1: %\documentstyle[12pt,amssym,aasms4,epsfig,flushrt]{article}
2: \documentstyle[amssym,aas2pp4,epsfig,flushrt]{article}
3: \begin{document}
4: \newcommand{\Mo}{\mbox{$\rm M_\odot$}}
5: \newcommand{\Lo}{\mbox{$\rm L_\odot$}}
6: \newcommand{\mic}{\mbox{$\rm \mu m$}}
7: \newcommand{\ivol}{\mbox{$\rm cm^{-3}$}}
8: \newcommand{\isup}{\mbox{$\rm cm^{-2}$}}
9: \newcommand{\isec}{\mbox{s$^{-1}$}}
10: \newcommand{\Av}{\mbox{$A_{\rm V}$}}
11: \newcommand{\Ak}{\mbox{$A_{\rm K}$}}
12: \newcommand{\Ne}{\mbox{$N_{\rm e}$}}
13: \newcommand{\Np}{\mbox{$N_{\rm p}$}}
14: \newcommand{\Te}{\mbox{$T_{\rm e}$}}
15: \newcommand{\ten}[1]{\mbox{$10^{#1}$}}
16: \newcommand{\xten}[1]{\mbox{$\times 10^{#1}$}}
17: \newcommand{\wl}{\mbox{$\lambda$}}
18: \newcommand{\forb}[2]{\mbox{$[{\rm #1\, #2}]$}}
19: \newcommand{\Ha}{\mbox{H$\alpha$}}
20: \newcommand{\PA}{\mbox{Pa$\alpha$}}
21: \newcommand{\Hb}{\mbox{H$\beta$}}
22: \newcommand{\QH}{\mbox{$Q(\rm H)$}}
23:
24: \lefthead{Capetti et al.}
25: \righthead{HST Infrared imaging polarimetry of Cen A}
26:
27: \title{HST infrared imaging polarimetry of Centaurus A:
28: implications for the unified scheme and the existence of a mis-directed
29: BL Lac nucleus\footnote{Based on
30: observations with the NASA/ESA Hubble Space Telescope, obtained at the
31: Space Telescope Science Institute, which is operated by AURA, Inc.,
32: under NASA contract NAS 5-26555 and by STScI grant GO-3594.01-91A}}
33:
34: \author{Alessandro Capetti}
35: \affil{Osservatorio Astronomico di Torino\\
36: Strada Osservatorio 20, I-10025 Pino Torinese, ITALY}
37:
38: \author{Ethan J. Schreier}
39: \affil{Space Telescope Science Institute\\
40: 3700 San Martin Drive, Baltimore, MD 21218, USA}
41:
42: \author{David Axon, Stuart Young, J. H. Hough, Stuart Clark}
43: \affil{Division of Physics and Astronomy, Dept. of Physical Sciences,\\
44: University of Hertfordshire, College Lane, Hatfield, Herts AL10 9AB, UK}
45:
46: \author{Alessandro Marconi}
47: \affil{Osservatorio Astrofisico di Arcetri\\
48: Largo E. Fermi 5, I-50125 Firenze, ITALY}
49:
50: \author{Duccio Macchetto\altaffilmark{2}}
51: \affil{Space Telescope Science Institute\\
52: 3700 San Martin Drive, Baltimore, MD 21218, USA}
53:
54: \and
55:
56: \author{Chris Packham}
57: \affil{Isaac Newton Group, Sea Level Office\\
58: Apartado de Correos, 321, 38780 Santa Cruz de La Palma,Islas Canarias, SPAIN}
59:
60: \altaffiltext{2}{Associated with Astrophysics Division, Space Science Dept.,
61: ESA}
62:
63: \begin{abstract}
64:
65: We report results from HST/NICMOS 2\mic\ imaging polarimetry of the central
66: region of Centaurus A.
67: In the vicinity of the nucleus we observe a complex polarization structure
68: which we explain by a combination of scattering of nuclear light
69: and dichroic polarization associated with
70: the dust lane. The scattered nuclear
71: radiation is found in an angular region which extends over $\gtrsim
72: 70^{\circ}$ and thus it does not originate from a highly collimated
73: beam, but is associated with more omni-directional nuclear
74: illumination.
75: These observations also show the presence of an
76: unresolved, highly polarized (P = 11.1\%)
77: nuclear source whose polarization angle $\theta = 148.2^{\circ}$
78: is perpendicular to the jet axis. We set an upper limit of $0\farcs04
79: (\sim$0.8 pc) to its extent.
80: The observed nuclear polarization is naturally accounted for
81: if we are observing scattered light from an otherwise obscured nucleus
82: provided that both the scattering region and the occulting torus
83: are extremely compact, with an outer radius of less than $\sim 1$ pc.
84: Alternatively, we might be directly seeing the infrared counterpart of the
85: radio core, similar to those found in other low luminosity radio-galaxies
86: observed with HST.
87: We discuss these results in the framework of the FR~I / BL~Lac unifying model.
88:
89: \end{abstract}
90:
91: \keywords{Galaxies - individual (NGC 5128=Centaurus A); Galaxies - active;
92: Polarization}
93:
94:
95: \section{Introduction}
96:
97: Centaurus A (= NGC5128) is the nearest active galaxy, and both the AGN and
98: the jet have been the subject of extensive studies in all wavelength bands
99: from radio through gamma rays. Radio observations provide strong upper
100: limits on the size of the central source - at 0.5 $\pm$ 0.1 milliarcsec
101: (0.01pc), it is the smallest known extragalactic radio source (Kellerman,
102: Zensus \& Cohen 1997). Because of the heavy obscuration due to the dust
103: lane, it is very difficult to study the AGN itself at high spatial
104: resolution in the optical. Ground-based IR observations have provided
105: evidence for a strong, highly polarized source at the nucleus (cf.
106: \cite{bailey:86}, \cite{packham:96}) and Bailey et al. suggested that Cen A
107: contained a low luminosity, mis-directed BL Lac source.
