Partial melting has been shown to be an important mechanism for intracrustal differentiation and granite petrogenesis. However, a series of compositional differences between granitic melt from experiments and natural granites indicate that the processes of crustal differentiation are complex. To shed light on factors that control the processes of crustal differentiation, and then the compositions of granitic magma, a combined study of petrology and geochemistry was carried out for granites (in the forms of granitic veins and parautochthonous granite) from a granulite terrane in the Tongbai orogen, China. These granites are characterized by high SiO2 (>72 wt%) and low FeO and MgO (<4 wt%) with low Na2O/K2O ratios (<0.7). Minerals in these granites show variable microstructures and compositions. Phase equilibrium modelling using P–T pseudosections shows that neither anatectic melts nor fractionated melts match the compositions of the target granites, challenging the conventional paradigm that granites are the crystallized product of pure granitic melts. Based on the microstructural features of minerals in the granites, and a comparison of their compositions with crystallized minerals from anatectic melts and minerals in granulites, the minerals in these granitoids are considered to have three origins. The first is entrained garnets, which show comparable compositions with those in host granulites. The second is early crystallized mineral from melts, which include large plagioclase and K-feldspar (with high Ca contents) crystals as well as a part of biotite whose compositions can be reproduced by crystallization of the anatectic melts. The compositions of other minerals such as small grained plagioclase, K-feldspar and anorthoclase in the granites with low Ca contents are not well reconstructed, so they are considered as the third origin of crystallized products of fractionated melts. The results of mass balance calculation show that the compositions of these granites can be produced by mixing between different proportions of crystallized minerals and fractionated melts with variable amounts of entrained minerals. However, the calculated modal proportions of different crystallized minerals (plagioclase, K-feldspar, biotite and quartz) in the granites are significantly different from those predicted by melt crystallization modelling. Specifically, some rocks have lower modes of biotite and plagioclase, whereas others show lower K-feldspar modes than those produced by melt crystallization. This indicates that the crystallized minerals would be differentially separated from the primary magmas to form the evolved magmas that produce these granites. Therefore, the crystal entrainment and differential melt-crystal separation make important contributions to the composition of the target granites. Compared with leucogranites worldwide, the target granites show comparable compositions. As such, the leucogranites may form through the crystal fractionation of primary granitic magmas at different extents in addition to variable degrees of partial melting.

Two eclogite samples from the Haut-Allier record a prograde evolution from ~20 kbar, 650°C to 750°C, 22–23 kbar followed by heating up to 850–875°C and partial melting. Incipient decompression in high-pressure granulite facies conditions (19.5 kbar, 875°C) was followed by exhumation to high-temperature amphibolite facies conditions (<9 kbar, 750–850°C). Following a detailed geochemical, petrological, and geochronological investigation using trace-element data and laser ablation inductively coupled plasma mass spectrometry U–Pb dating of zircon, apatite, and rutile, the eclogites reveal an Ordovician (c. 490 Ma) rifting event followed by Devonian (c. 370–360 Ma) subduction and Carboniferous (c. 350 Ma) exhumation in this part of the French Massif Central. The previously proposed Silurian age for the subduction, which strongly influenced many tectonic models, is definitively rejected. In the light of other geological data from the French Massif Central, including the lithological and geochemical zoning of calc-alkaline Devonian volcanism, we propose a southward polarity of the subduction and question the very existence of the so-called Massif Central Ocean. Furthermore, we infer that following subduction, the eclogites were relaminated to the upper plate and exhumed at the rear of the magmatic arc pointing to similarities with the geodynamics of the Bohemian Massif.

The trace-element composition of rutile is commonly used to constrain P–T–t conditions for a wide range of metamorphic systems. However, recent studies have demonstrated the redistribution of trace elements in rutile via high-diffusivity pathways and dislocation-impurity associations related to the formation and evolution of microstructures. Here, we investigate trace-element migration in low-angle boundaries formed by dislocation creep in rutile within an omphacite vein of the Lago di Cignana unit (Western Alps, Italy). Zr-in-rutile thermometry and inclusions of quartz in rutile and of coesite in omphacite constrain the conditions of rutile deformation to around the prograde boundary from high pressure to ultra-high pressure (~2.7 GPa) at temperatures of 500–565°C. Crystal-plastic deformation of a large rutile grain results in low-angle boundaries that generate a total misorientation of ~25°. Dislocations constituting one of these low-angle boundaries are enriched in common and uncommon trace elements, including Fe and Ca, providing evidence for the diffusion and trapping of trace elements along the dislocation cores. The role of dislocation microstructures as fast-diffusion pathways must be evaluated when applying high-resolution analytical procedures as compositional disturbances might lead to erroneous interpretations for Ca and Fe. In contrast, our results indicate a trapping mechanism for Zr.

