Abstract

We measure the evolution of the specific star formation rate (sSFR = SFR/M-stellar) between redshift 4 and 6 to assess the reported "constant" sSFR at z > 2. We derive stellar masses and star formation rates (SFRs) for a large sample of 750 z similar to 4-6 galaxies in the GOODS-S field by fitting stellar population models to their spectral energy distributions. Dust extinction is derived from the observed UV colors. We evaluate different star formation histories (SFHs, constant and rising with time) and the impact of optical emission lines. The SFR and M-stellar values are insensitive to whether the SFH is constant or rising. The derived sSFR is very similar (within 0.1 dex) in two M-stellar bins centered at 1 and 5 x 10(9) M-circle dot. The effect of emission lines was, however, quite pronounced. Assuming no contribution from emission lines, the sSFR for galaxies at 5 x 10(9) M-circle dot evolves weakly at z > 2 (sSFR(z) proportional to (1 + z)(0.6 +/- 0.1)), consistent with previous results. When emission lines are included in the rest-frame optical bands, consistent with the observed Infrared Array Camera [3.6] and [4.5] fluxes, the sSFR shows higher values at high redshift following sSFR(z) proportional to (1 + z)(1.0 +/- 0.1), i.e., the best-fit evolution shows a sSFR similar to 2.3 x higher at z similar to 6 than at z similar to 2. This is, however, a substantially weaker trend than that found at z 2 and even than that expected from current models for z > 2 (sSFR(z) proportional to (1 + z)(2.5)). Even accounting for emission lines, the observed sSFR(z) trends at z > 2 are still in tension with theoretical expectations.