Dehydration Of Phengite Research Articles

The Origin of the Lehmann Discontinuity Beneath the Ancient Craton: Insight From the High Pressure‐Temperature Elasticity Measurements of Topaz

AbstractIn this study, we concentrate on the seismic signature of subducted sediments and suggest the formation of the L‐discontinuity beneath the ancient craton related to migrated sediment dehydration. We first determined the single‐crystal elasticity of topaz, the product of sediment dehydration, at high pressures and temperatures by Brillouin scattering. Using the derived elastic parameters, we establish the velocity and density profiles of subducted sediments in the upper mantle. According to our modeling results, 8.5–17.5 vol.% sediments intruding into the upper mantle will induce a 2%–4% low‐VS anomaly at 210–260 km. Meanwhile, continuous heating will lead to the dehydration of phengite in sediments. The dehydration of this amount sediments can generate a 3%–6% ISS with negative Clapeyron slopes, satisfying the observed L‐discontinuity in northern Finland and northern America without the anisotropy changes but accompanied by low‐velocity anomalies. Our study thus provides new insights into the origin of the L‐discontinuity.

Omphacite Melting and the Destruction of Early High‐Pressure Rock Records

AbstractThe presence of eclogite marks high‐pressure metamorphism and deep, cold burial of crust, and is therefore a widely accepted indicator for the onset of modern‐style plate tectonics. However, eclogite is rarely preserved in many heavily granulite‐overprinted orogens, which are particularly common across ancient metamorphic terranes (i.e., Archean and Paleoproterozoic terranes). Here, we show that eclogite melting processes may be principally responsible for the poor preservation of high‐pressure records. We demonstrate eclogite melting via detailed petrological, geochronological, and geochemical analyses on eclogites and separated centimetric leucosomes from the central Himalaya. The central Himalayan eclogites were overprinted by strong granulite‐facies metamorphism, such that omphacite is only sparsely preserved. Thermodynamic modeling results indicate the eclogite experienced two types of anatectic reactions: phengite dehydration melting at high‐pressure in the presence of omphacite, and subsequent omphacite dominated melting during exhumation. Diagnostic features for these melting reactions include (a) the occurrence of Kfs‐Pl‐Qz multiphase solid inclusions and Cpx‐Pl‐Qz polymineralic inclusions; (b) the high Na2O/K2O melts and Na‐rich plagioclase in leucosomes; (c) the consistent trace elements patterns between omphacite and leucosomes; and (d) obvious Na2O zonation in clinopyroxene. Omphacite dominated melting is characterized by the jadeite breakdown, releasing Na and Al into the melts. This melting mechanism subsequently forms a less sodic clinopyroxene and high Na2O/K2O melts. These Himalayan findings appear relevant to early Earth explorations because of the high thermal gradients and intense granulite‐facies overprints, implying that partial melting of eclogite dominated by omphacite breakdown could have erased early high‐pressure records of modern‐style plate tectonics.

Peritectic minerals record partial melting of the deeply subducted continental crust in the Sulu orogen

