T42B-01
Variations in Seismic Anisotropy and Olivine LPO in Peridotites From Four Mid-Ocean Ridges: the Effect of Melt?
We present results of an electron backscatter diffraction (EBSD) study of abyssal peridotites from four mid- ocean ridges. Included in the study are peridotites from: a) the Mid-Atlantic Ridge (MAR) at 15°39' N (full spreading rate ~2.6 cm/year), b) the Atlantis II Fracture Zone (57°E) on the Southwest Indian Ridge (SWIR) (full spreading rate ~1.4 cm/year), c) the Gakkel Ridge at 84°38'N and 4°13'E (full spreading rate ~1.2 cm/year), and d) Hess Deep (2°N) on the East Pacific Rise (EPR) (full spreading rate ~12.3 cm/year). All samples preserve distinct crystal lattice-preferred orientations (LPOs) indicative of high-temperature (>1100°C) deformation in the dislocation creep regime. LPOs were used to predict seismic properties following the method of Mainprice (1990). The samples from the MAR and SWIR preserve olivine [010] axes perpendicular to the foliation, and olivine [100] axis maxima within the plane of foliation, indicative of slip in the (010)[100] or "Type A" slip system commonly observed in high- temperature environments of low differential stress and low water content. Predicted P-wave anisotropy ranges from 5.8 to 8.6%. The Gakkel Ridge sample also preserves an olivine [010] axis maxima perpendicular to the foliation, but has a girdle of olivine [100] axes within the plane of the foliation. Holtzman et al. (2003) observed this latter LPO pattern during deformation experiments of olivine + MORB, and suggested that it is characteristic of olivine deformation in the presence of oriented melt pockets. Predicted P-wave anisotropy is 6.8%. The EPR sample preserves both olivine [010] and [100] axis maxima within the plane of the foliation and olivine [001] axis maxima perpendicular to the foliation. We interpret the EPR olivine slip system as (001)[100], which is in agreement with findings by Boudier et al. (1996) from the same sample suite. This"Type E" slip system is thought to form in high-temperature low diferential stress environments, either with high water content (e.g., Katayama et al., 2004) or high melt content (Tommasi et al., 2007). Predicted P-wave anisotropy is 10.4%. The olivine [100] axis maximum is parallel to the seismically fast direction, and is also the slip direction corresponding to the mantle flow direction. Therefore, assuming a horizontal foliation in the oceanic lithosphere and a spreading-parallel flow direction, the seismically fast direction is horizontal and ridge- perpendicular in all cases. However, the observed LPOs require that the seismically slow direction is not consistent in all the samples. In samples from the Gakkel, SWIR, and MAR the seismically slow direction is oriented perpendicular to the foliation and hence vertically oriented in the lithosphere. However, in the EPR sample the seismically slow direction lies within the plane of the foliation and is therefore horizontally oriented in the oceanic lithosphere. Thus, we predict that the Gakkel, SWIR and MAR samples have the maximum seismic anisotropy in the vertical plane and the EPR samples have the maximum anisotropy in the horizontal plane. These results suggest that the mantle peridotite composing the oceanic lithosphere may preserve a variety of LPO patterns across the global range of spreading rates. The differences may reflect, for example, a higher melt content in the shallow mantle at fast spreading ridges than at slow, as well as inefficient melt extraction leaving small pockets of melt widely distributed at the slowest spreading rates.
T42B-02
Trace element poor clinopyroxene in mid-ocean ridge peridotites : is garnet required ?
Serpentinized spinel peridotites from recent ODP drilling (ODP legs 209 and 210) and elsewhere along mid- ocean ridges have preserved relic mineral compositions that are among the most depleted abyssal peridotites observed worldwide, A common feature of all these peridotite is their refractory mineralogy, with Cr number in spinel varying from 0.5-0.66 and cpx trace element compositions characterized by simultaneous low Yb and Lu contents and low MREE/HREE ratios. Application of simple melting models (either batch, fractional or critical) suggest that such signatures are aquired by a polybaric process, with a first step in the garnet stability field to generate the low MREE/HREE followed by a second step at lower pressure in the spinel stability field to decrease the HREE content. This results in a total degree of melting exceeding 20% and reaching up to 25%. These estimates are, however, on the high side of cpx stability in the residue during melting of peridotite; partial melting experiments indicate that cpx is exhausted from the residue at melt fractions of ca. 20% , even for the more fertile starting compositions. We present new data from ODP leg 210 that suggest the presence of exsolved pigeonite coexisting with opx in highly depleted spinel peridotite. Digital images converted to pyroxene phase proportions show up to 20% cpx is dissolved in opx. We determined the trace element compositions of pyroxenes by Laser Ablation coupled to a sector field - HR- ICP-MS that was newly established in our lab. By integrating cpx compositions into opx we obtain primary compositions that are pigeonitic with trace element characteristics similar to opx. Application of opx-melt partition coefficients shows that such pyroxenes are in equilibrium with MORB-like compositions. Cpx exsolution from pigeonitic pyroxenes during subsolidus cooling might produce cpx-bearing harzburgites that were originally cpx-free. We argue that low Yb and Lu, and very low MREE/HREE ratios in cpx could be produced by high degree melting in the spinel peridotite field at high temperatures and do not necessarily imply a first stage of melting in the garnet stability field followed by melting in the spinel stability field. At the meeting we will present some consequences for the interpretation of peridotites formed at ultra-slow spreading ridges and ocean-continent transitions.
