New insights on selected radio galaxy nuclei

New Astronomy Reviews 47 (2003) 219–223 www.elsevier.com / locate / newastrev

New insights on selected radio galaxy nuclei David Whysong*, Robert Antonucci Physics Department, University of California, Santa Barbara, CA 93106, USA

Abstract We are performing a survey of powerful 3CR radio galaxies and quasars in the mid-infrared. The purpose is to test for the presence of a powerful hidden quasar-like thermal nucleus by measuring the ‘waste heat’ that must be emitted by any obscuring dust. The dust is treated as a calorimeter for the central engine. Three early mid-IR detections are particularly interesting: Cygnus A, M87, and Centaurus A. These are notable both as a demonstration of our technique and as an example of the great variety of objects that are classified as radio galaxies. We confirm the presence of hidden quasar-like nuclei in Cyg A and Cen A, but M87 shows only weak mid-IR emission, indicating its AGN is non-thermal, and consistent with only synchrotron emission. We also present a new near-infrared adaptive optics image of Cygnus A which shows a secondary point source.  2003 Elsevier B.V. All rights reserved. Keywords: Quasars: general; Infrared: galaxies; Galaxies: individual: Cygnus-A; Centaurus-A; M87

1. Introduction In this review we highlight the pertinent points from our mid-IR observations of radio galaxies. Full details can be found in Whysong and Antonucci (2003). Quasars show powerful optical / UV emission thought to arise from the accretion flow (Koratkar and Blaes, 1999), and this continuum component is invariably associated with broad emission lines. The most powerful radio galaxies are thought to harbor hidden quasars. Many FR II, and most FR I radio galaxies have apparently weak, low-ionization emission lines (‘optically dull’ sources). Chiaberge et al. (1999, 2000) have shown that most of the optically dull sources show optical point sources in HST images. They argue that those with detectable unresolved optical *Corresponding author. E-mail address: [email protected] (D. Whysong).

sources cannot in general have thick obscuring tori: since we can see unresolved optical sources in most optically dull objects, and the selection criterion (radio lobe flux) is isotropic, then most lines of sight to their nuclei must be unobscured. In this scenario, the unresolved optical sources represent synchrotron emission from the bases of the jets. One caveat is that the unresolved optical sources might not represent truly compact nuclear emission. Of course many narrow line radio galaxies do have strong high ionization emission lines, and in some cases spectropolarimetric evidence for a hidden ‘thermal’ nucleus. In general such objects have no detectable optical point source, and are apparently obscured much like Seyfert 2 nuclei (for nearby examples, see Hurt et al., 1999 and Cohen et al., 1999). At z . 0.5, the density and projected linear size of quasars and radio galaxies are as expected if they are from the same parent population (Barthel, 1989); at lower redshift, the situation becomes more compli-

1387-6473 / 03 / $ – see front matter  2003 Elsevier B.V. All rights reserved. doi:10.1016 / S1387-6473(03)00029-0

220

D. Whysong, R. Antonucci / New Astronomy Reviews 47 (2003) 219–223

cated (Singal, 1993). The simplest explanation for the apparent statistical anomalies at low redshift might be the existence of a new type of small radio galaxy which does not harbor a hidden quasar. Gopal-Krishna et al. (1996) pointed out that this can also be explained if the torus opening angle increases with luminosity and radio emission decreases in time.

2. Observational program We are attempting to determine which radio galaxies in the 3CR catalog do not contain hidden nuclei by looking for radiation from warm dust that obscures, and is heated by, a hidden quasar-like nucleus. Various models of obscuring tori predict that the ‘waste heat’ will emerge in the mid-IR, and that this emission is roughly isotropic in all but the highest inclination cases (Pier and Krolik, 1992, 1993; Granato et al., 1997; Efstathiou and RowanRobinson, 1995; a related model in Konigl and Kartje, 1994 should also be consulted; Cen A is discussed specifically in Alexander et al., 1999). We assumed the intrinsic SED is similar to a PG quasar, and used a composite spectrum (Sanders et al., 1989) to estimate bolometric luminosity from the mid-IR flux. While our survey concentrates on FR IIs, we also have a small amount of data on FR Is. The preliminary results indicate that many radio galaxies of both FR types do not contain obscured quasar-like nuclei that emit powerful optical / UV radiation. The results of this survey, and the dependence of the results on optical spectrum, radio size, etc. will have significant implications for the unified model as well as AGN demographics and accretion mechanisms.

