6 Global properties of SBNGs

 

6.1 Morphological types, distances, magnitudes and environment

The global properties of our sample of SBNGs are displayed in Fig. 10, which presents the distribution of morphological types (according to RC3), inclinations, heliocentric radial velocities, distances, blue apparent and absolute magnitudes of the galaxies, using the values tabulated in Tables 1 and 4. In these diagrams, we show the global properties of SBNGs after removing the five HII galaxies and galaxies with ambiguous nuclear spectral classification. The global properties of AGNs (Seyfert 1 and 2, LINERs) are also shown for comparison.

The mean inclination of the galaxies is $\sim$45$^\circ$$\pm$ 16, both for SBNGs and AGNs. This value is comparable to that derived for barred galaxies with HII nuclei (Ho et al. 1997b). The heliocentric radial velocity of SBNGs is predominantly ($\sim$93%) lower than or equal to $10\,000$ km s^-1 with mean values around 4300 and 7000 km s^-1 for SBNGs and AGNs respectively. The SBNGs are located at a mean distance of 68 $\pm$ 38 Mpc, slightly farther than nearby HII nuclei ($\sim$22 Mpc; Ho et al. 1997b. The mean distance derived for the AGNs is 107 $\pm$ 50 Mpc. The mean value of the apparent blue magnitude is 14, with a large proportion ($\sim$90%) of giant spiral galaxies ($M_{\rm abs}$ $\leq -19.5$). The mean value of the absolute magnitude ($\sim -20.2$) is identical for our sample of SBNGs and for nearby HII nuclei (Ho et al. 1997b).

The SBNGs are equally distributed among early-type (S0-a to Sb; 55%) and late-type (Sbc to Sm; 42%) galaxies with only 3% elliptical galaxies as is expected in a sample of starburst galaxies. In fact, the proportion of SBNGs increases with the morphological type and reaches a maximum of 40% for Sbc/Sc galaxies. The distribution of AGNs is on the contrary more uniform from S0 to Sc. One can also note the deficiency of SBNGs with morphological types later than Sc, which confirms the low contamination of our sample by low-luminosity blue compact and irregular HII galaxies.

Coziol et al. (1995) found a majority of early-type galaxies among their sample of SBNGs, but this result should be considered with caution because only 39% of their galaxies are morphologically classified. On the contrary, Ho et al. (1997b) found a majority of late-type galaxies (62% of Sc-Sm) among their sample of HII nuclei. They note however that this effect is pronounced in barred galaxies (65%) whereas unbarred galaxies with HII nuclei are equally divided between early and late types. This may explain the relatively high frequency of Sc galaxies found in our sample of SBNGs since all our galaxies are barred.

  

Figure 11: a) Distribution of reddening-corrected H$\alpha$ luminosity for individual nuclear (solid line) and extranuclear HII region (dotted line). b) Distribution of total (nuclear and extranuclear) reddening-corrected H$\alpha$ luminosity for the SBNGs (solid line) and AGNs (dotted line). Mean values are indicated by vertical arrows

One of the most popular ideas for explaining powerful starbursts is that they must occur preferentially within galaxies undergoing gravitational interactions. We thus performed an analysis of environment and level of interaction for all our galaxies using CCD images (Contini 1996). We found that the majority (62%) of SBNGs are isolated galaxies. Only two galaxies (Mrk 617 and 960) are advanced mergers, 6% of the galaxies belong to close pairs (projected distance $\leq$ 1$^\prime$) and 23% to wide pairs (projected distance $\leq$ 15$^\prime$ and $\Delta V \leq$ 300 km s^-1). More than half of the SBNGs do not show any sign of past or present gravitational interaction. Asymmetries in the bar or spiral arms are observed in only 32% of the galaxies; among them, 12% have multiple bright knots along the bar. This does not indicate that bars are necessary for triggering starbursts in the absence of interactions; other samples of (barred and unbarred) SBNGs (Coziol et al.1997b) and HII galaxies (Telles & Terlevich 1995) have also been found to contain a low proportion ($\sim$20 to 25%) of interacting galaxies. Interactions are more frequent among luminous infrared galaxies. The level of interactions increases with the FIR luminosity, the proportion of mergers reaching a maximum among ultra-luminous infrared galaxies (Veilleux et al.1995).

6.2 H$\alpha$ luminosity

  

Figure 12: a) Distribution of total FIR luminosity for the SBNGs (solid line) and AGNs (dotted line). b) IRAS color-color diagram for SBNGs (open squares) and AGNs (filled squares), c) and d) Distributions of f25/f60 and f60/f100 IRAS colors for SBNGs (solid line) and AGNs (dotted line). Mean values are indicated by vertical arrows

In this section and the next, we compare the distribution of the H$\alpha$ and FIR luminosities of our sample of SBNGs with those derived for other samples of starburst galaxies. A more detailed and quantitative discussion on the relation between H$\alpha$, blue and FIR luminosities, and on the distribution of H$\alpha$ equivalent widths in terms of star formation history and age of the starbursts is given in Contini et al. (in preparation).

