Figure 3. Spectrum of extragalactic background light, based on COBE data after Hauser 2001 (diamonds with error bars, dotted and short-dash curves), Madau & Pozzetti 2000 (squares), Totani et al 2001 (crosses), Bernstein, Freedman & Madore 2001 (triangles) and Armand et al 1994 (asterisk). The broken-line curve (Biller et al 1998) and horizontal dash-dot line (Hauser 2001) show upper limits based on lack of attenuation of high-energy -rays from AGNs and the solid curve is from the model by Pei, Fall & Hauser (1999). The arrow showing an upper limit at 10µm is from unpublished thesis work by A. Barrau, cited by Gispard, Lagache & Puget (2000).
Fig 3 shows the spectrum of extragalactic background light (EBL) with the model fit by Pei, Fall & Hauser (1999). Gispert, Lagache & Puget (2000) have estimated the total EBL I d based on observation to lie within the following limits:
|6µm:||20 to 40||nwt||m-2||sterad-1|
|> 6µm:||40 to 50||"||"||"|
|Total:||60 to 90||"||"||"|
(The total from the model of Pei, Fall & Hauser (1999) is 55 in these units.)
We use the estimates of stellar and metal densities in Tables 1 and 2 together with eq (7) and an assumption about the mean red-shift of metal formation to derive the EBL contributions from:
Here we use eq (6) without the (1 + a), since most of the nuclear energy is already released on reaching this stage. Assuming most stars to belong to an old population so that
Madau & Pozzetti (2000) and Brusa, Comastri & Vignali (2001) have made estimates of the AGN contribution to EBL based on the the abundance of massive black holes and that of obscured hard X-ray sources, respectively. They agree that the contribution is quite small, of order 5 nwt m-2 sterad-1.
The upshot is that these readily identifiable contributions add up to 48 nwt m-2 sterad-1, well within range (given the obvious uncertainties in mean z and other parameters) of the lower estimate given at the beginning of this section. It is interesting to note that white dwarfs and intergalactic metals come out as the major contributors, either one predominating over metallicity in known stars. To reach the higher estimate may involve some more stretching of the parameters.
I thank Richard Bower, Jon Loveday and Max Pettini for helpful information and comments.