Molybdenum isotope evidence for extensive crustal extraction and recycling in Earth’s first billion years Alex J. McCoy-West1, 2, Priyadarshi Chowdhury2, Kevin W. Burton1, Paolo Sossi3, Geoff M. Nowell1, J. Godfrey Fitton4, Andrew C. Kerr5, Peter A. Cawood2 and Helen M. Williams1,6 1Department of Earth Sciences, Durham University, Elvet Hill, Durham DH1 3LE, UK 2School of Earth, Atmosphere and Environment, Monash University, Clayton, Victoria, 3800, Australia 3Institute of Geochemistry and Petrology, ETH Zürich 4School of GeoSciences, University of Edinburgh, Edinburgh EH9 3FE, UK 5School of Earth and Ocean Sciences, Cardiff University, Park Place, Cardiff CF10 3AT, UK 6Department of Earth Sciences, University of Cambridge, Downing Street, Cambridge CB2 3EQ, UK Number of words: 3,029 Number of references: 49 Number of figures: 4 First paragraph (words): 151 Caption Length (words): 100, 96, 64, 90 Methods number of words: 1,190 Corresponding Author Alex McCoy-West (alex.mccoywest@monash.edu) School of Earth, Atmosphere and Environment, Monash University, Clayton, Victoria, 3800, Australia 1 FIRST PARAGRAPH Estimates of the volume of the earliest crust based on zircon ages and radiogenic isotopes remain equivocal. Stable isotope systems, such as molybdenum (Mo), have the potential to provide further constraints but remain underused, due to the lack of complementarity between mantle and crustal reservoirs. We present Mo isotope data for Archean komatiites and Phanerozoic komatiites and picrites and demonstrate that their mantle sources all possess sub-chondritic signatures complementary to the super-chondritic continental crust. These results suggest confirm that the present-day degree of mantle depletion was likely achieved by 3.5 billion years ago and that the Earth has been in a steady state with respect to Mo recycling. Mass balance modelling shows that this early mantle depletion requires the extraction of a far significant greater volume of mafic-dominated proto-crust than previous thought, greater more than twice the volume of present-daythe continental crust today, , implying rapid crustal growth and destruction of crust in the first billion years of Earth’s history. MAIN TEXT The nature, extent and geodynamic settings of crustal formation and recycling are poorly constrained, particularly during Hadean-early Archean times for which the rock-record is scarce1,2. The growth of the crust is estimated to be either temporally skewed with >60-80% of the present-day volume of continental crust (PVCC) forming by 3 billion years ago (Ga)2-5, or much more gradual with time1,6. These growth curves are derived either from zircon formation ages1,2 or from radiogenic isotopic evolution within the crust-mantle system6-8. Zircon ages provide the lower bound on crustal growth as they cannot constrain the magnitude of recycling. In contrast, growth curves of radiogenic isotope systems track the evolution of mantle depletion and implicitly consider both crust extraction and recycling3,9. The complementarity of the crustal and mantle reservoirs for long-lived radiogenic isotopes (Sr-Nd-Hf) has long been established, with time-dependent models requiring that only ~25-50% of the mantle’s mass underwent melt extraction to balance the present-day compositions of the depleted mantle and crust7,8,10. Estimating crustal growth from a mantle-depletion perspective using time-invariant proxies provides an alternative approach4. The ratios of stable isotopes being independent of time, fit this criterion and can put quantitative constraints on differentiation processes occurring in the early Earth. However, this approach is hindered by the lack of resolvable isotopic variation in samples representative of the depleted mantle and crust for many non-traditional stable isotope systems. Molybdenum stable isotopes (δ98Mo = [(98Mo/95Mosample / 98Mo/95Mostandard) −1] × 1000; with the standard NIST3134 = 0‰) may be an exception, with a picture emerging of two complementary reservoirs in the crust and mantle. Chondritic meteorites, the purported building blocks of the terrestrial planets, have a relatively homogeneous average δ98Mo of −0.154 ±0.013‰11,12 (all errors on averages herein are 95% standard errors). Estimates of the composition of the modern continental crust based on molybdenites, granites and primitive arc-related basalts yield super-chondritic δ98Mo values ranging from +0.05‰ to +0.3‰13-15. If the bulk Earth is