Journal of Petrology Advance Access originally published online on January 31, 2008
Journal of Petrology 2008 49(3):393-420; doi:10.1093/petrology/egm084
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Petrology of a Late Archaean, Highly Potassic, Sanukitoid Pluton from the Baltic Shield: Insights into Late Archaean Mantle Metasomatism
1Institute of the Precambrian Geology and Geochronology Ras, st. Petersburg, Russia
2Department of earth sciences, sultan qaboos university, Box 36, Postal Code Al-Khodh-123, Muscat, Sultanate of Oman
3Laboratoire Magmas Et Volcans, Opgc–Université Blaise Pascal–Cnrs; 5 Rue Kessler, 63038 Clermont-Ferrand Cedex, France
RECEIVED FEBRUARY 18, 2007; ACCEPTED DECEMBER 10, 2007
 |
ABSTRACT |
|---|
The late Archaean Panozero pluton in Central Karelia (Baltic Shield) is a multi-phase high-Mg, high-K intrusion with sanukitoid affinities, emplaced at 2·74 Ga. The magmatic history of the intrusion may be subdivided into three cycles and includes monzonitic and lamprophyric magmas. Compositional variations are most extreme in the monzonite series and these are interpreted as the result of fractional crystallization. Estimates of the composition of the parental magmas to the monzonites and lamprophyres show that they are enriched in light rare earth elements, Sr, Ba, Cr, Ni and P but have low contents of high field strength elements. Radiogenic isotope data indicate a low U/Pb, high Th/U, high Rb/Sr, low Sm/Nd source. The magmatic rocks of the Panozero intrusion are also enriched in H2O and CO2; carbon isotope data are consistent with mantle values, indicating a fluid-enriched mantle source. The similarity in trace element character of all the Panozero parental magmas indicates that all the magmas were derived from a similar mantle source. The pattern of trace element enrichment is consistent with a mantle source enriched by fluids released from a subducting slab. Nd-isotope data suggest that this enrichment took place at c. 2·8 Ga, during the main episode of greenstone belt and tonalite–trondhjemite–granodiorite formation in Central Karelia. Sixty million years later, at 2·74 Ga, the subcontinental mantle melted to form the Panozero magmas. Experimental studies suggest that the monzonitic magmas originated by the melting of pargasite–phlogopite lherzolite in the subcontinental mantle lithosphere at 1–1·5 GPa. The precise cause of the melting event at 2·74 Ga is not known, although a model involving upwelling of asthenospheric mantle following slab break-off is consistent with the geochemical evidence for the enrichment of the Karelian subcontinental mantle lithosphere by subduction fluids.
KEY WORDS: Archaean; sanukitoid; monzonite; Karelia; mantle metasomatism
|
INTRODUCTION |
|---|
Highly potassic mafic magmas are rare in the Archaean, but where they are found they provide an insight into a mantle source region of a very particular type. This mantle source is thought to be very different from that from which the more common Archaean mafic and ultramafic rocks of greenstone belts are derived and also different from the source of the ubiquitous Archaean tonalite–trondhjemite–granodiorite (TTG) magmas. Here we report the geochemistry of a high-K, sanukitoid intrusion from Karelia and associated lamprophyres, which we use to characterize their source.
In an earlier study (Lobach-Zhuchenko et al., 2005a
; see also Bibikova et al., 2005) we showed that late Archaean sanukitoid magmas are an important constituent of the Baltic Shield. We argued that they formed after the cratonization of the Shield from an enriched source in the subcontinental lithospheric mantle. In this new study we have selected the Panozero intrusion for more detailed examination because it is the most potassic of the Karelian sanukitoids. Elsewhere, similar rocks are known only from the Clericy syenite pluton (Lafleche et al., 1991), the Kirland Lake syenites (Kerrich & Watson, 1984) or as dykes in the Abitibi Belt (Hattori et al., 1996). In addition, the Panozero intrusion is well exposed, and shows an extensive differentiation history, including mafic and ultramafic members and lamprophyre dykes. These mafic components provide important clues to the source of the magmas. Here we use a much larger sample set and a wider range of geochemical methods than in our earlier studies to identify the primary magmas from which the complex was derived and to infer a likely mantle source. We conclude that the Panozero intrusion contains a critical suite of rocks for the petrogenetic interpretation of the late Archaean sanukitoid magmatism of the Baltic Shield.
|
GEOLOGICAL BACKGROUND |
|---|
Tectonic setting
The 2·74 Ga Panozero sanukitoid pluton intrudes the central part of the middle to late Archaean granite–greenstone terrain of the Karelian Craton, in the SE of the Fennoscandian Shield. The pluton is located 2 km WSW of Lake Segozero in Central Karelia. It crops out around the margins of Lake Panozero as an oval composite body (4 km by 7 km) that intrudes the volcano-sedimentary sequence of the West Segozero greenstone belt (Fig. 1). The host rocks are greenschist-facies pelites and metabasites (Glebova-Kulbach et al., 1963
). The Panozero intrusion is probably related to the Elmus and Sharavalampi sanukitoid intrusions and the Syargozero and Hizhjarvi syenites; these are the same age as the Panozero intrusion and are situated along a common tectonic lineament, which probably controlled their emplacement (Ivanikov, 1997). These intrusions cut older (3·0–2·9 Ga) greenstone belts in the south, and were emplaced some 60 Myr after the formation of the younger (2·8–2·85 Ga) West Segozero greenstone belt in the north and TTG granitoids of similar age.
|
Age and composition of the Panozero pluton
The Panozero pluton is a central-type intrusion of potassic ultramafic to felsic plutonic igneous rocks. There are three magmatic cycles (Fig. 1, Table 1) intruded over a short time interval. Conventional zircon U–Pb age determinations on the pluton gave ages between 2737 ± 10 and 2727 ± 4 Ma (Chekulaev et al., 2003
). More recent ionprobe secondary ionization mass spectrometry (SIMS) U–Pb age determinations on the same rocks gave ages in the range 2741 ± 12 and 2742 ± 18 Ma (Bibikova et al., 2005). Here we use the SIMS age of 2·74 Ga for the emplacement age of the pluton, as it is based upon more representative samples than the conventional ages.
|
The first magmatic cycle
The rocks of the first magmatic cycle comprise a mafic complex, which includes pyroxenites, pyroxene hornblendites, monzogabbro and a monzonite, termed here monzonite-1.
The mafic complex forms a unit 300–350 m thick on the SE side of the pluton (Fig. 1). Ultramafic rocks form a marginal facies about 3–6 m thick. They are succeeded by a layered sequence of mafic and ultramafic rocks in which irregular bodies of ultramafic rock are interleaved with the mafic rocks. The mafic rocks make up about 70% of this sequence. These rocks have been deformed and in places biotite crystallization defines a schistosity.
The monzonitic phase contains abundant irregular mafic lenses and schlieren with sharp lobate to cuspate contacts with the host monzonite. These mafic lenses are from a few centimetres to tens of centimetres in length. An absence of chilled contacts between the mafic lenses and the host monzonite suggests that there was only a small temperature contrast between the two, indicative of mingling between two immiscible magmas (Fig. 2a and b) (Rollinson, 2003). Elsewhere, the contacts are more gradational, indicating mixing and assimilation between the mafic and felsic phases. Both these textures imply that at the time of monzonite emplacement a monzonitic melt coexisted with a mafic melt. In places the monzonites also contain large xenoliths (up to 6 m in length) of mafic volcanic rocks, which probably represent the former roof of the pluton (Chekulaev et al., 2003; Bibikova et al., 2005).
|
A further feature of the rocks of the first magmatic cycle is the presence of intrusive breccia dykes between 20 and 80 cm wide. These dykes contain abundant small xenoliths of monzogabbro and intermediate to felsic volcanic rocks from the host greenstone belt and a small amount of matrix (Lobach-Zhuchenko et al., 2005b
) (Fig. 2c).
The final stage of the first magmatic cycle was marked by the intrusion of lamprophyre stocks and dykes, varying in width from several centimetres to 1–2 m (Table 1, Figs 2d and 3a, b). Their distribution suggests that they probably formed within a restricted zone within the pluton. They contain small ultramafic xenoliths and sometimes ocelli, similar to those described from alkaline intrusions associated with carbonatites and lamprophyres (Le Roex & Lanyon, 1998).
|
The second magmatic cycle
The rocks of the second magmatic cycle are massive, medium- to fine-grained monzonites (termed monzonite-2) and monzogabbros. The monzonites are characterized by small (1–15 cm) inclusions of amphibolite, which in the SW part of the pluton have a shallow dip, probably indicating the direction of magma flow. The monzonite-2 phase is intruded by a second generation of lamprophyre dykes, which sometimes may contain small xenoliths of monzonite-1. Our field observations suggest that the rocks of the first cycle were deformed in shear zones before the magmas of the second cycle were emplaced. This deformation seems to have controlled the emplacement of the magmas of the second cycle, particularly in the southern part of the pluton, giving rise to a large-scale intrusion breccia.
