Mid-ocean Ridge Basalt Magmas Research Articles

Volatiles (He, C, N, Ar) in mid-ocean ridge basalts: assesment of shallow-level fractionation and characterization of source composition

To document the characteristics of volatiles in the terrestrial mantle, the abundances and the isotopic ratios of carbon, nitrogen, helium, and argon have been analyzed in 45 mid-ocean ridge basalts (MORB) glasses from the Mid-Atlantic Ridge between 24°N and 36°N, the East Pacific Rise at 21°N, 13°N, and 17–19°S, the Red Sea (18–20°N), the Indian Ocean near the Triple Junction, and the central North Fiji basin. Gases were extracted by crushing and subsequently were split for purification and analysis. Static mass spectrometry was used for N, He, and Ar, and conventional dynamic mass spectrometry was used for C. The data confirm the occurrence of near-constant He isotope ratios ( 3He/ 4He = 8.53 ± 0.79 Ra; n = 36) and C isotope ratios (δ 13C = −5.2 ± 0.7‰ vs. PDB; n = 21), and of a light nitrogen component in the convective mantle. Overall, the δ 15N signature of MORB (mean δ 15N = −3.3 ± 1.0‰ vs. ATM, for samples with 40Ar/ 36Ar > 1000; n = 19) does not drastically differ from that of diamonds, some of which being presumably ≥3 Ga, and is thought to represent the nitrogen isotopic ratio of the asthenospheric mantle. A common mantle source for diamond-bearing magmas and MORB magmas is unlikely, and this similarity may imply either active exchange of volatiles between the respective mantle sources or homogeneous distribution of nitrogen isotopes in these sources during most of Earth’s history. Variations of the 4He/ 40Ar∗ ratios as well as 40Ar/ 36Ar ratios are consistent with fractional crystallization–assimilation–degassing taking place in the depth range of 3 to 6 km. The corrected average C/N of the MORB mantle is 535 ± 224, significantly higher than potential cosmochemical and geochemical end-members. C/He and C/N ratios corrected for fractional crystallization–assimilation–degassing fractionation increase with the degree of MORB enrichment (e.g., increasing K 2O/TiO 2) and are thought to reflect carbon heterogeneities in the mantle source, as a result of fractional recycling of a (fluid?) component extremely enriched in carbon (C/N > 3000). The N/ 3He ratios are less variable than the C/ 3He ratios, suggesting limited recycling of nitrogen. Such limited recycling is required to preserve a N isotope ratio in the convective mantle distinct from that of the atmosphere–hydrosphere sediments. Overall, neon, argon, nitrogen, and carbon abundances in the mantle appear to be chondritic, rather than solar, although the neon isotopic signature of the mantle indicates contribution of a solar component. This apparent discrepancy may reflect mixing between a solar-type component mostly seen at present in light rare gases and a chondritic-type component and has strong implications for the origin of terrestrial matter and the processes of their accretion.

Plume-derived rare gases in 380 Ma carbonatites from the Kola region (Russia) and the argon isotopic composition in the deep mantle

