General characteristics of our sample edge-on galaxies are presented in Table 1. The columns of the table are: galaxy name (see Paper I for galaxy identification); morphological type according to NED 1 ; adopted distance ( 75 km/s/Mpc); corrected for Galactic absorption absolute blue luminosity and colour of the galaxy (apparent magnitudes and colours of galaxies are from Paper I); total HI mass, in , according to Huchtmeier & Richter (1989), de Vaucouleurs et al (1991) (RC3) and LEDA (for the members of Arp 242 we used associated with the central galaxies HI masses according to Hibbard & van Gorkom 1996); HI mass-to-blue luminosity ratio, in / (adopting 5.48); average scale height in the I passband; scalelength to scale height ratio in the I passband. For two objects - Arp 208 and UGC 11230 - we found no available redshifts and present in Table 1 scale height values in arcseconds.
Table 1. General characteristics of the sample galaxies
Two galaxies - Arp 124NE and Arp 127S - are not presented in the table. Arp 124NE demonstrates notable deviation from edge-on orientation (we see dust lane shifted from the galaxy nucleus, photometric profiles along the minor axis look very asymmetric). Arp 127S is embedded in a relatively bright extended envelope which contribute significantly to the observed surface brightness distribution along the minor axis. Moreover, according to NED the components of Arp 127 demonstrate very different radial velocities (Arp 127N - 5065 km/s, Arp 127S - 13650 km/s) so the nature of this double system is unclear.
The distribution of the scale height values in three passbands as a function of position along the major axis is shown in Fig.1 for all sample galaxies with more than two measurements of .
3.1. Scale heights
3.1.1. Radial distribution of
It is well established that the scale parameter is almost independent of radial distance for the disks of normal spiral galaxies (vKS, Shaw & Gilmore 1990, dGvK). Our results for seven non-interacting late-type spiral galaxies confirm this conclusion. As one can see in Fig.1, normal spirals demonstrate small variations of scale height with radius. We find that the mean dispersion of values in the I band is 7% % only (note that this value refer to relatively bright region of the galaxy disk - see item 2.2). This constancy level is consistent with the level of 10-15% found in dGvK.
In agreement with literature data (e.g., Shaw & Gilmore 1990, dGvK), we did not find variations with passband for our sample isolated spiral galaxies. The mean ratio of scale heights in the I and B passbands is for seven isolated spiral galaxies.
Visual inspection of Fig.1 shows that interacting spirals demonstrate more variety in behaviours than isolated galaxies. The ratio of scale height dispersion to the mean value of is equal to % for them in the I passband. This is almost twice as for non-interacting galaxies. The distribution of values for our sample interacting and isolated galaxies is compared in Fig.2. The distribution for interacting spirals is significantly wider and about half of interacting spirals demonstrate larger scale height variations than it was found for isolated spirals.
The most frequent feature of radial behaviour of interacting spirals is a systematical change of scale height along the disk (see Fig.1). One can attribute about 40-50% of the galaxies to such "slanting" disks. Probably. "slanting" disk is a transient phase of galaxy evolution caused by large-scale asymmetry of the potential due to proximity of massive companion. Four galaxies (Arp 121, Arp 295, VV 679, and K 14) show an increase of scale height to the central region. Apparently, such disk structure reflects the presence of a bar in the galaxies.
Only one galaxy - VV 490N - demonstrate clearly the radial increase of the disk scale height. Such disk behaviour is expected for the galaxies subject to accretion of small companions (see Fig.1 in Toth & Ostriker 1992, Fig.6 in Quinn et al 1993, Fig.9 in Walker et al 1996). The double system VV 490 resembles remarkably the numerical models studied in those works - it consists of a large spiral galaxy (VV 490N) and a small companion (VV 490S) probably settling to the plane of the primary (Fig.1 in Paper I). The strong flaring of the VV 490N disk begins at 2h and reaches about 50% at 3h. This is in general quantitative agreement with results of numerical simulations. One can note also that analogous flaring structure of stellar disks was found recently in two interacting galaxies - NGC 3808B and NGC 6286 - subject to strong matter accretion from the companions (Reshetnikov et al 1996).
3.1.2. Mean scale heights of interacting and isolated galaxies
The average scale heights in the I passband for the sample galaxies are presented in Table 1. As for isolated spirals, we found no colour dependence of values for interacting galaxies also. The mean ratio of scale heights in the I and B passbands is for 27 galaxies.
We compare mean values of for interacting and normal ("field") galaxies in the Table 2. (Note that according to Karachentsev et al 1993 two non-interacting galaxies in our sample - UGC 11301 and UGC 11841 - are "Malin 1" type galaxies. We excluded them from further consideration.) As one can see, interacting galaxies show larger mean value of in comparison with our and vKS samples of isolated spirals. The BD sample gives significantly larger average scale height with large dispersion.
