How to deal with spectral interferences when measuring water stable isotopes of plants?

Nowadays, a wide range of water extraction/vapor equilibration techniques for obtaining soil and plant water isotopic composition (δ¹⁸O and δ²H) is applied by various ecohydrological disciplines. Here, researchers need to rely on accurate and precise measurements of water isotope ratios for tracing water movement through the critical zone. Previous research has shown that utilizing isotope ratio infrared spectroscopy (IRIS) to analyze water or vapor samples containing co-extracted/-equilibrated organic contaminants (e.g., methanol, ethanol) has the potential to result in significant inaccuracies through spectral interferences. However, the scientific community and the manufacturers have not effectively addressed the inaccuracies caused by organic contaminants. While some hardware solutions for combusting organics as well as some software solutions exist for spectral interference detection during liquid water IRIS analysis, limited tools exist for the post-correction of direct vapor-mode IRIS data e.g., from in-situ water vapor measurements or from the direct water vapor equilibration laser spectrometry method (DVE-LS).

For our study, we applied three different water extraction and vapor equilibration techniques (i.e., DVE-LS, in-situ water vapor measurements and cryogenic vacuum extraction) to four types of vegetables (cauliflower, celery root, kohlrabi and potatoes). We investigated how co-extracted organic contaminants (i.e., methanol and ethanol) via the different methods affect the isotopic ratios between liquid and vapor CRDS measurements of our vegetable samples. Through applying different CRDS instrument-specific post-correction options, we could reduce isotopic discrepancies and maximize the accuracy and precision of CRDS measurements from vegetables.

We could show that all vegetables produced species-specific different amounts of organic contaminants, which affected the isotope ratios obtained via the different extraction or vapor equilibration techniques in different ways. Clear relationships between DVE-LS samples and spectral parameters indicated co-equilibrated contaminants which we used for a technique-specific ‘organics-correction’. Whereas, results obtained from in-situ water vapor measurements were the least affected by organic contaminants and showed the smallest data spread. Those were also comparable to results from cryogenic vacuum extraction for some type of vegetables.

Our study underlines the importance and necessity of plant water vapor isotope data post-correction and highlights the need for a definitive and general protocol in order to prevent ill-founded ecohydrological data interpretations.