Metasomatic Reactions between Archean Dunite and Trondhjemite at the Seqi Olivine Mine in Greenland

Abstract

A metasomatic zone formed between the contact of a 2940 ± 5 Ma intrusive trondhjemite sheet in the Archean dunite of the Seqi Ultramafic Complex, SW Greenland, consists of three distinct mineral zones dominated by (1) talc, (2) anthophyllite, and (3) phlogopite. These zones supposedly resulted from a process of dissolution of olivine by silica rich fluid residual from the trondhjemite magma, with crystallization of secondary minerals along a compositional gradient in the fluid phase. A zircon crystal inclusion in a large (4 cm) olivine porphyroblast was dated in situ via LA-ICP-MS U–Pb isotope analysis, yielding a weighted mean ²⁰⁷Pb/²⁰⁶Pb age of 2963 ± 1 Ma, which coincides with granulite facies metamorphism and potential dehydration. Considering phase relations appropriate for the dunite composition, we deduced the talc forming conditions to be at temperatures of 600–650 °C and at a pressure below 1 GPa. This is supported by oxygen isotope data for talc, anthophyllite and phlogopite in the metasomatic zone, which suggests formation in the temperature range of 600–700 °C from fluids that had a δ¹⁸O of ~8‰ and a Δ’¹⁷O_0.528 of about −40 ppm, i.e., from fluids that could have been derived from the late stage trondhjemite sheet.

1. Introduction

Metasomatism is the process of compositional change in a rock due to the removal and/or addition of chemical components during the interaction of a precursor rock with either melt or hydrous fluid. The main driver of metasomatic reactions is a chemical gradient between the rock and the fluid/melt phase, which leads to the exchange of components and therefore potentially phase transitions.

The Archean crust in the Nuuk region of SW Greenland ranges in age from 3.8 Ga tonalite to 2.5 Ga granite [1,2]. Interspersed with such evolved continental crust, are scattered enclaves of both mafic and ultramafic rocks, which range in size from decimeter to kilometer scale (Figure 1). These remnants of older crustal components in the Archean continental crust are dominated by mafic metavolcanic rocks [3], with lesser amounts of plutonic rocks, including anorthosite, gabbro, norite, and peridotite [4]. In addition, several crustal peridotite bodies, which represent large fragments and enclaves derived from layered complex of unknown age within the tonalitic orthogneiss, have been identified in SW Greenland in recent years (e.g., [5,6]).

The relationship between ultramafic rocks and the hosting continental crust is of specific interest due to their contrasting geochemical compositions and hence high chemical potential gradient at their contact. Contact zones between peridotite and felsic lithologies have previously been shown to result in unusual metasomatic minerals, including corundum, formed during high-grade metamorphic conditions [7,8].

In the present study, we investigated metasomatic reactions observed at the interface between Archean dunite and granitoid felsic sheets at the Seqi Olivine Mine, which led to the formation of phlogopite, anthophyllite, and talc with decreasing distance from the felsic intrusives. We combined mineralogical and geochemical data to examine the compositional changes and to characterize the element mobility during this metasomatic process at amphibolite facies conditions.

2. Geological Setting

The Nuuk region of SW Greenland hosts the largest expanse of Eoarchean continental crust anywhere on Earth, known as the Itsaq Gneiss Complex [1]. This orthogneiss complex forms the nucleus of the North Atlantic Craton, along which several Mesoarchean crustal terranes were amalgamated until their common stabilization at the end of the Archean Eon [10] (Figure 1).

One of these younger crustal blocks is the Mesoarchean Akia Terrane, which hosts the rocks of the present study. It formed during two distinct stages of crustal growth with an early 3.2 Ga dioritic core, and a ca. 3.0 Ga tonalitic phase [11,12,13]. The Akia Terrane was amalgamated to the Itsaq Gneiss Complex at around 2.8–2.7 Ga [10], which was then followed by suturing of the contact by the intrusion of the Qôrqut Granite Complex at around 2.5 Ga [2].

Intercalated within the granitoid crust of the Akia Terrane are remnants of supracrustal/greenstone belts and layered ultramafic complexes [9,14,15]. One of the most prominent of the ultramafic bodies is the Seqi Ultramafic Complex [5], which is dominated by dunite, mined recently (between 2005 and 2010) for industrial grade olivine by the Swedish company Minelco. The Seqi Ultramafic Complex covers an area of approximately 0.5 km². The olivine mine open pit covers about a quarter of the exposed dunite body.

The various types of dunite in the Seqi complex are exceptionally fresh and exhibit well-equilibrated granoblastic metamorphic textures. The dunites only display vague retrogression as seen by the formation of chlorite along grain boundaries, with breakdown of chromite grains. In addition to such relatively late features are also evidence of higher grade metamorphism in the form of a large poikilitic orthopyroxene that overgrows aggregates of olivine (Figure 2a), amphibole veins (Figure 2b), and euhedral olivine porphyroblasts, which commonly preserve cores of actinolite (Figure 2c,d). These features attest to an early stage of hydration of the Seqi dunites, which was followed by widespread dehydration and re-equilibration of the ultramafic rocks.

