Contributions to Mineralogy and Petrology

, Volume 161, Issue 6, pp 1027–1050

Archaean fluid-assisted crustal cannibalism recorded by low δ18O and negative εHf(T) isotopic signatures of West Greenland granite zircon

Authors

    • Research School of Earth SciencesAustralian National University
    • Division of Earth and Environmental SciencesKorea Basic Science Institute
    • NERC Isotope Geosciences LaboratoryBritish Geological Survey Keyworth
  • Vickie C. Bennett
    • Research School of Earth SciencesAustralian National University
  • Allen P. Nutman
    • Research School of Earth SciencesAustralian National University
    • School of Earth and Environmental SciencesUniversity of Wollongong
  • Ian S. Williams
    • Research School of Earth SciencesAustralian National University
Original Paper

DOI: 10.1007/s00410-010-0578-z

Cite this article as:
Hiess, J., Bennett, V.C., Nutman, A.P. et al. Contrib Mineral Petrol (2011) 161: 1027. doi:10.1007/s00410-010-0578-z
  • 262 Views

Abstract

The role of fluids during Archaean intra-crustal magmatism has been investigated via integrated SHRIMP U–Pb, δ18O and LA-MC-ICPMS 176Hf isotopic zircon analysis. Six rock samples studied are all from the Nuuk region (southern West Greenland) including two ~3.69 Ga granitic and trondhjemitic gneisses, a 3.64 Ga granitic augen gneiss, a 2.82 Ga granodioritic Ikkattoq gneiss, a migmatite with late Neoarchaean neosome and a homogeneous granite of the 2.56 Ga Qôrqut Granite Complex (QGC). All zircon grains were thoroughly imaged to facilitate analysis of magmatic growth domains. Within the zircon analysed, there is no evidence for metamictization. Initial εHf zircon values (n = 63) are largely sub-chondritic, indicating the granitic host magmas were generated by the remelting of older, un-radiogenic crustal components. Zircon from some granite samples displays more than one 207Pb/206Pb age, and correlated with 176Hf/177Hf compositions can trace multiple phases of remelting or recrystallization during the Archaean. Model ages calculated using Lu/Hf arrays for each sample indicate that the crustal parental rocks to the granites, granodiorites and trondhjemites segregated from a chondrite-like reservoir at an earlier time during the Archaean, corresponding to known formation periods of more primitive tonalite–trondhjemite–granodiorite (TTG) gneisses. Zircon from the ~3.69 Ga granite, the migmatite and QGC granite contains Eoarchaean cores with chondritic 176Hf/177Hf and mantle-like δ18O compositions. The age and geochemical signatures from these inherited components are identical to those of surrounding tonalitic gneisses, further suggesting genesis of these granites by remelting of broadly tonalitic protoliths. Zircon oxygen isotopic compositions (n = 62) over nine age populations (six igneous and three inherited) have weighted mean or mean δ18O values ranging from 5.8 ± 0.6 to 3.7 ± 0.5‰. The 3.64 Ga granitic augen gneiss sample displays the highest δ18O with a mildly supra-mantle composition of 5.8 ± 0.6‰. Inherited Eoarchaean TTG-derived zircon shows mantle-like values. Igneous zircon from all other samples, spanning more than a billion years of Archaean time, record low δ18O sub-mantle compositions. These are the first low δ18O signatures reported from Archaean zircon and represent low δ18O magmas formed by the remelting and metamorphism of older crustal rocks following high-temperature hydrothermal alteration by meteoric water. Meteoric fluid ingress coupled with crustal extension, associated high heat flow and intra-crustal melting are a viable mechanism for the production of the low δ18O granites, granodiorites and trondhjemites reported here. Both high and low δ18O magmas may have been generated in extensional environments and are distinct in composition from Phanerozoic I-type granitic plutonic systems, which are typified by increasing δ18O during intra-crustal reworking. This suggests that Archaean magmatic processes studied here were subtly different from those operating on the modern Earth and involved extensional tectonic regimes and the predominance of remelting of hydrothermally altered crystalline basement.

Keywords

ZirconOxygenHafniumArchaeanCrustGraniteGreenland

Introduction

Attempts to understand the origin of Archaean gneiss complexes in the Nuuk region, southern West Greenland have used a range of approaches including solely field geology (McGregor 1973; Chadwick and Nutman 1979), regional tectonics (Friend et al. 1987, 1988; Nutman et al. 1989; Crowley 2002; Friend and Nutman 2005a; Nutman and Friend 2007), use of U–Th–Pb geochronology (Baadsgaard 1973, 1976; Nutman et al. 1993, 1996, 1999, 2000, 2007a, b) or Rb–Sr and Sm–Nd radiogenic isotopes (Black et al. 1971; Moorbath et al. 1972; O’Nions and Pankhurst 1978; Bennett et al. 1993, 2007). These studies recognized that the tonalite–trondhjemite–granodiorite (TTG) components of the gneisses largely represent juvenile crust formation, whereas granites were produced from these older sources in a time frame ranging from a few million years to over a billion years later (Moorbath 1975; Taylor et al. 1980; Moorbath et al. 1981; Nutman and Bridgwater 1986; Friend and Nutman 2005b). The influence of juvenile versus recycling/remelting processes has implications for the budgets of continental growth, and the redistribution of radiogenic elements within the lithosphere. Determining the source of recycled crustal materials and the role and nature of crustal fluids participating in crustal reworking will permit identification of mechanisms producing voluminous granitoids during Earth’s early history.

The approach used here differs from that of previous studies noted above, in that the focus is on the use of integrated isotopic information contained within well-documented individual igneous zircons to reveal the petrogenetic history of their host granitoids. For example, hafnium isotopes in igneous zircon can be used to identify the involvement of evolved crustal components during magmatism and to constrain the timing of the segregation of those older components from the mantle (e.g. Amelin et al. 1999). Zircon oxygen isotopic compositions can be used to track incorporation of weathered supracrustal materials in igneous protoliths (Peck et al. 2001; Cavosie et al. 2005) or the cannibalization of country rock altered by hydrothermal meteoric fluids (Gilliam and Valley 1997; Bindeman and Valley 2000, 2001; Monani and Valley 2001; Wei et al. 2002).

Zircon is the ideal mineral to capture and retain this petrogenetic information, because it robustly records the isotopic composition of primary magmas from which it crystallized (Valley 2003; Kinny and Maas 2003). A common accessory phase in igneous rocks, zircon is a precise U–Pb chronometer (Parrish and Noble 2003; Ireland and Williams 2003). It is also relatively un-reactive through most geological processes, apart from highest grade metamorphism and magmatic systems, when a single zircon is capable of undergoing multiple phases of growth. Due to the exceptionally low diffusivity of zircon (Watson and Cherniak 1997; Cherniak et al. 1997a, b; Cherniak and Watson 2000, 2003, 2007; Page et al. 2007) such zoned grains can also reveal isotopically distinct sources and open-system processes operating during multiple phases of magmatism and metamorphism (Corfu et al. 2003).

Hiess et al. (2009) interpreted the oldest (3.7–3.85 Ga) Eoarchaean juvenile tonalitic magmas from the Itsaq Gneiss Complex (IGC; Nutman et al. 1996) in the Nuuk region as partial melts of hydrated oceanic crust via δ18O and 176Hf/177Hf isotopic compositions preserved within precisely dated zircon. Here, we apply these integrated techniques to younger, more evolved Archaean granites and granodiorites intruding the primitive tonalitic suites. This information is used to highlight the role of crustal reworking, the important contribution of fluids to magmatism and the influence of extensional tectonics. Also, the styles of granitoid formation operating during the Archaean are contrasted with that of Phanerozoic magmas derived from metaluminous, I-type igneous systems (Kemp et al. 2007; Bolhar et al. 2008).

Samples and their geological setting

All samples are from the Nuuk area, southern West Greenland (Fig. 1). The region is dominated by polyphase TTG orthogneisses of the Archaean North Atlantic craton (McGregor et al. 1991). Exposed crust is predominantly tonalites with lesser trondhjemites, quartz-diorites, diorites, granodiorites, granites, metagabbros, ultramafic rocks and supracrustal associations (McGregor et al. 1991). All have been metamorphosed to amphibolite, and locally granulite facies (e.g. Wells 1976; Griffin et al. 1980) accompanied by intensive and heterogeneous ductile deformation (Nutman et al. 2000). Estimated pressures range from 0.5 to 1.0 GPa (Wells 1976) indicating these lithologies represent the deep levels of ancient orogens. Orthogneiss units can be divided into several tectonostratigraphic terranes with unique ages, structural geometries and early metamorphic histories. Each terrane is separated by folded and recrystallised, amphibolite facies mylonites (Friend et al. 1987, 1988). Following continental collision and tectonic juxtaposition, all terranes experienced a common phase of Neoarchaean deformation and granite intrusion (Nutman et al. 1989; McGregor et al. 1991; Friend et al. 1996; Friend and Nutman 2005b; Nutman and Friend 2007).
https://static-content.springer.com/image/art%3A10.1007%2Fs00410-010-0578-z/MediaObjects/410_2010_578_Fig1_HTML.gif
Fig. 1

Sketch geological map of Nuuk region, southern West Greenland with major lithological units and samples localities indicated. Adapted after Nutman et al. (2007b)

~3.69 Ga granitic gneiss 248251 and trondhjemitic gneiss 248212

The Isukasia area in the northern part of the Nuuk region (Fig. 1) consists of the 3.8 and 3.7 Ga Isua supracrustal belt that separates the “central gneisses” to the north dominated by ~3.7 Ga tonalites, from the “southern gneisses” dominated by ~3.8 Ga tonalites (Nutman et al. 1996, 1997; Nutman and Friend 2009). The southern fringe of the central gneisses is the source of samples 248251 and 248212 discussed here. Minimal regional strain in the central gneisses allows the typically heterogeneous and banded Eoarchaean gneisses of the IGC (McGregor 1973; Nutman et al. 1996) to be resolved into their major components of predominantly grey tonalitic gneisses, intruded by paler granitic and less commonly granodioritic and trondhjemitic components (Nutman et al. 1983; Nutman and Bridgwater 1986). The central ~3.7 Ga grey gneisses (Baadsgaard et al. 1986a) are dominated by juvenile tonalitic compositions, formed by the partial melting of hydrated mafic rocks (Nutman et al. 1999; Hiess et al. 2009).

