Journal of Petrology | Volume 38 | Number 1 | Pages 85-113 | 1997
© Oxford University Press 1997
The Origins of Yakutian Eclogite Xenoliths
1 Planetary Geosciences Institute Dept. of Geological Sciences, University of Tennessee, Knoxville, TN 37996, USA
2 Earth and Planetary Sciences Department and McDonnell Center for the Space Sciences, Washington University St Louis, MO 63130, USA
3 Department of Geological Sciences, University of Michigan Ann Arbor, MI 48109, USA
4 Institute of Mineralogy and Petrography Siberian Branch, Russian Academy of Sciences, Novosibirsk, Russia
Received September 27, 1995; Revised typescript accepted August 5, 1996
 |
ABSTRACT |
|---|
 TOP ABSTRACT IntroductionAnalytical MethodsPetrographyMineral Chemistry and...Ion-Probe Trace-Element Mineral...Reconstructed Whole-Rock...Radiogenic Isotopic DataProtoliths for Eclogite...DiscussionSummaryREFERENCES |
|---|
Owing to the association with diamonds, eclogite xenoliths have received disproportionate attention given their low abundance in kimberlites. Several hypotheses have been advanced for the origin of eclogite xenoliths, from the subduction and high-pressure melting of oceanic crust, to cumulates and liquids derived from the upper mantle. We have amassed a comprehensive data set, including major- and trace-element mineral chemistry, carbon isotopes in diamonds, and Rb–Sr, Sm–Nd, Re–Os, and oxygen isotopes in ultrapure mineral and whole-rock splits from eclogites of the Udachnaya kimberlite pipe, Yakutia, Russia. Furthermore, eclogites from two other Yakutian kimberlite pipes, Mir and Obnazhennaya, have been studied in detail and offer contrasting images of eclogite protoliths. Relative to eclogites from southern Africa and other Yakutian localities, Udachnaya eclogites are notable in the absence of chemical zoning in mineral grains, as well as the degree of light rare earth element (LREE) depletion and unradiogenic Sr; lack of significant oxygen, sulfur, and carbon isotopic variation relative to the mantle; and intermineral radiogenic isotopic equilibration. Several of these eclogites could be derived from ancient, recycled, oceanic crust, but many others exhibit no evidence for an oceanic crustal protolith. The apparent lack of stable-isotope variation in the Udachnaya eclogites could be due to the antiquity of the samples and consequent lack of deep oceanic and biogenically diverse environments at that time. Those eclogites that are interpreted to be non-recycled have compositions characteristic of Group A eclogites from other localities that also have been interpreted as being directly from the mantle. At least two separate and diverse isotopic reservoirs are suggested by Nd isotopic whole-rock reconstructions. Most samples were derived from typical depleted mantle. However, two groups of three samples each indicate both enriched mantle and possible ultra-depleted mantle present beneath Yakutia during the late Archean and early Proterozoic. The vast majority of eclogites studied from the Obnazhennaya pipe also exhibit characteristics of Group A eclogites and are probably derived directly from the mantle. However, the eclogites from the Mir kimberlite are more typical of other eclogites world-wide and show convincing evidence of a recycled, oceanic crustal affinity. We concur with the late Ted Ringwood that eclogites can be formed in a variety of ways, both within the mantle and from oceanic crustal residues.
KEY WORDS: diamonds; eclogite xenoliths; isotopic composition; REE; Yakutia
|
Introduction |
|---|
|
TOPABSTRACTIntroductionAnalytical MethodsPetrographyMineral Chemistry and...Ion-Probe Trace-Element Mineral...Reconstructed Whole-Rock...Radiogenic Isotopic DataProtoliths for Eclogite...DiscussionSummaryREFERENCES |
|---|
Kimberlites are rare, small-volume, volatile-rich, mafic igneous rocks that typically are restricted to Proterozoic and Archean continental crust. Virtually the only kimberlites that have been extensively studied are those that contain diamonds. Among the suite of inclusions in kimberlites, eclogites generally make up only a small portion of the xenoliths at a given locality, the principal population being ultramafic xenoliths such as garnet- and spinel-bearing peridotites. Eclogite xenoliths may retain the geochemical signatures of the protoliths (e.g. subducted oceanic crust, ancient mantle), offering the opportunity to address mantle processes that have taken place at earlier times. Several workers have suggested that eclogites may be remnants of the early accretion of the Earth. McCulloch (1989)
claimed that two Yakutian eclogites (from the Obnazhennaya pipe) were derived from an extremely light rare-earth element (LREE)-depleted mantle. This early depletion may have been the result of a terrestrial magma ocean (Ohtani, 1985), similar to the origin proposed for the lunar high-Ti basalts (Snyder et al., 1994). Such a terrestrial mantle source would have a present-day 
Nd of approximately +45 and could have formed prior to 4 billion years ago. Conversely, Snyder et al. (1993) postulated that the ultimate source for at least one of the Yakutian eclogites could have been LREE-enriched mantle that formed over 4 billion years ago. Smyth et al. (1989) and Caporuscio & Smyth (1990, and references therein) presented a host of evidence in support of a deep upper-mantle, igneous-fractionation model for eclogite xenoliths. Eclogites may thus be a key with which to unlock the mysteries of mantle evolution during the earliest portions of Earth history. Conversely, most workers support a recycled crust model for the genesis of some eclogite xenoliths (Helmstaedt & Doig, 1975; MacGregor & Manton, 1986; Shervais et al., 1988; Neal et al., 1990; Taylor, 1993; Jerde et al., 1993a, 1993b).
Eclogite xenoliths from southern African kimberlites have been subdivided previously into three groups based on petrography, mineral chemistry, stable isotopes, and radiogenic isotopes. Shervais et al. (1988) made this subdivision in an attempt to remain consistent with the original mineralogical classification scheme of Coleman et al. (1965). Taylor & Neal (1989) and Neal et al. (1990) added trace-elements and radiogenic isotopes to this three-fold classification scheme:
Group A eclogite xenoliths were characterized by high whole-rock mg-numbers, occasional primary orthopyroxene or olivine as an accessory phase, low-jadeite and Cr-rich clinopyroxenes, Mg- and Cr-rich garnets, LREE-enriched clinopyroxenes and reconstructed whole rocks, relatively unradiogenic Sr and Nd isotopic compositions (87Sr/86Sr = 0.7042–0.7046; Nd = –14 to –16), and mantle-like 
18O.
Group B eclogites have moderate jadeite contents in clinopyroxene, Fe-rich garnets, low abundances of incompatible elements in whole rocks, LREE-depleted clinopyroxenes and extremely LREE-depleted and heavy rare-earth element (HREE)-enriched garnets, relatively radiogenic Sr isotopes (87Sr/86Sr = 0.7087–0.7100) and Nd isotopes (Nd = +40 to +219), and low 18O (3.0–3.3
).
Group C eclogites are characterized by high-jadeite content clinopyroxenes, CaO-rich garnets, low REE abundances accompanied by positive Eu anomalies in both garnets and clinopyroxenes, intermediate, albeit radiogenic, Sr and Nd isotopic compositions (87Sr/86Sr = 0.7083–0.7100; Nd = +46 to +110), and low 18O (4.3–4.9).
Group A eclogites were thought to be true mantle cumulates, whereas Groups B and C were considered fragments of subducted oceanic crust. Our work on Yakutian eclogites has led to a reevaluation of this classification scheme and the consequent combining of Groups B and C eclogites into one group, Group B–C (Jerde et al., 1993b). Furthermore, as will be shown in this paper, the isotopic parameters have been modified somewhat for these groups.
Eclogite xenoliths have been discovered in several kimberlites located within the Siberian Platform of north–central Siberia. The Siberian Platform occupies a large area in north–central Asia and forms the western two-thirds of the Province of Yakutia, extending northward from Lake Baikal to beyond the Arctic Circle (Fig. 1). The diamond-bearing kimberlites are mainly located in the central areas of ancient Archean–Proterozoic basement and range in age from 350 Ma to 
100 Ma. Importantly, metasomatism of the Siberian lithosphere is suspected to have been considerably less intense or extensive than for the Kaapvaal craton of southern Africa (Sobolev, 1977), hence, the igneous and metamorphic histories of Yakutian kimberlitic xenoliths should be better preserved. In addition, diamonds are rather common in the collected Yakutian eclogites, especially in the Mir pipe of the central region and the Udachnaya pipe at the Arctic Circle. After the first find and study of a xenolith of diamondiferous eclogite in the Mir pipe (Bobrievich et al., 1959; Sobolev, 1960), as well as several subsequent samples (Sobolev, 1977), a systematic approach was applied to collection of diamondiferous Yakutian eclogites. Most samples now come from the diamond company, Almazy Rossii Sakha, where diamond-bearing samples are identified by X-rays, and separated out after initial crushing and processing of the kimberlite. These diamond samples are then hand-processed to maximize recovery of large diamonds.
|
A variety of studies over the last 10 years have brought Yakutian eclogites to the forefront of xenolith research. Two diamond-free eclogite samples from the Obnazhennaya pipe on the northern part of the platform were studied in detail by McCulloch (1989)
, as well as six others by Qi et al. (1994). Beard et al. (1995, 1996) have investigated the mineral-chemical and stable isotopic compositions of 16 eclogites from the Mir kimberlite. Recently, several groups have studied many diamond-bearing eclogite xenoliths from a single pipe in Yakutia, Siberia—the Udachnaya pipe (Jerde et al., 1993b; Snyder et al., 1993, 1995; Rudnick & Green, 1994; Ireland et al., 1994; Jacob et al., 1994); 44 eclogite xenoliths have been studied in detail, including 33 by our group alone.
In this paper, we present new data for the Udachnaya eclogites and review all available data on eclogites from this pipe and from two other Yakutian kimberlites, Mir and Obnazhennaya, to formulate a generalized model for the genesis of eclogite xenoliths. The isotopic, trace- and major-element data on these xenoliths lead us to conclude that eclogite xenoliths can be formed in a variety of ways, including subduction and subsequent melting of oceanic crust leaving an eclogitic residue, as well as through the transformation of basaltic, mantle-derived melts into eclogite. Furthermore, some eclogites seem to be closely linked with fertilization of upper mantle, possibly by introduction of basaltic–eclogitic melts into depleted mantle peridotite [similar to Ringwood (1990)].
|
Analytical Methods |
|---|
|
TOPABSTRACTIntroductionAnalytical MethodsPetrographyMineral Chemistry and...Ion-Probe Trace-Element Mineral...Reconstructed Whole-Rock...Radiogenic Isotopic DataProtoliths for Eclogite...DiscussionSummaryREFERENCES |
|---|
Major-element mineral compositions were determined by electron microprobe analysis using the four-spectrometer CAMECA SX-50 in the Department of Geological Sciences at the University of Tennessee. The accelerating voltage was 15 kV, with beam current of 20 or 30 nA. The beam size was 5 or 10 µm, and counting times of 20 s were used for all elements and backgrounds, except for Na in garnet and K in clinopyroxene (40 s). All data were corrected using ZAF procedures adapted for the Cameca probe (PAP).
Rare-earth element (REE) and other trace-element analyses of garnet and clinopyroxene were made on individual minerals with a modified CAMECA IMS-3f ion microprobe (i.e. secondary-ion mass spectrometer, or SIMS) at Washington University in St Louis. Details of the analytical procedures have been given by Zinner & Crozaz (1986) and Lundberg et al. (1988, 1990). The reference element for garnet and clinopyroxene was Si. The relative abundances of 20 other elements were monitored.
