Pressure effect on Ti- or P-rich accessory mineral saturation in evolved granitic melts with differing K2O/Na2O ratios

The solubility of Ti- and P-rich accessory minerals has been examined as a function of pressure and K₂O/Na₂O ratio in two series of highly evolved silicate systems. These systems correspond to (a) alkaline, varying from alkaline to peralkaline with increasing K₂O/Na₂O ratio; and (b) strongly metaluminous (essentially trondhjemitic at the lowest K₂O/Na₂O ratio) and remaining metaluminous with increasing K₂O/Na₂O ratio (to 3). The experiments were conducted at a fixed temperature of 1000 °C, with water contents varying from 5 wt.% at low pressure (0.5 GPa), increasing through 5–10 wt.% at 1.5–2.5 GPa to 10 wt.% at 3.5 GPa. Pressure was extended outside the normal crustal range, so that the results may also be applied to derivation of hydrous silicic melts from subducted oceanic crust.

For the alkaline composition series, the TiO₂ content of the melt at Ti-rich mineral saturation decreases with increasing pressure but is unchanged with increasing K content (at fixed pressure). The P₂O₅ content of the alkaline melts at apatite saturation increases with increased pressure at 3.5 GPa only, but decreases with increasing K content (and peralkalinity). For the metaluminous composition series (termed as “trondhjemite-based series” (T series)), the TiO₂ content of the melt at Ti-rich mineral saturation decreases with increasing pressure and with increasing K content (at fixed pressure). The P₂O₅ content of the T series melts at apatite saturation is unchanged with increasing pressure, but decreases with increasing K content. The contrasting results for P and Ti saturation levels, as a function of pressure in both compositions, point to contrasting behaviour of Ti and P in the structure of evolved silicate melts. Ti content at Ti-rich mineral saturation is lower in the alkaline compared with the T series at 0.5 GPa, but is similar at higher pressures, whereas P content at apatite saturation is lower in the T series at all pressures studied. The results have application to A-type granite suites that are alkaline to peralkaline, and to I-type metaluminous suites that frequently exhibit differing K₂O/Na₂O ratios from one suite to another.     

The origin of medium-K ankaramitic arc magmas from Lombok (Sunda arc, Indonesia): Mineral and melt inclusion evidence

High-calcium, nepheline-normative ankaramitic basalts (MgO > 10 wt.%, CaO/Al₂O₃ > 1) from Rinjani volcano, Lombok (Sunda arc, Indonesia) contain phenocrysts of clinopyroxene and olivine (Fo_85–92) with inclusions of spinel (Cr# 58–77) and crystallised melt. Olivine crystals have variable but on average low NiO (0.10–0.23 wt.%) and high CaO (0.22–0.35 wt.%) contents for their forsterite number. The CaO content of Fo_89–91 olivine is negatively correlated with the Al₂O₃ content of enclosed spinel (9–15 wt.%) and positively correlated with the CaO/Al₂O₃ ratios of melt inclusions (0.9–1.5). Major and trace element patterns of melt inclusions are similar to that of the host rock, indicating that the magma could have formed by accumulation of small batches of melt, with compositions similar to the melt inclusions. The liquidus temperature of the magma was  1275 °C, and its oxygen fugacity ≤ FMQ + 2.5. Correlations between K₂O, Zr, Th and LREE in the melt inclusions are interpreted to reflect variable degrees of melting of the source; correlations between Al₂O₃, Na₂O, Y and HREE are influenced by variations in the mineralogy of the source. The melts probably formed from a water-poor, clinopyroxene-rich mantle source.     

Fluid-bearing alkaline carbonate melts as the medium for the formation of diamonds in the Earth''s mantle: an experimental study

We have experimentally studied the formation of diamonds in alkaline carbonate–carbon and carbonate–fluid–carbon systems at 5.7–7.0 GPa and 1150–1700 °C, using a split-sphere multi-anvil apparatus (BARS). The starting carbonate and fluid-generating materials were placed into Pt and Au ampoules. The main specific feature of the studied systems is a long period of induction, which precedes the nucleation and growth of diamonds. The period of induction considerably increases with decreasing P and T, but decreases when adding a C–O–H fluid to the system. In the range of P and T corresponding to the formation of diamonds in nature, this period lasts for tens of hours. The reactivity of the studied systems with respect to the diamond nucleation and growth decreases in this sequence: Na₂CO₃–H₂C₂O₄·2H₂O–C>K₂CO₃–H₂C₂O₄·2H₂O–C>>Na₂CO₃–C>K₂CO₃–C. The diamond morphology is independent of P and T, and is mainly governed by the composition of the crystallization medium. The stable growth form is a cubo-octahedron in the Na₂CO₃ melt, and an octahedron in the K₂CO₃ melt. Regardless of the composition of the carbonate melt, only octahedral diamond crystals formed in the presence of the C–O–H fluid. The growth rates of diamond varied in the range from 1.7 μm/h at 1420 °C to 0.1–0.01 μm/h at 1150 °C, and were used to estimate, for the first time, the possible duration of the crystallization of natural diamonds. From the analysis of the experimental results and the petrological evidence for the formation of diamonds in nature, we suggest that fluid-bearing alkaline carbonate melts are, most likely, the medium for the nucleation and growth of diamonds in the Earth's upper mantle.     

