Trace Element Partition Coefficients Research Articles

Amphibole–melt trace element partitioning of fractionating calc-alkaline magmas in the lower crust: an experimental study

Amphibole is one of the most important hydrous minerals of the middle and lower continental crust and plays a key role in the formation of intermediate to silica-rich magmas. This study reports a consistent set of amphibole trace element partition coefficients derived from fractional crystallization experiments at 0.7 GPa in a piston cylinder apparatus. Starting materials were doped with trace elements on the 20–40 ppm level and measured using laser ablation (LA)-ICP-MS. Amphibole is stable from 1010 to 730 °C and systematically changes its composition from pargasite to magnesiohornblende to cummingtonite, while coexisting liquids vary from andesite to dacite and rhyolite. Amphibole–liquid partition coefficients increase systematically with decreasing temperature and increasing SiO2 in the liquid. Potassium displays an inverse behavior and partitioning decreases with decreasing temperature. Rare earth element (REE) partition coefficients, assumed to occupy the M4 site within the amphibole structure, increase continuously up to one order of magnitude. The calculated lattice parameters, ideal cation radius (r 0) and Young’s modulus (E) remain nearly constant with decreasing temperature. The high-field strength elements Zr and Hf that occupy the M2 site of the amphibole structure reveal a fivefold increase in partition coefficients with decreasing temperature and constant lattice parameters r 0 and E. Partition coefficients correlate with edenite, tschermaks and cummingtonite exchange vectors indicating that the maximum partition coefficient (D 0) for an ideal cation radius increases with decreasing edenite component, while the latter decreases linearly with temperature. Regressing Amph/L D Ca against trace elements results in fair to excellent correlations (r 2 0.55–0.99) providing a predictive tool to implement the trace element partition coefficients in numerical geochemical modeling. Our data result in positive correlations between Amph/L D Nb/Ta and Amph/L D Ca, and Mg#. Spoon-shaped REE patterns and subchondritic Nb/Ta ratios in tonalitic to granodioritic plutonic rocks and andesitic to rhyolitic magmas directly constrained by measured trace element compositions of coexisting liquids are consistent with hornblende gabbro fractionation in the middle to lower crust. The systematic change of the measured trace element composition of fractionating calc-alkaline liquids indicates that hornblende gabbro formation in the middle to lower crust also exerts an important control on some commonly used trace element ratios such as Sr/Y, Sr/Ba or Nb/Zr.

Petrology, genesis and geodynamic implication of the Mesoproterozoic–Late Cretaceous Timmasamudram kimberlite cluster, Wajrakarur field, Eastern Dharwar Craton, southern India

New mineralogical and bulk-rock geochemical data for the recently recognised Mesoproterozoic (ca. 1100 Ma) and late Cretaceous (ca. 90 Ma) kimberlites in the Timmasamudram cluster (TKC) of the Wajrakarur kimberlite field (WKF), Eastern Dharwar Craton, southern India, are presented. On the basis of groundmass mineral chemistry (phlogopite, spinel, perovskite and clinopyroxene), bulk-rock chemistry (SiO2, K2O, low TiO2, Ba/Nb and La/Sm), and perovskite Nd isotopic compositions, the TK-1 (macrocrystic variety) and TK-4 (Macrocrystic variety) kimberlites in this cluster are here classified as orangeites (i.e. Group II kimberlites), with geochemical characteristics that are very similar to orangeites previously described from the Bastar Craton in central India, as well as the Kaapvaal Craton in South Africa. The remaining kimberlites (e.g., TK-2, TK-3 and the TK-1 microcrystic variant), are more similar to other 1100 Ma, Group I-type kimberlites of the Eastern Dharwar Craton, as well as the typical Group I kimberlites of the Kaapvaal Craton. Through the application of geochemical modelling, based on published carbonated peridotite/melt trace element partition coefficients, we show that the generation of the TKC kimberlites and the orangeites results from low degrees of partial melting of a metasomatised, carbonated peridotite.Depleted mantle (TDM) Nd perovskite model ages of the 1100 Ma Timmasamudram kimberlites show that the metasomatic enrichment of their source regions are broadly similar to that of the Mesoproterozoic kimberlites of the EDC. The younger, late Cretaceous (ca. 90 Ma) TK-1 (macrocrystic variant) and TK-4 kimberlites, as well as the orangeites from the Bastar Craton, share similar Nd model ages of 1100 Ma, consistent with a similarity in the timing of source enrichment during the amalgamation of Rodinia supercontinent. The presence of late Cretaceous diamondiferous orangeite activity, presumably related to the location of the Marion hotspot in southern India at the time, suggests that thick lithosphere was preserved, at least locally, up to the late Cretaceous, and was not entirely destroyed during the breakup of Gondwana, as inferred by some recent geophysical models.

