top

Abstract

Figures and Tables

top

Models of Craton Formation

An essential element to the study of the growth and longevity of Earth’s continental crust is the origin of the chemically buoyant and rigid lithospheric mantle that underpins ancient continental cratons. The vast majority of geochemical evidence for the origin of cratonic peridotites supports initial melt extraction at pressures ≤5 GPa, with most melting occurring in the spinel stability field. This depletion in a shallow, spinel-bearing environment is indicated by low concentrations of mildly incompatible elements and the Cr-rich nature of garnet in cratonic peridotites (Stachel et al., 1998

Stachel, T., Viljoen, K.S., Brey, G., Harris, J.W. (1998) Metasomatic processes in lherzolitic and harzburgitic domains of diamondiferous lithospheric mantle: REE in garnets from xenoliths and inclusions in diamonds. Earth and Planetary Science Letters 159, 1–12.

; Canil, 2004

Canil, D. (2004) Mildly incompatible elements in peridotites and the origins of mantle lithosphere. Lithos 77, 375–393.

; Brey and Shu, 2018

Brey, G.P., Shu, Q. (2018) The birth, growth and ageing of the Kaapvaal subcratonic mantle. Mineralogy and Petrology 1–19.

). If a moderate to low pressure origin is accepted, then two end member models remain. The first model (Fig. 1a) invokes the successive imbrication of subducted oceanic lithosphere (e.g., Helmstaedt and Schulze, 1989

Helmstaedt, H., Schulze, D.J. (1989) Southern African kimberlites and their mantle sample: Implications for Archean tectonics and lithosphere evolution. Kimberlites and Related Rocks 1, 358–368.

; Pearson and Wittig, 2008

Pearson, D.G., Wittig, N. (2008) Formation of Archaean continental lithosphere and its diamonds: the root of the problem. Journal of the Geological Society 165, 895–914.

). The second model – collisional compression via pure shear (Fig. 1b) - proposes that depleted mantle protoliths from Neoarchean sub-arc settings and/or in hot spreading centres are tectonically amalgamated, compressed, and gravitationally thickened during cooling (Jordan, 1978

Jordan, T.H. (1978) Composition and development of the continental tectosphere. Nature 274, 544–548.

; Lee et al., 2011

Lee, C.-T.A., Luffi, P., Chin, E.J. (2011) Building and Destroying Continental Mantle. Annual Review of Earth and Planetary Sciences 39, 59–90.

; Wang et al., 2016

Wang, H., van Hunen, J., Pearson, D.G. (2016) Making Archean cratonic roots by lateral compression: A two-stage thickening and stabilization model. Tectonophysics, doi: 10.1016/j.tecto.2016.12.001.

). The serpentinite-like major element signatures of some peridotites (Si-enriched bulk rock and subcalcic garnets) have been used to support the imbrication of hydrated oceanic lithosphere (Schulze, 1986

Schulze, D.J. (1986) Calcium anomalies in the mantle and a subducted metaserpentinite origin for diamonds. Nature 319, 483–485.

; Canil and Lee, 2009

Canil, D., Lee, C.-T.A. (2009) Were deep cratonic mantle roots hydrated in Archean oceans? Geology 37, 667–670.

). However, these cratonic peridotites have not been systematically tested for the distinctive oxygen isotopic ratios that are characteristic of serpentinites. Here, we search for these distinctive isotopic signatures in mantle-derived cratonic xenoliths.

Full size image | Download in Powerpoint

top

Approach and Results

To characterise the oxygen isotopic composition of cratonic mantle peridotites we present ion microprobe analyses (Supplementary Information) of olivines from coarse, “low-temperature” peridotite xenoliths (Table S-1), thought to represent the bulk of the cratonic lithospheric mantle. The sample suite is derived from the Slave, Rae, North Atlantic, Kaapvaal, and Siberian Archean cratons, and excludes cratonic peridotites that were modified after emplacement by high temperature deformation and melt metasomatism. We also report precise laser fluorination and ion probe measurements on cratonic mantle orthopyroxene (Opx) as well as olivine from forearc peridotites.

