Seeing Galaxy Formation in a Cold Light


In the absence of information about the redshift distribution, which is very poorly constrained as yet, the question of the relative fraction of the light arising in the submillimeter and in the optical is best addressed by looking at the total amount of light lying in the two wavelength ranges. The bolometric amount of light reflects the total amount of star formation radiated or reradiated at that wavelength divided by the (1+z) factor reflecting the expansion cooling. The bolometric optical light is best detemined from the deep counts in the Hubble Deep Field. This integrated light is close to the limits and possible measurements on the EBL at these wavelengths. The submillimeter EBL was discussed on the previous page. In the figure we show these two backgrounds, with the solid green line being the optical light (with the the x-axis in nanometers) and the solid black line the submillimeter background (with the x-axis in microns). Since we know the redshifts for many of the HDF galaxies we can further refine the optical EBL that can arise at higher redshifts by excluding known galaxies at z < 1. This latter quantity is shown as the dashed green line. If we assume that most of the galaxies giving rise to the submillimeter background are also at z > 1 --- still only a plausible hypothesis as yet --- then it is clear that the submillimeter bolometric light, which considerably exceeds the z > 1 limits on the optical light, must contain the bulk of the star formation.
The contribution of the observed 850µ counts is shown as the red square (the purple dots are the individual fields). The open red square is the extrapolation to fainter counts, assuming the shape follows the Guiderdoni models. The red solid and dashed lines show the extrapolation of this to other wavelengths for an observed dust temperature of T = 12°K, and the dotted lines for T = 15°K.