AbstractThe deeply subducted continental crust commonly undergoes partial melting during its exhumation from mantle to crustal depths. Peritectic minerals are commonly produced during decompressional melting, providing an important record of the timing, P–T conditions, and mechanisms of crustal anatexis in continental subduction zones. A combined study of petrography, mineral inclusions, mineral major and trace element compositions, phase equilibrium modelling, zircon and titanite U–Pb dating, and zircon Hf–O isotopes was carried out to characterize peritectic minerals in ultrahigh‐pressure (UHP) metamorphic gneisses from the Sulu orogen, China. Field and petrographic observations reveal that the crustal anatexis was extensive and is recorded by folded leucosomes, corroded phengite relicts, cuspate K‐feldspar and quartz, pseudomorphs of melt films along grain boundaries, and multiphase crystal inclusions in zircon, garnet, and phengite. The multiphase crystal inclusions represent the crystallized product of anatectic melt droplets trapped in the peritectic zircon and garnet as well as the residual UHP phengite. The peritectic origin of both zircon and garnet is identified by their compositional characteristics, and the peritectic origin of garnet is further supported by phase equilibrium modelling which shows an increase in garnet contents as partial melting proceeds. The modelled phase diagrams also indicate that the two peritectic minerals were produced through two episodes of phengite dehydration melting. The first occurred at 237 ± 3 Ma under P–T conditions of ≥2.6 GPa and 810–880°C, corresponding to thermobaric ratios of ~312–338℃/GPa. The second took place at 222 ± 2 Ma under P–T conditions of 1.5–2.1 GPa and 730–760℃, corresponding to thermobaric ratios of 362–487℃/GPa. In comparison to peak UHP metamorphic P–T conditions of 2.8–4.0 GPa and 720–870℃ for thermobaric ratios of 218–257℃/GPa, it appears that the decompressional melting is associated with an increase in thermal gradients during the exhumation in the different stages. Whereas the first is caused by decompression during initial exhumation of the deeply subducted continental crust, the second is caused by decompression during relamination of the UHP slices into the collision‐thickened orogenic lower crust. Peritectic amphibole and titanite were also recognized by comparing their petrographic textures with natural samples and experimental products. They were produced through water‐fluxed melting at 217 ± 3 Ma under lower P–T conditions, suggesting the liberation of aqueous fluids from the underlying exhuming UHP rocks. All of the peritectic minerals generally show compositional inheritance from their precursor minerals. Nevertheless, they commonly underwent dissolution when anatectic P–T conditions deviate significantly from those of peritectic reactions. This is indicated by the dissolution of peritectic zircon and garnet and by the occurrence of anhedral and embayed peritectic amphibole. As a consequence, the behaviour of peritectic minerals provides a snapshot of crustal anatexis for geochemical differentiation in collisional orogens.

Evidence for UHP anatexis in the Shuanghe UHP paragneiss from inclusions in clinozoisite, garnet, and zircon

AbstractThe Shuanghe garnet‐bearing paragneiss from the Dabie ultra‐high–pressure (UHP) orogen occurs as an interlayer within partially retrogressed eclogite. A first UHP metamorphic stage at 680°C, 3.8–4.1 GPa is documented by Zr‐in‐rutile temperatures coupled with phengite inclusions (Si = 3.55) in clinozoisite and grossular‐rich garnet. Relic matrix phengite and phengite inclusions in zircon rims display lower Si of 3.42. Combined with garnet compositions and Ti‐in‐zircon temperatures, they provide evidence for a second UHP metamorphic stage at 800–850°C, ~3.8 GPa. Such isobaric heating at UHP conditions has not been documented so far from the adjacent eclogites and other rock types in the Dabie orogen and indicates proximity to the hot, convecting mantle wedge. The dominant mineral assemblage consisting of plagioclase, epidote, biotite, and amphibole provides evidence for widespread retrogression during the exhumation of the UHP paragneiss. Several types of polyphase mineral inclusions were identified. Phengite inclusions hosted by clinozoisite are partially replaced by kyanite and K‐feldspar, whereas inclusions in host garnet consist of relic phengite, K‐feldspar, and garnet, indicating limited sub‐solidus dehydration of phengite by the reaction Ph→Kfs+Ky±Grt+fluid. Tightly intergrown K‐feldspar and quartz are preserved as inclusions with sharp boundaries and radial cracks in garnet. Analyses of whole inclusions also show small enrichments in light rare earth elements. These inclusions are interpreted to be derived from melting of an inclusion assemblage consisting of Ph+Coe±Czo. A third type of polyphase inclusion consists of typical nanogranite (Ab+Kfs+Qz±Ep) inclusions in recrystallized metamorphic zircon. Ti‐in‐zircon thermometry and the Si content of phengite included in these zircon domains indicate that melting occurred at 800–850°C and 3.8–4.0 GPa during isobaric heating at UHP conditions. The partial melting event led to an equilibration of trace elements in garnet, phengite, and apatite. Using published partition coefficients between these minerals and hydrous granitic melt, the trace element composition of the UHP anatectic melt can be constrained. The melts are characterized by high LILE contents and pronounced relative enrichments of U over Th and Ta over Nb. The REE are below primitive mantle values, likely due to the presence of residual clinozoisite and garnet during partial melting. So far, no major granitic bodies have been found that share the same trace element pattern as the partial melts from the UHP anatexis of the Shuanghe paragneiss.

