JOURNAL OF PETROLOGY
VOLUME 52
NUMBER 5
PAGES 985^1009
2011
doi:10.1093/petrology/egr014
Implications of the Nuvvuagittuq Greenstone
Belt for the Formation of Earth’s Early Crust
JONATHAN O’NEIL1*, DON FRANCIS2 AND RICHARD W. CARLSON1
1
DEPARTMENT OF TERRESTRIAL MAGNETISM, CARNEGIE INSTITUTION OF WASHINGTON, 5241 BROAD BRANCH
ROAD, N.W., WASHINGTON, DC 20015, USA
2
EARTH & PLANETARY SCIENCES DEPARTMENT, MCGILL UNIVERSITY, MONTREAL, QC, CANADA, H3A 2A7
RECEIVED APRIL 9, 2010; ACCEPTED MARCH 9, 2011
U^Pb geochronology for the Nuvvuagittuq greenstone belt put a minimum age constraint of 3·8 Ga for the supracrustal lithologies.
Recent 142Nd work raised the possibility that the dominant lithology
of the belt formed at 4·28 Ga, which would make it the only
known remnant of Hadean crust preserved on Earth. The dominant
lithology of the belt has a mafic composition that consists of gneisses
ranging from cummingtonite amphibolite to garnet^biotite schist
composed of variable proportions of cummingtonite þ biotite þ
quartz, plagioclase garnet anthophyllite cordierite. The
composition of this unit ranges from basalt to andesite and it is
divided into two distinct geochemical groups that are stratigraphically separated by a banded iron formation (BIF). At the base of the sequence, the mafic unit is mainly basaltic in composition and
generally has relatively low Al2O3 and high TiO2 contents, whereas
above the BIF, the unit is characterized by high Al2O3 and low
TiO2 contents and exhibits a wider range of compositions from basaltic to andesitic. The low-Ti unit can be further subdivided into a
trace element depleted and a trace element enriched subgroup. The
high-Ti unit is characterized by relatively flat REE patterns as
opposed to the low-Ti gneisses, which display light REE-enriched
profiles with flat heavy REE slopes. The incompatible element
depleted low-Ti rocks have U-shaped REE profiles.The geochemical
groups have compositional analogues in three types of ultramafic
sills that exhibit the same stratigraphic succession. Generally, the
mafic gneisses have low Ca, Na and Sr contents, with many samples
having CaO contents 51wt %. Such low Ca contents are unlikely
to represent the original composition of their igneous precursors and
are interpreted to reflect intensive alteration of plagioclase. These
compositional characteristics along with the presence of cordierite þ anthophyllite suggest that the protoliths of the mafic gneisses
were mafic volcanic rocks exhibiting variable degrees of hydrothermal
In this study we present the petrography and geochemistry
of the mafic^ultramafic rocks of the Nuvvuagittuq greenstone belt. With a minimum age of 3·75 Ga, the
Nuvvuagittuq greenstone belt is at least Eoarchean in age.
142
Nd deficits have recently been identified in the
Nuvvuagittuq mafic rocks (O’Neil et al., 2008) suggesting
an age of nearly 4·3 Ga, which would make it the only
known remnant of Hadean crust preserved on Earth.
*Corresponding author. Telephone: 202-478-8482. Fax. 202-478-8821.
E-mail. [email protected]
ß The Author 2011. Published by Oxford University Press. All
rights reserved. For Permissions, please e-mail: journals.permissions@
oup.com
alteration. The high-Ti compositional type shares geochemical characteristics with tholeiitic volcanic suites with low Al2O3 and high
TiO2 contents and is consistent with crystal fractionation at low
pressures under dry conditions. In contrast, the low-Ti compositional
group is geochemically similar to boninitic and calc-alkaline volcanic
suites.The high Al2O3 and lowTiO2 contents in the andesitic compositions suggest the early crystallization of Fe^Ti oxides and late
appearance of plagioclase, and are more consistent with fractionation
at elevated water pressures. The succession from ‘tholeitic’ to
‘calc-alkaline’ magmatism seen in the Nuvvuagittuq greenstone belt
is typical of the volcanic successions of many younger Archean greenstone belts. Regardless of the exact tectonic setting, this volcanic succession suggests that the geological processes responsible for the
formation and evolution of Archean greenstone belts were active at
3·8 Ga and perhaps as early as 4·3 Ga.
Nuvvuagittuq greenstone belt; Eoarchean; Hadean;
tholeitic magma; calc-alkaline magma; boninitic magma; Superior
Province; crustal evolution
KEY WORDS:
I N T RO D U C T I O N
JOURNAL OF PETROLOGY
VOLUME 52
The Nuvvuagittuq greenstone belt thus offers a unique
opportunity to constrain the composition of the Earth’s
early crust and the processes responsible for the formation
of crust in the Eoarchean^Hadean.
The Hadean Eon is defined by the absence of a rock
record on Earth. Although there is little consensus, the
Hadean Eon is typically taken as spanning the time between the accretion of the Earth (4·567 Ga; Amelin et al.,
2002) and the age of the formation of the Acasta Gneiss
(4·03 Ga; Bowring & Williams, 1999). Until recently, the
only available Hadean samples were detrital zircons from
the Jack Hills conglomerate (4·4 Ga) (Wilde et al., 2001).
Although these zircons provide invaluable information
about the early Earth, their host rocks have been destroyed, leaving no direct samples of the Earth’s primordial
crust. Other than these detrital zircons, rare occurrences
of Eoarchean crust provide the only compositional constraints on the nature of the Earth’s early crust. These include the Acasta Gneiss, Canada (4·03 Ga; Bowring &
Williams, 1999), the Itsaq Gneiss Complex, Greenland
(3·85^3·60 Ga; Nutman et al., 1996), the Napier complex,
Antarctica (3·95^3·8 Ga; Williams et al., 1986; Simon &
Nigel, 2007), the Saglek^Hebron block, Labrador (3·78^
3·73 Ga; Collerson, 1983; Schiotte et al., 1989), and the
Anshan area, China (3·8 Ga; Song et al., 1996; Liu et al.,
2008; Wu et al., 2008, 2009; Nutman et al., 2009). Thus a
period of over 500 Myr of the Earth’s history is unrepresented in the rock record and the nature of the Earth’s
early crust and the processes responsible for its formation
remain largely speculation. Moreover, although primary
crust forming today is mafic in composition, our knowledge of the Earth’s early crust is largely based on felsic
rocks because they are the most likely host rocks for the
zircons that provide reliable old U^Pb ages. Proposed
models for the Earth’s primordial crust range from felsic
to mafic in composition. Evidence suggesting a felsic composition for the early crust is based largely on the composition of the Jack Hills zircons, which indicates derivation
from an evolved granitic melt (Mojzsis et al., 2001;
Harrison et al., 2005, 2008; Blichert-Toft & Albare'de,
2008). A felsic primordial crust has been proposed to be
thermally unstable because the heat produced by the
radioactive decay of K, U and Th was considerably more
important (4 times higher) in the Hadean than it is
today (Kamber et al., 2005; Kamber, 2007). Kamber et al.
(2005) estimated that a felsic crust in the Hadean would
reach the solidus at a depth of 10 km and would reorganize itself into a layer enriched in incompatible elements on
top of a layer depleted in incompatible elements.
Moreover, any juvenile crust derived from the early
mantle would be mafic in composition. A felsic crust
would require differentiation or melting of a pre-existing
mafic body. Kamber et al., (2005) and Kamber (2007)
thus proposed a mafic early crust that was low in
NUMBER 5
MAY 2011
heat-producing radioactive elements. Even such a basaltic
crust would melt once it reached 40 km in thickness
(Kamber et al., 2005), forming evolved felsic granitoids
that could have been the source of the Hadean zircons.
This conclusion is supported by the Lu^Hf isotopic composition of the Jack Hills zircons, which suggests that their
host rock is compatible with a tonalite^trondhjemite^
granodiorite (TTG) suite derived from a basaltic precursor (Blichert-Toft & Albare'de, 2008; Kemp et al., 2010).
Despite these findings, the composition of the primordial
crust and how it formed remain poorly understood, in
part because the search for ancient crust is biased towards
felsic rocks because of the reliance on zircon as the mineral
of choice for dating ancient crust. For this reason, understanding the compositional characteristics of a mostly
mafic succession such as contained in the Nuvuaggittuq
belt may provide important new clues to the nature and
petrogenesis of Earth’s first crust.
GEOLOGIC A L S ET T I NG
The Nuvvuagittuq greenstone belt is located in the
Northeastern Superior Province (NESP) on the east coast
of Hudson Bay (Fig. 1). The NESP is dominated by
Neoarchean plutonic suites containing thin keels of highly
metamorphosed supracrustal rocks. The NESP has been
divided into two distinct terranes on the basis of U^Pb
geochronology and Nd isotopic signatures (Boily et al.,
2009). To the east, the Arnaud River terrane is composed
of Neo- to Mesoarchean rocks (53·0 Ga) and is isotopically
juvenile with Nd model ages (TDM) that are less than 3·0
Ga. To the west, the Hudson Bay terrane is characterized
by older TDM, lower initial Nd isotopic values, and zircon
inheritance ages suggesting the recycling of ancient
(2·9^3·8 Ga) crust. The Nuvvuagittuq greenstone belt lies
within the Hudson Bay Terrane and is composed of an
10 km2 Eoarchean to Hadean volcano-sedimentary sequence surrounded by younger 3·6 Ga tonalites
(David et al., 2009; O’Neil et al., 2007, 2008). The belt has
been isoclinally folded into a north-plunging synform and
then refolded into an open south-plunging synform (Fig. 1;
O’Neil et al., 2007; David et al., 2009). The large-scale
north-closing synform with a north^south axial trace that
dominates the map pattern of the region has a well-defined
cylindrical geometry with a fold axis plunging moderately
to the south. The tight isoclinal fold to the west of the belt
is defined by the attitude of the main schistosity. This fold
cannot be traced in detail because of the absence of
marker horizons but can be followed by the varying orientation of the foliation. This fold also has a well-defined cylindrical geometry, with a steeply east-plunging fold axis,
and therefore is almost vertical. The dominant lithology
of the Nuvvuagittuq Belt is a heterogeneous
cummingtonite-rich amphibolite. O’Neil et al. (2007)
called this compositionally and mineralogically
986
O’NEIL et al.
NUVVUAGITTUQ GREENSTONE BELT
Fig. 1. Geological map of the Nuvvuagittuq greenstone belt (modified from O’Neil et al., 2007, 2008). Dashed rectangle shows the location of
Fig. 8.
987
JOURNAL OF PETROLOGY
VOLUME 52
heterogeneous package of mafic gneisses the
‘Faux-amphibolite’, noting the dominance of cummingtonite instead of common hornblende in these mafic amphibolites. The amphibolites are intruded by a series of mafic
and ultramafic sills. The sills are more abundant in the
western limb of the synform, but also appear as smaller
bodies interior to the banded iron formation^Si-formation
horizon in the eastern limb (Fig. 1). The ultramafic sills
range from 5 to 30 m in thickness and are generally dismembered along strike, forming boudins a few tens to hundreds of meters in length. The interiors of the ultramafic
sills are serpentine-rich with lesser amounts of talc, tremolite, hornblende and chromite. The ultramafic sills also
contain thin amphibole-rich layers, and are characterized
by thin (1m) grey to dark green amphibole-rich margins
that are dominated by hornblende (O’Neil et al., 2007).
Locally, the ultramafic sills display gabbroic tops suggesting that they represent differentiated intrusions. The polarity of these gabbroic tops suggests that the isoclinal
synform is in fact a syncline. Gabbroic sills toward the
centre of the synform are highly deformed and display a
strong gneissic texture, in contrast to the more homogeneous coarse- to medium-grained texture of gabbros on
the limbs. The Nuvvuagittuq greenstone belt also contains
chemical sedimentary rocks that include a finely banded
iron formation (BIF) and a more massive silica-formation
(O’Neil et al., 2007). These together define an easily recognizable stratigraphic horizon that can be traced around
the entire belt. The belt has been metamorphosed to at
least upper amphibolite-facies conditions (O’Neil et al.,
2007; Cates & Mojzsis, 2009; Scher & Minarik, 2009),
with the possible exception of the southwestern corner of
the map area, where a typical mafic greenschist-facies mineral assemblage of chlorite þ epidote þ actinolite is
observed (O’Neil et al., 2007). Similar rocks that contain
relicts of garnet that have been completely retrograded to
chlorite are exposed in the southeastern edge of the belt,
suggesting that the greenschist-facies rocks may represent
a retrograde assemblage.
