Journal of Petrology Advance Access originally published online on January 9, 2007
Journal of Petrology 2007 48(3):589-625; doi:10.1093/petrology/egl074
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The Origin and Evolution of the Kaapvaal Cratonic Lithospheric Mantle
1Falw, Vu Amsterdam, De Boelelaan 1085, 1081Hv Amsterdam, Netherlands
2Department of Terrestrial Magnetism, Carnegie Institution of Washington, 5241 Broad Branch Road N.W., Washington, D.C. 20015, USA
3Department of Earth Sciences, Durham University, South Road, Durham DH1 3LE, UK
RECEIVED MARCH 15, 2006; ACCEPTED NOVEMBER 17, 2006
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONPETROGRAPHIC DESCRIPTION AND...ANALYTICAL METHODSRESULTSDISCUSSIONMODEL FOR THE FORMATION...SUPPLEMENTARY DATAREFERENCES |
|---|
A detailed petrological and geochemical study of low-temperature peridotite xenoliths from Kimberley and northern Lesotho is presented to constrain the processes that led to the magmaphile element depletion of the Kaapvaal cratonic lithospheric mantle and its subsequent re-enrichment in Si and incompatible trace elements. Whole-rocks and minerals have been characterized for Re–Os isotope compositions, and major and trace element concentrations, and garnet and clinopyroxene for Lu–Hf and Sm–Nd isotope compositions. Most samples are characterized by Archaean Os model ages, low Al, Fe and Ca contents, high Mg/Fe, low Re/Os, very low (< 0·1 x chondrite) heavy rare earth element (HREE) concentrations and a decoupling between Nd and Hf isotope ratios. These features are most consistent with initial melting at
3·2 Ga followed by metasomatism by hydrous fluids, which may have also caused additional melting to produce a harzburgitic residue. The low HREE abundances of the peridotites require that extensive melting occurred in the spinel stability field, possibly preceded by some melting in the presence of garnet. Fractional melting models suggest that 30% melting in the spinel field or 20% melting in the garnet field followed by 20% spinel-facies melting are required to explain the most melt-depleted samples. Garnet Nd–Hf isotope characteristics indicate metasomatic trace element enrichment during the Archaean. We therefore suggest a model including shallow ridge melting, followed by metasomatism of the Kaapvaal upper mantle in subduction zones surrounding cratonic nuclei, probably during amalgamation of smaller pre-existing terranes in the Late Archaean (2·9 Ga). The fluid-metasomatized residua have subsequently undergone localized silicate melt infiltration that led to clinopyroxene ± garnet enrichment. Calculated equilibrium liquids for clinopyroxene and their Hf–Nd isotope compositions suggest that most diopside in the xenoliths crystallized from an infiltrating kimberlite-like melt, either during Group II kimberlite magmatism at 200–110 Ma (Kimberley), or shortly prior to eruption of the host kimberlite around 90 Ma (northern Lesotho). KEY WORDS: Kaapvaal craton; lithospheric mantle; metasomatism; Nd–Hf isotopes; Re–Os isotopes
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONPETROGRAPHIC DESCRIPTION AND...ANALYTICAL METHODSRESULTSDISCUSSIONMODEL FOR THE FORMATION...SUPPLEMENTARY DATAREFERENCES |
|---|
Considerable progress has been made in the understanding of craton structure and history in recent years by integrating geophysical and geochemical data. Tomographic studies (e.g. James & Fouch, 2002
) and Os isotope analysis of mantle xenoliths (e.g. Walker et al., 1989; Pearson et al., 1995a; Carlson et al., 1999; Menzies et al., 1999; Irvine et al., 2001) from the Kaapvaal craton of southern Africa have confirmed and extended models for a thick, cold, lithospheric mantle root (e.g. Boyd & Nixon, 1973; Boyd & McCallister, 1976; Boyd & Gurney, 1986) by delineating the craton boundaries at depth and confirming the Archaean age of much of this 200 km thick keel. The mechanisms responsible for the formation, consolidation and modification of the cratonic lithosphere remain more controversial (Jordan, 1978; De Wit et al., 1992; Griffin et al., 2003b; Schmitz et al., 2004). Perhaps the most important feature of Archaean continental lithospheric mantle (CLM) is that its major element composition indicates that it is the residue of extensive partial melt extraction. Compared with melt residua produced in experiments, however, many samples of ancient CLM, and in particular the peridotites of the Kaapvaal and Siberian cratons, have unusually high Si/Mg and associated high modal orthopyroxene (opx) contents. The cause of the silica enrichment in Kaapvaal and Siberian low-temperature (low-T: < 1150°C) peridotites remains contentious. One model attributes Si enrichment to melting-related processes including addition of cumulate opx at high pressure and temperature in plumes rising from the deep mantle during the initial formation of cratons (Herzberg, 1999; Griffin et al., 2003a). The competing explanation is that the CLM peridotites are residua of low-pressure melting and that the Si enrichment is a signature of later enrichment events (e.g. Kesson & Ringwood, 1989; Rudnick et al., 1994; Kelemen et al., 1998).
To better understand the formation and modification processes of continental lithosphere, we present new whole-rock Re–Os isotope and whole-rock and mineral major and trace element and Nd and Hf isotope data for 26 low-T peridotite xenoliths from Kimberley in the centre of the western Kaapvaal craton (Fig. 1). These data are compared with a similar dataset for 19 well-characterized xenoliths from northern Lesotho (Irvine et al., 2001; Simon et al., 2003; Pearson et al., 2004) to address the following questions.
- How was the Kaapvaal CLM originally formed? If samples represent residua from melt extraction, when did the melting happen, how much melt was extracted, and under what temperature and pressure conditions did melting occur?
- How was the CLM modified over time and what process produced the present-day high modal opx contents of the CLM?
|
|
PETROGRAPHIC DESCRIPTION AND CLASSIFICATION |
|---|
|
TOPABSTRACTINTRODUCTIONPETROGRAPHIC DESCRIPTION AND...ANALYTICAL METHODSRESULTSDISCUSSIONMODEL FOR THE FORMATION...SUPPLEMENTARY DATAREFERENCES |
|---|
The Kimberley xenoliths were collected from the Old Boshoff Road dump (Bultfontein kimberlite) close to Kimberley, South Africa (Fig. 1). Details of northern Lesotho samples have been given by Irvine et al. (2001
), Irvine (2002) and Simon et al. (2003), who presented the whole-rock major and trace element compositions and modes of these samples. Sample selection focused on fresh samples thought to be representative of the bulk of the Kaapvaal lithospheric mantle; coarse-grained spinel harzburgites and relatively garnet- and clinopyroxene-poor lherzolites (Boyd & McCallister, 1976; Rudnick et al., 1998; James et al., 2004). A few samples that show direct evidence of modal metasomatism, i.e. that contain abundant coarse amphibole and/or phlogopite or veins or clusters of clinopyroxene (cpx) ± hydrous phases, were also collected to cover the more extreme end of the compositional spectrum. Detailed petrographic descriptions are given in Table 1 and photomicrographs can be found in Electronic Appendix 2 online at http://www.petrology.oxfordjournals.org.
|
Mineral modes for the Kimberley xenoliths were calculated from X-ray fluorescence (XRF) whole-rock major element and electron microprobe mineral data using a least-squares method (Table 2), following the approach of Boyd & Mertzman (1987
). Only olivine, opx, cpx, garnet and/or spinel were included in the calculation of modes. Residuals (root mean square; RMS) for calculated modes are usually small (< 0·05, Table 2), except for samples rich in minerals other than those used for the regression (e.g. mica and amphibole).
|
Samples are grouped based on modal mineralogy (Fig. 2) and opx, garnet and cpx morphology and grain-size distribution. This is to emphasize the differences between the samples and facilitate the discussion of the geochemical results by referring to the most extreme ‘end-members’.
|
Type I. This group is composed of refractory harzburgites and lherzolites. It consists of coarse granoblastic to porphyroclastic garnet- and spinel-lherzolites and -harzburgites that generally are texturally equilibrated. The garnet-free samples of this type have the lowest cpx + spinel + garnet contents (< 3%) and are characterized by equigranular–interlobate to mosaic olivine crystals, intergrown with varying amounts of opx and minor amounts of other phases [see Boyd et al. (1999
) for a detailed description of similar spinel-facies peridotites]. Opx crystals range in shape from interstitial vermicular (K2, K9) to centimetre-size tabular grains (K20), often showing exsolution lamellae and spinel–opx or spinel–cpx symplectites on opx rims. Generally, the size of opx increases with increasing modal opx. Some opx are poikilitic, enclosing round to ovoid olivine grains, and can have vermicular extensions, giving them a melt-like appearance. Samples may show minor deformation and contain traces of hydrous minerals.
Garnet-bearing Type I samples have variable modal mineralogies (Table 2, Fig. 2) and some contain minor phlogopite and/or amphibole (< 1%). Large (millimetre- to centimetre-size) interlobate olivine and opx coexist with 
4% cpx and 10% garnet. Opx vary in shape from large subhedral tabular grains to interlobate, vermicular and poikilitic (including olivine, garnet and cpx) crystals. Garnet grains display isometric–subhedral to ovoid shapes and often appear to be spatially associated with opx (on opx rims, surrounded by or included in opx). Round garnet grains sometimes have concave embayments, probably indicating resorption. Cpx can also be round to ovoid or have very irregular vermicular shapes.
Type II. Defined by garnet- and/or cpx-rich samples K3, K4, K19 and K21, these samples are coarse lherzolites like Type I samples, but are not texturally equilibrated. They contain much more cpx + spinel + garnet (>20%) than the other types. Cpx alone constitutes more than 12% in K21 and K3 (Fig. 2). Garnet (especially in K19) and cpx (in K21 and K4) form clusters and chains of irregularly shaped, vermicular grains, surrounding and overgrowing other minerals. Garnet and cpx grain sizes approach or even exceed those of olivine and opx (< 0·5 mm). K3 is a cpx-rich lherzolite, which is very distinct compared with all other samples. Mineral grains in K3 contain abundant inclusions of silicates and opaque phases (mainly spinel and ilmenite). K3 resembles xenoliths of the IRPS (Ilmenite–Rutile–Phlogopite–Sulphide) suite described from the Matsoku kimberlite pipe in northern Lesotho (Harte et al., 1987).
Type III. This type is characterized by higher abundance of hydrous minerals and/or alteration compared with the other types. Some Type III samples appear to be metasomatized former Type I rocks and are roughly equivalent to the phlogopite peridotites (PP) or phlogopite–K-richterite peridotites (PKP) of Erlank et al. (1987).
All samples contain tiny (<10 µm) cpx, phlogopite, amphibole and spinel crystals in cracks and along grain boundaries. Garnets usually have thin (maximum 0·1 mm) kelyphite rims. Samples may contain other late-stage secondary and/or alteration products such as calcite, barite, sulphides, etc. These features can be related to the infiltration of host kimberlite during eruption, or to weathering and alteration of the samples at the surface. The effects of these late-stage processes are clearly distinguishable from pre-eruption metasomatism.
|
ANALYTICAL METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONPETROGRAPHIC DESCRIPTION AND...ANALYTICAL METHODSRESULTSDISCUSSIONMODEL FOR THE FORMATION...SUPPLEMENTARY DATAREFERENCES |
|---|
XRF, solution inductively coupled plasma mass spectrometry (ICP-MS) and Re–Os isotope analyses were performed on fractions of 100–500 g (depending on hand-specimen size) powdered bulk-rock. Electron microprobe (EPMA), ion probe (secondary ionization mass spectrometry; SIMS) and laser ablation (LA) ICP-MS analyses were carried out in situ on polished thick sections. Typically 1–4 thick sections were prepared per sample, depending on sample homogeneity.
XRF analyses for major elements, Sr, Zr, Cr and V were carried out on 0·4 g of rock powder by S. Mertzman at the Department of Geology, Lancaster, PA, USA. The procedures for the XRF analyses, ferrous iron titrations and loss on ignition (LOI) determinations have been described by Boyd & Mertzman (1987).
