Cosmological Applications of H Surveys

The local ionising background radiation

From the studies of the WIM in our Galaxy and in nearby galaxies, it appears likely that a substantial fraction of ionising photons could escape from galaxies. Although at first surprising, the high escape fraction may be due to the patchy, clumpy nature nature of the ISM, fractal (Elmegreen 1997) or otherwise. Disc photons reach the Reynolds layer. Within the inner 60 kpc or so, the detection of diffuse H in the Magellanic Stream (Weiner & Williams 1996) implies that a large fraction of galactic ionising photons reach that distance (Bland-Hawthorn & Maloney 1997), while in nearby dwarf galaxies the large number of expanding cavities which may be fully ionised certainly allow a large fraction of ionising photons to escape from these low opacity regions (e.g. Puche et al. 1992, Puche & Westpfahl 1994). An obvious test is to take integrated spectra of these galaxies at wavelengths around 912 Å. This has been attempted with HUT on a few nearby starburst galaxies, with apparently negative results (Leitherer et al. 1995). However, the inclusion of the Galactic absorption, important for the low redshift galaxies studied, increased the escaping fraction, and lower limits ranging from 3% to 57% are allowed (Hurwitz, Jelinsky & Van Dyke Dixon 1997).

If many ionising photons escape galaxies, they will contribute significantly to the ionising background radiation, which is usually thought to be dominated by quasars. Yet locally quasars contribute much less than ten percent to the luminosity density in the B and U bands. Could galaxies dominate the background, at least locally? Again the proportionality between the emission measure, the number of recombinations and the H emission give a useful constraint on the local ionising background, as first pointed out by Sunyaev (1969). Assuming that the H is optically thin, the ionising photon flux is

where and are the total recombination coefficient and the H emission rate respectively, and the aspect ratio of projected to total area goes from 1/2 for a two-sided slab to 1/4 for a sphere. This expression actually assumes that the ionisation comes only from one side, the general expression for the ionisation from both sides not having been solved yet. The observations of isolated extragalactic clouds (e.g. Stocke et al. 1991, Vogel et al. 1995) provide limits of about 3 10 ionising photons s cm, which becomes a limit on the ionising background at the Lyman limit of erg s cm Hz sr assuming a slope of 1.4 for the background, typical for quasars. These values also assume case B recombination. However, as shown elsewhere (Valls-Gabaud & Vernet 1997) galaxies are the main component of the ionising background, and their soft spectra is consistent with the pattern of ionisation traced by SiIV, SiIII, CIV, CIII observed at z > 3. The consequence of this model is that at low redshift the background is also very soft, with an effective slope of the order of 3 or more. This relaxes the constraint on the amplitude of the background, since

then the limit becomes erg s cm Hz sr for case A recombination, which is more appropriate. Deharveng et al. (1997) obtained strong upper limits on the escape fraction using the H luminosity density at z=0, but assuming a background dominated by quasars. With a galaxy-dominated background, the present limits are consistent with an average escaping fraction larger than 30%. The model predicts that the local background must be very soft, and hence can be falsified because low ionisation stages must prevail over high ionisation species in isolated extragalactic clouds.

© Copyright Astronomical Society of Australia 1997