Intimate link between ammonium loss of phengite and the deep Earth's water cycle

Phengite, a high-pressure white mica, is stable down to ∼300 km depth in subduction environment and represents an important carrier of nitrogen and hydrogen to the deep Earth. Although the determination of nitrogen and hydrogen recycling efficiency to subduction zones is crucial for building global geodynamic models of volatile elements, no experimental data have so far evaluated the rate of ammonium release from phengite and its potential effect on phengite dehydration in a descending slab. Here, we investigated the kinetics of ammonium loss and dehydration of ammonium-bearing and ammonium-free phengite samples using FTIR spectroscopy. Single phengite grains were analyzed after annealing at high temperatures from 720 to 850°C for different periods of time. The results show that ammonium loss is one order of magnitude faster than dehydration at the same temperature (for instance, with diffusion coefficients of 1.5×10−15 m2/s and 1.0×10−16 m2/s at 800°C for ammonium loss and dehydration, respectively). Moreover, dehydration of ammonium-bearing phengite is one order of magnitude faster than that of ammonium-free phengite (with diffusion coefficients of 1.0×10−16 m2/s and 2.0×10−17 m2/s at 800°C for dehydration in ammonium-bearing and ammonium-free phengite, respectively). The activation energies of ammonium loss, dehydration in ammonium-bearing phengite and dehydration in ammonium-free phengite are around 150, 535, 548 kJ/mol, respectively. These results imply that ammonium release from phengite is not only easier than dehydration but also can promote dehydration. The geochemical behavior of ammonium in phengite is consistent with previous observations on natural samples from subduction zones of different geothermal gradients. Furthermore, since ammonium release can trigger dehydration, water may be released at depth locus shallower than theoretical phengite breakdown (i.e. <300 km). Thus, ammonium present in hydrous minerals may exert a strong control on the water flux released in subduction zones.

Dehydration of phengite inferred by electrical conductivity measurements: Implication for the high conductivity anomalies relevant to the subduction zones

Dehydration of phengite inferred by electrical conductivity measurements: Implication for the high conductivity anomalies relevant to the subduction zones

Modeling the exhumation path of partially melted ultrahigh-pressure metapelites, North-East Greenland Caledonides

Pseudosection modeling constrains the pressure–temperature (P–T) exhumation path of partially melted ultrahigh-pressure (UHP) metapelites exposed in the North-East Greenland UHP terrane. A robust peak P and T estimate of 3.6GPa and 970°C based on mineral assemblages in nearby kyanite eclogites is the starting point for the P–T path. Although the peak assemblage for the metapelite is not preserved, the calculated modeled peak assemblage contained substantial clinopyroxene, garnet, phengite, K-feldspar and coesite with minor kyanite and rutile. Combining the pseudosection and observed textures, the decompression path crosses the coesite-quartz transition before reaching the dry phengite dehydration melting reaction where phengite is abruptly consumed. In the range of 2.5 to 2.2GPa, clinopyroxene is completely consumed and garnet grows to its maximum volume and grossular content, matching the high grossular rims of relict megacrysts. Plagioclase joins the assemblage and the pseudosection predicts up to 12–13vol.% melt in the supersolidus assemblage, which contained garnet, liquid, K-feldspar, plagioclase, kyanite, quartz and rutile. At this stage, the steep decompression path flattened out and became nearly isobaric. The melt crystallization assemblage that formed when the path crossed the solidus with decreasing temperature contains phengite, garnet, biotite, 2 feldspars, kyanite, quartz and rutile. Therefore, the path must have intersected the solidus at approximately 1.2GPa, 825°C. The pseudosection predicts that garnet is consumed on the cooling path, but little evidence of late garnet consumption or other retrograde effects is observed. This may be due to partial melt loss from the rock. Isochemical PT–n and PT–X sections calculated along the P–T path display changes in mineral assemblage and composition that are consistent with preserved assemblages.

