Half a billion years of reworking of Hadean mafic crust to produce the Nuvvuagittuq Eoarchean felsic crust

Half a billion years of reworking of Hadean mafic crust to produce the Nuvvuagittuq Eoarchean felsic crust

Earth and Planetary Science Letters 379 (2013) 13–25 Contents lists available at ScienceDirect Earth and Planetary Science Letters www.elsevier.com/...

2MB Sizes 0 Downloads 44 Views

Earth and Planetary Science Letters 379 (2013) 13–25

Contents lists available at ScienceDirect

Earth and Planetary Science Letters www.elsevier.com/locate/epsl

Half a billion years of reworking of Hadean mafic crust to produce the Nuvvuagittuq Eoarchean felsic crust Jonathan O’Neil a,b,∗ , Maud Boyet b , Richard W. Carlson c , Jean-Louis Paquette b a b c

Department of Earth Sciences, University of Ottawa, 140 Louis Pasteur, Ottawa, K1N 6N5, Canada Clermont Université, Université Blaise Pascal, Laboratoire Magmas et Volcans, CNRS UMR 6524, BP 10448, F-63000 Clermont-Ferrand, France Department of Terrestrial Magnetism, Carnegie Institution of Washington, 5241 Broad Branch Rd. NW, Washington, DC, 20015, USA

a r t i c l e

i n f o

Article history: Received 1 November 2012 Received in revised form 3 July 2013 Accepted 24 July 2013 Available online 24 August 2013 Editor: T. Elliot Keywords: Nuvvuagittuq Hadean Eoarchean TTG crustal evolution Pb and Hf isotopes

a b s t r a c t The Nuvvuagittuq greenstone belt is dominated by mafic rocks, called the Ujaraaluk unit, that are mostly composed of cummingtonite–plagioclase–biotite with variable amounts of garnet. While the oldest zircons contained in thin intrusive trondhjemitic bands are ∼3.8 Ga, 146 Sm–142 Nd systematics suggest that the Ujaraaluk unit is as old as 4.4 Ga. The Nuvvuagittuq greenstone belt is surrounded by Eoarchean TTGs that have geochemical and isotopic compositions consistent with their derivation by partial melting of a source similar in composition and age to the Ujaraaluk unit. New zircon dates reported here show the Nuvvuagittuq TTGs to consist at least of four distinct age units of 3.76 Ga, 3.66 Ga, 3.5–3.4 Ga and 3.35 Ga. The Hf isotopic compositions of zircons from the TTG are consistent with derivation from Hadean mafic crust. The 3.66 Ga to 3.35 Ga TTGs appear to have been formed primarily from melting of a source compositionally similar to the 4.4 Ga Ujaraaluk unit, whereas the more radiogenic Hf of the zircons from the 3.76 Ga TTGs may suggest derivation from melting of a source compositionally similar to 4.1 Ga intrusive gabbros. Alternatively, the distinct rare earth element patterns of the 3.76 Ga and 3.66 Ga TTGs suggest their derivation from sources with variable amounts of residual garnet and hence formation at different depths. The composition of the older TTGs is indicative of a deeper source that may have involved a greater interaction between the melt and the mantle to explain the more radiogenic Hf isotopic compositions of their zircons. Sources compositionally similar to the Ujaraaluk unit and intrusive gabbros appear to be the most likely candidates for the Hadean precursor of the Nuvvuagittuq TTGs. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Investigation of Earth’s primitive crust, its formation mechanism and its evolution though time is unfortunately limited by the scarcity of Eoarchean/Hadean terranes. Most of the oldest Archean terranes are dominated by felsic rocks from the Tonalite– Trondhjemite–Granodiorite (TTG) series. These felsic rocks, however, cannot be directly produced by melting of peridotitic mantle, but must instead have been derived from the melting of an older mafic precursor. The Nuvvuagittuq greenstone belt (NGB) located in Northeastern Canada is mainly composed of mafic/ultramafic rocks with a minimum age of 3.75 Ga (Cates and Mojzsis, 2007; David et al., 2009). Such an early mafic precursor may represent a remnant of Earth’s primordial crust. The dominant lithology of the NGB, called the Ujaraaluk unit, is interpreted to be a metamorphosed hydrothermally altered mafic volcanic sequence (O’Neil et al., 2007, 2011) formed in the Hadean,

*

Corresponding author at: Department of Earth Sciences, University of Ottawa, 140 Louis Pasteur, Ottawa, K1N 6N5, Canada. E-mail address: [email protected] (J. O’Neil). 0012-821X/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.epsl.2013.07.030

between 4.3 and 4.4 Ga (O’Neil et al., 2008, 2012). Obtaining accurate ages on old terrestrial mafic rocks, however, is challenging. Because Zr-poor mafic magmas do not crystallize zircon, geochronological constraints on Archean mafic rocks commonly come from long-lived radiogenic isotope systems that are susceptible to partial or total resetting by younger metamorphic/metasomatic events. The short-lived 146 Sm–142 Nd isotope system that yields the Hadean age for the NGB is less susceptible to partial resetting (O’Neil et al., 2012) because 146 Sm became extinct before ∼4 Ga. The Hadean age of the NGB mafic rocks however has been questioned in part because the oldest U–Pb ages obtained on zircons in meta-igneous rocks from the NGB range 51 from 3758+ −47 Ma to 3817 ± 16 Ma (Cates and Mojzsis, 2007; David et al., 2009). These ∼3.8 Ga rocks, however are intrusive trondhjemitic bands and thus provide only a minimum age for the mafic rocks. In order to better constrain the geological relationship between the mafic and felsic rocks and the evolution of the NGB through time, we present whole-rock Lu–Hf data for the Nuvvuagittuq rocks as well as coupled U–Pb and Hf isotopic analyses in zircons from a series of surrounding and intrusive TTGs.

14

J. O’Neil et al. / Earth and Planetary Science Letters 379 (2013) 13–25

2. Geological setting 2.1. Geologic summary The NGB is a ∼10 km2 volcano-sedimentary belt located in Northeastern Canada (Fig. 1) mainly composed of a mappable unit called the Ujaraaluk unit consisting of a mafic cummingtonite– plagioclase–biotite ± garnet amphibolite (O’Neil et al., 2011). The Ujaraaluk unit is divided into distinct chemostratigraphic units with a high-Ti and a low-Ti geochemical group. The low-Ti Ujaraaluk is further subdivided into two subgroups based on their respective abundances of incompatible trace elements with a depleted low-Ti group and an enriched low-Ti group (O’Neil et al., 2011). The map pattern of the Ujaraaluk unit suggests that the different groups follow a chemical stratigraphy within the NGB subsequently deformed into a kilometer-scale refolded fold (Fig. 1). The NGB also contains 3 chemical groups of ultramafic bodies interpreted to be co-genetic cumulates to the mafic Ujaraaluk and following the same map pattern as their respective Ujaraaluk mafic groups (O’Neil et al., 2011). The NGB is bordered by Eoarchean TTGs dated at 3661 ± 4 Ma (David et al., 2009), further surrounded by younger 2750 ± 16 Ma heterogeneous tonalities (Simard et al., 2003). Predominantly along the western margin of the NGB, the Eoarchean TTG intrudes the NGB mafic lithologies. Despite the fact that they were produced after the extinction of 146 Sm, the TTGs yield a similar range in 142 Nd/144 Nd ratios compared to the Ujaraaluk unit, but with lower Sm/Nd ratios and no correlation between their 142 Nd/144 Nd and Sm/Nd ratios (O’Neil et al., 2012) as is also observed in new 142 Nd analyses reported in this study (Table 2 supplementary data, Fig. 6 supplementary data). The 142 Nd anomalies for the NGB TTGs are consistent with their derivation from melting of a Hadean precursor (i.e. formed before the extinction of 146 Sm) that is compositionally similar to the Ujaraaluk unit. Remelting of a Hadean Ujaraaluk-like basement to produce the NGB TTGs is also supported by experimental data showing that partial melting of rocks compositionally similar to the Ujaraaluk produces melts similar to the NGB Eoarchean TTG (Adam et al., 2012). Other felsic lithologies include a plagioclase–quartz–biotite schist dated at 3366 ± 3 Ma (David et al., 2009) and fine-grained trondhjemites locally intruding the NGB as thin 30–50 cm bands dated at 3751 ± 10 Ma (Cates and Mojzsis, 2007) and 3817 ± 16 Ma (David et al., 2009) establishing a minimum age for the NGB. 2.2. Geochronological debate Rocks from the Ujaraaluk unit and co-genetic ultramafic rocks yield a wide range of 142 Nd/144 Nd ratios, from samples with excesses in 142 Nd (up to +8 ppm) to samples with deficits in 142 Nd (down to −18 ppm) compared to terrestrial Nd standards (O’Neil et al., 2012). The variability in 142 Nd can only have been created in the Hadean as 146 Sm was extinct by the beginning of the Archean. Samples from the Ujaraaluk unit show a statistically significant correlation between their 142 Nd/144 Nd and Sm/Nd ratios in all Ujaraaluk groups taken together or individually by chemical group (O’Neil et al., 2012). The overall correlation provides a 15 slope that corresponds to an age of 4388+ −17 Ma (MSWD = 5.8,

