Mark Rehkämper

Ir, Ru, Pt, and Pd in basalts and komatiites: New constraints for the geochemical behavior of the platinum-group elements in the mantle

Abstract The concentrations of the platinum-group elements (PGE) Ir, Ru, Pt, and Pd were determined in 18 mantle-derived basalts from a variety of tectonic settings and six komatiites from three locations. All analyses were performed using isotope dilution, Carius tube digestion, and the precise technique of multiple collector inductively coupled plasma mass spectrometry. Multiple analyses of two samples indicate external reproducibilities, based upon separate dissolutions, of approximately 2–9% in the ppt to ppb concentration range. Mid-ocean ridge basalts from the Kolbeinsey Ridge, tholeiites from Iceland and alkali basalts from the Cameroon Line define three individual sample suites that are characterized by distinct major, trace, and platinum-group element systematics. All three-sample suites display correlations of the PGE with MgO, Ni, and Cr. The new analytical results are employed to constrain the geochemical behavior of the PGE during the formation and differentiation of mantle–derived melts. The PGE are inferred to be compatible in sulfides during partial melting with sulfide-silicate melt partition coefficients of ∼1 × 104. The fractionated PGE patterns of mantle melts are a consequence of the incompatibility of Pd in nonsulfide phases, whereas Ir and Ru must be compatible in at least one other mantle phase. Model calculations indicate that PGE alloys or spinel may be responsible for the higher compatibility of the latter elements during partial melting. It is further demonstrated that the shape of the melting regime has a profound effect on the PGE systematics of mantle magmas. The systematic trends of the three sample suites in plots of PGE against Ni and Cr are the result of magma differentiation processes that involve fractional crystallization of silicate minerals and the concurrent segregation of an immiscible sulfide liquid. The behavior of the PGE during magma fractionation indicates that the segregated sulfides probably equilibrate with >90% of the silicate magma and that PGE scavenging by sulfides is best described by a combination of batch and fractional equilibrium partitioning.

APPLICATIONS OF MULTIPLE COLLECTOR-ICPMS TO COSMOCHEMISTRY, GEOCHEMISTRY, AND PALEOCEANOGRAPHY

Multiple collector-inductively coupled plasma mass spectrometry (MC-ICPMS) is a new technique for the measurement of isotopic compositions at high precision, and is of great relevance to planetary, earth, ocean, and environmental sciences. The method combines the outstanding ionization efficiency of the ICP source with the superior peak shapes achievable from the ion optical focal plane of a large dispersion magnetic sector mass spectrometer, utilizing simultaneous multiple collection to achieve the most precise isotopic measurements yet made for many elements—particularly those with high first ionization potential. The addition of a laser facilitates studies for which spatial resolution is required. This method is still in its infancy, yet diverse applications have already led to a number of important scientific developments. Here we review some of these accomplishments and the potential for further work. The Lu-Hf isotopic system, for many years considered analytically challenging, is now relatively straightforward and offers great promise in fields as diverse as garnet geochronology, hydrothermal fluxes to the oceans, and crustal evolution. The age of the Earth’s core, the Moon, and Mars have been measured using a new short-lived chronometer 182Hf-182W. Other such new chronometers will follow. High precision isotope dilution measurements of the Earth’s inventory of many poorly understood elements such as In, Cd, Te, and the platinum group elements are providing tests for models for the accretion of the inner solar system. The small natural isotopic variations in elements such as Cu and Zn, produced by mass dependent fractionation, are now measurable at high precision with this method, and entirely new fields of stable isotope geochemistry can be developed. Similarly, measurements of small nucleosynthetic isotopic anomalies should be made easier for some elements. Measurements of U and Th isotopic compositions at very high sensitivity and reproducibility are now possible, allowing the development of higher resolution Quaternary geochronology. Finally, using laser ablation, the first precise in situ Sr, Hf, W, and Pb isotopic measurements have been made in natural materials, opening up a range of microanalytical isotopic studies in petrology and marine geochemistry. MC-ICPMS offers exciting times ahead in areas well beyond the bounds of geochemistry. Indeed, MC-ICPMS is likely to become the method of choice for many isotopic measurements because it is a more user friendly and efficient method for the acquisition of high precision data. It is also much more versatile, permitting elements to be measured that were previously considered intractable, and allowing the acquisition of data in situ, all with a single mass spectrometer. The limiting factor on the sensitivity is the transmission which is ≤2% for all instruments thus far designed. If it is found possible to improve the transmission still further, thermal ionization mass spectrometry, the technique that has, thus far, provided the high precision measurements necessary for most of the vast field of radiogenic isotope geochemistry, may be relegated to specialized applications.

