CC-BY 4.0
1. Introduction
40Ar/39Ar geochronology is one of the most efficient tools for determining the timing of deformation events at different crustal levels, thus it is very useful for constraining the entire development of an orogen from the first stages of its building until its final erosion. Since the first application of this method to terrestrial samples [Fitch et al. 1969], the dating of deformation in crustal rocks has been the focus of many studies, even if until the end of the 80s, few papers report 40Ar/39Ar ages directly on this topic [e.g. Dallmeyer 1975; Dallmeyer et al. 1985; Allen and Stubbs 1982; Monié et al. 1984; Avé Lallemant et al. 1985; Kligfield et al. 1986; McCaig and Miller 1986; Keppie and Dallmeyer 1987; Costa and Maluski 1988; Hubbard and Harrison 1989]. However, it is well known that in mountain belts, a large amount of the strain imposed by plate tectonics is accommodated by multiscale deformation zones formed under variable pressure, temperature and fluid regimes. Giving absolute time constraints on deformation is fundamental for determining the propagation of deformation through space and time and unraveling the complex evolution of the different units of an orogen and their juxtaposition [Oriolo et al. 2018]. However, dating of deformation zones remains challenging since most of them record variable P–T-strain conditions, resulting in the coexistence of several mineral generations which can be pre-, syn- or postkinematic with respect to the last deformation [e.g. Lanari et al. 2014]. In addition, the interpretation of 40Ar/39Ar ages of K-bearing minerals (most often phyllosilicates, more rarely feldspars or amphiboles) is complicated by the fact that deformation introduces microstructural defects in these minerals, favoring their partial or complete fluid-induced recrystallization that have a great effect on the distribution and the retention of argon in their crystalline lattice at the nanometer scale [Naumenko-Dèzes et al. 2021; Villa 2022]. It should also be stressed that for these deformed rocks the interpretation of the 40Ar/39Ar data in terms of cooling age or (re)crystallization age remains a subject of debate as well as the mechanisms of age reset during deformation [Villa et al. 2014; Villa 2016]. Not to mention that there is a wealth of evidence in the literature that the isotopic behaviour of argon is very different for deep/middle crustal and near-surface deformation.
In large and hot orogens, the 40Ar/39Ar dated minerals of deformed rocks usually reached a state of complete textural, chemical and isotopic equilibrium, and for very slow cooling rate, measured ages record cooling at temperature below the closure temperature of the isotopic systems (in the range 300–550 °C for mica and amphibole), which can occur hundreds of myrs after their crystallization due to slow heat dissipation. In this type of environment with important time-lag between the crystallization of K-bearing minerals and their isotopic closure for argon (e.g. old cratons), the 40Ar/39Ar method is considered essentially as a method of thermochronology [it dates when a mineral passes through a specific closure temperature, Dodson 1973; McDougall and Harrison 1999] that can be combined with other methods (e.g. U–Pb, Sm–Nd, Rb/Sr) to unravel the complex evolution of rocks during their progressive cooling/exhumation. In contrast, for colder (500–300 °C), and/or more localized deformations, the 40Ar/39Ar age interpretation in deformed zones is not so straightforward, in particular in the case of polydeformed rocks. Different issues have to be considered then: the partial or total resetting of inherited minerals through a combination of (i) thermally activated volume diffusion, deformation and fluid-assisted dissolution–recrystallization processes, (ii) the crystallization, cooling and closure of newly-formed, synkinematic and/or postkinematic minerals and (iii) eventually the presence of excess argon held by fluids. Considering the fulfilled assumptions of a complete reset of the inherited minerals and the absence of excess argon, 40Ar/39Ar ages are most often regarded as (re)-crystallization ages and the method is considered as a method of geochronology [it dates mineral (re)crystallization, not cooling; e.g. Mulch and Cosca 2004; Villa et al. 2014; Di Vincenzo et al. 2016; Villa 2016; Beaudoin et al. 2020 among many others]. Ideally, the method must be combined with other dating methods, such as the Rb–Sr internal isochron method [e.g. Faure and Mensing 2004; Rösel and Zack 2021] or the U–Pb method on monazite or apatite [e.g. Parrish 1990; Harrison et al. 2002; Chew et al. 2011] which can add additional constraints on the metamorphic and tectonic evolution of rocks deformed at intermediate temperatures. In all these deformed rocks, the 40Ar/39Ar dating of K-bearing phases is now conducted using mainly lasers with different approaches: step-heating of small mineral populations or single grains, fusion of individual grains, in situ ablation within a single grain or at the scale of a small rock section, with a spatial resolution of several tens of microns. As discussed below, this spatial resolution is still insufficient to get a very precise distribution of argon atoms in minerals from sheared zones. However, the combination of these approaches, which has benefited from an improvement in the analytical performance of mass spectrometers over the last ten years, can help to better understand the behavior of argon in deformed rocks and the meaning of 40Ar/39Ar ages.
For more superficial deformation conditions, many 40Ar/39Ar studies have successfully focused on fault gouges using authigenic phyllosilicates that form during the activity of these faults in a temperature range corresponding to the muscovite–illite transition (i.e. ductile–brittle transition, ∼300 °C) to the illite–smectite transition [∼140–90 °C, Freed and Peacor 1989]. Authigenic phyllosilicates in fault gouges are characterized by their small size (<2 μm), thus corresponding to clay minerals, and usually coexist with inherited K-bearing phases from the host rocks. The dating of these superficial deformations is often rather tricky for several reasons: (1) the size of the neoformed clays requires efficient mineral separation techniques; (2) the proportions of authigenic and inherited phases in the separated grain-size fractions have to be quantified; (3) due to their small size, clay minerals are subject to recoil of 39Ar during irradiation, resulting in an ageing of the samples. It is shown in the second part of this contribution that these limitations can be overcome and that the situation can sometimes be more complex due to the polyphased activity of faults and various fluid interactions.
Therefore, the present contribution is organized into two main parts addressing specific aspects of 40Ar/39Ar dating of ductile and brittle deformation in deep and superficial crustal domains through various examples illustrating the complexity of age interpretation in deformed rocks. These examples highlight in particular the benefit of coupling in situ laser 40Ar/39Ar analyses with detailed microstructural and petrological observations for the study of ductile deformations and the reconstruction of P–T–d–t (pressure–temperature–deformation–time) paths. The second section is focused on brittle deformation and fault gouges and shows the advantages of coupling geochronology and thermochronology in orogenic belts.
2. 40Ar/39Ar geochronology of ductile deformation
2.1. Some methodological aspects
It is beyond the scope of this contribution to present the technical details of the 40Ar/39Ar method that can be found in the reference book of McDougall and Harrison [1999] and more recently in Jourdan et al. [2014] or Schaen et al. [2021]. In a simplified way, the method is derived from the 40K–40Ar method, 39Ar being a proxy of the quantity of 40K after fast neutron irradiation in a nuclear reactor and the age directly dependent on the ratio between the two argon isotopes and the fluence of neutrons calculated from argon isotopes released by co-irradiated flux monitors. Technical aspects of the 40Ar/39Ar method, which concern the age of the monitors, decay constants, argon recoil effects, or the current 40Ar/36Ar atmospheric ratio, are still under debate but will not be addressed here [Schaen et al. 2021]. Soon after the inception of the 40Ar/39Ar method [Merrihue and Turner 1966], laser microprobe 40Ar/39Ar analyses were first performed on lunar breccias [Megrue 1973] and some years later on terrestrial samples [York et al. 1981; Maluski and Schaeffer 1981] using continuous or pulsed lasers for a sampling of small material amount (<milligram). From there on, many laboratories worldwide developed this technique using different lasers that can be coupled now with very sensitive mass spectrometers to produce very precise and reproducible 40Ar/39Ar ages (from 3.5 Ga to some ka) on K-bearing terrestrial samples. The method is under constant improvement and addresses a wide range of geological processes such as metamorphism, magmatism, hydrothermalism, diagenesis, weathering, deformation …. Its range of applications is then very large from geodynamics, paleomagnetism, mineral and geothermal resources, seismic risk but also paleontological or climate evolution.
