An experimental investigation of the partitioning of REE between garnet and liquid with reference to the role of defect equilibria

The partitioning of samarium and thulium between garnets and melts in the systems Mg₃Al₂-Si₃O₁₂-H₂O and Ca₃Al₂Si₃O₁₂-H₂O has been studied as a function of REE concentration in the garnets at 30 kbar pressure. Synthesis experiments of variable time under constant P, T conditions indicate that garnet initially crystallizes rapidly to produce apparent values of D _Sm (D _Sm=concentration of Sm in garnet/concentration of Sm in liquid) which are too large in the case of pyrope and too small in the case of grossular. As the experiment proceeds, Sm diffuses out of or into the garnet and the equilibrium value of D _Sm is approached. Approximate values of diffusion coefficients for Sm in pyrope garnet obtained by this method are 6 × 10^–13 cm² s^–1 at 1,300 ° C and 2 × 10^–12 cm² s^–1 at 1,500 ° C, and for grossular, 8.3 × 10^–12 cm² s^–1 at 1,200 ° C and 4.6 × 10^–11 cm² s^–1 at 1,300 ° C. The equilibrium values of D _Sm have been reversed by experiments with Sm-free pyrope and Sm-bearing glass, and with Sm-bearing grossular and Sm-free glass.Between 12 ppm and 1,000 ppm Sm in pyrope at 1,300 ° C and between 80 ppm and >2 wt.% Tm in pyrope at 1,500 ° C, partition coefficients are constant and independent of REE concentration. Above 100 ppm of Sm in garnet at 1,500 ° C, partition coefficients are independent of Sm concentration. At lower concentrations, however, D _Sm is dependent upon the Sm content of the garnet. The two regions may be interpreted in terms of charge-balanced substitution of Sm₃Al₅O₁₂ in the garnet at high Sm concentrations and defect equilibria involving cation vacancies at low concentrations. At very low REE concentrations (< 1 ppm Tm in grossular at 1,300 ° C) D_REE garnet/liquid again becomes constant with an apparent Henry's Law value greater than that at high concentrations. This may be interpreted in terms of a large abundance of cation vacancies relative to the number of REE ions.The importance of defects in the low concentration region has been confirmed by adding other REE (at 80 ppm level) to the system Mg₃Al₂Si₃O₁₂-H₂O at low Sm concentrations. These change D _Sm in the defect region, demonstrating their role in the production of vacancies.Experiments on a natural pyropic garnet indicate that defect equilibria are of importance to REE partitioning within the concentration ranges found in nature.     

Experimental rare earth element partition coefficients for garnet,clinopyroxene and amphibole coexisting with andesitic and basaltic liquids

The partitioning of La, Sm, Dy, Ho and Yb between garnet, calcic clinopyroxene, calcic amphibole and andesitic and basaltic liquids has been studied experimentally. Glasses containing one or more REE in concentrations of 500–2000 ppm were crystallized at pressures of 10–35 kbar, and temperatures of 900–1520°C. Water was added to stabilize amphibole and to allow study of partition coefficients over wide temperature ranges. Major element and REE contents of crystal rims and adjacent glass were determined by EPMA, with limits of detection for individual REE of 100–180 ppm. Measured partition coefficients, D_REE^Cryst-liq, are independent of REE concentration over the concentration ranges used.D-values show an inverse dependence on temperature, best illustrated for garnet. At a given temperature, they are almost always higher for equilibria involving andesitic liquid. Garnet shows by far the greatest range of D-values, with e.g. D_La < 0.05 and D_Yb ~ 44 for andesitic liquid at 940°C. D_Yb falls to ~ 12 at 1420°C. D_Sm^Ga-liq also correlates negatively with temperature and positively with the grossular content of garnet. Patterns of D_ree^Cryst-Liq for calcic clinopyroxenes and amphiboles are sub-parallel, with D-values for amphibole generally higher. Both individual D-values and patterns for the crystalline phases studied are comparable with those determined for phenocryst-matrix pairs in natural dacites, andesites and basalts.D-values and patterns are interpreted in terms of the entry of REE³⁺ cations into mineral structures and liquids of contrasted major element compositions. The significance of the partition coefficients for models of the genesis of andesitic and Hy-normative basaltic magmas is assessed. Most magmas of these types in island arcs are unlikely to be produced by melting of garnet-bearing sources such as eclogite or garnet lherzolite.     

Partitioning of rare earth elements between garnet and andesite melt: an autoradiographic study of P-T-X effects

High-pressure equilibrium studies were conducted in piston-cylinder apparatus to determine rare earth element (REE) partitioning between garnet and H₂O-vapor-saturated liquidus, from 20 kbar/980°C to 30 kbar/1060 °C. Ag capsules were employed to eliminate loss of Fe. Partition coefficients (K_D's) were determinined with autoradiographic techniques employing beta-active isotopes of Ce, Sm, and Tm. Major elements in garnet were determined by microprobe analysis. Synthesis and reversal runs of 24 hr or greater duration were used to bracket values of K_D's within analytical uncertainty.The K_D values for all three REE are constant over the radiogenic concentration range of 1 to 350 ppm, suggesting that the high abundance of natural REE in the starting materials may suppress possible deviations from Henry's Law behavior reported in similar autoradiographic studies of synthetic systems with no natural REE. Changes in K_D with increasing pressure and temperature at near-liquidus conditions suggest that the dominant control of K_D is the average size of cations occupying the 8-fold sites in garnet. Specifically, the substitution of 8-fold Ca for Mg and Fe causes an increase in K_D values greater than that attributable to the coincident effects of pressure and temperature. The inverse correlation of increasing K_D with REE ionic radius supports the interpretation that the average size of the 8-fold cation controls the relative variation of K_D among REE.These experimental K_D values produce less relative fractionation in melts between light REE and heavy REE than do previous K_D's derived from data on coexisting natural garnet phenocryst/rockmatrix pairs. Models for the derivation of orogenic andesites from partial melting of subducted basaltic eclogite are qualitatively improved by these new K_D's. Existing calculations of K_D values necessary for the viability of the eclogite fractionation are also in good agreement with these experimental K_D values.     