The third magmatic cycle
The rocks of the third magmatic cycle can be divided into three stages. The first of these is of relatively minor importance and is represented by small intrusive bodies both within and outside the pluton (Fig. 1). These rocks are porphyritic and contain fragments of earlier lamprophyre dykes. The second stage of this cycle is the emplacement of the quartz monzonite, which makes up a major part of the central and northern part of the Panozero pluton (Fig. 1). Although this quartz monzonite appears to represent a major volume of the Panozero intrusion it has shallow dipping contacts, and by analogy with the Kurgenlampi intrusion (Lobach-Zhuchenko et al., 2005a) is thought to have a sheet-like form. The final intrusive phase comprises dykes and veins of leucocratic granites, aplites, and pegmatites, some of which have been displaced by later faulting.
|
ANALYTICAL METHODS |
|---|
Sr and Nd isotopes
Sr and Nd isotopes were analysed at the Institute of Precambrian Geology and Geochronology, Russian Academy of Sciences, in St. Petersburg, on a Finnigan MAT-261 multicollector mass spectrometer equipped with eight cups. The samples were analysed for Rb, Sr, Nd concentrations and 87Sr/86Sr and 143Nd/144Nd ratios. Rb, Sr, Nd and Sm concentrations were determined by isotope dilution using 85Rb–84Sr and 150Nd–149Sm spikes. The accuracy ±0·5% for Rb and Sr, ±1·1% for 87Rb/86Sr and ±0·01% for 87Sr/86Sr (2
). 87Sr/86Sr ratios were corrected to 0·71024 for the SRM-987 Sr-standard. A weighted average of 87Sr/86Sr in the Sr-standard (nine runs) yielded 0·710249 ± 8 (2) when normalized to a 88Sr/86Sr ratio of 8·37521. Total laboratory blanks were 0·05–0·15 ng for Rb and 0·1–0·5 ng for Sr. The accuracy is ±0·5% for Sm and Nd, ±0·5% for 147Sm/144Nd and ±0·005% for 143Nd/144Nd (2). The 143Nd/144Nd ratios were corrected to the value 0·511860 for the La Jolla Nd-standard. A weighted average of 143Nd/144Nd in the Nd-standard (nine runs) yielded 0·511852 ± 8 (2) when normalized to a 146Nd/144Nd ratio of 0·7219. Total laboratory blanks were 0·1–0·2 ng for Sm and 0·1–0·5 ng for Nd. 
Nd(t) values were calculated using the present-day values for a chondritic uniform reservoir (CHUR) 143Nd/144Nd = 0·512638 and 147Sm/144Nd = 0·1967 (Jacobsen & Wasserburg, 1984).
Pb isotopes
Uranium and lead concentrations were determined by isotope dilution method using a 235U–208Pb spike. The magnitude of the mass fractionation in Pb and U was determined by running the NIST standards SRM-982 and SRM U-500, respectively (Todt et al., 1996). Measured lead and uranium isotope ratios were corrected for instrumental mass fractionation using a factor of 0·1% per a.m.u. for Pb and 0·34% per a.m.u. for U. The data are accurate to within 0·1% for 206Pb/204Pb, 0·15% for 207Pb/204Pb, and to within 0·2% for 208Pb/204Pb. Total blanks were 0·2 ng for Pb and 0·05 ng for U. U and Pb concentrations, U/Pb and Pb isotope ratios were calculated using PBDAT (Ludwig, 1993). Regression lines, 2 values and Stacey & Kramers (1975) model ages were calculated using ISOPLOT/Ex 2.2 (Ludwig, 2000).
Carbon and oxygen isotopes
CO2 was released using phosphoric acid. Carbon and oxygen isotope compositions were measured on a two-channel AEI MS-20 mass spectrometer at the VSEGEI Isotope Centre, St. Peterburg. Precision for C is about 0·1% and for O is 0·15%. Accuracy is about 0·15% for both elements.
Whole-rock analysis—major and trace elements
Whole-rock major element compositions were determined by X-ray fluorescence spectrometry (XRF) at VSEGEI, St. Petersburg; The trace elements Rb, Ba, Th, Ti, Sr, Zr, Y, Ni, Cr were analysed by XRF on a VRA-30 system at the IPGG, RAS, St. Petersburg. These measurements are accurate to 0·1–0·2%. Errors are 10% for values <5 ppm, and 5% for higher concentrations. Other trace elements and rare earth elements (REE) were determined by inductively coupled plasma mass spectrometry (ICP-MS) at VSEGEI, St. Petersburg and at the University of Kingston, UK. The analytical uncertainty is typically <3%.
Mineral analyses—major and trace elements
Electron microprobe mineral analyses were made using a LINK-AN10/85S detector attached to an ABT-55 electron microscope at IPGG RAS. Operating conditions were 25 kV. Mineral totals (Table 3) were normalized to 98% (amphiboles), 95% (biotites) and 100% (dry minerals) to facilitate comparisons between analyses. Mineral REE compositions were determined on a CAMECA-IMS-4f ion microprobe, using the NIST 610 standard, at the Institute of Microelectronic and Informatics, RAS, Yaroslavl, using the method of Sobolev & Shimizu (1993).
|
PETROGRAPHY, MINERAL CHEMISTRY AND GEOCHEMISTRY |
|---|
 TOPABSTRACTINTRODUCTIONGEOLOGICAL BACKGROUNDANALYTICAL METHODS PETROGRAPHY, MINERAL CHEMISTRY... ISOTOPE GEOCHEMISTRYTHE ORIGIN OF THE...TECTONIC IMPLICATIONSCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
The rocks of the Panozero pluton have high total alkali contents and plot in the alkaline field on a total alkali–silica variation diagram (Fig. 4; Tables 2 and 3). Representative mineral chemical data are given in Electronic Appendix 1 and details of sample localities are given in Electronic Appendix 2 (available for downloading at http://www.petrology.oxfordjournals.org). Typically most samples plot within the monzonite series (Fig. 4) given that they have approximately equal modal amounts of plagioclase and alkali feldspar. Some biotite-rich samples have a low alkali feldspar content, but are monzonites according to the chemical classification of Middlemost (1994
). In detail, we recognize two magmatic series, which we identify on the basis of their field relationship and their geochemistry. These are: (1) a monzonite series comprising pyroxenites, monzogabbros, monzonites and quartz monzonites; the monzonite series is present in all three of the magmatic cycles present at Panozero; (2) lamprophyre dykes. The magmas of the monzonite series make up the major part of the pluton.
|
|
The monzonite series
The rocks of the monzonite series vary in composition from mafic to felsic.
Petrography
Mafic rocks occur in two forms. Some are layered and occur as a marginal facies to the main pluton (Fig. 1). Others occur as mafic schlieren in monzonite-1 and represent an immiscible mafic melt that coexisted with the monzonite melt. Compositionally the mafic rocks vary from biotite-bearing pyroxenites to pyroxene–biotite–plagioclase monzodiorites. The principal chemical variations in these rocks reflect the relative proportions of mafic to felsic minerals (40–90% mafic) and the biotite/pyroxene ratio.
The primary minerals are clinopyroxene, magnesium-rich biotite, apatite (up to 5 wt %), and pargasite. Secondary minerals include actinolite (after pyroxene), epidote, chlorite and albite + zoisite after plagioclase. Accessory minerals include zircon, allanite, carbonate, pyrite, and chalcopyrite. Normative calculations suggest that the primary plagioclase composition was bytownite. Our field observations combined with the geochemical data described below lead us to the conclusion that these rocks were produced by the differentiation of a monzogabbro parent magma.
Monzonites are present in all three of the intrusive cycles and three monzonite phases have been identified (Table 1). Monzonite-1 is coarse-grained and contains clinopyroxene, abundant biotite, about 15% K-feldspar, and about 50% plagioclase (altered to albite + zoisite), which on the basis of normative calculations is 
An30. Coarse-grained clinopyroxene and plagioclase are present within a matrix of plagioclase and K-feldspar. Clinopyroxene phenocrysts have coronas of biotite, which may reflect the peritectic reaction diopside + melt = biotite. In one sample pyroxene is mantled by carbonate and elsewhere diopside is replaced by tremolite. Pargasite is rare. All minerals have irregular grain boundaries. Monzonite-2 has a hypidiomorphic–granular or porphyritic texture with phenocrysts of hornblende and plagioclase, partially replaced by actinolite and albite + zoisite, relics of pyroxene and about 5% K-feldspar. Monzonite-3 is medium-grained with a porphyritic texture and contains large euhedral grains of plagioclase, and amphibole relics surrounded by biotite in a quartz–microcline matrix.
Quartz monzonites are coarse-grained, homogeneous rocks, sometimes porphyritic. They contain euhedral grains of plagioclase and two types of K-feldspar, one of which has up to 10% Na2O. The mafic minerals are amphibole and biotite, which make up <10% of the rock. Apatite, zircon, titanite and opaque minerals are the main accessory phases.
Mineral chemistry
Clinopyroxene is the first crystallizing phase of the monzonite series. It is abundant in the mafic and ultramafic members and in the monzonites of the first cycle. Clinopyroxenes are unzoned and have a mg-number = 0·72–0·75, similar to those from the Roaring River sanukitoid pluton in the Superior Province, Canada (Stern & Hanson, 1991). There is little compositional variation within the clinopyroxenes in the monzogabbro–monzonites of the first cycle, apart from a small decrease in Al2O3 and MgO. In the pyroxenites clinopyroxene is lower in Na2O and higher in CaO and TiO2.
REE concentrations were measured in clinopyroxenes from a pyroxenite, a monzogabbro and a monzonite from the first cycle. All show curved light REE (LREE) patterns with normalized concentrations increasing towards either Ce or Nd, after which concentrations sytematically decrease (Fig. 5). The highest REE concentrations are found in the monzogabbro. The shape of these REE patterns is unusual for igneous clinopyroxenes, although similar patterns have been noted in clinopyroxenes from metasomatized mantle xenoliths (Ionov et al., 2002). The different REE patterns in these pyroxenes are thought to reflect the different REE chemistry of their host melts, although the clinopyroxenes in monzonite-204 have LREE concentrations that vary by a factor of six. These patterns all have a similar shape and it is possible that the differences reflect chemical disequilibrium during fractional crystallization. Van Orman et al. (2001) showed that REE diffusion coefficients in clinopyroxene are low at magmatic temperatures, allowing chemical disequilibrium to persist during crystallization.
|
Apatite is a primary mineral crystallizing along with the clinopyroxene and is more abundant in the pyroxenites than the quartz monzonites. There are vapour inclusions of H2O and CO2 in the cores of some grains.
