Elsevier

Precambrian Research

Volume 305, February 2018, Pages 479-488
Precambrian Research

Melting controls on the lutetium–hafnium evolution of Archaean crust

https://doi.org/10.1016/j.precamres.2017.12.026Get rights and content

Highlights

  • Crustal evolution arrays are defined by fixed 176Lu/177Hf.

  • Modelling shows melt Lu/Hf varies with degree and depth of melting of basalt.

  • However a melting event has greater impact on melt Lu/Hf than degree of melt.

  • Archaean felsic crust has lower Lu/Hf than that of the Proterozoic and Phanerozoic.

  • We show a modelled mafic source for TTG rocks has enriched Lu/Hf of 0.12.

Abstract

The lutetium–hafnium (Lu–Hf) isotope record, typically measured in zircon crystals, provides a major tool for the study of crustal growth and differentiation. Interpretations of Hf isotope datasets use an evolution array defined by source 176Lu/177Hf. However, the very process that drives crustal differentiation to produce such arrays – partial melting – is precisely that which may modify the trajectory of the array due to variable degrees of anatexis allied with the differing compatibilities of Lu and Hf in residual minerals. Further, Lu/Hf estimates derived from the composition of present-day continental crust may be inappropriate for modelling Archaean crustal evolution, where different geodynamic styles and magmatic sources prevailed.

Using an approach combining phase equilibria, and trace element and isotopic modelling, we quantify the effects of partial melting of both a modern (N-MORB) and Archaean (C-F2) mafic source on melt Lu/Hf. Melting N-MORB shows that the 176Lu/177Hf of the melt, which modelling predicts to be between 0.015 and 0.022, is sensitive to the degree of melting. This difference results in a variation of 4.5 epsilon units/billion years. By contrast, anatexis of C-F2 yields melts with 176Lu/177Hf ∼0.009 that are less affected by the degree of melting. Remelting TTG yields K-rich granitic melts (TTG + G) with 176Lu/177Hf ∼0.005. Thus, a partial melting event imposes a greater control on the resulting crustal reservoir Lu/Hf than the degree of melting. Archaean continental crust has a lower Lu/Hf than that of the average mid to upper continental crust, and therefore a lower 176Lu/177Hf (here 0.005–0.009) is appropriate to modelling its Hf isotopic evolution. There has been a secular change in average crustal Lu/Hf, with the median Lu/Hf of Proterozoic and Phanerozoic magmatic rocks being higher than that of Archaean TTG + G. We show that an enriched Archaean mafic source (C-F2) with a Lu/Hf ratio of ∼0.12 may produce TTG continental crust with a 176Lu/177Hf approaching that calculated in real rocks worldwide.

Introduction

The processes of partial melting, melt segregation, and assimilation, impose a major control on the chemical and rheological evolution of the lithosphere (Brown, 2007, Sawyer et al., 2007, Yakymchuk and Brown, 2014). The lutetium–hafnium (Lu–Hf) isotopic record, typically measured in zircon crystals, is sensitive to these processes, and thus has the potential to offer fundamental insights into crustal evolution (Belousova et al., 2010, Hawkesworth and Kemp, 2006, Kemp et al., 2005, Patchett, 1983, Scherer et al., 2007, Vervoort et al., 1996). Lutetium is generally more compatible than Hf during melting of both mantle and crust. Thus melt, as newly-formed crust, will have lower Lu/Hf relative to its source. These Lu/Hf compositions define evolution arrays along which different crustal (and mantle) reservoirs evolve with time through the ingrowth of radiogenic 176Hf arising from the isotopic decay of 176Lu (Fig. 1).

Evolution arrays have been used for interpreting the processes of reworking of a crustal source as recorded through Hf isotope data; they chart the isotopic evolution of melt products during these magmatic events (Kemp et al., 2006, Kirkland et al., 2013). During a crustal reworking event, the addition of new mantle-derived (juvenile) magma, or incorporation of existing (evolved) crust, may be inferred where the measured Hf isotopic composition of a magmatic rock departs from the evolution array. The evolution array may also be projected back to intercept a Hf evolution model for either the depleted mantle (DM), or that of new continental crust, establishing an Hf model age (TDM2) which may, under ideal circumstances, provide an estimate of the time of extraction of that crustal packet from the mantle (Dhuime et al., 2011, Kemp et al., 2006).

Hf isotope studies use estimates of 176Lu/177Hf to define evolution arrays for the continental crust. However, there are likely variable compositions of crustal melts generated within different settings, and the values commonly used (e.g., 0.012–0.015; Chauvel et al., 2014, Goodge and Vervoort, 2006, Griffin et al., 2002, Rudnick and Gao, 2014) may be inappropriate for modelling the evolution of Archaean continental crust. Such ancient crust is typified by low-K tonalite–trondhjemite–granodiorite (TTG) suites mostly formed from the partial melting of garnet-bearing hydrated mafic source rocks (i.e. garnet amphibolites), which are progressively reworked to generate increasingly K-rich granites (TTG + G suites; Champion and Smithies, 2007). The source of TTG + Gs is therefore distinct to that of modern andesitic crust.

