The following article appeared in:
Spectrum Newsletter of the Royal Observatories, June 1994, Issue no.2, p. 25-27

Dust in Spiral Galaxies

J.H. Knapen1, R.A.Jansen2, J.E. Beckman1, R.F. Peletier123 and R. Hes2

1 Instituto de Astrofísica de Canarias, E-38200 La Laguna, Tenerife, Spain
2 Kapteyn Astronomical Institute, Postbus~800, NL-9700~AV Groningen, The Netherlands
3 Observatorio del Roque de los Muchachos, Apartado 321, E-38700 Santa Cruz de la Palma, Tenerife, Spain

The issue of whether or not spiral galaxies possess significant quantities of dust in their discs is an important one, and has given rise to an interesting debate over the last few years. If, as some authors claim, the major parts of spiral discs are in general opaque, that would have important effects on their overall observational properties. It would also significantly change our ideas on the mass distribution in and around galaxies (the ``dark matter'' issue), since it implies that by measuring the optical emission from galaxies we greatly underestimate the amount of luminous matter in the universe.
Several tests have been described, and more are being published continuously, aimed at determining the dust extinction values in galaxies through studies of individual objects, or of samples of galaxies. However, no consensus has been reached yet on whether or not spiral discs are optically thick in the visible. In fact, in July of the present year a conference, to be held in Cardiff, will be devoted to the topic of dust in spiral galaxies.


Telescopes on La Palma
and at La Silla have been
used to compare dust
extinction in external
galaxies with our own


Fig.1 - Colour images of UGC 3214 (left) and UGC 11552 (right). Note the strong dust lanes blocking a significant fraction of the light from one half of the galaxy in the image of UGC 3214.

An important question is whether the extinction, absorption and scattering-properties of the dust in external galaxies differ widely, or even at all, from those in our own Galaxy. Not many studies have been published on this specific subject, and it is not obvious that one can assume that the Galactic extinction law is in fact universal when discussing dust in external spirals.
Some years ago, we published a study of the Sombrero galaxy where we looked at the dust properties in the dust lane that made the galaxy so well-known. The high degree of symmetry and the special geometry of the Sombrero, highly inclined, but not edge-on so that the dust lane does not obscure the centre of the galaxy, allowed us to obtain absolute dust extinction values in the dust lane by simply subtracting the unobscured half of the minor axis light profile from the half that is obscured by the dust lane, assuming intrinsic underlying symmetry in the light distribution. By doing this in optical and near-infrared (NIR) bands, and assuming a simple model where the dust is uniformly mixed with the stars (see below), we could conclude that the dust properties were in fact very similar to those of the dust in the Galaxy, lending more confidence to the idea that the Galactic extinction law is usable also in external galaxies.
To extend the basis of this earlier work to a larger sample, we have recently studied a few more highly inclined spiral galaxies each with a dust lane along the major axis, and slightly offset from the centre of the galaxy. The galaxies were selected from an initial sample of almost 40 galaxies that were selected using as criterium their high inclination angle and general appearance on the Palomar plates. We then observed these for 10 minutes through a V band filter with the JKT on La Palma. We found among them four galaxies that showed a geometry similar to that which allowed us to use the Sombrero galaxy in the work described above, and observed them, during the same JKT run, with deeper exposures in the B, V, R and I bands. During a later observing campaign at the ESO 2.2m telescope at La Silla we obtained images in the J and K' NIR bands.
The four galaxies, UGC 3065, UGC 3214, UGC 11552 and MCG 02-10-009, have been studied relatively little previously, and are angularly quite small. In Figure 1 we show colour images of UGC 3214 and UGC 11552, where the dust lanes are well visible, cutting out a significant part of the light from one half of the galaxies. Structure is visible inside the dust lane, showing that the dust is not distributed smoothly within the lane. The unobscured half of the bulge of UGC 3214 clearly shows that the isophotes deviate from elliptical shape: the bulge is ``boxy''. In Figure 2 we show contour plots in the B and K bands of another of our dust lane galaxies, UGC 3065, where it is clearly seen that, as in the other galaxies, the dust lane cuts out light from one half of the galaxy in the B band, but the effects of dust extinction have almost completely disappeared in the K band.


