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 H₂O 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.     

Rare earth element partitioning between clinopyroxene and silicate liquid at moderate to high pressure

Experimental determination of over seventy sets of clinopyroxene/silicate liquid (glass) partitition coefficients (D) for four rare earth elements (REE — La, Sm, Ho, Lu) in a range of REE-enriched natural rock compositions (basalt, basaltic andesite, andesite and rhyodacite) demonstrate a convex upward pattern, favouring the heavy REE (Ho, Lu) and markedly discriminating against the light REE (La). These patterns are consistent with previously documented clinopyroxene D values reported from natural phenocryst/matrix pairs and from experimental work using either REE-enriched compositions and electron microprobe analytical techniques (as in the present study) or natural or synthetic undoped compositions and mass spectrometric, ion probe or X-ray autoradiographic analytical techniques. However, the large data base in the present study allows evaluation of the effect of compositional and physical parameters on REE partitioning relationships. Considering D_Ho, it is shown that (1) D increases 6-fold with increasing SiO₂ content of the coexisting liquid from 50 to 70 wt% SiO₂ (2) D increases 4-fold with decreasing temperature from 1,120°C to 900° C (3) D increases 2-fold with increasing pressure from 2.5 to 20 kb. (4) D increases 2-fold fO₂ increases from approximately that of the MW buffer to the HM buffer (5) D remains unchanged within experimental error as the water content of the melt changes from 0.3 to 10% by weight H₂O.The absolute REE content of the clinopyroxene shows no consistent trend with temperature, but decreases slightly with increasing pressure, paralleling an increase in the jadeite component of the pyroxene. Thus the increase in D with increasing pressure is attributed to changes in the silicate liquid structure, which discriminate against accommodation of REE with increasing pressure. The clinopyroxene REE content increases with increasing fO₂, and in this case the increase in D with increasing fO₂ may be attributed mainly to this change in the clinopyroxene composition. Application of the present results to geochemical modelling allows a more appropriate choice of D values, according to the liquid composition and physical conditions applicable in the modelled system. They may also be used to evaluate cognate or xenocrystic relationships between clinopyroxene megacrysts and their host matrix.     

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 (D_Mg) and Si (D_Si) are related to the melt Mg/Si ratio such that D_Si becomes lower than D_Mg 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.     

Rare earth element partitioning between minerals from anhydrous spinel peridotite xenoliths

REE abundances in minerals from spinel peridotite xenoliths from West Germany, the south-western U.S. and Mongolia decrease in the order clinopyroxene > orthopyroxene > olivine > spinel. While clinopyroxenes are similar in absolute chondrite-normalized concentrations to those known from other studies, orthopyroxenes and olivines are significantly lower in LREE although comparable in HREE. Spinels are much lower in all REE than any previously reported values and are completely negligible for the REE budget of peridotites.Partition coefficients for most orthopyroxene/clinopyroxene pairs increase systematically from La to Lu. Olivine/clinopyroxene and spinel/clinopyroxene partition coefficients increase from the intermediate rare earth elements to Lu and normally are higher for La compared to Sm.The application of Nagasawa's (1966) elastic lattice model suggests that all heavy but only minor amounts of the light REE substitute into structural positions of orthopyroxene and olivine.Significant differences between orthopyroxene/clinopyroxene partition coefficients for various xenoliths may be assigned to dependences upon equilibration temperature and bulk chemistry.Apart from grain surface contaminations, fluid inclusions which are practically always present in mantle minerals, can highly concentrate light rare earth elements and thus may be responsible for unexpectedly high concentrations of incompatible elements frequently reported for mantle olivines or orthopyroxenes.     

