LAUSR

Abstract The use of LA-ICP-MS for uranium determination in the fission track dating technique is becoming increasingly popular because of several advantages over the classical external detector method and a variety of analytical and statistical protocols have been developed. However, two important issues remain unresolved in the context of the LA-ICP-MS approach (i) how to best deal with low track density (ρs) samples, and (ii) does a correlation between age and uranium content (or eU) reflect an annealing dependence or not? To assess the impact of the analytical methodology on these issues, we compare the multi-spot and more classical single spot methods on samples of known ages, variably zoned and/or with low track densities. To make the comparison we use an approach, implemented in a Python script, that randomly samples our multi-spot ICP-MS data to choose a single U measurement per grain, simulating the single spot approach. We then calculate the central age, p(χ2) and dispersion of the simulated single spot analysis and repeat this 2000 times. Our results show that the multi-spot approach is robust for low ρs and zoned samples, yielding both accurate and precise results without over-dispersion. Additionally, our random sampling approach shows that a single spot measurement can induce an overdispersion coupled to a relationship between single grain age and U content. This is at least partly attributable to zonation that creates a mismatch between the U in the counted area and the spot-measured U. Therefore, we recommend that if over-dispersion is observed for basement samples, when one typically expects a single age population, then multiple spot analysis should be carried out to assess if the excess dispersion is linked to undetected zoning and/or laser spot misplacement rather than to U dependent annealing behaviour.