Our samples span both sides of the average cratonic mantle olivine Mg# (molar Mg/Mg+Fe) of 92.6 (Fig. 2a). A subset of our samples are Si-enriched, manifested as high bulk rock SiO₂ and elevated modal Opx contents. We adopt the ΔMg/Si parameter of Canil and Lee, 2009

Canil, D., Lee, C.-T.A. (2009) Were deep cratonic mantle roots hydrated in Archean oceans? Geology 37, 667–670.

to quantify Si-enrichment (Fig. 2b), where negative ΔMg/Si denotes Si-enriched, or a lower Mg/Si and Al/Si than can be produced by experimental partial melts of peridotite.

Full size image | Download in Powerpoint

Our measurements reveal that cratonic mantle olivines are normally distributed (Shapiro-Wilk W = 0.99, p = 0.52) over an extremely restricted δ¹⁸O range of +5.01 to +5.53 ‰, with a mean of 5.26 ± 0.22 ‰ (2σ). There are no statistically significant deviations in δ¹⁸O with chemical proxies (Fig. S-1). Additionally, cratonic mantle Opx from variably Si-enriched xenoliths (Table S-2; mean δ¹⁸O of 5.74 ± 0.27 ‰) are in isotopic equilibrium with associated olivine (α_{enstatite–olivine} at 1300 °C = ~0.5 ‰; Rosenbaum et al., 1994

Rosenbaum, J.M., Kyser, T.K., Walker, D. (1994) High temperature oxygen isotope fractionation in the enstatite-olivine-BaCO3 system. Geochimica et Cosmochimica Acta 58, 2653–2660.

).

In order to determine if there are crustal sources of oxygen in the cratonic mantle, we compare our dataset to the well-defined MORB-source mantle (Fig. S-1). We adjust our olivine δ¹⁸O to whole rock peridotite values using a fractionation factor that accounts for the ultramafic composition. Oxygen isotope fractionation between an ultramafic peridotitic mantle source and olivine at high mantle temperatures has been estimated at ~0.2 ‰ (Valley et al., 1998

Valley, J.W., Kinny, P.D., Schulze, D.J., Spicuzza, M.J. (1998) Zircon megacrysts from kimberlite: oxygen isotope variability among mantle melts. Contributions to Mineralogy and Petrology 133, 1–11.

), based on a basalt-olivine fractionation of 0.32 ‰ at 1300 °C (Eiler et al., 2000

Eiler, J.M., Crawford, A., Elliott, T., Farley, K.A., Valley, J.W., Stolper, E.M. (2000) Oxygen Isotope Geochemistry of Oceanic-Arc Lavas. Journal of Petrology 41, 229–256.

) and the positive correlation between melt SiO₂ and D_melt-olivine (Eiler, 2001

Eiler, J.M. (2001) Oxygen Isotope Variations of Basaltic Lavas and Upper Mantle Rocks. Reviews in Mineralogy and Geochemistry 43, 319–364.

). A bootstrapped Kolmogorov-Smirnov (K-S) statistical evaluation of the MORB dataset and the 0.2 ‰ fractionation-corrected cratonic mantle olivines indicates that they are statistically indistinguishable (D = 0.11, p = 0.57), with both populations having analytical uncertainties equal to or larger than the population variance. In contrast, serpentinised peridotite and cratonic mantle olivine have clearly distinct distributions (D = 0.54, p << 0.05; Fig. 3a).

Full size image | Download in Powerpoint

top

Implications for Craton Formation

The striking similarity between oxygen isotope compositions of present day MORB and Archean cratonic peridotites documents the stability of the δ¹⁸O of Earth’s mantle over a 3 billion year interval. Despite this, the diverse δ¹⁸O values of cratonic eclogite xenoliths demonstrate that δ¹⁸O heterogeneities in the subcontinental lithospheric mantle can persist even at the metre scale, likely due to the slow diffusion of oxygen. Since hydrothermal alteration, or the serpentinisation of ultramafic rocks, is an effective method of producing highly variable δ¹⁸O (Wenner and Taylor, 1971

Wenner, D.B., Taylor, H.P. (1971) Temperatures of serpentinization of ultramafic rocks based on O18/O16 fractionation between coexisting serpentine and magnetite. Contributions to Mineralogy and Petrology 32, 165–185.