There are swarms of anastomising granite–granodiorite and pegmatite sheets that intrude the tonalites of the central gneisses (‘white gneisses’ of Nutman and Bridgwater 1986; Fig. 4d of Nutman et al. 2000). U–Pb zircon geochronology has indicated that most of these granitic phases are 3.65–3.60 Ga old (Baadsgaard et al. 1986a; Nutman et al. 1996, 2000, 2002; Crowley et al. 2002). Recently, the ~3.7 Ga tonalitic phases have been divided into two suites on the basis of field relationships and U–Pb zircon geochronology (Nutman and Friend 2009), with a volumetrically subordinate suite of 3.72–3.71 Ga more melanocratic tonalite and quartz diorite now recognized. This suite is most common along the southern and western fringes of the central gneisses, and locally intrudes coeval metavolcanic rocks of the Isua supracrustal belt (Nutman and Friend 2009). A second more volumetrically important suite of paler tonalites and low-K granodiorites has U–Pb zircon ages of 3.70–3.69 Ga (Nutman et al. 1996, 2000, 2002; Crowley et al. 2002). Rocks of this suite intruded the older 3.72–3.71 Ga suite prevalent along the southern and western fringes of the central gneisses, and both suites in turn were intruded by the granitic (sensu-stricto) white gneisses with U–Pb zircon ages of 3.65–3.64 Ga (Nutman et al. 1996, 2000; Crowley et al. 2002). The samples used here were collected by H. Baadsgaard in 1978, and are pale gneiss sheets intruding darker tonalites from the southern fringe of the central gneisses (Fig. 1).

Samples 248251 (Fig. 1; 65°07.72′N 49°57.89′W) and 248212 (Fig. 1; 65°06.15′N 50°01.50′W) were initially reported by Baadsgaard (1983) with bulk zircon, apatite and titanite U–Pb isotope data. U–Pb zircon analyses indicated an upper intercept age on a Concordia diagram of ~3,650 Ma. Unpublished whole rock major element geochemical data for samples 248251 and 248212 was acquired in 1980 with XRF on fused glass discs, at Grønlands Geologiske Undersøgelse. 248251 with SiO2 of 72.41 and K2O of 4.82 wt% is a granite. 248212 has more trondhjemitic composition with SiO2 of 71.07 and K2O of only 0.84 wt%. Samples 248251 and 248212 were originally regarded as part of the 3.65–3.64 Ga white gneiss suite (e.g. Baadsgaard et al. 1986a). However, on the basis of more recent data, including SHRIMP U–Pb dating presented below, they are actually ~3.69 Ga sheets intruded into primitive ~3.72–3.71 Ga tonalites and should not be correlated with the 3.65–3.64 Ga regional white gneiss sheet swarms.

~3.64 Ga granitic augen gneiss G97/111

South of the fjord Ameralik, 3.85 to 3.69 Ga IGC TTG gneisses are intruded by the ~3.64 Ga augen gneiss or iron-rich suite (McGregor 1973; Baadsgaard 1973; Nutman et al. 1984, 1996, 2000). Augen gneiss suite lithologies form part of an association of K-feldspar megacrystic, mafic granitic–granodioritic–quartz–monzonitic gneisses, with lesser ferrodiorites and gabbros. The suite is geochemically distinct from the regional banded gneisses (e.g. with Fe, Ti and REE enrichment, low SiO2 and Al2O3 contents; Nutman et al. 1984) and characterized as a ‘within plate granite suite’ based on general characteristics such as elevated Nb–Y concentrations (Nutman et al. 1996). The augen gneiss suite has been interpreted as the product of deep crustal melting (under granulite facies?) with the melting triggered by the emplacement of an underplate of mantle-derived gabbroic magma (Nutman et al. 1984). Fractionation of the gabbroic magma and its intermingling with crustally derived melts produced a broad range of generally high Fe, Ti rocks. Early isotopic evidence to support this interpretation came from data of Vervoort et al. (1996), where the bulk zircon analysis of a ferrogabbro sample had a more elevated initial εHf value than a sample of granitic augen gneiss.

Within the augen gneisses, sample G97-111 (Fig. 1; Honda et al. 2003), is an A-type meta-granite with vestiges of igneous K-feldspar preserved as phenocrysts. It is located at 64°01.96′N 51°36.60′W, from a low strain domain on the west of Narssaq peninsula, at the southern mouth of Ameralik. The sample is relatively mafic with abundant K-feldspar augen and probably represents a crustal melt with mantle contamination (Nutman et al. 1984, 1996). Zircons are prismatic, typically 150–300 μm long, have numerous ilmenite inclusions, oscillatory zonation parallel to grain boundaries, and are devoid of inherited cores in CL. Some exteriors are corroded while others have thin (<15 μm) overgrowths. The best oscillatory zoned domains yielded a weighted mean 207Pb/206Pb age of 3,642 ± 3 Ma (2SD, MSWD = 1.0) and record the crystallization age of the protolith (Honda et al. 2003).

~2.82 Ga granodioritic Ikkattoq gneiss VM97/01

The Mesoarchaean amphibolite facies Ikkattoq orthogneisses of the Tre Brødre Terrane (Friend et al. 1987; Nutman et al. 1989; McGregor et al. 1991; Nutman and Friend 2007; Friend et al. 2009) are found in tectonic contact with the IGC throughout the southern Nuuk region. Most are granodioritic in composition, and contain abundant gabbro-anorthosite inclusions and widely spaced, concordant pegmatite banding. Eleven SHRIMP U–Pb zircon dated samples of single-phase gneisses record magmatic ages within a narrow range from 2,817 ± 9 to 2,829 ± 11 Ma (Nutman and Friend 2007) consistent with ID-TIMS analysis of zircon from one sample yielding an age of 2,824 ± 2 Ma (Crowley 2002). Based on Sr, Nd and Pb isotopic data, the Ikkattoq gneisses contain a significant contribution of mixed older crustal materials (Friend et al. 2009). In respect to their generally granodioritic composition and their clear evidence of major contamination by older crust based on Sr, Nd and Pb whole rock isotopic data, they are an exception amongst the Eo-Neoarchaean TTG suites of the Nuuk region, where contamination by older crust is either absent or considerably muted (Moorbath et al. 1986; Bennett et al. 1993; Garde et al. 2000; Friend et al. 2009). Sample VM97/01, collected from the top of Hjortetakken, south of Nuuk (Fig. 1; 1,225 m; approximately 64°07.30′N 51°34.70′W), is a representative sample from the unit and was previously SHRIMP dated at 2,821 ± 8 Ma by Nutman and Friend (2007). Zircons are typically prismatic, ~200 μm in length, with fine scaled oscillatory zonation and Th/U >0.3.

Neoarchaean migmatite 195392 and granite 195376 of the Qôrqut granite complex (QGC)

The QGC (McGregor 1973; Brown et al. 1981; Friend et al. 1985) is a NE-SW elongated body of late Neoarchaean leucocratic, biotite composite granites, intruded as multiple granitic sheets into Eoarchaean, Mesoarchaean and Neoarchaean banded gneisses. The main part of the complex was not subject to major regional deformation following emplacement (Brown et al. 1981) and was emplaced at shallow to intermediate crustal levels into already evolved ‘continental’ crust. Age determinations by Baadsgaard (1976) and Moorbath et al. (1981) indicated formation at ~2.55 Ga while compositions indicate the granites approximate minimum melts formed by the partial melting of older surrounding crustal rocks (Brown et al. 1981; Moorbath et al. 1981). At deep structural levels in the main outcrop area of the QGC and at its northern fringes there are migmatites, which appear to contain neosome coeval with, or marginally older than the QGC (e.g. Friend et al. 1985; Nutman and Friend 2007). Additionally, major ductile shear zones along strike with the QGC were active when the granite was emplaced (Nutman et al. 1989, in press). Sample 195392 (Fig. 1; 64°16.33′N 51°04.00′W; Friend et al. 1985) is a migmatite sourced from a diatexite between regional biotite gneiss and a homogeneous leucocratic granite. Sample 195376 (Fig. 1; 64°16.50′N 51°00.00′W) is a sample of homogeneous Qôrqut granite previously dated by Nutman et al. (2007c) at 2,564 ± 12 Ma with Mesoarchaean and Eoarchaean inheritance.

Methods and results

Zircons were extracted using standard heavy liquid and magnetic separation techniques. The grains were cast into epoxy “megamounts” (Ickert et al. 2008) along with zircon reference materials. New generation SHRIMP megamounts have been found to greatly improve reproducibility of oxygen isotope analyses using SHRIMP II and details regarding their geometric design and performance can be found in Ickert et al. (2008). Following imaging using transmitted and reflected light and cathodoluminescence, U–Pb zircon ages were determined using SHRIMP RG at the Australian National University (ANU). The zircon mounts were then lightly re-polished and oxygen isotopic compositions were measured from spots placed on top of the original U–Pb sampling area for each zircon with SHRIMP II at ANU. Lutetium–hafnium isotopic compositions were determined by laser ablation MC-ICPMS (ANU Neptune). Although a larger spot size was used, care was taken to sample in the same growth domains of the zircon as for the U–Pb and O analyses. The details of the analytical methods used here and the statistical treatment of data (rejection of outliers and calculation of mean or weighted mean values) are the same as in Hiess et al. (2009) and are described fully in Online Resource 1. Mean (1σ) or weighted mean (95% confidence limit, c.l.) 207Pb/206Pb ages from this study typically have larger uncertainties, owing to fewer pooled analyses, but agree within uncertainties with previous U–Pb zircon age determinations on these samples, obtained using larger datasets where available. An integrated summary of the 90 U–Pb, 62 δ18O and 63 εHf(T) determinations is presented in Table 1. Complete reference material and sample data for oxygen and Lu–Hf analyses is given in Online Resource 2. A descriptive summary of the zircon characteristics and main geochemical results for each sample is presented in Online Resource 3. Representative cathodoluminescence zircon images with analysis locations and associated results for ~3.69 Ga granitic gneiss 248251 and trondhjemitic gneiss 248212 are given in Fig. 2a, ~3.64 Ga granitic augen gneiss G97/111 in Fig. 2b, ~2.82 Ga granodioritic Ikkattoq gneiss VM97/01 in Fig. 2c, and Neoarchaean migmatite sample 195392 and QGC granite 195376 in Fig. 2d. Tera–Wasserburg 238U/206Pb–207Pb/206Pb diagrams, summary plots of δ18O and εHf(T) values against corresponding 207Pb/206Pb crystallization age for each zircon analysis, and sample weighted means or means are illustrated in Fig. 3. The field for mantle zircon (δ18O = 5.3 ± 0.3‰) is the composition of zircon derived from the mantle (Valley et al. 1998). The field for Archaean and Hadean “supracrustal zircon” (δ18O = 6.5 to 7.5‰) reflects the composition of zircon from igneous protoliths whose source materials were altered by low temperature interaction with liquid water near Earth’s surface (i.e. weathered before incorporation back into an igneous system; Cavosie et al. 2005). The Lu–Hf chondritic uniform reservoir (CHUR) composition and uncertainty estimates are from Bouvier et al. (2008). Plots of δ18O against U, Th, Th/U, % common 206Pb and % discordance for samples analysed for 18O/16O are presented in Fig. 4.
Table 1

Summary of zircon and quartz U-Pb, δ18O and εHf(T) results with sample weighted mean and mean ages and compositions

Sample

spot

Grain

 

U

(ppm)

Th

(ppm)

Th/U

 

Comm.