Trace element partitioning between clinopyroxene and trachy-phonolitic melts: A case study from the Campanian Ignimbrite (Campi Flegrei, Italy)

The partitioning of trace elements between crystals and melts provides an important petrogenetic tool for understanding magmatic processes. We present trace element partition coefficients measured between clinopyroxene phenocrysts and trachy-phonolitic magmas at the Campi Flegrei (Italy), whose late Quaternary volcanism has been characterized by two major caldera-forming events (Campanian Ignimbrite at ~39ka, and Neapolitan Yellow Tuff at ~15ka). Our data indicate that the increase of trivalent rare earth elements and yttrium into the crystal lattice M2 site is facilitated by the charge-balancing substitution of Si4+ with Al3+ on the tetrahedral site. Higher concentrations of tetravalent and pentavalent high field strength elements on the M1 site are also measured when the average charge on this site is increased by the substitution of divalent cations by Alvi. In contrast, due to these charge balance requirements, divalent transitional elements become less compatible within the crystal lattice. On the basis of the lattice strain theory, we document that the incorporation of rare earth elements and yttrium in clinopyroxene is influenced by both compositional and physical parameters. Data from this study allow to update existing partitioning equations for rare earth elements in order to construct a self-consistent model for trachy-phonolitic magmas based on the lattice strain theory. The application of this model to natural products from the Campanian Ignimbrite, the largest caldera-forming eruption at the Campi Flegrei, reveals that the complex rare earth element pattern recorded by the eruptive products can be successfully described by the stepwise fractional crystallization of clinopyroxene and feldspar where the clinopyroxene–melt partition coefficient changes progressively as a function of the physicochemical conditions of the system.

Experimental determination of trace element partition coefficients between spinel and silicate melt: the influence of chemical composition and oxygen fugacity

We present new experimentally determined trace element partition coefficients between spinel and silicate melt. The experiments were performed at atmospheric pressure and at temperatures between 1220 and 1450 °C. To study the effect of redox conditions on trace element partitioning, we performed experiments under different redox conditions, with fO2 ranging from log −12 to log −0.7. The effect of different spinel compositions is also investigated. Our results show that spinel of all compositions readily incorporates the transition metals Ni, Co and Ga and the corresponding partition coefficients are >1. D Ni,Co,Ga are not significantly affected by changing melt composition, crystal composition or redox conditions. However, the multivalent trace elements V and Mo show a strong effect of redox conditions on their partitioning behavior with D V and D Mo highest at very reducing conditions and considerably lower at more oxidizing conditions. Partition coefficients for the high field strength elements Ti, Zr, Hf, Nb, and Ta and the elements Sc and Lu strongly depend on crystal composition, with D Ti and D Sc >1 for very Fe3+- or Cr-rich (and Al-poor) spinels, but one to two orders of magnitude lower in systems with Al-rich spinels. We present some examples on how our data may be used to reconstruct redox conditions of spinel formation. We also present some results on the partitioning of Pt and Rh between spinel and melt. D Rh depends strongly on redox conditions, while D Pt is not significantly affected.

Interface partition coefficients of trace elements in carbonate-silicate parental media for diamonds and paragenetic inclusions (experiments at 7.0-8.5 GPa)

Interface partition coefficients KD of a representative set of trace elements (TE) in the partly molten diamond-forming peridotite-eclogite- carbonatite system are experimentally determined at 7.0-8.5GPa. The experimental data evidence that trace-element partition does not depend on the melt composition, with heavy rare-earth elements (HREE) concentrating mainly in garnet. Model TE partition coefficients for the natural diamond-producing carbonatite melts of mantle chambers are calculated based on TE concentrations in minerals of peridotite and eclogite parageneses of diamond inclusions, on the one hand, and on the experimental KD coefficients, on the other. The results show that the TE of the parental media are mostly the mantle peridotite components, with the parental media being depleted in medium (Ba, La, Ce, Pr, Nd, Sm, Eu, Gd) and heavy (Tb, Dy, Ho, Er, Yb, Lu, Hf) rare-earth elements relative to the primitive peridotite. The elevated contents of Sr, Nb, and Ce in the completely miscible carbonate-silicate melts might be due to the participation of “metasomatic agents” in the formation of chambers of diamond-forming carbonatite magmas.

Effects of magma ocean crystallization and overturn on the development of 142Nd and 182W isotopic heterogeneities in the primordial mantle

One possible mechanism to explain the observed variability of the short-lived Sm146→Nd142 and Hf182→W182 systems recorded in some early Earth rocks is crystal–liquid fractionation and overturn in an early magma ocean. This process could also potentially explain the deviation between the 142Nd isotopic composition of the accessible Earth and the chondritic average. To examine these effects, the magma ocean solidification code of Elkins-Tanton (2008) and a modified Monte Carlo algorithm, designed to randomly choose physically reasonable trace element partition coefficients in crystallizing mantle phases, are used to model the isotopic evolution of early Earth reservoirs. This model, also constrained by the 143Nd composition of the accessible Earth, explores the effects of changing the amount of interstitial liquid trapped in cumulates, the half-life of 146Sm, the magnitude of late accretion, and the simplified model of post-overturn reservoir mixing. Regardless of the parameters used, our results indicate the generation of early mantle reservoirs with isotopic characteristics consistent with observed anomalies is a likely outcome of magma ocean crystallization and overturn of shallow, enriched, and dense (i.e., gravitationally unstable) cumulates. The high-iron composition and density of a hypothesized, early-formed enriched mantle reservoir is compatible with seismic observations indicating large, low-shear velocity provinces (LLSVPs) (e.g., Trampert et al., 2004) present in the mantle today. Later melts of an enriched reservoir are likely to have remained isolated deep within the mantle (e.g., Thomas et al., 2012), consistent with the possibility that the presently observed LLSVPs could be partially or fully composed of remnants of an early enriched reservoir.