The Partitioning of Tungsten bwtween Aqueous Fluids and Silicate Melts

An experimental study has been carried out to determine the partition coefficients of tungsten between aqueous fluids and granitic melts at 800 °C and 1.5 kb with natural granite as the starting material. The effects of the solutions on the partition coefficients of tungsten show a sequence of P > CO ₃ ²⁻ > B > H₂O. The effects are limited (generallyK _D < 0.3) and the tungsten shows a preferential trend toward the melt over the aqueous fluid. The value ofK _D increases with increasing concentration of phosphorus; theK _D increases first and then reduces with the concentration of CO ₃ ²⁻ when temperature decreases, theK _D between the solution of CO ₃ ²⁻ and the silicate melt increases, and that between the solution of B₄O ₇ ²⁻ and the silicate melt decreases. The partition coefficients of phosphorus and sodium between fluids and silicate melts have been calculated from the concentrations of the elements in the melts. TheK _D value for phosphorus is 0.38 and that for sodium is 0.56. Evidence shows that the elements tend to become richer and richer in the melts.     

Rare earth and high field strength element partitioning between iron-rich clinopyroxenes and felsic liquids

Rare earth elements are commonly assumed to substitute only for Ca in clinopyroxene because of the similarity of ionic radii for REE³⁺ and Ca²⁺ in eightfold coordination. The assumption is valid for Mg-rich clinopyroxenes for which observed mineral/melt partition coefficients are readily predicted by the lattice strain model for substitution onto a single site (e.g., Wood and Blundy 1997). We show that natural Fe-rich pyroxenes in both silica-undersaturated and silica-oversaturated magmatic systems deviate from this behavior. Salites (Mg# 48–59) in phonolites from Tenerife, ferrohedenbergites (Mg# 14.2–16.2) from the rhyolitic Bandelier Tuff, and ferroaugites (Mg# 9.6–32) from the rhyolitic Rattlesnake Tuff have higher heavy REE contents than predicted by single-site substitution. The ionic radius of Fe²⁺ in sixfold coordination is substantially greater than that of Mg²⁺; hence, we propose that, in Fe-rich clinopyroxenes, heavy REE are significantly partitioned between eightfold Ca sites and sixfold Fe and Mg sites such that Yb and Lu exist dominantly in sixfold coordination. We also outline a REE-based method of identifying pyroxene/melt pairs in systems with multiple liquid and crystal populations, based upon the assumption that LREE and MREE reside exclusively in eightfold coordination in pyroxene. Contrary to expectations, interpolation of mineral/melt partition coefficient data for heavy REE does not predict the behavior of Y. We speculate that mass fractionation effects play a role in mineral/melt lithophile trace element partitioning that is detectable among pairs of isovalent elements with near-identical radii, such as Y and Ho, Zr and Hf, and Nb and Ta.     

Rare earth element partitioning between hydrous ferric oxides and acid mine water during iron oxidation

Ferrous iron rapidly oxidizes to Fe (III) and precipitates as hydrous Fe (III) oxides in acid mine waters. This study examines the effect of Fe precipitation on the rare earth element (REE) geochemistry of acid mine waters to determine the pH range over which REEs behave conservatively and the range over which attenuation and fractionation occur. Two field studies were designed to investigate REE attenuation during Fe oxidation in acidic, alpine surface waters. To complement these field studies, a suite of six acid mine waters with a pH range from 1.6 to 6.1 were collected and allowed to oxidize in the laboratory at ambient conditions to determine the partitioning of REEs during Fe oxidation and precipitation. Results from field experiments document that even with substantial Fe oxidation, the REEs remain dissolved in acid, sulfate waters with pH below 5.1. Between pH 5.1 and 6.6 the REEs partitioned to the solid phases in the water column, and heavy REEs were preferentially removed compared to light REEs. Laboratory experiments corroborated field data with the most solid-phase partitioning occurring in the waters with the highest pH.     

Crystal chemical controls on rare earth element partitioning between epidote-group minerals and melts: an experimental and theoretical study

We have experimentally determined the partitioning of REE (rare earth elements) between zoisite and hydrous silicate melt at 1,100 °C and 3 GPa. All REE behave moderately compatible in zoisite with respect to the melt and all show a smooth parabolic dependence on ionic radius. The partitioning parabola peaks at Nd , and the compatibility slightly decreases towards La and decreases by half an order of magnitude towards Yb . Application of the elastic strain model of Blundy and Wood (1994) to the available zoisite and allanite REE mineral/melt partitioning data and comparison with partitioning pattern calculated from a combination of structural and physical data (taken from the literature) with the elastic strain model suggest that in zoisite REE prefer the A1-site and that only La and Ce are incorporated into the A2-site in significant amounts. In contrast, in allanite, all REE are preferentially incorporated into the large and highly co-ordinated A2 site. As a result, zoisite fractionates the MREE effectively from the HREE and moderately from the LREE, while allanite fractionates the LREE very effectively from the MREE and HREE. Consequently, the presence of either zoisite or allanite during slab melting will lead to quite different REE pattern in the produced melt.Editorial responsibility: J. Hoefs     