Discovery of coesite–eclogite from the Nordøyane UHP domain, Western Gneiss Region, Norway: field relations, metamorphic history, and tectonic significance

AbstractUltrahigh‐pressure (UHP) rocks from the Western Gneiss Region (WGR) of Norway record subduction of Baltican continental crust during the Silurian to Devonian Scandian continental collision. Here, we report a new coesite locality from the island of Harøya in the Nordøyane UHP domain, the most northerly yet documented in the WGR, and reconstruct the P–T history of the host eclogite. The coesite–eclogite lies within migmatitic orthogneiss, interpreted as Baltica basement, that underwent multiple stages of deformation and partial melting during exhumation. Two stages of metamorphism have been deduced from petrography and mineral chemistry. The early (M1) assemblage comprises garnet (Pyr38–41Alm35–37Grs23–26Spss1) and omphacite (Na0.35–0.40Ca0.57–0.60Fe2+0.08–0.10Mg0.53Fe3+0.01AlVI0.40–0.42)2(AlIV0.03–0.06Si1.94–1.97)2O6, with subordinate phengite, kyanite, rutile, coesite and apatite, all present as inclusions in garnet. The later (M2) assemblage comprises retrograde rims on garnet (Pyr38–40Alm40–44Grs16–21Spss1), diopside rims on omphacite (Na0.04–0.06Ca0.88–0.91Fe2+0.09–0.13Mg0.81–83Fe3+0.08AlVI0.03)2(AlIV0.07–0.08Si1.92–1.93)2O6, plagioclase, biotite, pargasite, orthopyroxene and ilmenite. Metamorphic P–T conditions estimated using thermocalc are ∼3 GPa and 760 °C for M1, consistent with the presence of coesite, and ∼1 GPa and 813 °C for M2, consistent with possible phengite dehydration melting during decompression. Comparison with other WGR eclogites containing the same assemblage shows a broad similarity in peak (M1) P–T conditions, confirming suggestions that large portions of the WGR were buried to depths of ∼100 km during Scandian subduction. Field relations suggest that exhumation, accompanied by widespread partial melting, involved an early phase of top‐northwest shearing, followed by subhorizontal sinistral shearing along northwest‐dipping foliations, related to regional transtension. The present results add to the growing body of data on the distribution, maximum P–T conditions, and exhumation paths of WGR coesite–eclogites and their host rocks that is required to constrain quantitative models for the formation and exhumation of UHP metamorphic rocks during the Scandian collision.

High-pressure partial melting and melt loss in felsic granulites in the Kutná Hora complex, Bohemian Massif (Czech Republic)