Determination of zircon/melt trace element partition coefficients from SIMS analysis of melt inclusions in zircon

Partition coefficients (^zircon/meltD_M) for rare earth elements (REE) (La, Ce, Nd, Sm, Dy, Er and Yb) and other trace elements (Ba, Rb, B, Sr, Ti, Y and Nb) between zircon and melt have been calculated from secondary ion mass spectrometric (SIMS) analyses of zircon/melt inclusion pairs. The melt inclusion-mineral (MIM) technique shows that D_REE increase in compatibility with increasing atomic number, similar to results of previous studies. However, D_REE determined using the MIM technique are, in general, lower than previously reported values. Calculated D_REE indicate that light REE with atomic numbers less than Sm are incompatible in zircon and become more incompatible with decreasing atomic number. This behavior is in contrast to most previously published results which indicate D > 1 and define a flat partitioning pattern for elements from La through Sm. The partition coefficients for the heavy REE determined using the MIM technique are lower than previously published results by factors of ≈15 to 20 but follow a similar trend. These differences are thought to reflect the effects of mineral and/or glass contaminants in samples from earlier studies which employed bulk analysis techniques.D_REE determined using the MIM technique agree well with values predicted using the equations of Brice (1975), which are based on the size and elasticity of crystallographic sites. The presence of Ce⁴⁺ in the melt results in elevated D_Ce compared to neighboring REE due to the similar valence and size of Ce⁴⁺ and Zr⁴⁺. Predicted ^zircon/meltD values for Ce⁴⁺ and Ce³⁺ indicate that the Ce⁴⁺/Ce³⁺ ratios of the melt ranged from about 10⁻³ to 10⁻². Partition coefficients for other trace elements determined in this study increase in compatibility in the order Ba < Rb < B < Sr < Ti < Y < Nb, with Ba, Rb, B and Sr showing incompatible behavior (D_M < 1.0), and Ti, Y and Nb showing compatible behavior (D_M > 1.0).The effect of partition coefficients on melt evolution during petrogenetic modeling was examined using partition coefficients determined in this study and compared to trends obtained using published partition coefficients. The lower D_REE determined in this study result in smaller REE bulk distribution coefficients, for a given mineral assemblage, compared to those calculated using previously reported values. As an example, fractional crystallization of an assemblage composed of 35% hornblende, 64.5% plagioclase and 0.5% zircon produces a melt that becomes increasingly more enriched in Yb using the D_Yb from this study. Using D_Yb from Fujimaki (1986) results in a melt that becomes progressively depleted in Yb during crystallization.     

REE in skarn systems: A LA-ICP-MS study of garnets from the Crown Jewel gold deposit

Metamorphic and magmatic garnets are known to fractionate REE, with generally HREE-enriched patterns, and high Lu/Hf and Sm/Nd ratios, making them very useful as geochemical tracers and in geochronological studies. However, these garnets are typically Al-rich (pyrope, almandine, spessartine, and grossular) and little is known about garnets with a more andraditic (Fe³⁺) composition, as frequently found in skarn systems. This paper presents LA-ICP-MS data for garnets from the Crown Jewel Au-skarn deposit (USA), discusses the factors controlling incorporation of REE into garnets, and strengthens the potential of garnet REE geochemistry as a tool to help understand the evolution of metasomatic fluids.Garnets from the Crown Jewel deposit range from Adr₃₀Grs₇₀ to almost pure andradite (Adr_>99). Fe-rich garnets (Adr_>90) are isotropic, whereas Al-rich garnets deviate from cubic symmetry and are anisotropic, often showing sectorial dodecahedral twinning. All garnets are extremely LILE-depleted, Ta, Hf, and Th and reveal a positive correlation of ΣREE³⁺ with Al content. The Al-rich garnets are relatively enriched in Y, Zr, and Sc and show “typical” HREE-enriched and LREE-depleted patterns with small Eu anomalies. Fe-rich garnets (Adr_>90) have much lower ΣREE and exhibit LREE-enriched and HREE-depleted patterns, with a strong positive Eu anomaly. Incorporation of REE into garnet is in part controlled by its crystal chemistry, with REE³⁺ following a coupled, YAG-type substitution mechanism , whereas Eu²⁺ substitutes for X²⁺ cations. Thermodynamic data (e.g., H_mixing) in grossular-andradite mixtures suggest preferential incorporation of HREE in grossular and LREE in more andraditic compositions.Variations in textural and optical features and in garnet geochemistry are largely controlled by external factors, such as fluid composition, W/R ratios, mineral growth kinetics, and metasomatism dynamics, suggesting an overall system that shifts dynamically between internally and externally buffered fluid chemistry driven by fracturing. Al-rich garnets formed by diffusive metasomatism, at low W/R ratios, from host-rock buffered metasomatic fluids. Fe-rich garnets grow rapidly by advective metasomatism, at higher W/R ratios, from magmatic-derived fluids, consistent with an increase in porosity by fracturing.     

