Core-mantle Exchange Research Articles

Origins of Os-isotope and platinum-group element compositions of metasomatized peridotite and cumulate pyroxenite xenoliths from Kharchinsky Volcano, Kamchatka

Platinum-group element (PGE) + Re abundances and Os isotopic compositions were determined for metasomatized peridotite and cumulate pyroxenite xenoliths from Kharchinsky Volcano, Kamchatka. Two peridotites have moderately high PGE concentrations with 2.9–3.3 ppb Os and smoothly variable PGE patterns similar to abyssal peridotites and encompassing primitive and depleted mantle compositions. One of these has unradiogenic Os (187Os/188Os = 0.1148) at the lower limit of compositions observed in abyssal peridotite which results in a time of Re depletion (TRD) model age of ∼1.8 Ga. These contrasts Kharchinsky peridotites that are Pt-Ir enriched with sharply fractionated PGE patterns with Pt/Pd = 132–168 and Ir/Ru ∼ 4.0. This pattern of enrichment in the heavier and more highly siderophile Pt and Ir over the lighter and more siderophile-chalcophile Ru and Pd (Fleet and Stone, 1991; Fleet et al., 1991) is consistent with sulfur depletion under oxidizing metasomatic conditions that leave a residue with PGE host minerals that are dominantly alloys. A single Ru-enriched peridotite with low PGE abundances is interpreted to reflect advanced stages of paragenesis driven by metasomatism and resulting in the loss of all PGE and Re relative to Ru. All Kharchinsky peridotites included in this study have relatively unradiogenic Os (187Os/188Os = 0.1148–0.1314) but some are coupled to PGE and lithophile element characteristics that require oxidizing metasomatic conditions (Ce/Ce*≪1.0) and desulfurization leading to a general loss of PGE and data patterns that are consistent overall with low abundances of relatively radiogenic Os in arc peridotites globally (Brandon et al., 1996; Widom et al., 2003; Saha et al., 2005). Kharchinsky pyroxenites, which are relatively undeformed and only weakly metasomatized, have similarly fractionated PGE patterns with enrichments in Pt (5.7–46 ppb) and Pd (0.5–2.7 ppb) but with Os, Ir, and Ru abundances < 0.23 ppb. The patterns resemble arc basalts and may reflect preferential mobility of Pt, Pd, and Re in slab fluids. High Pt and Pt/Pd up to 33 in the pyroxenites indicate that a Pt-rich alloy with high Pt/Pd is present in the early (olivine-rich) cumulate assemblage. Several Kharchinsky samples (peridotites and pyroxenites) are Pt-enriched with high Pt/Os and Pt/Re > 20 that will produce rapid ingrowth of 186Os relative to 187Os. Deep recycling and storage of such subduction-modified rocks might preclude the need for core-mantle exchange to explain 186Os-enrichments in komatiites and Hawaiian picrites originating in the deep mantle (Brandon et al., 1998; Brandon and Walker, 2005).

182W evidence for core-mantle interaction in the source of mantle plumes

Tungsten isotopes are the ideal tracers of core-mantle chemical interaction. Given that W is moderately siderophile, it preferentially partitioned into the Earth’s core during its segregation, leaving the mantle depleted in this element. In contrast, Hf is lithophile, and its short-lived radioactive isotope 182Hf decayed entirely to 182W in the mantle after metal-silicate segregation. Therefore, the 182W isotopic composition of the Earth’s mantle and its core are expected to differ by about 200 ppm. Here, we report new high precision W isotope data for mantle-derived rock samples from the Paleoarchean Pilbara Craton, and the Reunion Island and the Kerguelen Archipelago hotspots. Together with other available data, they reveal a temporal shift in the 182W isotopic composition of the mantle that is best explained by core-mantle chemical interaction. Core-mantle exchange might be facilitated by diffusive isotope exchange at the core-mantle boundary, or the exsolution of W-rich, Si-Mg-Fe oxides from the core into the mantle. Tungsten-182 isotope compositions of mantle-derived magmas are similar from 4.3 to 2.7 Ga and decrease afterwards. This change could be related to the onset of the crystallisation of the inner core or to the initiation of post-Archean deep slab subduction that more efficiently mixed the mantle.

Redox state of the Archean mantle: Evidence from V partitioning in 3.5–2.4 Ga komatiites

Oxygen fugacity of the mantle is a crucial thermodynamic parameter that controls such fundamental processes as planetary differentiation, mantle melting, and possible core-mantle exchange. Constraining the evolution of the redox state of the mantle is of paramount importance for understanding the chemical evolution of major terrestrial reservoirs, including the core, mantle, and atmosphere. In order to evaluate the secular evolution of the redox state of the mantle, oxygen fugacities of six komatiite systems, ranging in age from 3.48 to 2.41 Ga, were determined using high-precision partitioning data of the redox-sensitive element vanadium between liquidus olivine, chromite and komatiitic melt. The calculated oxygen fugacities range from −0.11 ± 0.30 ΔFMQ log units in the 3.48 Ga Komati system to +0.43 ± 0.26 ΔFMQ log units in the 2.41 Ga Vetreny system. Although there is a slight hint in the data for an increase in the oxygen fugacity of the mantle between 3.48 and 2.41 Ga, these values generally overlap within their respective uncertainties; they are also largely within the range of oxygen fugacity estimates for modern MORB lavas of +0.60 ± 0.30 ΔFMQ log units that we obtained using the same technique. Our results are consistent with the previous findings that argued for little change in the mantle oxygen fugacity since the early Archean and indicate that the mantle had reached its nearly-present day redox state by at least 3.48 Ga.

Partitioning experiments in the laser-heated diamond anvil cell: volatile content in the Earth's core

The present state of the Earth evolved from energetic events that were determined early in the history of the Solar System. A key process in reconciling this state and the observable mantle composition with models of the original formation relies on understanding the planetary processing that has taken place over the past 4.5Ga. Planetary size plays a key role and ultimately determines the pressure and temperature conditions at which the materials of the early solar nebular segregated. We summarize recent developments with the laser-heated diamond anvil cell that have made possible extension of the conventional pressure limit for partitioning experiments as well as the study of volatile trace elements. In particular, we discuss liquid-liquid, metal-silicate (M-Sil) partitioning results for several elements in a synthetic chondritic mixture, spanning a wide range of atomic number-helium to iodine. We examine the role of the core as a possible host of both siderophile and trace elements and the implications that early segregation processes at deep magma ocean conditions have for current mantle signatures, both compositional and isotopic. The results provide some of the first experimental evidence that the core is the obvious replacement for the long-sought, deep mantle reservoir. If so, they also indicate the need to understand the detailed nature and scale of core-mantle exchange processes, from atomic to macroscopic, throughout the age of the Earth to the present day.

Applications of PGE Radioisotope Systems in Geo- and Cosmochemistry

The platinum-group elements contain three radioisotope systems that have been used in many and varied ways in geo- and cosmochemistry. Unique chronological applications include dating the formation of such diverse materials as sulfides, gold, organic-rich sediments, iron meteorites, and sulfide inclusions in diamonds. These systems also serve as isotope tracers for processes such as continental erosion, the deposition of extraterrestrial materials on Earth's surface, crust-mantle differentiation, recycling of subducted crust into the mantle, core-mantle exchange, and volatile-element depletion of planets and planetesimals. Although these systems have been in use for only a short time, the discoveries they have provided bode well for their incorporation as staples in the geochemical toolbox.

Tungsten geochemistry and implications for understanding the Earth's interior

The concentration of tungsten (W) in basaltic melts provides a window into the behavior of this element during core–mantle separation, crust formation, silicate differentiation, and potentially core–mantle interaction. We have analyzed an extensive suite of modern basalts ( n = 86) for their trace element chemistry via laser ablation ICP-MS, with barium (Ba), thorium (Th), uranium (U), and W concentrations typically determined to ≤ 5% (2 σ) uncertainty. We find that the partitioning behavior of U mirrors that of W during basalt genesis, whereas Ba and Th both behave more incompatibly. The W/U ratio of our complete sample suite (0.65 ± 0.45, 2 σ) is representative of the mean modern mantle, and is indistinguishable from that of mid-ocean ridge basalts (W/U MORB = 0.65 ± 0.41, n = 52), ocean island basalts (W/U OIB = 0.63 ± 0.07, n = 10), and back-arc basin basalts (W/U BABB = 0.62 ± 0.09, n = 12). This ratio is also consistent with the W/U ratio of the continental crust, and thus represents the W/U ratio of the entire silicate portion of the Earth. Assuming a concentration of 20 ± 8 (2 σ) ng/g U in the bulk silicate Earth, the abundance of W in the silicate Earth is 13 ± 10 ng/g. Following mass balance, this implies a mean modern mantle and core composition of 8.3 ± 7.1 ng/g W and 500 ± 120 ng/g W, respectively. Additionally, the MORB source is modeled to contain approximately 3.0 ± 2.3 ng/g W, indicating a four-fold depletion of the highly incompatible elements in the MORB source relative to the silicate Earth. Although both the isotopic composition of W and the constancy of the silicate Earth W/U ratio allow for potential insight into core–mantle exchange, both of these proxies are extremely dependent on the chemical composition of the source. A case study of three Hawaiian picrites with enrichments in 186Os– 187Os but terrestrial ε 182W can be explained by: i) a lack of a core component in the Hawaiian “plume,” ii) crustal contamination, or iii) a source composition enriched in incompatible trace elements relative to the bulk silicate Earth.

