at a distance of $21''$ to $66''$, and the
208: outer region runs from $66''$ to $141''$. We call the region marked in blue,
209: near the outer extent of the inner region, the ``bright''
210: region.}
211: \label{fig:jetregions}
212: \end{figure}
213:
214: The \chandra\/ data used here are from the \source\ Very Large Project
215: (\source-VLP), comprising six deep ($\sim 100$~ks) exposures (OBSIDs
216: 7797, 7798, 7799, 7800, 8489, 8490) of the front-illuminated CCD chips
217: of the Advanced CCD Imaging Spectrometer (ACIS) made between 2007
218: March 7 and May 30. Standard
219: methods\footnote{http://asc.harvard.edu/ciao/} have been followed for
220: data re-processing and grade selection, and small astrometric
221: corrections were made to register the six exposures to a common frame
222: using detected point sources \citep{hard07, jordan07}. The 0.8-3~keV
223: merged image of the inner $\sim 4'$ of the X-ray jet is shown in
224: Figure~\ref{fig:jet}.
225:
226: We are interested here in the jet after it has become radiatively
227: bright in both radio and X-ray \citep[more than $21''$ from the core,
228: and after knots AX1A, AX1C in the terminology of ][]{hard03b} until
229: its major decline relative to the radio (at $\sim 141''$). Motivated
230: by \ntof, where the fast spine is approximately 60\% of the jet width,
231: we have defined regions shown in Figure~\ref{fig:jetregions}. While
232: not wishing to suggest {\it a priori\/} that the \source\ jet does
233: have a slower sheath and faster spine, for shorthand we call these
234: regions the {\it spine}, {\it sheath} and {\it background}, working
235: outwards from the jet axis. Compact sources excluded from the
236: background regions are also marked. A division of the spine and
237: sheath into two regions with length down the jet is shown. We refer
238: to these as the {\it inner} and {\it outer} regions, and the boundary
239: was chosen so that there is a comparable number of total counts in
240: each region.
241:
242: The jet is positioned on a single ACIS I chip in all six exposures.
243: The boundary between the outer and inner regions lies between 1.4 and
244: 4.0 arcmin off axis depending on OBSID, giving a point-spread-function
245: (PSF) half-power-diameter (HPD) at this location of better than about
246: 1 arcsec at the $\sim 1.5$-keV energy of maximum count rate per keV.
247: The readout streak from the \source\ core runs close to the SE sheath
248: region in OBSID 7797, and so we exclude this observation in
249: spectral fitting to the SE sheath and inner spine.
250: We extract data and responses for identical sky regions in the
251: different OBSIDs, and we fit the same model simultaneously to the data
252: sets. Spectra are binned to a minimum of 30 counts per bin so that
253: the $\chi^2$ statistic can be applied.
254:
255: In this work we are searching for statistical trends in spectral index
256: in knotty and diffuse emission transverse to the jet axis. Knots are
257: identified as bright enhancements, but we believe them to be composed
258: of multiple distinct physical structures, each potentially with a
259: different emission spectrum. The precise definition of knots is
260: therefore not crucial, and we have adopted the previously published
261: lists of \citet{kraft02} and \citet{kataoka}, which combined give 31
262: knots within our total spine and sheath regions. The knots were
263: examined visually, and small centroid shifts (within uncertainties)
264: were made to give best alignment in the new data. A circle of nominal
265: radius 2.4 arcsec was used to mask each knot. Four additional weak
266: but compact enhancements seen in the new data were also masked as
267: knots with circles of radii 1.2 arcsec. Because of our interest in
268: statistical trends, for knots that overlap regions we divide the
269: counts according to where they were detected. When extracting spectra
270: of the knots the diffuse (residual) emission in the corresponding
271: region is used as background. The outer sectors shown in
272: Figure~\ref{fig:jetregions} provide background for the diffuse
273: emission.