Carbonatite melt–CO2 fluid inclusions in mantle xenoliths from Tenerife, Canary Islands: a story of trapping, immiscibility and fluid–rock interaction in the upper mantle

Three types of fluid inclusions have been identified in olivine porphyroclasts in the spinel harzburgite and lherzolite xenoliths from Tenerife: pure CO₂ (Type A); carbonate-rich CO₂–SO₂ mixtures (Type B); and polyphase inclusions dominated by silicate glass±fluid±sp±silicate±sulfide±carbonate (Type C). Type A inclusions commonly exhibit a “coating” (a few microns thick) consisting of an aggregate of a platy, hydrous Mg–Fe–Si phase, most likely talc, together with very small amounts of halite, dolomite and other phases. Larger crystals (e.g. (Na,K)Cl, dolomite, spinel, sulfide and phlogopite) may be found on either side of the “coating”, towards the wall of the host mineral or towards the inclusion center. These different fluids were formed through the immiscible separations and fluid–wall-rock reactions from a common, volatile-rich, siliceous, alkaline carbonatite melt infiltrating the upper mantle beneath the Tenerife. First, the original siliceous carbonatite melt is separated from a mixed CO₂–H₂O–NaCl fluid and a silicate/silicocarbonatite melt (preserved in Type A inclusions). The reaction of the carbonaceous silicate melt with the wall-rock minerals gave rise to large poikilitic orthopyroxene and clinopyroxene grains, and smaller neoblasts. During the metasomatic processes, the consumption of the silicate part of the melt produced carbonate-enriched Type B CO₂–SO₂ fluids which were trapped in exsolved orthopyroxene porphyroclasts. At the later stages, the interstitial silicate/silicocarbonatite fluids were trapped as Type C inclusions. At a temperature above 650 °C, the mixed CO₂–H₂O–NaCl fluid inside the Type A inclusions were separated into CO₂-rich fluid and H₂O–NaCl brine. At T<650 °C, the residual silicate melt reacted with the host olivine, forming a reaction rim or “coating” along the inclusion walls consisting of talc (or possibly serpentine) together with minute crystals of NaCl, KCl, carbonates and sulfides, leaving a residual CO₂ fluid. The homogenization temperatures of +2 to +25 °C obtained from the Type A CO₂ inclusions reflect the densities of the residual CO₂ after its reactions with the olivine host, and are unrelated to the initial fluid density or the external pressure at the time of trapping. The latter are restricted by the estimated crystallization temperatures of 1000–1200 °C, and the spinel lherzolite phase assemblage of the xenolith, which is 0.7–1.7 GPa.     

The nature of the sub-continental mantle: constraints from the major-element composition of continental flood basalts

Many continental flood basalts (CFB) have isotope and trace-element signatures that differ from those of oceanic basalts and much interest concerns the extent to which these reflect differences in their upper mantle source regions. A review of selected data sets from the Mesozoic and Tertiary CFB confirms significant differences in their major- and trace-element compositions compared with those of basalts erupted through oceanic lithosphere. In general, those CFB suites characterised by low Nb/La, high (⁸⁷Sr/⁸⁶Sr)_i and low εNd_i tend to exhibit relatively low TiO₂, CaO/Al₂O₃, Na₂O and/or Fe₂O₃, and relatively high SiO₂. In contrast, those which have high Nb/La, low (⁸⁷Sr/⁸⁶Sr)_i and high εNd_i ratios, like the upper units in the Deccan Traps, have major- and trace-element compositions similar to oceanic basalts. It would appear that those CFB that have distinctive isotope and trace-element ratios also exhibit distinctive major-element contents, suggesting that major and trace elements have not been decoupled significantly during magma generation and differentiation.

When compared (at 8% MgO) with oceanic basalt trends, the displacement of many CFB to lower Na₂O, Fe₂O₃^*, TiO₂ and CaO/Al₂O₃, but higher SiO₂, at similar Mg#, is not readily explicable by crustal contamination. Rather, it reflects source composition and/or the effects of the melting processes. The model compositions of melts produced by decompression of mantle plumes beneath continental lithosphere have relatively low SiO₂ and high Fe₂O₃^*. In contrast, the available experimental data indicate that partial melts of peridotite have low TiO₂, Na₂O and Fe₂O₃^*CaO/Al₂O₃, if the peridotite has been previously depleted by melt extraction. Moreover, melting of hydrated, depleted peridotite yields SiO₂-rich, Fe₂O₃- and CaO-poor melts. Since anhydrous, depleted peridotite has a high-temperature solidus, it is argued that the source of these CFB was variably melt depleted and hydrated mantle, inferred to be within the lithosphere. Isotope data suggest these source regions were often old and relatively enriched in incompatible trace elements, and it is envisaged that H₂O±CO₂ were added at the same time as the incompatible elements. An implication is that a significant proportion of the new continental crust generated since the Permian reflected multistage processes involving mobilization of continental mantle lithosphere that was enriched in minor and trace elements during the Proterozoic.     

Basalt magma genesis and fractionation in collision- and extension-related provinces: A comparison between eastern, central and western Anatolia