The minimum age of the Seqi complex is constrained by an intrusive tonalitic sheet at 2978 ± 8 Ma [5], coincident with the transition from continental crust formation to regional granulite facies metamorphism in the Akia Terrane [16]. This event might indeed be responsible for the dehydration of the Seqi dunites. However, as we demonstrated in this study, there was also local infiltration of fluids and/melts during the intrusion of a felsic granitoid.

The present study focused on metasomatic reactions that are observed at the contact between late intrusive granitoid sheets and dunite within the Seqi Ultramafic Complex, such as the examples given in Figure 3. These contacts are associated with significant hydration of the protolith, with concomitant changes of the granitoid itself.

3. Materials and Methods

3.1. Samples and Petrography

The samples used in the present study are from the Seqi Olivine Mine [5], and include the contact zone between refractory cumulate dunite (represented by sample 186452) and intrusive granitoid pegmatite (represented by sample 186453, Figure 3c). The latter granitoid pegmatite was previously dated at 2940 ± 5 Ma by LA-ICP-MS (Laser Ablation Inductively Coupled Plasma Mass Spectrometry) U–Pb in zircon [5]. Our analyses showed that the pegmatite is dominated by albitic plagioclase, and is therefore classified as trondhjemite.

The studied metasomatic reaction consists of three mineral associations going from a mica-rich proximal zone at the pegmatite side, to a fibrous white amphibole zone, and finally a talc-dominated zone on the dunite side of the reaction front (Figure 4). Microscopy showed elongated, fibrous, colorless mineral with high relief, occurring in all three metasomatic zones (Figure 5c). This mineral was identified by X-Ray Diffraction (XRD) and Electron Microprobe Analysis (EMPA) as anthophyllite. A colorless mineral with a low relief, was also present in most of the samples and was identified as talc (Figure 5a). A brown to yellowish flaky mineral with a low relief, was also present throughout the metasomatic zone (Figure 5d), and was identified as phlogopite. Opaque minerals appeared as a minor phase (Figure 5b,c) and were identified as spinel (chromite and magnetite) by EMPA analysis.

The host dunite was previously described in detail by Szilas et al. [5]. Although several textural types occur in the Seqi olivine mine, the main one can be described as being texturally homogeneous with unserpentinized olivine having well-equilibrated granoblastic annealing texture with 120° angles between the grains (Figure 5a), consistent with the high-grade granulite facies metamorphic conditions that were documented for this region [12,16].

The talc zone (Figure 4 and Figure 5b) is predominately talc (plastically deformed finer grains) in combination with some lepidoblastic anthophyllite. Opaque minerals occur as small partially digested grains (Figure 6a). Flaky/nematoblastic mica is present (weak greenish-yellow pleochroism), but to a lesser degree than anthophyllite. Olivine relicts are rarely observed.

The amphibole zone (Figure 5c) is dominated by oriented aggregates of fibrous anthophyllite crystals growing perpendicular to the contact plane (Figure 4 and Figure 6b) with minor amounts of irregular to rounded opaque inclusions. This zone has a gradational transition into the mica zone, and anthophyllite mingles with broader phlogopite grains (green-colorless pleochroism) with successive disorientation of the grains.

The transitional mixture of anthophyllite and phlogopite with minor amounts of talc gives way to mica (phlogopite) with a weak greenish-yellow pleochroism. This mica zone (Figure 5d and Figure 6c) eventually becomes entirely dominated by phlogopite.

Additionally, a zircon grain was identified within a euhedral olivine porphyroblast in dunite sample 186460. The zircon was dated by in situ U–Pb isotope analyses using the LA-MC-ICP-MS method (laser ablation multi-collector inductively coupled plasma mass spectrometry).

3.2. Methods

3.2.1. Bulk-Rock Geochemical Data

Bulk-rock major and trace element data were acquired from the geochemical laboratory at Washington State University (WSU, Pullman, WA, USA). The procedure involves flux fusion digestion of rock powders followed by routine XRF (X-Ray Fluorescence) and ICP-MS analysis. Potential contamination from the Li-tetraborate flux was corrected for by a blank subtraction. The full analytical procedures are described on WSU’s website and the reader is referred to this for additional details [17].

3.2.2. Electron Microprobe Analysis

Electron Microprobe Analysis (EMPA) was conducted using a JEOL JXA8200 at the University of Copenhagen (Copenhagen, Denmark). A beam current of 15 nA, acceleration voltage of 15 kV, and spot size of 5μm was used. The counting time was set for 20 s for each element, except for sodium and potassium, for which 10 s was used. The elements chosen for the analysis were silicon (Si), manganese (Mn), potassium (K), chromium (Cr), sodium (Na), aluminum (Al), iron (Fe), calcium (Ca), nickel (Ni), magnesium (Mg), and titanium (Ti) using the standardization described by [18].

