4.1. Structural parameters
Here and in the next subsection we will discuss the structural parameters and colors of bulge dominated LSB galaxies and compare them to disk dominated LSB galaxies and HSB galaxies. We will focus on trends to explore the question whether bulge dominated LSB galaxies fit in with the general trends defined by HSB galaxies, and more importantly, whether they form the "missing link" between HSB and giant LSB galaxies.
The majority of disk dominated LSB and HSB galaxies has disk scale lengths between 2 and 6 kpc (McGaugh & Bothun 1994; de Blok 1997; dJ95). The galaxies in our sample have much larger disk scale lengths and the largest galaxies also have bulges. There appears a trend that the longest disk scale lengths appear in galaxies with the longest bulge scale lengths.
We use the Pearson correlation coefficient to determine the significance of the correlation and find r = 0.59. We thus find a restricted range for the bulge-to-disk ratios at the 99.8% probability level. The correlation between B band disk and bulge scale length is shown in Fig. 5 and is consistent with other studies (dJ95; Courteau et al. 1996).
However, the trend could be partly due to the selection criteria used, because we are discriminating against large bulge, small disk galaxies. Galaxies with compact, bright bulges and faint extended disks would comply to the criteria, but we do not have them in our sample. Furthermore, there are no large, pure disk, LSB galaxies known till now, so it is likely that only the area below the plotted trend is affected by selection effects. A correlation between disk and bulge scale length suggests that the formation of bulge and disk is coupled.
The ratio of disk-to-bulge scale length for both HSB and LSB galaxies has large scatter around and is illustrated in Fig. 6. The LSB galaxies continue the trend defined by HSB galaxies towards lower surface brightnesses.
In Fig. 7 we plot the distribution of disk central surface brightness with disk scale length for the samples of dJ95, dB95, Sprayberry and the current sample. All samples fit in with the general trend that there are no galaxies with high central surface brightnesses and large disk scale lengths. Some of the galaxies in our sample have such large disk scale lengths that they are found in the region of giant LSB galaxies and therefore could be classified as such.
To investigate whether disk and bulge central surface brightnesses are related we plot these parameters in Fig. 8 for the samples of dJ95 and our sample. The figure clearly shows how the LSB galaxies fill the low surface brightness region, but split into two groups; one near = 20 mag arcsec-2 and one near = 22 mag arcsec-2. These groups do not correspond to the two types of bulge LSB galaxies mentioned in Sect. 2, but the groups are a mixture of both these types. There seems to be a very broad general tendency for both HSB and LSB galaxies to have brighter bulge central surface brightnesses with increasing disk central surface brightnesses.
The color profiles (Figs. 1, bottom panels) show that in most galaxies the outer parts are bluer than the inner parts. It was shown by dB95 that this is also true for late-type LSB galaxies. It is hard to draw conclusions about the colors of the bulges judging from the color profiles alone. But one can reveal differences between disk and bulge colors by comparing bulge-to-disk (B/D) ratios in different wavelength bands. We determined B/D ratios by comparing the total light output of bulge and disk per passband. We noticed a clear tendency for the B/D ratios to increase towards the redder wavelengths (Table 4). This means that the bulges of LSB galaxies are redder than their disks, confirming a trend also observed for HSB galaxies (dJ95).
Table 4. Bulge-to-disk ratios in the different bands.
In Table 5 we present the integrated colors of the galaxies in our sample and those of disk dominated LSB (de Blok 1997) and giant LSB galaxies (Sprayberry et al. 1995). Prior to determining the mean nuclear color of the galaxies in our sample galaxies without a clear bulge were excluded. For the determination of the mean area weighted colors the galaxies without a clear bulge were included. The systems in our sample have redder area weighted colors than disk LSB galaxies and this would be more pronounced if the more or less bulgeless galaxies were left out of the sample.
Table 5. Comparison of integrated colors.
Fig. 9 shows the distribution of B-V color with disk scale length (top) as well as the distribution of B-V color (center) and B-R color (bottom) with disk central surface brightness. There is no trend of central surface brightness with color and the colors do not depend on size. HSB galaxies cannot be the progenitors of LSB galaxies since galaxies fade and redden . The bluest galaxies are the disk dominated LSB galaxies and are concentrated in a rather small area whereas the HSB galaxies scatter over the entire color range towards the redder colors. The bulge dominated LSB galaxies fill up the region between these two samples. The large scatter in color for the bulge dominated LSB sample is due to the wider range in morphological types. The disk dominated LSB galaxies are quiescent and form a fairly uniform sample. The galaxies in our sample have red bulges, but some of them also have bars and (blue) rings making the spread in color larger than for a more uniform sample.