), cratonic mantle olivines derived by prograde metamorphism of serpentinite (Canil and Lee, 2009

Canil, D., Lee, C.-T.A. (2009) Were deep cratonic mantle roots hydrated in Archean oceans? Geology 37, 667–670.

) should retain both serpentinite-like major element chemistry and distinct oxygen isotopic compositions (Fig 3a).

In contrast to the extreme δ¹⁸O variability of Phanerozoic serpentinite (0 to +8 ‰; Fig. 3a), cratonic peridotitic olivines are strikingly uniform, even over the full compositional range typical for Archean lithospheric mantle. Olivines co-existing with subcalcic garnets and from Si-enriched peridotites - proposed signatures of a serpentinite origin (Schulze, 1986

Schulze, D.J. (1986) Calcium anomalies in the mantle and a subducted metaserpentinite origin for diamonds. Nature 319, 483–485.

; Canil and Lee, 2009

Canil, D., Lee, C.-T.A. (2009) Were deep cratonic mantle roots hydrated in Archean oceans? Geology 37, 667–670.

) - have oxygen isotope compositions that are statistically indistinguishable from normal upper mantle (Fig. 3b). A key piece of evidence previously used in support of the meta-serpentinite hypothesis was a reported correlation between δ¹⁸O and Mg/Si from Colorado Plateau peridotitic xenoliths (Canil and Lee, 2009

Canil, D., Lee, C.-T.A. (2009) Were deep cratonic mantle roots hydrated in Archean oceans? Geology 37, 667–670.

). We note that this Proterozoic terrane is not under-pinned by an Archean cratonic root, and shows clear evidence of hydrous weakening and thinning (Li et al., 2008

Li, Z.-X.A., Lee, C.-T.A., Peslier, A.H., Lenardic, A., Mackwell, S.J. (2008) Water contents in mantle xenoliths from the Colorado Plateau and vicinity: Implications for the mantle rheology and hydration-induced thinning of continental lithosphere. Journal of Geophysical Research 113, 1–22.

). Furthermore, our replotting of these data resolves a correlation (r² = 0.89) between δ¹⁸O and Mg# (Fig. 3c). This trend is not consistent with a metaserpentinite origin, as serpentine minerals are commonly higher in Mg# than the initial olivine and Opx that they replace (Evans, 2008

Evans, B.W. (2008) Control of the Products of Serpentinization by the Fe2+Mg−1 Exchange Potential of Olivine and Orthopyroxene. Journal of Petrology 49, 1873–1887.

). Therefore, these samples likely do not represent the products of seafloor alteration processes (Canil and Lee, 2009

Canil, D., Lee, C.-T.A. (2009) Were deep cratonic mantle roots hydrated in Archean oceans? Geology 37, 667–670.

), but may instead be a consequence of the water-rich metasomatism of the Colorado Plateau lithospheric mantle by Si-enriched fluids or melts from the low-angle subducting Farallon Plate at 80-35 Ma (Li et al., 2008

Li, Z.-X.A., Lee, C.-T.A., Peslier, A.H., Lenardic, A., Mackwell, S.J. (2008) Water contents in mantle xenoliths from the Colorado Plateau and vicinity: Implications for the mantle rheology and hydration-induced thinning of continental lithosphere. Journal of Geophysical Research 113, 1–22.

). Finally, as the narrow variance in Archean cratonic mantle peridotite is statistically equivalent to that of MORB, there is no evidence for even small inputs of meta-serpentinite. Hence, meta-serpentinite is an unlikely protolith to the peridotitic root of the Colorado Plateau, and can be rejected as a protolith for Archean cratonic mantle.