CONTRASTING PARENTAL MAGMA COMPOSITIONS FOR THE HONGGE AND PANZHIHUA MAGMATIC Fe-Ti-V OXIDE DEPOSITS, EMEISHAN LARGE IGNEOUS PROVINCE, SW CHINA

Research Article| September 01, 2014 CONTRASTING PARENTAL MAGMA COMPOSITIONS FOR THE HONGGE AND PANZHIHUA MAGMATIC Fe-Ti-V OXIDE DEPOSITS, EMEISHAN LARGE IGNEOUS PROVINCE, SW CHINA* Zhong-Jie Bai; Zhong-Jie Bai 1State Key Laboratory of Ore Deposit Geochemistry, Institute of Geochemistry, Chinese Academy of Sciences, Guiyang 550002, China Search for other works by this author on: GSW Google Scholar Hong Zhong; Hong Zhong † 1State Key Laboratory of Ore Deposit Geochemistry, Institute of Geochemistry, Chinese Academy of Sciences, Guiyang 550002, China †Corresponding author: e-mail, zhonghong@vip.gyig.ac.cn Search for other works by this author on: GSW Google Scholar Chusi Li; Chusi Li 2Department of Geological Sciences, Indiana University, Bloomington, Indiana 47405 Search for other works by this author on: GSW Google Scholar Wei-Guang Zhu; Wei-Guang Zhu 1State Key Laboratory of Ore Deposit Geochemistry, Institute of Geochemistry, Chinese Academy of Sciences, Guiyang 550002, China Search for other works by this author on: GSW Google Scholar De-Feng He; De-Feng He 1State Key Laboratory of Ore Deposit Geochemistry, Institute of Geochemistry, Chinese Academy of Sciences, Guiyang 550002, China Search for other works by this author on: GSW Google Scholar Liang Qi Liang Qi 1State Key Laboratory of Ore Deposit Geochemistry, Institute of Geochemistry, Chinese Academy of Sciences, Guiyang 550002, China Search for other works by this author on: GSW Google Scholar Economic Geology (2014) 109 (6): 1763–1785. https://doi.org/10.2113/econgeo.109.6.1763 Article history received: 29 Jul 2013 accepted: 03 Jan 2014 first online: 09 Mar 2017 Cite View This Citation Add to Citation Manager Share Icon Share MailTo Twitter LinkedIn Tools Icon Tools Get Permissions Search Site Citation Zhong-Jie Bai, Hong Zhong, Chusi Li, Wei-Guang Zhu, De-Feng He, Liang Qi; CONTRASTING PARENTAL MAGMA COMPOSITIONS FOR THE HONGGE AND PANZHIHUA MAGMATIC Fe-Ti-V OXIDE DEPOSITS, EMEISHAN LARGE IGNEOUS PROVINCE, SW CHINA. Economic Geology 2014;; 109 (6): 1763–1785. doi: https://doi.org/10.2113/econgeo.109.6.1763 Download citation file: Ris (Zotero) Refmanager EasyBib Bookends Mendeley Papers EndNote RefWorks BibTex toolbar search Search Dropdown Menu toolbar search search input Search input auto suggest filter your search All ContentBy SocietyEconomic Geology Search Advanced Search Abstract Many ~260 Ma mafic-ultramafic layered intrusions, including Hongge and Panzhihua, in the Emeishan large igneous province, southwestern China host world-class Fe-Ti-V oxide ore deposits. These two most important ore-bearing intrusions show differences in lithology and mineral chemistry. The most important orebodies in the Hongge intrusion occur as concordant layers in the middle part of the intrusion, closely associated with clinopyroxenites. Titanomagnetite and Mg-rich ilmenite are the major ore minerals of the Hongge deposit. Coexisting clinopyroxene contains >1.7 wt % TiO2. These data indicate high Ti parental magma for the Hongge ore-bearing clinopyroxenites. In the Panzhihua intrusion, the most important orebodies also occur as concordant layers but in its lower part instead of middle part. In contrast with the Hongge deposit, the most important host rocks of the Panzhihua deposit are gabbros, not clinopyroxenites. In addition, ilmenite is rare and titanomagnetite is predominant in the Panzhihua deposit. Coexisting clinopyroxene in the Panzhihua deposit contains <1.6 wt % TiO2. The contrasting lithologic and mineral compositions indicate that the parental magma for the Hongge deposit has higher TiO2 than that for the Panzhihua deposit. The compositions of clinopyroxene from the Panzhihua and Hongge deposits indicate that their parental magmas are also different in MgO/FeO and trace element ratios as well. Calculations using average experimental Mg-Fe exchange coefficient and trace element partition coefficients between clinopyroxene and magma show that the Hongge parental magma has higher MgO/FeO ratios and more fractionated mantle-normalized trace element patterns than the Panzhihua parental magma. The estimated compositions of the parental magmas for the Hongge and Panzhihua ore-bearing lithologies resemble the average compositions of the Longzhoushan-type high Ti basalts and Ertan-type intermediate Ti basalts in the Emeishan large igneous province, respectively. The new trace element data from this study, together with available Sr-Nd isotope data for the ore-bearing intrusions from literature, support a new petrogenetic model involving selective assimilation of newly subducted, stagnant oceanic gabbroic slab above the deep-seated Emeishan mantle plume. This process and subsequent contamination with the upper crust played an important role in the variation of parental magma compositions between the Hongge and Panzhihua magmatic oxide ore deposits. Abundant Fe-Ti oxide ore deposits associated with less evolved basaltic magma in the Emeishan large igneous province than elsewhere in the world are attributed to selective assimilation of newly subducted, stagnant oceanic lithospheric slab by the ascending mantle plume-derived picritic magma that was originally undersaturated with Fe-Ti oxides. You do not have access to this content, please speak to your institutional administrator if you feel you should have access.