chondritic with respect to Mo stable isotopes and Mo is not fractionated during its partitioning into Earth’s core (cf. 16), then an isotopically light, sub-chondritic Mo reservoir must exist in the mantle17,18. Arc lavas show extremely variable δ98Mo (−0.88‰ to +0.24‰) but the consensus is that subduction zones appear to be fluxing isotopically light Mo into the mantle19-21. However, whether this material is efficiently recycled or has enough mass to affect the composition of the bulk mantle remains to be established. Previous Mo isotope analyses of Archean komatiites17 have slightly sub-chondritic compositions, but within error of chondrites11, while five of the most depleted (143Nd/144Nd >0.5131) mid-ocean ridge basalts (MORB) measured are resolvably sub-chondritic22. Therefore, it is possible that a complementary light sub-chondritic Mo isotope reservoir is present within the mantle18, but its composition and nature remains poorly constrained. Here, we focus on komatiite and picrite samples from four well characterized suites: two from the Archean, the 3.5 Ga Komati (South Africa) and 2.7 Ga Munro (Canada) komatiites23, and two from the Phanerozoic, the 89 Ma Gorgona (Colombia) komatiites24 and the 61 Ma Baffin Island (NE Canada) picrites25,26, to better constrain the Mo isotope composition of the mantle throughout Earth’s history. The selection of rock samples for this purpose is non-trivial due to the complex behaviour of Mo during mantle melting. None of the major mineral phases in the mantle host significant Mo27, and the presence of residual sulfides will strongly affect the Mo concentration of a melt owing given its chalcophile nature18. Furthermore, isotopic studies of Mo isotopes in ultramafic lithologies are hampered by the low concentration of Mo (<50 ng/g) and the significant isotopic variability in mantle lithologies12,17. Our studied ultramafic lavas formed at elevated temperatures (>1400 °C), by high-degrees of partial melting (>25%), which would have that lead to complete sulfide extraction from their source regions28, implying such that their Mo isotope compositions closely resemble that of their mantle source regions. Our new results combined with the existing data are used to constrain the δ98Mo of the Earth’s mantle, and subsequently global crustal volumes, during Hadean-Archean times. ESTABLISHING A SUB-CHONDRITIC MO ISOTOPE RESERVIOR Our measurements show sub-chondritic values for unaltered Archean Komati and Munro komatiites with δ98Mo varying from −0.22 to −0.18‰ (Fig. 1; Table S1). Previous analyses of Archean komatiites presented in Greber et al.17 define a wide range (−0.32‰ < δ98Mo < +0.07‰) with an average δ98Mo of the four investigated localities calculated as −0.210 ±0.098‰. Combing these results is not straightforward. For example, previously analysed samples from the Vetreny Belt, Fennoscandia have experienced significant crustal assimilation29 and consequently display resolvably heavier δ98Mo (−0.077 ±0.083‰). We thus disregard these samples in subsequent interpretations. In Greber et al.17, lavas from the Komati Formation that were undoubtedly modified by alteration were excluded (Fig. 1; δ98Mo up to +0.44‰), but no further filtering for alteration was attempted. Given the high mobility of Mo in fluids at low temperatures30, we have filtered the Archean komatiite Mo isotope data (Fig. S1), excluding samples that display major element mobility unrelated to magmatic differentiation and are thus considered to have been modified by alteration (see supplement). Our new data, along with the alteration-filtered dataset of 17, allows the calculation of the δ98Mo of Archean komatiites as −0.199 ±0.019‰. Samples from the Phanerozoic Gorgona komatiites, the freshest komatiite occurrence in the world, have a restricted range of δ98Mo from −0.18 to −0.25‰ and yield an average δ98Mo of −0.207 ±0.034‰, which is within error of their Archean equivalents. In contrast, the Phanerozoic Baffin Island picrites possess variable δ98Mo from −0.13 to −0.32‰, which at first glance suggests a lighter composition (Fig. 1). However, the Baffin Island picrites represent a special case of disequilibrium olivine accumulation26 and after correction the composition of the parental melt is calculated as δ98Mo = −0.210 ±0.010‰ (Table S2). The δ98Mo of the Baffin Island parental melts are thus within error of depleted MORB22, the Gorgona komatiites, and three Archean komatiite localities that span 800 