In an effort to document the source of the parental melts to carbonatites, we have measured rare gases in 380 Ma carbonatites and associated mineral assemblages from the Kola Peninsula, eastern part of the Baltic shield in Russia. These series were emplaced during widespread Devonian magmatism when several large ultrabasic–alkaline–carbonatite massifs were formed. 4He/ 3He ratios vary from 1×10 6 to 1×10 7 in the bulk He extracted by melting of samples from three localities, including the large Kovdor massif. A comparison of measured abundances of 3He and 4He with those expected from in-situ production revealed a large (up to 10 5 times) excess of 3He, implying a significant contribution from a mantle-derived 3He-bearing fluid. Crushing of these samples allowed extraction of fluids with 4He/ 3He ratios down to 38,000, lower than those of mid-ocean ridge basalts and in the range of 4He/ 3He observed in 3He-rich ocean island basalts (OIBs) related to mantle plumes. 20Ne/ 22Ne up to 12.1±0.2 are higher than the atmospheric value of 9.80, implying the occurrence of primordial (solar-type) neon in the carbonatite source. 20Ne/ 22Ne and 21Ne/ 22Ne ratios display a good correlation, with the regression line close to (slightly to the right of) the Loihi Seamount correlation. Extrapolation of the regression to solar 20Ne/ 22Ne of 13.8 gives a 21Ne/ 22Ne of 0.045 for the plume end-member, well below the mid-ocean ridge basalt (MORB) source (upper mantle) end-member of 0.07. The measured 40Ar/ 36Ar ratios up to 2790 correlate very well with the Ne isotopic ratios, and the best estimate of the 40Ar/ 36Ar ratio of the plume source is within 5000±1000. Although the 3He/ 22Ne ratio in the plume source appears to be comparable to the solar value within a factor of 2, the 22Ne/ 36Ar ratio, computed from Ne–Ar isotope correlation, is two orders of magnitude lower than the solar value. Such difference is unlikely to be due to magmatic fractionation since the observed 4He/ 40Ar * ratios are close to values expected for radiogenic production and accumulation in the mantle source. It may rather represent a characteristic of the plume source. The isotope composition of light noble gases in samples from ultrabasic–alkaline rocks of the Kola Peninsula, and associated carbonatites, indicate a contribution of material with lower time-integrated (U + Th)/( 3He, 22Ne) and ( 40K/ 36Ar) ratios than those in the asthenospheric upper mantle, the subcontinental lithosphere, and the continental crust. The location of such material is likely to be below the convective mantle supplying MORB magmas, and reflects the contribution of a plume source material to Kola carbonatitic magmatism. These data support models which advocate a structure of the Earth heterogeneous in its refractory/volatile content.

Graphite-melt equilibria during mantle melting: constraints on CO 2 in MORB magmas and the carbon content of the mantle

Melts produced under fluid-absent, graphite-saturated conditions have tightly constrained oxidation states and CO 2 contents. A model is presented for the generation of mid-ocean ridge basalt (MORB) magma by fluid-absent partial melting of graphite-saturated upper mantle. The model results in melt ferric/ferrous ratios and oxygen fugacities which are consistent with observed values in glassy MORBs and oceanic peridotites. Increasing the degree of partial melting causes decreases in the ferric/ferrous ratio of the melt, the oxygen fugacity, and the concentration of CO 2 in the melt. Sensitivity analysis indicates that the initial ferric-iron concentration in the source region has the greatest effect on the results. For the range of probable initial values, the oxygen fugacity of the graphite-saturated primary MORB magma at 1 GPa is 2.5 log 10 units below the Ni–NiO buffer, and the CO 2 content of the melt ranges from 900 to 1800 ppm (mass). The amount of source region carbon consumed during partial melting is between 40 and 80 ppm. The dissolved CO 2 in MORBs represents a flux of 3.5 to 7×10 10 kg/year which is significantly greater than estimates based on the measured CO 2 content of glassy MORB samples.

Geochemistry of Basalt from Escanaba Trough: Evidence for Sediment Contamination

Mid-ocean ridge basalts (MORBs) erupted in Escanaba Trough, the southernmost segment of the Gorda Ridge, have greater isotopic and geochemical variability than those from the remainder of the ridge. Samples from the sediment-free northern sites in the trough are depleted normal-MORB similar to those from the northern Gorda Ridge. However, samples from the sediment-covered portion of the trough are significantly enriched in incompatible elements, and Sr, Nd and Pb isotopic ratios are shifted toward the composition of local sediments. Pb isotopic ratios, in particular, lie on a mixing line with local sediments. Assimilation–fractional crystallization (AFC) calculations suggest that 25–40% fractional crystallization coupled with assimilation of 1–2% of local sediments can produce the enrichment observed for some Escanaba basalts. Variations in the amount of enrichment of elements with similar incompatibility for comparably differentiated basalts, and uncorrelated isotopic ratios suggest variable amounts of assimilation of compositionally heterogeneous assimilants by parental magmas. Assimilation of sediment by MORB magmas probably resulted when laccolithic intrusions ponded at the sediment–basement interface beneath the thick sediment cover in the southern part of Escanaba Trough. Geochemical and isotopic evidence for contamination can be detected in these lavas because of the strong compositional contrast between the parental magma and assimilant. The depleted MORB magma and highly enriched sediment assimilant are end-member compositions that provide a sensitive indicator for the smallest amount of contamination of oceanic magma by crustal processes.

