Figures

top

Introduction

Earth is the only terrestrial planet with voluminous felsic continental crust. The formation of early Archean crust is a fundamental issue for understanding planetary differentiation and habitability. Water is widely regarded as a decisive factor during the generation and evolution of silicic rocks (e.g., tonalite-trondhjemite-granodiorite, TTG). Water significantly lowers rock solidus temperature, reduces melt viscosity, enhances melt volume, thereby influencing the composition of residual mineral assemblages, melt compositions and differentiation pathways (Collins et al., 2020

Collins, W.J., Murphy, J.B., Johnson, T.E., Huang, H.Q. (2020) Critical role of water in the formation of continental crust. Nature Geoscience 13, 331–338. https://doi.org/10.1038/s41561-020-0573-6

). Based on the mode of water involvement during melting, previous studies have proposed two fundamental end member models with significant tectonic implications: fluid-present and fluid-absent melting (Beard and Lofgren, 1991

Beard, J.S., Lofgren, G.E. (1991) Dehydration Melting and Water-Saturated Melting of Basaltic and Andesitic Greenstones and Amphibolites at 1, 3, and 6.9 kb. Journal of Petrology 32, 365–401. https://doi.org/10.1093/petrology/32.2.365

; Pourteau et al., 2020

Pourteau, A., Doucet, L.S., Blereau, E.R., Volante, S., Johnson, T.E., Collins, W.J., Li, Z.X., Champion, D.C. (2020) TTG generation by fluid-fluxed crustal melting: Direct evidence from the Proterozoic Georgetown Inlier, NE Australia. Earth and Planetary Science Letters 550, 116548. https://doi.org/10.1016/j.epsl.2020.116548

; Tamblyn et al., 2023

Tamblyn, R., Hermann, J., Hasterok, D., Sossi, P., Pettke, T., Chatterjee, S. (2023) Hydrated komatiites as a source of water for TTG formation in the Archean. Earth and Planetary Science Letters 603, 117982. https://doi.org/10.1016/j.epsl.2022.117982

). The former typically involves the influx of external free fluids, enabling melting at relatively low temperatures and is commonly invoked to support subduction or water-rich tectonic environments. In contrast, the latter primarily occurs in the absence of a free fluid phase and is driven by heating or decompression that induces the breakdown of hydrous minerals, typically associated with thickened lower crust or a mafic plateau setting. It is suggested that hydrous mantle plateaus occurred occasionally in the Phanerozoic (Liu et al., 2017

Liu, J., Xia, Q.K., Kuritani, T., Hanski, E., Yu, H.R. (2017) Mantle hydration and the role of water in the generation of large igneous provinces. Nature Communications 8, 1824. https://doi.org/10.1038/s41467-017-01940-3

). Distinguishing between those two melting mechanisms is not only a fundamental petrological issue but also central to evaluate whether water-rich conditions within mafic plateaus existed in the Archean (Roman and Arndt, 2020

Roman, A., Arndt, N. (2020) Differentiated Archean oceanic crust: Its thermal structure, mechanical stability and a test of the sagduction hypothesis. Geochimica et Cosmochimica Acta 278, 65–77. https://doi.org/10.1016/j.gca.2019.07.009

; Smithies et al., 2021

Smithies, R.H., Lu, Y., Kirkland, C.L., Johnson, T.E., Mole, D.R., Champion, D.C., et al (2021) Oxygen isotopes trace the origins of Earth’s earliest continental crust. Nature 592, 70–75. https://doi.org/10.1038/s41586-021-03337-1

).

However, precisely identifying the specific mechanisms of partial melting remains challenging. On the one hand, traditional thermobarometry highly depends on pressure, melt composition, water activity, and equilibrium state, making it difficult to serve as a direct indicator for fluid-present melting (Schwindinger et al., 2019

Schwindinger, M., Weinberg, R.F., Clos, F. (2019) Wet or dry? The difficulty of identifying the presence of water during crustal melting. Journal of Metamorphic Geology 37, 339–358. https://doi.org/10.1111/jmg.12465

). On the other hand, Archean felsic terranes commonly underwent amphibolite to granulite facies metamorphic overprinting, obscuring primary magmatic signatures. Because the stability of plagioclase is highly sensitive to the water content of melts, fluid-present melting suppresses plagioclase stability, whereas fluid-absent melting stabilises plagioclase (Beard and Lofgren, 1991

Beard, J.S., Lofgren, G.E. (1991) Dehydration Melting and Water-Saturated Melting of Basaltic and Andesitic Greenstones and Amphibolites at 1, 3, and 6.9 kb. Journal of Petrology 32, 365–401. https://doi.org/10.1093/petrology/32.2.365

; Müntener and Ulmer, 2018

Müntener, O., Ulmer, P. (2018) Arc crust formation and differentiation constrained by experimental petrology. American Journal of Science 318, 64–89. https://doi.org/10.2475/01.2018.04

). Plagioclase is the primary reservoir of Sr in magmatic systems and preferentially incorporates heavy Sr isotopes relative to the bulk rock, such that its fractionation or residue produces resolvable stable Sr isotope fractionation (Charlier et al., 2012

Charlier, B.L.A., Nowell, G.M., Parkinson, I.J., Kelley, S.P., Pearson, D.G., Burton, K.W. (2012) High temperature strontium stable isotope behaviour in the early solar system and planetary bodies. Earth and Planetary Science Letters 329-330, 31–40. https://doi.org/10.1016/j.epsl.2012.02.008

