Home – Home – Jornals – Russian Geology and Geophysics 2009 number 12

Here we combine petrological-geochemical and thermomechanical modeling techniques to explain origin of primary magmas of both Maimecha-Kotui meimechites and the Gudchikhinskaya basalts of Norilsk region, which represent, respectively, the end and the beginning of flood magmatism in the Siberian Trap Province.
We have analyzed the least altered samples of meimechites, their olivine phenocrysts, and melt inclusions in olivines, as well as samples of dunites and their olivines, from boreholes G-1 and G-3 within the Guli volcanoplutonic complex in the Maimecha-Kotui igneous province of the northern Siberian platform. The Mn/Fe and Ni/MgO ratios in olivines indicate a mantle peridotite source of meimechites. Parental meimechite magma that rose to shallow depths was rich in alkalies and highly magnesian (24 wt.% MgO), largely degassed, undersaturated in sulfide liquid, and oxidized. At greater depths, it was, likely, high in CO₂ (6 wt.%) and H₂O (2 wt.%) and resulted from partial melting of initially highly depleted and later metasomatized harzburgite about 200 km below the surface. Trace-element abundances in primary meimechite magma suggest the presence of garnet and K-clinopyroxene in the mantle source and evidence a genetic relation to the sources of the early Siberian flood basalts (Gudchikhinskaya suite) and kimberlites. The analyzed dunite samples from the Guli complex have chemistry and mineralogy indicating their close relation to meimechites.
We have also computed a thermomechanical model of interaction of a hot mantle plume with the shield lithosphere of variable thickness, using realistic temperature- and stress-dependent visco-elasto-plastic rocks rheology and advanced finite-element solution technique.
Based on our experimental and modeling results, we propose that a Permian-Triassic plume with a potential temperature of about 1650 oC transported a large amount of recycled ancient oceanic crust (up to 15%) as SiO₂-supersaturated carbonated eclogite. Low-degree partial melting of eclogite at depths of 250-300 km produced carbonate-silicate melt that metasomatized the lithospheric roots of the Siberian shield. Further rise of the plume under relatively attenuated lithosphere (Norilsk area) led to the progressive melting of eclogite and formation of reaction pyroxenite, which then melted at depths of 130-180 km. Consequently, a large volume of melt (Gudchikhinskaya suite) penetrated into the lithosphere and caused its destabilization and delamination. Delaminated lithosphere that included fragments of locally metasomatized depleted harzburgite subsided into the plume and was heated to the temperatures of the plume interior with subsequent generation of meimechite magma. Meimechites showed up at the surface only under the thicker part of the lithosphere aside from major melting zone above, otherwise they would have been mixed up in more voluminous flood basalts. We further suggest that meimechites, uncontaminated Siberian flood basalts, and kimberlites all share the same source of strongly incompatible elements, the carbonated recycled oceanic crust carried up by hot mantle plume.

Four Precambrian metamorphic complexes in the vicinity of regional faults in the Transangarian region of the Yenisei Ridge were examined. Based on geothermobarometry and P-T path calculations, our geological and petrological studies showed that the Neoproterozoic medium-pressure metamorphism of the kyanite-sillimanite type overprinted regionally metamorphosed low-pressure andalusite-bearing rocks at about 850 Ma. A positive correlation between rock ages and P-T estimates for the kyanite-sillimanite metamorphism provide evidence for the regional structural and tectonic heterogeneity. The medium-pressure metamorphism was characterized by (1) the development of deformational structures and textures and kyanite-bearing blastocataclasites (blastomylonites) with sillimanite, garnet, and staurolite after andalusite-bearing regional metamorphic rocks; (2) insignificant apparent thickness of the zone of medium-pressure zonal metamorphism (from 2.5 to 7 km), which was localized in the vicinity of the overthrusts; (3) a low metamorphic field gradient during metamorphism (from 1-7 to 12 °C/km); and (4) a gradual increase in lithostatic pressure toward the thrust faults. These specific features are typical of collisional metamorphism during overthrusting of continental blocks and are evidence for near-isothermal loading. This event was justified within the framework of the crustal tectonic thickening model via rapid overthrusting and subsequent rapid uplifting and erosion. The results obtained allowed us to consider medium-pressure kyanite-bearing metapelites as a product of collision metamorphism, formed either by unidirectional thrusting of rock blocks from Siberian craton over the Yenisei Ridge in the zones of regional faults (Angara, Mayakon, and Chapa areas) or by opposite movements in the zone of splay faults of higher ranks (Garevka area).