274:
275: All our spectral models are single-component power laws with free
276: hydrogen column density, $N_{\rm H}$, over the energy range 0.4--10
277: keV. Intrinsic $N_{\rm H}$ in \source\ over and above the Galactic
278: value of $8.4 \times 10^{20}$ cm$^{-2}$ \citep{dlock90} causes
279: reductions in the count rate in soft X-ray images, consistent with the
280: dust lanes that cross the jet roughly transverse to its axis. Since
281: our regions extend hundreds of pc along the jet, we are combining data
282: with different intrinsic $N_{\rm H}$. This is expected to give a bias
283: towards finding slightly flatter power-law slopes than the true slope
284: if the jet emission is characterized by a single spectrum over its
285: length. However, the gradients in $N_{\rm H}$ are principally down the
286: jet and not across the jet, and so provide similar weightings when we
287: compare sections transverse to the jet axis. We can thus use the
288: fitted $N_{\rm H}$ and spectral index for our transverse sections of
289: the jet as proxies of systematic trends in the spectra. Appreciable
290: cross-jet differences could then allow us to conclude that there are
291: real variations in the X-ray spectrum across the jet.
292: Self-consistency tests (see \S\ref{sec:results}) verify this
293: assertion. We also find, through comparison with the spectra of
294: individual knots and sections of diffuse emission along the jet
295: \citep{hard07}, that any bias in spectral slope in our analysis is
296: small.
297:
298: In the following, spectral index $\alpha$ (one less than X-ray photon
299: index) is defined in the sense that flux density is proportional to
300: $\nu^{-\alpha}$.
301:
302:
303: \begin{deluxetable*}{clccccc}
304: \tablewidth{0pt}
305: \tablecaption{\chandra\ fits to a single-component power law
306: \label{tab:specresults}}
307: \tablehead{
308: \colhead{Row} & \colhead{Component/} &
309: \colhead{region \% counts} &
310: \colhead{region \% area} &
311: \colhead{$\alpha$\tablenotemark{a}} &
312: \colhead{$N_{\rm H}$\tablenotemark{a}($10^{21}$ cm$^{-2}$)} &
313: \colhead{$\chi^2_{\rm min}$/dof}
314: }
315: \startdata
316: 1 & Spine & 100 & 100
317: & $0.83^{+0.05}_{-0.04}$ & $1.54^{+0.18}_{-0.17}$ & 961/895 \\
318: 2 & Sheath& 100 & 100
319: & $0.97^{+0.08}_{-0.07}$ & $1.40\pm0.28$ & 527/512 \\
320: 3 & NW sheath & 100 & 100
321: & $0.98^{+0.12}_{-0.11}$ & $1.47^{+0.46}_{-0.44}$ & 318/277 \\
322: 4 & SE sheath & 100 & 100
323: & $0.99\pm0.11$ & $1.51^{+0.45}_{-0.42}$ & 295/282 \\
324: 5 & Spine knots & 64 & 25
325: & $0.60\pm0.06$ & $1.77^{+0.29}_{-0.28}$ & 602/657 \\
326: 6 & Sheath knots & 49 & 17
327: & $0.73^{+0.12}_{-0.11}$ & $1.12^{+0.53}_{-1.06}$ & 244/254 \\
328: 7 & Spine diffuse & 36 & 75
329: & $1.24\pm0.10$ & $1.79^{+0.37}_{-0.34}$ & 471/407 \\
330: 8 & Sheath diffuse & 51 & 83
331: & $1.20\pm0.14$ & $1.82^{+0.51}_{-0.50}$ & 338/308 \\
332: 9 & Inner spine knots & 77 & 47
333: & $0.63\pm0.08$ & $2.34^{+0.44}_{-0.42}$ & 474/520 \\
334: 10 & Inner sheath knots & 61 & 33
