1: \documentclass[12pt,preprint]{aastex}
2:
3: \begin{document}
4:
5: \title{\bf CH Cyg X-ray Jet Activity and Multi-component Structures}
6:
7: \author{Margarita Karovska\altaffilmark{1}}
8: \author{Christopher L. Carilli\altaffilmark{2}}
9: \author{John C. Raymond\altaffilmark{1}}
10: \author{Janet A. Mattei\altaffilmark{3}}
11: \altaffiltext{1}{Smithsonian Astrophysical Observatory, 60 Garden St.,
12: Cambridge, MA 02138; mkarovska@cfa.harvard.edu.}
13: \altaffiltext{2}{National
14: Radio Astronomy Observatory, PO Box O, Socorro, NM 87801}
15: \altaffiltext{3}{American Association of Variable Star Observers, 25 Birch
16: Street, Cambridge, MA 02138}
17:
18: \begin{abstract}
19:
20: In this paper we report detection of multiple component structures
21: in a Chandra
22: X-ray image obtained in March 2001 of the nearby symbiotic interacting
23: binary system CH Cyg.
24: These components
25: include a compact central
26: source, an arc-like
27: structure or a loop extending to 1.5'' (400 AU) from the central
28: source associated with
29: the 1997 jet activity,
30: and possibly a
31: newly formed jet extending to $\sim$ 150 AU from the central source.
32: The structures are also visible in
33: VLA and HST images obtained close in time to the Chandra
34: observations.
35: The emission from the loop
36: is consistent with optically thin thermal X-ray
37: emission originating from a shock resulting from interaction of the jet ejecta with
38: the dense circumbinary material.
39: The emission from the central source originates within $<$ 50 AU region, and is likely associated with the accretion disk around the
40: white dwarf.
41: CH Cyg is only the second symbiotic system
42: with jet activity detected at X-ray wavelengths, and the
43: Chandra high-angular resolution image,
44: combined with the VLA and HST images,
45: provides the closest view of the region of jet
46: formation and interaction with the circumbinary material
47: in a symbiotic binary.
48:
49:
50: \end{abstract}
51:
52: \keywords{accretion, accretion disks --- binaries: close --- stars:
53: individual (CH Cyg) --- binaries: symbiotic, stars: winds, outflows --- x-rays: general}
54:
55: \section{INTRODUCTION}
56:
57:
58: Symbiotic systems are interacting binaries
59: showing a composite spectrum with signatures of
60: a late-type giant and a high-temperature component, often a compact
61: accreting object. As with many other astronomical sources, accretion is thought to be the
62: primary energy source in these systems.
63: The interaction between the components occurs via wind accretion
64: and/or Roche lobe overflow.
65: Symbiotic systems are a very important subgroup of interacting
66: binaries because they are among the likely candidates for progenitors
67: of bipolar planetary nebulae, and they have also been invoked as potential
68: progenitors of at least a fraction of Supernovae type Ia, key
69: cosmological distance indicators (e.g. Chugai and Yungelson 2004).
70: So far there are only a couple hundred known members in this class.
71:
72:
73: CH Cyg is one of few symbiotic systems
74: showing jet activity (e.g. Corradi {\it et al.} 2001, Kellogg {\it et al.} 2001).
75: At a distance of 245$^{+}_{-}$50 pc (Perryman {\it et al.} 1997)
76: CH Cyg is one of the closest symbiotic systems, therefore providing a
77: unique opportunity for
78: high-angular resolution imaging of a close circumbinary environment in
79: an interacting binary,
80: including the region of jet formation and early propagation.
81:
82:
83: The CH Cyg system is composed of an evolved M6-7 III giant and
84: an accreting white dwarf.
85: There is also evidence for a third
86: body in the system, also an evolved giant
87: (e.g. Hinkle {\it et al.} 1993, Scopal {\it et al.} 2002).
88: This binary is one of
89: the most fascinating objects in its class because of its extremely complex
90: circumbinary environment, and its dramatic outbursts.
91:
92:
93:
94: In the past few decades CH Cyg has undergone several outbursts,
95: each preceded by extended intervals of quiescence (Karovska and
96: Mattei 1992, Sokoloski and Kenyon 2003,
97: Eyres {\it et al.} 2002, Skopal {\it et al.} 2002).
98: %The jet was preceded by a sharp drop in the V magnitude.
99: %A similar event was observed in 1996-1997 (Karovska {\it et al.} 1998)
100: %Since 1984 several radio outburst have been detected and the
101: Since 1984, when a powerful jet was formed in the midst of a prominent radio and optical
102: outburst (Taylor {\it et al.} 1986), several outbursts have been detected including
103: in 1996 and in 1998 (Karovska {\it et al.}
104: 1998a\&b, Eyres {\it et al.}
105: 2002).
106: Radio monitoring of the system has shown indications of a
107: possible precessing jet with enhanced activity associated with occasional
108: outbursts (Crocker {\it et al.} 2002).
109:
110:
111:
112: Observations of CH Cyg have been carried out over the past
113: two decades using several X-ray satellites.