The "unexpected" (the word is from H.G.F. Winkler, 1974) discovery of CO₂-rich inclusions in granulites has initiated a debate, which, after more than 35 years, is still an important issue in etamorphic petrology. Experimental and stable isotope data have led to the conception of a "fluid-absent" model, opposed to the "fluid-assisted" hypothesis, derived from fluid inclusion evidence. Besides CO₂, other fluids have been found to be of importance in these rocks, notably, concentrated aqueous solutions (brines), also able to coexist with granulite mineral assemblages at high P and T . Brines also occur in inclusions or, more impressively, have left their trace in large-scale metasomatic effects typical of a number of high-grade areas, e.g., intergranular K-feldspar veining and quartz exsolution (myrmekites), carbonate metasomatism along km-scale shear zones (Norway, India), "incipient charnockites" (India, Sri Lanka, Scandinavia), and highly oxidized Archean granulites. All together, this impressive amount of evidence suggests that the amount of fluids in the lower crust, under peak metamorphic conditions, was very large indeed, far too important to be only locally derived. Then, except for remnants contained in inclusions, these fluids have left the rock system during postmetamorphic uplift.
Fluid remnants identical to those occurring in deep crustal granulites are also found in mantle minerals, including diamonds. Major mantle fluid source is related to the final stages of melting processes: late magmatic emanations from alkalic basaltic melts, and carbonate-metasomatizing aqueous fluids issued from igneous carbonatites. Even if a local derivation of some fluids by crustal melting cannot be excluded, most lower-crustal granulite fluids have the same origin. They are transferred from the mantle into the crust by synmetamorphic intrusives, also responsible for the high thermal gradient typical of granulites, notably, HT- or UHT-types. These are mostly found in Precambrian times, generated during a small number of time intervals, e.g., around 500, 1000, 1800, 2500 Ma. High-temperature granulites forming events occur at world scale in supercontinents or supercratons, either at the end of amalgamation or shortly before breaking-off. They provide a mechanism for a vertical accretion of the continental slab, which complement the more classical way of lateral accretion above subduction zones at convergent boundaries.

We have studied a large (12 × 22 × 30 cm) spinel lherzolite xenolith with undeformed margins in alkali basalt (basanite) from the eroded crater of Late Cenozoic Shavaryn Tsaram-1 volcano in western Mongolia. The xenolith was sampled along its median transversal profile, at every 15-20 mm for bulk chemistry of lherzolite and basalt (ICP MS) and at 4-10 mm for the chemistry of olivine, orthopyroxene, clinopyroxene, and Cr-spinel minerals, and of material filling cracks (LA ICP MS). Incompatible elements (especially LREE) are distributed unevenly over the xenolith, both in lherzolite and in its constituent minerals, as well as in crack-filling material, with abnormal LREE enrichment in some specimens. Judging by the measured trace-element spectra compared with the model patterns, incompatible elements reside in different amounts as interstitial impurity in cracks inside and between mineral grains in lherzolite, also being a substitutional impurity in the lherzolite constituent minerals. Experimental acid leaching of specimens from sites of high crack density showed (La/Yb)_n ratios in the crack fill to be much higher than in the basalt host and more so in bulk lherzolite (180 against 33 and 1.5-3.6, respectively). The proportional contents of P and Ca in the leaching solution, especially in that from the xenolith's center, mark the presence of an apatite microphase, which can be a LREE repository.
The observed patterns of LREE and other incompatible elements in the xenolith and in the host alkali basalt fit a model implying that mobile elements residing as interstitial impurity came with fluids which were released from rising basaltic magma and percolated into the xenolith along cracks.