G E O C H RO N O L O G I C A L
CONST R A I N TS
The oldest zircon U^Pb dates from the Nuvvuagittuq
greenstone belt have been obtained from rare thin felsic
bands (10^50 cm in width). These bands have a tonalitic
composition (67^73 wt % SiO2) and consist of a
fine-grained assemblage of plagioclase, quartz, and biotite.
Zircons from one of these felsic bands have yielded a
U^Pb age of 3817 16 Ma that is interpreted to be the
crystallization age of the rock (David et al., 2009). These
felsic bands are found only near the southern nose of the
isoclinal synform within a limited area of at most 300 m2
that displays more intensive shearing and deformation
NUMBER 5
MAY 2011
than the rest of the Nuvvuagittuq greenstone belt.
Although these felsic bands appear to be concordant,
their exact relationship to the adjacent lithologies is unclear owing to the extent of shearing. Cates & Mojzsis
(2007) obtained a minimum age of 3750 Ma based on U^
Pb analyses of zircons from the same lithology in an outcrop that they suggested preserves a cross-cutting relationship with the supracrustal rocks. We were, however,
unable to find any clear evidence of the cross-cutting relationship between the felsic bands and their enclosing lithologies. Although the exact nature of these felsic bands
remains unclear, they are commonly found within what
we interpret to be gabbroic sills, suggesting that they may
also be intrusive. If indeed the felsic bands do intrude or
cross-cut the supracrustal rocks, their age would represent
only a minimum age for the Nuvvuagittuq greenstone
belt. Zircons from a conglomeratic unit yielding
207
Pb/206Pb ages ranging from 3770 Ma to 3258 Ma have
been reported by David et al. (2009). These zircons lie
along a discordia with an upper intercept of 3787 25 Ma
that would represent a maximum age for the deposition
of the conglomeratic unit. However, to put any age constraints on the supracrustal succession, the relationship between the conglomeratic unit and the other lithologies
needs to be known. These conglomeratic units, however,
are found only within the same high-strain zone that contains the felsic bands, making their relationship with the
adjacent lithologies similarly equivocal. Furthermore, as
pointed out by David et al. (2009), whether this lithology
indeed is a sedimentary conglomerate is also unclear. A definitive identification of the nature of the conglomeratic
unit in the Nuvvuagittuq greenstone belt and its stratigraphic relationship with the mafic gneisses thus requires
further study.
Recent work has shown that the mafic cummingtonite^
biotite gneisses have a 7^15 ppm deficit in 142Nd compared
with terrestrial standards (O’Neil et al., 2008). The only
other occurrence of 142Nd deficits in terrestrial rocks is
found in alkaline rocks from the Khariar nepheline syenite
complex in India that have been dated at 1·48 Ga and are
interpreted to represent a diluted remnant of a Hadean enriched reservoir (Upadhyay et al., 2009). The variation reported in the 142Nd/144Nd ratios of the Nuvvuagittuq
mafic rocks may reflect their derivation from the mantle
within 300 Myr of the Earth’s formation, while 146Sm was
still actively decaying (146Sm has a half-life of 103 Myr). A
correlation between the Sm/Nd ratios of the
Nuvuaggittuq mafic rocks and their 142Nd isotopic composition has a slope corresponding to an age of 4280
þ53/^81 Ma, which would make them the oldest known
rocks on the Earth (O’Neil et al., 2008). However, these
rocks also define a scattered 147Sm^143Nd isochron with
an age of 3·8 Ga. The considerable scatter of the data
about the 3·8 Ga best-fit line, however, appears to be the
988
O’NEIL et al.
NUVVUAGITTUQ GREENSTONE BELT
result of combining samples consistent with an older age
with samples for which the 147Sm^143Nd system has been
partially reset by a younger thermal event coinciding with
the emplacement of pegmatites around 2·7 Ga. For example, 147Sm^143Nd data only for a gneissic gabbro sill
intruding the amphibolites give an age of 4023 110 Ma
with significantly less scatter, which supports an older age
for the intruded mafic rocks (O’Neil et al., 2008). At 3·8
Ga, most Nuvvuagittuq rocks exhibiting a 142Nd deficit
also have negative initial e143Nd, resulting in 143Nd
depleted mantle model ages (TDM) older than 4 Ga consistent with a Hadean age. Several aspects of the Nd data
for the mafic rocks in the Nuvuaggittuq belt thus support
an age older than the 3·8 Ga minimum age provided by
zircon dates on rare felsic rocks whose stratigraphic relationship to the greenstones is unclear.
A N A LY T I C A L P RO C E D U R E S
Large samples (minimum 20 cm) were crushed, especially
for coarse-grained samples with large garnets, in the attempt to obtain representative whole-rock powders.
Weathered surfaces and quartz veins were sawn from samples prior to crushing. Samples were crushed in a steel jaw
crusher to centimeter-sized chips that were then ground to
powder in an alumina shatter box. Major and trace element abundances were analyzed by X-ray fluorescence
(XRF) at the McGill Geochemical Laboratories, using a
Philips PW2400 4 kW automated XRF spectrometer
system with a rhodium 60 kV end window X-ray tube.
Major elements, Cr, Ni and V were analyzed using 32 mm
diameter fused beads prepared from a 1:5 sample^lithium
tetraborate mixture. Sc, Rb, Sr, Zr, Nb and Y were analyzed using 40 mm diameter pressed pellets prepared at a
pressure of 20 tons from a mixture of 10 g sample powder
with 2 g Hoechst Wax C Micropowder.
Calibration lines were prepared using between 15 and
40 International Standard Reference Materials (more
details can be found at: http://www.eps.mcgill.ca/
Department/facilities.php?id¼trace).
Corrections for mass absorption effects were applied on
concentration values using a combination of alpha coefficients and/or Compton scatter. The accuracy for silica is
within 0·5% absolute, and is within 1% for other major
and within 5% for trace elements. Instrument precision is
within 0·3% relative, generally within 0·23% relative,
and the overall precision for beads and pressed pellets is
within 0·5% relative. Rare earth element (REE), Co, Zn,
Ga, Ge, Cs, Ba, Ta, Hf, and Th concentrations were determined by inductively coupled plasma mass spectrometry
(ICP-MS) at Activation Laboratories using a Perkin
Elmer SCIEX ELAN 6000 system, using a lithium metaborate^tetraborate fusion technique for digestion (code
4B2 research). Analytical techniques and detection limits
are available at: http://www.actlabs.com/page.aspx?
page¼555&app¼226&cat1¼549&tp¼12&lk¼no&
menu¼64. Duplicate analyses for major and trace elements
are shown in Electronic Appendix 3 (available for downloading
at
http://www.petrology.oxfordjournals.org).
Duplicate samples are not necessarily mafic amphibolite
samples. All duplicate samples are from the Nuvvuagittuq
greenstone belt and are selected to cover a wide range of
composition from felsic to ultramafic.
P E T RO G R A P H Y A N D
G E O C H E M I S T RY
Ujaraaluk unit (Faux-amphibolite)
By far the most abundant rock type in the Nuvuaggittuq
belt is heterogeneous gneisses composed of variable
amounts of cummingtonite þ biotite þ quartz, plagioclase garnet phlogopite anthophyllite cordierite
(O’Neil et al., 2007). The proportion of cummingtonite and
biotite in the gneisses varies widely, with lithologies ranging from amphibolite composed almost entirely of cummingtonite to garnet^biotite schist. Because of the large
variation in mineralogy at a wide range of spatial scales
in these rocks, rather than assigning many petrographically accurate rock names to each compositionally distinct
layer, O’Neil et al., (2007, 2008) instead defined this rock
package as a lithological mappable unit named the
‘Faux-amphibolite’. The term ‘Faux’ is a unit field name to
emphasize the fact that the unusual mineralogy of these
amphibolites gives them a light grey to beige colour in the
field, contrasting with the green^black colour typical
of most Archean amphibolites. Although the name
Faux-amphibolite has a historical meaning and has been
previously published, it may be ambiguous for those not familiar with the area and this lithology. Therefore, in an attempt to reduce the confusion, we use here the name
Ujaraaluk unit, a lithological unit comprising all variations of the Faux-amphibolite. It should be noted that the
exact same lithology has been referred to as the
Faux-amphibolite in previous publications (O’Neil et al.,
2007, 2008). Variations in the proportions of cummingtonite
and biotite also occur on a centimeter-scale, with the majority of the Ujaraaluk unit displaying an alternation of
biotite-rich and cummingtonite-rich banding (Fig. 2a).
The garnet content of the Ujaraaluk unit changes systematically across the Nuvvuagittuq greenstone belt. To the
west, the Ujaraaluk unit is practically devoid of garnet,
whereas in the east layers with up to 50% garnet are
common. The amount of garnet also varies considerably
within the garnet-bearing amphibolites, and there is
typically a meter-scale alternation of garnet-rich and
garnet-poor layers (Fig. 2b). The Ujaraaluk unit also contains discontinuous millimeter- to centimeter-scale quartz
ribbons that generally follow the schistosity (O’Neil et al.,
2007). Locally, the Ujaraaluk unit is characterized by an
989
JOURNAL OF PETROLOGY
VOLUME 52
(a)
(b)
Garnet-rich
Garnet-poor
(c)
Fig. 2. (a) Garnet-rich Ujaraaluk unit showing the common presence of biotite-rich and cummingtonite-rich centimeter-scale layers.
(b) Layered Ujaraaluk unit exhibiting garnet-rich zones and
garnet-poor zones. (c) Deformed pillow lavas with differential erosion
highlighting the selvages of the single pillows.
assemblage of cordierite, anthophyllite and biotite, with
little or no garnet. The cordierite is typically altered to
fine-grained pinite and chlorite.
Despite the large variation in mineralogy in the rocks
forming the Ujaraaluk unit, they generally have a mafic
NUMBER 5
MAY 2011
composition. Most have SiO2 contents between 42 and
58% and thus would be classified from basalt to andesite
in terms of their silica contents, with a few samples having
a dacitic composition (Table 1, Fig. 3, Electronic Appendix
1). These samples show a concomitant decrease in MgO
content (16·7^2·2 wt %) with increasing SiO2 (Fig. 4a).
The Ujaraaluk unit can be divided into two chemical
groups that define distinct trends in a plot of TiO2 vs Zr
(Fig. 4e), which are hereafter referred to as the high-Ti
and low-Ti Ujaraaluk units. These two chemical groups
are spatially separated in the Nuvvuagittuq greenstone
belt, with the high-Ti group occurring to the outside of
the BIF^silica-formation horizon in the limbs of the isoclinal synform, and the low-Ti group occurring within the
core of the synform, interior to the BIF^silica-formation
horizon. Although the high-Ti and the low-Ti groups are
similar in terms of major element composition, the low-Ti
unit exhibits a larger range of SiO2 and MgO contents
than the high-Ti unit (Figs 3b, c and 4 a). The low-Ti unit
ranges in composition from basalt through andesite to
rare dacite, whereas the high-Ti unit is largely basaltic to
basaltic andesite in composition (Fig. 3a). The low-Ti suite
has higher Al2O3 contents than the high-Ti suite at similar
MgO contents. Although garnet is occasionally found in
the high-Ti unit, it is significantly more abundant in the
low-Ti unit. Garnet-bearing Ujaraaluk unit samples generally have lower MgO and CaO, but higher SiO2 and
Al2O3 contents, than those without garnet (Fig. 4a, c and
d). Both the high-Ti and low-Ti units locally exhibit cordierite^anthophyllite^biotite
mineral
assemblages.
Samples containing cordierite and anthophyllite are characterized by higher MgO and lower SiO2 and CaO contents, with many samples containing less than 2 wt %
CaO (Fig. 4a and c). The cordierite^anthophyllite-bearing
rocks are also characterized by higher loss on ignition
(LOI) values, with most sample having values higher
than 5 wt %.