114: Spectral modeling indicates a multi-component
115: structure in the emission from this system
116: (Leahy and Taylor
117: 1987; Leahy and Volk 1995; Ezuka {\it et al.} 1998; Mukai {\it et
118: al.} 2006).
119: For example, Ezuka {\it et al.} (1998) detected a soft component in
120: the ASCA spectrum with
121: flux of 3.7$\times$ 10$^{-12}$ $ergs$ $cm^{-2}$$s^{-1}$ from 0.5 to 2.5
122: keV. They modeled this emission as
123: two optically thin thermal plasma emissions, the temperatures of which are 0.2 and
124: 0.7 keV. However, the origin of the soft component remained unclear.
125: A previously unknown hard component
126: was also detected in the spectrum with flux of 6.9$\times$ 10$^{-12}$ $ergs$ $cm^{-2}$$s^{-1}$ in the region from 2-10 keV.
127: This component was modeled with optically thin thermal plasma emission with temperature of 7.3 keV,
128: undergoing a strong photoelectric absorption with N$_H$ =$\sim$ 10$^{23}$ $cm^{-2}$.
129: In addition, several emission features
130: were detected including
131: iron emission lines between 6 and 7 keV.
132: Wheatley and Kallman (2006) re-analyzed the ASCA
133: spectrum and concluded that there is only one emission
134: component. They suggest that similar to Seyfert 2 galaxies,
135: the soft emission is due to scattering of
136: the hard component in a photoionized medium surrounding the white
137: dwarf. Mukai {\it et al.} (2006) analyzed recent Suzaku
138: observations and found that the large equivalent width of the
139: 6.4 keV fluorescence line implies that the hard X-rays are
140: indeed dominated by scattering. However, they concluded that there are two distinct
141: components responsible for the hard and soft X-ray emission.
142: They also confirmed that there is a significant long-term variability
143: in the X-ray emission in the system.
144:
145: Chandra observations carried out in 2001
146: detected an elongation in the CH Cyg image, extending 2-4'' to
147: the south from the central source. Galloway and Sokoloski (2004)
148: attributed this extension to jet activity associated
149: with the white dwarf. This is only
150: the second source from a population of few hundred known
151: symbiotics showing jet activity in X-rays.
152: However,
153: in the Galloway and Sokoloski (2004) analysis of the Chandra image the potential spatial resolution of $<$0.5''
154: on-axis
155: was significantly degraded by gaussian smoothing used
156: to enhance the faint extended emission.
157: Therefore, the resolution was
158: insufficient to study directly the spatial distribution of the
159: X-ray emission.
160:
161:
162:
163: In this paper we present results from our further detailed spatial
164: analysis of the Chandra observation of CH Cyg
165: combined with the results from
166: VLA and HST imaging of the system. Using the
167: full angular resolution potential of Chandra we identify several
168: distinct spatial components contributing to the X-ray emission and
169: resolve the ambiguity
170: left from previous X-ray observations regarding the
171: extended emission and the
172: contributors to the soft and the hard component of the X-ray
173: spectrum of CH Cyg.
174:
175:
176:
177: \section{CHANDRA OBSERVATIONS AND ANALYSIS}
178:
179: The Chandra pointed observations of CH Cyg were carried out over
180: 47 ks on 2001 March 27 (OBSID 1904) using the ACIS-S instrument in
181: combination with HETG grating (Weisskopf {\it et al.} 2002).
182: Significant signal was detected mostly in the
183: zeroth-order of the spectrum;
184: a total of 770 counts were detected form 0.2 to 8 keV.
185: We analyzed the {\it Chandra} observations using {\it CIAO} data
186: reduction and analysis routines. In our analyses we used Ciao 3 and cal database
187: 2.2 (similar to
188: Galloway and Sokoloski 2004), as well as the more recent
189: Ciao 3.3 version.
190: \footnote{CIAO is the Chandra Interactive Analysis of Observation's
191: data analyses system package (http://cxc.harvard.edu/ciao)}
192:
193:
194: {\it Chandra} data include information about the photon energies and
195: positions. This information was used to obtain energy-filtered images
196: and to carry out
197: sub-pixel resolution spatial analysis ({\it vis} Karovska {\it et
198: al.} 2005).
199: Based on the three spectral
200: emission components identified by Ezuka {\it et al.} (1998) and Galloway and
201: Sokoloski (2004)
202: (see for example Fig. 3 in Galloway and Sokoloski 2004)
203: we created images in the soft
204: band (0.2-2keV), the
205: hard band (2-6 kev), and in the spectral region around the strong
206: emission feature at 6.5-7 attributed to 6.67 keV iron line.
207: The total counts detected in these images using a circular region with
208: a radius of 5'' centered on the brightest central source are: 290,
209: 260, and 100 counts, respectively.
210: The expected X-ray background in the region is negligible.
211:
212:
213: We explored the spatial
214: distribution of the emission in the images of the system at 0.2''
215: resolution using deconvolution and smoothing
216: techniques.
217: The pixel size of the Chandra/ACIS detector is 0.492".
218: However, due to the telescope dither smaller
219: spatial scales are accessible as the image moves across the detector pixels.
220: This sampling of smaller than detector pixel scales allows sub-pixel
221: analysis and deconvolution.
