Previous Article Next Article Table of Contents
Ultrafaint H-alpha Emission from Intergalactic...(1Sep94) H I Clouds and Galaxy Halos (from NOAO HIGHLIGHTS!, NOAO Newsletter No. 39, 1 September 1994) Megan Donahue (STScI), Greg Aldering (Minnesota), and John Stocke (Colorado) went looking for H-alpha emission in all the wrong places. Their initial program was to search for faint H-alpha emission arising from surface photoionization of intergalactic 21-cm clouds, an approach that offers the most sensitive probe available of the metagalactic ionizing radiation field at the present epoch. Their observations set interesting constraints on the properties of this radiation field, but the clouds themselves remained invisible. Donahue and her collaborators, however, did find H-alpha emission coming from an extended halo around the galaxy NGC 4631, a place they didn't expect. From the start, Donahue, Aldering, and Stocke knew that any emission coming from the intergalactic H I clouds was likely to be extremely faint, so they used a telescope and techniques maximally suited for detection of low surface-brightness, highly extended, emission-line sources. Surprisingly, the telescope of choice is not the 4-m, but the Burrell-Schmidt, equipped with a 2048 X 2048 CCD. This system features a field of view of over one square degree, large (2 arcsec) pixels, and a sufficiently fast beam so that the bandpass and throughput of the narrow-band filters were not adversely affected by the converging angles of the light rays. Taking advantage of dark, stable nights, and using the "shift-and-stare" technique for constructing night sky super-flats, they were sensitive to H-alpha emission at surface-brightness levels unprecedented in an imaging experiment. (Similar limits have been achieved in Fabry-Perot experiments by Reynolds and others, but with limited spatial information.) Donahue, Aldering, and Stocke expected to detect or to set significant limits on faint, low-surface-brightness emission from the surfaces of the H I clouds, which had previously been identified in 21-cm emission. Indeed, they obtained interesting upper limits for the Leo Ring, a clearly intergalactic cloud not directly associated with any single galaxy. Their current surface- brightness detection limit for the Leo Ring is u H-alpha < 102 mR (for comparison, this is over 10^7 times fainter than the Orion nebula). This limit corresponds to a background ionizing intensity of JO(13.6eV) < 1.2 X 10^-22 erg s^-1 cm^-2 sr^-1 Hz^-1 or an ionizing flux at the face of the cloud of 5.6 X 10^4 photons cm^-2 s^-1. This upper limit is consistent with estimates of the local ionizing background from quasar counts, the "proximity" effect in local Lyman-alpha forest clouds, models to reproduce the sharp edges in the H I-profiles of disks, and extrapolations from the far-UV background. New data acquired this spring will either decrease this limit by a factor of 2-3 or reveal a detection. Donahue, Aldering, and Stocke, however, did not expect to detect H-alpha emission between two H I spurs extending nearly 30 kpc from the disk of the well-studied edge-on galaxy N4631. Nor did they expect to detect a halo of low surface-brightness H-alpha emission, nearly 10 kpc thick, around the same galaxy. Although extensive H-alpha images have been made of N4631, none of them had the field of view or the sensitivity to low surface brightness emission that were achieved by using the Burrell Schmidt and excellent night sky flats. The processed H-alpha image with the H I 21 cm contours from Rand and van der Hulst (1993, AJ, 105, 2098) is shown in the figure. In producing the image, bright objects such as stars and the interiors of the galaxies were masked out, and show up as sharp white features in the figure. [Figure not included] The entire image was then median-filtered on 2 X 2 arcmin scales. The scattering halos of extremely bright stars are still present in the images, and show up as boxy shapes with a rectangular imprint. These data were taken with the old filter-holder, which vignetted the field on the corners, so edge-effects are visible in a circular ring along the outer borders of the field (the new filter wheel eliminates this problem). The image has not been continuum-subtracted, but the continuum R and B images do not show emission at such large scale heights, except for extremely faint emission in the tidal tail that accounts for less than 50% of the counts detected in the narrowband filter. The emission measures seen in light grey in the inverse grey-scale image are approximately 0.1 cm^-6 pc, dark grey in the tidal tail between the H I tails corresponds to 0.3-0.5 cm^-6 pc; black corresponds to emission measures of 0.9 cm^-6 pc. The emission measures are too high to arise from metagalactic photoionization alone, since at such high levels, H-alpha would have been detected easily from the H I spurs and the Leo Ring. Therefore, most of the heating of the thick H-alpha halo must be internal to the galaxy. Since no other edge-on galaxy has been imaged to such sensitivity and wide angular scale in H-alpha, we do not know whether such extended H-alpha features are common, or somehow generated in N4631 tidally. Certainly the 30 kpc "tidal tail" feature detected between the two H I spurs (seen in the 21-cm map of Rand and van der Hulst has a tidal origin, since no explosive scenario can eject gas that far from a galaxy. Such observations have interesting consequences for the possible nature of absorption lines from "Lyman-alpha forest" at low redshift. If detectable ionized gas can exist at moderate radii, then it might not be far-fetched to expect that gas with extremely low H I column density, as detected in absorption against low- redshift background quasars by HST, might be still farther away from the galaxy, and yet still associated with it. The Burrell-Schmidt CCD system provided a new view of edge-on galaxies and new limits on the local metagalactic ionizing flux. Similar techniques directed at revealing faint extended emission around other targets will show whether spiral galaxies have larger cross-sections than we thought, as well as provide the first reliable estimate of the local metagalactic ionizing background.
Previous Article Next Article Table of Contents