14 n = 50), or 4406+ −17 Ma (MSWD = 1.0, n = 9) when considering only the samples with the least disturbed Sm–Nd. The Ujaraaluk unit has been affected by Neoarchean metamorphism as shown by the crystallization of garnet around 2.7 Ga (O’Neil et al., 2012). This thermal event has affected the 147 Sm–143 Nd systematics of the Ujaraaluk unit causing all chemical groups combined to define a scattered 147 Sm/144 Nd vs. 143 Nd/144 Nd correlation corresponding to an age of 3598 ± 200 Ma (MSWD = 134) with the individual chemical groups defining 147 Sm–143 Nd correlations that range in

age from 2517 to 3257 Ma. The least disturbed Ujaraaluk samples however yield a 147 Sm–143 Nd isochron age of 4321 ± 160 Ma (MSWD = 6.3, n = 9, O’Neil et al., 2012). Massive gabbro sills intruding the Ujaraaluk unit give a 147 Sm–143 Nd isochron age of 4115 ± 100 Ma (MSWD = 4.8, n = 13) interpreted by O’Neil et al. (2012) as a minimum age for the Ujaraaluk, consistent with their Hadean age. The Hadean age of the Ujaraaluk unit has been challenged. The 142 Nd anomalies in the Ujaraaluk unit have been confirmed by Roth et al. (2013), but they attributed the 142 Nd variation to be an inherited feature caused by mixing, at 3.8 Ga, with a hypothesized Hadean enriched source reservoir. We note that their hypothesized enriched Hadean end member has Sm/Nd and 143 Nd/144 Nd ratios essentially identical to the lowest Sm/Nd sample from the Ujaraaluk unit thus allowing the Ujaraaluk to be their hypothesized enriched Hadean crust. Consequently, this model does not rule out the simpler model proposed by O’Neil et al. (2012) that the 142 Nd isotopic variation in the Ujaraaluk unit is due to their formation while 146 Sm was still actively decaying. Guitreau et al. (2013) proposed an age of 3864 ± 70 Ma for the NGB based on a Lu–Hf isochron that includes rock types ranging from mafic amphibolite to TTG to quartzites and includes separated zircons. As their sample set includes data for only two Ujaraaluk samples, and given that our much more extensive Lu–Hf dataset for the Ujaraaluk, presented in this paper, provides Lu–Hf “isochron” slopes ranging from 2537 to about 4400 Ma depending on sample lithology, we consider the limited Lu–Hf data from Guitreau et al. (2013) to provide little or no constraint on the age of the Ujaraaluk. Cates et al. (2013) proposed a maximum age of 3780 Ma for the NGB based on the oldest 207 Pb/206 Pb age on zircon from a putative detrital fuchsitic quartzite. However, the 207 Pb/206 Pb ages believed to be non-metamorphic are as low as 3718 Ma and the U–Pb Concordia age of these zircons suggest they derive from a single source with an age of 3742 ± 14 Ma. This age 51 is indistinguishable from the 3758+ −47 Ma to 3817 ± 16 Ma trondhjemites intruding the NGB that provide a minimum age of emplacement for the NGB mafic rocks. The single age population and its agreement with the age of intruding trondhjemites calls into question the interpretation of this sample as a “detrital” quartzite. Darling et al. (in press) have rather suggested that these fuchsitic quartzites are metasomatically-altered orthogneiss intrusive bands. 3. Sample description and analytical procedures A series of mafic and felsic samples were selected and analyzed for their whole-rock Lu–Hf isotopic compositions (Table 1). Mafic samples comprise 24 Ujaraaluk samples including all three distinct geochemical groups, one ultramafic sample and 6 gabbro samples. Felsic samples comprise 9 TTGs, one pegmatite and the 3366 Ma plagioclase–quartz–biotite schist from David et al. (2009) called felsic schist in this study. The TTGs have compositions ranging from trondhjemites to granodiorites (Fig. 1 supplementary data). They comprise coarse grained rocks exhibiting a strong foliation found to the West of the NGB or inside the major NGB fold (Fig. 1), and thin finer grained bands intruding the NGB to the Southwest. For this study, we call the ∼3.8 Ga intrusive bands, the trondhjemitic bands, and the coarser-grained felsic rocks on both sides of the NGB, the TTGs. Of these felsic samples, zircon separates from 3 TTGs (PC-284, PC-285, PC-286), one trondhjemitic band (PC-287), and the felsic schist (PC-134) also were analyzed for Lu–Hf and U– Pb systematics (Tables 2, 3 supplementary data). Four TTG samples have been analyzed for their 146 Sm–142 Nd and 147 Sm–143 Nd isotope systematics (Table 2 supplementary data). Sample locations are shown on Fig. 1. All Hf isotopic compositions were determined using the Nu-Plasma ICP-MS at the Department of Terrestrial Magnetism (DTM) in Washington DC while the whole-rock Sm–Nd data

J. O’Neil et al. / Earth and Planetary Science Letters 379 (2013) 13–25

15

Fig. 1. Geological and geochronological map of the Nuvvuagittuq greenstone belt with sample locations. The different geochemical groups of the Ujaraaluk unit are shown (including their respective co-genetic ultramafic sills) in different shades of green. TTG samples are color-coded according to their ages. Sample numbers are indicated for the TTGs. All mafic and felsic samples have been analyzed for their whole-rock Lu–Hf isotopic compositions. Stars indicate the samples where zircons were also separated for Lu–Hf measurments by LA-ICP-MS. Coordinates are in UTM NAD 27. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

16 Table 1 Whole-rock Lu–Hf isotopic data for the NGB rocks. 2σ -errors for 176 Hf/177 Hf are in the last decimal place. Values for the depleted mantle used for TDM age calculations corresponds to a 176 Lu/177 Hf ratio = 0.03915 starting with a chondritic 176 Hf/177 Hf at 4.567 Ga (176 Hf/177 Hf = 0.2833). Values used for ε Hf calculations are 176 Hf/ 177 HfCHUR = 0.282785, 176 Lu/177 HfCHUR = 0.0336 (Bouvier et al., 2008) and decay constant for 176 Lu = 1.865 × 10−11 yr−1 (Scherer et al., 2001). Ages for Ujaraaluk and gabbro are from O’Neil et al. (2012), ages for the TTG, felsic schist and trondhjemtic band are from this study and age for the pegmatite is from David et al. (2009). Sample numbers marked “D” are duplicate analyses. Coordinates are in UTM NAD 27. Lithology

Easting

Northing

PC-033 PC-129 PC-421 PC-423 PC-540 PC-541 PC-227 PC-230 PC-230D PC-410 PC-433 PC-521 PC-546 PC-149 PC-162 PC-162D PC-225 PC-225D PC-251 PC-267 PC-267D PC-275 PC-275D PC-276 PC-282 PC-282D PC-282D2 PC-438 PC-442 PC-443 PC-515 PC-517 PC-221 PC-248 PC-508 PC-511 PC-522 PC-523 PC-039 PC-104 PC-106 PC-169 PC-284 PC-285 PC-286 PC-134 PC-101 PC-102 PC-287 R18

High-Ti Ultramafic High-Ti Ujaraaluk High-Ti Ujaraaluk High-Ti Ujaraaluk High-Ti Ujaraaluk High-Ti Ujaraaluk Depleted low-Ti Ujaraaluk Depleted low-Ti Ujaraaluk Depleted low-Ti Ujaraaluk Depleted low-Ti Ujaraaluk Depleted low-Ti Ujaraaluk Depleted low-Ti Ujaraaluk Depleted low-Ti Ujaraaluk Enriched low-Ti Ujaraaluk Enriched low-Ti Ujaraaluk Enriched low-Ti Ujaraaluk Enriched low-Ti Ujaraaluk Enriched low-Ti Ujaraaluk Enriched low-Ti Ujaraaluk Enriched low-Ti Ujaraaluk Enriched low-Ti Ujaraaluk Enriched low-Ti Ujaraaluk Enriched low-Ti Ujaraaluk Enriched low-Ti Ujaraaluk Enriched low-Ti Ujaraaluk Enriched low-Ti Ujaraaluk Enriched low-Ti Ujaraaluk Enriched low-Ti Ujaraaluk Enriched low-Ti Ujaraaluk Enriched low-Ti Ujaraaluk Enriched low-Ti Ujaraaluk Enriched low-Ti Ujaraaluk Central gabbro Central gabbro Central gabbro Central gabbro Central gabbro Central gabbro TTG TTG TTG TTG TTG TTG TTG Felsic schist Trondhjemitic band Trondhjemitic band Trondhjemitic band Pegmatite