Ultrapure mineral separates (and the whole-rock sample, U-86) were leached using a scheme modified from Zindler & Jagoutz (1988). Approximately 30–100 mg of samples were dissolved for each mineral separate analysis. Techniques for chemical separation and mass spectrometric analysis of the isotopes of Rb, Sr, Sm, and Nd and isotope dilution procedures have been outlined in detail by Lee et al. (1993) and Snyder et al. (1994). All Sr and Nd isotopic analyses are normalized to 86Sr/88Sr = 0.1194 and 146Nd/144Nd = 0.7219, respectively. Analyses of NIST-SRM 987 Sr and La Jolla Nd standards were performed throughout this study and gave weighted averages (at the 95% confidence limit, external precision) of 87Sr/86Sr = 0.710250 ± 0.000011, and 143Nd/144Nd = 0.511854 ± 0.000011, respectively. The data in Tables 7 and 8 (below) have been corrected for total procedural blanks as follows: Rb = 11 ± 5 pg; Sr = 127 ± 50 pg; Sm = 6 ± 3 pg; and Nd = 50 ± 20 pg. The blank corrections for Rb and Sr in clinopyroxene and Sm and Nd in both clinopyroxene and garnet were negligible. Some of the Rb and Sr data for garnet required correction; the uncertainties are incorporated in the Rb/Sr and Sr isotopic compositions. Errors in isochron ages are 2
(York, 1969) of the scatter, as calculated in the ISOPLOT program of Ludwig (1990).
|
|
|
Petrography |
|---|
|
TOPABSTRACTIntroductionAnalytical MethodsPetrographyMineral Chemistry and...Ion-Probe Trace-Element Mineral...Reconstructed Whole-Rock...Radiogenic Isotopic DataProtoliths for Eclogite...DiscussionSummaryREFERENCES |
|---|
Most eclogite xenoliths that we have studied from Yakutia are small (ranging from 3 to 7 cm, although some samples are >20 cm), coarse-grained, equigranular, and consist of a bimineralic assemblage of clinopyroxene and garnet varying in proportions from 70:30 to 30:70. One Udachnaya eclogite (U-73) exhibits an even higher proportion of garnet (80%), and four samples (U-1, U-108, U-112, and U-604) have 3–15% kyanite. Although rare, corundum-bearing samples occur in the Mir and Obnazhennaya pipes (Sobolev, 1977
); these have not been fully characterized and are not included in this paper. Most studied Udachnaya eclogites, except for U-1, U-5, U-25, U-236, U-237, and U-281, contain diamond.
Both microdiamonds (50–500 µm, with an average size of 80 µm) and larger diamonds up to 1.5 cm across were identified in the eclogites (Ponomarenko et al., 1976, 1980; Pokhilenko et al., 1982; Sobolev et al., 1991a, 1991b) (Fig. 2a); they are colorless to slightly pale yellow and exhibit layered octahedra (Sobolev, 1977), some with twinning (Jacob et al., 1994). Clinopyroxene is pale, ‘christmas-tree’ green in color and transparent, and is generally interstitial to larger, unaltered garnet grains. Clinopyroxenes are variably altered; 50–95% of the mineral is often altered to a white or very pale green color and consists of opaque, fine-grained clay minerals. Spongy reaction rims of finer (50–100 µm) clinopyroxenes are observed, though uncommon, and are attributed to metasomatic effects (Taylor & Neal, 1989). Garnets are orange to pale orange and anhedral, ranging from 0.2 to 8 mm in size. Kelyphitic rims on garnet, composed of phlogopite, amphibole, and interstitial microcrystalline material, are common. In some samples (e.g. U-41), books of phlogopite (1–3 mm) are also observed. Rutile and sulfides (pyrrhotite and pyrite) also occur.
|
Eclogites from the Udachnaya kimberlite pipe have been subdivided by Sobolev et al. (1994)
into five groups based on texture, mineralogy, and degree of alteration of clinopyroxene (Fig. 2b-2h). Type 1 eclogites (U-1, U-108, U-112, and U-604) are distinguished by highly altered clinopyroxene (90–97% altered), fractured garnets, and the presence of kyanite and rutile (Fig. 2b and 2c). Type 2 eclogites (U-25 and U-281) exhibit a recrystallized texture, including embayed and rounded garnets, and clinopyroxenes that are 70% altered (Fig. 2d), and are diamond free; they are equivalent to the Group A eclogites of Coleman et al. (1965). Type 3 eclogites (U-73, U-86, and U-236) are dominated by garnet and are distinguished by the occurrence of garnets in glomerocrysts, phlogopite in kelyphytic rims, and a low degree of alteration of clinopyroxene (Fig. 2e and 2f). Type 4 eclogites (U-5 and Zh-9) are distinguished by almost completely altered clinopyroxene and pale, pink–orange, ‘stretched’ garnets (Fig. 2g). Type 5 eclogites include the remainder of the samples, by far the largest proportion, and are composed of rounded and coarse-grained (2–8 mm in size), orange to dark orange garnets, and interstitial clinopyroxenes altered to an intermediate degree (70–80%) (Fig. 2h). |
Mineral Chemistry and Equilibration Conditions |
|---|
|
TOPABSTRACTIntroductionAnalytical MethodsPetrographyMineral Chemistry and...Ion-Probe Trace-Element Mineral...Reconstructed Whole-Rock...Radiogenic Isotopic DataProtoliths for Eclogite...DiscussionSummaryREFERENCES |
|---|
In contrast to eclogites from other Yakutian localities (e.g. Beard et al., 1995
), the Udachnaya eclogites are characterized by the virtual absence of both inter- and intra-grain compositional zoning (Sobolev et al., 1994), except for samples U-25 and U-281. These two samples are enriched in Cr and Mn (in garnet), and U-281 is also depleted in Al. The major-element compositions of eclogitic minerals vary considerably (Tables 1 and 2). Udachnaya garnets vary from pyrope- (or Mg-) rich to relatively pyrope-poor (or Ca- and Fe-rich; Fig. 3). It should be noted that garnets from samples U-281, U-25, U-236, and U-75 plot in the Mg-garnet-rich region (designated Group A; U-237 plots near this region) and that the remaining samples plot in the more Fe- and Ca-rich region (Table 1; Fig. 3). Two separate trends appear to converge in the middle of the ternary diagram at sub-equal proportions of pyrope, grossular, and almandine + spessartine; one array continues at constant almandine + spessartine from the pyrope apex toward the grossular apex, and the second projects from the almandine + spessartine apex toward the first trend. The clinopyroxene compositions again show four of the same samples, U-25, U-236, U-237, and U-281, that plot in the Group A region at the lowest Na2O and highest MgO (Fig. 4). Again, all of these so-called Group A eclogites are diamond free. Furthermore, the only Udachnaya eclogites to exhibit intermineral chemical variations are the distinctly Group A samples U-25 and U-281. It is also notable that eclogitic clinopyroxenes with the highest Na2O, which plot within the Group C region in Fig. 4, also contain kyanite.
|
|
|
|
Preliminary study of another, unique, sample (U-92) with Cr-rich garnet (and clinopyroxene) has been described by Sobolev et al. (1995)
. However, this sample also contains low-Cr garnet and clinopyroxene. In fact, Cr in these minerals ranges from 0.1 to 1.0 wt % Cr2O3. This enrichment in Cr in some garnets is reminiscent of low-Cr garnets in peridotites. A direct correlation of increasing Mg with Cr in garnets and clinopyroxenes across this sample is also significant. Rutile and ilmenite are common inclusions (20–40 µm) in all minerals of this eclogite. Furthermore, this eclogite contains minor orthopyroxene with mg-number [Mg2+/(Mg2+ + Fe2+)] = 94. All these features suggest that U-92 is transitional between more typical eclogites and mantle peridotites (Sobolev et al., 1995). We will return to the interpretation of this sample later in the discussion.
The presence of diamond in most eclogites from the Udachnaya and Mir pipes limits the minimum pressure to 4 GPa. Eclogites from the Obnazhennaya kimberlite pipe are diamond free; however, most contain orthopyroxene (so-called Group A eclogites) allowing independent pressures to be estimated (Wood & Banno, 1973) at 4–5 GPa (Qi et al., 1994), similar to those inferred from the Udachnaya and Mir eclogites.
Equilibration temperatures (Ellis & Green, 1979) for eclogites from the Udachnaya eclogites range from 770°C to 1300°C; only a few yield temperatures below 950°C (Jerde et al., 1993b; Sobolev et al., 1994). Those samples with the highest clinopyroxene/garnet KD values {= [Fe/(Fe + Mg)]cpx / [Fe/(Fe + Mg)]gnt}, 0.4–0.6, also yield the highest temperatures, 1182–1302°C (Sobolev et al., 1994). Others with lower KD values (0.2–0.35) yield lower equilibration temperatures.
Other Yakutian eclogites exhibit more restricted temperature ranges. The so-called low-Ca Mir eclogites yield the highest temperatures, 1020–1210°C, whereas high-Ca group eclogites give lower temperatures, 1050–1100°C (Beard et al., 1995, 1996). Group A eclogites from the Obnazhennaya kimberlite pipe give similar temperatures of 880–1100°C and the single Group B–C eclogite yields a temperature of 1080°C (Qi et al., 1994). It appears that the vast majority of Yakutian eclogites were last equilibrated 120–150 km deep in the upper mantle at temperatures from 1000°C to 1200°C.
|
Ion-Probe Trace-Element Mineral Chemistry |
|---|
|
TOPABSTRACTIntroductionAnalytical MethodsPetrographyMineral Chemistry and...Ion-Probe Trace-Element Mineral...Reconstructed Whole-Rock...Radiogenic Isotopic DataProtoliths for Eclogite...DiscussionSummaryREFERENCES |
|---|
The trace-element compositions of minerals from a suite of 19 Udachnaya eclogites (including both garnet and clinopyroxene) were determined using the ion microprobe and are presented in Tables 3 and 4. Although we have indicated the tendency for clinopyroxene to be altered, all analyses were on carefully chosen, crack-free, unaltered grains. Furthermore, we have shown that major-element zonation is not evidenced in either well-preserved clinopyroxenes or garnets. Thus, it is likely that these trace-element analyses reflect igneous processes in the deep mantle and are not affected by alteration and/or kimberlitic metasomatism. The trace-element data presented herein complement the data of 14 other Udachanaya eclogites presented by Jerde et al.
