5.5.2. Could the D/H which we observe be too high?

The answer to this question from Kirshner is, that the D/H could be slightly lower than we measure, but not by a large amount. We discuss two possibilities: measurement problems and biased sampling of the universe.

First we consider whether the D/H in the quasar absorbers could be less than observed. This can readily happen if the D is contaminated by H, but a large reduction in D/H is unlikely because the D line widths match those expected in Q1937-1009 and Q1009+2956. We do not know how the ISM D/H values could be too high, and Galactic chemical evolution requires primordial D/H to be larger than that in the ISM, and similar to the low value from quasars. Hence it is unlikely that the D/H is much below the observed value.

Second, we consider whether the absorbers seen in the quasar spectra are representative.�� The absorbers are biased in three ways: they represent regions of the universe with well above (100 - 1000 times) the average gas density at z = 3, and amongst such high density regions, which are observed as Lyman Limit absorption systems, they have relatively low temperatures (2 x 10⁴ K), and simple quiescent velocity structures. The last two factors are necessary to prevent the H absorption from covering up that from D, while the high density follows from the high density of neutral H which is needed to give detectable neutral D. It is likely that the gas in the absorbers at z = 3 has by today fallen into a Galaxy, though this is not required because some gas will be heated as galaxies form, preventing infall. The low temperatures and quiescent velocities argue against violent astrophysical events, and there are no reasons to think that the absorbers are any less representative than, say, the gas which made up our Galaxy.

We should also consider whether the quasar absorbers might be unrepresentative because of inhomogeneous BBN. In this scenario regions with above average density will have below average D/H, but the evolution of density fluctuations could be such that the low density regions fill more volume [108], [47], so that they are more likely to dominate the observed universe today. In that scenario the _b derived from the D/H would be below the universal average, and the observed (low) value of D/H would be ``high'' compared to expectation for SBBN with the same _b . This scenario will be tested when we have observations of many more quasars. measurements of D/H towards QSOs.