3 Properties of dwarf galaxies, GCs, and cD halo in the Fornax cluster center

The center of the Fornax cluster hosts the central galaxy NGC 1399 with an extraordinarily rich globular cluster system and an extended cD halo as well as a halo of X-ray emitting gas. In the following we give a short review on the properties of the different components that have to be considered in the picture of a common evolution. In Table 1 those properties are summarized: the slopes of the surface density profiles, the velocity dispersion, and the ranges of metallicities. Furthermore, the absolute V luminosities and estimated masses are given, if available.

  

Table 1: Overview of different parameters of Fornax components, as the population of dwarf elliptical and dwarf S0 galaxies, the central GCS, the unresolved stellar light of NGC 1399 and its cD halo, and the gas component in the central region of the cluster. The assumed distance modulus is (m - M)₀ = 31.3 mag. The profile slope $\alpha$ is the exponent in the density law $\rho \propto r^\alpha$,where $\rho$ is either the projected number density or the projected surface brightness profile

3.1 Dwarf galaxies in the Fornax cluster

The most complete investigation of the Fornax dwarf galaxies was done by Ferguson (1989, Fornax Cluster Catalog (FCC)) as well as by Davies et al. (1988, and following papers: Irwin et al. 1990; Evans et al. 1990). As we have shown in Paper I the morphological classification of Fornax members by Ferguson (1989) is very reliable and nearly no dE has been missed within the survey limits as far as we can judge from the comparison with our sample fields. Thus, the following properties of the Fornax dwarf galaxies are mainly based on the FCC plus the additional new members as presented in Paper I.

  

Table 2: Power law slopes of the surface density profiles of Fornax galaxies taken from the FCC (Ferguson 1989)

The spatial distribution of dEs in Fornax can be represented by a King profile with a core radius of $0\hbox{$.\!\!^\circ$}67\pm0\hbox{$.\!\!^\circ$}1$ and a center located about $25\hbox{$^\prime$}$ west of NGC 1399 (Ferguson 1989). In order to compare their surface density profile with that of the GCS and the cD halo light we fitted power laws to the radial distribution of the dEs and dS0s in the extended FCC adopting NGC 1399 as the center. For that we counted galaxies brighter than $B_{\rm T} =$ 19 mag in 7 equi-distant rings from 0 to $3\hbox{$^\circ$}$. We determined the slopes of the density profiles in the inner ($r < 0\hbox{$.\!\!^\circ$}8$) as well as in the outer ($0\hbox{$.\!\!^\circ$}8 < r < 3\hbox{$^\circ$}$) part. The dividing radius of $0\hbox{$.\!\!^\circ$}8$ is about the limit out to where the cD halo light and the gas envelope have been measured. The results are summarized in Table 2. In addition, we also give the mean slopes, when fitting a power law to the total profile, and the fitted values for the giant galaxies. The nucleated dwarf galaxies have the steepest slope and are more concentrated towards the central galaxy than the non-nucleated dE/dS0s.

  

Table 3: Faint end slopes of the luminosity function for different subsamples of the Fornax galaxy population

The luminosity function (LF) of the Fornax dwarf galaxies was studied by Ferguson & Sandage (1988) in a region with radius smaller than $2\hbox{$.\!\!^\circ$}4$, centered on NGC 1399. They found that the nucleated dwarf ellipticals (dE,Ns) as well as the dwarf lenticular (dS0) galaxies are brighter than the non-nucleated dEs. Further, the faint end slope of the dE/dS0 LF, fitted by a Schechter (1976) function, is quite flat ($\alpha = -1.08\pm0.10$) compared to other clusters like Virgo ($\alpha = -1.31\pm0.05$) or Centaurus ($\alpha = -1.68\pm0.56$, Jerjen & Tammann 1997). Table 3 summarizes the results for the faint end slopes of Schechter function fits to different subsamples of the extended FCC.