Experimental determination of REE partition coefficients between amphibole and basaltic to andesitic liquids at high pressure

Amphibole/liquid partition coefficients for the REE(D^amph/liq_REE) obtained from natural rocks increase systematically with bulk rock compositional change from basalt to rhyolite and are higher for the middle to heavy REE. Five new experimentally determined sets of D_REE (La, Sm, “Eu²⁺”, Ho, Lu)and 4 published sets are similar to these data and provide values for use in geochemical modelling of magmatic processes involving amphibole, over a range of temperature, pressure and oxygen fugacity. The partition coefficients increase significantly with decreasing temperature, and increase slightly with increasing oxygen fugacity. Pressure does not appear to have a major effect, although one data set suggests increased pressure lowers D^amph/liq_REE in a basaltic composition.     

Kiglapait geochemistry VII: Yttrium and the rare earth elements

Based on 51 wholerock analyses by XRF and summation over the layered group, the Kiglapait Intrusion contains 4.7_?1.6^+1.2 ppm Y, which resides principally in augite and apatite. Using liquid compositions calculated by summation, the partition coefficient D^AUG/L_Y is 0.95 ± 0.12 from 84 to 97 PCS (percent solidified) and 1.5 ± 0.4 above 97 PCS. For feldspar, the most likely value for D is 0.028 ± 0.02 (N = 6).REE analyses for 13 whole rocks were interpreted with the aid of yttrium models to yield trends for wholerocks and liquids vs PCS. Summations over the rocks of the layered group gave La = 2.5, Ce = 5.8, Nd = 3.9, Sm = 1.0, Eu = 0.8, Tb = 0.17, Yb = 0.37, and Lu = 0.06 ppm, with 2 s.d. errors near ± 30%. All these elements are highly incompatible until the arrival of augite, which affects chiefly the HREE, and apatite, which affects all (but more strongly, the LREE). The net result is that after apatite arrival at 94 PCS, the liquid compositions are nearly constant, hence D^WR/L_REE ≈ 1.0. These results are compatible with the mineralogy of the intrusion and the estimated partition coefficients for feldspar, olivine, augite, apatite, and Fe-Ti oxide minerals. For pre-apatite liquids, D^FSP/L_REE vary regularly with the normative di content of the liquid and change by an order of magnitude, hence the bulk liquid composition must be considered in any attempt to invert the compositions of feldspars to parent liquids.The Eu anomaly at first decreases in Kiglapait liquids due to plagioclase fractionation, but then increases due to removal of augite and apatite with negative Eu anomalies. The features dominantly responsible for Eu partitioning are liquid structure and, for monoclinic ternary feldspars, crystal structure. The former is best monitored by the augite or diopside content of the liquid and the latter, by the K content of the feldspar.The chondrite-normalized REE pattern for the intrusion has La_N = 7.4, Lu_N = 1.6, (Ce/Yb)_N = 3.6, and Eu/Eu* = 2.4, indicating its feldspar-rich nature. The chilled margin of the nearby Hettasch Intrusion has a similar but more evolved pattern, corresponding roughly to the Kiglapait liquid at 70 PCS. As with other data, those for the REE suggest source differences for the two intrusions rather than a relationship due to fractionation.     

Biogenic and abiogenic low-Mg calcite (bLMC and aLMC): Evaluation of seawater-REE composition,water masses and carbonate diagenesis

The ΣREE and shale-normalized (PAAS) REE_SN values of modern brachiopods (biogenic low-Mg calcite: bLMC) represented by several species from high- to low latitudes, from shallow- to deep waters and from warm- and cold-water environments, define three distinct average ‘seawater’ trends. The warm- and cold-water brachiopods define two indistinguishable (p < 0.050) groups that mimic open-ocean seawater REE chemistry, exhibiting the typical LREE enrichment with a slightly positive to negative Ce anomaly followed by an otherwise invariant series. Other recent brachiopods from an essentially siliciclastic seabed environment are distinct in both ΣREE and REE_SN trends from the previous two populations, showing a slight enrichment in the MREEs and an increasing trend in the HREEs. Other groups of modern brachiopods are characterized by elevated REE_SN trends relative to the ‘normal’ groups as well as by complexity of the series trends. The most characteristic feature is the decrease in the HREEs in these brachiopods from areas of unusual productivity (i.e., such as upwelling currents, fluvial input and aerosol dust deposition). Preserved brachiopods from the Eocene and Silurian exhibit REE_SN trends and Ce anomalies similar to that of the ‘open-ocean’ modern brachiopods, although, their enriched ΣREE concentrations suggest precipitation of bLMC influenced by extrinsic environmental conditions.Preservation of the bLMC was tested by comparing the ΣREE and REE_SN trends of preserved Eocene brachiopods to those of Oligocene brachiopods that were altered in an open diagenetic system in the presence of phreatic meteoric-water. The altered bLMC is enriched by approximately one order of magnitude in both ΣREE and REE_SN trends relative to that in bLMC of their preserved counterparts. Similarly, the ΣREE and REE_SN of preserved Silurian brachiopod bLMC were compared to those of their enclosing altered lime mudstone, which exhibits features of partly closed system, phreatic meteoric-water diagenesis. Despite these differences in the diagenetic alteration systems and processes, the ΣREEs and REE_SN trends of the bLMC of altered brachiopods and of originally mixed mineralogy lime mudstones (now diagenetic low-Mg calcite) are enriched by about one order of magnitude relative to those observed in the coeval and preserved bLMC.In contrast to the changes in ΣREE and REE_SN of carbonates exposed to phreatic meteoric-water diagenesis, are the REE compositions of late burial calcite cements precipitated in diagenetically open systems from burial fluids. The ΣREE and REE_SN trends of the burial cements mimic those of their host lime mudstone, with all showing slight LREE enrichment and slight HREE depletion, exhibiting a ‘chevron’ pattern of the REE_SN trends. The overall enrichment or depletion of the cement REE_SN trends relative to that of their respective host rock material reflects not only the openness of the diagenetic system, but also strong differences in the elemental and REE compositions of the burial fluids. Evaluation of the (Ce/Ce*)_SN and La = (Pr/Pr*)_SN anomalies suggests precipitation of the burial calcite cements essentially in equilibrium with their source fluids.     

