Factors influencing the precision and accuracy of Nd isotope measurements by thermal ionization mass spectrometry

co-auteur étranger; International audience; Taking the example of Nd, we present a method based on a 4-mass-step acquisition scheme to measure all isotope ratios dynamically by thermal ionization mass spectrometry (TIMS); the aim being to minimize the dependency of all mass fractionation-corrected ratios on collector efficiencies and amplifier gains. The performance of the method was evaluated from unprocessed JNdi-1 Nd standards analyzed in multiple sessions on three different instruments over a period of ~ 1.5 years (n = 61), as well as from standards (18 JNdi-1 and 19 BHVO-2) processed through different chemical purification procedures. The Nd isotopic compositions of standards processed through fine-grained (25–50 μm) Ln-spec resin show a subtle but clear fractionation caused by the nuclear field shift effect. This effect contributes to the inaccuracy of Nd isotope measurements at the ppm level of precision.Following a comprehensive evaluation of the mass spectrometer runs, we suggest several criteria to assess the quality of data acquired by TIMS, in particular to see whether the measurements were affected by domain mixing effects on the filaments. We define maximum tolerable Ce and Sm interference corrections and the minimum number of ratios to acquire to ensure the best possible accuracy and precision for all Nd isotope ratios. Changes in fractionation of Nd isotope ratios in between acquisition steps can result in significant inaccuracy and bias dynamic μ142 values by > 15 ppm. To correct for these effects, we developed a systematic drift-correction based on the monitoring of Nd isotope ratios through time. The residual components of scatter in the JNdi-1 and BHVO-2 datasets were further investigated in binary isotopic plots in which we modeled the theoretical effects of domain mixing on filaments, nuclear field shift and correlated errors from counting statistics using Monte-Carlo simulations. These plots indicate that the 4-step method returns precisions limited by counting errors only for drift-corrected dynamic Nd isotope ratios. Data acquired on three different TIMS instruments suggest the following composition for the JNdi-1 reference standard: 142Nd/144Nd = 1.141832 ± 0.000006 (2s), 143Nd/144Nd = 0.512099 ± 0.000005 (2s), 145Nd/144Nd = 0.348403 ± 0.000003 (2s), 148Nd/144Nd = 0.241581 ± 0.000003 (2s), and 150Nd/144Nd = 0.236452 ± 0.000006 (2s) when normalized to 146Nd/144Nd = 0.7219. Measurements performed on different instruments (Triton™ vs. Triton Plus™) show resolvable differences of about 10 ppm for absolute 143Nd/144Nd, 145Nd/144Nd and 148Nd/144Nd ratios. The different criteria and corrections developed in this study could be applied to other isotopic systematics to improve and better evaluate the quality of high-precision data acquired by TIMS.