Felsic granulites from the Kutná Hora complex in the Moldanubian zone of central Europe preserve mineral assemblage that records transition from early eclogite to granulite facies conditions, and exhibits leucocratic banding, which is interpreted as an evidence for melt loss during the decompression path. The granulites are layered and consist of variable proportions of quartz, ternary feldspar, garnet, biotite, kyanite, and rutile. In the mesocratic layers, garnet grains show relatively high Ca contents corresponding to 28–41mol% grossular end member. They have remarkably flat compositional profiles in their cores but their rims exhibit an increase in pyrope and a decrease in grossular and almandine components. In contrast, garnets from the leucocratic layers have relatively low Ca contents (15–26mol% grossular) that further decrease towards the rims. In addition to modeling of pressure–temperature pseudosections, compositions of garnet core composition, garnet rim-ternary feldspar–kyanite–quartz equilibrium, ternary feldspar composition, and the garnet–biotite equilibrium provide five constraints that were used to reconstruct the pressure–temperature path from eclogite through the granulite and amphibolite facies. In both layers, garnet cores grew during omphacite breakdown and phengite dehydration melting at 940°C and 2.6GPa. Subsequent decompression heating to 1020°C and 2.1GPa produced Ca- and Fe-poor garnet rims due to the formation of Ca-bearing ternary feldspar and partial melt. In both the mesocratic and leucocratic layer, the maximum melt productivity was 26 and 18vol.%, respectively, at peak temperature constrained by the maximum whole-rock H2O budget, ~1.05–0.75wt.%, prior to the melting. The preservation of prograde garnet-rich assemblages required nearly complete melt loss (15–25vol.%), interpreted to have occurred at 1000–1020°C and 2.2–2.4GPa by garnet mode isopleths, followed by crystallization of small amounts of residual melt at 760°C and 1.0GPa. Phase formation and melt productivity were independently determined by experiments in the piston-cylinder apparatus at 850–1100°C and 1.7–2.1GPa. Both the thermodynamic calculations and phase equilibrium experiments suggest that the partial melt was produced by the dehydration melting: muscovite+quartz=melt+K-feldspar+kyanite. The presence of partial melt facilitated attainment of mineral equilibria at peak temperature thus eliminating any potential relics of early high-pressure phases such as phengite or omphacite. By contrast, adjacent mafic granulites and eclogites, which apparently share the same metamorphic path but have not undergone partial melting commonly preserve relics or inclusions of eclogite-facies mineral assemblages.

Melting-induced fluid flow during exhumation of gneisses of the Sulu ultrahigh-pressure terrane

Hydrothermally altered rocks are products of fluid–rock interactions, and typically preserve numerous quartz veins that formed as chemical precipitates from fluids that fill up cracks. Thus, quartz veins are the record of the fluid system that involved fracture flow in the direction of changing temperature or pressure. In order to decipher the fluid activity in the Sulu ultrahigh-pressure (UHP) terrane in eastern China, quartz veins together with an adjacent eclogite lens and the host gneiss were studied. In one location a deformed quartz vein is located at the boundary between the host gneiss and the eclogite lens. The amphibolite-facies overprinting of the eclogite lens decreases from the rim to the core of the lens, with fresh eclogite preserved in the core. The foliated biotite gneiss contains felsic veins and residual phengites. Zircon rims from the gneiss are characterized by melt-related signatures with steep HREE patterns, high Hf contents and negative Eu anomalies, and a pool of weighted average 206Pb/ 238U analyses reveal an age of 219 ± 3 Ma (2σ), which is younger than the UHP metamorphic age (236 ± 2 Ma, 2σ) recorded by zircons from the eclogite lens. This suggests that the gneiss in the Sulu UHP terrane could have suffered from partial melting due to phengite dehydration during the “hot” exhumation stage. The formation age of the quartz vein (219 ± 2 Ma, 2σ) defined by zircon rims agrees well with the partial melting time (219 ± 3 Ma, 2σ) of the host gneiss. The initial 176Hf/ 177Hf ratios of zircon rims from the quartz vein are obviously lower than zircons from the eclogite lens, but overlap with the coeval zircon domains from the nearby granite dikes produced by partial melting of orthogneiss. These observations suggest that the quartz vein and corresponding fluid flow could be associated with partial melting of the host gneiss. On the other hand, amphibole-bearing and HREE-rich zircon rims from the amphibolite pool an amphibolite-facies metamorphic age of 217 ± 5 Ma (2σ), overlap with the formation age of the quartz vein. This implies that retrogression of the eclogite lens could have been caused by melting-induced fluid flow. Based on the above observations, we speculate that partial melting of the gneiss in the continental subduction-related UHP belt could have induced a significant fluid flow during the exhumation stage, and thus contributed significantly to the extensive retrogression of eclogites in the Sulu UHP terrane.

