Trace Element Partition Coefficients Research Articles

Trace element partitioning between apatite and silicate melts

We present new experimental apatite/melt trace element partition coefficients for a large number of trace elements (Cs, Rb, Ba, La, Ce, Pr, Sm, Gd, Lu, Y, Sr, Zr, Hf, Nb, Ta, U, Pb, and Th). The experiments were conducted at pressures of 1.0 GPa and temperatures of 1250 °C. The rare earth elements (La, Ce, Pr, Sm, Gd, and Lu), Y, and Sr are compatible in apatite, whereas the larger lithophile elements (Cs, Rb, and Ba) are strongly incompatible. Other trace elements such as U, Th, and Pb have partition coefficients close to unity. In all experiments we found D Hf > D Zr, D Ta ≈ D Nb, and D Ba > D Rb > D Cs. The experiments reveal a strong influence of melt composition on REE partition coefficients. With increasing polymerisation of the melt, apatite/melt partition coefficients for the rare earth elements increase for about an order of magnitude. We also present some results in fluorine-rich and water-rich systems, respectively, but no significant influence of either H 2O or F on the partitioning was found. Furthermore, we also present experimentally determined partition coefficients in close-to natural compositions which should be directly applicable to magmatic processes.

The partitioning of trace elements between ilmenite, ulvospinel, armalcolite and silicate melts with implications for the early differentiation of the moon

Mineral/melt trace element partition coefficients for a large number of trace elements (Zr, Hf, Ta, Nb, V, Co, Cu, Zn, Sr, REE, Sb, Sn, Mo, Cr, W, U, Th) were determined experimentally for ilmenite (FeTiO 3), ulvöspinel (Fe 2TiO 4), minerals of the armalcolite solid solution (Fe,Mg)Ti 2O 5). The measured partition coefficients indicate that the high field strength elements (Zr, Hf, Nb, and Ta) and some transition metals (V, Cr, Co, Sn) are moderately compatible in iron–titanium oxides, whereas other elements (REE, Y, Sr, U, Th, Mo, Sb, and W) are strongly incompatible. The partition coefficients for U are always significantly higher than the Th partition coefficients which makes Fe–Ti oxides efficient in fractionating U from Th. Our data also indicate that ilmenite, and other iron–titanium oxides, are able to fractionate W from Hf, the latter of which is much more compatible. This makes the iron–titanium oxides very effective agents to influence Hf–W isotopic ratios during the evolution of the moon. In addition, we also present some trace element partition coefficients for iron and titanium-bearing orthopyroxene.

A new method for the study of trace element partitioning between calcium carbonate and aqueous solution: A test case for Sr and Ba incorporation into calcite

A new experimental method (evaporation method) for calcium carbonate precipitation in aqueous solution was attempted in order to develop a convenient and controllable experimental technique for obtaining precise trace element partition coefficients. Calcite crystals were formed by evaporation of H2O from the aqueous mother solution using a dehumidifier, and the consumed Ca ions were supplied from a refill solution to maintain supersaturation. This method has been tested for Sr and Ba partition coefficients using NaCl (3.5%) solution under room temperature and pressure. The results of two runs give Sr partition coefficients of 0.0585 and 0.0546, and Ba partition coefficients of 0.0302 and 0.0275 for growth rates of 152 and 499 μmol · m-2 · hr-1, respectively. It was found that this method has several merits for partitioning experiments: (1) pH of the solution is close to the natural ones, for example, pH of 8.08 in the present case is close to that of seawater; (2) chemical composition of the mother solution can be chosen to imitate natural composition; (3) the crystal growth rate is controllable in a relatively small range and is measurable; and (4) temperature and PCO2 pressure can be controlled. The relative errors of the partition coefficients in the present study were estimated to be about 5%. The results seem to be promising, and suggest that more precise determination of partition coefficients could be possible via more careful experimental handling and improving the apparatus.

