U and Th content in magnetite and Al spinel obtained by wet chemistry and laser ablation methods: implication for (U–Th) ∕ He thermochronometer

Abstract. Magnetite and spinel thermochronological (U–Th) / He dates often display significantly dispersed values. In the present study, we investigated the contribution of analytical (and standardization) errors to this dispersion. U and Th content of magnetite (natural and synthetic) and natural Al spinel samples with U and Th concentrations between 0.02 and 116 µg g−1 were analyzed using both wet chemistry and in situ laser ablation sampling methods. New magnetite reference samples (NMA and NMB) were synthesized, consisting of U- and Th-doped nano-magnetite powders, whose U and Th concentrations were determined using a wet chemistry method (U and Th of NMA and NMB are ∼40 µg g−1 and ∼0.1 µg g−1, respectively). We show that, for both U and Th analyses, the reproducibility obtained with the wet chemistry protocol depends on the U and Th concentration. It is below 11 % for U–Th values higher than 0.4 µg g−1 and reaches 22 % for U–Th content lower than 0.1 µg g−1. This result implies that (U–Th) / He thermochronological ages cannot be more reproducible than 24 % for magnetite containing less than 0.1 µg g−1 of U and Th, thus explaining part of the natural ages variability. U and Th data obtained by laser ablation ICP-MS on natural magnetite and Al spinel samples were calibrated using both silicate glass standards and synthetic magnetite samples. The U and Th contents determined using NMA are consistent with those obtained by means of the wet chemistry method, but they are overestimated by 30 % when using the glass standard samples only. These results highlight the impact of the matrix effect on the determination of the U–Th content in magnetite. We thus recommend the use of a well-characterized magnetite reference for the calibration of the U–Th signals obtained by laser ablation. The scatter in the (U–Th) / He magnetite ages can be expected to be ∼20 % if the U and Th contents are determined by laser ablation. This level of precision is actually not significantly different from that obtained using the wet chemistry method, which paves the way for the use of laser ablation for determining (U–Th) / He ages. In the absence of a spinel reference for U and Th calibration using LA-ICP-MS, silicate glass references, along with NMA, were used. U and Th contents were found to be ∼30 % lower than the values obtained using wet chemistry. This discrepancy underlines the importance of using a standard with a composition close to that of the mineral of interest. Although magnetite and Al spinel have related crystal structures, the magnetite standard is not appropriate for U and Th analysis in Al spinel using LA-ICP-MS.