High precision 230Th/232Th and 234U/238U measurements using energyfiltered ICP magnetic sector multiple collector mass spectrometry

We have developed new chemical and mass spectrometric techniques for the precise determination of 234U/238U and 230Th/232Th ratios in geological materials. The isotope ratio measurements were performed using an inductively coupled plasma double focusing magnetic sector multiple-collector mass spectrometer (MC-ICPMS) equipped with an additional energy filter. Following sample dissolution, U and Th are first separated from the rock matrix by a new and highly efficient column chromatographic procedure utilizing TRU-Spec resin. The strong affinity of U and Th for this material allows the use of extremely small ( < 0.5 ml resin) columns, even for the processing of silicate samples as large as ∼5 g. Our new mass-spectrometric techniques permit precise corrections for mass discrimination and gain variation during analysis. As a consequence, the precision and external reproducibility of uranium and thorium isotopic analysis is improved by a factor of ∼3–5 compared with previous results by conventional thermal ionization mass spectrometry (TIMS). A sensitivity of ∼0.04% is routinely achieved for Th and this is comparable to the best values achieved by TIMS for large sample sizes. Recent instrumental improvements, however, have further increased our sensitivity by about a factor of five. Our Th and U isotope data for standard reference materials and other geological samples are in excellent agreement with previously published values from other laboratories, further highlighting the reliability and analytical capabilities of our new techniques.

Stable isotope compositions of cadmium in geological materials and meteorites determined by multiple-collector ICPMS

A new technique for the precise and accurate determination of Cd stable isotope compositions has been developed and applied to geological materials and meteorites. The Cd isotope analyses are performed by multiple collector inductively coupled plasma mass spectrometry (MC-ICPMS) using external normalization to Ag for mass bias correction. The accuracy of the new procedure was ascertained by the comparison of data for meteorites with published results acquired by thermal ionization mass spectrometry and double spiking. Some results were also confirmed by measurements using external normalization to Sb on a different MC-ICPMS instrument. A long-term reproducibility of ± 1.1 eCd/amu (2 sd) was obtained for separate dissolutions and multiple analyses of several rock and meteorite samples (eCd/amu represents the deviation of a Cd isotope ratio of a sample relative to the JMC Cd standard in parts per 104, normalized to a mass difference of 1 amu). As little as 5–20 ng of Cd are sufficient for the acquisition of precise and accurate data. Terrestrial rock and mineral samples display little variations in Cd isotope compositions (eCd/amu between −1 and +1.2), except for a tektite sample that was found to be enriched in the heavy Cd isotopes by +7.6 eCd/amu. The carbonaceous chondrites Orgueil, Murchison and Allende have Cd isotope ratios that are unfractionated relative to the JMC Cd standard and terrestrial rocks. The ordinary chondrites analyzed in this study and a Rumuruti chondrite display Cd isotope fractionations, ranging from −19 to +36 eCd/amu. These results suggest that substantial (inorganic) natural Cd isotope fractionations are generated only by evaporation and/or condensation processes. The lack of resolvable Cd isotope variations between the different carbonaceous chondrites, despite large differences in Cd concentrations, implies that the primary depletion of Cd in the early solar system did not involve Rayleigh evaporation. The Cd isotope fractionation in ordinary and Rumuruti chondrites is probably due to the redistribution of Cd by evaporation and condensation processes during thermal metamorphism on the parent bodies. Models that explain the enrichments of highly volatile elements in unequilibrated ordinary chondrites by primary equilibrium condensation appear to be inconsistent with the Cd isotope data.

Non-chondritic platinum-group element ratios in oceanic mantle lithosphere: petrogenetic signature of melt percolation?