Element signatures of subduction-zone fluids. An experimental study of the element partitioning (Dfluid/rock) of natural partly altered igneous rocks from the ODP drilling site 1,256

The trace element signatures of fluids were investigated by leaching experiments on natural samples of partly altered mafic igneous rocks recovered from the drilling site 1,256 of ODP Leg 206 on the Cocos plate (Central America). Experiments with ultrapure water were performed at 400 °C/0.4 GPa and 500 °C/0.7 GPa. Both fluids and residual solids were examined to obtain the partition coefficients (Dfluid/rock) of various trace elements. Element partition coefficients (Dfluid/rock) obtained at 500 °C/0.7 GPa are significantly lower compared to results obtained at 400 °C/0.4 GPa, which is in contrast to observations at higher pressures (2.2–6 GPa) and temperatures between 700 and 1,400 °C (Kessel et al. in Earth Planet Sci Lett 237: 873–892, 2005a; Spandler et al. in Chem Geol 239: 228–249, 2007). This finding may indicate a considerable pressure effect on the leaching processes and strongly divergent fluid–rock interactions in the upper part of a subduction zone at 0.4–0.7 GPa compared to deeper subduction areas with higher pressures. Furthermore, this may be interpreted as one of the earliest fractionation processes during the subduction of crustal material.

Carbonated sediment–peridotite interaction and melting at 7.5–12GPa

Abstract To gain a better insight into deep subduction-zone processes, the interaction between model sediment and peridotite was experimentally studied using a multianvil apparatus at 7.5–12 GPa and 900–1400 °C. Two H 2 O- and CO 2 -bearing synthetic materials similar to GLOSS of Plank and Langmuir (1998) with different K/Na ratios were used. The peridotite (harzburgite) mixture consisted of natural olivine, orthopyroxene and garnet from kimberlite-derived xenoliths. Sediment (below) and peridotite (above) were packed into a Re-lined Pt capsule and exposed to constant pressure and temperature for 24–72 h. Several experiments were conducted with pure sediment mixtures. The experimental products were analyzed by EPMA and LA ICP MS. Pure sediment melting produces Ca–alkali carbonatite melts near the solidus and liquids richer in SiO 2 at high temperatures. The residual mineral assemblage is dominated by coesite/stishovite, garnet, kyanite and jadeite. Peridotite–sediment interaction results in (1) replacement of olivine in the metaperidotite zone by low-Ca pyroxene and magnesite, (2) formation of a transitional garnet-rich zone containing low-Ca pyroxene and coesite/stishovite, and (3) disappearance of kyanite and stabilization of carbonate in the metasediment layer. The peridotite zone is transformed into a low-Ca pyroxene–garnet aggregate, and its garnet/pyroxene ratio increases with increasing temperature. The sediment zone evolves with increasing temperature from coesite / stishovite + jadeite + garnet + carbonate + rutile to coesite / stishovite + jadeite + garnet. Mass transfer is dominated by Si flux from sediment to peridotite and Mg and Fe from peridotite to sediment. The composition of the melts produced in the interaction experiments differs from that from pure sediment melting in higher Mg and Si contents and much lower Ca. Trace element partition coefficients (D) between residual minerals (pyroxenes, garnet, carbonate, zircon, rutile and SiO 2 polymorphs) and melts were determined in several experiments. Modeling using the obtained D values showed that interaction with peridotite enhances REE fractionation in sediment-derived melt owing to the replacement of kyanite by garnet. The retention of jadeitic pyroxene in the transitional and metasediment zones depresses Na mobility and leads to the formation of high-K/Na liquids. The increasing stability of carbonate in the metasediment may be favorable for deep CO 2 recycling and decoupling of H 2 O and CO 2 . It was shown that the garnet–clinopyroxene residua produced by sediment–peridotite interaction and melting are similar to high-MgO diamond-bearing eclogites reported from several kimberlite pipes worldwide.