3.2.3. Powder X-ray Diffraction

The powder x-ray diffraction (PXRD) instrument used in this investigation was a Bruker-AXS D8 Advance diffractometer with a primary Ge111 monochromator and a LinxEye silicon-strip detector with an active surface covering 3.3° located at the University of Copenhagen (Copenhagen, Denmark). The used radiation was from a sealed Cu tube (1.54059 Å). The measurements were made between 5 and 90° 2θ, in steps of 0.02° and a measuring time of 4 s. The reflection Bragg-Brentano technique was used with sample rotation and a fixed divergence at 0.25°. Sample preparation included crushing samples to a grain size of less than 45 μm, and mounting the powder in a metal sample holder with a cavity of 2 mm depth.

Mineral identification was done using an own database with mineral diffraction data and refined with a help of Topas6 Software (version 1, 2017, Bruker AXS, Billerica, MA, United States) using the Rietveld method, with a simultaneous quantitative analysis of mineral proportions.

3.2.4. Triple Oxygen Isotope Analysis

The triple O isotope compositions (δ¹⁷O, δ¹⁸O) of samples from the intrusive granitoid pegmatite, the talc zone, the antophyllite zone, and the mica zone were determined using the analytical setup at the University of Göttingen, Germany [19]. Two samples of the refractory dunite were analyzed for their triple O isotope compositions as well. For each sample, approximately 2.2 mg of material was reacted with excess BrF₅ by laser heating in a stainless-steel sample chamber. The O₂ that was produced by the reaction was cleaned from contaminant gasses by, respectively, distillation at −196 °C (e.g., BrF₅), reaction with NaCl (F₂) followed by distillation (Cl₂), and with a gas chromatograph (e.g., N₂, NF₃). The δ¹⁷O and δ¹⁸O values of the sample O₂ were then determined with a MAT 253 gas source mass spectrometer in dual inlet mode, relative to a reference gas of which the composition was determined relative to the VSMOW2 water standard [19].

The triple O isotope compositions of the samples are expressed below with the parameter Δ’¹⁷O_0.528, which denotes the deviation of the linearized δ¹⁷O at the linearized δ¹⁸O of the sample, relative to a reference line lambda (λ) with a slope of 0.528 (e.g., [20,21]. The samples were analyzed together with UWG-2 garnet [22] and San Carlos olivine as external reference materials. The average δ¹⁸O and Δ’¹⁷O_0.528 values of UWG-2 garnet (δ¹⁸O = 5.74‰; Δ’¹⁷O_0.528 = −56 ppm) and San Carlos olivine (δ¹⁸O = 5.45‰; Δ’¹⁷O_0.528 = −49 ppm) that were analyzed together with the samples are in good agreement with the long-term average values for these reference materials from our labs. The material of San Carlos olivine that was analyzed in this study was sampled from a different olivine crystal than the material that was analyzed by [19], and had a higher δ¹⁸O than the material in the [20] by ~0.2‰ relative to UWG-2 garnet.

3.2.5. Zircon U–Pb Isotope Analysis

U–Pb isotopes were measured using a Nu Plasma II multi-collector ICPMS with high purity Ar as the plasma gas (flow rate 0.98 L min⁻¹) at Curtin University, Bentley, Australia. During time-resolved processing, contamination resulting from inclusions and compositional zoning was monitored, and only the relevant part of the signal was integrated. The primary reference material used for U–Pb dating in this study was zircon standard OG1 (3465.4 ± 0.6 Ma [23]), which has an ablation response similar to the Archean grain of interest. Secondary standard zircon 91500 (1062.4 ± 0.4 Ma [24], GJ-1 (601.7 ± 1.4 Ma [25]), and Plešovice (337.13 ± 0.37 Ma [26]) yielded ²⁰⁶Pb/²³⁸U ages within 1% of the accepted value when reduced using appropriate matrix matched standards. The time-resolved mass spectra were reduced using the U–Pb Geochronology data reduction schemes in IoliteTM [27], and in-house Microsoft Excel macros. No common lead corrections were deemed necessary due to generally low ²⁰⁴Pb counts. All ages are reported with ±2σ uncertainties unless specifically stated otherwise.

4. Results

4.1. Bulk-Rock Geochemical Data

The three reaction zones (Figure 4) were measured for their bulk-rock major and trace element compositions for comparison with the intrusive trondhjemite sheet and the host dunite. The data for the latter were previously reported by Szilas et al. [5] and are thus not presented in Supplementary Table S1, but are included for reference in the geochemical diagrams.

