Astron. Astrophys. 325, 1259-1263 (1997)

4. Comparison with the ionizing background: constraints on the escape fraction

We have compared the predictions of Fig. 1 with the measurements of the ionizing background (at ) at the Lyman edge. These measurements are only indirect and include the contribution of all ionizing sources, chief among them the quasars. They are therefore upper limits, possibly very generous, to the contribution of galaxies. The measurements have been obtained so far through a variety of methods (Bechtold 1993, for a review), such as the H emission from 21-cm emitting clouds, the modeling of the sharp edges of HI clouds and, the proximity effect (Kulkarni & Fall 1993), following the observations of the Lyman forest at low redshift with the HST. From the recent works of Dove & Shull (1994), Vogel et al. (1995) and Donahue et al. (1995) and review of previous determinations therein, we conclude that the ionizing background at 900 Å should lie in the range  erg cm^-2  s^-1  Hz^-1  sr^-1. At its lower bound the ionizing background is comparable to the evaluations currently made for the contribution of quasars (e.g. Madau 1992), indicating a possible negligible role of the galaxies.

Fig 1 shows that, under most of the conditions, the escape fraction is smaller than 1%. This is smaller than the upper limits in the four star-forming galaxies observed by Leitherer et al. (1995), especially after the modifications advocated by Hurwitz et al. (1997) on the basis of unaccounted absorption by HI gas in our Galaxy. Our upper limit would be less restrictive than 1% if the spectral index is lower than -4, or the evolution milder () than found by Lilly et al. (1996). Insofar as our comparison was made with the intergalactic radiation field of all origin, our upper limit is probably generous but, even if small, does not exclude that galaxies make most of the ionizing intergalactic radiation field.

Our conclusion also depends crucially on the value of the total H luminosity density of the local universe in the sense that the larger the H luminosity density (or the associated number of photons), the stronger is the upper limit for the escape fraction. Our limit of 1% seems firm for two reasons. First, the H luminosity function built by Gallego et al. (1995) is based on star-forming galaxies with H N[II] equivalent widths larger than 10 Å and therefore should more likely lead to an underestimation of the total H luminosity density. Second, our conversion of the H luminosity density into a photon density under case B assumptions provides, as previously said, a lower limit on the latter quantity.

© European Southern Observatory (ESO) 1997

Online publication: April 28, 1998

helpdesk.link@springer.de