Final Report Summary - HISLA-DR (Hydrogen Incorporation in Subducting Lithosphere after Dehydration Reactions)
Introduction
Subduction zones (convergent places where lithosphere is underthrusted, sinking into the Earth's mantle) are key geological settings for the Earth dynamics and the recycling of volatiles is perhaps their most distinctive feature. Water is incorporated into hydrous minerals (such as serpentine and amphiboles) during seafloor alteration of the oceanic lithosphere. During subduction of the partially hydrated lithosphere, dehydration of these hydrous minerals produces a fluid phase. A part of this fluid phase will be recycled back to the Earth’s surface through hydrothermal aqueous fluids or through hydrous arc magmas, whereas another part of the water will be transported to the deep mantle by Nominally Anhydrous Minerals (NAMs) such as olivine, pyroxene and garnet. The partitioning of water between these two processes is crucial for the understanding the deep water cycle and has profound implications for the rheology of the upper and lower mantle. The HISLa-DR project has addressed the quantification of the water in NAMs as the result of dehydration reactions in the subducting lithosphere and mantle wedge.
Work carried out
The experimental part of the HISLa-DR project was conducted during the outgoing phase in the Research School of Earth Sciences (ANU, Australia). During the early stage of the project it was realized that in order to retrieve reliable hydrogen content data from natural samples and to design and optimize the high-pressure experiments it was essential to acquire new data on hydrogen diffusion in NAMs in particular olivine. We conduct a series of dehydroxilation experiments on forsterite crystals synthetized with well-constrained hydrous defect chemistry and subsequently dehydroxylated at different temperatures ranging from 800 to 1200ºC at atmospheric pressure.
In a second step, the hydrogen content of olivine, pyroxene and garnets in rocks from high-pressure (HP) metaperidorite (hydrous ultramafic rocks from the Betic Cordillera and Eastern Alps that experienced dehydration reactions) were systematically measured by means of infrared spectroscopy, constituting the most complete dataset on water-content in natural peridotites. To further constrain maximum water capacity of NAMs in subduction settings, a series of piston-cylinder experiments were performed using a broader PT window (corresponding to depths from 15 to 200 km, and temperatures from 700 to 1050°C). A novel capsule configuration was designed to produce dehydration reactions and concomitant hydroxylation of NAMs (using serpentine layers as fluid source and mineral separates as sensor layers). Double polished thick sections of the capsules were prepared and hydrogen contents in NAMs were measured using infrared spectroscopy.
Another issue investigated in the HISLa-DR project was the development of a crystallographic preferred orientation in prograde minerals after dehydration reactions. We performed experiments using cores of natural serpentinites and we compared the results with metaperidotites recording the same reaction in nature. Textural data was obtained during the return phase of the project at Géosciences Montpellier (France).
Main results
Early experiments showed that hydrogen is the fastest species able to diffuse through the olivine lattice. We have found, however, experimental and natural evidence suggesting that hydrogen diffusion can also be orders of magnitude slower. We link these disparate diffusion coefficients to different defects where hydrogen is bounded in the structure. Importantly the hydrogen associated to the most common defects expected for olivine in the upper mantle diffuses 1.5 up to 3 orders of magnitude slower than previously assumed.
Metamorphic olivines formed after dehydration reactions from the Alps preserve their original water contents despite long times of exhumation (2-3 Ma). Only slow hydrogen diffusion coefficients are able to explain the observed preservation of water content and its correlation with pressure and temperature. Olivine water content in these metaperidotite varies from <10 wt ppm (parts per million in weight) at low temperature (400-450 ºC) up to 30-40 wt ppm after chlorite breakdown (750-800ºC and 3 GPa). These values are in excellent agreement with the piston cylinder experiments at the same conditions. These experiments also reveal a systematic increase in water storage capacity with pressure and temperature for olivine and mostly with temperature for orthopyroxene and clinopyroxene. These results are significant as it clearly show that the partition coefficient between olivine and pyroxene are not constant over a broad PT range. Water capacity is maximum in the subsolidus mantle wedge (e.g. 490 wt ppm at 6 GPa, 1000ºC) whereas extrapolated values corresponding to the subducting slab are lower (200 wt ppm at 7 GPa and 600ºC).
Textural analyses of statically dehydrated serpentinites reveal that the CPO of the prograde minerals is partially inherited from the protolith. Moreover we show that the fabric strength varies from weak to strong depending on the degree of compaction and the mechanisms of fluid drainage.
Potential impact
A slow hydrogen diffusion in olivine implies that the water content in mantle xenoliths closely resemble original water content in the source. On the other hand the combination of natural observations and experiments enable to obtain for the first time a quantitative view of maximum water capacity in NAMs in subduction zones suggesting the main role of the mantle wedge as a water reservoir. In addition obtained solubility laws can be use to assess the extent of subsaturation in the lithosphere from other settings where fluid saturation conditions do not prevail.
The results from the HISLa-DR project provide fundamental information needed to better understand the effect that fluids have a on the physico-chemical properties of Earth’s materials, such as rheology, melting temperature, reaction kinetics, diffusion length scales, electrical and elastic properties. Data obtained in this project will be of great interest to other applied branches in Earth Sciences working in subduction zones. These setting are the most geologically active places on Earth and are associated with hazards with devastating social consequences such as high-magnitude Earthquakes and volcanoes. Eventually all these hazards have in common the effect that fluids have on the physical properties on Earth's materials.