108: Further evidence for this hypothesis came from the conclusions
109: of Morganti et al. (1991),
110: that the blue filaments in the jet are photo-ionized by a relativistically
111: beamed continuum source. {\sl Conversely, Antonucci and Barvainis (1990)
112: attributed the observed polarization to scattering
113: of nuclear light.} While similarly arguing
114: for scattering as the source of the polarization, Packham et al. (1996)
115: suggested that Cen A could still be a BL Lac type object.
116:
117: High resolution polarimetry of this nearest AGN is important not only to
118: the understanding of AGN in general, but to verifying the unified model
119: and more specifically, the suggestion that Fanaroff-Riley type I radio
120: galaxies provide the parent population for BL Lac objects (cf. Urry \& Padovani
121: 1995). Previous Hubble Space Telescope (WFPC-1) R band imaging polarimetry
122: of the inner region of NGC5128 (\cite{schreier:96})
123: failed to reveal any details of the polarization structure at the nucleus
124: because of its extreme reddening but identified a small
125: region of polarization with a scattering knot southwest of the
126: nucleus.
127: Subsequent Near Infrared Camera and Multi-Object
128: Spectrometer (NICMOS) observations revealed the existence of
129: a nuclear point source which could also be identified
130: in the optical images of the Wide Field Planetary Camera (WFPC2)
131: allowing its flux to be determined from V through K
132: (Marconi et al. 2000).
133: NICMOS narrow band imaging in Pa $\alpha$ and FeII
134: lead to the discovery of
135: an elongated structure which was interpreted as a 20 pc radius ionized
136: gas circumnuclear disk
137: (Schreier et al. 1998).
138:
139: We report here HST
140: (NICMOS) polarimetric observations at 2\mic\, which allow us to map the
141: polarization structure in the vicinity of the nucleus at much higher
142: resolution (0\farcs3) than previously possible.
143: HST's improved spatial resolution
144: over the ground, and NICMOS's sensitivity and polarimetric capability at
145: 2\mic\ allow us to isolate the polarized nuclear component, and to
146: study the effects of the AGN on its nearby environment. We summarize the
147: observations and data reduction in Section 2 and the analysis and observed
148: results in Section 3. The interpretation of the off-nuclear polarization
149: structure is presented in Section 4. The
150: origin of the polarized nuclear emission, and the
151: implications for the AGN unified scheme, are
152: discussed in Sections 5 and 6.
153: Summary and conclusions are given in Section 7.
154: Throughout, we assume a distance to
155: Centaurus A of 3.5 Mpc (\cite{hui:93}), whence 1\arcsec$\simeq$17pc.
156:
157: \section{Observations and Data Reduction}
158:
159: The nuclear region of NGC 5128 was observed on May 6, 1998 using NICMOS
160: Camera 2 with the long wavelength polarizers.
161: The bandpass of the polarizers covers the
162: spectral range 1.9 - 2.1 \mic\ and their principal axes
163: of transmission are
164: nominally oriented at PA 0, 120 and 240 degrees. The spatial resolution is
165: 0\farcs075/pixel and the field of view of the camera is 19\farcs4
166: $\times$ 19\farcs4 (256 $\times$ 256
167: pixels.) All observations were carried out with a MULTIACCUM sequence
168: (\cite{mackenty:97}), i.e. the detector is read out non-destructively
169: several times during each integration to facilitate removal of cosmic rays
170: and saturated pixels. The data were re-calibrated using the pipeline
171: software CALNICA v3.0
172: (\cite{bushouse:97}) and the best reference files in the Hubble Data
173: Archive to produce flux calibrated images. Bad pixels were removed
174: interpolating from values of neighboring pixels.
175:
176: The observing strategy used varying integration times from 16s to
177: 960s for each polarizer, to allow for potential saturation,
178: caused by the bright IR nucleus.
179: Repeating the observation sets with all polarizers in each of 4 orbits,
180: allowed the removal of time dependent effects. The highest count rate observed
181: is 330 c/s per pixel; even for the longest exposure times, the total counts
182: only reach $\sim 12000$ in each of the 25 individual read-outs, well below
183: the saturation limit of $2\times$ $10^5$ and within the linearity regime of
184: the camera. Thus, all the observations are
185: within the linear regime of the camera. This is confirmed by comparing the
186: short and long exposure images, which reveal no significant differences in
187: observed count rates. We present results from the analysis of the set of
188: 12 images obtained with the longest (960s) exposure times, one for each of
189: the three polarizers, taken in four subsequent orbits.
190:
191: A drift in the bias level is known to be present in the NICMOS images and
192: this results in spatially dependent residuals in the calibrated images (the
193: ``pedestal'' problem, \cite{pedestal}). We estimated the effects of the
194: pedestal on our observations by comparing images taken with
195: each polarizer in different
196: orbits. The differences between the images are always smaller than
197: $\sim$ 0.5 \%.
198: We thus estimate that the presence of the pedestal could
199: translate into a spurious polarization on an otherwise unpolarized source
200: of less than 0.6 \%.
201:
202: A bright star is clearly visible in the field of view, $\sim 8 \arcsec$
203: southwest of the nucleus, as in all previous HST images of Cen A. We used
204: the position of this star to check the alignment of the individual images.
205: Shifts between images taken with the different polarizers are negligible,
206: $\lesssim$ 0.03 pixels. There is a small drift between images taken in
207: different orbits, amounting to a total of 0.12 pixels from the first to the
208: fourth orbit. The point-spread functions in the three polarizers do not
209: show significant differences.
210:
211: After registration, images were co-added and combined to produce the final
212: polarization maps. Preflight thermal vacuum tests
213: showed that the transmission of the polarizers is not identical
214: and that they are offset from their nominal position angles
215: (Hines, Schmidt \& Lytle 1997).
216: Polarization parameters were then estimated using the
217: algorithm developed by Sparks and Axon (1999) which deals with this
218: non ideal instrumental configuration and which allowed us to derive the
219: Stokes parameters and, finally, the polarization images.
220:
221: The vacuum tests also showed that any instrumental NICMOS
222: polarization is $\lesssim 1 \%$. On-orbit calibration observations of the
223: polarized star CHA-DC-F7 produce values that agree with 2 \mic\ ground
224: based observations within 0.2\% at the 1\% polarization level.
