A. V. Girnis

The system MgO-Al2O3-Cr2O3 revisited reanalysis of Doroshev et al.'s (1997) experiments and new experiments

New experimental data are presented on the partitioning of Cr and Al between spinel-picrochromite and corundumeskolaite solid solutions at 5.0 GPa and 1300, 1500°C and garnet-spinel-orthopyroxene equilibria at 3.0 and 4.0 GPa and 1000°C. The experimental products of Doroshev et al. 9s (1997) experiments on mineral equilibria in the MASCr system were reanalysed in order to clarify the direction and mechanism of approach to equilibrium. Detailed analysis of mineral compositions was carried out by electron microprobe with high spatial resolution, which demonstrated complex zoning patterns of minerals. In experiments with pure pyrope in the starting material, garnet shows simple zoning with Cr-rich seams mantling relicts of aluminous garnet. In experiments with Cr-rich garnets in the starting mixture (Cr/(Cr+Al) = 0.3–0.8), more complex patterns were observed. Knorringiterich garnet cores are overgrown in sequence by Al-rich (X Cr = 0.1–0.2) and then Cr-rich zones. The analytical data allowed us to establish the direction of approach to equilibrium for particular grains, which eliminated the path-looping problem. The results were compared to published experimental data on garnet, orthopyroxene, spinel, and eskolaite equilibria in the MAS and MASCr systems. It was shown that all the data can be described by a simple thermodynamic model, if the uncertainty of analytical results is taken into account.

Average compositions of igneous melts from main geodynamic settings according to the investigation of melt inclusions in minerals and quenched glasses of rocks

We compiled a database containing more than 480000 determinations for 73 elements in melt inclusions in minerals and quenched glasses of volcanic rocks. These data were used to estimate the mean contents of major, volatile, and trace elements in igneous melts from main geodynamic settings. The following settings were distinguished: (I) oceanic spreading zones (mid-ocean ridges); (II) zones of mantle plume activity on oceanic plates (oceanic islands and plateaus); (III) and (IV) settings related to subduction processes, including (III) zones of island-arc magmatism generated on the oceanic crust and (IV) magmatic zones of active continental margins involving the continental crust into magma generation processes; (V) intracontinental rifts and continental hot spots; and (VI) back-arc spreading centers. The histogram of SiO2 contents in the natural igneous melts of all geodynamic settings exhibits a bimodal distribution with two maxima at SiO2 contents of 50–52 wt % and 72–74 wt %. The range 62–64 wt % SiO2 comprises the minimum number of determinations. Primitive mantle-normalized spidergrams were constructed for average contents of elements in the igneous melts of basic, intermediate, and acidic compositions from settings I–V. The diagrams reflect the characteristic features of melt compositions for each geodynamic setting. On the basis of the analysis of data on the composition of melt inclusions and glasses of rocks, average ratios of incompatible trace and volatile components (H2O/Ce, K2O/Cl, Nb/U, Ba/Rb, Ce/Pb, etc.) were estimated for the igneous melts of all of the settings. Variations of these ratios were determined, and it was shown that, in most cases, the ratios of incompatible elements are significantly different between settings. The difference is especially pronounced for the ratios of elements with different degrees of incompatibility (e.g., Nb/Yb) and for some ratios with volatile components (e.g., K2O/H2O).

Average compositions of magmas and mantle sources of mid-ocean ridges and intraplate oceanic and continental settings estimated from the data on melt inclusions and quenched glasses of basalts