3. Results

3.1. 3 C 405 ( Cygnus A) This is a very powerful FR II radio galaxy (Barthel and Arnaud, 1996) at a redshift of 0.056. It has strong high-ionization narrow lines, suggestive of a hidden AGN. A broad Mg II 2800 emission line is detectable in total flux (Antonucci et al., 1994),

and an extremely broad H-alpha line has been seen in polarized flux (Ogle et al., 1997; it is virtually invisible in total flux largely because its great width makes it hard to distinguish from continuum emission). A powerful hidden nucleus should manifest a mid-IR dust luminosity much larger than the observed optical luminosity. However, for this object (and virtually all others!) the only IR data available were taken with very large beam sizes. We (and Radomski et al., 2002)1 isolate the core much better with the | 0.3 arcsec ( | 1.1 kpc) resolution provided by the Keck I telescope. Since the emission is powerful and at the relatively short wavelength of 11.7 mm, it is very likely that this comes from nuclear dust rather than a starburst. The morphology is well resolved (the radial profile of the core has about twice the FWHM of a PSF star), and the total flux is about 150 mJy. It is entirely possible that the extended ( | 10) emission is thermal dust even at radii up to 1 kpc (Radomski et al., 2002). The temperature of nuclearheated dust can be estimated from the radius and luminosity (Barvainis, 1987). Adopting a Hubble constant of 75 km / s Mpc, the luminosity is n Ln 5 9.2 3 10 43 erg / s at 11.7 mm. Our estimated bolometric luminosity is 16.5 n Ln (11.7 mm) | 1.5 3 10 45 erg / s. An optical luminosity of ¯ 10 45 erg / s produces a dust temperature of ¯ 120 K at a 500 pc radius. For comparison the jet power is estimated several different ways (Carilli and Barthel, 1996; Sikora, 2001; Punsly, 2001). The values are rough, but generally lie in the * 10 45 ergs / s range. This is consistent with the finding that jet power and optical / UV luminosity are often comparable in double radio quasars (e.g., Falcke et al., 1995). The conclusion is simple and expected: Cyg A has a moderately powerful hidden nucleus. Fig. 1 shows a new near-IR J, H, K9 composite image of Cyg A obtained with the Keck II NIRC2 instrument. The morphology alone conveys substantial new information. There is a prominent spiral1

Our Keck I image appeared publicly before that of Radomski et al. (2002). We used our mid-IR flux to calculate the bolometric luminosity of the putative hidden quasar, and showed it was of the same order as calculated values for the jet kinetic power from the literature.

D. Whysong, R. Antonucci / New Astronomy Reviews 47 (2003) 219–223

221

AGN sources. In fact, much or all of the mid-IR flux could simply be synchrotron radiation associated with the innermost part of the jet, so the measured flux is an upper limit to the dust luminosity. This is just what Chiaberge et al. implicitly predicted. This rules out a powerful hidden nucleus unless the nuclear dust is too obscured to emit in the mid-IR. The observed 11.7 mm flux corresponds to n Ln 5 1.0 3 10 41 erg / s, for a distance of 15 Mpc. Suppose the mid-IR core is in fact all dust emission. Our estimate for the bolometric luminosity is | 1.6 3 10 42 ergs / s. For comparison, a lower limit to the jet kinetic luminosity in M87 is | 5 3 10 44 erg / s (Owen et al., 2000), so the jet is by far the dominant channel for energy release. If correct, this suggests that M87 is the true ‘misaligned BL Lac.’ Arguably this is the first near-proof of the existence of a ‘nonthermal AGN.’ Fig. 1. JHK9 composite image of the nucleus of Cygnus A. Note the dust lane and the secondary point source 0.420 W of the nucleus. The line at the bottom of the image is 20 across. See this article on astro-ph for a full color image.

shaped dust lane (mostly visible in the J-band data). Most notable is the prominent off-nuclear point source, unresolved at | 70 mas FWHM. While we have not determined the nature of this source, color data rule out the possibility of a foreground object. We have spectra, and the data will be presented fully in an upcoming paper by Canalizo et al.

