Phase equilibria modelling of kyanite‐bearing anatectic paragneisses from the central Grenville Province

Kyanite‐bearing paragneisses from the Manicouagan Imbricate Zone and its footwall (high‐P belt of the central Grenville Province) preserve evidence of partial melting with development of metamorphic textures involving biotite–garnet ± kyanite ± plagioclase ± K‐feldspar–quartz. Garnet in these rocks displays a variety of zoning patterns with respect to Ca. Pseudosection modelling in the Na₂O–CaO–K₂O–FeO–MgO–Al₂O₃–SiO₂–H₂O–TiO₂–O (NCKFMASHTO) system using measured bulk rock compositions accounts for the textural evolution of two aluminous and two sub‐aluminous samples from the presumed thermal peak to conditions at which retained melt solidified. The prograde features are best explained by pseudosections calculated with compositions to account for melt loss. The intersection of isopleths of grossular content and Fe/(Fe + Mg) relating to large porphyroblasts of garnet provide constraints on the P–T conditions of the metamorphic peak. These P–T estimates are considered to be minima because of the potential for diffusional modification of the composition of garnet at high‐T and during the early stages of cooling. However, they are consistent with textural observations and pseudosection topology, with peak assemblages best preserved in rocks for which the calculated pseudosections predict only small changes in mineral proportions in the P–T interval, in which retrograde reactions are inferred to have occurred between the thermal peak and the solidus. Maximum P–T conditions (14.5–15.5 kbar and 840–890 °C) and steep retrograde P–T paths inferred for rocks from the Manicouagan Imbricate Zone are comparable with those determined for mafic rocks from the same area. In contrast, maximum P–T conditions of 12.5–13 kbar and 815–830 °C and flatter P–T paths are inferred for the rocks of the footwall to the Manicouagan Imbricate Zone. The general consistency between textures, mineral compositions and the topologies of the calculated pseudosections suggests that the pseudosection approach is an appropriate tool for inferring the P–T evolution of high‐P anatectic quartzo‐feldspathic rocks.     

The reaction history of kyanite in high‐P aluminous granulites

Cathodoluminescence (CL) mapping of kyanite in high pressure, aluminous granulites from the central Grenville Province reveals internal structures that are linked to their metamorphic reaction history. In two samples, individual kyanite crystals are shown to be composite porphyroblasts comprising three distinct generations, defined by their CL intensity and Cr (±V, Ti, Fe and Ga) content, and each separated by resorbed interfaces. In contrast, a sub‐aluminous sample contains two types of kyanite, one as resorbed inclusions in garnet and another in the groundmass or replacing garnet. These textural variants of kyanite are interpreted within the framework of phase equilibria modelling. In P–T pseudosections, a first generation of kyanite, which is only present in the most aluminous samples, is potentially linked to staurolite breakdown, and its resorption is consistent with a subsequent increase in pressure. This kyanite represents the earliest remnant of prograde metamorphism identifiable in these rocks. The second generation, present in the porphyroblasts in the same samples and as inclusions in garnet in the sub‐aluminous sample, is interpreted to be the peritectic product of muscovite dehydration melting. Resorption of this kyanite is consistent with subsequent continuous dehydration melting of biotite, which is also inferred based on microstructural considerations. The final generation of kyanite, present as rims on the prograde kyanite porphyroblasts in aluminous samples and as part of the groundmass or replacing garnet in the sub‐aluminous rock, is interpreted to have grown during melt crystallization upon retrogression. The presence of retrograde kyanite implies that the melt crystallized over a wide range of temperatures, and provides an important constraint on the P–T conditions of the metamorphic peak and on the retrograde P–T path. CL mapping is crucial for identifying retrograde kyanite in aluminous samples, as it preferentially overgrows existing kyanite rather than replacing other prograde phases. The scarcity of kyanite in sub‐aluminous rocks allows retrograde kyanite to grow as discrete crystals that can be identified by optical microscopy. This work attests to the potential of unconventional tools such as CL imaging for deciphering the metamorphic history of rocks.     