The trondhjemite sample has the highest SiO₂ and Al₂O₃ contents of the samples, with values of 83.23 wt.% and 9.96 wt.%, respectively. This is consistent with the predominance of quartz and albite in this sample, which also explain the high Na₂O content of 4.64 wt.% (Figure 7). It should be noted that it is difficult to obtain an average sample of the sheet because of relatively large crystals of albite and quartz. The sample used for the whole rock chemical analysis evidently contained more quartz than the one used for the PXRD analysis. According to the latter one the amounts of SiO₂, Al₂O₃, and Na₂O are 61.58, 23.77, and 14.40 wt.%, respectively, due to a significantly higher content of albite (80.5 wt.%) and a lower content of quartz (19 wt.%), which is consistent with the pegmatitic nature of this sample. In terms of trace element abundances, the pegmatitic trondhjemite has low contents of compatible elements, as expected (Figure 8). The trace element pattern of the trondhjemite sample is quite typical of the continental crust, with the exception that this sample has a strong negative anomaly of P₂O₅, in addition to the typical negative anomalies of Nb and TiO₂ (Figure 9).

The three metasomatic zones (talc, anthophyllite, and phlogopite) have quite comparable bulk-rock major element compositions, with a few distinct features. In terms of their SiO₂ and MgO content, they plot intermediate between the dunite and granitoid endmembers (Figure 7). The amphibole zone is notably enriched in FeO^T (9.74 wt.%) and MnO (0.584 wt.%) in comparison to all other measured samples. In contrast, the mica zone is enriched in K₂O (4.62 wt.%) and TiO₂ (0.077 wt.%), as well as in all incompatible trace elements to the point where it exceeds the abundances in even the trondhjemite sample (Figure 9). However, this might be due to a non-representative choice of the sample of trondhjemite, as mentioned before. The talc zone is comparable to the host dunite in most aspects, except for having significantly higher SiO₂ (60.87 wt.%) and lower MgO (29.80 wt.%). Surprisingly, Ni, V, and Sc are lower in the talc zone, but Cr is high, in comparison with the Seqi dunite (Figure 8). The talc zone is also enriched in K₂O, Cs, Rb, Pb, as well as, U (Figure 9).

4.2. Electron Microprobe Analysis

The three reaction zones, and the intrusive trondhjemite were studied by EMPA. The dunites of the Seqi Ultramafic Complex were previously described by Szilas et al. [5], so we refer the reader to that study for more details on the host dunite. Below we outline the results of the main minerals that were observed in the different reaction zones (

Table 1. The average mineral compositions from the talc zone (see Figure 6a).

Formula	Sites	Mineral
Na0.01(Ca0.04Mg5.36Mn0.1Fe1.42Ni0.03)Si8.01O22(OH)2	Σ A(0.01), M(6.95), T(8.01)	Anthophyllite
$\left({\mathrm{Cr}}_{1.61}{\mathrm{Fe}}_{1.36}{\mathrm{Al}}_{0.08}{\mathrm{Mg}}_{0.05}{\mathrm{Mn}}_{0.04}{\mathrm{Ti}}_{0.01}\right){\mathrm{O}}_4$	Σ M(3.15)	Chromite
(K0.72Na0.05Ca0.02)(Mg2.67Fe0.25Al0.02Cr0.02Ni0.02Ti0.01)(Si3.09Al0.91)O10(OH)2	Σ A(0.79), M(2.99), T(4)	Phlogopite
$\left({\mathrm{Mg}}_{2.81}{\mathrm{Fe}}_{0.09}{\mathrm{Al}}_{0.01}{\mathrm{Ni}}_{0.03}\right){\mathrm{Si}}_{4.03}{\mathrm{O}}_{10}{\left(\mathrm{OH}\right)}_2$	Σ M(2.94), T(4.03)	Talc

,

Table 2. The average mineral compositions from the amphibole zone (see Figure 6b).

Formula	Sites	Mineral
Na0.01(Ca0.05Mg5.36Fe1.38Mn0.08Ni0.03Al0.01)Si8.05O22(OH)2	Σ A(0.01), M(6.85), T(8.05)	Anthophyllite
$\left({\mathrm{Cr}}_{1.51}{\mathrm{Fe}}_{1.45}{\mathrm{Al}}_{0.11}{\mathrm{Mg}}_{0.06}{\mathrm{Mn}}_{0.04}{\mathrm{Ti}}_{0.01}\right){\mathrm{O}}_4$	Σ M(3.18)	Chromite
(K0.82Na0.05Ca0.01)(Mg2.65Fe0.26Ni0.02Al0.02Cr0.03Ti0.01)(Al0.94Si3.06)O10(OH)2	Σ A(0.88), M(2.99), T(4)	Phlogopite
(Mg2.81Fe0.1Ni0.03)Si4.01O10(OH)2	Σ M(2.94), T(4.01)	Talc

,

Table 3. The average mineral compositions from the mica zone (see Figure 6c).

Formula	Sites	Mineral
(K0.84Na0.05Ca0.01)(Mg2.66Fe0.26Ni0.02Al0.03Ti0.01)(Al0.94Si3.06)O10(OH)2	Σ A(0.9), M(2.98), T(4)	Phlogopite
(Na0.01Mg2.85Fe0.11Ni0.02Al0.01)Si3.99O10(OH)2	Σ M(3), T(4)	Talc

and

Table 4. The average mineral compositions from the trondhjemite (see Figure 6d).

Formula	Sites	Mineral
$\left({\mathrm{Na}}_{0.90}{\mathrm{Ca}}_{0.12}{\mathrm{K}}_{0.01}\right)\left({\mathrm{Si}}_{2.88}{\mathrm{Al}}_{1.10}\right){\mathrm{O}}_8$	Σ A(1.03), T(3.98)	Albite
(K0.93 Na0.01)(Mg1.45Fe1.07Al0.20Ti0.12Mn0.01)(Si2.88Al1.12)O10(OH)2	Σ A(0.94), M(2.85), T(4)	Biotite
K0.01(Na0.05Ca1.89Mg3.87Fe1.04Al0.1)Si8O22(OH)2	Σ A(0.01), M(6.95), T(8)	Tremolite
	SiO2	Quartz

). The full EMPA data (wt.%) are presented in the Supplementary Table S2.

4.3. Powder X-ray Diffraction

The results of PXRD are represented in

Table 5. Mineral compositions and proportions obtained by a Rietveld refinement of the PXRD data.

Talc Zone	Formula	wt.%
Talc	Mg2.9Fe0.1Si4O10(OH)2	93.7
Anthophyllite	Mg5.76Fe1.24Si8O22(OH)2 *	0.5
Phlogopite	KMg2.76Fe0.24AlSi3O10(OH)2	5.8
Amphibole Zone	Formula	wt.%
Talc	Mg2.8Fe0.2Si4O10(OH)2	1
Anthophyllite	Mg5.76Fe1.24Si8O22(OH)2	92
Phlogopite	KMg2.76Fe0.24AlSi3O10(OH)2	7
Chromite	Cr2FeO4 *	<0.1
Mica Zone	Formula	wt.%
Talc	Mg2.9Fe0.1Si4O10(OH)2	4
Phlogopite	KMg2.93Fe0.07Al1.19Si2.81O10(OH)2	61.3
Tremolite	Ca1.84Fe0.08Mg5Si8O22(OH)2	34.5
Chlorite	Mg3Fe3Al1.39Si2.61O10(OH)8	0.2

* The composition was not refined due to a small quantity.

and diffractograms with results of the Rietveld refinement can be seen in the Supplementary Figures S1–S4. It is important to note that XRD gives a number of electrons for a structural site, without element speciation. It is thus possible to express simultaneous occupancies for only two chemical elements at a time unless crystal chemical parameters (e.g., bond lengths) are involved, which is practically feasible only for single crystal studies, not for PXRD on mineral mixtures. In the present cases, the main components in solid solutions at various structural sites were Mg and Fe, so the number of electrons for mixed sites was interpreted as due to Mg/Fe proportions and this is reflected in the calculated formulae.

It can be seen that mineral compositions and mineral proportions in the talc and amphibole zone correspond well with the EMPA (Table 1 and Table 2) and the bulk chemical analysis. In the mica zone; however, the amphibole registered by PXRD is tremolite and not anthophyllite. This can be explained by the fact that this zone was analyzed close to the amphibole zone in EMPA (both zones were present in the polished section), whereas for PXRD a sample further away from the contact was taken in order to get a cleaner picture of this zone. Bulk-rock geochemical analysis, which shows significant Ca content, supports the PXRD results.

4.4. Triple Oxygen Isotope Analysis

The oxygen isotope compositions of the samples are given in

Table 6. Triple oxygen isotope compositions. Measurement precisions (1 S.D.) are ±0.15‰ and ±10 ppm for δ¹⁸O and Δ’¹⁷O_0.528, respectively, based on the repeated analysis of UWG-2 garnet and San Carlos olivine.

Sample	δ17O [‰]	δ18O [‰]	Δ’17O0.528 [ppm]
186452 Dunite	2.71	5.19	−27
186478 Dunite	2.66	5.11	−43
208001 Talc zone	3.34	6.40	−37
208002 Amphibole zone	3.21	6.17	−40
208003 Mica zone	2.94	5.65	−33
186453 Trondhjemite	4.57	8.76	−43

. The δ¹⁸O and Δ’¹⁷O_0.528 values of the trondhjemite sample are typical for evolved igneous rocks, and so are the triple O isotope compositions of the minerals from the metasomatic zones (e.g., [21,29]). The δ¹⁸O and Δ’¹⁷O_0.528 of two dunite samples are similar to δ¹⁸O and Δ’¹⁷O_0.528 of olivine in mantle peridotites (e.g., [19,20,30]).