The high-Ti and low-Ti units have different trace element compositions and define distinct trends and groups in
plots of high field strength elements (HFSE), such as Zr,
Nb, Y, Ta, and REE (Fig. 4e^h) vs TiO2. The low-Ti unit
can be further divided into two sub-groups on the basis of
their trace element compositions. Rocks from the low-Ti
Ujaraaluk unit interior and adjacent to the BIF^
silica-formation horizon have distinctly lower HFSE and
REE concentrations than the innermost low-Ti rocks in
the core of the synform. In the following discussion, the
former are referred to as the depleted low-Ti Ujaraaluk
unit and the latter are termed the enriched low-Ti
Ujaraaluk unit, reflecting their relative incompatible trace
element contents.
The three chemical groups are best discriminated in a
plot of TiO2 vs Al2O3 (Fig. 4d), in which three distinct
trends are apparent. Each group is characterized by a
990
O’NEIL et al.
NUVVUAGITTUQ GREENSTONE BELT
Table 1: Major (wt %) and trace (ppm) element data for representative samples from the Ujaraaluk unit
High-Ti unit
Garnet-free
Sample:
PC-5
Garnet-bearing
Cordierite–anthophyllite
PC-132
PC-407
PC-418
PC-419
PC-421
PC-423
PC-429
PC-175
PC-408A
PC-408B
PC-424
PC-173A
PC-427
PC-453
SiO2
49·55
51·45
49·36
51·76
51·88
52·14
48·42
52·93
50·27
47·56
47·51
53·82
45·82
50·54
TiO2
1·13
0·90
1·04
0·95
0·94
0·91
1·26
0·80
1·03
1·02
0·92
1·21
0·73
0·56
45·36
0·96
Al2O3
15·21
14·42
15·27
15·05
15·88
16·76
14·65
14·06
18·07
15·34
14·70
14·73
15·26
11·19
14·02
MgO
8·08
8·69
7·39
10·84
8·73
8·61
10·98
6·62
6·88
8·70
9·12
6·81
15·62
15·41
15·64
FeO
12·06
10·58
14·42
10·97
10·16
9·49
12·52
10·87
12·95
15·41
15·92
12·58
12·11
12·50
13·47
MnO
0·24
0·22
0·18
0·24
0·23
0·25
0·25
0·30
0·14
0·23
0·23
0·13
0·18
0·28
0·18
CaO
8·65
8·28
8·36
3·83
4·68
4·96
6·18
8·82
1·06
7·07
7·29
4·49
1·03
4·62
1·55
Na2O
0·32
1·90
0·17
2·34
2·82
3·02
0·45
0·55
0·60
0·60
0·55
0·60
1·05
0·17
0·48
K2O
0·89
1·09
0·58
1·58
1·75
1·28
2·19
2·11
3·60
0·62
0·45
1·43
2·79
0·57
3·15
P2O5
0·08
0·08
0·08
0·08
0·10
0·07
0·09
0·08
0·04
0·09
0·11
0·12
0·08
0·02
0·07
LOI
2·62
1·83
2·32
1·65
1·55
1·55
1·89
2·26
4·02
2·05
1·77
2·98
4·47
3·34
39
31
22
62
63
40
Sr
97
75
75
104
126
120
96
84
11
70
75
73
63
40
44
Zr
55
56
50
55
50
49
67
61
49
60
51
73
38
28
53
Nb
Y
57
54
122
24
18
45
80
15
3·98
Rb
92
2·8
3·2
2·4
2·5
2·4
2·0
2·8
3·6
2·5
2·3
1·9
3·3
2·1
1·4
2·7
21·5
17·4
17·3
19·4
18·7
14·9
22·0
18·3
13·9
17·5
17·1
15·8
19·0
10·4
20·9
Ni
103
45
128
116
128
109
136
165
154
104
72
72
167
311
149
Cr
238
143
253
275
343
402
272
423
465
218
197
144
323
1533
300
V
290
280
270
184
200
150
233
214
293
258
235
239
224
190
276
Sc
39
39
34
33
29
30
38
29
38
33
32
37
27
36
35
Hf
1·5
1·3
1·6
Ta
0·19
0·18
0·21
Th
0·81
0·18
0·36
La
4·04
3·31
4·40
Ce
9·80
6·59
8·72
Pr
1·45
0·89
1·19
Nd
6·98
4·75
6·16
Sm
2·10
1·54
1·92
Eu
0·70
1·00
1·01
Gd
2·68
2·22
2·99
Tb
0·49
0·41
0·55
Dy
3·20
2·68
3·72
Ho
0·68
0·58
0·82
Er
2·04
1·82
2·56
Tm
0·30
0·28
0·39
Yb
1·96
1·80
2·47
Lu
0·31
0·26
0·37
(continued)
991
JOURNAL OF PETROLOGY
VOLUME 52
NUMBER 5
MAY 2011
Table 1: Continued
Depleted low-Ti unit
Garnet-free
Sample:
Garnet-bearing
Cordierite–anthophyllite
PC-131
PC-151
PC-222
PC-227
PC-410
PC-433
PC-163
PC-230
PC-313
PC-401
PC-402
PC-412
PC-135
PC-426
PC-430
SiO2
50·29
50·37
49·95
53·95
49·84
50·75
52·82
56·54
60·70
60·05
66·62
55·18
48·50
50·78
45·45
TiO2
0·34
0·38
0·37
0·29
0·38
0·37
0·42
0·41
0·43
0·33
0·32
0·38
0·36
0·29
0·45
Al2O3
16·39
17·39
16·84
14·56
17·70
16·71
19·19
18·76
18·87
16·06
15·30
18·51
16·41
14·79
18·57
MgO
11·43
9·35
9·82
11·35
10·48
11·03
6·97
4·39
3·01
4·96
3·01
6·28
13·61
13·59
14·38
FeO
8·87
9·30
9·92
9·75
9·30
9·34
9·41
9·25
8·97
11·84
5·07
10·70
11·57
11·02
9·39
MnO
0·18
0·21
0·22
0·22
0·20
0·19
0·29
0·39
0·17
0·42
0·12
0·27
0·13
0·11
0·21
CaO
5·17
7·49
8·93
6·58
8·84
6·84
2·30
5·08
0·43
2·38
4·72
2·53
0·46
0·25
0·78
Na2O
2·06
1·15
1·32
0·83
0·80
1·99
2·31
1·76
0·50
0·60
1·56
0·56
0·05
0·03
0·39
K2O
1·93
1·76
0·51
0·60
0·26
0·27
1·74
1·23
3·37
1·36
1·10
2·49
1·80
1·34
2·45
P2O5
0·03
0·03
0·03
0·06
0·03
0·04
0·03
0·04
0·02
0·02
0·03
0·04
0·03
0·03
0·03
LOI
2·91
2·02
1·75
1·82
1·95
2·10
4·00
2·18
3·00
1·76
2·00
2·29
6·37
6·62
Rb
64
44
14
14
7
7
52
40
86
49
40
73
Sr
126
53
59
30
64
40
52
56
12
29
57
30
Zr
22
24
25
19
23
24
27
26
24
23
21
24
Nb
Y
57
7·37
42
125
8
4
20
22
20
33
1·2
1·1
1·1
0·7
1·0
1·2
1·4
1·3
1·1
1·1
1·0
1·1
0·9
0·8
1·2
11·8
12·6
11·7
14·6
11·2
11·0
15·1
12·6
11·0
11·9
6·8
13·9
8·0
10·1
11·0
Ni
184
126
143
126
147
125
100
86
189
81
68
143
176
175
139
Cr
408
193
196
158
276
198
250
285
365
245
235
263
414
357
173
V
205
222
231
192
231
232
255
237
265
226
191
232
215
189
232
Sc
45
47
46
48
43
44
48
59
50
55
31
55
43
44
41
Hf
0·8
0·9
0·8
0·5
0·7
0·7
0·8
0·7
0·7
0·7
0·7
0·6
Ta
0·05
0·07
0·06
0·05
0·06
0·09
0·06
0·05
0·05
0·07
0·07
0·04
0·9
0·17
Th
0·46
0·36
0·28
0·33
0·52
1·36
0·16
0·20
0·19
0·19
0·43
0·45
1·37
La
2·8
2·3
2·2
1·9
3·3
3·0
2·4
2·1
3·8
2·5
1·4
2·5
8·1
Ce
5·7
4·9
4·8
4·0
5·0
4·6
4·8
2·7
6·0
3·4
2·9
3·1
13·4
Pr
0·7
0·6
0·6
0·6
0·6
0·5
0·6
0·3
0·7
0·4
0·3
0·4
1·4
Nd
2·9
3·1
2·8
2·8
2·7
2·2
2·6
1·9
2·9
2·0
1·5
1·8
5·6
Sm
0·8
0·9
0·8
0·9
0·7
0·6
0·7
0·7
0·7
0·6
0·4
0·5
1·3
Eu
0·4
0·4
0·4
0·3
0·3
0·2
0·4
0·2
0·4
0·2
0·1
0·2
0·3
Gd
1·2
1·2
1·2
1·4
1·2
1·0
1·3
1·4
1·0
1·4
0·6
1·0
1·5
Tb
0·2
0·3
0·3
0·3
0·2
0·2
0·3
0·3
0·2
0·3
0·2
0·2
0·3
Dy
1·8
2·2
2·0
2·2
1·8
1·7
2·5
2·3
1·3
2·5
1·3
1·5
1·9
Ho
0·4
0·5
0·5
0·5
0·4
0·4
0·6
0·6
0·3
0·6
0·3
0·4
0·4
Er
1·4
1·6
1·5
1·6
1·5
1·4
1·7
2·2
0·9
1·9
1·0
1·3
1·5
Tm
0·2
0·3
0·2
0·3
0·2
0·2
0·3
0·4
0·1
0·3
0·2
0·2
0·2
Yb
1·5
1·8
1·7
1·7
1·6
1·5
1·8
2·7
1·0
2·0
1·2
1·3
1·6
Lu
0·3
0·3
0·3
0·3
0·2
0·2
0·3
0·4
0·1
0·3
0·2
0·2
0·2
(continued)
992
O’NEIL et al.