339 608 339 550 340 205 340 125 340 909 340 947 339 703 340 296 340 296 340 208 340 515 340 134 340 052 339 686 339 961 339 961 339 903 339 903 339 743 339 857 339 857 339 624 339 624 339 725 339 751 339 751 339 751 339 996 339 840 339 682 339 868 339 755 339 910 339 722 339 761 339 596 339 955 339 920 339 571 339 451 339 478 340 204 340 782 340 876 339 847 339 620 339 769 339 777 339 950 340 328

6 464 731 6 464 197 6 463 745 6 463 634 6 46 4059 6 464 065 6 463 852 6 464 625 6 464 625 6 464 874 6 465 139 6 464 945 6 465 381 6 463 935 6 463 911 6 463 911 6 463 800 6 463 800 6 463 670 6 463 636 6 463 636 6 463 151 6 463 151 6 463 051 6 462 983 6 462 983 6 462 983 6 464 141 6 463 627 6 463 830 6 463 412 6 463 347 6 463 978 6 463 700 6 464 257 6 463 811 6 464 530 6 464 308 6 465 080 6 464 267 6 464 268 6 463 539 6 463 174 6 462 453 6 462 816 6 464 220 6 462 773 6 462 779 6 462 935 6 464 844

Garnet no no no no no no yes yes no no no no no no no yes yes no yes yes yes yes yes yes yes yes yes no no no no

Age (Ma)

[Lu] (ppm)

[Hf](ppm)

176

Lu/177 Hf

4406 4406 4406 4406 4406 4406 4406 4406 4406 4406 4406 4406 4406 4406 4406 4406 4406 4406 4406 4406 4406 4406 4406 4406 4406 4406 4406 4406 4406 4406 4406 4406 4115 4115 4115 4115 4115 4115 3657 3657 3657 3756 3657 3508 3756 3353 3757 3757 3757 2686

0.17 0.31 0.26 0.40 0.25 0.42 0.30 0.29 0.30 0.21 0.25 0.25 0.27 0.25 0.27 0.26 0.30 0.34 0.29 0.49 0.54 0.42 0.46 0.36 0.44 0.42 0.44 0.16 0.26 0.24 0.26 0.24 0.42 0.32 0.46 0.27 0.27 0.31 0.15 0.16 0.11 0.06 0.32 0.18 0.06 0.40 0.06 0.08 0.06 0.73

1.02 1.74 1.54 2.16 1.42 2.28 0.67 0.89 0.90 0.76 0.91 0.80 0.76 1.23 1.49 1.46 1.92 1.92 1.80 1.88 1.98 2.21 2.22 1.85 1.30 1.73 1.78 2.04 1.66 1.22 1.64 1.45 1.56 1.50 2.26 2.09 0.65 1.76 4.26 4.32 4.53 3.60 3.06 3.92 3.30 6.81 3.01 4.62 3.64 2.74

0.0234 0.0255 0.0243 0.0266 0.0250 0.0263 0.0642 0.0464 0.0479 0.0391 0.0387 0.0447 0.0508 0.0295 0.0256 0.0257 0.0227 0.0251 0.0233 0.0375 0.0388 0.0273 0.0295 0.0278 0.0347 0.0347 0.0356 0.0115 0.0226 0.0280 0.0227 0.0237 0.0373 0.0311 0.0169 0.0173 0.0604 0.0252 0.0049 0.0054 0.0034 0.0025 0.0151 0.0067 0.0027 0.0084 0.0030 0.0024 0.0022 0.0381

176

Hf/177 Hf

0.282185 0.282327 0.282336 0.282483 0.282355 0.282475 0.284813 0.283726 0.283787 0.283725 0.283387 0.283671 0.284058 0.282530 0.282136 0.282296 0.282064 0.282177 0.282009 0.282765 0.282808 0.282240 0.282361 0.282316 0.282651 0.282611 0.282670 0.281488 0.281922 0.282540 0.282030 0.282041 0.283037 0.282697 0.281584 0.281602 0.284412 0.282163 0.280757 0.280722 0.280605 0.280548 0.281494 0.280906 0.280571 0.281024 0.280541 0.280534 0.280558 0.282947

±

2σ -error

176

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

4 4 3 5 3 3 4 4 7 4 2 4 4 3 6 8 6 3 4 3 5 3 4 3 5 5 5 6 3 3 4 3 8 6 11 5 7 6 4 3 4 5 3 5 4 4 6 3 4 4

0.280181 0.280139 0.280256 0.280209 0.280210 0.280221 0.279317 0.279755 0.279688 0.280373 0.280070 0.279843 0.279709 0.280001 0.279943 0.280093 0.280120 0.280029 0.280014 0.279555 0.279485 0.279905 0.279838 0.279937 0.279681 0.279640 0.279618 0.280503 0.279987 0.280141 0.280084 0.280015 0.280061 0.280219 0.280238 0.280222 0.279596 0.280154 0.280411 0.280343 0.280362 0.280366 0.280428 0.280454 0.280372 0.280485 0.280323 0.280356 0.280395 0.280991

Hf/177 Hf(i)

ε Hf(i) 9 .8 8 .3 12.5 10.8 10.8 11.2 −21.1 −5.4 −7.8 16.6 5 .8 −2.3 −7.1 3 .4 1 .3 6 .6 7 .6 4 .3 3 .8 −12.6 −15.1 − 0.1 −2.5 1 .1 −8.1 −9.6 −10.4 21.3 2 .8 8 .3 6 .3 3 .8 −1.6 4 .1 4 .8 4 .2 −18.2 1 .7 − 0.1 −2.5 −1.8 0 .7 0 .5 −2.1 0 .9 −4.7 − 0.8 0 .4 1 .8 −2.4

TDM (Ga) 3.7 3.7 3.4 3.4 3.5 3.3 3.1 3.1 2.9

3.5 3.4 4.1 4.4 3.9 3.9 4.1 4.2

4.4

3.4 4.3 3.5 4.0 4.2 3.9 4.0 4.0 2.7 4.2 3.8 3.9 3.9 3.9 3.9 3.8 3.9 3.8 3.9 3.9 3.8