Based on the generalization of the compositions of melt inclusions and quenched glasses from basaltic rocks, the average compositions of magmas were estimated for mid-ocean ridges (MOR), intraplate continental environments (CR), and ocean islands and plateaus (OI). These compositions were used to constrain the average contents of trace and volatile elements in mantle sources. A procedure was developed for the estimation of the average contents of incompatible elements, including volatiles (H2O, Cl, F, and S), in the mantle. A comparison of the obtained average contents for the depleted mantle (DM) with the available published estimates showed that the contents of most incompatible trace elements (H2O, Cl, F, Be, B, Rb, Sr, Zr, Ba, La, Ce, Nd, Sm, Eu, Hf, Ta, Th, and U) can be reliably estimated from the ratio of K to the desired trace element in the MOR magmas and the average content of K in the DM. For Nb, Ti, P, S, Li, Y, and heavy REE, we used the ratios of their contents to an element with a similar degree of incompatibility in MOR magmas (U for Nb and Dy for the other elements). This approach was used to determine the average contents of incompatible elements in oceanic plume mantle (OPM) and the subcontinental mantle of intraplate settings or continental plume mantle (CPM). It was shown that the average composition of both suboceanic and subcontinental mantle plumes is moderately enriched compared with the DM in the most incompatible elements (K, U, Ba, and La) and volatile components (H2O, Cl, and F). The extent of volatile component enrichment in the plume mantle (500–1500 ppm H2O) is insufficient for a significant depression of the mantle solidus. Therefore, mantle plumes must be hotter than the ambient depleted mantle. The average contents of incompatible trace elements in the OPM are similar to those of the primitive mantle, which could be related either to the retention of primitive mantle material in the regions of plume generation or to DM fertilization at the expense of the deep mantle recycling of crustal materials. In the latter case, the negative anomaly of water in the trace-element distribution patterns of the OPM is explained by the participation of dehydrated crust in its formation. Variations in the compositions of magmas and their sources were considered for various geodynamic settings, and it was shown that the sources are heterogeneous with respect to trace and volatile components. The chemical heterogeneity of the magma sources and gradual transitions between them suggest that the mantle reservoirs interact with each other. Chemical variations in continental and oceanic plume magmas can be attributed to the existence of several interacting sources, including one depleted and at least two enriched reservoirs with different contents of volatiles. These variations are in agreement with the zoned structure of mantle plumes, which consist of a hot and relatively dry core, a colder outer shell with high contents of volatile components, and a zone of interaction between the plume and depleted mantle.

Average composition of basic magmas and mantle sources of island arcs and active continental margins estimated from the data on melt inclusions and quenched glasses of rocks

Based on the generalization of data on melt inclusions and quenched glasses, the average compositions of subduction (island arc and active continental margin settings) basic magmas were estimated. The main geochemical features of the average composition of these magmas are significant depletion in Nb and Ta, less significant depletion in Ti, Zr, and Sm, and enrichment in Cl, H2O, F, and P in the primitive mantlenormalized patterns. The average normalized contents of moderately incompatible HREE in these magmas are close to those in the basic magmas of other geodynamic settings. Subduction basic magmas exhibit negative correlation of Li, Y, Dy, Er, Yb, Lu, and Ti contents with MgO content. Most of incompatible elements (Nb, Ta, U, Th, LREE) do not correlate with MgO, but correlate with each other and K2O. Variations in element contents are related to crystallization differentiation, magma mixing, and possibly, participation of several sources. The water content in the island arc basic magmas varies from almost zero value to more than 6 wt %. Most compositions are characterized by weak negative correlation between H2O and MgO contents, but some compositions define a negative correlation close to that in magmas of mid-ocean ridges (MOR). Considered magmas demonstrate distinct positive correlation between MgO content and homogenization temperature, practically coinciding with that of MOR magmas. Modeling of phase equilibria revealed widening of crystallization field of olivine in the magmas of subduction zones compared to MOR magmas. This can be related to the high water content in subduction magmas. Simultaneous liquidus crystallization of olivine and clinopyroxene in subduction magmas occurs at pressure approximately 5 kbar higher than that of MOR magmas. Based on the average ratios of trace element to K2O content, we determined the average compositions for subduction magma sources. Relative to depleted mantle, they are enriched in all incompatible elements, with positive anomalies of U, Rb, Ba, B, Pb, Cl, H2O, F, and S, and negative anomalies of Th, K, Be, Nb, Ta, Li, Nd, Pb, and Ti. A general elevated content of incompatible elements indicates a reworking of the rocks of mantle wedge by fluids and melts that were released from the upper layers of subducted plate.

Composition and chemical structure of oceanic mantle plumes

The average compositions (including H2O, Cl, F, and S contents) and chemical structure of oceanic mantle plumes were estimated on the basis of the ratios of incompatible volatile components, potassium, and some other elements in the basaltic magmas of ocean islands (melt inclusions and quenched glasses). The following average concentrations were estimated for the plume mantle: 510 ppm K2O, 520 ppm H2O, 21 ppm Cl, 55 ppm F, and 83 ppm S; these values are significantly higher than those of the depleted mantle (except for S). The abundances of H2O, Cl, and S are lower than in the primitive mantle. The normalized H2O content in the plume mantle is similar to the concentrations of similarly incompatible La and Ce but lower than the concentrations of K2O, Cl, and Sr. This is at odds with the idea of wet mantle plumes. Three types of basaltic magmas corresponding to three types of plume sources (M1, M2, and M3) were distinguished. The concentrations of incompatible elements in these reservoirs were estimated using two models, assuming either an isochemical mantle or a moderately enriched composition of plume material. The latter model gave the following average concentrations of H2O, Cl, F, and S: 130, 33, 11, and 110 ppm for M1, 110, 12, 65, and 45 ppm for M2; 530, 29, 49, and 110 ppm for M3. The plume mantle is not homogeneous, and its heterogeneity is related to the existence of three main compositions, one of which (M1) is similar to the mantle of mid-ocean ridges, and two others (M2 and M3) are moderately enriched in K2O, TiO2, P2O5, F, and incompatible trace elements. The compositions of M2 and M3 are strongly different in H2O, Cl, and S contents. The M2 mantle reservoir is significantly poorer in these components and richer in incompatible trace elements than M3. The plume mantle was formed mainly by the mixing of three sources: ultradepleted mantle, moderately enriched relatively dry mantle, and moderately enriched H2O-rich mantle. In addition to the three main components of the plume mantle, there are probably minor components enriched in chlorine and depleted in fluorine. It is supposed that all these components are entrained into the plume mantle through the mantle recycling of components of the oceanic and continental crust. The established relationships are in agreement with the zonal model of a mantle plume, which includes a hot central part poor in H2O, Cl, and S; an outer part enriched in volatile and nonvolatile incompatible elements; and enclosing mantle material interacting with the plume.