206Pb%

238U/

206Pb

err

207Pb/

206Pb

err

207Pb/206Pb

Age (Ma)

err

Disc.

(%)

δ18O

VSMOW (‰)a

err

176Lu/177Hf

 

errb

176Hf/177Hf

Meas.

errb

εHf

Init.c

err

Abs.

err

248251 Granitic gneiss (65°07.72′N 49°57.89′W)

               

1.1

p m os

2556

850

0.34

0.01

1.346

0.014

0.3234

0.0013

3,586

6

0

         

3.1d

p m os

630

56

0.09

0.10

1.432

0.017

0.3419

0.0007

3,671

3

8

3.1

0.3

0.000950

25

0.280417

32

−1.8

1.1

2.4

4.1d

p m os

132

48

0.37

0.06

1.444

0.016

0.3477

0.0011

3,697

5

9

4.2

0.3

0.001008

45

0.280428

39

−1.0

1.4

2.5

8.2d

p f os

211

33

0.16

0.08

1.401

0.016

0.3450

0.0009

3,685

4

6

  

0.000537

14

0.280412

40

−0.7

1.4

2.5

10.1d

p c os

231

104

0.46

0.06

1.299

0.014

0.3517

0.0010

3,715

4

1

5.4

0.3

0.001151

17

0.280469

37

0.5

1.3

2.5

10.2d

p m os

722

93

0.13

0.02

1.300

0.014

0.3504

0.0005

3,709

2

1

4.3

0.3

0.001160

94

0.280454

32

−0.2

1.2

2.4

11.1

p c h

199

22

0.10

0.15

1.553

0.018

0.2977

0.0009

3,458

5

8

3.6

0.3

       

12.1

p e os

632

119

0.19

0.06

1.572

0.017

0.2699

0.0006

3,305

4

4

2.3

0.3

0.001350

26

0.280427

31

−10.9

1.1

2.4

13.1d

p m os

218

64

0.30

0.05

1.370

0.015

0.3492

0.0009

3,704

4

5

  

0.000792

27

0.280417

33

−0.7

1.2

2.4

13.2d

p c os

847

257

0.31

0.07

1.300

0.023

0.3390

0.0016

3,658

7

−1

         

14.1d

p m os

357

160

0.46

0.12

1.425

0.015

0.3499

0.0007

3,707

3

8

  

0.000850

31

0.280430

49

−0.3

1.7

2.7

18.1d

an m h

824

2

0.00

0.11

1.323

0.015

0.3386

0.0007

3,656

3

1

  

0.000827

8

0.280414

25

−2.0

0.9

2.3

20.1

an c t

3960

944

0.25

0.04

1.438

0.016

0.2864

0.0008

3,398

4

0

  

0.000884

50

0.280445

41

−7.0

1.5

2.6

22.1d

p e os

763

47

0.06

0.02

1.307

0.014

0.3457

0.0005

3,688

2

1

  

0.001217

31

0.280471

43

−0.2

1.5

2.6

23.1d

an m os

144

30

0.21

0.15

1.455

0.017

0.3384

0.0011

3,655

5

8

  

0.001471

137

0.280479

41

−1.3

1.5

2.6

Weighted Meane ± 95% confidence limits or Meanf ± 1σ

 

 

3,686f

22

 

4.2f

1.0

 

 

 

 

−0.8e

 

0.8

n

         

11 of 15

  

4 of 6

     

10 of 12

  

MSWD

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.4

 

 

248212 Trondhjemitic gneiss (65°06.15′N 50°01.50′W)

              

1.1d

p c os

215

104

0.50

0.01

1.329

0.016

0.3465

0.0008

3,692

3

2

5.0

0.3

0.001108

30

0.280408

34

−2.1

1.2

2.4

2.1d

p c os

771

261

0.35

0.01

1.280

0.015

0.3435

0.0008

3,678

3

−1

5.2

0.3

0.001014

16

0.280449

35

−0.7

1.3

2.4

3.1d

p m os

350

28

0.08

0.01

1.419

0.017

0.3455

0.0006

3,687

3

7

3.6

0.3

0.000659

3

0.280366

31

−2.5

1.1

2.4

4.1

p m os

2032

272

0.14

0.00

1.424

0.016

0.3277

0.0010

3,607

5

5

4.8

0.3

0.001022

97

0.280354

34

−5.8

1.2

2.4

5.1d

p m os

232

85

0.38

0.01

1.328

0.016

0.3488

0.0007

3,702

3

2

4.3

0.4

0.001040

20

0.280431

31

−0.9

1.1

2.4

6.1d

an f os

128

71

0.57

0.03

1.313

0.017

0.3458

0.0009

3,688

4

1

5.1

0.4

0.000609

17

0.280434

33

0.0

1.2

2.4

7.1

an m os

172

78

0.47

0.04

1.362

0.018

0.3305

0.0027

3,619

12

2

5.0

0.4

0.001171

58

0.280469

36

−1.8

1.3

2.5

9.1d

an m os

225

92

0.42

0.02

1.234

0.021

0.3473

0.0009

3,695

4

−3

4.8

0.4

0.001018

28

0.280452

31

−0.2

1.1

2.4

Weighted Meane ± 95% confidence limits or Meanf ± 1σ

 

 

3,690f

8

 

4.6f

0.6

 

 

 

 

−1.1e

 

1.0

n

         

6 of 8

  

6 of 8

     

6 of 8

  

MSWD

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.7

 

 

G97/111 Granitic augen gneiss (64°01.96′N 51°36.60′W)

              

1.1d

an m os

24

18

0.77

0.00

1.341

0.029

0.3313

0.0021

3,623

10

1

6.6

0.3

0.000602

5

0.280474

27

0.0

1.0

2.3

1.3d

an m os

323

158

0.51

0.01

1.317

0.015

0.3284

0.0007

3,610

3

−1

         

2.1d

an m os

225

104

0.48

0.03

1.445

0.017

0.3319

0.0011

3,626

5

7

4.7

0.3

0.000663

28

0.280456

27

−0.8

0.9

2.3

3.1d

an m os

323

166

0.53

0.06

1.387

0.016

0.3353

0.0008

3,641

4

4

6.0

0.3

0.000494

11

0.280430

22

−0.9

0.8

2.2

4.1d

p m os

178

112

0.65

0.01

1.308

0.016

0.3343

0.0009

3,637

4

−1

  

0.000642

4

0.280445

25

−0.9

0.9

2.3

B-1.1d

an f m os

225

112

0.51

0.02

1.326

0.011

0.3369

0.0008

3,649

4

1

6.5

0.3

0.000712

24

0.280419

33

−1.7

1.2

2.4

B-2.1d

an m os

89

40

0.46

0.05

1.390

0.014

0.3339

0.0015

3,635

7

4

5.8

0.3

0.000609

4

0.280436

31

−1.2

1.1

2.4

B-3.1d

an e os

160

107

0.69

0.04

1.304

0.012

0.3384

0.0010

3,656

5

0

5.6

0.3

0.000587

5

0.280413

29

−1.4

1.0

2.3

B-4.1

an m os

190

89

0.48

0.02

1.500

0.012

0.3009

0.0008

3,475

4

6

6.6

0.3

0.000591

4

0.280432

34

−5.0

1.2

2.4

B-5.1d

p f m os

327

148

0.47

0.19

1.337

0.010

0.3343

0.0008

3,637

4

1

5.6

0.3

0.000890

6

0.280441

26

−1.6

0.9

2.3

B-6.1

an e os

175

92

0.54

0.03

1.391

0.012

0.3254

0.0009

3,595

4

3

6.2

0.3

0.000626

24

0.280434

35

−2.2

1.2

2.4

B-7.1

an m os

125

59

0.49

0.13

1.436

0.013

0.3084

0.0011

3,513

6

3

6.2

0.3

0.000424

7

0.280425

33

−4.0

1.2

2.4

B-8.1

p e os

144

87

0.62

0.03

1.412

0.012

0.3216

0.0010

3,578

5

4

6.4

0.3

0.000702

11

0.280443

28

−2.5

1.0

2.3

Weighted Meane ± 95% confidence limits or Meanf ± 1σ

 

 

3,635f

14

 

5.8f

0.6

 

 

 

 

−1.1e

 

0.8

n

         

9 of 13

  

7 of 11

     

8 of 12

  

MSWD

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.2

 

 

VM97/01 Granodioritic Ikkattoq gneiss (64°07.30′N 51°34.70′W)

             

1.1d,g

p e os

258

105

0.40

0.10

1.729

0.050

0.1998

0.0010

2,825

9

4

         

2.1d,g

p m os

410

192

0.47

0.21

1.854

0.052

0.1976

0.0013

2,807

11

−1

3.8

0.5

0.000992

82

0.281017

32

−0.5

1.2

2.4

3.1d,g

p e os

431

127

0.29

0.10

1.737

0.044

0.1997

0.0009

2,824

7

4

4.7

0.5

0.000893

47

0.281000

43

−0.5

1.5

2.6

4.1d,g

p e os

424

181

0.43

0.45

1.644

0.045

0.1981

0.0027

2,811

23

9

         