(1993b). However, several other trace-elements were analyzed in all 33 samples and are presented here for the first time. These new analyses include Na, Sc, Ti, V, Cr, Mn, Co, and Zr in garnet and K, Ti, V, Cr, Sr, and Zr in clinopyroxene (Table 5) and are instrumental in understanding relative mineral–mineral partitioning of many of these elements (as well as the REE) at high pressure.
|
|
|
Clinopyroxene–garnet trace-element partitioning
The average clinopyroxene–garnet partition coefficients for Ti, V, Cr, Zr, and the REE in eclogites from three Yakutian kimberlites, Udachnaya, Mir, and Obnazhennaya, are given in Table 6. These data are compared with partition coefficients calculated from the Bellsbank eclogites (Jerde et al., 1993a
) and other eclogites from southern Africa and Australia (O’Reilly & Griffin, 1995), as well as those calculated from the high-pressure mineral–melt partitioning studies of Johnson (1994) and Hauri et al. (1994). The LREE partition coefficients are vastly different among the three Yakutian pipes and vary dramatically with the chemical grouping of the eclogite. For example, clinopyroxene–garnet partition coefficients for Ce in Obnazhennaya eclogites (six were analyzed; five are Group A) average 320, nearly an order of magnitude higher than that of Mir eclogites (36; all Mir eclogites are Group B–C) and 14 times higher than that of Udachnaya Group B–C eclogites (22). Four Udachnaya eclogites (U-25, U-281, U-236, and U-237) were classified as Group A and have extremely high clinopyroxene–garnet partititon coefficients, including an average, high value of 226 for La. Bellsbank, South Africa, Group A eclogites yield even higher partition coefficients (La = 1810; Jerde et al., 1993a), and Group C partition coefficients which are an order of magnitude or more lower than Group B–C eclogites from Udachnaya and Mir (Table 6). The difference in REE partitioning between Groups A and B–C world-wide is undoubtedly due to different processes involved in eclogite formation and metamorphism and/or crystal-chemical controls. However, these two eclogite groups (A and B–C) exhibit only minor differences in P–T estimation (see above); therefore, differing metamorphic or metasomatic histories are less likely as a mechanism for creating these vast differences.
|
The partition coefficients calculated for Ti, V, Cr, Zr, and Sr do not differ markedly between the Groups B–C and A eclogites and appear to be more consistent for eclogites from different localities world-wide (Table 6). Ti partition coefficients are near unity, as shown for all eclogites and including the experimental work of Johnson (1994)
. V and Cr partition coefficients are similar for eclogites from Udachnaya, Yakutia, and Bellsbank, South Africa, and to the experimental data of Hauri et al. (1994). Zr partition coefficients appear to be low for Udachnaya eclogites when compared with others from South Africa and Australia, but are similar to those found in the experimental work of Johnson (1994). Sr partition coefficients are variable and are probably an artifact of the method of analyses. Sr was determined on mineral separates by isotope dilution for the Udachnaya and Bellsbank samples and, therefore, no information is available about partitioning of adjacent minerals. When the partition coefficients of Hauri et al. (1994) are different from those in the measured eclogites, they are low, and by as much as an order of magnitude. However, it is interesting to note that the experimental LREE partitioning data of Hauri et al. (1994) are similar to those of Bellsbank Group C eclogites.
With the exception of La, Ce, and Sr, which are elevated by factors of 2.2, 1.6, and 4.3, respectively, the clinopyroxene–garnet partition coefficients in Group B–C Udachnaya eclogites are comparable with those of Johnson (1994). The mineral–mineral partition coefficients drop to very low values (Er–Lu; Table 6) with similar magnitudes. The experimental data of Hauri et al. (1994) are vastly different from either our observations or the experimental work of Johnson (1994). It should be noted that the values of Hauri et al. (1994) are lower than than those of Johnson (1994) and, thus, would not explain the elevated LREE partition coefficients in the Mir and Obnazhennaya Group B–C eclogites. The order of magnitude differences (in the case of La and Ce) in the experimental clinopyroxene–garnet partition coefficients between Johnson (1994) and Hauri et al. (1994) are difficult to explain. The Udachnaya data match most closely those of Johnson (1994), and their data will be used in further discussions.
There could be several explanations for the difference in La and Ce between the experimental work of Johnson (1994) and the Udachnaya Group B–C data; two will be discussed here. First, the pressures in the experimental work were 2–3 GPa; this is 1–2 GPa lower than that estimated for the Udachnaya eclogites (Jerde et al., 1993b; Sobolev et al., 1994). Johnson (1994) found that there was a pronounced pressure effect for clinopyroxene; the LREE were more enriched relative to the HREE at atmospheric pressure (i.e. McKay et al., 1986). Second, a small partial melt of the eclogite could further deplete the garnet in the LREE relative to clinopyroxene. High-pressure experiments indicate that clinopyroxene is almost always the liquidus phase (Johnson, 1994), suggesting that garnet is exhausted first during melting of eclogite. By inference, garnet may also melt first and thus be depleted in the most incompatible elements (La and Ce) with initial melting of the eclogite, causing an increase in the clinopyroxene–garnet partition coefficient for these elements. This melting scenario may also explain the more than an order of magnitude increase in LREE partition coefficients for Group A eclogites, such as those from Obnazhennaya. That Group A eclogites contain the most pyrope-rich garnet compositions is consistent with this speculation. In fact, garnet LREE abundances in the Obnazhennaya eclogites are the lowest yet analyzed from Yakutian eclogites (Qi et al., 1994).
As pointed out by McKay et al. (1986) for clinopyroxenes and Harte & Kirkley (1994) for clinopyroxene–garnet pairs, the REE partitioning into these minerals is correlated with major-element chemistry. Specifically, REE partitioning is well correlated with the Ca content in garnet (Harte & Kirkley, 1994). The REE patterns of Udachnaya eclogite clinopyroxenes are also linked to major-element compositional variations in the minerals; both Na and Al increase with increasing LREE depletion. Garnets also show a systematic increase in HREE depletion [i.e. increasing (Sm/Yb)n] with increasing Ca. Four eclogites, U-25, U-237, U-236, and U-281, have higher clinopyroxene–garnet partition coefficients for the REE than all other Udachnaya eclogites of similar garnet compositions. These samples are unique in that they are diamond free and contain the most Mg-enriched clinopyroxenes and Cr- and pyrope-rich garnets. These four samples may be classified as Group A eclogites in terms of the major-element compositions and trace-element clinopyroxene–garnet partitioning relationships. Could the higher partition coefficients be indicating a separate process (partial melting?) or environment (higher pressure?) in the mantle?
Mineral REE patterns
Clinopyroxenes in the Udachnaya eclogites have similar middle rare-earth element (MREE) abundances (i.e. Nd = (2–20) x C1 chondrites), but LREE and HREE abundances vary by over two orders of magnitude (Table 4). One clinopyroxene (U-86) exhibits a pronounced positive Eu anomaly; many others exhibit slight negative Eu anomalies. The clinopyroxenes were divided into three types based on the REE patterns (Figs 5 and 6): (1) the most abundant ones exhibit ‘typical’ patterns for clinopyroxene in other mantle xenoliths, (La/Nd)n = 0.4–0.7, (Sm/Yb)n <15 (Fig. 5a); (2) five samples are significantly more HREE depleted (Fig. 5c) than ‘typical’ clinopyroxene, (Sm/Yb)n = 15–50, and tend toward flat to LREE-enriched patterns, (La/Nd)n = 0.5–1.3; (3) five samples are both more LREE and HREE depleted (Fig. 5b) than ‘typical’ clinopyroxene; (La/Nd)n < 0.25; (Sm/Yb)n > 20. These clinopyroxenes are most depleted in LREE, have either chondritic or much lower Sm/Nd ratios, and have sinusoidal REE patterns (Fig. 5b). Sample U-5 has the most nonradiogenic 143Nd/144Nd determined from Udachnaya clinopyroxenes (Nd = –25), which may imply that this LREE enrichment was long lived (>1 Ga). With the exception of U-25, U-236, and U-237, the observed REE patterns are consistent with experimental partition coefficients, suggesting equilibrium was achieved between garnet and clinopyroxene.
|
|
There is no doubt that the original starting compositions of the eclogites exert a major control on the REE compositions of the clinopyroxenes and garnets. Thus, those minerals which are relatively depleted in the LREE probably also had LREE-depleted protoliths. However, the partitioning of the REE into clinopyroxene also may be, in part, crystal-chemically controlled. Al3+ can occur in the tetrahedral (T) site in pyroxenes, but it may also occur in the M1 site to counterbalance the addition of a univalent cation, such as Na+, in the M2 site. Such is the case for Group C, jadeitic clinopyroxenes. The addition of Al3+ to the M1 site causes ‘crimping’ of the M2 site (Cameron & Papike, 1982
) and a consequent reduction in M2 site size. Thus, the M2 site may become even less favorable for such large trivalent cations as the LREE (La, Ce), and more favorable for smaller trivalent cations such as the MREE (Nd and Sm) (Fig. 5b).
Garnets can also be subdivided based on REE patterns (Fig. 7a-7c) into three groups which are generally sympathetic to the three clinopyroxene groups: (l) those garnets with intermediate LREE depletion [(La/Nd)n = 0.03–0.06; n = chondrite normalized] and relatively enriched in the HREE [(Tb/Lu)n = 0.1–0.4] are considered to be ‘normal’ garnet compositions from mantle xenoliths; (2) those garnets that are the most LREE depleted [(La/Nd)n 
0.02], and, unlike the clinopyroxenes from this group which are uniformly the most HREE depleted; the garnets can be further subdivided into HREE-depleted [(Tb/Lu)n 
1.1] and HREE-enriched [(Tb/Lu)n 0.7] varieties; (3) three samples that are similar in degree of LREE depletion to ‘normal’ garnet, but are much more (3x more) HREE depleted. Finally, a single sample, U-119 (Fig. 7c), is the most LREE enriched (albeit still very depleted) of all the garnets analyzed, and coexists with the most LREE-enriched clinopyroxene (see Fig. 5c).
|
|
Reconstructed Whole-Rock Chemical Compositions |
|---|
|
TOPABSTRACTIntroductionAnalytical MethodsPetrographyMineral Chemistry and...Ion-Probe Trace-Element Mineral...Reconstructed Whole-Rock...Radiogenic Isotopic DataProtoliths for Eclogite...DiscussionSummaryREFERENCES |
|---|
In an effort to ‘see through’ the pervasive patent metasomatism in these eclogites to the original protolith(s), we have reconstructed the precursor whole-rock composition by assigning alteration products to either garnet or clinopyroxene (or, in some samples, kyanite) (Sobolev et al., 1994
). The major differences between analyzed whole-rock chemical compositions and reconstructed whole-rock compositions of these eclogites include enrichments in K2O, CaO, and the REE (in particular the LREE), and depletions in Na2O relative to reconstructed whole rocks (Sobolev et al., 1994). Thus, whole-rock chemical analyses will be virtually useless in determining the igneous precursors of these eclogites, and we must reconstruct the whole-rock compositions from major- and trace-element, and isotopic data of ultrapure minerals extracted from the eclogites. Owing to the small size of some of these xenoliths (down to 2–3 cm), whole-rock reconstructions must be considered approximate. Several eclogites were too small to determine meaningful modes; for these samples equal proportions of garnet and clinopyroxene are assumed. It is important to note that Jerde et al. (1993b) have shown that the overall REE patterns for these eclogites are fairly insensitive to the proportions of garnet and clinopyroxene in the sample. However, Sm/Nd ratios vary much more dramatically as the MREE is the approximate position of the ‘cross-over’ of garnet and clinopyroxene patterns, which are of comparable abundances.
REE patterns, relative to chondrites, are shown in Fig. 8a and 8b for these reconstructed whole rocks, and are compared with those reconstructed from two diamond inclusions (Ireland et al., 1994). Two samples, U-86 and U-237, have both the lowest MREE abundances and positive Eu anomalies. With the exception of sample U-237, which is slightly enriched in La and Ce relative to Pr, all reconstructed eclogite whole rocks are LREE depleted. The degree of LREE depletion and HREE enrichment of individual eclogites is not related to the modal proportion of garnet. MREE enrichment is also indicated for many of the reconstructed eclogites (Fig. 8a). This sub-set of Udachnaya eclogites has HREE abundances from 2x to 10x chondrites. Still other reconstructed whole rocks have REE patterns which indicate monotonic enrichments from the LREE to the HREE (Fig. 8b). This sub-set has HREE from 7x to 30x chondrites. These patterns are the most depleted in the LREE and also the most enriched in the HREE of the Udachnaya eclogites.
|
|
Radiogenic Isotopic Data |
|---|
|
TOPABSTRACTIntroductionAnalytical MethodsPetrographyMineral Chemistry and...Ion-Probe Trace-Element Mineral...Reconstructed Whole-Rock...Radiogenic Isotopic DataProtoliths for Eclogite...DiscussionSummaryREFERENCES |
|---|
It is conceivable that the five clinopyroxenes with chondritic to enriched LREEs (Fig. 5c) could have been metasomatized by an LREE-enriched kimberlitic fluid, as suggested by Ireland et al. (1994)
in their study of eclogitic mineral inclusions in diamond. However, although we have observed zoning in both trace and major elements in garnets and clinopyroxenes from other Yakutian eclogites (e.g. Fraracci, 1994; Beard et al., 1995, 1996), those from Udachnaya are remarkably homogeneous both between grains and within grains of a given sample. In fact, most (28 of the 33 we have studied) eclogites exhibit clinopyroxene and garnet REE patterns which are inconsistent with involvement of kimberlitic fluids. For instance, we do not consider it plausible that a kimberlitic fluid [typically with (La/Nd)n = 10–13, Nd > 100x chondrites] could be involved in the formation of clinopyroxenes with (La/Nd)n < 0.25, (Sm/Yb)n > 20 and Nd abundances <10x chondrites (e.g. Fig. 6, Table 4). Instead, we interpret these patterns as indicative either of sources or of processes involved in the formation of mantle eclogites, and not owing to any known form of metasomatism.
The low Rb contents and extremely low Rb/Sr ratios of the leached mineral separates (for clinopyroxene: 0.123–1.10 p.p.m.; for garnet: 0.0046–0.390 p.p.m. with most <0.05 p.p.m.; Table 7) are evidence of the purity of the minerals. Furthermore, we have compared the Sm/Nd ratios measured by SIMS and by isotope dilution in Fig. 9 to support the purity of the minerals. The SIMS analysis involves probing a spot chosen from the center of a grain specifically for its purity and lack of zoning profile. Isotope dilution involves the analysis of a bulk mineral separate, thus any, albeit slight, chemical zonation in individual grains will be ‘averaged out’. The phlogopite separate with the lowest Sm/Nd in the Yakutian eclogites (and the highest Rb at 363 p.p.m.) is a useful gauge of the contribution of metasomatic and/or kimberlitic material (Table 8). Thus, mineral separates that plot to the left of the 1:1 line in Fig. 9 could have been altered by a component with a lower Sm/Nd component (i.e. kimberlite or metasomatic fluid). With the exception of samples U-37 and U-604, all samples plot either on the line or to its right (more LREE-depleted isotopic signatures). This relationship suggests that these few samples indicate depletion in incompatible elements along the rims of the grains, possibly owing to very small degrees of partial melting—a depletion which might not be seen in the interior of the grain. Conversely, these few samples may have preserved trace-element heterogeneities in the minerals owing to rapid crystal growth in the sub-continental mantle (Shimizu & Sobolev, 1995). More importantly, with the exception of a few samples, most notably garnets in U-37 and U-604, and clinopyroxenes in U-281 and U-108, LREE contamination is contraindicated.
|
A second caveat in determining the usefulness of radiogenic isotopic data is that these garnet–clinopyroxene pairs must be shown to be in equilibrium. We have already indicated that the trace-element partitioning (as determined by ion probe) between these two minerals appears to be a primary, igneous feature. Sm/Nd ratios in garnet relative to clinopyroxene, as determined by isotope dilution for all Udachnaya eclogites (Snyder et al., 1993
; Jacob et al., 1994; Pearson et al., 1995a), define a line which indicates an (Sm/Nd)garnet/(Sm/Nd)clinopyroxene of 4.72 (4 of 19 excluded). The r2 value for this line is 0.918, consistent with equilibrium between garnet and clinopyroxene. A similar (Sm/Nd)garnet/(Sm/Nd)clinopyroxene ratio of 4.44 was found for the Mir eclogite data (Beard et al., 1995, 1996).
Nd–Sr isotopic compositions of primary minerals
Strontium isotopes in ultrapure clinopyroxene separates yield measured 87Sr/86Sr in the range of 0.7015–0.7034 for Udachnaya samples (Table 7) and up to 0.7046 for Mir samples (Snyder et al., 1997). These separates generally have lower Rb/Sr ratios (87Rb/86Sr = 0.001–0.1) than coexisting garnet. Mir eclogite clinopyroxenes (two analyzed to date) yield even lower 87Rb/86Sr (0.0001) by an order of magnitude (Snyder et al., 1997). All 11 clinopyroxene separates measured in this work and by Snyder et al. (1993) from the Udachnaya pipe give 87Sr/86Sr 0.7043. However, four of six clinopyroxenes measured by Jacob et al. (1994) have higher 87Sr/86Sr values. The single clinopyroxene separate analyzed by Pearson et al. (1995a) yields an 87Sr/86Sr of 0.7010. Garnet separates yield more variable 87Sr/86Sr (0.7021–0.7116), reflecting wide-ranging 87Rb/86Sr (0.0047–1.8) owing to the incompatibility and low abundances of both Rb and Sr ( 0.4 p.p.m. and 5 p.p.m., respectively). 143Nd/144Nd ratios in clinopyroxene separates vary from 0.5113 to 0.5187 and correlate with 147Sm/144Nd (0.094–0.26). The garnets yield 147Sm/144Nd ratios of 0.35–0.94 and 143Nd/144Nd from 0.5121 to 0.5206. A single garnet separate analyzed by Pearson et al. (1995a) yielded the highest 147Sm/144Nd measured for Udachnaya eclogites.
Mineral isochrons
The Sm–Nd isotopic systematics of garnet–clinopyroxene pairs behave differently from the Rb–Sr isotopic systems for these mineral pairs. In contrast to mineral separates from eclogites of South Africa, mineral separates from Yakutia always yield positive slopes on conventional Sm–Nd isochron diagrams, thus allowing speculation about age information. Athough there are a few exceptions, the Udachnaya eclogites yield Sm–Nd ages which approximate the time of emplacement of the host kimberlite (Table 9). Rb–Sr mineral ages vary from 793 ± 22 Ma to 111 ± 8 Ma, but bracket the suspected age of the kimberlite [380 Ma as stated by Jagoutz (1986) and 384 Ma as determined by 40Ar–39Ar analyses; Burgess et al. (1992); Table 9]. Three other Rb–Sr mineral ‘isochrons’ yield a late-Archean ‘age’ of 3000 ± 400 Ma (U-5), an impossibly old age, 5010 ± 150 Ma (U-112), and an early Proterozoic age of 1980 ± 110 Ma. Eclogite U-112 is the only sample where the clinopyroxene is more radiogenic than the coexisting garnet (a feature common to South African eclogites; Neal et al., 1990). It would appear that the Rb–Sr sytematics of these minerals have been reequilibrated at various times or to varying degrees. Sm–Nd garnet–clinopyroxene ages cluster much more tightly than do Rb–Sr ages. A weighted average for all Udachnaya eclogites gives an Sm–Nd age of 389 ± 4 Ma, remarkably similar to the 384 Ma 40Ar–39Ar age of Burgess et al. (1992). A single Obnazhennaya eclogite gave an Sm–Nd age of 1699 ± 35 (McCulloch, 1989).
|
Nd–Sr isotopic compositions of secondary minerals
The abundances of Sr in the three phlogopite separates are variable (109–1429 p.p.m.; Table 8). 87Sr/86Sr ratios are also highly variable and radiogenic (0.7109–0.7541). 147Sm/144Nd ratios are similar and comparatively low (0.081–0.099) and Nd isotopic compositions approximate those of CHUR.
Whole-rock isotopic compositions are variable and reflect the involvement of fluids that may have also precipitated phlogopite (Table 8). The Sr isotopic compositions of all whole rocks are more radiogenic than the garnet and clinopyroxene separates, and the Nd isotopic compositions are always similar to present-day CHUR. A single sample, U-86, was leached in a similar manner to the mineral separates and exhibits the lowest abundances of both Rb and Sr, and a 87Sr/86Sr ratio that also is rather low. However, this 87Sr/86Sr is higher than those of both clinopyroxene and garnet separates.
Two samples of the host kimberlite have been analyzed from the Udachnaya pipe by Jacob et al. (1994) and Pearson et al. (1995a). Although the Rb and Sr abundances and isotopic compositions are variable, the Sm and Nd isotopic compositions are similar, yielding 147Sm/144Nd = 0.0808–0.0812 and 143Nd/144Nd = 0.51258–0.51260. These ratios are also similar to those of phlogopite separates, possibly indicating the influence of this relatively minor accessory mineral upon the whole-rock isotopic composition.
Re–Os whole-rock isotopic systematics
A suite of seven eclogites—U-1, U-5, U-25, U-281, U-236, UV-179, and UV-464—also have been analyzed for Re and Os abundances and Os isotopic compositions (Pearson et al., 1995b). Re and Os abundances vary from 0.087 to 1.6 p.p.b. and from 0.028 to 0.346 p.p.b., respectively. 187Os/188Os ratios vary from 0.8296 to 9.808. These values are extremely radiogenic, with 10–55% of the total Os being from 187Re decay alone. Five of the seven samples yield model ages between 2.8 and 3.5 Ga. Two other samples, U-1 and U-25, yield model ages that are, respectively, much younger (1.2 Ga) and much older (6.7 Ga), and have been dismissed by Pearson et al. (1995b) as the result of contamination or metasomatism. Excluding these two samples, the other samples which yield Archean ages plot along a line which represents an age of 2.90 ± 0.38 Ga (MSWD = 15.8) and an initial Os isotopic composition, 
Os(i), of 82 ± 41 (Pearson et al., 1995b).
|
Protoliths for Eclogite Xenoliths |
|---|
|
TOPABSTRACTIntroductionAnalytical MethodsPetrographyMineral Chemistry and...Ion-Probe Trace-Element Mineral...Reconstructed Whole-Rock...Radiogenic Isotopic DataProtoliths for Eclogite...DiscussionSummaryREFERENCES |
|---|
Several disparate origins have been postulated for eclogite xenoliths. However, most recent models for the production of eclogite xenoliths can be divided into two broad groups based upon whether the postulated protoliths are of recycled oceanic material or are from a homogeneous mantle. First, we will evaluate evidence world-wide for these two distinctly different protoliths. Thereafter, we will consider specific evidence for the origins of the Yakutian eclogites.
Evidence in favor of recycled oceanic crustal protolith(s)
Ringwood and coworkers were the first to propose a subducted oceanic crustal origin for eclogite xenoliths in kimberlites (Ringwood & Green, 1967; Ringwood, 1975), based solely on the lack of olivine. He also pointed out that an oceanic tholeiite would transform to a quartz eclogite with increasing pressure. Subsequently, the quartz would be melted out during calc-alkaline magma genesis leaving an eclogitic residue. A similar model has recently been proposed for the production of eclogites and consequent formation of Archean tonalites (Ireland et al., 1994).
Further studies tended to support a subducted oceanic crustal origin for eclogites. Helmstaedt & Doig (1975) were the first to present convincing mineralogic and petrographic evidence for such an origin. Jagoutz et al. (1984), MacGregor & Manton (1986), and Shervais et al. (1988) noted that the chemical compositions of some eclogite xenoliths from southern Africa [type B and C of Taylor & Neal (1989)] favor an oceanic crustal protolith. Many others have since followed suit (Neal et al., 1990; Taylor, 1993; Jerde et al., 1993b; Ireland et al., 1994; Jacob et al., 1994). Evidence in favor of such an origin for some eclogites includes: (1) the occurrence of coesite and sanidine, ‘evolved’ phases in igneous crystallization models, throughout the entire compositional range of eclogites (Schulze & Helmstaedt, 1988); (2) positive Eu anomalies and high-Al whole-rock compositions, which point to a low-P anorthositic precursor (Taylor & Neal, 1989); (3) Sr and O isotopic ratios consistent with hydrothermal seawater alteration (Neal et al., 1990; Taylor, 1993); (4) 18O values for eclogitic garnets and clinopyroxenes which are both below and above the accepted mantle values (MacGregor & Manton, 1986; Shervais et al., 1988; Neal et al., 1990; Taylor, 1993; Snyder et al., 1995); (5) a broad range of 13C values (–35 to 0) for eclogitic diamonds world-wide, consistent with a biogenic (i.e. an oceanic ridge vent environment) precursor; (6) radiogenic Pb in eclogitic diamond inclusions (Eldridge et al., 1991; Rudnick et al., 1993); (7) positive correlations in FeO, Sm/Nd, 87Sr/86Sr, and 143Nd/144Nd, and negative correlations in CaO, vs 18O (Jacob et al., 1994). Still other workers have proposed that all eclogite xenoliths are derived from the subduction and high-pressure melting of oceanic crust (Rudnick & Green, 1994; Ireland et al., 1994; Jacob et al., 1994). They cited the high Na2O contents of some eclogitic clinopyroxenes as well as high-pressure experimental data on the composition of liquidus and subsolidus clinopyroxenes in basaltic systems (Thompson, 1974; Rapp, 1995), as evidence against a mantle origin and in favor of a crustal one (Ireland et al., 1994).
Evidence in favor of unrecycled mantle protolith(s)
Abundant evidence exists that at least some eclogites are of mantle derivation with little or no involvement of ocean-floor basalt. Shervais et al. (1988) and Taylor & Neal (1989) subdivided eclogite xenoliths from southern Africa into three groups, A, B, and C, and pointed out that Group A eclogites were consistent with a mantle precursor. Characteristics of Group A eclogites which are consistent with a mantle origin include high Mg/Fe, low Na in clinopyroxene, pyropic garnet, low incompatible element concentrations, moderate 18O (within the range of acceptable ‘mantle’ values), and low 87Sr/86Sr and 143Nd/144Nd. In addition, those workers pointed out higher Cr contents in garnets and clinopyroxenes, vs <0.1 wt % Cr2O3 in crustal-derived Groups B and C. Finally, Haggerty et al. (1994) showed that the bulk P contents of west African eclogites (250 p.p.m.) are similar to upper-mantle estimates (130–220 p.p.m.) and much less than average ocean floor basalts (750 p.p.m.). They used this observation as evidence that eclogites were ‘unlikely to have had subducted protoliths’.
Other groups have suggested that all eclogites represent cumulates from an evolving basaltic magma at high pressure (Smyth et al., 1989; Caporuscio & Smyth, 1990). Evidence for this origin includes: (1) compositional continuity of Groups B and C; (2) the observation of banded eclogites that contain two or more chemically distinct eclogites; (3) garnet and kyanite exsolution from clinopyroxene indicative of cooling from near-solidus temperatures above 3 GPa pressure; (4) reconstructed exsolved clinopyroxene compositions similar to the total kyanite eclogite host; (5) a crystal-chemical model of clinopyroxene–liquid fractionation which alleges to explain MREE enrichments and LREE and HREE depletions characteristic of many eclogites. However, much of this evidence for a mantle origin is also consistent with an oceanic crustal protolith.
What is evident from these and other studies of ‘accepted’ mantle xenoliths is the total lack of definitive indicators for direct mantle derivation. The designation of a xenolith or rock massif as mantle material is based on a lack of crustal indicators. Thus, it would seem that proving crustal involvement is much simpler, as it only involves finding one piece of evidence supportive of crustal material or processes. Conversely, proof of a mantle heritage involves lack of proof to the contrary. As demonstrated below, the Yakutian eclogites present us with evidence for both scenarios.
Protoliths for Yakutian eclogites
Udachnaya eclogite xenoliths are unusual compared with eclogites from other Yakutian localities, as well as eclogite xenoliths world-wide. First, Udachnaya eclogites exhibit inter- and intra-grain homogeneity not found in most other eclogites, including those from the Mir and Obnazhennaya pipes of Yakutia (e.g. Qi et al., 1994; Beard et al., 1995, 1996). Second, Udachnaya clinopyroxenes are generally more LREE depleted than others from South Africa and from Mir and Obnazhennaya (Fig. 10). In fact, several Udachnaya clinopyroxenes (see Fig. 6c) are at least an order of magnitude more LREE depleted than those from South Africa (Taylor & Neal, 1989; Jerde et al., 1993a). Third, the Sr isotopic compositions of Udachnaya clinopyroxenes are less radiogenic than those from South Africa (Fig. 11). With the exception of two samples, Udachnaya clinopyroxenes have 87Sr/86Sr 0.705, whereas clinopyroxenes from South Africa are generally 0.706. Fourth, garnet–clinopyroxene pairs from Udachnaya eclogites yield positive slopes in 147Sm/144Nd vs 143Nd/144Nd diagrams, approximating the age of the kimberlite (Snyder et al., 1993; and Table 9), whereas garnet–clinopyroxene pairs from South African eclogites often give negative slopes indicating disequilibrium (Neal et al., 1990). Fifth, garnets and clinopyroxenes from the Yakutian eclogites exhibit considerably less oxygen isotopic variability than eclogite xenoliths from South Africa (Snyder et al., 1995). Sixth, 13C and 34S values for Udachnaya eclogitic diamonds show little variation, with ranges (13C = –1 to –7, and 34S = –4 to +4) similar to mantle-derived peridotitic diamonds (Rudnick et al., 1993; Snyder et al., 1995). Therefore, it is the interpretation of the singular nature of the Udachnaya eclogites which is of paramount importance.
|
|
Oxygen isotopic variations in Udachnaya eclogites
The very restricted range of oxygen isotopic data from Udachnaya (i.e.
18O = 4.9–7.0) contrasts with that of the Kaapvaal eclogites described by Neal et al. (1990) and Taylor (1993). They envisaged a protolith of ancient, oceanic crust for many eclogites from southern Africa which yield much lower 18O values. The presence of 18O-enriched oxygen in eclogites has been ascribed to melting (or dehydrating) of the upper portion of a slab undergoing subduction (Gregory & Taylor, 1986a, 1986b). These workers also pointed out, however, that even though such enrichment occurs, the 18O of clinopyroxene does not deviate by more than 3 from the original composition under closed-system conditions. This lack of strong deviation suggests that an oceanic oxygen signature may be preserved in the mantle over long time intervals. However, only four of 19 Udachnaya samples show high 18O values, and no demonstrably low values were discovered.
Additional evidence for oceanic crustal involvement can be found in correlations between oxygen isotopes and other chemical parameters. Although Jacob et al. (1994) found support for an oceanic crustal precursor in sympathetic trends, from eight eclogites, of 18O with Fe, Sm/Nd, and 143Nd/144Nd, and a negative trend for Ca vs 18O, we do not find such trends in our much larger data set (which includes a total of 33 eclogites; Sobolev et al., 1994; Snyder et al., 1995). Instead, Udachnaya eclogite data, in toto, indicate a lack of correlation between 18O and other parameters (Snyder et al., 1995). The generally heavier oxygen isotopic values in our data set, compared with that of Jacob et al. (1994), are not due to sampling errors involved in using more conventional oxygen isotopic techniques (Snyder et al., 1995), as we have reproduced these data using laser microanalytical methods (Taylor et al., 1996b).
However, Udachnaya eclogites do display a broad trend of increasing 18O with increasing 87Sr/86Sr values (Snyder et al., 1995). Such a trend is consistent with an ancient ophiolite analog (e.g. McCulloch et al., 1981; Jacob et al., 1994). In this model, eclogites with lower 18O and 87Sr/86Sr might have come from a lower position in the oceanic crust (possibly in the cumulate gabbro section of an ophiolite), and those eclogites with higher 18O and 87Sr/86Sr might have come from the upper portions of the oceanic crust (possibly in the sheeted dike to pillowed lava section of an ophiolite). This ancient oceanic crust was then subducted during the Archean and converted to eclogite and finally returned in pieces to the surface by alkalic magmas (389 m.y. ago; Snyder et al., 1993). Whereas only the low 18O component of such a sequence was seen in South African Bellsbank eclogites (Neal et al., 1990), Udachnaya eclogites may afford a look at an intact Archean ophiolite sequence from cumulate gabbros to sheeted dikes and pillowed lavas (e.g. McCulloch et al., 1981).
Four Group A Udachnaya eclogites, U-25, U-236, U-237, and U-281, with high-mg-number, high-Cr, and ultra-LREE-depleted garnets and clinopyroxenes (Tables 1 and 2) are probably mantle derived. The oxygen isotopic composition of whole rocks and minerals in these samples are well within the accepted mantle range (Snyder et al., 1995). Relatively nonradiogenic Sr [87Sr/86Sr(390 Ma) = 0.70298–0.70315] and depleted whole-rock Nd isotopic signatures [Nd(390 Ma) = +4.3 to +6.7] are also consistent with depleted upper-mantle derivation.
Mir eclogites
In contrast to the Udachnaya eclogites, xenoliths from the Mir kimberlite pipe exhibit convincing evidence of an oceanic crustal precursor. On the basis of Ca content, Beard et al.(1995, 1996) divided the Mir eclogites into low Ca (2.65–5.39 wt % CaO), intermediate Ca (5.31–9.09 wt % CaO), and high Ca (11.7–13.6 wt % CaO) varieties. The low Ca group also contains lower Al than the high Ca group. They also found that these groups were confirmed by trace-element chemistry and oxygen isotopes. The low Ca group have LREE in chondritic to enriched proportions and garnets with 18O values of 7.18–4.90, whereas the high Ca group have LREE which are extremely depleted relative to the HREE and 18O values in garnet of 5.40–3.09. Beard et al. (1995, 1996) showed that the mineral and trace-element, and oxygen isotopic chemistry of the Mir eclogites is consistent with an Archean ophiolite precursor. They interpreted the chemical differences in the low Ca, intermediate Ca, and high Ca groups as the result of the location of the protolith in the ophiolite sequence. Low-Ca eclogites could have come from the extrusive section of the ophiolite, containing basalts which were altered by seawater at low temperatures, whereas the high-Ca eclogites could have come from the plutonic, gabbroic cumulate section that underwent high-temperature seawater alteration.
Important oddities
Other Yakutian eclogites may represent melts of a downgoing oceanic crustal slab. Evidence for this is found in sample U-92 (Sobolev et al., 1995). This ‘eclogite’ is mainly composed of garnet and clinopyroxene, with minor amounts of high mg-number orthopyroxene and tiny inclusions of rutile and ilmenite. The compositions of garnets and clinopyroxenes vary from high Cr and high Mg on one side of the sample to low Cr and low Mg on the other. These characterisitics led Sobolev et al. (1995) to postulate that this sample may be a hybrid of eclogite and mantle peridotite. The process that caused the juxtaposition of these two normally distinct rock types could be similar to that proposed by Ringwood (1990) for the fertilization of peridotite mantle sources. In this model, melts of subducted oceanic crust, saturated with garnet and clinopyroxene, escape from the downgoing slab and are intruded into the overlying, depleted, peridotitic mantle. These melts react with the olivine-rich, magnesian mantle and change compositions by assimilation of peridotitic minerals. The hybrid melts precipitate clinopyroxene, garnet, and orthopyroxene. This hybridization process converts depleted peridotite to fertile garnet lherzolite and may have produced rare eclogitic rocks such as U-92.
|
Discussion |
|---|
|
TOPABSTRACTIntroductionAnalytical MethodsPetrographyMineral Chemistry and...Ion-Probe Trace-Element Mineral...Reconstructed Whole-Rock...Radiogenic Isotopic DataProtoliths for Eclogite...DiscussionSummaryREFERENCES |
|---|
Individual mineral ion-probe analyses, as well as radiogenic isotope data on ultrapure mineral separates, indicate little if any kimberlitic and/or metasomatic overprint for most Udachnaya eclogite xenoliths [Snyder et al. (1993)
, and above]. Furthermore, analyses of diamond-hosted garnet–clinopyroxene pairs in concert with the host minerals from five eclogites from the Udachnaya and Mir kimberlite pipes have produced conflicting results (Ireland et al., 1994; Taylor et al., 1996a). Some mineral inclusions are more LREE depleted than the host eclogite minerals, indicating possible metasomatism of the host, whereas other mineral inclusions are more LREE enriched than the host minerals (Taylor et al., 1996a). Thus, one cannot state with confidence that all eclogites have been metasomatized after diamond formation (Ireland et al., 1994). In fact, primary minerals within many of the host eclogites exhibit little, if any, evidence of pervasive metasomatism. Therefore, the mineral chemistry of such unmetasomatized eclogites is useful in determining the ultimate origins of these rocks.
One interesting feature of the garnet chemisty of Udachnaya eclogites does deserve further attention—the presence of two trends on the ternary garnet compositional plot in Fig. 3. Helmstaedt & Schulze (1989) labeled the group of data projecting from the almandine–spessartine apex the ‘metabasic trend’, and that projecting from the pyrope apex the ‘ultrabasic’ trend. They also showed that their two trends for eclogite xenoliths of the Colorado Plateau converged in the Group B field, in a similar manner to those from Udachnaya (Fig. 3). Their ‘metabasic’ trend was considered to be due to prograde metamorphism of lawsonite-bearing oceanic crust during subduction. The breakdown of lawsonite with rising temperature and increasing pressure released Ca and H2O leading to more grossular-enriched garnets. Helmstaedt & Doig (1975) hypothesized that the lawsonite-bearing eclogites of the Colorado Plateau were the low-temperature equivalents of rodingites (grossular- and clinopyroxene-bearing, metamorphosed gabbros) found in ophiolite terranes and grospydite (kyanite-bearing eclogite) xenoliths found in kimberlites. In their model, the ‘ultrabasic trend’ was formed by hydration and Na-metasomatism of garnet clinopyroxenite and websterite lenses and/or dikes in ultramafic rocks within subcontinental mantle above the subducted slab. We believe that this hypothesis for the origin of Udachnaya eclogites has merit, and needs further evaluation.
When interpreting the protoliths of very old eclogites (e.g. >2 Ga), a lack of typical, some might say classical, crustal evidence may not allow one to rule out oceanic crustal involvement. Snyder et al. (1995) and McCandless & Gurney (1996) have indicated that 87Sr/86Sr and carbon isotopic data may not be particularly useful for distinguishing very old oceanic crust. First, the Sr isotopic composition of Archean seawater was much lower than at present, and other Earth reservoirs did not have sufficient time to diverge in terms of radiogenic isotopes. Second, biogenic reservoirs, which could have contained low-13C material, were probably more restricted in distribution. Third, although Archean seafloor may have had more ridges, the oceans were shallower. This prevented seawater from reaching its critical point, leading to extensive fracture systems, and altering oceanic crust at depth (McCandless & Gurney, 1996). The lack of extensive hydrothermal fracture systems also would not have allowed development of low-13C vent biota common in the Proterozoic and Phanerozoic. In support of this hypothesis, McCandless & Gurney (1996) have pointed out that most eclogitic diamonds are considered to be Proterozoic in age and consequently exhibit a broad range in 13C isotopic composition. Only the Udachnaya eclogites, with a known Archean protolith, exhibit such a narrow range in 13C. Although reliable ages have not been forthcoming for the Mir eclogites, we might expect, based on higher 87Sr/86Sr and more variable 18O, that these would have Proterozoic ages. Thus, although the evidence is considered weak, one could make a case that most Udachnaya eclogites have oceanic crustal precursors.
Melting of eclogites and the TTG hypothesis
Felsic intrusive rocks are the dominant component of Archean granite–greenstone terranes world-wide. These rocks include tonalites, trondhjemites, and granodiorites (TTG) and are thought to have formed by modest degrees of partial melting of mafic material in the lower crust or upper mantle, leaving behind a residue of eclogite (Drummond & Defant, 1990; Rapp et al., 1991; Rapp, 1995). Ireland et al. (1994) proposed that eclogite xenoliths in kimberlites are residues of such a major Archean melting event. An important part of this model is that melting of the mafic protolith occurred in the garnet stability field. Such melting produces a liquid that is LREE enriched with low HREE abundances and a residue with the opposite characteristics. If eclogites are the residues from extraction of a TTG melt, the samples should have high HREE contents and be LREE depleted. However, many samples are HREE depleted relative to the MREE (Fig. 8a), which precludes these samples as being examples of restites from TTG extraction.
Evidence of ancient enrichment and/or metasomatism
Most Udachnaya clinopyroxenes plot along a linear array on a plot of Sr vs Nd contents (Fig. 12). This array corresponds to a Sr/Nd abundance ratio of 20, similar to that expected for mantle-derived magmas (e.g. O’Nions, 1987; Galer et al., 1989). However, eight samples plot above this array to higher Sr abundances and define a possible secondary array on the order of 100:1. Analyses of several diamond inclusion clinopyroxenes from Yakutia confirm these two arrays (Sobolev et al., 1996) and that these arrays were developed before diamond formation. Sample U-5 has the highest Sr abundance; this sample also yields the most nonradiogenic Nd isotopic composition (Nd = –22 at 390 Ma) of any Udachnaya eclogitic clinopyroxene. Furthermore, those eclogites which have clinopyroxenes with the highest Sr abundances and Sr/Nd 20, also have the lowest Nd values (Fig. 13). This is generally true not only for eclogites from Udachnaya, but also for those from Mir (two samples to date). The majority of the remaining eclogites indicate a vertical array in Fig. 13, suggesting little or no correlation of Sr abundance with Nd isotopic composition. However, those samples with the lowest Sr abundances (<100 p.p.m.) also exhibit the highest Nd values (+20 to +114).
|
|
Most of the Yakutian eclogites exhibit uniformly nonradiogenic 87Sr/86Sr, as well as extremely low 87Rb/86Sr ratios. Mir eclogites analyzed to date are notably depleted in Rb compared with the Udachnaya eclogites (Snyder et al., 1997
). Generally, those Udachnaya eclogites with the highest Nd values (Fig. 13; Table 9) also have the lowest 87Rb/86Sr. These observations lead us to conclude that Udachnaya eclogites can most easily be explained by mixing of at least two distinct components. One of these components has the characteristics of a depleted mantle with uniform Sr/Nd, low 87Sr/86Sr and radiogenic Nd isotopic composition, and the other has elevated Sr/Nd (by about a factor of five), lower to equivalent Rb/Sr, similarly low 87Sr/86Sr and a very nonradiogenic Nd isotopic composition. What is the lithologic nature of this Archean, high Sr/Nd, low Rb/Sr, and nonradiogenic component? We speculate that this component could be an ancient carbonatite magma and/or fluid. This fluid could have been produced within the mantle or been ‘sweated’ off a descending oceanic slab during subduction.
Depleted and enriched mantle signatures
To gain an understanding of the time-integrated chemistry of the protolith(s) of these eclogites, we plotted all Yakutian samples, including those from Obnazhennaya and Mir, on a conventional Sm–Nd isochron diagram (Fig. 14). For comparison, we have calculated the reconstructed whole-rock isotopic compositions of all samples using equal proportions of clinopyroxene and garnet [i.e. the isotopic study of Jacob et al. (1994) did not include modal proportions of the minerals] (Fig. 14). Three separate groups of data are shown on this plot. Most data from all three pipes lie along a linear array that yields an ‘age’ of 1.2 Ga and an initial Nd of +2.6, indicating derivation from a LREE-depleted (mantle?) reservoir. However, this array is probably a mixing array, and probably reflects only an average age and isotopic composition of the eclogites. Three other samples (including U-5, with the most nonradiogenic Nd isotopic signature) yield a similar ‘age’, but suggest a reservoir that had time-integrated extreme LREE enrichment and would have had an Nd of –25 at 1.2 Ga. These eclogites point to an old enriched component, similar to that proposed by Snyder et al. (1993). Further evidence of an enriched mantle component is seen in the extremely radiogenic 187Os/188Os for the Udachnaya eclogites (Pearson et al., 1995b) and the Os (at 2.9 Ga) of +82 ± 41. Finally, a third group in Fig. 14 suggests severe LREE depletion and, for a similar age, would yield extremely positive Nd values. The involvement of an old ultra-depleted mantle component beneath Yakutia is supported by Nd isotopic studies of peridotites which yield mid-Proterozoic Nd values of +23 (1.7 Ga) and +24.5 (1.54 Ga), and a garnet pyroxenite with an Nd of +20 at 2.6 Ga (McCulloch, 1989; Zhuravlev et al., 1991). Isotopic psystematics alone point to two extreme end-member components: one ultra-depleted and the other enriched.
|
|
Summary |
|---|
|
TOPABSTRACTIntroductionAnalytical MethodsPetrographyMineral Chemistry and...Ion-Probe Trace-Element Mineral...Reconstructed Whole-Rock...Radiogenic Isotopic DataProtoliths for Eclogite...DiscussionSummaryREFERENCES |
|---|
Eclogites from the Udachnaya kimberlite pipe, taken as a suite, are unique in that (1) the individual minerals exhibit little zonation and are relatively homogeneous both within grains and between grains; (2) clinopyroxenes are generally more LREE depleted, by over an order of magnitude, than either those from southern Africa or those from other Yakutian localities; (3) the Sr isotopic compositions of clinopyroxenes are, overall, much less radiogenic (87Sr/86Sr
0.705) than those from southern Africa (87Sr/86Sr 0.706), and slightly less radiogenic than those from the Mir kimberlite pipe; (4) garnet–clinopyroxene pairs from Udachnaya eclogites yield positive slopes on typical 147Sm/144Nd vs 143Nd/144Nd diagrams appoximating the age of the kimberlite, whereas those pairs from South Africa often yield negative slopes indicating possible disequilibrium; (5) minerals exhibit less oxygen, carbon, and sulfur isotopic variability than either those from South Africa or those from other Yakutian localities.
Extremely low measured 87Sr/86Sr values, as well as reconstructed whole-rock Nd and Os model ages, provide evidence for the antiquity of the Yakutian eclogite protoliths. However, the nature of this precursor, whether of crustal or mantle origin, is still debatable. We believe that at least two diverse sources for Yakutian eclogites, one mantle and one crustal, are supported by the present data set. Based particularly on the consistency of trends in major- and trace-element mineral chemistry with oxygen isotopic compositions that vary above and below mantle values, Beard et al. (1996) have convincingly argued that the Mir eclogites are dominantly of a subducted ophiolite origin. However, the Udachnaya eclogites have oxygen isotopic values which lie only within and slightly above the mantle range; no convincingly low 18O values have been found (Snyder et al., 1995). Also, no consistent mineral-chemical trends exist with oxygen isotopic values for the Udachnaya eclogites, contrary to the claims of Jacob et al. (1994).