Melt inclusions (MIs) in high-temperature metamorphic rocks provide a unique window into crustal anatexis in collisional orogenic belts and have been widely used to characterize compositions of anatectic melts as well as melting mechanisms. In this study, MIs hosted by peritectic garnet were for the first time identified in an Al2SiO5-free graywacke-type paragneiss from the Namche Barwa Complex, the Eastern Himalaya, Southeast Tibet. These MIs occur as nanogranites in the rims of porphyroblastic garnet, exhibit negative crystal shapes with an average diameter of ~12 μm and consist of a mineral assemblage of biotite + quartz + plagioclase + K-feldspar ± muscovite. Re-homogenization experiments of these nanogranites were conducted at a pressure of 1.5 GPa and temperatures of 800°C, 850°C and 900°C and produced homogeneous glasses at 850°C. The homogenized glasses are strongly peraluminous and calc-alkalic in composition, with 66.43–71.31 wt.% SiO2, 12.64–15.06 wt.% Al2O3, high alkaline (5.41–7.22 wt.%) and low ferromagnesian (2.72–4.46 wt.%) contents. They are lower in silica and CaO but higher in K2O compared with MI produced by fluid-present melting of metasedimentary rocks, thus indicating fluid-absent melting. These glasses are also characterized by enrichment of large ion lithophile elements (particularly Cs and Rb), depletion of Ba and Sr, low contents of light rare earth elements (3.6 to 33.7 ppm), high Rb/Sr ratios (6.19–37.3) and low Nb/Ta ratios (2.55–18.7). In combination with phase equilibrium modelling, these compositional features suggest that a sequential dehydration melting of muscovite and biotite was responsible for the production of MI during prograde metamorphism of the studied paragneiss. By compiling MI data published in the literature, we show that dehydration melting of metasedimentary rocks from the Himalayan orogen can produce initial melts with various peraluminous and granitic compositions.

Recent studies have demonstrated that kinetic factors can significantly influence garnet nucleation, delaying its appearance with respect to the equilibrium predictions. Overstepping of garnet nucleation occurs in both contact and regional metamorphic settings, with extremely variable degrees: The factors controlling such highly variable degrees of overstepping are still unclear. This study focuses on garnet nucleation and growth in aluminous metapelites from the Barrovian inverted metamorphic pile of the upper portion of the Lesser Himalayan Sequence (Upper-LHS; central Nepal), with the aim of (i) investigating if (and how) the bulk-rock composition can influence overstepping of garnet nucleation and (ii) which are the implications of garnet nucleation overstepping for the P–T evolution of a Barrovian metamorphic sequence. Detailed petrographic, microstructural and compositional data are presented for six phyllitic schists, containing porphyroblasts of garnet, staurolite and/or kyanite. Their P–T evolution is constrained through thermodynamic forward modelling (i.e., isochemical phase diagrams combined with isopleth thermobarometry), assuming that equilibrium was attained at every stage of their metamorphic evolution. Comparison between the P–T conditions inferred for the growth of garnet core, the assemblages predicted to be stable at these P–T conditions and the modelled garnet-in reaction boundary suggests that the studied samples have experienced different degrees of apparent thermal (ΔT) and/or baric (ΔP) overstepping of garnet nucleation. We suggest that the bulk-rock MnO and CaO amounts might have influenced the apparent ΔT and ΔP overstepping of the garnet-in reaction: more specifically, the higher the bulk-rock MnO content, the more pronounced the apparent ΔT overstepping, whereas the lower the bulk-rock CaO content, the greater the ΔP overstepping. However, rather than an effective delay of garnet appearance with respect to equilibrium predictions, the apparent ΔT overstepping of garnet nucleation could reflect the attainment of the critical 0.5% threshold of garnet abundance, below which garnet is not readily detected in thin section. Kinetic factors seem much less critical in controlling the growth of the garnet rim at peak P–T conditions, confirming that peak metamorphic conditions constrained through equilibrium approaches based on the composition of garnet rim and of the matrix assemblage can be considered as reliable. Overall, the P–T paths of the studied samples are characterized by prograde heating coupled with tectonic overload (peak-P conditions of 9.5–10.5 kbar, 580–590°C), followed by heating during exhumation (peak-T conditions of 8.2–8.9 kbar, 610–630°C), supporting those thermo-mechanical models that predict a period of slowdown (or quiescence) of the Main Central Thrust activity.

