40Ar/39Ar dating of basaltic rocks and the pitfalls of plagioclase alteration
Introduction
Precise and accurate dating of basaltic terrestrial rocks is a prerequisite for understanding a large variety of geological and geodynamical phenomena, such as the link between mantle plumes, supercontinent breakups, and intraplate alkaline magmatism (e.g., Dalrymple and Clague, 1976, Turner et al., 1994, Marzoli et al., 1999, Olierook et al., 2019); the relationship between large igneous provinces (LIPs), impact events, and mass extinction and oceanic anoxic events (e.g., Jourdan et al., 2014, Burgess et al., 2015, Renne et al., 2015, Schoene et al., 2019); and the formation mechanisms of volcanic-hosted ore deposits (e.g, Camprubí et al., 2003, Barrote et al., 2020) on Earth. Understanding the early history and evolution of the Moon, Mars, and asteroids, such as their collisional/impact history, timing of volcanic activity, and metamorphic/cooling history, also depends on accurate and precise age constraints on basaltic rocks (e.g., Cohen et al., 2001, Iizuka et al., 2019, Thiessen et al., 2019, Jourdan et al., 2020, Merle et al., 2020), which constitute the most abundant crustal material in the solar system.
Materials from basalt that can be used for 40Ar/39Ar dating include plagioclase, pyroxene, groundmass (or whole rock for aphyric basalt), biotite and hornblende. Despite a few recent cases of successful dating on pyroxene separates in terrestrial basalt (Ware and Jourdan, 2018, Konrad et al., 2019, Zi et al., 2019, Jiang et al., 2021c), 40Ar/39Ar dating of pyroxene is still challenging due to its very low K content and reliance on the absence of thick exsolution lamellae which prevent irradiation-induced recoil redistributions (Ware and Jourdan, 2018). Hydrated minerals such as biotite and hornblende are rich in K and have the ability to yield high-precision 40Ar/39Ar ages (e.g., Reichow et al., 2009, Ware et al., 2018, Sahoo et al., 2020). However, these minerals are rarely present in mafic rocks and therefore plagioclase and groundmass are still the most commonly used materials in dating low-K basalt (e.g., Koppers et al., 2000, Renne et al., 2015, Gomes et al., 2020).
There is a growing number of publications demonstrating that groundmass of relatively old samples (e.g., >30 Ma) do not always provide reliable crystallization age for the basaltic rocks despite generating concordant plateau ages (Hofmann et al., 2000, Baksi, 2007a, Jourdan et al., 2007, Renne et al., 2015, Merle et al., 2019, Jiang et al., 2021a). This is because groundmass samples are prone to be affected by: (i) K leaching due to hydrothermal alteration and/or 39Ar recoil loss (Jourdan et al., 2007, Renne et al., 2015), which can lead to older apparent plateau ages, and/or (ii) 40Ar* (radiogenic Ar) loss and/or recrystallization from cryptic hydrothermal alteration (Hofmann et al., 2000), which will lead to younger apparent plateau ages (e.g., Merle et al., 2019). For groundmass, even though potential signs of alteration (e.g., brown, red or yellow discolouration) can be avoided during sample preparation, cryptic alteration cannot be easily observed since groundmass is not transparent. Plagioclase is transparent and thus, contrary to groundmass, fresh plagioclase grains can be separated from altered ones under a binocular petrographic microscope and, if only fresh plagioclase can be isolated, can yield reliable crystallization ages for basaltic rocks.
Nevertheless, the often very low K2O abundance (<0.1 wt.%) of plagioclase makes 40Ar/39Ar dating on plagioclase extremely sensitive to the alteration of plagioclase, as the latter process is accompanied with neo-crystallization of sericite, a fine-grained variety of muscovite with ~10 wt.% of K2O. The sericitization of plagioclase refers to the replacement of plagioclase (NaAlSi3O8–CaAl2Si2O8) by sericite [KAl3Si3O10(OH,F)2] through filling in microfractures within the plagioclase crystal and/or mineralogical replacement when hydrothermal fluids percolate through the plagioclase at low-temperature (100–350 °C) and low-pressure conditions (e.g., Morad et al., 2010, Li et al., 2013). The neo-crystallized sericite is a newly added phase and thus, does not retain any 40Ar* from the plagioclase and starts its own K/Ar clock at the time of formation. Furthermore, the process of alteration can be accompanied by 40Ar* diffusive loss associated with the temperature rise caused by the hydrothermal fluids. These processes would cause the loss of 40Ar* without a concomitant loss of K, and lead to lowered apparent age compared to the crystallization age. As a result, the 40Ar/39Ar ages provided by partially or fully sericitized plagioclase can only be regarded as strict minimum eruption ages (e.g., Rohde et al., 2013, Olierook et al., 2015).
Although one can exclude sericite grains and partially sericitized plagioclase grains, and select fresh plagioclase grains for 40Ar/39Ar dating during sample preparation processes, the sericitization of plagioclase below 1% is hard to visually detect using a petrographic microscope (Fig. 1). It is possible to measure the compositions of separated grains by using in situ analytical methods such as Scanning Electron Microscopy (SEM), Secondary Ion Mass Spectrometry (SIMS), or Electron probe micro-analyzer (EPMA), in order to ensure the purity and freshness of plagioclase grains, however, these analyses require the grains to be mounted in epoxy, which could contaminate the plagioclase with K and/or Cl. Alternatively, thin sections can be used for these analyses, however, it is impossible to assess the proportion of sericite in the very grains that are separated for dating. Furthermore, in order to obtain a precise age, several mg to tens of mg (i.e., hundreds to thousands of crystals) of plagioclase are needed, so measuring the grains individually to detect possible sericitization is impractical. Ultraviolet laser ablation can be used to measure a spot 40Ar/39Ar age from a single plagioclase grain (e.g., Kelley et al., 1994, Mark et al., 2008), but this technique is too imprecise for most applications related to LIPs, impact events, and extraterrestial rocks. Therefore, assessing the effect of alteration on plagioclase 40Ar/39Ar geochronology is very important in order to determine whether a given 40Ar/39Ar age truly reflects the crystallization age of the dated material or is biased by alteration.
Previously, Verati and Jourdan (2014) used theoretical modelling to demonstrate how the K/Ar and 40Ar/39Ar total fusion age can be lowered by different degrees of sericitization of plagioclase with various Ca/K ratios (proxy for K abundance). In practice however, one uses 40Ar/39Ar age spectra (40Ar/39Ar apparent ages plotted against cumulative percentage of 39Ar released in a step-heating experiment) rather than total fusion ages to date plagioclase. This is because for 40Ar/39Ar plateau ages, it is possible to assess if the K/Ar system stayed closed since formation, by assessing the reproducibility of the ages through robust statistical tests (e.g., MSWD and/or p; Baksi, 2007b); and to detect the presence of excess 40Ar by testing whether the trapped 40Ar/36Ar ratio is atmospheric, or if not, using an inverse isochron approach (McDougall and Harrison, 1999).
In this study, we illustrate the behaviour of 40Ar/39Ar age and Ca/K spectra of plagioclase crystals that have been affected by various degrees of sericitization, by using a theoretical simulation and step-heating experimental approach on mixtures of plagioclase and muscovite (an ideal proxy for sericite or fully sericitized plagioclase both chemically and regarding diffusion kinetics). The laboratory step-heating experiments were conducted on mixtures with known proportions of plagioclase and muscovite that aimed to reproduce the physical process of a plagioclase–sericite mixture, and its observable effect on 40Ar/39Ar age spectra. The theoretical simulations aim to model the geological processes that occur in nature, including the 40Ar* accumulation since the crystallization of the plagioclase, the 40Ar* loss associated with the increase of temperature due to hydrothermal fluids, and the Ar degassing behaviour from theoretical plagioclase and sericite in a step-heating experiment. The results are then used to discuss how ‘fresh’ the plagioclase needs to be in order to produce reliable crystallization ages, how sericite signatures can be identified in 40Ar/39Ar age and Ca/K spectra, and how ‘altered’ the sericitized plagioclase needs to be to yield the age of hydrothermal alteration events.