This lack of correlation between mineral-chemical compositions and oxygen isotopes, along with the unzoned nature of the minerals, can be explained in three ways. The Udachnaya kimberlite lies near the center of the Siberian craton and, thus, could have tapped a source which came from deeper in the mantle keel (reflected in the higher temperatures) than that for the more peripheral Mir kimberlite. Thus, the Udachnaya eclogite xenoliths could have remained at much higher temperatures within the mantle for much longer times, leaving the minerals open for extensive isotopic, and trace- and major-element diffusion. This could have allowed more complete exchange and equilibration with the surrounding mantle, if, indeed, the original protolith was oceanic crust. Therefore, any oceanic crustal signature could have been masked by later overprinting. Conversely, the Udachnaya eclogites, although exhibiting some crustal affinities, contain a significant mantle component not found in the Mir eclogites. Again, this mantle component could have been derived from the original eclogite source or might have been introduced through diffusion or melt interaction in the mantle over time. Alternatively, the lack of variation in oxygen, carbon, and sulfur isotopic ratios in the Udachnaya eclogites, and their similarity to mantle-derived peridotitic values, may be reflecting the relative antiquity of the protolith. Archean oceanic environments may not have had the distribution of biogenic material needed to cause fractionation of light from heavy isotopes in these elements.
It appears that we must concur with the late Ted Ringwood that eclogites do indicate a ‘multiplicity of origins’.
|
ACKNOWLEDGEMENTS |
|---|
This study was supported by NSF Grants EAR 91–18043, EAR 93–04053, and EAR-9505930 to L.A.T., NSF Grants EAR 90–04133 and EAR 91–04877 to A.N.H., and NSF Grants EAR 90–17587 and EAR 93–16328 to G.C. Allan Patchen helped maintain the electron microprobe in top working order and assisted in data collection. We thank Nick Pokhilenko for his friendship and insightful discussions of mantle xenoliths. We also thank Emil Jagoutz, Roberta Rudnick, and Graham Pearson for their interest and for thoughtful and critical discussions. Discussions with Herb Helmstaedt helped to clarify some outstanding problems in eclogite genesis. A continuing correspondence with Tom McCandless on carbon and sulfur isotopes in diamonds and inclusions has made us aware of another body of literature left relatively neglected. Formal reviews by Steve Haggerty, J. G. ‘Louie’ Liou, and Editor Mike Rhodes helped to tighten the prose, focus the purpose, and balance the discussion. The result is a much improved paper; much appreciation to the reviewers. For their hospitality and the supply of diamond-bearing eclogites, we are deeply indebted to our colleagues at the Yakutian Diamond Company—Almazy Rossii Sakha; most notable in this regard is our good friend Vladimir ‘Hoss’ Tsiganov—‘the Russian Larry Taylor’.
|
FOOTNOTES |
|---|
‘It appears that eclogites found in diamond pipes may possess a multiplicity of origins and, when correctly interpreted, may provide guides to fractionation processes which have occurred within the mantle.’ A. E. Ringwood (1975)
* Corresponding author
|
REFERENCES |
|---|
|
TOPABSTRACTIntroductionAnalytical MethodsPetrographyMineral Chemistry and...Ion-Probe Trace-Element Mineral...Reconstructed Whole-Rock...Radiogenic Isotopic DataProtoliths for Eclogite...DiscussionSummaryREFERENCES |
|---|
Beard B. L., Snyder G. A., Taylor L. A., Fraracci K. N., Sobolev V. N., Clayton R. N., Mayeda T. K., Sobolev N. V. Eclogites from the Mir kimberlite, Russia: evidence of an Archean ophiolite protolith. In: Extended Abstracts of 6th International Kimberlite Conference (1995) Novosibirsk: Siberian Branch, Russian Academy of Sciences. 41–43.
Beard B. L., Fraracci K. N., Taylor L. A., Snyder G. A., Clayton R. N., Mayeda T. K., Sobolev N. V. Petrography and geochemistry of eclogites from the Mir kimberlite, Russia. Contributions to Mineralogy and Petrology (1996) (in press).
Bobrievich A. P., Smirnov G. I., Sobolev V. S. An eclogite xenolith with diamonds. Doklady Akademii Nauk SSSR (1959) 126:3. (in Russian).
Burgess R., Turner G., Harris J. W. 40Ar–39Ar laser probe studies of clinopyroxene inclusions in eclogitic diamonds. Geochimica et Cosmochimica Acta (1992) 56:389–402.[Web of Science]
Cameron M., Papike J. J. Crystal chemistry of silicate pyroxenes. Pyroxenes. Mineralogical Society of America, Reviews in Mineralogy—Prewitt C. T., ed. (1982) 7:5–92.
Caporuscio F. A., Smyth J. R. Trace element crystal chemistry of mantle eclogites. Contributions to Mineralogy and Petrology (1990) 105:550–561.[Web of Science]
Coleman R. G., Lee E. D., Beatty L. B., Brannock W. W. Eclogites and eclogites: their differences and similarities. Geological Society of America Bulletin (1965) 76:483–508.
Drummond M. S., Defant M. J. A model for trondhjemite–tonalite–dacite genesis and crustal growth via slab melting: Archean to modern comparisons. Journal of Geophysical Research (1990) 95:21503–21521.
Eldridge C. S., Compston W., Williams I. S., Harris J. W., Bristow J. W. Isotope evidence for the involvement of recycled sediments in diamond formation. Nature (1991) 353:649–653.
Ellis D. G., Green D. H. An experimental study of the effect of Ca upon garnet–clinopyroxene Fe–Mg exchange equilibria. Contributions to Mineralogy and Petrology (1979) 71:13–22.[Web of Science]
Fraracci K. N. Diamondiferous eclogite xenoliths from the Mir kimberlite, Yakutia, Siberia: geochemistry and petrogenesis. (1994) Knoxville: University of Tennessee. Unpublished M.Sc. Thesis.
Galer S. J. G., Goldstein S. L., O’Nions R. K. Limits on chemical and convective isolation in the Earth’s interior. Chemical Geology (1989) 75:257–290.[Web of Science]
Gregory R. T., Taylor H. P. Possible non-equilibrium oxygen isotope effects in mantle nodules, an alternative to the Kyser–O’Neil–Carmichael 18O/16O geothermometer. Contributions to Mineralogy and Petrology (1986a) 93:114–119.[Web of Science]
Gregory R. T., Taylor H. P. Non-equilibrium, metasomatic 18O/16O effects in upper mantle mineral assemblages. Contributions to Mineralogy and Petrology (1986b) 93:124–135.[Web of Science]
Haggerty S. E., Fung A. T., Burt D. M. Apatite, phosphorus and titanium in eclogitic garnet from the upper mantle. Geophysical Research Letters (1994) 21:1699–1702.[Web of Science]
Harte B., Kirkley M. B. Clinopyroxene–garnet partition coefficients for REE, Sr, and Y, and their relationship to major–minor element composition. Mineralogical Magazine (1994) 58A:388–389. Abstracts V. M. Goldschmidt Conference.
Hauri E. H., Wagner T. P., Grove T. L. Experimental and natural partitioning of Th, U, Pb and other trace elements between garnet, clinopyroxene and basaltic melts. Chemical Geology (1994) 117:149–166.[Web of Science]
Helmstaedt H., Doig R. Eclogite nodules from kimberlite pipes of the Colorado Plateau—samples of subducted Franciscan-type oceanic lithosphere. Physics and Chemistry of the Earth (1975) 9:95–111.
Helmstaedt H., Schulze D. J. Eclogite-facies ultramafic xenoliths from Colorado Plateau diatrem breccias: comparison with eclogites in crustal environments, evaluation of the subduction hypothesis, and implications for eclogite xenoliths from diamondiferous kimberlites. In: Eclogite-Facies Rocks—Smith D., ed. (1989) Amsterdam: Elsevier.
Ireland T. R., Rudnick R. L., Spetsius Z. Trace elements in diamond inclusions from eclogites reveal link to Archean granites. Earth and Planetary Science Letters (1994) 128:199–213.[Web of Science]
Jacob D., Jagoutz E., Lowry D., Mattey D., Kudrjavtseva G. Diamondiferous eclogites from Siberia: remnants of Archean oceanic crust. Geochimica et Cosmochimica Acta (1994) 58:5191–5207.[Web of Science]
Jagoutz E. Sm–Nd systematics in eclogite from Siberia. In: Fourth International Kimberlite Conference, Extended Abstracts (1986) Perth: Geological Society of Australia. 265–266.
Jagoutz E., Dawson J. B., Hoernes S., Spettel B., Wanke H. Anorthositic oceanic crust in the Archean Earth (extended abstract). In: 15th Lunar and Planetary Science Conference (1984) Houston, TX: Lunar and Planetary Institute. 395–396.
Jerde E. A., Taylor L. A., Crozaz G., Sobolev N. V. Exsolution of garnet within clinopyroxene of mantle eclogites: major- and trace-element chemistry. Contributions to Mineralogy and Petrology (1993a) 114:148–159.[Web of Science]
Jerde E. A., Taylor L. A., Crozaz G., Sobolev N. V. Diamondiferous eclogites from Yakutia, Siberia: evidence for a diversity of protoliths. Contributions to Mineralogy and Petrology (1993b) 114:189–202.[Web of Science]
Johnson K. T. M. Experimental cpx/ and garnet/melt partitioning of REE and other trace elements at high pressures: petrogenetic implications. Mineralogical Magazine (1994) 58A:454–455. Abstracts V. M. Goldschmidt Conference.
Lee D.-C., Halliday A. N., Hunter R. H., Holden P., Upton B. G. J. Rb–Sr and Sm–Nd isotopic variations in dissected crustal xenoliths. Geochimica et Cosmochimica Acta (1993) 57:219–230.[Web of Science]
Ludwig K. L. ISOPLOT: a plotting and regression program for radiogenic-isotope data, for IBM-PC compatible computers, version 2.01. US Geological Survey Open-file Report (1990) 88–557.
Lundberg L. L., Crozaz G., McKay G., Zinner E. Rare earth element carriers in the Shergotty meteorite and implications for its chronology. Geochimica et Cosmochimica Acta (1988) 52:2147–2163.[Web of Science]
Lundberg L. L., Crozaz G., McSween H. Y. Rare earth elements in minerals of the ALHA77005 shergottite and implications for its parent magma and crystallization history. Geochimica et Cosmochimica Acta (1990) 54:2535–2547.[Web of Science]
MacGregor I. D., Manton W. I. Roberts Victor eclogites: ancient oceanic crust. Journal of Geophysical Research (1986) 91:14063–14079.
McCandless T. E., Gurney J. J. Diamond eclogites: comparison with carbonaceous chondrites, carbonaceous shales, and microbial carbon-enriched MORB. In: Proceedings of the Sixth International Kimberlite Conference (1996) in press.
McCulloch M. T. Sm–Nd systematics in eclogite and garnet peridotite nodules from kimberlites: implications for the early differentiation of the earth. In: Kimberlites and Related Rocks, Vol. 2—Ross J., ed. (1989) Perth: Geological Society of Australia. 864–876.
McCulloch M. T., Gregory R. T., Wasserburg G. J., Taylor H. P. Jr. Sm–Nd, Rb–Sr, and 18O/16O isotopic systematics in an oceanic crustal section: evidence from the Samail ophiolite. Journal of Geophysical Research (1981) 86:2721–2735.