Micas are high-Mg biotites (mg-number = 0·55–0·69) and similar in composition to those in the Roaring River sanukitoid pluton (Stern & Hanson, 1991).
Amphiboles are calcic and according to the classification of Leake et al. (1997) are Mg-hornblende, edenite and pargasite (hornblende group) and actinolite–tremolite. Hornblende is the alteration product of clinopyroxene and is itself altered to biotite and actinolite. We calculated the lithostatic pressure at the time of amphibole crystallization for monzonite-2 (sample 235, Table 2) and quartz-monzonite (sample 156, Table 2) using the amphibole Al-barometer of Johnson & Rutherford (1989) and obtained a pressure of 1· 6 ± 0·6 kbar. We note, however, that Blundy & Holland (1990) argued that Al exchange in edenite is also governed by temperature, in which case a lower pressure may be appropriate. Using the same method we estimated the emplacement pressure of the Roaring River pluton described by Stern & Hanson data (1991) to be 0·6 ± 0·6 kbar, indicating that both intrusions crystallized at a relatively shallow level in the crust.
Plagioclase compositions, estimated from normative calculations, vary from labradorite in monzogabbro to oligoclase in the quartz-monzonite. Typically, the plagioclase is made up of aggregates of albite and epidote.
Carbonate occurs in all rock units of the monzonite series and varies from rare separate grains to thin quartz–carbonate and quartz–microcline–carbonate (aplitic) veins with 10% of carbonate.
Whole-rock major element chemistry
Representative whole-rock analyses of the monzonite series are given in Table 2. The principal geochemical characteristic of these rocks is that they have very high alkali and incompatible element contents (see Fig. 4). On a K2O vs Na2O diagram many of the mafic samples plot in the ultrapotassic field whereas the more felsic rocks are shoshonitic (Fig. 6a). First cycle monzonites have higher K contents than the monzonites of cycles 2 and 3 (Fig. 6a and b), implying a decrease in shoshonitic character during magmatic evolution. In our earlier study (Lobach-Zhuchenko et al., 2005a) we showed that on Harker variation diagrams the rocks of the first cycle display negative trends for FeO, MgO, CaO and P2O5 (Fig. 7e) and positive trends for K2O, Na2O, and Al2O3 with increasing silica (Fig. 7a and b). These relationships are considered to be the product of crystal fractionation; a detailed discussion of the likely fractionating phases involved is given below.
|
|
Of particular importance in the context of this study are the mafic rocks of the first cycle. There is no compositional difference in either major element chemistry or REE chemistry between those mafic rocks that occur as lenses and schlieren in the monzonite-1 magma and those that form discrete layered sequences at the margins of the complex (Fig. 7f). The presence of mafic inclusions within monzonite-1 (Fig. 2a and b), which are interpreted as the product of magma mingling, demonstrates that mafic melts and felsic melts coexisted during the early stages of crystallization of the intrusion.
Second cycle monzonites are poorer in K2O and have lower Na2O, Al2O3, Sr, and Ba than the monzonites of the first cycle (Figs 6a and 7a–d). Monzonites of the third cycle are more leucocratic and richer in SiO2.
Trace element chemistry
In primitive mantle-normalized trace element diagrams the rocks of the monzonite series show a general trend of decreasing concentrations with increasing compatibility in the mantle source. They have negative Nb–Ta, Zr–Hf, P and Ti anomalies, high abundances of Sr, Ba and LREE, and small positive Pb anomalies (Figs 8a and 9b). The mafic members of the monzonite series are strongly enriched in LREE and have abundances between 200 and 600 times chondrite (Fig. 9). They plot in the shoshonite field on the Ce/Yb vs Ta/Yb diagram of Pearce (1982). Both mafic and felsic members of the monzonite series have strongly fractionated REE patterns [(La/Yb)N between 17 and 37; Fig. 9]. REE abundances decrease with increasing silica content, such that the monzonites have lower REE abundances than in the mafic rocks and quartz-monzonites have lower REE abundances than the monzonites (Fig. 9b). We note a correlation between Ce and P2O5 suggesting that apatite content exercises some control on the LREE abundances in these rocks. The negative P anomaly in mantle-normalized trace element diagrams could imply apatite fractionation, although, as this is common feature of the monzonite and lamprophyre magmas, it is possible that this is a characteristic of the mantle source.
|
|
Lamprophyres
Two groups of lamprophyre dykes have been identified at Panozero on the basis of their chemistry and field relationships. There is an early ocelli-bearing lamprophyre group, emplaced after the monzonites in the first magmatic cycle and a later group of lamprophyres. The later lamprophyres were emplaced in two stages, the first immediately after the ocelli-bearing lamprophyres in the first magmatic cycle and the second after the second cycle monzonites (Table 1). Representative chemical analyses of the two lamprophyre types are given in Table 3.
Early, ocelli-bearing lamprophyres
Early lamprophyres contain small spherical or ellipsoidal ocelli, 0·5–4·5 cm across, comprising 1–3% of the rock. This group of lamprophyres contain 60–80% tremolite–actinolite, 20–30% phlogopite, calcite and dolomite and, in the more silica-rich samples, about 10% albite. It is unlikely that this is the primary mineralogy of these rocks, and normative calculations show 23–35% diopside and 30–40% olivine. Accessory phases include titanite, allanite, apatite (1–2%), sulphides (pyrite, chalcopyrite, pyrrhotite, pentlandite), magnetite, chrome-spinel, and barite. Some lamprophyres also contain ultramafic and mafic inclusions that are compositionally distinct from the other rocks of the Panozero intrusion (Table 3).
|
Major element chemistry. The early ocelli-bearing lamprophyres are characterized by a high potassium content, low Al2O3 (< 10%), low Na2O (< 1%) and high MgO (14·4–19·2 wt %; mg-number = 0·68–0·79). They have leucite in their norm. On a K2O–Na2O diagram they plot in the ultrapotassic and shoshonitic fields (Fig. 6c). In detail we identify a low-silica group (45–47% SiO2) and a high-silica group (50–52% SiO2) (Table 3). Rocks of the low-silica group are ultrapotassic and show some similarities to late Archaean minettes from Abitibi described by Wyman & Kerrich (1993
) and Wyman et al. (2006) and the myaskitic lamproites described by Bogatikov et al. (1991) from the Aldan Shield. Trace element chemistry. On a primitive mantle-normalized trace element diagram the early ocelli-bearing lamprophyres show negative Nb, Zr, Ti and Sr anomalies (Fig. 8b), similar to those observed in the monzonite series. Both the high- and low-silica varieties are strongly enriched in LREE (Fig. 10). However, the low-silica group have flatter LREE patterns (La/YbN = 12·7–28·5) than the higher silica group (La/YbN = 13·2–35·61). An ultramafic inclusion has much lower REE abundances and a sigmoidal REE pattern.
|
Ocelli. The ocelli are zoned with radiating tremolite needles at their margins (possibly after clinopyroxene), an inner zone of phlogopite (mg-number = 0·70) and a core of phlogopite, tremolite–actinolite, carbonate, apatite, chrome-spinel, magnetite, sulphides, and barite. These textures are illustrated in Fig. 11. Tremolite-rich and carbonate-rich varieties are the most common. They have been interpreted as original immiscible droplets of a magnesium-rich melt, enriched in the volatiles H2O and CO2 (Kogarko et al., 1995
), and related to the late exsolution of magmatic gases (Rock, 1991), respectively.
|
The carbonates are either calcite or dolomite, with traces of Sr, Ba, Ce, and Cr. Chromites in the lamprophyre matrix have cr-number = 0·82, whereas in the ocelli they have cr-number up to 0·93. They are very Fe-rich and relatively high in Mn, Co and Zn, similar in composition to chromites reported from lamprophyres elsewhere (Wyman et al., 2006
). Relative to the rock matrix the ocelli are enriched in SiO2 (by about 4 wt %), MgO and CaO, and have a higher mg-number, but are depleted in Al, Ti, Fe, V, alkalis, Sr, Y, and the REE (Table 3). The ocelli have REE abundances that are lower with less fractionated patterns (La/YbN = 8·3–14·8) than their host. The REE patterns are sigmoidal in shape. Ocelli in the low-silica lamprophyres have slightly lower REE abundances than ocelli in the high-silica lamprophyres. The REE patterns of the ocelli are similar in shape to those of clinopyroxene in the pyroxenites (Figs 5 and 10), possibly suggesting some clinopyroxene control in their petrogenesis.
Late lamprophyres
The late lamprophyres are medium- to fine-grained melanocratic rocks, with Mg-biotite, amphibole (mg-number = 0·55–0·60) and plagioclase as a major phases and apatite, titanite, sulphides, and Fe-rich chromite and magnetite as accessory minerals. Representative whole-rock analyses are given in Table 3. Relative to the early ocelli-bearing group they are more aluminous (11· 0–15·7 wt %), less magnesian (6·38–12 wt %; mg-number = 0·55–0·72) and more sodic (1·7–3·7 wt %). On the Na2O vs K2O clasification diagram of Turner et al. (1996) they plot in the shoshonite field (Fig. 6c), whereas on the K2O vs SiO2 plot of Peccerillo & Taylor (1976) they plot as absarokites (Fig. 6d). When compared with the average calc-alkaline lamprophyre and average kersantite of Rock (1991) the Panozero lamprophyres have similar concentrations of alkalis and silica, but lower TiO2, P2O5, Al2O3, volatiles and incompatible elements, and higher FeO, MgO and CaO (Fig. 8c). At a given silica content they have a higher mg-number than the mafic rocks in the monzonite suite but have lower mg-number than the early ocelli-bearing lamprophyres (Fig. 7f).