222: Similar techniques were applied for resolving Mira AB 0.5" binary
223: system (Karovska {\it et al.} 2005),
224: as well as for SN1987A (e.g. Burrows {\it et al.} 2000, Park {\it et al.} 2002)
225: which allowed resolving sub-arcsecond structures
226: in the 1" scale SN remnant.
227: After processing of the data (including with pixel randomization turned off)
228: we binned the images with 0.1'' pixels.
229:
230: We used PSF simulations for comparison with the observations in order
231: to evaluate detected structures in
232: the observed images as well as to estimate the extent of the central source
233: in different spectral bands.
234: The PSF
235: simulations were carried out using information on the spectral
236: distribution and off-axis location of the system as inputs to ChaRT PSF
237: simulator\footnote{http://asc.harvard.edu/chart/index.html}
238: (Karovska {\it et al.} 2003).
239:
240:
241: To detect the low contrast diffuse emission we applied the adaptive
242: smoothing CIAO tool {\it csmooth}, based on the algorithm
243: developed by Ebeling {\it et al.} (2006) and implemented in {\it asmooth}.
244: {\it Csmooth} is an adaptive smoothing tool for
245: images containing multiscale complex structures,
246: and it preserves the spatial signatures
247: and the associated counts as well as significance estimates.
248: The smoothing is achieved by convolution with a
249: gaussian kernel. The size of the kernel increases from a small initial
250: value and the smoothing scale is increased
251: until the total number of counts under the kernel exceeds a
252: value that is determined from a preset significance and the
253: expected number of background counts in the kernel area.
254: Significance is computed at each pixel location
255: for each smoothing scale by
256: comparing the total counts under the kernel to the
257: background in the same area.
258: Pixels for which the significance exceeds
259: the threshold value set by the user are smoothed at the
260: current scale, and are excluded from subsequent smoothing
261: steps at larger scales.
262:
263:
264: In Figure 1 we display the smoothed images of CH Cyg in
265: the soft (0.2-2), and the
266: hard (2-6 keV) bands.
267: In these images we show structures smoothed above 3 $\sigma$
268: significance level, and the smoothing scales vary from 0.2 to 0.5 arcseconds.
269: A compact central source is detected in all three bands.
270: Our analysis
271: shows that the central sources in the hard band image and in the 6.5 to 7 keV image are consistent with unresolved
272: point sources (FWHM less then 0.2'').
273: In addition to a central source, several extended
274: structures are detected that are especially prominent in the soft
275: image.
276: For the central source,
277: within a circular region with 0.3'' radius,
278: we estimate a hardness ratio of $\sim$2,
279: (counts $>$2keV divided by the counts $<$ 2 keV).
280:
281:
282:
283: Figure 2 shows the smoothed soft image, as in Figure 1, with overplotted
284: contours of the simulated Chandra PSF.
285: When compared to the PSF this image shows a somewhat extended central
286: source to $\sim$ 0.7'' at P.A. $\sim$150/330 degrees, and an extended
287: emission to the south
288: consisting of at least two components:
289: a loop-like structure, and
290: a faint emission extending beyond the loop.
291: Figure 3 displays the soft
292: image after subtracting the PSF, showing the extended structure in
293: the central source at P.A. $\sim$ 330 degrees and the loop-like
294: structure to the south.
295:
296:
297: The loop-like structure
298: extends to
299: a distance of
300: $\sim$1.5'' from the central source.
301: The southern part of the loop appears brighter and resembles an arc.
302: There are regions of enhanced brightness (3-5 $\sigma$) in the
303: southern portion of this loop, smoothed on scales of 0.3 arcseconds.
304: We detect $\sim$40 counts in the region of the loop structure.
305: The emission in the loop region is fainter in the images above 2keV
306: and contains about half
307: of the counts in the soft image.
308: The low counts are insufficient to
309: extract detailed spectral information; however,
310: the loop clearly shows a soft X-ray
311: signature.
312: There is also a faint emission beyond the loop extending
313: to $\sim$4'' south of the central source.
314: This feature could be
315: associated with a diffuse emission with a low surface
316: brightness (total $\sim$20 counts).
317:
318: The multi-component structure in CH Cyg was not detected in
319: the Galloway and Sokoloski (2004) image because they applied
320: gaussian smoothing with
321: a FWHM of 2'' which
322: resulted in
323: degraded resolution.
324: In Figure 4 we compare our soft band high-angular resolution image
325: with
326: their result.
327: Although they did detect an extension of a few arcseconds from the central
328: source, their smoothed
329: image did not permit detection of the multi-component structures, including
330: the
331: loop-like structure
332: separated from the central source, which
333: are clearly seen in our adaptively smoothed images.
334:
335:
336:
337: We further explored the reality of these low-count structures
338: by applying a statistical deconvolution technique
339: EMC2 (Expectation
340: through Markov Chain Monte Carlo) specifically applicable to low-counts Poisson data (Esch {\it et al.} 2004).
341: This technique uses a wavelet-like multiscale representation of the true image
342: in order to achieve smoothing at all scales of resolution simultaneously.
