Annu. Rev. Astron. Astrophys. 1995. 33: 581-624 Copyright © 1995 by Annual Reviews. All rights reserved

5. GAS-DYNAMICAL BLACK HOLE SEARCHES

M87 (M_BH 3 x 10⁹ M)

M87 is the father of this subject. Well known for its radio and optical jet (Biretta 1993; Macchetto 1994), it has long been the prototypical AGN galaxy in which to test the BH paradigm. It was the first giant elliptical with CCD photometry; Young et al. (1978) found that it has a cuspy core. Sargent et al. (1978) showed that its velocity dispersion continues to rise inside the core radius from 278 km s^-1 at r = 10" to 350 km s^-1 at r 1".5. These observations could be fitted with isotropic dynamical models only if M87 contained an MDO of mass M_BH 3 x 10⁹ M.

But 1978 was the last year in which isotropy could be the default assumption. Illingworth (1977), Binney (1978), and others showed that most giant ellipticals are anisotropic. Models with _r > _tangential can reproduce the dispersion and brightness profiles of M87 without an MDO (Duncan & Wheeler 1980; Binney & Mamon 1982; Richstone & Tremaine 1985; Dressler & Richstone 1990). Also, almost all giant ellipticals that are close enough to show resolved cores have profiles like that of M87 (Lauer 1985a,b; Kormendy 1985, 1987a; HST photometry references in Section 4.1). If cuspy cores were enough to discover BHs, we would know many examples. Similarly, if central gas and dust disks perpendicular to jets (e.g., NGC 4261: Jaffe et al. 1993) were enough to reveal BHs, we would have many examples (Kotanyi & Ekers 1979; Kormendy & Stauffer 1987). Then this review could have been written seven years ago, and it would have been mercifully short.

The subsequent history of the BH search in M87 is reviewed by Sargent (1987), Richstone (1988), Filippenko (1988), Dressler (1989), Kormendy (1993), and van der Marel (1994a). Recently, HST has shown that I r^-0.26 near the center of M87 (Lauer et al. 1992a), exactly as predicted by Young et al. (1978). But this is still not sufficient evidence for a BH (Section 3). Finally, van der Marel (1994a) finds that 400 km s^-1 at r 0".5. But he can still explain this with anisotropic models. The stellar-dynamical BH case in M87 remains ambiguous.

There is hope for progress via LOSVDs. BHs make broad power-law wings that extend beyond the escape velocity of the stars (van der Marel 1994a,b). But the LOSVD wings are confused with continuum variations that we do not know how to model. That is, kinematic measurements and population synthesis are coupled. Exploiting LOSVDs will not be easy.

Gas, on the other hand, tells a clearer story. Using the FWPC2 and HST, Ford et al. (1994) have discovered that ionized nuclear gas in M87 looks like a disk with trailing spiral arms and a major axis that is perpendicular to the jet (Figure 12). Harms et al. (1994) used the HST Faint Object Spectrograph and 0".26 diameter aperture to get spectra centered at r = ± 0".25 = 19 pc along the disk major axis. At a luminosity-weighted mean radius or ± 0".22 = 16 pc, the spectra show emission lines separated by 2V = 916 km s^-1. If this gas is in circular motion and if the disk is inclined at 42° as implied by the apparent axial ratio, then M87 contains a central dark mass of M_BH = 3 x 10⁹ M (Table 1). This interpretion is supported by the following arguments, in decreasing order of importance. 1. Velocities at three other positions (r = 0".35, 0".56, and 2") agree with the Keplerian disk model. 2. At the center, the width of the emission lines is FWHM 1700 km s^-1. In fact, a large FWHM suggesting that M_BH 3 x 10⁹ M was already seen in ground-based data (van der Marel 1994a, and references therein). 3. The disk minor axis is within 11-20° of the position angle of the jet; the inclination of the disk is i = 42° ± 5°, nearly perpendicular to the jet inclination i = -40° inferred by Biretta (1993). 4. Finally, spiral structure implies disk shearing.

The caveat is that gas is easily pushed around. It may not be in circular orbits, in which case it cannot be used to measure M_BH. We emphasize that non-circular motions are not just an academic worry, they are seen in many active galaxies (e.g., Balick & Heckman 1982; Heckman et al. 1993) and non-AGN environments (e.g., Fillmore et al. 1986; Kormendy & Westpfahl 1989). The BH case in M87 can be strengthened (or falsified) by measuring V (r) closer to the center: 0".1 resolution is feasible, so a test for a Keplerian V (r) is possible.

Figure 12. HST H + [N II] image of M87 (Ford et al. 1994). The central ionized gas distribution is elongated perpendicular to the jet.