Compositions of partial melts formed in mantle peridotites at high pressures and their relation to those of primitive MORB

The compositions of partial melts formed from five different mantle peridotites (Mg# ranging from 85.3 to 90.5) have been determined using the diamond aggregate method. When melt compositions were compared with those of primitive mid-ocean ridge basalt (MORB), it was found that partial melts formed from relatively fertile peridotites at 10–15 kbar were close to, but distinctly lower in CaO than, primitive MORB. However, fractionation of 3–10 wt.% olivine can produce primitive MORB compositions. Partial melts from peridotites with Mg# between about 89 and 85 have FeO and MgO abundances close to those of primitive MORB at pressures of 10–15 kbar, and again, primitive MORB compositions can be reproduced with small amounts of olivine fractionation. However, peridotite with Mg# greater than about 90 cannot produce primitive MORB melts at pressures less than 15 kbar. From such peridotite, primitive MORB melts can be produced by partial melting at pressures greater than 15 kbar, but only after extensive olivine fractionation. Experimental melts produced by polybasic, stepwise fractional melting of relatively fertile mantle peridotite along a model adiabat indicate that the compositions of accumulated, incremental melts produced from 20 kbar to 10 and 5 kbar are close to, but slightly more olivine-rich than, those of primitive MORB. Primitive MORB magmas can be generated by fractional melting over a wide depth range; however, magmas formed over a wide depth range may re-equilibrate with mantle peridotite at shallow levels during their ascent. The fact that the major element abundances of the primitive MORB are rather close to those of partial melts formed by batch melting at about 10 kbar suggests that re-equilibration in major elements may occur at depths of about 30 km or the mean pressure of polybasic melting is close to 10 kbar.

Recent volcanism in the Siqueiros transform fault: picritic basalts and implications for MORB magma genesis

Small constructional volcanic landforms and very fresh-looking lava flows are present along one of the inferred active strike-slip faults that connect two small spreading centers (A and B) in the western portion of the Siqueiros transform domain. The most primitive lavas (picritic and olivine-phyric basalts), exclusively recovered from the young-looking flows within the A-B strike-slip fault, contain millimeter-sized olivine phenocrysts (up to 20 modal%) that have a limited compositional range (Fo 91.5-Fo 89.5) and complexly zoned CrAl spinels. High-MgO (9.5–10.6 wt%) glasses sampled from the young lava flows contain 1–7% olivine phenocrysts (Fo 90.5-Fo 89) that could have formed by equilibrium crystallization from basaltic melts with Mg# values between 71 and 74. These high MgO (and high Al 2O 3) glasses may be near-primary melts from incompatible-element depleted oceanic mantle and little modified by crustal mixing and/or fractionation processes. Phase chemistry and major element systematics indicate that the picritic basalts are not primary liquids and formed by the accumulation of olivine and minor spinel from high-MgO melts (10% < MgO < 14%). Compared to typical N-MORB from the East Pacific Rise, the Siqueiros lavas are more primitive and depleted in incompatible elements. Phase equilibria calculations and comparisons with experimental data and trace element modeling support this hypothesis. They indicate such primary mid-ocean ridge basalt magmas formed by 10–18% accumulative decompression melting in the spinel peridotite field (but small amounts of melting in the garnet peridotite field are not precluded). The compositional variations of the primitive magmas may result from the accumulation of different small batch melt fractions from a polybaric melting column.

Experimental and natural partitioning of Th, U, Pb and other trace elements between garnet, clinopyroxene and basaltic melts

Partition coefficients for Th, U, Pb, rare-earth elements (REE), high field strength elements (HFSE), alkaline-earth elements, Sc, Cr, V and K were measured by ion microprobe techniques in two experiments on a natural high-alumina basalt composition from Medicine Lake, California. All elements were measured at natural abundance levels except Th, U and Pb, which were each present in the starting mix at 1-wt% levels. The results show that garnet retains U preferentially over Th ( D gt melt U = 0.0059 , D gt melt Th = 0.0014 ), while clinopyroxene shows the opposite sense of partitioning ( D cpx melt U = 0.0127 , D cpx melt Th = 0.014 ). The experimental Th, U and Pb partition coefficients for garnet-melt and cpx-melt are consistent with garnet-cpx pairs from garnet-bearing ultramafic rocks which exhibit U-Pb isochrons, thus demonstrating equilibrium ( D gt cpx U = 0.30 , D gt cpx Th = 0.072 , D gt cpx Pb = 0.016 ). The partition coefficients for Th and U between clinopyroxene and basaltic melt vary systematically as a function of the tetrahedral Al content of clinopyroxene. Garnet/melt values for Th, U and Pb agree with previous determinations, indicating that mid-ocean ridge basalt (MORB) generation begins in the stability field of garnet lherzolite. However, high 226 Ra 230 Th ratios in MORB require very small porosities near the region where the melts lose chemical equilibrium with the mantle. Partitioning data for HFSE and REE suggest that this region of melt segregation is not in the spinel lherzolite field. This requires either rapid transport of MORB magmas from ϵ 70 km, or some degree of disequilibrium during melt generation and/or transport.

Basaltic melt evolution of the Hengill Volcanic System, SW Iceland, and evidence for clinopyroxene assimilation in primitive tholeiitic magmas

The thick oceanic crust of Iceland is formed by tholeiitic central volcanoes arranged in en echelon patterns along the 40–50 km wide rift zones. The Hengill central volcano in the southwestern rift zone has produced 25–30 km3 of hyaloclastites and lava during the last 0.11 m.y., with maximum productivity during the isostatic rebound following the degalciations 0.13 and 0.01 m.y. ago. The eruption units cover a compositional spectrum from picrite to rhyolite, but the volume of andesite, dacite, and rhyolite is small. The petrographic relations of pillow rim and hyaloclastite glass indicate that the basaltic melts were saturated with olivine and plagioclase, except for the most primitive ones that were undersaturated with plagioclase. Saturation with clinopyroxene was reached in some of the intermediate and evolved basaltic melts. Corroded and partly resorbed crystals of clinopyroxene and partly disintegrated gabbro nodules with resorbed clinopyroxene indicate that selective assimilation contributed to the evolution of the most primitive melts. The intermediate and evolved basaltic glass compositions fall along the low‐pressure cotectic for mid‐ocean ridge basalt (MORB) compositions saturated with olivine, plagioclase, and clinopyroxene, but the primitive glasses (9–9.7 wt % MgO) fall well inside the low‐pressure olivine + plagioclase primary phase volume. The primitive Hengill glasses have significantly higher CaO and lower Al2O3 than primitive glasses from oceanic spreading centers. Their low pressure undersaturation with respect to clinopyroxene and the absence of clinopyroxene phenocrysts indicate that they are not parental to the intermediate Hengill basalts, since fractionation modelling requires a large proportion of clinopyroxene in the fractionating assemblage. The most primitive melts could be produced by fractionation of olivine and plagioclase combined with 5–30 % assimilation of clinopyroxene, and the intermediate melts could be derived by mainly olivine and plagioclase fractionation, from melts equilibrated with peridotitic residues at pressures of 1–2 GPa. The further evolution of the basaltic spectrum can be explained by fractionation of olivine, plagioclase, and clinopyroxene combined with minor contamination by anatectic crustal melts. The rate of magma supply from the mantle to the crust is controlled by the isostatic conditions and is very low in periods of glacial loading of the crust. This leads to infrequent and small eruptions of dominantly evolved magmas. The dense picritic magmas (9–9.7 wt % MgO in the glass phase) were driven to the surface by magmatic overpressure in the mantle at an early deglaciation stage characterized by the absence of large, trapping magma chambers in the lower crust. The assimilation of clinopyroxene in these melts could proceed by direct contact with the solidified cumulate sequences and gabbro intrusions. Clinopyroxene assimilation in combination with olivine fractionation may also contribute to the chemical evolution of some of the most primitive MORB magmas.

Sulfide globules in mid-ocean ridge basalts (MORB), and the effect of oxygen abundance in FeSO liquids on the ability of those liquids to partition metals from MORB and komatiite magmas

