Proc. Natl. Acad. Sci. USA Vol. 92, pp. 11356-11359, December 1995 Colloquium Paper

This paper was presented at a colloquium entitled " and Active Galactic Nuclei: High Resolution Radio Imaging," organized by a committee chaired by Marshall Cohen and Kenneth Kellermann, held March 24 and 25, 1995, at the National Academy of Sciences Beckman Center, Irvine, CA.

Very-long-baseline radio interferometry observations of low power radio G. GIOVANNINI*t, W. D. COFTONt, L. FERETTI*t, L. LARAt§, T. VENTURIt, AND J. M. MARCAIDEI *Dipartimento di Astronomia, Universita di Bologna, Bologna, Italy; tIstituto di Radioastronomia, Consiglio Nazionale delle Richerche, Bologna, Italy; *National Radio Astronomy Observatory, Charlottesville, VA 22903-2475; §Instituto de Astrofisica de Andalucia, Consejo Superior de Investigaciones Cientificas, Andalucia, Spain; and 1Departamento de Astronomia, Universitat de Valencia, Valencia, Spain

ABSTRACT The parsec scale properties of low power Table 1. VLBI radio galaxies radio galaxies are reviewed here, using the available data on Log 12 Fanaroff-Riley type I galaxies. The most frequent radio P408,t VLBI structure is an asymmetric parsec-scale morphology-i.e., Radio z* W/Hz morphology4: Ref.§ core and one-sided jet. It is shared by 9 (possibly 10) of the 12 mapped radio galaxies. One (possibly 2) of the other galaxies 0055+30 NGC 315 0.0167 23.95 One-sided (7) has a two-sided jet emission. Two sources are known from 0104+32 3C31 0.0169 24.50 One-sided This paper published data to show a proper motion; we present here 0206+35 4C35.03 0.0375 24.28 One-sided This paper evidence for proper motion in two more galaxies. Therefore, in 0755+37 NGC 2484 0.0413 25.04 One-sided (8) the present sample we have 4 radio galaxies with a measured 0836+29 4C29.30 0.0790 25.08 One-sided (9) proper motion. One of these has a very symmetric structure 1142+20 3C264 0.0206 24.85 One-sided This paper and therefore should be in the plane of the sky. The results 1144+35 0.0630 24.15 One-sided? This paper discussed here are in agreement with the predictions of the 1222+13 3C272.1 0.0037 23.27 Two-sided? This paper unified scbeme models. Moreover, the present data indicate 1228+12 3C274 0.0037 25.07 One-sided (10) that the parsec scale structure in low and high power radio 1626+39 3C338 0.0303 25.25 Two-sided (11) and galaxies is essentially the same. this paper 1637+82 NGC 6251 0.0230 24.55 One-sided (5) 2335+26 3C465 0.0301 25.39 One-sided (9) The knowledge of the structure of radio galaxies on the *Galaxy . parsec scale is important to test current models of jet tTotal radio power at 408 MHz. dynamics and acquire new pieces of information to test the 4Parsec scale radio morphology. radio source unified schemes. Very-long-baseline radio in- §Reference for radio morphology. terferometry (VLBI) data on powerful radio galaxies and quasars show a strong evidence of relativistic jets and in many Radio galaxies in the same range of total radio power, but cases a proper motion with an apparent superluminal velocity unresolved on the arcsecond scale, have not yet been properly has been found (1, 2). VLBI observations of low power radio mapped. Therefore, they are not discussed here. Observations galaxies are also necessary to compare the parsec scale prop- to study this class of radio sources remain to be made, in order erties of radio sources with different radio powers and to test to understand their nature and connection with the more the unified scheme models, which predict also that low power radio galaxies should have parsec scale jets moving at a velocity powerful CSO and CSS sources (6). close to the speed of light. In this paper, we present and discuss A Hubble constant Ho = 100 km-sec- 'Mpc-1 (Mpc, the available VLBI data on extended low power radio galaxies megaparsec; 1 pc = 3.09 x 1016 m) and deceleration parameter [Fanaroff-Riley type I, hereafter referred to as FRI (3)]. We qo = 1 have been used throughout this paper. will use the radio galaxies of the sample currently under study by us (4) observed at 5 GHz with the very long baseline array Radio Morphology (VLBA) or the global array. This sample was obtained by selecting from the B2 and 3CR galaxy samples those objects The list of radio galaxies studied so far and discussed here is with a core flux density >100 millijansky (mJy; 1 Jy = 10-26 presented in Table 1. A morphological analysis based on the wm-2-Hz-1) at 6 cm at arcsecond resolution (4). The core flux available VLBI maps indicates that an asymmetric morphol- limit, imposed by observational constraints, could produce a ogy-i.e., core and one-sided jet-is the most frequent radio sample biased toward objects with jets pointing toward the structure (see, for example, Fig. 1). It is shared by 9 (possibly observer. This point is not important in discussing single 10) of the 12 mapped radio galaxies. A clear symmetric struc- objects but has to be taken in account for statistical studies. ture is found in 3C338 (Fig. 2), while 3C272.1 shows a complex The well known FRI galaxy NGC 6251 (see ref. 5 and structure with a possible counterjet close to the core (Fig. 3). references therein), not included in our sample, was added, for The one-sided jet is always well collimated and only small a total of 12 FRI radio galaxies. Some of them have been oscillations or bendings are visible. The nuclear emission is observed also at 1.6 and 8.4 GHz. Only 5 have observations at always the dominant component. When maps at two or more different epochs to search for a possible proper motion. frequencies are available, the core emission shows a flat or

