ARlogo Annu. Rev. Astron. Astrophys. 2004. 42: 603-683
Copyright © 2004 by Annual Reviews. All rights reserved

Next Contents Previous

4.9. Nuclei

Nuclei are compact star clusters 5 AGN light is expressly excluded. Caution: Not all authors do this. at galactic centers (see Kormendy & Djorgovski 1989 for a review). They should not be confused with steep density cusps that are the central parts of nearly featureless power-law profiles (Lauer et al. 1995). E.g., M 32 does not have a nucleus (Lauer et al. 1992, 1998). Rather, nuclei are clearly differentiated from the surrounding (pseudo)bulge and disk in the sense that they have much smaller effective radii re and much higher effective surface brightnesses µ(re) than their surroundings. Figure 19 illustrates the prototypical example in M 33 (Kormendy & McClure 1993). Tremaine & Ostriker (1982) show that the nucleus and bulge of M 31 are dynamically independent; we assume the same for other nuclei. Local Group examples occur in M 31 (e.g., Lauer et al. 1993, 1998), M 33, and NGC 205 (Jones et al. 1996). Not much farther away are nuclei in IC 342 (Böker et al. 1997, 1999) and in NGC 7793 (Díaz et al. 2002). The literature on nuclei is extensive; we review only the part of it that is relevant to this paper.

Late-type disk galaxies usually contain nuclei. Many of these have young stellar populations. They imply episodic star formation over long periods of time and so are consistent with secular growth in a manner that is similar to the proposed formation of pseudobulges. But the "smoking gun" that is most compelling is not the one that points at secular evolution. Rather, the observations are screaming that there is another physical process taking place that we do not understand. Because we will see that nuclei do not appear to be the low-luminosity limit of pseudobulges. Also, nuclei and pseudobulges often occur in the same galaxy. If they form similarly, why are they so different? Our review highlights a remarkable list of enigmas.

Surveys of late-type galaxies and detailed studies of individual objects provide the following list of properties:

  1. Nuclei very common in late-type spirals. Böker et al. (2002) find them in 75 % of 77 Scd - Sm galaxies in an I-band HST survey. Carollo et al. (2002) find nuclei in 30 % of S0 - Sa galaxies, 59 % of Sab - Sb galaxies, and 77 % of Sbc - Sm galaxies. Nuclei are also common in spheroidal galaxies (Binggeli et al. 1984, 1985; van den Bergh 1986).

  2. Nuclei are rare in irregulars (van den Bergh 1995); this is not understood.

  3. Nuclei are fairly homogeneous in their properties. Typical luminosities are 106 to 107 Lodot (Kormendy & McClure 1993; Matthews et al. 1999a; Carollo et al. 2001; Matthews & Gallagher 2002; Böker et al. 2002). Effective radii are typically 101 ± 0.5 pc (Carollo et al. 1999; Böker et al. 2003b). Observed central densities are high and limited by the spatial resolution of the images: 104 to 105 Lodot pc-3 in examples in Matthews et al. (1999a) and at least and 107 Lodot pc-3 in M 33 (Lauer et al. 1998). These values are enormously higher than the surrounding disk densities.

  4. Most nuclei are at the centers of their galaxies to within measurement errors (Böker et al. 2002). Exceptions are rare (Matthews et al. 1999a; Binggeli, Barazza, & Jerjen 2000; Carollo et al. 2002). This is hard to understand. In their absence, the center does not look like a special place. The gravitational potential gradient of the visible matter is shallow (Kormendy & McClure 1993; Matthews et al. 1999a; Böker et al. 2002). Why does a nucleus form at the center and why is its scale length so short compared to that of the rest of the galaxy (Figure 19; Böker, Stanek, & van der Marel 2003)? Could the reason be that cold dark matter halos are cuspier than the baryons (Navarro, Frenk, & White 1996, 1997), or, if they are not so now (Moore 1994), could they have been so in the past, before baryonic physics intervened (Navarro, Eke, & Frenk 1996)? Could nuclei be compact not because the galactic center is a special place but rather because galaxies know how to make compact clusters and they can sink to the center by dynamical friction (Tremaine, Ostriker, & Spitzer 1975)? Carollo (1999) argues that the timescale for dynamical friction against dark matter is interestingly short. This may explain the "seeds" of the observed nuclei, but it remains remarkable that subsequent star formation keeps them so compact.