Mineral physics data related to the deep dehydration of stagnant slabs are summarized. The hydrogen diffusion in minerals of the mantle transition zone is not fast enough to homogenize the transition zone on the geological time scale, and hydrogen is expected to be unevenly distributed there. The hydrous fluid formed in the transition zone tends to percolate into shallower depths (410 km) to form gravitationally stable hydrous magmas at the base of the upper mantle. We need further studies on the relation of intraplate volcanism above the stagnant slab and deep dehydration, because we expect the geochemical fingerprints of deep dehydration to be quite different from those of shallow dehydration from the subducting slabs.

A new physicochemical model is proposed to predict and calculate the viscosity of magmatic melts and water diffusion as functions of the following parameters: P _total ; P _fl ; T ; melt composition, including volatiles (H₂O, OH^-, CO₂, , F^-, Cl^-); cation ratios: Al³⁺/(Al³⁺ + Si⁴⁺), Fe²⁺/(Fe²⁺ + Fe³⁺), Al³⁺/(Na⁺ + K⁺ + Ca²⁺ + Mg²⁺ + Fe²⁺) and a volume content of crystals and bubbles (up to 0.45, as applied to magma viscosity). The new model is specified by: (1) structural chemical approach; (2) ultimate simplicity of analytical dependences; (3) high accuracy of the prediction (±30 rel.%), which is consistent with the uncertainties of the experimental data on viscosity and water diffusion, especially at high pressures. The generalized equations suggested for calculating and predicting concentration, temperature, and pressure dependences of the viscosity of magmas and water diffusion are as follows:
Viscosity: η_T ^P = η₀ exp (E_X^P/RT),
where η₀ is a pre-exponent constant characterising the viscosity of liquids at T → ∞, (η₀ = 10^-3.5 ± 10^0.1) (dPa · s), or poises); T is the temperature in K; E _X^P is the activation energy of viscous flow (cal/mole), which is a function of melt composition, including volatile components, and pressure; R = 1.987 (cal/(mole·K)) is universal gas constant; and η_T^P is melt viscosity at the given temperature and pressure, in dPa · s.
Water diffusion:
log D (H₂O) = -(0.69 log η + 3.74) (depolymerized melts: andesite, basalt),
log D (H₂O) = -(0.29 log η + 5.35) (polymerized melts: rhyolite, obsidian, albite, dacite),
where D(H₂O) is the coefficient of water diffusion in cm²/s and η is melt viscosity in dPa·s.
A simple computer program developed for calculating the viscosity of magmatic melts and water diffusion is recommended for modeling magmatic and volcanic processes as well as their dynamics.

Carbonatites associated with syenites and subalkalic mafic rocks (lamprophyres) occur in the Himalayan continental collisional zone, and this suggests their existence in other Phanerozoic collisional settings. The Early Paleozoic Ol'khon collisional system in Western Cisbaikalia is considered one of the possible occurrences. Subalkalic gabbroids as well as peculiar carbonate (brucite marbles) and calc-silicate rocks were found here, within the Tazheran massif of alkali and nepheline syenites. Alkali syenites, nepheline syenites, and calciphyres were dated at 471 Ma, 451-464 Ma, and 466 Ma, respectively, and their ages correspond to the main collisional events in the system. A geochemical description of igneous and carbonate rocks in the massif is provided. Close mapping showed unusual syenite and brucite marble combinations and the frequent vein or pipe-like form of carbonates and calciphyres corresponding to their magmatic intrusion. But carbonatite nature of the marble mentioned above does not fit their typical crustal geochemical features. Not ruling out the possibility of a total change of geochemical signatures of mantle carbonatite in the collisional medium, we propose two other possible explanations for these facts: (1) melting of carbonate masses by syenite and mafic magmas, followed by carbonate melt intrusion into the upper crust, and (2) protrusion of carbonates into syenites and gabbroids at the late stages of the contact action of a silicate magma. In this case, the above-mentioned carbonate structural features result from late recrystallization, whose mechanism is yet to be explained.

Nonisothermal equilibrium physicochemical dynamics has been numerically modeled to estimate the effect of reduced asthenosphere fluids on continental lithosphere profiles beneath the Siberian Platform (SP). When the over-asthenosphere continental mantle is metasomatically changed by reduced magmatic fluids, the following sequence of zones forms: (1) zone where initial rocks are intensively sublimated and depleted by most petrogenic components; the restite in this case becomes carbonated, salinated, and graphitized; (2) zone of Si and Fe enrichment and carbon deposition in initial rocks depleted in Na, K, P, Mn; (3) zone of diamond-bearing lherzolites enriched with Na; (4) zone of hydrated rocks enriched with K; (5) zone of hydrated rocks not enriched with petrogenic components. Zone 1 can be responsible for the formation of kimberlite melts, zones 3 and 4 can be substrates of alkaline magma melting, and zone 5 can be the source of mafic tholeiitic magma.

