Implications of liquid-liquid distribution coefficients to mineral-liquid partitioning

The effect of silicate liquid structure upon mineral-liquid partitioning has been investigated by determining element partitioning data for coexisting immiscible granitic and ferrobasaltic magmas. The resulting elemental distribution patterns may be interpreted in terms of the relative states of polymerization of the coexisting magmas. Highly charged cations (REE, Ti, Fe, Mn, etc.) are enriched in the ferrobasaltic melt. The ferrobasaltic melt is relatively depolymerized due to its low $\text{Si}\text{O}$ ratio. This allows highly charged cations to obtain stable coordination polyhedra of oxygen within the ferrobasaltic melt. The granitic melt is a highly polymerized network structure in which Al can occupy tetrahedral sites in copolymerization with Si. The substitution of Al⁺³ for Si⁺⁴ produces a local charge imbalance in the granitic melt which is satisfied by a coupled substitution of alkalis, thus explaining the enrichment of low charge density cations, the alkalis, in the granitic melt. P₂O₅ increases the width of the solvus and, therefore, the values of the distribution coefficients of the trace elements. This effect is attributed to complexing of metal cations with PO₄^?3 groups in the ferrobasaltic melt.The values of ferrobasalt-granite liquid distribution coefficients are reflected in distribution coefficients for a mineral and melts of different compositions. The mineral-liquid distribution coefficient for a highly charged cation is greater for a mineral coexisting with a highly polymerized melt (granite) than it is for that same mineral and a depolymerized melt (ferrobasalt). The opposite is true for low charge density cations. Mineralliquid and liquid-liquid distribution coefficients determined for the REE's indicate that fractionated REE patterns are due to mineral selectivity and not the state of polymerization of the melt.     

PbO-SiO2 melts: structure and thermodynamics of mixing

Thermodynamic properties of PbO-SiO₂ melts, obtained from published data and calculated from freezing point depressions, reflect the gradual polymerization of silicate anions in the melt as the $\text{SiO}{}_2\text{PbO}$ ratio is increased. The free energy of mixing curve at 1000°C has a minimum at 40 mole % SiO₂ and is convex-upward between 72 and 98 mole % SiO₂. The latter is an indication of metastable liquid immiscibility. The free energy minimum is correlated with the maximum in the distribution of nonbridging oxygens in the melt. In SiO₂-poor melts, the activities of PbO and SiO₂ (pure liquid standard states) show sharp negative deviations from ideality. The PbO activity reflects the paucity of free oxygen species in the melt whereas the SiO₂ activity reflects the depolymerized state of the silicate anions. In more SiO₂-rich melts, the activity of SiO₂ shows a positive deviation from ideality which is qualitatively correlated to a polymerization parameter. The heat of mixing term has a minimum of ?2000 cal at 35 mole % SiO₂ and a maximum of +200 cal at 90 mole % SiO₂. The minimum is associated with the exothermic heat effect obtained during the reaction (O⁰) + (O^2?) = 2(O^?), whereas the maximum corresponds to the endothermic heat effect obtained when coordination polyhedra of oxygens form around the Pb cation. The entropy of mixing curve has the same form but is systematically smaller than a theoretical curve calculated on the assumption of random mixing of oxygen species. The discrepancy is due to the entropy loss obtained by the clustering of oxygen species to form complex silicate species.     

Oxide solution mechanisms in silicate melts: Systematic variations in the activity coefficient of SiO2

Phase equilibria and spectroscopic data are used to develop a simple model for the interaction of various oxide components and molten SiO₂. Network modifying oxides, M_xO_y produce nonbridging oxygens thereby depolymerizing the SiO₂ network. The energetics of nonbridging oxygen formation are least favorable when the field strength of the metal cation is high. This produces relatively strong M-O and Si-O-Si bridging bonds at the expense of weaker Si-O-M bonds (De Jonget al., 1980). This relationship is manifested by an increase in positive deviations from ideality with increasing cation field strength in M_xO_y-SiO₂ systems; the activity coefficient of SiO₂ is inversely correlated with Si-O-M bond strength. Network forming oxides (aluminates, phosphates, titanates. zirconates, etc.) may copolymerize with the SiO₂ network. Mixing on the same quasi-lattice produces solutions which approach ideality. Deviations from ideality in such solutions can be linked to distortions in the SiO₂ network. Discrete anion formers (phosphates, titanates, chromates, zirconates) complex with metal oxides other than SiO₂ to form discrete structural units which do not copolymerize with SiO₂. The SiO₂ network is essentially shielded from the high charge density cations in such systems and unmixing is common. As a result, the relative deviations from ideality in such melts are high. It is important to recognize that oxides such as P₂O₅, TiO₂ and ZrO₂ may act as either network-formers or discrete anion formers depending upon melt composition, and are probably distributed between these two “sites” in most geologically important liquids. The latter structural role is favored in more basic compositions.     

