A method for determining isotopic composition of elements by thermal ionization source mass spectrometry. Application to strontium.

It is shown that in thermal ionization source mass spectrometry, if isotope fractionation of the element in the sample follows a linear law, straight-line distributions in xm versus xm/ym diagrams are observed, where xm and ym are two measured isotope ratios. The slopes and y-intercepts of these linear distributions are functions of the 'true' (starting) values xt and yt of the element in the sample and of the masses of the isotopes involved in ratios x and y. Since the masses of the nuclides are known, true ratios xt and yt can be calculated. This theoretical result is used to determine the non-radiogenic part of the isotopic composition of strontium in NBS SRM 987, one 84Sr-enriched isotopic tracer prepared at the Oak Ridge National Laboratory, and two natural samples (rocks from the metamorphic basement of the Italian Eastern Alps) without any assumption about the isotopic composition itself. Strontium was loaded as nitrate on single tungsten filaments, and 88Sr/86Sr and 84Sr/86Sr ratios were measured up to a fractionation of ~1% u-1 in a single-collector VG 54E mass spectrometer. For each run, 86Sr/88Sr, 84Sr/86Sr and 84Sr/88Sr ratios were calculated for all useful xm versus xm/ym distributions. The respective weighted average values are considered the true values of the isotope ratios in the sample. Four runs of isotopic standard NBS SRM 987 and one run of the isotopic tracer gave accurate and reproducible results which are identical, within error limits, to the respective certified values. The four determinations of NBS 987 resulted in the following weighted average values: 86Sr/88Sr = 0.11942 ± 0.00018; 84Sr/86Sr = 0.056485 ± 0.000075; 84Sr/88Sr = 0.006746 ± 0.000017 (error at 2? level). The values of the natural 86Sr/88Sr ratio (two rocks: 0.11956 ± 0.00017 and 0.11957 ± 0.00008; NBS 987: 0.11942 ± 0.00018) are identical within error limits, and identical or very close to the recommended value of 0.1194, the worldwide assumed 'true' 86Sr/88Sr value in the commonly used procedure of determining 87Sr/86Sr ratio by normalization. However, due to the accuracy of the above determinations, it is suggested that, in nature, significant differences exist in the non-radiogenic part of the isotopic composition of strontium.