This paper studies the petrology of K-alkaline lamproite-carbonatite complexes, which are widespread in Siberia. They are exemplified by the Murun and Bilibino massifs in West and Central Aldan. In these massifs, the entire range of differentiates was first found, from K-ultrabasic-alkaline rocks through basic and intermediate ones to alkali granites and unique residual calc-silicate rocks (benstonite Ba-Sr-carbonatites and charoite rocks). Also, intrusive equivalents of lamproites occur in these massifs, and the Murun massif was probably formed from highly differentiated lamproite magmas. In many K-alkaline complexes, silicate and silicate-carbonate magma layering takes place. Stages of magmatism are described for both massifs. Binary and ternary petrochemical diagrams exhibit the same compositional trend from early to late rocks.
In this paper, lamproites are considered from the chemical point of view; their diagnostic properties are described in terms of chemical and mineral composition. From geological, petrological, and geochemical data, formational analysis of alkaline complexes was performed, four formational types of world lamproites were first identified, and diamond content criteria were developed for them.
The carbonatite problem was studied from the petrological point of view, and four formational types of carbonatites were identified using geological, geochemical, and genetic criteria. It has been suggested that for dividing carbonatite complexes into four formational types, the following criteria should be used: the alkalinity type (Na or K) of alkaline rocks in the complex and the time when the carbonatite liquid separates from silicate melts in different stages of primary magma differentiation. These linked parameters influence the ore content type of carbonatite complexes.
A formation model for K-alkaline carbonatite complexes is given, and the Tomtor alkaline carbonatite massif with tuffaceous rare-metal ores is described to prove that they have ore reserves. The geochemistry of C, O, Sr, and Nd isotopes shows that K-alkaline complexes, depending on their geotectonic setting, can originate from three types of mantle sources: depleted mantle, enriched mantle 1 (EM1), and enriched mantle 2 (EM2). It is concluded that ore-bearing ultrabasic-alkaline complexes of lamproites and carbonatites can melt out of different types of mantle, whose composition only slightly influences their ore content. Apparently, the main factors are the low degree of selective mantle melting (less than 1%) and plumes supplying fluid and alkaline components, which stimulate this melting. Later on, the processes important for the accumulation of ore and trace elements are long-term magma differentiation and its layering during crystallization.

Experiments on water solubility in forsterite in the systems Mg₂SiO₄ - K₂Mg (CO₃)₂ - H₂O and Mg₂SiO₄-H₂O-C were conducted at 7.5-14.0 GPa and 1200-1600°C. The resulting crystals contain 448 to 1480 ppm water, which is 40-70% less than in the forsterite-water system under the same conditions. This can be attributed to lower water activity in the carbonate-bearing melt. The water content of forsterite was found to vary systematically with temperature and pressure. For instance, at 14 GPa, the H₂O content of forsterite in the system forsterite-carbonate-H₂O drops from 1140 ppm at 1200°C to 450 ppm at 1600°C, and at 8 GPa it remains constant or increases from 550 to 870 ppm at 1300-1600°C. Preliminary data for D-H-bearing forsterite are reported. Considerable differences were found between IR spectra of D-H- and H-bearing forsterite. The results suggest that CO₂ can significantly affect the width of the olivine-wadsleyite transition, i.e., the 410 km seismic discontinuity, which is a function of the water content of olivine and wadsleyite.

The electrical conductivity of the major upper mantle minerals, namely, olivine, wadsleyite, and ringwoodite, is reviewed in this paper. There are mainly three electrical conduction mechanisms for three upper mantle minerals, namely, hopping, ionic, and proton conductions. The charge carriers for these conduction mechanisms are an electron hole in Fe ion, a vacancy in Mg site, and a proton, respectively. Hopping conduction is the most essential conduction mechanism for the major upper mantle minerals. Because ionic conduction has high activation energy, it becomes a dominant conduction mechanism only at high temperatures. Proton conduction contributes at relatively low temperatures. If the mantle minerals contain a large amount of water (more than 0.1 wt.%), proton conduction can be a dominant conduction mechanism, even at high temperatures.