Experimental evidence for a relationship between liquid immiscibility and ore-formation in felsic magmas

Geological studies demonstrate that liquid immiscibility in felsic magma closely associates with the ore forming process. In order to obtain experimental evidence demonstrating the relationship between the ore forming process and liquid immiscibility in felsic magma, we carried out a series of experiments at high temperature and atmospheric pressure. The experimental results show that the granite ∼ KBF₄∼Na₂MoO₄ system is a homogeneous melt at high temperature. With decrease in temperature, however, the melt decomposes into two immiscible melts: silicate melt and ore-forming melt. The ore-forming melt exists as globules in the silicate phase. Molybdenm, Ca, Na, Mg, P, Mn, F, B, and OH are concentrated in these globules. The ore forming melt is characterized with very low SiO₂ and Al₂O₃ concentrations but the concentration of MoO₃ and CaO is very high. In contrast, the silicate melts are significantly enriched in SiO₂ and Al₂O₃, and depleted in MoO₃ and CaO. In the silicate melt the concentrations of network modifying elements (e.g. Mo, Ca, Na, P, Mg) and volatiles (F, OH) are very low. The differences between the two immiscible melts exist not only in chemical composition but also in structure. The ore-forming melt structurally consists of [MoO₄], [MoOF₄], [B(OH)₄], and OH, while the silicate melt is [Si0₄]. Because of the difference in composition and structure the two immiscible melts possess different physical properties. Compared to silicate melt, the ore-forming melt has a lower density and viscosity, which permits the globules to behave as bubbles in granite magma and to move and concentrate in the upper part of magma chamber. This process is probably responsible for the concentration of ore-forming elements in the upper part of granite bodies and their immediate aureoles. The present experimental results suggest that liquation in felsic magma can be the first step in the ore-forming process during granitoid evolution.     

Experiments on silicate melt immiscibility in the system Fe2SiO4–KAlSi3O8–SiO2–CaO–MgO–TiO2–P2O5 and implications for natural magmas

The effect of CaO and MgO, with or without TiO₂ and P₂O₅, on the two-melt field in the simplified system Fe₂SiO₄–KAlSi₃O₈–SiO₂ has been experimentally determined at 1,050°–1,240°C, 400 MPa. Despite the suppressing effect of MgO, CaO, and pressure on silicate melt immiscibility, our experiments show that this process is still viable at mid-crustal pressures when small amounts (0.6–2.0 wt%) of P₂O₅ and TiO₂ are present. Our data stress that the major element partition coefficients between the two melts are highly correlated with the degree of polymerisation (nbo/t) of the SiO₂-rich melt, whatever temperature, pressure, or exact composition. Experimental immiscible melt compositions in natural systems at 0.1 MPa from the literature (lunar and tholeiitic basalts) plot on similar but distinct curves compared to the simplified system. These relations between melt polymerisation and partition coefficients, which hold for a large range of compositions and fO₂, are extended to various volcanic and plutonic rocks. This analysis strengthens the proposal that silicate melt immiscibility can be important in volcanic rocks of various compositions (from tholeiitic basalts to lamprophyres). However, the majority of proposed immiscible compositions in plutonic rocks are at least not coexisting melts, but may have suffered accumulation of early crystallized minerals.     