NUVVUAGITTUQ GREENSTONE BELT
Table 1: Continued
Enriched low-Ti unit
Garnet-free
Sample:
PC-59
Garnet-bearing
Cordierite–anthophyllite
PC-150
PC-162
PC-250
PC-160
PC-225
PC-247
PC-275
PC-277
PC-278
PC-438
PC-452
PC-152
PC-442
PC-443
46·01
SiO2
52·04
54·92
52·42
57·92
55·41
52·76
56·24
55·73
63·11
57·69
57·80
55·09
48·70
46·82
TiO2
0·50
0·52
0·46
0·56
0·62
0·58
0·67
0·65
0·53
0·59
0·67
0·57
0·62
0·54
0·53
Al2O3
16·58
16·18
15·46
16·06
17·53
16·88
18·37
18·09
15·62
17·56
18·08
16·13
16·18
16·02
16·85
MgO
8·26
10·27
9·42
6·12
6·77
7·11
4·63
5·36
3·91
3·56
5·39
6·36
15·85
15·31
15·07
FeO
9·71
8·31
9·15
7·82
12·26
14·64
12·89
12·55
7·43
9·62
9·24
12·18
9·97
11·22
10·58
MnO
0·23
0·18
0·22
0·21
0·10
0·15
0·14
0·22
0·13
0·31
0·06
0·15
0·20
0·28
0·18
CaO
8·26
5·01
7·60
6·92
0·14
0·13
0·36
0·87
0·72
1·69
0·19
0·66
0·28
0·36
2·61
Na2O
1·00
1·79
1·20
1·53
0·16
0·07
0·12
0·29
0·12
0·21
0·16
0·03
0·35
0·04
0·77
K2O
0·40
0·10
1·01
0·94
1·81
2·70
3·16
3·25
4·42
4·41
3·75
3·63
0·43
1·71
1·56
P2O5
0·05
0·05
0·05
0·06
0·06
0·07
0·06
0·07
0·06
0·07
0·07
0·07
0·04
0·05
0·06
LOI
2·10
2·57
2·46
1·91
4·20
3·94
2·69
2·57
3·48
3·78
3·71
3·78
6·79
7·08
Rb
13
2
29
26
60
Sr
94
25
56
99
16
Zr
56
40
49
62
67
Nb
Y
86
125
107
131
144
101
108
8
7
12
10
14
5
8
67
72
72
56
67
71
60
10
4·89
43
36
4
4
48
65
55
38
3·2
2·0
2·8
3·5
4·3
3·8
4·0
4·0
3·1
4·0
4·0
3·8
3·5
2·9
1·4
13·6
12·6
14·0
13·7
14·7
16·5
20·9
19·3
5·6
15·1
8·6
10·4
15·1
11·3
11·6
Ni
98
81
136
63
133
162
134
88
93
74
172
148
88
101
67
Cr
177
240
351
139
180
162
194
167
238
155
189
263
146
237
204
V
185
206
181
189
228
213
263
246
199
214
245
216
219
205
207
Sc
40
47
38
38
36
44
59
55
39
40
20
39
38
40
38
Hf
1·4
1·7
2·0
1·9
2·1
2·0
1·9
1·6
1·9
1·6
1·0
Ta
0·23
0·29
0·34
0·32
0·27
0·34
0·31
0·44
0·36
0·27
0·13
Th
1·79
3·82
2·91
2·59
3·56
3·48
2·79
2·46
2·88
1·51
0·69
La
5·5
12·7
11·9
7·0
16·4
13·0
9·8
10·2
12·6
5·4
2·5
Ce
11·2
24·7
23·8
14·1
31·3
25·6
17·4
18·6
24·3
9·3
4·1
Pr
1·4
3·0
2·7
1·6
3·4
2·8
1·9
2·0
2·7
1·1
0·5
Nd
5·9
11·2
10·4
6·6
13·5
11·2
7·1
8·1
10·6
4·6
2·5
Sm
1·6
2·4
2·5
1·7
2·9
2·6
1·6
2·0
2·3
1·2
0·7
Eu
0·5
0·9
0·7
0·5
0·6
0·7
0·4
0·5
0·4
0·2
0·1
Gd
2·0
2·6
2·4
2·1
3·2
2·8
1·8
2·6
2·3
1·7
1·3
Tb
0·4
0·4
0·5
0·5
0·7
0·6
0·3
0·5
0·5
0·3
0·3
Dy
2·3
2·5
3·2
3·2
4·6
3·5
1·7
3·0
2·9
2·2
1·9
Ho
0·5
0·5
0·7
0·7
1·0
0·8
0·4
0·7
0·6
0·5
0·4
Er
1·6
1·5
2·1
2·3
3·1
2·4
1·1
2·1
1·9
1·6
1·4
Tm
0·2
0·2
0·3
0·4
0·5
0·4
0·2
0·3
0·3
0·2
0·2
Yb
1·7
1·5
2·1
2·5
3·1
2·5
1·0
2·0
2·0
1·6
1·4
Lu
0·3
0·2
0·3
0·4
0·5
0·4
0·2
0·3
0·3
0·2
0·2
Complete data table with additional trace element analyses is provided as Supplementary Data Electronic Appendix 1.
993
JOURNAL OF PETROLOGY
15
High-Ti
(a)
depleted Low-Ti
(b)
VOLUME 52
NUMBER 5
MAY 2011
Frequency
10
5
0
15
1
Grt-free Grt Crd-Bt
high-Ti
enriched low-Ti
depleted low-Ti
(d)
Com/Pant
Phonolite
Rhyolite
Frequency
10
Trachyte
.1
Zr /Ti
Rhyodacite/Dacite
TrachyAnd
Andesite
Bsn/Nph
.01
5
Alk-Bas
Andesite/Basalt
SubAlkaline Basalt
.001
.01
0
Frequency
15
enriched Low-Ti
.1
Nb/Y
1
10
(c)
10
5
0
42 44 46 48 50 52 54 56 58 60 62 64 66 68
SiO2 (wt. %)
Fig. 3. Histogram of the SiO2 (wt %) contents of (a) the high-Ti, (b) the depleted low-Ti Ujaraaluk unit and (c) the enriched low-Ti
Ujaraaluk unit. (d) Nb/Y vs Zr/Ti discrimination diagram for the Ujaraaluk unit. Symbols: open, high-Ti Ujaraaluk unit; filled black, enriched
low-Ti Ujaraaluk unit; filled grey, depleted low-Ti Ujaraaluk unit; vertical diamond, garnet-free Ujaraaluk unit; horizontal diamond,
garnet-bearing Ujaraaluk unit; triangle, cordierite^anthophyllite^biotite Ujaraaluk unit.
994
O’NEIL et al.
NUVVUAGITTUQ GREENSTONE BELT
20
100
(a)
Zr (ppm)
15
MgO (wt.%)
(e)
80
10
5
60
40
20
0
40
45
50
55
60
SiO2 (wt.%)
65
0
0.0
70
1.0
TiO2 (wt.%)
1.5
8
(b)
60
0.5
(f)
Nb (ppm)
Al2O3 / TiO2
6
40
20
0
2
0
10
MgO (wt.%)
0
0.0
20
12
0.5
1.0
TiO2 (wt.%)
1.5
8
(c)
10
(g)
6
8
La / Sm
CaO (wt.%)
4
6
4
4
2
2
0
0
5
10
MgO (wt.%)
15
0
0.0
20
25
0.5
1.0
400
(h)
300
20
V (ppm)
Al2O3 (wt.%)
(d)
200
15
10
0.0
1.5
TiO2 (wt.%)
0.5
1.0
TiO2 (wt.%)
1.5
100
0.0
0.5
1.0
TiO2 (wt.%)
Fig. 4. MgO and TiO2 vs selected major and trace element concentrations for the Ujaraaluk unit. Symbols as in Fig. 3.
995
1.5
JOURNAL OF PETROLOGY
VOLUME 52
distinct Al/Ti ratio, with the high-Ti unit characterized by
the lowest Al/Ti ratios, the depleted low-Ti unit having
the highest Al/Ti ratios, and the enriched low-Ti unit
having intermediate Al/Ti ratios (Fig. 4b).
The REE profiles for the three groups within the
Ujaraaluk unit are also distinctive. The high-Ti sub-group
has flat to mildly fractionated light REE (LREE) profiles,
whereas both low-Ti sub-groups have more strongly fractionated LREE profiles. The enriched low-Ti unit has,
however, higher LREE concentrations compared with the
depleted low-Ti unit and generally exhibits flat heavy
REE (HREE) profiles, whereas the depleted low-Ti unit
has Gd/Yb ratios 51 producing distinctive U-shaped REE
profiles (Fig. 5a, c and e).
Ultramafic sills
The Ujaraaluk unit is intruded by a series of ultramafic
sills. Two geochemically distinct groups of sills have been
identified previously and referred to as Type-1 and Type-2
sills, located stratigraphically outside and inside the BIF^
silica-formation horizon, respectively (O’Neil et al., 2007).
Cr increases with decreasing MgO within the Type-1 sills,
whereas Cr decreases with MgO in the Type-2 sills
(Fig. 6a, Table 2, Electronic Appendix 2). Type-1 sills are
relatively poor in Al and rich in Fe, whereas Type-2 sills
are relatively rich in Al and poorer in Fe (O’Neil et al.,
2007). Recent mapping and more detailed sampling of
these sills suggest that the Type-2 sills include a third chemically distinct sill type (Type-3) that is characterized by
Type-2 compatible element compositions, in terms of Cr
and Fe, but exhibits a distinct trend in a plot of TiO2 vs
Al2O3 (Fig. 6b). The three types of ultramafic sill exhibit
the same chemical features as the three groups in the
Ujaraaluk unit on a Zr vs TiO2 plot (Fig. 6c). As is the
case for the Ujaraaluk unit groups, the three types of
ultramafic sills also have distinct Al/Ti ratios (Fig. 6d).
Type-1 sills display low Al/Ti ratios, Type-2 sills have high
Al/Ti ratios, and Type-3 sills have intermediate Al/Ti
ratios.
DISCUSSION
Alteration, metamorphism, and mobility
of elements
The majority of the Nuvvuagittuq rocks from the
Ujaraaluk unit have relatively mafic compositions and
could be considered to range from basalt to basaltic andesite in terms of their SiO2 and MgO contents. They have,
however, undergone multiple phases of deformation and
high-grade metamorphism. Addressing the chemical
changes caused by the complex history of these rocks is
critical before using chemical signatures to constrain the
nature and petrogenesis of the protoliths of these gneisses.
The abundance of quartz ribboning within the Ujaraaluk
unit suggests that Si and perhaps other elements may have
NUMBER 5
MAY 2011
been mobile during metamorphism. The quartz ribbons
within the Ujaraaluk unit are, however, thin and discontinuous (Fig. 2a and b), suggesting that even if Si was
mobile during metamorphism, the source of Si was probably the Ujaraaluk unit rocks themselves. The presence of
quartz-rich lithologies such as the BIF and the
silica-formation may suggest that significant amounts of Si
were extracted from the basement and locally remobilized,
presumably by hydrothermal solutions. The most immobile
trace elements, such as Zr, Nb, Y, Ta, V, Hf, Th, and REE,
define well-correlated trends within all three chemical
groups of the Ujaraaluk unit that are consistent with their
relatively incompatible behaviour during magmatic processes, suggesting that they have been little affected by selective secondary mobilization. In terms of the more
mobile major elements, Ca displays an anomalously large
variation and reaches extremely low values in the cordierite^anthophyllite^biotite assemblages (1wt % CaO;
Fig. 4c, Table 1, Electronic Appendix 1). These very low
CaO contents are associated with low Na2O and Sr contents, and are unlikely to represent the composition of an
igneous precursor, indicating Ca mobility during
post-magmatic processes. The samples with low Ca contents do not have other major or trace element signatures
suggestive of the removal of fluid or melt during metamorphism, so we pursue below the possibility that these
chemical signatures are the result of alteration of the igneous precursors of these rocks prior to metamorphism. The
nature of this alteration can be evaluated by using indices
of chemical weathering. The Chemical Index of
Alteration (CIA; Nesbitt & Young, 1982) has been widely
used to address the question of weathering of rocks. The
garnet-bearing and the cordierite^anthophyllite rocks
from the Ujaraaluk unit have significantly higher CIA
values (68^94) than the garnet-free high-Ti and low-Ti
unit (55^70). All CIA values are higher than those typical
of fresh basalts, which range between 30 and 45 (Nesbitt
& Young, 1982). There are also strong correlations between
CIA values and oxides and elements such as CaO
(Fig. 7a), Na2O, Rb and Sr, which occur primarily in
plagioclase in basaltic and andesitic rocks. The high CIA
values would thus be consistent with extensive alteration
of the plagioclase in the igneous protolith of the
Nuvvuagittuq rocks from the Ujaraaluk unit. The cordierite^anthophyllite^biotite assemblages seen locally in both
the high and low-Ti groups are characteristically very low
in Ca and high in Mg (Fig. 4c), but are indistinguishable
from the cummingtonite-bearing high-Ti or low-Ti groups
in which they occur in terms of immobile trace elements.
This suggests that their anomalously low Ca and high Mg
contents represent post-magmatic alteration. Cordierite^
orthoamphibole rocks have been shown to reflect diverse
geological processes, ranging from the metamorphism of
paleosols (Gable & Sims, 1969; Young, 1973) to the restites
996
O’NEIL et al.
NUVVUAGITTUQ GREENSTONE BELT
100
(a)
100
Sample/Chondrite
(b)
Sample/Primitive Mantle
high-Ti
10
10
1
1
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Nb La Pr Zr Sm Ti Tb Y Er Yb
Th Ta Ce Nd Hf Eu Gd Dy Ho Tm Lu
100
(c)
(d)
Sample/Primitive Mantle
depleted low-Ti
Sample/Chondrite
100
10
10
1
1
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Nb La Pr Zr Sm Ti Tb Y Er Yb
Th Ta Ce Nd Hf Eu Gd Dy Ho Tm Lu
100
enriched low-Ti
(f)
Sample/Primitive Mantle
Sample/Chondrite
(e)
100
10
10
1
1
Sample/Chondrite
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
(g)
Nb La Pr Zr Sm Ti Tb Y Er Yb
Th Ta Ce Nd Hf Eu Gd Dy Ho Tm Lu
Crd-Anth
100
10
1
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Fig. 5. Chondrite-normalized REE profiles and primitive mantle-normalized trace profiles for the Ujaraaluk unit. (a, b) High-Ti Ujaraaluk
unit; (c, d) depleted low-Ti Ujaraaluk unit; (e, f) enriched low-Ti Ujaraaluk unit; (g) cordierite^anthophyllite Ujaraaluk unit. Symbols as in
Fig. 3. Values for chondrite and primitive mantle abundances are from Sun & McDonough (1989).