The publication costs of this article were defrayed in part by page charge Abbreviations: VLBI, very-long-baseline radio interferometry; FRI payment. This article must therefore be hereby marked "advertisement" in and -II, Fanaroff-Riley types I and II; Jy, Jansky; VLBA, very long accordance with 18 U.S.C. §1734 solely to indicate this fact. baseline array. 11356 Downloaded by guest on October 5, 2021 Colloquium Paper: Giovannini et al. Proc. Natl. Acad. Sci. USA 92 (1995) 11357

3C264 4974.989 MHz 3C338 4987.490 MHz I I ~I* I I 4 25 _A ;-> .I C~~-( 20 F 80 -2 15 F

20 10 0 -10 -20 10 F mas

FIG. 2. VLBA map of 3C338 at 5 GHz. The HPBW is 2.2 X 2.2 5 , milliarcsec (mas). The peak flux is 44.4 mJy per beam; contour levels are -0.5, 0.5, 0.7, 1, 1.5, 2, 3, 4, 6, 8, 10, 20, 30, and 40 mJy per beam. oF. Q Two-sided arrow shows the direction of the symmetric kiloparsec scale I 0 jet. -5 F strongly variable. Several measurements of the flux density (4) , '. I ,t-. at 1.4 and 5 GHz show that it was "300 mJy in 1974 while now -10 it is 540 mJy after reaching a maximum of 610 mJy in 1991. -15 Simultaneous multifrequency observations show that the core 15 10 5 0 -5 -10 spectrum is flat between 1.4 and 5 GHz but strongly steepens mas between 5 and 8.4 GHz. The VLBI structure consists of two main components (A and C) with an inverted spectrum 3C465 8417.990 MHz between 1.7 and 5 GHz and low brightness, jet-like features departing from them (Fig. 4). A comparison between our data and the 5-GHz VLBI map obtained in the second Caltech- Jodrell Bank VLBI survey (14) shows that (i) component A is probably variable, and therefore we tentatively identify it as the core; in this case the parsec scale structure would be in the direction of the fainter kiloparsec scale jet. However, owing to the complexity of this source and the slight asymmetry of the faint kiloparsec scale jets, the definition of a main jet may be ambiguous. (ii) The separation between A and C increases between the two observing epochs. This proper motion is clearly visible also in comparing the model given (14) with our visibilities. The data are consistent with a proper motion of component C with respect to A with an apparent superluminal velocity = 1.2c. The snapshot data obtained by us (4) are in agreement with this motion.