Literature data on the coefficients of heterovalent NaSi-CaAl interdiffusion in plagioclases are analyzed. It is shown that activation energy and frequency factor decrease as the partial pressure of water and plagioclase basicity increase. A general expression has been derived for the interdiffusion coefficients of albite and anorthite asa function of the plagioclase composition (N _Pl ), temperature T (K), and water pressure (kbar):
D [см²/c] = exp {13.14 - 0.354N_Pl + 40.72/(1 + P²_H2O) - (116095 - 1037N_Pl + 139732/(1 + P²_H2O))/RT }

Samples of poikoblastic garnets from the Escambray (Cuba), Maksyutov (Russia), and Sambagawa (Japan) eclogite complexes were heated to 700-1100°C at 3 to 4 GPa (30-40 kbar). Epidote, amphibole, and chlorite inclusions in the garnets underwent dehydration melting over the entire experimental PT -range, which is typical of ultrahigh-pressure (UHP) metamorphic complexes. In the presence of aqueous fluids, carbonate minerals in the inclusions began to melt at 800°C and 3 GPa. Melting gave rise to new garnet, with the composition controlled by the chemistry of the primary inclusions and by PT run conditions. Garnet either grew directly from the melt or formed by replacement of host garnet walls leaving residual melt at the substitution front in the latter case. Partial melting of inclusions decreased the mechanical strength of the garnet host and led to local shearing. The experimental results were used to interpret observed features in two samples of a diamond-bearing and a diamond-free carbonate-silicate rocks from the Kumdy-Kol deposit in the Kokchetav Massif. Multiphase inclusions in both samples contain newly formed garnet with morphologies and compositions consistent with those produced experimentally under the given PT -conditions. Minerals in the inclusions are compositionally similar to those in matrix, thus suggesting that melting may have occurred on a large scale.

This paper gives an analytical overview of the experimental data obtained by different authors at high P and T in the model system MgO-Al₂O₃-SiO₂-Cr₂O₃ (MASCr). A set of four simple polynomial equations is proposed for the temperature and pressure dependence of chromium content in garnet and spinel in the assemblage Gar + Opx + Es and Gar + Fo + Opx + Sp.
From the first equation, one can estimate the minimum pressure at a given temperature which is required for the formation of peridotite garnets of uncertain paragenesis with a known knorringite content. A combination of the second and third equations helps estimate P and T from the chromium content of garnet and spinel from assemblages containing both minerals. If the spinel composition is unknown but there is a reason to assign garnet to a spinel-bearing paragenesis, the fourth equation is applicable for estimating pressure at a given temperature.
Originally, the proposed garnet-spinel geothermobarometry was developed for a harzburgite paragenesis. However, it is applicable to garnets with CaO/Cr₂O₃ < 0.90 (including lherzolite ones), i.e., those within the Pyr-Kn-Uv triangle of the reciprocal quaternary diagram Pyr-Cros-Uv-Kn.
Using the above equations and an empirical P _CG geobarometer, comparative geothermobarometric estimates were obtained for a set of garnet and garnet-spinel inclusions in diamonds and intergrowths with diamond, as well as garnet inclusions in spinel. If garnet has CaO/Cr₂O₃ = 0.35-0.40, the results are in good accord. For Cr-richest and Ca-poorest garnets, the P _CG barometer shows pressures 10-15% higher compared with our estimates.

Residual pressure around mineral inclusions in diamond can provide useful information on the depth of diamond origin. Differential stress between an inclusion and host diamond arises from differences in thermal expansion and compressibility between host diamond and minerals. We determined residual pressure around mineral inclusions in a diamond from the Internatsionalnaya Pipe, Yakutia, Russia, using the three-dimensional Raman mapping system developed recently by our group. The maximum residual pressures around the olivine and chromite inclusions were determined to be 0.69 GPa and 0.75 GPa, respectively. We proposed an advanced method for determining simultaneously pressure and temperature conditions where the mineral inclusions were trapped in the host diamond. The obtained values were 3.0 GPa and 447 °C, but these values are lower than typical PT -conditions in the mantle. Several technical possibilities for the discrepancy are discussed.