Abstract The concentrations of the platinum-group elements (PGE) Ir, Ru, Pt and Pd were determined in 11 abyssal peridotites from ODP Sites 895 and 920, as well in six ultramafic rocks from the Horoman peridotite body, Japan, which is generally thought to represent former asthenospheric mantle. Individual oceanic peridotites from ODP drill cores are characterized by variable absolute and relative PGE abundances, but the average PGE concentrations of both ODP suites are very similar. This indicates that the distribution of the noble metals in the mantle is characterized by small-scale heterogeneity and large-scale homogeneity. The mean Ru/Ir and Pt/Ir ratios of all ODP peridotites are within 15% and 3%, respectively, of CI-chondritic values. These results are consistent with models that advocate that a late veneer of chondritic material provided the present PGE budget of the silicate Earth. The data are not reconcilable with the addition of a significant amount of differentiated outer core material to the upper mantle. Furthermore, the results of petrogenetic model calculations indicate that the addition of sulfides derived from percolating magmas may be responsible for the variable and generally suprachondritic Pd/Ir ratios observed in abyssal peridotites. Ultramafic rocks from the Horoman peridotite have PGE signatures distinct from abyssal peridotites: Pt/Ir and Pd/Ir are correlated with lithophile element concentrations such that the most fertile lherzolites are characterized by non-primitive PGE ratios. This indicates that processes more complex than simple in-situ melt extraction are required to produce the geochemical systematics, if the Horoman peridotite formed from asthenospheric mantle with chondritic relative PGE abundances. In this case, the PGE results can be explained by melt depletion accompanied or followed by mixing of depleted residues with sulfides, with or without the addition of basaltic melt.

Determination of ultra-low Nb, Ta, Zr and Hf concentrations and the chondritic Zr/Hf and Nb/Ta ratios by isotope dilution analyses with multiple collector ICP-MS

This study presents a new technique for the determination of precise and accurate concentrations of the high field strength elements (HFSE) Zr, Hf, Nb and Ta. The Ta concentration was determined for the first time by the isotope dilution (ID) technique using an isotopic tracer enriched in 180Ta. Zirconium and hafnium concentrations were also determined by ID, whereas the concentration of the mono-isotopic Nb was measured relative to Zr, after quantitative separation of the HFSE from the matrix. The analyses were performed on a Micromass Isoprobe multiple collector (MC) inductively coupled plasma source mass spectrometer (ICP-MS). Only about 0.5 ng of Zr, Hf and Ta are necessary to perform an ID analysis with an external reproducibility of better than 1% on the MC-ICP-MS using Faraday collectors. This new technique enables the precise and accurate determination of the HFSE concentrations even in ultra-depleted rocks like peridotites. The absolute uncertainties for ultra-depleted rocks, particular for Ta concentrations at the sub-ng level are limited by blanks and sample heterogeneities and not by the precision of the measurement. New and more precise Zr, Hf, Nb and Ta concentration data for the geological standard reference materials BHVO-2, BCR-2, BE-N, BIR-1 and the ultra-depleted standards PCC-1 and DTS-1 are presented. External reproducibilities of the concentration measurements are 0.4–5% for basalts and 2–10% for depleted peridotite samples (2 RSD), depending on element and concentration. The Zr/Hf and Nb/Ta ratio of the solar system was determined based on new data for two chondrites and six achondrites. The chondritic Nb/Ta of 17.6±1.0 determined in this study agrees with previous predicted values from the literature. However, the chondritic Zr/Hf of 34.2±0.3 determined in this study differs from previous literature values.

Thallium isotope variations in seawater and hydrogenetic, diagenetic, and hydrothermal ferromanganese deposits