On a mantle-normalized trace element diagram the late Panozero lamprophyres are similar to the early ocelli-bearing lamprophyres and the monzonite suite (Fig. 8). REE in the late lamprophyres are not as strongly fractionated (La/Ybn= 4·9–11· 6) as those in the early ocelli-bearing group (Fig. 10).
Summary
In this section we have described the main magmatic phases of the Panozero intrusion and have shown that they formed in three magmatic cycles. The main monzonite series is a high-Mg magma series with a wide range of SiO2 contents (39–67 wt %) and high alkali contents (K2O + Na2O = 5–10 wt %; K2O/Na2O >0·5) similar to volcanic rocks of the shoshonitic series (Morrison, 1980). During the evolution of the monzonite series, from monzonite-1 to quartz monzonite, SiO2 increases and there is a progressive reduction in K2O, Sr, Ba, Zr, REE and mg-number (Fig. 7).
The two groups of lamprophyres—the early ocelli-bearing group and the late lamprophyres—differ in composition but are part of a compositional continuum, implying a common parentage They are typically lower in alkalis and have a higher mg-number than the rocks of the monzonite series but have similar mantle-normalized multi-element patterns. They all show enrichment in the large ion lithophile elements (LILE; Rb, Ba, Th) of between 100 and 300 times primitive mantle, and show decreasing concentrations towards the high field strength elements. There are similar negative anomalies for the elements Nb, P, Zr and Ti (Fig. 8), suggesting some fundamental similarities in their source region.
|
ISOTOPE GEOCHEMISTRY |
|---|
Pb isotopes
U–Pb zircon dating shows that the rocks of the Panozero intrusion crystallized between 2·74 and 2·73 Ga (Chekulaev et al., 2003
; Bibikova et al., 2005). However, lead isotope measurements on K-feldspars indicate that their Pb-isotope system was disturbed after the emplacement of the intrusion. A plot of 207Pb/204Pb vs 206Pb/204Pb for feldspar residues after leaching defines a 2·75–1· 9 Ga palaeo-isochron, reflecting the influence of a 1·9 Ga Svecofennian thermal event on these rocks (Table 4, Fig. 12b). Stacey & Kramers (1975) model ages for these K-feldspars fall between 2·64 and 2·21 Ga, further supporting the view that some redistribution of lead occurred after the emplacement of the intrusion. A two-stage Stacey & Kramers (1975) model of isotopic evolution indicates a source with µ = 8·98; a single stage model gives µ = 7·6.
|
) and residues (
) for the Panozero pluton plotted on a 206Pb/204Pb vs 207Pb/204Pb diagram. (c) K-feldspar data for the Panozero pluton plotted on a 206Pb/204Pb vs 208Pb/204Pb diagram together with model evolution curve of Stacey & Kramers and the mantle, lower and upper crust model evolution curves of Zartman & Doe (1981
|
Model µ values calculated for individual samples vary between 9·06 and 8·69, indicating a low U/Pb ratio in the source (Table 4). In contrast, quartz monzonite sample 156, the last main magmatic phase, has a model age of 2·64 Ga and a µ of 10·43, higher than typical mantle, indicating a significant upper crustal lead component in the quartz monzonite (Fig. 12).
The Panozero K-feldspars also scatter on a 208Pb/204Pb vs 206Pb/204Pb isotope ratio plot and define a range of 232Th/238U (
) ratios (Fig. 12c). Feldspars in sample 205 (monzonite-1) have a model = 4·05 for a fixed µ two-stage Stacey & Kramers (1975) model and = 4·25 for a one-stage model. In contrast, sample 571 (monzonite-1, with strong hydrothermal alteration) has a lower 232Th/238U ratio ( = 0·50) and plots close to the Zartman & Doe (1981) upper crustal curve, probably as result of isotopic resetting at 1·9 Ga. For this reason we prefer sample 205 (monazite-1) as representative of the least modified 232Th/238U () ratio in this suite.
If the initial µ value (238U/204Pb ratio) for the mantle is taken as 11·2 (after Kramers & Tolstikhin, 1997) and the initial bulk Earth Th/U ratio is 4·0 ± 0·2 (Rocholl & Jochum, 1993) then the Panozero K-feldspar U–Th–Pb isotopic data show that the source of the Panozero magmas has (1) a low µ (U/Pb) relative to primitive mantle; (2) a high (Th/U) ratio relative to primitive mantle; this is also supported by the high Th/U (>0·5) ratios in Panozero zircons compared with zircons in Karelian tonalites (<0·5) (Bibikova et al. 2005).
These observations imply that the parental magmas of the Panozero pluton came from a mantle source that is different from typical ‘Zartman & Doe (1981) mantle’. In contrast, the K-feldspars in the late quartz monzonites (sample 156) have a much higher µ value, similar to that found in some other Karelian intrusions (marked as ‘H’ in Fig. 12), which requires a crustal contribution to the magma source (Halla, 2005).
Nd- and Sr-isotopes
Both the Sr- and Nd-isotopic systems also show evidence of having been disturbed after the emplacement of the Panozero intrusion. However, it is still possible to recover information from these data about the initial Sr and Nd isotopic compositions of the Panozero magmatic suite.
Sr isotope analyses for the Panozero sanukitoid intrusion and its volcano-sedimentary country rocks (Table 5) scatter between 2·73 and 1· 9 Ga reference isochrons, suggesting open-system behaviour. A sample of country rock (309) with the lowest Rb/Sr ratio, and so the least disturbed Sr-isotope ratio, was selected as indicative of the Sr-isotope composition of the country rocks at the time of emplacement (87Sr/86Sr(2700) = 0·7049). We also measured the Sr-isotopic composition of the minerals clinopyroxene and apatite, as both are capable of preserving primary magmatic Sr-isotope information (Hattori et al., 1996; Table 5). However, only one sample (236/1-03) preserves Sri values in apatite and clinopyroxene that are similar (0·7015 and 0·7017). All other samples display open-system behaviour.
|
Sm–Nd ages for apatite–clinopyroxene pairs in samples 204-4, 236/1-03 and 238/11-03 are 2613 ± 150, 2764 ± 130 and 2555 ± 200 Ma, and differ from the crystallization age of the intrusion. Sample 236/1-03 has an age closest to that of the U–Pb zircon age of the Panozero intrusion, and is also the sample that has near-concordant initial Sr-isotope ratios for clinopyroxene and apatite (0·7015–0·7017). This would seem to be the least disturbed sample, from which we infer an initial Sr isotope ratio for the intrusion of 0·7017. We have also estimated the initial
Nd for the intrusion from a Sm–Nd mineral isochron for apatites and the least altered clinopyroxene (236/1-03; age = 2809 ± 110 Ma, Nd(t) = +2·0] and from the clinopyroxene and apatite in this sample (+0·7 to + 1· 4, Table 5) leading us to infer an initial Nd of +1· 1 ± 0·4. This result is in reasonable agreement with the Nd-isotopic data for whole-rocks from the Panozero intrusion, which give initial Nd values between 1· 2 and 1· 9 (Kovalenko et al., 2005). Similar Nd values have been also calculated for clinopyroxenes from the syenites, lamprophyres and sanukitoids of the Superior Province (Stern & Hanson, 1991; Hattori et al., 1996).
The initial Sr-isotope ratio for the Panozero magma source is significantly greater than that for the depleted mantle at 2·74 Ga. Kovalenko et al. (2005) showed that Nd-model ages for the mantle source of the Panozero intrusion imply that it is about 60 Myr older than the intrusion. Assuming that the Panozero source evolved from a depleted mantle source over 60 Myr we can calculate that it had an Rb/Sr ratio of 0·12–0·32. This value is an order of magnitude greater than that proposed for depleted mantle (0·012–0·019, DePaolo, 1979) – and the primitive mantle (0·03, Sun & McDonough, 1989).
Carbon and oxygen isotopes
Most of the Panozero rocks contain calcite and/or dolomite and these minerals are particularly abundant in the ocelli found in the early lamprophyres, which contain as much as 40% carbonate by volume. Here we investigate the carbon and oxygen isotopic signatures of these carbonates to determine the extent to which the carbonate was derived from a mantle source (Ionov et al., 1993; Rudnick et al., 1993). Carbon and oxygen isotopic data are presented for carbonates from five samples of the monzonite series and from four lamprophyric ocelli in Table 6 and Fig. 13.
|
|
Carbonates in the monzonite series have
13C between –6·9 and –8·1
PDB and 18O between 9·0 and 11· 0 SMOW. Lamprophyre ocelli have 13C between –5·6 and –6·4 PDB and 18O between 10·4 and 11· 0 SMOW (Lobach-Zhuchenko et al., 2004). These values are much lower than those for Karelian greenstone belt carbonates (Tichomirova, 1991; Fig. 13). The 13C values resemble those of mantle-derived carbonatite carbonates (Fig. 13), although their elevated oxygen isotope ratios suggest that they have experienced some oxygen isotope exchange with high 18O fluids, perhaps during later, low-temperature alteration (Hoernle et al., 2002). Oxygen isotope ratios for zircons from Canadian sanukitoids average 6·4 ± 0·2, higher than the value for trondhjemites from the same province (5·5 ± 0·4, SMOW), suggesting that they are derived from a 18O-enriched source (King et al., 1998). In their study of Archaean shoshonitic lamprophyres from Abitibi, Canada, Wyman & Kerrich (1993) deduced a subducted sedimentary component as the enriched source.