5.1d,g

p e os

437

156

0.35

0.11

1.718

0.049

0.1999

0.0010

2,826

8

5

3.8

0.5

0.001671

34

0.280968

136

−3.1

4.8

5.3

5.2d,g

p e os

521

187

0.36

0.19

2.065

0.049

0.1970

0.0008

2,801

7

−9

3.1

0.5

       

6.1d,g

p f e os

368

174

0.47

0.09

1.790

0.064

0.1993

0.0018

2,820

15

2

         

7.1d,g

p e os

437

189

0.43

0.04

1.771

0.050

0.2000

0.0010

2,826

8

2

         

8.1d,g

p e os

345

150

0.43

0.25

1.837

0.059

0.2006

0.0012

2,831

10

−1

         

9.1d,g

p e os

497

254

0.51

0.06

1.816

0.050

0.2000

0.0011

2,827

9

0

         

10.1d,g

p m os

507

212

0.42

0.12

1.699

0.041

0.2015

0.0012

2,838

9

5

4.3

0.5

       

Weighted Meane ± 95% confidence limits or Meanf ± 1σ

 

 

2,821e

8

 

4.0e

0.8

 

 

 

 

−0.7e

 

1.6

n

         

11 of 11

  

5 of 5

     

3 of 3

  

MSWD

 

 

 

 

 

 

 

 

 

1.7

 

 

1.3

 

 

 

 

 

0.5

 

 

195392 Migmatite (64°16.33′N 51°04.00′W)

                

1.1d

p c os

248

119

0.48

0.00

1.388

0.041

0.3459

0.0016

3,689

7

5

5.4

0.5

0.000436

8

0.280437

33

0.6

1.2

2.4

4.1d

p c os

42

39

0.93

0.08

1.348

0.046

0.3305

0.0042

3,619

20

1

         

6.1

p c os

53

53

1.00

0.00

1.591

0.051

0.3046

0.0034

3,494

17

9

4.6

0.5

0.000440

21

0.280477

38

−2.6

1.4

2.5

10.1d

p c os

134

109

0.81

0.00

1.409

0.043

0.3274

0.0020

3,605

10

4

         

15.2d

p c os

94

37

0.39

0.33

1.447

0.043

0.3254

0.0017

3,596

8

6

         

16.1

p c os

183

46

0.25

0.14

1.321

0.041

0.3552

0.0042

3,729

18

3

4.5

0.5

       

B-1.1d

p c os

239

116

0.50

0.05

1.361

0.023

0.3401

0.0009

3,663

4

3

4.4

0.4

0.0007704

88

0.280475

28

0.5

1.0

2.3

B-1.2d

p m os

1316

494

0.39

0.00

1.262

0.019

0.3447

0.0003

3,684

1

−2

4.7

0.4

0.0008431

17

0.280416

23

−1.3

0.8

2.3

B-4.1d

p m os

166

260

1.62

0.27

1.410

0.033

0.3366

0.0009

3,647

4

6

5.7

0.4

0.0028683

88

0.280575

46

−1.6

1.7

2.7

B-6.2d

p c os

1026

37

0.04

0.15

1.459

0.021

0.3271

0.0008

3,603

4

7

4.2

0.4

0.0006074

15

0.280432

21

−2.0

0.7

2.2

B-8.2d

p c os

1244

86

0.07

0.00

1.321

0.021

0.3324

0.0006

3,628

3

0

5.3

0.4

0.0007310

14

0.280371

23

−3.9

0.8

2.3

B-8.4d

p c os

904

90

0.10

0.06

1.358

0.020

0.3306

0.0003

3,620

2

2

5.4

0.4

0.0007966

8

0.280417

19

−2.7

0.7

2.2

B-10.1d

an m h

1462

50

0.04

0.03

1.305

0.021

0.3351

0.0003

3,640

1

−1

5.1

0.4

       

B-11.1d

an c h

1476

56

0.04

0.00

1.365

0.020

0.3254

0.0009

3,595

4

1

5.0

0.4

0.0012015

27

0.280424

18

−4.0

0.7

2.2

B-12.1d

p m os

254

28

0.11

0.10

1.495

0.026

0.3168

0.0007

3,555

3

8

4.3

0.4

0.0004185

3

0.280416

23

−3.3

0.8

2.2

Weighted Meane ± 95% confidence limits or Meanf ± 1σ

 

 

3,627f

38

 

5.0e

0.4

 

 

 

 

−2.0f

 

1.7

n

         

13 of 15

  

10 of 12

     

9 of 10

  

MSWD

 

 

 

 

 

 

 

 

 

 

 

 

1.9

 

 

 

 

 

 

 

 

7.1d

p c os

43

32

0.75

0.12

1.811

0.060

0.2355

0.0051

3,090

35

8

3.5

0.5

0.000304

7

0.280434

56

−13.3

2.0

2.9

8.1d

p c os

396

31

0.08

0.29

1.866

0.055

0.2226

0.0012

2,999

9

8

         

9.1d

p c os

124

51

0.41

1.15

1.946

0.059

0.2202

0.0028

2,982

21

9

4.1

0.5

0.000310

26

0.280421

30

−16.3

1.1

2.4

11.1d

p c os

1291

296

0.23

0.01

1.798

0.052

0.2325

0.0006

3,069

4

7

         

13.1d

p c os

103

40

0.38

0.42

1.797

0.059

0.2253

0.0021

3,019

15

6

3.4

0.5

0.000684

22

0.280488

28

−13.8

1.0

2.3

15.1

p c os

64

25

0.39

0.58

1.727

0.055

0.2548

0.0048

3,215

30

8

         

Weighted Meane ± 95% confidence limits or Meanf ± 1σ

 

 

3,032f

46

 

3.7e

0.5

 

 

 

 

−14.6e

 

4.0

n

         

5 of 6

  

3 of 3

     

3 of 3

  

MSWD

 

 

 

 

 

 

 

 

 

 

 

 

0.7

 

 

 

 

 

1.7

 

 

B-2.2d

an c os

824

72

0.09

0.11

1.953

0.029

0.1878

0.0003

2,723

3

2

3.5

0.4

       

B-3.1

p e os

214

27

0.13

0.13

1.972

0.031

0.1808

0.0006

2,660

6

1

4.5

0.4

       

B-6.1d

p m os

309

20

0.07

0.50

2.034

0.031

0.1877

0.0009

2722

8

6

3.5

0.4

0.0005730

14

0.280521

22

−19.3

0.8

2.2

B-8.1d

p m os

357

24

0.07

0.86

1.918

0.029

0.1896

0.0012

2,739

10

1

4.8

0.4

0.0005417

6

0.280504

23

−19.5

0.8

2.2

B-8.3d

p e os

250

18

0.07

1.36

1.910

0.029

0.1891

0.0019

2,735

17

1

4.1

0.4

0.0006016

8

0.280489

20

−20.2

0.7

2.2

B-9.1d

p m os

205

20

0.10

0.25

1.918

0.029

0.1875

0.0007

2,721

6

1

3.9

0.4

0.0006371

19

0.280514

21

−19.7

0.7

2.2

B-9.2d

p c os

570

53

0.10

0.04

1.973

0.030

0.1886

0.0003

2,730

3

3

3.9

0.4

0.0006884

15

0.280551

22

−18.3

0.8

2.2

Weighted Meane ± 95% confidence limits or Meanf ± 1σ

 

 

2,726e

5

 

4.0e

0.5

 

 

 

 

−19.4e

 

1.0

n

         

6 of 7

  

6 of 7

     

5 of 5

  

MSWD

 

 

 

 

 

 

 

 

 

1.2

 

 

1.6

 

 

 

 

 

0.4

 

 

2.2d

p e h os

796

35

0.04

0.04

2.155

0.054

0.1707

0.0008

2,564

8

4

         

3.1d

p m os

180

67

0.37

0.19

2.087

0.062

0.1684

0.0017

2,541

16

1

         

5.1d

p m os

61

32

0.51

0.07

2.196

0.069

0.1688

0.0031

2,546

31

5

4.4

0.5

0.000608

29

0.280786

39

−14.0

1.4

2.5

5.2d

p m os

98

70

0.72

0.28

2.097

0.064

0.1704

0.0027

2,561

26

2

         

12.1d

p e h os

104

54

0.53

0.47

1.989

0.060

0.1714

0.0017

2,572

17

−2

4.5

0.5

0.000372

9

0.280526

35

−22.2

1.3

2.4

17.1d

p e h os

42

43

1.03

1.34

2.071

0.074

0.1704

0.0036

2,561

35

1

5.1

0.5

0.000354

15

0.280510

48

−23.0

1.7

2.7

18.1d

p e h os

151

61

0.41

0.39

2.140

0.063

0.1702

0.0013

2,559

13

3

         

B-5.2d

an e h

1258

56

0.05

0.01

2.008

0.033

0.1715

0.0002

2,572

2

−1

4.8

0.4

0.0005246

26

0.280651

18

−18.0

0.6

2.2

Weighted Meane ± 95% confidence limits or Meanf ± 1σ

 

 

2,570e

4

 

4.7e

0.4

 

 

 

 

−22.6e

 

1.8

n

         

8 of 8

  

4 of 4

     

2 of 4

  

MSWD

 

 

 

 

 

 

 

 

 

0.9

 

 

0.5

 

 

 

 

 

0.2

 

 

1

Qtz

           

8.7

0.1

       

2

Qtz

           

9.2

0.1

       

195376 Qôrqut granite complex granite (64°16.50′N 51°00.00′W)

             