Experimental determination of REE partition coefficients between zircon,garnet and melt: a key to understanding high‐T crustal processes

The partitioning of rare earth elements (REE) between zircon, garnet and silicate melt was determined using synthetic compositions designed to represent partial melts formed in the lower crust during anatexis. The experiments, performed using internally heated gas pressure vessels at 7 kbar and 900–1000 °C, represent equilibrium partitioning of the middle to heavy REE between zircon and garnet during high‐grade metamorphism in the mid to lower crust. The D_REE (zircon/garnet) values show a clear partitioning signature close to unity from Gd to Lu. Because the light REE have low concentrations in both minerals, values are calculated from strain modelling of the middle to heavy REE experimental data; these results show that zircon is favoured over garnet by up to two orders of magnitude. The resulting general concave‐up shape to the partitioning pattern across the REE reflects the preferential incorporation of middle REE into garnet, with D_Gd (zircon/garnet) ranging from 0.7 to 1.1, D_Ho (zircon/garnet) from 0.4 to 0.7 and D_Lu (zircon/garnet) from 0.6 to 1.3. There is no significant temperature dependence in the zircon–garnet REE partitioning at 7 kbar and 900–1000 °C, suggesting that these values can be applied to the interpretation of zircon–garnet equilibrium and timing relationships in the ultrahigh‐T metamorphism of low‐Ca pelitic and aluminous granulites.     

Trace-element partitioning between garnet,clinopyroxene and Fe-rich picritic melts at 3 to 7 GPa

We have determined mineral-melt partition coefficients (D values) for 20 trace elements in garnet-pyroxenite run products, generated in 3 to 7 GPa, 1,425–1,750°C experiments on a high-Fe mantle melt (97SB68) from the Paraná-Etendeka continental-flood-basalt (CFB) province. D values for both garnet (∼Py₆₃Al₂₅Gr₁₂) and clinopyroxene (∼Ca_0.2Mg_0.6Fe_0.2Si₂O₆) show a large variation with temperature but are less dependent on pressure. At 3 GPa, D ^cpx/liq values for pyroxenes in garnet-pyroxenite run products are generally lower than those reported from Ca-rich pyroxenes generated in melting experiments on eclogites and basalts (∼Ca_0.3–0.5Mg_0.3–0.6Fe_0.07–0.2Si₂O₆) but higher than those for Ca-poor pyroxenes from peridotites (∼Ca_0.2Mg_0.7Fe_0.1Si₂O₆). D ^grt/liq values for light and heavy rare-earth elements are ≤0.07 and >0.8, respectively, and are similar to those for peridotitic garnets that have comparable grossular but higher pyrope contents (Py_70–88Al_l7–20Gr_8–14). 97SB68 D _LREE^grt/liq values are higher and D _HREE^grt/liq values lower than those for eclogitic garnets which generally have higher grossular contents but lower pyrope contents (Py_20–70Al_10–50Gr_10–55). D values agree with those predicted by lattice strain modelling and suggest that equilibrium was closely approached for all of our experimental runs. Correlations of D values with lattice-strain parameters and major-element contents suggest that the wollastonite component and pyrope:grossular ratio exert major controls on 97SB68 clinopyroxene and garnet partitioning, respectively. These are controlled by the prevailing pressure and temperature conditions for a given bulk-composition. The composition of co-existing melt was found to have a relatively minor effect on 97SB68 D values. The variations in D values displayed by different mantle lithologies are subtle and our study confirms previous investigations which have suggested that the modal proportions of garnet and clinopyroxene are by far the most influential factor in determining incompatible trace-element concentrations in mantle melts. The trace-element partition coefficients we have determined may be used to place high-pressure constraints on garnet-pyroxenite melting models.     