; Klaver et al., 2020

Klaver, M., Lewis, J., Parkinson, I.J., Elburg, M.A., Vroon, P.Z., Kelley, K.A., Elliott, T. (2020) Sr isotopes in arcs revisited: tracking slab dehydration using δ^88/86Sr and ⁸⁷Sr/⁸⁶Sr systematics of arc lavas. Geochimica et Cosmochimica Acta 288, 101–119. https://doi.org/10.1016/j.gca.2020.08.010

; Chen et al., 2025b

Chen, X.Q., Deng, G.X., Jiang, D.S., Nan, X.Y., Huang, F. (2025b) Stable strontium isotope fractionation during crystal-melt separation in granitic magma evolution. Acta Geochimica 44, 731–739. https://doi.org/10.1007/s11631-024-00752-9

). This highlights stable Sr isotope (δ⁸⁸Sr) as a potentially powerful tracer of plagioclase behaviour and melting mechanism.

The Kaapvaal Craton is one of the classic regions for studying the formation of Archean continental crust. The ∼3.45 Ga Tsawela Gneiss (TG) from the Ancient Gneiss Complex (AGC) (Fig. S-1) is mainly composed of tonalite to trondhjemite (Fig. S-2a), providing a critical window into the debate over fluid-present versus fluid-absent melting. The TG exhibits calc-alkaline trends (Fig. S-2b), moderate Sr contents, and low Sr/Y ratios, consistent with a low pressure TTG origin (Fig. S-3). The TG displays weak deformation, lack migmatisation features, and has low loss on ignition (LOI) values, indicating relatively limited later modification (Hoffmann et al., 2016

Hoffmann, J.E., Kröner, A., Hegner, E., Viehmann, S., Xie, H., et al (2016) Source composition, fractional crystallization and magma mixing processes in the 3.48–3.43 Ga Tsawela tonalite suite (Ancient Gneiss Complex, Swaziland) – Implications for Palaeoarchaean geodynamics. Precambrian Research 276, 43–66. https://doi.org/10.1016/j.precamres.2016.01.026

). The source of TG is close to depleted mantle and lacks clear evidence for the involvement of ancient continental crustal materials (Hoffmann et al., 2016

Hoffmann, J.E., Kröner, A., Hegner, E., Viehmann, S., Xie, H., et al (2016) Source composition, fractional crystallization and magma mixing processes in the 3.48–3.43 Ga Tsawela tonalite suite (Ancient Gneiss Complex, Swaziland) – Implications for Palaeoarchaean geodynamics. Precambrian Research 276, 43–66. https://doi.org/10.1016/j.precamres.2016.01.026

). Based on rock assemblages, geochemistry and palaeomagnetic detection, these rocks are commonly interpreted to have formed in a mafic plateau tectonic setting during the early Archean (Hoffmann et al., 2016

Hoffmann, J.E., Kröner, A., Hegner, E., Viehmann, S., Xie, H., et al (2016) Source composition, fractional crystallization and magma mixing processes in the 3.48–3.43 Ga Tsawela tonalite suite (Ancient Gneiss Complex, Swaziland) – Implications for Palaeoarchaean geodynamics. Precambrian Research 276, 43–66. https://doi.org/10.1016/j.precamres.2016.01.026

; Brenner et al., 2026

Brenner, A.R., Fu, R.R., Foley, B.J., Lourenço, D.L., Palma-Gomez, J., et al (2026) Paleomagnetic detection of relative plate motions and an infrequently reversing core dynamo at 3.5 Ga. Science 391, 1278–1282. https://doi.org/10.1126/science.adw9250

).

Here, we investigate the ∼3.45 Ga TG exposed in the Kaapvaal Craton for high precision stable Sr isotopic analysis to distinguish the melting mechanisms. By integrating whole rock geochemistry and modelling, we aim to (1) evaluate plagioclase behaviour during TG petrogenesis, and (2) test the reliability of stable Sr isotopes in distinguishing fluid-present versus fluid-absent melting during ∼3.45 Ga TG formation.

top

Geological Background

Background. The Kaapvaal Craton preserves some of Earth’s representative Archean crustal fragments, comprising the AGC in Eswatini and the 3.6–3.2 Ga Barberton granitoid-greenstone terrane (BGGT) in South Africa (Hoffmann et al., 2016

Hoffmann, J.E., Kröner, A., Hegner, E., Viehmann, S., Xie, H., et al (2016) Source composition, fractional crystallization and magma mixing processes in the 3.48–3.43 Ga Tsawela tonalite suite (Ancient Gneiss Complex, Swaziland) – Implications for Palaeoarchaean geodynamics. Precambrian Research 276, 43–66. https://doi.org/10.1016/j.precamres.2016.01.026

). The AGC is primarily composed of the 3.66–3.2 Ga Ngwane Gneiss (NG), the 3.47–3.42 Ga TG and ∼3.46 Ga Dwalile Supracrustal Suite (DSS) (Hoffmann et al., 2016

Hoffmann, J.E., Kröner, A., Hegner, E., Viehmann, S., Xie, H., et al (2016) Source composition, fractional crystallization and magma mixing processes in the 3.48–3.43 Ga Tsawela tonalite suite (Ancient Gneiss Complex, Swaziland) – Implications for Palaeoarchaean geodynamics. Precambrian Research 276, 43–66. https://doi.org/10.1016/j.precamres.2016.01.026

). The NG exhibits a bimodal character with interlayered felsic and mafic components. It is intensely deformed and experienced upper amphibolite to granulite facies metamorphism, and represents the oldest basement component of the AGC. The TG is a weakly to well foliated TTG suite (Fig. S-2a) and intruded the older NG and Dwalile Greenstone remnants (Hoffmann et al., 2016