The coexistence of sulfide liquids (or more accurately, FeSO liquids) with silicate liquids in a variety of magma systems is well known, as is the ability of immiscible FeSO liquids to segregate Ni, Cu and platinum-group metals into orebodies. Evidence for association of FeSO liquids with magmas that form mid-ocean ridge basalts (MORB) comes from sulfide globules in basalt glass from Siqueiros Fracture Zone, East Pacific Rise; three fracture zones in the Indian Ocean; Lau Basin; and ∼ 23 previously published oceanic localities. Approximately 40% of the sulfur originally present in magma at Siqueiros Fracture Zone exsolves as a sulfide liquid during fractional crystallization. Virtually all of the Cu, and 20% of the Ni, originally in the magma is extracted by the sulfide liquid. Significant variations in the ability of FeSO liquids to partition Ni Cu exist among MORB and other mafic magmas, such as komatiites and layered mafic intrusions. Nernst partition coefficients (sulfide to silicate melt; D S L ) are 2–3 times higher for Ni in MORB than in komatiite, and ∼ 5 times higher for Cu. Reasons for this variation are not entirely clear, but a possible contributing factor is that FeSO liquids in MORB appear to have a higher oxygen content than those in komatiites. That FeSO liquids associated with MORB magma are oxygen-rich is suggested by low observed values at Siqueiros Fracture Zone for the NiFe exchange coefficient K D3 [1.21−4.56; K D3 = ( Fe Ni ) olivine ( Fe Ni ) sulfide ]. It has been previously shown that natural magmatic FeSO liquids are rich in O 2− [ O ( S + O) is apparently always greater than ∼ 0.49], and that K D3 is inversely correlated with O ( S + O) . K D3-values below ∼ 12 would result from FeSO liquids with O ( S + O) > 0.49 . K D3-value of> 20 would be impossible (values from 12 to 20 would be unlikely) because they would require an FeSO liquid poor in O 2−. This argument for low K D3 in magmas is strengthened by the low K D3-values of MORB, whose globules are inarguably of magmatic origin. K D3 in komatiites is higher than in MORB (indicating a lower oxygen fugacity in komatiite magma) and varies from ∼ 5 to > 20, with values of > 20 coming from highly metamorphosed komatiites. Fleet and MacRae argued on the basis of experimental results that K D3-values of < 20 could not result from equilibration at magmatic temperatures and that orebodies associated with komatiites would have to be of hydrothermal origin. But those experiments were carried out at O ( S + O) = 0 , an unrealistic value for magmatic FeSO liquids. K D3-values in komatiites of > 20 could not have resulted from magmatic processes, but might have come from metamorphism under oxygen-deficient conditions.

Mid-ocean ridge popping rocks: implications for degassing at ridge crests

The vesicle size distribution (VSD) and rare gas abundances in popping rocks from 14°N on the Mid-Atlantic Ridge provide constraints on the behavior of volatiles during ridge crest volcanism. These popping rocks, which contain 16–18 volume percent vesicles, are rare mid-ocean ridge basalt (MORB) magmas which appear to have retained much of their volatile inventory. The logarithm of vesicle population density displays the same linear correlation with decreasing size in two of the samples studied. This implies that continuous and simultaneous nucleation and bubble growth have occurred during magma ascent, with no significant perturbations due to accumulation, coalescence or loss of bubbles. In contrast, most MORB magmas display low vesicularities and we suggest that they have suffered some degree of pre-eruptive vesicle loss. We tentatively propose that large vesicles are produced by coalescence when MORB melt is at rest in chambers and conduits, and may be lost during early gas-rich episodes. Most MORB would represent residual liquids which erupt after vesicle loss has occurred, whereas popping rocks would represent a rare case where physical sorting of vesicles from melt did not occur, because storage in a magma chamber did not occur. The rare gas concentrations in the studied popping rocks are the highest yet measured in glassy ridge basalts ([He] > 50 μccSTP/g). The rare gas abundance pattern of these popping rocks probably resembles the pattern for non-vesiculated MORB magma and potentially reflects that of the depleted mantle source. This pattern is similar to the “mean MORB” pattern (computed from MORB glasses with 40Ar/ 36Ar > 10,000) although a higher enrichment in He (and possibly Ne) compared to the heavier rare gases is observed in MORB. The overall similarity in abundance patterns for MORB and popping rocks indicates that vesiculation and vesicle loss do not fractionate the Ar Kr Xe relative abundances from those in non-vesiculated magma, and that the modern flux ratios of these gases at ridges are similar to their elemental ratios in the depleted mantle. The degassing flux of He at ridge crests estimated from the MORB He deficit relative to popping rocks is comparable to the flux derived from the 3He budget for the abyssal ocean. This suggests that degassing at ridges may be strongly influenced by the dynamics and style of submarine volcanism.