Colors and metallicities of dwarf galaxies in Fornax have been studied by photometric as well as by spectroscopic means (e.g. Caldwell & Bothun 1987; Bothun et al. 1991). Spectroscopically determined metallicities seem to be consistent with the picture that the bluer dwarfs are the more metal-poor ones. The metallicity range for 10 bright dE,Ns is $-1.5 < {\rm [Fe/H]} < -0.8$ dex (Held & Mould 1994). The metallicities derived from Washington photometric indices for 15 LSB dwarfs are of the same order (Cellone et al. 1996). Concerning ages, all investigated dwarfs possess an old stellar population, some of them a contribution of intermediate-age stars, and only few have signs of recent or ongoing star formation (Held & Mould 1994; Cellone & Forte 1996). It seems that the Fornax dEs share the same characteristics as the Local Group dSph population (e.g. review by Grebel 1997).

Radial velocity measurements of 43 Fornax dwarfs ($18 \gt B_{\rm t} \gt 15$ mag) by Drinkwater et al. (1997) result in a velocity dispersion of $\sigma_v = 490$ km s^-1, significantly larger than that of 62 giants ($B_{\rm t} < 15$ mag), $\sigma_v =$ 310 km s^-1. According to the authors, this difference cannot be explained by measurement errors.

3.2 The central globular cluster system

The globular cluster system of NGC 1399 is one of the best investigated GCSs outside the Local Group. The total number of GCs is about $N_{\rm GC} = 5800\pm500$ (Kissler-Patig et al. 1997; Grillmair et al. 1998) within a radius of $10\hbox{$^\prime$}$ from the galaxy center. This is about 10 times the number of GCs in the other Fornax ellipticals, $300\pm60 < N_{\rm tot} < 700 \pm 100$.Adopting a distance of 18.2 Mpc or (m-M)₀ = 31.3 mag to NGC 1399 (Kohle et al. 1996, recalibrated with new distances of Galactic GCs, Gratton et al. 1997) the absolute magnitude of NGC 1399 is M_V = -21.75 mag when taking the apparent magnitude values from the literature (Faber et al. 1989, RC3: de Vaucouleurs et al. 1991). This corresponds to a specific frequency of $S_N = 11.6\pm2.0$. If the light of the cD halo within $10\hbox{$^\prime$}$is taken into account (see Sect. 3.3), S_N is reduced to $6.8\pm2.0$ ($M_{V,{\rm tot}} =$ -22.33 mag). However, distinguishing a cD halo and a bulge component in the galaxy light, S_N for the cD halo would be about $10\pm1$assuming, S_N=3.2 for the bulge, see Sect. 9, the average value of the other early-type Fornax galaxies (Kissler-Patig et al. 1997). Thus, the building up of the GCS of the cD halo component must have been very efficient.

The color distribution of the GCs around NGC 1399 is very broad compared to most other GCSs in Fornax ellipticals and can only be explained by a multimodal or perhaps just a bimodal GC population (e.g. Ostrov et al. 1993; Kissler-Patig et al. 1997, and Forbes et al. 1997). Spectroscopic analysis of 18 GCs by Kissler-Patig et al. (1998) shows a metallicity range between -1.6 and -0.3 dex (with possible peaks at -1.3 and -0.6 dex), and two exceptional GCs at about 0.2 dex, located in the red (metal rich) tail of the color distributions. The comparison of the line indices with theoretical evolutionary models suggests that most of the GCs are older than at least 8 Gyr. If one fits the GC color distribution with two Gaussians, the number ratio of metal rich (red) to metal poor (blue) GCs is about 1:1 (Forbes et al. 1997).