Whole-rock trace element concentrations were determined by ICP-MS on 100 mg of dissolved sample using the HP4500 Plus system at the Vrije Universiteit (VU) Amsterdam. A 10 ppb indium spike was added as an internal standard. The solutions were analysed up to three times and results averaged. A blank was measured at the beginning of each session. Abundance determinations were carried out by peak-height comparison with a solution of BHVO-2 (US Geological Survey Certificate of Analysis, 1998). Drift corrections were made by recalculating the measured values to literature-defined values (Eggins et al., 1997). MAG-1 (USGS standard sediment) was used as an internal standard. Most elements are reproducible to better than 5% and errors in inter-element ratios are better than 2%. The ICP-MS system was calibrated before the start of each session. Errors on sample analyses based on counting statistics (1
) are smaller than 10% for light rare earth elements (LREE), high field strength elements (HFSE), Pb, Th, U, Ba, Cs, Y and Sr, and usually <50% for heavy REE (HREE). The 1 standard deviations on HREE can increase to >100% for very low concentration samples such as Type I spinel harzburgite K6.
Mineral major element compositions were determined by electron microprobe at the Vrije Universiteit Amsterdam using a Jeol JXA-8800 system following Simon et al. (2003). Mineral compositions presented in Electronic Appendix Table A (http://www.petrology.oxfordjournals.org) represent averages obtained from 5–20 individual analyses.
Trace element contents in garnet, cpx and opx were measured by LA-ICP-MS at the University of Utrecht (most samples) or by SIMS (K18) at the Department of Terrestrial Magnetism, Washington, DC (DTM). Analytical procedures have been described in detail by Simon et al. (2003). Generally no trace element zonation could be detected. The concentrations are therefore average values of several ablation pits (1–5) on different grains (1–5). Where core and rim compositions differ significantly (e.g. in some opx), mineral core and rim compositions are presented separately (Table 3). Opx trace element analyses were complicated by a lack of fresh opx, exsolution lamellae, amphibole or phlogopite films on cleavage planes, and low trace element concentrations. We carefully selected spots for analysis and monitored the changes in concentration of critical elements (Ba, Sr, Nb, Ca) during ablation. Nevertheless, time-resolved analyses occasionally showed spikes in the signal intensities of these elements during analysis. Analyses showing these spikes have been eliminated, to minimize the effect of included phases on the opx analyses. Several trace element abundances in Kimberley opx are very close to, or even below, detection limits. Consequently, the observed variations in middle REE (MREE) and HREE (Eu–Lu), Ba, Y and sometimes Hf in opx are not statistically significant. LREE and large ion lithophile element (LILE) abundances generally exceed analytical errors.
|
Re–Os isotope analyses were performed on
1 g aliquots of whole-rock powder using procedures described previously (Carlson et al., 1999). For Nd and Hf isotope analyses, clean and transparent garnets and cpx from selected samples were separated from coarse crush by handpicking. In garnets where trace element zonation was observed we tried to avoid garnet rims by selecting larger fragments surrounded entirely by fresh conchoidal fractures. Minerals were washed in an ultrasonic bath for 15 min in 5 ml of 4N HCl followed by addition of 0·5 ml of concentrated HF and ultrasonication for another 5 min. The separates were then washed repeatedly in deionized water, re-examined under the binocular microscope and re-picked if necessary. Mineral dissolution and element separation were carried out using techniques described by Carlson et al. (2004). Nd and Hf separation and analysis of a subset of samples were performed at the University of Durham by Geoff Nowell (see Dowall et al., 2003; Pearson & Nowell, 2006, for analytical techniques). Sm and Nd isotope analyses on some samples were also carried out by thermal ionization mass spectrometry at the VU Amsterdam. Several mineral separates were analysed in both laboratories (DTM and VU) as a measure of inter-laboratory consistency. All mineral isotope data and the respective techniques and analysis locations are summarized in Table 4.
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONPETROGRAPHIC DESCRIPTION AND...ANALYTICAL METHODSRESULTSDISCUSSIONMODEL FOR THE FORMATION...SUPPLEMENTARY DATAREFERENCES |
|---|
Pressure–temperature conditions and depth of sampling
Pressures and temperatures were calculated for core and rim compositions of constituent minerals using a range of geothermometers and barometers (Table 5). The different thermometers agree within error for most garnet-bearing Type I samples [for an extensive review on mantle geothermobarometry, see Smith (1999
)] and yield temperatures between 900 and 1100°C at 3–4·5 GPa. With the exception of GP402, they fall on a P–T trend consistent with the average Kalahari geotherm of Rudnick & Nyblade (1999; Fig. 3). Type II samples K19 and K21 show considerable scatter in P–T estimates (Table 5) and tend to have lower equilibrium pressures and temperatures [850°C and 3·2 GPa for the combination of P and T from Al-in-opx barometry (Brey & Köhler, 1990) and garnet–olivine Fe–Mg thermometry (O'Neill & Wood, 1979)]. This might indicate disequilibrium between phases in these two samples, which is supported by textural evidence and trace element systematics (see below). The Ca-in-opx and Al-in-opx temperature ranges for Type I spinel peridotites (945–1100°C) overlap with the temperature range obtained for garnet peridotites, whereas Mg–Fe exchange thermometers give much lower temperatures (Table 5). This phenomenon was previously observed by Boyd et al. (1999) for Kimberley and Simon et al. (2003) for Lesotho spinel peridotites. Despite the extensive geochemical evidence for kimberlite modification of the peridotites discussed below, the mineral major element data appear to record coherent P–T information for Type I garnet peridotites, probably because of relatively rapid diffusion of the major elements on which the thermobarometers are based, relative to the slow diffusivities of trivalent trace elements (e.g. Ganguly et al., 1998; Van Orman et al., 2001, 2002).
|
|
Major and trace element compositions of the constituent minerals
A detailed description of the Kimberley mineral compositions is provided in Electronic Appendix 1 (available online at http://www.petrology.oxfordjournals.org). Representative major element data for Kimberley samples are presented in Electronic Appendix Table A and the most salient aspects of the data are discussed below. Mineral compositions for northern Lesotho xenoliths were discussed by Simon et al. (2003
).
Olivine
Olivines in the Kimberley samples have generally homogeneous compositions with mean and median Mg-number [=molar (Mg/(Mg + Fe)) x 100] of 92·4, within error of average cratonic mantle (median 92·1, Pearson et al., 2003).
Orthopyroxene
The opx are enstatites, with mean and median Mg-number (93·2 and 93·3, respectively) usually slightly higher than in the corresponding olivines (Electronic Appendix Table A). The majority of samples show minor compositional variations, restricted to the outermost rims (50 µm). Type I opx are more magnesian (Mg-number 93–94) and Type II are the least Mg-rich (90–92).
Opx have two distinct trace element patterns that seem to be unrelated to the petrographic types (Fig. 4a): (1) sinusoidal REE patterns characterized by depletion of LREE and HREE over MREE; (2) patterns with maxima at La–Ce and La/LuN > 1. The first pattern is more common, but the second occurs within the same sample and even within the rim of the same opx grain (e.g. in K17). Both opx patterns are depleted in all REE relative to chondrite, particularly in HREE (< 0·05). Sm/NdN varies between 0·5 in opx with LREEN maxima between La and Nd, and is up to 1·4 in opx with a REEN maximum at Sm (Fig. 4a, Table 3). Interestingly, those opx with super-chondritic Sm/NdN also have positive Hf anomalies (Fig. 5a). All opx show a strong positive Nb–Ta anomaly and a negative Y anomaly, and variable positive and negative Zr, Hf and Ti anomalies (Fig. 5a).
|
|
Clinopyroxene
Cpx porphyroblasts usually have homogeneous cores, with Type I samples having the highest and Type II the lowest Mg-number (93–96 and 87–92, respectively; Electronic Appendix Table A).
All cpx are significantly LREE enriched (Table 3, Fig. 4b, LaN = 10–80) and HREE depleted, leading to subchondritic Sm/Nd (Sm/NdN = 0·4–0·8) and Lu/Hf (Lu/HfN = 0–0·13) ratios. The cpx REE patterns of Type I samples are similar, especially in LREE concentrations (LaN = 10–20), and have extremely low HREE (LuN 0·1). They have humped patterns with maxima at Pr–Nd [(30–80) x C1]. In contrast, Type II samples K19 and K21 have cpx with REE maxima at Ce (CeN 80). They also have the highest LREE, Ba, Th and Nb–Ta and HREE concentrations of all cpx. Sr, Zr and Hf concentrations are superchondritic in all cpx (Fig. 5b). All cpx have negative HFSE anomalies, but Type II and III samples have higher Zr–Hf and Ti contents.
Garnet
Mg-number in garnets (80·4–87·8, mean = 84) correlate positively with Cr-number (6·4–15·5, mean = 12). The lowest Cr2O3 (< 3·1 wt% Cr2O3), highest Al (>21·5 wt% Al2O3) and relatively low CaO contents in garnet (<4·6 wt%) are found in Type II samples K3, K19 and K21 (and GP402), consistent with their differing textures and pyroxene major element chemistry (Electronic Appendix Table A). Only K3 (Type II) and K18 (Type I, deformed, high Cr-number and Mg-number) contain garnets with a significant amount of TiO2 and Na2O (0·3–0·4 wt% and 0·05–0·06 wt%, respectively). Garnets from GP402 and K15 are the only ones from Kimberley that have below 4·2 wt% CaO. All others have clearly lherzolitic Ca–Cr relations (Sobolev et al., 1973). No significant zonation in major elements was observed in Kimberley garnets, with the exception of deformed sample K18.
Type I garnets have REE patterns that increase markedly from La to Nd (La/NdN 0·05–0·015, Fig. 4c). The patterns are characterized by a maximum between Nd and Eu [(6·5–18) x C1], a decrease in concentrations from Eu to Dy–Tm and Dy–Tm/LuN < 1, resulting in sinusoidal REE patterns (Table 3, Fig. 4c). Type II display very distinct LREE-depleted (LaN = 0·1–0·3) and HREE-enriched patterns (Nd/LuN 0·2–0·08).
Garnet incompatible element patterns normalized to chondrite have low Ba and Sr, but high Hf and Zr (Fig. 5c). Type I and Type I–II garnets have strong negative Ti anomalies, whereas the anomaly is less pronounced in Type II garnets. All garnets except K1, K3 and K17 have negative Hf–Zr (and Ti) anomalies and Zr and Hf concentrations similar to, or lower than, a primitive mantle garnet (Simon, 2004). All samples except GP402 (Sm/NdN = 0·85) have superchondritic Sm/Nd ratios. Type I garnets (except GP402gt: Lu/HfN = 2·7) have Lu/HfN << 1, whereas Type II garnets (and GP402gt) have (Lu/Hf)N > 1.
Spinel
Primary spinels are magnesiochromites (Mg-number = 40–65, Cr-number 50) with homogeneous compositions and very low Ti contents (Electronic Appendix Table A). Primary spinels coexisting with garnet are generally more Cr-rich (e.g. Boyd & Nixon, 1975; Boyd et al., 1999).
Phlogopite and amphibole
Primary, texturally equilibrated phlogopite porphyroblasts have Mg-number varying from 93 to 95. TiO2 contents vary from 0·05 to 1·7 wt% (Electronic Appendix Table A).
Amphiboles are K-richterites with variable major element compositions (Electronic Appendix Table A). Mg-number varies from 87 to 95. TiO2 varies from 0 to 1·35, K2O from 0·35 to 4·6 and Na2O from 2·9 to 5·2 wt%. Amphibole was analysed for trace elements in Type III sample K11 (Table 3). The amphibole is LREE enriched (LaN = 16) with La/LuN 40, has a small positive Eu anomaly, is strongly enriched in Rb relative to Ba and Th (RbN = 300, Rb/ThN 100) and has positive Sr and HFSE anomalies, and HfN > ZrN and NbN > TaN. K11 amphibole has subchondritic Sm/NdN and Lu/HfN ratios (0·56 and 0·02, respectively), but a high Rb/Sr ratio compared with primitive mantle (Rb/SrPM = 7·2).
Whole-rock compositions of the xenoliths
Major and minor elements
Type I xenoliths are extremely depleted in magmaphile elements (Ca, Al, Na, Ti) and have high MgO contents and therefore Mg-number (Table 2). FeO contents are generally lower than in oceanic peridotites at a given Ca or Al, but typical for Kaapvaal low-T peridotites (e.g. Boyd, 1989). They are also characterized by low Mg/Si ratios, placing the Kimberley samples well above the oceanic trend of Boyd (1989; Fig. 6a). Type I and III samples have Cr2O3 contents within the range 0·20–0·44 wt%, with the garnet-bearing Type I samples having slightly higher Cr2O3 than the Type I spinel peridotites. Type II samples are variably enriched in CaO and/or Al2O3 (Fig. 6b), have the highest Cr2O3 (up to 0·6 wt%), and also tend to have lower Mg-number in olivine. In all samples, Cr2O3 increases with Al2O3 and decreases with MgO.
|
Although minor element contents might be affected by kimberlite contamination, most samples have very low TiO2 and Na2O contents (below 0·03 and 0·1 wt%, respectively). Elevated TiO2 and Na2O contents (up to 0·3 wt%) are found only in some Type II and III samples.
