Inflow of the Broad-Line Region and the Fundamental Limitations of Reverberation Mapping
aa r X i v : . [ a s t r o - ph . C O ] O c t Accretion and Ejection in AGNs: a Global ViewASP Conference Series, Vol. 000, 2010L. Maraschi, G. Ghisellini, R. Della Ceca & F. Tavecchio (eds)
Inflow of the Broad-Line Region and the FundamentalLimitations of Reverberation Mapping
C. Martin Gaskell
Astronomy Department, University of Texas, Austin, TX 78712-0259,USA
Abstract.
The evidence from velocity-resolved reverberation mapping show-ing a net infall of the broad-line region (BLR) of AGNs is reviewed. Differ-ent lines in many objects at different epochs give a consistent picture of BLRmotions. The motions are dominated by virialized motion (rotation plus turbu-lence) with significant net inflow. The BLR mass influx is sufficient to power theAGN. The increasing blueshifting of lines with increasing ionization potential isa consequence of scattering off infalling material. The high blueshiftings of theUV lines in Narrow-Line Seyfert 1s are due to enhanced BLR inflow rates ratherthan strong winds. Seemingly conflicting cases of apparent outflow reverbera-tion mapping signatures are a result of the breakdown of the axial-symmetryassumption in reverberation mapping. There are several plausible causes of thisbreakdown: high-energy variability tends to be intrinsically anisotropic, regionsof variability are necessarily located off-axis, and X-ray observations reveal ma-jor changes in line-of-sight column densities close to the black hole. Resultsfrom reverberation mapping campaigns dominated by a single event need to betreated with caution.
1. Introduction
At a conference on “Accretion and Ejection in AGNs” one could na¨ıvely expectroughly equal time for both accretion and ejection! However, inspection theAGN literature (and the conference program) shows that, while the theory ofejection in jets and winds, and observations of ejection from γ -rays to radiowaves are all well covered, there is almost nothing on observations of inflow .Yet we all believe that AGNs are ultimately powered by accretion onto blackholes, so something has to be inflowing! I argue here that the “something” isthe broad-line region (BLR) and that we can see consequences of the inflow.
2. The broad-line region – a “bird’s nest”
A BLR is a ubiquitous feature of all AGNs accreting at a high accretion rate(within two or three orders of magnitude of the Eddington limit). The linesare the dominant features of the UV/optical spectrum. The BLR consists ofrapidly moving ( v & − ) gas. The strong variability of the lines inresponse to continuum variations shows that they arise almost entirely fromphotoionization and that other energy input is negligible. Photoionization an-alyzes show that the gas is dense ( n H & cm − ). I have recently reviewedthe structure of the BLR elsewhere (Gaskell 2009) so I only give a brief sketch1 Gaskell here. From the large equivalent widths of the lines and the inferred shape of theionizing UV continuum, we can calculate that the BLR gas must have a highcovering factor of ∼
40% (Gaskell, Klimek, & Nazarova 2007). The absenceof observed absorption from the BLR tells us that the BLR has a flatteneddistribution and that we must be viewing it from near the axis of symmetry(Antonucci et al. 1989; Gaskell, Klimek, & Nazarova 2007). BLR transfer func-tions (the responses to delta function continuum flares) also show that at leastthe low-ionization BLR has a flattened distribution with almost no materialnear the line of sight (Maoz et al. 1991; Krolik et al. 1991; Horne et al. 1991).Gaskell, Klimek, & Nazarova (2007) point out that near the equatorial planethe covering fraction must be close to 100%, so the BLR will be self-shielding.They show that this explains the strong observed radial ionization stratifica-tion. Gaskell, Klimek, & Nazarova (2007) argue that the BLR will shield thetorus as well so that the BLR and torus cover similar fractions of 4 π steradi-ans. The overall appearance of the torus and BLR can best be described as abird’s nest (Mannucci et al. 1992). A cartoon and visualizations can be foundin Gaskell, Goosmann, & Klimek (2008) and Gaskell (2009).