The radial extension of the GCS around NGC 1399 can be traced out to about $10\hbox{$^\prime$}$ ($\simeq 53$ kpc). The slope of the GC surface density profile, $\rho \propto r^\alpha$, is about $\alpha = -1.5\pm0.2$, when taking the average of the published values. (Forbes et al. 1997) found that the distribution of the blue GC subpopulation is even flatter ($\alpha \simeq -1.0 \pm 0.2$), whereas the red GCs are more centrally concentrated ($\alpha \simeq -1.7 \pm 0.2$), comparable to the slope of the galaxy light ($\alpha = -1.6\pm0.1$). See Fig. 2 for a schematic overview.

Radial velocities of 74 GCs around NGC 1399 have been measured (Kissler-Patig et al. 1999; Minniti et al. 1998; Kissler-Patig et al. 1998). The velocity dispersion for the whole sample is $\sigma_v = 373\pm35$ km s^-1. No differences can be seen between the red and blue subpopulations. However, there exists a radial dependence of the velocity dispersion in the sense that $\sigma_v$ rises from $263 \pm 92$ to $408\pm107$ km s^-1 between $2\hbox{$^\prime$}$ and $8\hbox{$^\prime$}$ (Kissler-Patig et al. 1999).

3.3 cD halo and bulge

The galaxy light of NGC 1399 follows an extended cD profile (Schombert 1986; Killeen & Bicknell 1988) out to a radial distance of about 34 arcmin from the galaxy center ($\Sigma_B = 28$ mag isophotal surface brightness level). This is about 180 kpc in Fornax distance (18.2 Mpc) and comparable to the extent of the X-ray envelope (Ikebe et al. 1996; Jones et al. 1997).

  

Figure 1: The upper two panels show the surface brightness profile (open circles) of NGC 1399 plotted versus r1/4. The change of the slope to a flat cD halo is clearly visible. In the uppermost panel, the dashed line corresponds to the R1/4 law extrapolation of an aperture growth curve used to derive the apparent magnitude V = 9.55 mag (Faber et al. 1989, RC3). The largest aperture given by Burstein et al. (1984) is indicated. The dotted extension of the surface brightness profile at the faint end represents the slope between $2\hbox{$^\prime$}$ and $10\hbox{$^\prime$}$ by Killeen & Bicknell (1988). The dashed lines in the second panel represent the single components of the cD halo and the underlying bulge light, when fitting the sum of two de Vaucouleurs laws (solid line). The dotted lines give the ranges for possible fits. In the lowermost panel, r I(r) is plotted versus the radial distance to the center of NGC 1399. Again, the dashed curves describe the single components and the dotted ones the possible ranges. Within $10\hbox{$^\prime$}$ the total luminosity $I_{\rm tot}$ of the cD halo light is slightly higher than that of the bulge light

The determination of the stellar population parameters of the outer cD halo, like accurate photometric colors, metallicity or velocity dispersion, is very difficult due to the low surface brightness. Long slit spectra have been taken for the stellar bulge population within a radius of about $1\hbox{$.\mkern-4mu^\prime$}5$from the center of NGC 1399 (Franx et al. 1989; Bicknell et al. 1989). The velocity dispersion is about 200 km s^-1 at $1\hbox{$.\mkern-4mu^\prime$}5$ and rises within the central $10\hbox{$^{\prime\prime}$}$ to a central value of about 360 km s^-1. Besides the GCS, a useful tracer for the stellar population at larger radii is the population of planetary nebulae (PNe). Arnaboldi et al. (1994) studied the kinematics of 37 PNe out to a radius of $4\hbox{$.\mkern-4mu^\prime$}5$. They found an increase in the velocity dispersion with increasing radius from 269 km s^-1 for $r < 2\hbox{$.\mkern-4mu^\prime$}6$ to 405 km s^-1 for $2\hbox{$.\mkern-4mu^\prime$}6 < r < 4\hbox{$.\mkern-4mu^\prime$}5$ (18 of the 37 PNe).