Trace element mineral/melt partitioning for basaltic and basaltic andesitic melts: An experimental and laser ICP-MS study with application to the oxidation state of mantle source regions

Understanding magmatic processes such as crystallization and melting recorded in natural samples requires calibration of mineral–melt trace element partition coefficients (D) and their dependence on temperature, pressure, oxygen fugacity (fO2) and chemical composition. However, few experimental studies have focused on measuring trace element partition coefficients for a large number of trace elements, in the various minerals present in basaltic rocks, and under diverse conditions, particularly of variable fO2. Twenty-seven 0.1 MPa experiments provide partition coefficients for major elements and Sc, Ti, V, Mn, Co, Ni, Ga, Sr, Y, Nb, Ba, Ce, Nd, Eu, Gd, and Yb for the mineral phases olivine, plagioclase, orthopyroxene and clinopyroxene. The experimental conditions range from 1150 to 1190 °C with oxygen fugacities from QFM to NNO+2. Run products were analyzed by laser-ablation ICP-MS.The new partition coefficients, combined with previously published data, can be used to model crystallization processes at low pressure. Partitioning of multivalent cations V, Fe and Eu varies as a function of the redox conditions, consistent with previous work, and can be used to constrain oxidation states of magmatic source regions. The V/Yb ratio is shown to be a useful proxy for oxidation state. The V/Yb ratio varies during mantle melting as a function of oxidation state of the mantle source, and it is not modified during fractional crystallization of olivine ± plag ± cpx. V/Yb increases from MORB, BABB to arc lavas, suggesting a progressive increase of fO2 from QFM to NNO+2. Apparent fO2 of arc lavas, however, is quite variable. These results demonstrate that sub-arc mantle displays a larger range of redox conditions toward a more oxidized mantle than the MORB mantle.

Parameterizations of calcic clinopyroxene—Melt trace element partition coefficients

Semiempirical parameterizations of the values and trends of variation of Nernst trace element partition coefficient data for Ca‐rich clinopyroxene (cpx/liquidD) are provided, applicable mainly to common terrestrial magmatic suites. cpx/liquidD data for most trace elements show significant variability which cannot be neglected when modeling melting and crystallization. The influence of pressure on cpx/liquidD is strong for many elements, particularly Na and Sr, which increase as pressure rises, and most high‐field strength elements, which decrease with increasing pressure. Most cpx/liquidD values increase as temperature decreases, as wt % melt MgO, MgO# (MgO/MgO+FeOtotal), CaO, and FeO contents drop, as cpx molar Mg# (Mg/Mg+Fetotal) decreases, and as wt % melt SiO2 and Na2O+K2O increase. No clear trends are seen for variations of cpx/liquidD against melt H2O. For mafic melts, many elements show well‐defined trends of cpx/liquidD increase as the clinopyroxene tetrahedral Al content (cpx Aliv) increases. Many cpx/liquidD are well correlated against cpx/liquidDTi, and many “near‐neighbor” elements show good cpx/liquidD intercorrelations (e.g., Zr‐Hf, U‐Th, Nb‐Ta, La‐Ce, Yb‐Lu). Cpx/liquidD profiles calculated from these parameterizations can constrain changes of D values during melting or crystallization. Cpx/liquidD for the rare earth elements were fit to the lattice strain model to derive fits that can reproduce the cpx/liquidDREE profile shapes (REE = rare earth elements). These fits indicate that cpx/liquidDREE for melts more evolved than picritic basalts cannot be modeled assuming that all REE are in octahedral coordination in a single M2 site, but also require sixfold partitioning into an M1 site for Lu‐Yb‐Tm‐Er.

Experimental study of trace element partitioning between enstatite and melt in enstatite chondrites at low oxygen fugacities and 5 GPa

In order to investigate the influence of very reducing conditions, we report enstatite-melt trace element partition coefficients (D) obtained on enstatite chondrite material at 5GPa and under oxygen fugacities (fO2) ranging between 0.8 and 8.2log units below the iron–wustite (IW) buffer. Experiments were conducted in a multianvil apparatus between 1580 and 1850°C, using doped (Sc, V, REE, HFSE, U, Th) starting materials. We used a two-site lattice strain model and a Monte-Carlo-type approach to model experimentally determined partition coefficient data. The model can fit our partitioning data, i.e. trace elements repartition in enstatite, which provides evidence for the attainment of equilibrium in our experiments. The precision on the lattice strain model parameters obtained from modelling does not enable determination of the influence of intensive parameters on crystal chemical partitioning, within our range of conditions (fO2, P, T, composition). We document the effect of variable oxygen fugacity on the partitioning of multivalent elements. Cr and V, which are trivalent in the pyroxene at around IW−1 are reduced to 2+ state with increasingly reducing conditions, thus affecting their partition coefficients. In our range of redox conditions Ti is always present as a mixture between 4+ and 3+ states. However the Ti3+-Ti4+ ratio increases strongly with increasingly reducing conditions. Moreover in highly reducing conditions, Nb and Ta, that usually are pentavalent in magmatic systems, appear to be reduced to lower valence species, which may be Nb2+ and Ta3+. We propose a new proxy for fO2 based on D(Cr)/D(V). Our new data extend the redox range covered by previous studies and allows this proxy to be used in the whole range of redox conditions of the solar system objects. We selected trace-element literature data of six chondrules on the criterion of their equilibrium. Applying the proxy to opx-matrix systems, we estimated that three type I chondrules have equilibrated at IW−7±1, one type I chondrule at IW−4±1, and two type II chondrules at IW+3±1. This first accurate estimation of enstatite-melt fO2 for type I chondrules is very close to CAI values.

