Possible links between BL Lacertae objects and quasars from very long baseline interferometry radio data

Transcription

1 Proc. Natl. Acad. Sci. USA Vol. 92, pp , December 1995 Colloquium Paper This paper was presented at a colloquium entitled "Quasars 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. Possible links between BL Lacertae objects and quasars from very long baseline interferometry radio data D. C. GABUZDA* Department of Physics and Astronomy, University of Calgary, Calgary, AB, Canada ABSTRACT Systematic differences in the very long baseline interferometry (VLBI) radio polarization structure and average VLBI component speeds of BL Lacertae objects and quasars support the view that the observational distinction between these classes, based in large part on the strength of their optical line emission, is meaningful; in other words, this distinction reflects significant differences in the physical conditions in these sources. Possible models providing a link between the optical and VLBI properties of BL Lacertae objects and quasars are discussed. Most VLBI polarization observations to date have been global observations made at 6 cm; recent results suggest that the VLBI polarization structure of some sources may change dramatically on scales smaller than those probed by these 6-cm observations. I. Systematic Differences in the Very Long Baseline Interferometry (VLBI) Properties of BL Lacertae Objects and Quasars Differences in Polarization Structure. Our knowledge of the properties of the milliarcsecond (mas) scale jets in active galactic nuclei (AGN) has been greatly enhanced as the result of polarization-sensitive VLBI observations, mostly made with global arrays at 6 cm (see ref. 1-4 and references therein). BL Lacertae objects are highly variable, polarized AGN with compact, flat-spectrum radio emission, a nonthermal continuum, and comparatively weak emission lines; they are usually Fanaroff-Riley type I (FRI) radio sources (5). Similar properties are displayed by many core-dominated quasars, which have stronger emission lines and are FRII radio sources. Clear differences have emerged between the polarization properties of the parsec-scale jets in quasars and in BL Lacertae objects: the inferred magnetic fields in quasar jets tend to be parallel to the local jet direction, whereas those in BL Lacertae object jets tend to be orthogonal to it (Fig. 1). It is usually supposed that the longitudinal magnetic fields in quasars are due to shear. Perhaps the most natural interpretation of the transverse magnetic field structure observed in BL Lacertae objects is that the polarized jet components are relativistic shocks in which the transverse component of the magnetic field has been enhanced by compression (6, 7). It has sometimes been suggested that the relevant distinction is not between BL Lacertae objects and quasars but rather between low and high redshift objects (for example, see ref. 8). The available VLBI polarization information suggests quite strongly that this is not the case. The transverse magnetic fields characteristic of BL Lacertae objects have been observed in the VLBI jets of both low redshift (e.g., BL Lacertae, z =.7; , z =.11) and high redshift (e.g., , The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C solely to indicate this fact z =.68; , z =.66) sources in this class; in other words, in terms of their VLBI polarization structure, high redshift BL Lacertae objects more closely resemble low redshift BL Lacertae objects than they do quasars. Differences in Characteristic Superluminal Speeds. It has been suspected for several years (for example, see ref. 9) that the apparent speeds in the VLBI jets of BL Lacertae objects are systematically lower than those in the VLBI jets of quasars. It has recently become possible to show statistically that this is indeed the case, although there is clearly a fair bit of overlap between the two speed distributions; a Kolmogorov-Smirnov test indicates that the BL Lacertae object and quasar speed distributions are different with better than 99% confidence (2). Again, there does not appear to be a clear distinction between high and low redshift objects: if only quasars with redshifts within the observed range for BL Lacertae objects for which speeds have been measured (z <.8) are considered, the difference between the BL Lacertae object and quasar speed histograms persists (also with >99% confidence). If it is assumed that the origin for the (usually) superluminal motions seen on parsec scales is relativistic motion in a jet oriented reasonably close to the line of sight, the apparent speed of a component moving along the jet will be,b sin 1apfCos4V' where : is the intrinsic component speed and ' is the angle of the motion to the line of sight. For a given value of,b, this function is peaked at k = arccos3. It is clearly not trivial to uncover the origin for the systematic difference in VLBI component speeds of BL Lacertae objects and quasars, since the observed apparent speed depends on both angle to the line of sight and intrinsic speed for the moving feature. The jets of BL Lacertae objects and highly variable core-dominated quasars are usually considered to have fairly small angles to the line of sight to the earth. The idea that orientation of the VLBI jets of BL Lacertae objects especially close to the line of sight (4 appreciably less than arccosl3) might be the origin for the lower speeds observed in these sources is attractive in some ways, since BL Lacertae objects are generally believed to be among the most highly beamed sources (1, 11). There is considerable evidence, however, that the jets of BL Lacertae objects are not oriented especially close to the line of sight compared to those in quasars. For example, an angle to the line of sight of 35-4 has been derived for the jet in BL Lac by Hughes et al. (7) and Mutel et al. (12); Ghisellini et al. (13) Abbreviations: VLBI, very long baseline radio interferometry; AGN, active galactic nuclei; FR, Fanaroff-Riley; mas, milliarcsecond; pc, parsec. *Present address: Astro Space Center, P.N. Lebedev Physical Institute, Leninsky Prospekt 53, , Moscow, Russia.