3.1d

p m os

1930

1703

0.91

0.11

2.125

0.031

0.1716

0.0003

2,573

3

3

4.8

0.4

0.0006328

8

0.280712

22

−16.0

0.8

2.2

4.1

p m os

344

192

0.58

0.24

1.717

0.040

0.2327

0.0010

3,071

7

4

4.6

0.4

0.0005037

28

0.280847

30

0.5

1.1

2.3

5.1d

p m os

3274

4684

1.48

0.01

2.019

0.034

0.1719

0.0010

2,576

10

−1

3.7

0.4

0.0008013

21

0.280740

26

−15.3

0.9

2.3

5.2d

p c rx

1117

1201

1.11

-0.01

1.997

0.065

0.1701

0.0017

2,559

17

−2

4.9

0.4

0.0009886

30

0.280756

33

−15.4

1.2

2.4

6.1d

p m os

1493

1674

1.16

0.10

2.060

0.046

0.1699

0.0009

2,556

9

0

3.5

0.4

0.0005367

6

0.280824

23

−12.3

0.8

2.3

8.1

p m os

83

8

0.10

0.09

1.472

0.024

0.3268

0.0040

3,602

19

8

         

9.1

p m os

804

86

0.11

0.00

1.649

0.024

0.2348

0.0008

3,085

5

1

3.7

0.4

0.0009404

46

0.280867

38

0.7

1.4

2.5

Weighted Meane ± 95% confidence limits or Meanf ± 1σ

 

 

2,571e

10

 

4.2f

0.7

 

 

 

 

−14.7f

 

1.7

n

         

4 of 7

  

4 of 6

     

4 of 6

  

MSWD

 

 

 

 

 

 

 

 

 

1.3

 

 

 

 

 

 

 

 

 

 

 

1

Qtz

           

8.2

0.1

       

2

Qtz

           

8.9

0.1

       

a = [18O/16Osample / (18O/16Oreference measured / 18O/16Oreference true) - VSMOW] × 1000/VSMOW

b = ×10-6

c = (176Hf/177Hfinitial / 176Hf/177HfCHUR −1) × 10000

CHUR: 176Hf/177Hf = 0.282785±11, 176Lu/177Hf = 0.0336±1 (Bouvier et al. 2008)

λ176Lu = 1.867±8×10-11y-1 (Scherer et al. 2001; Söderlund et al. 2004)

d = Analysis used for weighted mean or mean calculations

e = Weighted mean ± 95% confidence limits

f = Mean ± 1σ

g = U-Pb age determined by Nutman and Friend (2007)

Grain descriptions

Habit: p prismatic, an anhedral, f fragment

Analysis site: c core, m middle, e edge

Zonation: os oscillatory, h homogeneous, t turbid, rx recrystallised

https://static-content.springer.com/image/art%3A10.1007%2Fs00410-010-0578-z/MediaObjects/410_2010_578_Fig2_HTML.jpg
Fig. 2

Representative CL images recording analysis locations, 207Pb/206Pb crystallization ages, % discordance, Th/U ratios, δ18O, εHf(T) for: a White gneisses 248251 and 248212, b Augen granite G97/111, c Ikkattoq gneiss VM97/01 and d Qôrqut Granite Complex samples 195392 and 195376. Scale bars are 100 μm

https://static-content.springer.com/image/art%3A10.1007%2Fs00410-010-0578-z/MediaObjects/410_2010_578_Fig3a-i_HTML.gif
https://static-content.springer.com/image/art%3A10.1007%2Fs00410-010-0578-z/MediaObjects/410_2010_578_Fig3j-r_HTML.gif
Fig. 3

Tera–Wasserburg diagrams and plots of δ18O and εHf(T) against corresponding 207Pb/206Pb crystallization ages for zircon analysis. Tera–Wasserburg data-point error crosses are at the 2σ level. δ18O and εHf(T) uncertainties are 1σ and 2σ, respectively, while 207Pb/206Pb ages are at 1σ level. Fields for mantle zircon, Archaean–Hadean “supracrustal zircon” and CHUR from Valley et al. (1998), Cavosie et al. (2005) and Bouvier et al. (2008). 176Lu/177Hf ratios for samples were determined by linear regression with R2 values indicating correlation coefficients

https://static-content.springer.com/image/art%3A10.1007%2Fs00410-010-0578-z/MediaObjects/410_2010_578_Fig4_HTML.gif
Fig. 4

Plots of δ18O against U–Th–Pb systematics for all samples. The lack of correlations is evidence that the δ18O values are primary features and not the result of alteration

Discussion

The origin of Archaean granites: insights from initial 176Hf/177Hf isotopic compositions

Hiess et al. (2009) previously determined zircon weighted mean initial εHf values of +0.5 ± 0.6 to −0.1 ± 0.7 for five Eoarchaean tonalite and felsic volcanic units from the IGC, implying derivation of these juvenile magmas from a largely unfractionated, chondritic source reservoir. In contrast, samples described here record largely sub-chondritic weighted mean or mean εHf(T) compositions of −0.7 ± 1.6 (2,821 ± 8 Ma), −0.8 ± 0.8 (3,686 ± 22 Ma), −1.1 ± 1.0 (3,690 ± 8 Ma), −1.1 ± 0.8 (3,635 ± 14 Ma), −2.0 ± 1.7 (3,627 ± 38 Ma), −14.6 ± 4.0 (3,032 ± 46 Ma), −14.7 ± 1.7 (2,571 ± 10 Ma), −19.4 ± 1.0 (2,726 ± 5 Ma) and −22.6 ± 1.8 (2,570 ± 4 Ma). These values consistently lie at the lower limit of, or clearly beneath the error envelope for the chondritic reference line of εHf(T) = 0.0 ± 0.4 (Bouvier et al. 2008; Figs. 3c, f, i, l, o and r) and imply that the magmas from which this zircon crystallized was derived, at least in part, from the remelting of more mature, crustal sources with lower Lu/Hf than CHUR.

The evolved source or sources for each sample cannot be precisely identified from these data. However, the timing of the source’s segregation from its ultimate mantle reservoir, here taken to be represented by a chondritic composition (Bouvier et al. 2008) and assumed to be equivalent to bulk silicate Earth (BSE), can be resolved by projecting the 176Lu/177Hf ratio defined by the zircon array to the intersection with the chondritic reference line. This is a form of model age, with the distinction that instead of assuming an average crustal 176Lu/177Hf for Hf isotopic evolution of the crustal source, prior to zircon crystallization as is commonly done for model age calculations of detrital zircons, the 176Lu/177Hf ratio as defined by the zircon populations within the given rock is used. Model ages are here calculated with reference to CHUR rather than a depleted mantle reference line on the basis of the initial Hf isotopic compositions determined from the oldest, most primitive Eoarchaean tonalites (Hiess et al. 2009), which show no evidence for long-term, high Lu–Hf mantle sources.

Data from granitic gneiss sample 248251 (εHf(T) = −0.8 ± 0.8, 3,686 ± 22 Ma, 176Lu/177Hf = 0.001) indicate segregation of its source from BSE composition at ~3,724 Ma (Fig. 3c), consistent with remelting of 3.72 Ga components of the northern Isua composite ca. 3.7 Ga package (Nutman and Friend 2009). Therefore, these data support the earlier interpretation of IGC granite petrogenesis by episodic partial melting of broadly tonalitic grey gneiss sources (Baadsgaard et al. 1986a).

The CHUR model age of A-type granitic augen gneiss G97/111 (εHf(T) = −1.1 ± 0.8, 3,635 ± 14 Ma, 176Lu/177Hf = 0.002) is ~3,681 Ma (Fig. 3i). This age is consistent with that of components of the surrounding TTG orthogneisses from the IGC gneisses south of the Ameralik Fjord (Nutman et al. 1996, 2000). While a precise source for the augen gneisses from the iron-rich suite has not been previously identified, they have been interpreted by Nutman et al. (1984) as products of incomplete mixing between lower crustal melts adjacent to fractionated, mantle-derived basic intrusions.

The range of initial εHf values from granodioritic Ikkattoq gneiss sample VM97/01 (−0.5 ± 1.2 to −3.1 ± 4.8) indicates that the zircons were probably derived from the melting of mixed crustal components, with a range of ages sourced from both near-chondritic and low εHf(T) reservoirs. This is in accord with the whole rock εNd (2,825 Ma) values for this suite, which range from −1.8 to −7.1 (Friend et al. 2009).

The crustal source of migmatite sample 195392 with a youngest 2,570 ± 4 Ma component, marginally older than, or coincident with the main QGC intrusion (εHf(T) = −22.6 ± 1.8, 176Lu/177Hf = 0.004) initially segregated from BSE at ~3,735 Ma (Fig. 3o). This age is broadly consistent with the timing of formation of surrounding Eoarchaean IGC orthogneisses. The Eoarchaean protolith is directly represented in the sample by inherited zircon cores (e.g. 195392-1.1, 3689 ± 7 Ma, εHf(T) = 0.6 ± 1.2, Fig. 2d). Similar chondritic initial εHf compositions are commonly observed within zircon from Eoarchaean meta-tonalite samples in the Nuuk region (Hiess et al. 2009) but are not observed within any younger zircon domains from this sample. The migmatite sample 195392 also contains a significant population of zircon domains of Mesoarchaean age (3,032 ± 46 Ma; Fig. 3o), that either reflects a rock component in the migmatite of that age (coeval with the orthogneisses of the Kapisillik or Akia Terranes (Friend and Nutman 2005a; Nutman and Friend 2007; Fig. 1), or alternatively, U–Pb isotopic resetting during recrystallization of Eoarchaean components in a tectonothermal event at that time. The sub-chondritic εHf(T) values measured in those zircon domains (weighted mean = −14.6 ± 4.0), however, indicate that if any Mesoarchaean rocks were involved, these were formed by the remelting of Eoarchaean rocks, and not as new juvenile additions to the crust. A Neoarchaean zircon population at 2,726 ± 5 Ma (weighted mean εHf(T) = −19.4 ± 1.0; Fig. 3o) possibly relates to zircon recrystallization and growth during a regional phase of crustal thickening and associated high-grade metamorphism at that time (Friend et al. 1996; Nutman and Friend 2007). While a Neoarchaean population at 2,570 ± 4 Ma (weighted mean εHf(T) = −22.6 ± 1.8), representing the age and composition of the granite sheet, largely falls on the same 176Lu/177Hf trajectory intersecting Eoarchaean and Mesoarchaean components (Fig. 3o). The observation that the different zircon populations fall on a single 176Lu/177Hf array implies the QGC largely formed by a relatively “closed” system cannibalistic remelting of the older, surrounding rocks. Two outlier analyses εHf(T) = −14.0 ± 1.4 (2,546 ± 31 Ma) and −18.0 ± 0.6 (2,572 ± 2 Ma) might represent a minor infiltrating component of young, higher εHf(T) melt during genesis of the QGC.