Trace elements
Measured whole-rocks show steep primitive mantle normalized (PM) incompatible element patterns. They exhibit strong enrichment in the most incompatible elements with a maximum at BaPM of 9, Ba/Lu 10, LuPM 0·1–1 and positive Nb and negative Hf and Ti anomalies (Table 6, Fig. 7). Whole-rock trace element contents reconstructed from mineral compositions and modes (Fig. 8) show markedly lower incompatible element concentrations and very distinct patterns compared with the measured whole-rock values, which reflects contamination of the xenolith by the extremely incompatible element-rich host magma (Hawkesworth et al., 1983; Boyd & Mertzman, 1987; Schmidberger & Francis, 2001; Pearson & Nowell, 2002; Grégoire et al., 2003). These reconstructed patterns do not include phlogopite and amphibole, because the modal abundance of these phases cannot be estimated accurately from either thin section or whole-rock major element concentrations. In addition, it is assumed that the incompatible trace element concentration of olivine is zero, because for many of these elements, the concentration was below detection by the analytical means employed. The lack of inclusion of phlogopite and amphibole may explain the large troughs in Ti concentration in the reconstructed patterns, and deficiencies in other HFSE such as Nb and Ta. The 30–180 ppm Ti concentrations measured in olivine would result in less than a doubling of the calculated whole-rock Ti concentration if included, and thus would not significantly erase the large Ti troughs observed in the reconstructed whole-rock patterns (Fig. 8). For other incompatible trace elements, olivine analyses by ion-probe showed abundances below detection limits, which implies that the contribution of olivine to the whole-rock abundances of these elements is minimal.
|
|
|
The reconstructed REE concentrations show several distinct patterns that are not always coherent with the classification derived from petrography and major element chemistry (Fig. 8). Most Type I whole-rock REE patterns are sinusoidal in shape, similar to the REE patterns of the garnets in these samples. Even though all samples show variable enrichment of MREE over HREE, all Type I rocks possess a positive slope in their HREE patterns (LuN > ErN). K21 is most enriched in LREE (CeN = 10). K3 has a REE pattern similar to those displayed by cpx. This reflects the dominance of this mineral in the bulk-rock trace element budget in these rocks. K21 and K3 are the only samples that are enriched in trace elements relative to primitive mantle (Hofmann, 1988
).
Re–Os, Sm–Nd and Lu–Hf isotopes
Whole-rock Re–Os isotope systematics
187Os/188Os in the Kimberley peridotites range from 0·1073 to 0·1189 (Table 7), which overlaps values reported previously for Kimberley (Carlson et al., 1999; Griffin et al., 2004). Type I samples generally have low kimberlite eruption age-corrected 187Os/188Os (maximum 0·1106, average 0·1086), translating into Archaean TRD model ages (Fig. 9) with a mean TRD of 2·80 ± 0·13 Ga [Re-depletion ages as defined by Walker et al. (1989)]. Conventional Re–Os model ages (TMA) show some scatter (Fig. 9; Table 7), mainly as a result of Re addition during entrainment of the xenoliths in the kimberlite (Simon et al., 2003). However, the median TMA (mantle model age) of 3·1 Ga approaches the 3·2 Ga age of the oldest basement rocks in the western Kaapvaal (Schmitz et al., 2004). Some samples from the metasomatized Types II and III also have Archaean TRD model ages, but others have younger ages, the youngest being 1·7 and 1· 9 Ga (Table 7). These samples, in general, also have higher Re contents (Table 7).
|
|
Garnet and cpx Sm–Nd and Lu–Hf isotope systematics
Kimberley garnets and cpx have Nd isotope compositions (
Nd ranges from –13·5 to 0·3 with a median of –4·0; Table 4, Fig. 10) generally lower than, but overlapping the low Nd end of the range displayed by South African Group I kimberlites (Smith, 1983; Nowell et al., 2004). All Kimberley cpx have substantially subchondritic Sm/Nd (mean 147Sm/144Nd = 0·095), whereas the garnets show a range in 147Sm/144Nd, but mostly superchondritic values (0·137–0·574, mean 0·353). The large difference in Sm/Nd, but the relatively similar Nd isotope compositions of the garnet and cpx from the same sample result in generally young two-point cpx–garnet Sm–Nd ‘ages’ for the Kimberley samples ranging from negative to 202 Ma (Table 8). Compared with the Kimberley samples, garnets from well-characterized Type I xenoliths from northern Lesotho (Simon et al., 2003) generally show much lower Sm/Nd ratios. Nd isotope compositions for the Lesotho cpx cluster in a relatively narrow range just slightly more radiogenic than those from Kimberley (Nd from –3·1 to 2·3, mean = 0·4) and fall completely within the field displayed by South African Group I kimberlites (Smith, 1983; Nowell et al., 2004; Fig. 10). The Lesotho cpx Nd isotope compositions show no systematic variation with Sm/Nd ratio or Nd contents. In contrast, Lesotho garnets span a wide range in Nd isotope compositions (Nd from –36·2 to 2·4, mean = –7·9), but are generally below chondritic (Fig. 10). Four out of seven of the Lesotho garnet–cpx isochrons provide negative ‘ages’ (Table 8).
|
|
In comparison with the generally subchondritic Nd isotope compositions, the Kimberley and Lesotho cpx and garnets tend towards superchondritic Hf isotope compositions (Table 4, Fig. 10). Lesotho garnets show a relatively restricted range in Hf isotope compositions with four out of five samples having
Hf between 12·8 and 22·8, well above the –6 to 7 range displayed by South African Group I kimberlites (Nowell et al., 2004). Kimberley garnets show a much wider range in Hf isotope compositions (Hf from –2·7 to 345, Fig. 10) that is accompanied by a wide range in Lu/Hf ratios. As a result, four of the Kimberley garnets provide depleted mantle Hf model ages between 0·87 and 1· 04 Ga, whereas two samples provide Archaean ages (GP402: 2·85 Ga; K27: 3·3 Ga; Table 4). In contrast, three out of five Lesotho garnets provide Hf model ages that are either negative or near zero, whereas only one sample provides an old model age (M9: 1·73 Ga). Hf isotope compositions in cpx scatter within the range observed for South African Group I kimberlites (Nowell et al., 2004), with similar median Hf values of 6·6 for Kimberley and 3·7 for Lesotho. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONPETROGRAPHIC DESCRIPTION AND...ANALYTICAL METHODSRESULTSDISCUSSIONMODEL FOR THE FORMATION...SUPPLEMENTARY DATAREFERENCES |
|---|
The trace element and Nd–Hf systematics of the xenolith minerals clearly are not consistent with these samples being residua of a single stage of melt extraction. The incompatible element enrichment and low Nd and Hf isotope compositions of the cpx reflect the well-known propensity of CLM to have experienced metasomatic enrichment in these elements (e.g. Menzies & Murthy, 1980
; Hawkesworth et al., 1983; Menzies & Hawkesworth, 1987; Griffin et al., 1999b). To determine how the Kaapvaal CLM was initially formed it is necessary first to understand how subsequent metasomatic events modified the CLM composition. We therefore explore the temporal evolution of the CLM by progressively stripping away the effects of metasomatism to provide a better estimate of the composition of the material left from the initial melting event that created the residual nature of the CLM.
Numerous studies have documented evidence for multiple metasomatic events affecting cratonic lithosphere [see the review by Pearson et al. (2003); and specifically for the Kaapvaal the studies by Carlson et al. (1999) and Griffin et al. (1999a, 2004)]. The large-scale geological events recorded in the Kaapvaal craton that may have resulted in metasomatism in the CLM include: (1) major crust formation in the Palaeoarchaean (e.g. Eglington & Armstrong, 2004); (2) Mesoarchaean accretion of the western and eastern Kaapvaal blocks, accompanied by subduction of the eastern Kaapvaal underneath the western Kaapvaal block at 2·9 Ga (Drennan et al., 1990; Schmitz et al., 2004); (3) Ventersdorp magmatism at 2·7 Ga (e.g. Armstrong et al., 1991); (4) accretion of the surrounding Proterozoic belts in the Mesoproterozoic (e.g. Hartnady et al., 1985); (5) Karoo and/or Type II kimberlite magmatism at 200–110 Ma (e.g. Hamilton et al., 1998; Konzett et al., 2000); (6) Group I kimberlite magmatism at 90 Ma (Allsopp & Barrett, 1975).
Interaction with kimberlite during entrainment and transport of the xenoliths
All xenoliths show evidence for infiltration of kimberlitic material [see petrography section and Simon et al. (2003)], as is reflected in the whole-rock trace element data, which show up to 10 times enrichment in incompatible elements relative to primitive mantle in comparison with the reconstructed whole-rock trace element patterns, which are markedly incompatible element depleted for all Type I and III samples (Table 6, Figs 7 and 8). The discrepancy between measured and reconstructed trace element contents of CLM xenoliths has been known for some time (e.g. Hawkesworth et al., 1983) and was most recently documented in cratonic mantle xenoliths from Siberia (Pearson & Nowell, 2002), Somerset Island (Schmidberger & Francis, 2001) and Kaapvaal (Pearson & Nowell, 2002; Grégoire et al., 2003). For the Kimberley samples studied here, there is relatively good agreement between measured and reconstructed whole-rock compositions for moderately incompatible elements (D > Hf), whereas all the measured LREE, LILE and Nb and Ta contents are five (Nd) to several hundred times (Ba) higher than the recalculated concentrations (Table 6, Figs. 7 and 8). Simple mass-balance mixing calculations show that the difference between calculated and measured trace element concentrations can be accounted for by 1·5% addition of an average Group I kimberlite for most elements, similar to estimates for Somerset Island (Schmidberger & Francis, 2001). The contribution of the kimberlite to the whole-rock budget for some of the highly incompatible elements (Ba, Rb, Th, Nb, Nd, Hf) can exceed 90% (e.g. Grégoire et al., 2003) and thus control whole-rock Nd and Hf isotope compositions. In contrast, 1·5% addition of a kimberlite with 1 ppb Os (e.g. Carlson et al., 1996) and 187Os/188Os = 0·13 to a residual peridotite with 3 ppb Os and 187Os/188Os = 0·107 would increase the measured whole-rock 187Os/188Os to only 0·1071, reducing the calculated Re-depletion age by only 10 Myr. Thus, kimberlite contamination is unlikely to have a significant effect on the whole-rock Re depletion ages measured for peridotite xenoliths.
Evidence for chemical disequilibrium between minerals within individual peridotites would provide unequivocal evidence that the peridotites had been modified by processes that occurred during or shortly before the entrainment by the host lavas. The mineral compositions of the Kimberley Type I samples provide temperature and pressure estimates using a variety of thermobarometers (Table 5) that plot along the ‘Kalahari geotherm’ determined for a wide variety of peridotite xenoliths from South Africa (e.g. Rudnick & Nyblade, 1999). This is suggestive of the attainment of major element chemical equilibrium between the phases while these rocks were resident at depth in the Kaapvaal CLM. Similarly, trace element partition coefficients between Kimberley garnet and cpx porphyroblast cores (Dgt/cpx) were calculated and compare well with equilibrium values (e.g. Zack et al., 1997). In contrast to the results from the northern Lesotho xenoliths (Simon et al., 2003), garnet and cpx in all Kimberley samples, except Type II samples K19 and K21, seem to be in trace element equilibrium. K19 and K21 (and to a lesser extent also K3), however, have Dgt/cpx for LREE, MREE and HFSE that are about three times lower than reference equilibrium values, and also show a range in pressure and temperature estimates from different thermobarometers (Table 5). These features all suggest that the minerals within these Type II samples were not in equilibrium at the time they were brought to the surface.
