HI Dominic--

The directly measured D/H values can be good, ca. +/- 0.5 per mil, but 
then the correction to the water D/H values means that error will be 
around +/- 2 per mil.  If the on-off pulses of the reference gas are 
giving good results, then its the equilibration that's causing 
problems.  I've taken to soaking our Pt sticks in 0.1 N nitric for ca. 
30 min, then lots of rinsing/soaking with nanopure water.  before 
letting them dry.  Test with your setup--I'm not too sure how well the 
glue holding the Pt beads onto the stick holds up.

take care,


On 9/7/2019 7:10 PM, Dennis Newell wrote:
> Hi Dominic -
> We used to make these measurements on the GB using the Pt rods, but 
> have moved over to the TC/EA because the precision is better. When 
> things were running well, we could achieve ~2 ‰, but it was not 
> uncommon to see some really unacceptable variability (up to 5 ‰ or 
> more). We found that baking the column prior to a run really improved 
> precision. Funky chemistry (salty, high organics or sulfide) has been 
> attributed to variability and the gasbench manual suggests adding some 
> Cu wire and/or activated C. We did this because we run some funky 
> waters, but were not convinced it made a measurable difference - it 
> made us feel better though.
> I don't remember what water backgrounds did, but I don't think the 
> increases in m18 that you observe should be a problem for the d2H 
> measurement, but maybe someone else knows differently.
> It is probably normal that your peak intensity drops as much as you 
> observe during a run due to dilution by the He flow. But 3 ‰ 
> variability within a run is too much (I think we were pretty happy 
> when it was 0.5). But we were running at much higher sample amplitude 
> - ours dropped from ~9 V to 7 V over ten peaks.
> best of luck sorting this out.
> Dennis L. Newell
> Associate Professor of Geology
> Department of Geology
> Utah State University
> 4505 Old Main Hill
> Logan, UT  84322
> 435-797-0479
> ------------------------------------------------------------------------
> *From:* Stable Isotope Geochemistry <[log in to unmask]> on 
> behalf of Dominic Ángel Becerra Serrato <[log in to unmask]>
> *Sent:* Saturday, September 7, 2019 2:25 PM
> *To:* [log in to unmask] <[log in to unmask]>
> *Subject:* [ISOGEOCHEM] Problems with high deviations in H2 results.
> Hello everyone, we´re running deuterium measurement tests with 
> GasBench and so far we have not been able to obtain the desired results.
> The deviations are very high, even among the ten peaks of the same 
> sample (between 0.8 - 3 ‰) and also when we compare the results of 
> different vials containing the same sample (we´re introducing 5 vials 
> per sample) the deviations reach values between 1 - 5 ‰.
> We´ve noticed that after finishing a test, the level of mass 18 
> increases considerably. For example, we started running the test with 
> 1400 mV in mass 18 and at the end of the test the intensity increased 
> to 3700 mV. Could this be affecting the deltas? If this is the case, 
> how can we fix it?
> We also observe that the intensity of the 10 peaks decreases 
> significantly, around 50% between the first and the last one (4500 mV 
> to 2300 mV). Is it normal for it to fall so abruptly?
> We would greatly appreciate your support. Regards.
> Dominic Ángel