As indicated in the introduction and schematized on Figure 1, different methods of argon extraction can be achieved with the different types of laser. This includes first the step-heating degassing of single grains or micro-populations of K-bearing samples by increasing progressively the power of a continuous infrared CO2 laser beam applied to the grains. An age is calculated at each argon release increment and the whole step data is portrayed in an age spectrum and inverse isochron plot. In some instances, Ca/K and Cl/K vs. age correlation plots can help to better differentiate the various argon reservoirs that can coexist in samples with a complex history of metamorphic-deformation processes and fluid interactions [e.g. Mulch et al. 2005; Simon-Labric et al. 2009; Allaz et al. 2011].
From left to right, (1) different modes of laser microprobe step-heating of a single grain or micro-population, (2) age and K/Ca spectra and (3) Cl/K spectrum and inverse isochron plot for a homogeneous amphibole micro-population [Fernandez et al. 2020].
However, it is well-known that hydrous phases such as micas and amphiboles do not release argon by volume diffusion when heated under vacuum because of their metastability and that age spectra do not reflect the true argon distribution in the dated minerals [e.g. Sletten and Onstott 1998; Lo et al. 2000; Villa 2021]. This is well supported by the different studies comparing step-heating and in situ laser probe experiments, for instance in the case of slow cooling terranes, excess argon contamination, thermal or tectonic overprint [e.g. Phillips 1991; Hodges et al. 1994; Monié et al. 1994; Kellet et al. 2016]. Caution should therefore be exercised when interpreting step heating results and, in case of doubt, a comparison with other modes of argon extraction or other isotope chronometers may be useful.
Continuous laser beams can also be used for the total fusion of individual grains [e.g. Uunk et al. 2018; Gemignani et al. 2018; Porkoláb et al. 2022]. Ages are usually reported in probability or Kernel density plots [Ludwig 2012; Vermeesch 2018; Schaen et al. 2021] in order to identify the different age populations in a rock sample (Figure 2), a method that requires a large number of analyses to get statistically meaningful peak ages. This approach, for which the main criterion of grain selection is their large size, can be used to better interpret discordant age spectra from complex polymetamorphic tectonites.
Distribution of total fusion ages for (a) a homogeneous population of biotite with few grains showing inherited or excess argon and (b) a strongly heterogeneous population of muscovite showing decreasing amount of argon resetting from left to right [Cap de Creus samples, Monié et al. 2018].
However, in deformation zones where phyllosilicates are the main target of 40Ar/39Ar analyses, micas of different origin, size and chemistry can occupy different microstructures and the distinction between not reset, partially reset, totally reset and newly crystallized populations is not always straightforward, without discarding the problem of excess argon contamination. The 40Ar/39Ar laser fusion dating of single grains does not always resolve these complex isotopic situations since, in general, few attention is paid to the grain-size effect on the age distribution or to the chemical heterogeneities within and between micas. By contrast, the in situ40Ar/39Ar method has the advantage to preserve the textural and chemical relationships between the different mineral generations in polydeformed rocks and to establish correlations between ages, mineral fabrics and petrology at the scale of several tens of microns, corresponding to a volume large enough to measure accurately the different argon isotopes on the mass spectrometer (typically a square of 150 × 150 μm and 20 to 50 μm deep). Different pulsed lasers have been used since the first dating of terrestrial samples 40 years ago, including infrared ruby lasers, argon-ion visible lasers, ultra-violet Nd–YAG lasers and more recently UV excimer lasers that provide the best spatial resolution for in situ analyses. These analyses that can be realized at the scale of a single grain if it is large enough to perform several ablations inside (Figure 3), have the potential to resolve core to rim zonations or other kinds of age distribution due to slow cooling or partial resetting by thermally activated diffusion or chemical exchanges and recrystallization [e.g. Phillips and Onstott 1988; Scaillet et al. 1990; Hames and Hodges 1993; Hames and Bowring 1994; Monié et al. 1994, for some of the oldest references].
In situ ages obtained inside single mica grains using an argon-ion visible laser (a,b) and a quintupled UV laser (c). (a) Homogeneous Triassic ages inside a high-pressure phengite from Turkey [Okay and Monié 1997]; (b) core-rim age zonation in a thermally overprinted Variscan magmatic biotite from the Eastern Pyrénées [Monié et al. 1994]; (c) irregular age zonation in a partially chloritized Variscan biotite from the Cévennes schists [Montmartin 2021].
However, for most in situ studies of deformed rocks, dating experiments are not realized at the scale of a single mineral, but at the scale of a portion of thin section (Figure 4) in order to deconvolute the complete history of deformation and metamorphism.
Different modes of in situ argon extraction with a quadrupled UV laser. (a) Typical surface of gneiss thick section after lasing. The micas aligned along the two schistosities S1 and S2 have been dated. (b) Detailed view of the surface of a mica after a fast laser rastering. The depth of the pits is less than 10 μm. (c) Laser drilling on a muscovite to a depth of about 50–70 μm. Note the white ejecta around the craters.
In an orogen, ductile shear zones concentrate a large part of deformation and, because of their efficient reactivity, phyllosilicates are ideal candidates for recording the short- and long-term evolution of faults through their chemistry (major elements, trace elements and fluid composition), grain size and isotopic ages. Thus, microscope observation of thin sections usually is a first step to identify the different generations of pre-, syn- and postkinematic minerals suitable for dating (Figure 5).
Examples of textural relationships in metamorphic tectonites, from left to right, (a) micaschist with prekinematic large muscovite (Musc1) in a fine-grained matrix of quartz–feldspar–muscovite (Musc2) [Cap de Creus massif, Monié et al. 2018]; (b) partially reworked garnet–sillimanite–quartz–biotite (Bio1) gneiss in a matrix of quartz and fine-grained biotite (Bio2). Note the different habits of prekinematic biotite as clasts in the new foliation, in a garnet pressure-shadow and in a crack inside garnet [Agly massif, Aumar et al. 2022]; (c) albite–quartz–epidote–phengite (Ph1) micaschist crosscut by a chlorite–phengite (Ph2) shear band [Andros island, Huet et al. 2015].
The examples of Figure 5 illustrate the importance of identifying the different textural habits of the minerals that can be dated by in situ 40Ar/39Ar laser analyses. Two samples from Cap de Creus and Agly massifs (Figure 5a,b) highlight the evolution of grain size in tectonites, with prekinematic micas being usually larger than synkinematic matrix micas. In most tectonites, the fine-grained matrix consists of two mica types: a first type resulting from a strong mechanical reduction in size of part of the inherited micas (comminution), and a second type that is newly formed by a fluid-enhanced dissolution–precipitation mechanism. Therefore, the matrix micas can only give homogeneous ages if those of the first type have undergone a complete isotopic resetting of their K–Ar chronometer during comminution, which is closely dependent on the crustal deformation conditions (temperature, pressure, duration, fluid behaviour, open/closed system behaviour of the host rock …). For the third sample from Andros island (Figure 5c), a first generation of synkinematic phengite Ph1 is aligned along the main foliation while synkinematic phengite Ph2 is also present in association with chlorite along a late shear band. However, it is likely that together with the neocrystallization of phengite Ph2, some early phengite Ph1 was caught by the shear band. Therefore, in situ 40Ar/39Ar mica dating in this kind of microstructure can constrain the last deformation age only if micas inherited from the main foliation have been fully reopened (or recrystallized) along the shear band, which is also dependent on the deformation conditions. Insofar as the ages of the different generations of synkinematic micas can be interpreted as reflecting crystallization ages (see below), they can provide an indication of how long a shear zone remains active [e.g. Schneider et al. 2012].
After this first step and before dating, it is fundamental to consolidate the microstructural observations by a petrological study of the mineral assemblages to constrain the P–T conditions of the different deformation episodes. Detailed EPMA analyses coupled with SEM imaging or EDS mapping are the main tools used to decipher the chemical composition of datable minerals, and to reveal at the scale of a thin section inter- and intra-grain variations related to more or less complex P–T–D paths recorded by the tectonites during the orogenic cycle. Figure 6 shows some examples of chemical data that can be served as a guide to perform texturally and chemically-controlled in situ 40Ar/39Ar analyses in deformed crustal rocks.