Sulfur saturation and second-stage melts; application to the Bushveld platinum metal deposits

A model is presented for the formation of platinum-group element ores by precipitation from sulfur-deficient, platinum-group element-rich, second-stage magmas. Numerous investigations indicate that basaltic magmas originating from undepleted or mildly depleted source regions are sulfur saturated at the time of eruption. Mantle-derived peridotites residual from midocean ridge basalt magma genesis have Pd abundances that are uniformly too high for it to have originated solely from silicate melt retained in the source region at the time of segregation; they imply the presence of an accessory immiscible sulfide melt to host the platinum-group elements. The very high sulfide melt-silicate melt partition coefficients of the platinum-group elements would result in strong fractionation of these metals into the accessory sulfide component during the melting event. The extreme platinum-group element enrichment of this accessory sulfide component is evidenced by the discovery ofplatinoid minerals in residual mantle nodules containing Pd (and Pt) abundances at the low ppb level (Keays et al., 1981).Magmas generated by low to moderate degrees of partial melting of undepleted mantle are therefore sulfur saturated at the time of segregation. These (first-stage) magmas become impoverished in platinum-group elements during the early stages of fractional crystallization because of coprecipitation of an immiscible sulfide phase (e.g., midocean ridge basalt glasses average

Oxidation states of mid-ocean ridge basalt glasses

Precise new analyses of ferrous and total iron for 78 hand-picked mid-ocean ridge basalt (MORB) glasses constrain the redox states of MORB magmas, and the systematics of those redox states with respect to geography and chemistry. The data indicate that MORB magmas at the point of eruption include some of the most reduced terrestrial lavas yet analyzed. Irrespective of their tectonic setting, geographic setting or chemical type, quench glasses from the outer surfaces of MORB lavas have Fe 3+/ΣFe ratios of 0.07 ± 0.03 (2σ), equivalent to relative oxygen fugacities 1–2 log 10 units below the fayalite-magnetite-quartz buffer (FMQ). These reduced values contrast markedly with those of cogenetic whole rocks, even for samples from the same pillow. Pillow cores (including virtually all published whole rock analyses) typically have Fe 3+/ΣFe close to 0.15, equivalent to oxygen fugacities close to FMQ. The post-eruptive oxidation responsible for this contrast apparently occurs rapidly, while the lava is still partially molten, most likely as a result of hydrogen loss. The striking contrast between glass and whole rock data suggests that two common assumptions with respect to treatment of MORB are inappropriate: (1) assumption of Fe 3+/ΣFe close to 0.15 for calculating norms and magnesium numbers; and (2) assumption of FMQ conditions for experimental studies of MORB crystallization. Oxygen fugacities in the mantle source regions of MORB have also been assumed to be close to FMQ. It is likely that, redox states determined from wholly or partially crystalline lavas are generally not representative of magmatic values. The new data reported herein define a new and rigid upper limit for the oxidation state of the sub-oceanic mantle.

Comment on ?Soret separation of mid-ocean ridge basalt magma? by D. Walker and S.E. DeLong

Comment on ?Soret separation of mid-ocean ridge basalt magma? by D. Walker and S.E. DeLong

Comment on “MORB—A Mohole misbegotten?” by M.J. O'Hara

One can only applaud O'Hara's call for further research into magmatic processes that take place at and near oceanic spreading centers (EOS, June 15, 1982). In particular, the desirability of having deep drill holes close to active spreading centers is indisputable and needs no elaboration. One of the major reasons for encouraging further research on midocean ridge basalt (MORB) petrogenesis is the need to resolve a controversy now being debated in the literature over the composition and depth of origin of primary MORB magmas [Rhodes, 1982]. O'Hara [1968a] and, recently, several others [Green et al., 1979; Stolper, 1980; Elthon and Scarfe, 1980] have argued that primary MORB magmas are picritic in composition and generated at various pressures from 15 to 30 kbar, depending on the author. Others contend that primary MORB magmas are close in composition to the least‐fractionated MORB glasses (which are not picritic but have mg numbers as high as 73) and are generated at more moderate pressures of about 8–10 kbar [Green and Ringwood, 1967; Kushiro, 1973; Presnall et al., 1979]. O'Hara did not mention any of the recent papers that disagreed with his concept of picritic primary magmas at spreading centers. In order to restore some balance, the following brief discussion is offered.