The role of phosphorus in rhyolitic liquids as determined from the homogeneous iron redox equilibrium

The redox ratio of iron is used as an indicator of solution properties of silicate liquids in the system (SiO–Al₂O₃–K₂O–FeO–Fe₂O₃–P₂O₅). Glasses containing 80–85 mol% SiO₂ with 1 mol% Fe₂O₃ and compositions covering a range of K₂O/Al₂O₃ were synthesized at 1400°C in air (fixed fO₂). Variations in the ratio FeO/FeO_1.5 resulting from the addition of P₂O₅ are used to determine the solution behavior of phosphorus and its interactions with other cations in the silicate melt. In 80 mol% SiO₂ peralkaline melts the redox ratio, expressed as FeO/FeO_1.5, is unchanged relative to the reference curve with the addition of 3 mol% P₂O₅. Yet, the iron redox ratio in the 85 mol% SiO₂ potassium aluminosilicate melts is decreased relative to phosphorus-free liquids even for small amounts of P₂O₅ (0.5 mol%). The redox ratio in peraluminous melts is decreased relative to phosphorus- free liquids at P₂O₅ concentrations of 3 mol%. In peraluminous liquids, complexing of both Fe⁺³–O–P⁺⁵ and Al⁺³–O–P⁺⁵ occur. The activity coefficient of Fe⁺³ is decreased because more ferric iron can be accommodated than in phosphorus-free liquids. In peralkaline melts, there is no evidence that P⁺⁵ is removing K⁺ from either Al⁺³ or Fe⁺³ species. In chargebalanced melts with 3 mol% Fe₂O₃ and very high P₂O₅ concentrations, phosphorus removes K⁺ from K–O–Fe⁺³ complexes resulting in a redox increase. P₂O₅ should be accommodated easily in peraluminous rhyolitic liquids and phosphate saturation may be suppressed relative to metaluminous rhyolites. In peralkaline melts, phosphate solubility may increase as a result of phosphorus complexing with alkalis. The complexing stoichiometry may be variable, however, and the relative influence of peralkalinity versus temperature on phosphate solubility in rhyolitic melts deserves greater attention.     

Aluminum coordination in metaaluminous and peralkaline silicate melts

Incremental amounts of Na₂O and K₂O added to immiscible melts in the MgO-CaO-TiO₂-Al₂O₃ SiO₂ system cause a decrease in critical temperature, phase separation and change in the pattern of Al₂O₃ partitioning. Al₂O₃, which is concentrated in the low SiO₂ immiscible melts in the alkali-free system, is increasingly partitioned into the high-SiO₂ immiscible melt as the alkali/aluminium ratio is increased. However, K₂O is more effective than Na₂O in stabilizing Al₂O₂ in the SiO₂-rich melt. The coordination changes occurring in the aluminosilicate melts upon the addition of the alkali oxides are described by CaAl₂O₄+2SiOK=2KAlO₂+SiOCaOSi where K (or Na) displaces Ca as the charge-balancing cation for the networkforming AlO₄ tetrahedra. The increased stability of the AlO₄ species in the highly polymerized SiO₂-rich melt and the consequent shrinkage of the miscibility gap is ascribed to positive configurational entropy and negative enthalpy changes associated with the formation of K, Na-AlO₄ species. Element partition systematics indicate that (Na, K)AlO₂ species favor the more polymerized, CaAl₂O₄ and TiO₂ species, the less polymerized silicate structure in the melt.     

Liquid immiscibility between silicate,carbonate and sulfide melts in melt inclusions hosted in co-precipitated minerals from Kerimasi volcano (Tanzania): evolution of carbonated nephelinitic magma

The evolution of a carbonated nephelinitic magma can be followed by the study of a statistically significant number of melt inclusions, entrapped in co-precipitated perovskite, nepheline and magnetite in a clinopyroxene- and nepheline-rich rock (afrikandite) from Kerimasi volcano (Tanzania). Temperatures are estimated to be 1,100°C for the early stage of the melt evolution of the magma, which formed the rock. During evolution, the magma became enriched in CaO, depleted in SiO₂ and Al₂O₃, resulting in immiscibility at ~1,050°C and crustal pressures (0.5–1 GPa) with the formation of three fluid-saturated melts: an alkali- and MgO-bearing, CaO- and FeO-rich silicate melt; an alkali- and F-bearing, CaO- and P₂O₅-rich carbonate melt; and a Cu–Fe sulfide melt. The sulfide and the carbonate melt could be physically separated from their silicate parent and form a Cu–Fe–S ore and a carbonatite rock. The separated carbonate melt could initially crystallize calciocarbonatite and ultimately become alkali rich in composition and similar to natrocarbonatite, demonstrating an evolution from nephelinite to natrocarbonatite through Ca-rich carbonatite magma. The distribution of major elements between perovskite-hosted coexisting immiscible silicate and carbonate melts shows strong partitioning of Ca, P and F relative to Fe_T, Si, Al, Mn, Ti and Mg in the carbonate melt, suggesting that immiscibility occurred at crustal pressures and plays a significant role in explaining the dominance of calciocarbonatites (sövites) relative to dolomitic or sideritic carbonatites. Our data suggest that Cu–Fe–S compositions are characteristic of immiscible sulfide melts originating from the parental silicate melts of alkaline silicate–carbonatite complexes.     

