I measured the δ13C and δ15N values of plant tissue and soil samples from a 13C and 15N pulse labeling experiment using MAT 253. The δ13C and δ15N values of the standard (UREA, add each 20 samples) exhibit some considerable variation and have no trend. How can I adjust the δ13C and δ15N values of these samples?
I guess that your samples are somehow enriched and what you see on your control samples is a carry-over/memory effect. Have you tried adding a blank after each measurement?
Dear Qi Li,
we usually work in collaboration with a facility service to analyze the isotopic composition of our samples, and they ususally place at least 3 references between each sample-batch (10-12 samples) and about 10 samples references at the beggining and at the end of the entire experiment analysis (about 5-6 batches). Furthermore, is UREA the only reference material you do use?. They usually use at several references (e.g. atropine, IAEA-NO3, USGS-40,... ). Probably you can get low variance using different reference samples with distinct isotopic compositions for N and C. Hope this can be useful for you.
Hi Qu Li,
I'm afraid that can be really tricky to be done, because, every sample would produce a different effect on the following one, according to its enrichment. What you could do is just to evaluate the effect of the last sample over the control sample, but you cannot assume that it can be the same for every other sample, and correct them accordingly. As a matter of fact, you don't see any trend in your control samples, which means that no build up of whatever regular effect can be detect in your series, that could be corrected with a linear correction factor, for example. In my opinion, you don't have other choice than repeat the measurements, increasing the flushing time and putting a dummy/blank to clean-up the system between the one and the other sample.
To scale calibrate measured d13C (and d15N) sample values to VPDB (and Air), one needs at least 2 different reference materials such as USGS40 and USGS41 or USGS40 and IAEA-600 (one needs at least 2 points to construct a linear regression equation). You could use two house standards instead provided they have been calibrated to VPDB (and Air) using certified reference materials supplied by the IAEA or NIST. Urea is not a good choice as an isotope standard since its elemental composition does not match the elemental composition of compounds one would typically wish to analyse. Urea's atypical C/N ratio can give rise to unrepresentative conversion into N2 and CO2 and thus result in some isotopic bias. Amino acids such as USGS40 (41) or caffeine (IAEA-600) are a much closer match from an elemental composition (C / N ratio) to your average sample (such as your plant tissue). On the subject of referencing strategies you might wish to consult Willi Brand's paper in Rapid Communication of Mass Spectrometry (2001; vol 15, 501-519).
You need to run 2 standards, 1 control, 6 to 10 samples, one control, 2 standards sequence. As previous authors noted, you need to chose the correct standard for your samples (“same” chemical composition) and you need also to have your isotope ratios between the two standards. Closer your standards are to your sample the better result you get. The second problem in IRMS (but generally in spectroscopy) is the memory effect. You can measure 10 times one sample and make a plot, having on X axis the sample sequence and on y axis the measured values. Then you can see how bad the memory effect is. Just as a caution. The memory effect depends on the value of the previous sample. Less memory effect you get if there are no big variations between samples.
Here is a good book on IRMS and is free. You can download from the folowing page:
http://www.forensic-isotopes.org/gpg.html
I might have got it wrong, but I thought that here UREA wasn't used as calibration standard, but as control sample that, once calibrated (against a number of standards, that Qi Li didn't mention), showed high variability compared to what it should have had. Too much of my immagination?
Hi, all, thank you so much. I'm very appreciate for your all replies.
As Sirignano said, UREA wasn't used as calibration standard, but as control sample (sorry for making it unclear). Once the UREA as control sample showed high variability compared to what it should have had, How can I adjust this?
Hi Qi Li, as mentioned in my previous post, urea is not a good choice for IRMS work due to its elemental composition. For every mol-equivalent of urea 1 mol-equivalent of N2 (and 1 molequivalent of CO2) is generated. Most organic compounds (materials) generate 0.5 mol-equivalent of N2 for 6 (or more) mol-equivalents of CO2. The disproportionally high amount of N urea is comprised seems to result in non-quantitative conversion of N into N2 (which will results in high variability of d15N data) while the excess NOx (NOx will tail into the CO2 peak) will cause problems with 13C analysis. I am afraid there is nothing you can do to retrospectively correct your data unless your sample batch run comprised other reference materials and not just urea. If have analysed a bracket of 2 or more end-member reference materials with your samples and urea is only used to check (quality control) the result of the scale calibration (based on 2 or more end-member reference materials), report your data on the basis of the 2 end-member scale calibration and "ignore" the urea data. For future use, instead of urea use a different compound as quality control; something like leucine or acetanilide. Best wishes, Wolfram
- Badges
- Science topic
- Similar topics
- Higher Education
- Universities