The partitioning of copper and molybdenum between silicate melts and aqueous fluids

The partitioning of copper and molybdenum between silicate melts and aqueous fluids has been determined at 750°C, and 1.4 Kb. The experiments were conducted in a $\text{1}\text{2}$ inch ID, rapid quench, cold seal pressure vessel. The aqueous and glass phase run products were analyzed by atomic absorption spectrophotometry and ion microprobe, respectively. The vapor/melt partition coefficient for copper, $\text{D}{}^{\mathrm{<mtext>v</mtext><mtext>l</mtext>}}{}_{\mathrm{Cu}}$, defined as the ratio of the concentrations of copper in the vapor to copper in the melt was found to be $\text{D}{}^{\mathrm{<mtext>v</mtext><mtext>l</mtext>}}{}_{\mathrm{Cu}}\text{= (9.1 ± 2.5)m}{}^{\mathrm{v}}{}_{\mathrm{Cl}}$ at NNO up to at least 4.5 moles of chlorine per kg of solution. The partition coefficient for molybdenum is equal to 2.5 ± 1.6 at NNO and QFM; its value is independent of the fluorine concentration of the melt up to at least 1.7 wt. percent fluorine, and of the chlorine concentration up to at least 4.5 moles of chlorine per kg of solution. Copper is probably present in the univalent state in both the silicate melt and in the associated aqueous phase at NNO; the most important aqueous complex of copper is probably CuCl⁰. Molybdenum is probably present in the aqueous phase as one or more molybdate species.     

Henry’s and non-Henry’s law behavior of Br in simple marine systems

Experimental studies for the partitioning of Br as a trace element between aqueous and solid solutions were carried out in simple marine systems. The evaporation experiments were performed at 25°C and 1 atm in the systems of halite (NaCl), sylvite (KCl), kainite (KMgClSO₄ · 2.75H₂O), carnallite (KMgCl₃ · 6H₂O), and bischofite (MgCl₂ · 6H₂O). The partition coefficients for the systems investigated are constant only at a restricted concentration range. For concentrations lower than 100 to 300 μg Br/g aqueous solutions, D_Br increases with decreasing concentrations. Various evaporation experiments indicate that this observation is not due to kinetic effects (evaporation rates). To find a link between the partition coefficient and the Henry’s law behavior, the activity coefficients of the trace components in the solid solutions were recalculated from the experimentally derived data. It can be shown from these calculations that constant activity coefficients or Henry’s law behavior is reached for higher mole fractions of the trace component in the solid solution in halite and sylvite and thus correspond to constant partition coefficients. For bischofite and carnallite, Henry’s law behavior is restricted to the lower mole fractions, where D_Br is not constant. This behavior is caused by the activity of the trace component in the aqueous solution, which is powered by the stoichiometric factor of this component in the Br-end-member solid solution. For halite, sylvite, and kainite, this factor equals 1 and is 2 for bischofite and 3 for carnallite. However, it is thus impossible to correlate Henry’s law behavior with constant partition coefficients for solid solution systems where the stoichiometric factor of the trace component is greater than 1.     

Effect of melt composition on the partitioning of trace elements between titanite and silicate melt

Isobaric and isothermal experiments were performed to investigate the effect of melt composition on the partitioning of trace elements between titanite (CaTiSiO₅) and a range of different silicate melts. Titanite-melt partition coefficients for 18 trace elements were determined by secondary ion mass spectrometry (SIMS) analyses of experimental run products. The partition coefficients for the rare earth elements and for Th, Nb, and Ta reveal a strong influence of melt composition on partition coefficients, whereas partition coefficients for other studied monovalent, divalent and most quadrivalent (i.e., Zr, Hf) cations are not significantly affected by melt composition. The present data show that the influence of melt composition may not be neglected when modelling trace element partitioning.It is argued that it is mainly the change of coordination number and the regularity of the coordination space of trace elements in the melt structure that controls partition coefficients in our experiments. Furthermore, our data also show that the substitution mechanism by which trace elements are incorporated into titanite crystals may be of additional importance in this context.     