335: & $0.94^{+0.17}_{-0.18}$ & $1.64^{+0.80}_{-0.76}$ & 221/185 \\
336: 11 & Outer spine knots & 34 & 18
337: & $0.98^{+0.17}_{-0.18}$ & $0.09^{+1.48}_{-0.06}$ & 209/207 \\
338: 12 & Outer sheath knots & 27 & 13
339: & $0.80^{+0.30}_{-0.40}$ & $0.05^{+1.90}_{-0.05}$ & 60/71 \\
340: 13 & Inner spine diffuse & 23 & 53
341: & $0.96^{+0.16}_{-0.17}$ & $2.85^{+0.83}_{-0.88}$ & 182/168 \\
342: 14 & Inner sheath diffuse & 39 & 67
343: & $0.90^{+0.25}_{-0.22}$ & $2.73^{+1.27}_{-1.16}$ & 136/121 \\
344: 15 & Outer spine diffuse & 66 & 82
345: & $1.35^{+0.12}_{-0.11}$ & $1.46^{+0.40}_{-0.36}$ & 318/336 \\
346: 16 & Outer sheath diffuse & 73 & 87
347: & $1.33^{+0.19}_{-0.17}$ & $1.41^{+0.63}_{-0.59}$ & 191/191 \\
348: 17 & ``Bright'' spine knots & 88 & 57
349: & $0.74^{+0.11}_{-0.10}$ & $1.36^{+0.51}_{-0.48}$ & 229/249 \\
350: 18 & ``Bright'' sheath knots & 82 & 60
351: & $0.89^{+0.23}_{-0.19}$ & $1.12^{+0.96}_{-0.90}$ & 147/124
352: \enddata
353: \tablenotetext{a}{Errors are 90\% for 2 interesting parameters}
354: \end{deluxetable*}
355:
356: \section{Results}
357: \label{sec:results}
358:
359: Our spectral results are summarized in Table~\ref{tab:specresults},
360: and we refer to them here by row number. The quoted uncertainties are
361: 90\% for two interesting parameters, so where pairs of rows disagree
362: in both parameters there is at most a 1\% probability of agreement by
363: chance. Uncertainties in $\alpha$ and $N_{\rm H}$ are correlated, and
364: so the $\chi^2$ contours must be examined to assess fully the spectral
365: disagreement. We show such plots only for the comparisons of most
366: interest (Figs.~\ref{fig:contours} \& \ref{fig:contours2}).
367:
368:
369: We see (Table~\ref{tab:specresults} rows 1 \& 2) that the sheath
370: spectrum is steeper than the spine. Further investigation allows us
371: to find the origin of this result. The NW and SE regions of the
372: sheath (rows 3 \& 4) give excellent agreement in $\alpha$ and $N_{\rm
373: H}$, indicating that it is unlikely that the result is from isolated
374: features, and confirming the absence of strong gradients in $N_{\rm
375: H}$ transverse to the jet. Separating the knot and diffuse emission
376: in the spine and sheath (rows 5 to 8) shows that while the knots have
377: flatter spectra than the diffuse emission, the steeper spectrum in the
378: sheath is caused solely by differences in the knots. In the
379: two-dimensional plane of $\alpha$ and $N_{\rm H}$, rows 5 and 6
380: disagree.
381:
382: \begin{figure}
383: \centering
384: %emulateapj%
385: \includegraphics[width=0.71\columnwidth,clip=true]{f3.eps}
386: %aastex
387: %\epsscale{0.8}
388: %\plotone{mkfig.ps}
389: \caption{Confidence contours ($1\sigma$, 90\% and 99\% for two
390: interesting parameters) showing a significant difference between
391: the spine inner knots (left contours)
392: and sheath inner knots (right contours): rows 9 \& 10 of
393: Table~\ref{tab:specresults}. The $y$ axis begins at Galactic
394: $N_{\rm H}$.