Experimental studies of synthetic and natural basalt systems suggest that conditions of magma genesis and fractionation depend fundamentally on mantle temperatures and lithospheric stress fields. In general, compressional settings are more conducive to polybaric fractionation than extensional settings and in this regard, the Anatolian magmatic province offers a natural laboratory for comparing near-coeval basalt eruptions as a function of regional tectonics — compressional (collision-related) régimes dominating in eastern Anatolia and extensional tectonics characterizing a western province related to Aegean Sea opening. Projection of Plio-Quaternary basalt normative compositions from the Western Anatolia Extensional Province (WAEP), the Central Anatolian ‘Ova’ Province (CAOP), and Eastern Anatolia Compressional Province (EACP) are projected onto Ol–Ne–Cpx and Pl–Cpx–Ol planes in the simplified basalt system (Ne–Cpx–Ol–Qz), each showing distinctive liquid lines of descent. WAEP basalts are mostly constrained by low-pressure (< 0.5 GPa) cotectics while CAOP and EACP compositions conform to moderate and/or high-pressure (0.8–3.0 GPa) cotectics. Overall, a quasi-linear shift from moderate and/or high-pressure to low-pressure equilibria matches the westward transgression from compressional east Anatolia to the extensional west Anatolian–Aegean region. Comparison of their respective primary (mantle-equilibrated) magmas–simulated by normalizing their compositions to MgO = 15 wt.% (Mg-15)–with parameterized anhydrous and H₂O-undersaturated experimental melts suggests they segregated from spinel- to garnet-lherzolite mantle facies at pressures between c. 2 and 3 GPa (c. 70–100 km depth) under H₂O-undersaturated conditions. Interpolated potential temperatures (T_p) and lithospheric stretching factors (β) range as follows: (1) eastern Anatolian basalts associated with the Arabian foreland show T_p varying between 1250 and 1400 °C (except for the Karacalidag alkali basalts, south of the Bitlis–Zagros fracture zone, for which T_p ranges up to 1450 °C), for β values of 1.2–1.8. T_p values for central Anatolia (e.g. Sivas) range between 1300 and 1375 °C (except for Karapinar, Egrikuyu and Hasandag, which show < 1150 °C), and β values of 1.3–1.4. For western Anatolian basalts, T_p range mostly between 1250 and 1330 °C, except for a single value for Canakkale of 1400 °C and Kula sample showing T_p < 1200 °C, and β values of 1.3–2.0. Variation of these conditions is as great or greater than that between provinces, although there are clearly significant constraints on the inferred polybaric to low-pressure isobaric fractionation régimes. Covariation of total FeO, TiO₂, La/Yb, Ce/Sm, Zr/Y and Zr/Nb reflects small but significant differences in bulk composition and ambient melt fraction while the covariance of Ce/Sm and Sm/Yb is consistent with the segregation of primitive melts at the spinel- to garnet-lherzolite transition.     

Melting relations of peridotite and the density crossover in planetary mantles

Melting relations of primitive peridotite were studied up to 25 GPa. The change of the liquidus phase from olivine to majorite occurs at 16 GPa. We confirmed the density crossover of the FeO-rich peridotite melt and the equilibrium olivine (Fo₉₀) at 7 GPa. Sinking of equilibrium olivine (Fo₉₅) in the primitive peridotite melt was observed up to 10 GPa. The compression curves of FeO-rich peridotitic and komatiite melts reported in this and earlier work suggest that the density crossover in the Earth's mantle will be located at 11–12 GPa at 2000°C, consistent with an previous estimation by C.B. Agee and D. Walker.

The density crossover can play a key role in the Moon and the terrestrial planets, such as the Earth, Venus and Mars. Majorite and some fraction of melt could have separated from the ascending diapir and sunk downwards at the depths below the density crossover. This process could have produced a garnet-rich transition zone in the Earth's mantle. The density crossover may exist in the FeO-rich lunar mantle at around the center of the Moon. The density crossover which exists at the depth of 600 km in the Martian mantle plays a key role in producing a fractionated mantle, which is the source the parent magmas of the SNC meteorites.     

Status report on stability of K-rich phases at mantle conditions

Experimental research on K-rich phases and observations from diamond inclusions, UHP metamorphic rocks, and xenoliths provide insights about the hosts for potassium at mantle conditions. K-rich clinopyroxene (Kcpx–KM³⁺Si₂O₆) can be an important component in clinopyroxenes at P>4 GPa, dependent upon coexisting K-bearing phases (solid or liquid) but not, apparently, upon temperature. Maximum Kcpx content can reach 25 mol%, with 17 mol% the highest reported in nature. Partitioning ^(K)D_(cpx/liquid) above 7 GPa=0.1–0.2 require ultrapotassic liquids to form highly potassic cpx or critical solid reactions, e.g., between Kspar and Di. Phlogopite can be stable to about 8 GPa at 1250 °C where either amphibole or liquid forms. When fluorine is present, it generally increases in Phl upon increasing P (and probably T) to about 6 GPa, but reactions forming amphibole and/or KMgF₃ limit F content between 6 and 8 GPa. The perovskite KMgF₃ is stable up to 10 GPa and 1400 °C as subsolidus breakdown products of phlogopite upon increasing P. ^(M4)K-substituted potassic richterite (ideally K(KCa)Mg₅Si₈O₂₂(OH,F)₂) is produced in K-rich peridotites above 6 GPa and in Di+Phl from 6 to 13 GPa. K content of amphibole is positively correlated with P; Al and F content decrease with P. In the system 1Kspar+1H₂O K-cymrite (hydrous hexasanidine–KAlSi₃O₈·nH₂O–Kcym) is stable from 2.5 GPa at 400 to 1200 °C and 9 GPa; Kcym can be a supersolidus phase. Formation of Kcym is sensitive to water content, not forming within experiments with H₂O2O>Kspar. Phase X, a potassium di-magnesium acid disilicate ((K_1−x−n)₂(Mg_1−nM_n³⁺)₂Si₂O₇H_2x), forms in mafic compositions at T=1150–1400 °C and P=9–17 GPa and is a potential host for K and H₂O at mantle conditions with a low-T geotherm or in subducting slabs. The composition of phase-X is not fixed but actually represents a solid solution in the stoichiometries □₂Mg₂Si₂O₇H₂–(K□)Mg₂Si₂O₇H–K₂Mg₂Si₂O₇ (□=vacancy), apparently stable only near the central composition. K-hollandite, KAlSi₃O₈, is possibly the most important K-rich phase at very high pressure, as it appears to be stable to conditions near the core–mantle boundary, 95 GPa and 2300 °C. Other K-rich phases are considered.     

