40Ar/39Ar dating of basaltic rocks and the pitfalls of plagioclase alteration

40Ar/39Ar geochronology is one of the most important techniques for constraining the timing of basaltic events due to the paucity of suitable minerals in basalts for other geochronological techniques such as U–Pb (e.g., zircon, baddeleyite). Among a variety of materials from basaltic rocks that have been used for 40Ar/39Ar dating, plagioclase is the most important due to its common presence in basalts as a primary crystallizing phase, and its transparency so that fresh grains can be selected during sample preparation. However, plagioclase 40Ar/39Ar geochronology has often been compromised by alteration (e.g., sericitization by hydrothermal events), which, in practice, is difficult to identify using a petrographic microscope when the amount of alteration is low (e.g., We used laboratory step-heating experiments and theoretical simulations to characterize the 40Ar/39Ar age and Ca/K spectra of altered plagioclase so that 40Ar/39Ar dating results on altered samples can be identified and better interpreted. The step-heating experiments and theoretical simulations yielded consistent results, and show that with the presence of even a tiny amount of sericite (~0.01% for K-poor samples and ~0.1% for K-rich samples), the plagioclase samples yielded alteration plateau ages that are 3%–4% younger than the crystallization age. The difference between the alteration age of sericitized plagioclase and its crystallization age is primarily controlled by the time lapse between the crystallization and sericitization events, but also by the Ca/K ratios of the plagioclase. For plagioclase samples that experienced the same alteration event, the higher the Ca/K ratio is, the more sensitive the 40Ar/39Ar age is to alteration. We propose that the alteration signatures of plagioclase can be effectively identified through inspecting the 40Ar/39Ar age spectra, the Ca/K spectra, and the degassing curves. We also investigated the effect of sericitization of plagioclase microliths in basaltic groundmass and modelled the 40Ar/39Ar age and Ca/K spectra of altered groundmass samples. We validate our approach by revisiting published 40Ar/39Ar dating results for large igneous provinces, and showed that these dates should have been interpreted as alteration ages (minimum eruption ages) rather than crystallization ages. Finally, we demonstrate that with high degrees of alteration (~50% for K-poor and >70% for K-rich plagioclase samples), the age of hydrothermal alteration can be successfully dated.