395: }
396: \label{fig:contours}
397: \end{figure}
398:
399:
400: \begin{figure}
401: \centering
402: %emulateapj%
403: \includegraphics[width=0.71\columnwidth,clip=true]{f4.eps}
404: %aastex
405: %\epsscale{0.8}
406: %\plotone{mkfig.ps}
407: \caption{Same as Fig.~\ref{fig:contours} for the
408: bright region spine knots (left contours)
409: and sheath knots (right contours): rows 17 \& 18 of
410: Table~\ref{tab:specresults}.
411: }
412: \label{fig:contours2}
413: \end{figure}
414:
415: Having established that the knots in the sheath have a steeper
416: spectrum than in the spine, we checked the inner and outer regions to
417: see if there is any trend with distance from the core. Comparing rows
418: 9 and 10 with 11 and 12, we see the expected preference for a higher
419: $N_{\rm H}$ in the inner regions, but more interestingly we see that
420: it is only for the knots in the inner regions that the spine and
421: sheath spectra differ (Fig.~\ref{fig:contours}). The diffuse
422: emission, in contrast, shows remarkable spectral agreement between the
423: sheath and spine in the inner region (rows 13 \& 14). To check that
424: the contour separation in Figure~\ref{fig:contours} is causal rather
425: than statistical we have drawn random sets of the required size from
426: the modelled underlying spectral-index distribution deduced from all
427: knots in the inner region. The average spectral index has an error on
428: the mean of $\pm 0.07$ for knots matching in number those in the inner
429: spine, and $\pm 0.11$ for the inner sheath. The observed
430: spectral-index difference, $\Delta\alpha = 0.31$
431: (Fig.~\ref{fig:contours}), therefore significantly exceeds the
432: difference expected from random subsamples of the distribution.
433:
434:
435: Since the spine inner jet knots tend to lie in regions of higher
436: $N_{\rm H}$ than the sheath inner knots (Figs~\ref{fig:jetregions} and
437: \ref{fig:contours}), we have repeated the analysis only for the
438: ``bright'' region, marked in blue on Figure~\ref{fig:jetregions}.
439: This region is beyond the main dust lanes that cross the jet
440: \citep[e.g., fig.~1 of][]{hard07}. Results are consistent with sheath
441: knots having steeper spectra than spine knots
442: (Fig.~\ref{fig:contours2}, and rows 17 \& 18). Simulations show that
443: the bias (\S 2) on the absolute value of mean spectral index from the
444: $N_{\rm H}$ distribution over the inner jet is $\sim 0.05$.
445:
446:
447: We investigated the possibility that the spectral flattening in the
448: inner spine might result from the inner knots being brighter, if
449: brightness is associated with particle-acceleration efficiency.
450: Figure~\ref{fig:inknotprofile} shows the knot surface brightness and
451: hardness across the inner jet. The bin-width is 1 degree in position
452: angle, which corresponds roughly to the HPD at an average jet-axis
453: distance, and so the points are not highly correlated. The hardness
454: ratio shows the decrease with increasing distance from the jet axis
455: that is anticipated from our spectral results, with a high degree of
456: confidence that the hardness is not constant. The gradual decrease of
457: hardness with distance from the jet axis shows that our finding of a
458: spectral difference in the spine and sheath regions was not dictated
459: by the {\it a priori\/} definition of the region boundaries.
460:
461:
462: \begin{figure}
463: \centering
464: %emulateapj%
465: \includegraphics[width=1.0\columnwidth,clip=true]{f5.eps}
466: %aastex
467: %\plotone{profile4-nocomments.ps}
468: \caption{Surface-brightness profiles with position angle
469: of the knot emission from the inner jet in a hard energy
470: band (upper panels) with the hardness ratio below. The X-ray
471: emission is softer
472: towards the jet edges
473: ($\chi^2 = 116$ for 11 dof for a fit to constant hardness). There is
474: no trend for regions of higher integrated knot surface brightness to
475: show harder spectra.