Factors affecting preservation of coesite in ultrahigh‐pressure metamorphic rocks: Insights from TEM observations of dislocations within kyanite

To understand the preservation of coesite inclusions in ultrahigh‐pressure (UHP) metamorphic rocks, an integrated petrological, Raman spectroscopic and focussed ion beam (FIB) system–transmission electron microscope (TEM) study was performed on a UHP kyanite eclogite from the Sulu belt in eastern China. Coesite grains have been observed only as rare inclusions in kyanite from the outer segment of garnet and in the matrix. Raman mapping analysis shows that a coesite inclusion in kyanite from the garnet rim records an anisotropic residual stress and retains a maximum residual pressure of ~0.35 GPa. TEM observations show quartz is absent from the coesite inclusion–host kyanite grain boundaries. Numerous dislocations and sub‐grain boundaries are present in the kyanite, but dislocations are not confirmed in the coesite. In particular, dislocations concentrate in the kyanite adjacent to the boundary with the coesite inclusion, and they form a dislocation concentration zone with a dislocation density of ~10⁹ cm^?2. A high‐resolution TEM image and a fast Fourier transform‐filtered image reveal that a tiny dislocation in the dislocation concentration zone is composed of multiple edge dislocations. The estimated dislocation density in most of the kyanite away from the coesite inclusion–host kyanite grain boundaries is ~10⁸ cm^?2, being lower than that in kyanite adjacent to the coesite. In the case of a coesite inclusion in a matrix kyanite, using Raman and TEM analyses, we could not identify any quartz at the grain boundaries. Dislocations are not observed in the coesite, but numerous dislocations and stacking faults are developed in the kyanite. The estimated overall dislocation density in the coesite‐bearing matrix kyanite is ~10⁸ cm^?2, but a high dislocation density region of ~10⁹ cm^?2 is also present near the coesite inclusion–host kyanite grain boundaries. Inclusion and matrix kyanite grains with no coesite have dislocation densities of ≤10⁸ cm^?2. Dislocation density is generally reduced during an annealing process, but our results show that not all dislocations in the kyanite have recovered uniformly during exhumation of the UHP rocks. Hence, one of the key factors acting as a buffer to inhibit the coesite to quartz transformation is the mechanical interaction between the host and the inclusion that lead to the formation of dislocations in the kyanite. The kyanite acts as an excellent pressure container that can preserve coesite during the decompression of rocks from UHP conditions. The search for and study of inclusions in kyanite may be a more suitable approach for tracing the spatial distribution of UHP metamorphic rocks.     

Dehydration of metapelites during high‐P metamorphism: The coupling between fluid sources and fluid sinks

Polymetamorphic metapelites and embedded eclogites share a complex, episodic interplay of dehydration and fluid infiltration at the eclogite type‐locality (Saualpe–Koralpe, Eastern Alps, Austria). The metapelites inherited a fluid content (i.e. mineral‐bound OH expressed in terms of mol.% H₂O) of ～6–7 mol.% H₂O from high‐T–low‐P metamorphism experienced during the Permian. At or near P_max of the subsequent Eoalpine event (～20 kbar and 680^°C), the breakdown of paragonite to Na‐rich clinopyroxene and kyanite in metapelites released a discrete pulse of hydrous fluid. Prior to the dehydration event, the rocks were largely fluid absent, allowing only limited re‐equilibration during the prograde Eoalpine evolution. Similarly, Permian‐aged gabbros have persisted metastably due to the absence of a catalyst prior to fluid‐induced re‐equilibration. The fluid triggered partial to complete eclogitization along a fluid infiltration front partially preserved in metagabbro. Near‐isothermal decompression to ～7.5–10 kbar and 670–690°C took place under fluid‐absent conditions. After decompression, a second breakdown of phengitic white mica and garnet produced muscovite, biotite, plagioclase and ～0.1–0.7 mol.% H₂O that enhanced extensive fluid‐aided re‐equilibration of the metapelites. Potential relicts of high‐P assemblages were largely obliterated and replaced by the recurrent amphibolite facies assemblage garnet+biotite+staurolite+kyanite+muscovite+plagioclase+ilmenite+quartz. The hydrous fluid originating from the metapelites infiltrated the embedded eclogites at these P–T conditions and induced the local breakdown of the peak assemblage omphacite and garnet to fine‐grained symplectites of diopside and plagioclase. Further fluid infiltration led to the formation of hornblende–quartz poikiloblasts at the expense of the symplectites. The metapelites re‐equilibrated until the growth of retrograde staurolite consumed any remaining free fluid, thereby terminating the process. Further re‐equilibration is inhibited by both the lack of a catalytic fluid and H₂O as a reactant essential for rehydration reactions. The interplay between fluid sources and fluid sinks describes a closed cycle for the rocks at the eclogite type‐locality. Final, near‐isobaric cooling is indicated by a slight increase of X_Fe in garnet rims. Post‐decompression dehydration and fluid‐aided re‐equilibration arrested by the introduction of staurolite might explain the apparently homogeneous retrogression conditions as well as the notorious absence of diagnostic high‐P assemblages in metapelites at the eclogite type‐locality.     