McKay G., Wagstaff J., Yang S.-R. Clinopyroxene REE distribution coefficients for shergottites: the REE content of the Shergotty melt. Geochimica et Cosmochimica Acta (1986) 50:927–937.[Web of Science]
Neal C. R., Taylor L. A., Davidson J. P., Holden P., Halliday A. N., Nixon P. H., Paces J. B., Clayton R. N., Mayeda T. K. Eclogites with oceanic crustal and mantle signatures from the Bellsbank kimberlite, South Africa, Part 2: Sr, Nd, and O isotope geochemistry. Earth and Planetary Science Letters (1990) 99:362–379.[Web of Science]
O’Nions R. K. Relationships between chemical and convective layering in the Earth. Journal of the Geological Society London (1987) 144:259–274.
O’Reilly S. Y., Griffin W. L. Trace-element partitioning between garnet and clinopyroxene in mantle-derived pyroxenites and eclogites: P–T–X controls. Chemical Geology (1995) 121:105–130.[Web of Science]
Ohtani E. The primordial terrestrial magma ocean and its implication for stratification of the mantle. Physics of the Earth and Planetary Interiors (1985) 38:70–80.[Web of Science]
Pearson D. G., Shirey S. B., Carlson R. W., Boyd F. R., Pokhilenko N. P., Shimizu N. Re–Os, Sm–Nd, and Rb–Sr isotope evidence for thick Archaean lithospheric mantle beneath the Siberian craton modified by multistage metasomatism. Geochimica et Cosmochimica Acta (1995a) 59:959–977.[Web of Science]
Pearson D. G., Snyder G. A., Shirey S. B., Taylor L. A., Carlson R. W., Sobolev N. V. Re–Os isotope evidence for a mid-Archean age of diamondiferous eclogite xenoliths from Siberia and constraints on Archean tectonics. Nature (1995b) 374:711–713.
Pokhilenko N. P., Sobolev N. V., Yefimova Y. S. Xenolith of deformed diamond-bearing kyanite eclogite from the Udachnaya pipe, Yakutia. Doklady Akademii Nauk SSSR (1982) 266:212–216. (in Russian).
Ponomarenko A. I., Sobolev N. V., Pokhilenko N. P., Lavrent’yev Y. G., Sobolev V. S. Diamond-bearing grospydite and diamond-bearing kyanite eclogites from the Udachnaya kimberlite pipe, Yakutia. Doklady Akademii Nauk SSSR (1976) 226:927–930. (in Russian).
Ponomarenko A. I., Spetsius Z. V., Sobolev N. V. New type of diamond-bearing rock—garnet pyroxenite (in Russian). Doklady Akademii Nauk SSSR (1980) 251:438–441.
Qi Q., Taylor L. A., Snyder G. A., Sobolev N. V. Eclogites from the Obnazhennaya kimberlite pipe, Yakutia, Russia. International Geological Review (1994) 36:1023–1037.
Rapp R. P. Is eclogite in the sub-continental lithosphere the residue from melting of subducted oceanic crust? Experimental constraints and implications for the origin of the Archean continents (extended abstract). In: Abstracts of 6th International Kimberlite Conference (1995) Novosibirsk: Siberian Branch, Russian Academy of Sciences. 457–459.
Rapp R. P., Watson E. B., Miller C. F. Partial melting of amphibolite/eclogite and the origin of Archean trondhjemites and tonalites. Precambrian Research (1991) 51:1–25.[Web of Science]
Ringwood A. E. Composition and Petrology of the Earth’s Mantle (1975) New York: McGraw–Hill. 618.
Ringwood A. E. Slab–mantle interactions: 3. Petrogenesis of intraplate magmas and structure of the upper mantle. Chemical Geology (1990) 82:187–207.[Web of Science]
Ringwood A. E., Green D. H. An experimental investigation of the gabbro–eclogite transformation and some geophysical implications. Tectonophysics (1967) 3:383–427.[Web of Science]
Rudnick R. L., Green D. H. Petrological and geochemical constraints on the origin of eclogites: implications for the composition of ancient oceanic crust. Transactions, American Geophysical Union (1994) 75:61.
Rudnick R. L., Eldridge C. S., Bulanova G. P. Diamond growth history from in situ measurement of Pb and S isotopic compositions of sulfide inclusions. Geology (1993) 21:13–16.
Schulze D. J., Helmstaedt H. Coesite–sanidine eclogites from kimberlite: products of mantle fractionation or subduction? Journal of Geology (1988) 96:435–443.[Web of Science]
Shervais J. W., Taylor L. A., Lugmair G. W., Clayton R. N., Mayeda T. K., Korotev R. L. Early Proterozoic oceanic crust and the evolution of subcontinental mantle: eclogites and related rocks from southern Africa. Geological Society of America Bulletin (1988) 100:411–423.
Shimizu N., Sobolev N. V. Young peridotitic diamonds from the Mir kimberlite pipe. Nature (1995) 375:394–397.
Smyth J. R., Caporuscio F. A., McCormick T. C. Mantle eclogites: evidence of igneous fractionation in the mantle. Earth and Planetary Science Letters (1989) 93:133–141.[Web of Science]
Snyder G. A., Jerde E. A., Taylor L. A., Halliday A. N., Sobolev V. N., Sobolev N. V. Nd and Sr isotopes from diamondiferous eclogites, Udachnaya kimberlite pipe, Yakutia, Siberia: evidence of differentiation in the early Earth? Earth and Planetary Science Letters (1993) 118:91–100.[Web of Science]
Snyder G. A., Lee D.-C., Taylor L. A., Halliday A. N., Jerde E. A. Evolution of the upper mantle of the Earth’s moon: Nd and Sr isotopic constraints from high-Ti mare basalts. Geochimica et Cosmochimica Acta (1994) 58:4795–4808.[Web of Science]
Snyder G. A., Taylor L. A., Jerde E. A., Clayton R. N., Mayeda T. K., Deines P., Rossman G. R., Sobolev N. V. Archean mantle heterogeneity and the origin of diamondiferous eclogites, Siberia: evidence from stable isotopes and hydroxyl in garnets. American Mineralogist (1995) 80:799–809.[Abstract]
Snyder G. A., Taylor L. A., Beard B. L., Halliday A. N., Sobolev N. V. A depth profile through a subducted Archean ophiolite sequence: Nd and Sr isotopic chemistry of eclogites from the Mir kimberlite. (1997) Siberia. In preparation.
Sobolev N. V. Deep-seated Inclusions in Kimberlites and the Problem of the Composition of the Upper Mantle (1977) Washington, DC: American Geophysical Union. 304.
Sobolev N. V., Bakumenko I. T., Yefimova E. S., Pokhilenko N. P. Morphological features of microdiamonds, sodium in garnet, and potassium in clinopyroxenes: contents of two eclogite xenoliths from the Udachnaya Kimberlite pipe (Yakutia) (in Russian). Doklady Akademii Nauk SSSR (1991a) 321:585–591.
Sobolev N. V., Zuev V. M., Bezborodov S. M., Ponomarenko A. I., Spetsius Z. V., Kuligin S., Yefimova E. S., Afanasiev V. P., Koptil V. I., Botkunov A. I. Eclogite paragenesis of diamonds from Udachnaya and Mir pipes, Yakutia. In: Proceedings of the Fifth International Kimberlite Conference (1991b) Araxa, Brazil: CPRM. 391. (extended abstracts).
Sobolev N. V., Yefimova E. S., Taylor L. A., Snyder G. A., Sobolev V. N. Major and trace element variations in multiple eclogitic inclusions in individual diamonds from Yakutia, Russia. Transactions, American Geophysical Union (1996) 77. in press.
Sobolev V. N., Taylor L. A., Snyder G. A., Sobolev N. V. Diamondiferous eclogites from the Udachnaya kimberlite pipe, Yakutia, Siberia. International Geological Review (1994) 36:42–64.
Sobolev V. N., Taylor L. A., Snyder G. A., Sobolev N. V. Diamondiferous eclogites from the Siberian platform: samples with peridotitic signatures? In: Proceedings of the 6th International Kimberlite Conference (1995) Novosibirsk: Siberian Branch, Russian Academy of Sciences. 552–554.
Sobolev V. S. Formation conditions of diamond deposits (in Russian). Geologiya i Geofizika (1960) 1:7–22.
Taylor L. A. Evolution of the subcontinental mantle beneath the Kaapvaal craton: a review of evidence for crustal subduction for Bellsbank eclogites. The V. S. Sobolev Memorial Volume, Russian Geology and Geophysics (1993) 34:21–39.
Taylor L. A., Neal C. R. Eclogites with oceanic crustal and mantle signatures from the Bellsbank kimberlite, South Africa, Part I: Mineralogy, petrography, and whole rock chemistry. Journal of Geology (1989) 97:551–567.[Web of Science]
Taylor L. A., Snyder G. A., Sobolev V. N., Crozaz G., Sobolev N. V. Eclogitic inclusions in diamonds: complex mantle processes over time. Earth and Planetary Science Letters (1996a) 142:535–551.[Web of Science]
Taylor L. A., Valley J. W., Clayton R. N., Mayeda T. K., Snyder G. A., Sobolev V. N., Sobolev N. V. Oxygen isotopes by laser-heating and conventional techniques: a study of Siberian diamondiferous eclogites. In: Proceedings of the 30th International Geological Congress (1996b) Beijing, China. 359.
Thompson R. N. Some high-pressure pyroxenes. Mineralogical Magazine (1974) 39:768–787.[Web of Science]
Wood B. J., Banno S. Garnet–orthopyroxene and orthopyroxene–clinopyroxene relationships in spinel and complex systems. Contributions to Mineralogy and Petrology (1973) 41:109–124.
York D. Least-squares fitting of a straight line with correlated errors. Earth and Planetary Science Letters (1969) 5:320–324.[Web of Science]
Zhuravlev A. Z., Laz’ko E. E., Ponomarenko A. I. Radiogenic isotopes and REE in garnet peridotite xenoliths from the Mir kimberlite pipe, Yakutia. Geokhimiya (1991) 7:982–984.
Zindler A., Jagoutz E. Mantle cryptology. Geochimica et Cosmochimica Acta (1988) 52:319–333.[Web of Science]
Zinner E., Crozaz G. A method for the quantitative measurement of rare earth elements in the ion microprobe. International Journal of Mass Spectrometry and Ion Processes (1986) 69:17–38.[Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  T. Usui, E. Nakamura, K. Kobayashi, S. Maruyama, and H. Helmstaedt Fate of the subducted Farallon plate inferred from eclogite xenoliths in the Colorado Plateau Geology, July 1, 2003; 31(7): 589 - 592. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Jacob, E. Jagoutz, D. Lowry, and E. Zinngrebe Comment on 'The Origins of Yakutian Eclogite Xenoliths' by G. A. Snyder,L. A. Taylor, G. Crozaz, A. N. Halliday, B. L. Beard, V. N. Sobolev and N. V. Sobolev J. Petrology, August 1, 1998; 39(8): 1527 - 1533. [Full Text] [PDF]
|
|
|
|
|
|
G. A. Snyder, L. A. Taylor, B. L. Beard, G. Crozaz, A. N. Halliday, V. N. Sobolev, and N. V. Sobolev Reply to a Comment by D. Jacob et al. on 'The Origins of Yakutian Eclogite Xenoliths' J. Petrology, August 1, 1998; 39(8): 1535 - 1543. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||