5.2 NGC 4258 (M_BH 4 x 10⁷ M)

NGC 4258 has an AGN of modest power (Table 1) but great interest. Its optical spectrum shows a low-ionization nuclear emission-line region (LINER, Heckman 1980) with a weak broad-line component (Filippenko & Sargent 1985). Between its normal spiral arms, NGC 4258 has ``anomalous'' arms seen in H (Courtès & Cruvellier 1961; Deharveng & Pellet 1970), radio synchrotron emission (van der Kruit et al. 1972; van Albada & van der Hulst 1982), and x-rays (Pietsch et al. 1994; Cecil et al. 1995). The H arms consist of braided components (Ford et al. 1986; Cecil et al. 1992); the velocity field is complicated (above papers; Rubin & Graham 1990; Dettmar & Koribalski 1990). The canonical explanation is that the nucleus emits jets in the disk plane; these interact with disk gas to produce the anomalous arm emission (above papers; Martin et al. 1989; Plante et al. 1991). NGC 4258 is therefore a good - albeit unusual - example of the BH paradigm.

A powerful new BH probe is provided by the detection of H₂O maser emission (Claussen et al. 1984; Claussen & Lo 1986) including components ± 900 km s^-1 from the systemic velocity (Nakai et al. 1993). Greenhill et al. (1995b) show that the systemic-velocity source consists of many masers along a line 0."00026 x 0".00005 = 0.009 x 0.002 pc in extent that is perpendicular to the inner radio jet. Velocities are tightly correlated with position: dV / dx = 270 km s^-1 (0".001)^-1 = 7430 ± 40 km s^-1 pc^-1 along the above line. Greenhill et al. (1995b), Watson & Wallin (1994), and Haschick et al. (1994) suggest that the masers are in a thin annulus of radius 0.1 pc. Individual maser velocities drift by ~ 9 km s^-1 y^-1, consistent with centripetal acceleration dV / dt = V² / r in the part of the annulus that is between us and the central source (Haschick et al. 1994; Greenhill et al. 1995a). Then the high-velocity components imply a rotation velocity of 900 km s^-1 and a mass M_BH = 2 x 10⁷ M. These conclusions are model-dependent, but they can be tested.

In an important paper, Miyoshi et al. (1995) find strong evidence in favor of the disk model. With the Very Long Baseline Array, they confirm that the high-velocity masers are located 0".005 to 0".008 on either side of the central line of sources. The masing physics suggests that they radially span the edge-on torus where it is tangent to the line of sight. Then the path length to us is long at constant velocity, allowing strong maser amplification. Based on this assumption, Miyoshi et al. (1995) derive the rotation curve shown in Figure 13. Remarkably, V (r) = 2180 (r / 0".001)^-1/2 km s^-1 = (832 ± 2) [r / (0.25 pc)]^-1/2 km s^-1 is Keplerian to high precision. This is a powerful argument for the annular disk model. If the rotation is circular, then the mass interior to 0".005 = 0.18 pc is M_BH = 4.1 x 10⁷ M.

If Figure 13 is correct, then the M_BH measurement is more secure than any previous one based on gas dynamics, including that of M87. This work is vulnerable on at least three counts: 1. The interpretation depends critically on the assumed geometry. The assumptions are conservative in the sense that they minimize M_BH. But the conclusion that V r^-1/2 depends on the geometry, too: if the maser sources are differently distributed along the line of sight, then V could vary differently r. Our confidence in the mass measurement depends on the Keplerian rotation curve. 2. As usual, it is not certain that gas velocities measure mass. 3. We do not know that the detected mass is dark. This work also has clear strengths: 1 The gas is molecular; being fragile, it is less likely to have noncircular velocities. 2. V (r) looks Keplerian. 3. In terms of (rotation velocity)/(speed of light), the measurements get closer to the BH than in any other candidate except M87. 4. There are no uncertainities about the velocity dispersion or its anisotropy. 5. The gas disk is perpendicular to the jet. 6. Finally, if M_BH were spread out over the central 0.18 parsec, the density would be ~ 1.6 x 10⁹ M pc^-3. This is 70 times higher than in the Galaxy (the closest competitor). On the other hand, the stellar density at the Galactic center is ~ 10⁸ M pc^-3 (Eckart et al. 1993). A 1.6 x 10⁹ M pc^-3 star cluster in not excluded, although it would evolve quickly and possibly violently.

If NGC 4258 is any indication, masers may be a potentially powerful tool for BH searches. They allow us to probe tiny radii in great physical detail. For example, acceleration and proper motion measurements can decisively test the NGC 4258 nuclear disk model. Because masers are preferentially found in AGN galaxies (Braatz et al. 1994), we can also look forward to closer connections between dynamical searches and other parts of the BH paradigm (Section 8).

Figure 13. Rotation curve for maser sources in NGC 4258, from Miyoshi et al. (1995). The line is the Keplerian V (r) given in the text.