The difference in 13C between carbonates in the lamprophyre ocelli and those in the monzonite series may reflect a difference in the carbon isotopic composition of their sources or, more probably, is the product of isotopic fractionation during magmatic degassing (Javoy et al., 1986).
|
THE ORIGIN OF THE PANOZERO MAGMAS |
|---|
The composition of the parental magmas and their subsequent magmatic evolution
The monzonite series
There is field and petrographic evidence in the Panozero intrusion for large-scale fractional crystallization (e.g. magmatic layering, zoned minerals). For this reason we propose that the the monzonite series magmas were produced by fractional crystallization from a common parent magma (Lobach-Zhuchenko et al., 2007). The process is first modelled using mass-balance calculations (Stormer & Nichols, 1978
) based only on major elements. The algorithm calculates both the modal and chemical compositions of the cumulate that fractionated from the magma and the degree of crystallization. As there are more analysed major elements than mineral phases there are more equations than unknowns and so the system is over-determined. The advantage of the Stormer & Nichols (1978) algorithm approach to over-determined systems is that it uses the least square of the residuals to select a best-fit solution.
A second step consists of reintroducing the computed modal compositions of the cumulate and the degree of fractionation into the trace element model, which uses the Rayleigh (1896) fractionation equation. Mineral compositions are taken from the cumulus phases in the layered mafic rocks and we use the mineral/melt partition coefficients from http:/earthref.sdsc.edu (Table 7). Plagioclase compositions were calculated from CIPW norm calculations.
|
On a K2O–SiO2 plot (Fig. 6b) there are two subparallel trends—a high-K2O trend from mafic rocks to monzonite-1 and a low-K2O trend from mafic rocks to monzonite-2 to monzonite-3 to quartz monzonite. Our modelling calculations are reported in Tables 7–9
. They show that it is possible to produce monzonite-1 from a parental monzogabbro magma by about 44% fractional crystallization of the assemblage diopside (46·3 wt %), phlogopite (31·5 wt %), plagioclase (12·2 wt %), titanomagnetite (7 wt %) with minor amounts of apatite, zircon and allanite (Table 9). The presence of phlogopite indicates the hydrous character of the magma. In a similar way, monzonite-2 could have been produced from a parental monzogabbro magma by about 42% fractional crystallization of diopside (42·9 wt %), phlogopite (35·4 wt %), plagioclase (11· 6 wt %), titanomagnetite (5·9 wt %) with minor amounts of apatite, zircon and allanite. In this case, the greater amount of phlogopite in the fractionate accounts for the lower-K2O character of this series, when compared with monzonite-1. Monzonite-3 could be produced from monzonite-2 by about 24% fractionation (diopside 22·5 wt %, phlogopite 34·5 wt %, plagioclase 30 wt %, titanomagnetite 9·2 wt % and apatite 4·2 wt %; Table 8). Our Pb-isotope data indicate that the petrogenesis of the quartz monzonite involves some crustal contamination and so the calculated fractionation assemblage given in Table 8 is probably an over simplification of the process.
|
|
Trace element behaviour, computed for the mineral assemblages calculated from the major elements, is in good agreement with measured REE and LILE contents in the studied rocks. Models based on the transition metals Cr, Co, V and Ni, even if not in contradiction with them, fit the analytical data less well (Table 9).
We have estimated the composition of the parental magma for the monzonite series from the composition of the layered mafic and ultramafic rocks found as a marginal facies in the first series of monzonites. We assumed that the ultramafic rocks are cumulates derived from parent mafic magmas. The parental magma composition has been calculated assuming that the proportions of ultramafic (30%) and mafic rocks (70%) observed in the field is representative of the initial magma. The main characteristics of this initial magma are as follows: 50·3 wt % SiO2, 3·4 wt % K2O, 8·0 wt % MgO, K2O/Na2O = 1·19, (K/Na)atomic = 0·79 and mg-number = 0·58 (Table 10). This parental melt, transitional between monozogabbro and monzodiorite, is silica-undersaturated (with normative nepheline), strongly enriched in LREE (La = 330 times chondrite) with (La/Yb)n = 24·9 (Fig. 14). It is also enriched in LILE (Rb 106 ppm, Sr 1400 ppm, Ba 2115 ppm, Th 9·6 ppm; Fig. 15). Alternatively, if we assume that the mafic schlieren in monzonite-1 are representative of the parental melt, a slightly more primitive composition is obtained with 46·3 wt % SiO2, 3·4 wt % K2O, 11· 8 wt % MgO, K2O/Na2O = 2·66, (K/Na)atomic = 1·75 and mg-number = 0·60 (Table 10). This melt is strongly enriched in LREE (La = 620 times chondrite) with (La/Yb)n = 31·1. It is also LILE-rich (Rb 108 ppm, Sr 840 ppm, Ba 1049 ppm, Th 13·9 ppm; Figs 14 and 15). In many respects these two, independent estimates of the parental magma for the monzonite series are similar, which confirms that the monzonite series was derived from a mafic magma enriched in LREE and LILE.
|
|
|
The lamprophyres
We have shown above that the Panozero lamprophyres can be divided into two groups—an early ocelli-bearing group and a late group. They have a broad range of compositions (44·4–52·0 wt % SiO2, 0·1–3·4 wt % Na2O, 6·38–19·2 wt % MgO) that form a compositional continuum (Fig. 7g and h). They also have similar primitive mantle-normalized trace element patterns, although their REE patterns differ—the late lamprophyres have less fractionated REE patterns than the early group (Fig. 10). Element correlations suggest that the compositional range within the lamprophyre group is controlled by fractional crystallization. The elements Mg–Cr–Ni–Co–K are positively correlated (e.g. Fig. 7g), as are the elements Al–Na–Sr, although the latter group show a negative correlation with Mg. These relationships suggest that phlogopite and plagioclase are fractionating phases. The element P correlates positively with Nd–Sm–Eu–Gd and weakly with Th, suggesting that REE concentrations are controlled by monazite fractionation. Fractionation trends differ from those in the mafic members of the monzonite series (Fig. 7g and h), but the compositional ranges do overlap.
The compositions of the Panozero lamprophyres appear to be controlled by both the addition and removal of the phases phlogopite, plagioclase and monazite. For this reason we have estimated the composition of the parental lamprophyre magma by taking the average composition of the lamprophyre suite. When recalculated on an anhydrous basis this ‘melt’ has 50·1 wt % SiO2, 13·1 wt % MgO (mg-number = 0·69) and 3·1 wt % K2O. This composition has a similar trace element and REE composition to that estimated for the monzonite series parent magma (Figs 14 and 15).
Fluids in the Panozero magmas
The Panozero magmas contain abundant hydrous minerals from which we infer that they were derived from a fluid-rich mantle source. The presence of primary H2O and CO2-rich gas inclusions in quartz and apatite in the rocks of the monzonite series also supports this view (S. K. Baltybaev, personal communication). In addition, breccia dykes specifically located within the first generation monzonites may be phreato-magmatic (Fig. 2; Lobach-Zhuchenko et al., 2003, 2005a; Rollinson, 2003) and the carbonate-rich ocellar textures in the early lamprophyres support the view that these magma were fluid-rich. The dominant fluid associated with the magmas of the monzonites and late lamprophyres is H2O and minor CO2, whereas the early lamprophyres contain H2O and abundant CO2.
It was shown above that the isotope signature for carbon in the Panozero carbonates is very similar to that of carbonatites and different from that in Archaean carbonate sediments, indicating a mantle origin for the carbon. Carbonatite metasomatism has been reported in some mantle xenoliths from the subcontinental mantle (e.g. Ionov et al., 1993; Lee et al., 2000), and Green & Wallace (1988) have shown that it is possible that carbonatite metasomatism can be responsible for increased LREE and and LILE concentrations in the mantle.
The geochemical, mineralogical and isotopic character of the Panozero mantle source
Major and trace elements
Our inferred primitive melt compositions are given in Table 10, and their trace element characteristics shown in Figs 14 and 15. The dominant character of these melts is that, relative to the primitive mantle, they are strongly enriched in the fluid-mobile elements Ba, Sr, Rb, K, and the LREE (Fig. 15). It was shown above that some rocks have disturbed radiogenic isotope signatures, suggesting that some fluid-mobile trace elements have also experienced post-magmatic alteration. However, the trace element patterns illustrated in Figs 8 and 15 show a broad self-consistency for the elements K, Rb, Sr and Ba, indicating that the high absolute abundances recorded here are real. Primitive mantle-normalized trace element patterns for both the monzonite and lamprophyre parental melts are similar and show negative anomalies for Nb–Ta, P, Zr–Hf and Ti (Fig. 15), suggesting that they originate from a common source.
Calculations based upon the partial melting of a spinel lherzolite-facies primitive mantle source at 12–20 kbar show that the concentrations of Y, Ti and Yb in the Panozero primitive melts can be reproduced by a small fraction of mantle melting (about 5%). In contrast, the concentrations of the elements Ce, Nd, Ba and Sr are higher than those expected from primitive mantle melting, indicating that the REE enrichment was more efficient for LREE than for heavy REE (HREE). We propose that, because LREE are more soluble in fluids than HREE, the mantle source has been enriched in LREE in a similar manner to the sanukitoid mantle source calculated by Shirey & Hanson (1984). Other geochemical characteristics of the Panozero primitive melts could be the result of small amounts of early crystal fractionation, such that the variation in Ni and Cr reflects fractionation of phases such as olivine, spinel and pyroxenes, and the covariation in P and the LREE reflects apatite or monazite fractionation.