997
JOURNAL OF PETROLOGY
7000
VOLUME 52
MAY 2011
25
(a)
(b)
20
6000
Gabbro Tops
Al2O3 (wt.%)
Cr (ppm)
NUMBER 5
5000
4000
3000
2000
15
10
5
1000
0
0
10
20
MgO (wt.%)
30
0
0.0
40
0.5
TiO2 (wt.%)
1.0
1.5
100
(c)
Al2O3 / TiO2
Zr (ppm)
(d)
60
80
60
40
40
20
20
Gabbro Tops
0
0.0
0.5
1.0
TiO2 (wt.%)
0
1.5
0
10
20
30
40
MgO (wt.%)
Fig. 6. MgO and TiO2 vs selected major and trace element concentrations for the ultramafic sills and Ujaraaluk unit. Symbols as in Fig. 3 for
diamonds and triangles; open circle, Type-1 sill; filled grey circle, Type-2 sill; filled black circle, Type-3 sill. Grey field, gabbro tops of Type-1, -2
and -3 sills. Samples from the sills include ultramafic rocks and pyroxenites.
of partial melting of metasediments (Grant, 1968; Hoffer &
Grant, 1980). The most commonly proposed model, however, suggests that cordierite^anthophyllite rocks represent
hydrothermally altered mafic volcanic rocks (Vallance,
1967; Smith et al. 1992; Peck & Smith, 2005). The localized
occurrence of the cordierite^anthophyllite assemblages in
both the high-Ti and low-Ti units in the Nuvvuagittuq
greenstone belt is more consistent with localized hydrothermal alteration rather than general weathering or regional metamorphism. The high Mg content of the
cordierite^anthophyllite Ujaraaluk rocks would be consistent with seawater being the hydrothermal agent (Spear,
1993). The REE profiles are also consistent with hydrothermal alteration, exhibiting large negative Eu anomalies
consistent with the alteration of plagioclase (Fig. 5g)
Although the cordierite^anthophyllite rocks are characterized by the lowest Ca contents, all varieties of rocks
from the Ujaraaluk unit have lower Ca compared with
the gabbro sills that intrude them. The abundance of cummingtonite in the Ujaraaluk unit, as opposed to
hornblende in the gabbro sills, is interpreted to reflect this
Ca deficit (O’Neil et al., 2007), as cummingtonite is a
common metamorphic phase found in Ca-poor
mafic rocks. The Ujaraaluk unit displays a well-developed
layering with alternation of
biotite-rich and
cummingtonite-rich layers as well as garnet-rich and
garnet-poor horizons (Fig. 2b). These features are reminiscent of sedimentary rocks, and the Ujaraaluk rocks were
originally interpreted to be metasediments (Simard et al.,
2003). However, the Ujaraaluk rocks are mafic in composition and are unlikely to represent mature sediments. A volcanic origin of the Ujaraaluk unit is also supported by the
presence of rare pillow lavas (Fig. 2c). We interpret the
well-developed layering and mafic bulk compositions of
the Ujaraaluk unit to reflect a mafic pyroclastic protolith,
as opposed to massive mafic volcanic rocks. Layering is
also a common feature in submarine volcanic deposits
(Fiske & Matsuda, 1964; Houghton & Landis, 1989;
Cashman & Fiske, 1991; and references therein). However,
metamorphism and deformation of the Ujaraaluk unit
998
O’NEIL et al.
NUVVUAGITTUQ GREENSTONE BELT
Table 2: Major (wt %) and trace (ppm) element data for representative samples of the Nuvvuagittuq sills
Ultramafic sills
Type-1 sills
Sample:
PC-29
Type-2 sills
PC-33
PC-45
PC-47
PC-125
PC-3
Type-3 sills
PC-92
PC-94
PC-166
PC-167
PC-76
PC-138
PC-142
PC-213
PC-216
SiO2
40·52
49·67
45·69
42·87
39·82
37·56
40·54
40·33
42·93
50·48
40·01
42·38
44·97
46·34
41·01
TiO2
0·27
0·66
0·80
0·67
0·23
0·10
0·08
0·11
0·24
0·11
0·20
0·24
0·47
0·27
0·15
Al2O3
2·20
5·50
6·18
5·62
1·77
4·72
3·83
5·16
11·26
5·49
5·07
6·27
11·72
7·30
4·11
MgO
30·54
17·59
18·82
25·31
35·34
29·28
32·75
31·42
21·75
16·86
30·83
28·34
22·24
27·10
31·99
FeO
12·02
12·02
14·56
11·65
10·11
8·42
8·13
8·03
9·91
12·22
8·31
8·63
9·96
10·37
7·93
MnO
0·20
0·30
0·21
0·20
0·18
0·21
0·14
0·15
0·20
0·38
0·14
0·15
0·13
0·14
0·14
CaO
2·73
11·40
9·51
5·67
0·52
5·51
2·12
3·06
7·04
11·22
3·59
4·86
5·55
2·52
3·04
Na2O
0·08
0·47
0·43
0·29
0·03
0·20
0·11
0·15
0·34
0·17
0·12
0·14
0·31
0·11
0·14
K2O
0·02
0·13
0·10
0·07
0·04
0·04
0·03
0·04
0·09
0·05
0·04
0·03
0·11
0·04
0·05
P2O5
0·02
0·03
0·04
0·03
0·01
0·01
0·02
0·01
0·03
0·02
0·02
0·03
0·05
0·03
0·02
LOI
9·89
1·45
2·08
6·53
11·28
12·43
10·67
9·85
4·98
1·65
9·96
8·05
3·49
4·46
9·95
Rb
1·6
1·2
1·8
1·3
3·0
1·1
1·0
0·6
1·5
1·1
1·2
1·1
2·4
0·9
0·4
Sr
14·2
15·3
12·8
10·7
3·2
14·3
8·1
7·3
4·8
3·3
20·1
16·4
35·9
5·4
20·0
Zr
13·7
32·4
43·4
34·0
10·0
4·5
3·8
4·6
11·9
4·7
21·5
25·6
51·1
28·8
12·7
Nb
0·7
1·7
1·9
2·1
0·6
0·3
0·3
0·1
0·8
2·5
1·4
1·4
2·5
2·0
0·8
Y
3·2
9·0
13·9
8·1
1·8
2·7
2·4
2·7
7·6
5·4
6·6
9·7
16·6
7·7
5·2
Ni
1775
206
408
1489
2326
1806
1610
1789
570
865
1941
1620
651
1605
2259
Cr
1689
2039
2035
2482
1274
3860
4554
5643
1748
2563
4431
4178
2113
4321
6933
67
181
190
147
44
73
57
83
162
85
74
93
168
91
71
V
Samples include ultramafic rocks and pyroxenites. The complete dataset is provided in Electronic Appendix 2.
has obliterated any primary features that could permit a
more definitive identification of the process of deposition.
The low Ca contents of the Ujaraaluk unit may simply reflect the fact that their pyroclastic protolith was more susceptible to weathering and alteration than the massive
sills intruded into these deposits. This interpretation is supported by the lower CIA values of the gabbro sills
(Fig. 7a). The CIA, however, assumes that CaO, Al2O3,
Na2O and K2O are present only in feldspars. Thus, the
variations in CIA values may not only be due to weathering or alteration but may be caused by primary chemical
variations in the unweathered protolith. CIA values for
the Ujaraaluk rocks increase more significantly for samples with CaO contents 54 wt % (Fig. 7a). CaO contents
for the Ujaraaluk unit are generally lower than the CaO
contents for the gabbro sills, even for the Ujaraaluk rocks
exhibiting the lowest CIA values. This suggests that the
low Ca contents of the Ujaraaluk unit are not only due to
weathering, but may also represent a primary chemical
feature of the Faux-amphibolites being generally Ca-poor
compared with the gabbro sills.
Ohta & Arai (2007) developed an empirical statistical
index of chemical weathering that is independent of the
composition of the unweathered protolith. In an ‘MFW’
plot (Fig. 7b), unaltered rocks define an igneous trend ranging from mafic (M) to felsic compositions (F). The
degree of weathering is evaluated by the W value, a measure of the deviation from the igneous trend towards the W
apex. A large number of Ujaraaluk unit samples plot
close to the igneous trend, suggesting that most Ujaraaluk
rocks are relatively unweathered or only mildly weathered.
The high-Ti group is displaced from the igneous trend towards the W apex, and could be consistent with mildly
weathered basaltic rocks. Garnet-free, low-Ti samples plot
closer to the igneous trend, and range in composition
from basalt to andesite consistent with relatively unweathered mafic protoliths. In contrast, garnet-bearing low-Ti
samples are strongly displaced towards the W apex, and
could be consistent with highly weathered, evolved andesite compositions. The gabbro sills and the tonalite fall on
the igneous trend, suggesting that the weathering mechanism responsible for the alteration of the Ujaraaluk unit
999
JOURNAL OF PETROLOGY
100
VOLUME 52
NUMBER 5
M
(a)
(b)
High-Ca
boninites
80
CIA 60
Low-Ca
boninites
Gabbro sills
6
8
10
12
CaO (wt.%)
M
14
Basalt
Nuvvuagittuq
gabbros
d
4
tren
2
Komatiite
ous
0
Igne
40
20
MAY 2011
Andesite
W
ea
nd
Nuvvuagittuq
tonalites
tre
th
er
ing
Ign
eo
us
Dacite
Granite
F
W
100
100
80
(c)
(d)
80
60
60
W
W
40
40
20
20
0
0
100
10
20
30
40
Al2O3 / TiO2
50
60
0
70
0
1
2
3
4
5
6
7
8
La / Sm
100
(e)
(f)
80
80
60
60
W
W
40
40
20
20
0
0.0
0.5
1.0
Eu/Eu*
1.5
2.0
0
0
5
10
15
20
MgO(wt.%)
Fig. 7. Alteration indices diagrams for the Ujaraaluk unit. (a) CaO (wt %) vs Chemical Index of Alteration (CIA) ¼ Al2O3/
(Al2O3 þ CaO þ Na2O þ K2O) 100. (b) MFW diagram. Values plotted in the MFW diagram are: exp(M), exp(F), exp(W), re-totaled to
100. Values for M, F and W calculation are normalized to 100 wt %. M ¼ ^0·395 ln(SiO2) þ 0·206 ln(TiO2) ^ 0·316 ln(Al2O3) þ 0·160
ln(Fe2O3) þ 0·246 ln(MgO) þ 0·368 ln(CaO) þ 0·073 ln(Na2O) ^ 0·342 ln(K2O) þ 2·266. F ¼ 0·191 ln(SiO2) ^ 0·397 ln(TiO2) þ 0·020
ln(Al2O3) ^ 0·375 ln(Fe2O3) ^ 0·243 ln(MgO) þ 0·079 ln(CaO) þ 0·392 ln(Na2O) þ 0·333 ln(K2O) ^ 0·892. W ¼ 0·203 ln(SiO2) þ 0·191
ln(TiO2) þ 0·296 ln(Al2O3) þ 0·215 ln(Fe2O3) ^ 0·002 ln(MgO) ^ 0·448 ln(CaO) ^ 0·464 ln(Na2O) þ 0·008 ln(K2O) ^1·374 (Ohta & Arai,
2007). (c) to (f) Selected trace and major elements vs Weathering Index values (W). Symbols as in Fig. 3. Data for boninites (grey fields) are
from Crawford et al. (1989).
1000
O’NEIL et al.
NUVVUAGITTUQ GREENSTONE BELT
did not affect these lithologies. For comparison, fields for
low-Ca and high-Ca boninites are also plotted in Fig. 7b.