Multi-stage pseudomorphic replacement of garnet during polymetamorphism: 1. Microstructures and their interpretation

Metapelites from the southern aureole of the Vedrette di Ries tonalite (eastern Alps) were variably overprinted by contact and earlier regional metamorphic events during pre-Alpine and Alpine metamorphic cycles. In these rocks, starting from a primary garnet mica-schist (garnet stage), a complex sequence of transformations, affecting the site of the garnet, has been recognized. In the outermost part of the aureole, the primary garnet sites are occupied by nodules of kyanite (kyanite stage). Closer to the tonalite, kyanite is replaced by staurolite (staurolite stage), which in turn is pseudomorphed by muscovite (muscovite stage). The aggregates of kyanite do not overgrow garnet directly; they post-date a stage (fibrolite stage) represented by the pseudomorphic alteration of garnet into fibrolitic sillimanite plus biotite. A further sericite stage is likely to have occurred between the fibrolite and kyanite stages. Preservation of the sub-spherical garnet shape during all these transformations and persistence of mineralogical and textural relicts from earlier stages were favoured by the very low strain experienced by the rocks since the garnet stage. The textural sequence is in agreement with the metamorphic history of this part of the Austroalpine basement of the Eastern Alps: the garnet and fibrolite stages, and the coeval main foliation of the samples, are referred to the high-grade Hercynian metamorphism; the kyanite stage to the Eo-Alpine metamorphism; the staurolite and muscovite stages to the Oligocene contact metamorphism. It is suggested that kyanite growth as microgranular aggregates took place in polymetamorphic rocks where static, high- P /low- T  metamorphism overprinted high- T  assemblages that contained sillimanite or andalusite.     

Regional distribution of Al-silicates and metamorphic zonation in the low-grade Verrucano metasediments from the Northern Apennines, Italy

Abstract Mineralogical and petrological studies of Triassic Verrucano metasediments of the Northern Apennines are reported. The widespread occurrence of Al-silicates allows the delineation of four metamorphic zones with increasing metamorphic grade: (1) kaolinite zone (well Perugia 2, Umbria); (2) kaolinite-pyro-phyllite zone (Monte Argentario and part of the Verrucano of the Monticiano-Roccastrada area and Monti Leoni); (3) pyrophyllite zone (Monti Pisani, Iano, Monti Leoni, the Monticiano-Roccastrada area and some wells in the Larderello region); (4) kyanite zone (Massa area and some wells in the Larderello area).
The four metamorphic zones correspond to temperatures ranging from 300°C to about 450°C. On the basis of the Si content of muscovite and geological arguments, pressures of between 3 and 5 kbar are estimated. The metamorphic zones are located more or less parallel to the bent north-west-south-east trending structural zonation of the Northern Apennines, with the concave side towards the Tyhrrenian Sea.
During the Alpine orogeny, the Verrucano metasediments underwent three folding phases each of which has produced an axial plane schistosity (S₁, S₂, S₃). During the first folding phase the Verrucano sediments were buried increasingly deeply within the crust from east to west. The climax of Alpine metamorphism was attained prior to the second folding phase with crystallization of porphyroblasts of kyanite and chloritoid in a central area located between Massa and Larderello. The inferred paleo-temperature distribution pattern resembles an asymmetric thermal high defined by the kyanite zone, and surrounded by the pyrophyllite zone. A similar pattern is still present in the Tuscan crust, as indicated by a series of geothermal anomalies passing through the Northern Apennines.     