J. O’Neil et al. / Earth and Planetary Science Letters 379 (2013) 13–25

Sample

J. O’Neil et al. / Earth and Planetary Science Letters 379 (2013) 13–25

were measured by TIMS at the Laboratoire Magmas et Volcans (LMV) in Clermont-Ferrand, France. Whole rocks for Hf analysis were digested in Savillex beakers and any remaining residues after centrifugation of the dissolved samples were dissolved in Parr bombs and combined with the sample after complete digestion. Zircon U–Pb analyses were done by laser ablation ICP-MS at LMV, while the zircon Hf analyses were measured by laser ablation on the Nu-Plasma at DTM. The complete analytical procedures are detailed in the supplementary data. 4. Results 4.1. Zircon U–Pb geochronology Analyzed zircons (n = 458) are 100 μm to 450 μm long and are generally prismatic with oscillatory or sector zoning (Fig. 2 supplementary data). They also contain small inclusions and/or exhibit different degrees of fracturing. Most of the analyzed zircon grains showed <10% discordance. The felsic schist (PC-134) generally contains smaller zircon grains with oscillatory zoning that is not as well developed as in zircons from the other felsic rocks. These zircons appear to represent a single population and yield a weighted average 207 Pb/206 Pb age of 3353 ± 4 Ma (n = 66) (Fig. 2) in agreement with data from David et al. (2009). Two zircon grains from the felsic schist contain an older core with 207 Pb/206 Pb ages of 3474 Ma and 3575 Ma respectively. The average Th/U ratio for zircons from the felsic schist is 0.85 ± 0.06 (2σ -mean) consistent with an igneous magmatic origin. Despite their relatively high Th/U ratios, zircons from this sample have been interpreted by David et al. (2009) to be metamorphic due to their low U concentration and absence of magmatic crystallization structures. The zircons from the same sample analyzed in this study, however, have U concentrations similar to the other TTGs zircons and, although their internal structure is not as well developed as in zircons from the other TTGs, the zircons from the felsic schist could have crystallized from a melt instead of metamorphic fluids. Zircons from TTG PC-285 are generally larger. Most grains analyzed were 250 to 450 μm in size and display a fine magmatic oscillatory zoning. This TTG sample exhibits a more complex zircon population that yields weighted average 207 Pb/206 Pb ages of 3508 ± 8 Ma (n = 20) and 3433 ± 7 Ma (n = 50) for the cores and the rims, respectively (Fig. 2). PC-285 also contains smaller populations of zircons of ∼3230 Ma and ∼3340 Ma. Zircons from this sample have higher U concentrations compared to the other analyzed TTGs, with the cores ranging from 300–400 ppm and some rims up to 1000 ppm U. The 3.51 Ga age can be interpreted as the age of primary magmatism, while the younger ages may have recorded successive magmatic and metamorphic events. Interestingly, some crystals in this sample also indicate an age of 3.34 Ga, which is similar to that obtained for the felsic schist (PC-134). TTG samples PC-284, PC-286 and the trondhjemitic band sample PC-287 each contain zircons from a single age population with weighted average 207 Pb/206 Pb ages of 3657 ± 4 Ma (n = 80), 3756 ± 4 Ma (n = 83) and 3757 ± 4 Ma (n = 80) respectively (Fig. 2). Darling et al. (in press) have obtained a similar age of 3781 ± 11 Ma for PC-286. Zircons with CL images that could suggest the presence of a core and rim were analyzed in both core and rim, but no age differences were found between rims and cores (Fig. 2 supplementary data). Zircons from these three samples have Th/U ratios consistent with an igneous origin (PC-284 = 0.74 ± 0.05, PC-286 = 0.71 ± 0.03, PC-287 = 0.58 ± 0.04; 2σ -mean). The age for the TTG sample PC-284 and for the trondhjemitic band sample PC-287 are in agreement with ages previously published on the same lithologies (Cates and Mojzsis, 2007; David et al., 2009). The high number of measurements (n > 360)

17

combined with the fact that all zircon grains have been analyzed in their center, make it improbable that no zircon populations (even smaller than 2–3%) older than 3.8 Ga could have been missed (Vermeesch, 2004). 4.2. Whole-rock Lu–Hf isotopic compositions The whole-rock Lu–Hf isotopic compositions of the NGB samples are presented in Table 1. The Ujaraaluk unit displays a wide range of 176 Lu/177 Hf (0.0115 to 0.0642) and 176 Hf/177 Hf ratios (0.281488 to 0.284813). All Ujaraaluk samples taken together define a highly scattered 176 Lu/177 Hf vs. 176 Hf/177 Hf array corresponding to an age of 3483 ± 320 Ma (MSWD = 992; Fig. 3(A)). As with the Ujaraaluk Sm–Nd systematics (O’Neil et al., 2012), the Lu– Hf array is composed of different clusters, with distinct slopes, for the different Ujaraaluk compositional units. The most distinctive group is the depleted low-Ti Ujaraaluk with significantly higher 176 Lu/177 Hf ratios compared to the other Ujaraaluk samples. Excluding one sample falling significantly off the array, the depleted low-Ti Ujaraaluk define a correlation corresponding to an age of 2995 ± 390 Ma (MSWD = 59; Fig. 3(B)). The correlation has a high initial ε Hf (+8.9) suggesting that the high Lu/Hf ratios characteristic of this group was present in these rocks long before 3 Ga. The enriched low-Ti Ujaraaluk unit appears to be divided into two populations with different Lu–Hf isotopic compositions, corresponding with the presence or absence of garnet. The samples containing garnet define the correlation with the least scatter of all Ujaraaluk groups, yielding an age of 2537 ± 90 Ma (MSWD = 24). The garnetfree enriched low-Ti Ujaraaluk and the high-Ti Ujaraaluk samples display a range in 176 Lu/177 Hf ratio too small to define a significant 176 Lu–176 Hf isochron, but align along steeper slopes similar to the 4300 to 4400 Ma ages (Fig. 3(C)) indicated by 146 Sm–142 Nd systematics of the Ujaraaluk (O’Neil et al., 2012). Samples from the central gabbro sills display a wide range of Lu–Hf isotopic compositions and define a scattered 176 Lu/177 Hf vs. 176 Hf/177 Hf array in which the best fit line corresponds to an age of 3400 ± 350 Ma (MSWD = 186; Fig. 3(D)). The 3.76 Ga trondhjemitic bands are characterized by a very limited spread in 176 Lu/177 Hf ratios with initial ε Hf values ranging from +1.8 to −0.8. The Hf isotopic compositions reported here for the trondhjemitic bands are similar to the data from Guitreau et al. (2013) (ε Hf(3.8 Ga) = +2 to −1.9) on the same lithology. Four 3.66 Ga TTG samples have been analyzed for whole rock Lu–Hf. Taken together, they scatter about a best fit line of 3917 ± 830 Ma (MSWD = 314; Fig. 3(E)). The scatter about this best fit line suggests that the Lu–Hf system has been disturbed in these samples. The 3.76 Ga TTGs and trondhjemitic bands have low 176 Lu/177 Hf ratios (0.0022–0.0030) compared to all the younger TTGs that have 176 Lu/177 Hf ratios ranging from 0.0034 to 0.0151. The older TTGs have similar REE profiles to the trondhjemitic bands, displaying more prominent heavy REE (HREE) depletion compared to younger TTGs, whereas the felsic schist has a similar REE profile to the 3.66 Ga TTGs but with higher trace element concentrations (Fig. 4). All NGB felsic rocks have Hf model ages (TDM Hf) between 3.8 and 3.9 Ga independent of their younger (up to 470 Ma younger) crystallization ages spanning more than 400 Ma, suggesting derivation from an older precursor. 4.3. Zircon Lu–Hf isotopic compositions Only zircons from the NGB felsic rocks with less than 3% discordance were analyzed for Lu–Hf systematics. The results are summarized in Table 2 and fully reported in supplementary data Table 4. Initial ε Hf values for each zircon grain have been calculated to their respective measured 207 Pb/206 Pb ages. Zircons from the 3.76 Ga TTG and trondhjemitic band samples have similar

18

J. O’Neil et al. / Earth and Planetary Science Letters 379 (2013) 13–25

Fig. 2. U–Pb Concordia diagrams and relative probability for the different population of zircons. (A) Felsic schist PC-134, (B) TTG PC-284, (C) TTG PC-285, (D) TTG PC-286, (E) trondhjemitic band PC-287. Grey ellipses on U–Pb Concordia diagrams are for zircons not included in the age calculation, except PC-285 where the 2 shades of grey ellipses are for the 2 main zircon populations and the open ellipses are for zircons not included in the age calculations. Red bars on histograms are the average ages for the main zircon populations. (For interpretation of the references to color in this figure, the reader is referred to the web version of this article.)

J. O’Neil et al. / Earth and Planetary Science Letters 379 (2013) 13–25

19

Fig. 2. (continued)

initial ε Hf values ranging from +1.3 to −4.9 and from +2.3 to −4.7 respectively, with most zircon grains yielding subchondritic ε Hf values at 3.76 Ga. Zircons from the 3.66 Ga TTG have initial ε Hf values ranging from −3.0 to −7.1, whereas zircons from the 3.4–3.5 Ga TTG exhibit initial ε Hf values ranging from −3.2 to −12.0. This TTG sample contains a complex zircon population and the initial zircon ε Hf values within this TTG also seem to decrease with zircon ages. The youngest sample, the 3.35 Ga felsic schist, yields relatively low initial zircon ε Hf values ranging from −3.4 to −12.0. 5. Discussion 5.1. Geochronology of the NGB TTGs Despite its relatively small size, the NGB includes lithologies spanning more than one billion years in age that record multiple thermal events responsible for producing several generations of TTG. The oldest NGB felsic rocks are the trondhjemitic bands with an age of 3.76 Ga. They are found in a restricted area in the NGB often within gabbro sills. Their occurrences in the field and their 142–143 Nd isotopic compositions led O’Neil et al. (2012) to propose that they could have been derived from the partial melting of the gabbros, representing leucosomes within older ∼4.1 Ga mafic

rocks. This is consistent with the common gneissic texture within these gabbros displaying pyroxene/amphibole rich layers that may represent residual melanosomes (Fig. 3 supplementary data). The TTG located at the southern contact, inside the NGB major fold, also is 3.76 Ga (Fig. 1). The geochemical and isotopic similarities between this dated TTG (PC-286) and the TTG sample located further north, along the same contact (PC-169), suggest that the inward contact of the NGB is bordered by a TTG contemporaneous to the production of the trondhjemitic bands (Fig. 1). The 3.76 Ga TTG suggests that this thermal event produced a greater volume of felsic magma than previously suggested by the scarcity of the thin localized contemporaneous intrusive trondhjemites. This border TTG also has similar geochemical features to the trondhjemitic bands suggesting, along with the identical ages, that they could have been produced from similar processes and perhaps the same precursor. The TTG PC-286, however, displays a significantly lower 142 Nd/144 Nd ratio compared to the trondhjemitic bands (O’Neil et al., 2012) and this 142 Nd deficit (−16.2 ppm compared to Nd standard) would be difficult to produce simply by partial melting of the gabbro sills as the analyzed sills have smaller 142 Nd deficits (>−9 ppm compared to Nd standard). The TTG to the West of the NGB was previously dated at 3.66 Ga by U–Pb in zircons (David et al., 2009) and yielded a similar 147 Sm–143 Nd age (O’Neil et al., 2012). The TTG sample PC-284

20

Fig. 3. 176 Lu/177 Hf vs. TTGs only.