Despite the large variation in CaO contents (3·55^
11·11wt %), both low-Ca and high-Ca boninites fall
within the unweathered igneous trend, suggesting that the
W index discriminates between low-Ca primary compositions and Ca depletion owing to weathering. The fact
that most Ujaraaluk unit samples lie close to the igneous
trend on the MFW plot, despite their low Ca contents, suggests that their igneous protolith was Ca-poor relative to
the gabbro sills. The W index appears to represent a good
indicator of weathering for the Nuvvuagittuq rocks. It can
be used to discriminate the chemical elements representing
original igneous chemical features from elements remobilized during post-magmatic alteration. The Al2O3/TiO2
ratios remain relatively constant within the three
Ujaraaluk unit geochemical groups independent of their
degree of weathering (Fig. 7c), suggesting that the distinct
Al2O3/TiO2 ratios represent a primary igneous feature of
the Ujaraaluk unit. The lack of correlation between the W
index and La/Sm ratios (Fig. 7d) is also consistent with
the distinct REE profiles of the various Ujaraaluk unit
groups being magmatic features rather than the products
of post-magmatic weathering. Figure 7e demonstrates that
although the Ujaraaluk rocks display a range of Eu/Eu*
ratios, from slight negative to slight positive anomalies, all
Ujaraaluk unit samples exhibiting high W values have
negative Eu anomalies (Eu/Eu*41), consistent with
alteration-induced loss of plagioclase. However, the weathering mechanism for the cordierite^anthophyllite
Ujaraaluk rocks appears to be associated with an increase
in MgO (Fig. 7f), which is more consistent with alteration
by hydrothermal seawater circulation.
Nature of the volcanic protolith
Regardless of the relative contributions of early alteration
versus later metasomatic processes to the chemistry of the
Ujaraaluk unit, the lack of correlation between Ca contents or the W index and relatively immobile HFSE and
REE suggests that the relative proportions of these immobile elements probably reflect the primary chemical characteristics of the Ujaraaluk unit protolith. The high-Ti,
the depleted low-Ti and the enriched low-Ti units represent
three chemically distinct magmatic groups that are directly analogous to the Type-1, Type-2 and Type-3 ultramafic
sills, suggesting that the Ujaraaluk unit and the sills are
cogenetic (Fig. 6b and d) and that the alteration experienced by the Ujaraaluk unit has not obliterated these igneous relationships. Estimated parental magmas for Type-1
and Type-2 ultramafic sills suggest that they were fractionated from distinct magmatic sources (O’Neil et al., 2007).
Both Type-1 and Type-2 sills have parental liquids that are
komatiitic in composition, but with distinct Al and Fe contents. The estimated parental magma for Type-1 sills is
Fe-rich and Al-poor and is similar to aluminum-depleted
komatiites, whereas the parental magma of the Type-2 sill
resembles aluminum-undepleted komatiites, with higher
Al and lower Fe contents.
Although the Nuvvuagittuq greenstone belt has been
highly deformed, the three chemical groups of the
Ujaraaluk unit define a consistent chemical stratigraphy,
with the high-Ti unit and Type-1 sills occurring in the
outer limbs of the synform and the low-Ti unit as well as
Type-2 and Type-3 sills occurring in the interior of the synform, with the BIF^silica-formation horizon between
Type-1 and Type-2 sills (Fig. 8). Furthermore, the enriched
low-Ti group is located in the core of the synform, whereas
the depleted low-Ti group is found in the limbs, adjacent
and interior to the BIF^silica-formation horizon. The presence of gabbroic tops within some sills in the western limb
can be used as tentative top indicators. Also, based on
rare facing directions, the east limb of the tight synform
faces west, in which case the synform would be a syncline.
However, local overturning during intense shearing also
could have occurred. Despite the paucity of definitive top
indicators in the Nuvvuagittuq volcanic succession, the
available polarity evidence indicates that the tight synform
is a syncline, in which case the high-Ti unit would represent the base of the volcanic succession, with the low-Ti
unit succeeding it in time. Although there is a possibility
that the high-Ti and the low-Ti groups are two distinct terranes that have been tectonically juxtaposed along the
BIF horizon, we consider it more likely that they represent
a continuous volcanic succession.
The high-Ti Ujaraaluk unit displays geochemical features similar to those of tholeiitic volcanic suites with flat
REE profiles and relatively low Al/Ti ratios, whereas the
low-Ti Ujaraaluk unit is geochemically similar to
calc-alkaline volcanic suites with LREE-enriched profiles
and higher Al/Ti ratios. The different fractionation trends
in terms of major elements for the high-Ti and low-Ti
Ujaraaluk units could be explained by crystal fractionation at differing water pressures. The low-Ti unit is characterized by increasing Al contents in the range from
basalt to andesite with Al2O3 contents falling only within
the most evolved dacitic compositions (Fig. 9a). It also displays low Ti concentrations at all SiO2 contents. The increase in Al2O3 content from basaltic to andesitic
compositions is consistent with crystal fractionation at elevated water pressure, where plagioclase fractionation is
suppressed relative to mafic phases. This feature is characteristic of calc-alkaline volcanic suites, which exhibit high
Al2O3 contents at andesitic compositions. In contrast, the
high-Ti unit is characterized by Al2O3 contents that
remain relatively low even for the most evolved compositions, which is more consistent with crystal fractionation
under low-pressure dry conditions (Fig. 9a). Experiments
suggest that these different trends probably reflect the
effect of water on the onset of crystallization of plagioclase
1001
JOURNAL OF PETROLOGY
VOLUME 52
NUMBER 5
MAY 2011
Fig. 8. Close-up of the geological map (see Fig. 1) showing the location of the high-Ti Ujaraaluk unit, the depleted low-Ti Ujaraaluk unit and
the enriched low-Ti Ujaraaluk unit. Symbols: open diamonds, high-Ti Ujaraaluk unit; grey diamonds, depleted low-Ti Ujaraaluk unit; black
diamonds, enriched low-Ti Ujaraaluk unit. Location of the map area is shown in Fig. 1. Coordinates are in UTM NAD27.
(Spulber & Rutherford, 1983; Sisson & Grove, 1993).
According to this explanation, the low Al2O3 contents of
the early high-Ti Ujaraaluk unit reflect the early crystal
fractionation of feldspar at low pressures, under dry conditions. In contrast, the higher Al2O3 contents of the later
low-Ti Ujaraaluk unit reflect the delayed crystallization of
plagioclase at elevated water pressures. This interpretation
is also consistent with the experimental differences in the
behaviour of Ti during fractionation under dry versus elevated water pressures. The high-Ti unit would reflect the
1002
O’NEIL et al.
25
(a)
20
Al2 O3
NUVVUAGITTUQ GREENSTONE BELT
2 kb H2O
Calc-Alkaline
15
Tholeiitic
dry
10
5
30
40
50
60
70
80
SiO2
(b)
dry
TiO 2
3
2
2 kb H2O
Tholeiitic
1
Calc-Alkaline
0
30
40
50
60
70
80
SiO2
Fig. 9. SiO2 vs (a) Al2O3, (b) TiO2 for the Ujaraaluk unit. Symbols as in Fig. 3. Fractionation trends are based on experimental data from
Spulber & Rutherford (1983) and Sisson & Grove (1993). Filled circles, starting experimental compositions; continuous lines, fractionation
trend at 1atm dry from experimental data; dashed lines, fractionation trend at 2 kbar H2O from experimental data; dark grey field, tholeiites;
light grey area, calc-alkaline lavas. Sources: Sigvaldson (1974); Hickey et al. (1986); Kay et al. (1987); Munoz & Stern (1989).
late crystallization of Fe^Ti oxides during dry fractionation, whereas the low-Ti unit would reflect the early crystallization of Fe^Ti oxides under wet conditions (Fig. 9b).
This interpretation is also supported by the chemical
trends displayed by the cogenetic ultramafic sills. As
shown in Fig. 6a, Cr decreases with MgO content in
Type-2 and Type-3 sills, which are cogenetic with the
low-Ti Ujaraaluk unit. This is consistent with the early
crystallization of Fe^Ti oxides under oxidizing wet conditions, causing the depletion in Cr with fractionation.
However, the Cr contents in Type-1 sills initially increase
with decreasing MgO content but start decreasing later
1003
JOURNAL OF PETROLOGY
VOLUME 52
during crystallization. This suggests fractionation under
dry reducing conditions where Cr behaves as an incompatible element in the absence of Fe^Ti oxide crystallization.
Tectonic setting
Attributing tectonic setting to Archean rocks based on
modern-day analogues must be done with caution because
the nature of Archean, and especially Hadean, tectonic
processes is poorly constrained. The high-Ti group displays
flat REE profiles as opposed to the enriched LREE profiles
of the low-Ti group (Fig. 5). These two different REE patterns and the two distinct trends on a plot of Zr vs TiO2
(Fig. 4e) may indicate derivation from two geochemically
distinct mantle sources for the high-Ti and the low-Ti
Faux-amphibolites. This is also supported by the distinct
composition for the parental magmas of their respective
cogenetic ultramafic sills. The low Al2O3/TiO2 ratios,
higher FeO, and flat to slightly enriched REE profiles of
the high-Ti unit are consistent with derivation from an
undepleted primitive mantle source, and crystal fractionation under dry, low-pressure conditions. The high-Ti unit
exhibits many chemical similarities to present-day
mid-ocean ridge basalts (MORB). In contrast, the low
TiO2 contents, high Al2O3/TiO2 ratios, and LREEenriched profiles of the depleted low-Ti group are consistent with a more refractory mantle source that has experienced secondary enrichment. The low Ti contents and the
positive slope of the HREE suggest that the source was
depleted in incompatible elements, whereas the
LREE-enriched profiles suggest re-enrichment of the
source by metasomatic fluids. These geochemical features
are similar to those of modern-day subduction-related volcanism. In fact, despite somewhat higher Al2O3 and
HREE contents, the depleted low-Ti Ujaraaluk rocks are
similar to modern-day boninites, which are interpreted to
be produced by the melting of metasomatized refractory
mantle in convergent margin environments (Crawford
et al., 1989; Tatsumi & Maruyama, 1989; Falloon &
Danyushevsky, 2000). Alternative models have, however,
been proposed for low-Ti volcanic rocks in Archean greenstone belts. For example, Smithies et al. (2004) proposed
that Whitney-Type volcanic rocks, with chemical characteristics similar to those of the depleted low-Ti Ujaraaluk
rocks, are second-stage melts related to a mantle plume,
rather than subduction, because of their close association
with komatiitic magmas and ultramafic rocks. Although
the Nuvvuagittuq Ujaraaluk unit is also associated with
ultramafic rocks derived from komatiitic parental liquids
(O’Neil et al., 2007), there is no evidence suggesting that a
mantle plume was involved in their petrogenesis.
HFSE have been widely used to discriminate basaltic
rocks and ratios such as Nb/Th often serve as indicators
of their tectonic setting (e.g. Pearce & Cann, 1971, 1973;
Pearce & Peate, 1995; Condie, 2005; Pearce, 2008; Furnes
et al., 2009). Th and Nb have similar geochemical
NUMBER 5
MAY 2011
behaviour but Th is more mobile in a subduction setting
whereas Nb is preferentially retained in the subducting
slab. Thus, subduction-related basaltic rocks generally
have low Nb/Th ratios. These proxies can be applied to
the Nuvvuagittuq Ujaraaluk unit to attempt to determine
the tectonic setting into which it was emplaced. The Nb/
Th ratios of the depleted low-Ti unit are relatively low,
except for the garnet-bearing samples, whereas the enriched low-Ti unit displays low Nb/Th ratios at any Zr
content (Fig. 10a). There is also some overlap between the
Ujaraaluk rocks displaying boninitic affinities and the 3·8
Ga Isua Garbenschiefer samples. The Garbenschiefer displays some chemical similarities to modern-day boninites
(Polat et al., 2002) and has been interpreted by some workers to have formed in a convergent margin setting. Pearce
(2008) has shown that modern MORB and ocean island
basalt (OIB) suites define an array that is distinct from
that of subduction-related basalts on a Th/Yb vs Nb/Yb
diagram. The Ujaraaluk rocks define an oblique trend
and are displaced to the high-Th^low-Nb side of the
MORB^OIB array (Fig. 10b), suggesting that Nb and Th
are decoupled, as expected in a convergent margin setting.