225:
226: \section{Results}
227:
228: Fig. \ref{fig:i} shows the total intensity map over the entire NICMOS
229: field of view. Not surprisingly, given the similarity between the long
230: wavelength polarizers spectral band and the medium band F222M ("K-band")
231: filter, this image displays the same basic features as the NICMOS image
232: presented by Schreier et al. (1998). A central unresolved
233: source is superimposed on a smooth galactic structure, punctuated by
234: regions of high absorption that are particularly evident toward the South
235: where the thickest regions of the dust lane are located.
236:
237: \subsection{The off-nuclear polarization}
238:
239: In Fig. \ref{fig:ipm} we present, from left to right, maps of the total
240: intensity, polarized flux and percentage polarization, respectively, in the
241: central 7\arcsec $\times$ 7\arcsec. The polarized flux clearly shows a
242: more complex structure than the total intensity image. Two elongated
243: features extend out from opposite sides of the nucleus along PA $\sim
244: 25^{\circ}$ to a distance of about 1\arcsec,
245: where the polarization reaches 4\%.
246: These features,
247: together with two fainter extensions oriented approximately perpendicularly
248: to the previous ones, suggest an S-shaped structure centered on the
249: nucleus; this structure is more apparent in the
250: percentage polarization map. The rest of the polarized emission is rather
251: diffuse, with polarization ranging between 2 and 3 \%.
252:
253: Fig. \ref{fig:vec} shows the polarization field for the same central
254: 7\arcsec\ $\times$ 7\arcsec\, re-binned at 4 $\times$ 4 pixels (0\farcs3
255: $\times$ 0\farcs3); the lengths of the vectors are proportional to the
256: polarized flux and the orientation of the vectors indicates the polarization position angle. Outside the central source, the overall polarization
257: pattern is dominated by a uniform structure oriented approximately along
258: the dust lane. The integrated polarization of the central 2\farcs25, after
259: subtracting the nuclear
260: component, is oriented at PA $\sim$ 106$^\circ$, with $P$ = 2.6\%. This is
261: consistent with the ground based measurements and with the HST/WFPC1 value
262: for the polarization angle, $\sim$ 110$^\circ$. However, there are clearly
263: large deviations from a uniform field which suggest a centro-symmetric
264: structure.
265:
266: To quantify this effect, we plot in Fig. \ref{fig:model} the polarization
267: angle at each pixel versus the position angle (plotted twice for clarity).
268: As usual, position angle 0 corresponds to North and the angle increases
269: towards East. In addition to a uniform polarization field oriented at
270: PA$\sim$105$^\circ$, large deviations with a quasi-periodic behavior
271: suggest that an additional component or components are present. We suggest
272: that the dominant effect can be understood as a nuclear-symmetric pattern
273: superimposed on the uniform field. We discuss the origins of the extended
274: polarization components in Section 4 below.
275:
276: \subsection{\label{sec:nucpol} Nuclear polarization and photometry }
277:
278: We performed aperture photometry on each of the 12 long exposure images
279: separately, to estimate the polarization parameters of the nucleus and to
280: derive direct estimates of the statistical errors. The total flux density
281: of the nucleus is 21.2 $\pm$ 0.2 mJy (1.6 \xten{-15} erg cm$^{-2}$ s$^{-1}$
282: \AA$^{-1}$). Its percentage polarization is 11.1$\pm$0.2\% and the
283: polarization position angle is 148.2$\pm$1.0$^\circ$. Systematic errors are
284: estimated at $\lesssim$ 5\% for the total intensity, due to the uncertainty
285: in the NICMOS calibration, while
286: the effects of instrumental polarization and pedestal are negligible on
287: such a highly polarized source.
288:
289: The size of the nuclear infrared source is estimated by fitting the radial
290: profiles of the nucleus and the bright star SW of the nucleus as a
291: comparison. The FWHM are 2.30 $\pm$ 0.02 and 2.32 $\pm$ 0.01 pixels for the
292: nucleus and the star respectively (errors derived from the spread of the
293: 12 independent images). We conclude that the central IR source in Cen A is
294: unresolved at HST resolution. An upper limit to its extent can be
295: determined assuming that the PSF and the nuclear angular dimension add in
296: quadrature. The three-sigma upper limit for the nuclear FWHM is then 0.6
297: pixels, or 0.8 pc. The polarized source is also unresolved, with a formal
298: FWHM = 2.34 $\pm$ 0.06 pixels; the corresponding three-sigma upper limit on
299: its size is 1.2 pixels, or 1.4 pc.
300:
301: A direct comparison with previous (ground-based) IR imaging polarimetry of
302: Cen A cannot be performed since the NICMOS polarizers bandpass differs
303: significantly from standard infrared filters. The mismatch is particularly
304: important due to the steepness of the Cen A nuclear infrared
305: spectrum (Marconi et al. 2000).
306: We note, however, that within a 2\farcs25 synthetic aperture, our values
307: for the polarization parameters fall between those measured by Packham et
308: al. (1996) in the H and K bands, bracketing the wavelength range of our
309: observations (see Tab. 1).
310:
311: \section{ The extended components: scattering and the dust lane }
312:
313: The polarization map presented above suggests a superposition of two
314: components - a constant field and a centro symmetric component.
315: The field resulting from this superposition
316: depends on the relative intensity $R$ of the two polarizing
317: mechanisms. If the uniform field dominates, the reflection pattern would
318: produce periodic oscillations around the mean value, whose amplitude
319: increases with $R$. In the limit of pure scattering (no uniform field), we
320: would expect straight lines representing an offset of 90 degrees with
321: respect to the polar angle. In Fig. \ref{fig:model} we illustrate two
322: models, for $R = 2$ (dotted line) and for $R= 1/2$ (solid line),
323: respectively. Note that the intensity ratio $R$ of the two components is
324: the only free parameter; the average value of the uniform polarization
325: field can be fixed from ground based observations.