3.3. Cen A The observation and calibrations were a little more complicated for Cen A; refer to the main paper for details. The Cen A images show only an unresolved source at both 11.7 and 17.75 mm. Photometric results are F(11.7 mm)51.6 Jy and F(17.75 mm) 5 2.3 Jy. Once again the published larger aperture mid-IR fluxes are much higher (Grasdalen and Joyce, 1976). Fig. 2 shows the SED for Cen A adopted from Chiaberge et al., 2001, with our data added.

3.2. M87 This is one of the FR I radio galaxies with an unresolved optical / UV source (Chiaberge et al., 1999, 2000). We detected an unresolved source with 1362 mJy flux.2 Published large aperture (Puschell, 1981) data leave plenty of room for waste heat from a hidden AGN, but our higher resolution data do not. The mid-IR luminosity n Ln is only of order that in the optical rather than much greater as for the hidden 2

Preliminary reports of this work were published in Antonucci (2002) and Whysong and Antonucci (2001); shortly thereafter, Gemini mid-IR images were published and analyzed by Perlman et al. (2001). Their measurements and conclusions were similar to ours. Since their images were very deep, they were also able to detect extended jet emission.

Fig. 2. Spectral energy distribution for Cen A, adopted from Chiaberge et al. (2001). Additional points are from Keck I / LWS at 17.75 mm (small cross) and 11.7 mm (large cross).

222

D. Whysong, R. Antonucci / New Astronomy Reviews 47 (2003) 219–223

The mid-IR flux of | 1.6 Jy at 11.7 mm corresponds to n Ln | 6.0 3 10 41 erg / s. That translates to a bolometric luminosity of 1 3 10 43 ergs / s for the hidden AGN. Cen A demonstrates that radio galaxies with weak or low-ionization lines (the ‘optically dull’ ones; see Storchi-Bergmann et al. (1997) for optical spectra of Cen A), and FR I radio galaxies, can have hidden Type 1 nuclei. The nature of the mid-IR emission is important. The SED has been fit to a synchrotron self-Compton model (Chiaberge et al., 2001) which leads to a classification for Cen A as a ‘misaligned BL Lac’. The fit was to the mid-IR continuum slope instead of the ISO broadband fluxes because Chiaberge et al. felt that the slope would be less sensitive to the strong absorption and emission features (M. Chiaberge, 2001, private communication). Chiaberge argues that the mid-IR spectrum might be consistent with nonthermal emission absorbed by dust. However, small aperture mid-IR spectra show Si absorption and PAH emission features (Alexander et al., 1999; Mirabel et al., 1999), and the SED in # 40 apertures is consistent with predominantly dust rather than synchrotron. These features indicate that most of the mid-IR flux is thermal emission as expected for the torus model. Note also that the ISOCAM CVF flux from the Mirabel et al. (1999) spectrum is in good agreement with our 11.7 mm flux measurement, suggesting that the | 40 ISOCAM spectrum is in fact nuclear, and representative of our smaller aperture. That the high ionization emission features seen in the ISOCAM spectrum are only present in the (admittedly 4 arcsec aperture) nuclear spectrum and not in the surrounding star formation regions suggests the presence of a hidden ionizing continuum. Further evidence to support the thermal emission hypothesis comes from near-IR polarization images (Capetti et al., 2000; Marconi et al., 2000). The bright near-IR peak is highly polarized, and the electric vector position angle is perpendicular to the radio jet axis. Empirically, this is normal for all reflected light objects but would be unusually fortituous for synchrotron jets. Also, the extended, centrally symmetric optical and near-IR scattered light on | 10 pc scales would not be expected in the synchrotron jet hypothesis, but is consistent with light scattered from the central point source.

Acknowledgements The adaptive optics data on Cygnus A were obtained in collaboration with Claire Max, Gabriela Canalizo, Bruce Macintosh, and Alan Stockton. Canalizo in particular deserves a great deal of credit for processing the spectacular color image of Cygnus A. Randy Campbell at Keck deserves our eternal thanks for obtaining the Cen A mid-IR data on our behalf.