Lutetium and hafnium have different compatibilities in different minerals and so, during partial melting, melt Lu/Hf will be mediated by the presence of certain minerals in the residual source (Vervoort and Patchett, 1996). In particular, the mineral garnet, a phase whose stability is pressure-dependent, and which is a proxy for melting depth, strongly controls the partitioning of Lu. Consequently, the trajectories of consecutive evolution arrays based on source Lu/Hf, arising from successive melt events of a package of crust, are influenced by the nature of these successive melting events. Modelling a single evolution array projecting through successive melt events as linear is simplistic.

In summary, interpretation of Hf isotope datasets and calculation of Hf model ages using evolution arrays is potentially compromised through uncertainties arising from the variability in source Lu/Hf composition, especially with respect to formation of felsic continental crust in the early Earth, and changes in melt Lu/Hf due to different degrees of partial melting mediated by the presence of key minerals in the residue. It is important to quantify these effects, and to establish the role of melting in the development of Hf isotope evolution arrays.

In this contribution, we demonstrate an approach linking phase equilibria modelling, with trace element and isotopic modelling, and how this may be used to explore the sensitivity of melt Lu/Hf to partial melting processes. Phase equilibria modelling is a powerful approach to investigate the processes of partial melting (Johnson et al., 2017, White and Powell, 2002), and here we use new thermodynamic solution models applicable to the melting of mafic rocks (Green et al., 2016). We first model melting of N-MORB. We then focus on the early Earth scenario where different geodynamic styles and magmatic sources may have prevailed, and where the production of felsic continental crust can be more directly tied to the melting of mafic source rocks (Bédard, 2006, Johnson et al., 2014, Martin, 1986, Martin et al., 2014, Nagel et al., 2012, Palin et al., 2016, Rapp and Watson, 1995, Sizova et al., 2015).

During the process of partial melting, Lu/Hf is fractionated between the source rocks and the melt (e.g., the mantle melting event M1 on Fig. 1). Primary (basaltic) crust is ultimately derived from partial melting of the mantle, and will have a lower Lu/Hf than mantle (Sun and McDonough, 1989; Workman and Hart, 2005). This lower Lu/Hf results in the crust evolving along a shallower evolution trajectory than mantle (Fig. 1). Melting of this basaltic crust (the M2 melting event in Fig. 1) will in turn yield melt which may form first generation felsic crust (TTGs in an early Earth scenario). This second melting event will also result in fractionation of Lu/Hf between the basaltic (crustal) source and TTGs. The felsic crust will then evolve along an evolution array defined by its bulk 176Lu/177Hf (Fig. 1). If this crustal package is again melted at a later stage (the M3 melting event in Fig. 1), the crystallization of zircon will reflect the bulk 176Hf/177Hf of its parental magma at this point in time. Thus, the effect of the M2 melting event resulting in different bulk 176Lu/177Hf ratios, will be manifested in differences in the analyzed M3 zircon Hf isotope record.

Section snippets

Approach

We assess the sensitivity of melt Lu/Hf to partial melting of a hydrated mafic source through a combination of phase equilibria, trace element, and isotopic, modelling. Phase equilibria modelling allows calculation of the composition and abundance of phases (stable minerals, melt, and volatile species) as a function of pressure, temperature and bulk rock composition, assuming thermodynamic equilibrium. Trace element modelling uses mineral/melt partition coefficients and the abundances of

Phase control on melt Lu/Hf

The significantly higher compatibility of Lu in garnet compared to Hf means that retention of garnet in the source rocks during anatexis will significantly fractionate melt Lu/Hf (Salters and Hart, 1989, Vervoort and Patchett, 1996). Our modelling reproduces this effect. Fig. 3 also shows the modal proportions of the residual mineral phases (excluding quartz) formed in equilibrium with melt for different degrees of melting along each modelled geothermal gradient. We calculate garnet to be

Conclusions

Variations in melt Lu/Hf produced from the anatexis of hydrated basalts is sensitive to both the degree of partial melt, and to the depth of melting. This variance is controlled to a significant degree by the presence of garnet, an important phase in the lower continental crust, and one whose stability is itself highly sensitive to source XMg. Predicted variations in melt Lu/Hf are mediated by the geological processes of melting and segregation, notably the fertility of the melt as crystallized

Acknowledgements

NJG, TEJ and CLK acknowledges Curtin University for financial support. RHS publishes with permission of the Executive Director, Geological Survey of Western Australia. We thanks Jesse Reimink and two anonymous reviewers for their comments which have greatly enhanced this paper. Randy Parrish is thanked for his editorial handling.

References (74)

  • W.L. Griffin et al.

    Zircon chemistry and magma mixing, SE China: in-situ analysis of Hf isotopes, Tonglu and Pingtan igneous complexes

    Lithos

    (2002)
  • C.J. Hawkesworth et al.

    Using hafnium and oxygen isotopes in zircons to unravel the record of crustal evolution

    Chem. Geol.

    (2006)
  • C. Herzberg et al.

    Thermal history of the Earth and its petrological expression

    Earth Planet. Sci. Lett.