Determination of trace element partition coefficients between water and minerals by high‐pressure and high‐temperature experiments: Leaching technique

A refined approach for the experimental determination of fluid composition at elevated pressure and temperature conditions is developed. It effectively determines partition coefficients of trace elements between aqueous fluid and minerals. This approach is developed on top of many previously known technical achievements. However, it differs from the previous ones by the use of calibrated base and acid leaching of quenched samples to strip off the precipitates formed from the fluid during the quench. This permits the reconstruction of a fluid composition from an experimental charge that consists of a complete mixture of solid and fluid phases. This is a significant difference from the previous analogous studies, in which the fluid was separated from the solids all through the experiment. Mass balance calculations and error propagation assess the merits of the present method applied to a fluid‐antigorite system.

The effect of temperature on the equilibrium distribution of trace elements between clinopyroxene, orthopyroxene, olivine and spinel in upper mantle peridotite

The abundance of 30 trace elements has been determined in the minerals of 16 well-equilibrated spinel lherzolite xenoliths by laser-ablation inductively-coupled-plasma mass-spectrometry (LA-ICP-MS). Major elements were analysed by electron microprobe. The xenoliths span a range of equilibration temperatures from 1150 to 1500 K (calculated at an assumed pressure of 1.5 GPa from two-pyroxene geothermometry), allowing the trace-element partitioning relationships among the phases (olivine, orthopyroxene, clinopyroxene, spinel, and in some lower temperature xenoliths, amphibole) to be quantified as a function of temperature. Most elements show smooth partitioning trends among all phases that depend primarily on temperature but with some influences from bulk composition, particularly the amount of Na in clinopyroxene. Although most incompatible trace elements are concentrated into clinopyroxene, the effect of increasing temperature is to redistribute these elements into orthopyroxene and even olivine, such that these latter phases hold non-negligible proportions of many trace elements at the temperature at which peridotite would be in equilibrium with basaltic melts. The inter-crystalline trace-element partition coefficients reported in this study can be used to reconstruct the trace-element abundances in clinopyroxene at melting temperatures, and should also prove useful in elucidating the histories of more complex mantle peridotites with unequilibrated mineral compositions and textures.

Partitioning of trace elements between rutile and silicate melts: Implications for subduction zones

Mineral/melt trace element partition coefficients were determined for rutile (TiO 2) for a large number of trace elements (Zr, Hf, Nb, Ta, V, Co, Cu, Zn, Sr, REE, Cr, Sb, W, U, Th). Whilst the high field strength elements (Zr, Hf, Nb, Ta) are compatible in rutile, other studied trace elements are incompatible (Sr, Th, REE). In all experiments we found D Ta > D Nb, D Hf > D Zr and D U > D Th. Partition coefficients for some polyvalent elements (Sb, W, and Co) were sensitive to oxygen fugacity. Melt composition exerts a strong influence on HFSE partition coefficients. With increasing polymerization of the melt, rutile/melt partition coefficients for the high field strength elements Zr, Hf, Nb and Ta increase about an order of magnitude. However, D Nb/D Ta and D Hf/D Zr are not significantly affected by melt composition. Because D U ≫ D Th, partial melting of rutile-bearing eclogite in subducted lithosphere may cause excesses of 230Th over 238U in some island arc lavas, whereas dehydration of subducted lithosphere may cause excesses of 238U over 230Th. From our partitioning results we infer partition coefficients for protactinium (Pa) which we predict to be much lower than previously anticipated. Contrary to previous studies, our data imply that rutile should not significantly influence observed 231Pa- 235U disequilibria in certain volcanic rocks.

Rutile stability and rutile/melt HFSE partitioning during partial melting of hydrous basalt: Implications for TTG genesis

