In situ analysis of Th-230-Th-232-U-238 ratios in titanite by fs-LA-MC-ICPMS

The potential of femtosecond laser ablation multi-collector inductively coupled plasma mass spectrometry (fs-LA-MC-ICPMS) for in situ analysis of U-Th disequilibria in titanite was investigated. The aim of the study was to resolve spatial variations in (Th-230/U-238) ratios (where parentheses denote activity) in titanite from slowly cooled magma bodies. An in-house titanite glass (TG2), determined to be in secular equilibrium by solution mode MC-ICPMS (i.e. (Th-230/U-238) = 1), was used to correct for U-Th elemental fractionation by sample standard bracketing. The effect of instrument operating conditions on the accuracy and reproducibility of (Th-230/U-238), (Th-232/U-238) and (Th-230/Th-232) ratios was studied by analyses of titanite minerals with known composition and a secondary titanite glass standard. The (Th-230/Th-232) data were found to be accurate and reproducible, independent of the instrument setting used, suggesting that corrections made for SEM-Faraday gain and abundance sensitivity were appropriate. However, plasma conditions, laser ablation mode, laser energy and wavelength, and titanite material properties were all found to variably influence the U-Th elemental fractionation and compromise the accuracy of the (Th-230/U-238) data to different extents. Hot plasma conditions significantly reduce the fractionation between U and Th. A drift in elemental fractionation was observed during single spot analyses using NIR laser ablation and results in errors of up to 29% on the (Th-230/U-238) data. The magnitude of the drift in the elemental fractionation was different for different laser wavelengths and energies. Ablation using the UV single spot mode was significantly less affected by variable elemental fractionation compared to NIR spot analyses, but precision was limited by lower sample uptake. Scanning mode analyses were not compromised by temporal variation of the U-Th intensity ratios but the degree of elemental fractionation was variable between analyses of different materials (e.g. glass versus minerals). This observation suggests material-dependent differences in U-Th fractionation even for near identical titanite compositions. Analyses of the secondary titanite glass standard TG1 bracketed by TG2 yield the most reproducible and accurate (Th-230/U-238) data, indicating more adequate correction for elemental fractionation when the calibration standard is matched in terms of material composition and structure.