326: Clearly, the largest deviations
327: from uniform polarization, which occur around PA$\sim50$ and $\sim250$ are
328: well fitted by models in which the scattering field dominates, i.e. for
329: large values of $R$.
330:
331: In Fig. \ref{fig:dev} we plot the deviations from constant polarization
332: angle as a function of position with respect to the nucleus. The deviations
333: range up to 70$^\circ$, and the largest occur from PA 35$^\circ$ to
334: 60$^\circ$ and from PA 190$^\circ$ to 280$^\circ$. Much smaller amplitude
335: deviations of $\sim$ 15$^\circ$ degrees are present at PA 130$^\circ$, and
336: no significant deviations are seen in the fourth quadrant (PA
337: $\sim$300$^\circ$). This shows that significant scattering occurs in the
338: nuclear region of Cen A, in a broad bi--cone
339: centered on the jet axis (PA $\sim$55$^\circ$).
340:
341: The two elongated features seen in the polarization images, which form the
342: body of the S-shaped structure we pointed out in Section 3.1
343: (labeled S1 and S2 in Fig. 2b) are
344: not aligned along the position angle of the jet (at the VLBI scale,
345: PA=51$^\circ$, Tingay et al. 1998).
346: On the other hand, they are co-spatial with the small,
347: ionized gas disk discovered in Pa $\alpha$ by Schreier et al. (1998)
348: and the polarization vectors are perpendicular to the radius vector to the
349: nucleus.
350: The most likely explanation of these structures is therefore
351: that they are the result of a concentration of scattering material
352: associated with the
353: disk of ionized gas rather than being directly associated with
354: a collimated nuclear beam.
355:
356: The faint outer extensions of the S-shape structure show a rather different
357: behavior than the inner parts: 1) they are not associated with line
358: emission, and 2) their polarization angle is aligned with the large-scale
359: dust lane. Both results can be accounted for if these features lie outside
360: the region illuminated by the nucleus, with their polarization therefore
361: being accounted for by dichroic transmission rather than scattering.
362:
363:
364: \section{\label{sec:nucorigin} Origin of the polarized nuclear emission}
365:
366: In this section, we examine three possible explanations for the highly
367: polarized source observed at the nucleus of Centaurus A - 1) dichroic
368: transmission through dust, 2) scattering, and 3) synchrotron radiation - as
369: other authors have previously done (cf. Packham et al. 1996). In the next
370: section, we discuss the implications of our results.
371:
372: \subsection{\label{sec:dichroic} Dust dichroic transmission}
373:
374: If we assume the observed polarization arises from dichroic transmission
375: through aligned dust grains, we can calculate the implied extinction. We
376: empirically compare our data to observations of the center of our own
377: galaxy, where a polarization of 6.4\% at K is seen (Bailey, Hough \& Axon
378: 1984), and the extinction is \Av $\sim$ 30 mag. We conclude that \Av
379: $\sim$ 50 mag (i.e. \Ak $\sim$ 6.6 mag)
380: is required to produce the observed polarization of 11.1\% at
381: the nucleus of Cen A. {\sl The intrinsic 2\mic\ flux of Cen A would then be
382: of $\sim 9$ Jy. As a comparison, the mid-infrared nuclear flux of Cen A
383: at, e.g., 15 \mic\, is 1.2 Jy (Mirabel et al. 1999).
384: Such a large flux increase toward shorter wavelengths
385: is not expected regardless on the origin of the nuclear emission.
386: In fact, broad band spectra of radio quiet quasars
387: (Barvainis 1990), which we take as
388: representative of the case in which
389: the nuclear emission is thermally dominated, as well as
390: optically thin synchrotron emission, show the opposite behaviour.
391: The discrepancy with the predicted intrinsic Cen A spectrum is
392: sufficiently large that cannot simply be ascribed to, e.g., a higher
393: efficiency of dichroic polarization and leads us to rule out this
394: explanation.}
395:
396:
397: \subsection{\label{sec:scattering} Scattered nuclear light }
398:
399: If the nuclear polarization originates from scattering of light
400: from an otherwise obscured nucleus, our
401: observations put a strong constraint on the size of the
402: scattering region and the obscuring torus, both of which must have an
403: outer radius smaller than 0.8 pc.
404: To test this possibility we have used the
405: scattering model presented in Young (2000) to determine if the implied
406: compact region can scatter enough flux to account for the observed
407: unresolved source. The model, itself an expansion of the model of Young
408: et al. (1995), allows for a conical {\sl electron}
409: scattering region within a torus-like
410: geometry of a flared disk. The extinction through the torus from any
411: point within the scattering region is calculated in-situ.
412:
413: The outer radius of the scattering cone was taken to be 0.8 pc, the largest
414: size compatible with our unresolved source.
415: The inner and outer radii of the torus were chosen to be 0.03 pc and
416: 1 pc, respectively. The model included a vertical scale-height for
417: the dust density away from the torus equatorial plane, taken to be 1 pc.
418:
419: Typically for active galaxies with obscured broad line regions
420: about 1 per cent of the incident flux is scattered into the line of sight
421: (Miller and Goodrich 1990; Young et al. 1996).
422: With an inner scattering radius of 0.03 pc, and an extinction
423: through the torus of A$_V$ = 47 magnitudes along the direct line of sight, it
424: was possible to reproduce the observed degree of polarization while
425: maintaining an optically thin scattering region with a number density of
426: scatterers of $3 \, 10^{5}$ cm$^{-3}$ at the inner scattering radius.
427: Such a density
428: is comparable to that implied from the scattering modelling of some
429: IRAS galaxies (Young et al. 1996).
430:
431: While the inner radius of obscuring tori in AGN is set
432: by the dust sublimation radius, the appropriate value for its outer radius
433: is matter of debate but it has been usually taken between 10 and 100 pc
434: (Young 2000 and references therein),
435: much larger than the limit set by our observations.