High temperature calorimetric studies of heat of solution of NiO, CuO, La2O3, TiO2, HfO2 in sodium silicate liquids

The enthalpies of solution of La₂O₃, TiO₂, HfO₂, NiO and CuO were measured in sodium silicate melts at high temperature. When the heat of fusion was available, we derived the corresponding liquid-liquid enthalpies of mixing. These data, combined with previously published work, provide insight into the speciation reactions in sodium silicate melts. The heat of solution of La₂O₃ in these silicate solvents is strongly exothermic and varies little with La₂O₃ concentration. The variation of heat of solution with composition of the liquid reflects the ability of La(III) to perturb the transient silicate framework and compete with other cations for oxygen. The enthalpy of solution of TiO₂ is temperature-dependent and indicates that the formation of Na-O-Si species is favored over Na-O-Ti at low temperature. The speciation reactions can be interpreted in terms of recent spectroscopic studies of titanium-bearing melts which identify a dual role of Ti⁴⁺ as both a network-former end network-modifier. The heats of solution of oxides of transition elements (Ni and Cu) are endothermic, concentration-dependent and reach a maximum with concentration. These indicate a charge balanced substitution which diminishes the network modifying role of Na⁺ by addition of Ni²⁺ or Cu²⁺. The transition metal is believed to be in tetrahedral coordination, charge balanced by the sodium cation in the melts.     

Experiments on liquid immiscibility along tholeiitic liquid lines of descent

Crystallization experiments have been conducted on compositions along tholeiitic liquid lines of descent to define the compositional space for the development of silicate liquid immiscibility. Starting materials have 46–56 wt% SiO₂, 11.7–17.7 wt% FeO_tot, and Mg-number between 0.29 and 0.36. These melts fall on the basaltic trends relevant for Mull, Iceland, Snake River Plain lavas and for the Sept Iles layered intrusion, where large-scale liquid immiscibility has been recognized. At one atmosphere under anhydrous conditions, immiscibility develops below 1,000–1,020°C in all of these compositionally diverse lavas. Extreme iron enrichment is not necessary; immiscibility also develops during iron depletion and silica enrichment. Variations in melt composition control the development of silicate liquid immiscibility along the tholeiitic trend. Elevation of Na₂O + K₂O + P₂O₅ + TiO₂ promotes the development of two immiscible liquids. Increasing melt CaO and Al₂O₃ stabilizes a single-liquid field. New data and published phase equilibria show that anhydrous, low-pressure fractional crystallization is the most favorable condition for unmixing during differentiation. Pressure inhibits immiscibility because it expands the stability field of high-Ca clinopyroxene, which reduces the proportion of plagioclase in the crystallizing assemblage, thus enhancing early iron depletion. Magma mixing between primitive basalt and Fe–Ti–P-rich ferrobasalts can serve to elevate phosphorous and alkali contents and thereby promote unmixing. Water might decrease the temperature and size of the two-liquid field, potentially shifting the binodal (solvus) below the liquidus, leading the system to evolve as a single-melt phase.     

Behavior of molybdenum in alkali silicate melts: Effects of excess SiO2 and CO2

The effects of excess SiO₂ and CO₂ on the solubility of molybdenite in hydrous sodium disilicate melts were experimentally determined at 680 bars and 650°C. The molybdenite solubility decreases with increasing SiO₂ and CO₂. Under the experimental conditions, the MoS₂ content of the vapor-saturated liquid decreases from 10 wt.% to 2.5 wt.% at SiO₂ saturation. In the presence of CO₂, the solubility decreases to 4.6 wt.% MoS₂ and becomes negligible at high P_CO2. These results are explained as deriving from the increased polymerization and hence decreased NBO/Si ratio of the melt with increasing SiO₂ content and CO₂, respectively. Sulfur dissolves principally as SO₄^?2 at the relatively high fo₂ of the experiments. Consequently, the effect of sulfur is to lower the Mo solubility by effectively decreasing the NBO/Si ratio of the melt. Sulfur saturation is, therefore, likely to be a limiting factor in the Mo content of alkali silicate melts because of the chalcophile affinities of molybdenum.     