Homogeneous granite sample 195376 of the QGC with εHf(T) = −14.7 ± 1.7 was probably formed by remelting of juvenile Mesoarchaean and Eoarchaean rocks as represented by zircon inheritance with primitive Mesoarchaean compositions, distinct from un-radiogenic Mesoarchaean components of sample 195392. This demonstrates that the migmatite sample 195392 might not be genetically related to the main body of the QGC. This observation from sample 195376 that the QGC was largely derived by the remelting of regional Eoarchaean (IGC) and Mesoarchaean (possibly Kapisillik or Akia Terrane) banded gneisses is entirely consistent with the earlier Rb–Sr and Pb/Pb whole rock isochrons and interpretation of Moorbath et al. (1981).

The slope of zircon εHf(T)207Pb/206Pb arrays for samples of Eoarchaean granite 248251, Eoarchaean augen gneiss granite G97/111 and Neoarchaean migmatite 195392 define 176Lu/177Hf ratios of 0.001, 0.002 and 0.004, respectively, and are lower than that of Eoarchaean tonalites ranging from 0.007 to 0.012 (Hiess et al. 2009). This is consistent with the expectation that the granitic crustal reservoirs from which these zircons crystallized would become progressively more fractionated to lower Lu/Hf, following partial melting of a broadly tonalitic source.

Granite 248251 records a 176Lu/177Hf array of 0.001 (Fig. 3c), identical to the 176Lu/177Hf (~0.001) measured within the individual zircon for that sample (Table 1). This indicates for 248251 that 207Pb/206Pb ages younger than the magmatic age of 3,686 ± 22 Ma simply represent resetting of the zircon U–Pb systematics (Pb loss) during recrystallization, without resetting of the Lu–Hf isotopic composition. Augen gneiss granite G97/111 and migmatite 195392 have 176Lu/177Hf arrays of 0.002 and 0.004 (Fig. 3i, o) that are slightly higher than the 176Lu/177Hf ratios ~0.0005 measured in zircon (Table 1). This indicates that zircon 207Pb/206Pb ages younger than the magmatic ages could represent metamorphic resetting of zircon U–Pb systematics combined with the growth of new magmatic zircon from remelting the surrounding low Lu/Hf crust. Whole rock data from 3.81 to 3.70 Ga gneisses from the region record low 176Lu/177Hf compositions of ~0.001 to ~0.002 (Vervoort and Blichert-Toft 1999). The low 176Lu/177Hf ratios in these evolved rocks and zircons represent strong Lu/Hf fractionation in the garnet source region of the initial TTG magmas, further enhanced by progressive remelting of Hf-enriched granitic crust.

Significantly, no samples within this study record distinctly positive values for initial εHf. As their parental melts formed by the remelting of older crustal components, these zircon compositions do not reflect that of the Archaean mantle at the time of their formation. However, they also provide no evidence that those older crustal protolith sources were derived from a radiogenic, high Lu/Hf mantle reservoir with respect to chondrites. Although there is no evidence for super-chondritic Hf isotopic compositions in the samples from southwest Greenland, it is unclear at present how representative the Archaean terranes of Greenland are of wider processes of crustal growth during the Archaean. Consequently, while these data cannot exclude other models, it would support those for slow (Hurley and Rand 1969) or progressive (Bennett 2003) continental crust growth that may have been balanced by rapid subduction recycling.

The role of fluids in Archaean granitoid complex evolution: insights from δ18O compositions of igneous zircons

Zircon from juvenile Eoarchaean tonalite and felsic volcanic samples recorded weighted mean δ18O values of 5.1 ± 0.4 to 4.9 ± 0.7‰ and largely imply derivation from a mantle or gabbroic source, with little influence from weathered crustal material with its elevated δ18O compositions (Hiess et al. 2009). In contrast (with exception of augen gneiss granite G97/111—see below), the younger, more evolved samples in this study typically record mean or weighted mean δ18O values that are distinctly below the field for mantle zircon with values of 4.2 ± 1.0‰ (3,686 ± 22 Ma, Fig. 3b), 4.6 ± 0.6‰ (3,690 ± 8 Ma, Fig. 3e), 4.0 ± 0.8‰ (2,821 ± 8 Ma, Fig. 3k), 3.7 ± 0.5‰ (3,032 ± 46 Ma, Fig. 3n), 4.0 ± 0.5‰ (2,726 ± 5 Ma, Fig. 3n), 4.7 ± 0.4‰ (2,570 ± 4 Ma, Fig. 3n) and 4.2 ± 0.7‰ (2,571 ± 10 Ma, Fig. 3q). These compositions suggest that the parental magmas from which these samples crystallized included rocks that had been hydrothermally altered at high temperatures. In contrast, augen gneiss sample G97/111 is an exception with a δ18O of 5.8 ± 0.6‰ (3,635 ± 14 Ma, Fig. 3h) and lying above the field for mantle zircon but beneath the field for Archaean–Hadean “supracrustal zircon”. This suggests the involvement of mildly evolved crustal components during the formation of G97/111, but does not strictly require protoliths that were altered near the Earth’s surface by liquid water at low temperatures. The rare, older Eoarchaean inherited zircon components of Eoarchaean granitic gneiss 248251, trondhjemite 248212, and the 3,627 ± 38 Ma population from migmatite sample 195392 at 5.0 ± 0.4‰ (Fig. 3n) all preserve values that lie within the range of mantle zircon. Similar mantle-like δ18O compositions are commonly observed within zircon from Eoarchaean meta-tonalite samples (Hiess et al. 2009) but are not observed within any younger zircon domains from these samples.

Do these zircons preserve their magmatic δ18O?

It is important to investigate whether the variable zircon δ18O values measured represent the primary composition of the magmas from which the zircon crystallized, are analytical artefacts, or are products of secondary hydrothermal alteration and isotopic exchange (e.g. Cavosie et al. 2005). In consideration of potential instrumental effects, we note that sample HfO2 concentrations for the zircons measured here are comparable with those of the standardizing reference materials, suggesting that variable matrix effects within the zircon lattice are not significant (see Online Resource 1). Additionally, zircons recording the entire δ18O compositional range were measured sequentially during single analytical sessions, arguing against a unidirectional analytical fractionation, related to mount geometry or instrument tuning during a single analytical session. For example, samples G97/111 (6.1 ± 0.3‰, n = 8, Online Resource 2b), FC1 (5.3 ± 0.3‰, n = 10, Online Resource 2a) and 248212 (4.6 ± 1.1‰, n = 4, Online Resource 2b) were all part of analytical session 3. Also, several samples analysed in multiple analytical sessions record the same range of values, indicating the experiments are reproducible (e.g. 248212 in sessions 3 and 4, G97/111 in sessions 1 and 3, 195392 in sessions 6 and 7; Online Resource 2b).

The possibility of post-magmatic grain alteration must also be considered. We note that the analysed zircons generally display euhedral habits, well-preserved oscillatory growth zonation (e.g. Cavosie et al. 2005) and a high level of U–Pb concordance that is expressed in their coherent initial εHf behaviour. Calculated εHf(T) values are highly sensitive to inaccuracy or disturbance in U–Pb systematics, which result in erratic compositions and highly scattered 176Lu/177Hf arrays. The presence of such features would also likely indicate disturbance of oxygen isotopic compositions. However, for the zircons analysed here, even the inherited Eoarchaean zircon cores typically retain distinct, original, mantle-like compositions, e.g. 248251-10.1 (Fig. 2a), 195392-1.1, 195392B-8.2 and 195392B-8.4 (Fig. 2c) characteristic of their assumed tonalitic source rocks (Hiess et al. 2009). This suggests the grains were not subject to pervasive secondary alteration and demonstrates the robustness of the oxygen isotopic values. Inherited Eoarchaean components with lower U and Th concentrations (10–100 ppm) are less likely to experience secondary alteration when compared to Mesoarchaean and Neoarchaean overgrowths. However, no correlation exists between δ18O and U, Th, Th/U, % common 206Pb, and % discordance for any sample (Fig. 4) suggesting that the range in δ18O cannot be directly associated with differences in radiation dose or U–Th–Pb systematics. Consequently, we contend that the range of compositions measured within the sample zircons accurately reflects that of their primary magmas.

Comparison between zircon and quartz δ18O

Neoarchaean migmatite sample 195392, which includes components coeval with the QGC, records distinct measured δ18OZr values for different zircon age populations (Fig. 3n). The weighted mean composition of the oldest population at 5.0 ± 0.4‰ (3,627 ± 38 Ma) lies within the lower limit of the field for mantle zircon. This probably reflects the original composition of an inherited Eoarchaean tonalitic gneiss component within the 195392 migmatite unit. The compositions of Mesoarchaean and Neoarchaean zircon populations at 3.7 ± 0.5, 4.0 ± 0.5 and 4.7 ± 0.4‰ are systematically lower than those of the inherited Eoarchaean component and imply a different petrogenetic process operating during their crystallization. This is interpreted to be the melting of hydrothermally altered, broadly tonalitic rocks. We assume that the youngest (2,570 ± 4 Ma) zircon population at 4.7 ± 0.4‰ was in equilibrium with the rest of the 195392 migmatite during the emplacement, minus unmelted inherited components e.g. zircon cores. Using the empirical zircon–quartz fractionation calibration of Trail et al. (2009) the 2.57 Ga population would correspond to a calculated δ18OQtz composition of 7.0 ± 0.4‰. Measured duplicate δ18OQtz values of 8.7 ± 0.1‰ and 9.2 ± 0.1‰ indicate a discrepancy of ~1.7 to ~2.2‰ between measured and calculated δ18OQtz that is likely to reflect the recrystallization of quartz during regional metamorphism (e.g. Valley and Graham 1996; King et al. 1997). Homogeneous Qôrqut granite sample 195376, with a mean δ18OZr composition of 4.2 ± 1.2‰ would correspond to a calculated δ18OQtz value of 6.5 ± 1.2‰. Duplicate measured δ18OQtz at 8.2 ± 0.1‰ and 8.9 ± 0.1‰ again indicates disequilibrium on the order of ~1.7 to ~2.4‰ that is also attributed to quartz recrystallization. Overall, it may be the case that even the “freshest” Archaean samples reflect the effects of recrystallization during metamorphism. Under these conditions, highly robust zircon is capable of retaining its primary igneous composition (Page et al. 2007) while other phases, such as feldspar and quartz recrystallise during metamorphism, are more prone to diffusive exchange and will re-equilibrate with the surrounding rock (Wei et al. 2002).

The generation of low δ18O magmas

The intrusion of new granitoid magmas can fracture the surrounding country rock and drive groundwater through lateral temperature gradients (Taylor 1977). Low δ18O meteoric fluids become heated and can undergo isotopic exchange with wall rocks at heterogeneous scales that are dictated by fracture permeability. Convection cells eventually collapse following the crystallization and cooling of the intrusion. In addition, multiple episodes of magmatism can lead to the melting of earlier altered products and the acquisition of averaged compositions from fossilized hydrothermal systems (Bacon et al. 1989; Bindeman et al. 2001, 2008). The significant amount of isotopic and age heterogeneity within some samples from the Greenland suite probably relate to the operation of such open fluid system processes (e.g. assimilation, magma mixing) during zircon crystallization. Previous work on the IGC leucogranites has also provided strong evidence that crystallization took place within a water-rich environment (Nutman and Bridgwater 1986). In such settings and possible analogous ones for the Ikkattoq gneiss, Neoarchaean migmatite and QGC samples, there is a likely potential for large isotopic fractionations and shifts towards low δ18O magmatic compositions. Low δ18O magmas can also form by the melting of 18O-depleted oceanic slabs (Eiler 2001). However, whole rock and isotopic compositions from these granitic samples are inconsistent with direct melting of a basaltic source and require an intermediate compositional step matched by the regional tonalites (Brown et al. 1981; Moorbath et al. 1981; Nutman and Bridgwater 1986).

Major Archaean–Proterozoic faults and regional mylonites cut the orthogneiss complexes of the Nuuk Region (Fig. 1). Local hydrothermal alteration is often associated with those major shear zones, as well as other intrusive contacts (Glassley et al. 1984; Nutman 1982). Measured whole rock δ18O values typically demonstrate significant fractionations within close proximity to these features. For example, emplacement of mid to late Archaean Tarssartôq dikes in the Isua area (possibly equivalent to the Ameralik dike swarms in Godthåbsfjord; Nutman and Bridgwater 1986) results in their host TTG gneisses showing lower δ18O in proximity to the dikes (Read 1976, Baadsgaard et al. 1986b). Also, an altered mylonite from within the Ataneq fault zone near Isua records a whole rock composition of −1.3‰ (Baadsgaard et al. 1986b). This alteration is attributed to low δ18O fluids such as meteoric water emanating from such structures (Longstaffe 1979; Baadsgaard et al. 1986b). These relationships provide a clear mechanism by which low δ18O fluids have been able to access and locally alter the TTG gneisses during the Archaean.

It is also noted that hydrothermal alteration is generally restricted to geothermal systems at relatively shallow crustal levels, above the brittle–ductile transition. Therefore, it could be difficult to reconcile the occurrence of low δ18O zircon with these processes in mid-crustal levels granitoids. However, these samples are collected from terranes that have been assembled into their current configuration along Neoarchaean mylonites, probably reactivated during several tectonic switching events through the Archaean (Friend et al. 1987, 1988; Nutman et al. 1989; Crowley 2002; Friend and Nutman 2005a; Nutman and Friend 2007). Consequently, it is feasible that the broadly tonalitic protolith sources that these melts were derived from could have been at medium to high crustal levels at some early stage during their history. At lower pressures, these rocks may have been exposed to hydrothermal alteration by fluids percolating down from the surface or near surface. The most efficient conduits for this would be major faults with extensional components. Following alteration, they could be transported back to mid-crustal levels, prior to the partial melting that formed low δ18O granitic magmas, which in turn crystallized low δ18O zircon.

Worldwide occurrences of low δ18O magma and zircon, and their petrogenetic significance

Occurrences of low δ18O (<5‰) zircons are common in the Phanerozoic, but Archaean and Proterozoic zircon with low δ18O is a significantly minor component of the global zircon δ18O compilation of Valley et al. (2005), with only a scattering of analyses falling within this compositional range (Fig. 5). Here, a new field is established for Archaean “low δ18O zircon” that falls within a compositional range from δ18O = 2.0 to 4.0‰. The upper limit is conservatively set to resolve values from that of mantle zircon, given the limited analytical precision of ion probe measurements which are on the order of ±0.5‰ or better. The lower limit to this low δ18O field may be much lower than 2‰, however, Archaean zircon compositions <2‰ have not been reported to date. The prevalence of low δ18O Greenland zircon values and the absence of mantle-like compositions for the granitic-granodioritic samples <3.7 Ga is striking, particularly in comparison with the Eoarchaean tonalitic dataset of Hiess et al. (2009), in which all zircon falls within error of the mantle field. This suggests a different style of crustal petrogenesis, which facilitated the incorporation of meteoric water and the formation of hydrothermally altered materials, was active during the later parts of the Archaean that is not seen in the earlier dataset. The high δ18O value of sample G97/111 and others in the global compilation, however, indicate that processes involving low δ18O fluids and magmas were not ubiquitous, and crustal components with δ18O above the mantle were clearly also involved in Archaean magmatism. The dataset for this study, however, reinforces the earlier observation in Hiess et al. (2009) that supracrustal materials that had been involved in low temperature weathering cycles near the Earth’s surface leading to the production of δ18O compositions >7.5‰, were not involved to any significant degree in the Archaean magmatism that formed the Greenland suite samples.
https://static-content.springer.com/image/art%3A10.1007%2Fs00410-010-0578-z/MediaObjects/410_2010_578_Fig5_HTML.gif
Fig. 5

δ18O values of dated zircon from Archaean Greenland in this study, Hiess et al. (2009) and the global compilation of Valley et al. (2005). Field for mantle zircon from Valley et al. (1998) and Archaean–Hadean “supracrustal zircon” from Cavosie et al. (2005). A new field is defined for Archaean “low δ18O zircon”

Low δ18O zircons are commonly observed in samples <150 Ma and have generally been targeted for analysis due to the previous identification of their low δ18O parent magmas (e.g. Gilliam and Valley 1997; Bindeman and Valley 2000, 2001; Monani and Valley 2001). The apparent rarity of low δ18O magmas and zircon before about 150 Ma may plausibly relate to a preservation or sampling bias (Valley et al. 2005) or alternatively may accurately reflect the rarity of such melt compositions (Balsley and Gregory 1998). In previously studied low δ18O samples, fractionated oxygen isotopic zircon compositions are typically interpreted to reflect the remelting or assimilation of hydrothermally altered wall rocks in shallow, sub-volcanic, felsic magma chambers, e.g. The British Tertiary Igneous Province (Gilliam and Valley 1997; Monani and Valley 2001), Yellowstone (Bindeman and Valley 2000, 2001) and Mesozoic granitoids of eastern China (Wei et al. 2002).

Modern basalts from Iceland show anomalous δ18O values that are systematically lower than basalts of other oceanic islands or oceanic ridges, and cannot be explained by processes of secondary alteration (Muehlenbachs et al. 1974). Mechanisms offered to produce these magmas include exchanges with hydrothermally altered rocks or meteoric water or as the result of a distinct mantle source beneath Iceland (Muehlenbachs et al. 1974). Glaciated Pleistocene silicic volcanic provinces such as Kamchatka and the Aleutians also recorded low δ18O magmas and phenocrysts (Bindeman et al. 2001, 2004). Here, at high latitudes, shallow hydrothermal systems are interpreted to be charged with highly fractionated, glacially derived, melt waters (δ18O < −25‰). These extreme values require significantly less fluid–rock interaction to lever the bulk composition of altered wall rock.

Neoproterozoic igneous zircon, with extremely low δ18O has also been measured within coesite-bearing Triassic rocks of the Dabie-Sulu terrane, China (Rumble et al. 2002; Zheng et al. 2003, 2004, 2007; Zhao et al. 2008). Despite experiencing ultrahigh pressure metamorphic conditions, the oxygen isotopic value of these zircons has been interpreted to reflect the composition of their primary magmas. In these samples, the zircon protolith is believed to have experienced high temperature interaction with glacially derived melt water formed during Neoproterozoic Snowball Earth events (Rumble et al. 2002). The records of ancient hydrothermal systems preserved in the zircons can thus be used as palaeoclimate proxies.

It is interesting to speculate on the source of low δ18O Archaean hydrothermal fluids that interacted with Greenland tonalitic protoliths to produce the low δ18O granitoids. Glacial melt water has been previously argued as an effective agent to charge low δ18O magmatic systems and provide a regional fingerprint on rocks, magmas, zircons and other phases (Bindeman et al. 2001; Rumble et al. 2002; Bindeman et al. 2004). In these studies, arguments for glaciated conditions have been substantiated by various other independent lines of evidence. For example, in Kamchatka, several phases of voluminous glaciation during the Pleistocene (Grosswald 1998) are demonstrated to have covered volcanic edifices across the peninsular (Savoskul 1999) and reached the coastline (Prueher and Rea 2001). The hypothesis of a global glaciation during the Cryogenian period has been based on a worldwide association of glacial deposits and limestones on Neoproterozoic platforms, with associated isotopic evidence (Hoffman et al. 1998).

The Earth’s surface conditions during the Archaean have been extensively debated (e.g. Knauth and Lowe 2003; Perry and Lefticariu 2003; Kasting and Ono 2006). One argument has suggested the existence of a cold, glaciated climate, forced by a faint young sun (Ringwood 1961; Sagan and Mullen 1972; Zahnle et al. 2007). Other competing environmental factors include a CO2- and CH4-rich atmosphere that was outgassed during vigorous phases of volcanic activity (Walker 1977; Kasting 1987, 1993; Kasting et al. 2006) and is suggested to have driven an Archaean temperate or greenhouse Earth. Although it is interesting to speculate on the role of early climate conditions in generating the pervasive and long-lived low δ18O signatures of Archaean granitoids presented here, the paleolatitude of these Greenland rocks during the Archaean cannot be established, nor can the degree of fluid interaction or δD, to confidently constrain fluid compositions or sources. Therefore, we conservatively interpret these signatures as originating in, shallow hydrothermal systems fed by meteoric fluids at moderate temperatures.

