Fig. 1 shows the distributions of the radio emission perpendicular to the major axis in the observed galaxies. For all three objects the distribution could be well fitted by a two-component exponential function, convolved with the beam size. Considering also inclination effects, the scale-heights of both components are similar in all galaxies: h_{thin disk} ~ 300 pc and h_{thick disk} ~ 1.8 kpc, independently of the star forming activity in the underlying disk and the interaction state. Since all three galaxies exhibit a large-scale magnetic field predominantly parallel to the disk (Dumke et al. 1995), the field configuration could possibly be the determinant factor for the emission scale-heights. Differing results in previous investigations may be caused by a lack of sensitvity/resolution and the lack of large-scale structure information in interferometric data. Strong star forming activity, however, increases the absolute intensities of both components and may be responsible for a relatively high intensity of the thick disk component compared to the thin disk, as it is the case in NGC 891.
|
| Fig. 1. Distribution of the radio continuum emission perpendicular to the major axis of the galaxies. The dotted lines show the telescope beam, the solid lines the fitted function for a two-component distribution model |
The spectral index distribution perpendicular to the plane shows a similar behaviour in all three galaxies: an increase with galactic height z. We see (with the available resolution of 2 - 3 kpc) no evidence for different transport mechanisms for relativistic electrons. The non-interacting objects NGC 891 and NGC 4565, however, have flatter spectra in the disk and in the halo. Especially the spectra in the disk of these two objects can only be explained with a very high thermal fraction of the total emission, about 50 % at 6.2 cm. This is also suggested by a spatial correlation of flat spectra and low fractional polarizations in localized regions in the disk. The spectra observed in face-on galaxies are usually steeper, because those objects must have significant halo emission (which has a steeper spectrum) to yield an observable surface brightness.
The two galaxies NGC 891 and NGC 4565 seem to be very similar, and the first one appears to be just an upscaled version of the latter. On the other hand the spectral indices in the interacting galaxy NGC 3628 indicate just a very small amount of thermal emission in the galactic disk, and strong star formation seems to be restricted to the central region. Hence the relatively strong thick disk component cannot be due to recent star formation in the thin disk. Possible explanations may be the gravitational interaction with other galaxies and/or that the observed thick disk is a remnant of higher star forming activity in earlier times.
Fig. 2 shows the behaviour of the fractional polarization with increasing distance from the galactic plane, as a function of wavelength and beam size. From the high-resolution 6.2 cm data (white circles) we find that p increases over the first ~ 2 kpc, but then decreases again. This suggests strong depolarization effects in the disk and in the upper halo of the galaxies. In general, the degree of polarization is highest in NGC 4565, the galaxy with lowest star forming activity, and smallest in NGC 891. This is expected, because star formation leads to enhanced turbulence and more efficient depolarization. At 20 cm (which is not shown in Fig. 2) the polarized intensity is always very weak, due to Faraday depolarization. In case of strong star formation, as in the central region of NGC 3628 and over the whole disk of NGC 891, there is already strong Faraday depolarization at 6.2 cm.
|
| Fig. 2. Degree of polarization in the disk and halo of the observed galaxies for different wavelengths and resolutions |
Next: References
Up: Radio and polarization properties ...
Previous: Combination of interfermeter and ...