Dear Attila and all,
the ring test for D/H measurement that I organized using the Isogeochem
list server in 1996 had a clear result regarding VSMOW/SLAP scaling:
there are multiple sources for a non ideal SMOW/SLAP difference, some of
which are due to the preparation of hydrogen, some due to the mass
spectrometer. These effects need to be addressed separately. The mass
spec may overrestimate the difference and the preparation at the same
time produce too small a difference resulting in an overall compensation
and the feeling that the procedure works all right when indeed it does
not.
With the Cr reduction method we in general see the need for scaling at
the 1% level at the most, often less. The gas produced from the reaction
interacts with the hot Cr; the lighter isotope dissolves preferentially
leaving the gas phase heavy by about 10 per mille. Provided the amount
of analyte is kept constant and the timing of the preparation is
maintained by a computer or stop watch, the fractionation almost cancels
and the measured difference is the correct one.
Klogging of the furnace can be prevented by selecting a uniform grain
size >100µm. With the method we routinely run 60 analyses per night with
precisions of about 0.5 per mille and accuracies of about 1 per mille.

Willi A. Brand
Manager Stable Isotope Lab
Max Planck Institute for Biogeochemistry
PO Box 100164
07701 Jena, Germany
Tel xx49-3641-643718
lab xx49-3641-643823
fax xx49-3641-643710
email [log in to unmask]