References Alexander, D.M., Efstathiou, A., Hough, J.H., Aitken, D.K., Lutz, D., Roche, P.F., Sturm, E., 1999. MNRAS 310, 78. Antonucci, R., Hurt, T., Kinney, A., 1994. Nature 371, 313. Antonucci, R., 2002. Polarization insights for active galactic nuclei, in: Trujillo-Buone, et al. (Eds.), Astrophysical Spectropolarimetry, Cambridge UP, Cambridge, pp. 151–175. Barthel, P., 1989. ApJ 336, 606. Barthel, P.D., Arnaud, K.A., 1996. MNRAS 283, L45. Barvainis, R.E., 1987. ApJ 320, 537. Carilli, C.L., Barthel, P.D., 1996. A&AR 7, 1. Capetti, A. et al., 2000. ApJ 544, 269. Chiaberge, M., Capetti, A., Celotti, A., 1999. A&A 349, 77. Chiaberge, M., Capetti, A., Celotti, A., 2000. A&A 355, 873. Chiaberge, M., Capetti, A., Celotti, A., 2001. MNRAS 324, 33. Cohen, M.H., Ogle, P.M., Tran, H.D., Goodrich, R.W., Miller, J.S., 1999. AJ 188, 1963. Efstathiou, A., Rowan-Robinson, M., 1995. MNRAS 273, 649. Falcke, H., Malkan, M.A., Biermann, P.L., 1995. A&A 298, 375. Gopal-Krishna, Kulkarni, V.K., Wiita, P.J., 1996. ApJ, 463, L1 Granato, G.L., Danese, L., Franceschini, A., 1997. ApJ 486, 147. Grasdalen, G.L., Joyce, R.R., 1976. ApJ 208, 317. Hurt, T., Antonucci, R., Cohen, R., Kinney, A., Krolik, J., 1999. ApJ 514, 579. Konigl, A., Kartje, J.F., 1994. ApJ 434, 446. Koratkar, A., Blaes, O., 1999. PASP 111, 1. Marconi, A., Schreier, E.J., Koekemoer, A., Capetti, A., Axon, D., Macchetto, D., Caon, N., 2000. ApJ 528, 276. Mirabel, I.F. et al., 1999. A&A 341, 667. Ogle, P.M., Cohen, M.H., Miller, J.S., Tran, H.D., Fosbury, R.A.E., Goodrich, R.W., 1997. ApJ 482, L37. Owen, F.W., Eilek, J.A., Kassim, N.E., 2000. ApJ 543, 611. Perlman, E.S., Sparks, W.B., Radomski, J., Packham, C., Fisher, ˜ R., Biretta, J., 2001. ApJ 561, 51L. R.S., Pina, Pier, E.A., Krolik, J.H., 1992. ApJ 401, 99. Pier, E.A., Krolik, J.H., 1993. ApJ 418, 673. Punsly, B., 2001. Black Hole Gravitohydromagnetics. Springer, Berlin. Puschell, J.J., 1981. ApJ 247, 48. ˜ R.K., Packham, C., Telesco, C.M., TadhunRadomski, J.T., Pina, ter, C., 2002. ApJ 566, 675.

D. Whysong, R. Antonucci / New Astronomy Reviews 47 (2003) 219–223 Sanders, D.B., Phinney, E.S., Neugebauer, G., Soifer, B.T., Matthews, K., 1989. ApJ 347, 29. Sikora, M., 2001. Jets in quasars, in: Padovani, P., Urry, C.M. (Eds.), Blazar Demographics and Physics, ASP Conf. Ser. 227, ASP, p. 95 (astro-ph / 0101381).

223

Singal, A., 1993. MNRAS 262, L27. Storchi-Bergmann, T., Bica, E., Kinney, A.L., Bonatto, C., 1997. MNRAS 290, 231. Whysong, D., Antonucci, R., 2001. astro-ph / 0106381 Whysong, D., Antonucci, R., 2003. astro-ph / 0207385