J. O’Neil et al. / Earth and Planetary Science Letters 379 (2013) 13–25

176

Hf/177 Hf isochron diagrams for (A), (B) and (C) the Ujaraaluk unit, (D) the gabbros and (E) the TTGs. The isochron data for the TTGs is for the 3.66 Ga

located within the main fold of the NGB mafic rocks (Fig. 1) yields an identical age of 3657 ± 4 Ma. This suggests that the 3.66 Ga thermal event may have been the most prominent episode of felsic crust formation from the partial melting of pre-existing mafic rocks in the NGB. The NGB also comprises younger felsic rocks. PC-285 is located at the southwest edge of the NGB (Fig. 1). Its zircon population suggests felsic magma generation at ∼3.5 Ga followed by a complex thermal history for the NGB that produced a younger population of zircons (Fig. 2(B)). The small population of zircons in PC-285 that yield a younger age of ∼3.34 Ga is coeval with the formation of the felsic schist at 3.35 Ga. This suggests production of felsic magmas in the Paleoarchean, although TTGs of this age seem to be only a minor component of the NGB. At least five thermal events thus have been recorded in the NGB TTGs covering almost 400 Ma of Earth history and the geochronological data suggest that these episodes of felsic magma generation occurred

relatively regularly every 100–150 Ma throughout the Eoarchean and Paleoarchean (3.8 Ga to 3.3 Ga). After that, no thermal events were recorded in the NGB for a period of ∼500 Ma, before the intrusion of the Neoarchean pegmatites and regional upper amphibolite metamorphism. 5.2. Post-magmatic disturbance of the whole-rock Lu–Hf system O’Neil et al. (2012) showed that the long-lived 147 Sm–143 Nd system was affected by Neoarchean metamorphism/metasomatism in the NGB rocks. The significant scatter on every Lu–Hf isochron diagram (Fig. 3) and the wide range in initial Hf isotopic compositions (Table 1) suggests that the Lu–Hf isotopic system in the NGB rocks also has been affected at the whole rock scale, perhaps even more than the Sm–Nd system. The different Ujaraaluk groups display a wide range of ε Hf values at any age except for

J. O’Neil et al. / Earth and Planetary Science Letters 379 (2013) 13–25

Fig. 4. Chondrite-normalized REE profiles for the Nuvvuagittuq felsic rocks. Symbols are as for Fig. 3 and open hexagons are for the felsic schist. Blue and red areas are model magmas derived from melting of garnet-free and garnet-bearing amphibolites respectively (Martin, 1987). The top of the colored areas is for 10% melting and the bottom is for 50% melting. The light grey line is the source composition used by Martin (1987) and the dark grey lines are representative primitive Ujaraaluk compositions. Normalizing chondrite values are from Sun and McDonough (1989). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

the garnet-bearing Ujaraaluk with a tight cluster of ε Hf values at 2537 Ma (Fig. 4 supplementary data), which is expected from the 176 Lu–176 Hf isochron they define (Fig. 3(C)). Garnet can be a major constituent of the NGB mafic rocks and its effect on the Lu–Hf isotopic system of the Ujaraaluk unit is evident. Fig. 3(C) shows that rocks from the enriched low-Ti Ujaraaluk define two distinct arrays based on their garnet contents. Garnet-bearing samples define the Lu–Hf isochron with the least scatter. The slope of this isochron corresponds to a Neoarchean age contemporaneous with the Sm– Nd age of the garnets (O’Neil et al., 2012). The very low initial ε Hf of this isochron (−7.5 at 2537 Ma) is consistent with this age reflecting resetting of the Lu–Hf system in a much older rock with a

21

relatively low Lu/Hf ratio. The poorly defined Lu–Hf trends of the garnet-free rocks are more consistent with the Hadean 142 Nd age for these rocks (O’Neil et al., 2012). The variation in initial ε Hf values of the whole rocks and zircons from the TTGs also reflects the disturbance of the Lu–Hf system at the whole-rock scale. The initial ε Hf values for the whole rock TTG samples are systematically higher than their zircon ε Hf values (Tables 1 and 2). Zircons are chemically stable and less prone to alteration than the other rock-forming minerals in the TTG whole rocks. The strong affinity of Hf for zircon and the fact that zircon contains almost no Lu, also makes it very difficult to affect their Lu–Hf isotope systematics. In contrast, if the Lu/Hf ratios of the whole rocks were lowered by only ∼10% by metamorphism occurring anytime significantly after rock formation, this would lead the calculated whole rock initial ε Hf to exceed those of the zircons by the amounts observed. Given the amount of Lu–Hf disturbance present in the NGB mafic rocks and the demonstrated sensitivity (Fig. 3(C)) of this system to garnet growth, we suggest that the chronological information from Lu–Hf, particularly for the mafic rocks of the NGB, must be interpreted with considerable caution. 5.3. Nuvvuagittuq crustal evolution and implications for Earth’s early crust Archean felsic crust is dominated by rocks from the Tonalite– Trondhjemite–Granodiorite series that likely were produced from the partial melting of older basaltic crust (Moyen and Martin, 2012 and references therein). Fig. 5 shows the Hf isotopic evolution of the Nuvvuagittuq TTG zircons through time. Almost all the zircons have subchondritic initial ε Hf values except for a few zircons from the 3.76 Ga rocks. Zircon ε Hf values also decrease as their host rocks get younger. Most zircons analyzed from the 3.35 Ga to 3.66 Ga TTGs in this study have initial ε Hf values within the array corresponding to the evolution of a ∼4.4 Ga precursor with

Fig. 5. Initial ε Hf isotopic evolution through time for the NGB zircons. Circles are for zircons from TTGs and the felsic schist and triangles are for zircons from the trondhjemitic band. Grey symbols are for 3.76 Ga rocks and white symbols are for 3.66 Ga rocks. Yellow stars are the ε Hf modes for each TTG sample calculated at their respective whole-rock U–Pb age. The error bars on the stars represent the total range of zircon ε Hf values calculated at the same U–Pb ages. The blue square is the Lu–Hf age and initial ε Hf data from Guitreau et al. (2013). Most of the NGB zircons fall between the evolution lines of 176 Lu/177 Hf = 0.020 (trend from the “least disturbed” Jack Hills zircons with chondritic Hf at ∼4430 Ma; Kemp et al., 2010) and 176 Lu/177 Hf = 0.026 (average MORB; Blichert-Toft and Albarède, 2008) derived from the depleted mantle curve at 4400 Ma. The evolution line of 176 Lu/177 Hf ratio = 0.029 represents the average oceanic plateau (Blichert-Toft and Albarède, 2008). The Jack Hills detrital zircon area includes data from Amelin et al. (1999), Harrison et al. (2005, 2008) and Blichert-Toft and Albarède (2008). The “least disturbed” Jack Hills area is from Kemp et al. (2010). Acasta data is from Iizuka et al. (2009). The depleted mantle evolution line corresponds to a 176 Lu/177 Hf ratio = 0.03915 starting with a chondritic 176 Hf/177 Hf at 4.567 Ga. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

J. O’Neil et al. / Earth and Planetary Science Letters 379 (2013) 13–25

Total ε Hf(i) range

−12.3 to −3.7 −7.1 to −1.3 −9.7 to −3.0 −3.6 to +1.0 −4.0 to +1.8 −7.5 −4.5 −6.4 −1.6 −1.0 0.0013–0.0023 0.00008–0.1078 0.0000459 (AVG) 0.000005–0.005054 0.282498 (AVG) 0.282127 (AVG)

0.0175–0.0537 0.0155–0.0498 0.0125–0.0451 0.0074–0.0232 0.0009–0.0323 0.00059–0.00273 0.00068–0.00210 0.00007–0.00186 0.00031–0.00089 0.00026–0.00128 0.280320–0.280563 0.280300–0.280419 0.280367–0.280508 0.280296–0.280431 0.280263–0.280459