Phase Transformations and Differentiation in Subducted Lithosphere: Implications for Mantle Dynamics, Basalt Petrogenesis, and Crustal Evolution

Mantle pyrolite differentiates at mid-oceanic ridges to form a layered lithosphere consisting of a basaltic crust, immediately underlain by harzburgite and further underlain by pyrolite which has experienced depletion only of highly incompatible elements (e.g., Rb, light REE). The body forces driving subduction are concentrated mainly in the upper cool, brittle layer of lithosphere, comprised of basalt and harzburgite. The lower layer of relatively ductile pyrolite is stripped off during subduction and resorbed into the upper mantle. This material, which is depleted in highly incompatible elements, provides a future source region of mid-ocean ridge basalt magmas on a timescale of 10⁹ years. The Nd, Sr, and Pb isotopic characteristics of MORBs are explained on the basis of this model. The slab, which sinks to ~600 km, is comprised mainly of former basalt and harzburgite. These differentiated layers undergo a significantly different series of phase transformations to those experienced by mantle pyrolite. The characteristics of these phase transformations and their influence on the density contrast between the slab and surrounding pyrolite are reviewed in detail. They cause the cool subducted lithosphere to remain denser than surrounding mantle to 600-650 km. Below 650 km, former basaltic crust remains denser than surrounding mantle, whereas former harzburgite becomes relatively buoyant. The resulting non-uniformity in stress distribution causes the slab to buckle and to accrete to form a large, relatively cool, ovoid "megalith" of mixed former harzburgite and basaltic crust, sitting on the seismic discontinuity at 650 km. The megalith (dimension >300 km) is heated mainly by thermal conduction, and thermally equilibrates with surrounding mantle on a timescale of 1-2 b.y. Partial melting of entrained former basaltic crust ensues. The resultant liquids contaminate surrounding regions of former harzburgite, rendering them fertile in the sense of future capacity to produce basaltic magmas. As the megalith warms up, its viscosity falls, and large dense blocks of former oceanic crust (now depleted in incompatible elements by partial melting) sink into the lower mantle. Newly fertile, former harzburgite is now buoyant. Diapirs of this material separate and rise into the upper mantle, becoming incorporated in the lithosphere, and experience small degrees of partial melting to produce the alkaline basaltic suite. In oceanic regions, the rising diapirs are responsible for "hot-spot" alkaline volcanism whereas, in continental regions, the upwelling diapirs cause doming and rifting, also accompanied by alkaline activity. The residual components of the diapirs become permanently incorporated into the sub-continental lithosphere. This is a cumulative process and is ultimately responsible for the development of the chemical, physical, and isotopic characteristics of the sub-continental lithosphere. The U-Pb, Nd-Sm, and Rb-Sr characteristics of alkaline (and calcalkaline) associations can be explained in terms of the model. The geochemical evolution of the continental crust and its formation from the mantle by multistage irreversible differentiation processes are also examined within the framework of the model. The extraction of highly incompatible elements (e.g., Rb, K, Ba, U) is essentially decoupled from the extraction of the major elements of the crust (Si, Al, Ca, Fe, Mg, Na). The incompatible elements derive via MORB source regions from a reservoir comprising the upper 650 km of the mantle. The sialic elements, on the other hand, are derived from much more localized reservoirs within this region, namely the mantle wedges overlying subduction zones. The final products of long-term irreversible differentiation of mantle pyrolite are the sialic continental crust, the sub-continental lithosphere of depleted peridotite, and the former oceanic crust, which is consigned ultimately to the lower mantle.

Mineralogy and geochemistry of volcanic and plutonic rocks from the boundaries of the Caroline plate: Tectonic implications