Examples of mica chemical variations that can be observed in metamorphic rocks: (a) EPMA analyses of phengites in a retrogressed high-pressure (HP) micaschist fron Iran with two populations that can correspond to the situation in Figure 6b [core-rim zoning, Rossetti et al. 2017] or that in Figure 5c (HP foliation crosscut by lower pressure shear band); (b) SEM imaging of a zoned phengite with a celadonite-rich core and a poorer rim; (c) EPMA analyses of TiO2 content of muscovite in an orthogneiss from the South Armorican Massif. Note the high Ti content in the mica core compared to the rim and to the newly crystallized muscovite along the foliation; (d) X-ray mapping of Ti content in muscovites from Figure 6c [Turrillot et al. 2011].
Therefore, before in situ 40Ar/39Ar dating, a detailed petro-structural study of the samples that are the most representative of the study area is highly recommended and depending on the goals pursued, the study should focus on texturally and chemically homogeneous samples dominated by a single K-bearing mineral population (early or late) or on more complex samples that record several phases of deformation–(re)crystallization (Figure 6). As shown on Figure 4, the in situ 40Ar/39Ar method has the ability to provide ages with a spatial resolution of about 100–150 μm2 that is enough to get first-order data on the homogeneity or heterogeneity of dates at the scale of a thick section. It is likely that this spatial resolution will be improved in the future with the advent of increasingly sensitive mass spectrometers. However, many studies have shown that reliable textural–chemical–age relationships can be already established with current 40Ar/39Ar spatial resolution [e.g. Maluski and Monié 1988; Kramar et al. 2001; Agard et al. 2002; Mulch and Cosca 2004; Mulch et al. 2002; Augier et al. 2005; Mulch et al. 2005; Di Vincenzo et al. 2007; Schneider et al. 2012; Angiboust et al. 2015; Bellanger et al. 2015; Di Vincenzo et al. 2016; Kellet et al. 2016; Scharf et al. 2016; Berger et al. 2017; Laurent et al. 2017; Angiboust et al. 2018; Tan et al. 2019; Beaudoin et al. 2020; Ghignone et al. 2021; Di Vincenzo et al. 2022].
2.2. Age interpretation
From the earliest applications of the method to the dating of magmatic and metamorphic rocks, a notion of isotopic system closure temperature emerged and was formulated by Dodson [1973] on the basis of volume diffusion laws. During cooling, a K-bearing mineral will start to accumulate radiogenic argon at a certain temperature T1 and will become a closed system at lower temperature T2, the temperature interval T1–T2 defining what is called the partial retention zone. The closure temperature Tc of this mineral lies within this partial retention zone whose range is dependent on the cooling rate. For fast cooling rates (e.g. volcanic rocks), T1, T2 and Tc are very close and the 40Ar/39Ar ages can be interpreted to date crystallization. This is the case of the sanidine standards from volcanic rocks used to monitor the fast neutron flux during sample irradiation whose 40Ar/39Ar ages agree with U–Pb zircon ages [e.g. Fleck et al. 2019]. It is also common to observe this concordance of U–Pb and 40Ar/39Ar ages for granites emplaced in the upper crust, with micas revealing no core to rim age zonation due to fast cooling [e.g. Brichau et al. 2007]. Note that in deformed rocks from the mid-upper crust, synkinematic minerals formed at temperatures close to 500–550 °C can also provide similar monazite U–Pb and 40Ar/39Ar mica ages interpreted as crystallization ages [e.g. Aumar et al. 2022]. For very slow cooling low- to mid-crustal deformed rocks (e.g. large Proterozoic cratons), the partial retention zone is usually very large and the time elapsed between the crystallization of K-bearing minerals and their complete closure for argon diffusion can reach several hundred million years. This results in core to rim age zonations reflecting a progressive closure of the K–Ar chronometer [e.g. Hodges et al. 1994, for muscovite] that can be modeled using thermal volume diffusion laws [e.g. Warren et al. 2012; Skipton et al. 2018]. Indeed, by contrast with the fast cooling setting, the core to rim age range represents a segment on the cooling path which length is dependent on grain size, cooling rate and residence time of the crustal rocks at given temperature and pressure, diffusion parameters of the mineral species under consideration and the open behavior of the host rock. Available K–Ar systems permit to cover a continuous range of closure temperatures from about 550 °C (amphibole) to 150 °C (K-feldspar) that have been determined empirically, theoretically and experimentally. To reconstruct the T–t cooling trajectory above and below this temperature range, complementary geochronological and thermochronological studies using U–Pb, Sm–Nd, Rb–Sr, fission tracks and U–Th–Sm/He methods are required. Therefore, in very slowly cooled high-temperature rocks with a single episode of deformation at peak temperature, the ages provided by the 40Ar/39Ar method can only be considered as minimum estimates of the deformation age. When different deformation episodes occur during cooling, thermo-barometric analysis of synkinematic mineral assemblages allows in principle to assign the different deformation stages along the P–T path. If these deformations occur at temperatures close to the isotopic closure temperatures of argon chronometers, in situ 40Ar/39Ar dating of the synkinematic K-bearing minerals can constrain the timing and duration of tectonic events and the rate of deformation propagation or localization, provided that the dated minerals do not contain inherited or excess argon which can be resolved by in situ intra- and inter-grain analyses. However, at present, large uncertainties remain on the closure temperature of the mineral species used for 40Ar/39Ar dating, due to the lack of precision of the experimentally determined diffusion parameters, the validity of diffusion experiments themselves as well as the influence of the multiple parameters that can affect argon retentivity in the minerals during cooling (chemical composition, pressure, lithology, open/closed behavior of host rock, fluid composition and behaviour …). This is true in particular for the white mica group for which a large range of closure temperature is suggested by different studies, from 350 to 550 °C or more [e.g. Hames and Bowring 1994; Harrison et al. 2009; Nteme et al. 2022], a range which remains an area for future exploration. As a consequence, the reconstruction of P–T–D–t path in polydeformed high-temperature rocks using 40Ar/39Ar data can suffer from this indetermination of the retention properties of the dated minerals and for this purpose the use of multiple geochronometers is highly recommended. It is important also to mention that the mechanisms involved during the closure of an isotopic system may be different from those leading to its resetting. In particular, various 40Ar/39Ar laser probe studies of polydeformed high-temperature rocks showed that a mineral can survive temperatures well in excess of its closure temperature without loss of radiogenic argon [e.g. Kelley et al. 1997; Di Vincenzo et al. 2001; Maurel et al. 2003; Schneider et al. 2008]. This is interpreted to reflect a closed system behaviour of the host rock with no or limited fluid-rock interaction due to intrinsic rock properties (mineralogy, porosity, deformation, interstitial fluid amount) and conditions of overprinting (temperature, pressure, duration, fluid behavior). This is to say that in a polyphased context, the interpretation of ages strictly in terms of crossing a given temperature barrier in one direction or the other can be debated without a proper understanding of the variable parameters that control the closure and opening of the chronometers being used. This is a highlight made several years ago by Chopin and Maluski [1980] in their studies of some high-pressure rocks of the Western Alps and subsequently validated by many geochronological works [see Agard et al. 2002; Ghignone et al. 2021 for in situ 40Ar/39Ar studies in the same area, or McDonald et al. 2016 for the HP gneisses of Norway].
For deformations that occur at lower temperature, from 500 °C to the ductile–brittle transition at 300 °C, it is frequent that different generations of K-bearing minerals (re-)crystallized along the P–T–D–t path which renders the interpretation of 40Ar/39Ar ages somewhat difficult due to the interaction of several effects of grain deformation during shearing, chemical exchanges enhanced by fluid circulation, dissolution–precipitation reactions that can result in isotopic heterogeneities at a very fine scale. However, as mentioned above, there is now compelling evidence that texturally and chemically-controlled in situ40Ar/39Ar analyses of such rocks can provide significant data to decipher the succession of deformation episodes in tectonites. This is also due to the fact that for rocks deformed below 500 °C, it is now widely accepted that 40Ar/39Ar ages of synkinematic minerals represent (re)-crystallization ages rather than cooling ages [e.g. Agard et al. 2002; Augier et al. 2005; Rolland et al. 2008; Sanchez et al. 2011; Di Vincenzo et al. 2016; Villa et al. 2014; Bosse and Villa 2019]. However, the interpretation of in situ40Ar/39Ar analyses in polydeformed rocks can be faced with different problems: the isotopic inheritance in relict phases that are not fully reset during shearing, the chemical complexity of newly formed minerals (zonation, mica interlayering) and their fine grain size, the presence of short circuit pathways for argon migration, the presence of excess argon trapped at the time of mineral crystallization or later, the late interaction with fluids. Below, we present some examples that highlight the relevance of in situ40Ar/39Ar analyses compared to more conventional argon degassing modes for dating mid-upper crustal deformation in orogens.