Synthesis experiments were conducted on a natural basalt (with 2 or 5 wt.% H 2O added) at 1.0–2.5 GPa and 900–1100 °C to investigate the stability field of rutile and rutile/liquid HFSE partitioning during partial melting of hydrous basalt. The basalt chosen has TiO 2 content close to average N-MORB. 100 ppm of Ta, Nb, Hf, Zr, etc., were added to the starting composition in order to improve analytical precision with the LAM-ICP-MS and the electron microprobe. Rutile occurs in the partial melting field of hydrated basalt at pressures higher than approximate 1.5 GPa, depending on H 2O content and bulk composition (especially TiO 2 and K 2O). Its stability increases with increasing pressure and decreasing temperature. H 2O helps produce a more mafic melt and so results in dissolution of rutile and shrinkage of the P– T field of rutile crystallization. The rutile/melt partitioning results confirm previous observations [ Green, T.H., Pearson, N.J., 1986. Ti-rich accessory phase saturation in hydrous mafic–felsic compositions at high P, T. Chem. Geol. 54, 185–201; Jenner, G.A., Foley, S.F., Jackson, S.E., Green, T.H., Fryer, B.J., Longerich, H.P., 1993. Determination of partition coefficients for trace elements in high pressure–temperature experimental run products by laser ablation microprobe-inductively coupled plasma-mass spectrometry (LAM–ICP–MS). Geochim. Cosmochim. Acta 57, 5099–5103; Foley, S.F., Barth, M.G., and Jenner, G.A., 2000. Rutile/melt partition coefficients for trace elements and assessment of the influence of rutile on the trace element characteristics of subduction zone magmas. Geochim. Cosmochim. Acta 64, 933–938; Schmidt, M.W., Dardon, A., Chazot, G., and Vannucci, R., 2004. The dependence of Nb and Ta rutile–melt partitioning on melt composition and Nb/Ta fractionation during subduction processes. Earth Planet. Sci. Lett. 226, 415–432 ], including that rutile is a dominant carrier for Nb and Ta, and that rutile favours Ta over Nb with D Nb always lower than D Ta for each rutile/melt pair. In addition our experiments demonstrate that both D Nb and D Ta decrease with increasing H 2O content but increase with decreasing temperature. Rutile is a necessary residual phase during the generation of Archean tonalite– trondhjemite–granodiorite (TTG) magmas to account for the negative Nb–Ta anomaly of the magmas. The depth for TTG production via melting of subducted oceanic crust must be more than 45–50 km based on the approximate 1.5 GPa minimum pressure for rutile appearance. Rutile fractionates Nb from Ta and will result in slightly higher Nb/Ta in coexisting liquids. Archean TTG magmas with subchondritic Nb/Ta must, therefore, have been derived from low Nb/Ta source regions [cf. Rapp, R.P., Shimizu, N., Norman, M.D., 2003. Growth of early continental crust by partial melting of eclogite. Nature 425, 605–609] unless alternative magmatic processes have lowered the Nb/Ta ratio. Also rutile-bearing residues should display lower Nb/Ta after TTG liquids are extracted. Hence, the present data do not support the view that subducted rutile-bearing eclogitic oceanic crust is a superchondritic Nb/Ta reservoir on Earth.

Compositional effects on element partitioning between Mg-silicate perovskite and silicate melts

High-pressure melting experiments were performed at ~26 GPa and ~2,200–2,400°C on synthetic peridotite compositions with varying FeO and Al2O3 contents and on a synthetic CI chondrite analogue composition. Peridotite liquids show a crystallisation sequence of ferropericlase (Fp) followed down temperature by Mg-silicate perovskite (MgPv) + Fp, which contrasts a sequence of MgPv followed by MgPv + Fp observed in the chondritic composition. The difference in crystallisation sequence is a consequence of the different bulk Mg/Si ratios. MgPv/melt partition coefficients for major, minor and trace elements were determined by electron microprobe and secondary ion mass spectrometry. Partition coefficients of tri- and tetravalent elements increase with increasing Al concentration in MgPv. A lattice strain model indicates that Al3+ substitutes predominantly onto the Si-site in MgPv, whereas most elements substitute onto the Mg-site, which is consistent with a charge-compensating coupled substitution mechanism. MgPv/melt partition coefficients for Mg (DMg) and Si (DSi) are related to the melt Mg/Si ratio such that DSi becomes lower than DMg at low Mg/Si melt ratios. We use a crystal fractionation model, based on upper mantle refractory lithophile element ratios, to constrain the amount of MgPv and Ca-silicate perovskite (CaPv) that could have fractionated during a Hadean magma ocean event and could still be present as a chemically distinct heterogeneity in the lower mantle today. We show that a fractionated crystal pile composed of 96% MgPv and 4% CaPv could comprise up to 13 wt% of the entire mantle.

