Serpentinization at mid-ocean ridges occurs in slow spreading rate contexts and affects mantle-derived peridotites next to large-offset axial normal faults (detachment faults). Microstructural, mineralogical and oxygene isotope studies of serpentinized samples from the footwall of these detachments indicate that their serpentinization involved an initial stage of formation of the serpentine mesh texture, followed by several stages of veining and recrystallization of the initial serpentine, and occurred at relatively high temperatures (200-350°C), within the range that is most kinetically favorable for olivine serpentinization. Microcracks that are used as pathways for hydrothermal fluids at the mesh texture formation stage are tightly-spaced (< a few hundred microns) and kinetic constrains indicate that this stage probably occurs at low fluid-rock ratio within a few years of hydrothermal fluids gaining access to the peridotite. Mesh texture serpentinization at mid-ocean ridges is therefore expected to be fast relative to geological displacement rates. This leads to a conceptual model of mesh texture formation in the axial lithosphere at depths corresponding to the 200-350°C range of temperature, in and next to axial detachment fault zones and in association with deep and weak hydrothermal convection. Seismic constraints suggest that this may occur at depths down to 15 km below seafloor at very slow spreading rates, and up to 4-5 km into the detachment footwall. Episodes of extensive serpentinization involving black-smoker type fluids also occur locally in more superficial tectonically damaged domains that channel the deep parts and the upwelling limbs of vigourous hydrothermal cells cooling magmatic intrusions. These two modes of serpentinization are expected to combine in detachment-dominated ridge regions to create an upper lithosphere layer made of partially serpentinized peridotites and isolated gabbroic bodies, with crustal-type seismic velocities and densities.

Serpentinization of forearc mantle in nonaccretionary convergent plate margins provides a mechanism for cycling of oceanic plate constituents. Fluid seeps and serpentinite mud volcanism in such environments opens a window into subduction channel processes at depths far too deep to be accessed by any known technology. Samples from the subducted oceanic plate, from guyots subsiding with it, and from forearc lithosphere have all been recovered in cores, dredges, and submersible/ROV dives on Mariana forearc serpentinite mud volcanoes and fluid seeps. Recently IODP drilling on Expedition 366 recovered cores from the summits and flanks of three mud volcanoes that extended sampling of subduction channel materials from ~13 to 19 km depth to slab, from ~55 to 86 km west of the Mariana Trench, from ~80°C to 350°C, and from an along-strike distance of over 650 km. These samples show a pattern of compositional variation consistent with models of subducting slab dehydration. Extensional tectonic deformation in the Mariana forearc, caused by rollback of the Pacific plate, creates fault-controlled conduits for escape of slab-derived fluids. The seeps offer a foothold for microbial and macrofauna communities and the mudflows record a history of eruption as old as 50 Ma. Similar serpentinite deposits worldwide, as old as 3.5 Ga, suggest an even greater history of convergent margin cycling. The details of dynamics of such processes are still poorly known. IODP Expedition 366 emplaced cased boreholes at the active summits of three serpentinite mud volcanoes to be instrumented and thus provide future long-term monitoring.

Approximately 15,000 km2 of partially serpentinised peridotite of the Samail ophiolite is actively undergoing serpentinization at temperatures <60°C. This is evidenced by the presence of highly reducing, pH>11, H2, CH4 and Ca-OH-bearing fluids in springs and in water wells, and calcite travertine terraces on the surface. The progress of serpentinization depends on the ingress capability of water into the rock formation, which is a function of permeability, porosity, and connectivity of the rocks as well as the fluid pathways and residence time of fluids in the rocks. There is little information about these parameters in the mantle peridotite aquifers of the Samail ophiolite. Previous studies developed a conceptual hydrogeological and reaction-path model of the peridotite [1,2], describing an intensely fissured, permeable zone of about 50 m below surface and a deeper, less permeable zone, dominated by rare discrete fractures. The upper zone is characterized by a Mg-HCO3-rich, pH 8 to 9 groundwater and the deeper zone by a Ca-OH-rich, pH 11-12 fluids. Although this general hydrogeological model is widely accepted, it is not quantitative and has not been tested by direct observations. We are investigating the different hydrogeological regimes and the coupling between these regimes and water-rock reactions in order to quantify the permeability, porosity and connectivity of fluid pathways as well as the residence times of fluids in the peridotite aquifers. Borehole pumping and packer testing, combined with wireline logging and detailed geochemical analysis of fluids and dissolved gases reveal a general stratification of the aquifers with regard to permeability, fluid types and fluid residence times as a function of depth. Well-defined transitions zones, with steep geochemical gradients were detected that coincide with changes in permeability. More important, these transition zones may be key locations for thriving subsurface microbial communities, deriving energy from ongoing water-rock reactions.