476: }
477: \label{fig:inknotprofile}
478: \end{figure}
479:
480: Perhaps surprisingly, a comparison of the panels of
481: Figure~\ref{fig:inknotprofile} shows that there is no trend for
482: regions of higher integrated knot surface brightness to be harder in
483: spectrum. We have made plots similar to
484: Figure~\ref{fig:inknotprofile} for the inner and outer diffuse
485: emission, and the outer knot emission. The surface-brightness
486: profiles of diffuse emission show some lumpiness, suggesting that some
487: of what we call diffuse emission is composed of
488: lower-surface-brightness knots, but all three hardness-ratio profiles
489: are consistent with being constant across the jet ($\chi^2$ values of
490: 10.5, 7.9, and 14.8 for 11 degrees of freedom (dof), as compared with
491: $\chi^2$ =116 in Fig.~\ref{fig:inknotprofile}).
492:
493:
494: \section{Conclusions}
495: \label{sec:conclude}
496:
497: If the knots in \source\ were predominantly a surface feature in a
498: shear layer between the spine and sheath, as we suggested might be the
499: case for \ntof, one would expect the knot X-ray spectra to be similar
500: across the jet. This is not what is seen. In the inner jet, between
501: $21''$ and $66''$ from the core, we have observed in a statistical
502: sense that the X-ray knot spectra steepen farther from the jet axis.
503: No such trend is seen in the more diffuse X-ray emission. The results
504: suggest that knots are distributed throughout the jet volume, with a
505: trend towards steeper spectra in the outer regions. The trend will
506: then be underestimated in our results because of projection effects.
507:
508: Although the spectral trends disfavor knot confinement in a shear
509: layer, they imply that the electron energy distributions in knots
510: towards the center and edge of the jet are different. Electrons with
511: sufficient energy to emit synchrotron X-rays must be accelerated
512: locally to the knots, since their energy loss timescale is short
513: compared with any dynamical timescale associated with fluid motion.
514: Thus the conditions for particle acceleration differ across the
515: jet. The knots might arise from intruders from outside travelling at
516: speeds close to the stellar velocity dispersion, or internal flow
517: irregularities which are most likely to arise in the central part of
518: the flow and might be expected to move quickly towards the edge of the
519: jet under the influence of circulation in the shear layer. In either
520: case a transverse velocity structure might cause the strength of
521: shocks near the knots to vary across the jet, and since the kinetic
522: energy density of the flow should be higher nearer the jet axis, we
523: might also expect velocity irregularities to generate stronger
524: turbulent cascades near the axis than near the jet edge, with
525: consequent efficient particle acceleration \citep{manolakou} and a
526: flatter on-axis X-ray spectrum.
527:
528: As the jet expands, the lower on-axis flow velocity, as suggested by
529: the less collimated appearance of the jet further from the core, would
530: then imply less X-ray spectral difference between sheath and spine,
531: consistent with our results. Indeed, a spectral steepening between
532: the inner and outer regions in diffuse X-ray emission may relate to
533: such a decline in flow speed along the jet (rows 13 to 16 of
534: Table~\ref{tab:specresults}).
535:
536: It is interesting to ask whether it will be possible to detect knot
537: migration in \source. Expansion of the jet along its axis, with
538: apparent motions of about $0.5 c$, has been measured in radio data
539: from the Very Large Array over 10 years \citep{hard03b}, but lateral
540: motion is more difficult to measure. It has been detected in the
541: radio in M\,87 \citep{biretta95}. We speculate that transverse
542: motions should be present in the knots and are potentially detectable
543: if knots are flow irregularities, although future observations may be
544: required.
545:
546: These new observations of \source\ have permitted the first X-ray
547: spectral study of a radio jet in the transverse direction, separating
548: knotty and more diffuse emission. We have found the striking result
549: that where the X-ray emission is brightest, at $\sim 0.4$ to 1.2 kpc
550: from the core, the spectrum in the knotty emission is steeper at
551: larger distances from the jet axis. We speculate that this is the
552: result of lateral knot migration.