Evolution of a kyanite-bearing belt within a HT-LP orogen: the case of NW Variscan Iberia

Kyanite replaces andalusite in a belt of Ordovician and Silurian pelitic rocks that form a narrow synform pinched between high-grade antiforms in NW Variscan Iberia. Kyanite occurs across the belt in Al-rich, black pelites in assemblages I: kyanite–chloritoid–chlorite–muscovite and II: kyanite–staurolite– chlorite–muscovite. In I, kyanite occurs in the matrix and in kyanite–muscovite aggregates that pseudomorph earlier andalusite porphyroblasts. The aggregates are found across the belt and can still be recognized in assemblage II and even in III: andalusite–staurolite–biotite–muscovite, this latter being a hornfelsic Silurian schist where kyanite is relic and staurolite occurs in the matrix, and is resorbed inside new massive pleochroic andalusite. KFMASH and MnKFMASH pseudosections have been constructed using Thermocalc for Al-rich and Al-poorer compositions from the belt. Chloritoid zoning in Al-rich rocks containing assemblage I, plus chloritoid–chlorite thermometry complemented with garnet–chlorite thermometry in Al-poorer lithologies, mean that the path is one of increasing pressure and temperature. Conditions prior to assemblage I, with earlier andalusite stable, are those of the andalusite–chloritoid– chlorite field as testified by chloritoid enclosed in andalusite porphyroblast rims. The passage from assemblage I to II implies a prograde path within the kyanite field. Assemblage III represents peak conditions, indicating a prograde staurolite-consuming reaction across a KFMASH field, leading eventually to a locally found andalusite–biotite–muscovite hornfels. The lowest pressure stages are recorded by cordierite–biotite in Al-poor pelites. Garnet-bearing MnKFMASH assemblages in Al-poorer pelites record conditions similar to assemblages II and III. The replacement of andalusite by kyanite in assemblage I is attributed to downdragging of andalusite-bearing rocks into a synform as testified by the strained andalusite porphyroblasts affected by a subvertical crenulation cleavage. Prograde metamorphism in the eastern contact of the belt is due to heat transferred to the belt from the ascending high grade antiform across the Vivero fault.     

Kyanite deformation in whiteschist of the ultrahigh-pressure metamorphic Kokchetav Massif, Kazakhstan

Microstructures in minerals from ultrahigh‐pressure metamorphic (UHPM) terranes are keys to understanding the rheological properties and the exhumation mechanisms of rocks from subduction zones. Kyanite‐bearing whiteschist, associated with eclogite lenses, is part of UHPM unit II located south‐west of Lake Zheltau in the Kulet region of the Kokchetav Massif. The equilibrium assemblage is kyanite + garnet + talc + phengite + coesite/quartz. Previously reported peak pressure–temperature (P–T) conditions are ～3.5 GPa at 750 °C. A strong foliation is defined by the talc and phengite, with a corresponding weak shape preferred alignment of kyanite. Crystallographic orientation maps and analysis of kyanite blades were performed using electron backscatter diffraction methods. The data are consistent with a (100)[001] slip system for the formation of undulose extinction and kink bands in kyanite. Rotations measured across individual kink bands are 10–50° about <010>, and rotations along kyanite with undulose extinction are up to 50° about <010> with variations between adjacent points typically <2°. The undulose extinction is interpreted to have developed through crystal plastic deformation by dislocation creep. Kink bands mark the development of high‐angle grain boundaries by dislocation climb. The deformation of kyanite occurred in the fault‐bounded terrane during the exhumation of the Kokchetav Massif.     