Experimental studies of diamond formation in the alkaline silicate-carbon system Na₂O-K₂O-MgO-CaO-Al₂O₃-SiO₂-C were carried out at 8.5 GPa. In accordance with the diamond nucleation criterion, a high diamond generation efficiency (spontaneous mass diamond crystallization) has been confirmed for the melts of the system Na₂SiO₃-carbon and has been first established for the melts of the systems CaSiO₃-carbon and (NaAlSi₃O₈)₈₀ (Na₂SiO₃)₂₀-carbon. It is shown that in completely miscible carbonate-silicate melts oversaturated with dissolved diamond-related carbon, a concentration barrier of diamond nucleation (CBDN) arises at a particular ratio of carbonate and silicate components. Study of different systems (eclogite-K-Na-Mg-Ca-Fe-carbonatite-carbon, albite-K₂CO₃-carbon, etc.) has revealed a dependence of the barrier position on the chemical composition of the system and the inhibiting effect of silicate components on the nucleation density and rate of diamond crystal growth. In multicomponent eclogite-carbonatite solvent, the CBDN is within the range of carbonatite compositions (<50 wt.% silicates). Based on the experimental criterion for the syngenesis of diamond and growth inclusions in them, we studied the syngenesis diagram for the system melanocratic carbonatite-diamond and determined a set of the composition fields and physical parameters of the system that are responsible for the cogeneration of diamond and various mineral and melt parageneses. The experimental results were applied to substantiate a new physicochemical concept of carbonate-silicate (carbonatite) growth media for most of natural diamonds and to elaborate a genetic classification of growth mineral, melt, and fluid inclusions in natural diamonds of mantle genesis.

Sulfide inclusions in diamonds, the most common of all inclusions, contain critical evidence about the timing and physical/chemical conditions prevailing during diamond formation. Typically, sulfide inclusions are encapsulated as a monosulfide solid solution (Mss) in the Fe-Ni-S system, with a minor amount of Cu. This Mss and the enclosing diamond have sufficiently different thermal expansion properties, so that, after encapsulation, the Mss creates a series of cracks in the diamond radiating from the sulfide. On cooling, this increase in volume permits the Mss to undergo exsolution to an assemblage of pyrrhotite + pentlandite + chalcopyrite ± ± pyrite. The kinetics of this exsolution is so rapid that practically no Mss remains in nature. Instead, in recovered diamonds, all sulfides that originally were Mss now consist of this fine-grained assemblage. Chalcopyrite prefers to form around the edges of the inclusions and also migrates into the minute cracks in the diamonds. It is the bulk composition of the Mss as encapsulated that is important for interpretation of diamond petrogenesis (P- versus E-type diamonds) and to the commonly used Re-Os dating technique. However, this bulk composition is definitely not attainable with polished sections cut through the inclusions. The assumption that the kernel of the sulfide inclusion for Re-Os age dating represents the entire original Mss may also be incorrect, depending what has been lost, mostly chalcopyrite, which has migrated into the surrounding cracks within the diamond host.

Manganoan ilmenite was identified in Juina, Brazil kimberlitic rocks among other megacrysts. It forms oval, elongated, rimless grains comprising 8-30 wt.% of the heavy fraction. Internally the grains are homogeneous. The chemical composition of Mn-ilmenite is almost stoichiometric for ilmenite, except for an unusually high manganese content, with MnO = 0.63-2.49 wt.% (up to 11 wt.% in inclusions in diamond) and an elevated vanadium admixture (V₂O₃ = 0.21-0.43 wt.%). By the composition, Mn-ilmenite megacrysts and inclusions in diamond are almost identical. The concentrations of trace elements in Mn-ilmenite, compared to picroilmenite, are much higher, and their variations are very wide. Chondrite-normalized distribution of trace elements in Mn-ilmenite megacrysts is similar to the distribution in Mn-ilmenites included in diamond. This confirms that Mn-ilmenite in kimberlites is genetically related to diamond. The finds of Mn-ilmenite known before in kimberlitic and related rocks are late- or postmagmatic metasomatic phases. They either form reaction rims on grains of picroilmenite or other ore minerals or compose laths in groundmass. In contrast to those finds, Mn-ilmenite megacrysts in Juina kimberlites are a primary mineral phase with a homogeneous internal structure obtained under stable conditions of growth within lower mantle and/or transition zone. In addition to pyrope garnet, chromian spinel, picroilmenite, chrome-diopside, and magnesian olivine, manganoan ilmenite may be considered another kimberlite diamond indicator mineral.

