3.1. Deuterium

Since deuterium is completely burned away whenever it is cycled through stars, and there are no astrophysical sites capable of producing deuterium in anywhere near its observed abundance [2], any D-abundance derived from observational data provides a lower bound to its primordial abundance. Thus, without having to correct for Galactic evolution, the `here-and-now' deuterium abundance inferred from UV observations along 12 lines-of-sight in the local interstellar medium (LISM) bounds the `there-and-then' primordial abundance from below ((D/H)_P (D/H)_LISM = (1.5 ± 0.1) x 10^-5). As may be seen from Figs. 1 and 2, any lower bound to primordial D will provide an upper bound to the baryon-to-photon ratio [4].

Solar system observations of he permit an independent, albeit indirect determination of the pre-solar system deuterium abundance [5]. This estimate of the Galactic abundance some 4.5 Gyr ago, while somewhat higher than the LISM value, has a larger uncertainty (D/H = (2.1 ± 0.5) x 10^-5 [6]). Within the uncertainty it is consistent with the LISM value suggesting there has been only modest destruction of deuterium in the last 4.5 Gyr. There is also a recent measurement of deuterium in the atmosphere of Jupiter using the Galileo Probe Mass Spectrometer [7], (D/H = (2.6 ± 0.7) x 10^-5).

To further exploit the solar system and/or LISM deuterium determinations to constrain/estimate the primordial abundance would require corrections for the Galactic evolution of D. Although the simplicity of the evolution of deuterium (it is only destroyed) suggests that such correction might be very nearly independent of the details of specific chemical evolution models, large differences remain between different estimates. It is therefore fortunate that data on D/H in high-redshift (nearly primordial), low-metallicity (nearly unevolved) Lyman- absorbers has become available in recent years ([8, 9]). It is expected that such systems still retain their original, primordial deuterium, undiluted by the deuterium-depleted debris of any significant stellar evolution. That's the good news. The bad news is that, at present, D-abundance determinations are claimed for only three such systems, and that the abundances inferred for two of them (along with a limit to the abundance for a third) appear to be inconsistent with the abundance estimated for the remaining one. Here we have a prime illustration of ``precise'', but possibly inaccurate cosmological data. Indeed, there is a serious obstacle inherent to using absorption spectra to measure the deuterium abundance since the isotope-shifted deuterium absorption lines are indistinguishable from the corresponding lines in suitably velocity-shifted (-81 km s<sup>-1</sup>) hydrogen. Such ``interlopers'' may have been responsible for some of the early claims [8] of a ``high'' deuterium abundance [10]. Indeed, an interloper may be responsible for the one surviving high-D claim [11]. At present it seems that only three good candidates for nearly primordial deuterium have emerged from ground- and space-based observations. It may be premature to constrain cosmology on the basis of such sparse data. Nonetheless, the two ``low-D'' systems suggest a primordial deuterium abundance consistent with estimates of the pre-Galactic value inferred from LISM and solar system data ((D/H)_P = (3.4 ± 0.25) x 10^-5 [12]). To illustrate the confrontation of cosmological theory with observational data, this D-abundance will be adopted in the following. However, the consequences of choosing the ``high-D'' abundance ((D/H)_P = (20 ± 5) x 10^-5 [11]) will also be discussed.