T42B-03
Chromite-Hosted Hydrous Melt Inclusions in Oceanic Dunites
Dunites are commonly interpreted to serve as melt transport channels to the crust at mid-ocean ridges. This interpretation is based on field observations in the mantle section of ophiolites, and is supported by petrological studies and numerical physical models. Both observations and models emphasize that such dunites are replacive features, which are formed by focused reactive melt transport at high melt-rock ratios. On the ocean floor, mantle rocks are directly exposed, but are confined mainly to fracture zones of slow to ultraslow spreading ridges. Along the Sparsely Magmatic Zone (SMZ) of the ultraslow spreading Gakkel Ridge (Arctic Ocean), partly serpentinized mantle peridotites represent the dominant lithology exposed along the axial valley walls. Orthopyroxene-free dunites are significantly more abundant at Gakkel Ridge compared to other mid-ocean ridges, which is apparently inconsistent with the high melt production that is required to form dunites. If recent melting and focused melt transport generated the Gakkel Ridge dunites, these melts should have generated a magmatic crust of near-normal thickness, rather than leading to widespread exposures of mantle rocks along the axial valley walls. We present preliminary results of a petrological investigation of dunites and dunite-hosted melt inclusions from the central part of the SMZ. These homogeneous chromian spinels have a very limited compositional range at intermediate Cr-nrs (30-40) and can contain abundant multiphase inclusions, which may offer a clue to the formation of these dunites. Among the included phases detected as yet are olivine, albite, cpx, amphibole, micas (K- and Na-phlogopite, muscovite), magnetite, pentlandite and apatite. The presence of such hydrous phases (which have been observed in oceanic dunites elsewhere) is in apparent contrast to the moderately LREE-depleted interstitial cpx in the dunite matrix, which is equilibrium with a MORB-type liquid. These cpx postdate the formation of the chromite-hosted melt inclusions. Thus, the trace element composition of homogenized melt inclusions should provide us with an earlier stage of focused reactive melt transport. On a local scale, this will help us understanding the genetic relationship between these dunites and LREE-depleted lherzolites, and LREE-enriched harzburgites collected from the same location.
T42B-04 INVITED
The importance of silica activity on melt-rock reaction processes in the oceanic mantle
Plagioclase peridotites are widespread on the ocean floor along slow-spreading ridges and in the mantle section of ophiolites. They are usually characterized by (a) LREE depleted plagioclase, with variable but low Sr concentrations and high anorthite contents (~70-95), (b) LREE depleted and MREE-HREE enriched cpx rims relative to their cores, (c) highly variable spinel compositions on grain and thin section scale, (d) textures indicating cpx+ol dissolution and concomitant opx+plag crystallization. Their origin has usually been attributed to subsolidus breakdown, melt entrapment and/or fractional crystallization from a transient melt. However, existing models cannot account for all observed features combined, and seem to suggest extensive melt-rock reaction at high melt-rock ratios prior to the formation of plagioclase peridotites. Gakkel Ridge plagioclase peridotites originate from the "Sparsely Magmatic Zone", a magma-starved environment where only mantle rocks have been recovered. The exceptional freshness of these samples allowed a very detailed, texturally controlled study of the major and trace element mineral compositions. The studied samples share all the aforementioned characteristics. Furthermore, complex zoning has been preserved in all minerals (probably because of the early onset of rapid cooling at the ultraslow-spreading Gakkel Ridge) that record various stages of reaction progress. We show that all salient features of these plagioclase peridotites are best explained by partial equilibration with infiltrating melt(s) characterized by relatively high aSiO2 (inherited from relatively fertile peridotite) into a more depleted peridotite (or harzburgite) with lower silica activity. Depending on the reactivity of the system, this may continue to complete consumption of the melt (leading to reactive melt stagnation), or until equilibrium is reached. This process can operate at minimal melt-rock ratios and will occur at all pressures. It can be applied to other environments as well, with potentially wider applications to mantle petrology.