Zircon Synthetic zircons MudTank STD MUNZirc (0, 1, 2) STD

39 14

3353 3657 3508 3756 3757 Felsic schist TTG TTG TTG Trond. band PC-134 PC-284 PC-285 PC-286 PC-287

26 38 36 45 44

Lithology

nb zircon (n)

Age (Ma)

Range of measured 176 Hf/177 Hf

Range of measured 176 Lu/177 Hf

176

Range of measured Yb/177 Hf

Modal

0.280406 0.280287 0.280334 0.280302 0.280318

177

Hf/176 Hf(i)

Modal ε Hf(i)

176

Sample

Table 2 Summarized Lu–Hf isotopic data for the NGB zircons. Complete Lu–Hf data for single zircons are in supplementary data Table 4. Range values are the lowest and highest measured values for each rock sample. 176 Hf/177 Hf(i) and ε Hf(i) values are calculated at the same U–Pb age for each rock sample. Histograms to determine the mode values are presented in Fig. 5 supplementary data. Values used for ε Hf calculations are 176 Hf/177 HfCHUR = 0.282785, 176 Lu/177 HfCHUR = 0.0336 (Bouvier et al., 2008) and decay constant for 176 Lu = 1.865 × 10−11 yr−1 (Scherer et al., 2001). Values marked “AVG” are average measured values.

22

Lu/177 Hf ratio characteristic of a mafic crust like the Ujaraaluk unit (Fig. 5). The yellow stars on Fig. 5 are the ε Hf values of the mode for each TTG sample (Table 2, Fig. 5 supplementary data) calculated at their respective whole-rock U–Pb age with the error bars on the stars representing the total range of zircon ε Hf values calculated at the same U–Pb ages. The trend defined by the 3.35 Ga to 3.66 Ga TTGs modal ε Hf values is also consistent with derivation from a Hadean mafic crust. As discussed in Section 5.2, the whole-rock Lu–Hf isotopic compositions of the NGB whole rocks seem to have been affected by post-magmatic processes, primarily the metamorphic event that resulted in garnet growth. Because of garnet’s ability to strongly fractionate Lu from Hf, the garnet-free samples most likely provide the best estimate on the original Lu–Hf systematics of the Ujaraaluk unit, assuming that garnet growth did not occur isochemically at the whole-rock scale. The garnet-free Ujaraaluk have a restricted range of 176 Lu/177 Hf ratios (0.0226 to 0.0295) that would lead to Hf isotopic compositions similar to those measured in the TTG zircons (Fig. 5). Therefore, a mafic Hadean crustal basement with Lu–Hf systematics similar to the Ujaraaluk is a likely progenitor of the NGB felsic rocks. The deficits in 142 Nd of the TTGs also indicate their derivation by melting of an light REEenriched Hadean component, and again, a crustal basement compositionally and temporally similar to the Ujaraaluk has appropriate characteristics to serve as the TTG source. The TTGs were formed after the extinction of 146 Sm, thus their low 142 Nd/144 Nd ratios were inherited from their source rocks and their Sm/Nd ratios changed during partial melting such that the TTGs now display no correlation between their Sm/Nd and 142 Nd/144 Nd ratios and fall to the low Sm/Nd side of the Ujaraaluk correlation (Fig. 6 supplementary data). Experimental data also have shown that partial melting of rocks compositionally similar to the Ujaraaluk unit produces melts that are consistent with the Nuvvuagittuq TTGs compositions (Adam et al., 2012). Although the isotopic composition of the NGB TTGs does not constrain the age of the Ujaraaluk unit itself, a 4.4 Ga Ujaraaluk-like crustal basement has all the characteristics (geochemical composition, 176 Lu/177 Hf, 147 Sm/144 Nd, 142 Nd/144 Nd needed to produce the 3.66 Ga NGB TTGs. If indeed the Eoarchean NGB TTGs are derived from a 4.3–4.4 Ga Ujaraaluk-like mafic crust, it would imply that this early mafic precursor did not remelt to produce felsic magmas for a period of ∼500 Ma. This long quiescence period is unusual for what is observed in most Archean cratons, but is consistent with the scarcity of Hadean zircons. Despite the fact that we have evidence for >4.3 Ga primitive crust on Earth (e.g. Wilde et al., 2001; Kemp et al., 2010), zircons older than 4.0 Ga are only found in the Jack Hills conglomerates and a few zircon cores in the Acasta gneisses (Iizuka et al., 2006). Only after ∼3.8 Ga does the zircon record become more prominent. The zircon age distribution in itself suggests that the mafic primitive crust did not produce a significant amount of felsic magma for ∼500 Ma, which is consistent with the proposed NGB evolution. The volumetrically most important felsic unit in the Nuvvuagittuq region (the 3.6 Ga TTG) suggests a petrogenesis by remelting of a 4.3–4.4 Ga crustal source similar in composition to the Ujaraaluk unit. The 3.76 Ga TTG and trondhjemitic band however yield initial zircon ε Hf values that overlap the Hf isotope evolution trend of the Hadean mafic precursor, but with most ε Hf values falling above it (Fig. 5). This could suggest a different protolith for these older TTGs. Fig. 5 also shows that the evolution of a 4.1 Ga mafic rock with a 176 Lu/177 Hf ratio of 0.026 (average ratio for the NGB gabbros, excluding one sample with a significantly higher 176 Lu/177 Hf ratio) could explain the higher ε Hf zircon values for the 3.76 Ga TTGs. One could argue that the 3.66 Ga TTGs were produced by remelting of the 3.76 Ga felsic rocks instead of from a Hadean mafic precursor. The 3.66 Ga TTGs, however,

J. O’Neil et al. / Earth and Planetary Science Letters 379 (2013) 13–25

23

do not show the heavy-REE depletion characteristic of the older TTG’s and hence have relatively higher 176 Lu/177 Hf ratios (0.0034 to 0.0151) which makes it unlikely that the younger TTG’s are derived by melting of the older felsic rocks. Alternative models put a maximum age of ∼3.8 Ga on the NGB mafic lithologies (Cates et al., 2013; Roth et al., 2013; Guitreau et al., 2013). Guitreau et al. (2013) proposed an age of 3864 Ma for the NGB rocks based on a Lu–Hf isochron for all lithologies in the NGB, including separated zircons and quartzites that yielded an initial ε Hf value of +0.4. An NGB TTG precursor starting at 3.86 Ga with ε Hf = +0.4 would require a 176 Lu/177 Hf ratio of ∼0.0015 to reach the initial ε Hf of the <3.66 Ga TTGs, which is significantly lower than for any measured Ujaraaluk or TTG samples (Table 1). The rate of change of ε Hf with time for the 3.66 Ga NGB felsic rocks is also inconsistent with a very low Lu/Hf ratio precursor and instead suggests a progenitor with Lu/Hf ratios within the range of the ∼4.4 Ga Ujaraaluk-like mafic crust. A 3.8 Ga age for the Ujaraaluk was argued to be supported by Hf–142–143 Nd mixing hyperbolas within the NGB amphibolites (Guitreau et al., 2013). The mixing curves shown by Guitreau et al. (2013) however are defined by 6 samples (including only 2 Ujaraaluk) with the enriched end member being a rock collected more than 5 km outside the NGB without any clearly established geological relationship to the NGB. Fig. 6 shows similar plots of 176 Hf/177 Hf(3.8 Ga) vs. 143 Nd/144 Nd(3.8 Ga) and 176 Hf/177 Hf(3.8 Ga) vs. 142 μ Nd presented by Guitreau et al. (2013) for an extended set of Ujaraaluk and gabbro data. The dark grey and green lines in Fig. 6 show the Hf–Nd correlations expected for a compositionally diverse suite of rocks formed at 4.4 Ga that have Sm/Nd and Lu/Hf ratios that would result in a Nd–Hf isotope correlation with slope similar to that seen in old crustal rocks (Vervoort and Patchett, 1996). In all parts of Fig. 6, the expected Nd–Hf isotopic variability at 3.8 Ga in compositionally diverse rocks formed at 4.4 Ga reproduces the observed range of variability in Nd and Hf isotopic composition of the NGB rocks at 3.8 Ga. In other words, the general isotopic trends interpreted by Guitreau et al. (2013) as mixing lines also would be expected for a compositionally diverse suite of mafic rocks, like the Ujaraaluk, if they formed at 4.4 Ga. 5.4. Petrogenesis of the Nuvvuagittuq TTGs The Nuvvuagittuq felsic rocks include TTG that show a wide range in slopes of their REE patterns (Fig. 4). The HREE-depletion characteristic of the older 3.76 Ga TTGs is consistent with partial melting of a mafic parent with residual garnet (e.g. Halla et al., 2009; Moyen, 2011), whereas the younger Nuvvuagittuq TTGs appear to have been formed from melting of a garnet-poor precursor. The source amphibolite used by Martin (1987) to model the origin of TTGs is an Archean tholeiite with a similar trace element composition to typical primitive Ujaraaluk (Fig. 4) suggesting that melting of Ujaraaluk-like progenitors at different pressures (i.e. with variable amounts of residual garnet) can explain the difference in HREE between the 3.76 Ga and 3.66 Ga Nuvvuagittuq TTGs. This may suggest that the depth of partial melting shallowed with time with the 3.76 Ga TTGs forming at higher pressures than the 3.66 Ga TTGs, assuming a common source for both TTG types. Variable amounts of garnet in the source rocks also could explain the differences in Hf isotopic compositions between the two types of Nuvvuagittuq TTGs if the garnet-forming event resulted in a group of source rocks with distinct Lu/Hf ratios long enough before TTG genesis to evolve distinct Hf isotopic compositions. The garnet-rich Ujaraaluk show a wider range and tendency to higher Lu/Hf ratios than the garnet-free samples. Consequently, the wider range and trend to higher ε Hf of the 3.76 Ga TTG/trondhjemitic bands is consistent with a source similar to the garnetrich Ujaraaluk. Because the 3.66 Ga TTG melts were derived from

