A systematic evaluation of titanite reference materials for optimizing trace element and U-Pb analysis by LA-ICP-MS

Titanite (CaTiSiO5) is a commonly occurring and versatile accessory mineral with broad applications in petrochronology. In situ U-Pb and trace element analyses via SIMS and LA-ICPMS are routinely performed using a matrix-matched reference material for U-Pb and standard glasses (non-matrix matched reference material) for elemental abundance determination. We report U-Pb isotopic ratios and major and trace element concentrations for three titanite samples (Ecstall, McClure and FCT) which are commonly used as reference materials in petrochronology studies. In addition, we characterize two new samples which can potentially serve as matrix-matched reference materials for titanite trace element geochemistry (BLR-2 and BRA-1). Based on electron microprobe analysis, samples BLR-1 and BLR-2 are homogeneous and suitable for use as a primary reference material for trace element concentrations. Whereas Ecstall, McClure, and FCT titanite reference materials show high intra-grain heterogeneity, yielding relative standard deviations for most trace elements between 5% and 40%, with higher standard deviations for U of 70% for Ecstall (n = 26), 265% for McClure (n = 22), and 202% for FCT (n = 26). Therefore, we suggest that these grains are unsuited to serve as reference materials for trace element quantification. The BRA-1 titanite has low trace element concentrations and is chemically heterogeneous (total REE abundances of 40 ppm for the rim and 95 ppm for the core of the grain), thus is not suitable for standardization of chemical composition using LA-ICPMS. It is commonly asserted that a matrix-matched standardization provides a more robust downhole fractionation correction compared to a non-matrix matched standardization. However, it remains unclear which standardization approach (matrix-matched vs non-matrix matched/glass) is more accurate for titanite trace element quantification. To resolve this, we tested several standardization approaches for trace element quantification, comparing matrix-matched (BLR-1) and non-matrix-matched (NIST612) standardizations with different internal elemental standards (IES; Ca, Si and Ti) and without internal standardization (semi-quantitative). To provide an independent constraint on the accuracy of the various trace element standardization techniques we compared results to trace element concentrations obtained via solution Q-ICPMS on crushed BLR-2 and BRA-1 aliquots. The matrix-matched standardization using Si as the IES yields the best reproducibility of trace element concentrations followed by the matrix-matched reduction using Ti as the IES. Moreover, the matrix-matched semi-quantitative correction yielded the lowest weighted percentage of difference compared to reference trace composition quantified by solution ICPMS. Finally, in this contribution we also benchmark sampling-size for precise U-Pb dating of common-Pb rich phases like titanite.