2 11394 Colloquium Paper: Gabuzda Proc. Natl. Acad. Sci. USA 92 (1995) A Total Intensity I Peak(mJy/beam)=672. 3C345 Total Intensity Peak(mJy/beam)=694. La 53 8 _Z. co C C, a) a) a4 ui _ o a -4._i -4. 8) al B >~. D 3C345 Complex Polarization Peak(mJy/beam)=167. o F.I _ 6q 58 8 CJ 8i a) ap: I I C FIG. 1. Typical VLBI polarization structure of BL Lacertae objects and quasars, indicated by 6-cm global observations. The BL Lacertae object at epoch : total intensity (A) and linear polarization (B), with contours of polarized intensity and electric field vectors superimposed. Quasar 3C345 at epoch : total intensity (C) and linear polarization (D), with contours of polarized intensity and electric field vectors superimposed. Assuming the jet components are optically thin, the inferred magnetic fields are perpendicular to the electric field vectors shown. suggest that the jets in BL Lacertae objects as a class lie within -3 to the line of sight, while those in quasars lie within -4 to the line of sight. There also does not appear to be evidence for an anticorrelation between app and the core dominance ratio R, as might be expected if the jets in BL Lacertae objects were oriented especially close to the line of sight (2). Thus, it seems more likely that the observed systematic difference in VLBI component speeds of BL Lacertae objects and quasars is associated with systematic differences in the intrinsic speeds for the components; this possibility is discussed below. It should also be pointed out that if some of the components for which speeds have been measured are relativistic shocks-as is almost certainly true-the measured speed is the pattern speed for the shock, which can in some cases differ significantly from the underlying physical speed of the flow (for example, see ref. 14). Thus, it is possible that the

3 Colloquium Paper: Gabuzda systematic difference in the component speeds for BL Lacertae objects and quasars is associated with different characteristic pattern speeds for these components compared to the underlying flow. II. Models Providing a Link Between BL Lacertae And Quasars By definition, the emission line equivalent widths of BL Lacertae objects are small (15, 16). A better indicator of the "strength" of the line emission is probably the emission line luminosity. The emission lines in BL Lacertae objects are nearly always an order of magnitude or more less luminous than those in quasars, although the emission line luminosities for a few BL Lacertae objects approach those for quasars (ref. 5 and references therein; C. Lawrence, personal communication). Thus, although the division in emission line luminosities separating BL Lacertae objects and quasars is not sharp, a systematic difference in emission line luminosity between the two classes does clearly exist. The fact that certain VLBI properties-polarization structure and apparent component speed-are systematically different for BL Lacertae objects and quasars offers further evidence that the optical observational distinction between these sources is not arbitrary but rather has some significant physical basis. It is useful to try to devise models in which it would be natural for sources with weaker emission lines to have the observed VLBI properties of BL Lacertae objects. Two obvious possible origins for the lower line luminosity in BL Lacertae objects are that these sources have a lower gas content or a less powerful ionizing UV continuum than do quasars. Let us suppose that the origin of the weaker line emission in BL Lacertae objects is simply a smaller quantity of emission line gas and that the central engine powering the AGN is a massive black hole (2). It seems plausible that a lower gas content in the host galaxy could lead to a lower accretion rate onto the black hole; this could in turn lead to the ejection of relatively weak and turbulent jets of relativistic material. In this picture, quasars would have a higher gas content, higher accretion rates, and correspondingly stronger, more stable jets. We might expect transverse shocks to form more easily in the