As summarized by Kormendy (2004), spatially resolved kinematical observations have convincingly measured BH masses in a sizable number of nearby galaxies, to the point that important inferences on their demographics can be drawn (Richstone 2004). Following an argument first due to Sotan (1982), comparison of the integrated radiation density from quasars to the integrated mass density in local BHs shows that BHs have grown mostly via a radiatively efficient mode of accretion during their bright AGN phase (e.g., Yu & Tremaine 2002). Nearby galaxies, therefore, should be home to AGN relics.
The LLAGNs summarized in this review provide an important confirmation of this basic prediction. Not only are weakly accreting BHs found in great abundance in the local Universe, but they are found where prevailing wisdom says that they should be found, namely in the centers of galaxies that contain bulges. Among E, S0, and Sb galaxies, the AGN detection rate is ~ 50%, increasing to over 70% among Sa galaxies. Since sensitivity and confusion impact the detection rates, these statistics are not inconsistent with the notion that BHs are ubiquitous in essentially all E-Sb galaxies.
Notably, the incidence of AGNs plummets for galaxies with Hubble types Sc and later, precisely at the point where classical (r1/4 profile) bulges effectively cease to exist and secular dissipation processes begin to kick in (Kormendy & Kennicutt 2004). Although the AGN fraction is low for late-type spirals, it is also not zero. Careful scrutiny of this minority population addresses two important questions: (1) Are there central (nonstellar) BHs with masses below 106 M? (2) Must central BHs always be encased in a bulge?
Two remarkable galaxies give the clearest testimony that low-mass BHs do, in fact, exist. Within the Palomar survey, the nearby (~ 4 Mpc) galaxy NGC 4395 contains all the usual attributes of a self-respecting AGN: broad optical and UV emission lines (Filippenko & Sargent 1989; Filippenko, Ho & Sargent 1993), a compact radio core (Ho & Ulvestad 2001) of high brightness temperature (Wrobel, Fassnacht & Ho 2001; Wrobel & Ho 2006), and rapidly variable hard X-ray emission (Shih, Iwasawa & Fabian 2003; Moran et al. 2005). Contrary to expectations, however, NGC 4395 is an extremely late-type (Sdm) spiral (Figure 11, left), whose central stellar velocity dispersion does not exceed ~ 30 km s-1 (Filippenko & Ho 2003). If NGC 4395 obeys the MBH - relation, its central BH should have a mass 105 M. This limit agrees surprisingly well with the value of MBH estimated from its broad H line width or X-ray variability properties (~ 104-105 M; Filippenko & Ho 2003), or from reverberation mapping (3.6 × 105 M; Peterson et al. 2006). POX 52 (Figure 11, right) presents another interesting case. As first noted by Kunth, Sargent & Bothun (1987), the presence of a Seyfert-like nucleus in POX 52 is unusual because of the low luminosity of the host galaxy. Barth et al. (2004) show that POX 52 bears a close spectroscopic resemblance to NGC 4395. Based on the broad profile of H, these authors derive a virial BH mass of 1.6 × 105 M for POX 52, again surprisingly close to the value of 1.3 × 105 M predicted from the MBH - relation given the measured central stellar velocity dispersion of 35 km s-1.
Figure 11. Two examples of AGNs in late-type galaxies. The left panel shows an optical image of NGC 4395, adapted from the Carnegie Atlas of Galaxies (Sandage & Bedke 1994); the image is ~ 15' (17 kpc) on a side. The right panel shows an HST I-band image of POX 52, adapted from C.E. Thornton, A.J. Barth, L.C. Ho, R.E. Rutledge, J.E. Greene (in preparation); the image is ~ 11" (5 kpc) on a side.
The two objects highlighted above demonstrate that the mass spectrum of nuclear BHs indeed does indeed extend below 106 M, providing great leverage for anchoring the MBH - at the low end. Furthermore, they shed light on the variety of environments in which nuclear BHs can form, providing much-needed empirical clues to the conditions that fostered the formation of the seeds for supermassive BHs. NGC 4395 has little or no bulge, but it does have a compact, central star cluster in which the BH is embedded (Filippenko & Ho 2003). Interestingly, G1, a massive star cluster in M31 that to date contains the best direct detection of an intermediate-mass (~ 20,000 M) BH (Gebhardt, Rich & Ho 2002, 2005; Ulvestad, Greene & Ho 2007), is thought to be the stripped nucleus of a once small galaxy. The same holds for the Galactic cluster Cen, for which Noyola, Gebhardt & Bergmann (2008) report a central dark mass of 5 × 104 M. POX 52 is equally striking. Deep images reveal POX 52 to be most akin to a spheroidal galaxy (Barth et al. 2004; C.E. Thornton, A.J. Barth, L.C. Ho, R.E. Rutledge, J.E. Greene, in preparation), to date an unprecedented morphology for an AGN host galaxy. This is quite unexpected because spheroidal galaxies, while technically hot stellar systems, bear little physical resemblance to classical bulges. Spheroidals occupy a distinct locus on the fundamental plane (e.g., Geha, Guhathakurta & van der Marel 2002; Kormendy et al. 2008), and they may originate from harassment and tidal stripping of late-type disk galaxies (e.g., Moore et al. 1996). Thus, like NGC 4395, POX 52 stands as testimony that a classical bulge is not a prerequisite for the formation of central BH.
But how common are such objects? AGNs hosted in high-surface brightness, late-type spirals appear to be quite rare in the nearby Universe. Within the comprehensive Palomar survey, NGC 4395 emerges as a unique case of an unambiguous broad-line AGN hosted in a late-type system. The majority of late-type spirals do possess compact, photometrically distinct nuclei (Böker et al. 2002), morphologically not dissimilar from that in NGC 4395, but, with few exceptions (Shields et al. 2008), these nuclei are compact star clusters with no compelling evidence for an accompanying accreting central BH (Walcher et al. 2006). Nuclear star clusters do not appear to directly impact a galaxy's ability to host an AGN (Seth et al. 2008). Several serendipitous cases of AGNs in late-type galaxies have recently been found from analysis of Spitzer mid-IR spectra (Satyapal et al. 2007, 2008), as well as a number of AGN candidates from inspection of Chandra images (Desroches & Ho 2008). Among earlier Hubble types, Gallo et al. (2008) report the detection of X-ray nuclei in two low-luminosity early-type galaxies.
To assess the true incidence of AGNs like NGC 4395 and POX 52 requires a spectroscopic survey much larger than Palomar. Greene & Ho (2007b) performed a systematic spectral analysis of over 500,000 SDSS spectra with z < 0.35 to search for broad-line AGNs, producing not only a detailed BH mass function for low-redshift AGNs (Greene & Ho 2007a) but also a comprehensive catalog of ~ 200 low-mass (MBH < 106 M) objects. Not much is known yet about the host galaxies, except that on average they are about 1 mag fainter than L*. HST imaging of the initial sample of 19 objects discovered by Greene & Ho (2004) reveal that the host galaxies are either mid- to late-type spirals (although none seems as late-type as NGC 4395) or compact, spheroidal-looking systems not unlike POX 52 (Greene, Ho & Barth 2008). When projected onto the galaxy fundamental plane, the "bulge" component in some systems resides within the locus of spheroidal galaxies. Follow-up high-dispersion spectroscopy shows that these objects approximately follow the MBH - relation (Barth, Greene & Ho 2005).