Hoffmann, J.E., Kröner, A., Hegner, E., Viehmann, S., Xie, H., et al (2016) Source composition, fractional crystallization and magma mixing processes in the 3.48–3.43 Ga Tsawela tonalite suite (Ancient Gneiss Complex, Swaziland) – Implications for Palaeoarchaean geodynamics. Precambrian Research 276, 43–66. https://doi.org/10.1016/j.precamres.2016.01.026

). Within the TG, SiO₂ contents range from 58.67 to 71.48 wt. %. They originated from a juvenile source supported by Nd and Hf isotopic compositions (Zeh et al., 2011

Zeh, A., Gerdes, A., Millonig, L. (2011) Hafnium isotope record of the Ancient Gneiss Complex, Swaziland, southern Africa: evidence for Archaean crust-mantle formation and crust reworking between 3.66 and 2.73 Ga. Journal of the Geological Society 168, 953–963. https://doi.org/10.1144/0016-76492010-117

; Hoffmann et al., 2016

Hoffmann, J.E., Kröner, A., Hegner, E., Viehmann, S., Xie, H., et al (2016) Source composition, fractional crystallization and magma mixing processes in the 3.48–3.43 Ga Tsawela tonalite suite (Ancient Gneiss Complex, Swaziland) – Implications for Palaeoarchaean geodynamics. Precambrian Research 276, 43–66. https://doi.org/10.1016/j.precamres.2016.01.026

). Importantly, the low initial ⁸⁷Sr/⁸⁶Sr ratios reported for the AGC (Davies and Allsopp, 1976

Davies, R.D., Allsopp, H.L. (1976) Strontium isotopic evidence relating to the evolution of the lower Precambrian granitic crust in Swaziland. Geology 4, 553–556. https://doi.org/10.1130/0091-7613(1976)4<553:SIERTT>2.0.CO;2

) suggest negligible involvement of pre-existing siliceous crust. Interpretation of major and trace element proxies is complicated by source heterogeneity, fractional crystallisation, and melting depth (Kendrick and Yakymchuk, 2020

Kendrick, J., Yakymchuk, C. (2020) Garnet fractionation, progressive melt loss and bulk composition variations in anatectic metabasites: Complications for interpreting the geodynamic significance of TTGs. Geoscience Frontiers 11, 745–763. https://doi.org/10.1016/j.gsf.2019.12.001

). We integrate stable Sr isotopes with modelling to constrain the petrogenetic regime of the TG suite.

Samples and methods. Six TG and one NG samples were analysed; sample location (Fig. S-1), petrography, major and trace element data, U-Pb zircon ages, and Hf-Nd isotopes were reported in Hoffmann et al. (2016)

Hoffmann, J.E., Kröner, A., Hegner, E., Viehmann, S., Xie, H., et al (2016) Source composition, fractional crystallization and magma mixing processes in the 3.48–3.43 Ga Tsawela tonalite suite (Ancient Gneiss Complex, Swaziland) – Implications for Palaeoarchaean geodynamics. Precambrian Research 276, 43–66. https://doi.org/10.1016/j.precamres.2016.01.026

. TG samples are medium to coarse grained tonalitic gneisses with weak deformation; sample AGC502T exhibits more pronounced fabric development. The studied NG sample (AGC470) is a strongly deformed migmatitic orthogneiss.

Stable and radiogenic Sr isotope measurements were performed at the State Key Laboratory of Lithospheric and Environmental Coevolution, University of Science and Technology of China, following the protocol of Chen et al. (2022)

Chen, X.Q., Zeng, Z., Yu, H.M., Sun, N., Huang, F. (2022) Precise measurements of δ^88/86Sr for twenty geological reference materials by double-spike MC-ICP-MS. International Journal of Mass Spectrometry 479, 116883. https://doi.org/10.1016/j.ijms.2022.116883

. Long term external reproducibility was 0.03 ‰ (2 s.d.) for δ⁸⁸Sr and 0.000025 (2 s.d.) for ⁸⁷Sr/⁸⁶Sr. Procedural blanks were <0.18 ng Sr. Initial ⁸⁷Sr/⁸⁶Sr ratios were calculated at 3.45 Ga. Results for international standards and replicate analyses (Table S-1) are in agreement with certified values, ensuring data robustness. For the TG samples, replicate analyses show differences of ⁸⁷Sr/⁸⁶Sr but a consistent δ⁸⁸Sr value, which were attributed to the slight heterogeneity of the Rb/Sr ratio among the different mineral phases within the bulk rock powders, while δ⁸⁸Sr remains homogeneous at the hand specimen scale (Chen et al., 2025b

Chen, X.Q., Deng, G.X., Jiang, D.S., Nan, X.Y., Huang, F. (2025b) Stable strontium isotope fractionation during crystal-melt separation in granitic magma evolution. Acta Geochimica 44, 731–739. https://doi.org/10.1007/s11631-024-00752-9

).

top

Results

The δ⁸⁸Sr and ⁸⁷Sr/⁸⁶Sr ratios for the AGC gneisses are summarised in Table S-1. TG samples exhibit ⁸⁷Sr/⁸⁶Sr ratios of 0.723139–0.773669 and initial (⁸⁷Sr/⁸⁶Sr)_t values of 0.69799–0.71722. Their δ⁸⁸Sr values range from 0.19 ‰ to 0.30 ‰ (2 s.d. < 0.04 ‰), with a mean of 0.25 ± 0.07 ‰ (2 s.d.) (Figs. 1, S-4), largely overlapping with the reported Bulk Silicate Earth (BSE) values (0.27 ± 0.05 ‰, Moynier et al., 2010