553: \smallskip
554:
555: %\acknowledgments
556:
557: This work was partially supported by NASA grant GO7-8105X.
558: We thank the referee for constructive comments and MJH
559: thanks the Royal Society for support.
560: Facilities: \facility{CXO}
561:
562: \begin{thebibliography}{}
563: \bibitem[Biretta et al.(1995)]{biretta95} Biretta, J.A., Zhou, F., \&
564: Owen, F.N.~1995, \apj, 447, 582
565: \bibitem[Canvin et al.(2005)]{canvin05} Canvin, J.R., Laing, R.A.,
566: Bridle, A.H., \& Cotton, W.D.~2005, \mnras, 363, 1223
567: \bibitem[Chiaberge et al.(2000)]{chiaberge00} Chiaberge, M., Celotti,
568: A., Capetti, A., \& Ghisellini, G.~2000, \aap, 358, 104
569: \bibitem[Dickey \& Lockman(1990)]{dlock90} Dickey,
570: J.M., \& Lockman, F.J.~1990, ARA\&A, 28, 215
571: \bibitem[Ferrarese et al.(2007)]{ferrarese07} Ferrarese, L., Mould,
572: J.R., Stetson, P.B., Tonry, J.L., Blakeslee, J.P., Ajhar, E.A.~2007,
573: \apj, 654, 186
574: \bibitem[Hardcastle et al.(2003a)]{hard03a} Hardcastle M.J., Worrall,
575: D.M., Birkinshaw, M., \& Canosa,
576: C.M.~2003a, \mnras, 338, 176
577: \bibitem[Hardcastle et al.(2003b)]{hard03b} Hardcastle M.J., Worrall,
578: D.M., Kraft R.P., Forman, W.R., Jones, C., \& Murray,
579: S.S.~2003b, \apj, 593, 169
580: \bibitem[Hardcastle et al.(2007)]{hard07} Hardcastle, M.J.~et
581: al.~2007, \apj, 670, L81
582: \bibitem[Jord\'an et al.(2007)]{jordan07} Jord\'an, A.~et al.~2007, \apjl,
583: in press (arXiv/0711.2303)
584: \bibitem[Kataoka et al.(2006)]{kataoka} Kataoka, J., Stawarz, L.,
585: Aharonian, F., Takahara, F., Ostrowski, M. \& Edwards, P.G.~2006,
586: \apj, 641, 158
587: \bibitem[Kraft et al.(2002)]{kraft02} Kraft, R.P., Forman, W.R.,
588: Jones, C., Murray, S.S., Hardcastle, M.J. \& Worrall,
589: D.M.~2002, \apj, 569, 54
590: \bibitem[Laing(1996)]{laing96} Laing, R.A.~1996,
591: in ASP Conf.~Ser.~100, Energy transport in radio galaxies and quasars,
592: ed.~P.E.~Hardee,A.H.~Bridle \& J.A.~Zensus (San Francisco: ASP), 241
593: \bibitem[Laing \& Bridle(2002)]{laing02} Laing, R.A., \&
594: Bridle, A.H.~2002, \mnras, 336, 1161
595: \bibitem[Manolakou et al.(1999)]{manolakou} Manolakou, K.,
596: Anastasiadis, A., \& Vlahos, L.~1999, \aa, 345, 653
597: \bibitem[Schwartz et al.(2006)]{schwartz06} Schwartz, D.A.~et al.~2006, \apj,
598: 640, 592
599: \bibitem[Worrall et al.(2001)]{worrall01} Worrall
600: D.M., Birkinshaw, M. \& Hardcastle, M.J.~2001,
601: \mnras, 326, L7
602: \bibitem[Worrall et al.(2007)]{worrall07} Worrall
603: D.M., Birkinshaw, M., Laing, R.A., Cotton, W.D., Bridle, A.H.~2007,
604: \mnras, 380, 2
605: \end{thebibliography}
606:
607: \end{document}
608:
609: