Next Contents Previous

11. AN OBSCURED QSO IN CYGNUS A?

Cygnus A might well be an important nearby example of a powerful radio galaxy harboring a QSO in its nucleus, confirming the (Fanaroff and Riley Class II) QSR - radio galaxy unification scheme (Scheuer 1987, Peacock 1987, Barthel 1989). Simply put, this scheme argues that FRII radio galaxies are the parent population of radio loud quasars, with the jets in quasars oriented closer to our line of sight than those in the radio galaxies. This scheme is based on the observation of weaker cores and jets and the larger angular sizes for radio galaxies as compared to quasars, coupled to their similar space density distributions, as well as the statistics of superluminal motion in these classes of objects. An essential ingredient is an opaque torus with a polar axis roughly aligned with the axis of the radio jet, obscuring the broad emission line region from direct view in the radio galaxy class. Although cone angle evolution (with redshift and/or luminosity) cannot be excluded, available data seem to imply a dividing cone angle of approximately 45° (Barthel 1989, Padovani and Urry 1991). Barthel (1993) and Urry and Padovani (1995) summarize arguments pro and contra this unification scheme.

As will be described below, a simple symmetric twin relativistic jet model suggests the Cygnus A radio source axis to be oriented at an intermediate angle to the line of sight, consistent with the simple FRII unification scheme. Tadhunter et al. (1994) point out that the opening half angle of the QSO cone in Cygnus A may be more like ~ 60°, with the radio axis inclination therefore greater than ~ 60°. In addition, we have seen that dust is abundant in the nuclear regions. In this respect it is also interesting to recall the study of Hes et al. (1993), investigating optical emission line luminosities and morphologies within the FRII unification scheme. Hes et al. (1993) find [OII] luminosities for the classes in accordance with the scheme, and infer larger sizes for the [OII] nebulae compared to the [OIII] nebulae, attributing these facts to dust in the narrow line region. Although [OII] images for Cygnus A have not been obtained, the original long slit spectroscopy of Baade and Minkowski (1954) indicated an [OII] extent of at least 30 arcsec, which should be compared to ~ 6 arcsec for the [OIII] nebula (Baum et al. 1988, Stockton et al. 1994)! However, proving that Cygnus A indeed contains a QSO obscured from direct view is another matter.

Following Pierce and Stockton's (1986) demonstration of an extended source of featureless blue continuum radiation, Goodrich and Miller (1989) presented spectropolarimetric measurements, in search for the nature of this continuum component. They observed a rather low percentage of polarization indicating that most of the blue light which Osterbrock (1983) had estimated to make up a large fraction of the light in the central regions does not arise via scattering, but is seen directly. Goodrich and Miller (1989) proposed that hot stars near the galaxy center must be responsible for the blue light component. Despite the low level of polarization, Tadhunter et al. (1990) were able to map the polarized V-band radiation, and discovered a polarization pattern indicative of bipolar reflection. This centro-symmetric pattern, which extended over the central 3 arcsec, was evidence that at least some part of the radiation was due to a continuum source obscured from direct view. This part (FC1, following Tran 1993) combines with directly seen blue radiation (FC2), possibly from hot stars, and the normal stellar population to make up the blue continuum in the central region of the Cygnus A galaxy. Near-infrared imaging by Djorgovski et al. (1991) and near-infrared spectroscopy by Ward et al. (1991) resulted in more evidence for the obscured QSO hypothesis. Whereas the former study reported a bright, very red nucleus at 2.2 and 3.7 µ which could be combined with data at other frequencies for a rough extinction estimate (AV ~ 50 ± 30 mag), the latter study used established quasar correlations to obtain a more precise estimate for the nuclear continuum extinction: AV = 54 ± 9 mag. Recent near-IR imaging by Stockton et al. (1994) indicated the presence of a marginally resolved second K-band nucleus, located about 1 arcsec NW of the true (radio core) nucleus.

Hard X-rays could also penetrate through the postulated circumnuclear dust, and hence facilitate the search for an obscured QSO. The question of X-ray emission from the active nucleus of Cygnus A has been addressed both through spectroscopy, and high resolution imaging. Based on X-ray spectra taken with EXOSAT, Arnaud et al. (1987) were the first to discuss evidence for a hard X-ray tail to the spectrum of Cygnus A, perhaps indicative of emission from the active nucleus. Ueno et al. (1994) have recently obtained GINGA spectra of Cygnus A, confirming the results of Arnaud et al. (1987). For a single component thermal model they find an unreasonably high temperature of 18 x 107 K. The data is best fit using two-components: thermal emission from cluster gas at T = 8.5 x 107 K, and a heavily absorbed (N(HI) = 4 x 1023 cm-2) power-law emission component (presumably the nucleus) with photon index of 2 and total luminosity between 2 keV and 10 keV of 6 x 1044 h-2 erg sec-1. These conclusions concerning the nuclear X-ray spectrum of Cygnus A are supported by recent ASCA observations (Arnaud et al. 1996). Although large, the total absorption corrected X-ray luminosity is somewhat smaller than that of radio-loud quasars of the same radio luminosity, being typically in the range 1045 - 1046 erg sec-1 (e.g. Zamorani et al. 1981). Harris et al. (1994b) find evidence for a point X-ray source at the center of Cygnus A on the ROSAT HRI image. This point source is robust in terms of modeling and subtracting the extended cluster distribution. On the other hand, using the the absorption column and core spectrum derived by Ueno et al. implies that the core source should be completely obscured in the HRI bandpass, i.e. the obscuring gas is opaque below about 4 keV. Investigation of the spectrum of the compact X-ray core source must await future spatially resolving spectroscopic observations.

The absorption corrected X-ray luminosity for the nucleus of Cygnus A is furthermore comparable to that seen for broad line radio galaxies, and agrees with the rough correlation between far-infrared and hard X-ray emission for nearby broad-line AGN (Ward et al. 1988). Ueno et al. (1994) point out that the implied absorbing column density is very high, comparable to that found for some Seyfert 2 galaxies. The ultimate test of the existence of an obscured X-ray nucleus of quasar-type luminosity at the center of Cygnus A will be possible with the next generation of X-ray satellites, which will combine high spectral resolution with high spatial resolution.

Barvainis and Antonucci (1994) and McNamara and Jaffe (1994) searched for CO absorption towards the radio nucleus, arising in the postulated dusty molecular torus. Optical depth limits of approx 0.5 were set for both CO 0-1 and CO 1-2 at velocity resolutions of order 1 km sec-1. The lack of CO absorption presents a challenge to the torus model for Cygnus A. Barvainis and Antonucci suggest three possible solutions. First, it may be that there is no such torus. Second, the sizes of the molecular clouds in the torus may be smaller than the size of the background continuum source. Models of Krolik and Begelman (1988) and Krolik and Lepp (1989) suggest cloud sizes approx 0.03 pc, which is still well below the smallest scale yet probed by VLBI observations. And third, Maloney et al. (1994) discuss the possibility that the radio continuum emission from the nucleus radiatively excites the CO, `rendering the lower-J CO transitions undetectable in absorption.' This depends critically on the radio flux-to-particle density ratio, with high ratios leading to substantial radiative excitation. The extreme limit is a CO excitation temperature approaching the brightness temperature of the radio continuum nucleus. Current data cannot rule out any of these three options.

The most exciting recent result in the search for absorption towards the active nucleus of Cygnus A is the recent detection of HI 21cm absorption at the systemic redshift by Conway and Blanco (1995). The absorption line seen towards the nucleus is broad (FWHM = 270 km sec-1), perhaps consisting of two components, with a peak optical depth of 0.05. These figures are in contrast to regular dust lane HI absorption: the NGC 5128 (Centaurus A) absorption for instance measures 30 km sec-1. The implied neutral hydrogen column density towards Cygnus A is 2 x 1019 TS / f, where TS is the spin temperature of the gas, and f is the covering factor. Neither OH nor H2CO absorption is detected, at 1% optical depth and 54 km sec-1 resolution. Conway and Blanco (1995) compare their results to absorption seen towards other radio loud AGN, and conclude that they are seeing absorption by the hypothesized dense circumnuclear torus. The opacity and velocity structure are comparable to those predicted by Krolik and Begelman (1988), and Krolik and Lepp (1989). Given the lack of molecular absorption, Conway and Blanco discuss the possibility of a dusty, atomic torus, and show that such a situation is physically reasonable, although a molecular torus is by no means precluded. Comparing their results to the X-ray studies, they suggest: TS = 104 K, and f = 1, numbers which can be tested through future VLBI spectral line imaging.

A final curious point is that the spectrum of the VLBI nucleus is inverted between 5 GHz and 43 GHz. The straight-forward interpretation of this inverted spectrum nucleus is synchrotron self absorption. However, it is worth noting that in their standard model for dusty molecular tori, Krolik and Lepp (1989) predict that such a torus will be optically thick to free-free absorption at around 10 GHz. If the inverted spectrum core is due to free-free absorption, then the width of the obscuring torus is set by the fact that the inverted spectrum is seen only for the nucleus in Cygnus A, not the VLBI jet (Carilli et al. 1994a). The implied maximum width to the torus is approx 1 pc, in the case of torus axis oriented in the sky plane. This can be further tested with multifrequency VLBI observations.

Overall, it seems likely there is a continuum source of quasar-like luminosity in the nucleus of the Cygnus A galaxy, hidden from direct view by a dense column of dusty gas. Is there also an obscured broad emission line region? Ward et al. (1991), following earlier measurements by Saunders (1984) and Lilly and Hill (1987) determined a Paalpha line width of 510 km/sec, in good agreement with the optical Balmer line widths. More importantly, the extinction towards broad Paalpha emission was determined using the relationship between X-ray and broad line luminosity (Ward et al. 1988) resulting in AV (BLR) > 24. It thus appears that even in the near-IR broad lines (as QSO characteristics) will be difficult to detect. In an attempt to determine the level of scattering using spectropolarimetry Jackson and Tadhunter (1993) found increased polarization towards the blue, but no sign of scattered broad lines, confirming similar work by Miller and Tran (1993). Stockton et al. (1994) examined their high S/N spectra in search for broad Hbeta. The strong featureless continuum just SE of the radio core predicts broad Hbeta equivalent widths of at least 50 Å if all of the continuum is scattered QSO radiation. Stockton et al. (1994) measured an upper limit to such broad Hbeta of 15 Å, demonstrating that the featureless continuum is most likely not dominated by scattered QSO radiation.

A high S/N ultraviolet spectrum obtained with HST was presented by Antonucci et al. (1994). This spectrum, taken of the SE part of the Cygnus A nuclear region (which is dominated by continuum radiation) does show a broad MgII lambda2795 line, with FWHM ~ 7500 km sec-1 and equivalent width ~ 25 Å. Invoking a Rayleigh scattering origin (propto lambda-4) or even a less steep scattering law, the measured MgII flux of 8 x 10-16 erg cm-2 sec-1 would be consistent with the absence in the optical band of broad Halpha in available spectra. Also with reference to the ultraviolet polarization properties (Antonucci et al. 1994), these data comprise strong evidence that Cygnus A does harbour an obscured broad line region. It will be of great interest to compare the ultraviolet line and continuum polarization properties. Present Cygnus A data seem therefore to support the quasar-radio galaxy unification model. On the assumption of the (narrow) Balmer line emission to be due to quasar ionization Stockton et al. (1994) calculate the quasar blue magnitude to be ~ 16. A similar computation by Metz et al. (1994) leads to MV ~ -23.5 for the central QSO. It follows that the obscured QSO in Cygnus A may not be very luminous optically. We note however that some fraction of the featureless blue continuum and the line emission in the Cygnus A central regions may be due to young stars, so there is still uncertainty as to the true `QSO content.' In fact, the quoted QSO luminosities are lower limits, since the adopted value of the ionization parameter is a lower limit, given the possibility of young stars contributing to the ionization of the gas.

Next Contents Previous