Moynier, F., Agranier, A., Hezel, D.C., Bouvier, A. (2010) Sr stable isotope composition of Earth, the Moon, Mars, Vesta and meteorites. Earth and Planetary Science Letters 300, 359–366. https://doi.org/10.1016/j.epsl.2010.10.017

; 0.29 ± 0.07 ‰, Charlier et al., 2012

Charlier, B.L.A., Nowell, G.M., Parkinson, I.J., Kelley, S.P., Pearson, D.G., Burton, K.W. (2012) High temperature strontium stable isotope behaviour in the early solar system and planetary bodies. Earth and Planetary Science Letters 329-330, 31–40. https://doi.org/10.1016/j.epsl.2012.02.008

; 0.30 ± 0.02 ‰, Amsellem et al., 2018

Amsellem, E., Moynier, F., Day, J.M.D., Moreira, M., Puchtel, I.S., Teng, F.Z. (2018) The stable strontium isotopic composition of ocean island basalts, mid-ocean ridge basalts, and komatiites. Chemical Geology 483, 595–602. https://doi.org/10.1016/j.chemgeo.2018.03.030

). Compared with the wide range for granitoids δ⁸⁸Sr (−1.51 ‰ to 0.54 ‰) (Chen et al., 2025a

Chen, X.Q., Quan, Y., Liu, X.C., Deng, G., Nan, X., Huang, F. (2025a) Stable strontium isotope fractionation by fluid-melt interaction recorded in the Himalayan leucogranites. Chemical Geology 690, 122869. https://doi.org/10.1016/j.chemgeo.2025.122869

, 2025b

Chen, X.Q., Deng, G.X., Jiang, D.S., Nan, X.Y., Huang, F. (2025b) Stable strontium isotope fractionation during crystal-melt separation in granitic magma evolution. Acta Geochimica 44, 731–739. https://doi.org/10.1007/s11631-024-00752-9

) (Fig. S-4), the stable Sr isotope fractionation in the TG sample is negligible (Fig. 1).

Full size image

The NG sample shows δ⁸⁸Sr = 0.10 ± 0.02 ‰, significantly lighter than both TG samples and BSE. Its ⁸⁷Sr/⁸⁶Sr ratio is 0.758083, with a calculated initial (⁸⁷Sr/⁸⁶Sr)_t value of 0.70597.

top

Discussion

What controls δ⁸⁸Sr in the early Archean felsic magmas? Available data for Archean mantle derived rocks indicate that the stable Sr isotopic composition of the convecting mantle has remained constant since the Paleoarchean (Amsellem et al., 2018

Amsellem, E., Moynier, F., Day, J.M.D., Moreira, M., Puchtel, I.S., Teng, F.Z. (2018) The stable strontium isotopic composition of ocean island basalts, mid-ocean ridge basalts, and komatiites. Chemical Geology 483, 595–602. https://doi.org/10.1016/j.chemgeo.2018.03.030

). Importantly, mafic igneous processes have limited effect on stable Sr isotope fractionation (Amsellem et al., 2018

Amsellem, E., Moynier, F., Day, J.M.D., Moreira, M., Puchtel, I.S., Teng, F.Z. (2018) The stable strontium isotopic composition of ocean island basalts, mid-ocean ridge basalts, and komatiites. Chemical Geology 483, 595–602. https://doi.org/10.1016/j.chemgeo.2018.03.030

). Stable Sr isotopes in felsic magmas are primarily governed by feldspar and apatite, either through crystallisation or as residual phase. Plagioclase and apatite preferentially incorporate heavy Sr isotopes, while K feldspar favours light isotopes (Charlier et al., 2012

Charlier, B.L.A., Nowell, G.M., Parkinson, I.J., Kelley, S.P., Pearson, D.G., Burton, K.W. (2012) High temperature strontium stable isotope behaviour in the early solar system and planetary bodies. Earth and Planetary Science Letters 329-330, 31–40. https://doi.org/10.1016/j.epsl.2012.02.008

; Andrews and Jacobson, 2018

Andrews, M.G., Jacobson, A.D. (2018) Controls on the solute geochemistry of subglacial discharge from the Russell Glacier, Greenland Ice Sheet determined by radiogenic and stable Sr isotope ratios. Geochimica et Cosmochimica Acta 239, 312–329. https://doi.org/10.1016/j.gca.2018.08.004

; Klaver et al., 2020

Klaver, M., Lewis, J., Parkinson, I.J., Elburg, M.A., Vroon, P.Z., Kelley, K.A., Elliott, T. (2020) Sr isotopes in arcs revisited: tracking slab dehydration using δ^88/86Sr and ⁸⁷Sr/⁸⁶Sr systematics of arc lavas. Geochimica et Cosmochimica Acta 288, 101–119. https://doi.org/10.1016/j.gca.2020.08.010

; Chen et al., 2025b

Chen, X.Q., Deng, G.X., Jiang, D.S., Nan, X.Y., Huang, F. (2025b) Stable strontium isotope fractionation during crystal-melt separation in granitic magma evolution. Acta Geochimica 44, 731–739. https://doi.org/10.1007/s11631-024-00752-9

). In this study, the TG exhibits an average δ⁸⁸Sr of 0.25 ± 0.07 ‰, consistent with the mantle/BSE composition (Moynier et al., 2010