Based on the nature of ferric sillimanite, an activity model for sillimanite containing Fe3+ is constructed, tested and adopted to calculate phase equilibria of pelitic compositions under ultrahigh-temperature (UHT) conditions. The calculated P–T projections and pseudosections suggest that the incorporation of Fe3+ into sillimanite can fairly solve the current imperfectly topological match between thermodynamic calculations and synthetic experiments, especially at high oxygen fugacity. Fe3+ in sillimanite remarkably elevates the temperature to switch the parageneses of orthopyroxene + sillimanite (Opx + Sil) and sapphirine + quartz (Spr + Qz) in oxidized metapelites, with an increment around 50–70°C. The calculated compatibility diagrams show that the widely approbatory UHT-diagnostic mineral assemblages of Opx + Sil and Spr + Qz usually occur in metapelites with high Mg/Fe2+ ratios, which depend on both bulk-rock MgO and oxygen fugacity, whereas the metapelites with low Mg/Fe2+ ratios are characterized by the assemblage of Garnet + sillimanite (Grt + Sil) with or without spinel (Spl) in UHT conditions. Moreover, comprehensive comparisons suggest that the essential petrogenetic framework of natural UHT metapelites is mostly governed by the two metamorphic reactions of Opx + Sil = Spr + Grt and Grt + Sil = Spl + Spr.

Melt migration through high-strain zones in the crust fundamentally influences their rheological behaviour and is important for the transfer of fluids to upper crustal regions. The inference of former melt-present deformation, based on field observations, may be hampered if the high-strain zone experience a low time-integrated melt flux or high melt volume expulsion during deformation. In these cases, typical macro-scale field evidence of former melt presence limits interpretations. In this contribution, we investigate igneous field evidence ranging from obvious to cryptic in the Gough Dam shear zone (central Australia), a 2- to 4-km-wide high-strain zone shown to have acted as a significant melt pathway during the Alice Springs Orogeny. Within bands of the high-strain zone, granitic lenses are easily discernible in the field and are inferred to have formed during melt present deformation. Related coarse K-feldspar is observed in biotite-rich (>75 vol%) schist (glimmerite) as either isolated grains, forming trails (sub)parallel to the main foliation, or in aggregates with subordinate quartz. Detailed characterization of the granitic lenses shows that pockets of phenocrysts may be entrained in the shear zone. If melt expulsion and melt-rock interaction is severe, isolated K-feldspar grains in glimmerite may form. These grains exhibit (i) partially preserved crystal faces; (ii) a lack of internal grain deformation; (iii) reaction textures preferentially formed along the main crystallographic axes showing dissolution of K-feldspar and precipitation of dominantly biotite; (iv) low-strain domains between multiple K-feldspar grains being inferred to enclose crystallized melt pockets, with some apparently isolated grains showing connectivity in three dimensions; and (v) a weak quartz and K-feldspar crystallographic preferred orientation. These observations suggest an igneous phenocrystic origin for the isolated K-feldspar grains hosted in glimmerite, which is consistent with the observed REE concentration patterns with positive Eu anomaly. We propose that the K-feldspar phenocrysts are early-formed crystals that were entrained into the glimmerite rocks as reactive melt migrated through the actively deformatting high-strain zone. Previously entrained K-feldspar phenocrysts were trapped during the collapse of the melt pathway when melt flux-related fluid pressure waned while confining pressure and tectonic stress were still significant. The active deformation facilitated expulsion or loss of the melt phase but retainment and trapping of phenocrysts. Hence, the presence of isolated or ‘trains’ of K-feldspar phenocrysts is a cryptic signature of syndeformational melt transfer. If melt transfer occurs in an open chemical system, phenocrysts will be entrained within the reaction product of melt-rock interaction. We suggest that these so-called trapped phenocrysts are a viable indicator of former syntectonic melt passage through rocks.