Some of the large bulge dominated LSB galaxies in our sample could easily be classified as giant LSB galaxies, judging from their sizes and luminosities. It is therefore instructive to compare the colors with those of giant LSB galaxies. The disk colors of the biggest galaxies in our sample are significantly bluer than giant disks (Table 5) and the B/D ratios are smaller than for typical giants (Sprayberry et al. 1995). So although they have bulges and very large low surface brightness disks, they are not as evolved as the giants.
The comparison of our sample with a large sample of typical "Freeman galaxies" (de Jong & van der Kruit 1994) not only shows that on average the bulge dominated LSB galaxies are bluer, but also that one can distinguish a LSB from a HSB galaxy of the same type by its bluer color.
4.3. Bulge formation scenarios for LSB galaxies
We will take a more detailed look at the two types of bulge LSB galaxies as described in Sect. 2 to see if any systematic differences show up in their properties. We will also try to make some rough quantitative statements about the bulges of the galaxies in our sample.
Fig. 10 shows the relation between various galaxy properties as derived from the photometry for both types of LSB galaxies present in our sample. In the two top panels both types follow roughly the expected trend of having redder nuclear colors than area weighted colors. In the bottom-left panel it is seen that the bulges do not follow the same trend as the disks; there are galaxies with large bulge scale lengths and high bulge central surface brightnesses. The galaxies with bars appear to have relatively high bulge central surface brightnesses, while the ones with round bulges show a large(r) spread. The bottom-right panel shows the division in high and lower luminosity galaxies, but both types are mixed.
On the basis of disk or bulge colors, scale lengths, surface brightnesses or absolute magnitudes no clear distinction can be made between the two types.
The correlation between bulge and disk scale lengths suggests that the formation of these two components is closely coupled. The restricted range of the ratio between bulge and disk scale lengths of is used as an argument for secular evolution models (Courteau et al. 1996). In this picture bulges form from small dynamical instabilities of disks which induce formation of bars. Secular accretion or satellite accretion in the center in turn can dissolve the bar into a spheroidal component (Norman et al. 1996). If this scenario is correct then the bulges are relatively young.
However, the observed color gradients suggest that at larger radii the emitted light is dominated by a relatively young population of stars. This and the fact that the bulges are redder than the disks favors a scenario in which the bulge forms before the disk. It is hard to pin down the exact causes of the color gradients without metallicity and extinction information. Andredakis (1998) has shown that a correlation between bulge and disk scale lengths also quite naturally arises if the bulge were to be formed before the disk. If this scenario is correct then the bulges are relatively old.
By examining the bottom panels of Figs. 1 we see that the R-I gradient is in general less steep than the B-V and B-R gradients and the R-I profile is in most cases fairly constant. Comparison of bulge and disk colors shows that the color difference is smallest in R-I . This is reflected in the fact that the differences in B/D ratios are also smallest between R and I band for most of the galaxies in our sample. This could indicate that the old stellar populations in the bulge and disk are of approximately the same age. But without knowledge of metallicity, extinction, initial mass function, and/or star formation history it is hard to make any definite statements about the difference in age between bulge and disk. The effects that differences in these properties have on the colors of disk galaxies are extensively discussed in van den Hoek et al. (1998). Peletier & Balcells (1996) conclude from their study of optical and near-infrared colors of 30 galaxies of types S0 to Sbc that the inner disks are only slightly younger than bulges. The difference in age generally lies somewhere between 0 and 3 Gyrs.
A natural question that arises is whether the bulges of LSB galaxies are different than those of HSB galaxies. To explore this question we compared the bulge parameters, central surface brightness (R band) and scale length, with a sample of 33 Sb,Sc type HSB galaxies (Andredakis 1997, see upper panel of Fig. 11). The bulge colors were compared with another sample of 30 S0-Sbc type galaxies (Peletier & Balcells 1996, see bottom panel of Fig. 11). Andredakis' sample uses r band photometry (Thuan & Gunn 1976), but that has no effect on our conclusions. Peletier & Balcells take the bulge color at 0.5 or at 5", whichever is larger and the disk color at 2 K band scale lengths. We take the nuclear color as bulge color and the color at 2 R band disk scale lengths as disk color. No distinction can be made between bulges of HSB and LSB galaxies on basis of their structural parameters. The distribution of bulge colors of the LSB galaxies has a larger spread, but on average the bulges of HSB and LSB galaxies have similar colors. LSB disks are bluer making the differences between bulge and disk colors larger for LSB galaxies.
We note that all our conclusions are based on a small sample of LSB galaxies and it would clearly help if data for more bulge LSB galaxies would become available.
© European Southern Observatory (ESO) 1999
Online publication: November 16, 1999