Fischer-Tropsch type (FTT) processes are commonly invoked to explain abiotic CH₄ formation in serpentinization systems; however, established models fail to explain the occurrence of radiocarbon-free CH₄ in seafloor hydrothermal systems and ophiolitic gas seeps, which suggests that CH₄ formation is decoupled from dissolved inorganic carbon in circulating fluids^1-3. This study examines serpentinization processes and the formation of CH₄ in olivine-hosted secondary fluid inclusions from different geologic settings. Raman spectroscopy of inclusion contents reveals chrysotile, brucite, and magnetite as the dominant alteration mineral assemblage and CH₄ or H₂ (or both) as the dominant volatiles in most samples. The formation of CH₄ is envisioned as follows: 1) Cooling of igneous rocks to below the ductile-brittle transition temperature allows fracturing and creates pathways for hydrothermal and magmatic fluids. 2) At temperatures > ~ 400°C olivine is stable in the presence of H₂O, which can cause healing of fractures and entrapment of fluids. 3) Cooling to lower greenschist facies conditions destabilizes olivine, which reacts with trapped H₂O to form daughter minerals and H₂ ⁴. 4) During this process, aH₂O and f_O2 decrease within the inclusion, creating conditions conducive to CH₄ formation. The olivine host appears to be sufficiently impermeable with respect to CH₄ and H₂ at low temperatures to store these volatiles over geologic timescales until they are released by fracturing or dissolution of their host. The range of δ¹³C of trapped CH₄ overlaps with the measured range of δ¹³C_CH4 in submarine vent fluids and continental gas seeps in both mafic and ultramafic substrates5. The concept of CH₄ synthesis in fluid inclusions, storage  over geological timescales, and subsequent mining of trapped CH₄ provides a plausible mechanism for the occurrence of radiocarbon-dead abiotic CH₄ in a wide range of geological settings, including mafic and ultramafic seafloor hydrothermal systems and ophiolitic gas seeps.

Molecular hydrogen (H2) and magnetite are two of the most significant products of serpentinisation, both of which result from oxidation of iron as the reaction proceeds. Natural serpentinites exhibit substantial variability in the extent of iron oxidation and magnetite production, implying considerable variation in hydrogen production as well. However, the underlying reason for this variability remains highly uncertain. I will present results of a series of laboratory experiments designed to investigate how different environmental parameters impact the extent of iron oxidation during serpentinisation and the reaction pathways involved. Rates of hydrogen generation drop off significantly with decreasing temperature, both because the overall reaction slows and because iron increasingly partitions into brucite rather than magnetite, which does not involve oxidation. Conversely, higher pH increases rates of reaction and hydrogen production in most cases. Conditions that result in a greater contribution from orthopyroxene relative to olivine can limit magnetite production as iron gets preferentially partitioned into serpentine minerals, but this does not necessarily decrease hydrogen generation since ferric iron can be incorporated into serpentines. While chemical thermodynamics appear to be a major contributor to controlling the fate of iron during serpentinisation, many reactions in experimental systems remain out of thermodynamic equilibrium indicating kinetic factors must also contribute.

Iron is a key component of the dynamic chemistry of serpentinizing systems. Oxidation of Fe(II) in primary minerals and accommodation of Fe(III) in secondary phases drives production of energy-rich H2 gas and the creation of reducing conditions. The effect of reaction conditions (e.g. protolith composition, temperature, fluid composition/flux) on the transformation and accommodation o Fe in the alteration mineral assemblage is not yet clearly understood. This is despite the influence that Fe chemistry has on many geological and (astro)biological processes including: the potential for and activity of microbial life in the subsurface ocean crust and hydrothermal vents, the potential for sequestration of carbon in ultramafic rocks, the origin of life, and the interpretation of the significance of mineral signatures detected on other planets. Integrated Raman hyperspectral, quantitative elemental, and x-ray absorption (sensitive to Fe oxidation state) spectral images and point analyses illustrate that serpentinized rocks obtained from IODP Expedition 357 to the Atlantis Massif have undergone multiple episodes of water/rock interaction; producing a variety of alteration textures and phases including Fe-rich and Fe-poor varieties of serpentine that have distinct distributions. The difference in Fe content may be indicative of different reaction conditions associated with each episode of water/rock interaction. The ferric component of all serpentines ranges from Fe3+/FeTotal ~40-80%, suggesting H2 could have been produced during serpentine formation. However, the difference in reaction conditions could affect the potential development of a subsurface, in-situ microbial biosphere supported by the serpentinization process.

Mantle peridotite is far from equilibrium in near surface conditions leading to rapid, extensive serpentinization, carbonation and oxidation, as well as other geochemical changes. In the Samail ophiolite in the Sultanate of Oman this geochemical changes have likely occurred through the entire history of the ophiolite, from formation near the axis of a spreading ridge to present day low temperature serpentinization/carbonation. Alteration products occur across the entire mantle section of the ophiolite, primarily as serpentine minerals, brucite and carbonates varying in composition from magnesium rich to calcium rich as water flows through the peridotites. The nature of these alteration products suggests that alteration is not entirely isochemical and provides evidence of magnesium mobility during carbonation and serpentinization. Each of the alteration products exhibits different fluid-mineral Mg isotopic fractionation deviating significantly from an initially homogenous mantle Mg isotopic composition (26Mg = −0.25‰ DSM3). Partially serpentinized harzburgites and dunites have mantle-like 26Mg (-0.23‰). Weathered serpentinized peridotites with magnesium depletion have heavy 26Mg (0.94‰), among the heaviest 26Mg values that have been reported in altered peridotite. Massive magnesite veins with very light compositions (-3.3‰) distributed within peridotite massifs are potential sinks for light magnesium isotopes removed from altered peridotites. The fractionation of Mg isotopes observed in the mantle section of the ophiolite spans more than 50% of the known terrestrial fractionation.