436: However, modelling of the Spectral Energy Distribution (SED) of Cen A
437: also suggests a quite compact torus, with an outer radius of $\sim$ 3.6 pc
438: (Alexander et al. 1999), when compared to other sources
439: (e.g. Efstathiou, Hough \& Young 1995). Furthermore the radio observations
440: by Jones et al. (1996) require the presence of dense material
441: confined to within 1 pc from the nucleus. Taken together these results
442: imply that any obscuring torus in Cen A must be
443: very compact and surrounded by a high density scattering cloud.
444:
445: \subsection{\label{sec:synchro} Synchrotron emission}
446:
447: Analysis of a complete sample of HST images of Fanaroff-Riley type I
448: (FR~I) radio galaxies shows a strong linear correlation between optical and radio
449: core emission (Chiaberge, Capetti \& Celotti 1999). The average
450: radio-to-optical spectral index of the sample, defined by the best-fit
451: correlation, is 0.75. The correlation extends over four orders of magnitude
452: in luminosity, with a dispersion of only 0.4 dex.
453: This lead Chiaberge et al. to argue for a
454: common, non-thermal synchrotron origin for both the radio and optical
455: emission.
456:
457: The very high column density estimated from soft X-ray
458: photoelectric absorption of $9.42\times 10^{22}$ cm$^{-2}$ (Rothschild et
459: al. 1999) is often thought to be too large to allow a direct view
460: of the nucleus of Cen A at optical wavelengths. In fact,
461: it corresponds to a visual extinction of $A_V\sim
462: 47$ mag, assuming a standard gas-to-dust ratio $A_V=5\times 10^{-22} N_H$
463: {\sl which would completely obscure the nuclear source in the optical
464: and, as discussed above, require an untenably high intrinsic infrared flux}.
465: However, Granato, Danese \& Franceschini (1997) argue that most of the
466: X-ray absorption in optically obscured AGNs is produced within the dust
467: sublimation radius, where the gas is essentially dust free. A conversion
468: of column density to extinction based on the galactic gas-to-dust ratio may
469: thus lead to an overestimate of $A_{\rm V}$. A value of only one third
470: standard would reduce the extinction to $A_V \sim15$ mag, making the optical
471: nucleus directly observable even in the optical band.
472: We note this value agrees with that
473: derived from the 10\mic\ silicate absorption feature
474: and with the observations of the nuclear spectral energy
475: distribution by Marconi et al. (2000).
476:
477: In this context, the radio core
478: flux for Cen A of 9.1 Jy at 15 GHz (Clarke, Burns \& Norman 1992)
479: predicts a flux of 10 mJy at 2\mic \,
480: which, seen through a reddening of \Av$\sim$15 mag,
481: is only 0.7 dex smaller than our measurement of 21 mJy.
482: {\sl Similarly, the optical V flux of 6.3 \xten{-20} erg cm$^{-2}$ s$^{-1}$
483: \AA$^{-1}$ (Marconi et al. 2000),
484: when corrected for such an amount of extinction falls
485: again only 0.7 dex below the FR~I radio/optical nuclear correlation
486: \footnote{As discussed in detail by Marconi et al.
487: the central wavelength of the F555W image, 5440 \AA,
488: is significantly offset from its effective value, 5990 \AA, due to the
489: steepness of Cen A nuclear spectrum, wavelength
490: at which fluxes and extinction must by then evaluated.}.}
491:
492: Potentially, a more substantial problem for the synchrotron interpretation
493: of the continuum is that while the nucleus of Cen A is heavily polarized in
494: the IR, it is unpolarized at 800 and 1100 \mic\
495: (with a 3 $\sigma$ upper limit of 0.7 \% at 1100 \mic\ Packham et al. 1996)
496: {\sl and it shows little polarization in the radio band (0.46 \%,
497: Burns, Feigelson and Schreier, 1983).
498: However, a similar behaviour is seen in the two FR~I nuclei
499: for which nuclear polarization measurements exist
500: from HST observations.
501: In M~87 the nucleus is polarized at 2.9 \% in the UV (Capetti et al. 1995)
502: while it shows a polarization level of only 0.95 \%
503: in the radio (Eric Perlman, private communication); similarly, analysis
504: of archival UV images of 3C~264 revealed a polarization of 23 \%,
505: while the radio polarization is only 3.5 \% (Lara et al. 1997).
506: All polarization angles are,
507: within the errors, consistent with being perpendicular to the
508: respective radio jet axis, and this is the case also for Cen A
509: whose radio polarization is oriented at PA = 149 (Rusk and Seaquist 1985).
510: The observed IR polarization properties
511: of the nucleus of Cen A nicely fit in this scheme which suggests a common
512: synchrotron origin for the polarization in the radio and
513: in the IR/optical/UV bands.
514: Certainly intriguing is the location and origin of the medium responsible for
515: the depolarization observed at radio wavelengths.}
516:
517: \section{Implications for the FR~I / BL Lac unified model}
518:
519: In this section we explore the implication of our results for Cen A
520: as a representative member of the FR~I population in the framework of
521: the unifying model for low luminosity radio-sources.
522: The role of obscuration in this scheme is still a matter of debate.
523: Although
524: circumnuclear tori are crucial for the unification of broad and narrow line
525: Seyfert galaxies
526: and powerful radio sources, there is as yet no evidence in favour
527: of their existance in FR~I. Indeed they are not required implicitely
528: for the FR~I / BL Lac unification (see Urry and Padovani 1995).
529: The results of Chiaberge et al.
530: suggest that tori are either geometrically thin (height over
531: radius $\simeq 0.1$) or present only in a minority of FR~I.
532:
533: It is clear from our previous discussion that our IR polarization
534: observations of Cen A can be well explained in a model in which the nucleus is
535: surrounded by an obscuring torus
536: and the infrared nuclear source (as well as its optical counter part)
537: is purely due to scattering.
538: By allowing the scattering region to be partially obscured,
539: the extinction observed between
540: the V-band and the near-IR can also be reproduced.
541:
542: Alternatively, as discussed in the previous section,
543: we might have a direct view of the synchrotron emission from the
544: nucleus of Cen A.