mas 3C272.1 4973.241 MHz

FIG. 1. (A) VLBI map of 3C264 at 5.0 GHz. The half-power beam 40 width (HPBW) is 3.5 x 2.1 milliarcsec (mas) in position angle 11°. The peak flux is 135 mJy per beam; contour levels are -0.3, 0.3, 0.5, 0.8, 1.5, 3, 5, 10, 25, 50, and 100 mJy per beam. Arrow shows the direction of the main kiloparsec scale jet. (B) VLBI map of 3C465 at 8.4 GHz. 30 The HPBW is 2.52 x 0.83 mas in position angle -9.7°. The peak flux '0 is 132 mJy per beam; contour levels are -0.75, 0.75, 1.5, 2, 3, 5, 10, 20, 50, and 100 mJy per beam. Arrow shows the direction of the main kiloparsec scale jet. 20 _

inverted spectrum while the jet emission has a spectral index Io1 of ;0.5. A few sources are described in detail below. peculiar 10 _ 3C264. The one-sided jet detected at parsec resolution (Fig. 1A) is oriented within a few degrees of the optical jet visible in the Hubble Space Telescope map (12). Unfortunately, the resolution of the radio map is too high for a detailed compar- O _ ison with optical data. 3C272.1. The core radio power in this nearby galaxy is low with respect to the total radio power, so we expect that this -10 source is very close to the plane of the sky (see next section). IIAIL The VLBI map shows a complex structure with a curved jet on 10 5 0 -5 -10 the side of the main kiloparsec jet (North) and a possible short mas counterjet (Fig. 3) on the opposite side. FIG. 3. VLBI map of 3C272.1 at 5 GHz. The HPBW is 5.3 x 1.2 for a detailed discussion of this 3C274. See Biretta (10) milliarcsec (mas) in position angle -5°. The peak flux is 168 mJy per source. beam; contour level are -0.7, 0.7, 1, 2, 3, 5, 7, 10, 20, 50, 100, and 150 1144 + 35. This source shows a kiloparsec scale dominant mJy per beam. Arrow shows the direction of the main kiloparsec scale core and two slightly asymmetric faint jets (13). The core is jet. Downloaded by guest on October 5, 2021 11358 Colloquium Paper: Giovannini et al. Proc. Natl. Acad. Sci. USA 92 (1995)

1144 + 35 4991.990 MHz kiloparsec jet. A faint counterjet is visible only in very large array (VLA) maps at arcsecond resolution. 15 Discussion

10 Jet Velocity and Orientation. In all the sources discussed here, the parsec scale jet is oriented on the same side of the A main kiloparsec scale jet with the exception of 1144+35, which is still uncertain (see above). This correlation implies either b that jets are intrinsically asymmetric or that parsec and kilo- 0 parsec scale jets are both relativistic. The presence of relativ- istic jets in strong radio sources is now widely accepted (1, 15). -5 Furthermore, a detailed study of the inner kiloparsec scale properties of low power radio galaxies (16) and the evidence I -10 of proper motion at high velocities in some galaxies (see below) in FRI are 5 0 -5 -10 -15 -20 suggest that radio jets radio galaxies initially mas relativistic. For these reasons, we interpret the radio structures presented here as affected by Doppler favoritism and will use FIG. 4. VLBI map of 1144 + 35 at 5.0 GHz. The HPBW is 2.5 X the available data to constrain the possible values of the jet 0.8 milliarcsec (mas) in position angle 5°. The peak flux is 255 mJy per velocity (/3 = v/c) and of the orientation of the radio source beam; contour levels are -1.5, 1.5, 3, 5, 7, 10, 15, 20, 50, and 200 100, with respect to the line of sight (6). We can constrain these two mJy per beam. Arrow shows the direction of the main kiloparsec scale a to coun- jet (but see text). parameters in few different ways: (i) from the jet terjet brightness ratio, (ii) from the prominence of the core radio power with respect to the total radio power, (iii) from 3C338. This source has a very steep global spectrum and is comparing the observed x-Ray nuclear emission with that associated with the multiple-nuclei cD galaxy NGC 6166. Even expected by the Self Compton Model, and (iv) from imposing source core for this the arcsecond flux density is strongly an upper limit on 0 to restrict the maximum intrinsic radio variable in time. At parsec resolution, this source shows a flat source size to 1.5 Mpc. A detailed discussion on these methods spectrum core with two symmetric jets oriented in the East- can be found in (8). The allowed values for the jet velocity /3 West direction (Fig. 2). We observed this source at 1.6, 5, and and its orientation with respect to the line of sight 0 are given 8.4 GHz, and second epoch observations are available at 5 and in Table 2. 8.4 GHz. While analysis of the 5-GHz data remains to be Proper Motion. While a proper motion of well defined completed, a preliminary comparison between the 8.4-GHz features inside the radio jets is firmly established and well maps of the two epochs (Fig. 5) shows a clear change in the studied in strong radio galaxies, quasars, and BL Lac type source structure. It is not obvious in such a complex structure objects (1, 2), the situation is still unclear for low power radio how to determine an unambiguous proper motion; however, galaxies. A few galaxies (five) among those presented here the present data suggest a possible motion corresponding to an have at least two observations at different epochs that can be apparent velocity of 0.5 c. used to look for the existence of a proper motion. 3C465. This giant Wide Angle Tail has a total The galaxy 3C274 shows evidence of stationary knots as well radio power intermediate between low and high power radio as structures moving at a subrelativistic velocity; moreover, galaxies. The parsec scale map (Fig. 1B) shows a core emission some substructures seem to move with an apparent velocity and an asymmetric jet in the same direction of the main larger than c (10). NGC 6251 could have both a stationary knot