Experimental synthesis and phase relations of phengitic muscovite from 6.5 to 11 GPa in a calcareous metapelite from the Dabie Mountains, China

The stability and phase relations of phengitic muscovite in a metapelitic bulk composition containing a mixed H₂O+CO₂ fluid were investigated at 6.5–11 GPa, 750–1050°C in synthesis experiments performed in a multianvil apparatus. Starting material consisted of a natural calcareous metapelite from the coesite zone of the Dabie Mountains, China, ultrahigh-pressure metamorphic complex that had experienced peak metamorphic pressures greater than 3 GPa. The sample contains a total of 2.1 wt.% H₂O and 6.3 wt.% CO₂ bound in hydrous and carbonate minerals. No additional fluid was added to the starting material. Phengite is stable in this bulk composition from 6.5 to 9 GPa at 900°C and coexists with an eclogitic phase assemblage consisting of garnet, omphacite, coesite, rutile, and fluid. Phengite dehydrates to produce K-hollandite between 8 and 11 GPa, 750–900°C. Phengite melting/dissolution occurs between 900°C and 975°C at 6.5–8 GPa and is associated with the appearance of kyanite in the phase assemblage. The formation of K-hollandite is accompanied by the appearance of magnesite and topaz-OH in the phase assemblage as well as by significant increases in the grossular content of garnet (average X_grs=0.52, X_py=0.19) and the jadeite content of omphacite (X_jd=0.92). Mass balance indicates that the volatile content of the fluid phase changes markedly at the phengite/K-hollandite phase boundary. At P≤8 GPa, fluid coexisting with phengite appears to be relatively CO₂-rich (XCO₂/XH₂O=2.2), whereas fluid coexisting with K-hollandite and magnesite at 11 GPa is rich in H₂O (XCO₂/XH₂O=0.2). Analysis of quench material and mass balance calculations indicate that fluids at all pressures and temperatures examined contain an abundance of dissolved solutes (approximately 40 mol% at 8 GPa, 60 mol% at 11 GPa) that act to dilute the volatile content of the fluid phase. The average phengite content of muscovite is positively correlated with pressure and ranges from 3.62 Si per formula unit (pfu) at 6.5 GPa to 3.80 Si pfu at 9 GPa. The extent of the phengite substitution in muscovite in this bulk composition appears to be limited to a maximum of 3.80–3.85 Si pfu at P=9 GPa. These experiments show that phengite should be stable in metasediments in mature subduction zones to depths of up to 300 km even under conditions in which aH₂O1. Other high-pressure hydrous phases such as lawsonite, MgMgAl-pumpellyite, and topaz-OH that may form in subducted sediments do not occur within the phengite stability field in this system, and may require more H₂O-rich fluid compositions in order to form. The wide range of conditions under which phengite occurs and its participation in mixed volatile reactions that may buffer the composition of the fluid phase suggest that phengite may significantly influence the nature of metasomatic fluids released from deeply subducted sediments at depths of up to 300 km at convergent plate boundaries.     

Mineral chemistry and thermobarometry of inclusions from De Beers Pool diamonds, Kimberley, South Africa

Silicate and oxide mineral inclusions in diamonds from the geologically and historically important De Beers Pool kimberlites in Kimberley, South Africa, are characterised by harzburgitic compositions (>90%), with lesser abundances from eclogitic and websteritic parageneses. The De Beers Pool diamonds contain unusually high numbers of inclusion intergrowths, with garnet+orthopyroxene±chromite±olivine and chromite+olivine assemblages dominant. More unusual intergrowths include garnet+olivine+magnesite and an eclogitic assemblage comprising garnet+clinopyroxene+rutile. The mineral chemistry of the De Beers Pool inclusions overlaps that of most worldwide localities. Peridotitic garnet inclusions exhibit variable CaO (<5.8 wt.%) and Cr₂O₃ contents (3.0–15.0 wt.%), although the majority are harzburgitic with very low calcium concentrations (<2 wt.% CaO). Eclogitic garnet inclusions are characterised by a wide range in CaO (3.3–21.1 wt.%) with low Cr₂O₃ (<1 wt.%). Websteritic garnets exhibit intermediate compositions. Most chromite inclusions contain 63–67 wt.% Cr₂O₃ and <0.5 wt.% TiO₂. Olivine and orthopyroxene inclusions are magnesium-rich with Mg-numbers of 93–97. Olivine inclusions in chromite exhibit the highest Mg-numbers and also contain elevated Cr₂O₃ contents up to 1.0 wt.%. Peridotitic clinopyroxene inclusions are Cr-diopsides with up to 0.8 wt.% K₂O. Eclogitic and websteritic clinopyroxene inclusions exhibit overlapping compositions with a wide range in Mg-numbers (66–86).

Calculated temperatures for non-touching inclusion pairs from individual diamonds range from 1082 to 1320 °C (average=1197 °C), whereas pressures vary from 4.6 to 7.7 GPa (average=6.3 GPa). Touching inclusion assemblages are characterised by equilibration temperatures of 995 to 1182 °C (average=1079 °C) and pressures of 4.2–6.8 GPa (average=5.4 GPa). Provided that the non-touching inclusions represent equilibrium assemblages, it is suggested that these inclusions record the conditions at the time of diamond crystallisation (1200 °C; 3.0 Ga). The lower average temperatures for touching inclusions are attributed to re-equilibration in a cooling mantle (1050 °C) prior to kimberlite eruption at 85 Ma. Pressure estimates for touching garnet–orthopyroxene inclusions are also skewed towards lower values than most non-touching inclusions. This apparent difference may be an artefact of the Al-exchange geobarometer and/or the result of sampling bias, due to limited numbers of non-touching garnet–orthopyroxene inclusions. Alternatively pressure differences could be caused by differential uplift in the mantle or possibly variations in thermal compressibility between diamond and silicate inclusions. However, thermodynamic modelling suggests that thermal compressibility differences would cause only minor changes in internal inclusion pressures (<0.2 GPa/100 °C).     