3. The direction of motion of the BLR
Inflow, outflow, and virialized motions in either random orbits or more planarKeplerian orbits in a disk have all been considered as possibilities for BLRmotions. The presence of broad absorption lines (BALs) in some AGNs provesthat some gas is outflowing. Spatially-resolved outflow of some of the narrow-line region can also be seen. However, BALs do not necessarily mean that theBLR is outflowing because BALs commonly extend to velocities several timeshigher than those observed for the BLR in the same objects (see, for example,Turnshek et al. 1988). It is thus not clear that there is necessarily any connectionbetween BALs and BLRs. I will argue here that BLR kinematics are consistentwith the “bird’s” nest geometry and that most of the BLR gas is not outflowing.The question of the direction of motion can be settled by determining whichsides of the black hole the blueshifted and redshifted wings of the line profilesarise from. It has long been recognized that the most unambiguous way of doingthis for emission-line gas is through velocity-resolved light echoes (e.g., Fabrika1980; Ulrich et al. 1984) but early results were ambiguous. Extensive “reverber-ation mapping” began with the introduction of the cross-correlation technique(Gaskell & Sparke 1986; Gaskell & Peterson 1987) which has now been used todetermine the effective radii of dozens of BLRs from line and continuum timeseries. The technique can also be applied to the red and blue wings of a BLRline to determine the direction of motion (Gaskell 1988). If a line arises from aradiatively-driven outflow we can confidently predict that the peak of the cross-correlation function of the red-wing and blue-wing time series will show the bluewing leading the red by about twice the light-crossing time. With only a coupleof exceptions (discussed in section 5) this is never seen.Gaskell (1988) found that for NGC 4151 the wings of both the high-ionizationC IV line and the low-ionization Mg II line varied almost simultaneously, but the red wings of both C IV and Mg II led the blue wings by ∼ ∼ . nflow of the Broad-Line Region . ± .
03 days. The profiles of Mg II and H β generally agree inAGNs (Grandi & Phillips 1979) and their widths are well correlated from objectto object (e.g., Gaskell & Mariupolskaya 2002), so we expect similar kinematicsfor the two lines. Maoz et al. (1991) found the red wings of both the broad H α and H β lines to be leading their respective blue wings by 4 – 5 days (see theirFig. 13) in good agreement with the Gaskell (1988) Mg II result.Similar behavior has been found in other objects. Koratkar & Gaskell(1989) found the same combination of virialized motion plus inflow in Fairall 9.For both C IV and Mg II pure outflow was again excluded at a high confidencelevel and the red wing of C IV led the blue wing by 75 ±
45 days (note thatthe BLR of Fairall 9 is more than 10 times bigger than that of NGC 4151).Examination of
IUE spectra of NGC 5548 taken over the 10-year period of1978–1988 excluded pure radial outflow of the high-ionization BLR at > ± Inter-national AGN Watch ( IAW ) campaign showed the red wing leading by 5 ± IAW campaign (Korista et al. 1995) showed the red wing leading by 4 . ± . ±
300 km s − (about a third of the line FWHM). For H β Kollatschny & Dietrich (1996) found the red wing leading the blue wing by 5 ± IAW optical spectra from 1988-1993 showed the red wing of H β leading the blue wing during this period by5 . ± . ±
29 days. Even thoughin early studies the significance of the inflow signature was not often not high,the combined significance was high ( > all objects were consistent with virialized motions plus infall. Thisresult has strengthened considerably with monitoring of more objects. These in-clude 3C 390.3 (O’Brien et al. 1998; Dietrich et al. 1998), Mrk 6 (Sergeev et al.1999), Akn 120 (Doroshenko et al. 1999, 2008), Mrk 110 (Kollatschny 2003),Mrk 40 (Bentz et al. 2008), NGC 3227, and NGC 3516 (Denney et al. 2009),and Mrk 1310, NGC 4748, and NGC 6814 (Bentz et al. 2009). For 3C 390.3 thered wings of Ly α and C IV led the corresponding blue wings by 19 and 17 ± β led by 14 . ± . β leads the blue. For Mrk 110 Kollatschny (2003) the wings at ± − show the red sides leading the blue by ∼
13 days for H β , ∼ Gaskell L a g ( d a y s ) Mrk 6 L a g ( d a y s ) Mrk 6 L a g ( d a y s ) Mrk 6 L a g ( d a y s ) Radial velocity (km/sec)
Mrk 6 L a g ( d a y s ) Radial velocity (km/sec)
Mrk 6 L a g ( d a y s ) Radial velocity (km/sec)
Mrk 6