Trace element geochemistry of Amba Dongar carbonatite complex, India: Evidence for fractional crystallization and silicate-carbonate melt immiscibility

Carbonatites are believed to have crystallized either from mantle-derived primary carbonate magmas or from secondary melts derived from carbonated silicate magmas through liquid immiscibility or from residual melts of fractional crystallization of silicate magmas. Although the observed coexistence of carbonatites and alkaline silicate rocks in most complexes, their coeval emplacement in many, and overlapping initial87Sr/86Sr and143Nd/144Nd ratios are supportive of their cogenesis; there have been few efforts to devise a quantitative method to identify the magmatic processes. In the present study we have made an attempt to accomplish this by modeling the trace element contents of carbonatites and coeval alkaline silicate rocks of Amba Dongar complex, India. Trace element data suggest that the carbonatites and alkaline silicate rocks of this complex are products of fractional crystallization of two separate parental melts. Using the available silicate melt-carbonate melt partition coefficients for various trace elements, and the observed data from carbonatites, we have tried to simulate trace element distribution pattern for the parental silicate melt. The results of the modeling not only support the hypothesis of silicate-carbonate melt immiscibility for the evolution of Amba Dongar but also establish a procedure to test the above hypothesis in such complexes.

Element partitioning between minerals and melt, melt composition, and melt structure

Mineral/melt trace element partition coefficients, D i mineral/melt (=concentration ratio, mineral/melt), can vary by up to two orders of magnitude in the melt composition range of natural magmatic liquids. As most geochemically interesting minor and trace elements are network-modifiers in silicate melts, one would expect, as observed, positive correlation between mineral/melt partition coefficients, D i mineral/melt, and the degree of polymerization of the silicate melt, NBO/T. Linear relationships between D i mineral/melt and NBO/T sometimes exist over restricted NBO/T-ranges where the type of Q n -species in the melt does not change and their abundance does not vary greatly. Other experimental D i mineral/melt vs. NBO/T data suggest log-linear or log–log relations in some cases, whereas in other studies, there are no simple relationships between D i mineral/melt and NBO/T of the melt. When relating mineral/melt partition coefficients to degree of melt polymerization, NBO/T, it is assumed that a principal melt structural control on D i mineral/melt is the activity of nonbridging oxygen in the melt, a NBO. The a NBO is related to NBO/T. The NBO/T is not, however, a quantitative measure of a NBO because the nonbridging oxygens in coexisting Q n -species in melts are energetically non-equivalent. This is evident in relationships between activity coefficient ratios of melt network-modifying cations, γ i / γ j , where the ionization potentials of i and j differ, and NBO/T of the melt. Such relationships are parabolic with minima or maxima at NBO/T near 1 and resemble the relations of abundance ratio, X Q3/ X Q2, vs. NBO/T of the melt. Interestingly, this NBO/T-value is near that of natural basalt melt. These observations suggest that i-NBO(Q 3) and j-NBO(Q 2) bond characteristics differ from one another, govern trace element solution behavior in silicate melts and, therefore, control the effect of melt composition on mineral/melt partitioning in silicate systems.

Partitioning of Sr, Ba, Rb, Y, and LREE between alkali feldspar and peraluminous silicic magma

Samples from well-known rhyolitic systems in the western USA were analyzed to determine trace-element partition coefficients between sanidine and coexisting melt. Both D Sr and D Eu correlate positively with An concentration in sanidine. Barium, not surprisingly, shows more correlation with Or in sanidine. All three show negative correlations with whole-rock Al 2 O 3 . Partition coefficients for REE show a decreasing trend with strong positive DEu. DLREE correlate negatively with the whole-rock Al 2 O 3 . In peraluminous rocks, Al 2 O 3 in magma seems to exert a major influence on partition coefficients of trace elements. With increasing Al 2 O 3 concentration, the melt composition becomes increasingly similar to that of feldspar. Incompatible trace elements (Rb, Y, Nb, and Zr) in sanidine show positive trends whereas compatible trace elements (Sr, Eu, and Ba) show negative correlations when plotted against whole-rock Al 2 O 3 . LREEs, which are expected to substitute for Ca in the sanidine structure, behave like Sr and Eu. Because trace-element partitioning is influenced by both crystal and magma compositions, multiple regression analysis has been used to predict trace-element partition coefficients. The equations will assist workers in the selection of partition coefficients used in petrogenetic models of silicic systems.