Ma, suggesting that the Mo isotope composition of the accessible mantle has changed little over the last 3.5 Ga. These data for magmatic rocks are augmented by mantle xenoliths to calculate the average composition of the depleted mantle as δ98Mo = −0.204 ±0.008‰ (Table S3). These results place several new constraints on the evolution of Earth’s mantle, notably: 1) the Mo isotope composition of the accessible mantle is unambiguously sub-chondritic (an analysis of variance test confirms that the mantle samples are a resolvably different population to chondritic meteorites at the 99% significance level; p-value <0.001); 2) the formation of this reservoir must have occurred before ~3.5 Ga, 3) it must have had a substantial volume (magmas generated at a range of melting depths are affected); and 4) no resolvable temporal variations are observed with Archean komatiites ranging in age from 3.5−2.7 Ga having identical δ98Mo to Cretaceous Gorgona komatiites and Paleogene Baffin Island picrites and modern MORB (an analysis of variance test confirms that the means of these populations are identical; p-value ~0.42). Together these constraints suggest demonstrate that most of the present-day depletion of the mantle was probablymust have been completed by the Paleoarchean. This finding is in agreement with . Analogous results can be obtained from the independent assessment constraints onf the temporal chemical evolution of continental basalts, which indicates a nearly constant amount of mantle depletion, since ~3.8 Ga31. However, the amount of mantle depletion, and hence the volume of early continental crust produced and subsequently destroyed, remain under-constrained. Although estimates are quite variable,Nonetheless, most studies agree that 30−50% melt depletion of the whole mantle can reproduce most of the radiogenic and incompatible element signatures of the crust and depleted mantle, assuming they represent complementary reservoirs7,8,10. This has significant implications for the growth of early crust given that the proto-crust and depleted mantle should chemically complement each other, if no other processes have perturbed the system. We explore this further below. COMPOSITION OF THE SILICATE EARTH Due to the refractory nature of Mo in the solar nebula, we assume that the proto-Earth inherited the δ98Mo of chondritic meteorites (Fig. 2). Soon after accretion, core formation occurred (≈ 34 Ma32) resulting in the efficient removal of the highly siderophile elements into the Fe-Ni metal core, including 95% of the Earth’s original Mo33 (Table S5). The near quantitative removal of Mo to the core means isotope ratios in the metallic phase are unlikely to be fractionated from those in bulk chondrites, as observed in iron meteorites11. Early experimental work suggested this sequestration of Mo may have been associated with a small but resolvable isotopic fractionation of the silicate portion of the planet34. However, recent metal-silicate experiments which incorporate the effect of Mo valence state16 suggest a significantly reduced ∆98Mometal-silicate of as little as −0.008‰ (assuming Mo6+/∑Mo = 0.1; T = 2500 °C), which means the mantle would remain within the error of the composition of chondrites following core formation. Subsequent modification of the residual bulk silicate Earth (BSE) may have occurred during: 1) the Moon-forming impact: where an planet-sized body impacted Earth and added volatiles, including significant sulfur, which were then sequestered to the outer core in the “Hadean matte” (<1% of core mass; this sulfide-enriched phase is expected to have preferentially incorporated isotopically light Mo35,36); or 2) late accretion: since geochemical modelling suggests that all of the Mo in Earth’s mantle was added during the last 10% of accretion37, with N-body simulations require only ~1% of the Earth’s mass was accreted following the Moon-forming impact38. Ultimately, due to the chondritic composition of the new materials these processes will not significantly change the δ98Mo of the BSE, which should be around δ98Mo ≈ −0.154 ‰. Therefore, the only remaining global-scale mechanism that can modify the Earth’s Mo isotope budget and account for the Earth’s super-chondritic crust and sub-chondritic mantle is the extraction of the crust (Fig. 2). Furthermore, the presence of positive Nb anomalies and radiogenic Nd isotope compositions in some komatiite suites suggest that their source regions have