V33E-1496 1340h
Re-Os systematics of komatiites and komatiitic basalts at Dundonald Beach, Ontario, Canada: Evidence for a complex alteration history and implications of a late-Archean chondritic mantle source
Re-Os concentrations and isotopic compositions have been examined in one komatiite unit and one komatiitic basalt unit at Dundonald Beach, which is part of the spatially-extensive 2.7 Ga Kidd-Munro volcanic assemblage in the Abitibi greenstone belt, Ontario, Canada. The komatiitic rocks in this locality record at least three episodes of alteration of Re-Os elemental and isotope systematics. First, an average of 40% and as much as 75% Re was lost due to shallow degassing during eruption and/or hydrothermal leaching during or immediately after the lava emplacement. Second, the Re-Os isotope systematics of the rocks with $^{187}$Re/$^{188}$Os ratios $>$1 were reset at $\sim$2.5 Ga, most likely due to a regional metamorphic event. Finally, there is evidence for relatively recent gain and loss of Re. The variations in Os concentrations in the Dundonald komatiites yield a relative bulk distribution coefficient for Os (D$_{Os}$ $^{solid/liquid}$) of 2-4, consistent with those obtained for stratigraphically-equivalent komatiites in the nearby Alexo area and in Munro Township. This suggests that Os was moderately compatible during crystal-liquid fractionation of the magma parental to the Kidd-Munro komatiitic rocks. Furthermore, whole-rock samples and chromite separates with low $^{187}$Re/$^{188}$Os ratios ($<$1) yield a precise chondritic average initial $^{187}$Os/$^{188}$Os ratio of 0.1083 $\pm$ 0.0006 (\gamma$_{Os}$ = 0.0 $\pm$ 0.6). The chondritic initial Os isotopic composition of the mantle source for the Dundonald rocks is consistent with that determined for komatiites in the Alexo area and in Munro Township. Our Os isotope results for the Dundonald komatiitic rocks, combined with those in the Alexo and Pyke Hill areas suggest that the mantle source region for the Kidd- Munro volcanic assemblage had evolved along a long-term chondritic Os isotopic trajectory until their eruption at $\sim$2.7 Ga. The chondritic initial Os isotopic composition of the Kidd-Munro komatiites is indistinguishable from that of the projected contemporaneous convective upper mantle. The uniform chondritic Os isotopic composition of the $\sim$2.7 Ga mantle source for the Kidd-Munro komatiites contrasts with the typical large-scale Os isotopic heterogeneity in the mantle sources for komatiites from the Gorgona Island, present-day ocean island basalts or arc-related lavas. This suggests a significantly more homogeneous mantle source in the Archean compared to the presentday mantle.
V33E-1497 1340h
Layered Mantle Structure Beneath the Western Guyana Shield, Venezuela: Evidence From Diamonds and Xenocrysts in Guaniamo Kimberlites
The Neo-Proterozoic (712 Ma) kimberlites from Guaniamo, Venezuela contain abundant low-Ca (harzburgitic) Cr-pyrope garnet xenocrysts, as well as lherzolitic Cr-pyropes, but few diamonds belonging to the peridotite-suite. Trace element compositions suggest three groupings of garnet. Most Cr-pyropes, both lherzolitic and harzburgitic, are highly depleted in magmaphile elements (e.g., Y $<$ 5 ppm, Zr $<$ 31 ppm) and have sinusoidal REE patterns. A second group (primarily lherzolites) has relatively undepleted characteristics (e.g., 9 - 14 ppm Y, Zr $<$ 22 ppm), and typical LREE-depleted REE patterns. A few garnets show enrichment typical of peridotites metasomitised by relatively low-temperature mantle fluids (e.g., 7 - 9 ppm Y, 30 - 90 ppm Zr). Ni-in-garnet thermometry allows depths of origin to be estimated for the garnet xenocrysts by projecting temperatures onto the conductive geothermal gradient (40 Mw/m2 heat flow equivalent) obtained from compositions of Guaniamo Cr-diopside xenocrysts using the Nimis and Taylor method. Within the limited depth range of lithosphere sampled by the kimberlites (100-150 km) compositions vary considerably, but regularly, defining a strongly layered mantle section. Major and trace element compositions suggest the following lithologic sequence: highly-depleted lherzolite from 100 to 115 km, mixed ultra-depleted harzburgite and lherzolite from 115 to 120 km, relatively undepleted lherzolite from 120 to 135 km, mixed depleted harzburgite and relatively-undepleted lherzolite from 135 to 150 km. Based on comparison with well-documented mantle peridotites and xenocrysts from elsewhere, we conclude that the Meso-Proterozoic Cuchivero Province of the Guyana Shield (host to the Guaniamo kimberlites) is underlain by depleted and ultra-depleted shallow Archean mantle which was under-plated, and uplifted, by Proterozoic subduction, perhaps more than once. These Proterozoic subduction events introduced less-depleted oceanic lithosphere beneath the Archean section, which remained there and was the source of the abundant Guaniamo eclogite-suite diamonds that have ocean-floor carbon and oxygen isotope signatures. Although diamond-indicative low-Ca Cr-pyropegarnets are abundant, they were derived primarily from the shallow depleted layer within the field of graphite stability, and the rare peridotite-suite diamonds were either metastably preserved at these shallow depths, or were derived from the small amount of depleted lithosphere sampled by these kimberlites that remained within the diamond stability field (the mixture of Archean and Proterozoic mantle in the depth range 135-150 km).
V33E-1498 1340h
Peraluminous websterite and granulite xenoliths from the Chyulu Hills volcanic field, Kenya: Plagioclase-rich cumulates re-equilibrated at uppermost mantle and crustal conditions?
Basanites of the Chyulu Hills volcanic field, Kenya, contain a suite of meta-igneous peraluminous spinel-garnet olivine websterite, Mg-Al sapphirine-bearing and Ca-Al hibonite-bearing granulite xenoliths. The websterites are the most mafic and magnesian members of this sequence. The Mg-Al sapphirine-bearing granulites are more Si- and Al-rich and less magnesian. They consist of ortho- and clinopyroxene, corundum, spinel, sapphirine, sillimanite, plagioclase and garnet. The Ca-Al hibonite-bearing granulites are most enriched in Si and Al. They are dominated by clinopyroxene and plagioclase and may contain hibonite closely associated with spinel, mullite, sapphirine and sillimanite. Hibonite, which is very rare in terrestrial rocks, is the earliest mineral in the crystallization sequence. All rocks are poor in REE, HFSE and Cr and enriched in LILE and Ni. They follow a magmatic fractionation trend and form a cumulate sequence finally equilibrated in the range of uppermost mantle (websterites) and crustal (granulites) depths. The websterites could have formed by high-pressure metamorphism of low-pressure troctolite-like cumulates and appear similar to some websteritic lithologies from exhumed high-pressure ultramafic complexes (Kornprobst, 1990; Morishita et al., 2003). The final P-T conditions for most websterites correspond to 920-1000 C / 17-22 kbar, whereas the final equilibration in granulites occurred at ca. 600-740 C / <8 kbar, the pressure limit being defined by the stability field of sillimanite. The geodynamic interpretation of spatially close granulite terranes (Moeller et al., 1998) coupled with the petrological evidence suggest that the studied rocks were metamorphosed and may have formed in the environment of a Pan-African active continental margin. Moeller, A., Mezger, K., Schenk, V. (1998). Journal of Petrology 39, 749-783. Kornprobst, J., Piboule, M., Roden, M., Tabit, A. (1990). Journal of Petrology 31, 717-745. Morishita, T., Arai, S., Gervilla, F., Green, D. (2003). Geochimica et Cosmochimica Acta 67, 303-310.