comparatively less stable jets of BL Lacertae objects. In addition, it would be natural for the component speeds in the jets of these sources to be on average lower than those observed in quasars, since the jets themselves would be less powerful. A similar picture may be derived from recent interpretation of the optical and radio observational differences between FRI and FRII radio sources by Baum et al. (17). They point out that the emission line luminosity for FRI sources is substantially less than for FRII sources of the same total radio luminosity or same radio core powert and suggest that the collected evidence indicates that this is associated with a lack of ionizing UV continuum in FRI sources, rather than a lack of cold gas. They suggest that the most likely origin for the weaker UV continuum in FRI sources is a lower accretion rate, possibly coupled with a lower central black hole spin rate, compared to FRII sources, and that one consequence of this might be lower Mach numbers in FRI jets. If we suppose that BL Lacertae objects and quasars should be primarily unified with FRI and FRII sources, respectively, this picture suggests that the jets of BL Lacertae objects should be slower than those in quasars; the origin for the comparatively weak emission lines in BL Lacertae objects would in this case be a weak UV continuum. As above, the lower jet speeds in BL Lacertae objects would tthe observation that emission line luminosities in FRI sources are systematically lower than those in FRII sources provides intriguing support for the unification of BL Lacertae objects with FRI sources and quasars with FRII sources. Proc. Natl. Acad. Sci. USA 92 (1995) lead to systematically lower apparent VLBI component speeds and a higher prevalence of transverse shocks in these sources. The idea that jet speed and stability on VLBI scales could play an important role in the relationship between BL Lacertae objects and quasars has also been suggested by Duncan and Hughes (18) on the basis of hydrodynamic simulations of relativistic jets. Somewhat to their surprise, they found evidence that even within the relativistic regime, jets with comparatively low y (-5, for example) are considerably less stable than jets with comparatively high y ('1, for example) and suggested that the dominance of transverse shocks in the VLBI jets of BL Lacertae objects but not in quasars could be understood if BL Lacertae objects have intrinsically lower ry than quasars. In addition, there is some theoretical support for a connection between low accretion rate and low jet speed; for example, Blandford (19) has suggested that a low accretion rate around a slowly rotating black hole may give rise to jets that rapidly decelerate. In such pictures, one would probably expect more or less a continuum of sources with differing accretion rates and emission line strengths; it is possible that the substantial minority of BL Lacertae objects that have FRII arcsecond-scale structure and/or luminosity (2) could be sources with "intermediate" accretion rates. If the accretion rates in BL Lacertae objects are in fact systematically lower than those in quasars, it is also possible that quasars could evolve into BL Lacertae objects if the accretion rate decreases in time, due, for example, to a decreasing supply of gas to feed the black hole (see ref. 17). III. Recent Results Possible Shocks in 3C345 and 3C Six-centimeter VLBI polarization images have shown the dominant magnetic field in the jet of 3C345 to be longitudinal (21). Preliminary analysis of recent 2.8-cm images of this source (D.C.G. and A. Mioduszewski, unpublished data; Fig. 2) suggests that the polarization electric vectors in jet components roughly 1 mas from the core align with the local jet direction. Assuming these components are optically thin, the inferred magnetic field would be transverse, indicating that the polarization from these components may be dominated by transverse shocks. Similar results have been obtained in recent 1.3-cm polarization studies and spectral index studies conducted by Lepannen, Lobanov, and