Evidence that complete equilibration was not achieved in many of the samples comes from the observation that tie-lines connecting the garnet and cpx Sm–Nd data for seven samples (Table 8) provide ages younger than the time of kimberlite eruption (86 Ma: Allsopp & Barrett, 1975; Davis, 1977), including a number of samples that give negative ages. For example, K27 gives a negative Sm–Nd garnet–cpx age because the K27 garnet has markedly lower 143Nd/144Nd than the K27 cpx despite the higher Sm/Nd of the garnet. Negative Sm–Nd garnet–cpx ages are not uncommon for peridotite xenoliths (Günther & Jagoutz, 1991; Carlson et al., 2004) and were first described in samples from Kimberley (Richardson et al., 1985).
The Lu–Hf system has previously been shown to be more robust towards metasomatic processes and probably has a higher closure temperature than the Rb–Sr and Sm–Nd systems, particularly in garnet (Schmidberger et al., 2002; Bedini et al., 2004; Pearson & Nowell, 2004). Nevertheless, samples K1, K13 and M9 yield negative Lu–Hf garnet–cpx ages whereas the ages for the other seven garnet–cpx pairs range from 659 to 1995 Ma (Table 8). There is no clear correlation between equilibration temperature and Lu–Hf ages. Indeed, the samples with the highest equilibration temperatures yield the highest garnet–cpx Lu–Hf isochron and garnet Hf depleted mantle model ages, which is the opposite of what would be expected if the age differences are due solely to cooling through a Lu–Hf closure temperature as suggested by Bedini et al. (2004). As with Nd, the Hf depleted mantle model ages for the majority (21 out of 24) of the cpx range from 0·5 to 1·5 Ga, overlapping the model ages of Group I kimberlites (Smith, 1983; Nowell et al., 2004). This overlap, the negative garnet–cpx ages, the calculated trace element compositions of liquids that could be in equilibrium with the cpx in Lesotho xenoliths (Simon et al., 2003), and the identification of trace element zonation in garnet grains in the Lesotho samples (Simon et al., 2003), all suggest that the Nd–Hf systems in these minerals have been affected by recent chemical exchange with a metasomatic agent that had trace element and Nd–Hf isotope compositions similar to those of Group I kimberlites (Simon et al., 2003; Simon, 2004) as has been suggested previously (Carswell, 1975; Shimizu, 1999; Van Achterbergh et al., 2001; Grégoire et al., 2002, 2003; Simon et al., 2003).
Evidence for localized silicate melt metasomatism from Type II peridotites
Type II samples record some melt depletion in that they retain high Mg-numbers in olivine, but have been extensively modified by incompatible element addition and growth of new, metasomatic silicate phases such as cpx and garnet. This is evident from their major element variations compared with melt residua, their trace element concentrations and their petrographic appearance. K19 and K21 have garnet and cpx that are not in textural and chemical equilibrium with olivine and opx, which suggests that the metasomatic event occurred relatively close in time to kimberlite eruption. Type II sample K3 has a distinct texture, is much more Fe-rich and Mg-poor than all other samples, has a very high cpx content (19%) and contains minerals of the IRPS suite. This sample clearly has been extensively melt-modified and is suitable as an end-member for Fe + Ca addition to the lithosphere. Using the distribution coefficients given in Electronic Appendix Table B2, calculated melts in equilibrium with cpx and garnet in Type II samples K3, K19 and K21 have distinctively higher HREE and Ti contents than the Type I equilibrium melts (Fig. 11), indicating reaction with a silicate melt. The melts in equilibrium with Type II cpx have markedly higher incompatible element concentrations than the garnet melts (Fig. 11b), confirming chemical disequilibrium in these samples.
|
Type II samples also possess more radiogenic Os (but generally lower Os contents; Table 7) than Type I rocks, and their Os isotope compositions correlate positively with CaO and Al2O3 contents (Fig. 12). Types II and III, in general, have higher Re contents than the Type I samples (Table 7), consistent with Re addition during metasomatism. The Type II samples also are shifted away from the opx (SiO2)–187Os/188Os trend shown by Type I samples (Fig. 13), indicating a distinct metasomatic history for these samples. All of these characteristics suggest that the Type II samples were affected by interaction with mafic, instead of kimberlitic or carbonatitic, melts.
|
|
The origin of orthopyroxene introduction and garnet enrichment
To constrain the nature of possible metasomatic agents that led to incompatible element enrichment of the garnets and introduction of opx, we calculated the trace element compositions of hypothetical liquids in equilibrium with opx, garnet and cpx in the xenoliths. The range in published Dmin/melt values is large, especially for Dpx/melt (Electronic Appendix Table B2), but the major consequence of this variation on the calculated patterns is to shift similar-shaped patterns up and down in Fig. 11. In most samples, the calculated liquids in equilibrium with opx (Fig. 11a) and garnet (not shown) are within error of one another. In the samples where opx cores and rims were analysed (e.g. K17, LET64), the less enriched cores overlap with the garnet liquid compositions, whereas the more LREE-enriched rims usually match the cpx equilibrium liquids (Simon, 2004
). The liquids in equilibrium with garnet and opx (cores) display strong enrichment of Ba and LREE (LaPM up to >1000) over HREE (YbPM = 0·02–0·3), positive Nb and Ti, and negative Hf (and, less pronounced, Zr) anomalies (Fig. 11a). These characteristics are incompatible with equilibration with normal mafic melts. The high Ba, Nb and LREE might indicate re-equilibration with a kimberlitic melt, whereas MREE and HREE contents are too low to be in equilibrium with a Group I or II kimberlite. One alternative explanation is that a kimberlitic melt or fluid that percolates through a garnet-bearing peridotite will have its HREE preferentially removed by partial equilibration with garnet, leaving the residual fluid depleted in these elements through the process of chromatographic fractionation (Godard et al., 1995). Another alternative notes that the LREE to HREE ratio of the hypothetical opx liquids fits remarkably well those of the hydrous fluid determined experimentally by Stalder et al. (1998) in equilibrium with eclogite-assemblage minerals (garnet, clinopyroxene, and rutile) at 900–1200°C and 3·0–5·7 GPa (Electronic Appendix Table B2). The positive Ti and negative Hf anomalies in the calculated liquids are pronounced, but also agree with the experimental fluid composition reported by Stalder et al. (1998). More recent experiments on hydrous mid-ocean ridge basalt (MORB) at 4 and 6 GPa and 700–1200°C, including a larger range of trace elements, yield similar fluid compositions, but also show that there is a continuous transition between aqueous fluids and hydrous silicate melts at these conditions (Kessel et al., 2005). The fluids or melts of the experiments of Kessel et al. (2005) are characterized by a negative Y anomaly, a feature that is also shown by many of the opx from Lesotho (Simon et al., 2003) and Kimberley (Fig. 5a) and the hypothetical melts that were calculated to be in equilibrium with those opx (Fig. 11a). The strongly positive Nb anomaly shown by the melts that would be in equilibrium with opx, however, can only be accounted for by the involvement of kimberlites.
If the LREE re-enrichment of the samples was due to an aqueous fluid, this fluid must also have enriched the CLM in other fluid-mobile elements such as LILE. Ti is assumed to be immobile in fluids, whereas the HFSE content of the fluid depends strongly on the presence or absence of HFSE-rich minerals, such as rutile, in the subducting slab (e.g. Stalder et al., 1998; Foley et al., 2000). The garnets in samples least affected by recent metasomatism (those with the lowest Nd: K27, M13, LET2) tend to have high LREE, Rb and Sr, and low Hf (high Zr/Hf), Ti and HREE contents (and high LREE/HFSE and low HREE/HFSE ratios), relatively high Cr-number, low FeO, high Sm/Dy, high Nb and low whole-rock 
Osi (–17 to <–14). The FeO and Ti contents in the garnets increase with decreasing LREE/HREE fractionation, hence the least LREE-enriched garnets have the highest Ti and FeO concentrations. Garnets also have less sinusoidal REE patterns (lower Sm/Dy) with increasing FeO and TiO2 (Fig. 14) and show positive correlations between LREE, LILE and Cr2O3, but a negative correlation between LREE and HREE (Fig. 15). Thus, it appears that the process that led to the LREE enrichment of garnets also leads to enrichment in LILE and a relative depletion in HFSE and possibly HREE. Ti, FeO and Cr2O3 contents in the garnets from Type I xenoliths are apparently not affected by this process, whereas Ti and FeO increase and Cr2O3 decreases in the garnets from samples that experienced metasomatism by a mafic or kimberlitic melt (most pronounced in Type II garnets). Similar characteristics (extreme depletion in HREE accompanied by enrichment in LREE relative to a residue of dry melting) have been described by Bizimis et al. (2000) for ophiolitic peridotites from the Hellenic Peninsula and were interpreted by those workers to reflect fluid-fluxed melting and simultaneous highly incompatible element enrichment in a supra-subduction zone setting.
|
|
The above chemical trends are consistent with previous findings by other workers (e.g. Pearson et al., 1995b
; Griffin et al., 1999b; Shimizu et al., 1999; Zhang et al., 2003), but are interpreted differently here. Whereas previous studies related the LREE enrichment and sinusoidal REE patterns of garnets to metasomatism by carbonatitic melts (e.g. Pearson et al., 1995b; Griffin et al., 1999b), or the melts that formed the ilmenite–rutile–phlogopite–sulphide veins in the mantle (Zhang et al., 2003), we support the suggestion that the observed geochemical characteristics are most consistent with metasomatism by a hydrous fluid (e.g. Bizimis et al., 2000; Stachel et al., 2004; Bell et al., 2005). These fluids might have also caused some additional fluid-fluxed melting (e.g. Bizimis et al., 2000), probably in Archaean subduction zones. This conclusion is supported by oxygen isotope data for garnets from polymict xenoliths. Zhang et al. (2003) found distinctly lower 
18O (< 5) in dark-coloured garnets with sinusoidal REE patterns than in light-coloured garnets with ‘normal’, HREE-enriched patterns, from the same samples. Peridotitic garnet inclusions in diamonds also show a relatively wide range of 18O (Lowry et al., 1999) extending to lower values (4·5) than those typical for peridotite xenoliths (18O = 5–5·5; Mattey et al., 1994), and Pearson et al. (2003) showed that low-Ca garnets in general have low 18O. These lower than normal 18O values might indicate metasomatism of the sinusoidal garnets by slab-derived metasomatic agents, whereas the garnets with low LREE/HREEN were formed or overprinted by reaction with mantle-derived melts such as alkali basalts or kimberlites.
A slab-derived hydrous fluid was discussed as a potential source for crystallization of opx in the mantle wedge (Kesson & Ringwood, 1989; Smith et al., 1999; Morishita et al., 2003; Bell et al., 2005) because in addition to the trace element characteristics discussed above it could be a carrier of Si and Mg into the mantle without addition of many other major elements (such as Al, Ca, Fe, etc.). Si and Mg have been shown experimentally to be soluble in a hydrous fluid under high pressures (Stalder et al., 2001; Mibe et al., 2002). In addition, hydrous fluids might be able to migrate over long distances through the mantle (Mibe et al., 1998), making fluids more suitable agents for introduction of Si into the lithosphere than highly siliceous slab melts, because Si-rich melts from subducted crust are likely to react instantaneously with the surrounding mantle and form opx-rich layers at the interface between eclogite and peridotite (Yaxley & Green, 1998). Si-rich melts such as trondjhemites also have significant Al contents, and this led Carlson et al. (1999) to reject such granitoid melts as likely metasomatic agents (Fig. 12).
