Subducted oceanic crust plays an important role in controlling the chemical budget of the crust and mantle and in the composition of arc lavas. Oceanic eclogites represent fragments of oceanic crust that have been through a subduction zone and were subsequently exhumed and exposed. A range of geochemical signatures in oceanic eclogites have been studied to unravel the fluid-rock interaction processes. A key tracer is the trace-element boron and its isotope ratio (¹¹B/¹⁰B) in oceanic eclogites with its great potential for quantifying mass transfer processes at convergent margins. Current models predict a strong decrease of δ¹¹B values in the subducting crust with progressive dehydration to values much below that of the depleted mantle (appr. -7 ‰).

We have analyzed elemental abundances and boron isotopic compositions of oceanic metamorphic rocks, from Zambezi Belt, Cabo Ortegal complex, Raspas Complex, Syros Island, and Tian Shan. Whole-rock B/Pr ratios were used to quantify the progress of dehydration. The boron isotopic composition of almost all samples (approximately -10 to +5 ‰) ranges from δ¹¹B values close to that of fresh MORB to that of typical altered oceanic crust. Also, the sample set shows no correlation between δ¹¹B values and B/Pr, as would be expected from current theoretical models. Our results, thus, demonstrate that B isotopic fractionation in subducted oceanic crust is much smaller than predicted. We suggest that this discrepancy can be resolved by accepting high pH values in high-pressure hydrous fluids, which show a much smaller boron isotope fractionation in equilibrium with B-bearing silicates.

Dehydration of oceanic lithosphere during subduction is a key process in the Earth's deep volatile cycle. Field observations and studies of obducted meta-serpentinites that underwent dehydration at depth often show an interconnected, channelized vein network that formed during dehydration and served as pathway for fluid release from the dehydrating rock. Previous studies show that chemical heterogeneities in the bulk-rock composition lead to channelization of fluid flow from the onset of the dehydration process and how subsequent reactive fluid flow in the porous network causes further dehydration and channelization. On larger scales, fluid release from the rock is governed by mechanical processes such as porosity waves. While the micro- and meso-scale have been studied so far, this large-scale fluid release mechanism has not yet been explored for dehydrating serpentinite.

Here, we present a model for reactive porosity waves that investigates the large-scale fluid release from a dehydrating serpentinite. The model combines viscous rheology with the transport of dissolved silica in the fluid which has been shown to be a key metasomatic agent in the dehydration process. As input for our model, we use a multi-scale dataset of fully hydrated serpentinite from an ophiolite taken as representative for serpentinized oceanic lithosphere entering a subduction zone. We use the data to explore the formation of the vein network during dehydration and the behavior during large-scale fluid escape by reactive porosity waves.

Rutile is potentially very important for the transport of water during subduction metamorphism after the breakdown of hydrous phases, as it is one of the most hydrous nominally anhydrous minerals (up to several 1000 μg/g H₂O) and commonly occurs in a variety of lithologies and P-T conditions across subduction zones.

We present results from quantitative in-situ Fourier Transform Infrared (FTIR) spectroscopy of rutile from different subduction settings (P-T, lithology, geothermal gradients) and high-resolution FTIR mapping to evaluate the variability of H₂O contents in rutile and retention of H⁺.

Observed H₂O contents are highly variable. Granulite facies rutile have low H₂O contents (<150 μg/g) with high-P granulites showing up to ~400 μg/g H₂O. The highest average H₂O contents were observed in low-T eclogite facies rutile (500–1700 μg/g). Amphibolite- and high-T – high-P eclogite facies rutile has intermediate H₂O contents (~200–400 μg/g). Rutile from UHP shows greatly varying H₂O contents (<10–700 μg/g).

FTIR maps of high-P granulite-, high-T – high-P eclogite facies and UHP rutile show evidence for diffusive H⁺ loss, while low-T eclogite- and amphibolite facies rutile are homogeneous or show growth zoning and thus retain their original H₂O contents. Therefore, the typically variable and lower H₂O contents at higher P-T conditions result from H⁺ loss at temperatures above ~650–700 °C.

Generally, H₂O contents are distinctive for specific subduction zone conditions, especially when coupled with Zr-in-rutile thermometry and trace-element geochemistry, e.g. high H₂O contents above 500 μg/g coupled with Zr contents below 200 μg/g indicate cold subduction geotherms.

The recycling of volatiles like H₂O from the surface into the interior is critical as it influences the physical and chemical properties of the mantle. One of the essential effects H₂O has on the mantle is the reduction of the solidus temperature, as this can trigger partial melting. On the one hand, the process of partial melting is vital for arc and ore deposit formation. On the other hand, it is thought to be a requirement for the generation of the continental crust.

Even though intensive work has been done on investigating the stability fields of hydrous phases experimentally and thermodynamically depending on pressure, temperature, and composition, it is still debated to which depths and in which quantities H₂O can be recycled. However, the parameters influencing the mineral stability and break-down (e.g., pressure, temperature, composition, oxygen fugacity) change constantly during the recycling process. To include these changes, global convection models are required which take the varying conditions over time into account. While numerous studies exist, that model the subduction process itself, the recycling of H₂O is not considered to date in most of the simulations, or strongly simplified models are applied.

We model the quantity of H₂O that can be recycled as well as the depth to which the recycling is possible. For this purpose, we test individual parameters affecting the H₂O recycling and dehydration from the subducting slab. To benchmark our simulations, we compare our results with seismological observations of the Pacific plate subducting beneath northeast Japan.