Continental subduction and collision are recorded by ultrahigh-pressure (UHP) terranes; UHP terranes that form at early stages of an orogeny tend to be small and experience short residence at eclogite-facies depths, whereas terranes that form at mature stages of an orogeny tend to be larger and experience longer residence at these depths, but accurately determining eclogite-facies residence time requires a large geochronologic dataset tied to metamorphic conditions (via trace elements and/or inclusions). In the Dulan area, North Qaidam UHP terrane, China, it remains unclear whether the terrane experienced a long residence at eclogite-facies depths, marking the mature stage of an orogeny or two distinct (ultra)high pressure ([U]HP) events (with short residence times), interpreted as the transition from oceanic subduction to continental collision, where one (U)HP event is related to the former and second (U)HP event to the latter. To address this issue, we report new zircon U–Pb ages and trace-element data from eclogite and host paragneiss from the Dulan area and show that this terrane records ~42 Myr of eclogite-facies metamorphism at (U)HP conditions, similar to other large UHP terranes. Zircon from 11 eclogite and 2 gneiss samples yields weighted mean ages of 463–425 Ma, flat heavy rare earth element (HREE) patterns without negative Eu anomalies, and eclogitic mineral inclusions, indicating eclogite-facies conditions. Paragneiss metamorphic ages overlap with ages from eclogite but are generally younger, suggesting that a lack of internally generated fluids may have inhibited zircon growth and/or recrystallization until early decompression and white mica consumption in felsic gneiss generated fluids; thus, we interpret that these felsic rocks record the later stages of continental collision. Dataset patterns from all new and previously published analyses for the Dulan area (34 eclogite and 14 gneiss) suggest that metamorphic zircon in eclogite records prograde, peak and possibly early retrograde conditions, in contrast to the prediction from mass balance models that metamorphic zircon should only grow during exhumation and cooling. We reconcile our observations with these model predictions by recognizing that differential solubility can lead to grain-scale zircon growth or recrystallization over a large segment of the pressure–temperature (P–T) path even where zircon abundance decreases at the whole-rock scale.

Accurately defining the peak ages and timescales of high-temperature metamorphism is fundamental to unravelling tectonic dynamics. However, metamorphic constraints are frequently hampered by a large spread of zircon U–Pb ages without explicit textural relationships. Integrated garnet and zircon petrochronology may clarify ambiguous ages retrieved from ancient high-temperature metamorphic rocks. There is a long-standing debate on the interpretation of the spread of zircon ages from c. 2.5–1.8 Ga for the granulites of the North China Craton. In order to clarify the timing and duration of (ultra)high-temperature metamorphism in the North China Craton, we investigated a mafic granulite and the adjoining gneiss from the Yinshan Block of the North China Craton using zircon and titanite U–Pb geochronology combined with garnet Lu–Hf and Sm–Nd geochronology. Pseudosection modelling and conventional thermobarometric calculations constrain the peak metamorphic conditions to be ~1.0 GPa and ~850°C. The near-complete lack of major-element zoning in garnet, aside from ~2 μm diffusion profiles at crystal rims, suggests complete re-equilibration at peak temperatures followed by fast cooling from high temperatures. The Lu–Hf garnet age of 1870 ± 4 Ma and Sm–Nd age of 1870 ± 7 Ma, determined on the same garnet fractions, are indistinguishable from the zircon U–Pb age of 1866 ± 11 Ma obtained from zircon that grew contemporaneously with garnet, evidenced by the chemical equilibrium of coexisting garnet and zircon, and are additionally consistent with a titanite U–Pb age of 1876 ± 7 Ma. We interpret this close agreement of ages, within uncertainty, coupled to the existence of flat Sm–Nd–Hf profiles in garnet that also has well-preserved Lu zoning, to reflect a short-lived high-temperature metamorphic event that was terminated by rapid exhumation and cooling. The short-lived (<4 Myr) high-temperature metamorphism may be generated in the lowermost parts of the crust through magmatic underplating/intraplating during extension that follows collision of the Ordos and the Yinshan Blocks.

We report the first occurrence of poly-cyclic high-pressure low-temperature (HP-LT) rocks from the easternmost Indus-Yarlung suture zone, formed during subduction of Neo-Tethyan oceanic lithosphere. Petrology, mineral composition and P–T pseudosection modelling reveal two low-temperature eclogite facies metamorphic events with an initial high-pressure P–T condition of 16.4–18.7 kbar and 510–520°C, exhumation to 10.5–12.0 kbar and 580–590°C and a subsequent second high-pressure P–T condition of ~16 kbar and ~560°C and exhumation to ≤9 kbar and ≤600°C. This history implies a complex ‘yo-yo type’ P–T path. In situ monazite dating and textural relationships show that late-stage exhumation, cooling and garnet breakdown occurred at c. ~25–22 Ma. We interpret the first burial event to represent subduction of the Neo-Tethys Ocean at the eastern Indus-Yarlung suture zone. Initial exhumation, reburial and final exhumation represent material transport in a large-scale convective circulation system in the subduction channel. Convective overturn in the subduction channel evidently serves both as a mechanism to produce poly-cyclic metamorphism and to exhume LT eclogite facies rocks.