While the Colorado Plateau lithosphere is not a good proxy for pristine cratonic mantle, it is a geometric analogue for the Archean slab imbrication model. The low angle subduction that occurred underneath the Colorado Plateau in the Cretaceous to Eocene likely injected slab melts/fluids directly into the overlying lithospheric mantle, offsetting δ¹⁸O, ΔMg/Si, and Mg#. The similar geometry required for the Archean slab imbrication model would likely produce the same geochemical traces in cratonic mantle olivine. Since our extensive suite of cratonic mantle olivines does not show “Colorado Plateau-like” systematics (between δ¹⁸O and metasomatic proxies - ΔMg/Si or Mg#; Figs. 3c, S-2), it is unlikely that cratonic mantle thickening occurred by slab imbrication. This lack of a crustal alteration signature would be in line with suggestions of higher pressure melting in plumes, but such a model can be rejected from other considerations that place melt depletion in the spinel stability field (see discussion in Introduction). Instead, we hypothesise that cratonic mantle protoliths were initially depleted at MOR and/or at water fluxed sub-arc mantle wedges, environments that produce material with canonical mantle-like δ¹⁸O (Table S-3). These protoliths were then thickened, without significant isotopic modification, by pure shear compression to form cratonic mantle roots (Jordan, 1978

Jordan, T.H. (1978) Composition and development of the continental tectosphere. Nature 274, 544–548.

; McKenzie and Priestley, 2016

McKenzie, D., Priestley, K. (2016) Speculations on the formation of cratons and cratonic basins. Earth and Planetary Science Letters 435, 94–104.

; Wang et al., 2016

Wang, H., van Hunen, J., Pearson, D.G. (2016) Making Archean cratonic roots by lateral compression: A two-stage thickening and stabilization model. Tectonophysics, doi: 10.1016/j.tecto.2016.12.001.

).

While our data is consistent with the construction of cratons by collisional compression of un-altered, depleted peridotite protoliths, we still require a means to enrich these regions heterogeneously in high-Si Opx. Many explanations for cratonic mantle Opx enrichment rely on the infiltration of slab derived melts (e.g., Rudnick, 1994

Rudnick, R.L. (1994) Northern Tanzanian peridotite xenolith: a comparison with Kaapvaal peridotites and evidence for carbonatite interaction with ultra-refractory residues. Proceedings of the 5th International Kimberlite Conference, 336–353.

; Kelemen et al., 1998

Kelemen, P.B., Hart, S.R., Bernstein, S. (1998) Silica enrichment in the continental upper mantle via melt/rock reaction. Earth and Planetary Science Letters 164, 387–406.

). However, lithosphere that has been thinned by crustal melt infiltration is marked by low Mg# and exotic δ¹⁸O (Wang et al., 2018

Wang, C.-G., Xu, W.-L., Yang, D.-B., Liu, Y.-S., Pei, F.-P., Li, Q.-L., Zhou, Q.-J. (2018) Olivine Oxygen Isotope Evidence for Intracontinental Recycling of Delaminated Continental Crust. Geochemistry, Geophysics, Geosystems 19, 1–12.

). Using pMELTS isenthalpic modelling (Ghiorso et al., 2002

Ghiorso, M.S., Hirschmann, M.M. Riners, P.W., Kress, V.C. (2002) The pMELTS: A revision of MELTS for improved calculation of phase relations and major element partitioning related to partial melting of the mantle to 3 GPa. Geochemistry, Geophysics, Geosystems 3, 1–36.

) of assimilation-fractional crystallisation (AFC) for 100 g of slab melt at 3 GPa, we demonstrate that these melts fractionate garnet at slab melt/assimilated rock ratios of  >5, followed by Opx (ratios of 1-5), and olivine, once the melt SiO₂ drops to ~50 wt. % (ratios of <1). Given the low MgO and FeO of experimentally-derived slab melts (Table S-4), high Mg# minerals crystallise after minimal assimilation of peridotite. However, producing mantle-like δ¹⁸O is more difficult (Fig. 4). Mass balance calculations that include fractionation factors (Eiler, 2001

Eiler, J.M. (2001) Oxygen Isotope Variations of Basaltic Lavas and Upper Mantle Rocks. Reviews in Mineralogy and Geochemistry 43, 319–364.