    (2010)
  • C. Herzberg et al.

    Formation of cratonic lithosphere: an integrated thermal and petrological model

    Lithos

    (2012)
  • B.S. Kamber

    The evolving nature of terrestrial crust from the Hadean, through the Archaean, into the Proterozoic

    Precambrian Res.

    (2015)
  • A.I.S. Kemp et al.

    Hadean crustal evolution revisited: New constraints from Pb–Hf isotope systematics of the Jack Hills zircons

    Earth Planet. Sci. Lett.

    (2010)
  • C.L. Kirkland et al.

    Constraints and deception in the isotopic record; the crustal evolution of the west Musgrave Province, central Australia

    Gondwana Res.

    (2013)
  • A. Kröner et al.

    Zircon and muscovite ages, geochemistry, and Nd–Hf isotopes for the Aktyuz metamorphic terrane: Evidence for an Early Ordovician collisional belt in the northern Tianshan of Kyrgyzstan

    Gondwana Res.

    (2012)
  • H. Martin

    Adakitic magmas: modern analogues of Archaean granitoids

    Lithos

    (1999)
  • H. Martin et al.

    Why Archaean TTG cannot be generated by MORB melting in subduction zones

    Lithos

    (2014)
  • H. Martin et al.

    An overview of adakite, tonalite–trondhjemite–granodiorite (TTG), and sanukitoid: relationships and some implications for crustal evolution

    Lithos

    (2005)
  • J.-F. Moyen

    The composite Archaean grey gneisses: petrological significance, and evidence for a non-unique tectonic setting for Archaean crustal growth

    Lithos

    (2011)
  • R.M. Palin et al.

    Partial melting of metabasic rocks and the generation of tonalitic–trondhjemitic–granodioritic (TTG) crust in the Archaean: constraints from phase equilibrium modelling

    Precambrian Res.

    (2016)
  • P.J. Patchett

    Importance of the Lu-Hf isotopic system in studies of planetary chronology and chemical evolution

    Geochim. Cosmochim. Acta

    (1983)
  • E. Sizova et al.

    Generation of felsic crust in the Archean: a geodynamic modeling perspective

    Precambrian Res.

    (2015)
  • R.H. Smithies et al.

    Formation of Paleoarchean continental crust through infracrustal melting of enriched basalt

    Earth Planet. Sci. Lett.

    (2009)
  • R.H. Smithies et al.

    It started with a plume – early Archaean basaltic proto-continental crust

    Earth Planet. Sci. Lett.

    (2005)
  • R.H. Smithies et al.

    The Mesoarchean emergence of modern-style subduction

    Gondwana Res.

    (2007)
  • J.D. Vervoort et al.

    Behavior of hafnium and neodymium isotopes in the crust: constraints from Precambrian crustally derived granites

    Geochim. Cosmochim. Acta

    (1996)
  • R.K. Workman et al.

    Major and trace element composition of the depleted MORB mantle (DMM)

    Earth Planet. Sci. Lett.

    (2005)
  • A. Zeh et al.

    The oldest zircons of Africa—their U-Pb–Hf–O isotope and trace element systematics, and implications for Hadean to Archean crust–mantle evolution

    Precambrian Res.

    (2014)
  • Y. Amelin et al.

    Nature of the Earth’s earliest crust from hafnium isotopes in single detrital zircons

    Nature

    (1999)
  • Berry, A.J., Danyushevsky, L.V., St C. O’Neill, H., Newville, M., Sutton, S.R., 2008. Oxidation state of iron in...
  • M. Brown

    Crustal melting and melt extraction, ascent and emplacement in orogens: mechanisms and consequences

    J. Geol. Soc.

    (2007)
  • Brown, M., Johnson, T.E., 2017. Secular change in metamorphism and the onset of global plate tectonics. Am. J. Sci., in...
  • K.C. Condie

    Archean Greenstone Belts

    (1981)
  • M.J. Defant et al.

    Derivation of some modern arc magmas by melting of young subducted lithosphere

    Nature

    (1990)
  • Cited by (36)

    • Isotopic modelling of Archean crustal evolution from comagmatic zircon–apatite pairs

      2021, Earth and Planetary Science Letters
      Citation Excerpt :

      As such, when estimating SiO2 by the method described here, the most weight should be placed on Rb/Sr, which can then be used to check whether the 176Lu/177Hf ratio of the precursor chosen for model age determination is appropriate. In Hf isotope studies, this ratio is either inferred from the slope of reworking arrays in εHf space, or assumed, with values of 0.022 commonly used for mafic precursor crust (Amelin et al., 1999; Pietranik et al., 2008), 0.015 used for average continental crust (Griffin et al., 2002), and 0.005–0.009 for Archean TTGs (Blichert-Toft and Albarède, 2008; Gardiner et al., 2018). Because of the ability of triple isotope system models to link the Lu/Hf and SiO2 of the precursor, we can improve on the practice and compare modelled values directly to the average Lu/Hf and SiO2 composition of crustal rocks, using data extracted from a global geochemical compilation (Gard et al., 2019).

    View all citing articles on Scopus
    View full text