Geochemical and mineralogical characteristics of ion-adsorption type REE mineralization in Phuket, Thailand

Geochemical and mineralogical studies were conducted on the 12-m-thick weathering profile of the Kata Beach granite in Phuket, Thailand, in order to reveal the transport and adsorption of rare earth elements (REE) related to the ion-adsorption type mineralization. The parent rock is ilmenite-series biotite granite with transitional characteristics from I type to S type, abundant in REE (592 ppm). REE are contained dominantly in fluorocarbonate as well as in allanite, titanite, apatite, and zircon. The chondrite-normalized REE pattern of the parent granite indicates enrichment of LREE relative to HREE and no significant Ce anomaly. The upper part of the weathering profile from the surface to 4.5 m depth is mostly characterized by positive Ce anomaly, showing lower REE contents ranging from 174 to 548 ppm and lower percentages of adsorbed REE from 34% to 68% compared with the parent granite. In contrast, the lower part of the profile from 4.5 to 12 m depth is characterized by negative Ce anomaly, showing higher REE contents ranging from 578 to 1,084 ppm and higher percentages from 53% to 85%. The negative Ce anomaly and enrichment of REE in the lower part of the profile suggest that acidic soil water in an oxidizing condition in the upper part mostly immobilized Ce⁴⁺ as CeO₂ and transported REE³⁺ downward to the lower part of the profile. The transported REE³⁺ were adsorbed onto weathering products or distributed to secondary minerals such as rhabdophane. The immobilization of REE results from the increase of pH due to the contact with higher pH groundwater. Since the majority of REE in the weathered granite are present in the ion-adsorption fraction with negative Ce anomaly, the percentages of adsorbed REE are positively correlated with the whole-rock negative Ce anomaly. The result of this study suggests that the ion-adsorption type REE mineralization is identified by the occurrence of easily soluble REE fluorocarbonate and whole-rock negative Ce anomaly of weathered granite. Although fractionation of REE in weathered granite is controlled by the occurrence of REE-bearing minerals and adsorption by weathering products, the ion-adsorption fraction tends to be enriched in LREE relative to weathered granite.     

Adsorption of REE(III)-humate complexes onto MnO2: Experimental evidence for cerium anomaly and lanthanide tetrad effect suppression

Experiments were conducted to evaluate the impact of organic complexation on the development of Ce anomalies and the lanthanide tetrad effect during the adsorption of rare-earth elements (REE) onto MnO₂. Two types of aqueous solutions—NaCl and NaNO₃—were tested at pH 5 and 7.5. Time-series experiments indicate that a steady-state is reached within less than 10 h when REE occur as free inorganic species, whereas steady state is not reached before 10 d when REE occur as REE-humate complexes. The distribution coefficients (K_d^REE) between suspended MnO₂ and solution show no or only very weak positive Ce anomaly or lanthanide tetrad effect when REE occur as humate complexes, unlike the results obtained in experiments with REE occurring as free inorganic species. Monitoring of dissolved organic carbon (DOC) concentrations show that log K_d^REE_organic/K_d^DOC ratios are close to 1.0, implying that the REE and humate remain bound to each other upon adsorption. Most likely, the Ce anomaly reduction/suppression in the organic experiments arises from a combination of two processes: (i) inability of MnO₂ to oxidize Ce(III) because of shielding of MnO₂ surfaces by humate molecules and (ii) Ce(IV) cannot be preferentially removed from solution due to quantitative complexation of the REE by organic matter. We suggest that the lack of lanthanide tetrad effect arises because the adsorption of REE-humate complexes onto MnO₂ occurs dominantly via the humate side of the complexes (anionic adsorption), thereby preventing expression of the differences in Racah parameters for 4f electron repulsion between REE and the oxide surface. The results presented here explain why, despite the development of strongly oxidizing conditions and the presence of MnO₂ in the aquifer, no (or insignificant) negative Ce anomalies are observed in organic-rich waters. The present study demonstrates experimentally that the Ce anomaly cannot be used as a reliable proxy of redox conditions in organic-rich waters or in precipitates formed at equilibrium with organic-rich waters.     

Experimental study of polybaric REE partitioning between olivine, pyroxene and melt of the Yamato 980459 composition: Insights into the petrogenesis of depleted shergottites

A synthetic composition representing the Yamato 980459 martian basalt (shergottite) has been used to carry out phase relation, and rare earth element (REE) olivine and pyroxene partitioning experiments. Yamato 980459 is a sample of primitive basalt derived from a reduced end-member among martian mantle sources. Experiments carried out between 1-2 GPa and 1350-1650 °C simulate the estimated pressure-temperature conditions of basaltic melt generation in the martian mantle. Olivine-melt and orthopyroxene-melt partition coefficients for La, Nd, Sm, Eu, Gd and Yb (D_REE values) were determined by LA-ICPMS, and are similar to the published values for terrestrial basaltic systems. We have not detected significant variation in D-values with pressure over the range investigated, and by comparison with previous studies carried out at lower pressure.We apply the experimentally obtained olivine-melt and orthopyroxene-melt D_REE values to fractional crystallization and partial melting models to develop a three-stage geochemical model for the evolution of martian meteorites. In our model we propose two ancient (∼4.535 Ga) sources: the Nakhlite Source, located in the shallow mantle, and the Deep Mantle Source, located close to the martian core-mantle boundary. These two sources evolved distinctly on the ε¹⁴³Nd evolution curve due to their different Sm/Nd ratios. By partially melting the Nakhlite Source at ∼1.3 Ga, we are able to produce a slightly depleted residue (Nakhlite Residue). The Nakhlite Residue is left undisturbed until ∼500 Ma, at which point the depleted Deep Mantle Source is brought up by a plume mechanism carrying with it high heat flow, melts and isotopic signatures of the deep mantle (e.g., ε¹⁸²W, ε¹⁴²Nd, etc.). The plume-derived Deep Mantle Source combines with the Nakhlite Residue producing a mixture that becomes a mantle source (herein referred to as “the Y98 source”) for Yamato 980459 and the other depleted shergottites with the characteristic range of Sm/Nd ratios of these meteorites. The same hot plume provides a heat source for the formation of enriched and intermediate shergottites. Our model reproduces the REE patterns of nakhlites and depleted shergottites and can explain high ε¹⁴³Nd in depleted shergottites. Furthermore, the model results can be used to interpret whole rock Rb-Sr and Sm-Nd ages of shergottites.     

