Galaxies and the ionising background radiation
David Valls-Gabaud
1,2, Joël Vernet3
1: Observatoire de Strasbourg, 11 Rue de l'Université,
F-67000 Strasbourg, France
2: Royal Greenwich Observatory, Madingley Road, Cambridge,
CB3 0EZ, UK
3: ESO, Karl-Schwarzschild-Strasse
2, 85748 Garching, Germany
Abstract.
Recent observations of the HeII Gunn-Peterson effect and
of ionised species at high redshift indicate that the ionising background
is softer than previously thought. The spectra from quasars being too hard,
a background dominated by galaxies may be a solution. Using the observed
evolution of HI, and the first self-consistent model atmospheres and evolutionary
tracks for massive stars, we predict the emissivity of galaxies as a function
of redshift from 200 to 3000 angströms. The predictions are
consistent with the local background, the CFRS and FOCA observations as
well as with the HDF constraints. When compared with the contribution from
quasars, galaxies dominate the emissivity at the Lyman break, but fail
to contribute enough to increase the softness SL
to the observed levels.
Previous estimates of the ionising background produced by galaxies have
been hampered by either ad hoc SFR prescriptions for galaxies (Bechtold
et al., 1987) or by a 'galaxy formation rate' (Miralda-Escude & Ostriker,
1991) which were both poorly constrained by observations, particularly
at high redshift.
Figure 1: Evolution of the emissivity of
galaxies (black) and quasars (green).
Here we overcome the problem by using the evolution of the cosmological
density of HI as traced by DLAs (Lanzetta et al., 1995) to infer the evolution
of the density of the star formation rate as a function of z. We then use
the recent self consistent stellar evolution models from Schaerer and De
Koter (1997), which combine for the first time stellar interiors and atmospheres.
This is particularly important because they give a much harder spectrum
for the He ionisation, in comparison with the previous models from Kurucz
(1979).
We then predict the evolution of the luminosity density at FUV wavelengths,
and compare them with observations (Fig. 1). The predictions
are robust, since the FUV emission depends only on the current SFR, and
not on the past history of the SFR. Note that the emissivity at 2800 angströms
above z > 1.5 is larger than the HDF limits, although consistent
with the recent observations from Treyer et al. (1997). These limits are
however uncertain, and subject to corrections for completeness and dust
extinction. The models presented here are likely to give the maximum UV
emissivity (since we assume fesc = 1) consistent with the constraint
given by the present density of stars .
Figure 2: Evolution of the resulting ionising
background for galaxies and quasars.
The resulting background radiation, taking into account the opacity of
the Lyman alpha absorbers and the recombination radiation is indicated
in Fig. 2. Although galaxies dominate the background
at 912 angströms, and produce a value of SL ~8000
at z = 3-4, they fail to contribute signifcantly since fesc
< 1 in realistic conditions.