previously undergone melt extraction23. EXTRACTION OF AN ISOTOPICALLY HEAVY CRUST The sub-chondritic mantle δ98Mo signature may be the result of partial melting22 or continental crust extraction17 or both, but the exact magnitude of fractionation remains uncertain. Here we have developed a partial melting model to assess the direction and magnitude of fractionation of δ98Mo between melt and residual mantle (Fig. 3). This modelling demonstrates several important points: 1) high-MgO partial melts are accurate recorders of the Mo isotope composition of their mantle sources because at high temperatures ∆98Momelt-solid <0.012‰ at 30% melting (Fig. 3a); 2) melting of a chondritic reservoir to form basalt reproduces the average basalt used in modelling (δ98Mo = −0.10 ‰) with ~12% melting at 1300 °C. This ~0.05‰ difference in δ98Mo is comparable to that observed between N-MORB22 and the depleted mantle composition (herein); 3) the composition of modern upper continental crust or Phanerozoic granites (Fig. 1; ∆98Mogranite-mantle +0.36‰) cannot be generated by direct melting of the mantle. The majority of the enrichment of these samples in heavy δ98Mo must instead result from intracrustal differentiation, either through the addition of isotopically heavy subduction zone fluids19 or hydrothermal fluids39 or the removal of isotopically light hydrous phases (biotite or amphibole)13 into cumulates in the lower crust. Molybdenum isotope fractionation during melt extraction may be driven by both changes in Mo oxidation state and co-ordination number. Given that Mo6+ is significantly more incompatible than Mo4+27 residues of melting will have lower Mo6+/∑Mo than melt in addition to higher mean co-ordination number, and hence will display lighter δ98Mo consistent with sense of fractionation observed in the komatiites measured herein (Fig. 1). The oxidation state of Mo in the modern mantle remains uncertain, however, partitioning studies indicate Mo is predominantly hexavalent in melts at typical upper mantle conditions (Mo6+/∑Mo ≈0.9916,27,40). Although mantle oxygen fugacity is generally considered to have been constant for the last ~3.5 Ga41, recent work using V partitioning provides strong evidence of increasing oxygen fugacity with time42, therefore here we impose Mo6+/∑Mo = 0.95 for early mantle melting (Fig. 3b). Creation of felsic components of the Hadean-Eoarchean crust such as tonalite-trondhjemite-granodiorite (TTG) granitoids, requires remelting of metabasalt (mafic amphibolite)43, which will further enrich this felsic component in heavier isotopes by up to 0.08‰ (at 900 °C and F = 20%), but cannot explain the full range of heavy δ98Mo observed. The models presented here evaluate mantle melting only and should be considered minimum estimates and approximate until Mo isotope fractionation factors can be independently determined for accessory phases that may retain isotopically light Mo (e.g. garnet, amphibole, sulfide). Nonetheless, they demonstrate that there is no need to invoke subduction zone processes in the early Earth to form the mafic crusts discussed below, which can instead be generated solely through mantle melting processes. EXTENSIVE EXTRACTION AND RECYCLING OF EARLY CRUST Assuming a two-reservoir model involving a proto-crust(C) and depleted mantle(DM), we have estimated the crustal volume that is required to have formed by ~3.5 Ga to reconcile the δ98Mo and Mo-concentration of the mantle that sourced the Archean komatiites using the mass-balance equation: where , and represents the mass, Mo concentration and Mo isotope composition, respectively, of the various reservoirs. It is important to note the mass balance modelling presented here does not reflect the instantaneous removal of melts from the mantle, but rather the effect of the time-integrated isolation of the proto-crust from the convecting mantle. Calculations of continental growth based on the zircon archive and mantle depletion commonly use the present-day continental crust as the crustal endmember. However, there are two major compositional differences between the early continents and their modern analogues2,8 . These are: 1) TTG granitoids were the dominant felsic rocks with true potassic (K) granites subordinate in abundance43 and, 2) mafic lithologies were more abundant than their felsic counterparts44,45. Here we assume the BSE had an initial δ98Mo equal to chondritic meteorites (for alternate scenarios Fig. S6) and have investigated