Isovalent trace element partitioning between minerals and melts: A computer simulation study

We present a new approach for the rationalisation of trace element partitioning between silicate melts and minerals, which is not based on the empirical, parameterised continuum models in common use. We calculate the energetics of ion substitution using atomistic simulation techniques, which include an explicit evaluation of the relaxation energy (strain energy) contribution to this process. Solution energies are estimated for isovalent impurities in CaO, diopside, orthoenstatite, and forsterite. These show a parabolic dependence on ionic radius, similar to the variation of mineral-melt partition coefficients with ionic radius. The success of the empirical models, which often include only the strain energy, appear to have been due to the partial cancellation of energy terms, and to the empirical fitting of the parameters included in these models. Our approach can be readily extended to aliovalent substitution.     

Rare earth element geochemistry and petrogenesis of miles (IIE) silicate inclusions

An ion probe study of rare earth element (REE) geochemistry of silicate inclusions in the Miles IIE iron meteorite was carried out. Individual mineral phases among inclusions have distinct REE patterns and abundances. Most silicate grains have homogeneous REE abundances but show considerable intergrain variations between inclusions. A few pyroxene grains display normal igneous REE zoning. Phosphates (whitlockite and apatite) are highly enriched in REEs (50 to 2000 × CI) with a relatively light rare earth element (LREE)-enriched REE pattern. They usually occurred near the interfaces between inclusions and Fe host. In Miles, albitic glasses exhibit two distinctive REE patterns: a highly fractionated LREE-enriched (CI normalized La/Sm ∼15) pattern with a large positive Eu anomaly and a relatively heavy rare earth element (HREE)-enriched pattern (CI-normalized Lu/Gd ∼4) with a positive Eu anomaly and a negative Yb anomaly. The glass is generally depleted in REEs relative to CI chondrites.The bulk REE abundances for each inclusion, calculated from modal abundances, vary widely, from relatively depleted in REEs (0.1 to 3 × CI) with a fractionated HREE-enriched pattern to highly enriched in REEs (10 to 100 × CI) with a relatively LREE-enriched pattern. The estimated whole rock REE abundances for Miles are at ∼ 10 × CI with a relatively LREE-enriched pattern. This implies that Miles silicates could represent the product of a low degree (∼10%) partial melting of a chondritic source. Phenocrysts of pyroxene in pyroxene-glassy inclusions were not in equilibrium with coexisting albitic glass and they could have crystallized from a parental melt with REEs of ∼ 10 × CI. Albitic glass appears to have formed by remelting of preexisting feldspar + pyroxene + tridymite assemblage. Yb anomaly played an important role in differentiation processes of Miles silicate inclusions; however, its origin remains unsolved.The REE data from this study suggest that Miles, like Colomera and Weekeroo Station, formed when a molten Fe ball collided on a differentiated silicate regolith near the surface of an asteroid. Silicate fragments were mixed with molten Fe by the impact. Heat from molten Fe caused localized melting of feldspar + pyroxene + tridymite assemblage. The inclusions remained isolated from one another during subsequent rapid cooling.     

Solution of rare earth elements in silicate solid phases; henry's law revisited in light of defect chemistry

It is shown that experimentally observed departures from Henry's law at high dilution conditions for rare earth elements (REE) in garnet, clinopyroxene and plagioclase can be referred to stabilization in the lattice point defects. The solution process involves negative interaction parameters (?500/?3,000 cal for garnet). On the basis of the experimental data, the activity coefficients of the trace REE can be approximated by a model based on trace cation/cationic vacancy binary solid solution, taking into account association phenomena between them as well.     