Moynier, F., Agranier, A., Hezel, D.C., Bouvier, A. (2010) Sr stable isotope composition of Earth, the Moon, Mars, Vesta and meteorites. Earth and Planetary Science Letters 300, 359–366. https://doi.org/10.1016/j.epsl.2010.10.017

), suggesting that plagioclase was a negligible residual phase in the melting source. The lack of correlations between δ⁸⁸Sr and SiO₂, Sr contents, Eu/Eu* (Fig. 1) suggests negligible plagioclase, K feldspar and apatite fractionation. In addition, Rayleigh fractionation modelling (Fig. 2) suggests that 10 wt. % plagioclase fractionation would decrease the δ⁸⁸Sr value of whole rock to 0.14 ‰ (for plagioclase-melt fractionation factor α_Pl-Melt = 1.0003 from Chen et al., 2025b

Chen, X.Q., Deng, G.X., Jiang, D.S., Nan, X.Y., Huang, F. (2025b) Stable strontium isotope fractionation during crystal-melt separation in granitic magma evolution. Acta Geochimica 44, 731–739. https://doi.org/10.1007/s11631-024-00752-9

) and 0.00 ‰ (for α_Pl-Melt = 1.0007 from Charlier et al., 2012

Charlier, B.L.A., Nowell, G.M., Parkinson, I.J., Kelley, S.P., Pearson, D.G., Burton, K.W. (2012) High temperature strontium stable isotope behaviour in the early solar system and planetary bodies. Earth and Planetary Science Letters 329-330, 31–40. https://doi.org/10.1016/j.epsl.2012.02.008

). This demonstrates that δ⁸⁸Sr is sufficiently sensitive to detect even a small amount of plagioclase crystallisation. However, this deviation is not observed in the TG suite. Also, the coarse grained texture of the TG precludes rapid cooling as a potential mechanism for suppressed plagioclase fractionation.

Full size image

Secondary processes are unlikely to account for the restricted δ⁸⁸Sr values observed in the TG samples. First, the TG samples are fresh, as indicated by low LOI values, and no correlation exists between LOI and δ⁸⁸Sr (Fig. S-5). Second, the major Sr hosting minerals (feldspar, apatite) are relatively stable during amphibolite facies metamorphism. The narrow δ⁸⁸Sr range despite variable initial Sr isotopes (Fig. S-5b) and bulk compositions argue against substantial metamorphic resetting or crustal contamination. Previous hybridisation models require mixing between mafic and felsic end members (Hoffmann et al., 2016

Hoffmann, J.E., Kröner, A., Hegner, E., Viehmann, S., Xie, H., et al (2016) Source composition, fractional crystallization and magma mixing processes in the 3.48–3.43 Ga Tsawela tonalite suite (Ancient Gneiss Complex, Swaziland) – Implications for Palaeoarchaean geodynamics. Precambrian Research 276, 43–66. https://doi.org/10.1016/j.precamres.2016.01.026

); however, reproducing the observed Sr concentrations through such a process would require the mafic component to contain an extremely high Sr content (800 ppm), which is inconsistent with typical mafic melts (Fig. S-10b). Moreover, magma mixing between isotopically distinct end members would be expected to generate resolvable δ⁸⁸Sr heterogeneity (Fig. S-10a), which is not observed. Therefore, the narrow δ⁸⁸Sr range of the TG samples likely reflects their primary magmatic compositions, consistent with the preservation of their relatively low initial Sr isotope compositions (Davies and Allsopp, 1976

Davies, R.D., Allsopp, H.L. (1976) Strontium isotopic evidence relating to the evolution of the lower Precambrian granitic crust in Swaziland. Geology 4, 553–556. https://doi.org/10.1130/0091-7613(1976)4<553:SIERTT>2.0.CO;2

; this study).

However, the Sr content of TG samples decreases with increasing SiO₂ (Fig. S-5a) (Hoffmann et al., 2016

Hoffmann, J.E., Kröner, A., Hegner, E., Viehmann, S., Xie, H., et al (2016) Source composition, fractional crystallization and magma mixing processes in the 3.48–3.43 Ga Tsawela tonalite suite (Ancient Gneiss Complex, Swaziland) – Implications for Palaeoarchaean geodynamics. Precambrian Research 276, 43–66. https://doi.org/10.1016/j.precamres.2016.01.026

). This trend may indicate heterogeneous Sr contents in their mafic precursors, consistent with the observed heterogeneity in whole rock Hf-Nd and initial (⁸⁷Sr/⁸⁶Sr)_t values (Davies and Allsopp, 1976

Davies, R.D., Allsopp, H.L. (1976) Strontium isotopic evidence relating to the evolution of the lower Precambrian granitic crust in Swaziland. Geology 4, 553–556. https://doi.org/10.1130/0091-7613(1976)4<553:SIERTT>2.0.CO;2

; Fig. S-6) (Hoffmann et al., 2016

Hoffmann, J.E., Kröner, A., Hegner, E., Viehmann, S., Xie, H., et al (2016) Source composition, fractional crystallization and magma mixing processes in the 3.48–3.43 Ga Tsawela tonalite suite (Ancient Gneiss Complex, Swaziland) – Implications for Palaeoarchaean geodynamics. Precambrian Research 276, 43–66. https://doi.org/10.1016/j.precamres.2016.01.026

). This interpretation is further supported by the wide variation in Sr contents (∼25–250 ppm), Sr/Y (1.2–9.8), and Eu/Eu* (0.78–1.02) of the ∼3.46 Ga DSS greenstone remnants (Fig. S-7) (Hoffmann et al., 2020

Hoffmann, J.E., Musese, E., Kröner, A., Schneider, K.P., Wong, J., et al (2020) Hafnium-Neodymium isotope, trace element and U-Pb zircon age constraints on the petrogenesis of the 3.44–3.46 Ga Dwalile greenstone remnant, Ancient gneiss Complex, Swaziland. Precambrian Research 351, 105970. https://doi.org/10.1016/j.precamres.2020.105970

).

