Kormendy & Richstone (1995) review BH search techniques and summarize the ground-based detections. Recent reviews (e.g., Richstone et nuk. 1998) concentrate on BH astrophysics. There has not been a comprehensive review of BH discoveries made with the Hubble Space Telescope (HST), so we provide a summary here.
There are ground-based BH detections in 10 galaxies, including the nearest (our Galaxy, M31, and M32) and the best (our Galaxy and NGC 4258) candidates. References are given in Kormendy & Richstone (1995) and Table 1. Of the 7 stellar-dynamical cases, 6 have been reobserved with HST. In all cases, the BH detection was confirmed and the ground-based BH mass agrees with the HST result to within a factor of ~ 2. The HST papers are: M31: Statler et al. (1999), Bacon et al. (2001); M32: van der Marel et al. (1998); NGC 3115: Kormendy et nuk. (1996a), Emsellem et al. (1999); NGC 3377: Richstone et nuk. (2001), NGC 4486B (Green et al. 2001), and NGC 4594: Kormendy et nuk. (1996b).
Our Galaxy is the strongest BH case, based on observations of velocities
in the plane of the sky of stars in a cluster within
0."5 = 0.02 pc of the central radio source Sgr A*
(Eckart & Genzel
1997;
Genzel et al.
1997,
2000;
Ghez et al. 1998,
2000).
The fastest star is moving at 1350 ± 40 km
s-1. Acceleration vectors have been measured for three stars;
they intersect, to within the still-large errors, at Sgr A*, supporting the
identification of the radio source with the inferred central mass of
M
= (2.6 ± 0.2) × 106
M
(Ghez et al.
2000).
The stellar orbital
periods could be as short as several decades, so we can look forward
to seeing the Galactic center rotate in our lifetimes! Most important,
the mass
M
is
constrained to live inside such a small radius that alternatives
to a supermassive black hole are ruled out by astrophysical
constraints. Brown
dwarf stars would collide, merge, and become luminous, and clusters of white
dwarf stars, neutron stars, or stellar-mass black holes would evaporate too
quickly
(Maoz 1995,
1998;
Genzel et al.
1997,
2000).
The next-best BH case is NGC 4258. In it, a water maser disk shows
remarkably Keplerian rotation velocities inward to a radius of 0.2 pc
(Miyoshi et al.
1995).
The implied central mass,
M =
4 × 107
M
, again
is confined to a small enough volume to exclude the above BH alternatives
(Maoz 1998).
Such arguments cannot yet be made for any other galaxy. Nevertheless,
they increase our confidence that all of the dynamically detected
central dark objects are BHs.
The BH search has now largely moved to HST. With the aberrated HST, BH work was based on indirect arguments that have serious problems (Kormendy & Richstone 1995). But with COSTAR, HST has become the telescope of choice for BH searches, and the pace of detections has accelerated remarkably.
The HST era is divided into two periods. Before the installation of the
Space Telescope Imaging Spectrograph (STIS) in 1997, the main instrument
used was the Faint Object Spectrograph (FOS). It was inefficient,
because it used an aperture instead of a slit. Nevertheless, the first
HST BH detections were made with the FOS (M87:
Harms et al.
1994;
NGC 4261:
Ferrarese et al.
1996;
NGC 7052:
van der Marel & van
den Bosch 1998).
It is often suggested that
HST was required to make BH cases convincing. This is an exaggeration. HST
beats ground-based resolution by a factor of 5, but the first BH
detections made with HST were in Virgo cluster galaxies or in ones that
are 2 - 4 times farther
away. Virgo is ~ 20 times farther away than M31 and M32.
Therefore the ground-based BH discoveries in M31 and M32 had better
spatial resolution (in pc) than the HST BH discoveries in the above
galaxies. Of course, the distant BHs have higher masses. Therefore a
better measure of relative resolution is the ratio of the radius
rcusp = G
M
/
2 of the BH
sphere of influence to the resolution. Table 1 lists
rcusp for all BH detections. Since the PSF in the
ground-based discovery observations had a radius of
~ 0."3 - 0."5 while the FOS observations used a 0."09
square aperture
(NGC 4261),
Table 1 shows that the FOS BH detections in
M87, NGC 4261, and NGC 7052 had comparable or lower relative resolution
than the ground-based observations of M31, M32, NGC 3115, and NGC 4594. As HST spatial
resolution improved (especially with STIS), BH cases have indeed gotten
stronger. But the main thing that HST has provided is many more detections.
STIS has begun a new period in the BH search. With the efficiency of a
long-slit spectrograph and CCD detector, the search has become feasible for
most nearby galaxies that have unobscured centers and old stellar
populations. It is still not easy; finding a
106-M BH is difficult at the
distance of the Virgo cluster and impossible much beyond. But the
pace of
discoveries has accelerated dramatically. At the 2000 Summer AAS meeting, 14
new BH detections were reported, and several more have been published
since. As a result, about 37 BH candidates are now available. We say
"about" because not all cases are equally strong: which ones to include
is a matter of judgment. Table 1 provides a census.