In fact, both sub-groups of the low-Ti Ujaraaluk unit compositionally overlap Mariana arc samples used by Pearce
et al. (2005) and Pearce (2008) to characterize
subduction-related basaltic rocks. The Th^Nb composition
of the low-Ti unit thus is similar to that observed in a
modern-day convergent margin setting and may suggest a
similar tectonic setting for the Nuvvuagittuq rocks. This
possibility is also supported by the negative Nb and Ti
anomalies in the primitive mantle normalized trace element patterns of the low-Ti Ujaraaluk unit (Fig. 5d and f),
in contrast to the high-Ti unit, which does not exhibit Ti
anomalies and has only a slight negative Nb anomaly
(Fig. 5b). Interestingly, the Nuvvuagittuq high-Ti
Ujaraaluk unit overlaps with the Mariana arc samples.
Regardless of their respective tectonic setting, the major
and trace element geochemistry of the three Ujaraaluk
groups suggests that they represent three distinct magmatic suites. The depleted and the enriched low-Ti units share
many geochemical characteristics and define a common
trend in a plot of Zr vs TiO2, consistent with derivation
from a similar mantle source. However, the enriched
low-Ti rocks do not display U-shaped REE profiles and
have higher Gd/Yb ratios than the depleted low-Ti rocks.
Moreover, the enriched low-Ti group has higher concentrations of incompatible elements and lower Al/Ti ratios
compared with the depleted low-Ti group. This may suggest that the mantle source of the depleted low-Ti rocks
had experienced more extensive depletion compared with
the mantle source of the enriched low-Ti rocks, but that
both experienced a similar secondary enrichment process
responsible for their similar LREE-enriched REE profiles.
Different degrees of partial melting of similar geochemical
1004
O’NEIL et al.
NUVVUAGITTUQ GREENSTONE BELT
15
(a)
Isua
Garbenschiefer
Nb/Th
10
5
0
50
Zr
0
100
(b)
ra
y
10
M
Mariana
Arc Rift
O
RB
-O
IB
ar
Mariana Arc
OIB
rb Isu
en a
sc
hie
f
Th/Yb
er
1
Ga
0.1
E-MORB
N-MORB
0.01
0.1
1
Nb/Yb
10
100
Fig. 10. (a) Zr vs Nb/Th and (b) Nb/Yb vs Th/Yb (after Pearce, 2008) for the Ujaraaluk unit. Symbols as in Fig. 3. Isua Garbenschiefer data
are from Polat et al. (2002), Polat & Hofmann (2003) and Furnes et al. (2009). Mariana arc data are from Pearce et al. (2005) and Pearce (2008).
1005
JOURNAL OF PETROLOGY
fractionation at
elevated water
pressures
VOLUME 52
hydrothermal
alteration
Type-2 sill
NUMBER 5
MAY 2011
"calc-alkaline" affinities
Enriched low-Ti
Ujaraaluk unit
Type-3 sill
Depleted low-Ti
Ujaraaluk unit
Type-1 sill
High-Ti
Ujaraaluk unit
Refractory mantle
re-enrichment
chemical sediments
Type-1 sill
High-Ti
Ujaraaluk unit
Refractory mantle
fractionation under
low pressure
dry conditions
hydrothermal
alteration
"tholeiitic" affinities
Type-1 sill
High-Ti
Ujaraaluk unit
Primitive mantle
Fig. 11. Schematic section of the Nuvvuagittuq greenstone belt. Temporal evolution is from bottom to top. At the base of the sequence, emplacement of the high-Ti Ujaraaluk unit with tholeiitic affinities and fractionation under dry conditions along with cogenetic sills. Above the
high-Ti Ujaraaluk unit, deposition of chemical sediments followed by the emplacement of the low-Ti Ujarraluk unit (depleted low-Ti and enriched low-Ti units) and cogenetic sills. The low-Ti Ujarraluk units are consistent with the melting of a re-enriched refractory mantle source
and fractionation at elevated water pressure. Hydrothermal alteration of the Ujarraluk unit induces Ca depletion and Mg enrichment to form
the cordierite^anthophyllite rocks.
reservoirs, however, may also in part explain the geochemical differences between the depleted and enriched low-Ti
units. The more enriched low-Ti rocks may reflect a smaller degree of partial melting than that responsible for the
depleted low-Ti Ujaraaluk unit.
The succession from ‘tholeiitic to ‘calc-alkaline’
magmatism in the Nuvvuagittuq greenstone belt is in fact
typical of the volcanic successions of many younger
Archean greenstone belts (Jolly, 1975, 1980; Condie,
1981; Jolly & Hallberg, 1990; Wyman, 1999; Ayers et al.,
2002; Mtoro et al., 2009; Polat, 2009), suggesting that the
geological processes responsible for the formation and evolution of Archean greenstone belts were already active at
3·8 Ga and perhaps as early as 4·3 Ga. The geochemical
characteristics of the Ujaraaluk unit suggest that the
Nuvvuagittuq greenstone belt represents an Eoarchean^
Hadean volcanic succession derived from at least two
mantle sources, one relatively undepleted and the other
relatively refractory but re-enriched by metasomatic fluids
(Fig. 11).
1006
O’NEIL et al.
NUVVUAGITTUQ GREENSTONE BELT
CONC LUSIONS
AC K N O W L E D G E M E N T S
With a possible age as old as 4·3 Ga, the Ujaraaluk unit
of the Nuvvuagittuq greenstone belt may represent the
only remnant of Hadean crust yet found on the surface of
the Earth. Even if they are closer in age to the 3·8 Ga
felsic rocks, the Nuvvuagittuq rocks are among the oldest
supracrustal rocks on Earth. The Nuvvuagittuq
Ujaraaluk unit comprises a volcanic succession of largely
mafic composition that can be subdivided into three distinct chemical groups that are cogenetic with three geochemical types of ultramafic sill. The earlier high-Ti unit
displays ‘tholeiitic’ affinities with flat REE profiles and is
consistent
with
derivation
from
a
relatively
undepleted primitive mantle source followed by crystal
fractionation under dry, low-pressure conditions. In
contrast, the later low-Ti unit has ‘calc-alkaline’ affinities
consistent with differentiation at elevated water pressures.
The depleted low-Ti Ujaraaluk sub-group displays
U-shaped REE profiles and is consistent with the
melting of a depleted mantle source that had previously
experienced
secondary
incompatible
element
enrichment. The enriched low-Ti Ujaraaluk unit has fractionated LREE profiles, also consistent with secondary
enrichment, but with flat HREE profiles perhaps suggesting a lesser degree of depletion of their mantle source.
The geochemistry of the mafic Ujaraaluk unit and occurrence of low-CaO cordierite^anthophyllite assemblages
is consistent with the protolith of the Ujaraaluk unit
being hydrothermally altered basaltic to andesitic volcanic rocks. The Nuvvuagittuq Ujaraaluk unit may represent the only remnant of Hadean crust on Earth. So far,
the only confirmed Hadean samples are the Jack Hills
detrital zircons. The relatively unradiogenic Hf
(Amelin et al., 1999; Harrison et al., 2005, 2008;
Blichert-Toft & Albare'de, 2008; Kemp et al., 2010) and
heavy oxygen isotopic compositions (Mojzsis et al., 2001;
Wilde et al., 2001; Valley et al., 2002; Cavosie et al., 2005;
Harrison et al., 2008) of the Jack Hills zircons suggest that
their host rock was derived from the melting of a Hadean
mafic protolith that had interacted with the
proto-hydrosphere (Blichert-Toft & Albare'de, 2008; Kemp
et al., 2010). The Nuvvuagittuq Ujaraaluk unit may represent such a protocrust.
The Nuvvuagittuq greenstone belt preserves a volcanic
succession with tholeiites at the base of the sequence
overlain by BIF and more evolved calc-alkaline suites towards the top of the succession. Regardless of the exact tectonic setting, this transition from ‘tholeiitic’ to ‘calcalkaline’ magmatism appears to be typical of volcanic successions documented in many younger Archean greenstone
belts, suggesting that the geological processes responsible
for the formation and evolution of Archean greenstone
belts were active in the Eoarchean and perhaps as early as
4·3 Ga.
We would like to acknowledge the people of the Pituvik
Landholding Corporation, especially Mike Carroll,
Johnny Mina, Minnie Palliser, Simeonie Elyasiapik,
Arthur Elyasiapik, Aliva Epoo and Simeonie, for their
critical logistical support in the field. We thank Minnie
Nowkawalk and Noah Echalook for their precious help in
2009 field camp. We are also grateful to the people and municipality of Inukjuak for their hospitality. We would particularly like to thank Andrew Hynes for his contribution
to our understanding the complex structure of the highly
deformed Nuvvuagittuq greenstone belt. We also acknowledge Glenna Keating and Bill Minarik, who performed
the whole-rock XRF analyses. Finally, we thank the three
anonymous reviewers and the editor Dominique Weis for
their useful comments, which greatly improved the present
paper.
FU N DI NG
This work was supported by the National Science and
Engineering Research Council of Canada (Discovery
Grants RGPIN 7977^00 to D.F.), the National Science
Foundation (NSF-EAR-0910442 to R.W.C.) and the
Carnegie Canada Foundation.
S U P P L E M E N TA RY DATA
Supplementary data for this paper are available at Journal
of Petrology online.
R EF ER ENC ES
Amelin, Y., Lee, D., 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.
Amelin, Y., Krot, A. N., Hutcheon, I. D. & Ulyanov, A. A. (2002).
Lead isotopic ages of chondrules and calcium^aluminum-rich inclusions. Science 297, 1678^1683.
Ayer, J., Amelin, Y., Corfu, F., Kamo, S., Ketchum, J., Kwok, K. &
Trowell, N. (2002). Evolution of the southern Abitibi greenstone
belt based on U^Pb geochronology; autochthonous volcanic construction followed by plutonism, regional deformation and sedimentation. Precambrian Research 115, 63^95.
Blichert-Toft, J. & Albare'de, F. (2008). Hafnium isotopes in Jack Hills
zircons and the formation of the Hadean crust. Earth and Planetary
Science Letters 265, 686^702.
Boily, M., Leclair, A., Maurice, C., Be¤dard, J. H. & David, J. (2009).
Paleo- to Mesoarchean basement recycling and terrane definition
in the Northeastern Superior Province, Que¤bec, Canada.
Precambrian Research 168, 23^44.
Bowring, S. A. & Williams, I. S. (1999). Priscoan (4·00^4·03 Ga)
orthogneisses from northwestern Canada. Contributions to
Mineralogy and Petrology 134, 3^16.
Cashman, K. V. & Fiske, R. S. (1991). Fallout of pyroclastic debris from
submarine volcanic eruptions. Science 253, 275^280.
Cates, N. L. & Mojzsis, S. J. (2007). Pre-3750 Ma supracrustal rocks
from the Nuvvuagittuq supracrustal belt, northern Quebec. Earth
and Planetary Science Letters 255, 9^21.
1007
JOURNAL OF PETROLOGY
VOLUME 52
Cates, N. L. & Mojzsis, S. J. (2009). Metamorphic zircon, trace elements and Neoarchean metamorphism in the ca. 3·75 Ga
Nuvvuagittuq supracrustal belt, Quebec (Canada). Chemical
Geology 261, 98^113.
Cavosie, A. J., Valley, J. W. & Wilde, S. A. (2005). Magmatic d18O in
4400^3900 Ma detrital zircons; a record of the alteration and recycling of crust in the early Archean. Earth and Planetary Science
Letters 235, 663^681.
Collerson, K. D. (1983). Ion microprobe zircon geochronology of the
Uivak gneisses: implications for the evolution of early terrestrial
crust in the North Atlantic Craton. Lunar and Planetary Institute
Technical Report 83, 28^33.
Condie, K. C. (1981). Archean Greenstone Belts. Amsterdam: Elsevier.
Condie, K. C. (2005). High field strength element ratios in Archean
basalts: a window to evolving sources of mantle plumes? Lithos 79,
491^504.
