Fluid-saturated subsolidus experiments from 2·0 to 6·5 GPa, and from 680 to 800°C have been performed on three model peridotites in the system Na₂O–CaO–FeO–MgO–Al₂O₃–SiO₂–H₂O (NCFMASH). Amphibole and chlorite coexist up to 2·4 GPa, 700°C. Chlorite persists to 4·2 GPa at 680°C. Starting from 4·8 GPa, 680°C a 10 Å phase structure replaces chlorite in all compositions. The 10 Å phase structure contains significant Al₂O₃ (up to 10·53 wt %) deviating from the MgO–SiO₂–H₂O 10 Å phase (MSH 10 Å phase). A mixed layered structure (chlorite–MSH 10 Å phase) is proposed to account for aluminium observed. In the Tinaquillo lherzolite amphibole breakdown occurs via the reaction Thermal stability of chlorite (chlorite + orthopyroxene = forsterite + garnet + H₂O) is shifted towards lower temperatures, compared with the system MASH. Furthermore, the chlorite thermal breakdown is also related to the degenerate reaction Chlorite and the Al-10 Å phase structure contribute significantly to the water budget in subduction zones in the depth range relevant for arc magmatism, whereas amphibole-related fluid release is restricted to the forearc region. Chlorite and Al-10 Å phase breakdowns might explain the occurrence of a double seismic zone by dehydration embrittlement.

Plagioclase peridotites are important markers of processes that characterize the petrological and tectonic evolution of the lithospheric mantle in extensional tectonic settings. Studies on equilibrated plagioclase peridotites have documented continuous chemical changes in mantle minerals in response to plagioclase crystallization, potentially tracing the re-equilibration of mantle peridotites up to very low pressure. This experimental study provides new constraints on the stability of plagioclase in mantle peridotites as a function of bulk composition, and the compositional and modal changes in minerals occurring within the plagioclase stability field as a function of P–T–bulk composition. Subsolidus experiments have been performed at pressures ranging from 0·25 to 1·0 GPa, and temperatures ranging from 900 to 1200°C on fertile and depleted anhydrous lherzolites modelled in the system TiO₂–Cr₂O₃–Na₂O–CaO–FeO–MgO–Al₂O₃–SiO₂ (Ti,Cr-NCFMAS). In the fertile lherzolite (Na₂O/CaO = 0·08; X_Cr = 0·07) a plagioclase-bearing assemblage is stable up to 0·7 GPa, 1000°C and 0·8 GPa, 1100°C, whereas in the depleted lherzolite (Na₂O/CaO = 0·09; X_Cr = 0·10) the upper limit of plagioclase stability is shifted to lower pressure. The boundary between plagioclase lherzolite and spinel lherzolite has a positive slope in P–T space. In a complex chemical system, the plagioclase-out boundary is multivariant and sensitive to the X_Cr value [X_Cr = Cr/(Cr + Al)] of spinel. This latter is controlled by the reaction MgCr₂O₄ + CaAl₂Si₂O₈ = MgCrAlSiO₆ + CaCrAlSiO₆, which is a function of the Cr–Al partitioning between spinel and pyroxenes, and varies with the X_Cr value and chromite/anorthite normative ratio of the bulk composition. Within the plagioclase stability field, the Al content of pyroxenes decreases, coupled with an increase in the anorthite content in plagioclase, and Ti and X_Cr in spinel with decreasing pressure; these chemical variations are combined with systematic changes in modal mineralogy governed by a continuous reaction involving both plagioclase and spinel. As a consequence, the composition of plagioclase varies significantly over a rather narrow pressure range and is similar at the same P–T conditions in the investigated bulk-rocks. This suggests the potential application of plagioclase composition as a geobarometer for plagioclase peridotites.

We experimentally investigated phlogopite- and C–O–H-bearing lherzolite to model the mantle wedge fluxed by volatiles released from a subducting crustal slab. Experiments have been carried out at 900–1050°C and 1·6–3·2 GPa, at fluid- and carbon-saturated conditions. We used an end-loaded piston cylinder apparatus and a conventional double-capsule technique to constrain the redox state of the experiments, using the nickel–nickel oxide oxygen buffer (NNO). Following thermodynamic calculations, we expect inner fO₂ values to be systematically below NNO, with fluids that are mixtures of CO₂ and H₂O. Estimated fO₂ in the runs are between ΔFMQ ≈ –0·7 at 3 GPa and ΔFMQ ≈ –1·1 at 1·8 GPa, values that have been reported for natural mantle-wedge xenoliths. At the conditions investigated, the hydrous phases are phlogopite and pargasitic amphibole. Whereas phlogopite is ubiquitous, amphibole disappears at 3·1 GPa at 900°C and 2·7 GPa at 1050°C, where the solidus is encountered. The amphibole-out reaction also consumes orthopyroxene and liberates water. From low to high P, we observed first carbonate-free, amphibole-bearing assemblages, then carbonate + amphibole-bearing assemblages, and finally amphibole-free, carbonate-bearing assemblages. Carbonate-free assemblages melt to produce trachyandesite at T > 1050°C, whereas dolomitic carbonatites have been found beyond the solidus of carbonate-bearing assemblages. Carbonates occur as dolomite at <1·9 GPa, 900°C and at 2·4 GPa, 900°C and >2·7, 1050°C; between these limits, a magnesite + dolomite-bearing assemblage constitutes a two-carbonate field. P–T pseudosections fail to reproduce the experimental results concerning amphibole breakdown and reaction positions involving carbonates. The amount of COH fluid is thought to have a major role, even in fluid-saturated peridotites. Clinopyroxene and olivine are not expected at fluid-oversaturated conditions, for which dolomite or magnesite are stable respectively. The presented results are useful for unravelling the exhumation history of orogenic lherzolites bearing COH phases and to suggest a way to transfer carbon species to the mantle wedge. We suggest that once carbon-bearing fluids react with mantle-wedge peridotites, a sort of buoyant ‘cold plume’ will form containing low-density phases such as amphibole, carbonates and carbonatitic melt. This plume could represent an important source of CO₂ and H₂O, and it is one of a series of processes that ultimately lead to arc magmatism.