| Abstract | Zircon U/Pb geochronology using Laser Ablation-Inductively Coupled Plasma-Mass Spectrometry (LA-ICP-MS) is rapidly being adopted in the earth sciences. Challenges in the accurate determination of ages
and uncertainties, however, remain, and include finding the optimal ways to calculate and calibrate U/Pb and Pb/Pb measurements of unknowns. Of particular importance is the determination of U/Pb ratios that are complicated by laser-induced
fractionation of U relative to Pb. In order to address these challenges and provide best practice recommendations, this study systematically explores and fully documents three algorithms for calculating U/Pb and Pb/Pb ratios (ratio of mean signals,
mean of replicate ratios, and intercept of ratio vs. ablation time relationship), four different placements of calibration materials within a series of unknowns, and three different least-squares approaches for interpolation of drift correction
factors (mean, linear, and quadratic). The lowest mean relative standard errors (2?) for single uncalibrated laser analyses were obtained for the ratio of means and mean of ratios algorithms, and were ?1% for 206Pb/238U and 207Pb/ 206Pb. The
experimental uncertainty for the ratio of the mean and the mean of ratios algorithms was nearly three times higher for 206Pb/238U and 1.5 times higher for 207Pb/ 206Pb compared to that predicted from counting statistics. The experimental uncertainty
for the intercept method, however, for 206Pb/238U was nearly six times higher than that predicted due to counting statistics. These results suggest that additional and unaccounted for sources of error are present. The best reproducibility (2?) for
both 206Pb/238U (?2 to 3% 2RSD) and 207Pb/206Pb (?1.5% 2RSD) was also obtained using the ratio of means and mean of ratios algorithms. The most precise fits to calibration materials placed in various positions within a 'run' of twenty data
acquisitions were obtained using a quadratic fit to the data and when the calibration materials were closely spaced, as would be expected. There was no significant gain in accuracy, however, using a quadratic fit compared to other methods, suggesting
that the more complex fits are simply following noise in the data. This clearly demonstrates that drift corrections based on mean or linear least squares are most appropriate for optimizing precision and accuracy for U/Pb dating of zircon using
LA-ICP-MS. Close spacing of calibration materials between single or small groups of unknowns (6 or less) was shown to produce the most accurate results, demonstrating that long, continuous sequences of unknowns introduce significant deviation from
the true age of the samples. The mean precision (2?) of individual age analyses of the 1065 Ma 91500 zircon obtained in this study range from 16-38 Ma (1.5-3.5%, 2? RSE), improving as more acquisitions of calibration materials are utilized. The
accuracy was 1% or better. © 2010 The Royal Society of Chemistry. |