Water-melt interaction in hydrous magmatic systems at high temperature and pressure

Experimental data on the structure and properties of melts and fluids relevant to water-melt interaction in hydrous magmatic systems in the Earth's interior have been reviewed. Complex relationships between water solubility in melts and their bulk composition [Al/Si-ratio, metal oxide/(Al + Si) and electron properties of metal cations] explain why water solubility in felsic magmas such as those of rhyolite and andesite composition is significantly greater than the water solubility in basalt melts. The silicate solubility in aqueous fluid is also significantly dependent on composition with metal oxide/(Al + Si) and electron properties of the metal cations, the dominant variables. Hydrogen bonding is not important in hydrous fluids and melts at temperatures above 500°C to 550°C and does not, therefore, play a role in hydrous magmatic systems. The properties of hydrous melts and aqueous solutions are governed by how the silicate speciation (Qn species, where n is the number of bridging oxygen in an individual species) varies with bulk composition, silicate composition, temperature, and pressure. The reactions that describe the interactions are similar in melts, fluids, and supercritical fluids. The degree of melt polymerization caused by dissolved water varies with melt composition and total water content. Silicate- and alkali-rich felsic magmatic melts are more sensitive to water content than more mafic magmas. Transport and configurational properties of hydrous magmatic melts can be modeled with the aid of the Qn speciation variations. Liquidus and melting phase relations of hydrous systems also can be described in such terms, as can minor and trace element partition coefficients. Stable isotope fractionation (e.g., D/H) can also be rationalized in this manner. Critical to these latter observations is the high silicate concentration in aqueous fluids. These components can enhance solubility of minor and trace elements by orders of magnitude and change the speciation of H and D complexes so that their fractionation factors change quite significantly. Data from pure silicate-H2O systems cannot be employed for these purposes.

The partitioning of trace elements between clinopyroxene and trachybasaltic melt during rapid cooling and crystal growth

We present the variation in trace element partition coefficients measured at the interface between rapidly cooled clinopyroxene crystals and co-existing melts. Results indicate that, as the cooling rate is increased, clinopyroxene crystals are progressively depleted in Si, Ca and Mg counterbalanced by enrichments in Al (mainly tetrahedral Aliv), Na and Ti. Partition coefficients (Ds) for rare earth elements (REE), high field strength elements (HFSE) and transition elements (TE) increase with increasing cooling rate, in response to clinopyroxene compositional variations. The entry of REE into the M2 site is facilitated by a coupled substitution where either Na substitutes for Ca on the M2 site or Aliv substitutes for Si in the tetrahedral site. The latter substitution reflects an increased ease of locally balancing the excess charge at M2 as the number of surrounding Aliv atoms increases. Due to the lower concentration of Ca in rapidly cooled clinopyroxenes, divalent large ion lithophile elements (LILE) on M2 decrease at the expense of monovalent cations. Conversely, higher concentrations of HFSE and TE on the M1 site are facilitated as the average charge on this site increases with the replacement of divalent-charged cations by Alvi. Although crystallization kinetics modify clinopyroxene composition, deviations from equilibrium partitioning are insufficient to change the tendency of a trace element to be compatible or incompatible. Consequently, there are regular relationships between ionic radius, valence of the trace element and D. At both equilibrium and cooling rate conditions, Ds for isovalent cations define parabola-like curves when plotted against ionic radius, consistent with the lattice strain model, demonstrating that the partitioning of trace elements is driven by charge balance mechanisms; cation substitution reactions can be treated in terms of the energetics of the various charge-imbalanced configurations.

Experimental partitioning of halogens and other trace elements between olivine, pyroxenes, amphibole and aqueous fluid at 2GPa and 900-1,300°C.

