Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 354, 823-835 (2000)

Previous Section Next Section Title Page Table of Contents

5. The distribution of the ionized gas

The major part of the H[FORMULA] emission comes from the HII regions, but there is substantial flux which comes from zones where there are no peaks in the widespread, low level emission. These zones do not present HII regions, but a diffuse component, which is often associated with the most luminous HII regions, but is detected across the whole face of the galaxy, including the interarm, and the central zones. The HII regions are strongly concentrated in the arms and the bar, but not distributed symmetrically, as one arm has far more of them than the other. The HII region distribution coincides in large measure with the HI distribution analyzed by Ball (1986). The two arms both stand out in both types of emission, though the HI distribution is notably more symmetric. Most notable are the string of very luminous HII regions along the bar, and the strong bursts of star formation at its ends.

In Fig 5 we show the flux density distribution of the HII regions across the face of the galaxy, with the H[FORMULA] flux distribution, normalized to the value in the innermost bin, overlaid on that in the I-band. In Fig 6 we show the number density distribution of the HII regions. Both of these plots are on linear scales, in ordinate (intensity) and in abscissa (distance from the centre). The fluxes were computed dividing the image into rings, using the P.A. and i of the galaxy to deproject, each of width 4", and calculating the flux per unit area in each ring. The central value was not plotted as the radial interval gave too small an area for a reliable point. As expected for a disc galaxy, there is a decline in flux density radially outwards, but another notable feature of Fig 5 is the presence of peaks and troughs within the general trend. There is a prominent peak at 50 arcsec from the centre, due to the quasi-annular zone of star formation formed by the inner spiral arms, while the smaller peaks between 70 and 120 arcsec radius correspond to peaks in the star formation density along the most prominent spiral arm.

[FIGURE] Fig. 5. Flux density distribution (normalized to the central value) of all the HII regions of NGC 3359, as a function of deprojected radius from the centre of the galaxy. I-band normalized flux density distribution (slanted shading) is overlaid on the normalized flux density for the H[FORMULA] image (horizontal shading).

[FIGURE] Fig. 6. Number density distribution of all the HII regions of NGC 3359 as a function of deprojected distance from the centre of the galaxy.

The underlying radial flux density distribution can be fitted by an exponential of the form


from which we derive a value for the H[FORMULA] scale length, [FORMULA]= (2.3[FORMULA]0.2) kpc.

In the case of the emission in the I-band we derive a value for the I scale length, [FORMULA]= (1.95[FORMULA]0.05) kpc.

With the broad band images aligned (see Sect. 2.2), colour images were produced in the standard way by subtraction in magnitude of the calibrated individual single-band images. Since the observational resolution varied from image to image, in each case the sharper had to be convolved with an appropriate gaussian to yield the resolution of the other before combining the two. We present here three images with complementary morphological information: U-I with a resolution of 1.1", in Fig 7; an overlay of I on H[FORMULA] in Fig 8, an overlay of U on I in Fig 9, I-K with a resolution of 1.9" in Fig 10. The resolutions have not been matched in the overlays, which are for morphological illustration.

[FIGURE] Fig. 7. Grey scale representation of the U-I image of NGC 3359.

[FIGURE] Fig. 8. Grey scale representation of the H[FORMULA] image of NGC 3359, with overlaid I intensity contours, showing position angle lag of the bar between I and H[FORMULA]. I intensity countours correspond to 23.5, 22.6 magnitude followed by 21.7 to 18.4 I magnitude in steps of -0.3 magnitude.

[FIGURE] Fig. 9. I-band image of the bar of NGC 3359 overlaid with intensity contours in U, showing the rotation of the position angle of the bar with wavelength, due to a combination of star-formation and dust geometries (see text for details).

[FIGURE] Fig. 10. Grey scale representation of the I-K image of NGC 3359. Darker tones correspond to stronger dust extinction. The underlying two-armed spiral is better picked out in the dust than in star-formation zones (cf. Fig 1).

The U-I image is a clear representation of recent star formation across the galaxy, and the star-forming zones correspond perfectly to those picked out in H[FORMULA] in Figs. 1 and 8. In Fig 7 the white areas in U-I (the bluest in U-I) are more extensive than the bright H[FORMULA] areas in Fig 1, but this is due mainly to the higher S:N ratio in the colour image. The diffuse component in U-I delineates continuum scattered preferentially in U from the dust in the interstellar medium, and is strongest where there is enough dust to scatter well, but not enough to extinguish the U-band light, i.e. in the zones surrounding the spiral arms. It is also notable that the star forming zones along the bar are less prominent in U-I than in H[FORMULA], an effect undoubtedly due to differential extinction by the dust concentration along the bar. Any similar effect along the arms is not appreciable, because of the displacement of the dust lanes towards the edges of the arms, i.e. away from the star formation zones, (cf. Fig 10) and because the dust optical depth is in any case less in the disc than in the bar. We can infer that as the U extinction in the arms is small, that at H[FORMULA] it can be safely neglected when constructing the LF's. The effect of the dust in the bar can be seen in Fig 10 where stronger dust absorption yields blacker tones in the image, as expected from an I-K representation, since K is much less affected by dust than I. The dust along the bar covers the centre ridge where the star formation is occurring, and although it is displaced in position angle from the line of maximum emission in H[FORMULA] (comparing Fig 10 with Fig 1 or with Fig 8), it covers the nuclear zone, and the zones of maximum star formation, where it reduces the U intensity significantly; the H[FORMULA] along the bar is much less affected by this extinction, due to the wavelength difference U-H[FORMULA] as would be expected. In fact we can use the U-I and H[FORMULA] maps to make a first order estimate of the extinction along the bar at H[FORMULA], above the HII regions observed. This does not exceed one stellar magnitude in the densest portion of the bar dust lane, and in general will affect the measured H[FORMULA] luminosity of the bar HII regions by less than 0.2 dex, i.e. by less than a bin width in Fig 3. The total number of HII regions affected by the bar dust is small, but five of the most prominent have measured luminosities above the [FORMULA] = 38.6 erg s-1 break. These are well specified, and only two at the most could be shifted into a higher luminosity bin as a result of dust extinction, so although the slope of the upper part of the LF in Fig 3 would be very slightly affected, the net resulting error is very small indeed, and has not been included in Fig 3.

In Fig 10 we can see that the dust distribution along the two spiral arms is more symmetric than the emission distribution in H[FORMULA] or in U-I, indicating that the neutral interstellar medium is a cleaner indicator of the underlying two-armed spiral structure than is the star formation distribution. Of particular interest is the overlay of I on H[FORMULA] in Fig 8, where we can see clearly that there is a lag in position angle between the bar in H[FORMULA] and the bar in I. The morphology of the bar in the I image is that of the full underlying stellar population, while that in H[FORMULA] follows current massive star formation, which is favoured along the leading edge of the bar at either side of the nucleus, with a projected cross-over at the nucleus itself. In fact the bar isophotes are twisted, so that the position-angle lag between I and H[FORMULA] is an increasing function of isophote major axis length, because the I light closer to the nucleus must contain a component from the red supergiants in the young population, and farther from the nucleus begins to take in some emission from the young arm population. A similar position angle lag, due to the massive star formation at the leading edge of the bar, is seen between the bar in the I image and in the U image, as represented in the overlay in Fig 9.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 2000

Online publication: February 25, 2000