Galaxy | Type | MB,bulge | M![]() |
![]() |
D | rcusp | Reference |
(M![]() |
(km/s) | (Mpc) | (arcsec) | ||||
Galaxy | Sbc | -17.65 | 2.6 (2.4-2.8) e6 | 75 | 0.008 | 51.40 | See notes |
M 31 | Sb | -19.00 | 4.5 (2.0-8.5) e7 | 160 | 0.76 | 2.06 | Dressler + 1988; |
Kormendy 1988a | |||||||
M 32 | E2 | -15.83 | 3.9 (3.1-4.7) e6 | 75 | 0.81 | 0.76 | Tonry 1984, 1987 |
M 81 | Sb | -18.16 | 6.8 (5.5-7.5) e7 | 143 | 3.9 | 0.76 | Bower + 2001b |
NGC 821 | E4 | -20.41 | 3.9 (2.4-5.6) e7 | 209 | 24.1 | 0.03 | Gebhardt + 2001 |
NGC 1023 | S0 | -18.40 | 4.4 (3.8-5.0) e7 | 205 | 11.4 | 0.08 | Bower + 2001a |
NGC 2778 | E2 | -18.59 | 1.3 (0.5-2.9) e7 | 175 | 22.9 | 0.02 | Gebhardt + 2001 |
NGC 3115 | S0 | -20.21 | 1.0 (0.4-2.0) e9 | 230 | 9.7 | 1.73 | Kormendy + 1992 |
NGC 3377 | E5 | -19.05 | 1.1 (0.6-2.5) e8 | 145 | 11.2 | 0.42 | Kormendy + 1998 |
NGC 3379 | E1 | -19.94 | 1.0 (0.5-1.6) e8 | 206 | 10.6 | 0.20 | Gebhardt + 2000a |
NGC 3384 | S0 | -18.99 | 1.4 (1.0-1.9) e7 | 143 | 11.6 | 0.05 | Gebhardt + 2001 |
NGC 3608 | E2 | -19.86 | 1.1 (0.8-2.5) e8 | 182 | 23.0 | 0.13 | Gebhardt + 2001 |
NGC 4291 | E2 | -19.63 | 1.9 (0.8-3.2) e8 | 242 | 26.2 | 0.11 | Gebhardt + 2001 |
NGC 4342 | S0 | -17.04 | 3.0 (2.0-4.7) e8 | 225 | 15.3 | 0.34 | Cretton + 1999a |
NGC 4473 | E5 | -19.89 | 0.8 (0.4-1.8) e8 | 190 | 15.7 | 0.13 | Gebhardt + 2001 |
NGC 4486B | E1 | -16.77 | 5.0 (0.2-9.9) e8 | 185 | 16.1 | 0.81 | Kormendy + 1997 |
NGC 4564 | E3 | -18.92 | 5.7 (4.0-7.0) e7 | 162 | 15.0 | 0.13 | Gebhardt + 2001 |
NGC 4594 | Sa | -21.35 | 1.0 (0.3-2.0) e9 | 240 | 9.8 | 1.58 | Kormendy + 1988b |
NGC 4649 | E1 | -21.30 | 2.0 (1.0-2.5) e9 | 375 | 16.8 | 0.75 | Gebhardt + 2001 |
NGC 4697 | E4 | -20.24 | 1.7 (1.4-1.9) e8 | 177 | 11.7 | 0.41 | Gebhardt + 2001 |
NGC 4742 | E4 | -18.94 | 1.4 (0.9-1.8) e7 | 90 | 15.5 | 0.10 | Kaiser + 2001 |
NGC 5845 | E | -18.72 | 2.9 (0.2-4.6) e8 | 234 | 25.9 | 0.18 | Gebhardt + 2001 |
NGC 7457 | S0 | -17.69 | 3.6 (2.5-4.5) e6 | 67 | 13.2 | 0.05 | Gebhardt + 2001 |
NGC 2787 | SB0 | -17.28 | 4.1 (3.6-4.5) e7 | 185 | 7.5 | 0.14 | Sarzi + 2001 |
NGC 3245 | S0 | -19.65 | 2.1 (1.6-2.6) e8 | 205 | 20.9 | 0.21 | Barth + 2001 |
NGC 4261 | E2 | -21.09 | 5.2 (4.1-6.2) e8 | 315 | 31.6 | 0.15 | Ferrarese + 1996 |
NGC 4374 | E1 | -21.36 | 4.3 (2.6-7.5) e8 | 296 | 18.4 | 0.24 | Bower + 1998 |
NGC 4459 | SA0 | -19.15 | 7.0 (5.7-8.3) e7 | 167 | 16.1 | 0.14 | Sarzi + 2001 |
M 87 | E0 | -21.53 | 3.0 (2.0-4.0) e9 | 375 | 16.1 | 1.18 | Harms + 1994 |
NGC 4596 | SB0 | -19.48 | 0.8 (0.5-1.2) e8 | 136 | 16.8 | 0.22 | Sarzi + 2001 |
NGC 5128 | S0 | -20.80 | 2.4 (0.7-6.0) e8 | 150 | 4.2 | 2.26 | Marconi + 2001 |
NGC 6251 | E2 | -21.81 | 6.0 (2.0-8.0) e8 | 290 | 106 | 0.06 | Ferrarese + 1999 |
NGC 7052 | E4 | -21.31 | 3.3 (2.0-5.6) e8 | 266 | 58.7 | 0.07 | van der Marel + 1998 |
IC 1459 | E3 | -21.39 | 2.0 (1.2-5.7) e8 | 323 | 29.2 | 0.06 | Verdoes Kleijn + 2001 |
NGC 1068 | Sb | -18.82 | 1.7 (1.0-3.0) e7 | 151 | 15 | 0.04 | Greenhill + 1996 |
NGC 4258 | Sbc | -17.19 | 4.0 (3.9-4.1) e7 | 120 | 7.2 | 0.36 | Miyoshi + 1995 |
NGC 4945 | Scd | -15.14 | 1.4 (0.9-2.1) e6 | 3.7 | Greenhill + 1997 | ||
Notes - BH detections are based on
stellar dynamics (top group), ionized gas dynamics (middle)
and maser dynamics (bottom). Column 3 is the B-band absolute
magnitude of the bulge part of the galaxy. Column 4 is the BH mass
M |
