Now that there is a better understanding of mass spectrometry problems in
the analysis of hydrogen isotope ratios, attention can be turned to the
problem of what technique to use in preparing H2 gas standards:

In the "good old days" when uranium was used almost exclusively to reduce
water (and some, like Simon Sheppard, happily still use uranium), we could
take a sample of H2 of known isotopic composition, oxidize it to H2O, and
then reduce this water back to H2 while retaining its original dD value to
about +/- 1 per mil or so.  Torsten Vennemann tried this at Michigan a few
times with zinc reduction but I don't recall his results.

In any event someone out there should be able to do these manipulations very
carefully with the reducing agent of their choice, and perhaps with improved
precision.  Once someone can produce some data proving that you can go back
and forth between H2O and H2 with isotopic fidelity (to some accepted error
like +/- 1 per mil or better), we are there.  

Thus we could either prepare gas standards from waters of known isotopic
composition or reduce aliquots of tank H2 to water.  Assuming the
availability of appropriate sources of tank hydrogen, the latter direction
would be most convenient in preparing large amounts of hydrogen gas for
distribution as reference standards.
James R. O'Neil
Department of Geological Sciences
University of Michigan
Ann Arbor, MI  48109-1063
Fax: (313) 763-4690