An experimental study of dehydration melting of phengite-bearing eclogite at 1.5–3.0 GPa

Dehydration melting experiments were performed on ultrahigh-pressure eclogite from Bixiling in the Dabie orogen at 1.5–3.0 GPa and 800–950°C using piston cylinder apparatus. The results show that (1) eclogite with ∼5% phengite started to melt at T⩽800–850°C and P = 1.5–2.0 GPa and produced about 3% granitic melt; (2) the products of dehydration melting of phengite-bearing eclogite vary with temperature and pressure. Fluid released from dehydration of phengite and zoisite leads to partial melting of eclogite and formation of plagioclase reaction rim around kyanite at pressures of 1.5–2.0 GPa and temperatures of 800–850°C; (3) phengite reacted with omphacite and quartz and produced oligoclase, kyanite and melt at elevated temperatures. Oligoclase is the primary reaction product produced by partial melting of phengite in the eclogite; and (4) the dehydration melting of phengite-bearing eclogite at pressures of 1.5–3.0 GPa and temperatures ⩾900°C results in formation of garnets with higher molar fraction of pyrope (37.67 wt.%–45.94 wt.%). Potassium feldspar and jadeite occur at P = 2.4–3.0 GPa and T⩾900°C, indicating higher pressure and fluid-absent conditions. Our results constrain the solidus for dehydration melting of phengite-bearing eclogite at pressures of 1.5–3.0 GPa. Combining experimental results with field observations of partial melting in natural eclogites, we concluded that phengite-bearing eclogites from the Dabie-Sulu orogen were able to partially molten at P = 1.5–2.0 GPa and T = 800–850°C during exhumation. The ultrahigh-high pressure eclogites would have experienced partial melting in association with metamorphic phase transformation under different fluid conditions.

Experimental evidence on phlogopitic mantle metasomatism induced by phengite dehydration

To simulate metasomatism at convergent plate boundaries, the process of phengite dehydration, followed by reactions between the expelled fluid and a model mantle material consisting of forsterite + enstatite, was investigated at subduction zone conditions using a piston cylinder apparatus. An experimental technique has been devised in which the fluid, which should be set free during phengite decomposition at about 950°C and 3.5 GPa, would be triggered by a temperature gradient of about 50 K to infiltrate through a perforated Au-foil the hotter part of the capsule, containing the model mantle material. Crystallization of phlogopitic mica and Al-rich enstatite at the boundary phengite - (forsterite + enstatite), representing the „slab - mantle” contact region, indicates successful experimental simulation of mantle wedge metasomatism. Cs-Rb exchange coefficients K D , determined from compositions of newly formed phlogopitic micas, recrystallized phengite and of an amorphous phase, representing the quenched liquid, agree well with recently determined alkali-fractionation data between micas and fluid. The composition of the quenched liquid is rich in alkali, aluminium, silicon, and poor in magnesium, which is in agreement with fluid solubility data determined in earlier experimental studies at similar conditions.

