Hydrous Peridotite Research Articles

Partial melting of mantle wedge and evolution of island arc crust

Evolution of the crust in the northeast Honshu arc is discussed on the basis of the melting experiments of a peridotite under hydrous conditions and the crustal compositions estimated from deep crustal xenoliths and seismic velocity data. The chemical compositions of melts formed by partial melting of a hydrous peridotite within the pressure‐temperature conditions for shallow parts of the mantle wedge (1.2–2.0 GPa) have been experimentally determined with the “sandwich experiment” technique. The melts range from olivine tholeiitic to magnesian andesitic compositions (SiO2 48–56 wt %; MgO 9–14 wt %) beneath the volcanic front in the northeast Honshu arc and are slightly less silicic beneath the coast of the Sea of Japan. The crustal compositions of the volcanic zone in the northeast Honshu arc estimated from deep‐seated crustal inclusions in volcanic rocks and the velocity structures of the crust are andesitic in the central zone and high‐alumina basaltic near the coast of the Sea of Japan. The compositions of possible primary melts formed in the mantle wedge are significantly more mafic than these crustal compositions. Consequently, if the major part of the crustal materials in the volcanic zone was formed by magmas generated in the mantle wedge, thick (up to 30 km) ultramafic cumulates including olivine pyroxenite must have been formed beneath Moho during fractional crystallization and would constitute the upper part of the mantle wedge. The volume of the mantle wedge consisting entirely of even fertile peridotite was not sufficient to produce crustal materials and cumulates in the volcanic zone by such processes in the northeast Honshu arc. A static mantle wedge model is, therefore, not satisfactory for the growth of the crust in northeast Honshu arc. A dynamic model is suggested in which the mantle materials have been supplied into the mantle wedge from the back arc side by a convective flow.

Phlogopite and K‐amphibole in the upper mantle: Implication for magma genesis in subduction zones

High‐pressure phase relations have been examined for phlogopite + diopside with and without enstatite under vapor absent conditions in the pressure range of 5 to 13 GPa and in the temperature range of 1000 to 1300°C. Phlogopite in these systems can be stable up to 6–7 GPa and decomposes through pressure‐dependent reactions to crystallize phases including potassic amphibole. The experimental results suggest that phlogopite, which is one of main hydrous phases in the downdragged hydrated peridotite at the base of mantle wedge, plays an important role in the formation of magmas at the backarc side of a volcanic arc. The existence of potassic amphibole at higher pressure regions may imply the involvement of subduction component in magma generation in the region far away from the trench axis.

Opening of the Sea of Japan back-arc basin by asthenospheric injection

The position of a volcanic front on an arc is governed by dehydration reactions of amphibole + chlorite in down-dragged hydrated peridotite at the base of the mantle wedge. As these reactions are essentially pressure-dependent (ca. 3.5 GPa), the paleo-position of a volcanic front offers an estimate for the depth of the subducted oceanic lithosphere beneath it. This paper evaluates volcanism in the Northeast Japan arc since the Oligocene and discusses the mass transportation in the mantle wedge during the episode of back-arc extension in the Sea of Japan with reference to the migration of the volcanic front. The volcanic front during the Oligocene (ca. 30 Ma) was located along the western coast of the present Northeast Japan are and migrated about 200 km trenchwards at 23 Ma before the opening of the Sea of Japan back-arc basin. It follows that the angle of subduction became steeper during the period of 30-23 Ma, triggered by the injection of the asthenosphere into the mantle wedge. The back-arc extension of the Sea of Japan between 21 and 14 Ma is a manifestation of the dynamic process of the lithosphere due to the asthenospheric injection. The temperature of the injecting asthenospheric mantle was higher than that of the surrounding upper mantle or lithosphere. This caused the characteristic volcanism at 23-22 Ma: (1) High-magnesian andesite volcanism took place in the near-trench region; (2) across-arc compositional variation was not observed; (3) the volume of volcanic materials increases towards the back-arc side of volcanic arc. The volcanic front has migrated over a short distance inwards (less than 100 km) since 15 Ma, possible as a result of cooling in the mantle wedge, because both amphibole and chlorite break down at slightly higher pressures at lower temperatures. The upper mantle materials which injected into the mantle wedge and caused back-arc spreading are identical to those of the mid-oceanic ridge basalt source in Sr-Nd isotopic compositions, because volcanic rocks in the back-arc region of the Northeast Japan arc show secular variation in isotopic compositions from enriched to depleted isotopic characteristics during the last 30 Ma.

