Seeing Galaxy Formation in a Cold Light


The sum of all the light from star formation outside our galaxy is usually called the extragalactic background light or EBL. The EBL is hard to measure because we need to subtract many foreground sources of emission to find the remaining light of the EBL. These foreground emissions include light from our atmosphere (for ground-based experiments), light from the scattering and reradiation of sunlight by dust in our own Solar System, starlight and dust-scattered and reradiated starlight from our own Galaxy, and finally the cosmic microwave background itself.
Because of these problems there are only upper limits on the EBL at optical and near infrared wavelengths, but recently the EBL has been measured at submillimeter wavelengths using data from the FIRAS and DIRBE experiments on the COBE satellite. The first results of Puget et al. (1996) have been followed by a number of analyses in the last two years (Guiderdoni et al. (1997), Schlegel et al. (1997), Fixsen et al. (1998), Hauser et al. (1998). This measurement is easier at longer wavelengths near one millimeter, where the contamination from Galactic dust is smaller, and there is extremely good agreement on the EBL at wavelengths above about 400µ. Residual uncertainty on how to deal with dust emission associated with ionized gas in our Galaxy means there are larger uncertainties associated with the shorter wavelengths, and Puget et al.'s estimated value is about a factor of two lower than that obtained by Fixsen et al. and Hauser et al. at 200µ.
The EBL is an integrated history of star formation, and in order to interpret it, and to turn it into a map of the formation as a function of redshift, we need to identify the individual objects which compose it and determine their redshifts. The current SCUBA results are the first step in this process since they allow us to identify a large fraction of the objects making up the submillimeter EBL, and now we know what these objects are, we can, at least where they have optical emission, determine their redshifts using larger ground-based optical telescopes such as Keck, the VLT or Gemini.
The figure shows Fixsen et al.'s analytic approximation to the quantity nu times S(nu), where S(nu) is the EBL in frequency (nu) units, shown as a solid black line. Because of the small size of each SCUBA field (about 5.6 square arcminutes) there is large field-to-field scatter in the measurements --- the green dot is the measurement of Hughes et al. for the contribution of sources brighter than 3 mJy in their HDF image, and the two purple dots are the same quantity in SSA13 and the Lockman Hole. The average of the three fields is shown as the large red square. Roughly 25% of the background light at 850µ is resolved at this sensitivity level. A model extrapolation to lower fluxes is shown as the open red square, and suggests that the direct observations are quite consistent with the EBL determination. The average observed temperature in the observed 850µ sources must be low in order for them not to show too much light at the shorter wavelengths. The dashed red curve is for an observed temperature of 50K, the solid 25K and the dotted 20K, where each is based on a lambda-weighted Planck function. The low temperature required (< 20K) is probably a result of the galaxies being at high redshift, so higher intrinsic temperatures in the rest frame are redshifted to lower observed temperatures at the present time.