XAFS characterization of the structural site of Yb in synthetic pyrope and grossular garnets

The incorporation and site preference of minor amounts (about 1 wt%) of Yb³⁺ in synthetic pyrope (Mg₃Al₂Si₃O₁₂) and grossular (Ca₃Al₂Si₃O₁₂) garnet were studied by X-ray Absorption Fine-Structure (XAFS) Spectroscopy. The measurements, performed in the temperature range 77–343 K at both Yb L_I- and L_III-edges, demonstrate that Yb³⁺ enters the garnet structure and is located in the dodecahedral site in both samples. The coordination environment of Yb³⁺ in the two samples was compared to that of the X-site cation in end-member synthetic pyrope and grossular and in Yb₃Al₅O₁₂ as determined by single-crystal X-ray diffraction. The local geometry around Yb³⁺ is different from that of Mg and Ca in the bulk of the garnet, and also from that of Yb³⁺ in Yb₃Al₅O₁₂. Τhe XAFS results indicate that, (1) structural relaxation occurs around Yb³⁺ in the garnet structure; (2) the host garnet matrix exerts a major structural control on the incorporation of Yb³⁺, and (3) minor amounts of Yb³⁺ in garnet are located in structural sites and not in ill-defined defects. Received: 15 January 1998/ Revised, accepted: 21 July 1998     

Partitioning of Ni between olivine and silicate melt: The ‘Henry's Law problem’ reexamined

The partitioning of Ni between olivine and silicate melt has been investigated experimentally at atmospheric pressure in air. Beta track autoradiography using ⁶³Ni and direct microprobe analysis of polished run products were employed. At constant temperature and bulk composition, the $\text{olivine}\text{liquid}$ partition coefficient for Ni in the system Di₇₀Fo₂₅Qtz₅ remains independent of concentration from approximately 10 ppm to 40,000 ppm Ni in olivine. Similar experiments by Mysen (1979) in the system Jd₈₀Fo₂₀ utilizing beta track autoradiography alone indicated that Henry's Law was followed only in the concentration interval of approx. 10–1000 ppm Ni in olivine. Above 1000 ppm, olivine/liquid partition coefficients decreased monotonically to about half of the value observed below 1000 ppm. We have performed experiments in the system Jd₈₀Fo₂₀, but are unable to replicate Mysen's results. While in agreement with Mysen below 1000 ppm Ni in olivine, we do not observe the decrease in partition coefficient value at higher concentrations. We conclude from our reversed experiments that, at constant temperature and bulk composition, the $\text{olivine}\text{liquid}$ partition coefficient for Ni in the system Jd₈₀Fo₂₀ remains independent of concentration from approx. 10–60,000 ppm Ni in olivine. Attempts to resolve these differing conclusions by changing experimental techniques have been unsuccessful.     

Stabilisation of divalent rare earth elements in natural fluorite

Summary ?The occurrence of divalent rare earth elements (Sm²⁺, Yb²⁺, Tm²⁺, and Ho²⁺) in natural fluorite is evaluated using a suite of 37 samples deriving mainly from Sn–W deposits in the Erzgebirge (Germany), Central Kazakhstan, and the Mongolian Altai. Trace element composition was determined by ICP-AES and ICP-MS. The defect structure of the samples was studied by cathodoluminescence (CL), electron paramagnetic resonance (EPR), and optical absorption spectroscopy. Reduction of cubic Sm³⁺, Yb³⁺, Tm³⁺, and Ho³⁺ under radioactive irradiation produces the corresponding divalent centres. Our data suggest a preferable formation of Sm²⁺ and Yb²⁺ under thorium and of Tm²⁺ and Ho²⁺ under uranium irradiation. Irradiation (indicated by intense brownish (thorium) and deep purple (uranium) coloration of fluorite) gives rise to a population of divalent centres in equilibrium with their decay. However, sporadic radioactive irradiation and stabilisation of the divalent state of the REE by other electron defects were found in most cases. Three models of stabilisation of Sm²⁺, Yb²⁺, Tm²⁺, and Ho²⁺ are discussed. The most effective mechanism for Sm, Yb, Tm, and Ho is coupling with Fe³⁺ centres (REE³⁺+Fe²⁺ → REE²⁺+Fe³⁺). Accordingly, the occurrence of Fe³⁺ centres in natural fluorite is regarded to indicate not an oxidising, but rather a reducing environment during fluorite precipitation. Originally incorporated in the divalent form, Fe²⁺ was converted to Fe³⁺ by radioactive irradiation. Such a conclusion is in agreement with the finding of high contents of interstitial fluorine providing tetragonal local compensation of trivalent REE centres in crystals with high Fe³⁺. If Fe is not present, compensation of divalent Sm, Yb, and Tm is achieved by radiogenic oxidation of Ce(Pr, Tb)³⁺ accompanied by charge transfer (REE³⁺+Ce(Pr, Tb)³⁺ → REE²⁺+ Ce(Pr, Tb)⁴⁺). Ho²⁺ is sometimes stabilised by a hole trapped by an electron localised on a F vacancy (Ho³⁺+e⁻ on □_F → REE²⁺+ self-trapped exciton). Because Sm²⁺ is optically active, the stabilisation by Fe³⁺ (stable up to temperatures above 350 °C) or Ce(Pr, Tb)⁴⁺ (unstable even under visible light) in samples may be determined by careful observations in the field. Institut für Geotechnik, ETH Zürich, ETH-H?nggerberg, Zürich, Switzerland Stanford Linear Accelerator Center, Menlo Park, CA, USA Received January 8, 2002; revised version accepted June 10, 2002     

Non-Henry’s Law behavior of REE partitioning between fluorapatite and CaF2-rich melts: Controls of intrinsic vacancies and implications for natural apatites

The partitioning of rare-earth elements (REEs: Gd and multiple REEs), Sr, and Mn between fluorapatite and CaF₂-rich melts was investigated over a wide range of REE concentrations (i.e., from 0.8 ± 0.1 to 25,000 ± 2600 ppm Gd in fluorapatite) in two different sample assemblies (i.e., tightly covered Pt crucibles and sealed Pt capsules) at 1220 °C and atmospheric pressure. Attainment of equilibrium is indicated by selected reversal experiments. The partition coefficient D(Gd) decreases from ∼2 to ∼0.5 with increasing Gd in fluorapatite, hence a marked non-Henry’s Law behavior, but becomes independent of composition at and above ∼5000 and ∼1000 ppm Gd for experiments in Pt crucibles and Pt capsules, respectively. Non-Henry’s Law behavior is also observed in experiments involving multiple REEs. All REE patterns are convex upward in shape with maxima between Nd and Gd, and D(La)/D(Nd) and D(Nd)/D(Yb) decrease systematically with increasing total REEs in fluorapatite, suggesting that REE fractionations are partly related to non-Henry’s Law behavior. These experimental results and local structural data from previous electron paramagnetic resonance spectroscopic studies suggest that the non-Henry’s Law behavior of REE partitioning between fluorapatite and melt is controlled by intrinsic Ca²⁺ vacancies in the c-axis channels. The D(Sr) and D(Mn) values are independent of composition and, therefore, do not deviate from the Henry’s Law in their respective compositional ranges investigated in this study.Nonstoichiometry, such as Ca²⁺ and F⁻ vacancies in the c-axis channels, is well known in natural apatites, particularly in biogenic apatites. Therefore, the observed non-Henry’s Law behavior of REE partitioning is expected to have important implications for REE geochemical modeling involving apatites and for the uptake of REEs by natural apatites. Particularly, the non-Henry’s Law behavior of REE partitioning is at least partly responsible for the commonly observed, bell-shaped REE patterns in fossil biogenic apatites.     

Sm-Nd isotope systematics in garnet from different lithologies (Eastern Alps): age results, and an evaluation of potential problems for garnet Sm-Nd chronometry [Chem. Geol. 185 (2002) 255-281]

Interpretation of Sm-Nd garnet ages is frequently impaired by one of the following restrictions: (a) high-LREE inclusions, (b) isotopic disequilibrium, and (c) the uncertainty about the closure temperature. These issues are addressed by way of an evaluation of garnet Sm-Nd data from different rock types of the Austroalpine basement units, Eastern Alps, including metabasic eclogites, mica schist and paragneiss, metapegmatite and metagranite.Nd concentration in handpicked garnet varies between 0.021 and 23.1 ppm in metabasites, 0.49 and 17.4 ppm in metapelites and between 0.024 and 4.6 ppm in metapegmatites and metagranites. The overall range of ¹⁴⁷Sm/¹⁴⁴Nd is 0.15-2.5 in garnet from metabasites, 0.12-3.03 in metapelite garnet and 0.66-7.21 in Mn-rich garnet from metapegmatites and metagranites. A clear negative correlation between Nd concentration and Sm/Nd is observed in garnets from all these lithologies. Therefrom, it is concluded that even optically “clean” garnet separates may contain high-LREE microinclusions, such as epidote-allanite, zoisite, apatite, sphene, monazite or zircon. However, very low Nd concentrations correlated with low Sm/Nd as well as high Nd concentrations (>5 ppm) correlated with fairly high Sm/Nd ratios (0.8) have also been observed. Apart from replicate analyses within as well as between samples with a common PT-history, leaching experiments are a useful technique to elucidate any distorting influence of unequilibrated inclusions on the garnet age, especially if the observed Sm/Nd ratio is low (<0.5). Leaching of garnet separates with HCl (2.5, 5.8 M) produces no obvious element fractionation, but may improve Sm/Nd, and hence age precision, considerably. Isotopic disequilibrium between garnet and other matrix minerals is observed preferentially in basic eclogites, derived from gabbroic precursors.Sm-Nd garnet analysis allows the recognition of several distinct garnet-forming events in the Eastern Alps.(a) A Variscan high-P event is documented in metabasites from the northern-central Ötztal basement around 360-350 Ma, whereas garnet from sillimanite-bearing gneisses dates the Variscan thermal peak in the western part of the same subunit around 345-330 Ma.(b) A long-lived, Permian to Triassic event (285-225 Ma), correlated with crustal extension and low-P metamorphism, is documented by spessartine-rich garnet from metapegmatites as well as almandine-rich garnet cores from mica schist.(c) Age data of garnet from eo-Alpine (Cretaceous) deeply subducted rocks of the southern/eastern Austroalpine units are related to near-peak PT, eclogite- to amphibolite-facies metamorphic conditions (peak: 2 GPa/685 °C), and/or incipient isothermal decompression, due to fast, tectonically driven exhumation (110/100-85 Ma). At cooling rates of 20-30 °C/Ma (exhumation rates: 3-5 km/Ma), the Sm-Nd closure temperature (T_c) for mm-sized garnet in these rocks is estimated at 650-680 °C.     