Qp‐melting temperature relation in peridotite at high pressure and temperature: Attenuation mechanism and implications for the mechanical properties of the upper mantle

The anelastic properties of compressional waves in a peridotite have been determined in the laboratory at sufficiently high temperatures (to 1280°C) and pressures (to 0.73 GPa) to warrant comparison with seismic measurements of the Earth. A substantial decrease of Qp is observed at temperatures well below the onset of partial melting. Qp systematically increases with increasing pressure over the entire temperature range. Of major significance is the finding that Qp is dependent on the ratio of the temperatures to the melting (solidus) temperature; i.e., Qp depends on the homologous temperature. The pressure dependence of Qp appears through the pressure dependence of the solidus of the peridotite. Within the uncertainties of measurement of both Qp and the phase diagram, it appears that melting and high‐temperature anelastic properties have a common origin in peridotite. The homologous temperature dependence of Qp suggests that we may estimate the temperature and pressure dependence of Qp for peridotites of different compositions and possibly even for hydrous peridotites, if solidus temperatures are known as a function of pressure (a far easier measurement than elastic and anelastic properties). The activation volume of Qp is greatly reduced at high pressure, since the slope of solidus versus pressure rapidly decreases with increasing pressure. Pressure dependence of seismic velocity and melt fraction in peridotite also appears to be related to the homologous temperature. The Qp‐homologous temperature relation suggests a connection between Qp and the properties of the grain boundaries; that is, the major loss of seismic energy occurs at the grain boundaries. Grain boundary relaxation or high‐temperature background attenuation is a possible mechanism for the grain boundary damping. No frequency dependence of Qp is resolved (0<α<0.2 in Qp ∝ fα) over the pressure, temperature and frequency ranges of the measurement. The present results and the model of grain boundary relaxation suggest that an appropriate choice of grain size may give an ultrasonic Q that is applicable to the Earth. Experimentally determined anelastic properties of a peridotite are critical for modeling mechanical properties of the upper mantle. Implications of the results are as follows: (1) Seismic data commonly interpreted as indicating a partially molten asthenosphere may instead reflect a hot solid asthenosphere at 90–100% of the solidus temperature. (2) Partial melting may not produce any abrupt change of seismic velocity and Q; rather, elastic and anelastic properties of the upper mantle will change gradually at the boundary where the geotherm crosses the solidus. (3) There may be no sharp mechanical boundary between the lithosphere and the asthenosphere.

Migration of fluid phases and genesis of basalt magmas in subduction zones

Dehydration and hydration reactions in both the downgoing lithosphere and the overlying mantle wedge have been examined in order to understand the role of H2O in the production of magmas at convergent plate boundaries. The subduction of oceanic lithosphere, occurring with increasing pressures and rising temperatures, causes liberation of H2O from the slab. Amphibole, which can be stable to the highest PT conditions among hydrous phases in the slab, decomposes at around 90 km depth. It follows that the subducted lithosphere is essentially anhydrous beneath volcanic arcs lying more than 110 km above the slab and that the supply of slab‐derived H2O is not a direct trigger for the production of arc magmas. Instead, the H2O released from downgoing lithosphere reacts with the forearc mantle wedge to crystallize hydrous minerals (serpentine, talc, amphibole, chlorite, and phlogopite). This hydrated peridotite is dragged downward on the slab toward higher PT regions and releases H2O to shallower potential magma source regions in the mantle wedge. Combining experimental data on the stability of serpentine and talc with the thermal structure in the mantle wedge, it is concluded that those minerals decompose beneath the forearc region. On the other hand, high PT experimental and thermodynamic data suggest that dehydration of amphibole and chlorite in the downdragged hydrated peridotite can take place just beneath a volcanic front. Phlogopite in the peridotite decomposes to release H2O at a deeper level (about 200 km). H2O liberated from the hydrated peridotite causes partial melting of overlying mantle wedge peridotites. Along with the migration of H2O through the above processes, subduction components, especially large ion lithophile elements, can be overprinted on the magma source region, which governs the geochemical characteristics of arc magmas.

Secular variation of magma source compositions beneath the Northeast Japan arc

Sr and Nd isotopic compositions were determined for Tertiary volcanic rocks with ages of 22, 16 and 8 Ma along the volcanic front in the NE Japan arc in order to examine the secular variation of magma source compositions in the mantle wedge. These rocks show no through-time variation in both Sr and Nd isotope ratios (0.70412–0.704759 and 2.09–4.12 ∈-units, respectively). It follows that the subduction components do not increasingly pollute the magma source region during the continuous subduction of oceanic lithosphere. The degree of pollution by the subduction components is controlled by the amount of clinopyroxene and aluminous phases in mantle wedge peridotites, because H 2O which causes partial melting of the mantle wedge beneath a volcanic front is supplied by the breakdown of both amphibole and chlorite in the dragged hydrated peridotite formed by the addition of slab-derived fluid beneath the fore-arc region.