Diamond crystallization in multicomponent melts of variable composition is studied. The melt carbonates are K₂CO₃, CaCO₃ · MgCO₃, and K-Na-Ca-Mg-Fe-carbonatites, and the melt silicates are model peridotite (60 wt.% olivine, 16 wt.% orthopyroxene, 12 wt.% clinopyroxene, and 12 wt.% garnet) and eclogite (50 wt.% garnet and 50 wt.% clinopyroxene). In the experiments carried out under the PT -conditions of diamond stability, the carbonate-silicate melts behave like completely miscible liquid phases. The concentration barriers of diamond nucleation (CBDN) in the melts with variable proportions of silicates and carbonates have been determined at 8.5 GPa. In the system peridotite-K₂CO₃/CaCO₃ · MgCO₃/carbonatite they correspond to 30, 25, and 30 wt.% silicates, respectively, and in the analogous eclogite-carbonate system, 45, 30, and 35 wt.%. In the silicate-carbonate melts with higher silicate contents the seed diamond growth occurs, which is accompanied by the crystallization of thermodynamically unstable graphite phase. In the experiments with melts compositionally corresponding to the CBDN at 7.0 GPa and 1200-1700 °C, a full set of silicate minerals of peridotite (olivine, orthopyroxene, clinopyroxene, garnet) and eclogite (garnet, clinopyroxene) parageneses was obtained. The minerals occur as syngenetic inclusions in natural diamonds; moreover, the garnets contain an impurity of Na, and the pyroxenes, K. The experimental data indicate that peridotite-carbonate and eclogite-carbonate melts are highly effective for the formation of diamond (or unstable graphite) together with syngenetic minerals and melts, which agrees with the carbonate-silicate (carbonatite) model for the mantle diamond formation.

We discuss the chemistry of exceptionally rare phlogopite inclusions coexisting with ultramafic (peridotitic) and eclogitic minerals in kimberlite-hosted diamonds of Yakutia, Arkhangelsk, and Venezuela provinces. Phlogopite inclusions in diamonds are octahedral negative crystals following the diamond faceting in all 34 samples (including polymineralic inclusions). On this basis phlogopite inclusions have been interpreted as syngenetic and in equilibrium with the associated minerals. In ultramafic diamonds phlogopites coexist with subcalcic high-Cr₂O₃ pyrope and/or chromite, olivine, and enstatite (dunite/harzburgite (H) paragenesis) or with clinopyroxene, enstatite, and/or olivine, and pyrope (lherzolite (L) paragenesis). Ultramafic phlogopites have high Mg# [100 ?Mg/(Mg+Fe)] from 92.4 to 95.2 and Cr₂O₃ higher than TiO₂ in H-phlogopites (1.5-2.5 wt.% versus 0.1-0.4 wt.%, respectively) but lower in L-phlogopites (0.15-0.5 wt.% versus 1.3-3.5 wt.%, respectively). Eclogitic (E) phlogopites show Mg# from 47.4 to 85.3 inclusive, and very broad ranges of TiO₂ up to 12 wt.%. The primary syngenetic origin of phlogopite is indicated, besides other factors, by its compositional consistency with the associated minerals. The analyzed phlogopites are depleted in BaO (0.10-0.79 wt.%), and their F and Cl contents are highly variable, reaching 1.29 and 0.49 wt.%, respectively. The latter is in line with high Cl enrichment in some unaltered kimberlites and in nanometric fluid inclusions from diamonds. The presence of syngenetic phlogopite in kimberlite-hosted diamonds provides important evidence that volatiles participated in diamond formation and that at least part of diamonds may have been related to early stages of kimberlites formation.