Na2O solubility in CaO-MgO-SiO2 melts

The sodium solubility in silicate melts in the CaO-MgO-SiO₂ (CMS) system at 1400 °C has been measured by using a closed thermochemical reactor designed to control alkali metal activity. In this reactor, Na_(g) evaporation from a Na₂O-xSiO₂ melt imposes an alkali metal vapor pressure in equilibrium with the molten silicate samples. Because of equilibrium conditions in the reactor, the activity of sodium-metal oxide in the molten samples is the same as that of the source, i.e., aNa₂O_(sample) = aNa₂O_(source). This design also allows to determine the sodium oxide activity coefficient in the samples. Thirty-three different CMS compositions were studied. The results show that the amount of sodium entering from the gas phase (i.e., Na₂O solubility) is strongly sensitive to silica content of the melt and, to a lesser extent, the relative amounts of CaO and MgO. Despite the large range of tested melt compositions (0 < CaO and MgO < 40; 40 < SiO₂ < 100; in wt%), we found that Na₂O solubility is conveniently modeled as a linear function of the optical basicity (Λ) calculated on a Na-free basis melt composition. In our experiments, γNa₂O_(sample) ranges from 7 × 10⁻⁷ to 5 × 10⁻⁶, indicating a strongly non-ideal behavior of Na₂O solubility in the studied CMS melts (γNa₂O_(sample) ? 1). In addition to showing the effect of sodium on phase relationships in the CMS system, this Na₂O solubility study brings valuable new constraints on how melt structure controls the solubility of Na in the CMS silicate melts. Our results suggest that Na₂O addition causes depolymerization of the melt by preferential breaking of Si-O-Si bonds of the most polymerized tetrahedral sites, mainly Q⁴.     

Structure of H2O-saturated peralkaline aluminosilicate melt and coexisting aluminosilicate-saturated aqueous fluid determined in-situ to 800 °C and ∼800 MPa

The structure of H₂O-saturated silicate melts and of silicate-saturated aqueous solutions, as well as that of supercritical silicate-rich aqueous liquids, has been characterized in-situ while the sample was at high temperature (to 800 °C) and pressure (up to 796 MPa). Structural information was obtained with confocal microRaman and with FTIR spectroscopy. Two Al-bearing glasses compositionally along the join Na₂O•4SiO₂-Na₂O•4(NaAl)O₂-H₂O (5 and 10 mol% Al₂O₃, denoted NA5 and NA10) were used as starting materials. Fluids and melts were examined along pressure-temperature trajectories of isochores of H₂O at nominal densities (from PVT properties of pure H₂O) of 0.85 g/cm³ (NA10 experiments) and 0.86 g/cm³ (NA5 experiments) with the aluminosilicate + H₂O sample contained in an externally-heated, Ir-gasketed hydrothermal diamond anvil cell.Molecular H₂O (H₂O°) and OH groups that form bonds with cations exist in all three phases. The OH/H₂O° ratio is positively correlated with temperature and pressure (and, therefore, fugacity of H₂O, f_H2O) with (OH/H₂O°)^melt > (OH/H₂O°)^fluid at all pressures and temperatures. Structural units of Q³, Q², Q¹, and Q⁰ type occur together in fluids, in melts, and, when outside the two-phase melt + fluid boundary, in single-phase liquids. The abundance of Q⁰ and Q¹ increases and Q² and Q³ decrease with f_H2O. Therefore, the NBO/T (nonbridging oxygen per tetrahedrally coordination cations), of melt is a positive function of f_H2O. The NBO/T of silicate in coexisting aqueous fluid, although greater than in melt, is less sensitive to f_H2O.The melt structural data are used to describe relationships between activity of H₂O and melting phase relations of silicate systems at high pressure and temperature. The data were also combined with available partial molar configurational heat capacity of Qⁿ-species in melts to illustrate how these quantities can be employed to estimate relationships between heat capacity of melts and their H₂O content.     