Petrology and geochemistry of the Kamloops Group volcanics, British Columbia

The Kamloops Group is an alkali-rich calc-alkaline volcanic suite of Early to Middle Eocene age, widespread in south-central British Columbia. Rock types in the suite range from high-K basalt through andesite to rhyolite. The suite is characterized by relatively high K2O, Sr, and Ba, but low Zr, Ti, and Ni concentrations, only moderate Ce enrichment, and little or no Fe enrichment. Initial ratios 87Sr/86Sr are about 0.7040 in the western half, and about 0.7060 in the eastern half of the study area. No difference in chemistry or mineralogy marks this sharp transition. Chemically similar suites include the Absaroka–Gallatin suite in Wyoming and the lower San Juan (Summer Coon) suite in Colorado. The content of K2O at 60% SiO2 increases regularly eastward across southern British Columbia. The chemical data support the subduction-related continental arc origin of the Kamloops Group volcanics.The volcanic rocks consist in the main of augite–pigeonite andesites ranging from 52 to 62% silica, with subordinate quantities of olivine–augite–pigeonite basalt and biotite rhyodacite and rhyolite. The andesites and basalts were derived by a combination of low-pressure fractional crystallization, higher pressure fractional crystallization, and variable parental magmas, whereas low-pressure fractional crystallization of plagioclase, biotite, and apatite from parental basalt and andesite produced the rhyolites. The parental magmas were basalts and basaltic andesites with high K, Sr, and Ba. The primary source of these magmas is inferred to have been an alkali-enriched hydrous peridotite with neither plagioclase nor garnet present in the residuum.

Tectono-magmatic evolution of the West Philippine-Mariana region and the origin of boninites

Re-examination of published data on the tectomagmatic evolution of the West Philippine-Mariana region indicate that arc magmatism and back-arc extensional pulses are not synchronous but are largely asynchronous. Arc volcanism ceases within a few million years of the development of a back-arc basin, and recommences oceanward on a new arc during the final stages in the development of the back-arc basin. Following Karig's model, we believe that ascent beneath the arc axial chain of a line of diapirs of MORB-source mantle interferes with processes of arc magma generation and arc magmatism wanes. Partial melting of the diapirs produces voluminous MORB-type tholeiites which split the arc and form the crust of a widening back-arc basin. We show for both the West Philippine-Mariana region, and occurrences of boninite-like lavas in ophiolites, that boninites appear to be erupted after arc magmatism and immediately before eruption of MORB-type lavas. Passage of a diapir of hot MORB-source mantle through refractory sub-arc peridotite which has been enriched in LILE and H 2O, will result in limited partial melting of the hydrous peridotite and the restricted production of boninitic magmas. These are more likely to erupt modified or only slightly modified by crystal fractionation through the thin forearc crust. Continued ascent of the MORB-source mantle diapir will result in it partially melting to yield MORB tholeiites which erupt shortly after, and in considerably more voluminous amounts, than the boninites. Finally we suggest that forearcs regions are likely to be incorporated into foldbelts and that boninite (or low-Ti lava)-bearing ophiolites may characterize such regions.

Experimental phase relations of hydrous peridotites modelled in the system H2O-CaO-MgO-Al2O3,-SiO2

The hydration of peridotites modelled by the system H2O-CaO-MgO-Al2O3-SiO2 has been treated theoretically after the method of Schreinemakers, and has been investigated experimentally in the temperature range 700°–900° C and in the pressure range of 8–14 kbar. In the presence of excess forsterite and water, the garnet- to spinel-peridotite transition boundary intersects the chlorite dehydration boundary at an invariant point situated at 865±5° C and 15.2±0.3 kbar. At lower pressures, a model spinel lherzolite hydrates to both chlorite- and amphibole-bearing assemblages at an invariant point located at 825±10° C and 9.3±0.5 kbar. At even lower pressures the spinel-to plagioclase-peridotite transition boundary intersects the dehydration curve for amphibole+forsterite at an invariant point estimated to lie at 855±10° C and 6.5±0.5 kbar. Both chlorite and amphibole were characterized along their respective dehydration curves. Chlorite was found to shift continuously from clinochlore, with increasing temperature, to more aluminous compositions. Amphibole was found to be tremolitic with a maximmum of 6 wt.% Al2O3. The experimentally determined curves in this study were combined with the determined or estimated stability curves for hydrous melting, plagioclase, talc, anthophyllite, and antigorite to obtain a petrogenetic grid applicable to peridotites, modelled by the system H2O-CaO-MgO-Al2O3-SiO2, that covers a wide range of geological conditions. Direct applications of this grid, although quite limited, can be made for ultramafic assemblages that have been extensively re-equilibrated at greenschist to amphibolite facies metamorphism and for some highgrade ultramafic assemblages that display clear signs of retrogressive metamorphism.