Fig.2 - Contour plus grey-scale plots of the blue (B) and NIR (K' band) images of the dust-lane galaxy UGC 3065. Note that the effects of dust extinction, obvious in the blue, are almost completely absent from the NIR image.

In order to compare the minor axis profiles of each galaxy, all the different frames available have to be precisely aligned. This is achieved by making Gaussian fits to foreground stars in the images. In the case of MCG 02-10-009, no stars were available in the frame, the frames could thus not be aligned, and the galaxy could not be used in the subsequent analysis. Note that the centres of the galaxies cannot be used for the alignment, because dust extinction due to the outer parts of the dust lanes may obscure the central region slightly, and thus change the central position in especially the bluer bands.
Once the images are aligned, and after determining the correct major axis position, we can subtract the unobscured part of the light profile along the minor axis from the dust-obscured part, and obtain dust extinction profiles across the dust lanes, for each photometric band studied. In Figure 3 we show these resulting profiles for one of the galaxies, UGC 3065, in the B, V, R, I, J and K' bands (from top to bottom). The extinction is given in absolute magnitudes, which are derived even without the need for photometric calibration of the images, as an intrinsic product of the differential method used to produce the extinction profiles. The extinction is strongest, as expected, in the B band, where a peak value of AV = 2.0 mag occurs between 4 and 5 arcsec from the centre (the centre position was determined in K), and gradually weakens while moving towards longer wavelengths. From these dust extinction profiles, we can, for each galaxy, determine ratios A\lambda / AV at different distances from the centre. In a schematic picture, if the dust had the exact properties of Galactic dust, the ratios A\lambda / AV would be exactly those as determined in our Galaxy (extinction law) at all points.

Extinction profiles

Fig.3 - Extinction profiles over the dust lane, along the minor axis, as functions of distance r from the centre of UGC 3065 in (from top to bottom) the optical B,V,R and I and the NIR J and K' bands.

In fact, when plotting the measured values of A\lambda against AV for these three galaxies, we see that the points lie along lines that significantly deviate from the ``Galactic'' line in some cases. To explain these deviations, we have invoked a simple ``uniform'' model, where we assume that the dust and the stars in the dust lane are uniformly mixed. The results of fitting the data points with such a model are shown graphically in Figure 4, where we plotted the fitted values of the extinction ratios A\lambda / AV as a function of the logarithm of the wavelength, indicating along the abscissa the location of the photometric bands used in our imaging. We have also included in the plot the Galactic extinction law, and the results found in our earlier work for the Sombrero galaxy. Within the uncertainties, all galaxies follow the Galactic extinction law. There are some deviations, notably in the blue, which could well be due to scattering of light within the galaxies. In the case of UGC 11552, the slight deviations from the Galactic law visible are possibly a result of a less-than-perfect determination of the position of the centre. In that galaxy, the dust lane shows its effects on the light distribution up to quite small distances from the centre. But in general the results confirm that we can use the Galactic reddening law in external galaxies, so that it is almost justifiable, and certainly tempting, to talk of a ``universal dust extinction law'', at least in the visible and NIR ranges.

Extinction ratios

Fig.4 - Extinction ratios A\lambda / AV as functions of the logarithm of the wavelength for the observed data, fitted with a model where stars and dust are uniformly mixed in the dust lane, for UGC 3065, UGC 3214 and UGC 11552, compared to the Galactic extinction law (solid line) and that for the Sombrero galaxy.

Although the images in visible light of these dust lane galaxies show that there is a significant amount of dust in the discs of these galaxies, this cannot be simply taken as proof that these galaxies would be optically thick in dust if seen face-on. In the central positions we find maximum values for the extinction of about 2 V-mag in the three galaxies studied here. The peaks in the extinction quickly decrease at increasing radial distances from the centre. Considering typical ratios (of ~10) between scale lengths and heights of dust and stars, we can conclude that these galaxies would be optically thin in dust over most off their discs when seen face-on.

Last updated: Nov 15, 2000 (minor style change)