An important HST contribution has been to enable BH searches based on ionized gas dynamics (middle part of Table 1). The attraction of gas is simplicity - unlike the case of stellar dynamics, velocity dispersions are likely to be isotropic and projection effects are small unless the disks are seen edge-on. Especially important is the fact that gas disks are easy to observe even in giant ellipticals with cuspy cores. These galaxies are a problem for stellar-dynamical studies: they are expensive to observe because their surface brightnesses are low, and they are difficult to interpret because they rotate so little that velocity anisotropy is very important. It is no accident that most BH detections in the highest-luminosity ellipticals are based on gas dynamics. Without these, we would know much less about the biggest BHs.
At the same time, the uncertainties in gas dynamics are often
underestimated. Most studies assume that disks are cold and in circular
rotation. But the gas masses are small, and gas is easily pushed around. It
would be no surprise to see velocities that are either slower or faster than
circular. Faster-than-circular motions can be driven by AGN or starburst
processes, while motions that are demonstrably slower than circular are
observed in many bulges (see
Kormendy & Westpfahl
1989).
A separate issue is the large emission-line widths seen in many
galaxies. If these are due to pressure support, then the observed
rotation velocity is less than the circular velocity
and M
is underestimated if the line width is ignored. The situation
is like that in any stellar system that has a significant velocity
dispersion, and the cure is similar. In the context of a not-very-hot
disk like that in our Galaxy, the correction from observed to circular
velocity is called the "asymmetric drift correction", and in the context
of hotter stellar systems
like ellipticals, it is handled by three-integral dynamical models. For gas
dynamics, the state of the art is defined by
Barth et al.
(2001),
who discuss the line broadening problem in detail. They point out that
asymmetric drift corrections may be large or they may be inappropriate
if the line width is due
to the internal microturbulence of gas clouds that are in individual, nearly
circular orbits. We do not understand the physics of line broadening, so it
adds uncertainty to BH masses. But it is likely that
M
will be underestimated if the line width is ignored. In contrast,
Maciejewski & Binney
(2001)
emphasize that
M
can be overestimated by as much as a factor of three if we neglect
the smearing effects of finite slit sizes. The best gas-dynamic
M
estimates (e.g.,
Sarzi et al.
2001;
Barth et al. 2001)
are thought to be
accurate to ~ 30%. In future, it will be important to take all of
the above effects into account. It is not clear a priori whether they
are devastating or small. The best sign that they are manageable is the
observation that stellar- and gas-dynamical analyses imply the same
M
correlations (compare the squares and circles in
Figure 2).
Gas-dynamical BH searches are limited mainly by the fact that suitable gas
disks are rare.
Sarzi et al.
(2001)
found gas disks with well-ordered, nearly
circular velocities in only about 15% of their sample of galaxies that were
already known to have central gas. So
10% of a complete
sample of bulges is likely to have gas disks that are usable for BH
searches. Nevertheless, within the next year, we should have
gas-kinematic observations of
30-40 galaxies from a variety of groups. They will provide a wealth of
information both on nuclear gas disks and on BHs.