Dear dr. Pavese,
As I understand your problem, its solution would be very simple: you
just have to send small aliquots of the Hydrogen gas you use to a lab
performing isotope ratio measurements (such as ours). Such a lab can
then determine the D-content relative to VSMOW with a precision high
enough for your goal, and you could correct your results accordingly.
As for the required precision: you state that the full range from
zero to VSMOW concentration of deuterium (or roughly 0 to 0.015 %
abundance, or in our terms -1000 to 0 o/oo) your error could be 1 mK,
whereas you aim at precisions < 0.1 mK. This means that you would be
satisfied if an isotope lab could give you the relative D/H ratio of
the hydrogen gas you use to better than, say, 50 o/oo on the absolute
scale.
Now, there are some problems for isotope ratio labs to perform
"semi-absolute" measurements: (1) our standard material VSMOW is
water, we have to convert this somehow to hydrogen gas. Since we are
primarely interested in water samples, it does not harm us if we do
not 100% transform the water to hydrogen, as long as we have the same
deviation all the time. The point is that we do not know whether we
deviate or not. However, a ring test of pure hydrogen in 1997 (Brand
and Coplen, see the Isogeochem web site archives) has shown as a side
result that "conversion errors" typically stay within 10 o/oo or so.
(2) Isotope ratio measurements suffer from all kind of "scale
contraction and extraction" errors, such that what we claim to be 1
o/oo is not really 1 o/oo. The typical size of error is up to 5% of
the deviation between the sample and VSMOW. Suppose the hydrogen gas
you use would be severely depleted (suppose -600 o/oo, equivalent to
a D-content of only 40% of that in VSMOW), this error would amount 30
o/oo at maximum. There are ways to decrease this error, based on
using another standard water called SLAP, which has a known deuterium
content as well (some 40% less than VSMOW). This procedure, called
normalisation, would decrease this error considerably, as long as the
deuterium content of the sample lies in the range between SLAP and
VSMOW. When it is depleted even further than SLAP, things get
uncertain again.
In conclusion, it would not be difficult for an isotope ratio lab to
perform the "semi-absolute" D-content measurement to within the +/-
50o/oo (and most likely considerably better) that you would need. The
best way to increase the precision would be to use hydrogen gas which
is close to VSMOW in its D-content. Prescribing such a demand in your
new guidelines (including a isotope ratio measurement) would be the
way to go, I think.
I hope I have been clear enough. For any more questions, feel free to
contact me again, either via isogeochem, or directly
([log in to unmask])
Sincerely,
Harro Meijer
>I have freshly subscribed this list having heard of it from a
>colleague of NIST (Wes Tew) in US.
>I am not a geochemist, but instead a metrologist working in low
>temperature thermometry at IMGC, the equivalent in Italy of NIST.
>
>In thermometry, the International Temperature Scale of 1990 (ITS-90)
>is using the triple point of hydrogen as a definition fixed point. We
>are able to reproduce the temperature of this state for any
>high-purity (typ 6N) sample to better than 0.1 mK. At this level, the
>D content in "hydrogen" is critical: a change from the VSMOW value to
>zero would change the triple-point temperature by about 1 mK.
>
>Until a few weeks ago no one of us, thermometrists, was aware of the
>fact the the actual content of D in H is not necessarily the VSMOW
>and can be so variable. Everybody was assuming the VSMOW as fixed and
>in the ITS-90 we did not put any specific warning and any
>specification about that, meaning that we assumed "natural" abundance
>for "hydrogen" as an isotopic mixture.
>
>By chance, talking recently with a gas supplier, I went aware of the
>fact that D content can be very different in commercial hydrogen of
>different sources and batches. I was also informed by Wes of the
>existance of this discussion list and I took a look of your archives
>on the web.
>
>As thermometrists we are EXTREMELY interested in interacting with all
>people, like you, that have experience and data on this subject
>matter: I am talking of gaseous hydrogen with a purity from 5N to 7N
>that can be obtained from commercial supplier around the world, that
>you probably use as reference material in your studies.
>
>We will need to update the definition of the ITS-90 about this point
>as soon as possible (we have a formal meeting in April 2000) to avoid
>a very large increase of the uncertainty in this part of the ITS-90
>and, consequently, we will need to perform new measurements on
>different H-D mixtures at least in the range of the values that can
>be found in commercial gases. Not all the gas manufacturers are
>measuring the D content, and even less of them routinely. Not many of
>them can provide acertificate of analysis specific of the lecture
>bottle, especially for the isotopic composition.
>
>I will be grateful to everybody of you that could supply me
>information about the actual state-of-the-art.
>
>Sincerely
>Franco Pavese
>
>
>--ATTENTION: please do not use anymore the addresses:
> [log in to unmask],
>[log in to unmask]
> nor the phones:
> +39 11 3977 341 nor FAX +39 11 3977 347
>------------------------------------------------------------------------
>Dr. Franco Pavese
>CNR - Istituto di Metrologia "G.Colonnetti"
>strada delle Cacce 73
>10139 Torino, Italy
>
>tel. ATTENTION +39 011 3977 341
>FAX ATTENTION +39 011 3977 347
>e-mail: [log in to unmask]
>e-mail home: [log in to unmask]
>cellular-phone: 0348 8130101
Prof. dr. Harro A.J Meijer
Centrum voor IsotopenOnderzoek (CIO)
Nijenborgh 4
9747 AG Groningen
Netherlands
tel +31-50-3634760
fax +31-50-3634738
e-mail [log in to unmask]
|