Volatiles in basaltic magmas of ocean islands and their mantle sources: I. Melt compositions deduced from melt inclusions and glasses in the rocks

Statistical analysis of a data bank of the compositions of glasses and melt inclusions in minerals from ocean-island basalts. The initial database contains more than 45 000 published analyses of ocean-island igneous rocks from around the world. Much attention was given to the contents of volatiles (H2O, Cl, F, and S) and their ratios to one another and to nonvolatile components of close incompatibility (Ti, P, K, and Ce). The average compositions of melt inclusions are similar to those of glasses of the rocks, including volatiles, with consideration for a somewhat higher degree (by approximately 20%) of the differentiation of glasses. The average compositions of ocean-island melts differ from those of mid-ocean basalts in having wider variations and elevated contents of some of the most incompatible elements (Sr, Nb, Ta, Ba, U, Th, and others), as well as H2O, F, and Cl. Based on the correlation of volatiles to one another and to incompatible elements, three groups of ocean-island basalts are distinguished: (I) low-K, P, Ti magma compositions approximating mid-ocean ridge magmas, (II) high-K, Ce, P, and Ti magmas that resemble continental rift magmas but differ from them in low H2O content, and (III) high-K, H2O, Ce, P, and Ti magmas close to continental rift magma. All three types of the melts were found only in the Hawaiian Archipelago, whereas other ocean islands are dominated by any one of these types. The distinguished melt types presumably reflect the differences (heterogeneity) in the compositions of the sources.

Mechanisms of formation of barium-rich phlogopite and strontium-rich apatite during the final stages of alkaline magma evolution

Carbonatite veinlets in fergusite from the Dunkeldyk potassium-rich basaltoid complex (southeastern Pamirs) are composed of clinopyroxene, phlogopite, and apatite phenocrysts embedded in a crystallized calcite-bearing groundmass. The examination of back-scattered electron images revealed areas of significantly different compositions in fluorapatite and fluorphlogopite. The content of BaO in the phlogopite ranges from 0.68 to 10.9 wt %. There are also variations in MgO and F contents. The maximum BaO content corresponds to high mole fractions of the Ba end member kinoshitalite (up to 0.24) in the phlogopite. The zoned fluorapatite phenocrysts are rich in SrO (0.77–25.4 wt %). An increase in SrO content is accompanied by an increase in Ce2O3, La2O3, and BaO and a distinct decrease in CaO. Most of the apatite grains are rimmed by elongated colorless crystals showing the maximum SrO contents. Based on the experimentally determined Ba and Sr partition coefficients between these minerals, silicate and carbonate melts, and fluid, a model was proposed for the enrichment of phases in these trace elements. It was shown that the mineral-forming media of the Ba-rich phlogopites was a residual melt enriched in volatiles (including F) and fluid-mobile elements. During that stage, the decomposition reactions of early Ba-bearing feldspars with subsequent incorporation of BaO in Ba-rich phlogopites played an important role. The mechanism of formation of Sr-rich apatites is fundamentally different: early apatite grains with moderate Sr contents recrystallized under the influence of Sr-rich fluids released during the late magmatic stage. Thus, despite their close association in a single rock, the Ba-bearing phlogopite and Sr-rich apatite were formed by significantly different mechanisms. Our previous investigations of melt and fluid inclusions in minerals from the rocks of the Dunkeldyk complex and the results obtained in this study allowed us to suggest that the barium, fluorite-carbonatite, and rare metal mineralization occurring in the region developed owing to the prolonged evolution of primary magmas, resulting in the formation of melt-solutions (brines) and hydrothermal systems.