3C338 8404.990 MHz

5 4 3 2 F o0 I _II O I I I --- I i I -1 -2 -3 -4

20 15 10 5 0 -5 -10 -15 -20 mas 3C338 8417.490 MHz I I p I I % I I I 4 2 - O o D - N l30_ i~ -.

-2

I I' 0 -4 N0A

20 15 10 5 0 -S -10 -15 mas

FIG. 5. (Upper) VLBI map at 8.4 GHz of 3C338 obtained on 1991.3. The peak flux is 62.7 mJy per beam; contour levels are -0.3, 0.3, 0.5, 0.7, 1, 1.5, 2, 5, 10, 40, and 60 mJy per beam. (Lower) VLBA map of 3C338 at 8.4 GHz obtained on 1994.92. The peak flux is 25.8 mJy per beam; contour levels are -0.3, 0.3, 0.5, 0.7, 1, 1.5, 2, 5, 10, 15, and 20 mJy per beam. The HPBW is 2 x 1 milliarcsec (mas) in position angle 00 in both maps, which have been plotted in the same scale.

Downloaded by guest on October 5, 2021 ~~~~~~~ Colloquium Paper: Giovannini et al. Proc. Natl. Acad. Sci. USA 92 (1995) 11359 Table 2. Jet velocity and orientation kiloparsec scale differences seem to arise from conditions far from the nuclei and could be related to a different interaction Radio galaxy Allowed (3* Allowed O,t 0" if y , with the surrounding medium. A similar result was deduced by 0055+30 NGC 315 0.7-1.0 30-40 30-40 De Young (18) in an optical study of FRII and FRI galaxies 0104+32 3C31 0.5-1.0 25-60 50-60 and by Maraschi and Rovetti (19), who compared BL Lac type 0206+35 4C35.03 0.6-1.0 0-50 35-50 objects and flat spectrum radio quasars. 0755+37 NGC 2484 0.5-1.0 0-55 40-55 0836+29 4C29.30 0.7-1.0 0-40 30-40 We thank the staffs at the telescopes for their contribution to these 1142+20 3C264 0.5-1.0 0-55 40-55 observations and the staffs at the Bonn and VLBA correlator where 1144+35 0.4-1.0 0-60 45-60 the data have been correlated in absentia. The National Radio 1222+13 3C272.1 0.4-1.0 60-80 75-80 Astronomy Observatory is operated by Associated Universities, Inc., 1626+39 3C338 0.0-1.0 0-90 85-90 under contract with the National Science Foundation. 1637+82 NGC 6251 0.9-1.0 40-45 40-45 2335+26 3C465 0.6-1.0 0-50 35-50 1. Zensus, J. A., Krichbaum, T. R. & Lobanov, A. P. (1995) Proc. = Natl. Acad. Sci. USA 92, 11348-11355. *Allowed range for (3 v/c in parsec scale jets. 2. Vermeulen, R. C. (1995) Proc. Natl. Acad. Sci. USA 92, 11385- tAllowed range for source angle with respect to the line of sight 0 11389. corresponding to the jet velocity range given in *. ' 3. Fanaroff, B. L. & Riley, J. M. (1974) Mon. Not. R. Astron. Soc. tAs t but for a jet with Lorentz factor y 3. 167, 31-37. 4. Giovannini, G., Feretti, L. & Comoretto, G. (1990) Astrophys. J. and one moving at v = 1.2c, the same velocity we have found 358, 159-163. for the component C in 1144 + 35. 3C338 has certainly 5. Jones, D. L. & Wehrle, A. E. (1994) Astrophys. J. 427, 221-226. changed its structure, and a proper motion with v 00.5c is 6. Readhead, A. C. S. (1995) Proc. Natl. Acad. Sci. USA 92, 11447- compatible with data but it needs to be confirmed. For NGC 11450. 