Timing of orthopyroxene introduction and ancient metasomatism
The spinel and garnet peridotites (Type I) and most phlogopite and phlogopite–amphibole peridotites (Type III) follow a steep trend of increasing 187Os/188Os90 Ma with increasing modal orthopyroxene (Fig. 13). The garnet and/or cpx-rich samples (Type II) show a shallower trend of moderately increasing orthopyroxene with rapidly increasing 187Os/188Os90 Ma. The trends overlap at 187Os/188Os = 0·106, which is equivalent to a TRD of 3·2 Ga, the same age as the oldest crustal rocks in the western Kaapvaal. This convergence in isotopic composition suggests that the samples were modified from the same primary composition, an orthopyroxene-poor, olivine-rich harzburgite, and that the olivine-rich parental rock experienced melt extraction in the Palaeoarchaean. The trend of increasing 187Os/188Os90 Ma with increasing modal orthopyroxene (a proxy for SiO2; Fig. 13) is not accompanied by increases in Re/Os (Table 7), suggesting that the increase in 187Os/188Os reflects the addition of radiogenic Os through interaction with the fluid or melt that deposited the opx. A key feature that constrains both the composition of the fluid or melt involved and possibly also the tectonic setting of its genesis is that the Type I samples do not show increasing CaO, TiO2 or Re concentrations with increasing orthopyroxene and 187Os/188Os. These features are similar to those displayed by modern xenoliths from subduction zones (Brandon et al., 1996; Widom et al., 2003), where the metasomatic agent appears to be an Os-rich hydrous fluid (e.g. Brandon et al., 1996; McInnes et al., 1999; Widom et al., 2003; Becker et al., 2004). The 187Os/188Os of the metasomatic agent in Type I samples is poorly constrained by the data, but the most radiogenic of these samples has a 187Os/188Os of 0·1105, compared with a fertile mantle composition at 2·9 Ga of 0·108. An estimate of the Os isotope composition of subduction-derived fluids present beneath Kimberley at 2·9 Ga can be derived from the initial Os isotope composition of eclogite-paragenesis sulphides found in Kimberley diamonds of 187Os/188Os = 0·156 (Richardson et al., 2001). If the metasomatic agent for Type I samples had an Os isotope composition similar to the initial ratio of the diamond inclusion sulphides, then to raise the Os isotope composition of a 3·2 Ga residual peridotite with 3 ppb Os and 187Os/188Os of 0·1055 to 187Os/188Os = 0·1105 would require fluid/rock ratios of 4 and 0·7 for fluid Os concentrations of 0·1 and 1 ppb, respectively.
The Os data do not tightly constrain the timing of the opx addition, but they are suggestive of a Palaeoarchaean protolith for the Kaapvaal CLM, in agreement with numerous previous studies (Walker et al., 1989; Pearson et al., 1995a; Carlson et al., 1999; Carlson & Moore, 2004; Griffin et al., 2004). We have made the argument above that the opx addition was associated with the addition of LREE and LILE, in which case the Sm–Nd and Lu–Hf systems should provide chronological information on this event. Unfortunately, extracting chronological information from the Nd and Hf data is compromised by the obvious evidence for recent interaction of the samples with highly incompatible element-rich melts discussed previously. The most compelling evidence for ancient metasomatism of the Kaapvaal CLM is provided by the very low 143Nd/144Nd and Palaeoarchaean model ages for diamond inclusion garnets from Finsch and Kimberley (Richardson et al., 1984). Of the samples studied here, those that are considered to be least affected by interaction with the kimberlite (K27, M13 and LET2) have garnets with 143Nd/144Nd approaching the low values measured for Kimberley diamond inclusion garnets, but the >4·5 Ga model ages of K27gt and M13gt indicate that even these samples have inconsistent Sm–Nd isotope relationships, probably reflecting later metasomatic events and re-equilibration with (late metasomatic) cpx. Of these three samples, in part because of the very low Hf contents of their garnets, we were able to obtain Hf isotope data only for K27, where a depleted mantle model age of 3·3 Ga was obtained. This age supports a Palaeoarchaean origin for this sample, as does the 2·85 Ga Hf model age of GP402 garnet, but these are the only two samples in the dataset that provide Archaean Hf model ages. An important observation for the K27 garnet data, however, is that, for a garnet, this sample has a relatively low Lu/Hf ratio, and hence its model age probably dates the time of LILE enrichment, not the original depletion event. For example, in comparison with the relatively low 176Hf/177Hf (Hf = 36·6) measured for K27 garnet, a garnet in a Palaeoarchaean strongly melt-depleted residue would have Hf of many tens of thousands (Simon, 2004).
Depth and extent of the initial melting of the protolith
Provided that the effects of metasomatic mineral formation and compositional modification can be adequately accounted for, it is possible to constrain the original partial melting environment of the Kaapvaal mantle using the major and trace element systematics of the CLM peridotites. The major element and HREE compositions are most severely affected by the metasomatic event that led to the crystallization of opx. This process probably involved complex melt– or fluid–rock interaction, but we do not have enough constraints to adequately model this event. We therefore use a simplified approach of bulk addition of opx. Because of the low abundances of HREE in opx, opx addition does not significantly affect the whole-rock HREE budget. Trace element modelling was performed to evaluate the depth and extent of melting experienced by the Kaapvaal Type I samples. We used perfect fractional polybaric non-modal melting of a pyrolitic source adapted after Johnson et al. (1990) and Hellebrand et al. (2002). As all the peridotites studied here are variably enriched in strongly incompatible elements by metasomatic processes, we model the HREE of the least metasomatized Type I samples. To do so, we assume that the positive slope in the HREE and the low HREE concentrations in these samples are relicts from the original melting event (e.g. Kelemen et al., 1998). The results from the modelling are shown together with the reconstructed REE patterns of garnet-facies Type I samples and the measured REE pattern of K6 (spinel-facies Type I) in Fig. 16. The extremely low HREE contents of Type I samples require a significant proportion of melting to take place in the absence of garnet. The model shows that the HREE concentrations of K24, or an average Type I sample, can be reproduced by >25% melting in the spinel stability field. However, we cannot exclude the possibility that some melting took place in the garnet field, followed by 15–20% (or more) melting in the spinel stability field. Garnet-free Type I samples (e.g. K6) have even lower HREE contents than garnet-bearing Type I. The very low YbN (0·035) of this sample requires even more extensive melting in the absence of garnet. Together with the depleted major element characteristics of K6 (< 0·3 wt% CaO and <0·85 wt% Al2O3), this points to an origin of this (and other spinel peridotites) as a residue of >30% partial melt extraction at relatively shallow levels. After melting, the residue would have been a spinel- and cpx-free harzburgite with approximately >75% modal olivine and <25% modal opx [model residue after 28% melt extraction in spinel stability field only, applying the melting modes of Johnson et al. (1990)]. In contrast, K6 contains 65% olivine and 34% opx (plus < 1% exsolved spinel and cpx), indicating opx addition some time after melt extraction.
|
The above conclusions are supported by Yb–Ca and moderately incompatible element systematics (Kelemen et al., 1998
; Canil, 2004). Samples LET2, K18 and GP402 have higher whole-rock Yb than the main Type I cluster. This might indicate that they melted at higher average pressures or experienced later Yb (HREE) addition.
In Fig. 17, bulk-rock major element compositions of the Kimberley and Lesotho samples are plotted together with experimental melting trends at 1, 3 and 7 GPa (Jaques & Green, 1980; Walter, 2003) and trajectories for 5–35% opx addition to an experimental residue of 37% melt extraction at 3 GPa. In the Kaapvaal xenoliths, SiO2 is high and FeO is low at a given MgO content relative to non-cratonic peridotites and experimental residua, features that were previously recognized as typical for Kaapvaal (and some Slave and Siberian) low-T peridotites (e.g. Boyd & Mertzman, 1987; Boyd, 1989; Canil, 1991; Herzberg, 1993, 1999; Rudnick et al., 1994; Kelemen et al., 1998; Walter, 1998, 1999; Griffin et al., 1999a). The high SiO2 contents can be explained by 5–35% of opx addition to a residue of 30–40% melting (Fig. 17; Herzberg, 1993, 1999; Walter, 1998). The low FeO and high MgO contents have previously been explained by melting at very high average pressures (30 to >50% melt extraction at 5 to >7 GPa; e.g. Walter, 2003). Al2O3–MgO systematics, however, strongly argue against extensive melting at pressures >3 GPa (Fig. 17b) and are consistent with 20–40% melting at average pressures of 1–3 GPa, followed by opx addition. Opx addition also lowers the FeO content in the bulk-rock by up to 1 wt%, but this is not sufficient to explain the extremely low FeO of the samples (Fig. 17c). It is possible that melt– or fluid–rock reaction (instead of just bulk addition of opx) might lead to a stronger reduction in FeO, in particular if this process takes place under more oxidizing conditions, where iron becomes more incompatible in peridotite as a result of a higher proportion of ferric iron (Canil, 2002).
|
The Cr/Al ratios of the Kimberley xenoliths (Fig. 18) also indicate that the average depth of melting was equivalent to <3 GPa (Bulatov et al., 1991
) and that the garnet-free Type I samples probably experienced melting at shallower depths (highest Cr2O3/Al2O3 ratios) than the garnet-bearing Type I samples.
|
Kelemen et al. (1998
) noted the apparent paradox that, despite evidence of extensive low-P melting, most of the samples they studied contain abundant garnet. Accepting that the modelling above demonstrates that melt extraction from the now garnet-rich samples ceased in the spinel or even plagioclase stability field, we concur with Kelemen et al. (1998) in invoking a second process that (tectonically) re-emplaces them into the garnet stability field. Although some garnet will form during cooling of residual spinel peridotites formed at high temperatures (e.g. Canil, 1991) this process cannot explain the almost ubiquitous presence of garnet in Kaapvaal xenoliths.
We note that the results presented here point to the importance of low-pressure melting in the genesis of the protolith of the cratonic lithosphere. The high-pressure plume melting alternative for the formation of CLM requires very high degrees of melt removal to explain the lack of an aluminous phase on the solidus (Griffin et al., 2003a). As shown by Walter (1998), garnet stability on the solidus increases with increasing pressure and garnet would be removed from the residue only at >67% melt extraction at 7 GPa. In contrast, garnet is exhausted by 14% melting at 3 GPa (Walter, 1998). Residua of high-P melting are even more deficient in opx than residua of melting at P < 3 GPa. Hence any high-P model also must rely on a secondary process of opx addition to explain the silica-rich nature of the CLM (Herzberg, 1999). Because the opx involved in the high-pressure models is the crystallization product of high-degree melts, it would have none of the trace element characteristics that we report here for CLM opx. Instead, the observed opx trace element characteristics match those expected for fluids or melts released from subducting slabs into the shallow upper mantle. Such fluids, through melt–rock interaction with previously highly depleted peridotite, would introduce opx in the manner proposed by Kesson & Ringwood (1989), Rudnick et al. (1994) and Kelemen et al. (1998). Thus, a high-P model (e.g. Griffin et al., 2003a) requires extraordinary conditions (mantle potential temperatures >1700°C) for both melting and the process of silica enrichment, yet still fails to explain the trace element characteristics of the Kaapvaal low-T peridotites.
|
MODEL FOR THE FORMATION OF KAAPVAAL CLM |
|---|
|
TOPABSTRACTINTRODUCTIONPETROGRAPHIC DESCRIPTION AND...ANALYTICAL METHODSRESULTSDISCUSSIONMODEL FOR THE FORMATION...SUPPLEMENTARY DATAREFERENCES |
|---|
Here we attempt to place the results in a geological context and propose a simplified model that explains the most important of the observed characteristics (Fig. 19). The geochemical evidence is summarized in Table 9.
- Melting in an intra-oceanic setting in the Palaeoarchaean results in a residual mantle that is highly depleted in magmaphile major elements and incompatible trace elements. The intra-oceanic setting is suggested by the extensive shallow melting, which implies that melting occurred in an area of very thin lithosphere.
- Subduction and collision of small blocks of early crust lead to the formation of a first continental shield (De Wit et al., 1992). This setting provides the opportunity for Si-rich fluids released from the subducted slab to enrich the overlying mantle wedge in Si, LREE and LILE. As a result, the HFSE contents decrease relative to other trace elements, producing the observed high LREE/HFSE ratios. This process can account for the enrichment in Si over Mg and hence the high opx contents in the Kaapvaal upper mantle, and explain the observed correlation between the modal opx content and the Os isotope compositions. The presence of Mesoarchaean eclogitic inclusions in diamonds confirms the presence of subducted material in the Kimberley area at 2·9 Ga (Richardson et al., 2001; Shirey et al., 2002) as do Archaean ages for Kaapvaal eclogites (Shirey et al., 2001) and for high Re/Os ratio sulphides widely distributed in Kaapvaal peridotites (Griffin et al., 2004; Carlson et al., 2005). Cessation of subduction, attachment of the mantle wedge to the lithosphere and thickening of the whole lithospheric section could have been accomplished by the collision of the western and eastern blocks of the Kaapvaal Craton at 2·9 Ga (e.g. Jordan, 1988; Schmitz et al., 2004).
- Subsequent to the stabilization of the Kaapvaal craton in the late Archaean (2·8 Ga), the CLM was only locally affected by metasomatic events. With the exception of the Bushveld intrusion (e.g. Carlson et al., 1999), none of these events was sufficient to lead to major modifications of the bulk chemistry of the lithospheric mantle by, for example, the introduction of >10% modal opx, or completely reset the Re–Os system.
|
|
Kimberlite magmatism leads to another phase of metasomatism resulting in, amongst other effects, the introduction of cpx. In Lesotho, this event appears to be restricted to Group I kimberlite magmatism in the Cretaceous, whereas multiphase events, including Group II kimberlite volcanism, affect the craton further west in the Kimberley area.
|
SUPPLEMENTARY DATA |
|---|
|
TOPABSTRACTINTRODUCTIONPETROGRAPHIC DESCRIPTION AND...ANALYTICAL METHODSRESULTSDISCUSSIONMODEL FOR THE FORMATION...SUPPLEMENTARY DATAREFERENCES |
|---|
Supplementary data for this paper are available at Journal of Petrology online.
|
ACKNOWLEDGEMENTS |
|---|
We thank Gordon Irvine for sharing his Lesotho samples. D. Bell, J. J. Gurney, D. Smith and J. Robey provided invaluable help for successful sampling of the Kimberley dumps. We acknowledge technical help from Wim Lustenhouwer (VU EPMA), Saskia Kars (VU SEM), Bas van der Wagt (VU solution ICP-MS), Paul Mason (Utrecht LA-ICP-MS), Mary Horan (DTM chemistry), Tim Mock (DTM mass spectrometry) and Geoff Nowell (Nd isotopes, Durham). Comments by E.-R. Neumann and the editor M. Wilson on an earlier version of the manuscript helped to improve and shorten it. Thanks also go to J. Davies for redrawing some of the figures. A helpful review by M. Grégoire and two extremely thorough reviews by D. Canil and an anonymous reviewer helped to focus the arguments. The first author was supported by a Ph.D. grant from the Dutch research council (NWO project no. 809-31.001).
*Corresponding author. Present address: Physics of Geological Processes, University of Oslo, PO Box 1048—Blindern, 0316 Oslo, Norway. Telephone: +47-22856922. Fax: +47-22855101. E-mail: n.s.c.simon{at}fys.uio.no
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONPETROGRAPHIC DESCRIPTION AND...ANALYTICAL METHODSRESULTSDISCUSSIONMODEL FOR THE FORMATION...SUPPLEMENTARY DATAREFERENCES |
|---|
Allsopp HL and Barrett DR. (1975) Rb–Sr age determinations on South African kimberlite pipes. Physics and Chemistry of the Earth 9:605–617.[CrossRef]
Armstrong RA, Compston W, Retief EA, Williams IS, Welke HJ. (1991) Zircon ion microprobe studies bearing on the age and evolution of the Witwatersrand triad. Precambrian Research 53:243–266.[CrossRef][Web of Science]
Ballhaus C, Berry RF, Green DH. (1991a) High-pressure experimental olivine–orthopyroxene–spinel–oxygen geobarometer—implications for the oxidation-state of the upper mantle. Contributions to Mineralogy and Petrology 107:27–40.[CrossRef][Web of Science]
Ballhaus C, Berry RF, Green DH. (1991b) Erratum, High-pressure experimental calibration of the olivine–orthopyroxene–spinel–oxygen geobarometer—implications for the oxidation-state of the upper mantle (Vol 107, Pg 27, 1991). Contributions to Mineralogy and Petrology 108:384.[CrossRef]
Becker H, Carlson RW, Shirey SB. (2004) Slab-derived osmium and isotopic disequilibrium in garnet pyroxenites from a Paleozoic convergent plate margin (lower Austria). Chemical Geology 208:141–156.[CrossRef][Web of Science]
Bedini R-M, Blichert-Toft J, Boyet M, Albarède F. (2004) Isotopic constraints on the cooling of the continental lithosphere. Earth and Planetary Science Letters 223:99–111.[CrossRef][Web of Science]
Bell DR, Grégoire M, Grove TL, Chatterjee N, Carlson RW, Buseck PR. (2005) Silica and volatile-element metasomatism of Archean mantle: a xenolith-scale example from the Kaapvaal Craton. Contributions to Mineralogy and Petrology 150:251–267.[CrossRef][Web of Science]
Bertrand P and Mercier J-CC. (1985) The mutual solubility of coexisting ortho- and clinopyroxene: toward an absolute geothermometer for the natural system. Earth and Planetary Science Letters 76:109–122.[CrossRef][Web of Science]
Bizimis M, Salters VJM, Bonatti E. (2000) Trace and REE content of clinopyroxenes from supra-subduction zone peridotites. Implications for melting and enrichment processes in island arcs. Chemical Geology 165:67–85.[CrossRef][Web of Science]
Boyd FR. (1989) Compositional distinction between oceanic and cratonic lithosphere. Earth and Planetary Science Letters 96:15–26.[CrossRef][Web of Science]
Boyd FR and Gurney JJ. (1986) Diamonds and the African lithosphere. Science 232:472–477.
Boyd FR and McCallister RH. (1976) Densities of fertile and sterile garnet peridotites. Geophysical Research Letters 3:509–512.[Web of Science]
Boyd FR and Mertzman SA. (1987) Composition and structure of the Kaapvaal lithosphere, southern Africa. In Mysen BO (Ed.). Magmatic Processes: Physicochemical PrinciplesGeochemical Society, Special Publications 1: pp. 13–24.
Boyd FR and Nixon PH. (1973) Structure of the upper mantle beneath Lesotho. Carnegie Institution of Washington Yearbook 72:431–445.
Boyd FR and Nixon PH. (1975) Origins of the ultramafic nodules from some kimberlites of Northern Lesotho and the Monastery Mine, South Africa. Proceeding of the First International Conference on Kimberlites1973Cape Town (Physics and Chemistry of the Earth) (Pergamon PressIn Ahrens LH, Dawson JB, Duncan AR, Erlank AJ (Eds.). , Oxford) pp. 431–454.
Boyd FR, Pearson DG, Mertzman SA. (1999) Spinel-facies peridotites from the Kaapvaal root. Proceedings of the 7th International Kimberlite Conference1998Cape Town(Red Roof DesignIn Gurney JJ, Gurney JL, Pascoe MD, Richardson SH (Eds.). , Cape Town) pp. 40–48.
Brandon AD, Creaser RA, Shirey SB, Carlson RW. (1996) Osmium recycling in subduction zones. Science 272:861–863.[CrossRef][Web of Science][Medline]
Brey GP and Köhler T. (1990) Geothermobarometry in four-phase lherzolites II. New thermobarometers, and practical assessment of existing thermobarometers. Journal of Petrology 31:1353–1378.
Bulatov V, Brey GP, Foley SF. (1991) Origin of low-Ca, high-Cr garnets by recrystallization of low-pressure harzburgites. Fifth International Kimberlite Conference, Extended Abstracts, CPRM Special Publication 2/91: pp. 29–31.
Canil D. (1991) Experimental evidence for the exsolution of cratonic peridotite from high-temperature harzburgite. Earth and Planetary Science Letters 106:64–72.[CrossRef][Web of Science]
Canil D. (2002) Vanadium in peridotites, mantle redox and tectonic environments: Archean to present. Earth and Planetary Science Letters 195:75–90.[CrossRef][Web of Science]
Canil D. (2004) Mildly incompatible elements in peridotites and the origins of mantle lithosphere. Lithos 77:375–393.[CrossRef][Web of Science]
Carlson RW and Moore RO. (2004) Age of the Eastern Kaapvaal mantle: Re–Os isotope data for peridotite xenoliths from the Monastery kimberlite. South African Journal of Geology 107:81–90.
Carlson RW, Esperanca S, Svisero DP. (1996) Chemical and Os isotopic study of Cretaceous potassic rocks from southern Brazil. Contributions to Mineralogy and Petrology 125:393–405.[CrossRef][Web of Science]
Carlson RW, Pearson DG, Boyd FR, Shirey SB, Irvine G, Menzies AH, Gurney JJ. (1999) Re–Os systematics of lithospheric peridotites: implications for lithosphere formation and preservation. Proceedings of the 7th International Kimberlite Conference1998Cape Town(Red Roof DesignIn Gurney JJ, Gurney JL, Pascoe MD, Richardson SH (Eds.). , Cape Town) pp. 99–108.
Carlson RW, Irving AJ, Schulze DJ, Hearn J, Carter B. (2004) Timing of Precambrian melt depletion and Phanerozoic refertilization events in the lithospheric mantle of the Wyoming Craton and adjacent Central Plains Orogen. Lithos 77:453–472.[CrossRef][Web of Science]
Carlson RW, Pearson DG, James DE. (2005) Physical, chemical and chronological characteristics of continental mantle. Reviews of Geophysics 43: doi:10.1029/2004RG000156.
Carswell DA. (1975) Primary and secondary phlogopites and clinopyroxenes in garnet lherzolite xenoliths. Proceeding of the First International Conference on Kimberlites1973Cape Town (Physics and Chemistry of the Earth) (Pergamon PressIn Ahrens LH, Dawson JB, Duncan AR, Erlank AJ (Eds.). , Oxford) pp. 417–429.
Davis CL. (1977) The ages and uranium contents of zircons from kimberlites and associated rocks. Carnegie Institution of Washington Yearbook 76:631–654.
De Wit MJ, Roering C, Hart RJ, Armstrong RA, de Ronde CEJ, Green RWE, Tredoux M, Peberdy E, Hart RA. (1992) Formation of an Archaean continent. Nature 357: pp. 553–562.[CrossRef]
Dowall DP, Nowell GM, Pearson DG. (2003) Chemical pre-concentration procedures for high-precision analysis of Hf–Nd–Sr isotopes in geological materials by plasma ionisation multi-collector mass spectrometry (PIMMS) techniques. In Holland JG and Tanner SD (Eds.). Plasma Source Mass Spectrometry: Applications and Emerging Technologies(Royal Society of Chemistry, Cambridge) pp. 321–337.
Drennan GR, Robb LJ, Meyer FM, Armstrong RA, de Bruiyn H. (1990) The nature of the Archaean basement in the hinterland of the Witwatersrand Basin; II, A crustal profile west of the Welkom Goldfield and comparisons with the Vredefort crustal profile. South African Journal of Geology 93:41–53.[Abstract]
Eggins SM, Woodhead JD, Kinsley LPJ, Mortimer GE, Sylvester P, McCulloch MT, Hergt JM, Handler MR. (1997) A simple method for the precise determination of 
40 trace elements in geological samples by ICPMS using enriched isotope internal standardisation. Chemical Geology 134:311–326.[CrossRef][Web of Science]
Eglington BM and Armstrong RA. (2004) The Kaapvaal Craton and adjacent orogens, southern Africa: a geochronological database and overview of the geological development of the craton. South African Journal of Geology 107:13–32.
Ellis DJ and Green DH. (1979) An experimental study of the effect of Calcium upon garnet–clinopyroxene iron magnesium exchange equilibria. Contributions to Mineralogy and Petrology 95:499–511.[CrossRef]
Erlank AJ, Waters FG, Hawkesworth CJ, Haggerty SE, Allsopp HL, Rickard RS, Menzies M. (1987) Evidence for mantle metasomatism in peridotite nodules from the Kimberley pipes, South Africa. In Menzies M and Hawkesworth CJ (Eds.). Mantle Metasomatism(Academic Press, London) pp. 221–309.
Foley SF, Barth MG, Jenner GA. (2000) Rutile/melt partition coefficients for trace elements and an assessment of the influence of rutile on the trace element characteristics of subduction zone magmas. Geochimica et Cosmochimica Acta 64:933–938.[CrossRef][Web of Science]
Ganguly J, Cheng W, Chakraborty S. (1998) Cation diffusion in aluminosilicate garnets: experimental determination in pyrope–almandine diffusion couples. Contributions to Mineralogy and Petrology 131:171–180.[CrossRef][Web of Science]
Godard M, Bodinier J-L, Vasseur G. (1995) Effects of mineralogical reactions on trace element redistributions in mantle rocks during percolation processes: a chromatographic approach. Earth and Planetary Science Letters 133:449–461.[CrossRef][Web of Science]
Grégoire M, Bell DR, Le Roex AP. (2002) Trace element geochemistry of phlogopite-rich mafic mantle xenoliths: their classification and their relationship to phlogopite-bearing peridotites and kimberlites revisited. Contributions to Mineralogy and Petrology 142:603–625.[Web of Science]
Grégoire M, Bell DR, Roex APL. (2003) Garnet lherzolites from the Kaapvaal Craton (South Africa): trace element evidence for a metasomatic history. Journal of Petrology 44:629–657.