Liquid Immiscibility and the Evolution of Basaltic Magma

This experimental study examines relationships between alternativeevolution paths of basaltic liquids (the so-called Bowen andFenner trends), and silicate liquid immiscibility. Syntheticanalogues of natural immiscible systems exhibited in volcanicglasses and melt inclusions were used as starting mixtures.Conventional quench experiments in 1 atm gas mixing furnacesproved unable to reproduce unmixing of ferrobasaltic melts,yielding instead either turbid, opalescent glasses, or crystallizationof tridymite and pyroxenes. In contrast, experiments involvingin situ high-temperature centrifugation at 1000g (g = 9·8m/s²) did yield macroscopic unmixing and phase separation. Centrifugationfor 3–4 h was insufficient to complete phase segregation,and resulted in sub-micron immiscible emulsions in quenchedglasses. For a model liquid composition of the Middle Zone ofthe Skaergaard intrusion at super-liquidus temperatures of 1110–1120°C,centrifugation produced a thin, silicic layer (64·5 wt%SiO₂ and 7·4 wt% FeO) at the top of the main Fe-richglass (46 wt% SiO₂ and 21 wt% FeO). The divergent compositionsat the top and bottom were shown in a series of static runsto crystallize very similar crystal assemblages of plagioclase,pyroxene, olivine, and Fe–Ti oxides. We infer from theseresults that unmixing of complex aluminosilicate liquids maybe seriously kinetically hampered (presumably by a nucleationbarrier), and thus conventional static experiments may not correctlyreproduce it. In the light of our centrifuge experiments, immiscibilityin the Skaergaard intrusion could have started already at thetransition from the Lower to the Middle Zone. Thus, magma unmixingmight be an important factor in the development of the Fe-enrichmenttrend documented in the cumulates of the Skaergaard LayeredSeries. KEY WORDS: liquid immiscibility; Skaergaard; layered intrusions; experimental petrology     

Minor and trace element partitioning between immiscible ocelli-matrix pairs from lamprophyre dikes and sills,Monteregian Hills petrographic province,Quebec

Many lamprophyre dike and sill rocks in the Monteregian Hills petrographic province of southwestern Quebec contain felsic segregations (ocelli) which have been interpreted as globules of immiscible liquid (Philpotts 1976). Ocelli and matrix material were separated from a number of these rocks and analyzed for major and trace elements. The major element data, when plotted on a Greig diagram, outline a field of possible silicate-liquid immiscibility at higher alumina+alkali content than that previously mapped in iron-rich experimental systems. The trace element data support a liquid immiscibility hypothesis for the formation of these ocelli since high-charge density cations are preferentially concentrated in the matrix (mafic) material, a result which is consistent with theoretical and experimental studies.The distribution of minor and trace elements between ocelli and matrix indicates that several factors control the partitioning of these elements between immiscible felsic and mafic liquids. These factors include the difference in relative polymerization (as measured by the Si∶O ratio) of the two liquids, with an increase in this difference favoring partitioning of the high-charge density cations into the mafic liquid; the concentration of P₂O₅ in the mafic liquid which favors the partitioning of high-charge density cations into this liquid; the presence of a CO₂ vapor (?) phase which favors the partitioning of high-charge density cations into the CO₂ enriched phase; and the presence of solid phases at the onset of immiscibility. These observations indicate that the chemical compositions of two possibly immiscible melts should be known if minor and trace element data are to be used as evidence for silicate-liquid immiscibility.     

Two-liquid partition coefficients: Experimental data and geochemical implications

Partition coefficients for Cs, Ba, Sr, Ca, Mg, La, Sm, Lu, Mn, Ti, Cr, Ta, Zr, and P between immiscible basic and acidic liquids in the system K₂O-Al₂O₃-FeO-SiO₂ were experimentally determined at 1,180 °C and 1 atm. Phosphorus is most strongly enriched in the basic melt (by a factor of 10), followed by rare earth elements, Ta, Ca, Cr, Ti, Mn, Zr, Mg, Sr, and Ba (enriched by a factor of 1.5). Of the elements studied, only Cs is enriched in the acidic melt. The two-liquid partition coefficients of Zr, Ta, Sm, and Mn are constant for concentrations ranging from <0.1% to as high as 1 wt.-%, suggesting that Henry's law is applicable in silicate melts (at least for these elements) to concentrations well above typical trace element levels in rocks. The strong relative preference of many elements for the basic melt implies that the structural characteristics of basic melts more readily permit stable coordination of cations by oxygen. Partitioning of elements between crystal and liquid in a magma must therefore be influenced by the composition (and consequent structure) of the liquid.Application of the two-liquid partition coefficients to possible occurrences of liquid immiscibility in magmas reveals that typical basalt-rhyolite associations are probably not generated by two-liquid phase separation. However, liquid immiscibility cannot be discounted as a possible origin for lamprophyric rocks containing felsic segregations.     

