We have put this equation into practice to re-evaluate the carbon isotope delta values for the CRMs BEET-1, GALT-1, FRUT-1, VANA-1, and VANB-1. Seven laboratories provided measurements for BEET-1, GALT-1, and FRUT-1, as well as for the reference materials USGS40, USGS62, and IAEA-CH-6, which were used for calibration to the VPDB-LSVEC scale 1. The consensus values were determined using errors-in-variables regression, and the Monte Carlo method was used to evaluate the uncertainty due to characterization, which took into account the correlations between the laboratory results 5. Two separate analysis sequences performed at the NRC were used to evaluate the uncertainties due to homogeneity and short-term stability. Additional carbon isotope delta measurements of BEET-1, GALT-1, and FRUT-1 were performed at the NRC and at Geotop, and were calibrated with the carbonate reference materials IAEA-603, IAEA-610, IAEA-611, and IAEA-612, whose carbon isotope delta values are reported on the VPDB scale. Conversion of these values from the VPDB scale to the VPDB-LSVEC scale was performed using the equation given by Hélie et al. 4 with the uncertainty propagation performed using the NIST Uncertainty Machine 6. These additional datasets were combined with the original measurements (Tables S1–S3). Similarly, to obtain the carbon isotope delta values of BEET-1, FRUT-1, and GALT-1 on the VPDB scale, the values for the calibrants USGS40, USGS62, and IAEA-CH-6 were converted from the VPDB-LSVEC scale to the VPDB scale. The consensus carbon isotope delta values and associated uncertainties were determined using the statistical model described in Chartrand et al. for the CRMs VANA-1 and VANB-1 7 (Table S4). Briefly, a linear multipoint errors-in-variables calibration curve was established for each measurement sequence from each laboratory on each carbon isotope delta scale (Tables S5–S10). A consensus value from these carbon isotope delta values was obtained by fitting a random effects model, which considers random laboratory effects, correlations between the laboratory results, and bottle-to-bottle homogeneity (Table 1). As expected, the carbon isotope delta values for these materials differ between the two scales, with the largest difference of 0.11 ‰ observed for BEET-1. There was limited information regarding the stability of the certified carbon isotope delta values with respect to short-term transportation or long-term storage in our 2018 publication 1. Since then, an experimental assessment spanning 6 years with unopened and opened CRM units stored under recommended conditions was conducted (Table S11). All the materials were found to be stable. As such, the uncertainty due to stability was set to zero in the reevaluated assessment. The uncertainties associated with the carbon isotope delta values of BEET-1, GALT-1, and FRUT-1 have been reduced by a factor of two from their original values established in 2018, an improvement that can be largely attributed to the inclusion of high-quality carbonate calibrants, even with the additional uncertainty from the conversion of these carbon isotope delta values to the VPDB-LSVEC scale. This is demonstrated in Figure 1, which shows improved uncertainties for both laboratories and all materials. Indeed, the uncertainties obtained from the analysis sequences calibrated using the IAEA carbonate reference materials are approaching that of the calibrants (Figure 1 and Tables S5–S10). However, this improvement is reflected only partially in the consensus value (Table 1), as only three of the 12 measurement sequences were performed using these reference materials. Moossen et al. 3 report carbon isotope delta values for BEET-1, GALT-1, and FRUT-1 on the VPDB scale using NBS 19 and USGS44 for calibration, then convert those results to the VPDB-LSVEC scale using the equation given by Hélie et al. 4. Instead of performing scale conversion calculations on the final results, we convert the carbon isotope delta values associated with all calibrators to the same scale. Results from both studies are in agreement (Table 1). While the results for FRUT-1 are within their stated uncertainties, the larger than expected difference warrants further investigation. Carbon isotope delta measurements of VANA-1 8 and VANB-1 9 are reported on the VPDB-LSVEC scale only, even though the suite of IAEA carbonate materials was part of the measurement campaign 7. Following the approach described above, the isotope delta values for VANA-1 and VANB-1 were established on the VPDB scale (Table 1 and Tables S12 and S13). The incorporation of additional carbon isotope delta measurements calibrated against the high-quality carbonate reference materials has resulted in significantly improved uncertainties for BEET-1 10, GALT-1 11, and FRUT-1 12. Our study demonstrates that the VPDB and VPDB-LSVEC scale conversion equation is fit for purpose without a significant increase in the measurement uncertainty. Michelle M. G. Chartrand: conceptualization, methodology, investigation, formal analysis, writing – original draft, writing – review and editing, project administration. Jean-Francois Hélie: conceptualization, methodology, investigation, writing – review and editing. Agnieszka Adamowicz-Walczak: investigation, writing – review and editing. Zoltán Mester: conceptualization, methodology, writing – review and editing. Juris Meija: conceptualization, methodology, formal analysis, writing – review and editing. The data that supports the findings of this study are available in the supplementary material of this article. Table S1: Raw input data for BEET-1. Table S2: Raw input data for GALT-1. Table S3: Raw input data for FRUT-1. Table S4: Laboratory-specific errors-in-variables statistical model to determine the carbon isotope delta value and associated uncertainties, as well as correlations between analyses. Table S5: Results from the errors-in-variables model to determine the consensus carbon isotope delta value, standard uncertainties due to characterization and homogeneity, and the individual laboratory results for BEET-1 characterization on the VPDB-LSVEC scale. Table S6: Results from the errors-in-variables model to determine the consensus carbon isotope delta value, standard uncertainties due to characterization and homogeneity, and the individual laboratory results for BEET-1 characterization on the VPDB scale. Table S7: Results from the errors-in-variables model to determine the consensus carbon isotope delta value, standard uncertainties due to characterization and homogeneity, and the individual laboratory results for GALT-1 characterization on the VPDB-LSVEC scale. Table S8: Results from the errors-in-variables model to determine the consensus carbon isotope delta value, standard uncertainties due to characterization and homogeneity, and the individual laboratory results for GALT-1 characterization on the VPDB scale. Table S9: Results from the errors-in-variables model to determine the consensus carbon isotope delta value, standard uncertainties due to characterization and homogeneity, and the individual laboratory results for FRUT-1 characterization on the VPDB-LSVEC scale. Table S10: Results from the errors-in-variables model to determine the consensus carbon isotope delta value, standard uncertainties due to characterization and homogeneity, and the individual laboratory results for FRUT-1 characterization on the VPDB scale. Table S11: Long-term stability assessment of BEET-1, GALT-1, and FRUT-1 for unopened and opened units stored under recommended conditions. Table S12: Results from the errors-in-variables model to determine the consensus carbon isotope delta value, standard uncertainties due to characterization and homogeneity, and the individual laboratory results for VANA-1 characterization on the VPDB scale. Table S13: Results from the errors-in-variables model to determine the consensus carbon isotope delta value, standard uncertainties due to characterization and homogeneity, and the individual laboratory results for VANB-1 characterization on the VPDB scale. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.
Chartrand et al. (Fri,) studied this question.