343: Wavelet decomposition of the image allows the Poisson likelihood to be factored into
344: separate parts, corresponding
345: to the wavelet basis.
346: Each of these factors in the likelihood can be reparametrized as a split of the intensity
347: from the previous, coarser factor.
348: A {\it prior} is assigned to these splits (which can be viewed as smoothing parameters),
349: and then a model is fit using Markov Chain Monte Carlo (MCMC) methods.
350: In this way small and large-scale structures in the image are captured, e.g. diffuse emission
351: as well as point sources and sharp features in the image.
352: This Bayesian model-based analysis allows assessment of the uncertainties
353: in the reconstructed image.
354: Details of this technique and examples are described in Esch {\it et al.} 2004.
355: The EMC2 has been applied to various data sets including to Chandra observations
356: (e.g. Esch {\it et al.} 2004, Karovska {\it et al.} 2005).
357:
358:
359: We used EMC2 and the simulated PSF to perform deconvolution of
360: the CH Cyg soft X-ray image.
361: Figure 5 displays the EMC2 deconvolved image showing
362: complex multi-component structures
363: with significance above 3 $\sigma$. The structures are
364: similar to those
365: resulting from adaptive smoothing, but with improved resolution.
366: In addition to the loop-like structure to the south of the central source, the central source appears to have a compact core
367: and an elongated structure extending to $\sim$ 0.7'' at the P.A. of 330 deg.
368: A fainter structure is seen to the south at about same distance from
369: the source at P.A. of 150 deg.
370:
371:
372: The structures detected in the X-ray images of CH Cyg are very similar
373: to the structures detected in radio and optical observations obtained
374: close in
375: time to the Chandra observations as described in the following
376: section.
377:
378:
379: \section{COMPARISON WITH HST AND VLA OBSERVATIONS}
380:
381:
382:
383:
384: Optical/UV and radio images of CH Cyg
385: obtained since 1996
386: show a complex circumbinary environment including ionized outflow,
387: and structures associated with
388: jet or outburst activity in the system (e.g. Corradi {\it et al.} 2001;
389: Eyres {\it et al.} 2002).
390:
391:
392: Although there are no observations contemporaneous with the 2001 Chandra
393: observation, we identified a set of VLA and HST observations that were carried out close in
394: time to the Chandra observations.
395: Radio maps of CH Cyg were obtained using the VLA at 5GHz and 8GHz on
396: 2000 November 9 about 4.5 months before the Chandra observation.
397: HST observations were carried out
398: on 1999 August 13 (in the optical, with WFPC2 ), 1.6 years
399: before the Chandra observations.
400:
401:
402: The loops detected in the soft X-ray image are
403: similar to the arc-like extensions
404: observed in the HST WFPC2 images (Eyres {\it et al.} 2002, and Crocker
405: {\it et al.} 2002).
406: The HST/WFPC2 images obtained
407: 1.6 years before the Chandra observations
408: are dominated by the
409: UV continuum, [OIII], H~$\alpha$ and H~$\beta$ lines. They
410: show extended emission
411: to the north and to the south of the central region.
412: Several structures, including an arc-like structure 1.2''
413: to the south
414: of the
415: central source, have been detected in
416: this extended emission (e.g. Eyres {\it et al.} 2002).
417: The loop-like structure is most clearly visible in the 502N [OIII]
418: image.
419:
420: In Figure 5 we compare the deconvolved Chandra
421: soft X-ray image with the HST [OIII]
422: image which has a similar resolution.
423:
424: We note that the HST observation is the same as shown in Figure 1
425: in Galloway and
426: Sokoloski (2004).
427: The extended [OIII] structures show
428: three
429: distinct components: a central source, $\sim$0.5'' extensions to the north
430: well as at $\sim$ PA160 deg to the south, and
431: a loop-like feature to the south at 1.2''.
432: The [OIII] loop-like feature fits
433: within the loop in the Chandra soft image, as shown in Figure 5.
434: The south side of the [OIII] loop appears brighter as
435: seen in the X-ray image.
436:
437:
438:
439: The VLA map shown in Figure 6 was made
440: from data taken on November 9, 2000 (VLA in the ``A'' configuration).
441: The resolution is
442: $0.55" \times 0.39"$ with major axis position angle = $-82^o$, and
443: the
444: rms = 35 $\mu$Jy beam$^{-1}$. This map shows a very similar
445: morphology when compared to the soft band Chandra image; it shows a
446: central structure extended at P.A.$\sim$ 330 degrees, as well a loop like structure
447: $\sim$1.5'' to the south. The total flux density from the source is
448: 7.2 mJy, with a peak surface brightness of 2.4 mJy beam$^{-1}$ at
449: 19$^h$ 24$^m$ 33.03$^s$ +50$^o$ 14$'$ 27.1$''$ (J2000). The loop
450: structure shows enhancement in the south side, similar to that of
451: the X-ray loop. The southern loop has an integrated radio flux of 1.6
452: mJy.
453: The core source has an inverted spectrum of index $\alpha = +0.6$,
454: between 5 and 8 GHz, suggesting optically thick emission.
455: We define $\alpha$ as $S_\nu \propto \nu^{\alpha}$.