545: Packham et al. (1996) pointed out that the nucleus of Cen A has rather
546: different properties from BL Lac objects. In particular, they noted: a) lower
547: polarization at mm wavelengths than typical for BL Lacs; and b) a
548: polarization angle which is approximately perpendicular to the jet axis,
549: while in BL Lacs it is usually parallel to it.
550: In addition our observations do not show
551: a highly collimated polarized region cospatial with the jet axis
552: which again argues against the presence of a misdirected
553: BL Lac nucleus in Cen A.
554:
555: However convincing these arguments appear
556: when taken together with our discussion about scattering,
557: they do not strictly rule out a synchrotron origin for the
558: nuclear polarization, since there is growing evidence that the emission
559: which dominates in BL Lacs
560: does not originate from the same region of that seen in FR~I. First,
561: the polarization of radio cores in all radio galaxies is much lower than in
562: BL Lacs (Saikia 1999).
563: Second, the intensity ratio between BL Lacs and FR~I
564: nuclei is typically $10^3 - 10^4$, indicative of Lorentz factors $\Gamma
565: \sim 5 $ again smaller that typical values for BL Lacs, i.e.
566: $\Gamma \sim 10 - 15$ (Ghisellini et al. 1993, Chiaberge et al.
567: 2000). In fact Chiaberge et al. concluded that the core emission in FR~I comes
568: from a less anisotropic component than in BL Lacs, due to the
569: presence of a velocity structure within the jet, which allows slower
570: components to dominate the faster ones at large viewing angles.
571: A similar conclusion has been derived by Laing et al. (1999)
572: discussing the jet asymmetry of FR~I.
573: If we adopt this view then the nuclear polarization of Cen A can
574: still be due to synchrotron from an unresolved slow jet halo,
575: surrounding a high $\Gamma$ jet.
576:
577: A key constraint provided by our observations is that one has to
578: explain not only the nuclear polarization but also the nature
579: of the off nuclear scattered light.
580: While this is readily accounted for if there is a shadowing torus,
581: in the synchrotron picture the polarized light is reflecting
582: the intrinsic anisotropy of the nuclear radiation field.
583: Our images show that the angular region covered by scattered light has an
584: opening angle of $\sim 70-90$$^\circ$ and therefore requires only
585: modest anisotropy in the nuclear emission.
586: The opening angle (FWHM) of the illuminating beam is
587: related to the jet bulk Lorentz factor as $\theta \sim 2 \Gamma^{-1}$.
588: If this is due to Doppler beaming it requires
589: $\Gamma \sim 2$ in agreement with the estimates of Lorentz factors
590: for the jet halo derived by Chiaberge et al. (2000).
591:
592: Because the source of illumination is intrinsically polarized
593: in principle one should expects deviations from circular symmetry
594: in the polarization pattern.
595: The detailed modeling of this effect shows that the source polarization
596: must be comparable with the maximum polarization caused by scatterers
597: (60 \%)
598: in order to have a significant impact on the observed pattern.
599: Unfortunately in Cen A the foreground dichroic sheet prevents any such
600: effect to be seen.
601:
602: However, we also have to explain the absence of a collimated
603: scattered counter part to the fast synchrotron core of the radio jet.
604: The only obvious way to do it is to make
605: this region essentially devoid
606: of scatterers by clearing them with the passage of the arcsec scale jet.
607:
608: \section{Summary and conclusions}
609:
610: The large scale polarized component seen in the nuclear regions of Cen A
611: is consistent with previous ground-based and HST WF/PC-I
612: observations and is explained by dichroic transmission through the
613: foreground dust lane. Significant quasi-periodic deviations in the
614: polarization angle within a few arcseconds of the nucleus
615: are caused by scattering of nuclear radiation.
616: The strongest effects are seen in the quadrants that
617: contain the jet symmetry axis, but there is no detailed correlation with
618: the jet itself.
619:
620: The nucleus is found to be unresolved, with a 3-sigma upper limit on
621: the FWHM of 0.8 pc. It is highly polarized, at 11.1$\pm$0.2\%,
622: with a position angle of 148.2$\pm$1.0$^\circ$, perpendicular to the jet axis.
623:
624: The nuclear polarization cannot be explained by
625: dichroic transmission through dust, as this would require an absorption of
626: \Av$\sim$50 mag. {\sl The intrinsic 2 \mic\ flux would then exceed by a factor
627: $\sim$ 10 the mid infrared measurements and this is not expected
628: regardless on the origin (thermal or non thermal)
629: of the nuclear emission.}
630:
631: Our modeling shows that it is possible to explain the nuclear
632: polarization as optically thin scattering from a $\sim 1$ pc cone.
633: This requires the existence of an extremely compact torus.
634: However, this is in agreement with the modeling of the SED of the Cen A
635: nucleus and also with the radio observations.
636: This model also naturally explains the large scale biconical scattering
637: morphology.
638:
639: A synchrotron origin for the nuclear radiation requires
640: that it originates from a low $\Gamma \sim 2$
641: unresolved jet halo surrounding the faster jet core so that the angular size
642: of extended scattering can be matched.
643: This kind of jet morphology is expected on the basis
644: of recent results on the statistical properties of FR~I radio-galaxies.
645: In order to explain the absence of a bright scattered BL Lac beam,
646: i.e. a high $\Gamma$ jet core, one has to assume that the jet
647: clears away the scattering material along its trajectory.
648:
649: It would be very important at this stage
650: to establish whether circumnuclear tori are common in FR~I
651: or whether their observed properties are dominated by their jet
652: emission. Unfortunately,
653: previous measurements of the nuclear polarization in FR~I have always
654: been hampered by the dominant dilution of starlight
655: as their nuclear sources only account for a few per cent
656: of the total flux in typical ground based apertures
657: (see e.g. Impey, Lawrence \& Tapia 1991).
658: A high resolution HST study of the sort present here for the FR~I class
659: as a whole is needed to establish the polarization
660: properties of their IR and optical cores and their pattern of illumination.
661:
662: \acknowledgements
663: We would like to thank the referee, R. Antonucci, for his useful comments
664: and suggestions.