Figure 1. The lag relative to the optical continuum of H β emission in Mrk 6as a function of radial velocity for 1993–1995. Note (a) that the red wingresponds before the blue wing, and (b) that the high-velocity gas at bothpositive and negative velocities responds faster than the slower velocity gas.The dashed line shows the lag of the whole line. Data from Sergeev et al.(1999). λ ∼ λ β led by 4 ± ∼
4. Accretion and the blueshifting of high-ionization lines
Gaskell (1982) discovered that the high-ionization C IV line was blueshiftedwith respect to the low-ionization BLR lines and proposed that this was aresult of the high-ionization lines arising in a wind while the low-ionizationlines arose from a disk hiding the receding side of the wind. Strong blueshift-ings of high-ionization lines have subsequently been widely interpreted as ev-idence of strong winds (e.g., in Narrow-Line Seyfert 1 galaxies (NLS1s)– seeLeighly & Moore 2004). The big problem with the disk-wind explanation, how-ever, is that velocity-resolved reverberation mapping of C IV shows no evidence ofoutflow . Gaskell & Goosmann (2008) show that blueshifting from scattering offinfalling material solves this problem, and that the observed infall velocities andgeometry naturally explain the shifted profiles. It has long been recognized thatif the BLR is inflowing, the mass influx is sufficient to account for AGN energygeneration (Padovani & Rafanelli 1988). In NLS1s the accretion is higher so the nflow of the Broad-Line Region
5. The limitations of reverberation mapping
Despite the substantial evidence for BLR infall in addition to virialized motion,a significant fraction of monitoring campaigns have not shown a clear infall sig-nature. Examples include NGC 4593 (Kollatschny & Dietrich 1997), NGC 3227and NGC 5548 (Denney et al. 2009), Mrk 1310, NGC 4748, and NGC 6814(Bentz et al. 2009). The 1989 NGC 5548 monitoring reveals what is going on.For the whole campaign Kollatschny & Dietrich (1996) found the same infallsignature for the Balmer lines that Crenshaw & Blackwell (1990) had found forC IV, but for the first outburst alone , both H α and H β showed a strong outflow signature! Given this, it is not surprising that Welsh et al. (2007) found substan-tially different inflow signatures for different observing seasons. For example, in1990 the blue/red lag was 0 . ± . . ± . not meanthat the BLR is outflowing. Conversely, a very strong inflow signature (e.g., inMrk 40; Bentz et al. 2008) does not necessarily mean strong inflow.I believe the problems arise because real AGNs are messy objects! Differentparts of line profiles show different responses to observed continuum variability(e.g., Sergeev et al. 2001; Shapovalova et al. 2004). For Mrk 6 Sergeev et al.(1999) pointed out that some changes in the Balmer line profiles shapes cannotbe caused by matter redistribution or light-travel time effects, but are probablycaused by changes in the anisotropy of the ionizing continuum. An anisotropicionizing continuum could arise in several ways. Gaskell & Klimek (2003) andGaskell (2006) argue that variable components of AGN continuum emission athigh energies are likely to be intrinsically anisotropic. Even if the intrinsicvariability is not anisotropic, X-ray observations show that large variations inthe absorbing column are common (Risaliti et al. 2002; see Turner & Miller 2009for an extensive review) – especially when viewing the system near edge on, asthe BLR is doing. These changes can be large and rapid. NGC 1365, forexample, changed from being Compton-thick to Compton-thin (Risaliti et al.2005) in only a few weeks! There will be a huge effect on the response of aregion of BLR gas as optically-thick material passes through its line of sight toan ionizing continuum source. A third source of anisotropy is non-axisymmetricvariability of the disk which has already been proposed as a cause of differingreverberation responses to differing outbursts (Gaskell 2008).I believe that it is these deviations from the standard reverberation-mappingassumption of an isotropic, centrally-located ionizing continuum source that arenow limiting reverberation mapping campaigns rather than poor sampling orpoor signal-to-noise ratios . In particular there is a significant risk of gettingerroneous reverberation mapping results from observing campaigns of short du- Gaskell ration where the results are dominated by single events.
It is more important tohave longer monitoring campaigns rather than denser sampling.
Acknowledgments.