Enriched Pt-Re-Os Isotope Systematics in Plume Lavas Explained by Metasomatic Sulfides

To explain the elevated osmium isotope (186Os-187Os) signatures in oceanic basalts, the possibility of material flux from the metallic core into the crust has been invoked. This hypothesis conflicts with theoretical constraints on Earth's thermal and dynamic history. To test the veracity and uniqueness of elevated 186Os-187Os in tracing core-mantle exchange, we present highly siderophile element analyses of pyroxenites, eclogites plus their sulfides, and new 186Os/188Os measurements on pyroxenites and platinum-rich alloys. Modeling shows that involvement in the mantle source of either bulk pyroxenite or, more likely, metasomatic sulfides derived from either pyroxenite or peridotite melts can explain the 186Os-187Os signatures of oceanic basalts. This removes the requirement for core-mantle exchange and provides an effective mechanism for generating Os isotope diversity in basalt source regions.

CORE MANTLE CHEMICAL ISSUES

Is the core’s chemical isolation from the mantle complete? This is an important question for the balancing of the persistently enigmatic budgets of the highly siderophile elements and PGE, to which Fleet’s work has contributed so much. On redox criteria alone, it is clear that the core and the mantle cannot be in bulk equilibrium, and therefore that there are chemical incentives for core–mantle exchange. Recent Os isotope evidence suggests that the core may indeed be leaking chemically (although some W isotope evidence does not encourage this view). Possible mechanisms of chemical transfer between core and mantle include the following. (1) Cooling of the Earth provides a temperature perturbation that affects solubility in the outer core of the oxides, sulfides, carbides, hydrides, and silicates with which inevitably the outer core becomes saturated. Cooling and buoyancy lead to a progressive transfer of liquid (or crystalline) precipitates to the mantle. Solubilities of nonmetals in the core need determination. (2) Electromagnetic effects arising from core dynamo action can also drive core–mantle transfers. Interesting phase-changes, electrowetting, and dramatic trace-element chemical fractionation effects are driven by ~1 V potentials. The electronic character of, and voltages across, the core–mantle boundary (CMB) need determination. (3) Chemical transfers can be forced by emplacement of chemically incompatible assemblages into the CMB region. The redox imbalance between the core and oxidized slab material subducted to the CMB will certainly require chemical re-adjustments, most plausibly leading to core-to-mantle transfer of mass. The current debate about whether the chemical signature seen in some plumes is recycled crust or core need not have an exclusive “either/or” resolution. Recycled material could be key in refluxing core material into the mantle.

The debate over core–mantle interaction

Some hotspot volcanism is likely the surface manifestation of hot upwelling mantle plumes that rise from the core–mantle boundary (CMB) or from compositional boundaries within the mantle. Plumes that rise from the CMB may generate magmas that carry compositional information about the diversity of materials present in the lowermost mantle. Some plume-derived materials have coupled enrichments in 186Os/188Os and 187Os/188Os relative to upper mantle materials. The coupled enrichments are consistent with Os isotopic compositions predicted for the liquid outer core that might result from elevated Pt/Os and Re/Os generated via progressive crystallization of a solid inner core over Earth history. If these enrichments are a reflection of core–mantle interaction, then the mechanisms of material transport between the core and the mantle, rates of whole mantle convection at present and over time, and cooling histories for the core and the mantle can potentially be constrained. However, fractionation may also occur between Pt/Re/Os via mantle and crustal processes and in particular, Pt and Pt/Re may be elevated in some portions of ancient recycled slabs thought to be present in certain plume sources. This possibility brings into question whether the coupled Os isotope enrichments are a product of core–mantle exchange. Hence, a lively debate over the causes of Os isotopic heterogeneities in plume-derived materials has ensued. Resolution of this debate has been hampered by limited constraints on the behavior of these elements in the Earth's core, mantle and crust. A variety of analytical and experimental tasks will have to be pursued to test different aspects of both the core–mantle interaction and crustal recycling models, and to further constrain the behaviors of these elements. These future endeavors may help to unravel the compositional complexities of plume-derived materials and their mantle sources, and ultimately provide important insights to processes operating at the CMB.

Laser-ablation microprobe (LAM)-ICPMS unravels the highly siderophile element geochemistry of the oceanic mantle

The highly siderophile element (HSE) contents of base-metal sulphides have been determined by laser-ablation microprobe (LAM)-ICPMS in abyssal peridotites from the Mid-Atlantic and South West Indian ridges. (Pd/Ir)N (0.007–505, N: CI-chondrite-normalised), (Pt/Ir)N (0.001–0.77) and (Rh/Ir)N (0.159–273) vary significantly between both grains and samples, irrespective of indicators of melt removal, but in line with bulk-rock platinum-group element (PGE) ratios and sulphide modal abundances. Positive deviations of PGE abundance ratios in whole-rock analyses are due to late-precipitated Cu–Ni-rich magmatic sulphides from incompletely extracted partial melts. These results contradict explanations of the HSE systematics of the oceanic mantle as reflecting global scale processes such as core–mantle exchange.

Rhenium abundances and systematics in oceanic basalts

We have examined the systematics of Re abundances in concert with new and published data on major and trace elements in oceanic basalts. Re abundances in MORB average 927 ppt (parts per trillion), a factor of 2.6 higher than the OIB average (350 ppt). The variability of Re abundances in both MORB and 01B is only 45%, similar to moderately incompatible major and trace elements such as Ca, Al, and the heavy rare earths. The inferred mantle/melt partitioning behavior of Re is similar to Yb, and average Re/Yb ratios in MORB, O1B, and the continental crust are within 50% of each other. The Re/Yb ratios of the depleted mantle and continental crust are both lower than the primitive mantle. The Re/Yb systematics place constraints on core-mantle exchange. The incorporation of material from the outer core into mantle plumes, and isolation of the upper mantle from the core, would result in Re/Yb ratios in OIB elevated by factors of 5–24 over MORB. Although the scatter of the OIB Re/Yb data may allow core addition up to the 1.2% level, the lack of a correlation between Re/Yb and Os isotopes in OIB, and the small difference in average Re/Yb ratios of MORB and OIB suggests an upper limit of 0.2 wt% core material in mantle plumes. A mass balance calculation between continental crust, depleted mantle, and primitive mantle indicates that the crustal abundance of Re is too low (by a factor of ten) to balance the Re depletion of the depleted MORB mantle (DMM). Anoxic sediments and black scales are an important crustal reservoir for Re, but the estimated abundance of these rocks in the continental crust cannot be increased substantially without upsetting the crustal U budget and ratios of Th/U, Nb/U, and U/Pb in the depleted mantle. The Re missing from DMM must therefore reside in recycled oceanic crust stored in the deep mantle. We propose that elevated Os isotope ratios in OIB are a direct measure of a ubiquitous component of recycled MORB crust in mantle plumes. The Re-Os mass balance allows a simple estimate of the total size of the OIB reservoirs to be made. A mass of recycled MORB equivalent to 4.4% of the mantle is required to balance the Re depletion of DMM. The total mass of the OIB reservoirs is estimated at 4.4% to 22% of the mantle mass, depending on whether the low- 187Os/ 188Os component(s) in OIB sources are external or intrinsic, respectively. Using an estimate of the present-day plume flux, the average residence time of subducted slabs in the convecting mantle is estimated at 700–3500 Ma; however, the slab residence time scales linearly with the size of the OIB reservoir (and thus the proportion of recycled crust in mantle plumes). We prefer a model in which individual mantle plumes incorporate slabs with distinct ages. Consideration of Os and He isotope covariations in OIB suggest a model in which subducted slabs are recycled through the whole mantle on timescales of < 1.8 Ga.