Results are presented for the first in-depth investigation of Tl isotope variations in marine materials. The Tl isotopic measurements were conducted by multiple collector-inductively coupled plasma mass spectrometry for a comprehensive suite of hydrogenetic ferromanganese crusts, diagenetic Fe–Mn nodules, hydrothermal manganese deposits and seawater samples. The natural variability of Tl isotope compositions in these samples exceeds the analytical reproducibility (±0.05‰) by more than a factor of 40. Hydrogenetic Fe–Mn crusts have ϵ205Tl of +10 to +14, whereas seawater is characterized by values as low as −8 (ϵ205Tl represents the deviation of the 205Tl/203Tl ratio of a sample from the NIST SRM 997 Tl isotope standard in parts per 104). This ∼2‰ difference in isotope composition is thought to result from the isotope fractionation that accompanies the adsorption of Tl onto ferromanganese particles. An equilibrium fractionation factor of α∼1.0021 is calculated for this process. Ferromanganese nodules and hydrothermal manganese deposits have variable Tl isotope compositions that range between the values obtained for seawater and hydrogenetic Fe–Mn crusts. The variability in ϵ205Tl in diagenetic nodules appears to be caused by the adsorption of Tl from pore fluids, which act as a closed-system reservoir with a Tl isotope composition that is inferred to be similar to seawater. Nodules with ϵ205Tl values similar to seawater are found if the scavenging of Tl is nearly quantitative. Hydrothermal manganese deposits display a positive correlation between ϵ205Tl and Mn/Fe. This trend is thought to be due to the derivation of Tl from distinct hydrothermal sources. Deposits with low Mn/Fe ratios and low ϵ205Tl are produced by the adsorption of Tl from fluids that are sampled close to hydrothermal sources. Such fluids have low Mn/Fe ratios and relatively high temperatures, such that only minor isotope fractionation occurs during adsorption. Hydrothermal manganese deposits with high Mn/Fe and high ϵ205Tl are generated by scavenging of Tl from colder, more distal hydrothermal fluids. Under such conditions, adsorption is associated with significant isotope fractionation, and this produces deposits with higher ϵ205Tl values coupled with high Mn/Fe.

Cadmium, indium, tin, tellurium, and sulfur in oceanic basalts: Implications for chalcophile element fractionation in the Earth

Concentrations of S, Cd, In, Sn, and Te are reported for 80 samples of mid-ocean ridge basalt (MORB), submarine and subaerial ocean island basalt (OIB) and submarine arc lavas. Cadmium, In, and Sn are moderately incompatible, and Te is compatible during partial melting. Cadmium is particularly uniform, consistent with a homogeneous distribution in the mantle. Tellurium is more variable (1–6 ppb) and is notably higher in Loihi, ranging up to 29 ppb, the most likely explanation for which is accumulation of Cu-bearing sulfide. The average Cd/Dy ratio is the same (0.027) for OIB glasses, MORB glasses and the continental crust, yielding a primitive mantle Cd concentration of ∼18 ppb. Indium, despite being more volatile, is less depleted than Cd and the other very volatile chalcophile elements Pb, Bi, Tl, and Hg. From the depletion of In we deduce that core formation depleted the silicate Earth in Cd, Pb, Bi, Tl, and Hg by between factors of 5 and 10. The In depletion yields concentrations of C, S, Se, and Te in the core of C ∼1.2%, S > 2.4%, Se > 7.1 ppm, and Te > 0.89 ppm. The Moon appears to be enriched in Te relative to the silicate Earth. Either a significant fraction of the Moon was derived from a more Te-rich body or the silicate Earths inventory of chalcophile and siderophile elements was depleted by further terrestrial core growth after formation of the Moon.

Early evolution of the Earth and Moon: new constraints from Hf-W isotope geochemistry

