Causes and mitigation of U-Pb fractionation during LA-ICPMS analyses of zircon using nanosecond excimer laser systems

Abstract. This work investigates the causes of measurement biases in U-Pb dating of zircon by laser ablation inductively coupled mass spectrometry (LA-ICPMS) and possible methods for correction. Plasma-induced biases include oxidative loss of U, which is normally minimized by restricting nebulizer Ar flow at the cost of lower sensitivity, mass bias and ionization efficiency, which together normally produce a negative bias of several tens of percent for ²⁰⁶Pb/²³⁸U. There is also an ablation bias due to the relative volatility of Pb, which depends to some extent on the structural state of the zircon. We suggest that oxidation of elements in the plasma is caused by turbulent entrainment of oxygen from the surrounding air and point out a flaw in the design of some movable two-stage ablation chambers that can result in variable degrees of oxidation as position is changed. Measurements of NIST glass and zircon include significant ablation bias even when scanned. The main cause of increasing ²⁰⁶Pb/²³⁸U ablation bias seen in zircon and other minerals appears to be sequestration of Pb-depleted melt in fallback and along the sides of the pit. Integrated signal profiles measured using laser pulses at 0.2 Hz combined with modelling of U/Pb fractionation suggest that the earliest Pb/U measurements (first 10 pulses) are affected by decreasing fractionation from a melt pool as it becomes increasingly depleted in Pb. This trend is opposed by deposition of depleted material as fallback, which dominates signal loss for the first 10 pulses but rapidly decreases. The fractionation sequence from the first 10 or so pulses is therefore chaotic. Ratios from the following 50 or so pulses show an approximately linear increase in fractionation. Normalized data from these pulses give trends with higher intercepts and lower slopes for standards with higher radiation damage during the same session. Fractionation and signal decay subsequently rise more slowly but remain linear, probably because fractionation is dominated by deposition in the deepening pit. An ablation fractionation model is proposed based on the drop in measured signal but this cannot be used to estimate accurate bias-free ²⁰⁶Pb/²³⁸U ratios because of the number of unconstrained parameters. The best approach for calibrating against an unknown is a direct comparison of all or part of the standard ratio profile with the sample profile after multiplication by a calibration factor. The factor that results in the best fit should represent a ratio of unbiased ²⁰⁶Pb/²³⁸U between standard and sample. Software is included to process and calibrate data. Data from Precambrian zircon with well-established ²⁰⁷Pb/²⁰⁶Pb ages suggest that radiation damage below the metamict state results in little bias to discordance. Reverse discordance from metamict zircon appears to be approximately proportional to U concentration. Increased accuracy of ²⁰⁶Pb/²³⁸U ages using nanosecond laser systems can most likely be achieved through design improvements rather than data processing. Sensitivity should be increased and ablation bias decreased by reducing the volume in the sampling cup to minimize fallback. Plasma cooling in a flow of N₂ gas from a liquid nitrogen dewar should enable increased sensitivity without oxidation of U. The use of H₂, instead of He as a carrier gas should reduce fallback and cost but would require design changes to prevent exposure to the plasma before complete venting of O₂ from the ablation chamber.