2.3. Examples of in situ 40Ar/ 39Ar dating of shear zones
2.3.1. In situ 40Ar/ 39Ar dating of amphibole from shear zones
In crustal deformation zones, micas are the main target of in situ 40Ar/39Ar investigations because of their ability to accommodate strain constraints by different deformation mechanisms (kinking, gliding parallel to cleavage) and different (re-)crystallization processes in the presence of fluids. On rare occasions, shear zone amphiboles were dated by this technique and an example from the upper Neoproterozoic shear zone of In Ouzzal in southern Algeria is given below [from Ferkous and Monié 2002]. In south Algeria, the Hoggar shield belongs to a Neoproterozoic (800–600 Ma) orogenic belt that was juxtaposed to the Archean West African craton by large-scale N–S striking subvertical strike-slip shear zones under upper greenschist–amphibolite metamorphic conditions. The studied shear zone is a 3 km wide corridor made of protomylonites, mylonites and ultramylonites and contains abundant pre- to synkinematic igneous rocks, as well as syn- to postkinematic auriferous quartz veins. Amphibole from a prekinematic granophyre is of two types, brown magmatic porphyroclastic pargasite and blue-green ferrotschermakite in the pressure shadow of pargasite. In situ 40Ar/39Ar analyses (Figure 7a) provide dates between 749 and 849 Ma for the magmatic amphibole and between 626 and 650 Ma for the metamorphic one, consistent with the step-heating analyses of amphibole (623 Ma) and biotite (613 Ma) from a synkinematic granodiorite in the shear zone (Figure 7b). The estimated P–T conditions of shearing are about 0.4 GPa and 450–500 °C, suggesting that the dates of synkinematic ferrotschermakite can be interpreted as crystallization ages close to the peak of metamorphism related to the main episode of fault activity along the In Ouzzal shear zone, in agremeent with the muscovite age of synkinematic gold-rich quartz veins (611 Ma). In this sample, the primary magmatic amphibole is partially reset and the date of 849 Ma can be taken as a minimum estimate for the age of granophyre emplacement and cooling. This large deformation corridor was locally reactivated about 100 Ma later by a second episode of strike-slip tectonics as attested by in situ40Ar/39Ar dating of fine grained biotite in a narrow ultramylonitic band (Figure 7c).
In situ and step-heating 40Ar/39Ar data from the In Ouzzal shear zone. (a) In situ dates on prekinematic and synkinematic amphibole from a sheared granophyre; (b) age spectra of amphibole and biotite from a syntectonic gabbro; (c) in situ date distribution (Kernel density plot) on fine biotites from a narrow mylonitic band [Ferkous and Monié 2002].
This study highlights (1) the interest of the laser microprobe approach to differentiate the ages of synkinematic amphibole from prekinematic amphibole, which is not always possible with the dating of mineral populations; (2) the common characteristic of faults to be weakness zones that can have a long term activity, and which can result in the coexistence of multiple generations of newly formed minerals during diachronous episodes of shearing.
2.3.2. In situ 40Ar/ 39Ar dating of white mica from shear zones
Three examples are presented in this section, the Oisans external massif (western Alps), and two massifs of the eastern Pyrenees, the Agly massif (North Pyrenean Zone) and the Cap de Creus massif (Axial Zone). All these massifs are made of a Variscan basement overprinted to various extent by Alpine deformations under greenschist to upper greenschist metamorphic conditions, resulting in a mixture of Variscan and Alpine mineral assemblages dominated by Variscan minerals in the weakly deformed rocks and sometimes vice-versa. For these massifs, there were few geochronological studies of the deformation zones themselves and debates persisted about the relative importance of pre-Alpine or Alpine deformation on the present structuration of these two belts. This is particularly true for the two Eastern Pyrenees massifs for which in the absence of systematic geochronological studies focused on shear zones, controversies still remain since several years on the age of their activity, varying by authors from the late Paleozoic to the Eocene.
In the Oisans massif, in situ40Ar/39Ar studies were carried on different W-directed reverse shear zones that accommodate E–W shortening during the Alpine collision. The main goal of this study was to document the westward propagation of the deformation on variously strained samples and to evaluate the effect of mixture between partially or totally recrystallized Variscan muscovite and newly-formed synkinematic Alpine phengite in protomylonites, mylonites and phyllonites formed at temperature below 350 °C [Bellanger et al. 2015]. EPMA analyses of white-micas yield two distinct end-members with very few intermediate values. The first end-member corresponds to prekinematic muscovite clasts (Si = 3.15; Na = 0.06–0.08) and the second one to synkinematic phengite (Si = 3.2–3.4; Na = 0–0.02). In situ40Ar/39Ar analyses were obtained on three samples selected over a single strain gradient of a major shear zone. The results are synthesized in Figure 8 which shows for each protomylonite, mylonite and phyllonite sample, the textural aspect of the dated rock section, the distribution of in situ dates on this section and the Kernel density plot for these 40Ar/39Ar ages.
In situ 40Ar/39Ar analyses on three samples of Variscan gneiss showing, from left to right, an increase of Alpine deformation over a major shear zone of the Oisans massif. The red line corresponds to the maximum age of Alpine deformations imposed by stratigraphic constraints [modified from Bellanger et al. 2015].
In the less deformed sample containing discrete shear bands (protomylonite), in situ ages are scattered between 207 ± 5 Ma and 24 ± 2 Ma (Triassic to Miocene). Of the 29 dates, only 6 are younger than 34 Ma, the maximum age of Alpine deformations imposed by stratigraphic constraints. Interestingly, these dates come from small and low Cl/K (<0.01) phengites from the core of the shear bands. 12 analyses provide the oldest dates (>50 Ma) outside the shear bands and are correlated with relatively high Cl/K ratio (0.01–0.02). They correspond to partially recrystallized inherited muscovites with no peak age in the cumulative correlation plot, which indicates that Variscan muscovite experienced variable argon loss during Alpine thermotectonic overprint and interaction with Alpine fluids. The data do not enable to discount the possibility that an initial argon resetting may have occurred as early as the Permian or Triassic. The mylonite sample partially preserved its Variscan fabrics (high-temperature foliation) and is cut by a 3 mm-thick shear band where Alpine synkinematic recrystallizations are almost complete. In the shear band, phengite yields a range of in situ dates between 30 and 40 Ma with a peak age at about 32 Ma for the micas from the core of the shear band (12 of 30 analyses) while those from its inner edge are older than 34 Ma (red line on Figure 8). Outside the shear band, muscovite clasts yield older dates (48–129 Ma) with higher Cl/K ratios than phengite, as observed in the protomylonite. These relationships, only detectable by in situ analyses, indicate that synkinematic crystallization of phengite by dissolution–precipitation processes at the expense of K-bearing Variscan minerals (muscovite and feldspar) was more complete in the core of the shear band than at its edge and even more than in the host rock. In the phyllonite almost devoid of muscovite clasts, of the 29 dates of phengite (Cl/K < 0.01), only 3 are older than 34 Ma while the remaining dates define two peaks, a major peak at 26 Ma and a minor one at 32 Ma consistent with the peak date given by phengite in the mylonite. These in situ data are consistent with those provided by a previous step-heating study focused on ultramylonites that also highlighted some contamination by excess argon related to the presence of Ca-rich impurities in the dated mineral populations and limited argon loss possibly due to hydrothermal alteration [Simon-Labric et al. 2009]. In our example, the correlation of in situ 40Ar/39Ar ages and deformation intensity within and between the three dated sections shows that the distribution of Alpine ages is primarily controlled by deformation textures. The data also show that the presence of Alpine ages in the three dated rock sections is correlated with the intensity of phengite blastesis, particularly important in high-strain zones along which fluids circulated. Considering that temperatures reached during the Alpine orogeny in the Oisans massif did not exceed 300–350 °C, it is very likely that the ages retained by newly-formed phengites represent true crystallization ages, provided that they evolved in a closed system, without being contaminated by excess (extraneous or inherited) argon. Phengite dates are distributed on two Oligocene peak ages that are interpreted to record the progressive localisation of deformation along the highly strained tectonites in the Oisans massif during E–W shortening [see Bellanger et al. 2015, for a more complete discussion].