Fluid-present melting and tectonic implications. A mafic plateau setting has been proposed as a viable tectonic environment for the formation of voluminous felsic crust during the early Archean (Johnson et al., 2017

Johnson, T.E., Brown, M., Gardiner, N.J., Kirkland, C.L., Smithies, R.H. (2017) Earth’s first stable continents did not form by subduction. Nature 543, 239–242. https://doi.org/10.1038/nature21383

). However, this view has been challenged by the argument that the lower crust of thick mafic plateaus is predominantly dry, thereby limiting the production of large volumes of silicic melt (Roman and Arndt, 2020

Roman, A., Arndt, N. (2020) Differentiated Archean oceanic crust: Its thermal structure, mechanical stability and a test of the sagduction hypothesis. Geochimica et Cosmochimica Acta 278, 65–77. https://doi.org/10.1016/j.gca.2019.07.009

). Resolving the presence or absence of fluids in mafic plateaus is therefore critical for understanding the mechanisms of early continental crust formation and differentiation.

Thermochemically, magmatic water reduces melt polymerisation by breaking Si-O-Si bonds (Stolper, 1982

Stolper, E. (1982) The speciation of water in silicate melts. Geochimica et Cosmochimica Acta 46, 2609–2620. https://doi.org/10.1016/0016-7037(82)90381-7

), which significantly depresses the plagioclase liquidus while expanding the stability field of amphibole (Beard and Lofgren, 1991

Beard, J.S., Lofgren, G.E. (1991) Dehydration Melting and Water-Saturated Melting of Basaltic and Andesitic Greenstones and Amphibolites at 1, 3, and 6.9 kb. Journal of Petrology 32, 365–401. https://doi.org/10.1093/petrology/32.2.365

; Pichavant and Macdonald, 2007

Pichavant, M., Macdonald, R. (2007) Crystallization of primitive basaltic magmas at crustal pressures and genesis of the calc-alkaline igneous suite: experimental evidence from St Vincent, Lesser Antilles arc. Contributions to Mineralogy and Petrology 154, 535–558. https://doi.org/10.1007/s00410-007-0208-6

; Müntener and Ulmer, 2018

Müntener, O., Ulmer, P. (2018) Arc crust formation and differentiation constrained by experimental petrology. American Journal of Science 318, 64–89. https://doi.org/10.2475/01.2018.04

). The resulting melts are consequently enriched in Al₂O₃, follow calc-alkaline differentiation trends, exhibit low Sr/Y ratios (Beard and Lofgren, 1991

Beard, J.S., Lofgren, G.E. (1991) Dehydration Melting and Water-Saturated Melting of Basaltic and Andesitic Greenstones and Amphibolites at 1, 3, and 6.9 kb. Journal of Petrology 32, 365–401. https://doi.org/10.1093/petrology/32.2.365

; Pichavant and Macdonald, 2007

Pichavant, M., Macdonald, R. (2007) Crystallization of primitive basaltic magmas at crustal pressures and genesis of the calc-alkaline igneous suite: experimental evidence from St Vincent, Lesser Antilles arc. Contributions to Mineralogy and Petrology 154, 535–558. https://doi.org/10.1007/s00410-007-0208-6

; Müntener and Ulmer, 2018

Müntener, O., Ulmer, P. (2018) Arc crust formation and differentiation constrained by experimental petrology. American Journal of Science 318, 64–89. https://doi.org/10.2475/01.2018.04

) and most importantly, preserve δ⁸⁸Sr values indistinguishable from mantle compositions because plagioclase, the primary Sr-rich phase that preferentially incorporates heavy Sr, is largely absent from the fractional crystallisation assemblages. The restricted, mantle-like δ⁸⁸Sr values of the TG samples require minimal isotopic fractionation, which in turn demands suppression of plagioclase crystallisation. While high pressure (>1.5 GPa) can also destabilise plagioclase, the TG samples are characterised by low pressure TTG signatures (Fig. S-3), ruling out a deep origin. We therefore ascribe the suppression of plagioclase crystallisation to fluid-present melting. This interpretation is independently supported by their calc-alkaline differentiation trend (Hoffmann et al., 2016

Hoffmann, J.E., Kröner, A., Hegner, E., Viehmann, S., Xie, H., et al (2016) Source composition, fractional crystallization and magma mixing processes in the 3.48–3.43 Ga Tsawela tonalite suite (Ancient Gneiss Complex, Swaziland) – Implications for Palaeoarchaean geodynamics. Precambrian Research 276, 43–66. https://doi.org/10.1016/j.precamres.2016.01.026

), low pressure TTG affinity (Fig. S-3; Pourteau et al., 2020

Pourteau, A., Doucet, L.S., Blereau, E.R., Volante, S., Johnson, T.E., Collins, W.J., Li, Z.X., Champion, D.C. (2020) TTG generation by fluid-fluxed crustal melting: Direct evidence from the Proterozoic Georgetown Inlier, NE Australia. Earth and Planetary Science Letters 550, 116548. https://doi.org/10.1016/j.epsl.2020.116548

) and no obvious Eu anomaly (Hoffmann et al., 2016

Hoffmann, J.E., Kröner, A., Hegner, E., Viehmann, S., Xie, H., et al (2016) Source composition, fractional crystallization and magma mixing processes in the 3.48–3.43 Ga Tsawela tonalite suite (Ancient Gneiss Complex, Swaziland) – Implications for Palaeoarchaean geodynamics. Precambrian Research 276, 43–66. https://doi.org/10.1016/j.precamres.2016.01.026

). In contrast, fluid-absent melting favours plagioclase crystallisation, producing melts with lighter δ⁸⁸Sr and tholeiitic differentiation trends (Müntener and Ulmer, 2018

Müntener, O., Ulmer, P. (2018) Arc crust formation and differentiation constrained by experimental petrology. American Journal of Science 318, 64–89. https://doi.org/10.2475/01.2018.04

).