Subduction zones (convergent places where lithosphere is underthrusted, sinking into the Earth's mantle) are key geological settings for the Earth dynamics and the recycling of volatiles is perhaps their most distinctive feature. Water is incorporated into hydrous minerals (such as serpentine and amphiboles) during seafloor alteration of the oceanic lithosphere. During subduction of the partially hydrated lithosphere, dehydration of these hydrous minerals produces a fluid phase. A part of this fluid phase will be recycled back to the Earth’s surface through hydrothermal aqueous fluids or through hydrous arc magmas, whereas another part of the water will be transported to the deep mantle by Nominally Anhydrous Minerals (NAMs) such as olivine, pyroxene and garnet. The partitioning of water between these two processes is crucial for the understanding the deep water cycle and has profound implications for the rheology of the upper and lower mantle. The HISLa-DR project has addressed the quantification of the water in NAMs as the result of dehydration reactions in the subducting lithosphere and mantle wedge.
Work carried out
The experimental part of the HISLa-DR project was conducted during the outgoing phase in the Research School of Earth Sciences (ANU, Australia). During the early stage of the project it was realized that in order to retrieve reliable hydrogen content data from natural samples and to design and optimize the high-pressure experiments it was essential to acquire new data on hydrogen diffusion in NAMs in particular olivine. We conduct a series of dehydroxilation experiments on forsterite crystals synthetized with well-constrained hydrous defect chemistry and subsequently dehydroxylated at different temperatures ranging from 800 to 1200ºC at atmospheric pressure.
In a second step, the hydrogen content of olivine, pyroxene and garnets in rocks from high-pressure (HP) metaperidorite (hydrous ultramafic rocks from the Betic Cordillera and Eastern Alps that experienced dehydration reactions) were systematically measured by means of infrared spectroscopy, constituting the most complete dataset on water-content in natural peridotites. To further constrain maximum water capacity of NAMs in subduction settings, a series of piston-cylinder experiments were performed using a broader PT window (corresponding to depths from 15 to 200 km, and temperatures from 700 to 1050°C). A novel capsule configuration was designed to produce dehydration reactions and concomitant hydroxylation of NAMs (using serpentine layers as fluid source and mineral separates as sensor layers). Double polished thick sections of the capsules were prepared and hydrogen contents in NAMs were measured using infrared spectroscopy.
Another issue investigated in the HISLa-DR project was the development of a crystallographic preferred orientation in prograde minerals after dehydration reactions. We performed experiments using cores of natural serpentinites and we compared the results with metaperidotites recording the same reaction in nature. Textural data was obtained during the return phase of the project at Géosciences Montpellier (France).
Main results
Early experiments showed that hydrogen is the fastest species able to diffuse through the olivine lattice. We have found, however, experimental and natural evidence suggesting that hydrogen diffusion can also be orders of magnitude slower. We link these disparate diffusion coefficients to different defects where hydrogen is bounded in the structure. Importantly the hydrogen associated to the most common defects expected for olivine in the upper mantle diffuses 1.5 up to 3 orders of magnitude slower than previously assumed.
Metamorphic olivines formed after dehydration reactions from the Alps preserve their original water contents despite long times of exhumation (2-3 Ma). Only slow hydrogen diffusion coefficients are able to explain the observed preservation of water content and its correlation with pressure and temperature. Olivine water content in these metaperidotite varies from <10 wt ppm (parts per million in weight) at low temperature (400-450 ºC) up to 30-40 wt ppm after chlorite breakdown (750-800ºC and 3 GPa). These values are in excellent agreement with the piston cylinder experiments at the same conditions. These experiments also reveal a systematic increase in water storage capacity with pressure and temperature for olivine and mostly with temperature for orthopyroxene and clinopyroxene. These results are significant as it clearly show that the partition coefficient between olivine and pyroxene are not constant over a broad PT range. Water capacity is maximum in the subsolidus mantle wedge (e.g. 490 wt ppm at 6 GPa, 1000ºC) whereas extrapolated values corresponding to the subducting slab are lower (200 wt ppm at 7 GPa and 600ºC).
Textural analyses of statically dehydrated serpentinites reveal that the CPO of the prograde minerals is partially inherited from the protolith. Moreover we show that the fabric strength varies from weak to strong depending on the degree of compaction and the mechanisms of fluid drainage.
Potential impact
A slow hydrogen diffusion in olivine implies that the water content in mantle xenoliths closely resemble original water content in the source. On the other hand the combination of natural observations and experiments enable to obtain for the first time a quantitative view of maximum water capacity in NAMs in subduction zones suggesting the main role of the mantle wedge as a water reservoir. In addition obtained solubility laws can be use to assess the extent of subsaturation in the lithosphere from other settings where fluid saturation conditions do not prevail.
The results from the HISLa-DR project provide fundamental information needed to better understand the effect that fluids have a on the physico-chemical properties of Earth’s materials, such as rheology, melting temperature, reaction kinetics, diffusion length scales, electrical and elastic properties. Data obtained in this project will be of great interest to other applied branches in Earth Sciences working in subduction zones. These setting are the most geologically active places on Earth and are associated with hazards with devastating social consequences such as high-magnitude Earthquakes and volcanoes. Eventually all these hazards have in common the effect that fluids have on the physical properties on Earth's materials.