Section snippets
Experimental sample selections
We use plagioclase samples separated from basalts from the Greater Kerguelen LIP in Bunbury, Western Australia (BN21; Ca/K = 143; Olierook et al., 2016), and from the Deccan Traps in the Western Ghats, India (JW4; Ca/K = 52; Hofmann et al., 2000). These plagioclase samples were selected because they both yielded good 40Ar/39Ar plateaus, and have different Ca/K ratios. As a proxy for sericite, we select a muscovite sample from the granitic dykes intruding the peridotitic Beni Bousera unit
Step-heating experimental 40Ar/39Ar age and Ca/K spectra of sericitized plagioclase
The laboratory step-heating experimental results of plagioclase samples BN21 and JW4 with 0%, 0.01%, 1%, 5%, 10% and 20% of sericite are shown in Fig. 3, Table 2, and Supplementary Table 4.
Discussion
Hereafter, for an easier reading we use the superscripts Mo (model) and Ex (experiment) to differentiate between the two types of results when appropriate.
Conclusions
Using a combination of laboratory step-heating experiments and theoretical simulations, we modelled the effect of alteration (sericitization) on plagioclase 40Ar/39Ar geochronology, and demonstrated that:
- (1)
Plagioclase 40Ar/39Ar ages can be detrimentally affected by even a tiny amount of sericite. For plagioclase samples that are relatively rich in K (low Ca/K), and relatively poor in K (high Ca/K), the existence of 0.1% and 0.01% of sericite, respectively, can lead to a ‘fake’ 40Ar/39Ar plateau
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
We thank A. Frew for technical help with the sample analysis. Jiang acknowledges support of the CSC–CIPRS scholarship. Helpful comments from C. Sprain and an anonymous reviewer, and editorial handling by D. Shuster and J. Catalano are gratefully acknowledged.
References (88)
- et al.
4D history of the Nimbus VHMS ore deposit in the Yilgarn Craton, Western Australia
Precambrian Res.
(2020) - et al.
Middle Permian paleomagnetism of the Sydney Basin, Eastern Gondwana: Testing Pangea models and the timing of the end of the Kiaman Reverse Superchron
Tectonophysics
(2017) - et al.
Geochemistry of early Mesozoic tholeiites from Morocco
Earth Planet. Sci. Lett.
(1982) - et al.
High-precision geochronology links the Ferrar large igneous province with early-Jurassic ocean anoxia and biotic crisis
Earth Planet. Sci. Lett.
(2015) - et al.
Systematic variations of argon diffusion in feldspars and implications for thermochronometry
Geochim. Cosmochim. Acta
(2013) - et al.
40Ar/39Ar geochronology and volume estimates of the Tasmantid Seamounts: Support for a change in the motion of the Australian plate
J. Volcanol. Geoth. Res.
(2017) - et al.
Age of the Hawaiian-Emperor bend
Earth Planet. Sci. Lett.
(1976) - et al.
40Ar/39Ar dating of the Jurassic volcanic province of Patagonia: migrating magmatism related to Gondwana break-up and subduction
Earth Planet. Sci. Lett.
(1999) - et al.
40Ar/39Ar dating of volcanism and subsequent very low-grade metamorphism in a subsiding basin: example of the Cretaceous lava series from central Chile
Chem. Geol.
(2005) - et al.
Timescales and mechanisms of plume–lithosphere interactions: 40Ar/39Ar geochronology and geochemistry of alkaline igneous rocks from the Paraná-Etendeka large igneous province
Earth Planet. Sci. Lett.
(2006)