4.5. Zircon U-Pb Isotope Analysis

One large zircon grain (length = 250 µm), included within olivine from Seqi dunite sample 186460, was targeted for U–Pb geochronology and a concordia plot is presented in Figure 10 below (see sample image in Supplementary Figure S5). This grain was spot ablated 26 times using a 22 µm spot across the exposed polished surface. Analyses indicated moderate to high U content of 220 to 609 ppm and relatively consistent Th/U ratios of around 0.2. All analyses were within 5% of Concordia and yield a weighted mean ²⁰⁷Pb/²⁰⁶Pb age of 2963 ± 1 Ma (n = 22; MSWD = 1.8; see Supplementary Figure S6). Because we did not observe any systematic age variation across the zircon grain, this age was interpreted as the age of magmatic crystallization of a metasomatic component incorporated into this rock. The U–Pb isotope data are reported in the Supplementary Table S3.

5. Discussion

5.1. Pressure-Temperature-Time Constrains of Talc Fomation

Studies of the Ulamertoq ultramafic body (Figure 1) indicated that the peridotites of the Fiskefjord region, which Seqi is part of, potentially experienced serpentinization followed by complete dehydration [31,32]. This is supported by the occurrence of humite-group minerals, and the commonly observed magnetite inclusions in olivine. Thus, the metamorphic history of the studied rocks is potentially quite complicated. The various textural types of Seqi dunite also contain magnetite inclusions, and previous serpentinization of these rocks can therefore not be excluded. However, it should be noted that the oxygen isotope compositions of the Seqi olivine is within the range of magmatic rocks, which places constraints on the degree of potential early serpentinization. The timing of talc formation at Seqi is not well defined, but must certainly postdate the 2990 Ma high temperature granulite facies event in the Akia Terrane, which caused extensive dehydration [11].

Kirkland et al. [33] used U–Pb isotope dating of apatite, which has a closure temperature of 375 to 600 °C [34,35], to provide constraints on the lower temperature evolution of the Akia Terrane. That study found evidence for thermal events ranging in age from ca. 2.8 to 1.8 Ga. We can rule out talc formation during the younger part of this period, because regional events from about 2.4 to 1.8 Ga were marked by mafic dyke swarms, which only had very localized alteration that was associated with low grade epidote and sericite [36].

The Nuuk region was characterized by terrane amalgamation from around 2.8 to 2.7 Ga [10,37], followed by crustal scale shear zone development at ca. 2.6 Ga [38,39], and finally deep crustal anataxis and injection of the Qôrqut Granite Complex at 2550 Ma [2]. The two latter events would both have been likely to cause significant fluid infiltration in the crust causing metasomatism; however, only the granite injection has documented meteoric water signatures in the form of low δ¹⁸O in zircon.

The U–Pb isotope data for zircon included in olivine presented in this study demonstrated a thermal event at 2963 ± 1 Ma, which is consistent with previous zircon ages obtained for granitoid sheets within the Seqi Olivine Mine of 2978 ± 8 Ma (host tonalite), 2963 ± 6 Ma (intrusive tonalite sheet), and 2940 ± 5 Ma (granitoid pegmatite) [5]. We interpreted this zircon grain to have crystallized during metasomatism and melt infiltration into the ultramafic body, because preliminary Re–Os age of ca. 3.3 Ga of the Seqi Ultramafic Complex [40] rules out a xenocrystic origin of the zircon grain. This is also obvious, because the tonalitic protolith of the regional TTG (tonalite–trondhjemite–granodiorite) gneiss formed between 3050–3000 Ma [13]. The lack of any systematic age correlation across the zircon grain also rules out resetting of the grain by Pb-loss. The fact that this zircon crystal was found within a porphyroblastic olivine grain indicates that the dunites of the Seqi Ultramafic Complex experienced metasomatism associated with the intrusion of granitoid sheets at temperatures above 700 °C, which represents the boundary between olivine and serpentine/talc stability for the Seqi bulk-rock compositions [5]. These conditions are in agreement with the regional P–T–t evolution indicating granulite facies metamorphism from around 2990–2940 Ma [16].

The transition temperature from talc to anthophyllite in ultramafic rocks in the Bergell tonalite area (Italy) was determined to be around 650 °C by Trommsdorff and Connolly [41]. This estimate is compatible with the occurrence of talc for the given Seqi bulk-rock composition, which outlines a P–T range of less than 700 °C at below 1 GPa [5]. This field also allows for orthoamphibole to be stable within the same bulk rock as talc, which is indeed consistent with the observed petrography (Figure 6a).

From the above evidence, we can deduce that the talc formation in the Seqi olivine mine must have occurred at temperatures of 600–650 °C below 1 GPa at 2940 ± 5 Ma, and was associated with the intrusion of the late stage pegmatitic trondhjemite sheets.