456: The northern
457: extension has flat spectrum, $\alpha \sim 0$, suggesting optically thin
458: free-free emission. The southern loop has a steep spectrum, $\alpha \sim
459: -0.8$, implying optically thin non-thermal emission.
460: The
461: minimum energy magnetic field in the southern loop based on the
462: non-thermal radio surface brightness is about 1mG, with a minimum
463: pressure of $6\times 10^{-8}$ dynes cm$^{-2}$.
464:
465:
466: In Figure 6 we display an overlay of the radio contours on
467: the smoothed soft X-ray
468: image,
469: which has a similar resolution. The radio and the X-ray loops are
470: cospatial within the resolution limit of the VLA data of $\sim$0.4''.
471: We note that the resolution in the X-ray and radio images
472: is insufficient to determine if the regions of X-ray and radio emission
473: are separated.
474:
475:
476:
477:
478:
479: \section{DISCUSSION}
480:
481:
482:
483: Our analysis of the Chandra data shows that there are several
484: spatially distinct
485: X-ray emission components in the CH Cyg system.
486: The X-ray emission
487: in the CH Cyg image
488: is dominated by an unresolved central core, however, there are several
489: extended components that contribute as well.
490: The emission from the central core
491: originates within a region of less then 50 AU,
492: which is only few times larger
493: than the binary separation.
494: This is as close as we can get
495: to the central region of an interacting binary
496: with today's X-ray imaging capabilities.
497:
498: The hard emission from the central core, including in the iron line,
499: probably originates in
500: the inner region of the
501: accretion disk or in a boundary layer between the accretion disk and the
502: white dwarf.
503: The soft emission from the unresolved core may be attributed to the
504: accretion disk itself or to scattered hard X-rays from
505: the inner disk region by the outer disk or corona. The scattering
506: could be due to a non isotropically distributed
507: ionized photoionized medium similar to Seyfert 2 galaxies
508: (e.g. Wheatley and Kallman 2006).
509: The X-ray emission in CH Cyg core appears harder than in other symbiotics,
510: eg. in R Aqr and Mira AB (Kellogg et al 2001, Karovska
511: et al 2005), which have
512: relatively soft spectra, close to 1 kev.
513: Geometry of the surrounding material may play a role since part
514: of the hard X-rays in these
515: systems may be absorbed or scattered
516: in a different way.
517:
518:
519: In addition to the unresolved core, the soft X-ray image shows
520: a second component in the central region, an extended
521: structure to $\sim$ 150 AU at 330/150 degrees.
522: This structure was also
523: detected in the radio, and HST O[III] images.
524: The emission could be due to a
525: shocked gas formed in the region of collision between the
526: fast wind from a white
527: dwarf with the red giant wind.
528: However, since the orientation of this structure
529: is similar to that of radio and optical
530: jets observed in
531: the past, and to the general orientation of the loop
532: structures further to the south (e.g. Eyres {\it et al.} 2002, Corradi {\it et al.} 2001),
533: we believe that the structure is more likely
534: associated with a recent jet activity, e.g., during
535: the 1998-2000 active phase.
536: The emission could
537: result from jet interaction with the close
538: circumbinary material including circumstellar dust shells surrounding
539: the giant.
540: In fact, Mid-IR photometry does show the presence of a significant dust in the vicinity of the central region of CH Cyg
541: (Teranova and Shenavrin 2004). Furthermore, high-angular resolution Mid-IR
542: imaging of CH Cyg
543: detected a dusty
544: circumstellar feature $\sim$0.7'' to the North of central source,
545: similar to the extension of the feature detected in Chandra, Radio and
546: HST images (Biller {\it et al.} 2006). Follow up observations of this
547: extended X-ray structure are needed to
548: determine its nature and kinematics.
549:
550:
551:
552:
553: Our imaging shows that although the X-ray emission from
554: the
555: central region
556: dominates the soft emission in CH Cyg,
557: a significant fraction ($\sim$20$\%$)
558: originates further away,
559: from the loop at $\sim$ 400 AU and from the extended diffuse emission beyond.
560: The loop
561: is likely associated with the 1997 radio outburst which was followed
562: by jet activity (Karovska {\it et al.}
563: 1998a, Karovska {\it et al.}
564: 1998b).
565: The 1997 radio outburst in CH Cyg was detected using the VLA at several frequencies (Karovska et al
566: 1998a).
567: Nearly contemporaneous
568: continuum measurements at millimeter and submillimeter wavelengths
569: with the IRAM
570: interferometer and with SCUBA on the James Clerk Maxwell Telescope
571: also showed evidence for a strong radio outburst
572: (Mikolajewska {\it et al.} 1998).
573: The 1997 radio
574: outburst occurred following
575: an unprecedented drop in the V magnitude in 1996 (Karovska et al
576: 1998a), similar to a
577: previous event in 1984,
578: following which a powerful jet was formed (Taylor {\it
579: et al.} 1986).
580: In early 1998 we detected a radio jet
581: using the VLA in its highest resolution observing mode (Karovska {\it
582: et al.} 1998b).