collaborators using the very long baseline array (22). This suggests that although the effects of such transverse shocks are usually not apparent in the jet emission of quasars on the scales probed by 6-cm global VLBI observations, they are more evident on smaller scales. The possibility that shocks can also form in the VLBI jets of quasars is supported as well by 6-cm global VLBI images of 3C454.3 (23), which indicate that, although the dominant jet magnetic field is longitudinal, the polarization electric vectors in a recently emerged component align with the local jet direction, implying the local magnetic field to be transverse. In addition, although the dominant magnetic field in knots in 6-cm images of 3C345 is longitudinal, the interknot emission appears to be more highly polarized than the knot emission, suggesting that the knots are weak shocks in which the degree of polarization has been decreased by partial cancellation of the underlying longitudinal magnetic field and the transverse field due to compression (24). These results therefore imply that the longitudinal fields observed in the VLBI jets in quasars at 6 cm do not necessarily indicate that shocks do not form in these jets, but rather that on the scales that have been sampled by the 6-cm observations the longitudinal field is sufficiently strong to dominate the transverse field in any shock components. The VLBI Polarization Structure of X-Ray BL Lacertae Objects. In the past few years, increasing attention has been

4 11396 Colloquium Paper: GabuzdaPr Proc. Natl. Acad. Sci. USA 92 (1995) 'A. 3C345 Total Intensity I Peak(mJy/beam)=42O.B 3C345 Complex Polarization P Peak(mJy/beam)-147. cv a) Q) N I C? I Relative Right Ascension (mas) Relative Right Ascension (mas) FIG. 2. VLBI hybrid maps of the quasar 3C345 at 2.8 cm, epoch : total intensity (A) and linear polarization (B), with contours of polarized intensity and electric field vectors superimposed. Assuming the jet components are optically thin, the inferred magnetic fields are perpendicular to the electric field vectors shown. paid to the question of the relationship between BL Lacertae objects detected through x-ray and radio surveys (XBLs and RBLs). Although XBLs tend to have weaker radio emission than RBLs, a number of XBLs detected by the HEAO-1 Large Area Sky Survey are being studied using polarization VLBI by Kollgaard and collaborators. The first results from this effort have recently been accepted for publication (25): mas-scale polarization was detected in two of five x-ray BL Lacertae objects. In both sources, polarization was detected in the VLBI jet. In one, , the polarization electric vectors bear no obvious relation to the VLBI jet direction, but in the other, (Fig. 3), the electric vectors in two knots in the inner jet are transverse to the local jet direction, indicating the underlying magnetic field to be longitudinal, as is typical of quasars rather than BL Lacertae objects. At first, it seems that the magnetic field structure observed in indicates that conditions in the VLBI jets of XBLs and RBLs are quite different. However, there is a hint that the magnetic field in may change to being tranverse further down the jet (Fig. 3B); unfortunately, the dynamic range of the polarization image is insufficient to allow this to be asserted with certainty. If this transition from longitudinal to transverse magnetic field is present, however, it is reminiscent of behavior exhibited by the low redshift RBL , in which the dominant jet magnetic field is transverse, but a new component has emerged with a clearly longitudinal mag Total Intensity I Peak(mJy/beamn)=63.3 Cntrs(%)= Complex Polarization P Peak(mJy/beam)=4.97 A -B LO C.) at) ) C) al) al) Q) ao.i 5-5 uc)l 15 I ~~.1 4\ -5-1 FIG. 3. VLBI hybrid maps of the x-ray BL Lacertae object at 6 cm, epoch : total intensity (A) and linear polarization (B), with contours of polarized intensity and electric field vectors superimposed. Assuming the jet components are optically thin, the inferred magnetic fields are perpendicular to the electric field vectors shown.