In the eastern Pyrenees, the mylonitic shear zones occurring in the Variscan basement at Cap de Creus have been the subject of detailed structural studies [e.g. Carreras et al. 1977, 2010; Druguet et al. 1997, 2009] and are generally attributed to the end of the Variscan orogeny. This assumption was recently challenged by a 40Ar/39Ar geochronological study that argues for a Jurassic age for these relatively low-temperature (∼450 °C) shear zones, with possible Tertiary reactivation [Vissers et al. 2016, 2020]. However, the significance of these ages remains largely uncertain due to the large grain size of the dated muscovite (180–250 μm) compared to the size of synkinematic micas (<100 μm). Some new in situ and step-heating 40Ar/39Ar geochronological data are presented below that allow us to propose an Eocene age for the most obvious mylonitic deformation in the Cap de Creus massif [Monié et al. 2018].
The first sample is an ultramylonitic Variscan pegmatite [299 Ma U–Pb zircon age, van Lichtervelde et al. 2017] that was collected in the core of a 2 m wide NW–SE dextral shear zone in the Cap de Creus peninsula, close to the locality sampled by Vissers et al. [2016]. The mica population contains some rare muscovite clasts but is dominated by fine-grained synkinematic phengitic muscovite (Si = 3.1–3.2) formed at the expense of Variscan muscovite by fluid-assisted dissolution–precipitation mechanisms (Figure 9a).
Ultramylonitic pegmatite from the Cap de Creus peninsula; (a) thin section with a scarce sheared and recrystallized clast of muscovite in a fine matrix of quartz, albite and phengitic muscovite; (b) age spectra on two granulometric fractions representative of the two mica populations; (c) distribution of situ 40Ar/39Ar dates obtained on a thick section [Monié et al. 2018].
This population was studied using stepwise and in situ heating experiments on two size fractions and on a thick section made perpendicular to the foliation and parallel to the lineation, respectively. The age spectrum of a single muscovite clast (300–400 μm) shows a saddle-shaped pattern with dates of 101–133 Ma while about 30 grains of synkinematic phengitic muscovite (80–100 μm) give a plateau-like date of 42 ± 1 Ma (Figure 9b). In situ analyses of micas along the foliation plane yield dates ranging from 43 Ma to 83 Ma with a peak at 45 Ma in the cumulative probability diagram (Figure 9c). This peak age validates the ∼42 Ma plateau-like date of the fine-grained mica fraction as no value below 42 Ma was obtained. Older in situ dates confirm that small relics of Variscan muscovite, still containing some inherited argon, are present along the foliation, but are not optically distinguishable from synkinematic micas. All these new dates are much younger than the Jurassic ages (159–175 Ma) reported by Vissers et al. [2016] and point to an Eocene age for the main ductile activity of shear zones in the Cap de Creus peninsula. They also show how critical the choice of samples and minerals to date is in mylonites (strain intensity, grain size, mineral habit …) and how useful the in situ approach is to identify the different mica generations compared to the other methods of argon extraction. This is particularly true for the total melting approach which is biased by the selection of the largest mica grains in the sheared rocks, which is not the case for the fabric-preserving in situ approach that allows to target micas that differ in size, chemistry or textural habit.
This is confirmed by the study of mylonitic shear zones in the Roses granodiorite, 15 km southwest of the previous shear zone. This Variscan magmatic body [291 Ma U–Pb zircon age, Druguet et al. 2014] is affected by zones of heterogeneous and localised mylonitic deformation resulting from the coexistence of brittle and ductile deformation mechanisms and fluid-induced mineralogical softening reactions. In the most deformed facies, these reactions induce a complete replacement of plagioclase by epidote ± phengite, and magmatic biotite by more magnesian metamorphic biotite and white mica (Figure 10a) at a temperature of 450–500 °C.
Sheared granodiorite from the Roses massif; (a) thin section showing the coexistence of prekinematic magmatic and synkinematic metamorphic biotite; (b) distribution of situ 40Ar/39Ar dates across a gradient of deformation showing some biotite relics; (c) distribution of situ 40Ar/39Ar dates within two sections of mylonitic granodiorite (red square) and shear band (blue square) [Monié et al. 2018].
In situ 40Ar/39Ar dates were obtained on two protomylonitic and mylonitic samples, both showing an increasing gradient of deformation and metamorphic recrystallization. In the protomylonite, in situ biotite dates range from 69 Ma to 47 Ma from the least to the most deformed domains, with a peak age of about 49 Ma for the more sheared zone (Figure 10b). In the mylonite, biotite from the least deformed portion yields a range of values between 41.5 Ma and 47.5 Ma with a peak date of 44 Ma for newly crystallized biotite. In the shear band with abundant synkinematic phengite, in situ 40Ar/39Ar dates are mainly between 39 Ma and 42 Ma with a peak at ∼40 Ma (Figure 10c). These texturally and chemically controlled in situ 40Ar/39Ar analyses indicate that deformation and associated metamorphic recrystallizations play a prominent role in the distribution of dates at the scale of a centimeter section, with younger dates in the most deformed domains. Furthermore, the preservation of these deformation–recrystallization–chemistry–age relationships again suggests that the 40Ar/39Ar dates recorded by biotite and phengite in these mylonites should be considered as crystallization ages rather than cooling ages. In the mylonite, the 4 Myr time lag between the crystallization of metamorphic biotite in the host rock matrix and the crystallization of phengite in the shear band may represent the time required to switch from a diffuse to a localized deformation regime. These data suggest an Eocene age for the mylonitisation of the Roses granodiorite, in agreement with the data reported above in the Cap de Creus peninsula. On a regional scale, it is proposed that these NW–SE striking mylonitic strike-slip structures accommodated the end of the Pyrenean compression [e.g. Teixell et al. 2018; Ternois et al. 2019].
Recently, new Th–U/Pb and 40Ar/39Ar geochronological data were obtained in the Agly massif [Aumar et al. 2022], which has long been the subject of debate as to the influence of Cretaceous deformations on its architecture. The study focused on the Lower Gneiss Unit that shows a highly ductile extensional deformation with a NNE–SSW oriented stretching lineation that is also found in the Mesozoic sedimentary series to the north. In the basement gneisses, the original high-temperature Variscan deformation was overprinted by this new extensional deformation that evolved continuously from a ductile regime at relatively high temperature (⩾500 °C) to a brittle regime at a very low temperature (150/200 °C). Synkinematic minerals located within shear bands provided consistent Th–U/Pb monazite and 40Ar/39Ar mica dates with a 94–127 Ma age bracket interpreted as dating mylonitic deformation during the Cretaceous rifting episode that preceded the building of the chain. An example of stepwise heating and in situ40Ar/39Ar data is shown in Figure 11.
(a) Mylonitic leucogranite [291 Ma U–Pb monazite age, Vanardois et al. 2022] with a magmatic muscovite clast in a matrix of synkinematic fine-grained biotite and muscovite; (b) age spectra of muscovite and biotite clasts from a protomylonitic gneiss; (c) distribution of situ 40Ar/39Ar dates obtained on four thick rock sections of mylonites. A Kernel density bell-shaped curve has been drawn for each mica population [Aumar et al. 2022].
Several samples of proto- to ultramylonites (Figure 11a) were collected throughout the massif and regardless of the degree of mylonitisation, muscovite and biotite yield a cluster of well-defined plateau ages of 104–105 Ma (Figure 11b). These ages were obtained on mica clasts preserved in the mylonitic matrix and the lack of intra- and inter-sample age gradient suggests that the dated micas experienced a full reset of their K–Ar chronometer, which is supported by the U–Th/Pb age of 104–127 Ma for synkinematic monazite in the same samples [see Aumar et al. 2022]. In situ40Ar/39Ar analyses were carried on 4 thick sections of proto- to ultra-mylonites and the distribution of ages obtained on both large prekinematic mica clasts and tiny synkinematic micas is shown on Figure 11c. The main observation is that texturally earlier micas show older dates than newly crystallized micas in the foliation, mainly between 102 and 112 Ma for the clasts (n = 12) and between 94 and 102 Ma (n = 17) for the new micas. This age range (112–94 Ma) is interpreted to cover a large part of the ductile deformation in the Agly massif. Lastly, it can also be noted that these in situ 40Ar/39Ar analyses provide no evidence for the significant presence of an inherited Variscan argon component in the dated micas. These results from the Agly massif are contrasting with those of the Oisans and Cap de Creus massifs for which some micas in the high strain zones still contain inherited argon (Figures 8b,c and 9c). There are several explanations for the results obtained in the Agly massif: a higher temperature (⩾500 °C) at the beginning of the deformation supported by the micro-fabric of quartz, feldspar and garnet in the mylonites, a more penetrative and long-lived mylonitic deformation (∼40 myrs) in an extensional tectonic context, more extensive fluid-rock interactions highlighted by the widespread Cretaceous hydrothermal activity in the Northern Pyrenees [e.g. Boutin et al. 2016] and attested by a notable contribution of chlorine-derived 38Ar in some 40Ar/39Ar analyses.
