Quantification of Stable Strontium Isotope Variability in Nature Using MC- ICPMS

Recent findings of stable isotope variations in elements as heavy as Tl raise the possibility that there may be significant naturally induced isotopic variation in Sr. Convention in traditional 87Sr/86Sr measurements is to use 86Sr/^{88}Sr = 0.1194 to correct instrumental fractionation. Although the absolute value used for this correction is unimportant, implicit in its widespread utility is the assumption that the 86Sr/^{88}Sr ratio in standard materials is the same as that in samples, and that it is constant. Instrumental mass bias normalization to 86Sr/^{88}Sr thereby eliminates any scope for examining mass dependent fractionation of 86Sr/^{88}Sr. In this study, we present a new technique for measuring both the radiogenic Sr isotope variations (87Sr/86Sr) and possible small-scale mass dependent isotopic fractionation of the 86Sr/^{88}Sr ratio. The technique involves the use of admixed Zr for mass bias correcting Sr isotope ratios using the Durham Neptune MC-ICPMS. Analyses, in static multi-collection mode, comprised 1 block of 50 cycles with 4 sec. integrations. Data were processed offline for mass bias and interfering element corrections. Rather than using an `accepted' ^{90}Zr/^{91}Zr ratio for mass bias correcting measured Sr isotope ratios we determined a `true' ^{90}Zr/^{91}Zr value on our instrument for each analytical session by doping a 200 ppb NBS 987 solution with (50ppb) Zr. The measured ^{88}Sr/86Sr (after correction for 86Kr interference on mass 86Sr) was used to mass bias correct the ^{90}Zr/^{91}Zr using an exponential law and an ^{88}Sr/86Sr ratio of 8.375209 (equivalent to 86Sr/^{88}Sr = 0.1194). The ^{90}Zr/^{91}Zr value thus determined was then used to mass bias correct the 86Sr/^{88}Sr (and 87Sr/86Sr) in samples and standards. Long-term external precisions of <100ppm and <30ppm respectively were achieved. Using this technique we present 86Sr/^{88}Sr data for a range of natural rock standards, NBS 987 and some extraterrestrial samples (for which 86Sr/^{88}Sr fractionation has already been documented using double spike TIMS). Incomplete recovery of Sr from column separation can significantly fractionate Sr isotopes and shows that excluding overlapping Rb and Sr tails from column procedures may cause inaccuracies. We clearly demonstrate that NBS987 is fractionated relative to high-T magmatic rocks (possibly resulting from the manufacturing process), and establish a precise "high-T" reference line. Due to documented variation in low-T and biological samples we suggest that this high-T reference should be adopted for Sr isotope work.