I am grateful to Ren´e Goosmann and Ski Antonuccifor discussions, and to the NSF for support through grant AST 08-03883.
References
Antonucci, R. R. J., Kinney, A. L., & Ford, H. C. 1989, ApJ, 342, 64Bentz, M. C. et al. 2008, ApJ Lett, 689, L21Bentz, M. C. et al. 2009, ApJ submitted [arXiv:0908.0003]Bochkarev, N. G. & Gaskell, C. M. 2009, Ast. Lett., 35, 287Clavel, J. et al. 1991, ApJ, 366, 64Crenshaw, D. M. & Blackwell, J. H., Jr. 1990, ApJ. Letts., 358, L37Denney, K. D. et al. 2009, ApJ Letts., 704, L80Dietrich, M. et al. 1998, ApJS, 115, 185Done, C. & Krolik, J. H. 1996, ApJ, 463, 144Doroshenko, V. T. et al. 1999, Astron. Lett., 25, 569Doroshenko, V. T., Sergeev, S. G., & Pronik, V. I. 2008, Astron. Rep., 52, 442Fabrika, S. N. 1980, Astron. Tsirk., 1109, 1Gaskell, C. M. 1982, ApJ, 263, 79Gaskell, C. M. 1988, ApJ, 325, 114Gaskell, C. M. 2006, Astron. Soc. Pacific Conf. Ser., 360, 111Gaskell, C. M. 2008, Rev. Mexicana Astron. Ap. Conf. Ser., 32, 1Gaskell, C. M. 2009, New Ast. Rev., 53, 000 [arXiv:0908.0386]Gaskell, C. M. & Goosmann, R. W., ApJ, submitted [arXiv:0805.4258]Gaskell, C. M., Goosmann, R. W., & Klimek, E. S. 2008, Mem. Soc. Astron. Italiana,79, 1090Gaskell, C. M. & Klimek, E. S. 2003, Astron. Ap. Trans., 22, 611Gaskell, C. M., Klimek, E. S., & Nazarova, L. S. 2007, ApJ, submitted [arXiv:0711.1025]Gaskell, C. M. & Mariupolskaya, V. Y. 2002, ASP Conf. Ser., 255, 261Gaskell, C. M. & Peterson B. M., 1987, ApJS, 65, 1Gaskell, C. M. & Snedden, S. A. 1997, BAAS, 29, 1252Gaskell, C. M. & Sparke, L. S. 1986, ApJ, 305, 175Grandi, S. A. & Phillips, M. M. 1979, ApJ, 232, 659Horne, K., Welsh, W. F., & Peterson, B. M. 1991, ApJ Letts, 367, L5Kollatschny, W. 2003, A&Ap, 407, 461Kollatschny, W. & Dietrich, M. 1996, A&A, 314, 43Kollatschny, W. & Dietrich, M. 1997, A&A, 323, 5Koratkar, A. P. & Gaskell, C. M. 1989, ApJ, 345, 637Koratkar, A. P. & Gaskell, C. M. 1991a, ApJ, 375, 85Koratkar, A. P. & Gaskell, C. M. 1991b, ApJS, 75, 719Korista, K. T. et al. 1995, ApJS, 97, 285Krolik, J. H. et al. 1991, ApJ, 371, 541Leighly, K. M. & Moore, J. R. 2004, ApJ, 611, 107Mannucci, F., Salvati, M., & Stanga, R. M. 1992, ApJ, 394, 98Maoz, D. et al. 1991, ApJ, 367, 493O’Brien, P. T. et al. 1998, ApJ, 509, 163Padovani, P., & Rafanelli, P. 1988, A&Ap, 205, 53Risaliti, G., Elvis, M., Fabbiano, G., Baldi, A., & Zezas, A. 2005, ApJ Lett, 623, L93Risaliti, G., Elvis, M., & Nicastro, F. 2002, ApJ, 571, 234Sergeev, S. G., Pronik, V. I., Sergeeva, E. A., & Malkov, Yu. F. 1999, ApJS, 121, 159Sergeev, V. I., Pronik, S. G., & Sergeeva, E. A. 2001, ApJ, 554, 245Shapovalova, A. I. et al. 2004, A&A, 422, 925Turner, T. J. & Miller, L. 2009, Astron. Ap. Rev., 17, 47 nflow of the Broad-Line Region7