SIMS analyses of oxygen isotopes: Matrix effects in Fe-Mg-Ca garnets

Large instrumental mass fractionation (IMF) may occur during measurements of oxygen isotope ratios by SIMS. Part of this fractionation depends on crystal structure and mineral composition. In order to improve the accuracy of SIMS measurements, we gathered 6 commonly used garnet standards and prepared 6 others to adequately cover the composition range Alm_0-73, Prp_0-99, Grs_0-20. Electron microprobe analyses were performed at UBP-Clermont to check the chemical homogeneity of these standards. Oxygen isotope compositions were determined by laser fluorination and mass spectrometry at UW-Madison. Ten SIMS sessions and 336 δ¹⁸O measurements at CRPG-Nancy, on a Cameca IMS1270 instrument, demonstrate that the standards are homogeneous with external reproducibility of 0.3‰ (1σ). In terms of δ¹⁸O, SIMS measurements indicate that, during a single session, IMF can vary up to 6.3‰ from one garnet standard to another. In most of the sessions, IMF can be correlated with the grossular content. However, for a satisfactory correction scheme, we suggest the combination of the 3 main components (Ca, Fe, Mg). This is done using a simple least square calculation routine. The correction coefficients determined for each session can be used to calculate the IMF and correct the measured isotopic ratio of a garnet of known chemical composition. This way, we were able to reproduce the δ¹⁸O values of most of the Fe-Mg-Ca garnet standards within ± 0.6‰. Interestingly, the use of only 3 end-member standards (AlmCMG, PrpMM, GrsSE) plus a standard of intermediate composition (e.g. UWG-2) is sufficient to reproduce δ¹⁸O within the same precision. Thus, linear interpolation among end-member standards is satisfactory in the case of the garnet solid-solutions. Two studies carried out on zoned garnets from the Alps and the Pyrenees indicate that matrix effects become significant when variations in grossular contents are important (> 10%). In order to obtain reliable isotope ratio measurements on Fe-Mg-Ca garnets using a SIMS, we suggest a correction scheme using at least 3 reliable end-member standards plus a standard of intermediate composition (a garnet standard closest to the average composition of the analysed garnet). This allows cross-checking and incorporates a correction based on the variations in composition of zoned crystals.