It appears clearly in Fig. 11a that the average H$\alpha$ luminosity [log(L(H$\alpha$)/erg s^-1)] is higher (by a factor $\sim$10) in starburst nuclei $(41.0\, \pm \,0.7)$ than in extranuclear HII regions (40.2 $\pm$ 0.6). However, the H$\alpha$ luminosity of the HII regions, which are mainly located along the bar of our galaxies, is higher than that of typical disk HII regions (39.5; Kennicutt et al. 1989). The H$\alpha$ luminosities estimated in our starburst nuclei are typical of starburst galaxies (40.7; Balzano 1983). They are clearly higher than in nearby HII nuclei (39.2; Ho et al. 1997a) but slightly lower than in starbursts in luminous infrared galaxies (42.0; Veilleux et al.1995). Contrary to what occurs in nearby HII nuclei (Ho et al. 1997a), we do not find any significant difference between the total H$\alpha$ luminosities of early-type (41.3 $\pm$ 0.4) and late-type (41.0 $\pm$ 0.7) SBNGs.

In terms of H$\alpha$ luminosity, our sample of SBNGs is thus intermediate between nearby HII nuclei and luminous infrared galaxies. These luminosities are typical of starburst galaxies and comparable to other samples of SBNGs (i.e. Coziol et al.1994).

As shown in Fig. 11b, the total (nuclear and extranuclear) H$\alpha$ luminosity derived for SBNGs (41.2 $\pm$ 0.6) is slightly lower than that derived for AGNs (41.7 $\pm$ 0.5). These luminosities are very close to those observed in other samples of Seyfert galaxies ($\sim$42.0; Dahari & De Robertis 1988; Veilleux et al.1995).

6.3 FIR properties

We computed the FIR luminosities of the galaxies from the IRAS flux densities at 60 and 100 $\mu$m (Table 1) using the following relation which approximates well the total FIR luminosity between 42 and 122 $\mu$m (Helou et al.1988)

$$\log(L_{\rm FIR}) = 5.5378 + 2\log(D) + \log(2.58f_{60} + f_{100})$$ (1)

where the flux densities at 60 and 100 $\mu$m are expressed in Janskys, D is the distance of the galaxy in Mpc and $L_{\rm FIR}$ is the FIR luminosity in solar units.

The distribution of FIR luminosities [log($L_{\rm FIR}$/$L_{\odot}$)] is shown in Fig. 12a. Our sample of SBNGs has moderate FIR luminosities (10.1 $\pm$ 0.5), slightly higher than those observed in samples of HII nuclei ($\sim$9.4) and HII galaxies ($\sim$8.9), similar to other samples of SBNGs ($\sim$9.9) (see Coziol 1996 and references therein) but rather low compared to samples of luminous ($\sim$11) or ultra-luminous ($\sim$12) infrared galaxies (Veilleux et al.1995). There is no significant difference between the average FIR luminosity of the SBNGs and AGNs (10.4 $\pm$ 0.4). The AGNs in our sample follow the trend observed in infrared-bright galaxies, their proportion increases with FIR luminosity and reaches a maximum of 62% for $\log(L_{\rm FIR}/L_{\odot}) \geq 12$ (Veilleux et al.1995).

We did not use the fluxes at 12 and 25 $\mu$m to compute the FIR luminosities, because of the strong contribution of non-thermal radiation to dust heating at these wavelengths. This well-known phenomenon (e.g. Miley et al. 1985; de Grijp et al.1985) is illustrated in Figs. 12c,d where we show the distribution of the two IRAS colors, f₂₅/f₆₀ and f₆₀/f₁₀₀, for both SBNGs and AGNs. While no difference is seen in the distribution of f₆₀/f₁₀₀ (mean value $\sim$0.55) for AGNs and starburst galaxies, a clear excess of emission at 25 $\mu$m is observed for AGNs [$\log(f_{25}/f_{60}) = -0.5\,\pm \,0.2$] when compared to SBNGs [$\log(f_{25}/f_{60}) = -0.8\, \pm \,0.1$]. This indicates that the infrared emission at short wavelengths (12 and 25 $\mu$m) is mainly due to a "warm dust" component heated by the non-thermal ionizing radiation from AGNs. Figure 12b also shows that, regardless of the spectral classification, there is a clear tendency for galaxies with "warmer" 60/100 colors to have "cooler" 12/25 colors, illustrating the need for a multicomponent model to describe the nature of IRAS infrared emission (e.g. Helou 1986). Such a model requires the presence of a "warm dust" component of infrared emission associated with star formation regions, and a "cool dust'' component associated with the neutral interstellar medium.