665: \newpage
666:
667: \begin{thebibliography}{}
668: %
669: \bibitem[Alexander et al. 1999]{1999}
670: Alexander, D. M., Efstathiou, A., Hough, J. H., Aitken, D.,
671: Lutz, D., Roche, P., \& Sturm, E., 1999, MNRAS 310, 78
672: %
673: \bibitem[Antonucci and Barvainis (1990)]{1990ApJ...363L..17A} Antonucci, R.
674: and Barvainis, R. 1990, \apjl, 363, L17
675: %
676: \bibitem[Bailey Hough \& Axon 1984]{1984MNRAS.208..661B} Bailey, J., Hough,
677: J. H. \& Axon, D. J. 1984, \mnras, 208, 661
678: %
679: \bibitem[Bailey et al. 1986]{bailey:86}
680: Bailey, J., Sparks, W. B., Hough, J. H., \& Axon, D. J.,
681: 1986, Nature, 322, 150
682: %
683: \bibitem[Barvainis (1990)]{1990ApJ...353..419B} Barvainis, R. 1990, \apj,
684: 353, 419
685: %\bibitem[Burns, Feigelson and Schreier (1983)]{1983ApJ...273..128B} Burns,
686: J. O., Feigelson, E. D. and Schreier, E. J. 1983, \apj, 273, 128
687: %
688: \bibitem[Bushouse et al. 1997]{bushouse:97}%
689: Bushouse, H., Skinner, C.J., MacKenty, J.W, 1997,
690: NICMOS Instrument Science Report, 97-28 (Baltimore STScI)
691: %
692: \bibitem[Capetti, et al. (1995)]{1995ApJ...452L..87C} Capetti, A.,
693: Macchetto, F., Axon, D. J., Sparks, W. B. and Boksenberg, A. 1995, \apjl,
694: 452, L87
695: %
696: \bibitem[Chiaberge Capetti \& Celotti 1999]{1999A&A...349...77C} Chiaberge,
697: M., Capetti, A. \& Celotti, A. 1999, \aap, 349, 77
698: %
699: \bibitem[Chiaberge et al. 2000]{2000} Chiaberge,
700: M., Celotti, A., Capetti, A. \& Ghisellini, G. 2000, \aap, 358, 104
701: %
702: \bibitem[Clarke Burns \& Norman 1992]{1992ApJ...395..444C} Clarke, D. A.,
703: Burns, J. O. \& Norman, M. L. 1992, \apj, 395, 444
704: %
705: \bibitem[Efstathiou Hough \& Young 1995]{1995MNRAS.277.1134E} Efstathiou,
706: A., Hough, J. H. \& Young, S. 1995, \mnras, 277, 1134
707: %
708: \bibitem[Ghisellini et al. 1993]{ghisellini:93}
709: Ghisellini, G., Padovani, P., Celotti, A., Maraschi, L., 1993,
710: ApJ, 407, 65
711: %
712: \bibitem[Granato Danese \& Franceschini 1997]{1997ApJ...486..147G} Granato,
713: G. L. , Danese, L. \& Franceschini, A. 1997, \apj, 486, 147
714: %
715: \bibitem[Hawarden et al. 1993]{1993MNRAS.260..844H} Hawarden, T. G.,
716: Sandell, G., Matthews, H. E., Friberg, P., Watt, G. D. \& Smith, P. A.
717: 1993, \mnras, 260, 844
718: %
719: \bibitem[Hines Schmidt \& Lytle 1997]{1997hstc.work..217H} Hines, D. C.,
720: Schmidt, G. D. \& Lytle, D. 1997, The 1997 HST Calibration Workshop with a
721: New Generation of Instruments, p. 217, 217
722: %
723: \bibitem[Hui et al. 1993]{hui:93}
724: Hui, X., Ford, H.C., Ciardullo, R., Jacoby, G.H. 1993, \apj, 414, 463
725: %
726: \bibitem[Impey Lawrence \& Tapia 1991]{1991ApJ...375...46I} Impey, C. D.,
727: Lawrence, C. R. \& Tapia, S. 1991, \apj, 375, 46
728: %
729: \bibitem[Jones et al. 1996]{jones:96}%
730: Jones, D.L., Tingay, S.J., Murphy, D.W., et al., 1996, \apj, 466, L63
731: %
732: \bibitem[Laing Parma de Ruiter \& Fanti 1999]{1999MNRAS.306..513L} Laing,
733: R. A., Parma, P., de Ruiter, H. R. \& Fanti, R. 1999, \mnras, 306, 513
734: %
735: \bibitem[Kellermann Zensus \& Cohen 1997]{1997ApJ...475L..93K} Kellermann,
736: K. I., Zensus, J. A. \& Cohen, M. H. 1997, \apjl, 475, L93
737: %
738: \bibitem[MacKenty et al. 1997]{mackenty:97}%
739: MacKenty, J.W., et al., 1997, NICMOS Instrument Handbook, Version 2.0
740: (Baltimore STScI)
741: %
742: \bibitem[Marconi et al. 2000]{2000}
743: Marconi, A., Schreier, E. J., Koekemoer, A., Capetti, A., Axon, D. J.,
744: Macchetto, F. D., Caon, N.,
745: 2000, ApJ 528, 276
746: %
747: \bibitem[miller 1990]{mg90}
748: Miller, J. S., Goodrich, R. W., 1990, ApJ, 355, 456
749: %
750: \bibitem[Mirabel, et al. (1999)]{1999A&A...341..667M} Mirabel, I. F., et al.,
751: 1999, \aap, 341, 667
752: %
753: \bibitem[Morganti et al. 1991]{1991MNRAS.249...91M} Morganti, R., Robinson,
754: A., Fosbury, R. A. E., di Serego Alighieri, S., Tadhunter, C. N. \& Malin,
755: D. F. 1991, \mnras, 249, 91
756: %
757: \bibitem[Packham et al. 1996]{packham:96}%
758: Packham, C., Hough, J.H., Young, S., Chrysostomou, A., Bailey, J.A.,
759: Axon, D.J., Ward, M.J., 1996, \mnras, 278,406
760: %
761: \bibitem[Rothschild et al. 1999]{1999ApJ...510..651R} Rothschild, R. E., et
762: al. 1999, \apj, 510, 651
763: %
764: \bibitem[Rusk and Seaquist (1985)]{1985AJ.....90...30R} Rusk, R. and
765: Seaquist, E. R. 1985, \aj, 90, 30
766: %
767: \bibitem[Saikia 1999]{1999MNRAS.302L..60S} Saikia, D. J. 1999, \mnras, 302,
768: L60
769: %
770: \bibitem[Schreier et al. 1996]{schreier:96}%
771: Schreier, E.J., Capetti, A., Macchetto, F., Sparks, W.B., Ford, H.C.,
772: 1996, \apj, 459, 535
773: %
774: \bibitem[Schreier et al. 1998]{schreier:98}%
775: Schreier, E.J., Marconi, A., Capetti, A., Caon, N., Axon, D., Macchetto, F.,
776: in preparation.