3. 40Ar/39Ar geochronology of brittle deformation
Classically, the dating of brittle deformation is indirectly inferred from the dating of geological units that predate and postdate the fault activity. More recently, with the development of low temperature thermochronometric tools, fault motion is also inferred from differential thermal histories of the footwall and the hangingwall rocks [Malusà and Fitzgerald 2019, and references therein]. Rapid cooling or exhumation of wall rocks are often explained by fault motion responsible for isotherm advection. However direct 40Ar/39Ar (or K–Ar) dating of fault motion is possible even in the low temperature upper crust corresponding to the partial annealing zone of apatite fission tracks (Figure 12). 40Ar/39Ar (K–Ar) dating method has been successfully applied on fault gouges using synkinematic/authigenic clay minerals that form during the activity of shallow crustal faults from above the ductile–brittle transition [e.g. van der Pluijm et al. 2001], i.e. in a range of temperature between ∼300 °C, corresponding to the muscovite–illite transition, to ∼140–60 °C corresponding to the temperature range of illite-rich illite–smectite (I/S) interstratified mineral formation [Haines and van der Pluijm 2012]. The 40Ar/39Ar (K–Ar) dating method is very useful to provide direct constraints on the timing of shallow crust faulting but also it can yield information regarding the thermal history of the fault-zone rocks. Note that during the last decade, the U–Pb geochronology of calcite has been widely developed to address the timing of brittle faulting and fracturing [Roberts and Holdsworth 2022, for a review].
Fault-rock types and evolution of related deformation mechanisms with depth. Brittle deformation corresponds to the temperature range of low temperature thermochronometers: Zircon Fission Tracks (ZFT), Zircon (U–Th)/He (ZHe), Apatite Fission Tracks (AFT) and Apatite (U–Th)/He (AHe). Closure temperature of each thermochronometric system is reported; PAZ: partial annealing zone; PRZ: partial retention zone [modified after Malusà and Fitzgerald 2019].
3.1. Some methodological aspects
Synkinematic/authigenic clay minerals are characterized by their micron size (2 to 0.2 μm size fractions are commonly analyzed, and sometimes down to the 0.05 or up to 20 μm size fractions) that favors a 39Ar recoil effect during irradiation [Reuter and Dallmeyer 1987; Onstott et al. 1995]. The distance of 39Ar recoil during irradiation was estimated to be a mean of 0.08 μm [Onstott et al. 1995], which has significant effects on small crystallites. One must keep in mind that 39Ar recoil distance is also a function of fast neutron energy in the irradiation facility, and therefore the amount of argon loss is also dependent on this energy. Although the fraction of 39Ar lost by recoil effect generally increases with decreasing grain size, Dong et al. [1995] also showed that the amount of 39Ar recoil is rather dependent on the degree of crystallization of clay particles. During irradiation, the amount of 39Ar loss can reach 30% of the total amount of 39Ar produced which results in apparent 40Ar/39Ar ages older than K–Ar ages. This limitation favored K–Ar dating of clay minerals for a long time [Clauer et al. 2012; Clauer 2013], despite this method presents two main drawbacks: argon and potassium are measured on separated clay aliquots, and it does not permit to discriminate illite-rich mineral generations resulting from multistage crystallization. In the 90s, a sample encapsulation technique was developed in order to measure the 39Ar lost by clay particles during irradiation and to correct 40Ar/39Ar ages [Foland et al. 1992; Dong et al. 1995; Onstott et al. 1995]. Each sample (∼1 mg) is encapsulated and sealed under vacuum within a quartz tube that traps the 39Ar released during neutron irradiation. The 39Ar loss can then be measured by breaking or laser drilling the quartz tube under vacuum in the sample chamber to correct the bulk 40Ar/39Ar age obtained by classical step-heating 40Ar/39Ar method. Dong et al. [1995] defined a “retention age” (RA) corresponding to the 40Ar/39Ar total age that omits the fraction of 39Ar lost during irradiation and a “total gas age” (TGA) including 39Ar recoil. Dong et al. [1995] also showed that the fraction of 39Ar lost by recoil is a linear function of illite crystallinity. For thick crystallite, i.e. >50 nm, the “retention age” and the “total gas age” are close or identical. For thin crystallites, i.e. <10–15 nm, the “total gas age” can be considered as the best estimate of the crystallization age. For crystallites with intermediate thickness, crystallization ages fall between the “retention age” and the “total gas age”. Fitz-Diaz et al. [2016] proposed to correct 40Ar/39Ar ages by taking into account the average illite crystallite thickness (ICT)—considered as the number of TOT layer and related to illite crystallinity—and the corresponding recoiled 39Ar. This correction is particularly useful for samples containing authigenic illite with typical ICT between 10 and 50 nm.
At last, Lin et al. [2000] showed that the recoil effect is also influenced by the incident angle of the neutron flux with the c-axis of clay particles. A soft compaction of clay particles before in vacuo sealing is thus recommended to significantly reduce the recoil effect [Dong et al. 1995; Clauer et al. 2012]. During encapsulation there is a risk to heat the samples while sealing the quartz tube. The sealing is realized with a gas torch and in order to prevent heating the sample, the base of the tube containing the sample is placed either in iced water [Clauer et al. 2012; Hall 2015] or in a metallic holder cooled with liquid nitrogen. Clauer et al. [2012] questioned the accuracy of 40Ar/39Ar ages compared to K–Ar ages and argued that the lack of standardization of encapsulation processes may be a source of bias in 40Ar/39Ar dating of clay minerals and this should be addressed. To the whole, the encapsulation systems have been constantly improved and are now used routinely in several laboratories. For example, at Géosciences Montpellier, we developed an encapsulation system allowing encapsulating clay samples under very low vacuum before their irradiation, i.e. vacuum of 10−9 torr is currently reached (Figure 13) [Abd Elmola et al. 2018; Münch et al. 2021]. Reproducibility tests of 40Ar/39Ar ages on different clay aliquots and comparisons with K–Ar ages show the reliability of this encapsulation technique, with smaller error margins than for the K–Ar method.
Sketch of the encapsulation device developed at the Géosciences Montpellier noble gas laboratory. The quartz tube containing the charcoal is immersed in liquid nitrogen to improve the vacuum in the encapsulation line.
A second issue with 40Ar/39Ar (or K–Ar) dating of clay minerals from gouges is related to tectonic processes occurring within shallow crust faults: fragmentation of host rocks followed by grain-size reduction and authigenesis of clay minerals, resulting in a mixture of inherited/detrital and authigenic/synkinematic minerals and in the measurement of a mixed age. Depending on the lithology of the wall rock, authigenesis will correspond either to the illitisation of interstratified illite/smectite minerals present in sedimentary host rocks or to the neocrystallization of discrete illite in the case of metamorphic or magmatic host rocks, devoid of clay minerals [Haines and van der Pluijm 2008] (Figure 14). From the ductile/brittle transition, i.e. at temperature around 300 °C, newly formed illite is mainly of 2M1 polytype (polytypism is a variety of polymorphism common in layered minerals) with minor 1M∕1Md polytype and their relative proportion are inverted with decreasing temperatures (Figure 14). Below ∼200 °C, 1M∕1Md polytype is the only newly formed polytype and with decreasing temperatures the proportion of illite in I/S minerals decreases and ordering of illite layers within I/S also decreases (Figure 14). Thus at shallow depth, the proportion of inherited/detrital and newly formed phases can be resolved through quantitative X-ray diffraction analysis of different clay size fractions.