  5. Stellar populations often imply young ages for the stars that contribute most of the light. The M 33 nucleus is typical. It has a composite, late-A to early-F spectrum dominated by younger stars at bluer wavelengths (e.g., van den Bergh 1976a, 1991; Gallagher, Goad, & Mould 1982; O'Connell 1983; Schmidt, Bica, & Alloin 1990; Gordon et al. 1999). Population synthesis by Long, Charles, & Dubus (2002) gives a best fit to the spectrum between 1150Å and 5700Å for two starbursts, one with a mass of 9000 Modot, 40 Myr ago and the other with a mass of 76,000 Modot, 1 Gyr ago. (The total nuclear mass is 2 × 106 Modot; Kormendy & McClure 1993). The spectra are insensitive to still older starbursts.

    Additional examples of nuclei with blue colors indicative of young stars are discussed in Díaz et al. (1982); Bica, Alloin, & Schmidt (1990); Shields & Filippenko (1992); Böker et al. (1997, 1999) Matthews et al. (1999a); Davidge & Courteau (2002); Böker et al. (2001, 2003b); see also Ho, Filippenko, & Sargent (1997). Carollo et al. (2001) observe colors that are consistent with a range of ages; there is some tendency for bluer nuclei to be associated with bluer surrounding disks. In fact, "brighter nuclei (MV ltapprox -12) are typically found ... in the centers of galaxies with circumnuclear rings/arms of star formation or dust and an active, i.e., H II or AGN-type ... spectrum" (Carollo 1999). Since many nuclei contain young stars, star formation does not happen over only a small fraction of the life of the cluster but rather is secular.

  6. Nuclei are not more common in barred galaxies (Carollo et al. 2002; Böker et al. 2003b). Evidently, supplying gas to feed their star formation does not require a bar.

  7. In the Fundamental Plane parameter correlations, nuclei are more similar to large Galactic globular clusters and to compact young clusters in interacting and merging galaxies than they are to (pseudo)bulges (Carollo 1999; Geha, Guhathakurta, & van der Marel 2002; Böker et al. 2003b). There is no sign that nuclei form the faint end of the sequence of (pseudo)bulge properties (see also Section 4.10).

  8. The luminosities of nuclei correlate with the luminosities and central surface brighnesses of their host galaxies (Böker et al. 2003b).

What does all this mean? Point (5) - the prevalence of young stars in nuclei and their correlation with the surrounding star formation - is the strongest evidence that nuclei are built by secular processes like those that we suggest make pseudobulges. Point (8) also seems consistent. So is observation (1) that nuclei are more common in later-type galaxies; they are approximately as common as pseudobulges. However, observations (2), (3), (4), (6), and (7) either are major puzzles or suggest that nuclei and pseudobulges are fundamentally different.

The prudent conclusions at present are these: Nuclei are not a problem for our picture of pseudobulge formation by secular inward transport of gas. In fact, many authors quoted above - including one author of this paper - have argued that this is how they grow. But nuclei are not a secure argument for secular evolution, either. We find it compelling that nuclei and pseudobulges are very different in their parameters but occur together in the same galaxies. Nuclei appear to be related to globular clusters and to young clusters in merger starbursts. Several mysteries would be easier to understand if they got their start as such clusters and then sank to the center by dynamical friction. In particular, our problem with observation (3, 4) that nuclei are tiny and dense compared to pseudobulges and disks would vanish. This does not mean that we understand how star-forming clusters get so compact. But galaxies clearly know how to make them. Also, they are common in late-type disks, so it is easy to engineer a correlation with pseudobulges. Further work is needed to clarify how much nuclei then grow by secular evolution.


5 AGN light is expressly excluded. Caution: Not all authors do this. Back.

Next Contents Previous