Radiogenic isotopes
We have shown that the Panozero source has a low µ (time-integrated U/Pb ratio) and a higher (time-integrated Th/U ratio), relative to primitive mantle. These observations imply that the parental magmas of the Panozero monzonite series came from a mantle source that was enriched in Th relative to U, but depleted in U relative to Pb.
The initial Sr isotope ratio for the Panozero monzonite series source is thought to be 0·7017. This is greater than that for the depleted mantle value at 2·74 Ga. Above, we calculated that such a source would have had an Rb/Sr ratio of 0·12–0·32. These values are higher than those in the calculated parental magmas, suggesting that during magma genesis Rb was retained in the source, perhaps indicating the presence of residual phlogopite.
We have estimated the initial Nd for the monzonite source to be + 1· 1 ± 0·4. This result is about 1· 0 unit lower than that calculated for the depleted mantle by Nagler & Kramers (1998) and implies a Sm/Nd ratio in the source that is lower than in the depleted mantle.
Each of these trace elements ratios inferred from the isotopic character of these rocks—low U/Pb, high Th/U, high Rb/Sr and low Sm/Nd—is consistent with the trace element patterns shown in Fig. 15 and supports the view that these patterns are indicative of derivation of the magmas from an enriched mantle source.
The mineralogical character of the mantle source
Xenoliths derived from the subcontinental mantle lithosphere beneath the Karelian Craton have been described, from Finnish kimberlites, 300 km west of the study area, by Kukkonen & Peltonen (1999). They reported that the principal lithologies are garnet lherzolite and garnet harzburgite containing small amounts (<5%) of clinopyroxene. The phases present in these lithologies (c. 70% olivine, 20% orthopyroxene, 5% clinopyroxene, 5% garnet) cannot host the high levels of K and LILE found in the Panozero magmas, which demonstrates that the mantle beneath the Karelian Craton must be highly variable in mineralogy.
The variation of Rb/Ba vs Ba/Sr in the parental magmas of the Panozero suite, after Hawkesworth et al. (1987) and Pearson et al. (2003), shows that their characteristics can be explained by the presence of both phlogopite and amphibole in the source (Fig. 16). The high chromium content and the high Cr/Ni ratios in the monzonite and lamprophyre parental melts (2·4–3·8) indicate that Cr probably entered the melt through pyroxene melting. The high Y and Yb concentrations in these magmas (4–7 times primitive mantle concentrations) preclude residual garnet. These geochemical observations are consistent with the experimental studies of Conceição & Green (2004) who found that shoshonitic melts can be formed by the dehydration melting of a phlogopite–pargasite lherzolite at c. 1–1·5 GPa and 1050–1150°C. Kovalenko et al. (2005) also confirmed the importance of phlogopite in the Karelian sanukitoid source.
|
Agents of mantle metasomatism
There are three potential agents of metasomatism, capable of producing the enriched mantle source of the Panozero magmas. These include siliceous melts, H2O-rich fluids or melts and carbonate-rich fluids or melts. Siliceous melts, normally of basaltic composition, may react at shallow depths with previously depleted mantle. However, these melts are not noted for being enriched in K or the LILE and are unlikely to the prime agent of mantle metasomatism in the pre-Panozero mantle. However, it has been suggested that silicate melts of adakitic composition react with mantle peridotite to form phlogopite and richterite (Maury et al., 1996
), and this mechanism is thought to be responsible for the genesis of low-silica adakites with Nb–Ta anomalies typical of the Panozero assemblage (Martin & Moyen, 2005; Martin et al., 2005).
In contrast, however, it is well known from studies of arc magmas that fluids released from a subducting slab are hydrous and enriched in alkalis and LILE. This is largely due to the breakdown of hydrous phases in the subducting basaltic crust (Schmidt & Poli, 1998; Scambelluri & Philippot, 2001). The result of this process is to create the phases phlogopite, K-richterite and pargasite in the upper mantle. In addition, carbonatite melts have been reported from many mantle xenoliths. Wyllie et al. (1983) showed experimentally that magmas generated in CO2-bearing mantle below 80 km can contain up to 40 wt % dissolved CO2 , which, as pressure reduces, reacts with other phases to form carbonates. Hence we conclude from the presence of high K, Sr, LREE and CO2 in the Panozero rocks that the main agents of mantle metasomatism are H2O- and CO2-rich fluids, although a contribution from silicic (adakitic, TTG) melts cannot be completely excluded.
|
TECTONIC IMPLICATIONS |
|---|
The geochemical characteristics of the Panozero magmas demonstrate that they are enriched in K, LILE, LREE, H2O and CO2 and are from a metasomatized, enriched mantle source, which melted at 2·74 Ga. This enriched source was created some 60 Myr earlier, at about 2·8 Ga (Kovalenko et al., 2005
). TTG crust and greenstone belts were also formed in this central Karelian region at 2·8 Ga. Thus, we consider that the mantle sampled to form the Panozero magmas was created as subcontinental lithosphere (SCLM) to the Archaean TTG crust (Lobach-Zhuchenko et al., 2005a). The mechanism of mantle enrichment is not certain, although there are similarities between the mafic melts produced here and those produced from modern sub-arc mantle regions. Suprasubduction-zone magmas are enriched in fluid-mobile elements, have a strong Nb–Ta depletion, and have HREE abundances similar to those in mid-ocean ridge basalts. The trace element characteristics of the Panozero primary magmas show many similarities to those of arc magmas—strong fluid-mobile element enrichment, a strong negative Nb–Ta depletion and similar Ti, Y and Yb concentrations. However, they are strongly enriched in the LREE (high Ce/Yb and Sm/Yb ratios) and P relative to most arc lavas. This might reflect the melting of either apatite or monazite in the source, as both these phases have been reported from mantle xenoliths (Rudnick et al., 1993). Support for the idea that modern arc-style processes are responsible for the enrichment of Archaean subcontinental lithosphere is found in the work of Bell et al. (2005), who described a 3·0 Ga metasomatized garnet harzburgite xenolith from beneath the Kaapvaal Craton and likened the metsomatism to that produced by modern subduction-related fluids, and from King et al. (1998), who argued that 18O-enriched zircons in sanukitoids from the Superior Province were derived from a source enriched by the input of high 18O subduction fluids.
The Nd-isotopic evidence presented by Kovalenko et al. (2005) and discussed here shows that the enrichment of the subcontinental mantle occurred at the same time as crust formation, 60 Myr before the melting event from which the highly potassic magmas were generated. A similar chronology has been reported from the Archaean Superior Province by Beakhouse & Davis (2005), who showed that sanukitoids in this area post-date greenstone belt volcanism and TTG plutonism. Experimental studies of shoshonite melting (Conceição & Green, 2004) suggest that the melting event took place at 1–1·5 GPa and 1050–1150°C, in the subcontinental lithospheric mantle. This is well above the P–T conditions recorded by the kimberlite geotherm of Schmidberger & Francis (1999) and above the temperature of TTG melting (Moyen & Stevens, 1995; Rollinson, 1996; Fig 17).
|
The precise cause of this later melting event is the subject of some discussion. Possible mechanisms include the return to equilibrium mantle thermal gradients after the cessation of subduction, or a hot mantle upwelling related either to mantle plume activity or ‘slab break-off’ (Beakhouse & Davis, 2005
). Slab break-off was advocated as an appropriate mechanism for the genesis of high K–Ba–Sr–Zr Palaeozoic granitoids in Scotland (Atherton & Ghani, 2002), and by O’Neill & Wyman (2005) for the generation of late Archaean minette lamprophyres in the southern Superior Province, Canada. Given that there is no evidence for plume-related magmatism at 2·74 Ga in Central Karelia (i.e. no komatiites), and that the melting temperatures are higher than the kimberlite (equilibrium) geotherm, we prefer a model in which an enriched and hydrated subcontinental mantle was heated through mantle upwelling following slab break-off. |
CONCLUSIONS |
|---|
- The Panozero pluton is a high-Mg, high-K central-type intrusion with sanukitoid affinities that formed over three magmatic cycles separated by deformation and dyke intrusion. Lithologies vary in composition from pyroxenite to quartz monzonite. The intrusion is cut by lamprophyre dykes.
- Much of the compositional variation within the monzonite series can be explained in terms of the fractional crystallization of a mafic parent. The last magmas may have experienced some crustal contamination. The compositional variation within the lamprophyre series is also the result of fractional crystallization.
- All the major rock types of the Panozero pluton are enriched in incompatible elements relative to primitive mantle. Calculated parental melt compositions indicate that both the monzonite series and the lamprophyres were derived from a compositionally similar, enriched mantle source.
- The high proportion of hydrous and carbonate minerals in the Panozero magmas indicates that they were derived from a fluid-enriched source. Mantle-like 13C values for Panozero carbonates confirm that the CO2-bearing fluids are mantle derived.
- Nd-isotopes show that the enriched mantle source of the Panozero magmas was created at 2800 Ma, the time of generation of the young greenstone belts and TTGs of the Karelian Province, 60 Myr before the emplacement of the Panozero complex.