Experimental determination of trace element partitioning between garnet and silica‐rich liquid during anhydrous partial melting of MORB‐like eclogite

We present new experimentally determined trace element partition coefficients for nine garnet/melt and two clinopyroxene/melt pairs at 2.9–3.1 GPa and 1325°–1390°C, applicable to anhydrous partial melting of MORB‐like eclogite in the upper mantle. Phase compositions are similar to those documented in partial melting experiments of eclogite at these conditions: garnets with 16–25% grossular component and 0.4–2.0 wt % TiO2 coexist with siliceous partial melts having 52.4–57.1 wt % SiO2 and 1.1–6.7 wt % TiO2. Observed garnet/melt partitioning depends on TiO2 concentrations in garnet. Among the high‐TiO2 garnets (1.4–2.0 wt % TiO2), partition coefficients (Dgt/melt) increase with increasing grossular content of garnet (16–24%), but for the low‐TiO2 garnets (0.4–0.6 wt % TiO2) there is no difference in partitioning behavior at 19–24% grossular. In general, partition coefficients for the low‐TiO2 garnets tend to be higher than for the high‐TiO2 garnets. DZr and DHf increase with higher grossular content, as does DZr/DHf, but DZr and DHf always remain smaller than unity, and TiO2 in garnet appears to have little effect on DZr/DHf. DTh and DU have average values of 0.0012 and 0.0111 for the high‐TiO2 garnets and 0.0074 and 0.0376 for the low‐TiO2 garnets (DU/DTh of 9.1 and 5.1, respectively), affirming that partial melting of eclogite in the upper mantle can produce liquids with significant (230Th)/(238U) excess. D values of highly charged cations (Nb, Ta, Th, U) are lower in high‐TiO2 garnets. This is likely because Ti4+ occupying the garnet Y site requires charge balance by 2+ cations in Y or 3+ cations in Z, thereby limiting the availability of these charge‐balancing substitutions to accommodate other highly charged cations. The overall range of our new garnet/melt partition coefficients, regardless of Ti and grossular content, is similar to literature garnet/melt partitioning data applicable to garnet peridotite melting, and therefore bulk partition coefficients for eclogite and peridotite depend largely on mineral modes and choice of partitioning data for clinopyroxene. Using our new data and previously published data from the literature, we calculate bulk partition coefficients appropriate for two MORB‐like eclogites (82% clinopyroxene + 18% garnet and 75% clinopyroxene + 25% garnet) and a garnet peridotite (60% olivine + 17% orthopyroxene + 13% clinopyroxene + 10% garnet). Bulk partition coefficients are generally much higher for MORB‐like eclogite when compared to garnet peridotite, with the exception of Th and U. However, the ratios of bulk partition coefficients of eclogite and garnet peridotite are similar for pairs such as DSm/DYb, DZr/DHf and DU/DTh, indicating that both lithologies can induce similar trace element fractionations during partial melting in the presence of garnet.

Trace element partitioning in lamproitic magmas—the Gaussberg olivine leucitite

Partition coefficients for compatible and incompatible trace elements for use in modelling the origin and evolution of lamproitic and other potassic rocks are reported. These are based on values obtained by in situ determination using laser ablation microprobe-inductively coupled plasma-mass spectrometry (LAM-ICP-MS) on mineral/glass pairs in examples of the Gaussberg olivine leucitite, a classic leucite lamproite. Leucite/melt partition coefficients for 32 elements show values above 1 only for Cs and Rb, whereas values for Ni, Ga and Co are 0.4–0.1. D Lc/Lq for most other elements are in the range 0.07–0.0005, and delineate a gently sloping pattern to lower values for elements generally considered to be “incompatible”. Olivine/melt partition coefficients show higher values for the compatible elements Ni, Co, Mn and Cr than in all basaltic systems investigated experimentally to date. D Ol/Lq for heavy rare earth elements (HREE) are slightly lower than in basalts, whereas most other elements give similar values. Clinopyroxene/melt partition coefficients appear to vary widely as a function of the Al 2O 3 content of the rock. Values obtained here for the Gaussberg rock consistently show appreciably higher D Cpx/Lq than in basaltic systems for the incompatible elements Cs, Rb, Ba, Th, U, Nb, Ta, La and Ce, which is attributed to a larger M2 site. Early salitic cores have patterns with lower D-values for these elements and higher values for HREE, similar to those known from alumina-rich potassic rocks. Thus, the values obtained here should be applied only with caution to nonlamproitic rocks. Other notable characteristics are high D U/ D Th and low D Zr and D Hf, the latter of which is attributed to structural characteristics of peralkaline melts.