583: At that time, the radio maps detected two structures with a roughly
584: North-South orientation, with the southern structure
585: separated from the central source by
586: about 0.35''.
587: The VLA observations of the loop carried out later on in 2000
588: show that the structure has
589: expanded to a distance close to 1.5'', or $\sim$ 375AU.
590: This corresponds to an expansion velocity of $\sim$ 400 km/s.
591:
592:
593: Assuming that the loops we detected in the X-rays and radio, and in [OIII], are a
594: consequence of the same outburst event,
595: the [OIII] loop would have had to expand at a velocity of
596: several hundred km/s to reach the position seen in the Chandra
597: images 1.6 years later.
598: It would also have taken a few years for ejecta from the central region with
599: an expansion velocity of few
600: hundred km/s to reach the
601: apex of the [OIII] loop, and then to expand to the region
602: of the radio and X-ray loops (assuming that the ejecta
603: have not been slowed down since the outburst).
604: This is
605: in agreement with the apparent expansion velocity derived from the
606: 1998 VLA observations.
607:
608:
609:
610: Kinematic studies of outflows in CH Cyg do show velocities of the expanding
611: material on the order of few hundred km/s (Corradi {\it et al.} 2001)
612: which is in agreement with the velocity of expansion that we calculated.
613: On the other hand, spectroscopic observations also show evidence for
614: mass
615: ejections in CH
616: Cyg that indicate
617: much higher velocities, well over 1000 km/s (e.g. Skopal {\it et al.} 2002).
618: Skopal {\it et al.} (2002) suggested that this difference could be due to a
619: significant deceleration of the ejecta due to interaction with the
620: circumstellar medium, which would be also in agreement with our observations.
621: On the other hand, if the ejecta have not
622: been slowed down and are moving with a speed over 1000 km/s, there remains a
623: possibility that the X-ray and radio emission in the region of the
624: loop may be a consequence of an outburst during the more recent 1998-2000 active phase.
625:
626:
627:
628: The X-ray emission from the loop is consistent
629: with a shock generated by ejecta
630: colliding with the
631: surrounding pre-existing circumstellar material.
632: The X-ray emission beyond the loop could be associated with
633: continuous interaction of
634: the expanding wind with the pre-existing circumbinary material,
635: or it could result from more intermittent shocks associated with
636: a pulsed jet (e.g., Stute 2006). The non-thermal spectral index of the
637: radio emission could be explained by particle acceleration at the
638: shock.
639:
640: The small number of counts in the loop do not permit us to fit for
641: the temperature and absorbing column of the emitting gas. If we make
642: the most favorable assumptions given the 60 counts detected and
643: the hardness ratio, the temperature is about 1 keV and the luminosity
644: is about $5 \times 10^{28}~\rm erg~s^{-1}$. For a 400 AU sphere
645: and an X-ray emissivity of $2 \times 10^{-23}~\rm erg~cm^3~s^{-1}$
646: (e.g. Raymond, Cox \& Smith 1976), the density is about 50 $\rm cm^{-3}$.
647: This estimate is probably on the low side, given that the spherical
648: volume is not filled and that absorption will reduce the observed
649: count rate. Nevertheless, it implies a pressure about 2.5 times the
650: minimum magnetic pressure estimated from the radio emission. The mass of the
651: X-ray emitting gas is about $10^{26}$ g, or 0.1, 0.05 and 1.0 times the
652: masses of jet components A, B and C measured by Taylor et al. (1986),
653: respectively. Given that the velocity ellipses measured by Corradi
654: et al. (2001) suggest shock speeds of about 100 km/s, while the
655: X-rays require shock speeds near 1000 km/s, the simplest interpretation
656: is that the jet is underdense by a factor of 100 compared to the
657: ambient gas. The slower shock in the ambient gas produces Balmer line,
658: [O II] and [O III] emission by a combination of shock heating and
659: photoionization of the compressed gas. Because the gas cools quickly
660: the compression is very strong. The fast shock produces X-rays,
661: but a shock speed closer to 2000 $\rm km~s^{-1}$ than 1000 $\rm km~s^{-1}$
662: may be required, since electron temperatures are generally well below
663: proton temperatures in such fast shocks (e.g. Rakowski 2006).
664: The radio emission could come from either shock, but the
665: strong magnetic field suggests that it originates in gas that
666: has been strongly compressed by the slower shock. If the first shock
667: generates the radio emission, a significant magnetic field, on the
668: order of 25 $\mu$G, would be required in the jet.
669:
670: A high speed (precessing) jet may also clear a cavity in the pre-existing
671: circumstellar material (e.g. Ybarra {\it et al.} 2006). The X-ray and the radio emission in that case
672: originate from a loop-like structure delineating a shock
673: region in the boundary between the walls of the cavity and the
674: surrounding material.
675: A similar scenario has been suggested for the loop-like
676: structure observed previously further away ($\sim$5'') from the central source
677: in the UV O[II] HST images of CH Cyg(Corradi {\it et al.} 2001).
678:
679:
680:
681:
682:
683:
684:
685: \section{CONCLUSIONS}
686:
687: Our study of the CH Cyg close circumbinary environment using a
688: combination of multi-wavelength
689: X-ray, radio and optical high-angular resolution observations
690: identified several distinct components in this system. These
691: include a compact central
692: source, and several extended structures likely associated with jet
693: activity.