two scenarios to encompass the variability of δ98Mo in Archean felsic rocks (granites or TTGs represent the felsic endmember; Fig. 4). These scenarios thus provide the minimum and maximum estimates of the extent of pre-3.5 Ga crust extraction. We have then calculated crustal volumes for three different model proto-crusts: a hypothetical purely felsic crust, Mafic crust-A (minimum based on a mafic crust) and Mafic crust-B (a likely Eoarchean crustal composition). Calculations based on the purely felsic crusts suggest a minimum of 0.5-1.5 times the PVCC (~7.2x109 km3) existed prior to 3.5 Ga based on 30 % depletion of the whole mantle (Fig. 4). This range is consistent with the growth model calculated using Nb/U ratios of the crust-mantle system4, but is higher than those calculated using the crustal zircon formation ages (<50% of PVCC at 3.5 Ga;2). This suggests that time-invariant proxies of mantle depletion record similar volumes of early crust extraction, whereas their difference with the zircon-based models reflects the influence of crustal recycling. More realistic calculations based on dominantly mafic crust types require crustal volumes greater than the PVCC by ~3.5 Ga (Fig 4). For example, in the preferred Eoarchean scenario with a TTG felsic component the crustal volumes based on Mafic crust-A and –B will be 2.5 and 3.8 times the PVCC, respectively, assuming the minimum likely amount of mantle depletion (30%; 7,8,10; Fig. 4b). These higher values are mostly a consequence of the lower Mo concentration (and to a minor extent the lighter isotopic compositions) of these model crusts (see Table S5). It is debatable whether to consider dominantly mafic crust as continental or not44,45, but our calculations show that even the volume of a hypothetical TTG crust would have been greater than the PVCC, provided the depleted mantle size exceeds ~20% of the whole mantle. Thus, it is highly likely that a greater volume of crust than the PVCC was extracted in the first billion years of Earth’s history, most of which was then subsequently recycled into the mantle. Large-scale crust extraction is consistent with the prediction of voluminous melting of the mantle owing to its hotter thermal structure during Hadean-Archean times46. However, our calculated crustal volumes represent the amount of crust extracted from the mantle and not its net growth, which is determined by the difference between extracted (generated) and recycled volumes of the crust9. Nevertheless, high rates of crust formation should result in rapid crustal growth unless the recycling rates equal or exceed extraction rates. Several independent continental growth models2,3,5 do suggest extremely rapid continental growth consistent with the idea that extensive crust formation may have happened on the early Earth. Given the dearth of such old rocks in the present rock record, it is unequivocal that much of the >3.5 Ga crust has been recycled. Mantle-derived isotopic heterogeneities are widespread in modern basalts, reflecting sluggish mantle mixing. Modelling of stagnant lid regimes shows that mixing was up to an order of magnitude slower under these conditions47 therefore it is expected that this recycled crustal material will not have mixed back completely into the accessible mantle. Although difficult to constrain, recent studies on Archean continental recycling48,49 suggest extensive recycling (but not exceeding the formation rates) of the crust, with a volume equivalent to the PVCC probably recycled during the late Archean48. If the recycling rates were similar during most of the Hadean-Archean, twice the PVCC could have been recycled back into the mantle during that period. Consequently, we have not only been significantly underestimating the volumes of early formed crust, but also the amount of material that was being recycled back into the mantle. Acknowledgements Dave Selby is thanked for access to carius tube facilities. This project was funded by a European Research Council Starting Grant (“HabitablePlanet” 306655) to HMW and a NERC Grant (NE/M0003/1) to KWB. While at Monash AMW, PC and PAC were supported by ARC grant FL160100168. Author Contributions AMW and HMW conceived the study. AMW undertook the chemistry and mass spectrometry with assistance from GMN. JGF, ACK and PS provided the samples. AMW and PC developed the mass balance modelling. AMW and PS developed the Mo isotope partial melting model. All authors contributed to discussions on early crustal volumes and writing and editing the paper. REFERENCES 1 Condie, K. C. Episodic continental growth and supercontinents: a mantle avalanche connection? Earth and Planetary Science Letters 163, 97-108 (1998). 