Dehydration and melting experiments constrain the fate of subducted sediments

Geochemical tracers demonstrate that elements are cycled from subducted sediments into the arc melting regime at subduction zones, although the transfer mechanism is poorly understood. Are key elements (Th, Be, Rb) lost during sediment dehydration or is sediment melting required? To investigate this question, we conducted phase equilibria and trace element partitioning experiments on a pelagic red clay for conditions appropriate to the slab beneath arc volcanoes (2–4 GPa, 600°–1000°C). Using both piston cylinders and multianvils, we determined the solidus, phase stabilities, and major element compositions of coexisting phases. The solidus (H2O + Cl fluid‐saturated) was located at 775 ± 25°C at 2 GPa, 810 ± 15°C at 3 GPa, and 1025 ± 25°C at 4 GPa with noevidence for complete miscibility between melt and fluid. This sediment composition produces a profusion of phases both above and below the solidus: garnet, jadeitic pyroxene, alkali‐rich amphibole, phengite, biotite, magnetite, coesite, kyanite, apatite, zircon, Cl‐rich fluids, and peraluminous to peralkaline granitic melts. At 2 GPa the phengite dehydration solidus is at 800°–825°C, while biotite breaks down between 850° and 900°C. To explore trace element partitioning across the solidus at 2 GPa, we used diamonds to trap fluids and melts. Both the bulk sediment residues and diamond traps were analyzed postexperiment by inductively coupled plasma–mass spectrometry (ICP‐MS) and inductively coupled plasma–atomic emission spectrometry (ICP‐AES) for 40 elements for which we calculated bulk partition coefficients (D = Csolid/Cfluid). Below the solidus, Rb, Sr, Ba, and Pb showed the greatest mobility (D ∼ 0.5–1.0), while at the solidus, Th and Be became notably partitioned into the melt (D values changing from &gt;2.0 to &lt;1.0). K and Rb D values fall below 1.0 when the micas breakdown. Only at the solidus do Th and Rb attain similar partition coefficients, a condition required by arc data. Taken together, the experimental results indicate that critical elements (Th and Be) require sediment melting to be efficiently transferred to the arc. This conclusion is at odds with most thermal models for subduction zones, which predict slab temperatures more than 100°C lower than sediment solidi. Thus the condition of sediment melting (with oceanic crust dehydration) may provide new constraints on the next generation of thermal/geodynamical models of subduction zones.

The stability and composition of phengitic muscovite and associated phases from 5.5 to 11 GPa: Implications for deeply subducted sediments

The stability and composition of phengitic muscovite was investigated from 5.5–11 GPa, 700–1150°C in synthesis experiments performed in a multianvil apparatus. Starting materials consisted of natural minerals with a bulk composition similar to that of a K-deficient, intermediate dioctahedral-trioctahedral mica. Phengitic muscovite was found to be stable from 5.5–11 GPa at 900°C. Phengite melting occurs between 1075–1150°C at 7–8 GPa and 1000–1050°C at 10 GPa. At 10–11 GPa, 800°C octahedral cation deficient (OCD) muscovite and K-hollandite are observed rather than phengite.The average phengite content of muscovite is positively correlated with pressure ranging from 3.65 Si pfu at 5.5 GPa to 3.81 Si pfu at 11 GPa. The maximum phengite content of stable muscovite appears to be 3.80–3.85 Si pfu. Most phengite examined exhibits at least minor solid solution towards phlogopite, averaging 2.04 ± 0.06 (2σ) octahedral cations pfu.The hydrous phases phengite, lawsonite, topaz-OH, and Mg-pumpellyite occur between 6–8 GPa, 700–900°C. With increasing pressure lawsonite and Mg-pumpellyite dehydrate to form garnet between 8–9 GPa. With increasing temperature Mg-pumpellyite, lawsonite, and topaz-OH devolatilize to form garnet and kyanite between 900–1000°C at 7–8 GPa.Phengitic muscovite and topaz-OH would be stable in hydrous sediments in cool mature subduction zones to depths exceeding 360 km while lawsonite and Mg-pumpellyite would be stable to 240–300 km allowing the transport of H2O contained in these phases deep into the upper mantle. In warmer subduction zones the dehydration of Mg-pumpellyite, lawsonite, and topaz-OH; as well as the melting of phengite would result in fluid release at depths of 180–240 km.The presence of phengite at far greater depths than the zone of melt generation beneath arcs (100–150 km) requires a mechanism such as the partitioning of K, Be, B, Ba, and Rb into migrating fluids rather than the simple dehydration of phengite beneath arcs in order to provide for the transfer of these slab signature elements from phengite into arc magmas.