5.2. Oxygen Isotope Constraints on the Origin and Temperature of Metasomatic Fluids

To further study the temperature of talc formation, we calculated δ¹⁸O values of aqueous fluids that could have been in equilibrium with the talc, anthophyllite, and phlogopite samples at a range of temperatures, using available talc-water, anthophyllite-water, and phlogopite-water fractionation factors [42,43]. We also calculated δ¹⁸O of aqueous fluids that could have been in equilibrium with the trondhjemite sample at different temperatures, based on the modal abundances of 80.61% albite, 18.99% quartz, and 0.4% biotite in the trondhjemite sample as determined by XRD analysis, using the available biotite-water, albite-water, and quartz-water fractionation factors by [44], [45] and [43], respectively. Depending on the talc-water fractionation factors that are used in the calculation, the calculated compositions of fluids that were potentially in equilibrium with the talc sample either range in δ¹⁸O from 7 to 8‰ (c.f. the semi-empirical fractionation factors by Richter and Hoernes [42]) or, alternatively, they range from 5 to 6.5‰ (c.f., the theoretical fractionation factors by Zheng [43]). Whereas fluids with a δ¹⁸O of 7 to 8‰ could have been derived from an intrusion that had a similar δ¹⁸O as the trondhjemite sample, fluids with a δ¹⁸O < 6.5 could not have been derived from the trondhjemite sample at any temperature (Figure 11). We therefore suggest that the semi-empirical talc-water fractionation factors by Richter and Hoernes [42] more accurately correspond to the fractionation factors for the talc-water system that is studied here than the theoretical fractionation factors by Zheng [43]. The amount of early serpentinization experienced by the Seqi rocks is difficult to assess based on these oxygen isotope data, because the water/rock ratios at which serpentinization occurred are unknown. The δ¹⁸O values of the dunite samples are similar to the values of typical mantle olivine, which would suggest that either low degrees of serpentinization could have occurred at high water/rock ratios or, alternatively, that high degrees of serpentinization could have occurred at low water/rock ratios.

The compositions of fluids that could have been in O isotope equilibrium with the talc, anthophyllite and phlogopite samples intersect at a δ¹⁸O of ~8‰ within a temperature range of 600–700 °C (c.f. the talc-water fractionation factors by Richter and Hoernes [42] (Figure 11). The compositions of fluids that could have been derived from the trondhjemite sample also intersect this temperature range at a δ¹⁸O of ~8‰. We therefore suggest that the talc, anthophyllite, and phlogopite samples could have formed at a temperature range of 600–700 °C from fluids that had a δ¹⁸O of ~8‰, i.e., from fluids that could have been derived from the trondhjemite intrusion. The range of formation temperatures for the talc, anthophyllite, and phlogopite samples as reconstructed from their δ¹⁸O values (600–700 °C) is in good agreement with the temperature constraints from phase relations of the dunite bulk-rock composition (600–650 °C; see above).

The Δ’¹⁷O_0.528 values of the talc, anthophyllite, and phlogopite samples are similar to those of most igneous rocks and minerals (e.g., [21]), including the trondhjemite sample. This observation confirms that fluids that were involved in the formation of these particular samples were derived from magmatic waters with a Δ’¹⁷O_0.528 of about −40 ppm. In contrast, rocks that have become hydrothermally altered by fluids that were derived from seawater or from meteorite waters, for example, typically have δ¹⁸O and Δ’¹⁷O_0.528 values that are lower and higher than igneous rocks, respectively (e.g., [21,46]).

5.3. Metasomatic Element Mobility

The mica zone stands out from the rest of the samples by having extreme enrichment in K₂O, and CaO even when compared to the granitoid pegmatite. This likely reflects local stabilization and scavenging from a fluid phase. However, the relatively immobile TiO₂, Al₂O₃, and heavy rare earth elements are also strongly enriched in the mica zone (Figure 9), which points to residual enrichment by dissolution of the pegmatite.

The anthophyllite and the talc zones, on the other hand, appear to have been derived directly from the dunite host, mainly by the addition of SiO₂ and loss of MgO and FeO^T. The relative loss of Ni in these metasomatic zones in comparison to the dunite precursor, likely reflects the breakdown of olivine and thus liberation of the compatible Ni hosted by this mineral. In contrast, Cr is enriched, which appears to be related to the formation of chromite-magnetite solid solution by an immobilization of part of FeO^T from the dissolved olivine.

To quantify these compositional changes, we carried out isocon mass balance modelling [47,48] of the relevant sample pairs. It should be noted, however, that such modelling inherently has large uncertainties and only work under the assumption that the starting composition were identical for the samples that are being compared. Using TiO₂, Al₂O₃, and Y for the immobile reference line we find that the talc zone represents a net mass gain of about 37%, which was dominated by the addition of SiO₂, K₂O, Rb, Cs, as well as U, Th, and Cr, whereas MgO, FeO^T, and CaO were lost together with Ni and Sc.