Trace element partitioning between mantle wedge peridotite and hydrous MgO-rich melt

Magmas erupted at convergent margins consist of components from both mantle wedge and subducted slab. In an effort to quantify the relative contributions of these sources, we have determined 19 trace element partition coefficients (D values) for orthopyroxene and clinopyroxene, in equilibrium with spinel and hydrous high-MgO melt under conditions appropriate to melting in the mantle wedge, i.e., 1.3 GPa, 1245 °C, and f O2 of NNO + 1. All trace elements are more incompatible in clinopyroxene during hydrous melting than during anhydrous melting of fertile and depleted peridotite. Orthopyroxene D values are relatively insensitive to pressure, temperature, and phase composition. The new D values are used to calculate the trace-element composition of the mantle wedge, which produced primitive South Sandwich Islands and St. Vincent (Lesser Antilles) arc basalts. Both sources correspond to previously depleted mantle that has been enriched in LILEs and LREEs by slab-derived fluids. In the case of the South Sandwich Islands, the calculated source is in very close agreement with dredged fore-arc lherzolites. Our partitioning data confirm that hydrous melting of wedge peridotite itself cannot produce the characteristic enrichments of LILEs over REEs and HFSEs. Our estimates of the slab component in South Sandwich Islands and St. Vincent are consistent with estimates from other arcs, derived by alternative methods.

Partitioning of Sr, Ba, Rb, Y, and LREE between plagioclase and peraluminous silicic magma

Trace-element partition coefficients between plagioclase and coexisting glass/matrix have been determined for twenty-nine rhyodacite-rhyolite samples from nine volcanic centers. Strontium partition coefficients form two clear positive trends when plotted against the An content of plagioclase. One trend has a steep slope with D Sr between 5.6 and 15.8 (T1), whereas the other trend has a gentle slope with D Sr between 1.2 and 7.6 (T2). Barium partition coefficients show similar patterns: a steeply sloped trend is formed by samples with D Ba between 1.6 and 8.8, and a gently sloped trend by samples with D Ba <1. D Sr and D Ba in plagioclase correlate positively, and both D Sr and D Ba correlate positively with temperature. Samples with large D Sr and D Ba are all from rocks with <1 wt% CaO. These rocks are peraluminous and also have low total Sr and Ba. Partition coefficients for Sr and Ba are influenced by whole-rock CaO in high-Al rhyolite systems. Where these systems have low CaO, the opportunity increases for other divalent cations to enter the plagioclase structure. Strontium and Ba will then enter the M position of plagioclase to balance the charge deficiency caused by the substitution of Al for Si. Because of the potent influence of whole-rock composition on Sr and Ba partitioning for low CaO rocks, values of partition coefficients in petrogenetic modeling needs to be selected carefully.

Trace-element partitioning at conditions far from equilibrium: Ba and Cs distributions between alkali feldspar and undercooled hydrous granitic liquid at 200 MPa