Conditions of pocket formation in the Oktyabrskaya tourmaline-rich gem pegmatite (the Malkhan field, Central Transbaikalia, Russia)

Coexisting melt (MI), fluid-melt (FMI) and fluid (FI) inclusions in quartz from the Oktaybrskaya pegmatite, central Transbaikalia, have been studied and the thermodynamic modeling of PVTX-properties of aqueous orthoboric-acid fluids has been carried out to define the conditions of pocket formation. At room temperature, FMI in early pocket quartz and in quartz from the coarse-grained quartz–oligoclase host pegmatite contain crystalline aggregates and an orthoboric-acid fluid. The portion of FMI in inclusion assemblages decreases and the volume of fluid in inclusions increases from the early to the late growth zones in the pocket quartz. No FMI have been found in the late growth zones. Significant variations of solid/fluid ratios in the neighboring FMI result from heterogeneous entrapment of coexisting melts and fluids by a host mineral. Raman spectroscopy, SEM EDS and EMPA indicate that the crystalline aggregates in FMI are dominated by mica minerals of the boron-rich muscovite–nanpingite CsAl₂[AlSi₃O₁₀](OH,F)₂ series as well as lepidolite. Topaz, quartz, potassium feldspar and several unidentified minerals occur in much lower amounts. Fluid isolations in FMI and FI have similar total salinity (4–8 wt.% NaCl eq.) and H₃BO₃ contents (12–16 wt.%). The melt inclusions in host-pegmatite quartz homogenize at 570–600 °C. The silicate crystalline aggregates in large inclusions in pocket quartz completely melt at 615 °C. However, even after those inclusions were significantly overheated at 650±10 °C and 2.5 kbar during 24 h they remained non-homogeneous and displayed two types: (i) glass+unmelted crystals and (ii) fluid+glass. The FMI glasses contain 1.94–2.73 wt.% F, 2.51 wt.% B₂O₃, 3.64–5.20 wt.% Cs₂O, 0.54 wt.% Li₂O, 0.57 wt.% Ta₂O₅, 0.10 wt.% Nb₂O₅, 0.12 wt.% BeO. The H₂O content of the glass could exceed 12 wt.%. Such compositions suggest that the residual melts of the latest magmatic stage were strongly enriched in H₂O, B, F, Cs and contained elevated concentrations of Li, Be, Ta, and Nb. FMI microthermometry showed that those melts could have crystallized at 615–550 °C.

Crystallization of quartz–feldspar pegmatite matrix leads to the formation of H₂O-, B- and F-enriched residual melts and associated fluids (prototypes of pockets). Fluids of different compositions and residual melts of different liquidus–solidus P–T-conditions would form pockets with various internal fluid pressures. During crystallization, those melts release more aqueous fluids resulting in a further increase of the fluid pressure in pockets. A significant overpressure and a possible pressure gradient between the neighboring pockets would induce fracturing of pockets and “fluid explosions”. The fracturing commonly results in the crushing of pocket walls, formation of new fractures connecting adjacent pockets, heterogenization and mixing of pocket fluids. Such newly formed fluids would interact with a primary pegmatite matrix along the fractures and cause autometasomatic alteration, recrystallization, leaching and formation of “primary–secondary” pockets.     

Derivation of potassic (shoshonitic) magmas by decompression melting of phlogopite+pargasite lherzolite

A model metasomatized lherzolite composition contains phlogopite and pargasite, together with olivine, orthopyroxene, clinopyroxene and spinel or garnet as subsolidus phases to 3 GPa. Previous works established that at ≥1.5 GPa, phlogopite is stable above the dehydration solidus, determined by the melting behaviour of pargasite and coexisting phases. At 2.8 GPa, melts with residual phlogopite+garnet lherzolite mineralogy at 1195 °C and with garnet lherzolite mineralogy at 1250 °C are both olivine nephelinite with K/Na (atomic)=0.51 and K/Na=0.65, respectively. Recent work shows that melting along the dehydration (fluid-absent) solidus of the phlogopite+pargasite lherzolite at pressures <1.5 GPa is very different with the presence of phlogopite, decreasing the solidus below that of pargasite lherzolite. At 1.0 GPa, both phlogopite and pargasite disappear at temperatures at or slightly above the solidus. The compositions of two melts at 1.0 GPa, 1075 °C (with different water contents), in equilibrium with residual spinel lherzolite mineralogy are silica-saturated trachyandesite (5% melt fraction, 3% H₂O) to silica-oversaturated basaltic andesite (8% melt fraction, 4.5% H₂O). Both compositions may be classified as ‘shoshonites’ on the basis of normative compositions, silica-saturation, and K/Na ratio. Decompression melting of metasomatized lithospheric lherzolite with minor phlogopite and pargasite may produce primary ‘shoshonitic’ magmas by dehydration melting at 1 GPa, 1050–1150 °C. Such magmas may be parental to Proterozoic batholithic syenites occurring in Brazil.     

Melting relations and major element partitioning in an oxidized bulk Earth model composition at 15–26 GPa

Melting experiments were performed on an FeO-rich bulk Earth model composition in the CMFAS system in order to investigate the partitioning of major elements between coexisting minerals and melts. The starting material (34.2% SiO₂, 3.86% Al₂O₃, 35.2% FeO, 25.0% MgO and 1.88% CaO), contained in Re-capsules, was a mixture of crystalline forsterite and fayalite, and a glass containing SiO₂, Al₂O₃, and CaO. Olivine is the first liquidus phase at 10 GPa but is replaced by majoritic garnet (ga) in the 15–26 GPa range. Magnesiowüstite (mw) crystallizes close to the liquidus and is joined by perovskite (pv) at 26 GPa.