694: These results are important since
695: CH Cyg is only the second symbiotic system after R Aqr (Kellogg {\it et
696: al.} 2001)
697: with jet activity detected at X-ray wavelengths, and
698: the Chandra image combined with the VLA and HST data
699: provides the closest view of
700: the region where jets are being formed and interacting with the
701: surrounding material
702: as they propagate
703: through the circumbinary environment.
704: Furthermore, our analysis of the Chandra data resolves the ambiguity as to whether there is only one
705: (Wheatley and Kallman 2006), or multiple components (e.g. Ezuka {\it et
706: al. 1998}, Mukai {\it et al.} 2006) contributing to
707: the X-ray emission from this system.
708:
709:
710: Continuing multiwavelength monitoring of CH Cyg is necessary
711: to determine the characteristics of the multi-component
712: structures and their evolution as they
713: interact with the surrounding
714: circumbinary environment. It is especially important to obtain longer
715: exposure X-ray observations of this system which will increase the
716: signal-to-noise in the images and allow deriving
717: spectral
718: characteristics of the emissions from individual components.
719:
720:
721: \acknowledgments
722:
723:
724: Janet Mattei passed away before the completion of this article, and her coauthors respectfully
725: dedicate it to her memory.
726: We are grateful to W. Hack for providing the drizzled HST image of
727: CH Cyg, and to AAVSO for providing the CH Cyg light curve.
728: We thank Y. Butt for useful discussions.
729: We would like to thank the anonymous referee
730: for helpful comments and suggestions.
731: MK is a member of the Chandra X-ray Center,
732: which is operated by the Smithsonian Astrophysical Observatory under
733: NASA Contract NAS8-03060. CC would like to acknowledge support from
734: the Max-Planck Society and the Alexander von Humboldt Foundation
735: through the Max-Planck-Forschungspreis 2005.
736:
737: \begin{thebibliography}{}
738:
739: \bibitem{Biller} Biller {\it et al.} 2006, astro-ph 0604336
740:
741: \bibitem{Burrows2000} Burrows, D.N. {\it etb al.} 2000, ApJ, 543,149
742:
743:
744: \bibitem[\protect\citeauthoryear{Chugai and Yungelson}{2004}]{Chugai2004}
745: Chugai, N.N., and Yungelson, L.R.,, 2004, Astron. Lett., 30, 65.
746:
747: \bibitem[\protect\citeauthoryear{Corradi {\it et al.}}{2001}]{Corradi2001}
748: Corradi R.L.M. et al., 2001, ApJ, 560, 912.
749:
750: \bibitem{Crocker2002} Crocker, M.M., Davis, R.J., Spencer, R.E.,
751: Eyres, S.P.S., Bode, M.F., Skopal, A., 2002, MNRAS, 335, 1100
752:
753: \bibitem{Ebeling 2006} Ebeling, H.; White, D. A.; Rangarajan,
754: F. V. N. 2006, MNRAS, 368, 65
755:
756: \bibitem{Eyres2002} Eyres, S. P. S., Bode, M. F., Skopal, A., Crocker,
757: M. M.; Davis, R. J., Taylor, A. R., Teodorani, M., Errico, L.,
758: Vittone, A. A., Elkin, V. G. 2002, MNRAS, 335, 526
759:
760: \bibitem{Esch2004} Esch, D.N., Connors, A., Karovska, M., van Dyk,
761: D.A., 2004, ApJ, 610, 1213
762:
763: \bibitem{Ezuka 1998} Ezuka {\it et al.} 1998, ApJ, 499, 388
764:
765: \bibitem{GallowaySokoloski2004} Galloway, D.K. Sokoloski, J.L. 2004,
766: ApJ Letters, 613, 61
767:
768: \bibitem{Hinkle1993} Hinkle, K.H. et al 1993, AJ, 105, 1074.
769:
770: \bibitem{Karovska1998} Karovska, M., Carilli, C., and Mattei, J.,
771: 1998a,
772: IAU Circ. No. 6970.
773:
774: \bibitem{Karovska1998} Karovska, M., Carilli, C., and Mattei, J., 1998b, JASVO,
775: 26, 97.
776:
777: \bibitem{Karovska1992} Karovska, M. and
778: Mattei, J., 1992, $JAAVSO$, {\bf 21}, 23
779:
780: \bibitem[]{Karovska2006} Karovska, M., Schlegel, E., Hack, W., Raymond, J. C., \&
781: Wood, B. E. 2005, ApJ, 623, L137
782:
783: \bibitem[]{Karovska2006} Karovska, M. 2003, Chandra Newsletter, Vol. 10, p.20-21
784:
785: \bibitem{Kellogg} Kellogg, E., Pedelty, J., and Lyon, R. 2001, ApJ, 563,
786: L151.