Evidence from olivine/melt element partitioning that nonbridging oxygen in silicate melts are not equivalent

Partitioning of Ca, Mn, Mg, and Fe²⁺ between olivine and melt has been used to examine the influence of energetically nonequivalent nonbridging oxygen in silicate melts. Partitioning experiments were conducted at ambient pressure in air and 1400°C with melts in equilibrium with forsterite-rich olivine (Fo >95 mol%). The main compositional variables of the melts were NBO/T and Na/(Na+Ca). In all melts, the main structural units were of Q⁴, Q³, and Q² type with nonbridging oxygen, therefore, in the Q³ and Q² units.For melts with high Q³/Q²-abundance ratio (corresponding to NBO/T near 1), increasing Na/(Na+Ca) [and Na/(Na+Ca+Mn+Mg+Fe²⁺)] results in a systematic decrease of the partition coefficients, K_Ca^ol/melt, K_Mn^ol/melt, K_Mg^ol/melt, and K_Fe2+^ol/melt, because of ordering of the network-modifying Ca, Mn, Mg, and Fe²⁺ among nonbridging oxygen in Q³ and Q² structural units. This decrease is more pronounced the smaller the ionic radius of the cation. With decreasing Q³/Q² abundance ratio (less-polymerized melts) this effect becomes less pronounced.Activity-composition relations among network-modifying cations in silicate melts are, therefore, governed by availability of energetically nonequivalent nonbridging oxygen in individual Qⁿ-species in the melt. As a result, any composition change that enhances abundance of highly depolymerized Qⁿ-species will cause partition coefficients to decrease.     

Determination of melt influence on divalent element partitioning between anorthite and CMAS melts

We propose a theory for crystal-melt trace element partitioning that considers the energetic consequences of crystal-lattice strain, of multi-component major-element silicate liquid mixing, and of trace-element activity coefficients in melts. We demonstrate application of the theory using newly determined partition coefficients for Ca, Mg, Sr, and Ba between pure anorthite and seven CMAS liquid compositions at 1330 °C and 1 atm. By selecting a range of melt compositions in equilibrium with a common crystal composition at equal liquidus temperature and pressure, we have isolated the contribution of melt composition to divalent trace element partitioning in this simple system. The partitioning data are fit to Onuma curves with parameterizations that can be thermodynamically rationalized in terms of the melt major element activity product (a_Al2O3)(a_SiO2)² and lattice strain theory modeling. Residuals between observed partition coefficients and the lattice strain plus major oxide melt activity model are then attributed to non-ideality of trace constituents in the liquids. The activity coefficients of the trace species in the melt are found to vary systematically with composition. Accounting for the major and trace element thermodynamics in the melt allows a good fit in which the parameters of the crystal-lattice strain model are independent of melt composition.     

Rare earth partitioning between immiscible carbonate and silicate liquids and CO2 vapor: Results and implications for the formation of light rare earth-enriched rocks

Melting relations at 5 and 20 kbar on the composition join sanidine-potassium carbonate are dominated by a two-liquid region that covers over 60% of the join at 1,300 ° C. At this temperature, the silicate melt contains approximately 19 wt% carbonate component at 5 kbar and 32 wt% carbonate component at 20 kbar. The conjugate carbonate melt contains less than 5 wt% silicate component, and it varies less as a function of temperature than does the silicate melt.Partition coefficients for Ce, Sm, and Tm between the immiscible carbonate and silicate melts at 1,200 ° and 1,300 ° C at 5 and 20 kbar are in favor of the carbonate melt by a factor of 2–3 for light REE and 5–8 for heavy REE. The effect of pressure on partitioning cannot be evaluated independently because of complementary changes in melt compositions.Minimum REE partition coefficients for CO₂ vapor/carbonate melt and CO₂ vapor/silicate melt can be calculated from the carbonate melt/silicate melt partition coefficients, the known proportions of melt, and maximum estimates of the proportion of CO₂ vapor. The vapor phase is enriched in light REE relative to both melts at 20 kbar and enriched in all REE, especially the light elements, at 5 kbar. The enrichment of REE in CO₂ vapor relative to both melts is 3–4 orders of magnitude in excess of that in water vapor (Mysen, 1979) at 5 kbar and is approximately the same as that in water vapor at 20 kbar.Mantle metasomatism by a CO₂-rich vapor enriched in light REE, occurring as a precursor to magma genesis, may explain the enhanced REE contents and light REE enrichment of carbonatites, alkali-rich silicate melts, and kimberlites. Light REE enrichment in fenites and the granular suite of nodules from kimberlites attests to the mobility of REE in CO₂-rich fluids under both mantle and crustal conditions.     