2 Dhuime, B., Hawkesworth, C. J., Cawood, P. A. & Storey, C. D. A change in the geodynamics of continental growth 3 billion years ago. Science 335, 1334-1336, doi:10.1126/science.1216066 (2012). 3 Korenaga, J. Estimating the formation age distribution of continental crust by unmixing zircon ages. Earth and Planetary Science Letters 482, 388-395, doi:https://doi.org/10.1016/j.epsl.2017.11.039 (2018). 4 Campbell, I. H. Constraints on continental growth models from Nb/U ratios in the 3.5 Ga Barberton and other Archaean basalt-komatiite suites. American Journal of Science 303, 319-351, doi:10.2475/ajs.303.4.319 (2003). 5 Armstrong, R. L. Radiogenic isotopes: the case for crustal recycling on a near-steady-state no-continental-growth Earth. Philosophical Transactions of the Royal Society of London 301, 443-472, doi:10.1098/rsta.1981.0122 (1981). 6 McCulloch, M. T. & Bennett, V. C. Progressive growth of the Earth's continental crust and depleted mantle: Geochemical constraints. Geochimica et Cosmochimica Acta 58, 4717-4738, doi:http://dx.doi.org/10.1016/0016-7037(94)90203-8 (1994). 7 DePaolo, D. J. Crustal growth and mantle evolution: inferences from models of element transport and Nd and Sr isotopes. Geochimica et Cosmochimica Acta 44, 1185-1196, doi:https://doi.org/10.1016/0016-7037(80)90072-1 (1980). 8 Jacobsen, S. B. Isotopic and chemical constraints on mantle-crust evolution. Geochimica et Cosmochimica Acta 52, 1341-1350, doi:https://doi.org/10.1016/0016-7037(88)90205-0 (1988). 9 Cawood, P. A., Hawkesworth, C. J. & Dhuime, B. The continental record and the generation of continental crust. GSA Bulletin 125, 14-32, doi:10.1130/B30722.1 (2013). 10 O'Nions, R. K., Evensen, N. M. & Hamilton, P. J. Geochemical modeling of mantle differentiation and crustal growth. Journal of Geophysical Research: Solid Earth 84, 6091-6101, doi:doi:10.1029/JB084iB11p06091 (1979). 11 Burkhardt, C., Hin, R. C., Kleine, T. & Bourdon, B. Evidence for Mo isotope fractionation in the solar nebula and during planetary differentiation. Earth and Planetary Science Letters 391, 201-211, doi:http://dx.doi.org/10.1016/j.epsl.2014.01.037 (2014). 12 Liang, Y.-H. et al. Molybdenum isotope fractionation in the mantle. Geochimica et Cosmochimica Acta 199, 91-111, doi:https://doi.org/10.1016/j.gca.2016.11.023 (2017). 13 Voegelin, A. R., Pettke, T., Greber, N. D., von Niederhäusern, B. & Nägler, T. F. Magma differentiation fractionates Mo isotope ratios: Evidence from the Kos Plateau Tuff (Aegean Arc). Lithos 190–191, 440-448, doi:http://dx.doi.org/10.1016/j.lithos.2013.12.016 (2014). 14 Yang, J. et al. The molybdenum isotopic compositions of I-, S- and A-type granitic suites. Geochimica et Cosmochimica Acta 205, 168-186, doi:https://doi.org/10.1016/j.gca.2017.01.027 (2017). 15 Greber, N. D., Pettke, T. & Nägler, T. F. Magmatic–hydrothermal molybdenum isotope fractionation and its relevance to the igneous crustal signature. Lithos 190-191, 104-110, doi:https://doi.org/10.1016/j.lithos.2013.11.006 (2014). 16 Hin, R. C., Burnham, A. D., Gianolio, D., Walter, M. J. & Elliott, T. Molybdenum isotope fractionation between Mo4+ and Mo6+ in silicate liquid and metallic Mo. Chemical Geology 504, 177-189, doi:https://doi.org/10.1016/j.chemgeo.2018.11.014 (2019). 17 Greber, N. D., Puchtel, I. S., Nägler, T. F. & Mezger, K. Komatiites constrain molybdenum isotope composition of the Earth's mantle. Earth and Planetary Science Letters 421, 129-138, doi:https://doi.org/10.1016/j.epsl.2015.03.051 (2015). 18 Willbold, M. & Elliott, T. Molybdenum isotope variations in magmatic rocks. Chemical Geology 449, 253-268, doi:https://doi.org/10.1016/j.chemgeo.2016.12.011 (2017). 19 Freymuth, H., Vils, F., Willbold, M., Taylor, R. N. & Elliott, T. Molybdenum mobility and isotopic fractionation during subduction at the Mariana arc. Earth and Planetary Science Letters 432, 176-186, doi:http://dx.doi.org/10.1016/j.epsl.2015.10.006 (2015). 20 Gaschnig, R. M. et al. The Molybdenum Isotope System as a Tracer of Slab Input in Subduction Zones: An Example From Martinique, Lesser Antilles Arc. Geochemistry, Geophysics, Geosystems 18, 4674-4689, doi:doi:10.1002/2017GC007085 (2017). 