A New Measurement of the Primordial Abundance of Deuterium: Toward Convergence with the Baryon Density from the CMB?

From the analysis of the near-UV spectrum of the QSO 2206-199, obtained with a long series of exposures with STIS on the HST, we deduce a value (D/H) = (1.65 +/- 0.35) x 10(-5) (1 sigma error) for the abundance of deuterium in the z(abs) = 2.0762 damped Lyman alpha system (DLA) along this sight-line. The velocity structure of this absorber is very simple and its neutral hydrogen column density, N(H I), is accurately known; the error in D/H is mostly due to the limited signal-to-noise ratio of the spectrum. Since this is also one of the most metal-poor DLAs, with metal abundances of about 1/200 of solar, the correction due to astration of D is expected to be insignificant and the value we deduce should be essentially the primordial abundance of deuterium. When all (six) available measurements of D/H in high redshift QSO absorbers are considered, we find that the three DLAs---where N(H I) is measured most reliably---give consistently lower values than the three Lyman limit systems. We point out that the weighted mean of the DLA measurements, D/H = (2.2 +/- 0.2) x 10(-5), yields a baryon density Omega_B h^2 = 0.025 +/- 0.001 which is within 1 sigma of the value deduced from the analysis of the CMB angular power spectrum, and is still consistent with the present-day D/H and models of Galactic chemical evolution. Future observations of D I absorption in other DLAs are needed to establish whether our finding reflects a real advantage of DLAs over other classes of QSO absorbers for the measurement of D, or is just a statistical fluctuation.