4.2. Results

The unsharp-masked greyscale images of the survey galaxies are shown in Figs. 17 - 20 (the galaxies from our first sub-sample Rossa & Dettmar, 2000) and all others in Figs. 22-54 (only available electronically at EDP Sciences) in the central rows (between the R-band and H images) in order of their increasing R.A.. Each figure consists of 2 × 3 sub-panels, with two galaxies, each one accompanying one column. The scale is marked by a black horizontal bar in the lower corner of each figure, and the orientation is N to the top, and E to the left.

Figure 17. Unsharp-masked R-band image of IC2135, showing very patchy dust filaments. No continuous dust lane is visible. N is on top, and E is to the left.

The following results are obtained from the simple analysis of the unsharp-masked R-band images. We find extraplanar dust (eDust) filaments in 26 of our 74 galaxies. 48 galaxies lack in showing extraplanar dust. In one case (ESO274-1) we cannot state with confidence whether there are eDust filaments visible or not, as this galaxy is superimposed onto a crowded field of foreground stars. The unsharp-masked process caused many artifacts, therefore no clear statement can be given in this particular case. We count this galaxy as a negative detection in our statistics, which is summarized in Table 4. Typical distances of the extraplanar dusty filaments from the galactic midplane are | z| ~ 0.6 - 1.5 kpc.

The dust structures at high galactic latitudes are diverse in morphology. In IC2135 ⁶, where no prominent dust lane is detected, only patchy features are seen in the disk-halo interface on both sides of the disk (cf. Fig. 17). Very spectacular filamentary structures are discovered in NGC4302 (Fig. 18). Many of the discovered dusty filaments are strongly bended, suggesting that magnetic fields may act upon the charged dust particles. The Virgo Cluster spiral NGC4402 (see Fig. 19) shows strongly winded filaments, which are detectable out to | z| ~ 1.7 kpc. The dusty structures show a disturbed morphology as well as the eDIG morphology. This suggests that ram-pressure stripping, caused by the interaction of the ISM with the ambient Virgo Cluster Intracluster Medium (ICM), is acting upon NGC4402. This has been concluded for other Virgo Cluster galaxies as well (e.g., Veilleux et al., 1999; Vollmer et al., 2000). In Fig. 20 we show the remaining six of the nine galaxies studied in H by Rossa & Dettmar (2000).

Figure 18. Unsharp-masked R-band image of NGC4302, showing spectacular dust features above/below the galactic plane. The face-on spiral to the right is NGC4298. N is on top, and E is to the left.

Figure 19. Unsharp-masked R-band image of NGC4402. Several dust filaments, which reach high galactic latitudes, are visible. N is on top, and E is to the left.

Candidate galaxies with prominent dusty features at high galactic latitudes include two starburst galaxies NGC3628, NGC5775, and also the galaxy NGC7090. The unsharp-masked images can be seen in the Figs. 35, 41, 48, respectively.

The distribution of the high-| z| dust features reveals that usually extraplanar dust is visible in those galaxies, where also eDIG is detected. The dust features, however, reach much lower galactic latitudes, except in two cases - NGC360, NGC4302 - where dust is detected at larger distances from the galactic midplane than eDIG. Generally, the distribution of the high-| z| dust is restricted to | z| 1.5 kpc.

Figure 20. Unsharp-masked R-band images. Upper left: NGC3044, upper right: IC2531, middle left: NGC4634, middle right: NGC5170, lower left: IC4351, lower right: UGC10288. The corresponding H images can be found in Rossa & Dettmar (2000). Orientation: N is to the top and E to the left.

We find two cases, where eDust was detected (NGC360, ESO240-11), but where no eDIG has been detected. On the other side, we find six cases where the opposite is true. The majority (89%), however, shows a clear correlation between eDIG/eDust detections and non-detections. Although a correlation of the presence/non-presence of high-| z| gas and dust exists, it should be noted that in general no 1:1 correlation exists (i.e. individual gas and dust filaments are not spatially correlated)! This might indicate that other mechanisms are responsible for the transport of the dust into the halo, than for the gas transport.

⁶ IC2135 was inadvertently identified with NGC1963 in Rossa & Dettmar (2000), due to a general confusion in the literature and in electronic databases that existed back then. Back.