21 König, S., Wille, M., Voegelin, A. & Schoenberg, R. Molybdenum isotope systematics in subduction zones. Earth and Planetary Science Letters 447, 95-102, doi:https://doi.org/10.1016/j.epsl.2016.04.033 (2016). 22 Bezard, R., Fischer-Gödde, M., Hamelin, C., Brennecka, G. A. & Kleine, T. The effects of magmatic processes and crustal recycling on the molybdenum stable isotopic composition of Mid-Ocean Ridge Basalts. Earth and Planetary Science Letters 453, 171-181, doi:http://dx.doi.org/10.1016/j.epsl.2016.07.056 (2016). 23 Sossi, P. A. et al. Petrogenesis and Geochemistry of Archean Komatiites. Journal of Petrology 57, 147-184, doi:10.1093/petrology/egw004 (2016). 24 Kerr, A. C. et al. The petrogenesis of Gorgona komatiites, picrites and basalts: new field, petrographic and geochemical constraints. Lithos 37, 245-260, doi:http://dx.doi.org/10.1016/0024-4937(95)00039-9 (1996). 25 Starkey, N. A. et al. Helium isotopes in early Iceland plume picrites: Constraints on the composition of high 3He/4He mantle. Earth and Planetary Science Letters 277, 91-100, doi:https://doi.org/10.1016/j.epsl.2008.10.007 (2009). 26 McCoy-West, A. J., Godfrey Fitton, J., Pons, M.-L., Inglis, E. C. & Williams, H. M. The Fe and Zn isotope composition of deep mantle source regions: Insights from Baffin Island picrites. Geochimica et Cosmochimica Acta 238, 542-562, doi:https://doi.org/10.1016/j.gca.2018.07.021 (2018). 27 Leitzke, F. P. et al. Redox dependent behaviour of molybdenum during magmatic processes in the terrestrial and lunar mantle: Implications for the Mo/W of the bulk silicate Moon. Earth and Planetary Science Letters 474, 503-515, doi:https://doi.org/10.1016/j.epsl.2017.07.009 (2017). 28 McCoy-West, A. J., Bennett, V. C., O’Neill, H. S. C., Hermann, J. & Puchtel, I. S. The interplay between melting, refertilization and carbonatite metasomatism in off-cratonic lithospheric mantle under Zealandia: An integrated major, trace and platinum group element study. Journal of Petrology 56, 563-604, doi:10.1093/petrology/egv011 (2015). 29 Puchtel, I. S. et al. Petrology of a 2.41 Ga remarkably fresh komatiitic basalt lava lake in Lion Hills, central Vetreny Belt, Baltic Shield. Contributions to Mineralogy and Petrology 124, 273-290, doi:10.1007/s004100050191 (1996). 30 Keppler, H. & Wyllie, P. J. Partitioning of Cu, Sn, Mo, W, U, and Th between melt and aqueous fluid in the systems haplogranite-H2O−HCl and haplogranite-H2O−HF. Contributions to Mineralogy and Petrology 109, 139-150, doi:10.1007/bf00306474 (1991). 31 Keller, B. & Schoene, B. Plate tectonics and continental basaltic geochemistry throughout Earth history. Earth and Planetary Science Letters 481, 290-304, doi:https://doi.org/10.1016/j.epsl.2017.10.031 (2018). 32 Kleine, T. & Walker, R. J. Tungsten Isotopes in Planets. Annual Review of Earth and Planetary Sciences 45, 389-417, doi:10.1146/annurev-earth-063016-020037 (2017). 33 McDonough, W. F. in Treatise on Geochemistry (First Edition) Vol. 2 (eds Heinrich D. Holland & Karl K. Turekian) 547-568 (Elsevier, 2003). 34 Hin, R. C., Burkhardt, C., Schmidt, M. W., Bourdon, B. & Kleine, T. Experimental evidence for Mo isotope fractionation between metal and silicate liquids. Earth and Planetary Science Letters 379, 38-48, doi:http://dx.doi.org/10.1016/j.epsl.2013.08.003 (2013). 35 Savage, P. S. et al. Copper isotope evidence for large-scale sulphide fractionation during Earth’s differentiation. Geochemical Perspectives Letters 1, 53-64, doi:http://dx.doi.org/10.7185/geochemlet.1506 (2015). 36 McCoy-West, A. J., Millet, M.-A. & Burton, K. W. The neodymium stable isotope composition of the silicate Earth and chondrites. Earth and Planetary Science Letters 480, 121-132, doi:https://doi.org/10.1016/j.epsl.2017.10.004 (2017). 37 Dauphas, N. The isotopic nature of the Earth’s accreting material through time. Nature 541, 521, doi:10.1038/nature20830 https://www.nature.com/articles/nature20830#supplementary-information (2017). 38 Jacobson, S. A. et al. Highly siderophile elements in Earth’s mantle as a clock for the Moon-forming impact. Nature 508, 84, doi:10.1038/nature13172 (2014). 39 Neely, R. A. et al. Molybdenum isotope behaviour in groundwaters and terrestrial hydrothermal systems, Iceland. Earth and Planetary Science Letters 486, 108-118, doi:https://doi.org/10.1016/j.epsl.2017.11.053 (2018). 40 O'Neill, H. S. C. & Eggins, S. M. The effect of melt composition on trace element partitioning: an experimental investigation of the activity coefficients of FeO, NiO, CoO, MoO2 and MoO3 in silicate melts. Chemical Geology 186, 151-181, doi:http://dx.doi.org/10.1016/S0009-2541(01)00414-4 (2002). 