Beyond their spectacular spiral arms and their bright central bulges, spiral galaxies are embedded in extremely faint stellar haloes, that can reveal tantalizing clues to how the galaxies formed. This elusive compoment has long defied astronomers' efforts to characterize it, but scientists at the Max-Planck-Institut fuer Astrophysik have recently developed new techniques which exploit the wealth of data in the Sloan Digital Sky Survey to pin down the properties of the luminous haloes around spiral galaxies with unprecedented accuracy. Their results have now been confirmed by the Hubble Ultra Deep Field data, the deepest view of the Universe ever.

In the currently favoured cosmological framework galaxy formation proceeds hierarchically. Small protogalaxies form first and then merge, forming stars and accreting material until they eventually reach the status of "adult" galaxies, as seen around us in the local universe. Adult galaxies come in many different shapes, from elliptical star piles, to disc-like spiral galaxies and irregular, almost chaotic systems. As a by-product of their violent assembly, a significant fraction of the stars in a galaxy is expected to be thrown out to large distances, forming a stellar halo. Observing the properties of these halo stars can reveal when and how the galaxy was assembled and clarify the physical mechanisms that govern galaxy formation.

The stellar halo around our own Galaxy, the Milky Way, has been studied since the 1950's, and progress in observational techniques over the last few years has provided a detailed and rather surprising picture of its structure. Rather than being smooth and regular many of the stars lie in streams which seem to be the remnants of disrupted dwarf companion galaxies. This structure has strongly shaped current ideas about how the Milky Way formed. But how far can we generalize such models to other galaxies than the Milky Way? Do all spiral galaxies have a stellar halo? If so, do all stellar haloes have similar properties? Due to the extreme faintness of the halo (its surface brightness, i.e. the luminosity emitted per unit area, is several hundred times fainter than in the central regions of the galaxy) observations are very difficult and are so far available only for a few nearby galaxies.

Using the wealth of data provided by the Sloan Digital Sky Survey (SDSS), we have developed a new technique which stacks images of a large number of galaxies in order to increase sensitivity, on average, to the faint light from their outer parts. Our sample consists of 1054 spiral galaxies seen edge-on. The light from their haloes is then uncontaminated by superposed emission from the bright discs. We clearly detect a mean stellar halo in the stacked images, and its structural properties are quite similar to those of the few, known, individually observed spiral haloes. The statistical nature of our method enables us to quote results for "typical" galaxies and to assess the ubiquity of stellar haloes around spirals.

Fig. 1: The "mean" galaxy image which results from the stacking of 1054 galaxies in the SDSS. In the left panel the images taken in three different bands are composed in a pseudo-colour image that mimics the colour the human eye would see. The very thin disc, seen in the edge-on orientation, dominates the emission of the galaxy. In the right panel we have streched the intensity levels and used false colours to emphasize the faint halo emission surrounding the disc.

Are our results consistent with the hierarchical scenario of galaxy formation? Careful theoretical calculations of the structure expected will be needed to get a definitive answer. Our results include some definite surprises. For example, the colours of the extended halo emission are not easily explained by standard models of stellar populations. Although this might, in principle, indicate a problem with our stacking analysis, a recent study of an edge-on disc galaxy in the Hubble Ultra Deep Field, the deepest view of the Universe ever, indicates similar colours for the stellar halo in this individual system.

Fig. 2: As in Fig. 1, in the left panel we show the pseudo-colour image of an edge-on disc galaxy (z=0.32) in the Hubble Ultra Deep Field. Most of the light comes from bright thin disk which displays spiral arms and dust lanes. In the right panel we stretch the intensity levels in order to show the faintest emission (note that the regions corresponding to sources other than the galaxy have been blanked out). Although this image is shallower than the SDSS stacking by a factor of approximately 5, the halo emission is apparent and the similarity between the two observations is striking.

Stefano Zibetti, Simon D. M. White, and Annette M. N. Ferguson

Further reading:
S. Zibetti, S. D. M. White, J. Brinkmann
"Haloes around edge-on disc galaxies in the Sloan Digital Sky Survey",
2004, MNRAS, 347, 556

S. Zibetti and A. M. N. Ferguson
"A faint stellar halo around an edge-on disc galaxy in the Hubble Ultra Deep Field",
MNRAS accepted (astro-ph/0406207)