Full size image

To quantitatively evaluate the impact of magmatic H₂O content on mineral crystallisation sequences during magma cooling, we performed thermodynamic forward modelling at 4–10 kbar using the bulk compositions of sample AGC502T (primitive proxy) and AGC500 (evolved). Phase equilibrium modelling (Figs. 3, S-8) quantitatively confirms these experimental relationships linking water content to the stability fields of plagioclase and amphibole (Beard and Lofgren, 1991

Beard, J.S., Lofgren, G.E. (1991) Dehydration Melting and Water-Saturated Melting of Basaltic and Andesitic Greenstones and Amphibolites at 1, 3, and 6.9 kb. Journal of Petrology 32, 365–401. https://doi.org/10.1093/petrology/32.2.365

; Müntener and Ulmer, 2018

Müntener, O., Ulmer, P. (2018) Arc crust formation and differentiation constrained by experimental petrology. American Journal of Science 318, 64–89. https://doi.org/10.2475/01.2018.04

). A progressive increase in melt H₂O content lowers the plagioclase liquidus from >1200 °C to ∼750 °C while simultaneously extending the stability field of amphibole to ∼1050 °C (Figs. 3, S-8). Crucially, the effect of water activity on depressing the plagioclase liquidus far exceeds that of temperature and pressure variations in the upper to lower crustal regime, serving as the principal control on plagioclase stability, independent of the bulk composition (Fig. 3a,b).

Plagioclase fractionation may be suspended due to insufficient gravitational settling in a viscous magma chamber, which would drive magma to higher Sr contents, δ⁸⁸Sr values and positive Eu anomalies in the cumulate portions. However, those features are absent in the TG suite (Figs. 1, S-5). Thus, plagioclase was neither a significant fractionating nor a cumulate-forming phase, consistent with suppressed plagioclase crystallisation under water-rich conditions.

These findings challenge the prevailing assumption that the lower crust of mafic plateaus is inherently water-poor (Roman and Arndt, 2020

Roman, A., Arndt, N. (2020) Differentiated Archean oceanic crust: Its thermal structure, mechanical stability and a test of the sagduction hypothesis. Geochimica et Cosmochimica Acta 278, 65–77. https://doi.org/10.1016/j.gca.2019.07.009

). Their “wet” geochemical signature suggests that fluid activity was pivotal in non-plate tectonic regimes. Water not only enhances melt productivity by depressing the solidus but also promotes calc-alkaline differentiation, facilitating rapid generation of voluminous felsic crust. Consequently, early continental growth may not require modern style subduction; instead, it could have occurred within hydrated mafic plateaus (Fig. 4).

The ultimate source of these fluids remains uncertain, but can be constrained. Seawater hydrothermal alteration is a potential source (Vezinet et al., 2018

Vezinet, A., Pearson, D.G., Thomassot, E., Stern, R.A., Sarkar, C., Luo, Y., Fisher, C.M. (2018) Hydrothermally-altered mafic crust as source for early Earth TTG: Pb/Hf/O isotope and trace element evidence in zircon from TTG of the Eoarchean Saglek Block, N. Labrador. Earth and Planetary Science Letters 503, 95–107. https://doi.org/10.1016/j.epsl.2018.09.015

); however, intensely altered oceanic crust is characterised by elevated δ⁸⁸Sr values (0.27–0.34 ‰), higher ⁸⁷Sr/⁸⁶Sr, and correlated enrichments in fluid mobile element ratios such as K/La and K/Th ratios (Klaver et al., 2020

Klaver, M., Lewis, J., Parkinson, I.J., Elburg, M.A., Vroon, P.Z., Kelley, K.A., Elliott, T. (2020) Sr isotopes in arcs revisited: tracking slab dehydration using δ^88/86Sr and ⁸⁷Sr/⁸⁶Sr systematics of arc lavas. Geochimica et Cosmochimica Acta 288, 101–119. https://doi.org/10.1016/j.gca.2020.08.010

; Liu and He, 2021

Liu, C.T., He, Y. (2021) Rise of major subaerial landmasses about 3.0 to 2.7 billion years ago. Geochemical Perspectives Letters 18, 1–5. https://doi.org/10.7185/geochemlet.2115

). In contrast, the TG samples display mantle-like stable Sr isotopic compositions and lack systematic correlations with K/La and K/Th ratios (Fig. S-9), suggesting that the involvement of hydrothermal alteration fluids was likely minimal. Altered komatiites within oceanic plateaus represent another potential fluid source through dehydration (Tamblyn et al., 2023

Tamblyn, R., Hermann, J., Hasterok, D., Sossi, P., Pettke, T., Chatterjee, S. (2023) Hydrated komatiites as a source of water for TTG formation in the Archean. Earth and Planetary Science Letters 603, 117982. https://doi.org/10.1016/j.epsl.2022.117982