176 Fig. 6. (A) Hf/177 Hf vs. μ142 Nd (μ142 Nd = ((142 Nd/144 Nd)Sample / (142 Nd/144 Nd)JNdi −1)×106 ), (B) 176 Hf/177 Hf vs. 143 Nd/144 Nd and (C) 143 Nd/144 Nd vs. μ142 Nd for the NGB Ujaraaluk unit and gabbro. The 176 Hf/177 Hf and 143 Nd/144 Nd ratios are calculated at 3.8 Ga. Symbols are as for Fig. 3. Blue crosses and line are data and mixing curve proposed to fit these data, by Guitreau et al. (2013). Dark grey and green lines represent the isotopic composition at 3.8 Ga of a 4.4 Ga mafic crust derived from a chondritic reservoir (dark grey) or an early depleted reservoir (green; EDR from Boyet and Carlson, 2005) including samples with variable 147 Sm/144 Nd ratios as indicated by the numbers on the lines. The Lu/Hf ratios for these samples are calculated from the Sm/Nd ratios in order to produce a Hf–Nd isotope correlation with the slope measured for old continental crust by Vervoort and Patchett (1996). The EDR line is almost identical to the chondritic line on the diagram (B). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

24

J. O’Neil et al. / Earth and Planetary Science Letters 379 (2013) 13–25

a garnet-poor source, the Hf isotope evolution of their source may have been more similar to that of the garnet-free Ujaraaluk generally characterized by lower Lu/Hf ratios than the garnet-bearing Ujaraaluk. The point here is not to suggest that the Ujaraaluk unit itself is the source of the TTG’s, but that the compositional and mineralogical characteristics seen in the Ujaraaluk overlap those required for the source of the two chronologically and compositionally distinct groups of TTGs. Garnet formation in the Ujaraaluk is dated in the Neoarchean by both Sm–Nd and Lu–Hf, so the exposed Ujaraaluk mafic rocks would not carry the “garnet signature” during the 3.6 to 3.8 Ga formation of the TTGs. Nevertheless, the composition of the Ujaraaluk obviously is suitable to stabilize variable amounts of garnet, given the right pressure and temperature conditions. Consequently, a mafic crust with the compositional spectrum seen in the Ujaraaluk, with garnet formation occurring in the Hadean, provides a possible source model to explain the compositional and isotopic characteristics of the Nuvvuagittuq TTGs. The Lu–Hf isotopic composition of the Jack Hills detrital zircons have been used to constrain the evolution of the early crust (Kemp et al., 2010; Blichert-Toft and Albarède, 2008; Harrison et al. 2005, 2008; Amelin et al., 1999). Most Jack Hills detrital zircons have subchondritic initial Hf (Fig. 5). Similarly, the majority of NGB zircons have subchondritic initial Hf and no zircons from this study have yielded initial Hf more radiogenic than the depleted mantle. A large proportion of the Hadean and Eoarchean detrital Jack Hills zircons also plot below the array defined by the evolution of the NGB zircons, perhaps suggesting a different composition and/or a different age for their early crustal precursor. However, the “least disturbed” Jack Hills Hadean zircon (Kemp et al., 2010) array points to a precursor with a 176 Lu/177 Hf ratio consistent with derivation from a basaltic crust similar to the NGB mafic precursor but perhaps slightly older. The initial Hf isotopic compositions of the Acasta zircons (Iizuka et al., 2009) generally overlap with the NGB array and the same trend is also observed for the younger (∼3.3 Ga to 3.7 Ga) Jack Hills detrital zircons. This suggests that the Jack Hills mafic protocrust and the crustal precursor for the Acasta gneisses may have been similar in composition to the Ujaraaluk unit. The elevated δ 18 O compositions of the Jack Hills zircons also suggest that the protolith for their host magma was altered by low temperature interactions with the hydrosphere (Mojzsis et al., 2001; Wilde et al., 2001; Valley et al., 2002; Cavosie et al., 2005). Similarly, O’Neil et al. (2011, 2012) interpret the NGB Ujaraaluk unit as a ∼4.4 Ga hydrothermally altered basaltic volcanic sequence. Both the Jack Hills and the NGB basaltic protocrust, as well as the Acasta precursor appear to have been reworked for an extended period of time, producing TTGs over several hundred million years. Earth’s primitive crust was thus most likely a >4.3 Ga hydrothermally altered basaltic crust that remelted to generate the younger Nuvvuagittuq TTGs. The same type of mafic primordial crust could also have produced the Acasta gneisses, as well as the TTGs that served as sources for the Jack Hills Eoarchean/Hadean zircons. 6. Conclusion Uranium-lead zircon geochronology on the Nuvvuagittuq TTGs indicates multiple episodes of felsic crust formation at 3.76 Ga, 3.66 Ga, 3.5–3.4 Ga and 3.35 Ga. The secular Hf isotopic evolution of the zircons from the TTGs is consistent with derivation from a 4.3–4.4 Ga mafic crust of similar composition to the Ujaraaluk mafic rocks. The 3.76 Ga trondhjemitic bands show higher initial ε Hf suggestive of a different source, perhaps either the 4.1 Ga gabbros or garnet-rich varieties of the Ujaraaluk. The Nuvvuagittuq belt also contains both low pressure and high pressure TTGs suggesting a change in the depth and mechanism of their formation between 3.76 and 3.66 Ga. Even though the Lu–Hf system of the