Peralkaline silicic melts of island arcs, active continental margins, and intraplate continental settings: Evidence from the investigation of melt inclusions in minerals and quenched glasses of rocks

Based on the analysis of data on the composition of melt inclusions in minerals and quenched glasses of igneous rocks, we considered the problems of the formation of peralkaline silicic magmas (i.e., whose agpaitic index, the molar ratio AI = (Na2O + K2O)/Al2O3, is higher than one). The mean compositions of peralkaline silicic melts are reported for island arcs and active continental margins and compared with the compositions of melts from other settings, primarily, intraplate continental areas. Peralkaline silicic rocks are rather common in the latter. Such rocks are rare in island arcs and active continental margins, but agpaitic melts were observed in inclusions in phenocrysts of plagioclase, quartz, pyroxene, and other minerals. Plagioclase fractionation from an alkali-rich melt with AI < 1 is considered as a possible mechanism for the formation of peralkaline silicic melts (Bowen’s plagioclase effect). However, the analysis of available experimental data on plagioclase-melt equilibria showed that natural peralkaline melts are almost never in equilibrium with plagioclase. For the same reason, the melting of the majority of crustal rocks, which usually contain plagioclase, does not produce peralkaline melts. The existence of peralkaline silicic melt inclusions in plagioclase phenocrysts suggests that plagioclase can crystallize from peralkaline melts, and the plagioclase effect may play a certain role. Another mechanism for the formation of peralkaline silicic magmas is the melting of alkali-rich basic and intermediate rocks, including the spilitized varieties of subalkali basalts.

Origin of carbonatite magma during the evolution of ultrapotassic basite magma

Clinopyroxene phenocrysts in fergusite from a diatreme in the Dunkel’dyk potassic alkaline complex in the southeastern Pamirs, Tajikistan, and from carbonate veinlets cutting across this rock contain syngenetic carbonate, silicate, and complex melt inclusions. The homogenization of the silicate and carbonate material of the inclusions with the complete dissolution of daughter crystalline phases and fluid in each of them occur simultaneously at 1150−1180°C. The pressures estimated using fluid inclusions and mineral geobarometers were 0.5–0.7 GPa. The behavior of the inclusions during their heating and their geochemistry are in good agreement with the origin of carbonate melts via liquid immiscibility. Carbonatite magma was segregated at the preservation of volatile components (H2O, CO2, F, Cl, and S) in the melt, and this resulted in the crystallization of H2O-rich minerals and carbonates and testifies that the magma was not intensely degassed during its ascent to the surface. The silicate melts are rich in alkalis (up to 4 wt % Na2O and 12 wt % K2O), H2O, F, Cl, and REE (up to 1000 ppm), LREE, Ba, Th, U, Li, B, and Be. The diagrams of the concentrations of incompatible elements of these rocks typically show deep Nb, Ta, and Ti minima, a fact making them similar to the unusual type of ultrapotassic magmas: lamproites of the Mediterranean type. These magmas are thought to be generated in relation to subduction processes, first of all, the fluid transport of various components from a down-going continental crustal slab into overlying levels of the mantle wedge, from which ultrapotassic magmas are presumably derived.

Fluoride and chloride melts included in phenocrysts of agpaitic acid volcanic rocks from Pantelleria Island

978 Many deposits of rare elements (Be, Zr, Nb, REE, Y, and others) are genetically related to acid agpaitic (with the mole ratio (Na2O + K2O)/Al2O3 > 1) mag� mas [1, 2]. The concentration of these elements most likely results from crystallization differentiation lead� ing to their accumulation in the residual melts at the expense of low coefficients of minor element parti� tioning between silicate minerals and melts [3, 4]. Gradual accumulation of incoherent minor elements in melts may be violated by separation of saline melts from magma. According to the experimental data [5], minor elements are most effectively extracted by fluo� ride melts. Because of this, silicate–saline liquid immiscibility is an important factor of geochemical and metallogenic evolution of magmas. The formation of chloride melts during the evolu� tion of magmatic systems was established for some objects [6 and others]. The presence of fluoride melts in acid magmatic systems was described in individual cases [7, 8], and each new finding is of intense interest. In this paper we discuss the results of the study of melt inclusions in phenocrysts from magmatic rocks of Pantelleria Island (Central Mediterranean, Italy) [9], in which liquid immiscibility of alkaline acid magmas with the formation of saline melt was observed. Chlo� ride melts in inclusions and glasses from rocks of Pan� telleria Island were first discovered in 1987 [10] and hereafter were described in glasses of the rock ground� mass [11]. The melt and fluid inclusions were studied in phe� nocrysts from ignimbrite and pantellerite collected in different parts of Pantelleria Island and related to the precaldera and postcaldera stages of volcanism evolu� tion.