3.3.1 Luminosity and surface brightness profile

In this subsection we divide the light profile of NGC 1399 into a cD halo and a bulge component in order to compare their characteristics with those of the GCS and the dwarf galaxy population. We determined the absolute luminosity of the cD halo in the following way: in the $\mu-r^{1/4}$ plot (Fig. 1, upper panels) one can see that the SB profile of NGC 1399 (determined from the NE CCD field F2) changes its slope at about $50\hbox{$^{\prime\prime}$}$. We fitted the total profile by the sum of two de Vaucouleurs laws:

$$\mu (r) = ZP_{\rm cal} - 2.5\cdot \log [I^0_{\rm gal}\cdot ... ...I^0_{\rm cD}\cdot\\ \exp(-(r/\alpha_{\rm cD})^{1/4})].\nonumber$$

The steeper, more concentrated profile represents the luminosity of the bulge without cD halo, whereas the flatter, more extended profile contains the light of the cD halo. The total luminosity of each component is $I^{\rm tot}_{\rm gal,cD} = 2\pi \int r I_{\rm gal,cD}(r){\rm d}r$. We restricted our calculations to within a radius of $10\hbox{$^\prime$}$ where the number of detected GCs fades into the background. The dashed lines in Fig. 1 represent the "best'' fit (data points inside $1\hbox{$.\!\!^{\prime\prime}$}5$ radius have been omitted). The dotted lines give the ranges for possible fits. The surface density slopes of both profiles are $\alpha = -2.0\pm0.2$ and $\alpha = -1.0\pm0.2$ respectively, if $\Sigma(r) \propto r^\alpha$. The slope of the combined profile is $\alpha = -1.6\pm0.1$ (see also Fig. 2).

In the literature one finds an apparent magnitude for NGC 1399 of V = 9.55 mag (Faber et al. 1989, RC3: de Vaucouleurs et al. 1991, adopting a mean (B - V) color of 1.0 mag, Goudfrooij et al. 1994). This magnitude is derived from an aperture growth curve extrapolation. with a maximum aperture of diameter $1\hbox{$.\mkern-4mu^\prime$}5$(Burstein et al. 1984). A 1-component fit of an R^1/4 law within $1\hbox{$.\mkern-4mu^\prime$}5$ (the largest aperture in Burstein et al. 1984) is shown in Fig. 1 (uppermost panel). Adopting an absolute magnitude of M_V = -21.75 mag for the integrated light under this profile, the total luminosity of the bulge light from the 2-component fit (middle panel) is $M_{V,{\rm bulge}} = -21.50\pm0.20$ mag (about 80% of the luminosity given in the literature). The luminosity for the cD halo is $M_{V,{\rm cD}} = -21.65\pm0.2$ mag and for the whole system within $10\hbox{$^\prime$}$ $M_{V,{\rm tot}} = -22.33 \pm$ 0.2 mag.

Another check for the correct proportion of the luminosities of the different components can be made by comparing the integrated flux within an aperture of $1\hbox{$.\mkern-4mu^\prime$}5$ with the total flux within $10\hbox{$^\prime$}$. Adopting V = 10.30 mag for the $1\hbox{$.\mkern-4mu^\prime$}5$aperture (Burstein et al. 1984), we derive $M_{V,{\rm tot}} = -22.27\pm0.2$ mag for the whole system, in excellent agreement with the value given above.

Note that the total luminosity of the cD halo is about 180 times the luminosity of a typical dwarf galaxy with M_V = -16.0 mag or 2.2 times the total luminosity of the present dEs and dS0s in Fornax.

  

Figure 2: Schematic overview of the surface density or brightness profiles of different subpopulations in the Fornax cluster and their extensions. The profiles are arbitrarily shifted in the ordinate axis. Also the units are arbitrarily chosen. The profiles of the GCs and galaxies are surface density profiles, the one of the galaxy light is a surface brightness profile, and the one of the X-ray gas is again a particle (number) density profile (see text for more details). The slopes $\alpha$ ($\rho \propto r^\alpha$) and their uncertainties are indicated on the right hand side