5 netic field (ref. 2; D.C.G. and T. Cawthorne, unpublished data). Evidence for a "Transition Zone" Several Parsecs from the Core? These results may point toward the existence of a transition zone 2-4 parsecs (pc; 1 pc = 3.9 x 116 m) in projected distance from the core. In 3C345 and 3C454.3, the jet polarization appears to be dominated by transverse shocks within -1 mas (-4 pc) of the core, but by a longitudinal field further down the jet. Other evidence for substantially different conditions in the small-scale jet of 3C345 is suggested by preliminary VLBI Faraday rotation maps (26, 27), which suggest that there may be substantial Faraday rotation near the core but not further down the jet. One possible interpretation of such a transition is that the magnetic fields in the jets in quasars are initially tangled but that over some distance the longitudinal component to the field grows and eventually becomes predominant. There is also some evidence for the presence of a transition zone in BL Lacertae objects, although this evidence is weaker. In and , the magnetic field within 2-3 mas (2-3 pc) has a substantial longitudinal component, which appears to become weaker further down the jet. This possible transition appears to be at about the same projected distance from the core as in quasars 3C345 and 3C454.3, but this could be a coincidence. One possible origin for such a transition from longitudinal to transverse jet magnetic field in BL Lacertae objects in the context of the models discussed in Section II above may be that, in association with a low accretion rate, the jets are initially comparatively fast but rapidly decelerate as they propagate from the core. In this case, we might expect the magnetic field to initially be longitudinal due to shear where the flow is rapid but switch to being transverse further from the core after the jet has decelerated and it becomes easier for shocks to form. These results suggest that there may be a significant change in conditions a few mas from the cores of these sources, which manifests itself in a fairly abrupt change in characteristic magnetic field structure. To uncover the physical origin for such a transition, it is necessary to estimate the actual linear distance in the sources at which this transition occurs; this is difficult, since in the absence of knowledge of the angle of the jets to the line of sight, we may infer only projected distance. Nonetheless, if we assume that the angles to the line of sight for the jets in these sources are 5-3, the deprojected distance for an observed distance of -3 pc would be of the order of a few to a few tens of parsecs. One obvious possibility is that this corresponds to a transition in the surrounding medium-from the broad line to the narrow line region, for example; this would perhaps be consistent with the presence of substantial Faraday rotation inside this zone but not outside it. It is of interest to determine whether there is evidence for a transition zone at a characteristic distance from the core in a larger number of sources and to determine whether the evidence for such a zone exists for both BL Lacertae objects and quasars. Higher resolution VLBI polarization observations are clearly needed to address these issues. IV. Summary Colloquium Paper: Gabuzda The wealth of global 6-cm VLBI polarization data that is now available has established that the characteristic magnetic field structures in BL Lacertae objects and quasars on the size scales probed by these observations are systematically different; the inferred jet magnetic fields in BL Lacertae objects are transverse to the local jet direction (a signature of relativistic shocks) while those in quasars are longitudinal (possibly a signature of shear). In addition, the apparent VLBI component speeds observed in BL Lacertae objects are on average lower than those observed in quasars. These results indicate that the observational distinction between these two types of Proc. Natl. Acad. Sci. USA 92 (1995) AGN, based in large part on the strength of their optical line emission, is a reflection of significant physical differences between these sources. If this is so, there may well be some reasonably direct connection between the optical and VLBI characteristics of these sources. One possibility, for example, is that the accretion rates in BL Lacertae objects