777: %
778: \bibitem[Skinner at al. 1997]{pedestal}%
779: Skinner, C.J., Bergeron, L.E., Daou, D., 1997, HST Calibration Workshop,
780: Eds. S. Casertano et al. (Baltimore STScI), in press
781: %
782: \bibitem[Sparks \& Axon 1999]{1999PASP..111.1298S} Sparks, W. B. \& Axon,
783: D. J. 1999, \pasp, 111, 1298
784: %
785: \bibitem[Urry \& Padovani 1995]{1995PASP..107..803U} Urry, C. M. \&
786: Padovani, P. 1995, \pasp, 107, 803
787: %
788: \bibitem[Tingay et al. 1998]{1998AJ....115..960T} Tingay, S. J., et al.
789: 1998, \aj, 115, 960
790: %
791: \bibitem[Tubbs 1980]{tubbs:80}%
792: Tubbs, A.D., 1980, \apj, 241, 969
793: %
794: \bibitem[Young 2000]{}
795: Young, S., 1999, MNRAS, in press
796: %
797: \bibitem[Young 1995]{}
798: Young, S., Hough, J. H., Axon, D. J., Bailey, J. A., Ward, M. J., 1995, MNRAS,
799: 272, 513
800: %
801: \bibitem[Young 1996]{}
802: Young, S., Hough, J. H., Efstathiou, A., Wills, B. J., Bailey, J. A., Ward, M.
803: J., Axon, D. J., 1996, MNRAS, 281, 1206
804:
805: \end{thebibliography}
806:
807:
808:
809:
810:
811:
812: \newpage
813:
814: \begin{figure}
815: \centering
816: \psfig{figure=i.ps,width=0.9\linewidth}
817: \vskip 0.5cm
818: \caption{\label{fig:i} Gray scale image
819: of the nuclear region of Cen A in the wavelength range 1.9 -- 2.1 \mic.
820: The field of view is 19\arcsec $\times$ 19\arcsec. For all images, North is up and east is to the left.}
821: \end{figure}
822:
823: \begin{figure*}
824: \centering
825: \psfig{figure=ipm.eps,width=0.3\linewidth,angle=-90}
826: \vskip 0.5cm
827: \caption{\label{fig:ipm} Gray scale total intensity image of the innermost
828: 7\arcsec $\times$ 7\arcsec of Centaurus A.
829: Dots surrounding the nucleus are artifacts of the NICMOS PSF.
830: From left to right the images are total intensity, polarized intensity
831: and percentage of polarization.
832: The long dash marks the radio-jet orientation.}
833: \end{figure*}
834:
835: \begin{figure}
836: \centering
837: \epsfig{figure=vec.epsi,width=1.0\linewidth,angle=-90}
838: \vskip 0.5cm
839: \caption{\label{fig:vec} Polarization vectors
840: of the innermost 7\arcsec $\times$ 7\arcsec of Centaurus A.
841: Vectors are plotted every 0\farcs3\ and the length is proportional
842: to the polarized flux.
843: The central source has been masked out
844: to show more clearly the off nuclear polarization pattern.}
845: \end{figure}
846:
847: \begin{figure}
848: \centering
849: \psfig{figure=model.ps,width=1.0\linewidth}
850: \vskip 0.5cm
851: \caption{\label{fig:model} Polar diagram of the polarization position angle
852: with the origin centered on the nuclear IR source. Two periods are plotted
853: for clarity. Deviations from a uniform pattern (oriented at PA $\sim 105$)
854: are seen and are centered at the position angle of the jet and counter jet.
855: Lines represent polarization models
856: resulting from the superposition of a centrosymmetric and a uniform pattern
857: for two values of their intensity ratio $R$ = 2 (dotted line) and $R$ = 0.5
858: (solid line).}
859: \end{figure}
860:
861: \begin{figure}
862: \centering
863: \psfig{figure=dev.ps,width=1.0\linewidth}
864: \vskip 0.5cm
865: \caption{\label{fig:dev} Location of the largest deviations of the
866: polarization field from the uniform pattern. Empty circles are deviations
867: between 10 and 20 degrees, filled circles between 20 and 30, squares
868: above 30 degrees.
869: The dashed line marks the radio-jet orientation.}
870: \end{figure}
871:
872: \newpage
873:
874:
875: \begin{table}{\label{tab:photo}}
876:
877: \caption{Comparison of HST and ground based polarimetry}
878:
879: \begin{tabular}{c l c c c c}
880:
881: & & & & & \\
882: \hline
883: & & & & & \\
884: & Band & $\rm{I_P}[\%]$ & $\theta$ & I [mJy] & P [mJy]\\
885: & & & & & \\
886: \hline
887: & & & & & \\
888: & H & 3.45 & 129.7 & 73 & 2.5 \\
889: & HST & 3.83 & 135.6 & 69 & 2.6 \\
890: & K$_n$ & 5.43 & 142.2 & 88 & 4.8 \\
891: & & & & & \\
892:
893: \hline
894: \end{tabular}
895: \end{table}
896:
897: \end{document}
898:
899:
900:
901:
902:
903:
904:
905:
906:
907: