Identification and correction of spectral contamination in 2H/1H and 18O/16O measured in leaf, stem, and soil water

Plant water extracts typically contain organic materials that may cause spectral interference when using isotope ratio infrared spectroscopy (IRIS), resulting in errors in the measured isotope ratios. Manufacturers of IRIS instruments have developed post-processing software to identify the degree of contamination in water samples, and potentially correct the isotope ratios of water with known contaminants. Here, the correction method proposed by an IRIS manufacturer, Los Gatos Research, Inc., was employed and the results were compared with those obtained from isotope ratio mass spectrometry (IRMS). Deionized water was spiked with methanol and ethanol to create correction curves for delta O-18 and delta H-2. The contamination effects of different sample types (leaf, stem, soil) and different species from agricultural fields, grasslands, and forests were compared. The average corrections in leaf samples ranged from 0.35 to 15.73 parts per thousand for delta H-2 and 0.28 to 9.27 parts per thousand for delta O-18. The average corrections in stem samples ranged from 1.17 to 13.70 parts per thousand for delta H-2 and 0.47 to 7.97 parts per thousand for delta O-18. There was no contamination observed in soil water. Cleaning plant samples with activated charcoal had minimal effects on the degree of spectral contamination, reducing the corrections, by on average, 0.44 parts per thousand for delta H-2 and 0.25 parts per thousand for delta O-18. The correction method eliminated the discrepancies between IRMS and IRIS for delta O-18, and greatly reduced the discrepancies for delta H-2. The mean differences in isotope ratios between IRMS and the corrected IRIS method were 0.18 parts per thousand for delta O-18, and -3.39 parts per thousand for delta H-2. The inability to create an ethanol correction curve for delta H-2 probably caused the larger discrepancies. We conclude that ethanol and methanol are the primary compounds causing interference in IRIS analyzers, and that each individual analyzer will probably require customized correction curves. Copyright (C) 2011 John Wiley & Sons, Ltd.