The quenched melt compositions are homogeneous throughout the melt region of the charges and are only slightly enriched in Si, Ca and Fe, and depleted in Mg, relative to the starting composition. The Fe/Mg and Ca/Al ratios in all of the minerals increase rapidly below the liquidus to become compatible with the bulk composition at the solidus. At 26 GPa, a relative density sequence of mw>pv>melt>ga is observed. This indicates that majorite floating, combined with the sinking of magnesiowüstite and perovskite can be expected during the solidification of a Hadean magma ocean and in hot mantle plumes early in the Earth's history. The mineral–melt partitioning relations indicate that fractional crystallization or partial melting in the transition zone and the upper part of the lower mantle would increase the Fe/Mg and Ca/Al ratios of the melt, even if magnesiowüstite was predominant in the solid fraction. A significant contribution of accumulated mw to the segregation of the protocore is therefore unlikely. The suggested process of perovskite fractionation to the lower mantle is not capable of increasing the Mg/Si ratio in the residual melt, and the combined fractionation of perovskite and magnesiowüstite produces a melt with elevated ratios of Si/Mg, Ca/Al and Fe/Mg.     

Fluid inclusions in mantle xenoliths

Fluid inclusions in olivine and pyroxene in mantle-derived ultramafic xenoliths in volcanic rocks contain abundant CO₂-rich fluid inclusions, as well as inclusions of silicate glass, solidified metal sulphide melt and carbonates. Such inclusions represent accidentally trapped samples of fluid- and melt phases present in the upper mantle, and are as such of unique importance for the understanding of mineral–fluid–melt interaction processes in the mantle. Minor volatile species in CO₂-rich fluid inclusions include N₂, CO, SO₂, H₂O and noble gases. In some xenoliths sampled from hydrated mantle-wedges above active subduction zones, water may actually be a dominant fluid species. The distribution of minor volatile species in inclusion fluids can provide information on the oxidation state of the upper mantle, on mantle degassing processes and on recycling of subducted material to the mantle. Melt inclusions in ultramafic xenoliths give information on silicate–sulphide–carbonatite immiscibility relationships within the upper mantle. Recent melt-inclusion studies have indicated that highly silicic melts can coexist with mantle peridotite mineral assemblages. Although trapping-pressures up to 1.4 GPa can be derived from fluid inclusion data, few CO₂-rich fluid inclusions preserve a density representing their initial trapping in the upper mantle, because of leakage or stretching during transport to the surface. However, the distribution of fluid density in populations of modified inclusions may preserve information on volcanic plumbing systems not easily available from their host minerals. As fluid and melt inclusions are integral parts of the phase assemblages of their host xenoliths, and thus of the upper mantle itself, the authors of this review strongly recommend that their study is included in any research project relating to mantle xenoliths.     

Pressure dependence of melt viscosities on the join diopside-albite

The pressure dependence of melt viscosities on the join diopside-albite has been studied using falling-sphere viscometry. The five melt compositions investigated are: diopside, Ab₂₅Di₇₅, Ab₅₀Di₅₀, Ab₇₅Di₂₅ and albite. Experiments were performed at 1500° and 1600°C and at pressures of 5, 10, 15, 20 and 25 kbar. The positive and negative pressure dependence of the viscosity of diopside and albite, respectively, were confirmed. All intermediate compositions show an initial decrease in viscosity with increasing pressure; however, melt of Ab₂₅Di₇₅ composition passes through a minimum viscosity at approximately 12 kbar and 1600°C. This behavior is analogous to the variation in the viscosity of water with pressure at low temperature.

It is suggested that the three-dimensional, fully polymerized, albite structure dominates flow at low pressures. With increasing pressure, disruption of this structure and decrease in the average size of the flow units leads to domination by the diopside structure. The variation in viscosity with composition along the join at one atmosphere can be adequately modelled using the and (1965) configurational entropy model with an additional two-lattice configurational entropy of mixing term. The pressure dependence of viscosity in the diopside-albite system, however, cannot be predicted by the model, because there is an absence of information on the pressure dependence of the model parameters.

It is probable that relatively polymerized magmas (e.g. rhyolites to SiO₂-saturated basalts) show a negative pressure dependence of viscosity to depths where they originate in the lower crust or upper mantle. In contrast, the most depolymerized, naturally-occurring melts, such as strongly SiO₂-undersaturated basalts and picrites, may exhibit a viscosity minimum. The viscosity of these melts may be sufficiently high at depths within the upper mantle to inhibit their segregation, rise and eventual eruption at the surface.     