Partitioning coefficients between olivine and silicate melts

Variation of Nernst partition coefficients (D) between olivine and silicate melts cannot be neglected when modeling partial melting and fractional crystallization. Published natural and experimental ^{olivine/liquid}D data were examined for covariation with pressure, temperature, olivine forsterite content, and melt SiO₂, H₂O, MgO and MgO/MgO + FeO^total. Values of ^{olivine/liquid}D generally increase with decreasing temperature and melt MgO content, and with increasing melt SiO₂ content, but generally show poor correlations with other variables. Multi-element ^{olivine/liquid}D profiles calculated from regressions of D REE–Sc–Y vs. melt MgO content are compared to results of the Lattice Strain Model to link melt MgO and: D₀ (the strain compensated partition coefficient), E_M³⁺ (Young's Modulus), and r₀ (the size of the M site). Ln D₀ varies linearly with Ln MgO in the melt; E_M³⁺ varies linearly with melt MgO, with a dog-leg at ca. 1.5% MgO; and r₀ remains constant at 0.807 Å. These equations are then used to calculate ^{olivine/liquid}D for these elements using the Lattice Strain Model. These empirical parameterizations of ^{olivine/liquid}D variations yield results comparable to experimental or natural partitioning data, and can easily be integrated into existing trace element modeling algorithms. The ^{olivine/liquid}D data suggest that basaltic melts in equilibrium with pure olivine may acquire small negative Ta–Hf–Zr–Ti anomalies, but that negative Nb anomalies are unlikely to develop. Misfits between results of the Lattice Strain Model and most light rare earth and large ion lithophile partitioning data suggest that kinetic effects may limit the lower value of D for extremely incompatible elements in natural situations characterized by high cooling/crystallization rates.     

Experimental determination of rare earth element partitioning between hydrous silicate melt,amphibole and garnet peridotite minerals at upper mantle pressures and temperatures

Partition coefficients of Ce, Sm and Tm involving garnet peridotite minerals, amphibole and hydrous silicate melt have been determined experimentally in the temperature and pressure ranges 950–1075°C and 10–25 kbar.Only several parts per million to several tens of parts per million of rare earth element (REE) can dissolve in the minerals before the crystal-liquid partition coefficients begin to vary as a function of REE content. The concentration ranges of constant partition coefficient increase with increasing temperature and are also positively correlated with the magnitude of the crystal-liquid partition coefficients. The upper concentration limits of constant partition coefficient and the value of the crystal-liquid partition coefficient for REE decrease in the order garnet > clinopyroxene > amphibole > orthopyroxene > olivine.Partition coefficients may vary by at least an order of magnitude as a function of bulk composition of the liquid phase (e.g. changing from basaltic to andesitic). The approximate ranges of the values of the partition coefficients as a function of bulk liquid composition are as follows:

$\text{Ce}\text{Sm}\text{Tm}\text{K}{}^{\mathrm{<mtext>ga-liq</mtext>}}\text{0.01–0.1}\text{0.3–3.4}\text{1–10}\text{K}{}^{\mathrm{<mtext>cpx-liq</mtext>}}\text{0.05–0.4}\text{0.09–0.7}\text{0.04–0.4}\text{K}{}^{\mathrm{<mtext>amph-liq</mtext>}}\text{0.04–0.4}\text{0.08–0.8}\text{0.07–0.7}\text{K}{}^{\mathrm{<mtext>opx-liq</mtext>}}\text{0.04–0.1}\text{0.05–0.1}\text{0.08–0.1}\text{K}{}^{\mathrm{<mtext>ol-liq</mtext>}}\text{0.01–0.02}\text{0.01–0.02}\text{0.01–0.02}$

where the values increase with increasing acidity of the melt.