Imaging Polarimetry of the Seyfert 2 Active Galactic Nucleus in NGC1068
Stuart Young1, J.H Hough1, Chris Packham2,
Antonio Chrysostomou1,3, Jeremy Bailey4, Stuart Lumsden4
1 : Dept. of Pjysical Sciences, University of Hertfordshire, Hatfield,
HERTS AL10 9AB, UK
2 : Isaac Newton Group, Apartado 321, 38780 Santa Cruz de la Palma, Islas
Canarias, Spain
3 : Joint Astronomy Centre, 660 N. A'ohoku Place, Hilo, Hawaii 96720, USA
4 : Anglo-Australian Observatory, PO Box 296, Epping, NSW 2121, Australia
The study of NGC 1068, one of the nearest and brightest Seyfert 2 galaxies,
has proved to be the cornerstone upon which the unified theory of active
galactic nuclei (AGN) has been built. From the discovery of a scattered
type 1 components in the polarized flux spectrum of this object, Antonucci
& Miller (1985, ApJ, 297, 621) proposed that all AGN contain the same
basic components of a featureless continuum and broad line region (BLR),
making the type 1 core, together with a narrow line region (NLR).
In some objects the type 1 components are obscured from direct view by
an encircling geometrically and optically thick dusty torus. Indirect
evidence for the torus in NGC 1068 comes from observations of an ionization
cone, formed by the shadowing by the torus body (Pogge, 1988, ApJ, 328,
519), and from evidence for large amounts of gas at the centre of the galaxy
from CO (Planesas, Scoville & Myers, 1991, ApJ, 369, 364) and HCN observations
(Jackson et al., 1993, ApJ, 418, L13; Tacconi et al., 1994, ApJ, 426, L27).
To account for the mid- to far-infrared emission of NGC 1068, a substantial
amount of dust is required in a compact region about the nucleus.
The spectral energy distribution (SED) has been successfully modelled using
a torus-type geometry. Different groups, however, infer different
sizes for the torus, from ~ 10 pc (Pier & Krolik, 1993, ApJ, 418, 673)
to ~ 200 pc (Efstathiou, Hough & Young, 1995, MNRAS, 277, 1134).
At optical wavelengths the emission line/scattering region for NGC 1068
is one-sided owing to the extinction through the host galaxy, however,
at near-IR wavelengths the true bi-conical structure is apparent (Packham
et al., 1997, 288, 375). This leads to a more direct observational
test for the torus size, as a large torus will be visible in silhouette
against the counter-cone.
It was with this aim that imaging polarimetry of NGC 1068 was obtained
using UKIRT with IRCAM3+IRPOL2 during the commissioning of the polarimeter,
built by the University of Hertfordshire, in August 1995. These observations
proved to be completely consistent with the large torus model, and additionally
were also consistent with modelling of the polarized flux spectrum (Miller,
Goodrich & Mathews, 1991, ApJ, 378, 47; Young et al., 1995, MNRAS,
272, 513). Derived parameters were a position angle on the sky of
32 +/- 3 degrees, an inclination of the torus polar axis to the line of
sight of 42 +/- 4 degrees and a minimum diameter for the torus of 200 pc.
 |
FIGURE 1 : The H-band polarized flux intensity of NGC 1068. The contours
represent the fainter features of the image, whilst the false colours emphasise
the detail in the central nuclear region. North is to the top and east
to the left.
The recent upgrades at UKIRT, with significant improvement in image
quality and stability promised a better determination of the torus parameters,
and so NGC 1068 was observed again on 1997 October 1. The resultant
H-band (1.6 microns) polarized flux image is displayed in Fig. 1.
The contours represent the fainter features of the image and the false-colour
image is of the brighter features. Several new features are visible
in this image (resolution 0.25 arcsec), the most striking of which is that
the scattering cones have been resolved into a marked X-shaped pattern.
The north pointing component of the forward cone seems to be spatially
associated with the narrow-line emission features observed with the HST,
with the resolved knot of polarized flux immediately to the north of the
nucleus (in contour) being co-spatial with the knots D and E of the Evans
et al (1991, ApJ, 369, L27) emission line maps. The polarized flux
extends approximately 3 arcsec to the north, which is equal to the diffuse
filaments in the H-alpha image of Capetti, Axon & Macchetto (1997,
ApJ, 487, 560: middle image of their Fig 12; be aware that their scale
bars for the right two images of Fig 12 are incorrect and should be 3 and
1 arcsec as in Fig 4 Plate 11). Any polarized flux associated with
the emission line knots closer to the nucleus are unresolved from the nucleus
in Fig 1. The dark lane to the north-east is associated with the
radio jet, which is also coincident with the dark lane through the emission
line region (Capetti, Macchetto & Lattanzi, 1997, ApJ, 476, L67).
The bulk of the scattered polarized flux follows the centre line of the
torus axis, and is probably from the large but more diffusely filled scattering
cone; it may also result from the interception of the radiation cone and
the host galaxy disk.
 |
FIGURE 2 : A false colur image of the fainter regions of the H-band polarized
intensity image of NGC 1068 (left) and a repeat of Figure 1 (right). The
named regions are features of the polarized intensity related to physical
structures of the active nucleus, as discussed in the text.
The counter-cone, to the south-west, also shows a prominent dark lane
pointing away from the nucleus, co-spatial with the inner counter-jet,
however, emission line images for the counter regions do not exist to compare
features. The absorption owing to the torus is clearly evident on
both arms of the counter-cone polarized flux, and a reliminary
geometrical analysis along the lines of Young et al. (1996) result
in similar values for the position angle of the torus axis on the sky and
the inclination to the line of sight. The better image quality compared
to the August 1995 observations results in a better estimate of the minimum
size of the torus at 180 pc, close to that calculated from the SED model
fits (Efstathiou, Hough & Young 1995).
The nuclear region shows a D-shape structure with an extension along
the postulated torus axis. The D-shape seems to consist of an unresolved
component comprising of a dichroic view of the infrared emitting region
through the torus, with a possible scattering component as well. For approximately
180 degrees around the nucleus to the north-east there is a fainter C-shaped
arc, the shape of which suggests it may result from scattering off the
top surface of the torus, with any scattering from the near-side forward
funnel being attenuated by the bulk of the torus. if the torus is
scattering radiation directly from the near-IR emission region then the
top-side of the torus, within ~ 40 pc, must have a convex shape, again
in excellent agreement with the torus models of Efstathiou, Hough &
Young (1995).
The UKIRT upgrades have revealed a wealth of new information for NGC
1068, through imaging polarimetry. Preliminary analysis of this new
data strengthens the unified theory, and provides additional support for
our scattering and radiation transfer models for NGC 1068.
|