41 Hibbert, K. E. J., Williams, H. M., Kerr, A. C. & Puchtel, I. S. Iron isotopes in ancient and modern komatiites: Evidence in support of an oxidised mantle from Archean to present. Earth and Planetary Science Letters 321–322, 198-207, doi:http://dx.doi.org/10.1016/j.epsl.2012.01.011 (2012). 42 Nicklas, R. W. et al. Secular mantle oxidation across the Archean-Proterozoic boundary: Evidence from V partitioning in komatiites and picrites. Geochimica et Cosmochimica Acta 250, 49-75, doi:https://doi.org/10.1016/j.gca.2019.01.037 (2019). 43 Moyen, J.-F. & Laurent, O. Archaean tectonic systems: A view from igneous rocks. Lithos 302-303, 99-125, doi:https://doi.org/10.1016/j.lithos.2017.11.038 (2018). 44 Dhuime, B., Wuestefeld, A. & Hawkesworth, C. J. Emergence of modern continental crust about 3 billion years ago. Nature Geoscience 8, 552, doi:10.1038/ngeo2466 https://www.nature.com/articles/ngeo2466#supplementary-information (2015). 45 Tang, M., Chen, K. & Rudnick, R. L. Archean upper crust transition from mafic to felsic marks the onset of plate tectonics. Science 351, 372-375, doi:10.1126/science.aad5513 (2016). 46 Herzberg, C., Condie, K. & Korenaga, J. Thermal history of the Earth and its petrological expression. Earth and Planetary Science Letters 292, 79-88, doi:https://doi.org/10.1016/j.epsl.2010.01.022 (2010). 47 O'Neill, C., Debaille, V. & Griffin, W. Deep earth recycling in the Hadean and constraints on surface tectonics. American Journal of Science 313, 912-932, doi:10.2475/09.2013.04 (2013). 48 Chowdhury, P., Gerya, T. & Chakraborty, S. Emergence of silicic continents as the lower crust peels off on a hot plate-tectonic Earth. Nature Geoscience 10, 698, doi:10.1038/ngeo3010 https://www.nature.com/articles/ngeo3010#supplementary-information (2017). 49 Johnson, T. E., Brown, M., Kaus, B. J. P. & VanTongeren, J. A. Delamination and recycling of Archaean crust caused by gravitational instabilities. Nature Geoscience 7, 47, doi:10.1038/ngeo2019 https://www.nature.com/articles/ngeo2019#supplementary-information (2013). Figures Figure 1: Variation of δ98Mo in komatiites, picrites and major mantle and crustal reservoirs. Filled symbols are data analysed herein with hollow symbols data taken from Greber et al. 17. All individual analyses are plotted with the 2 standard deviation long term error, with the shaded areas for different formations and reservoirs the being 95% standard errors. The dark grey band represents chondritic meteorites (δ98Mo = −0.154 ±0.013‰; 11,12) with the green bar representing the resolvable lighter depleted mantle (δ98Mo = −0.204 ±0.008‰; herein). Average Archean komatiites (δ98Mo = −0.199 ±0.019‰; herein) with other reservoirs from 12,14,22 (see Table S3). Figure 2: Schematic Mo evolution of Earth’s mantle and crust during planetary differentiation. Earth accretes from chondritic meteorites thus the bulk Earth initial δ98Mo will be chondritic. During core formation 95 % of Earth’s Mo is sequestered into the core trapping isotopically light Mo in the metal phase, possibly making the residual BSE heavier. Subsequent extraction of Earth’s isotopically heavy crust prior to 3.5 Ga resulted in a bulk mantle that is lighter than the building blocks of Earth. Earth’s earliest crust was more mafic than modern crust and therefore had a different Mo concentration and isotopic composition. Figure 3: Partial melting model showing that the degree of enrichment of heavy Mo isotopes in the melt phase is controlled by both temperature and the valance state of Mo. (a) the effect of varying temperature at a constant oxygen fugacity (Mo6+/∑Mo = 0.95). Shaded areas represent varying the temperature by ±100 °C. (b) The effect of varying oxygen fugacity at a constant temperature (1300 °C). Figure 4: Results of Mo isotope mass balance calculations which estimate the mass of crust extraction required to balance the composition of the depleted mantle. This mass of crust can then be converted into a volume of crust (VC) relative to the present volume of continental crust (VPCC) and varies depending on the proportion of the total BSE that has undergone melt depletion (MDM/MBSE). Mafic crust-A and -B contain mafic and felsic rocks in 50:50 and 75:25 ratios, respectively. The shaded areas represent varying the proportions of the two endmembers by ±5%. 13 image1.png image2.png image3.png image4.png