Liquid immiscibility in deep-seated magmas and the generation of carbonatite melts

This paper reviews the results of investigations of melt inclusions in minerals of carbonatites and spatially associated silicate rocks genetically related to various deep-seated undersaturated silicate magmas of alkaline ultrabasic, alkaline basic, lamproitic, and kimberlitic compositions. The analysis of this direct genetic information showed that all the deep magmas are inherently enriched in volatile components, the most abundant among which are carbon dioxide, alkalis, halides, sulfur, and phosphorus. The volatiles probably initially served as agents of mantle metasomatism and promoted melting in deep magma sources. The derived magmas became enriched in carbon dioxide, alkalis, and other volatile components owing to the crystallization and fractionation of early high-magnesium minerals and gradually acquired the characteristics of carbonated silicate liquids. When critical compositional parameters were reached, the accumulated volatiles catalyzed immiscibility, the magmas became heterogeneous, and two-phase carbonate-silicate liquid immiscibility occurred at temperatures of ≥1280–1250°C. The immiscibility was accompanied by the partitioning of elements: the major portion of fluid components partitioned together with Ca into the carbonate-salt fraction (parental carbonatite melt), and the silicate melt was correspondingly depleted in these components and became more silicic. After spatial separation, the silicate and carbonate-silicate melts evolved independently during slow cooling. Differentiation and fractionation were characteristic of silicate melts. The carbonatite melts became again heterogeneous within the temperature range from 1200 to 800–600°C and separated into immiscible carbonate-salt fractions of various compositions: alkali-sulfate, alkali-phosphate, alkali-fluoride, alkali-chloride, and Fe-Mg-Ca carbonate. In large scale systems, polyphase silicate-carbonate-salt liquid immiscibility is usually manifested during the slow cooling and prolonged evolution of deeply derived melts in the Earth’s crust. It may lead to the formation of various types of intrusive carbonatites: widespread calcite-dolomite and rare alkali-sulfate, alkali-phosphate, and alkali-halide rocks. The initial alkaline carbonatite melts can retain their compositions enriched in P, S, Cl, and F only at rapid eruption followed by instantaneous quenching.     

General characteristics of the geochemical evolution of silicic melts during the development of massive sulfide magmatic ore systems: Evidence from the investigation of melt inclusions

The investigation of melt inclusions in the minerals of volcanic rocks from the massive sulfide deposits of Siberia and the Urals revealed some specific features in the development of their magmatic ore systems. It was shown that the petrochemical and rare earth element compositions of melt inclusions reflect the geodynamic conditions of their formation: island arc conditions for the massive sulfide deposits of Rudny Altai, eastern Tuva, and the Salair Range and a back arc basin environment for the Yaman-Kasy deposit. The silicic melts of inclusions from the volcanic rocks of massive sulfide deposits show some specific features with respect to the contents of volatile components. In all of the ore deposits studied, fluorine content was always low (0.03–0.08 wt %), whereas chlorine content (0.13–0.28 wt %) was higher than the average value for silicic melts (0.17 wt %). There is a strong differentiation of water content in melt inclusions, both between deposits and between various volcanics from a single deposit. Ore-bearing melts show the highest water contents of 3.34–4.07 wt %. High Cu contents in the silicic melts of the Yubileinoe and Kyzyl-Tashtyg deposits (up to 7118 and 3228 ppm, respectively) may indicate the affinity of some ore components to particular silicic magmas. This is supported by the elevated contents of Cu in the porphyry Cu deposits of Romania (Valea Morii), Mongolia (Bayan Ula), and Bolivia. On the other hand, the silicic melts of inclusions from the molybdenum-uranium deposit of the Strel’tsovka ore field show high contents of another group of ore components (U and F).     

Experimental study of fluorine and chlorine contents in mica (biotite) and their partitioning between mica,phonolite melt,and fluid

This paper reports the results of a study of the composition of mica (biotite) crystallizing in the system of phonolite melt-Cl- and F-bearing aqueous fluid at T ～ 850°C, P = 200 MPa, and \(f_{O_2 } \) = Ni-NiO, as well as data on F and Cl partitioning between coexisting phases. It was established that Cl content in mica is significantly lower than in phonolite melt and, especially, in fluid. Fluorine shows a different behavior in this system: its content in mica is always higher than in phonolite melt but lower than in fluid. The mica-melt partition coefficients of Cl and F also behave differently. The Cl partition coefficient gradually increases from 0.17 to 0.33 with increasing Cl content in the system, whereas the partition coefficient of F sharply decreases from 3.0 to 1.0 with increasing total F content. The apparent partition coefficients of F between biotite and groundmass (melt) in various magmatic rocks are usually significantly higher than the experimental values. It was supposed that the higher ^Bt/glassD_F values in natural samples could be related to the influence of later oxidation reactions, reequilibration of biotite at continuously decreasing \(f_{H_2 O} \)/f _HF ratio, and an increase in this coefficients with decreasing total F content in the system.     

