Peter,

If I remember correctly, much of this discussion on
scales and on oxygen-isotope acid fractionation
factors originated with R. Mark Larson's question
about how to convert his d18O values of CO2 from
PDB to SMOW.  Were it my project, I would have
processed NBS 19 calcite as a sample and would have
submitted the CO2 thus produced to the laboratory
along with CO2 from my samples.  Knowing that the
d18O value of NBS 19 solid (not the CO2) is -2.2 per
mill exactly or +28.65 per mill on the VPDB and
VSMOW scales, respectively, I would be able to
"adjust" the isotopic data received from the
laboratory, as required.

I think anyone requesting isotopic results from any
laboratory should request the delta value of NBS 19
(or other internationally distributed reference
material, as appropriate for the analytical method)
had it been submitted and analyzed.  Any competent
laboratory should be able to provide such information.

In my view, because of the small variations in marine
carbonate d18O and d13C values (and the discordance in
scales between some laboratories), every paper dealing
with marine carbonate stable isotope ratios should
provide d18O and (or) d13C value(s) of NBS 19 calcite
had it been analyzed with reported samples.  Readers
can then convert delta values to VPDB (or whatever
scale they choose) to ensure data from different
laboratories are on the same d18O and d13C scales.

A tip about purchasing isotopic reference materials
for d13C and d18O measurements:  If there are two
equivalent vendors and one vendor expresses delta
values relative to VPDB (or states delta value(s) of
NBS 19), choose it.  The other vendor doesn't deserve
your business because they have not properly
calibrated their isotopic reference materials.

It would appear to me that the IAEA and IUPAC are
in full agreement with your suggestion that authors
report delta values of internationally distributed
reference materials (as appropriate to the
analytical method) along with their isotopic data.
In particular, recent guidelines request the
following.

For hydrogen, "it is recommended that the author's
measured delta 2H of NBS 22 oil, NBS 30 biotite, or
other internationally distributed reference material
be reported, as appropriate to the analytical
procedure."

For carbon, "it is recommended that the author's
measured delta 13C of NBS 22 oil, USGS24 graphite,
or other internationally distributed reference
material be reported, as appropriate to the
analytical method."

For oxygen of carbonate, the author's measured delta
18O of NBS 18 or other internationally distributed
reference material should be reported, as appropriate.

For oxygen of other substances, the author's measured
delta 18O of NBS 28 quartz, NBS 30 biotite, NBS 127
barium sulfate, or atmospheric oxygen should be
reported, as appropriate to the analytical method.

The full Guidelines for Reporting of Stable H, C, and
O Isotope-Ratio Data can be found in:

Paleoceanography, v. 11, p. 369-370, 1996.

Eos, Transactions, AGU, v. 77, p. 255, 1996.

GCA, v. 60, p. 3359-3360, 1996.


Tyler B. Coplen
U.S. Geological Survey
431 National Center
Reston, VA 20192
Tel: 703 648 5862



>Dear IsoGeochem-ists,
>
>why not add another penny's worth on a hot summer day:
>
>the genius of the delta technique of McKinney et al is its
>pragmatism. We don't worry about accuracy, and in return we get
>almost any degree of precision. HOWEVER, as soon as we change to
>other types of sample, or, as Paul points out, even variants of
>procedure, the high precision is blunted by a bias of unknown
>sign and magnitude. We are then on a different delta scale, which
>requires different reference materials. For example, there is no
>way of comparing delta-O values of silicates to delta-O values of
>carbonates with coherent high precision. Such a comparison, if it
>is to be precise, is conditional on the quoting of two reference
>materials - one for each scale.
>
>The use of adopted fractionation factors (whose importance Kim
>and O'Neil, GCA 1997 have re-explained) does not remove possible
>bias of 0.5 permil or so between carbonates and silicates (or
>carbonates and  other carbonates, if the digestion is changed).
>It is easy to see why. Factors like alpha(CO2-H2O) or the kinetic
>factors, are very difficult to determine. To assume calcite delta
>values recalculated as SMOW to be precise to 0.1 permil for
>comparison with silicates, is to assume that the fractionation
>factors, if re-determined by the given laboratory at the same
>time, would turn out to within 0.1 permil of the internationally
>adopted values...!
>
>If 0.1 permil precision (or better) is required, we have to give
>away the idea of a unified SMOW scale for all kinds of oxygen
>isotope analyses - in spite of the fact that "it's only one
>world"! The delta technique needs abundant well prepared
>reference materials (like NBS 18, 19, 28 and the new CO2 samples
>of NIST) and we cannot avoid to analyse them often and to quote
>the values obtained - before or after adjusting to one or another
>preferred or required level.
>
>The "floating" and purely relative nature of the delta scales
>does NOT detract from useful things such as accurate thermometer
>calibrations. However, the reference values for each scale must
>become part of the calibration (- not so dissimilar to certain
>factors entered into isotope fractionations calculated
>theoretically by statistical mechanics).
>
>Maybe this would be a topic for ICOG?
>
>Peter
>
>([log in to unmask] )
>
>P.S. Normalization of scale length comes on top of this.
>