We present new partition coefficients for various trace elements including Cl between olivine, pyroxenes, amphibole and coexisting chlorine-bearing aqueous fluid in a series of high-pressure experiments at 2 GPa between 900 and 1,300 °C in natural and synthetic systems. Diamond aggregates were added to the experimental capsule set-up in order to separate the fluid from the solid residue and enable in situ analysis of the quenched solute by LA–ICP–MS. The chlorine and fluorine contents in mantle minerals were measured by electron microprobe, and the nature of OH defects was investigated by infrared spectroscopy. Furthermore, a fluorine-rich olivine from one selected sample was investigated by TEM. Results reveal average Cl concentrations in olivine and pyroxenes around 20 ppm and up to 900 ppm F in olivine, making olivine an important repository of halogens in the mantle. Chlorine is always incompatible with Cl partition coefficients D Clolivine/fluid varying between 10−5 and 10−3, whereas D Clorthopyroxene/fluid and D Clclinopyroxene/fluid are ~10−4 and D Clamphibole/fluid is ~5 × 10−3. Furthermore, partitioning results for incompatible trace element show that compatibilities of trace elements are generally ordered as D amph/fluid ≈ D cpx/fluid > D opx/fluid > D ol/fluid but that D mineral/fluid for Li and P is very similar for all observed silicate phases. Infrared spectra of olivine synthesized in a F-free Ti-bearing system show absorption bands at 3,525 and ~3,570 cm−1. In F ± TiO2-bearing systems, additional absorption bands appear at ~3,535, ~3,595, 3,640 and 3,670 cm−1. Absorption bands at ~3,530 and ~3,570 cm−1, previously assigned to humite-like point defects, profit from low synthesis temperatures and the presence of F. The presence of planar defects could not be proved by TEM investigations, but dislocations in the olivine lattice were observed and are suggested to be an important site for halogen incorporation in olivine.Electronic supplementary materialThe online version of this article (doi:10.1007/s00410-013-0902-5) contains supplementary material, which is available to authorized users.

The magma plumbing system of Bezymianny Volcano: Insights from a 54 year time series of trace element whole-rock geochemistry and amphibole compositions

Samples from 33 individual eruptions of Bezymianny volcano between 1956 and 2010 provide an opportunity to study in detail the temporal evolution of an arc volcano. Major element and ICP-MS trace element analyses show that the eruptive products shifted progressively from relatively silicic magma in 1956 (~60.4% SiO2) to more mafic compositions (e.g. 56.8% SiO2 in 2010). Amphibole compositions changed concurrently from low-Al2O3 to high-Al2O3. Whole rock element-element variation diagrams show tight compositional arrays, some with a distinct kink in the late 1970s, which cannot be reproduced by fractionation of a single magma along a liquid line of descent. Amphibole thermobarometry indicates amphibole crystallization in two separate reservoirs, one between 200 and 300MPa, and another between 500 and 750MPa. Liquid compositions calculated from the amphibole analyses show that liquids stored in each reservoir become increasingly mafic from 1956 to 2010, suggesting that each reservoir received magma inputs from more mafic sources throughout the eruptive cycle.End member mixing analysis of the dataset allows calculation of three end member compositions that can be combined in varying proportions to reproduce major and trace element whole rock compositions. The end-member mixing proportions vary systematically between 1956 and 2010, with maxima for end-members A, B and C during 1956, 1977, and 2010, respectively. Major element compositions of phenocrysts, combined with published trace element partition coefficients, show that each end member may have evolved from a common parental magma by fractionation of three different mineral assemblages, possibly due to different pressures of crystallization and volatile contents. The petrologic data are consistent with three magma reservoirs at different depths. Magmas from the shallowest reservoir erupted first, with increasing proportions of deeper reservoirs over time. Past studies have demonstrated similarities in eruptive style between Bezymianny and both Mount St. Helens and Soufrière Hills volcanoes. Contrasting time series data show that the magma plumbing systems of Mount St. Helens and Bezymianny have little in common, while Soufrière Hills may be a better analog. Similar eruptive styles at the surface do not necessarily reflect similar systems of magma supply and evolution at depth.

Trace element partitioning between majoritic garnet and silicate melt at 10–17 GPa: Implications for deep mantle processes

Melting experiments were performed on a silica-rich peridotite composition at 10–17GPa to determine majoritic garnet–melt partition coefficients (D) for major and trace elements. Our results show that D for many elements, including Na, Sc, Y and rare earth elements (REE), varies significantly with increasing pressure or proportion of majorite component. Lu and Sc become incompatible at 17GPa, with D decreasing from 1.5 at 10GPa to 0.9 at 17GPa. As predicted from lattice strain, log D for isovalent cations entering the large site of majoritic garnet exhibits a near-parabolic dependence on ionic radius. Our data are used to refine a previously published predictive model for garnet–melt partitioning of trivalent cations, which suffered from a lack of calibration in the 10–20GPa range. Our results suggest that Archean Al-depleted komatiites from Barberton (South Africa) may have been generated by partial melting of dry peridotite at depths between 200 and 400km. We also speculate that transition zone diamonds from Kankan (Guinea), which contain inclusions of majoritic garnet, may have formed from the partial reduction of CO2-rich magmas that subsequently transported them to the surface. This hypothesis would provide an explanation for the REE patterns of majoritic garnet trapped within these diamonds, including Eu anomalies. Finally, we show that segregation of majoritic garnet-bearing cumulates during crystallisation of a deep Martian magma ocean could lead to a variety of Lu/Hf and Sm/Nd ratios depending on pressure, leading to a range of ε143Nd and ε176Hf isotope signatures for potential mantle sources of Martian rocks.