Inclusions of silicate and sulfate melts in chrome diposide from the Inagli deposit,Yakutia, Russia

Melt inclusions were studied in chrome diopside from the Inagli deposit of gemstones in the Inagli massif of alkaline ultrabasic rocks of potassic affinity in the northwestern Aldan shield, Yakutia, Russia. The chrome diopside is highly transparent and has an intense green color. Its Cr₂O₃ content varies from 0.13 to 0.75 wt %. Primary and primary-secondary polyphase inclusions in chrome diopside are dominated by crystal phases (80–90 vol %) and contain aqueous solution and a gas phase. Using electron microprobe analysis and Raman spectroscopy, the following crystalline phases were identified. Silicate minerals are represented by potassium feldspar, pectolite [NaCa₂Si₃O₈(OH)], and phlogopite. The most abundant minerals in the majority of inclusions are sulfates: glaserite (aphthitalite) [K₃Na(SO₄)₂], glauberite [Na₂Ca(SO₄)₂], aluminum sulfate, anhydrite (CaSO₄), gypsum (CaSO₄ × 2H₂O), barite (BaSO₄), bloedite [Na₂Mg(SO₄)₂ × 4H₂O], thenardite (NaSO₄), polyhalite [K₂Ca₂Mg(SO₄)₄ × 2H₂O], arcanite (K₂SO₄), and celestite (SrSO₄). In addition, apatite was detected in some inclusions. Chlorides are probably present among small crystalline phases, because some analyses of aggregates of silicate and sulfate minerals showed up to 0.19–10.3 wt % Cl. Hydrogen was identified in the gas phase of polyphase inclusions by Raman spectroscopy. The composition of melt from which the chrome diopside crystallized was calculated on the basis of the investigation of silicate melt inclusions. This melt contains 53.5 wt % SiO₂, considerable amounts of CaO (16.3 wt %), K₂O (7.9 wt %), Na₂O (3.5 wt %), and SO₃ (1.4 wt %) and moderate amounts of Al₂O₃ (7.5 wt %), MgO (5.8 wt %), FeO (1.1 wt %), and H₂O (0.75 wt %). The content of Cr₂O₃ in the melt was 0.13 wt %. Many inclusions were homogenized at 770–850°C, when all of the crystals and the gas phase were dissolved. The material of inclusions heated up to the homogenization temperature became heterogeneous even during very fast quenching (two seconds) producing numerous small crystals. This fact implies that most of the inclusions contained a salt (rather than silicate) melt of sulfate-dominated composition. Such inclusions were formed from salt globules (with a density of about 2.5 g/cm³) occurring as an emulsion in the denser (2.6 g/cm³) silicate melt from which the chrome diopside crystallized.     

Metamorphic rocks of the Murmansk domain (Kola Peninsula) as compared with the oldest rock associations of the northeastern Baltic shield,Canada, and Greenland

Rock associations characterized by heterogeneous sets of petrogeochemical parameters were compared by quantifying the degree of their similarity-dissimilarity and searching for discrimination trends between them. Using procedures specially developed for this purpose, it was demonstrated for the first time that the lithotectonic complexes of the Murmansk domain are fundamentally different from those of typical granulite-gneiss terrains and resemble the granite-greenstone terrains of the Baltic shield, Greenland, and Canada. Based on the whole data set, the Murmansk domain can be considered as a deeply eroded Archean granite-greenstone terrain retaining only the tonalite-trondhjemite-gneiss basement with abundant supracrustal enclaves. A trend of the compositional difference between the older and younger rock associations is similar to that between the tholeiitic and boninitic volcanic series. It was suggested that the petrogeochemical “age” trend reflects the initial stage of the compositional evolution of the metamagmatic rocks of the region from metamorphic rocks similar to tholeiites at the early stages to the Paleoproterozoic boninite-like rocks, which are believed to be linked to the unique PGE-bearing province of the northeastern Baltic shield. This implies that the specific metallogenic features of the region emerged already in the Archean, which supported the suggestion on the long duration of geological processes in the Early Precambrian.     