Mineral transformations within fault gouges to form illite-rich clay minerals suitable for 40Ar/39Ar (K/Ar) dating [modified after Haines and van der Pluijm 2008]. (A) Illitization of illite–smectite (I/S) interstratified minerals present within the wall rock; this process results in higher contents of I/S minerals within the gouge than in the wall rock. (B) Growth of authigenic discrete 1Md illite in fault gouges; this process results in a mixture of authigenic/synkinematic 1Md illite and inherited/detrital 2M1 illite ( = muscovite) from the wall rock.
For less shallow crustal faults, but still above the brittle–ductile transition, the problem of dating a mixture of inherited/detrital and authigenic/synkinematic illite type minerals is much more difficult to solve. Indeed, with X-ray diffraction techniques, muscovite forming at temperatures above 300 °C is undistinguishable from 2M1 illite polytype forming at temperatures down to ∼200 °C (Figure 12). As a consequence, if muscovite is present within host rocks both authigenic/synkinematic and inherited/detrital 2M1 polytypes are mixed within the gouge rocks and their relative proportion cannot be estimated. At temperatures near 200 °C, 1M polytype authigenesis can also occur together with 2M1 but the proportion of detrital/authigenic can still not be estimated.
Like for 40Ar/39Ar dating of ductile deformation, a very thorough and precise mineralogical study of gouge samples must be conducted before dating them. In the case of clay minerals, this mineralogical study requires X-ray diffraction techniques on both oriented and non-oriented powders in order to determine the mineralogy and quantify the proportion of different mineralogical phases. Scanning and transmission electron microscope studies are also necessary to determine the mineralogy of clay minerals and in some case their chemistry that can also be analyzed by electron microprobe but only for well-crystallized coarse clay minerals.
3.2. Age interpretation
40Ar/39Ar age spectra obtained on encapsulated clay minerals differ from classical age spectra of magmatic or metamorphic minerals due to 39Ar recoil effect during irradiation [Dong et al. 1995; Hall et al. 1997, 2000] but also to complex mixture of minerals with different age, polytype and crystallinity, without mentioning the possibility that the dated fractions may contain other types of potassium-rich minerals than phyllosilicates. In classical 40Ar/39Ar studies, samples have a size above 50 μm in order to avoid recoil induced 39Ar loss and hopefully yield spectra with a plateau comprising more than 50% of total 39Ar released and at least 3 steps. An illite spectrum shows different features: (1) a first zero age step corresponding to the recoil gas fraction, (2) staircase steps with increasing apparent ages from near zero at low temperature (3) a more or less hump-shaped age pattern with apparent ages that exceed the “retention age” and sometimes an age drop at highest temperatures (Figure 15A). This third part of the spectrum sometimes looks like a seemingly good plateau but should be interpreted with great caution and as providing a meaningless intermediate age as in the case of hump-shaped spectra obtained for mixtures of two mica generations [Wijbrans and McDougall 1986]. To estimate the age of a fault gouge motion, only the TGA’s, eventually corrected for an ICT factor, of the different clay size fractions with well-identified polytype proportions must be used. Usually, TGA’s decreases with grain size and with the relative proportion of inherited/detrital and newly formed/synkinematic minerals of a given polytype. By plotting TGA’s against the percentage of detrital polytype, the age of the pure authigenic phase can be determined with the ordinate intercept (0% of detrital component) of the best fitting line. This is the so-called “Illite Age Analysis” of Pevear [1999] and van der Pluijm et al. [2001] that has been applied on numerous shallow crust faults in various tectonic settings [e.g. Duvall et al. 2011; Haines and van der Pluijm 2008, 2010, 2012; Münch et al. 2021; Ring et al. 2017; Zwingmann et al. 2010a,b; Figure 15A]. The ages calculated reflect the most recent period of activity of the faults along their exposure.
Examples of 40Ar/39Ar illite step-heating age spectra for vacuum-encapsulated clay gouge size fractions and illite age analysis plots [Pevear 1999; van der Pluijm et al. 2001] of 40Ar/39Ar illite TGA’s vs. % 2M1 (detrital illite) of different clay-size fractions of two samples from late shallow crustal faults occurring within metamorphic units of the Rif chain [modified after Münch et al. 2021]. (A) sample M10-29 showing classical spectra for all fractions and the corresponding illite age analysis plot. (B) Sample CEU12-03 showing complex spectra for two coarse fractions with low temperature staircase steps largely overshooting the “total gas” and “retention” ages as well as the age of muscovite from host rocks; the TGA’s of corresponding fractions have been removed from illite age analysis (dots between brackets). On illite age analysis plots, the black dot on the right axis corresponds to the mean age of muscovite within the metamorphic host rocks.
However, Solum et al. [2005] demonstrated that the illite age intercept calculated using the illite age analysis also varies from the fault core to the damage zone of a single fault, as a function of the intensity of the deformation and the proportion of authigenic illite. This result highlights that a thorough structural study of the fault zone that is to be dated is crucial and that samples selected for illite dating must come from the most deformed part of the fault. As authigenic illite or I/S growth in fault gouge occurs in a well constrained temperature range, it is important to combine fault gouge dating with thermochronological study of host rocks that constrains their cooling history. Low-temperature thermochronometric studies often emphasize one or more tectonic phases as the cause of accelerated exhumation and highlight the role of major faults. The combined use of thermochronometric tools and fault gouge dating allows the age of fault activity to be better constrained [Duvall et al. 2011; Münch et al. 2021; Milesi et al. 2022; Parry et al. 2001; Tagami 2012].
In some cases, the 40Ar/39Ar age spectra of illite are more complex and the illite age analysis is not so straightforward. For example, some spectra from coarse fractions of gouge samples exhibit rather a saddle shape for the high temperature steps, i.e. after the staircase increase of apparent ages from zero, and in some cases the “total gas age” of these fractions has even been found older than the mean age of muscovite from the metamorphic host rocks [Münch et al. 2021] (Figure 15B). The shape of the spectra may be interpreted in terms of excess of radiogenic 40Ar [McDougall and Harrison 1999] that probably originates from inherited minerals because these spectra were only obtained for coarser fractions in which the proportion of inherited/detrital muscovite is higher. On the other hand, 40Ar in excess could also have been incorporated during the growth of authigenic illites from fluids circulating within the fault [e.g. Roques et al. 2020]. As the shape of the spectra allows detecting such excess of radiogenic 40Ar, it must be considered before integrating the “total gas age” of each fraction in an illite age analysis. In some cases, with particularly complex shape of spectra, this can result in the rejection from the illite age analysis of the TGA(s) of the fraction(s) exhibiting such excess 40Ar (Figure 15B).
In order to better interpret the 40Ar/39Ar age spectrum of authigenic illites, Clauer et al. [2012] applied a correction of 39Ar recoil on step-heating experiments by distributing the recoiled 39Ar on each step in proportion to its representative percentage on the retention age spectrum. The goal was to reconstruct an original spectrum, not altered by the 39Ar recoil loss during the sample irradiation, which can be interpreted using the plateau age definition. The recoil-corrected spectrum of Clauer et al. [2012] is still similar to the spectra obtained from mixtures from which these authors defined one or two “limited plateaus”, not necessarily corresponding to 3 concordant and consecutive steps. Clauer et al. [2012] claim that the “limited plateaus” represent different 39Ar reservoirs and several illite generations for which they calculate a corrected age. Even if we agree that the complex spectra can reflect the mixture of several illite generations, other processes (fluid interaction, excess argon, mixture with other clay minerals or K-bearing inherited phases) can also lead to complex spectra. Above all, the correction applied is rather basic and the calculation of plateau ages for the different illite generations on a single spectrum is highly questionable. According to Fitz-Diaz et al. [2016] an age interpretation based on 40Ar/39Ar plateau ages is meaningless for most fine-grained clays because the grain size highly influences the shape of the spectra. However, the shape of the age spectra of different clay-size fractions from a single sample is obviously useful to depict the sample crystallization history in terms of mixture, excess 40Ar incorporation and is a crucial point to take into consideration before performing an illite age analysis of a fault gouge. This is a huge advantage of the step-heating 40Ar/39Ar method compared to the complementary K–Ar method that cannot provide such information.