315, an upper limit of 0.5c on the jet velocity was derived (7). 7. Venturi, T., Giovannini, G., Feretti, L., Comoretto, G. & Wehrle, It looks as if we see in the jets of low power radio galaxies A. E. (1993) Astrophys. J. 108, 81-91. stationary as well as subluminal and superluminal knots. This 8. Giovannini, G., Feretti, L., Venturi, T., Lara, L., Marcaide, J., could reflect the presence of oblique shocks and complex Rioja, M., Spangler, S. R. & Wehrle, A. E. (1994) Astrophys. J. situations where the measured velocity could be much lower 435, 116-127. than the jet velocity (17). More data are therefore necessary 9. Venturi, T., Castaldini, C., Cotton, W. D., Feretti, L., Giovan- to in case nini, G., Lara, L., Marcaide, J. M. & Wehrle, A. E. (1995) properly discuss this point, but any the detection of Astrophys. J., in press. proper motion in some low power radio galaxies confirms that 10. Biretta, J. A. & Junor, W. (1995) Proc. Natl. Acad. Sci. USA 92, in this class of radio galaxies parsec scale jets are relativistic. 11364-11367. UnifiedModels. The present data and the derived values for 11. Feretti, L., Comoretto, G., Giovannini, G., Venturi, T. & Wehrle, (3and 0 are in agreement with the expectations from unified A. E. (1993) Astrophys. J. 408, 446-451. models. In fact, in all the low power radio galaxies presented 12. Sparks, W. B., Biretta, J. A. & Macchetto, F. (1994) Astrophys. J. here, observational data are in agreement with the presence of Suppl. Ser. 90, 909-916. a parsec scale jet with a Lorentz factor y ' 3, viewed at angles 13. Parma, P., de Ruiter, H. R., Fanti, C. & Fanti, R. (1986) Astron. larger than 300 with respect to the line of sight. This is Astrophys. Suppl. 64, 135-171. consistent with FRI sources being the parent population of BL 14. Henstock, D. R., Browne, I. W. A., Wilkinson, P. N., Taylor, Lac type objects. G. B., Vermeulen, R. C., Pearson, T. J. & Readhead, A. C. S. we note that the scale FRI (1995) Astrophys. J. Suppl. Ser. 100, 1-25. Moreover, parsec properties of 15. Antonucci, R. R. J. (1993) Annu. Rev. Astron. Astrophys. 31, radiogalaxies are very similar to the parsec scale properties of 473-521. FRII radio galaxies and quasars. The large morphological 16. Laing, R. A. (1995) Proc. Natl. Acad. Sci. USA 92, 11413-11416. difference between FRI and FRII radio galaxies at the kilo- 17. Begelman, M. C. (1995) Proc. Natl. Acad. Sci. USA 92, 11442- parsec scale does not exist at the parsec scale. This similarity 11446. suggests that the nature and the power of the nuclear engine 18. De Young, D. S. (1993) Astrophys. J. 405, L13-L16. is the same in low power and high power sources. The 19. Maraschi, L. & Rovetti, F. (1994) Astrophys. J. 436, 79-88. Downloaded by guest on October 5, 2021