This study examines the effects of increasing supersaturation, attained by single-step liquidus undercooling (ΔT), on the partitioning of barium and cesium between potassic alkali feldspar (Afs) and hydrous granitic liquid at 200 MPa. The investigation is motivated by trace-element distribution patterns in granitic pegmatites which cannot be simulated by fractionation models using "equilibrium" partition coefficients, and thus its purpose is to assess if, how, and why partition coefficients for compatible and incompatible trace elements may vary when crystal growth commences far from the crystal–melt equilibrium boundary. Barium expands the liquidus stability field of potassic feldspar to higher temperatures, such that liquidi for the Ba-rich (~0.5 wt% BaO) compositions used are ~100 °C higher than for Ba-absent analogues. At low degrees of undercooling (ΔT~50 °C), values of DBaAfs/m. (~10–20) fall within the range of previous investigations, as do values of DCsAfs/m. (≤0.10) from experiments at all temperatures. Progressively greater undercooling is manifested in the run products by increasingly skeletal to cuneiform crystal morphologies, increased compositional zonation of Afs, and the development of compositional boundary layers in glass. Whereas the partitioning behavior of Cs (incompatible) is not measurably affected, strong undercooling apparently causes the partitioning of Ba (highly compatible) to deviate from equilibrium behavior. Feldspars produced by strong undercooling (ΔT≥100 °C) are heterogeneous, such that DBaAfs/m. versus K/K+Na varies linearly between the average value at 850 °C and the equilibrium value appropriate to the temperature of growth. Hence, high supersaturation accompanying undercooling produces feldspar compositions by isothermal growth which record a vestige of the liquid line of descent (i.e., an ontogeny within zoned crystals which approximately tracks the feldspar liquidus from high temperature to the final low temperature of actual crystal growth). Such zoning patterns may mimic normal patterns produced by fractionation with decreasing temperature under near-equilibrium (near-liquidus) conditions. Increasing fluorine contents tend to exacerbate the effects of undercooling by inhibiting feldspar nucleation, causing both K/K+Na ratios and compatible trace-element partitioning behavior in feldspar to deviate from equilibrium values. This effect continues until nucleation lag is overcome, whereupon a period of rapid growth from supersaturated liquid follows which is similar to that in F-absent compositions. Electronic Supplementary Material to this paper can be obtained by using the Springer LINK server located at http://dx.doi.org/10.1008/s00410-002-425-y.

Trace-element partitioning between alkali feldspar and peralkalic quartz trachyte to rhyolite magma. Part II: Empirical equations for calculating trace-element partition coefficients of large-ion lithophile, high field-strength, and rare-earth elements

Multivariate linear regression (MLR) analysis has been used to develop equations for predicting alkali feldspar/melt trace-element partition coefficients ( D -values) for certain large-ion lithophile elements (Rb and Eu), high field-strength elements (Zr and Nb), and rare-earth elements (Y, La, Ce, Pr, Nd, Sm, Eu, Gd, Dy, Er, and Yb) in peralkalic felsic systems. A partitioning equation for Sr based solely on data from Pantelleria is also presented. D Rb, D Zr, and D Nb each can be calculated given whole-rock SiO2 and agpaitic index [A.I. = mol (Na + K) /Al]; alternatively, D Rb can be calculated given feldspar Or content and agpaitic index. D Eu can be calculated given whole-rock SiO2. agpaitic index, and whole-rock Na2O/K2O. Other D REE can be calculated given whole-rock SiO2, agpaitic index, and whole-rock CaO. These D -values all increase with increasing whole-rock SiO2 and decreasing agpaitic index. D Sr for samples from Pantelleria increase with increasing feldspar Ab and can be calculated given feldspar Ab content. The equations formulated will help to constrain better the choice of D -values used in petrogenetic models of peralkalic systems.

An experimental test of Henry's Law in solid metal‐liquid metal systems with implications for iron meteorites

Abstract— Experimental solid metal‐liquid metal partition coefficients have been used to model the crystallization of magmatic iron meteorites and understand the evolution of asteroid cores. However, the majority of the partitioning experiments have been conducted with trace elements doped at levels that are orders of magnitude higher than measured in iron meteorites. Concern about Henry's Law and the unnatural doping levels have been cited as one reason that two recent iron meteorite studies have dismissed the experimental partition coefficients in their modeling. Using laser ablation ICP‐MS analysis, this study reports experimentally determined solid metal‐liquid metal trace element partition coefficients from runs doped down to the levels occurring in iron meteorites. The analyses for 12 trace elements (As, Co, Cr, Cu, Ga, Ge, Ir, Os, Pd, Pt, Re, and W) show no deviations from Henry's Law, and these results support decades of experimental work in which the partition coefficients were assumed to be independent of trace element concentration. Further, since our experiments are doped with natural levels of trace elements, the partitioning results are directly applicable to iron meteorites and should be used when modeling their crystallization. In contrast, our new Ag data are inconsistent with previous studies, suggesting the high Ag‐content in previous studies may have influenced the measured Ag partitioning behavior.