Liquid immiscibility in basic alkaline magmas

On the assumption that very basic silicate magmas contain appreciable amounts of O^2? ions, non-ideal mixing of the melt species may be due to strongly preferred ionic associations. The anions in the melt, O^2? on the one hand, and SiO₄^4? and polymerized aluminosilicate ionic species on the other, discriminate between the cations according to their field strength and polarizing power. K⁺, Na⁺, Ba²⁺, Sr²⁺ and Ca²⁺ associate with singly bonded oxygen, and Mg²⁺, Fe²⁺ and Ni²⁺ attach themselves to O^2? rather than to silicate or aluminosilicate anions. In extreme cases these relations may lead to unmixing in the silicate melt, especially in the presence of water which may lower the liquidus relative to the solvus temperature. Liquid—liquid phase separation in ultrabasic magmas may be related to the present model.     

Minor and trace element partitioning between immiscible ocelli-matrix pairs from lamprophyre dikes and sills,Monteregian Hills petrographic province,Quebec

Many lamprophyre dike and sill rocks in the Monteregian Hills petrographic province of southwestern Quebec contain felsic segregations (ocelli) which have been interpreted as globules of immiscible liquid (Philpotts 1976). Ocelli and matrix material were separated from a number of these rocks and analyzed for major and trace elements. The major element data, when plotted on a Greig diagram, outline a field of possible silicate-liquid immiscibility at higher alumina+alkali content than that previously mapped in iron-rich experimental systems. The trace element data support a liquid immiscibility hypothesis for the formation of these ocelli since high-charge density cations are preferentially concentrated in the matrix (mafic) material, a result which is consistent with theoretical and experimental studies. The distribution of minor and trace elements between ocelli and matrix indicates that several factors control the partitioning of these elements between immiscible felsic and mafic liquids. These factors include the difference in relative polymerization (as measured by the Si∶O ratio) of the two liquids, with an increase in this difference favoring partitioning of the high-charge density cations into the mafic liquid; the concentration of P₂O₅ in the mafic liquid which favors the partitioning of high-charge density cations into this liquid; the presence of a CO₂ vapor (?) phase which favors the partitioning of high-charge density cations into the CO₂ enriched phase; and the presence of solid phases at the onset of immiscibility. These observations indicate that the chemical compositions of two possibly immiscible melts should be known if minor and trace element data are to be used as evidence for silicate-liquid immiscibility.     

Experiments on liquid immiscibility in silicate melts with H2O, P, S, F and Cl: implications for natural magmas

Isobaric (200 MPa) experiments have been performed to investigate the effects of H₂O alone or in combination with P, S, F or Cl on liquid-phase separation in melts in the systems Fe₂SiO₄–Fe₃O₄–KAlSi₂O₆–SiO₂, Fe₃O₄–KAlSi₂O₆–SiO₂ and Fe₃O₄–Fe₂O₃–KAlSi₂O₆–SiO₂ with or without plagioclase (An₅₀). Experiments were heated in a rapid-quench internally heated pressure vessel at 1,075, 1,150 or 1,200 °C for 2 h. Experimental fO₂ was maintained at QFM, NNO or MH oxygen buffers. H₂O alone or in combination with P, S or F increases the temperature and composition range of two-liquid fields at fO₂ = NNO and MH buffers. P, S, F and Cl partition preferentially into the Fe-rich immiscible liquid. Two-liquid partition coefficients for Fe, Si, P and S correlate well with the degree of polymerization of the SiO₂-rich liquid and plot on similar but distinct power-law curves compared with equivalent anhydrous or basaltic melts. The addition of 2 wt% S to the system Fe₃O₄–Fe₂O₃–KAlSi₂O₆–SiO₂ stabilizes three immiscible melts with Fe-, FeS- and Si-rich compositions. H₂O-induced suppression of liquidus temperatures in the experimental systems, considered with the effects of pressure on the temperature and composition ranges of two-liquid fields in silicate melts, suggests that liquid-phase separation may be stable in some H₂O-rich silicate magmas at pressures in excess of 200 MPa.