T42B-05
Postcumulus Processes in Oceanic-Type Olivine-Rich Cumulates: the Role of Melt Entrapment vs. Melt-Rock Interaction
Evaluation of postcumulus processes in cumulate rocks can provide insights on the mechanisms and scales of melt migration and interaction within the crust, thus contributing to define crustal accretion models. Here we present a microstructural-geochemical study on MORB-type primitive olivine-rich cumulates intruded in the Erro-Tobbio (ET) mantle peridotites (Voltri Massif, Ligurian Alps, Italy), an on-land analogue of (ultra-) slow spreading settings. Postcumulus crystallization is indicated by the occurrence of accessory interstitial minerals (Ti- pargasite, opx , Fe-Ti oxides), and by chemical zoning in intercumulus clinopyroxene related to its textural occurence, i.e. marked REE, Ti, Zr enrichment at almost constant Mg-numbers (88-90) and LREE depletion, from core to rim of coarse anhedral clinopyroxene, to thin vermicular cpx grains. Interstitial pargasites have high Nb, Zr, REE contents although preserving "primitive" major element compositions (Mgvalue = 0.86-0.89) and LREE depletion. Significant trace element enrichment at almost constant LREE fractionation in interstitial clinopyroxenes and Ti-rich pargasites argue against the infiltration of exotic evolved melts and indicate that interstitial minerals were mainly related to close-system trapped melt crystallization. Geochemical modeling shows that crystallization of less than 5% trapped melt is sufficient to produce the REE enrichments observed in cpx. The progressive increase in (REE-Zr-Ti) abundances from core to rim of large clinopyroxene grains is accounted by an "in-situ" fractional crystallization process. Anomalous Zr enrichment is observed in thin (< 400 µm) interstitial and vermicular clinopyroxenes and pargasitic amphiboles, representing the very last melt fraction (reasonably < 2%). At this final crystallization stage, the low residual porosity likely inhibited large-scale melt migration. AFC modeling indicates that the high Zr/Nd ratios in cpx could be related to small-scale migration and interaction between residual, evolved, low melt fractions and the olivine cumulus matrix, consistent with textural evidence of lobate contacts between vermicular clinopyroxenes and cumulus olivine, indicative of partial olivine resorption. Our study points that small amount of melt trapped in an olivine-rich matrix can induce significant trace element enrichment in minerals. Such an evidence argue against large-scale residual melt migration, this latter being controlled by porosity and permeability of the crystal mush, and indicates that compaction was very efficient. This could have been favoured by the interplay of different factors, i.e. slow cooling rates and low thermal gradients, and synkinematic deformation (Natland and Dick, 2001, J. V. G. Res, 110, 191-233). The ET cumulates could thus represent the base of a stack of cumulates (3-5 Kb intrusion depth) from which differentiated melts migrated upwards and crystallized at shallower lithospheric environments. Overall, this indicates that different melt transport and interaction processes may act during crust accretion at very slow spreading settings, e.g. low melt fraction entrapment of indigenous melts in compaction-dominated deep-seated gabbroic intrusions vs. migration of exotic evolved melts through uppermost crustal levels.
T42B-06
Serpentinization Of The Leka Ophiolite Complex, Norway: Geochemical And Physical Implications
The Leka Ophiolite Complex (LOC) contains all the principle components of an ophiolite and is a part of the Upper Allochthon of the Scandinavian Caledonides. The ultramafic lithologies (harzburgites, dunites, wehrlites and orthopyroxenite dykes) of the LOC undergo hydration dependent on the constituent primary mineral assemblages thus preserving the hydration history over a wide range of temperatures, where the orthopyroxenite dykes record high temperature reactions (~650°C) while the dunites and harzburgites undergo serpentinization reactions at lower temperatures (<400°C). The ubiquity of bastites and mesh textures in the lithologies suggests that hydration occurred under static conditions. One of the important implications of hydration of ultramafic rocks is the effect on element mobility during the process. Although concentrations of some major elements (Mg and Si) in the lithologies of the LOC do not change with hydration, there is ample evidence for transfer of Fe, Mn and Ca within the rocks at the grain- scale. Fe and Mn are mobilized during the alteration of orthopyroxene and results in the formation of secondary olivine with high Fe and Mn contents which later forms serpentine and ferroan-brucite with high Fe and Mn contents. However, Mn is lost during the subsequent oxidation of ferroan-brucite to magnetite. Ca is also mobilized during the serpentinization of primary clinopyroxene, some of which forms secondary diopside after olivine. The release and uptake of elements during hydration processes in ultramafic rocks, therefore, plays an important role in modifying fluid properties and may have a significant impact on vent-fluid chemistry. Another point of interest is the density change occurring during serpentinization and its impact on deformation. Observations suggest that hydration of the orthopyroxenite dykes took place at temperatures higher than serpentinization of the surrounding dunites. The serpentinization of dunites results in a volume change of 25% which 'squeezes' the embedded orthopyroxenite dykes resulting in fracturing. The geometrical and statistical characteristics of the 2-D fracture networks in the dykes are typical of patterns generated during hierarchical fracturing where the layer is progressively broken up into smaller domains. The mechanism of fluid-migration, stress-inducing reactions and fracturing is an important process which creates fresh, reactive surfaces and new pathways for fluid infiltration.