Griffin WL, O'Reilly SY, Ryan CG. (1999a) The composition and origin of sub-continental lithospheric mantle. In Fei Y, Bertka CM, Mysen BO (Eds.). Mantle Petrology: Field Observations and High Pressure Experimentation: a Tribute to Francis R. (Joe) Boyd. Geochemical Society, Special Publications 6: pp. 13–45.
Griffin WL, Shee SR, Ryan CG, Win TT, Wyatt BA. (1999b) Harzburgite to lherzolite and back again: metasomatic processes in ultramafic xenoliths from the Wesselton kimberlite, Kimberley, South Africa. Contributions to Mineralogy and Petrology 134:232–250.[CrossRef][Web of Science]
Griffin WL, O'Reilly SY, Abe N, Aulbach S, Davies RM, Pearson NJ, Doyle BJ, Kivi K. (2003a) The origin and evolution of Archean lithospheric mantle. Precambrian Research 127:19–41.[CrossRef][Web of Science]
Griffin WL, O'Reilly SY, Natapov LM, Ryan CG. (2003b) The evolution of lithospheric mantle beneath the Kalahari Craton and its margins. Lithos 71:215–241.[CrossRef][Web of Science]
Griffin WL, Graham S, O'Reilly SY, Pearson NJ. (2004) Lithosphere evolution beneath the Kaapvaal craton: Re–Os systematics of sulfides in mantle-derived peridotites. Chemical Geology 208:89–118 doi:10.1016/j.chemgeo.2004.04.002003.[CrossRef][Web of Science]
Günther M and Jagoutz E. (1991) Isotopic disequilibria (Sm/Nd, Rb/Sr) between minerals of coarse grained, low temperature peridotites from the Kimberley floors, southern Africa. Proceedings of the 5th International Kimberlite Conference(Companhia de Pesquisa de Recursos MineraisIn Meyer HOA and Leonardos OH (Eds.). , Brazilia) pp. 122–124.
Hamilton MA, Pearson DG, Stern RA, Boyd FR. (1998) Constraints on MARID petrogenesis: SHRIMP II U–Pb zircon evidence for pre-eruption metasomatism at Kampfersdam. Extended Abstracts of the 7th International Kimberlite Conference, Cape Town pp. 296–298.
Harley SL. (1984) An experimental study of the partitioning of iron and magnesium between garnet and orthopyroxene. Contributions to Mineralogy and Petrology 86:359–373.[CrossRef][Web of Science]
Harte B. (1983) Mantle peridotites and processes – the kimberlite sample. In Hawkesworth CJ and Norry MJ (Eds.). Continental basalts and mantle xenoliths(Shiva Publishing Ltd, Nantwich) pp. 46–91.
Harte B, Winterburn PA, Gurney JJ. (1987) Metasomatic and enrichment phenomena in garnet peridotite facies mantle xenoliths from the Matsoku kimberlite pipe, Lesotho. In Menzies M (Ed.). Mantle Metasomatism(Academic Press, London) pp. 145–220.
Hartnady C, Joubert P, Stowe C. (1985) Proterozoic crustal evolution in southwestern Africa. Episodes 8:236–244.[Web of Science]
Hawkesworth CJ, Erlank AJ, March JS, Menzies MA, van Calsteren P. (1983) Evolution of the continental lithosphere: evidence from volcanics and xenoliths from Southern Africa. In Hawkesworth CJ and Norry MJ (Eds.). Continental Basalts and Mantle Xenoliths(Shiva, Nantwich) pp. 111–138.
Hellebrand E, Snow JE, Hoppe P, Hofmann AW. (2002) Garnet-field melting and late-stage refertilization in ‘residual’ abyssal peridotites from the Central Indian Ridge. Journal of Petrology 43:2305–2338.
Herzberg C. (1999) Phase equilibrium constraints on the formation of cratonic mantle. In Fei Y, Bertka CM, Mysen BO (Eds.). Mantle Petrology: Field Observations and High Pressure Experimentation: a Tribute to Francis R. (Joe) Boyd. Geochemical Society, Special Publications 6: pp. 241–257.
Herzberg CT. (1993) Lithosphere peridotites of the Kaapvaal Craton. Earth and Planetary Science Letters 120:13–29.[CrossRef][Web of Science]
Hofmann A. (1988) Chemical differentiation of the Earth: the relationship between mantle, continental crust, and oceanic crust. Earth and Planetary Science Letters 90:297–314.[CrossRef][Web of Science]
Irvine GJ. (2002) Time constraints on the formation of lithospheric mantle beneath cratons: a Re–Os isotope and platinum group element study of peridotite xenoliths from Northern Canada and Lesotho. Ph.D. thesis (Durham University, Durham).
Irvine GJ, Pearson DG, Carlson RW. (2001) Lithospheric mantle evolution of the Kaapvaal Craton: a Re–Os isotope study of peridotite xenoliths from Lesotho kimberlites. Geophysical Research Letters 28:2505–2508.[CrossRef][Web of Science]
James D and Fouch M. (2002) Formation and evolution of Archaean cratons: insights from southern Africa. In Fowler CMR, Ebinger CJ, Hawkesworth CJ (Eds.). The Early Earth: Physical, Chemical and Biological Development. Geological Society, London, Special Publications 199: pp. 1–26.
James DE, Boyd FR, Schutt D, Bell DR, Carlson RW. (2004) Xenolith constraints on seismic velocities in the upper mantle beneath southern Africa. Geochemistry, Geophysics, Geosystems 5: doi:10.1029/2003GC000551.
Jaques AL and Green DH. (1980) Anhydrous melting of peridotite at 0–15 kbar pressure and the genesis of tholeiitic basalts. Contributions to Mineralogy and Petrology 73:287–310.[CrossRef][Web of Science]
Johnson KTM, Dick HJB, Shimizu N. (1990) Melting in the oceanic upper mantle; an ion microprobe study of diopsides in abyssal peridotites. Journal of Geophysical Research 95:2661–2678.
Jordan TH. (1978) Composition and development of continental tectosphere. Nature 257:745–750.
Jordan TH. (1988) Structure and formation of the continental tectosphere. Journal of Petrology 29:11–37.
Kelemen PB, Hart SR, Bernstein S. (1998) Silica enrichment in the continental upper mantle via melt/rock reaction. Earth and Planetary Science Letters 164:387–406.[CrossRef][Web of Science]
Kessel R, Schmidt MW, Ulmer P, Pettke T. (2005) Trace element signature of subduction-zone fluids, melts and supercritical liquids at 120–180 km depth. Nature 437:724–727.[CrossRef][Medline]
Kesson SE and Ringwood AE. (1989) Slab–mantle interactions 2. The formation of diamonds. Chemical Geology 78:97–118.[CrossRef][Web of Science]
Köhler T and Brey G. (1990) Calcium exchange between olivine and clinopyroxene calibrated as geothermobarometer for natural peridotites from 2 to 60 kb with applications. Geochimica et Cosmochimica Acta 54:2375–2388.[CrossRef]
Konzett J, Armstrong RA, Günther D. (2000) Modal metasomatism in the Kaapvaal craton lithosphere: constraints on timing and genesis from U–Pb zircon dating of metasomatized peridotites and MARID-type xenoliths. Contributions to Mineralogy and Petrology 139:704–719.[CrossRef]
Krogh EJ. (1988) The garnet–clinopyroxene iron–magnesium geothermometer—a reinterpretation of existing experimental data. Contributions to Mineralogy and Petrology 99:44–48.[CrossRef][Web of Science]
Lahaye Y, Arndt N, Byerly G, Chauvel C, Fourcade S, Gruau G. (1995) The influence of alteration on the trace-element and Nd isotopic compositions of komatiites. Chemical Geology 126:43–64.[CrossRef][Web of Science]
Lowry D, Mattey DP, Harris JW. (1999) Oxygen isotope composition of syngenetic inclusions in diamond from the Finsch Mine, RSA. Geochimica et Cosmochimica Acta 63:1825–1836.[CrossRef][Web of Science]
MacGregor I. (1974) The system MgO–Al2O3–SiO2: solubility of Al2O3 in enstatite for spinel and garnet peridotite compositions. American Mineralogist 59:110–119.[Web of Science]
Mattey D, Lowry D, Macpherson C. (1994) Oxygen-isotope composition of mantle peridotite. Earth and Planetary Science Letters 128:231–241.[CrossRef][Web of Science]
McDonough WF and Sun S-S. (1995) The composition of the Earth. Chemical Geology 120:223–253.[CrossRef][Web of Science]
McInnes BIA, McBride JS, Evans NJ, Lambert DD, Andrew AS. (1999) Osmium isotope constraints on ore metal recycling in subduction zones. Science 286:512–516.[CrossRef][Web of Science][Medline]
Meisel T, Walker RJ, Irving AJ, Lorand J-P. (2001) Osmium isotopic compositions of mantle xenoliths: a global perspective. Geochimica et Cosmochimica Acta 65:1311–1323.[CrossRef][Web of Science]
Menzies AH, Carlson RW, Shirey SB, Gurney JJ. (1999) Re–Os systematics of Newlands peridotite xenoliths: implications for diamond and lithosphere formation. Proceedings of the 7th International Kimberlite Conference(Red Roof DesignIn Gurney JJ, Gurney JL, Pascoe MD, Richardson SH (Eds.). , Cape Town) pp. 566–573.
Menzies M and Murthy V. (1980) Enriched mantle: Nd and Sr isotopes in diopsides from kimberlite nodules. Nature 283:634–636.[CrossRef]
In Menzies MA and Hawkesworth CJ (Eds.). Mantle Metasomatism (1987) (Academic Press, London).
Mibe K, Fujii T, Yasuda A. (1998) Connectivity of aqueous fluid in the Earth's upper mantle. Geophysical Research Letters 25:1233–1236.[CrossRef][Web of Science]
Mibe K, Fujii T, Yasuda A. (2002) Composition of aqueous fluid coexisting with mantle minerals at high pressure and its bearing on the differentiation of the Earth's mantle. Geochimica et Cosmochimica Acta 66:2273–2285.[CrossRef][Web of Science]
Morishita T, Arai S, Green DH. (2003) Evolution of low-Al orthopyroxene in the Horoman Peridotite, Japan: an unusual indicator of metasomatizing fluids. Journal of Petrology 44:1237–1246.
Nickel KG and Green DH. (1985) Empirical geothermobarometry for garnet peridotites and implications for the nature of the lithosphere, kimberlites and diamonds. Earth and Planetary Science Letters 73:158–170.[CrossRef][Web of Science]
Nowell GM, Pearson DG, Bell DR, Carlson RW, Smith CB, Kempton PD, Noble SR. (2004) Hf isotope systematics of kimberlites and their megacrysts: new constraints on their source regions. Journal of Petrology 45:1583–1612.
O'Neill HSC. (1980) An experimental study of the iron–magnesium partitioning between garnet and olivine and its calibration as a geothermometer: corrections. Contributions to Mineralogy and Petrology 72:337.[CrossRef][Web of Science]
O'Neill HSC and Wall VJ. (1987) The olivine–orthopyroxene–spinel oxygen-geobarometer, the nickel precipitation curve and the oxygen fugacity of the Earth's upper mantle. Journal of Petrology 28:1169–1191.
O'Neill HSC and Wood BJ. (1979) An experimental study of the iron–magnesium partitioning between garnet and olivine and its calibration as a geothermometer. Contributions to Mineralogy and Petrology 70:59–70.[CrossRef][Web of Science]
Pearson DG and Nowell GM. (2002) The continental lithospheric mantle: characteristics and significance as a mantle reservoir. Philosophical Transactions of the Royal Society of London 360:2383–2410.
Pearson DG and Nowell GM. (2004) Re–Os and Lu–Hf isotope constraints on the origin and age of pyroxenites from the Beni Bousera peridotite massif: implications for mixed peridotite–pyroxenite mantle sources. Journal of Petrology 45:439–455.
Pearson DG and Nowell GM. (2005) Accuracy and precision in plasma ionisation multi-collector mass spectrometry: Constraints from neodymium and hafnium isotope measurements. In Holland JG and Bandura DR (Eds.). Plasma Source Mass Spectrometry: Current Trends and Future Developments(Royal Society of Chemistry, London) pp. 284–314.