V33E-1499 1340h
Mantle metasomatism to form kosmochlor-bearing diopside in peridotite xenoliths from North Island, New Zealand
Pale green Cr- and Na-rich diopside (kosmochlor-bearing diopside) was observed in anhydrous Group I mantle xenoliths (dunite, wehrlite, harzburgite, clinopyroxenite) hosted by Pliocene-Quaternary hawaiite from the Ngatutura volcanic field of the North Island, New Zealand. The diopside is characterized by high Cr2O3 (up to 5.87 wt%) and Na2O (up to 3.31 wt%) and low Al2O3 (1.09 wt%, on average), and contains the highest kosmochlor component (ca. 18 mol%) known for diopsides in spinel peridotite xenoliths (Ikehata and Arai, Am Min. in press). It commonly rims around partly dissolved chromian spinel and replaces secondary Al-poor (less than 0.57 wt% Al2O3) and Si-rich (around 58.1 wt% SiO2) orthopyroxene crystallized from slab-derived fluid/melt at the expense of olivine in dunite. It also occurs as a marginal part of zoned diopside in a wehrlite xenolith. The mode of occurrence indicates a metasomatic origin for the Ko-bearing diopside. The kosmochlor-bearing diopside was analyzed for 33 trace elements by LA-ICP-MS. In all rock types it shows high REE concentrations, is slightly enriched in LREE relative to HREE, and is strongly depleted in Hf, Zr, Ta, Nb, and Ti relative to REE. These characteristics are similar to those of diopsides produced by carbonatite metasomatism in the upper mantle (e.g. Neumann et al. 2002). The Ko-poor diopside part of the zoned diopside in wehrlite is similar in trace element concentrations to the Ko-bearing one. The kosmochlor-bearing diopside may have been formed through metasomatism by a Na-bearing carbonatite melt, which selectively reacts with orthopyroxene and can carry a large amount of REE. Cr-rich spinel in refractory peridotites that had been formed at an arc setting is prerequisite for genesis of this kind of kosmochlor-bearing diopside upon carbonatite metasomatism. The carbonatite melt impacted the mantle wedge at a supra-subduction zone setting until 15 Ma, from the deeper part due to the absence of slab cover at the intraplate setting (10 Ma-) in the North Island. Ongoing work examines the importance of carbonatite metasomatism within the mantle wedge.
V33E-1500 1340h
Trace Element and Isotopic (Re-Os, O) Systematics of Roberts Victor Eclogites: Evidence for 3 Ga Subduction-Incorporation of Archean Oceanic Lithosphere into the South African Kaapvaal Craton Keel
Eclogite xenoliths from the 125 Ma old, Group II, Roberts Victor kimberlite have long been of interest in studies of Kaapvaal craton evolution because of their diversity, abundance, availability in large size, occurrence with peridotite and their sometimes high carbon/diamond content. From among the coesite, corundum, kyanite, Ca-, Mg-, and Fe- rich eclogites available, we have chosen to work on those that can be classified as Group I, Group II or diamondiferous with the goal to better understand their petrogenesis, the evolution of the Kaapvaal craton keel, the role of eclogites in diamond formation and the behavior of the Re-Os system. Group I vs Group II eclogites can be distinguished by texture (isolated gt in a cpx matrix vs subhedral, interlocking gt-cpx) and mineral chemistry (higher Na$_{2}$O in gt and K$_{2}$O in cpx for GI vs lower for GII). Such differences have been thought to result from higher vs lower pressures of equilibration. Recent laser fluorination oxygen isotopes ($\delta^{18}$O) on gt (GI = 5.8 to 6.9; GII = 2.1 to 5.1), ion-probe trace elements (e.g. Ce chondrite normalized) on gt (G1 = 0.2 to 0.5; GII = 0.002 to 0.07) and cpx (G1 = 7 to 20; GII = 0.2 to 2) and whole-rock Re-Os (G1 Re = 0.19 to 3.41 ppb; GII Re = 0.006 to 0.38 ppb) highlight even more distinct differences between Groups I and II. These differences are likely a pre-metamorphic signature of their original protoliths and not just due to pressure differences or magmagenesis during emplacement into the lithosphere. Using the stratigraphic variation of O isotopic composition and trace element content of ophiolites and drill core from DSDP/ODP holes 735B and 504B as a guide, Group I eclogites might represent the volcanic rocks of Layer 2 of Archean oceanic crust whereas Group II might represent the cumulate, intrusive rocks of Layer 3. This idea is supported by the presence, only in Group II eclogites, of positive Eu anomalies in reconstructed ion-probe whole rock rare earth element patterns. The Re-Os sytematics of the oceanic lithosphere is poorly known, especially in the Archean, but Roberts Victor eclogite Re-Os and trace element abundances and major element compositions suggest a basaltic komatiitic protolith as might typify slightly hotter ocean ridges in the Archean. A U-Pb age of 3.061$\pm$0.006 Ga on zircon grains separated from a Group I Roberts Victor eclogite and a same-age but scattered whole-rock Re-Os isotope array containing the diamondiferous and some Group I eclogites (including the zircon-bearing eclogite), firmly date the eclogite protoliths as Meso-Archean. The direct covariation of Re content (and hence $^{187}$Re/$^{188}$Os) with O isotopic composition allows the low-T alteration process occurring on the seafloor to be firmly dated at that time. For Kaapvaal craton evolution, this age is interesting because it predates by about 100 Ma the terrane collision that sutured the Kimberley block to the Eastern Kaapvaal along the Colesburg magnetic lineament, the stabilization of a thickened lithospheric mantle keel, and the generation of a widely distributed suite of eclogitic diamonds. Incorporation of these eclogites in the lithosphere is further evidence for Meso-Archean plate tectonics and the role of subduction in cratonization and diamond genesis.