The W isotopic composition of the bulk silicate Earth (BSE) is chondritic within current analytical uncertainties, indicating that terrestrial core formation commenced more than 50 Myr after the differentiation of the earliest planetesimals in the solar system. This is consistent with UPb data and holds true whether core formation is modelled as a single catastrophic event or as a continuous process that started late, unless accretion was more than about twice as slow as recently estimated. The chondritic W isotopic composition of the BSE provides support for the assumption that the overall lithophile/siderophile refractory element ratio of the Earth is close to chondritic, but requires mixing to remove early heterogeneities introduced by the accretion of any planetesimals already segregated into silicate and metallic portions with distinct W isotopic compositions. The same applies to any later accreted material, such as the putative Moon-forming giant impactor. Many models of terrestrial accretion and core formation involve a core that developed during the first 90% of accretion history, generally considered to correspond to a time span significantly shorter than that permitted by the W isotopic data. These models are difficult to reconcile with the W isotopic data unless the proto-Earth was re-homogenized by a major impactor, or accretion took longer than currently estimated. The W isotopic compositions of early lunar rocks provide the best hope of determining which model of accretion and core formation is correct for the Earth. A conservative assessment of isotopic ages for lunar highlands rocks, combined with the constraints from W isotopic data, indicate that the onset of major terrestrial core segregation, the formation of the Moon and the development of a lunar magma ocean all took place within < 80 Myr at 4.47 ± 0.04 Ga. Certain isotopic ages for lunar rocks would be consistent with a more restricted time window of 4.50 ± 0.01 Ga. Potassium and Cr isotopic data indicate early volatile depletion of the material from which the Earth and Moon formed and constrain models of pre-core Pb isotopic evolution. The various estimates for the Pb isotopic composition of the BSE seem best explained by strongU/Pb fractionation accompanying terrestrial core formation. Using the 4.47 ± 0.04 Ga age of the core, the second stage Pb isotopic evolution reproduces reasonable estimates for the present day BSE Pb isotopic composition if the second stage238U204Pb (μ) is in the range of 8.9 ± 0.5. The ‘lead paradox’ is entirely predictable from the 4.47 ± 0.04 Ga age of the core. The similarities between the late ages of the Earths core, the Moon and the degassing of Xe from the terrestrial mantle are consistent with an accretion history which is more protracted than currently modelled. Alternatively, late impacts may have triggered all of these events. If a single late impact is invoked as an explanation, the Moon must have been derived primarily from the silicate portion of the impactor. Otherwise, the Hf-W data may define the age of a core that formed as a result of another impact, shortly prior to that which formed the Moon.

THE PRECISE MEASUREMENT OF TL ISOTOPIC COMPOSITIONS BY MC-ICPMS: APPLICATION TO THE ANALYSIS OF GEOLOGICAL MATERIALS AND METEORITES

The precision of Tl isotopic measurements by thermal ionization mass spectrometry (TIMS) is severely limited by the fact that Tl possesses only two naturally occurring isotopes, such that there is no invariant isotope ratio that can be used to correct for instrumental mass discrimination. In this paper we describe new chemical and mass spectrometric techniques for the determination of Tl isotopic compositions at a level of precision hitherto unattained. Thallium is first separated from the geological matrix using a two-stage anion-exchange procedure. Thallium isotopic compositions are then determined by multiple-collector inductively coupled plasma-mass spectrometry with correction for mass discrimination using the known isotopic composition of Pb that is admixed to the sample solutions. With these procedures we achieve a precision of 0.01–0.02% for Tl isotope ratio measurements in geological samples and this is a factor of ≥3–4 better than the best published results by TIMS. However, without adequate precautions, experimental artifacts can be generated that result in apparent Tl isotopic fractionations of up to one per mil. Analysis of five terrestrial samples indicate the existence of Tl isotopic variations related to natural fractionation processes on the Earth. Two of the three igneous rocks analyzed in this study display Tl isotopic compositions indistinguishable from our laboratory standard, the reference material NIST-997 Tl. A third sample, however, is characterized by eTl ≈ 2.5 ± 1.5, where eTl represents the deviation of the 205Tl/203Tl ratio of the sample relative to NIST-997 Tl in parts per 104. Even larger deviations were identified for two ferromanganese crusts from the Pacific Ocean, which display eTl-values of +5.0 ± 1.5 and +11.7 ± 1.3. We suggest that the large variability of Tl isotopic compositions in the latter samples are caused by low-temperature processes related to the formation of the Fe-Mn crusts by precipitation and scavenging from seawater. Thallium is the heaviest element for which such a natural isotopic fractionation has been verified. For a bulk sample of the C3V chondrite Allende we have determined a Tl isotopic composition of eTl = −2.1 ± 1.4, the lowest value measured in this study. By combining our data for Allende with a previously published Tl isotopic composition for the L3 chondrite Mezo-Madaras we obtain an upper limit of 9.2 × 10−5 for the 205Pb/204Pb abundance ratio of the bulk solar system at the time of volatile element depletion. If the chemical depletion of volatile elements occurred early, during the condensation of the solar nebula, our result would indicate that the initial Solar System abundance of 205Pb is significantly lower than some current astrophysical estimates.