Crawford, A. J., Falloon, T. J. & Green, D. H. (1989). Classification,
petrogenesis and tectonic setting of boninites. In: Crawford, A. J.
(ed.) Boninites and Related Rocks. London: Unwin Hyman, pp. 1^49.
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. Geological Society of America Bulletin 121, 150^163.
Falloon, T. J. & Danyushevsky, L. V. (2000). Melting of refractory
mantle at 1·5, 2 and 2·5 GPa under anhydrous and
H2O-undersaturated conditions; implications for the petrogenesis
of high-Ca boninites and the influence of subduction components
on mantle melting. Journal of Petrology 41, 257^283.
Fiske, R. S. & Matsuda, T. (1964). Submarine equivalents of ash flows
in the Tokiwa Formation, Japan. American Journal of Science 262,
76^106.
Furnes, H., Rosing, M., Dilek, Y. & de Wit, M. (2009). Isua supracrustal belt (Greenland); a vestige of a 3·8 Ga suprasubduction zone
ophiolite, and the implications for Archean geology. Lithos 113,
115^132.
Gable, D. J. & Sims, P. K. (1969). Geology and regional metamorphism of some high-grade cordierite gneisses, Front Range,
Colorado. Geological Society of America, Special Papers 128, 87 p.
Grant, J. A. (1968). Partial melting of common rocks as a possible
source of cordierite^anthophyllite bearing assemblages. American
Journal of Science 266, 908^931.
Harrison, T. M., Blichert-Toft, J., Mueller, W., Albare'de, F., Holden, P.
& Mojzsis, S. J. (2005). Heterogeneous Hadean hafnium; evidence
of 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,
d18O, and Ti thermometry results for Hadean zircons. Earth and
Planetary Science Letters 268, 476^486.
Hickey, R. L., Frey, F. A., Gerlach, D. C. & Lopez-Escobar, L. (1986).
Multiple sources for basaltic arc rocks from the Southern Volcanic
Zone of the Andes (348^418S); trace element and isotopic evidence
for contributions from subducted oceanic crust, mantle, and continental crust. Journal of Geophysical Research 91, 5963^5983.
Hoffer, E. & Grant, J. A. (1980). Experimental investigation of the formation of cordierite^orthopyroxene parageneses in pelitic rocks.
Contributions to Mineralogy and Petrology 73, 15^22.
Houghton, B. F. & Landis, C. A. (1989). Sedimentation and volcanism
in a Permian arc-related basin, southern New Zealand. Bulletin of
Volcanology 51, 433^450.
Jolly, W. T. (1975). Subdivision of the Archean lavas of the Abitibi area,
Canada, from Fe^Mg^Ni^Cr relations. Earth and Planetary Science
Letters 27, 200^210.
NUMBER 5
MAY 2011
Jolly, W. T. (1980). Development and degradation of Archean lavas,
Abitibi area, Canada, in light of major element geochemistry.
Journal of Petrology 2, 323^363.
Jolly, W. T. & Hallberg, J. A. (1990). Role of crustal contamination in
heterogeneous Archean volcanics from the Leonora region,
Western Australia. Precambrian Research 48, 75^98.
Kamber, B. S. (2007). The enigma of the terrestrial protocrust: evidence for its former existence and the importance of its complete
disappearance. In: Kranendonk, M. J., Smithies, R. H. &
Bennett, V. C. (eds) Earth’s Oldest Rocks. Amsterdam: Elsevier,
pp. 75^89.
Kamber, B. S., Whitehouse, M. J., Bolhar, R. & Moorbath, S. (2005).
Volcanic resurfacing and the early terrestrial crust; zircon U^Pb
and REE constraints from the Isua greenstone belt, southern West
Greenland. Earth and Planetary Science Letters 240, 276^290.
Kay, S. M., Maksaev, V., Moscoso, R., Mpodozis, C. & Nasi, C.
(1987). Probing the evolving Andean lithosphere; mid^late Tertiary
magmatism in Chile (298^308300) over the modern zone of subhorizontal subduction. Journal of Geophysical Research 92, 6173^6189.
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 and
Planetary Science Letters 296, 45^56.
Liu, D., Wilde, S. A., Wan, Y., Wu, J., Zhou, H., Dong, C. & Yin, X.
(2008). New U^Pb and Hf isotopic data confirm Anshan as the
oldest preserved segment of the North China Craton. American
Journal of Science 308, 200^231.
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.
Mtoro, M., Maboko, M. A. H. & Manya, S. (2009). Geochemistry
and geochronology of the bimodal volcanic rocks of the Suguti
area in the southern part of the Musoma-Mara greenstone belt,
northern Tanzania. Precambrian Research 174, 241^257.
Munoz, B. J. & Stern, C. R. (1989). Alkaline magmatism within the
segment 388^398S of the Plio-Quaternary volcanic belt of the
southern South American continental margin. Journal of
Geophysical Research 94, 4545^4560.
Nesbitt, H. W. & Young, G. M. (1982). Early Proterozoic climates and
plate motions inferred from major element chemistry of lutites.
Nature 299, 715^717.
Nutman, A. P., McGregor, V. R., Friend, C. R. L., Bennett, V. C. &
Kinny, P. D. (1996). The Itsaq gneiss complex of southern West
Greenland; the world’s most extensive record of early crustal evolution (3900^3600 Ma). Precambrian Research 78, 1^39.
Nutman, A. P., Wan, Y. & Liu, D. (2009). Zircon U/Pb and Hf
isotopic constraints on the early Archean evolution in Anshan of
the North China Craton; discussion. Precambrian Research 172,
357^360.
Ohta, T. & Arai, H. (2007). Statistical empirical index of chemical
weathering in igneous rocks: A new tool for evaluating the degree
of weathering. Chemical Geology 240, 280^297.
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. Amsterdam:
Elsevier, 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.
1008
O’NEIL et al.
NUVVUAGITTUQ GREENSTONE BELT
Pearce, J. A. (2008). Geochemical fingerprinting of oceanic basalts
with applications to ophiolite classification and the search for
Archean oceanic crust. Lithos 100, 14^48.
Pearce, J. A. & Cann, J. R. (1971). Ophiolite origin investigated by discriminant analysis using Ti, Zr and Y. Earth and Planetary Science
Letters 12, 339^349.
Pearce, J. A. & Cann, J. R. (1973). Tectonic setting of basic volcanic
rocks determined using trace element analyses. Earth and Planetary
Science Letters 19, 290^300.
Pearce, J. A. & Peate, D. W. (1995). Tectonic implications of the composition of volcanic arc magmas. Annual Review of Earth and
Planetary Sciences 23, 251^285.
Pearce, J. A., Stern, R. J., Bloomer, S. H. & Fryer, P. (2005).
Geochemical mapping of the Mariana arc^basin system; implications for the nature and distribution of subduction components.
Geochemistry, Geophysics, Geosystems 6, 27 p, doi: 10.1029/2004G
C000895.
Peck, W. H. & Smith, M. S. (2005). Cordierite^gedrite rocks from the
Central Metasedimentary Belt boundary thrust zone (Grenville
Province, Ontario); Mesoproterozoic metavolcanic rocks with affinities to the Composite Arc Belt. Canadian Journal of Earth Sciences
42, 1815^1828.
Polat, A. (2009). The geochemistry of Neoarchean (ca. 2700 Ma) tholeiitic basalts, transitional to alkaline basalts, and gabbros, Wawa
Subprovince, Canada; implications for petrogenetic and geodynamic processes. Precambrian Research 168, 83^105.
Polat, A. & Hofmann, A. W. (2003). Alteration and geochemical patterns in the 3·7^3·8 Ga Isua greenstone belt, West Greenland.
Precambrian Research 126, 197^218.
Polat, A., Hofmann, A. W. & Rosing, M. T. (2002). Boninite-like volcanic rocks in the 3·7^3·8 Ga Isua greenstone belt, West
Greenland; geochemical evidence for intra-oceanic subduction
zone processes in the early Earth. Chemical Geology 184, 231^254.
Scher, S. & Minarik, W. (2009). Thermobarometry in the Hadean:
The Nuvvuagittuq Greenstone Belt. EOS Transactions, American
Geophysical Union, Joint Assembly Supplement 90, Abstract V23C-05.
Schiotte, L., Compston, W. & Bridgwater, D. (1989). Ion probe U^Th^
Pb zircon dating of polymetamorphic orthogneisses from northern
Labrador, Canada. CanadianJournal of Earth Sciences 26, 1533^1556.
Sigvaldason, G. E. (1974). Basalts from the centre of the assumed
Icelandic mantle plume. Journal of Petrology 15, 497^524.
Simard, M., Parent, M., David, J. & Sharma, K. N. M. (2003).
Ge¤ ologie de la re¤ gion de la rivie' re Innuksuac (34K et 34L). Que¤bec:
Ministe're des Ressources naturelles, RG 2002^10, 46 pp.
Simon, L. & Nigel, M. K. (2007). Ancient Antarctica: The Archean of
the east Antarctic Shield. In: Kranendonk, M. J., Smithies, R. H.
& Bennett, V. C. (eds) Earth’s Oldest Rocks. Amsterdam: Elsevier,
pp. 149^186.
Sisson, T. W. & Grove, T. L. (1993). Experimental investigations of the
role of H2O in calc-alkaline differentiation and subduction zone
magmatism. Contributions to Mineralogy and Petrology 113, 143^166.
Smith, M. S., Dymek, R. F. & Schneiderman, J. S. (1992).
Implications of trace element geochemistry for the origin of
cordierite^orthoamphibole rocks from Orijarvi, SW Finland.
Journal of Geology 100, 545^559.
Smithies, R. H., Champion, D. C. & Sun, S. (2004). The case for
Archaean boninites. Contributions to Mineralogy and Petrology 147,
705^721.
Song, B., Nutman, A. P., Liu, D. & Wu, J. (1996). 3800 to 2500 Ma
crustal evolution in the Anshan area of Liaoning Province, northeastern China. Precambrian Research 78, 79^94.
Spear, F. S. (1993). Metamorphic Phase Equilibria and Pressure^
Temperature^Time Paths. Washington, DC: Mineralogical Society of
America.
Spulber, S. D. & Rutherford, M. J. (1983). The origin of rhyolite and
plagiogranite in oceanic crust; an experimental study. Journal of
Petrology 24, 1^25.
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, London, Special Publications 42,
313^345.
Upadhyay, D., Scherer, E. E. & Mezger, K. (2009). 142Nd evidence for
an enriched Hadean reservoir in cratonic roots. Nature 459,
1118^1121.
Tatsumi, Y. & Maruyama, S. (1989). Boninites and high-Mg andesites:
tectonics and petrogenesis. In: Crawford, A. J. (ed.) Boninites and
Related Rocks. London: Unwin Hyman, pp. 50^71.
Vallance, T. G. (1967). Mafic rock alteration and isochemical development of some cordierite^anthophyllite rocks. Journal of Petrology 8,
84^96.
Valley, J. W., Peck, W. H., King, E. M. & Wilde, S. A. (2002). A cool
early Earth. Geology 30, 351^354.
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.
Williams, I. S., Black, L. P. & Compston, W. (1986). Four zircon ages
from one rock: the evolution of a 3930 Ma-old granulite from
Mount Sones, Antarctica. Terra Cognita 6, 153.
Wu, F., Zhang,Y.,Yang, J., Xie, L. & Yang,Y. (2008). Zircon U/Pb and
Hf isotopic constraints on the early Archean crustal evolution in
Anshan of the North China Craton. Precambrian Research 167,
339^362.
Wu, F., Zhang,Y.,Yang, J., Xie, L. & Yang,Y. (2009). Zircon U/Pb and
Hf isotopic constraints on the early Archean crustal evolution in
Anshan of the North China Craton; reply. Precambrian Research 172,
361^363.
Wyman, D. A. (1999). A 2·7 Ga depleted tholeiite suite: evidence of
plume^arc interaction in the Abitibi greenstone belt, Canada.
Precambrian Research 97, 27^42.
Young, G. M. (1973). Tillites and aluminous quartzites as possible time
markers for middle Precambrian (Aphebian) rocks of North
America. In: Grant, M. Young (ed.) Huronian Stratigraphy
and Sedimentation. Geological Association of Canada, Special Papers 12,
97^127.
1009