A comparison of geochronological methods commonly applied to kimberlites and related rocks: Three case studies from Finland

Abstract

Despite their economic and scientific significance, the tectonic drivers that lead to the genesis of kimberlites and related rocks remain enigmatic. The presence of any spatio-temporal relationship(s) between the rift, drift and collision history of continents and the emplacement of kimberlites can only be elucidated with accurate geochronological constraints. In this study, we evaluated the three most common radiogenic dating methods used to determine the emplacement ages of kimberlites and related rocks - Rb-Sr phlogopite, U-Pb perovskite and ⁴⁰Ar/³⁹Ar phlogopite. We selected minimally altered samples from the Kaavi-Kuopio kimberlites, ultramafic lamprophyres and kimberlites from Kuusamo, and olivine lamproites from the Lentiira-Kuhmo-Kostomuksha cluster, all located in Finland. Our 33 new age determinations indicate that the olivine lamproites and ultramafic lamprophyres were emplaced on the Karelian craton during the interval between ~1180–1210 Ma, the Kuusamo kimberlites between ~730–750 Ma, and the Kaavi-Kuopio kimberlites between ~585–620 Ma.

Our results show that the Rb-Sr method regularly yields reproducible eruption ages from either magmatic or macrocrystic (xenocrystic) mica. In addition, the method can be employed on a single population of mineral separates using a sequential acid-leaching technique, without the need for an estimate of ⁸⁷Sr/⁸⁶Sr_(i). The U-Pb perovskite technique typically requires significant correction for ‘common-Pb’ i.e. Pb incorporated at the time of mineral formation and unrelated to subsequent radiogenic ingrowth. Here, we utilised isotope dilution analyses of co-existing spinel to extend the ‘spread’ in U-Pb isochron space obtained by in situ LA-ICP-MS analyses of perovskite and enhance the accuracy and precision of intrusion ages. Our tests also show that two-point U-Pb isochrons based on solution analyses of perovskite and spinel generally provide anomalously older ages. Relative to Rb-Sr and U-Pb results from the same kimberlite, we show that apparently precise ⁴⁰Ar/³⁹Ar plateau-ages, determined from step-heating experiments, are often anomalously ‘old’, which is attributable to ³⁹Ar recoil effects. In cases where the ⁴⁰Ar/³⁹Ar step-heating spectra are even marginally ‘disturbed’, ⁴⁰Ar/³⁹Ar total-gas ages compare favourably with ages from other geochronometers and appear to provide accurate estimates for the emplacement age of kimberlites and related rocks.

Overall, there is good agreement between the three geochronometers using the above approaches, which indicates that Rb-Sr phlogopite, U-Pb perovskite and ⁴⁰Ar/³⁹Ar phlogopite dating can provide robust age constraints for the emplacement of kimberlites and related rocks. The geochronometer of choice will rely on the freshness, abundance and size of datable material. Wherever possible, the application of multiple geochronometers is recommended.

Introduction

Kimberlites are rare, small-volume, volatile-rich igneous rocks that derive from deep within the Earth (Mitchell, 1986; Giuliani and Pearson, 2019). These rocks have attracted considerable economic and scientific interest because they: 1) are the principal primary economic host rocks to diamonds at the Earth's surface; 2) provide a window into the Earth's deep mantle; and 3) entrain and transport a cargo of mantle and crustal xenoliths which provide unique insights into the nature of the subcontinental lithosphere. Although kimberlite magmas have been emplaced on every continent over the last 2.8 billion years (Henning et al., 2003; Tappe et al., 2018; Heaman et al., 2019), there remains debate concerning the sources of kimberlites and what triggers mantle melting to form these enigmatic rocks. Robust determination of the timing of kimberlite eruption is a crucial prerequisite if we are to identify the presence of any spatiotemporal relationships between kimberlite emplacement and plumes from the core-mantle boundary or specific tectonic processes including the rifting, drifting and collision history of continents (e.g., Heaman and Kjarsgaard, 2000; Jelsma et al., 2009; Torsvik et al., 2010; Tappe et al., 2018).

Obtaining robust emplacement ages for kimberlites is commonly complicated because kimberlites are readily altered by syn- and post-emplacement interaction with hydrothermal fluids (e.g., Sparks, 2013; Giuliani et al., 2017) and are heterogeneous, ‘hybrid’ rocks consisting of both magmatic and xenocrystic (crust and mantle-derived) components (Mitchell et al., 2019). As a result, the abundance and grain-size of minerals of known origin, which are amenable to radiometric dating techniques (e.g., mica, perovskite, zircon, apatite) is variable and can be limited. A number of approaches have been applied to kimberlite geochronology (see Heaman et al., 2019), and the most routinely employed geochronometers are Rb-Sr phlogopite, ⁴⁰Ar/³⁹Ar phlogopite and U-Pb perovskite. Additional techniques include U-Pb dating of zircon (e.g., Davis et al., 1976; Kinny et al., 1989; Kamenetsky et al., 2014; Giuliani et al., 2015), (U-Th)/He dating (e.g., Blackburn et al., 2008; McInnes et al., 2009; Stanley et al., 2013; Stanley and Flowers, 2016) and fission-track analysis (e.g., Haggerty et al., 1983) of zircon and apatite. These latter two techniques are seldom used to constrain the age of emplacement because these systems have low closure temperatures. Consequently, (U-Th)/He and fission-track methods are more amenable to constraining events that post-date kimberlite emplacement, such as erosion and uplift (Stanley et al., 2013).

In this study, we employed U-Pb perovskite, Rb-Sr phlogopite and ⁴⁰Ar/³⁹Ar phlogopite methods to determine the emplacement ages of 12 kimberlites, one ultramafic lamprophyre and three olivine lamproites from the Kaavi-Kuopio field, Kuusamo cluster and the Lentiira-Kuhmo-Kostomuksha cluster in the Karelian Craton (Finland and north-western Russia; Fig. 1). Although age determinations exist for the Lentiira-Kuhmo-Kostomuksha cluster (Belyatskii et al., 1997; O'Brien et al., 2007), the emplacement ages of the other two volcanic fields are poorly constrained. By comparing data for up to three isotopic systems on single samples from the same kimberlite intrusion, this study provides a comprehensive evaluation of the dating systems typically employed in kimberlite and related rock geochronology.

The application of Rb-Sr dating of phlogopite mica to determine kimberlite emplacement ages has a long tradition (e.g., Allsopp and Barrett, 1975; Smith et al., 1985, Smith et al., 1994; Brown et al., 1989; Creaser et al., 2004). Phlogopite in kimberlites occurs as a groundmass mineral (<100 μm) with occasional larger phenocrysts up to 200–300 μm in size, and as xenocrystic or ‘antecrystic’ macrocrysts (>1 mm; e.g., Mitchell, 1986; Reguir et al., 2009; Giuliani et al., 2016), which originate from the disaggregation of mantle lithologies during kimberlite magma ascent. An additional complexity is that these mantle-derived grains often exhibit magmatic overgrowths. Even though xenocrystic/antecrystic phlogopites predate the host kimberlite, their Rb-Sr systems are expected to be reset during residence in the host magma (Sr closure temperature ~400 °C; Jenkin et al., 2001; Willigers et al., 2004) and should then record the kimberlite eruption age (e.g., Allsopp and Barrett, 1975; Fitzpayne et al., 2020). This has practical advantages as phlogopite macrocrysts can be readily extracted and usually contain elevated levels of radiogenic strontium and higher Rb/Sr ratios than magmatic phlogopite phenocrysts (Allsopp et al., 1989). Potential complications from alteration (e.g., chlorite) or Sr-rich mineral impurities (e.g. carbonates) can be minimised by careful handpicking and/or mild acid leaching e (e.g., Brown et al., 1989; Creaser et al., 2004).

Different approaches can be used to derive a Rb-Sr age from kimberlitic mica. Data for several phlogopite fractions or individual grains may be regressed with or without data for the bulk kimberlite to produce an isochron (e.g., Brown et al., 1989; Creaser et al., 2004; Larionova et al., 2016). Alternatively, when limited material is available, a 2-point Rb-Sr model age may be obtained by anchoring Rb-Sr data for a single phlogopite fraction to an initial ⁸⁷Sr/⁸⁶Sr estimated from typical mantle ⁸⁷Sr/⁸⁶Sr values (e.g., Creaser et al., 2004; Heaman et al., 2004, Heaman et al., 2006) or from coexisting low-Rb/Sr perovskite (e.g., Sarkar et al., 2015), with whole-rock values used where perovskite is unavailable. A third approach utilises phlogopite with calcite impurities to determine a 2-point model age by combining Rb-Sr data for an acid-leached mica ‘residue’ with those for the respective acid leachate. A possible third point can be included using data for the unleached (or ‘bulk’) mica, which also provides a check for internal consistency (Maas, 2003; Yaxley et al., 2013; Dalton et al., 2019). This last approach was employed in this study.

One of the key advantages of the ⁴⁰Ar/³⁹Ar method is the remarkably high precision that can be achieved with modern analytical techniques (<0.1%; e.g., Phillips and Matchan, 2013; Matchan and Phillips, 2014; Phillips et al., 2017a; Schmieder et al., 2018), particularly relative to other methods such as U-Pb perovskite (1–5%; e.g., Wu et al., 2010, Wu et al., 2013). Despite the benefit offered by such precision, and the presence of K-bearing groundmass phlogopite in many (but not all) kimberlites, this technique has seldom been applied to kimberlites and related rocks. Early work (e.g., Kaneoka and Aoki, 1978; Fitch and Miller, 1983; Allsopp and Roddick, 1984; Phillips and Onstott, 1986) highlighted a number of issues with the technique, related to the presence of extraneous argon in mica macrocrysts and phenocrysts which yields anomalously old or maximum emplacement ages (see Phillips, 2012). Nonetheless, apparently reliable age results have been obtained on magmatic mica from kimberlites and related rocks in southern Africa (e.g., Fitch and Miller, 1983; Allsopp and Roddick, 1984; Phillips et al., 1999), as well as India, Canada, Finland and Russia (e.g., Kent et al., 1998; Chalapathi Rao et al., 1999; O'Brien et al., 2007; Tappe et al., 2014; Yudin et al., 2014). A disadvantage of this approach is the inherently fine grainsize of groundmass phlogopite, i.e. the textural type of mica preferred for this dating method, and its susceptibility to alteration, which provides challenges for mineral separation.

Comparison of new, precise ⁴⁰Ar/³⁹Ar ages with other independent age constraints is used here to rigorously assess the instances where ⁴⁰Ar/³⁹Ar dating produces older apparent ages.

Since the pioneering work of Kramers and Smith (1983), Smith et al. (1989) and Heaman (1989) there have been over 50 published studies utilising U-Pb systematics of perovskite to date kimberlitic and related magmatism. The first in-situ U-Pb dating of kimberlitic perovskite was published by Smith et al. (1994) with analyses carried out using SIMS (SHRIMP) on kimberlites from the Prieska area, southern Africa. Batumike et al. (2008) presented the first in-situ LA-ICP-MS U-Pb dating results for kimberlites from South Africa and the Congo, building on the earlier work on carbonatitic perovskite by Cox and Wilton (2006). These studies demonstrated the potential of in-situ U-Pb dating of perovskite to provide robust kimberlite emplacement ages, with numerous subsequent investigations adopting this approach (e.g., Yang et al., 2009; Reguir et al., 2010; Wu et al., 2010; Donnelly et al., 2012; Griffin et al., 2014; Castillo-Oliver et al., 2016). At the same time, perovskite U-Pb dating by isotope dilution remained the favoured method by other authors (e.g., Heaman and Kjarsgaard, 2000; Heaman et al., 2004; Kumar et al., 2007; Tappe et al., 2011; Sarkar et al., 2015).

One common issue associated with the U-Pb system is the presence of ‘common’ Pb; that is, the Pb incorporated into the system at the time of crystallization and unrelated to subsequent radioactive decay. The presence of small amounts of uncorrected common Pb in a datable mineral will increase its apparent U-Th-Pb age (Andersen, 2002). Initial Pb correction based on ²⁰⁴Pb is problematic when employing LA-ICP-MS because of the near-ubiquitous high ²⁰⁴Hg background in the mass spectrum (Košler and Sylvester, 2003).

The most widely used correction method for perovskite is to plot uncorrected U-Pb isotope data in Tera-Wasserburg Concordia diagrams (²⁰⁷Pb/²⁰⁶Pb versus ²³⁸U/²⁰⁶Pb; Tera and Wasserburg, 1972). A number of studies have used this approach and applied a ²⁰⁷Pb correction to each individual perovskite analysis based on the ²⁰⁷Pb/²⁰⁶Pb intercept as defined by the original, uncorrected, Tera-Wasserburg regression (e.g., Batumike et al., 2008; Griffin et al., 2014; Castillo-Oliver et al., 2016). However, Tappe and Simonetti (2012) and Ludwig (2012) have criticised this approach on statistical and technical grounds. Alternatively, the common Pb content can be estimated from Pb evolution models (e.g., Stacey and Kramers, 1975; Zartman and Doe, 1981) and used to anchor the ²⁰⁷Pb/²⁰⁶Pb intercept on the aforementioned Tera-Wasserburg plot. Arguably, rather than a ‘modelled’ or ‘estimated’ common Pb, the best proxy for common Pb contents is that derived from analysis of a low-U/Pb co-genetic phase that contains negligible in-grown radiogenic Pb (Corfu and Dahlgren, 2008; Chew et al., 2014; Stamm et al., 2018). We rigorously evaluate this latter approach in the current study through the use of in-situ and solution analyses of perovskite in conjunction with isotope dilution analysis of spinel.