Pearson DG, Carlson RW, Shirey SB, Boyd FR, Nixon PH. (1995a) Stabilization of Archean lithospheric mantle: a Re–Os isotope study of peridotite xenoliths from the Kaapvaal Craton. Earth and Planetary Science Letters 134:341–357.[CrossRef][Web of Science]
Pearson DG, Shirey SB, Carlson RW, Boyd FR, Pokhilenko NP, Shimizu N. (1995b) Re–Os, Sm–Nd, and Rb–Sr isotope evidence for thick Archean lithospheric mantle beneath the Siberian Craton modified by multistage metasomatism. Geochimica et Cosmochimica Acta 59:959–977.[Web of Science]
Pearson DG, Canil D, Shirey SB. (2003) Mantle samples included in volcanic rocks: xenoliths and diamonds. In Carlson RW (Ed.). The Mantle and Core. (Treatise on Geochemistry)(Elsevier, Amsterdam) pp. 171–275.
Pearson DG, Irvine GJ, Ionov DA, Boyd FR, Dreibus GE. (2004) Re–Os isotope systematics and Platinum Group Element fractionation during mantle melt extraction: a study of massif and xenolith peridotite suites. Chemical Geology Special Volume: Highly Siderophile Elements 208:29–59.
Powell R. (1985) Regression diagnostics and robust regression in geothermometer/geobarometer calibration: the garnet–clinopyroxene geothermometer revisited. Journal of Metamorphic Geology 3:231–243.[Web of Science]
Rapp RP, Shimizu N, Norman MD, Applegate GS. (1999) Reaction between slab-derived melts and peridotite in the mantle wedge: experimental constraints at 3·8 GPa. Chemical Geology 160:335–356.[CrossRef][Web of Science]
Richardson S, Gurney J, Erlank A, Harris J. (1984) Origin of diamonds in old enriched mantle. Nature 310:198–202.[CrossRef]
Richardson S, Erlank A, Hart S. (1985) Kimberlite-borne garnet peridotite xenoliths from old enriched subcontinental lithosphere. Earth and Planetary Science Letters 75:116–128.[CrossRef][Web of Science]
Richardson SH, Shirey SB, Harris JW, Carlson RW. (2001) Archean subduction recorded by Re–Os isotopes in eclogitic sulfide inclusions in Kimberley diamonds. Earth and Planetary Science Letters 191:257–266.[CrossRef][Web of Science]
Ringwood A. (1975) Composition and Petrology of the Earth's Upper Mantle(McGraw–Hill, London).
Rudnick RL and Nyblade AA. (1999) The thickness and heat production of Archaean lithosphere: constraints from xenolith thermobarometry and surface heat flow. In Fei Y, Bertka CM, Mysen BO (Eds.). Mantle Petrology: Field Observations and High Pressure Experimentation: a Tribute to Francis R. (Joe) Boyd. Geochemical Society, Special Publications 6: pp. 3–12.
Rudnick RL, Donough WL, Orpin A. (1994) Northern Tanzanian peridotite xenoliths: a comparison with Kaapvaal peridotites and inferences on metasomatic interactions. In Meyer HOA and Leonardos OH (Eds.). Kimberlites, Related Rocks and Mantle Xenoliths. Companhia de Pesquisas de Recursos Minerais Special Publication Jan/94: pp. 336–353.
Rudnick RL, McDonough WF, O'Connell RJ. (1998) Thermal structure, thickness and composition of continental lithosphere. Chemical Geology 145:395–411.[CrossRef][Web of Science]
Schmidberger SS and Francis D. (2001) Constraints on the trace element composition of the Archaean mantle root beneath Somerset Island, Arctic Canada. Journal of Petrology 42:1095–1117.
Schmidberger SS, Simonetti A, Francis D, Gariépy C. (2002) Probing Archean lithosphere using the Lu–Hf isotope systematics of peridotite xenoliths from Somerset Island kimberlites, Canada. Earth and Planetary Science Letters 197:245–259.[CrossRef][Web of Science]
Schmitz MD. (2002) Geological and thermochronological evolution of the lower crust of southern Africa. Ph.D. thesis (Massachusetts Institute of Technology, Cambridge, MA)269.
Schmitz MD, Bowring SA, de Wit MJ, Gartz V. (2004) Subduction and terrane collision stabilize the western Kaapvaal craton tectosphere 2·9 billion years ago. Earth and Planetary Science Letters 222:363–376.[CrossRef][Web of Science]
Shimizu N. (1999) Young geochemical features in cratonic peridotites from Southern Africa and Siberia. In Fei Y, Bertka CM, Mysen BO (Eds.). Mantle Petrology: Field Observations and High-pressure Experimentation: a Tribute to Francis R. (Joe) Boyd. Geochemical Society, Special Publications 6: pp. 47–55.
Shimizu N, Pokhilenko NP, Boyd FR, Pearson DG. (1999) Trace element characteristics of garnet dunites/harzburgites, host rocks for Siberian peridotitic diamonds. Proceedings of the 7th International Kimberlite Conference1998Cape Town(Red Roof DesignIn Gurney JJ, Gurney JL, Pascoe MD, Richardson SH (Eds.). , Cape Town) pp. pp. 773–782.
Shirey S, Harris J, Richardson S, Fouch M, James D, Cartigny P, Deines P, Viljoen F. (2002) Seismic structure, diamond geology and evolution of the Kaapvaal–Zimbabwe craton. Science 297:1683–1686.[CrossRef][Web of Science][Medline]
Shirey SB, Carlson RW, Richardson SH, Menzies A, Gurney JJ, Pearson DG, Harris JW, Wiechert U. (2001) Archean emplacement of eclogitic components into the lithospheric mantle during formation of the Kaapvaal Craton. Geophysical Research Letters 28:2509–2512.[CrossRef][Web of Science]
Simon NSC. (2004) The formation and modification of cratonic lithospheric roots—a petrological and geochemical study of xenoliths from the Kaapvaal craton. Ph.D. thesis (Vrije Universiteit, Amsterdam)251.
Simon NSC, Irvine GJ, Davies GR, Pearson DG, Carlson RW. (2003) The origin of garnet and clinopyroxene in ‘depleted’ Kaapvaal peridotites. Lithos 71:289–322.[CrossRef][Web of Science]
Smith CB. (1983) Pb, Sr and Nd isotopic evidence for sources of southern African Cretaceous kimberlites. Nature 304:51–54.[CrossRef]
Smith D. (1999) Temperatures and pressures of mineral equilibration in peridotite xenoliths: review, discussion, and implications. In Fei Y, Bertka CM, Mysen BO (Eds.). Mantle Petrology: Field Observations and High-pressure Experimentation: a Tribute to Francis R. (Joe) Boyd. Geochemical Society, Special Publications 6: pp. 171–188.
Smith D, Riter JCA, Mertzman SA. (1999) Water–rock interactions, orthopyroxene growth, and Si-enrichment in the mantle: evidence in xenoliths from the Colorado Plateau, southwestern United States. Earth and Planetary Science Letters 165:45–54.[CrossRef][Web of Science]
Sobolev NV, Lavrent'ev YUG, Pokhilenko NP, Usova LV. (1973) Chrome-rich garnets from the kimberlites of Yakutia and their paragenesis. Contributions to Mineralogy and Petrology 40:39–52.[CrossRef][Web of Science]
Stachel T, Aulbach S, Brey GP, Harris JW, Leost I, Tappert R, Viljoen KSF. (2004) The trace element composition of silicate inclusions in diamonds: a review. Lithos 77:1–19.[CrossRef][Web of Science]
Stalder R, Foley SF, Brey GP, Horn I. (1998) Mineral aqueous fluid partitioning of trace elements at 900–1200°C and 3·0–5·7 GPa: new experimental data for garnet, clinopyroxene, and rutile, and implications for mantle metasomatism. Geochimica et Cosmochimica Acta 62:1781–1801.[CrossRef][Web of Science]
Stalder R, Ulmer P, Thompson A, Günther D. (2001) High pressure fluids in the system MgO–SiO2–H2O under upper mantle conditions. Contributions to Mineralogy and Petrology 140:607–618.[Web of Science]
Van Achterbergh E, Griffin WL, Stiefenhofer J. (2001) Metasomatism in mantle xenoliths from the Letlhakane kimberlites: estimation of element fluxes. Contributions to Mineralogy and Petrology 141:397–414.[Web of Science]
Van Orman JA, Grove TL, Shimizu N. (2001) Rare earth element diffusion in diopside: influence of temperature, pressure, and ionic radius, and an elastic model for diffusion in silicates. Contributions to Mineralogy and Petrology 141:687–703.[Web of Science]
Van Orman JA, Grove TL, Shimizu N, Layne GD. (2002) Rare earth element diffusion in a natural pyrope single crystal at 2·8 GPa. Contributions to Mineralogy and Petrology 142:416–424.[Web of Science]
Vervoort JD and Blichert-Toft J. (1999) Evolution of the depleted mantle: Hf isotope evidence from juvenile rocks through time. Geochimica et Cosmochimica Acta 63:533–556.[CrossRef][Web of Science]
Walker RJ, Carlson RW, Shirey SB, Boyd FR. (1989) Os, Sr, Nd, and Pb isotope systematics of southern African peridotite xenoliths: implications for the chemical evolution of subcontinental mantle. Geochimica et Cosmochimica Acta 53:1583–1595.[CrossRef][Web of Science]
Walter MJ. (1998) Melting of garnet peridotite and the origin of komatiite and depleted lithosphere. Journal of Petrology 39:29–60.[CrossRef][Web of Science]
Walter MJ. (1999) Melting residues of fertile peridotite and the origin of cratonic lithosphere. In Fei Y, Bertka CM, Mysen BO (Eds.). Mantle Petrology: Field Observations and High-pressure Experimentation: a Tribute to Francis R. (Joe) Boyd. Geochemical Society, Special Publications 6: pp. 225–239.
Walter MJ. (2003) Partial melting. In Carlson RW (Ed.). Treatise on Geochemistry(Elsevier, Amsterdam) pp. 1–33.
Wasylenki LE, Baker MB, Kent AJR, Stolper EM. (2003) Near-solidus melting of the shallow upper mantle: partial melting experiments on depleted peridotite. Journal of Petrology 44:1163–1191.
Wells PRA. (1977) Pyroxene thermometry in simple and complex systems. Contributions to Mineralogy and Petrology 62:129–139.[CrossRef][Web of Science]
Widom E, Kepezhinskas P, Defant M. (2003) The nature of metasomatism in the sub-arc mantle wedge: evidence from Re–Os isotopes in Kamchatka peridotite xenoliths. Chemical Geology 196:283–306.[CrossRef][Web of Science]
Witt-Eickschen G and Seck HA. (1991) Solubility of Ca and Al in orthopyroxene from spinel peridotite: an improved version of an empirical geothermometer. Contributions to Mineralogy and Petrology 106:431–439.[CrossRef][Web of Science]
Yaxley GM and Green DH. (1998) Reactions between eclogite and peridotite: mantle refertilization by subduction of oceanic crust. Schweizerische Mineralogische und Petrographische Mitteilungen 78:243–255.
Zack T, Foley SF, Jenner GA. (1997) A consistent partition coefficient set for clinopyroxene, amphibole and garnet from laser ablation microprobe analysis of garnet pyroxenites from Kakanui, New Zealand. Neues Jahrbuch für Mineralogie, Abhandlungen 172:23–41.[Web of Science]
Zhang H-F, Menzies MA, Mattey D. (2003) Mixed mantle provenance: diverse garnet compositions in polymict peridotites, Kaapvaal craton, South Africa. Earth and Planetary Science Letters 216:329–346.[CrossRef][Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  W. L. Griffin, S. Y. O'Reilly, J. C. Afonso, and G. C. Begg The Composition and Evolution of Lithospheric Mantle: a Re-evaluation and its Tectonic Implications J. Petrology, July 1, 2009; 50(7): 1185 - 1204. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. A. Gibson, J. Malarkey, and J. A. Day Melt Depletion and Enrichment beneath the Western Kaapvaal Craton: Evidence from Finsch Peridotite Xenoliths J. Petrology, October 22, 2008; (2008) egn048v1. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. S. C. Simon, E.-R. Neumann, C. Bonadiman, M. Coltorti, G. Delpech, M. Gregoire, and E. Widom Ultra-refractory Domains in the Oceanic Mantle Lithosphere Sampled as Mantle Xenoliths at Ocean Islands J. Petrology, June 1, 2008; 49(6): 1223 - 1251. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||