In contrast, the amphibole zone represents a net mass loss of about −50% in comparison with the dunite precursor. The losses in this zone were dominated by the depletion in MgO, CaO, SiO₂ and FeO^T (in that order), despite relative gains in MnO, K₂O, Rb, Cs, and U.

Using only TiO₂ and Al₂O₃ to constrain the isocon regression line for the pegmatite and mica zone alteration pair, we found no meaningful solution. This is consistent with the extreme enrichment observed for immobile elements like TiO₂, Zr and the heavy rare earth elements.

The calculated mass changes from the dunite to the talc and amphibole zones are consistent with the chemical gradient of silica between a felsic pegmatite and the dunite host, which would promote the diffusion of SiO₂ from the amphibole zone into the talc zone. Both metasomatic zones experienced enrichment in K₂O and associated large ion lithophile elements. Interestingly, U was also enriched in both zones, which point to relatively oxidizing fluids that were able to move uranyl-complexes from the granitoid into the ultramafic host rock.

Regarding that both dunite and trondhjemite at Seqi are “dry” rocks, the metasomatic zone situated on their contact is most probably the fluid residual of the trondhjemite sheet crystallization. Its mineralogical characteristic is a transition from a phlogopite-dominated zone on the trondhjemite side to a talc-dominated one on the dunite side, with a specific thin middle zone formed of almost pure anthophyllite with oriented growth of crystals suggesting that it formed during the last stage of the process, filling the remaining narrow fracture. We noted that the observed characteristic mineralogy can be understood from the action of a silica rich fluid expelled from a crystallizing felsic melt on the dunite. Its reaction with dunite resulted in dissolution of olivine and crystallization of talc, with a net enrichment of residual fluid in Mg and Fe. This enrichment enabled nucleation of phlogopite on the trondhjemite side and, subsequently, with a decrease in water content, a crystallization of contemporary and ultimately the sole anthophyllite in the middle of the metasomatic zone. The gradients of some characteristic major elements in the metasomatic zone can be understood from this mineral result. Iron is incorporated in talc only in relatively small amounts, whereas both phlogopite and anthophyllite form solid solutions with corresponding Fe-rich end-members and readily took Fe released from dunite dissolution. Like Fe, anthophyllite readily incorporates Mn, as reflected in its gradient. Among the three main minerals of the metasomatic zone, phlogopite is the only host for Al and K, which explains their concentration in the mica zone. An interesting detail is a concentration of Ca in the metasomatic zone, resulting in crystallization of tremolite, rather than anthophyllite, close to the trondhjemite side. At the same time, Na is mostly absent from the metasomatic zone, obviously fully absorbed in the highly albitic plagioclase in trondhjemite.

6. Conclusions

The investigated metasomatic zone at the Archean Seqi Ultramafic Complex, SW Greenland, consists of three distinct mineral zones dominated by (1) talc, (2) anthophyllite, and (3) phlogopite, which were developed at the contact of an intrusive trondhjemite sheet in dunite. The following conclusions can be made based in the presented data:

(1) A zircon crystal inclusion in a large (4 cm) olivine porphyroblast was dated by in situ U–Pb isotope analysis, yielding a weighted mean ²⁰⁷Pb/²⁰⁶Pb age of 2963 ± 1 Ma, demonstrating dehydration and recrystallization of the ultramafic rocks at that time. In contrast, the 2940 ± 5 Ma zircon age of the intrusive pegmatite was likely coincident with the development of the studied metasomatic talc zone at the Seqi Olivine Mine.

(2) Considering phase relations appropriate for the dunite bulk-rock composition, we deduced the talc forming conditions to be at temperatures of 600–650 °C and at a pressure below 1 GPa.

(3) Triple oxygen isotope data for talc, anthophyllite, and phlogopite in the metasomatic zones indicate formation in relation to fluids with a δ¹⁸O of ~8‰ and a Δ’¹⁷O_0.528 of about −40 ppm at a temperature range of 600–700 °C. This conclusion supports a formation scenario involving fluids that were derived from the late stage pegmatitic trondhjemite intrusion.

(4) Absolute element mobility in the metasomatic zone points to a net mass gain of about 37% for the talc zone, involving the addition of SiO₂, K₂O, Rb, Cs, as well as U, Th, and Cr, whereas MgO, FeO^T, and CaO were lost together with Ni and Sc. In contrast, the amphibole zone had a net mass loss of about −50% in comparison with the dunite precursor. These losses were dominated by the depletion in MgO, CaO, SiO_2, and FeO^T, and relative gains in MnO, K₂O, Rb, Cs, and U. Thus, the overall mass transfer demonstrates uptake and transport of SiO₂, large ion lithophile elements (LILE) and K₂O into the dunite body, whereas MgO and FeO^T were lost during this metasomatic process. The mobility of U indicates relative oxidizing fluid conditions during talc formation in contrast to the strongly reduced fluids that characterize serpentinization processes.