Chemical and mineralogical data from rocks dredged at six sites along plate boundaries around the Caroline Sea define at least three separate igneous suites. Three sites were located in the Sorol Trough, along the Caroline-Pacific plate boundary, and three sites were located in the Ayu Trough along the Caroline-Philippine plate boundary. Geophysical data suggest that both features have originated through processes of crustal extension. Weathered basalts with mid-ocean ridge basalt (MORB) affinities were recovered from the base of the northern margin of the Sorol Trough. Fresh transitional-to-alkalic pillow basalts from a seamount in the central part of the trough have geochemical and Sr-isotopic characteristics similar to enriched MORB or volcanics associated with off-ridge volcanism. A suite of tholeiitic rocks consisting of high-Mg basalt, ferrobasalt and a variety of mafic to ultramafic rocks dredged from the western section of the Sorol Trough has chemical and mineralogical characteristics of other intraplate tholeiitic suites like those from Hawaii. The ferrobasalts could have been derived from a high-Mg basaltic liquid through fractional crystallization of olivine, clinopyroxene and plagioclase. The geochemical data indicate that both MORB and “hot-spot” type sources have been involved in the generation of magmas beneath the Sorol Trough. Furthermore, the highly metamorphosed and tectonized mafic to ultramafic rocks recovered with the tholeiites suggest that a significant component of shear accompanied crustal accretion along “leaky” transform faults within the Sorol Trough. The chemical and isotopic characteristics of fresh to weathered pillow-lava fragments from the axial rift and eastern margin of the Ayu Trough, suggest that a MORB-type source was tapped during crustal accretion along a central spreading centre. An assemblage of metabasalts, metagabbros and silicic plutonic rocks (plagiogranites), recovered from a scarp along the western margin of the trough, have petrologic and chemical features analogous to rocks from plutonic complexes in ophiolites and some oceanic ridges. The plagiogranites could be differentiates from a relatively hydrous MORB magma, but cannot be related to the slightly LIL-enriched basalts in the Ayu Trough by crystal fractionation. Two separate mantle sources are implicated. Metamorphic features in the plutonic rocks are inferred to be the result of hydrothermal and dynamic events associated with the inception of spreading and to subsequent tectonism which exposed them. The petrochemistry of rocks from the Ayu and Sorol troughs is consistent with the hypothesis that crustal accretion occurred along the northern and western boundaries of the Caroline plate and further suggests that a variety of sub-oceanic mantle sources have been involved in the genesis of the crust in each region.

Chemical characteristics of island-arc basalts: Implications for mantle sources

Major-element, trace-element and isotopic compositions of approximately 1200 basalts (< 53 wt. % SiO 2) from intra-oceanic island arcs have been compiled to assess the nature and possible sources of primitive island-arc basalts (IAB). The chemical characteristics of IAB are examined with reference to those of mid-ocean ridge basalts (MORB) and intraplate oceanic basalts (IPB). Major-element compositions of primitive [ Mg ( Mg +Fe 2+) > 65 ] IAB and MORB are similar, but differ significantly from IPB. In general, IAB do not have higher Al 2O 3, lower TiO 2 or a lack of Fe enrichment compared to primitive MORB but many do have greater K 2O contents. Differences in major- and minor-element contents between more evolved IAB and MORB result from the dominance of plagioclase + olivine crystal fractionation in MORB magmas vs. clinopyroxene + olivine controlled fractionation in IAB suites. This difference in crystallization history may be related to the higher P H 2O or greater depth of crystallization of IAB magmas compared to those inferred for MORB. IAB are characteristically enriched in large-ion-lithophile (LIL) elements and depleted in high-field-strength ions (e.g., Zr, Nb and Hf) relative to normal MORB (N-type) and IPB. The enrichment of some LIL elements (e.g., Sr, Rb, Ba and Pb) relative to the rare-earth elements in IAB is difficult to explain by simple partial melting alone and suggests a multistage petrogenesis involving an LIL-enriched component. Low abundances of high-field-strength ions in evolved IAB are explicable in terms of fractional crystallization, but the cause for consistently low abundances in primitive IAB remains problematic. Island-arc lavas contain greater concentrations of volatiles and have higher CO 2 H 2 O and Cl/F ratios than either MORB or IPB, suggesting involvement of a slab-derived volatile component. However, this is not consistent with 3 He 4 He data which indicate that only near-trench volcanics have been significantly affected by dehydration of the oceanic crust. Sr-, Nd-, Pb- and O-isotopic data, in conjunction with the trace-element data, clearly indicate that IAB are derived from heterogeneous, LIL-depleted mantle sources most similar to those which give rise to enriched MORB (E-type). The marked shift towards higher 87 Sr 86 Sr in IAB compared to oceanic lavas with similar 143 Nd 144 Nd values cannot be explained simply by the addition of radiogenic Sr from the slab. Variable degrees of contamination from a crustally-derived sedimentary component is consistent with the isotopic and trace-element data from a number of arcs. However, the lack of correlation between LIL/REE ratios and more radiogenic isotopic ratios suggests that this enrichment/contamination process is complex. A multi-stage petrogenetic model involving subducted oceanic crust (± sediments), dehydration/volatile transfer, and partial melting of metasomatized mantle beneath island arcs is considered the most reasonable, although least constrained, method to generate a variety of primitive IAB.