More complex cases come from the deeper brittle fault zones that were active at temperatures comprised between ∼200 and ∼300 °C (Figure 16). There, 40Ar/39Ar (K–Ar) ages decrease with decreasing clay-size fraction, illite crystallinity and increasing 39Ar recoil loss [Milesi 2020]. The illite age analysis cannot be applied because inherited and neo-formed illites are 2M1 polytype and the age of the finest fraction can only be considered as a maximum age as it is frequently a mixture of mainly 2M1 with minor 1M polytype. The age of coarse size fraction is potentially also a mixed age but the proportion of detrital muscovite and authigenic illite cannot be determined. However, the shape of the spectra can help to resolve if there is a mixture or not. The age of the medium size fraction may either represent a mixed age with a higher proportion of authigenic 2M1 illite or a cooling age of the same mixture as in the coarse fractions. This issue can be even more complicated if the host rock is detrital and was deeply buried at equivalent temperatures, between 200 and 300 °C, before faulting. In this case, the gouge can contain a complex mixture of detrital muscovite, 2M1 illite formed in anchizonal conditions during burial and authigenic/synkinematic 2M1 illite. The age of fault motion is then very difficult to resolve and in such a case the stratigraphic age of host rocks as well as their clay mineralogy and the regional geological constraints have to be considered in order to correctly interpret the age of fault motion [Solum et al. 2005; Abd Elmola et al. 2022].
Examples of 40Ar/39Ar illite step-heating age spectra for vacuum-encapsulated clay gouge size fractions mainly constituted of 2M1 polytype (1M polytype is solely present in the <0.2 μm fraction) from the Têt fault in Eastern Pyrénées [Milesi 2020].
4. Pseudotachylytes
The long-term activity of faults is marked by a succession of events that are difficult to decipher precisely in a continuous way, particularly in the seismogenic upper crust. Among the various fault rocks, pseudotachylytes are witnesses of frictional flash melting associated with high-velocity co-seismic ruptures most often in the shallow and brittle domains of the lithosphere. They occur as glassy veins of variable thickness and are frequently associated with cataclasites and ultra-cataclasites, resulting in a complex, frequently asymmetric, zonation of the veins. Their melt origin is supported by the presence of a glassy material and different microstructures such as microlites, spherulites, vesicles, flow structures, vein injections, and sulfide droplets that are decisive criteria to differentiate them from crush-origin pseudotachylytes [Li et al. 2022]. Because of the high cooling rates implied by their rapid solidification and their high potassium content, pseudotachylytes represent prime targets to date co-seismic displacement by the 40Ar/39Ar method [e.g. Kelley et al. 1997; Sherlock and Hetzel 2001; Müller et al. 2002; Di Vincenzo et al. 2004; Souquière et al. 2011; Menant et al. 2018; Mittempergher et al. 2022]. Step-heating analysis of pseudotachylyte fragments has produced variously successful results due to the very heterogeneous character of the veins (abundant host rock clasts, partial devitrification, inclusions …) and the often polyphased history of fault zones [e.g. Torgersen et al. 2022]. The use of the in situ laser probe 40Ar/39Ar dating method has partially overcome some of these drawbacks by focusing the analyses on particular zones rich in molten glass and poor in clasts, such as the vein margins or the injection veins. Below (Figure 17) are some examples showing some characteristics of pseudotachylytes and some data obtained in the western Alps (Ivrea zone, Dent Blanche) using both step-heating and in situ 40Ar/39Ar analyses.
(a) Pseudotachylyte in a gabbro of the Ivrea zone. Note the injection vein perpendicular to the main vein composed of cataclasites on the left and glassy, black pseudotachylyte on the right; (b) Cathodoluminescence image of the contact between a pseutotachylyte vein and a gneiss host-rock cut by an injection vein (Dora Maira). The melt glass (in blue) is concentrated at the very contact with the undeformed host-rock and within the injection vein; (c) BSE image of a pseudotachylyte in a paragneiss of the Ivrea zone showing disoriented mica microlites around a corroded quartz; (d) 40Ar/39Ar age spectra of two pseudotachylyte samples (injection veins) from the Valpelline unit [Dent Blanche, Menant et al. 2018] showing low and high devitrification effects respectively. Note that the age spectrum of the pseudotachylyte (in red) overlaps the Rb–Sr age of phengites in a nearby shear zone; (e) distribution of in situ 40Ar/39Ar dates within the injection vein shown in (a). Relatively large errors are due to the low amount of potassium in the gabbro host-rock [Souquière et al. 2011]; (f) inverse isochron diagram of the same sample showing an agreement between peak and intercept ages and an initial atmospheric 40Ar/36Ar composition (no excess argon).
Therefore, the 40Ar/39Ar dating of pseudotachylytes and associated tectonites appears essential to better understand the mechanics of earthquakes and the links with tectonic activity recorded by ductile and brittle faults in the mid to upper crust. The examples of Figure 17 highlight the microstructural complexity of pseudotachylytes and the difficulty to use them as precise 40Ar/39Ar geochronometer of seismic ruptures. The behavior of this chronometer remains to be better understood and the effects of various parameters will require consideration in future studies, including slip velocity at rupture and amount of generated melt, mechanical and chemical zonation of the veins, lithology and age of the host rock, nature and distribution of clasts and fragments extracted from the host rock, nature and distribution of neo-formed phases, presence of fluids, alteration phases or devitrification effects.
5. Conclusion
This overview of some recent developments in the 40Ar/39Ar dating of deformation in fault zones shows that the interpretation of geochronological data is not so straightforward and that it must be supported by a good knowledge of the field, and detailed (micro-) structural and petrological data. For ductile deformations at temperature <500–550 °C, the different examples reported in Part 3 show that the 40Ar/39Ar ages recorded by newly formed synkinematic micas (and more rarely amphibole) have to be considered as crystallization ages rather than cooling ages. Frequently, these newly formed minerals are mixed with prekinematic minerals inherited from the original host rock or from an earlier phase of deformation along the fault zone. In this particular case, the in situ laser probe 40Ar/39Ar method is very powerful since it allows the dating of the different K-bearing mineral generations identified on microstructural and chemical grounds. The behavior of the inherited minerals is strongly dependent on the P–T-fluids conditions that prevail during deformation, on the size and chemistry of these minerals, on the presence or not of short diffusion pathways, on the closed/open system behavior of the studied tectonites. In some favorable geological setting, the K–Ar chronometer of inherited minerals can be fully reset while in other cases, only a limited resetting can be observed. For deeper conditions of deformation (e.g. high-pressure low temperature terranes), the closed-open system behavior of the host rock can be an critical parameter to consider since it can result in excess argon contamination of the newly crystallized minerals due to limited fluid mobility in a closed system [Laurent et al. 2017]. For all these situations, the integration of 40Ar/39Ar dates with independent geochronometers (Rb–Sr, U–Pb) may be particularly useful for reconstructing P–T–d–t paths in deformed rocks.
For brittle faults, clays from fault gouges are the first choice target for 40Ar/39Ar investigations. However, due to 39Ar recoil effect during irradiation, to complex mixture in the fault gouge of minerals with different age, polytype and crystallinity, to complex fluid rock interactions, the interpretation of 40Ar/39Ar data is not always simply a matter of processing a binary mixture between a single generation of 2M1 phyllosilicate inherited from the host rock and an authigenic 1M clay synkinematic fraction. By contrast with the K–Ar, the 40Ar/39Ar step-heating method has the potential to reveal the different argon reservoirs that contribute to the argon release and to eliminate samples with a too complex isotopic signature, such as those contaminated by excess argon. The step-heating approach can also bring information on complex mixtures of inherited and authigenic phyllosilicates in fault gouges with a multiphase activity. The coupling of 40Ar/39Ar studies in fault gouges with multimethod thermochronological studies (fission tracks, U–Th–Sm/He) in the host rocks may shed light on this repeated motion of brittle faults during exhumation processes using thermal and tectonic modeling (QTQt, Pecube). This aspect remains to be more systematically explored in orogens.
Conflicts of interest
Authors have no conflict of interest to declare.
Acknowledgements
Part of this work (pseudotachylytes, Têt fault) has received fundings from different research programs (3F, TelluS) of the Institut National des Sciences de l’Univers, CNRS. Stéphane Scaillet, Michel Faure and an anonymous reviewer are thanked for their comments that improved the quality of a first manuscript.