Mantle petrology and mineralogy of the Thetford Mines Ophiolite Complex

The Ordovician Thetford Mines ophiolite complex (TMOC) formed by boninite-fed seafloor-spreading, probably in a fore-arc environment. The mantle section is dominated by foliated harzburgite (≤ 5–6% clinopyroxene), cut by dunitic (± chromitite cores) and orthopyroxenitic veins and dykes. Contrasting structures, textures and mineral compositions allow us to subdivide the mantle. The granular-textured rocks of the Duck Lake Block (DLB) have two steeply-dipping foliations. The older foliation strikes NW, is sub-perpendicular to the Moho, and is interpreted to have resulted from upflow of the asthenosphere beneath the spreading ridge. This fabric is overprinted by a 2nd ductile foliation striking ENE, oriented sub-parallel to the Moho, which we interpreted as having formed by crust–mantle shear as the lithosphere migrated away from the spreading ridge. The DLB mantle has a limited range of spinel Cr# (100Cr / (Cr + Al) = 51–71). Comparison with experimentally determined residual spinel compositions (equilibrium melting) implies a maximum loss of 27–38% melt if the protolith had a fertile MORB mantle composition. However, interstitial-textured clinopyroxene may have high TiO₂ (< 0.04wt.%) and Na₂O (< 0.27wt.%), and some interstitial spinel has higher TiO₂ (< 0.09wt.%), suggesting interaction with (or crystallization from) an “impregnating” melt. Interstitial tremolitic amphibole also indicates the passage of late hydrous fluids. The harzburgite in the Caribou Mountain Block (CMB) has a porphyroclastic texture, with a strong, locally mylonitic foliation striking roughly N–S, parallel to the orientation of seafloor-spreading related paleo-normal faults in the crust. These fabrics and textures imply a colder, lithospheric deformation, possibly related to tectonic denudation (oceanic core complex). This would explain problematic lava/mantle contacts, favour infiltration of seawater, serpentinization, and reduced fO₂ conditions. The CMB mantle shows a wider range of mineral compositions than the DLB, with spinel Cr# (28–86) implying ≤ 15–45% of equilibrium melting. Locally higher TiO₂ in spinel (< 0.05wt.%) and clinopyroxene (< 0.11wt.%), a local rimward decrease in spinel Cr#, clinopyroxene Cr#, and olivine Fo-content, and traces of interstitial amphibole, are attributed to the circulation of an evolved hydrous melt during peridotite deformation. This suggests that the lower limit to the extent of melting inferred for the CMB (15%), established on the basis of Al-rich spinel rims and neoblasts, is probably too low. On the other hand, the higher inferred degree of depletion of the CMB is probably unaffected by the metasomatic overprint and is a more robust conclusion.     

Origin of lamprophyres by the mixing of basic and alkaline melts in magma chamber in Beiya area, western Yunnan, China

Lamprophyres consisting mainly of diopside, phlogopite and K-feldspar formed in the early Tertiary around 60 Ma in the Beiya area and are characterized by low SiO₂ ± 46–50 wt.%), Rb (31–45 ppm) and Sr (225–262 ppm), high Al₂O₃, (11.2–13.1 wt.%), CaO (8.0–8.7 wt.%), MgO (11.5–12.1 wt.%), K₂O(4.9–5.5 wt.%), TiO₂ (2.9–3.3 wt.%) and REE (174–177 ppm), and compatible elements (e.g. Sc, Cr and Ni) and HSF elements (e.g. Th, U, Zr, Nb, Ta, Ti and Y), and low ¹⁴³Nd/¹⁴⁴Nd 0.512372–0.512536, middle ⁸⁷Sr/⁸⁶Sr 0.707322–0.707395, middle ²⁰⁶Pb/²⁰⁴Pb 18.50–18.59, ²⁰⁷Pb/²⁰⁴Pb 15.60–15.65 and ²⁰⁸Pb/²⁰⁴Pb 38.75–38.8. These rocks developed peculiar quartz megacrysts with poly-layer reaction zones, melt inclusions, and partial melted K-feldspar and plagioclase inclusions, and plastic shapes. Important features of these rocks include: (1) hybrid composition of elements, (2) abrupt increase of SiO₂ content of the melt, recorded by zoned diopside, (3) development of sanidine and aegirine-augite reaction zones, (4) alkaline melt and partial melted K-feldspar and plagioclase inclusions, (5) deformed quartz inclusions associated with quartz megacrysts, (6) the presence of quartz megacrysts in plastic shape with their parent melts, (7) the occurrence of olivine, high-MgO ilmenite and spinel inclusions within earlier formed diopside, phlogopite and magnetite. Median ⁸⁷Sr/⁸⁶Sr values between Tertiary alkaline porphyries in the Beiya area and the western Yunnan and Tertiary basalt in the western Yunnan indicate that the Beiya lamprophyre melts were derivative and resulted from the mixing between basic melts that were related to the partial melting of phenocrysts of spinel iherzolite from a mantle source. The alkaline melts originated from partial melting along the Jinshajiang subduction ductile shear zone at the contact between the buried Palaeo-Tethyan oceanic lithosphere and the upper mantle lithosphere. The alkaline melts are composed of 65% sanidine (Or₇₀Ab₂₈An₂) and 35% SiO₂. The melt mixing occurred in magma chambers in the middle-shallow crust at 8–10 km before the derivative lamprophyre melts intruded into the shallow cover in Beiya area. This mixing of basic and alkaline melts might represent a general process for the formation of lamprophyre in the western Yunnan.     

Geochemical constraints on restite composition and unmixing in the Velay anatectic granite, French Massif Central

The main anatectic granite of the Velay complex is unique among major French Massif Central Hercynian granitoids in that rather than having an entirely lower crustal source, it formed by mixing between partial melts of the meta-igneous lower crust and ‘upper crustal’ country rock schists and orthogneisses. The geochemical variations in the Velay main anatectic granites cannot, however, be explained by mixing alone as their compositions range to lower SiO₂, with higher Al₂O₃, Fe₂O₃ and TiO₂ and lower Na₂O and CaO, than either end member in mixing. The variations are interpreted as being due to the presence of up to 35% restite in minimum melts of country rock compositions. Primary restites form equilibrium assemblages represented by biotite, ilmenite and surmicaceous enclaves which consist of biotite ± apatite, zircon and almandine. The main anatectic granites more rarely contain schist and gneiss enclaves, quartz resisters and plagioclase restites. Secondary restites are mainly represented by cordierite, and possibly K-feldspar, which formed by recrystallisation of primary biotite-rich restites. The unique characteristics of the Velay main anatectic granites are likely to be due, in part, to its late formation close to the end of the Hercynian orogeny. The metasedimentary lower crust may have become too refractory to yield large volumes of melt following partial melting to form the other major Massif Central granitoids. The heat necessary for partial melting at higher crustal levels was transferred from the lower crust by the intrusion of I-type granites and low volume diorites from the mantle. Upper crustal anatexis was mainly controlled by muscovite breakdown reactions (< 830 to 850 °C) and the liberation of water due to the recrystallisation of biotite to cordierite. The temperatures necessary for biotite breakdown were only achieved locally and resulted in the formation of high-LREE granites.     

