Bother River He (using a continuous flow Micromass system) and I using a
H/Device with Delta Plus have seen a similar problem. Both use chromium
reduction of water to produce H2. Running different isotope ratio
standards over time, different standards drift at different rates, and not
in a standard decay curve. You might want to check over you data and see
if this is the case for you as well. I had thought this phenomena was just
due to my instrument, but see River He's data where there is no bellows has
similar effects so we conclude it must be some artifact of the chromium,
perhaps the amount it is oxidized?
The good news is that by fitting a polynomial drift curve to both a high
and low standard, delta value vs. sample number, and then interpolating
between them for every sample number it is possible to get excellent data
from either instrument. I will be presenting a poster on this at Jesium
http://chemsrv0.pph.univie.ac.at/jesium/ . Using this method I get a
precision on quality control samples at different isotope ratio than the
standard, of better than 1 delta unit, so this correction seems to
work. River He's data is similar.
I have worked up a spreadsheet that does the corrections automatically for
two or three standards, and I will have copies at Jesium that I can talk
people through if anyone is interested.
At 05:14 PM 7/14/04 +0200, you wrote:
>One question concening the drift correction of hydrogen values.
>Since our MS is not equipped with an automatic reference refill system
>I observe a drift during a sequence of 50 samples measured for dD
>using a Finnigan H/Device coupled to a delta S. The drift is the
>result of a fractionation of the reference gas in the standard
>bellow. To correct the data for this drift we carry out repeated
>measurements of the same standard sample during the sequence.
>Theoretically, the drift should follow a Rayleigh fractionation and
>one should observe a (more or less) continuous drop of the measured
>standard sample values. The sequence values can be corrected for the
>drift by fitting a function to the data points of this "drift
>standard". This is exactly what I observe in most of the sequences.
>However, on some days the drift seems not to follow the theoretical
>model. Instrument parameters (filling of the bellow, temperature,
>etc.) are essentially the same as on the other days but in some
>sequences I observe a convex shaped function instead of a concave
>shaped Rayleigh function or a step-like function with a fast drop of
>the values at beginning a stable phase for several hours in the
>middle and a second drop at the end of the sequence. Memory effects
>can not account for the observed trends.
>I expect I am not the only one observing such a phenomena from time
>to time. The question is how you handle this kind of data? The first
>option is to discard the whole sequence. The second would be to fit a
>(polynomial) correction function to the measured values assuming that
>my data points represent the original drift of the gas (even if the
>measured drift differs from theoretical assumptions). The third
>option would be to fit a strict Rayleigh function (or alternatively a
>linear regression) to the data points which results in a very bad
>fitting of the function to the measured values.
>Robert van Geldern
>Leibniz Institute for Applied Geosciences (GGA)
>S3: Geochronology and Isotopehydrology
>Stilleweg 2, 30655 Hannover, Germany
>phone: +49-(0)511-643-2313, fax: +49-(0)511-643-3665
>mailto:[log in to unmask]
Paul D. Brooks,
Center for Stable Isotope Biogeochemistry,
Dept. Integrative Biology MC3140,
3060 Valley Life Sciences Building,
UC Berkeley, Ca. 94720-3140.
[log in to unmask]