5.5.1. Was the primordial D high but depleted in the absorbers?
The idea here is that the average BBN D/H was high, and it has been depleted in the three absorbers which show low D. There are two options: local depletion in some regions of the universe, and uniformly global depletion. We conclude that there is no known way to deplete D locally, and global depletion seems unlikely.
First we list seven observations which together rule out local depletion, including that suggested by Rugers & Hogan .
1. We note that D/H is also low in our Galaxy, and that Galactic chemical evolution accounts for the difference from the low primordial D. Hence we know of four places where D is low and consistent with a single initial value.
2. If the BBN D/H was high, let us say ten times larger at 34 x 10-5 , then the depletion in all four, widely separated in space, must be by a similar factor: Q1937-1009: 0.90 ± 0.02; Q1009+2956: 0.88 ± 0.02; Q0130-4021: > 0.80; local ISM in our Galaxy: 0.86 - 0.93, where for the Galaxy alone we assume that Galactic chemical evolution reduced the initial D/H by a factor of 1.5 - 3 .
3. The quasar absorption systems are large - a few kpc along the line of sight , far larger than can be influenced by a single star or supernovae. The gas today in the local ISM is a mixture of gas which was also distributed over a similar large volume prior to Galaxy formation.
4. The abundance of the metals in the quasar cases are very low; too low for significant (> 1%) destruction of D in stars .
5. The quasar absorbers are observed at high redshifts, when the universe is too young for low mass stars (< 2 solar masses) to have evolved to a stage where they eject copious amounts of gas.
6. The quasar absorbers are observed at about the time when old stars in the halo of our Galaxy were forming. These stars may have formed out of gas like that seen in the quasar spectra, but with high density. We expect that much of the gas seen in absorption is in the outer halo regions of young galaxies, and that some of it was later incorporated into galaxies and halo stars.
7. The ratio of the abundances of Si/C in the quasar absorbers is similar to that in old stars in the halo of our Galaxy. This abundance ratio is understood as the result of normal chemical evolution.
Global destruction of D prior to z = 3, or in the early universe, remains a possibility, but it seems contrived.
Gnedin & Ostriker  discuss photons from early black holes. Sigl et al.  show that this mechanism creates 10 times more 3He than observed, and Jedamzik & Fuller  find the density of gamma ray sources is improbably high.
Holtmann, Kawasaki & Moroi ,  showed that particles which decay just after BBN might create photons which could photodissociate D. With very particular parameters, the other nuclei are not changed, and it is possible to get a D/H which is lower than from SBBN with the same b. Hence low D and low Yp can be concordant. An exception is 6Li which is produced with 6Li / H 10-12, which is about the level observed in two halo stars. There is no conflict with the usual conclusion that most 6Li is made by Galactic cosmic rays prior to star formation, because the observed 6Li has been depleted by an uncertain amount. This scenario has two difficulties: Burles (private communication) notes that there would be a conflict with the b measured in other ways, and it seems unlikely that the hypothetical particle has exactly the required parameters to change some abundances slightly, within the range of measurement uncertainty, but not catastrophically.
Most conclude that there are no likely ways to destroy or make significant D.