T42B-07 INVITED
Interactions Between Serpentinization, Hydrothermal Activity and Microbial Community at the Lost City Hydrothermal Field
Seafloor investigations of slow- and ultraslow-spreading ridges have reported many occurrences of exposed mantle peridotites and gabbroic rocks on the ocean floor. Along the Mid-Atlantic Ridge, these uplifted portions of oceanic crust host high-temperature black smoker-type hydrothermal systems (e.g., Rainbow, Logatchev, Saldanha), and the more distinct low-temperature Lost City Hydrothermal Field (LCHF). Built on a southern terrace of the Atlantis Massif, the LCHF is composed of carbonate-brucite chimneys that vent alkaline and low-temperature (40-90°C) hydrothermal fluids. These fluids are related to serpentinization of mantle peridotites, which together with minor gabbroic intrusions form the basement of the LCHF. Long-lived hydrothermal activity at Lost City led to extensive seawater-rock interaction in the basement rocks, as indicated by seawater-like Sr- and mantle to unradiogenic Nd-isotope compositions of the serpentinites. These high fluid fluxes in the southern part of the massif influenced the conditions of serpentinization and have obliterated the early chemical signatures in the serpentinites, especially those of carbon and sulfur. Compared to reducing conditions commonly formed during the first stages of serpentinization, serpentinization at Lost City is characterized by relatively oxidizing conditions resulting in a predominance of magnetite, the mobilization/dissolution and oxidation of igneous sulfides to secondary pyrite, and the incorporation of seawater sulfate, all leading to high bulk-rock S-isotope compositions. The Lost City hydrothermal fluids contain high concentrations in methane, hydrogen, and low-molecular weight hydrocarbons considered as being produced abiotically. In contrast, organic compounds in the serpentinites are dominated by the occurrences of isoprenoids (pristane, phytane, and squalane), polycyclic compounds (hopanes and steranes), and higher abundances of C16 to C20 n-alkanes indicative of a marine organic input. We propose that the high seawater fluxes in the basement rocks favour the transport of marine organic carbon in the serpentinites and overprint any earlier abiotic signature. Serpentinites forming the basement of active hydrothermal chimneys have carbon and sulfur signatures, i.e. negative S-isotope compositions and high content of squalane biomarker, reflecting the influence of microbial activity in the subseafloor. Our geochemical study of the basement rocks, based on multiple isotopic systems, reveals the close relationships and reciprocal interactions between serpentinization, hydrothermal activity, and microbial community at Lost City. In addition, it sheds new light on the consequences of long-lived peridotite-hosted hydrothermal system on the chemical compositions of the oceanic lithosphere and global geochemical cycles.
T42B-08
Carbon and Sulphur Geochemistry of Rift Valley Sediments and Hydrothermal Fluids at the Ultra-Slow Spreading Southern Knipovich Ridge
The rift valley of the ultra-slow spreading southern Knipovich Ridge in the Norwegian-Greenland Sea (73°N) is partly buried by a thick sediment cover, as at Middle Valley, Escanaba Trough and Guaymas Basin. These glacial and post-glacial sediments (12000-20000 years) derived from the nearby Bear Island fan likely act as a thermal and hydrogeological boundary to heat and fluid flow and influence hydrothermal fluid compositions. Geochemical studies of the rift valley sediments and the hydrothermal vent fluids of the recently discovered black smoker vent field Loki's Castle provide insights into the influence of the sediment cover on the composition of the hydrothermal fluids at the southern Knipovich Ridge. Here we present an overview of preliminary data on the carbon and sulphur geochemistry of the sedimentary and hydrothermal components at Loki's Castle and compare these with other sedimented and un-sedimented mid-ocean ridge hydrothermal systems. The hydrothermal vent fluids have a pH of 5.5 and are characterized by elevated concentrations of hydrogen, methane, hydrogen sulphide and ammonia, which reflect a strong sedimentary input. Short gravity cores of the rift valley sediments show relatively constant total carbon contents of approximately 1 wt%, but locally reach up to 4 wt%. Varying carbon isotope compositions reflect a mix of marine carbonates with organic carbon. Extracted sediment pore fluids show an increase in alkalinity and dissolved inorganic carbon (DIC) concentrations with increasing burial depth. The corresponding δ13CDIC values show a clear depletion with increasing alkalinity and DIC concentrations. The vent fluid compositions and carbon and sulphur isotope geochemistry provide constraints on redox conditions and thermocatalysis of organic carbon during fluid-sediment interaction, and are distinct from un- sedimented mid-ocean ridge hydrothermal fluids.