787:
788: %\bibitem{Koekemoer2002} Koekemoer, A., Fruchter, A., Hook, R., Hack, W., 2002, in
789: %``HST Calibration Workshop'', p. 337
790:
791: \bibitem{LeahyTaylor} Leahy, D.A. \& Taylor, A.R. 1987,
792: AA, 176, 262
793:
794: \bibitem{LeahyVolk} Leahy, D.A. \& Volk, 1957,
795: ApJ, 440, 847
796:
797: \bibitem{Mikolajewska1998} Mikolajewska {\it et al.} 1998, $IAU Circ.$ No. 6968
798:
799: \bibitem{Mukai2006} Mukai, K., Ishida, M., Kilbourne, C., Mori, H.,
800: Terada, Y., Chan, K-W, 2006, PASJ, in press
801:
802: \bibitem{Park2002} Park, S. {\it etb al.} 2002, ApJ, 567, 314
803:
804:
805: \bibitem{Perryman1997} Perryman {\it etb al.} 1997, A\&A, 176, 262
806:
807: \bibitem{rakowski} Rakowski, C.E. 2006, Ad. Sp. Res. 35, 6, 1017
808:
809: \bibitem{Raymondetal} Raymond, J.C., Cox, D.P. \& Smith, B.W. 1976, ApJ, 204, 290
810:
811: \bibitem{Skopal2002} Skopal, A., Bode, M.F., Crocker, M.M., Drechsel, H., Eyres,S.P.S., Komzík,R. 2002, MNRAS, 335, 1109
812:
813: \bibitem{Sokoloski and Kenyon} Sokoloski, J.L. and Kenyon, S. 2003, ApJ,
814: 584, 1021)
815:
816: \bibitem{Stute} Stute, M. 2006, A\&A, 450, 645
817:
818: \bibitem{Taylor} Taylor, A.R., Seaquist, E.R., \& Mattei, J.A. 1986,
819: Nature, 319, 38
820:
821: \bibitem{Teranova} Teranova, O.G., \& Shenavrin, V. I. 2004,
822: Astronomy Reports, 48, 813
823:
824: %\bibitem{Viotti} Viotti, R. {\et al.} , in {\it Hipparcos-Venice 1997,
825: %ed. Battrick, B., ESA SP-402, ESA Publ. Div., Noordwijk, p. 405
826:
827: \bibitem{Weisskopf} Weisskopf, M.C. {\it et al.} 2002, PASP, 114, 1
828:
829: \bibitem{Wheatley2006} Wheatley, P.J, \& Kalmman T.R. 2006, MNRAS, 372,
830: 1606
831:
832: \bibitem{Ybarra} Ybarra {\it et al.} 2006, astro-ph/0608162
833:
834: \end{thebibliography}
835:
836: %{\bf FIGURES}
837:
838:
839: \begin{figure}
840: \epsscale{.85}
841: \plotone{f1.eps}
842: \caption{Adaptively smoothed
843: ACIS-S images of
844: CH Cyg in two spectral bands:
845: left - soft band (0.2-2 keV) image,
846: right - hard band (2-6 keV) image. North is up, East
847: is to the left.}
848: \end{figure}
849:
850:
851: \begin{figure}
852: \epsscale{.85}
853: \plotone{f2.eps}
854: \caption{
855: CH Cyg soft image with overlayed contours of the simulated PSF showing the loop at 1.5'' from the
856: central source. The central source appears elongated in the N-W direction.
857: North is up, East
858: is to the left.}
859: \end{figure}
860:
861:
862: \begin{figure}
863: \epsscale{.85}
864: \plotone{f3.eps}
865: \caption{Soft image residuals after subtraction of the PSF show the
866: loop-like structure to the south and extension toward the N-W at
867: $\sim$330 degrees}
868: \end{figure}
869:
870: \begin{figure}
871: \epsscale{.85}
872: \plotone{f4.eps}
873: \caption{Comparison of the Chandra soft band adaptively smoothed image
874: showing the central source and the loop with
875: the Galloway and Sokoloski (2004) image (contours). The latter was smoothed
876: using a gaussian with FWHM of 2''.}
877: \end{figure}
878:
879: \begin{figure}
880: \scalebox{0.998}{\rotatebox{0}{\includegraphics{f5.eps}}}
881: \epsscale{.85}
882: \caption{Chandra soft band image of CH Cyg deconvolved using simulated
883: PSF. The image shows
884: a complex multi-component structures
885: similar to the structures detected
886: in the adaptively smoothed image shown in Figure 2 but with
887: improved resolution.
888: In addition to the loop-like structure to the south of the central
889: source, an
890: elongated structure is detected at $\sim$ 0.7'' from the central source
891: in the direction of 330/150 degrees. Overlayed are contours of an OIII
892: HST image obtained 1.6 years before the Chandra image showing similar
893: morphology.}
894: \end{figure}
895:
896:
897:
898: \begin{figure}
899: \epsscale{.85}
900: \plotone{f6.eps}
901: \caption{Overlay of 5GHz radio image contours on the soft X-ray
902: smoothed image showing that the X-ray and radio loops are cospatial
903: within the resolution of 0.3-0.4''.
904: The radio image was obtained 4.5 months before the Chandra image.}
905: \end{figure}
906:
907:
908:
909:
910:
911: \end{document}
912:
913: