1. STUDY OF THE HIGHLY SIDEROPHILE ELEMENTS IN THE TERRESTRIAL MANTLE

The elements Re, Os, Ir, Ru, Pt and Pd are termed “highly-siderophile” (siderophile = iron loving) because in metal-silicate systems, these elements are strongly partitioned into the metal phase. Consequently, planetary core formation likely stripped >99.99% of these elements (and other highly siderophile elements) from the silicate portions of the Earth, Mars and other differentiated bodies. Despite the removal of highly siderophile elements from the silicate portion of the Earth, the estimated abundances of these elements in Earth’s upper mantle are higher than expected from metal-silicate equilibria.

Orange symbols are the highly siderophile elements. Their abundances in the Earth’s mantle are about 200 times lower than primitive chondritic meteorites, and occur in approximately chondritic relative abundances (Figure from Cottrell, 2004).

Of greater consequence than the absolute abundances of highly siderophile elements in the mantle, ¹⁸⁷Os/¹⁸⁸Os and ¹⁸⁶Os/¹⁸⁸Os ratios are broadly chondritic, requiring long-term chondritic Re/Os and Pt/Os ratios (¹⁸⁷Re decays to ¹⁸⁷Os by ^-β decay with a 42 billion year half-life; ¹⁹⁰Pt decays to ¹⁸⁶Os by α emission with a half life of approximately 430 billion years). Chondritic ratios of these elements in the mantle are a very unlikely result of metal-silicate equilibration, so this observation provides strong support to the hypothesis that the highly siderophile elements residing in the upper mantle today were mostly added to the Earth via a process that has been termed “late accretion”, e.g. an additional 0.3 to 1% planetary accretion following the final stages of core segregation. Late accretion would result in chondritic ratios of the highly siderophile elements because the bulk planetessimals added by this process would likely have had chondritic absolute and relative abundances. Highly siderophile elements added in this manner would not be subsequently extracted into the core because segregation had ceased. Late accretion is recorded on the face of the Moon by the large impact basins that were created between about 4.5 and 3.9 billion years ago (see below). Late accretion, therefore, likely set the stage for the generally chondritic evolution of Os isotopes in the terrestrial mantle.

Despite the chondritic starting point for highly siderophile elements in the terrestrial mantle, processes such as melt extraction, melt percolation, crustal recycling, and possibly core-mantle interaction can alter the chondritic relative abundances of these elements. Rhenium and Os are especially sensitive to these processes because they behave very differently during mantle melting (Re is moderately compatible and Os is highly compatible), and even during solid metal-liquid metal fractionation. Thus, the ¹⁸⁷Re-¹⁸⁷Os isotope system provides a means of examining the chemical evolution of highly siderophile elements in mantle domains (and in other planetary mantles) and, perhaps, including the outer core (see below).

Although the ¹⁸⁷Os/¹⁸⁸Os ratio of the bulk upper mantle is estimated to be within the range of chondritic meteorites, the isotopic composition of Os in the modern mantle is actually quite variable, ranging from ¹⁸⁷Os-depleted in ancient subcontinental lithospheric mantle, relative to chondrites, to enriched in some plume-derived rocks, including certain ocean island basalts. Determining how and when these heterogeneities were created will help us to more fully understand the chemical evolution of different portions of the mantle. Specific issues that Re-Os isotopes may be particularly useful in elucidating are: 1) the nature of material added to the Earth via late accretion, 2) the rate at which late accreted materials were homogenized throughout the mantle, 3) the nature and rate at which oceanic crust has been recycled back into the mantle, 4) the existence of, and level of chemical exchange between the mantle and outer core, and 5) the chronology of the development of subcontinental lithospheric keels.

One major research target of our lab is to characterize the Os isotopic composition of various mantle domains through time, focusing on the analysis of ultramafic rocks that are otherwise geochemically and isotopically well-characterized, and that have relatively high Os concentrations and low Re/Os ratios. Specific current targets of research are from a variety of locations worldwide including the Taitao ophiolite, Chile, peridotites from the Appalachians of North America, and komatiites from the Abitibi Belt of Canada, and Lapland, Finland. Both whole rock samples and mineral separates are analyzed from the studied areas. The resulting information is leading to a better understanding of how Os isotopic heterogeneities have developed in the mantle. These results may ultimately provide additional constraints on the nature of highly siderophile elements in the mantle and early mantle mixing rates.

Photo shows M.S. student Tracey Centorbi pointing at a chromite layer within a massive dunite from the Blue Ridge Province of North Carolina. These unusual dunitic bodies are often characterized by very fresh olivine. They may represent portions of dismembered ophiolite, thus recording isotopic characteristics of the mantle during the Paleozoic. Our Os isotopic work suggests the rocks were generated and emplaced as part of the Taconic orogeny, and may sample the mantle underlying the Iapetus Ocean.

a. Re-Os Isotope and Highly Siderophile Element Study of the Taitao Ophiolite, Chile.

Despite a very large amount of accumulated geochemical data, there is still considerable debate regarding the age, magnitude, causes and extent of chemical heterogeneity in the convecting upper mantle. Intensive trace element and isotopic studies of MORB indicate local through worldwide variations, but provide limited information regarding the distribution of components within the mantle. Studies of ophiolites, presumed to be tectonically emplaced slivers of oceanic lithosphere, have provided useful information regarding the distribution of chemical domains in the upper mantle and lower oceanic crust, but ophiolites by their nature are typically highly tectonized and metamorphosed.

In some circumstances, the Re-Os isotope system can be used to constrain the ages of melt depletion in mantle peridotites, and thus, can potentially provide unique insights to the ages of chemical heterogeneity (as caused by melting processes) in the convecting mantle. Although much Os work has been conducted on ophiolites, most work has focused on chromitites. Chromitites are advantageous in that they typically are well preserved and have very low Re/Os ratios, and consequently normally require only modest age corrections to obtain initial ¹⁸⁷Os/¹⁸⁸Os. However, the origin of chromitites in ophiolites is debated and a key interest of ours regarding the convecting mantle is the chemical/isotopic variability present within a relatively small mantle domain, and how the heterogeneities are manifested among the different mafic and ultramafic rocks present. For example, do harzburgites and cumulate dunites of the convecting (oceanic) mantle have similar ¹⁸⁷Os/¹⁸⁸Os, and if so, are these compositions similar to those of spatially associated chromites? How do these compositions vary relative to associated gabbros and basalts? Can any detected variations in the ¹⁸⁷Os/¹⁸⁸Os ratios of the different rocks be attributed to specific processes, such as melt percolation or interactions with highly radiogenic seawater? Are “old” lithologies (at least from an Os isotopic standpoint) common in ophiolites? Some previous studies of Phanerozoic and Precambrian ophiolites have attempted to address aspects of these questions, however, even modest open system behavior of Re-Os can lead to erroneous interpretations regarding the magnitude of heterogeneity present in the initial ¹⁸⁷Os/¹⁸⁸Os ratios of different materials because of age corrections.

One way to circumvent the effects of modest post-crystallization open system behavior on calculated initial ratios (excluding the addition of isotopically distinct Os) is to examine a very young ophiolite. Consequently, with collaborators at the University of Tsukuba and the University of Chile, we have begun an intensive study of the Taitao ophiolite (also known as the Bahia Barrientos ophiolite) of southern Chile. The Taitao ophiolite is located on the Taitao peninsula, 50 km southeast of the Chile triple junction (CTJ) and only 17 km from the Chile trench.

This ophiolite is likely related to ridge subduction and collision (e.g. Bourgois et al., 1993), and lies close to the CTJ where the Nazca, Antarctic and South American plates are juxtaposed. The chemical affinities of at least some of the ophiolite-related mafic rocks are of N-type and E-type MORB. Recent dating of minerals contained within gabbros reveal that the ophiolite is only approximately 6 million years old, resulting in very minor age corrections for Os in most rocks present.

Photo of the Taitao ophiolite provided by collaborator Prof. Ryo Anma (University of Tsukuba).

Geological map of the transition between peridotite and gabbroic units, showing the Os isotope ratios of the studied rocks (peridotites and gabbros in white and grey boxes respectively). Also shown are the different units recognized in Taitao peninsula (after Lagabrielle et al., 2000) and the geotectonic setting of the area.

Microphotograph of a Taitao peridotite with some well preserved olivine. The olivine is highly fractured and individual grains range in size from about 1 to 4 mm across. Roughly 10% of the rock consists of olivine crystals. Serpentine, chromite, and pyroxene comprise the bulk of this rock. Photo was taken with polarizers crossed, and the scale is approximately 3 mm across.

Plot of ¹⁸⁷Re/¹⁸⁸Os versus ¹⁸⁷Os/¹⁸⁸Os for Taitao peridotites. The ¹⁸⁷Re/¹⁸⁸Os ratios of peridotites normally decrease as a function of melt depletion, so the generally positive trend for Taitao peridotites indicates an ancient melting event for at least some of the peridotites. Ancient melt depletion is common in the subcontinental lithospheric mantle, and has been documented for the oceanic mantle via studies of abyssal peridotites. The relation between Os isotopes and major and trace elements in abyssal peridotites, however, is normally obscured by extensive serpentinization. The excellent preservation of some Taitao peridotites (see photo above) allows a consideration of the petrologic history of the rocks studied.

Plot of Mg# of whole rocks (WR) and olivines versus initial ¹⁸⁷Os/¹⁸⁸Os for Taitao peridotites. The generally negative trends are consistent with variable ancient melt depletion and indicate little change in bulk composition of the rocks since the melt depletion event. The depleted ¹⁸⁷Os/¹⁸⁸Os ratios indicate melt depletion occurred a minimum of about 1.6 billion years ago and that, although part of the convecting upper mantle, these rocks were not subsequently processed via further melting or melt percolation.

Highly siderophile element plots for Taitao peridotites normalized to chondrites (times 1000). Of note, the harzburgites have the lowest Re/Os ratios in this suite (and are accompanied by the lowest ¹⁸⁷Os/¹⁸⁸Os). These patterns show no evidence for melt percolation, with the possible exception of the tectonite.

b. Komatiites as Probes of Os Isotope Evolution in the Mantle.

Mantle sources of modern ocean island basalts (OIB), associated with hot spots or plumes, display large variations in ¹⁸⁷Os/¹⁸⁸Os ratios, ranging from chondritic ratios of about 0.13 to as high as 0.16, or _Os from 0 to about +25. The suprachondritic ¹⁸⁷Os/¹⁸⁸Os isotopic compositions have most often been ascribed to mantle sources that have incorporated variable proportions of mafic recycled oceanic crust and associated sediments. This interpretation is consistent with the lithophile isotope and trace element systematics in some OIB.

The Re-Os isotope system is an especially valuable tracer of ancient recycled mafic crust in mantle domains because MORB and oceanic gabbros, as noted above, are normally enriched by factors of 3 to 4 in Re relative to mantle peridotite, yet are highly depleted in Os. Over time, hybrid crust-mantle mixtures will generate suprachondritic ¹⁸⁷Os/¹⁸⁸Os. Because of the high concentration of Os in the mantle relative to the crust, however, the Re-Os system is rather insensitive to the addition of recycled crust, so high ¹⁸⁷Os/¹⁸⁸Os (_Os > +5) in a modern mantle source normally requires a relatively large proportion of recycled mafic crust that has aged either separately, or as part of a hybrid crust-mantle mixture, for at least 1 to 3 billion years.

Crustal recycling is probably not responsible for all ¹⁸⁷Os-enriched mantle sources, and alternate mechanisms for generating such enrichments have been proposed. One possibility raised by the presence of coupled ¹⁸⁶Os-¹⁸⁷Os enrichments in two large mantle volcanic systems, the Hawaiian and Siberian plumes, is that some mantle sources may derive a portion of their Os from the putative ¹⁸⁷Os- and ¹⁸⁶Os-enriched outer core (see below).

