Critical evaluation of internal and external calibration methods for accurate strontium isotope ratio determination

The isotopic composition of strontium n(⁸⁷Sr)/n(⁸⁶Sr) provides important information relevant to earth, life, food, archaeological and anthropological sciences. To determine accurate n(⁸⁷Sr)/n(⁸⁶Sr) isotope ratios, it is necessary to perform instrumental isotopic fractionation (mass bias) correction, which is a popular metaphor for isotope ratio calibration. Despite the importance of calibration in n(⁸⁷Sr)/n(⁸⁶Sr) analysis, experimental studies which critically evaluate the accuracy of conceptually different calibration methods remain scarce. Therefore, this work 1) investigates if n(⁸⁷Sr)/n(⁸⁶Sr) isotope ratios differ depending on the calibration method used and 2) discusses whether these differences might impact the outcome of a study on past human mobility.

This work compares n(⁸⁷Sr)/n(⁸⁶Sr) results of the same physical sample, which were determined using internal calibration (internal mass bias correction) and external calibration (standard-sample-bracketing) in a large sample set. Reference materials and environmental samples showed consistent results, while 66 % of investigated tooth enamel samples showed significantly different n(⁸⁷Sr)/n(⁸⁶Sr) isotope ratios depending on the calibration method used. In-depth evaluation ruled out sample preparation or measurement errors. Instead, internal calibration delivered biased results, because the δ(⁸⁸Sr/⁸⁶Sr)_SRM987 values in the affected samples deviated by more than -0.60 ‰ from the - assumingly constant - reference value. The bias amounted up to 0.00043 and was 2-times larger than theoretically predicted. Ultimately, using internally calibrated n(⁸⁷Sr)/n(⁸⁶Sr) data, led to false classification of 6 % of individuals in a provenance study focusing on past human migration. In contrast, external calibration yielded accurate data leading to classification results that were in alignment with existing knowledge on the population under investigation.

This work empirically confirms theoretical considerations that the traditionally used calibration method in n(⁸⁷Sr)/n(⁸⁶Sr) isotope ratio analysis i.e.; internal calibration can produce significantly biased data. Moreover, it shows for the first time that the choice of calibration method has an impact on the outcome of a provenance study. Therefore, future investigations should use a more reliable calibration method in case high accuracy n(⁸⁷Sr)/n(⁸⁶Sr) isotope ratios are needed. External calibration via standard-sample-bracketing represents an attractive option.