Successive hydration and dehydration of high-P mafic granofels involving clinopyroxene–kyanite symplectites, Mt Daniel, Fiordland, New Zealand

Three texturally distinct symplectites occur in mafic granofels of the Arthur River Complex at MtDaniel, Fiordland, New Zealand. These include symplectic intergrowths of clinopyroxene and kyanite, described here for the first time. Pods of mafic granofels occur within the contact aureole of the Early Cretaceous Western Fiordland Orthogneiss batholith. The pods have cores formed entirely of garnet and clinopyroxene, and rims of pseudomorphous coarse‐grained symplectic intergrowths of hornblende and clinozoisite that reflect hydration at moderate to high‐P. These hornfelsic rocks are enveloped by a hornblende–clinozoisite gneissic foliation (S1). Narrow garnet reaction zones, in which hornblende and clinozoisite are replaced by garnet–clinopyroxene assemblages, developed adjacent to fractures and veins that cut S1. Fine‐grained symplectic intergrowths of (1) clinopyroxene and kyanite and (2) clinozoisite, quartz, kyanite and plagioclase form part of the garnet reaction zones and partially replace coarse‐grained S1 hornblende and clinozoisite. The development of the garnet reaction zones and symplectites was promoted by dehydration most probably following cooling of the contact aureole. Maps of oxide weight percent and cation proportions, calculated by performing matrix corrections on maps of X‐ray intensities, are used to study the microstructure of the symplectites.     

Using calculated chemical potential relationships to account for replacement of kyanite by symplectite in high pressure granulites

Calculated mineral equilibria are used to account for the formation of sapphirine–plagioclase, spinel–plagioclase and corundum–plagioclase symplectites replacing kyanite in quartz–plagioclase–garnet–kyanite granulite facies gneisses from the Southern Domain of the Athabasca granulite terrane, a segment of the Snowbird tectonic zone in northern Saskatchewan, Canada. Metamorphic conditions of >14 kbar and 800 °C are established for the high pressure, garnet–kyanite assemblage using constraints from P–T pseudosections and Zr‐in‐rutile thermometry. Replacement of kyanite by symplectites reflects the reaction of kyanite with the matrix following near‐isothermal decompression to <10 kbar. The chemical potential gradients developed between the kyanite and the matrix led to diffusion that attempted to flatten the gradients, kyanite persisting as a stable phase while it is consumed by symplectite from its edge. In this local equilibrium model, the mineral and mineral compositional spatial relationships are shown to correspond to paths in μ(Na₂O)–μ(CaO)–μ(K₂O)–μ(FeO)–μ(MgO) in the model chemical system, Na₂O–CaO–K₂O–FeO–MgO–Al₂O₃–SiO₂ (NCKFMAS), with SiO₂ and Al₂O₃ taken to be completely immobile. The values of μ(Na₂O) and μ(CaO) are constrained by fixing P–T conditions and choosing appropriate μ(Na₂O) and μ(CaO) values that correspond to the observed plagioclase compositions. μ(FeO)–μ(MgO) diagrams show the corresponding spatial relationships with kyanite and the symplectite phases. These results demonstrate that the replacement of kyanite by sapphirine–plagioclase and spinel–plagioclase appears to be metastable with respect to replacement by corundum–plagioclase. Replacement by corundum–plagioclase does also occur, apparently overprinting pre‐existing symplectite and also kyanite. Ignoring corundum, the resulting diagrams account for the spatial relationships and compositions observed in the spinel–plagioclase and sapphirine–plagioclase symplectites. They are predicted to occur over both a wide range of P–T conditions (6–11 kbar, 650–850 °C) and plagioclase compositions (X_An = 0.5–0.9). The wide range of P–T conditions that may result in identical spatial and compositional relationships suggests that such reaction textures may be of limited use in accurately quantifying the P–T conditions of retrograde metamorphism.