- This enrichment event involved the addition of alkalis, LILE, LREE and volatiles (H2O) to a subcontinental mantle source. The pattern of enrichment is similar to that found in modern sub-arc mantle regions and so might have taken place in a subduction setting.
- Results of partial melting experiments are consistent with the derivation of the monzonite series by the melting of a phlogopite–pargasite lherzolite at 1–1· 5 GPa.
- The precise cause of the melting event at 2·74 Ga is not known, although the hot upwelling of asthenospheric mantle following slab break-off is consistent with the geochemical evidence for the enrichment of the Karelian subcontinental mantle by subduction fluids.
|
SUPPLEMENTARY DATA |
|---|
Supplementary data for this paper are available at Journal of Petrology online.
|
ACKNOWLEDGEMENTS |
|---|
This research was supported by INTAS grant 2001-0073, a Royal Society–NATO fellowship to A. Kovalenko and a grant from the Russian Academy of Science (Program N4 of the Earth Sciences Department of the RAS). We are grateful to Valery Ivanikov for discussions and for some rock and mineral analyses, to Gennady Artemenko for help with the fieldwork, and to Michael Tolkachev for help with the microprobe analyses. We thank Steve Foley, Simon Johnson and Derek Wyman for helpful comments on an earlier version of the manuscript.
*Corresponding author. Telephone: +968 24141444. E-mail: hrollin{at}squ.edu.om
|
REFERENCES |
|---|
Atherton MP, Ghani AA. Slab breakoff: a model for Caledonian, late granite syn-collisional magmatism in the orthotectonic (metamorphic) zone of Scotland and Donegal, Ireland. Lithos (2002) 62:65–78.[CrossRef][Web of Science]
Beakhouse GP, Davis DW. Evolution and tectonic significance of intermediate to felsic plutonism associated with the Helmo greenstone Belt, Superior Provimce, Canada. Precambrian Research (2005) 137:61–92.[CrossRef][Web of Science]
Bell DR, Gregoire M, Grove TL, Chatterjee N, Carlson RW, Buseck PR. Silica and volatile-element metasomatism of Archaean mantle: a xenolith-scale example from the Kaapvaal craton. Contributions to Mineralogy and Petrology (2005) 150:251–267.[CrossRef][Web of Science]
Bibikova E, Petrova A, Claesson S. The temporal evolution of sanukitoids in the Karelian Craton, Baltic Shield: an ion microprobe U–Th–Pb isotopic study of zircons. Lithos (2005) 79:129–145.[CrossRef][Web of Science]
Blundy JD, Holland JB. Calcic amphibole equilibria and a new amphibole–plagioclase geothermometer. Contributions to Mineralogy and Petrology (1990) 104(2):208–224.[CrossRef][Web of Science]
Bogatikov OA, Ryabchikov ID, Kononova VA, eds. Lamproites (1991) Moscow: Nauka. 301. (in Russian).
Chekulaev VP, Levchenkov OA, Ivanikov VV, Arestova NA, Kovalenko AV, Gooseva NS, Komarov AN. Composition, an age and Sm–Nd systematics of the sanukitoids of the Panozero massif. Geochimiya (2003) 8:817–828. (translated into English).
Conceição RV, Green DH. Derivation of potassic (shoshonitic) magmas by decompression melting of pholgopite + pargasite lherzolite. Lithos (2004) 72:209–229.[CrossRef][Web of Science]
DePaolo DJ. Inferences from correlated Nd and Sr isotopic variations for the chemical evolution of the crust and mantle. Earth and Planetary Science Letters (1979) 43:201–211.[CrossRef][Web of Science]
Elliott T, Plank T, Zindler A, White W, Bourdon B. Element transport from slab to volcanic front at the Mariana arc. Journal of Geophysical Research (1997) B102:14991–15019.[CrossRef]
Glebova-Kulbakh GO, Pinaeva NI, Lobach-Zhuchenko SB. Granites of South Karelia. In: Granites of Kola Peninsula and Karelia—Polkanov AA, ed. (1963) Leningrad: Academy of Science Press. 161–315. (in Russian).
Green DH, Wallace ME. Mantle metasomatism by ephemeral carbonatite mantle. Nature (1988) 336:459–462.[CrossRef]
Halla J. Late Archean high-Mg granitoids (sanukitoids) in southern Karelian Domain, eastern Finland: Pb and Nd isotopic constraints on crust–mantle interactions. Lithos (2005) 79:161–178.[CrossRef][Web of Science]
Hattori K, Hart SR, Shimizu N. Melt and source mantle compositions in the Late Archaean: study of strontium and neodymium isotope and trace elements in clinopyroxenes from shoshonite alkaline rocks. Geochimica and Cosmochimica Acta (1996) 60:4551–4562.[CrossRef]
Hawkesworth CJ, van Calsteren P, Rogers NW, Menzies MA. Isotope variations in recent volcanics: a trace-element perspective. In: Mantle Metasomatism—Menzies MH, Hawkesworth CJ, eds. (1987) New York: Academic Press. 365–388.
Hoernle K, Tilton G, Le Bas MJ, Duggen S, Garbe-Schönberg D. Geochemistry of oceanic carbonatites compared with continental carbonatites: mantle recycling of oceanic crustal carbonate. Contributions to Mineralogy and Petrology (2002) 142:520–542.[Web of Science]
Ionov A, Dupuy C, O’Reilly SY, Kopylova MG, Genshaft YS. Carbonated peridotite xenoliths from Spitsbergen: implication for trace element signature of mantle carbonate metasomatism. Earth and Planetary Science Letters (1993) 119:283–297.[CrossRef][Web of Science]
Ionov DA, Bodinier J-L, Mukasa SB, Zanetti A. Mechanisms and sources of mantle metasomatism: major and trace element compositions of peridotite xenoliths from Spitsbergen in the context of numerical modelling. Journal of Petrology (2002) 43:2219–2259.
Ivanikov VV. Archaean syenites and monzonites of Karelia. Vestnik SPbU (1997) 7(4):3–15. (in Russian).
Jacobsen SB, Wasserburg GJ. Sm–Nd evolution of chondrites and achondrites. Earth and Planetary Science Letters (1984) 67:137–150.[CrossRef][Web of Science]
Javoy M, Pineau F, Delorme H. Carbon and nitrogen isotopes in the mantle. Chemical Geology (1986) 57:26–41.
Johnson MC, Rutherford MJ. Experimental calibration of the aluminium in hornblende geobarometer with application to Long Valley caldera (California) volcanic rock. Geology (1989) 17:837–841.
Keller J, Hoefs J. Stable isotope characteristics of recent natrocarbonatites from Oldoinyo Lengai. In: Carbonatite Volcanism: Oldoinyo Lengai and the Petrogenesis of Natrocarbonatites—Bell K, Keller J, eds. (1995) Berlin: Springer. 113–123.
Kerrich R, Watson DP. The Macassa Mine Archean lode gold deposit, Kirkland Lake, Ontario; geology, patterns of alteration, and hydrothermal regimes. Economic Geology (1984) 79:1104–1130.
King EM, Valley JW, Davis DW, Edwards GR. Oxygen isotope ratios of Archean plutonic zircons from granite–greenstone belts of the Superior Province: indicator of magmatic source. Precambrian Research (1998) 92:365–387.[CrossRef][Web of Science]
Kogarko LN, Henderson CMB, Pacheco H. Primary Ca-rich carbonatite magma and carbonate–silicate–sulfide liquid immiscibility in the upper mantle. Contributions to Mineralogy and Petrology (1995) 121:267–274.[Web of Science]
Kovalenko A, Clemens JD, Savatenkov V. Petrogenetic constraints for the genesis of Archaean sanukitoid suites; geochemistry and isotopic evidence from Karelia, Baltic Shield. Lithos (2005) 79:147–160.[CrossRef][Web of Science]
Kramers JD, Tolstikhin IN. Two terrestrial lead isotope paradoxes, forward transport modeling, core formation and the history of the continental crust. Chemical Geology (1997) 139:75–110.[CrossRef][Web of Science]
Kukkonen IT, Peltonen P. Xenolith-controlled geotherm for the central Fennoscandian Shield: implications for lithosphere–asthenosphere relations. Tectonophysics (1999) 304:301–315.[CrossRef][Web of Science]
Lafleche MR, Dupuy C, Dostal J. Archaean orogenic ultrapotassic magmatism: an example from the southern Abitibi greenstone belt. Precambrian Research (1991) 52:71–96.[CrossRef][Web of Science]
Leake BE, et al, (Commission on New Minerals and Mineral Names). Nomenclature of amphiboles: report of the Subcommittee on Amphiboles of the International Mineralogical Association, Commission on New Minerals and Mineral Names. Canadian Mineralogist (1997) 35:219–246.[Web of Science]
Lee CT, Rudnick RL, McDonough WF, Horn I. Petrologic and geochemical investigation of carbonates in peridotite xenoliths from northeastern Tanzania. Contributions to Mineralogy and Petrology (2000) 139:470–484.[CrossRef][Web of Science]
Le Roex AP, Lanyon R. Isotope and trace element geochemistry of Cretaceous Damaraland lamprophyres and carbonatites, northwestern Namibia: evidence for plume–lithosphere interaction. Journal of Petrology (1998) 39:1117–1146.
Lobach-Zhuchenko SB, Chekulaev VP, Arestova NA, Kovalenko AV, Ivanikov VV, Guseva NS, Rollinson HR. High-Mg granitoids (sanukitoids) of the Baltic Shield—geological setting, geochemical characteristics and implications for origin of mantle-derived melts. In: EAE03-A-03744; VGP7-1FR1P-0628 (2003) 494. Geophysical Research Abstracts, 5, 3097.