NGB mafic rocks seems to be disturbed by the regional Neoarchean metamorphism, the 176 Lu/177 Hf ratios of the garnet-free Ujaraaluk are in accord with the trend defined by the Hf isotopic composition of the TTG zircons. The 142 Nd composition of the Eoarchean TTGs also is consistent with derivation from the Ujaraaluk unit. Therefore, the Nuvvuagittuq belt appears to have recorded over one billion years of early crust formation consisting of the formation of a Hadean enriched basaltic crust that remelted to produce TTGs over a period of ∼500 Ma. The crustal history recorded in the NGB seems to be similar to the early crustal evolution recorded in the Jack Hills Hadean zircons where reworking of an enriched basaltic precursor produced TTG-type magmatism sustained for several hundred million years. The Nuvvuagittuq Ujaraaluk unit may therefore represent the first protocrust to stabilize after the moon-forming impact and be the closest analogue to the primordial crustal source of the Jack Hills zircons. Acknowledgements We thank the people of Inukjuak for their hospitality and people from the Pituvik Landholding Corporation, especially Mike Carroll, Johnny Mina, Minnie Palliser, Simeonie Elyasiapik, Arthur Elyasiapik, Aliva Epoo, Simeonie, Minnie Nowkawalk and Noah Echalook, for their critical logistical support in the field. We acknowledge C. Bosq and D. Auclair for support in the LMV chemistry lab and Timothy Mock for assistance with the LA-ICP-MS at DTM. We also thank J.-M. Hénot and B. Devouard for assistance with CL imaging on the SEM. We acknowledge M. Majnoon and W. Minarik for providing the pegmatite sample. The manuscript benefited from useful discussions with H. Martin and H. Rizo and was greatly improved after comments from A. Kemp, an anonymous reviewer and the editor T. Elliott. We also thank J. Vervoort for useful technical advices about LA-ICP-MS Hf-zircon measurements and C. Fisher who provided synthetic Hf-REE doped zircons for analytical standards. This work was supported by the Blaise Pascal University and the European Research Council under the European Community’s Seventh Framework Program (FP7/2007–2013 Grant Agreement No. 209035), the National Science Foundation [NSF-EAR-0910442 to R.W.C.] and the Carnegie Canada Foundation. This is Laboratory of Excellence ClerVolc contribution No. 42. Appendix A. Supplementary material Supplementary material related to this article can be found online at http://dx.doi.org/10.1016/j.epsl.2013.07.030. References Adam, J., Rushmer, T., O’Neil, J., Francis, D., 2012. Hadean greenstones from the Nuvvuagittuq fold belt and the origin of the Earth’s early continental crust. Geology 40 (4), 363–366. Amelin, Y., Lee, D.-C., Halliday, A.N., Pidgeon, R.T., 1999. Nature of the Earth’s earliest crust from hafnium isotopes in single detrital zircons. Nature 399, 252–255. Blichert-Toft, J., Albarède, F., 2008. Hafnium isotopes in Jack Hills zircons and the formation of the Hadean crust. Earth Planet. Sci. Lett. 265, 686–702. Bouvier, A., Vervoort, J.D., Patchett, J.P., 2008. The Lu–Hf and Sm–Nd isotopic composition of CHUR: Constraints from unequilibrated chondrites and implications for the bulk composition of terrestrial planets. Earth Planet. Sci. Lett. 273, 48–57. Boyet, M., Carlson, R.W., 2005. 142 Nd evidence for early (>4.53 Ga) global differentiation of the silicate Earth. Science 309, 576–581. Cates, N.L., Mojzsis, S.J., 2007. Pre-3750 Ma supracrustal rocks from the Nuvvuagittuq supracrustal belt, northern Quebec. Earth Planet. Sci. Lett. 255, 9–21. Cates, N.L., Ziegler, K., Schmitt, A.K., Mojzsis, S.J., 2013. Reduced, reused and recycled: Detrital zircons define a maximum age for the Eoarchean (ca. 3750–3780 Ma) Nuvvuagittuq Supracrustal Belt, Québec (Canada). Earth Planet. Sci. Lett. 362, 283–293.

J. O’Neil et al. / Earth and Planetary Science Letters 379 (2013) 13–25

Cavosie, A.J., Valley, J.W., Wilde, S.A., 2005. Magmatic δ 18 O in 4400–3900 Ma detrital zircons: A record of the alteration and recycling of crust in the Early Archean. Earth Planet. Sci. Lett. 235, 663–681. Darling, J.R., Moser, D.E., Heaman, L.M., Davis, W.J., O’Neil, J., Carlson, R., in press. Eoarchean to Neoarchean evolution of the Nuvvuagittuq supracrustal belt: New insights from U–Pb zircon geochronology. Amer. J. Sci. David, J., Godin, L., Stevenson, R., O’Neil, J., Francis, D., 2009. U–Pb ages (3.8–2.7 Ga) and Nd isotope data from the newly identified Eoarchean Nuvvuagittuq supracrustal belt, Superior Craton, Canada. Geol. Soc. Am. Bull. 121, 150–163. Guitreau, M., Blichert-Toft, J., Mojzsis, S.J., Roth, A.S.G., Bourdon, B., 2013. A legacy of Hadean silicate differentiation inferred from Hf isotopes in Eoarchean rocks of the Nuvvuagittuq supracrustal belt (Québec, Canada). Earth Planet. Sci. Lett. 362, 171–181. Hallaa, J., van Hunen, J., Heilimo, E., Hölttä, P., 2009. Geochemical and numerical constraints on Neoarchean plate tectonics. Precambrian Res. 174, 155–162. Harrison, T.M., Blichert-Toft, J., Muller, W., Albarède, F., Holden, P., Mojzsis, S.J., 2005. Heterogeneous Hadean Hafnium: Evidence for continental crust at 4.4 to 4.5 Ga. Science 310, 1947–1950. Harrison, T.M., Schmitt, A.K., McCulloch, M.T., Lovera, O.M., 2008. Early (4.5 Ga) formation of terrestrial crust: Lu–Hf, δ 18O, and Ti thermometry results for Hadean zircons. Earth Planet. Sci. Lett. 268, 476–486. Iizuka, T., Horie, K., Komiya, T., Maruyama, S., Hirata, T., Hidaka, H., Windley, B.F., 2006. 4.2 Ga zircon xenocryst in an Acasta gneiss from northwestern Canada: Evidence for early continental crust. Geology 34, 245–248. Iizuka, T., Komiya, T., Johnson, S.P., Kon, Y., Maruyama, S., Hirata, T., 2009. Reworking of Hadean crust in the Acasta gneisses, northwestern Canada: Evidence from in-situ Lu–Hf isotope analysis of zircon. Chem. Geol. 259, 230–239. Kemp, A.I.S., Wilde, S.A., Hawkesworth, C.J., Coath, C.D., Nemchin, A., Pidgeon, R.T., Vervoort, J.D., DuFrane, S.A., 2010. Hadean crustal evolution revisited; new constraints from Pb–Hf isotope systematics of the Jack Hills zircons. Earth Planet. Sci. Lett. 296, 45–56. Martin, H., 1987. Petrogenesis of Archaean trondhjemites, tonalites and granodiorites from eastern Finland; major and trace element geochemistry. J. Petrol. 28, 921–953. Mojzsis, S.J., Harrison, T.M., Pidgeon, R.T., 2001. Oxygen-isotope evidence from ancient zircons for liquid water at the Earth’s surface 4,300 Myr ago. Nature 409, 178–181.

25

Moyen, J.-F., 2011. The composite Archaean grey gneisses: Petrological significance, and evidence for a non-unique tectonic setting for Archaean crustal growth. Lithos 123, 21–36. Moyen, J.-F., Martin, H., 2012. Forty years of TTG research. Lithos 148, 312–336. O’Neil, J., Maurice, C., Stevenson, R.K., Larocque, J., Cloquet, C., David, J., Francis, D., 2007. The geology of the 3.8 Ga Nuvvuagittuk (Porpoise Cove) greenstone belt, northern Superior Province, Canada. In: Kranendonk, M.J., Smithies, R.H., Bennett, V.C. (Eds.), Earth’s Oldest Rocks. Elsevier BV, Amsterdam, pp. 219–250. O’Neil, J., Carlson, R.W., Francis, D., Stevenson, R.K., 2008. Neodymium-142 evidence for Hadean mafic crust. Science 321, 1828–1831. O’Neil, J., Francis, D., Carlson, R.W., 2011. Implications of the Nuvvuagittuq greenstone belt for the formation of Earth’s early crust. J. Petrol. 52, 985–1009. O’Neil, J., Carlson, R.W., Paquette, J.-L., Francis, D., 2012. Formation age and metamorphic history of the Nuvvuagittuq greenstone belt. Precambrian Res. 220–221, 23–44. Roth, A.S.G., Bourdon, B., Mojzsis, S.J., Touboul, M., Sprung, P., Guitreau, M., BlichertToft, J., 2013. Inherited 142Nd anomalies in Eoarcheanprotoliths. Earth Planet. Sci. Lett. 361, 50–57. Scherer, E.E., Münker, C., Mezger, K., 2001. Calibration of the Lutetium–Hafnium clock. Science 293, 683–687. Simard, M., Parent, M., David, J., Sharma, K.N.M., 2003. Géologie de la région de la rivière Innuksuac (34K et 34L). Ministère des Ressources naturelles, Québec; RG 2002-10, 46 pp. Sun, S.S., McDonough, W.F., 1989. Chemical and isotopic systematics of oceanic basalts; implications for mantle composition and processes. In: Saunders, A.D., Norry, M.J. (Eds.), Magmatism in the Ocean Basins. Geological Society of London, London, pp. 313–345. Valley, J.W., Peck, W.H., King, E.M., Wilde, S.A., 2002. A cool early Earth. Geology 30, 351–354. Vermeesch, P., 2004. How many grains are needed for a provenance study? Earth Planet. Sci. Lett. 224 (3–4), 441–451. Vervoort, J.D., Patchett, P.J., 1996. Behavior of hafnium and neodymium isotopes in the crust: Constraints from Precambrian crustally derived granites. Geochim. Cosmochim. Acta 60, 3717–3733. Wilde, S.A., Valley, J.W., Peck, W.H., Graham, C.M., 2001. Evidence from detrital zircons for the existence of continental crust and oceans on the Earth 4.4 Gyr ago. Nature 409, 175–178.