SIMS determination of trace element partition coefficients between garnet, clinopyroxene and hydrous basaltic liquids at 2–7.5 GPa and 1080–1200°C

Trace element partition coefficients (D's) for up to 13 REE, Nb, Ta, Zr, Hf, Sr and Y have been determined by SIMS analysis of seven garnets, four clinopyroxenes, one orthopyroxene and one phlogopite crystallized from an undoped basanite and a lightly doped (200 ppm Nb, Ta and Hf) quartz tholeiite. Experiments were conducted at 2–7.5 GPa, achieving near-liquidus crystallization at relatively low temperatures of 1080–1200°C under strongly hydrous conditions (5–27 wt.% added water). Garnet and pyroxene D_REE show a parabolic pattern when plotted against ionic radius, and conform closely to the lattice strain model of Blundy and Wood (Blundy, J.D., Wood, B.J., 1994. Prediction of crystal–melt partition coefficients from elastic moduli. Nature 372, 452–454). Comparison, at constant pressure, between hydrous and anhydrous values of the strain-free partition coefficient (D₀) for the large cation sites of garnet and clinopyroxene reveals the relative importance of temperature and melt water content on partitioning. In the case of garnet, the effect of lower temperature, which serves to increase D₀, and higher water content, which serves to decrease D₀, counteract each other to the extent that water has little effect on garnet–melt D₀ values. In contrast, the effect of water on clinopyroxene–melt D₀ overwhelms the effect of temperature, such that D₀ is significantly lower under hydrous conditions. For both minerals, however, the lower temperature of the hydrous experiments tends to tighten the partitioning parabolas, increasing fractionation of light from heavy REE compared to anhydrous experiments.

Three sets of near-liquidus clinopyroxene–garnet two-mineral D values increase the range of published experimental determinations, but show significant differences from natural two-mineral D's determined for subsolidus mineral pairs. Similar behaviour is observed for the first experimental data for orthopyroxene–clinopyroxene two-mineral D's when compared with natural data. These differences are in large part of a consequence of the subsolidus equilibration temperatures and compositions of natural mineral pairs. Great care should therefore be taken when using natural mineral–mineral partition coefficients to interpret magmatic processes.

The new data for strongly hydrous compositions suggest that fractionation of Zr–Hf–Sm by garnet decreases with increasing depth. Thus, melts leaving a garnet-dominated residuum at depths of about 200 km or greater may preserve source Zr/Hf and Hf/Sm. This contrasts with melting at shallower depths where both garnet and clinopyroxene will cause Zr–Hf–Sm fractionation. Also, at shallower depths, clinopyroxene-dominated fractionation may produce a positive Sr spike in melts from spinel lherzolite, but for garnet lherzolite melting, no Sr spike will result. Conversely, clinopyroxene megacrysts with negative Sr spikes may crystallize from magmas without anomalous Sr contents when plotted on mantle compatibility diagrams. Because the characteristics of strongly hydrous silicate melt and solute-rich aqueous fluid converge at high pressure, the hydrous data presented here are particularly pertinent to modelling processes in subduction zones, where aqueous fluids may have an important metasomatic role.     

Solubility of cassiterite in evolved granitic melts: effect of T, fO2, and additional volatiles

The aim of this experimental study was to determine the solubility of cassiterite in natural topaz- and cassiterite-bearing granite melts at temperatures close to the solidus. Profiles of Sn concentrations at glass–crystal (SnO₂) interface were determined following the method of (Harrison, T.M., Watson, E.B., 1983. Kinetics of zircon dissolution and zirconium diffusion in granitic melts of variable water content. Contributions to Mineralogy and Petrology 84, 66–72). The cassiterite concentration calculated at the SnO₂–glass interface is the SnO₂ solubility. Experiments were performed at 700–850 °C and 2 kbar using a natural F-bearing peraluminous granitic melt with 2.8 wt.% normative corundum. Slightly H₂O-undersaturated to H₂O-saturated melt compositions were chosen in order to minimize the loss of Sn to the noble element capsule walls. At the nickel–nickel oxide assemblage (Ni–NiO) oxygen fugacity buffer, the solubility of cassiterite in melts containing 1.12 wt.% F increases from 0.32 to 1.20 wt.% SnO₂ with an increasing temperature from 700 to 850 °C. At the Ni–NiO buffer and a given corundum content, SnO₂ solubility increases by 10% to 20% relative to an increase of F from 0 to 1.12 wt.%. SnO₂ solubility increases by 20% relative to increasing Cl content from 0 to 0.37 wt.% in synthetic granitic melts at 850 °C. We show that Cl is at least as important as F in controlling SnO₂ solubility in evolved peraluminous melts at oxygen fugacities close to the Ni–NiO buffer. In addition to the strong effects of temperature and fO₂ on SnO₂ solubility, an additional controlling parameter is the amount of excess Al (corundum content). At Ni–NiO and 850 °C, SnO₂ solubility increases from 0.47 to 1.10 wt.% SnO₂ as the normative corundum content increases from 0.1 to 2.8 wt.%. At oxidizing conditions (Ni–NiO +2 to +3), Sn is mainly incorporated as Sn⁴⁺ and the effect of excess Al seems to be significantly weaker than at reducing conditions.