Lobach-Zhuchenko SB, Lochov KI, Prasolov EM. Isotopic composition of carbon and oxygen in carbonates of the Panozero Pluton (Central Karelia). In: A. P. Vinogradov XVII symposium on isotope geochemistry, 6-9 December 2004, Moscow. (Abstract) (2004) 149–150. (in Russian).
Lobach-Zhuchenko SB, Rollinson HR, Chekulaev VP, Arestova NA, Kovalenko AV, Ivanikov VV, Guseva NS, Sergeev SA, Matukov DI, Jarvis KE. The Archaean sanukitoid series of the Baltic Shield: geological setting, geochemical characteristics and implications for their origin. Lithos (2005a) 79:107–128.[CrossRef][Web of Science]
Lobach-Zhuchenko SB, Chekulaev VP, Krylov IN, Kovalenko AV, Guseva NS. The Archaean automagmatic breccias of the Panozero Pluton, Central Karelia, Baltic Shield. Doklady RAS (2005b) 401(2):1–5. (in English).
Lobach-Zhuchenko SB, Rollinson H, Chekulaev VP, Guseva NS, Arestova NA, Kovalenko A. Geology and petrology of the Archaean high-K and high-Mg Panozero massif, Central Karelia. Petrology (2007) 15:459–487.[CrossRef][Web of Science]
Ludwig KR. User's Manual for a Computer Program for Processing Pb–U–Th Isotope Data. Berkeley Geochronology Center, Open-file Report. (1993) 88–542.
Ludwig KR. User's Manual for Isoplot/Ex, Version 2.2, A Geochronological Toolkit for Microsoft Excel. Berkeley Geochronology Center, Special Publications. (2000) 1a.
Martin H, Moyen J-F. The Archaean–Proterozoic transition: sanukitoid and Closepet type magmatism. Mineralogical Society of Poland—Special Papers (2005) 26:57–68.
Martin H, Smithies RH, Rapp R, Moyen J-F, Champion D. An overview of adakite, tonalite–trondhjemite–granodiorite (TTG) and sanukitoid relationships and some implications for crustal evolution. Lithos (2005) 79:1–24.[CrossRef][Web of Science]
Maury RC, Sajona FG, Pubellier M, Bellon H, Defant MJ. Fusion de la croute océanique dans les zones de subduction/collision récentes: l’exemple de Mindanao (Philippines). Bulletin de la Société Géologique de France (1996) 167:579–595.[Abstract]
Middlemost EAK. Naming materials in the magma/igneous rocks system. Earth-Science Reviews (1994) 37:215–224.
Miyashiro A. Nature of alkalic volcanic rock series. Contributions to Mineralogy and Petrology (1978) 66:91–104.[CrossRef][Web of Science]
Morrison GW. Characteristics and tectonic setting of the shoshonite rock association. Lithos (1980) 13:97–108.[CrossRef][Web of Science]
Moyen J-F, Stevens G. Experimental constraints on TTG petrogenesis: implications for Archaean geodynamics. In: Archaean Geodynamics and Environments. American Geophysical Union, Geophysical Monograph—Benn K, Mareschal J-C, Condie K, eds. (2005) 164:141–175.
Nagler TF, Kramers JD. Nd isotopic evolution of the upper mantle during the Precambrian: models, data and the uncertainty of both. Precambrian Research (1998) 91:233–252.[CrossRef][Web of Science]
O’Neill C, Wyman D. Geodynamic modelling of late Archean subduction: P–T constraints from greenstone belt diamond deposits. In: Archean Geodynamics and Environments. American Geophysical Union, Geophysical Monograph—Benn K, Mareschal J-C, Condie KC, eds. (2005) 164:177–188.
Pearce JA. Trace element characteristics of lavas from destructive plate boundaries. In: Andesites—Thorpe RS, ed. (1982) New York: John Wiley. 525–548.
Pearson DG, Canil D, Shirey SB. Mantle samples included in volcanic rocks: xenoliths and diamonds. In: Treatise on Geochemistry, 2.05—Holland HD, Turekian KK, eds. (2003) Amsterdam: Elsevier. 171–275.
Peccerillo A, Taylor SR. Geochemistry of Eocene calc-alkaline volcanic rocks from the Kastomonou area, northern Turkey. Contributions to Mineralogy and Petrology (1976) 58:63–81.[CrossRef][Web of Science]
Rayleigh JWS. Theoretical considerations respecting the separation of gases by diffusion and similar processes. Philosophical Magazine (1896) 42:77–107.
Rocholl A, Jochum KP. Th, U and other trace elements in carbonaceous chondrites: implications for the terrestrial and solar system Th/U ratios. Earth and Planetary Science Letters (1993) 117:265–278.[CrossRef][Web of Science]
Rock NMS. Lamprophyres (1991) Glasgow: Blackie. 285.
Rollinson HR. Tonalite–trondhjemite–granodiorite magmatism and the genesis of the Lewisian crust during the Archaean. In: Precambrian Crustal Evolution in the North Atlantic Region. Geological Society, London, Special Publications—Brewer TS, ed. (1996) 112:25–42.
Rollinson HR. Magma mingling in the Panozero sanukitoid intrusion, Baltic Shield. Geophysical Research Abstracts (2003) 5:3065.
Rudnick RL, McDonough WF, Chappel BW. Carbonatite metasomatism in the northern Tanzanian mantle: petrographic and geochemical characteristics. Earth and Planetary Science Letters (1993) 114:463–475.[CrossRef][Web of Science]
Scambelluri M, Philipott P. Deep fluids in subduction zones. Lithos (2001) 55:213–227.[CrossRef][Web of Science]
Schmidberger SS, Francis D. Nature of the mantle roots beneath the North American Craton: mantle xenolith evidence from Somerset Island kimberlites. Lithos (1999) 48:195–216.[CrossRef][Web of Science]
Schmidt MW, Poli S. Experimentally based water budgets for dehydrating slabs and consequences for arc magma generation. Earth and Planetary Science Letters (1998) 163:361–379.[CrossRef][Web of Science]
Shirey S, Hanson GN. Mantle-derived Archaean monzodiorites and trachyandesites. Nature (1984) 310:222–224.[CrossRef]
Sobolev AV, Shimizu N. Ultra-depleted primary melt included in an olivine from the Mid-Atlantic Ridge. Nature (1993) 363:151–154.[CrossRef]
Stacey JS, Kramers JD. Approximation of terrestial lead isotope evolution by two-stage model. Earth and Planetary Science Letters (1975) 26:207–221.[CrossRef][Web of Science]
Stern RA, Hanson GN. Archaean high-Mg granodiorite; a derivation of light rare earth element-enriched monzodiorite of mantle origin. Journal of Petrology (1991) 32:201–238.
Stevenson R, Henry P, Gariepy C. Assimilation–fractional crystallization origin of Archean sanukitoid suites: Western Superior Province, Canada. Precambrian Research (1999) 96:83–89.[CrossRef][Web of Science]
Stormer JC, Nichols J. XLFRAC, a program for the interactive testing of magmatic differentiation models. Computers and Geosciences (1978) 4:143–159.[CrossRef]
Sun S, McDonough WF. Chemical and isotopic systematic of oceanic basalts: implications for mantle composition and processes. In: Magmatism in the Ocean Basins. Geological Society, London, Special Publications—Saunders AD, Norry MJ, eds. (1989) 42:313–345.
Tichomirova M. Variation of oxygen and carbon in carbonates of sedimentary and metamorphic in origin. In: Ph.D. thesis (1991) Leningrad. 31.
Todt W, Cliff RA, Hanser A, Hoffman AW. Evaluation of a 202Pb–205Pb double spike for high precision lead isotope analysis. In: Earth Processes: Reading the isotopic code. Geophysical monograph. American Geophysical Union—Basu A, Hart SR, eds. (1996) 95:429–437.
Turner S, Arnand N, Lin J, Rogers N, Hawkesworth C, Harris N, Kelley S, Van Calsteren P, Deng W. Post-collision shoshonitic volcanism of the Tibetan Plateau: implications for convective thinning of the lithosphere and the source of ocean island basalts. Journal of Petrology (1996) 37:45–71.
Van Orman JA, Grove TL, Shimizu N. Rare element diffusion in diopside: influence of temperature, pressure and ionic radius, and an elastic model for diffusion in silicates. Contributions to Mineralogy and Petrology (2001) 141:687–703.[Web of Science]
Veizer J, Hoers J, Ridler RH, Jensen LS, Lowe DR. Geochemistry of Precambrian carbonates: I. Archean hydrothermal systems. Geochimica et Cosmochimica Acta (1989) 53:845–857.[CrossRef][Web of Science]
Wyllie PJ, Huang WL, Otto J, Byrnes AP. Carbonation of peridotites and decarbonation of siliceous dolomites represented in the system CaO–MgO–SiO2–CO2 to 30 kbar. Tectonophysics (1983) 100:359–388.[CrossRef][Web of Science]
Wyman DA, Kerrich R. Archaean shoshonitic lamprophyres of the Abitibi sub-province, Canada: petrogenesis, age and tectonic setting. Journal of Petrology (1993) 34:1067–1109.
Wyman DA, Ayer AJ, Conceição RV, Sage RP. Mantle processes in an Archaean orogen: Evidence from 2·67 Ga diamond-bearing lamprophyres and xenoliths. Lithos (2006) 89:300–328.[CrossRef][Web of Science]
Zartman RE, Doe BR. Plumbotectonics—the model. Tectonophysics (1981) 75:135–162.[CrossRef][Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||