Given the geochemical requirements for generating suprachondritic ¹⁸⁷Os/¹⁸⁸Os in the mantle domains noted above, examination of Proterozoic and Archean mantle domains may provide important constraints on the rate and timing of early crustal recycling, or alternately, inner core formation and subsequent core-mantle interaction. A lack of, or limited occurrence of suprachondritic ¹⁸⁷Os/¹⁸⁸Os in Archean mantle domains might be suggestive of either the minimal presence of significantly older crust in the sources of the domains sampled, or a lack of substantial inner core formation (or lack of core-mantle interaction). Conversely, the presence of mantle domains with resolvable suprachondritic ¹⁸⁷Os/¹⁸⁸Os might be attributed to the recycling of crust generated within the first 500 Ma of Earth’s history, or core-mantle interaction.

As a means of characterizing the Os isotopic composition of ancient mantle reservoirs, studies of komatiites and picrites are ideal. It is generally agreed that at least some komatiites were generated as a consequence of hot spot or plume volcanism. In addition, because these lavas normally are generated from magmas with high Os concentrations, the effects of crustal contamination during transport to the surface are minimized. Nonetheless, metamorphic effects can seriously affect the Re-Os systematics of some whole rock samples. Consequently, it is critical in studies of all but the youngest rocks that the closed-system behavior of Re-Os be demonstrated. This is usually accomplished via the generation of an isochron of the correct age, as determined from other geochronometers.

Approximately 2.7 Ga komatiite flow, Pyke Hill, Ontario, Canada. Note the impressive spinifex texture.This komatiite was derived from a mantle source with a chondritic ¹⁸⁷Os/¹⁸⁸Os ratio.

As shown below, the sources of most komatiites and picrites evolved along an average chondritic trajectory. This may be representative of the evolution of the upper mantle, the lower mantle or possibly the bulk mantle. A subset of komatiites, however, show evidence for derivation from sources with long-term enrichment in Re/Os. This may indicate that the sources of some komatiites contain substantial, ancient recycled mafic crust, or include chemical or isotopic contributions from the outer core.

Compilation plot of Os isotopic composition (in _Os units: the percentage deviation of an Os isotopic composition at any time from that of the chondritic average) through time for ultramafic systems. Each datum or box represents a lot of data. The horizontal black line represents the evolution of the chondritic average. The boxes show the range of compositions for isotopically heterogenous systems. Notice that there are significant enrichments in ¹⁸⁷Os relative to chondrites in several Archean systems. The evolution of some mantle domains away from chondritic is evident in the data. Though the reasons for this diversification of isotopic composition are highly debated (e.g. crustal recycling versus lower mantle derivation versus core-mantle exchange), there is no doubt the trends reflect early fractionation of Re/Os in some substantial domains within the mantle. The letters are abbreviations of the locale name. B-Belingwe, Zimbabwe (Walker and Nisbet, 2002), N- Nori’lsk, Siberia flood basalt province (Horan et al., 1995); Ke-Keweenowan, Canada (Shirey, 1997); D-Deccan, India (Allegre et al., 1999); TNB-Thompson Nickel Belt, Canada (Hulbert et al., 2005); P-Pechenga, Russia (Walker et al., 1997); O-Onega, Russia (Puchtel et al., 1999); V-Vetreny, Russia (Puchtel, 2000.); RW-Ruth Well, Australia (Meisel et al., 2001); Kam.-Kambalda, Australia (Foster et al., 1996); Kidd-Munro, Canada (Gangopadhyay and Walker, 2003; Puchtel et al., 2005; Gangopadhyay et al., 2005); BC-Boston Creek, Canada (Walker and Stone, 2001); Kostomuksha, Russia (Puchtel et al., 2001); G-Gorgona Island, Colombia (Walker et al., 1999); Viet- Song La, Vietnam (Hanski et al., 2004); Isua – Greenland and Pil – Pilbara, Australia (Bennett et al., 2003), Com- Commondale, South Africa (Wilson et al., 2003); Jees – Jeesiörova, Finland (Gangopadhyay et al., 2006) . Shown for comparison is the range of data defined by modern ocean island basalts (OIB).