Magmatism during extensional tectonics

Eoarchaean tonalitic magmas from the IGC have been associated with the formation of continental crust at convergent plate boundaries (e.g. Nutman et al. 2007a; Hiess et al. 2009). For example, well-preserved 3.7 Ga tonalites north of Isua (Nutman and Bridgwater 1986) are separated from 3.8 Ga tonalites to the south (Nutman et al. 1999) by early Archaean mylonites along the Isua supracrustal belt (Nutman 1984; Nutman et al. 1997). Such crust would be highly unstable following a relaxation of the shortening tectonic regime, leading to orogenic collapse (e.g. Hermann et al. 2001). These processes may be repeated during multiple cycles over the duration of a large collisional event, with such extensional settings known to have high heat flow and granite production by intra-crustal melting (e.g. Rubatto et al. 1998; Beltrando et al. 2007).

Within the Isua area, the timing of formation for the ~3.7 Ga grey and ~3.65 Ga white gneisses can be clearly resolved by U–Pb dating (Baadsgaard et al. 1986a). Phases of extensional tectonics and continental breakup can produce abundant mafic magmas that can intrude the crust as dike swarms (McKenzie and Bickle 1988). It has been previously argued that the Inaluk dikes of the Isua region are such evidence for a high heat flow and regional extension during the formation of the granitic white gneisses 3,660–3,640 Ma (Nutman and Bridgwater 1986; Friend and Nutman 2005b). Extensional tectonics would be highly favourable for the generation of low δ18O magmas (Taylor 1977) as major faults can act as conduits for meteoric fluids to enter the crust. Crustal thinning can lead to asthenospheric upwelling and the compression of local geotherms that promote melting and metamorphism of the lower crust (Wickham and Oxburgh 1985; Sandiford and Powell 1986). A schematic model depicting the generation of evolved granitoid magmas in the Nuuk region is presented in Fig. 6.
https://static-content.springer.com/image/art%3A10.1007%2Fs00410-010-0578-z/MediaObjects/410_2010_578_Fig6_HTML.gif
Fig. 6

Schematic diagram depicting key components and processes involved in the generation of evolved granitoids in the Nuuk region

An extension related environment has been envisaged for formation of the ~3.64 Ga iron-rich suite (Nutman et al. 1984). Compositionally, the suite resembles Proterozoic rapakivi granite–ferrodiorite–norite (anorthosite) associations, which characteristically form in rifted, recently thickened, sialic crust following collisional tectonics (Emslie 1978). This tectonic regime would again provide sufficient heat from the mantle to melt the lower continental crust. Anatectic melts mixed and mingled with residual basic magmas that largely stayed in the ductile zone in the deep crust. The restriction of this suite to the ductile zone may explain why meteoric water had not penetrated to depress δ18O values. Underplating by a cushion of mafic magma is envisaged and melting of deep crust that may have included some sedimentary rocks.

Other cycles of juvenile crust formation and subsequent differentiation with tectonic switching may be represented later in the Archaean. Dated deformational fabrics in the Kapisillik and Færingehavn areas suggest that during the formation of the QGC, the crust was partitioned by active ductile shear zones with a component of extensional displacement (Nutman and Friend 2007; Nutman et al. in press). This suggests that ductile deformation and amphibolite facies metamorphism continued until the end of the Neoarchaean and was synchronous with, or outlasted, the intrusion of the QGC.

Comparison of Archaean and Phanerozoic magmatic styles

Recent studies have reported the combined O and Hf isotopic composition of zircon from classic Phanerozoic I-type plutonic systems in Eastern Australia and New Zealand (Kemp et al. 2007; Bolhar et al. 2008). Granitoids from the Lachlan Fold Belt in southeastern Australia record correlated shifts in εHf(T) and δ18O within individual suites. Mafic dioritic, basaltic and gabbroic lithologies anchor the arrays at mantle-like δ18O, and chondritic or super-chondritic 176Hf/177Hf zircon isotopic compositions. Each suite then progressively trends towards higher δ18O (>7‰), and sub-chondritic or lower initial εHf in their tonalitic and granitic samples. These arrays were interpreted to reflect the assimilation of surrounding 18O-enriched metasedimentary country rocks by the primitive mantle-derived melts. It was argued that this provides direct evidence that I-type magmatism drives the coupled growth and differentiation of the continental crust.

The Palaeozoic Australian suites demonstrate some overall similar evolutionary behaviour to that seen within the Archaean Greenland samples (Fig. 7). Compositions from both studies become progressively less radiogenic in εHf(T) as the bulk compositions become more evolved, reflecting the incorporation of older, low Lu/Hf crustal materials and they diverge away from mantle δ18O through time. However, they differ significantly in that the reworked crustal materials from the Lachlan Fold Belt were sedimentary rocks, none of which were hydrothermally altered prior to magmatism, such that their O isotopic fractionations are in a reversed sense compared to that of most of the Greenland rocks. That is, they trend towards higher δ18O, above mantle values (similar to that of granitic augen gneiss G97/111), rather than the lower δ18O seen in the granitic, trondhjemitic, granodioritic and migmatitic gneisses of the other Greenland granitic suites (Fig. 7). All arrays are anchored to the mantle-like oxygen and near-chondritic hafnium isotopic composition of the Eoarchaean inherited components reflective of the juvenile tonalitic protolith to the evolved granites. It is emphasized that the Greenland trends of decreasing δ18O and more negative initial εHf are repeated in multiple rock suites spanning over a billion year time span.
https://static-content.springer.com/image/art%3A10.1007%2Fs00410-010-0578-z/MediaObjects/410_2010_578_Fig7_HTML.gif
Fig. 7

Correlated shifts in δ18O and εHf(T) for Greenland zircon. Fields for mantle zircon from Valley et al. (1998) and CHUR from Bouvier et al. (2008). Greenland sample trends anchored at compositions from Eoarchaean tonalites of Hiess et al. (2009). Comparative isotopic arrays for zircons from Phanerozoic granitoid suites from Kemp et al. (2007) in grey

Granitoids from the Cretaceous Separation Point Suite of New Zealand record heterogeneous zircon εHf(T) compositions ranging from +12 to −4 units indicating the sources for these magmas were less radiogenic than the depleted mantle at ~120 Ma (εHf(T) = ~+16). Variable δ18O values range from 8 to 0‰ indicating their magmas incorporated both 18O-enriched supracrustal material, and 18O-depleted country rocks, hydrothermally altered by meteoric water. In this latter respect, the New Zealand rocks show clear similarities to the samples from the Greenland suite. The heterogeneity of isotopic compositions within individual samples from the Separation Point Suite has also been associated with open-system processes such as assimilation during magma evolution and zircon crystallization.

Although the details of reservoirs contributing to each magmatic system differ from case to case, the data for southwest Greenland samples presented in Hiess et al. (2009), and this study suggest that the styles of crustal petrogenesis operating during the Archaean were different than those on the modern Earth. The rates and styles of crustal recycling and subduction may have been different, with a greater role for extensional tectonic regimes in crustal reworking during the Archaean. This resulted in the more common development of hydrothermal systems associated with felsic magma genesis, and perhaps enhanced by low temperature climatic conditions, lead to the widespread formation of low δ18O magmas. Early crustal formation and evolution, however, need not be attributed to massive super-event cycles or unusual processes, but rather to progressive episodic behaviour of continental growth and reworking through time.

Conclusions

  • Initial εHf(T) values for zircons from 3.69 to 2.56 Ga granitic rocks range from chondritic to highly negative values and indicate granitoid formation by the remelting of surrounding TTG rocks and documenting intensive processes of crustal reworking throughout the Archaean.

  • Inherited Eoarchaean zircon cores record chondritic initial εHf values and mantle-like δ18O isotopic values, reflecting derivation from surrounding older tonalites. In contrast, younger zircons, recording their host magma composition, diverge from these primitive values.

  • Five of six analysed evolved granitoids, spanning a large (>1 billion year) temporal and geographic range, have low δ18O compared to mantle compositions, indicating the incorporation of an 18O-depleted source component. These are the first reported low δ18O values from Archaean zircons. The signatures are interpreted to reflect the influence of crustal protoliths that were hydrothermally altered by meteoric fluids.

  • Extensional regimes, promoting the formation of hydrothermal cells, are suitable tectonic environments for the generation of the evolved granitic magmas and are a geologically plausible scenario for the origin of the analysed samples.

  • Archaean plutonic systems have distinct geochemical differences from those from the Phanerozoic reflecting subtle changes in crust formation and evolution processes through time, with a likely more pronounced role for extensional tectonics in early crustal reworking.

Acknowledgments

We thank two anonymous reviews for helpful comments that improved the clarity of the paper. Apart from VM97/01, the samples were collected during field work funded by ANU or the Geological Survey of Denmark and Greenland who we thank for permission to publish data on these samples. The late Vic McGregor is acknowledged for collecting sample VM97/01. Clark Friend is acknowledged for providing samples 159352 and 159376. We thank Bud Baadsgaard for supplying zircon separates of 248251 and 248212. All analytical work was supported by the Australian Research Council grants DP0342798 and DP0342794 and was undertaken while Hiess was a PhD student at ANU supported by APA and Jaeger scholarships. We thank Shane Paxton and Jon Mya for zircon separations; Ryan Ickert and Peter Holden for contributions to SHRIMP oxygen analysis development; Malcolm McCulloch for access to the Neptune; Les Kinsley for assistance with running LA-MC-ICPMS; Steve Eggins for providing a template for Hf data reduction; Chuck McGee for technical assistance with LA-ICPMS analysis; Antti Kallio for providing LABRAT software; Yuri Amelin, Bob Rapp, Joerg Herman and Trevor Ireland for helpful discussions. We thank John Eiler, Carsten Munker and Pete Kinny for helpful reviews of an earlier version of this manuscript as a chapter in Hiess’ PhD thesis.

Supplementary material

410_2010_578_MOESM1_ESM.pdf (1.5 mb)
Supplementary material 1 (PDF 1,500 kb)

Copyright information

© Springer-Verlag 2010