are lower than in quasars, leading to the ejection of jets that are either intrinsically slow or rapidly decelerate and in which it is easy for transverse shocks to form. If the origin for the relatively weak emission lines in BL Lacertae objects is a relatively small mass of gas, this could perhaps naturally lead to low accretion rates, providing a connection between the optical and VLBI properties of these sources. Alternatively, low accretion rates could lead to comparatively weak UV continua, which would then provide the reason for the weakness of the line emission in BL Lacertae objects. In both of these pictures, the accretion rates in quasars are higher, leading to the ejection of comparatively strong, stable jets in which the dominant magnetic field is longitudinal due to shear. The apparent VLBI component speeds in BL Lacertae objects would naturally be lower than in quasars, since their jets would be slower. Higher resolution VLBI polarization observations are beginning to become available. Images made recently at 2.8 and 1.3 cm suggest that transverse shocks may be more dominant in the jets of quasars nearer the active nucleus than on the scales probed by 6-cm global VLBI observations. There is also some evidence that, in some cases, longitudinal magnetic fields are present in the jets of BL Lacertae objects on smaller scales. Together, these results may be pointing toward the existence of a transition zone a few to a few tens of parsecs from the core. I would like to thank my collaborators T. Cawthorne, R. Kollgaard, A. Mioduszewski, D. Roberts, and J. Wardle; John Wardle is due special thanks for a critical reading of this manuscript. I would also like to thank J. Vetukhnovskaya and B. Komberg for their interest and encouragement. Much of the work described here is currently being supported by an American Astronomical Society Henri Chretien International Research Grant. 1. Gabuzda, D. C., Cawthorne, T. V., Roberts, D. H. & Wardle, J. F. C. (1992) Astrophys. J. 388, Gabuzda, D. C., Mullan, C. M., Cawthorne, T. V., Wardle, J. F. C. & Roberts, D. H. (1994) Astrophys. J. 435, Cawthorne, T. V., Wardle, J. F. C., Roberts, D. H., Gabuzda, D. C. & Brown, L. F. (1993) Astrophys. J. 416, Cawthorne, T. V., Wardle, J. F. C., Roberts, D. H. & Gabuzda, D. C. (1993) Astrophys. J. 416, Kollgaard, R. I. (1994) Vistas Astron. 38, Laing, R. (198) Mon. Not. R. Astron. Soc. 193, Hughes, P. A., Aller, H. D. & Aller, M. F. (1989) Astrophys. J. 341, Burbidge, G. & Hewitt, A. (1989) in BL Lac Objects, eds. Maraschi, L., Maccacaro, T. & Ulrich, M.-H. (Springer, Berlin), p Mutel, R. L. (199) in Parsec-Scale Radio Jets, eds. Zensus, J. A. & Pearson, T. J. (Cambridge Univ. Press, Cambridge, U.K.), p Roberts, D. H., Gabuzda, D. C. & Wardle, J. F. C. (1987) Astrophys. J. 323, Vermeulen, R. C. & Cohen, M. H. (1994) Astrophys. J. 43, Mutel, R. L., Phillips, R. B., Bumei Su & Buccifero, R. R. (199) Astrophys. J. 352, Ghisellini, G., Padovani, P., Celotti, A. & Maraschi, L. (1993) Astrophys. J. 47, Cawthorne, T. V. & Wardle, J. F. C. (1988) Astrophys. J. 332,

6 11398 Colloquium Paper: Gabuzda 15. Stickel, M., Padovani, P., Urry, C. M., Fried, J. W. & Kuhr, H. (1991) Astrophys. J. 374, Kuhr, H. & Schmidt, G. D. (199) Astron. J. 99, Baum, S. A., Zirbel, E. L. & O'Dea, C. P. (1995) Astrophys. J. in press. 18. Duncan, G. C. & Hughes, P. A. (1994) Astrophys. J. 436, L119- L Blandford, R. D. (1994) in Proceedings of "The Physics of Active Galaxies," ASP Conference Series, eds. Bicknell, G. V., Dopita, M. A. & Quinn, P. J. (ASP), Vol. 54, p Kollgaard, R. I., Wardle, J. F. C., Roberts, D. H. & Gabuzda, D. C. (1992) Astron. J. 14, Brown, L. F., Roberts, D. H. & Wardle, J. F. C. (1994)Astrophys. J. 437, Proc. Natl. Acad. Sci. USA 92 (1995) 22. Zensus, J. A., Krichbaum, T. P. & Lobanov, A. P. (1995) Proc. Natl. Acad. Sci. USA 92, Cawthorne, T. V. & Gabuzda, D. C. (1996) Mon. Not. R. Astron. Soc., in press. 24. Wardle, J. F. C., Cawthorne, T. V., Roberts, D. H. & Brown, L. F. (1995) Astrophys. J. 437, Kollgaard, R. I., Gabuzda, D. C. & Feigelson, E. D. (1996) Astrophys. J., in press. 26. Wardle, J. F. C. & Roberts, D. H. (1994) in Compact Extragalactic Radio Sources, eds. Zensus, J. A. & Kellermann, K. I. (National Radio Astronomy Observatory, Socorro, NM), p Ochs, M. F. (1995) Ph.D. thesis (Brandeis University, Waltham, MA).