Trace-element partitioning between alkali feldspar and peralkalic quartz trachyte to rhyolite magma. Part I: Systematics of trace-element partitioning

New alkali feldspar/felsic magma trace-element partition coefficients (D-values) for Rb, Sr, Ba, Eu, Y, Zr, Nb, Ga, Zn, trivalent REE, Be, Cs, Hf, Pb, Th, and U for 30 samples of peralkalic quartz trachyte and rhyolite are presented. D-values of incompatible elements vary systematically with melt polymerization parameters, increasing with whole-rock silica, but decreasing in rocks with higher agpaitic indices [A.I. = mol (Na + K)/Al]. D-values for Sr and Ba (evaluated to be accurate) vary systematically with crystal chemistry, probably substituting for Na in the Ca-poor alkali feldspar phases. Apparent D-values for Sr and Ba from pre-Quaternary systems are fraught with contamination and analytical errors, respectively, and should be used with caution. D Eu decreases exponentially with A.I., ranging from compatible in weakly peralkalic (A.I. &lt; 1.1) rocks to strongly incompatible in very peralkalic (A.I. &gt; 1.2) rocks. These regular variations strongly suggest that partition coefficients for these elements may be predicted accurately if whole-rock and crystalchemical parameters are known.

The effect of H 2O on crystal-melt partitioning of trace elements

Many experimental data demonstrate distinct differences between crystal-liquid partition coefficients D i measured under high temperature, anhydrous conditions and those determined at lower temperatures in the presence of H 2O. In this study we develop a thermodynamic method of separating the effects of H 2O from those of temperature. We then apply the method to predict partitioning of REE between clinopyroxene, garnet and silicate melt over wide ranges of temperature, pressure and H 2O content. Our initial inputs are the model of Wood and Blundy (1997) for REE partitioning between clinopyroxene and anhydrous melt and the melting temperatures of diopside on the join CaMgSi 2O 6-H 2O (Eggler and Rosenhauer, 1978). We then make the hypothesis that the effect of H 2O on the activity of REE clinopyroxene component (REEMgAlSiO 6) in the melt is the same as the measured effect on CaMgSi 2O 6 component. This leads to predictions of REE partition coefficients for clinopyroxene coexisting with hydrous melt at any P,T and H 2O content up to 45 weight %. The results agree with observed REE partition coefficients with a standard deviation which is the same as that for the anhydrous data. We conclude, therefore, that the ‘H 2O-effect’ may, in this case, be accurately predicted. We extend the approach to garnet by using the join Mg 3Al 2Si 3O 12-H 2O to estimate the effects of H 2O on all ‘garnet-like’ components in the melt. This enables calculation of garnet-melt REE partition coefficients for melts containing up to 25% H 2O. The observation that H 2O influences major and trace component activities in a similar manner enables us to make some generalisations about the combined effects on partitioning of decreasing temperature and increasing water content of the melt. The relative enthalpies of fusion ΔH f of major and trace components dictate whether trace element partition coefficients increase or decrease with H 2O addition: ΔH f trace > ΔH f major gives increasing D trace with addition of H 2O and ΔH f trace < ΔH f major leads to decreasing D trace with addition of H 2O. Note that D trace is strictly the ratio of mole fractions of charge-balanced components such as REEMgAlSiO 6 in solid to melt phases. In the cases considered here, however, D trace closely approximates D = [trace] solid [trace] melt where [trace] refers to weight concentration. For clinopyroxene ΔH f REE < ΔH f major and H 2O addition leads to substantial decreases in REE partition coefficients. For garnet, ΔH f REE ≈ ΔH f major and addition of H 2O results in little change in D REE. The consequences are illustrated with respect to Yb contents of melts of spinel lherzolite. For partial melting under anhydrous conditions, Yb is relatively compatible in clinopyroxene and the Yb contents of the product melts remain close to 5 times source values at all F (fraction of equilibrium batch melting) between 0 and 0.25. Hydrous peridotite produces melts, because of low D Yb, which are strongly enriched in Yb (up to at least 10 times source) and whose Yb concentrations depend strongly on F. This means that the low Yb concentrations observed in many island arc tholeiites cannot be due to the addition of water to fertile peridotite.