V33E-1501 1340h
Mantle End-Members: The Trace Element Perspective
On the basis of their isotopic composition, ocean island basalts (OIB) have been classified into three to four end-members; HIMU with the most radiogenic Pb isotope ratios of OIB and Enriched Mantle 1 and 2 (EM1, EM2) with less radiogenic but variable Pb isotope and highly radiogenic Sr isotope signatures. It has also been argued that each of these isotopic families has common trace element characteristics that distinguish them from one another and so substantiated this classification. Here, we present new high-precision trace element data for samples from St. Helena, Tristan da Cunha and Gough in the Atlantic Ocean. The overall data-set is augmented by OIB data from the GEOROC database and includes data from all major isotopic families (HIMU: St. Helena, Mangaia, Tubuai, and Rururtu; EM1: Tristan da Cunha, Gough, Pitcairn; and EM2: Samoa, Marquesas, and Society). For each locality we use only islands defining the most extreme isotopic compositions. The entire data-set has been screened to exclude altered and highly differentiated samples. HIMU basalts have a very uniform trace element composition. Compared to HIMU-type basalts, EM-type basalts are enriched in Rb, Ba, and K, and depleted in U, Nb, and Ta, relative to La. Different EM-type OIBs from the same isotopic family (EM1 or EM2), have distinct trace element characteristics that can ultimately only be caused by different source compositions. For example, Ba/Th ratios in samples from both Tristan da Cunha (EM1) and Samoa (EM2) are similarly high (ca. 110) whereas Ba/Th ratios in samples from Pitcairn (EM1) and Society (EM2) samples are consistently lower (ca. 70). Thus on the basis of their trace element composition, EM-type OIB cannot be classified into EM1 and EM2 type basalts, nor can any other grouping be identified. The remarkably uniform isotopic and trace element composition of HIMU-type basalts suggests derivation from a single common source reservoir, most likely subduction-modified oceanic crust. Although there are some trace element characteristics common to all EM-type basalts, which distinguish them from HIMU-type basalts (e.g. uniformly high Th/U ratios of 4.7 $\pm$ 0.3, and enrichment in Cs-U), each suite of EM-type basalts has unique trace element signatures that distinguish them from any other suite of EM-type basalts. This is especially obvious when comparing the trace element composition of EM basalts from one isotopic family, for example EM1-type basalts from Tristan, Gough and Pitcairn. Consequently, the trace element systematics of EM-type basalts suggest that there are many different EM-type sources, whereas the isotopic composition of EM-type basalts suggest derivation from two broadly similar sources, i.e. EM1 and EM2. The large variability in subducting sediments with respect to both parent-daughter (e.g. Rb/Sr, Sm/Nd, U/Pb, Th/Pb,...) and other trace element ratios makes it unlikely that there are reproducible mixtures of sediments leading to two different isotopic evolution paths (EM1 and EM2) while preserving a range of incompatible element contents for each isotopic family, as would be required to reconcile the isotopic and trace element characteristics of EM-type basalts. Although this does not a priori argue against sediments as possible source components for OIB, it does argue against two distinct groups of sediments as EM1 and EM2 sources. Further characterization of sources with the same general origin (e.g. a certain type of crust or lithosphere) or identification of processes leading to reservoirs with similar parent-daughter ratio characteristics but different incompatible trace element contents could resolve the apparent conundrum.