Element partitioning between immiscible carbonatite–kamafugite melts with application to the Italian ultrapotassic suite

Trace element partition coefficients between conjugate carbonatite–silicate melts are a suitable tool to test whether such natural magmas originated by liquid immiscibility. In the Intra-Apennine Magmatic Province, Central Italy, highly silica-undersaturated ultrapotassic kamafugites coexist with carbonate-rich rocks of controversial nature: a mantle origin and magmatic nature of the carbonatites are indicated by the primitive nature of some kamafugites and by mantle-debris occurring within the kamafugite–carbonate/carbonatite deposits. An alternative hypothesis suggests that the CO2 originates from crustal carbonate assimilation by a basaltic magma, resulting in silica undersaturation of the kamafugite rocks. This experimental study investigates whether carbonatite and kamafugite melts such as the Italian ones can be immiscible at upper mantle conditions. Partition coefficients between conjugate melts determined in this study are compared with calculated element partitioning between natural carbonatite/kamafugite pairs to test a possible origin of the carbonatites by liquid immiscibility.The experiments demonstrate that kamafugite and alkali-rich carbonatite melts are immiscible at 1.7GPa, 1220°C. Trace element partition coefficients obtained from static- and centrifuging piston cylinder experiments are within factor 5 from unity and deviate less than a factor of 2 from unity for 26 out of 37 trace elements investigated. The alkali- and earth alkali-elements have Di≥1, indicating that carbonatite melts in equilibrium with alkaline silicate melts should be at least as alkali-rich as the silicate melts. HFSE partition preferentially into the silicate melt.A comparison of the experimental partition coefficients with natural carbonatite–kamafugite pairs from the Intra Apennine Province reveals good agreement (except for the alkalis), supporting that liquid immiscibility could indeed explain the observed rock suites. Original carbonatite compositions are not well preserved due to secondary leaching that almost completely removed the alkalis leading to significant uncertainties and compositional variations and potentially modified trace element concentrations too. Based on the observed trace element concentrations, large volume assimilation of sedimentary calcite is unlikely. We thus propose that CO2 was already present in the parental primitive magmas forming the kamafugites. The carbonatites formed by unmixing from the silicate magmas and the related CO2-saturation were most probably responsible and triggered the explosive eruptions in the Intra Apennine Province.

Trace element partitioning between perovskite and kimberlite to carbonatite melt: New experimental constraints

We present experimentally determined trace element partition coefficients (D) between perovskite (CaTiO3) and several different silicate and carbonate melt compositions. Our data indicate that perovskite is able to strongly fractionate trace elements during the evolution of SiO2-undersaturated magmas. In silicate magmas, perovskite efficiently fractionates Zr and Hf from Nb and Ta, with DZr and DHf coefficients below and DNb and DTa, significantly above unity, respectively. All trace element partition coefficients between perovskite and the diverse silicate melts, which primarily differ in their CaO and SiO2 contents, are fairly similar within propagated uncertainties. We find that D perovskite/silicate melt for the high field strength elements is much lower than corresponding perovskite/carbonate melt partition coefficients for the same elements. Another key difference between perovskite/melt trace element partitioning in silicate and carbonate systems is the relative behavior of U and Th. Whereas all silicate system experiments show that DTh>DU, the relationship is different in the carbonatite experiment. This implies that perovskite may fractionate U and Th in silicate magmas whereas perovskite cannot fractionate U and Th in carbonate magmas. The Th/U ratio of magmatic perovskite may, therefore, provide a tool for a better understanding of the evolution of hybrid carbonate–silicate magmatic systems.The rare earth elements (REE) show similar perovskite/melt partitioning in the silicate and carbonate systems investigated. Our data indicate that the REE are highly compatible in perovskite in both silicate and carbonatite compositions, which explains the high rare earth element concentrations in naturally occurring magmatic perovskites. In particular, our new trace element partition coefficients help to better understand and to quantify the strong control of perovskite on the bulk REE budget of kimberlites and related rocks.

In situ ion-microprobe determination of trace element partition coefficients for hornblende, plagioclase, orthopyroxene, and apatite in equilibrium with natural rhyolitic glass, Little Glass Mountain Rhyolite, California

Partially crystalline hornblende gabbro inclusions from the Little Glass Mountain Rhyolite contain euhedral plagioclase, orthopyroxene, hornblende, and apatite crystals in contact with interstitial rhyolitic (71-76% SiO 2 ) glass. Textural and mineral compositional data indicate that the gabbros crystallized sufficiently slowly that surface equilibrium was closely approached at the interface between crystals and the liquid. This rare occurrence represents a natural dynamic crystallization experiment with a “run time” that is not realistically achievable in the laboratory. SIMS analysis of mineral rim-glass pairs have permitted the determination of high-quality, equilibrium trace-element partition coefficients for all four minerals. These data augment the limited partition coefficient database for minerals in high-SiO 2 rhyolitic systems. For all minerals, the D values are consistent with those anticipated from crystal-chemical considerations. These data further support a liquid SiO 2 control on the REEs (and presumably other elements) partitioning wherein D values systematically increase with increasing liquid SiO 2 content.