We examine the radius-luminosity (R-L) relation for blue galaxies in the Team Keck Redshift Survey (TKRS) of GOODS-N. We compare with a volume-limited, Sloan Digital Sky Survey sample and find that the R-L relation has evolved to lower surface brightness since z=1. Based on the detection limits of GOODS this can not be explained by incompleteness in low surface-brightness galaxies. Number density arguments rule out a pure radius evolution. It can be explained by a radius dependent decline in B-band luminosity with time. Assuming a linear shift in M_B with z, we use a maximum likelihood method to quantify the evolution. Under these assumptions, large (R_{1/2} > 5 kpc), and intermediate sized (3 < R_{1/2} < 5 kpc) galaxies, have experienced Delta M_B =1.53 (-0.10,+0.13) and 1.65 (-0.18, +0.08) magnitudes of dimming since z=1. A simple exponential decline in star formation with an e-folding time of 3 Gyr can result in this amount of dimming. Meanwhile, small galaxies, or some subset thereof, have experienced more evolution, 2.55 (+/- 0.38) magnitudes. This factor of ten decline in luminosity can be explained by sub-samples of starbursting dwarf systems that fade rapidly, coupled with a decline in burst strength or frequency. Samples of bursting, luminous, blue, compact galaxies at intermediate redshifts have been identified by various previous studies. If there has been some growth in galaxy size with time, these measurements are upper limits on luminosity fading.
Read full abstract