). Although Barberton komatiites exhibit relatively heavy δ⁸⁸Sr values (0.16–0.97 ‰) (Amsellem et al., 2018

Amsellem, E., Moynier, F., Day, J.M.D., Moreira, M., Puchtel, I.S., Teng, F.Z. (2018) The stable strontium isotopic composition of ocean island basalts, mid-ocean ridge basalts, and komatiites. Chemical Geology 483, 595–602. https://doi.org/10.1016/j.chemgeo.2018.03.030

), the isotopic composition of the fluids released remains poorly constrained. By comparison, the low Sr content of komatiites (average Sr contents = 23 ppm; Chavagnac, 2004

Chavagnac, V. (2004) A geochemical and Nd isotopic study of Barberton komatiites (South Africa): implication for the Archean mantle. Lithos 75, 253–281. https://doi.org/10.1016/j.lithos.2004.03.001

) limits their ability to significantly modify the bulk Sr isotopic compositions of the mafic source. On the other hand, natural observations reveal that low temperature, slab derived fluids are enriched in light Sr isotopes (δ⁸⁸Sr = 0.122–0.157 ‰) (Kani et al., 2023

Kani, T., Misawa, K., Morikawa, N., Kazahaya, K., Kusuhara, F., Yoneda, S., Terakado, Y. (2023) Strontium Isotope Characteristics (δ^88/86Sr, ⁸⁷Sr/⁸⁶Sr) of Arima-Type Brines Originated from Slab-Fluids. Geophysical Research Letters 50. https://doi.org/10.1029/2022GL100309

), whereas high temperature fluids (>700 °C) tend to equilibrate with their source rocks, preserving a mantle-like isotopic signature (Klaver et al., 2020

Klaver, M., Lewis, J., Parkinson, I.J., Elburg, M.A., Vroon, P.Z., Kelley, K.A., Elliott, T. (2020) Sr isotopes in arcs revisited: tracking slab dehydration using δ^88/86Sr and ⁸⁷Sr/⁸⁶Sr systematics of arc lavas. Geochimica et Cosmochimica Acta 288, 101–119. https://doi.org/10.1016/j.gca.2020.08.010

). Therefore, the mantle-like δ⁸⁸Sr values exclude shallow, fractionated fluids but are consistent with high temperature fluid-rock equilibrium. Collectively, these observations suggest that the TG magmas were influenced by deep seated, high temperature fluids, partially derived from hydrated komatiites from the middle-lower crust (Tamblyn et al., 2023

Tamblyn, R., Hermann, J., Hasterok, D., Sossi, P., Pettke, T., Chatterjee, S. (2023) Hydrated komatiites as a source of water for TTG formation in the Archean. Earth and Planetary Science Letters 603, 117982. https://doi.org/10.1016/j.epsl.2022.117982

) and possibly from mantle plumes underplating the oldest continental fragments (Liu et al., 2017

Liu, J., Xia, Q.K., Kuritani, T., Hanski, E., Yu, H.R. (2017) Mantle hydration and the role of water in the generation of large igneous provinces. Nature Communications 8, 1824. https://doi.org/10.1038/s41467-017-01940-3

; Pourteau et al., 2020

Pourteau, A., Doucet, L.S., Blereau, E.R., Volante, S., Johnson, T.E., Collins, W.J., Li, Z.X., Champion, D.C. (2020) TTG generation by fluid-fluxed crustal melting: Direct evidence from the Proterozoic Georgetown Inlier, NE Australia. Earth and Planetary Science Letters 550, 116548. https://doi.org/10.1016/j.epsl.2020.116548

; Smithies et al., 2021

Smithies, R.H., Lu, Y., Kirkland, C.L., Johnson, T.E., Mole, D.R., Champion, D.C., et al (2021) Oxygen isotopes trace the origins of Earth’s earliest continental crust. Nature 592, 70–75. https://doi.org/10.1038/s41586-021-03337-1

), facilitating efficient crustal differentiation without the prerequisite of modern style plate subduction, although these sources cannot be uniquely distinguished.

Full size image

top

Conclusions

Stable Sr isotopes demonstrate that the ∼3.45 Ga Tsawela Gneiss formed through fluid-present melting of mafic crust. The mantle-like δ⁸⁸Sr values, consistency with calc-alkaline trends and phase equilibrium modelling provide direct Sr isotopic evidence for water-rich conditions in a Paleoarchean mafic plateau setting. Voluminous felsic crust can be generated efficiently under hydrous conditions without modern style subduction. Water is a critical driver of early continental growth, enabling differentiation of mafic plateaus and shaping the thermal volatile evolution of the early Earth.

top

Acknowledgements

This study was supported by the National Natural Science Foundation of China (Grant No. 42403002). JEH thanks the German Research Foundation (DFG) for grants HO 4697/1-1 and HO 4697/1-2. We thank Chun-Hui Li (from Chengdu University of Technology) for helpful discussion. We thank editor Helen Williams and Ambre Luguet for patient handling of the manuscript. We are also grateful to Renée Tamblyn and one anonymous reviewer for their constructive comments and suggestions that greatly improved the manuscript.

Editor: Helen Williams and Ambre Luguet

top

References

top

Supplementary Information

The supplementary material includes:

• Thermodynamic modeling
• The metamorphic overprinting of ∼3.45 Ga NG sample
• Rayleigh fractionation modeling
• Binary mixing modeling of mafic and felsic melts
• Discussion: evaluation of potential non-primary factors
• Figures S-1to S-11
• Tables S-1 and S-2
• Supplementary materials references

Download the Supplementary Information (PDF)

top