To learn more about this research, please refer to the following publications:

Walker R.J., Hanski E.J., Vuollo J. and Liipo J. (1996) The Os isotopic composition of Proterozoic upper mantle:evidence for chondritic upper mantle from the Outokumpu ophiolite, Finland. Earth Planet. Sci. Lett. 141, 161-173.

Meisel T., Walker R.J. and Morgan J.W. (1996) The osmium isotopic composition of the Earth’s primitive upper mantle. Nature 383, 517-520.

Walker R. J., Morgan J.W., Hanski E.J. and Smolkin V. (1997) Re-Os systematics of Early Proterozoic ferropicrites, Pechenga Complex, NW Russia:evidence for ancient ¹⁸⁷Os-enriched plumes. Geochim. Cosmochim. Acta 61, 3145-3160.

Walker R.J., Storey M., Kerr A., Tarney J. and Arndt N.T. (1999) Implications of ¹⁸⁷Os heterogeneities in mantle plumes: evidence from Gorgona Island and Curaçao. Geochim. Cosmochim. Acta 63, 713-728.

Tsuru A., Walker R.J., Kontinen A., Peltonen P. and Hanski E. (2000) Re-Os isotopic systematics of the Jormua Ophiolite Complex, NW Finland. Chemical Geology 164, 123-141.

Meisel T., Walker R.J., Irving A.J., and Lorand J.-P. (2001) Osmium isotopic compositions of mantle xenoliths: a global perspective. Geochim. Cosmochim Acta. 65, 1311-1323.

Walker R.J. and Stone W.R. (2001) Os isotope constraints on the origin of the 2.7 Ga Boston Creek flow, Ontario, Canada. Chem. Geol. 175, 567-579.

Walker R.J., Prichard H.M., Ishiwatari A. and Pimentel M. (2002) The osmium isotopic composition of convecting upper mantle deduced from ophiolite chromitites. Geochim. Cosmochim. Acta 66, 329-345.

Walker R.J. and Nisbet E. (2002) ¹⁸⁷Os isotopic constraints on Archean mantle dynamics. Geochim. Cosmochim. Acta 66, 3317-3325.

Gangopadhyay A. and Walker R.J. (2003) Re-Os systematics of the ca. 2.7 Ga Alexo komatiites, Ontario, Canada. Chem. Geol. 196, 147-162.

Hanski E.J., Walker R.J. Huhma H., Polyakov, G.V., Glotov, A.I., Tran Trong Hoa, Ngo Thi Phuong (2004) Origin of Permo-Triassic komatiites, northwestern Vietnam. Contrib. Mineral. Petrol. 147, 453-469.

Gornostayev S.S., Walker R.J., Hanski E.J. and Popovchenko S.E. (2004) Evidence for the emplacement of ca. 3.0 Ga lithospheric mantle in the Ukrainian Shield. Precamb. Res. 132, 349-362.

Walker R. J., Brandon A.D., Bird J.M., Piccoli P.M., McDonough W.F. and Ash R.D. (2005) ¹⁸⁶Os-¹⁸⁷Os systematics of Os-Ir-Ru alloy grains, southwestern Oregon. Earth Planet. Sci. Lett 230, 211-226.

Gangopadhyay A., Sproule R.A., Walker R.J. and Lesher M. (2005) 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. Geochim. Cosmochim. Acta 69, 5087-5098.

Gangopadhyay A., Walker R.J., E. Hanski, and P. Solheid (2006) Origin of Paleoproterozoic Ti-enriched komatiitic rocks from Jeesiörova, Kittilä Greenstone Complex, Finnish Lapland. Journal of Petrology 47, 773-789.

Puchtel I., Humayun M. and Walker R.J. Os-Nd-Pb isotope and lithophile and siderophile trace element systematics of komatiitic rocks from the Volotsk suite, SE Baltic Shield. Precambrian Res., in press.

Van Acken D., Becker H. and Walker R.J. Mantle refertilization processes and their impact on the Re-Os systematics of peridotites: a case study from the Totalp ultramafic massif, eastern Switzerland. Earth Planetary Sci. Lett., in review.

Last Revised July 2007