Clinoenstatite in boninites from the Bonin Islands, Japan

Highly magnesian andesitic lavas were considered a rare and unusual rock type, before boninites proved to be so extensive among the Tertiary volcanic rocks in the Benin–Mariana region1–5 and some ophiolitic basalts6,7. Protoenstatite is also a rare mineral and its terrestrial occurrence has been confined to the high-Mg andesites from Cape Vogel, eastern Papua8, and the Mariana Trench3. Some boninites from the Bonin Islands were found to contain abundant multiply-twinned clinoenstatite (inverted protoenstatite) with or without olivine. The clinoenstatite, coexisting with olivine with a similar forsterite content to that in the mantle, shows that protoenstatite crystallized later than the olivine, whereas in the olivine-free, clinoenstatite-bearing rock which has more SiO2 than the clinoenstatite-bearing or -free boninites with olivine, the protoenstatite may have been the liquidus mineral. We show here that the presence of inverted protoenstatite in the Bonin Islands confirms our earlier suggestion1,9,10 that boninites may have been formed by extensive partial melting of hydrous peridotite at relatively low pressures and rapid quenching from high temperatures at shallow depths.

Water in Some Hypersthenic Magmas

The concentrations of in inclusions of glass in phenocrysts are estimated indirectly by summation difference using electron probe microanalysis. The method probably is valid if inclusions are glassy (not devitrified) and larger than about 30 μm diameter and if more than about is present. Inclusions of melt may lose by diffusion and/or oxidation after initial entrapment. The concentrations of quenched in basaltic and andesitic glasses in olivine phenocrysts are mostly between 2 and 4 wt % for Pavlof Volcano, Alaska; Asama Volcano, Japan; Fuego Volcano, Guatemala; and Mt. Shasta, California. The revised estimate of the rate of igneous outgassing of in subduction zones is twice that for fresh submarine ridge basalts and suggests that most of the added to oceanic crust during alteration and weathering either is recycled by some means other than magmatism or is stored deep in the earth. The ratios of associated silicic glasses commonly are less than for the low-silica glasses. Silicic glasses commonly have equal or lesser concentrations of compared to associated less silicic glasses. Probably many silicic hypersthenic magmas develop in vapor saturated bodies of magma within a few kilometers of the surface. Some silicic hypersthenic magmas appear to have about and probably developed at greater depths either by differentiation from hydrous basaltic parents or by contamination with -rich dacitic melts derived from partial fusion of hydrous peridotites within the crust.

The Rare Earth Content and Origin of Alkali-Rich Basalts

The rare earth content of alkali-rich, silica-undersaturated basalts erupted onto oceanic and continental plates is consistent with their derivation by slight amounts of partial melting of hydrous garnet peridotite at the top of the asthenosphere, within the low-velocity, low-Q zone of the mantle. The rare earths are fractionated during partial melting by retention of heavy rare earths relative to light rare earths in garnet and clinopyroxene in the mantle residue. We conclude from a melting model that the amount of melting of hydrous peridotite at the top of the asthenosphere and its clinopyroxene-garnet ratio increase with decreasing thickness of the overlying dry lithospheric plate. Thus, potassic basalts derived by 0.5% melting of peridotite with a clinopyroxene-garnet ratio of 0.3 come from deeper than nephelinites (1% melting, ratio of 1), and alkali basalts (2% melting, ratio of 5). Convection within the partly molten asthenosphere establishes an adiabatic geothermal gradient and localizes renewable magma sources into updwelling plumes, or hot spots, that remain relatively stationary with respect to the overlying lithospheric plates. Unlike the source regions of midoceanic ridge tholeiite basalts, the peridotite source regions of the alkali-rich basalts do not appear to have been previously depleted in large cations (Ba, Rb, light rare earths) and generally have more radiogenic Sr and Pb. If the peridotite source regions of the alkali-rich basalts have chondritic proportions of the rare earths, the partial melting model indicates two to five times chondritic abundance of the rare earths in the peridotite.