Table 2. Comparison of the samples
The cumulative distribution of scale height values for normal galaxies (we summarized our, vKS and BD samples) is compared with distribution for interacting spirals in Fig.3a,b. Both distributions show the same widths with global peaks at 1.5 kpc for interacting galaxies and about 1 kpc for field spirals. Trying to understand the origin of the secondary peak in Fig.3a at 2 kpc, we inspected normal galaxies falling in this region. Using the DSS 2 images, we found that among 8 normal galaxies with 2 kpc (7 of them are from the BD sample) at least 6 have comparable size companions within 5 optical diameters. Moreover, several galaxies demonstrate signs of significant non-edge-on orientation. For instance, the most striking galaxy in the BD sample - ESO 460-G31 - with 3.9 kpc has a companion at one optical radius and noteably shifted from the nucleus dust lane. Therefore, the secondary peak at Fig.3a consists of non-isolated galaxies mainly. Normal non-interacting galaxies possess disks with scale height about 1 kpc. This conclusion is in agreement with results obtained from the dGvK sample. According to de Grijs & van der Kruit (1996), the exponential scale height of thick disk of spiral galaxies range from about 470 pc to 620 pc. Transforming exponential scale heights to , we have range 0.94-1.24 kpc that is in good agreement with Fig.3a distribution.
We did not take into consideration possible biases due to different morphological composition and luminosity distribution of the samples of normal and interacting galaxies in our previous analysis. As one can see in Table 1, only half of interacting galaxies have estimated morphological types. Within limits of poor statistics, the distributions of morphological types of our sample edge-on interacting galaxies and normal galaxies in cumulative sample (our+vKS+BD) are undistinguishable. Moreover, both samples do not show statistically significant correlations of scale height on morphological type. From the other side, the absolute luminosity distributions in both samples are close also (for instance, mean absolute blue luminosities of the samples galaxies are -19.1 and -19.6 for normal and interacting galaxies correspondingly with dispersion about ).
In Fig.4 we compare distributions of normal and interacting galaxies in the plane absolute blue luminosity - scale height. As one can see, both samples occupy approximately the same region in this plane. Solid line in Fig.4 shows dependence expected for normal face-on galaxy with 21.65 (Freeman 1970) and 5 (vKS, Bottema 1993). The data for real edge-on galaxies follow this relation quite acceptably with some systematic shift. This shift is a measure of total absorption in the disk of edge-on galaxy in comparison with face-on orientation. Considering only bulgeless galaxies, we find an estimation of total absorption in the B band as mag. This estimation is in remarkable agreement with the RC3 value of 1.5 mag.
Summarizing the results of direct comparison of distributions for normal and interacting galaxies, one can conclude that there is an evidence of moderately enhanced (at about 50%) disk thickness in interacting spirals.
One can expect that normalized thicknesses - ratios - can give more distinct evidence of enhanced thicknesses of interacting disks than absolute values of scale heights. Indeed, as one can see in Table 2, interacting galaxies demonstrate 1.5-2 times lower mean value of in comparison with all considered samples of normal spirals.
The distribution of ratios for normal galaxies in the joint sample (our+vKS+BD+dGvK) is compared with distribution for interacting spirals in Fig.5a,b. (We neglect in this figure the possible colour dependence of the ratios. But note that exclusion of the vKS data obtained in a relatively blue (close to the B) passband does not change the general shape of distribution for normal galaxies.) As one can see in the figure, both samples demonstrate remarkably different distributions - normal galaxies are peaked at 4-5 while interacting spirals show concentration around 2-3. Analysing with the DSS spatial environment of non-interacting galaxies having 3 (left wing of distribution in Fig.5a), we found that among 7 such galaxies 5 are non-isolated (see also item 3.1.2). From the other side, among 5 interacting spirals with 4, 3 (Arp 208W, K 3W, and K 14SW) demonstrate very regular symmetric optical morphology (see Fig.1 in Paper I) indicative of relatively weak interaction with companions. Exclusion of such contaminated galaxies makes the difference between two distributions in Fig.5 significant at any level of confidence.
In Fig.6 we compare scalelength distributions of normal galaxies in the joint sample and of our interacting spirals. Both distributions are statistically undistinguishable although one can note a somewhat shorter, on average, disks in interacting galaxies. Therefore, taking into account that interacting and normal galaxies demonstrate statistically undistinguishable distributions of morphological types and absolute luminosities also, we can conclude that the difference in ratio must reflect real thickening of interacting disks.
3.3. General observational conclusions
In our analysis of the vertical structure of edge-on interacting galaxies we found that strong tidal influence increases significantly scale height variations along galactic disks even within relatively bright parts of them. The most typical feature of interacting disks is the systematical change of along the radius.
From the direct comparison of scale height and distributions in the samples of interacting and isolated spirals we found an evidence of moderate (in 1.5-2 times) thickening of galactic disks in interacting systems. This thickening refer to the region of exponential disk between 1 and 2.4 of exponential scalelength (or between 0.6 and 1.4 of effective radius).
The mean characteristics of edge-on interacting galaxies in our sample are: absolute blue luminosity (so "face-on" magnitude must be about ), exponential scalelength kpc. Therefore, typical galaxy in our sample is comparable with the Galaxy and M 31. Most edge-on galaxies have comparable luminosity companions within one optical diameter (see Fig.1 in Paper I). One can conclude from this that tidal interaction between large spiral galaxies like the Milky Way and the Andromeda galaxy at the mutual distance about one optical diameter leads to 1.5-2 times thickening of their stellar disks at half-mass radius of the luminous disk.
© European Southern Observatory (ESO) 1997
Online publication: May 26, 1998