) and the compositional and mineralogical outputs of pMELTS (Table S-4), indicate that at the required melt/rock ratios, δ¹⁸O of Opx and olivine remains higher than mantle. Even melts with modest initial δ¹⁸O (e.g., δ¹⁸O of +4.0 or +6.5 ‰) cease to crystallise Opx before mantle-like δ¹⁸O is achieved, due to low melt SiO₂ (<50 wt. %). Since the measured δ¹⁸O of olivine and Opx from Si-enriched cratonic peridotites do not reflect those modelled in slab melt AFC models, we do not support the slab melt infiltration hypothesis for Si-enrichment.

Since Si-enrichment cannot be achieved by eclogite-derived melts and, as previously discussed in the Introduction, does not show evidence for a high pressure origin, we instead suggest two possible low pressure origins: 1) asthenospheric melts that have become Si-enriched due to the assimilation of wall rock pyroxene and fractionation of olivine during ascent (Kelemen et al., 1992

Kelemen, P.B., Dick, H.J.B., Quick, J.E. (1992) Formation of harzburgite by pervasive melt/rock reaction in the upper mantle. Nature 358, 635–641.

), or 2) the infiltration of moderately Si-rich melts (~56 wt. %) produced from the melting of a depleted harzburgitic mantle fluxed with ~2.5 wt. % H₂O from a devolatilising slab (Mitchell and Grove, 2015

Mitchell, A.L., Grove, T.L. (2015) Melting the hydrous, subarc mantle: the origin of primitive andesites. Contributions to Mineralogy and Petrology 170, 13.

). When 100 g of this +5.6 ‰ harzburgitic melt is input into our AFC model, it crystallises Opx and olivine with cratonic mantle-like Mg# and δ¹⁸O (Fig. 4). The proportion of precipitating Opx, and thus, the Si-enrichment of the surrounding rock, is dependent on the evolving melt-rock ratio. The final stage of thickening, via collisional compression of this heterogeneously Si-enriched material, leads to a cratonic mantle root with variable Opx content, but homogeneous oxygen isotopes.

Full size image | Download in Powerpoint

In summary, the δ¹⁸O homogeneity of a wide range of cratonic mantle peridotites demonstrates that their variably Si-enriched, sometimes subcalcic garnet-bearing nature, cannot have been inherited from protoliths that experienced seafloor serpentinisation (e.g., Schulze, 1986

Schulze, D.J. (1986) Calcium anomalies in the mantle and a subducted metaserpentinite origin for diamonds. Nature 319, 483–485.

; Canil and Lee, 2009

Canil, D., Lee, C.-T.A. (2009) Were deep cratonic mantle roots hydrated in Archean oceans? Geology 37, 667–670.

). Instead, these peridotites must have 1) lain at depths in the oceanic lithosphere where circulating water was minimal and 2) avoided extensive metasomatism from imbricating oceanic lithosphere. Based on these observations, we suggest that cratonic mantle must have been depleted either in thick Archean MOR sequences (Herzberg and Rudnick, 2012

Herzberg, C., Rudnick, R. (2012) Formation of cratonic lithosphere: An integrated thermal and petrological model. Lithos 149, 4–15.

) or in sub-arc fluxed mantle wedges (Parman et al., 2004

Parman, S.W., Grove, T.L., Dann, J.C., de Wit, M.J. (2004) A subduction origin for komatiites and cratonic lithospheric mantle. South African Journal of Geology 107, 107–118.

). The redistribution of Si with mantle-like δ¹⁸O was caused by peridotite-derived melt percolation. Finally, these buoyant, depleted proto-cratonic peridotites were thickened by collisional compression (Wang et al., 2016) into the roots that have protected Earth’s >2.5 billion year old crust.

top

Acknowledgements

We thank the Northwest Territories Geological Survey for funding the analytical time necessary for this project, as well as Dr. Dante Canil and an anonymous reviewer for their constructive comments, and Dr. Ambre Luguet for her editorial comments.

Editor: Ambre Luguet

top

References

top

Supplementary Information

The Supplementary Information includes:

• Supplementary Methods
• Tables S-1 to S-4
• Figures S-1 and S-2
• Supplementary Information References

Download the Supplementary Information (PDF).