Conditions of Quaternary magmatism at Spitsbergen Island

Petrological and geochemical data obtained on the Quaternary lavas of volcanoes at Spitsbergen Island indicate that the rocks were produced via the deep-seated crystallization of parental alkaline magmas at 8–10 kbar. The character of clinopyroxene enrichment in incompatible elements indicates that the mineral crystallized from more enriched melts than those inferred from the composition of the host lavas. These melts were close to the parental melts previously found as veinlets in mantle hyperbasite xenoliths in the lavas. According to the character of their enrichment in Pb and Sr radiogenic isotopes and depletion in Nd, the basalts from Spitsbergen Island define a single trend with the weakly enriched tholeiites of the Knipovich Ridge, a fact suggesting the closeness of the enriched sources beneath the continental margin of Spitsbergen and beneath the spreading zone. Magmatic activity at Spitsbergen was related to the evolution of the Norwegian-Greenland basin, which evolved in pulses according to the shift of the spreading axes. The most significant of the latter events took place in the Neogene, when the Knipovich Ridge obtained its modern position near the western boundary of Spitsbergen. Early in the course of the evolution, the emplacement of alkaline melts generated at Spitsbergen into the oceanic mantle could form the enriched mantle, which was later involved in the melting process beneath the spreading zone.     

Zircon coronas around Fe–Ti oxides: a physical reference frame for metamorphic and metasomatic reactions

Ilmenite in coronitic gabbros from the Bamble and Kongsberg sectors, southern Norway, is surrounded by zircons ranging in diameters from a fraction of a micrometer to 10 μm across. The zircons are inert during subsequent metamorphism (amphibolite- to pumpellyite–prehnite facies) and metasomatism (scapolitization and albitization) and can be found as trails in silicates (phlogopite, talc, chlorite, amphibole, albite, and tourmaline) in the altered rocks. The trails link up to form polygons outlining the former oxide grain boundary. This 3-dimensional framework of zircons is used to (a) recognize metasomatic origin of rocks, (b) quantify the mobility of elements during mineral replacement, (c) establish the growth direction of reaction fronts and to identify the reaction mechanism as dissolution–reprecipitation. Zircon coronas on Fe–Ti oxides have been described from a number of terrains and appear to be common in mafic rocks (gabbros and granulites) providing a tool for a better understanding of metasomatic and metamorphic reactions.     

Application of precise 142Nd/144Nd analysis of small samples to inclusions in diamonds (Finsch,South Africa) and Hadean Zircons (Jack Hills,Western Australia)

¹⁴⁶Sm–¹⁴²Nd and ¹⁴⁷Sm–¹⁴³Nd systematics were investigated in garnet inclusions in diamonds from Finsch (S. Africa) and Hadean zircons from Jack Hills (W. Australia) to assess the potential of these systems as recorders of early Earth evolution. The study of Finsch inclusions was conducted on a composite sample of 50 peridotitic pyropes with a Nd model age of 3.3 Ga. Analysis of the Jack Hills zircons was performed on 790 grains with ion microprobe ²⁰⁷Pb/²⁰⁶Pb spot ages from 3.95 to 4.19 Ga. Finsch pyropes yield 100 × ?¹⁴²Nd = ? 6 ± 12 ppm, ?¹⁴³Nd = ? 32.5, and ¹⁴⁷Sm/¹⁴⁴Nd = 0.1150. These results do not confirm previous claims for a 30 ppm ¹⁴²Nd excess in South African cratonic mantle. The lack of a ¹⁴²Nd anomaly in these inclusions suggests that isotopic heterogeneities created by early mantle differentiation were remixed at a very fine scale prior to isolation of the South African lithosphere. Alternatively, this result may indicate that only a fraction of the mantle experienced depletion during the first 400 Myr of its history. Analysis of the Jack Hills zircon composite yielded 100 × ?¹⁴²Nd = 8 ± 10 ppm, ?¹⁴³Nd = 45 ± 1, and ¹⁴⁷Sm/¹⁴⁴Nd = 0.5891. Back-calculation of this present-day ?¹⁴³Nd yields an unrealistic estimate for the initial ?¹⁴³Nd of ? 160 ?-units, clearly indicating post-crystallization disturbance of the ¹⁴⁷Sm–¹⁴³Nd system. Examination of ^146,147Sm–^142,143Nd data reveals that the Nd budget of the Jack Hills sample is dominated by non-radiogenic Nd, possibly contained in recrystallized zircon rims or secondary subsurface minerals. This secondary material is characterized by highly discordant U–Pb ages. Although the mass fraction of altered zircon is unlikely to exceed 5–10% of total sample, its high LREE content precludes a reliable evaluation of ¹⁴⁶Sm–¹⁴²Nd systematics in Jack Hills zircons.