Origins of Nd–Sr–Pb isotopic variations in single scheelite grains from Archaean gold deposits,Western Australia

Accessory gangue scheelite (CaWO₄) from the Archaean Mt. Charlotte lode Au deposit can be divided into two types with different rare earth element (REE) signatures. In some scheelite grains, specific REE signatures are reflected by different cathodoluminescence colours, which can be used to map their often complex oscillatory intergrowths. Domains with specific REE contents from two grains were sampled for Sm/Nd, Rb/Sr and Pb isotopic analyses using a micro-drilling technique.Type I scheelite is strongly enriched in middle REE (MREE) and Eu anomalies are either absent or slightly positive. Four fragments collected from Type I regions of two crystals have initial ⁸⁷Sr/⁸⁶Sr and ε_Nd values ranging from 0.70141 to 0.70163 and +2.5 to +3.5, respectively, and Pb isotope ratios reflecting the composition of greenstone sequence. This may indicate that Nd and Pb have their source, either locally or regionally, in the greenstones. Basic greenstone lithologies have ⁸⁷Sr/⁸⁶Sr<0.7015, and the radiogenic Sr signatures indicate that part of the Sr originated from felsic lithologies located either within or beneath the host greenstone pile. Alternatively, the Sr signature may have evolved from preferential leaching of a Rb-rich mineral during hydrothermal alteration of the greenstone.The REE patterns of Type II scheelite are either flat or MREE-depleted and have strong positive Eu anomalies. Three fragments collected from Type II regions of the same two crystals have initial ⁸⁷Sr/⁸⁶Sr ratios and ε_Nd values between 0.70130 and 0.70146, and +1.1 to +2.6, respectively, and Pb isotope signatures that are once again similar to that of the greenstone. This implies that ⁸⁷Sr/⁸⁶Sr ratios in Type II fluids were closer to those of the host dolerite (0.7008–0.7013), due to more extensive fluid interaction with the dolerite.A positive correlation between Na and REE suggests that REE³⁺ are accommodated by the coupled substitution REE³⁺+Na⁺=2 Ca²⁺ into both Type I and Type II scheelite. This is consistent with a fractional crystallisation model to explain the change in REE patterns from Type I to Type II, but not with a model involving different coupled substitutions and fluids from different origins. We propose that the complex REE and isotopic signatures of scheelite at Mt. Charlotte are related to small (<m) to medium (<km) scale processes involving mixing between “fresh” batches of hydrothermal fluid with fluids that had already been involved in extensive wall-rock alteration.The very high-ε_Nd values measured in some scheelites have been previously used to link gold mineralisation with komatiites containing unusually high Sm/Nd ratios. However, tiny (<20 μm) grains of secondary hydroxyl-bastnäsite were found within micro-fractures of one scheelite grain containing an extremely high-ε_Nd signature. The hydroxyl-bastnäsite probably formed during recent REE redistribution within the scheelite as a result of meteoric fluid circulation. The scale of this cryptic low-temperature alteration is sufficient to explain the anomalously high-ε_Ndi values observed in scheelite from Western Australia.     

Pressure-Dependence of Rare Earth Element Distribution in Amphibolite- and Granulite- Grade Garnets. A LA-ICP-MS Study

Using an excimer (KrF) laser ablation ICP-MS system, we studied the distribution of REE in garnets from metapelites and metabasites from Ivrea-Verbano (Western Alps, Italy) and from the Peña Negra Anatectic Complex (Central Iberia), finding systematic variations that correlate well with the metamorphic grade. Chondrite-normalized REE patterns of garnets from amphibolite-grade metapelites have lower-than-chondrite levels from La to Sm, a very small or no Eu anomaly, and a steep rise in the abundance of heavy REE as the atomic number increases. Metapelitic garnets from the amphibolite-granulite transition have a marked Eu negative anomaly and are enriched in MREE such that Sm is 10-15 times chondrite and the pattern is almost flat from Dy to Yb-Lu. In garnets from granulite-grade metapelites, the intensity of the Eu anomaly and the relative concentration of Nd, Sm, Gd and Tb increase, with almost flat chondrite-normalized patterns from Sm to Lu. Garnets from mafic granulites are remarkably similar to those of metapelitic garnets equilibrated at the same pressure, except for the Eu anomaly. The apparent paradox of enhanced uptake of larger REE ions with increasing pressure is attributed to the 3M²⁺ 2REE³⁺+ vacancy substitution, which produces a net decrease in the dimensions of the unit-cell of garnet. Variations in REE patterns depend essentially on the pressure and have little dependence on either temperature, bulk-composition of garnet, or REE whole-rock composition, so they could represent a new approach for geobarometric studies. The best numerical parameter to express pressure-related variations of REE distribution in garnets is the Gd/Dy ratio which does not seem perceptibly affected by disequilibrium partitioning. The regression equation between GASP pressure and the average Gd/Dy^garnet is P = 3.6 + 5.6 Gd/Dy. This equation seems to be reliable for garnets: (1)equilibrated within a pressure range of 4-9 kbar, (2) coexisting with modal monazite; and (3) with unit-cell dimensions under 11.46 Å.