Jürgen Gose

Water in enstatite from Mid-Atlantic Ridge peridotite: Evidence for the water content of suboceanic mantle?

We present the first water concentration data for oceanic mantle peridotite. The rocks come from the Mid-Atlantic Ridge and were sampled by drilling during Leg 153 of the Ocean Drilling Program. Earlier studies of peridotite xenoliths suggest that orthopyroxene is the most important host of water in the upper mantle. However, an unresolved question is whether the measured water content in orthopyroxene is equal to its original concentration in the mantle or possibly lower because of decompression-related water loss. Our investigations of oceanic mantle rocks yield water contents in orthopyroxene in the range between 160 and 270 ppm by weight (wt). These new data, compared with those from xenoliths, experimental studies, and geophysical investigations, imply that the measured values reflect the original mantle contents. The preservation of mantle water contents is ascribed to an exhumation mechanism that provides for cooling during ascent of the suboceanic mantle slices.

Olivine from spinel peridotite xenoliths: Hydroxyl incorporation and mineral composition

Abstract Traces of water in mantle minerals strongly influence mantle melting and viscosity that, in turn, governs large-scale processes like mantle convection, plate tectonics, and the stabilization of cratons. One way of estimating the mantle’s water content is by analyzing mantle xenoliths brought to the Earth’s surface. A major problem in estimating the mantle’s water budget from xenoliths arises from decompression-induced water loss during uplift. Mantle-derived xenoliths from numerous occurrences worldwide have been investigated with respect to water. However, little is known about water in the mantle beneath most parts of Europe and Asia. This study presents water contents for mantle olivine from Germany, Austria, Mongolia, and Nigeria and suggests a possibility to assess water loss. It also addresses the question whether or not water contents are related to olivine composition and/or the presence of coexisting amphibole. The highest water concentrations are present in olivine from the Eifel, Germany (up to 21 ppm H2O), whereas the Fichtelgebirge xenoliths, Germany, reveal the lowest contents (<1 ppm). All water-bearing olivines show three dominant infrared absorption bands: bands at 3572, 3525, and 3560 cm-1 and weaker bands at 3485, 3542, and 3597 cm-1. The peaks at 3572 and 3525 cm-1 are ascribed to Ti-related substitutions of H. Additional peaks, related to H substitutions involving trivalent cations, occur in the 3300-3400 cm-1 range. However, their intensity does not correlate with the content of trivalent cations. The olivines show a pronounced correlation of Al + Cr and Ca implying that incorporation of Al and Cr is governed by pressure and temperature and primarily attributed to Tschermak’s substitution. This study confirms that a coupled substitution involving Ti is the most important mode of water storage in shallow upper mantle olivine. The Ti content and the fraction of water bound to “Ti defects” are related and support the substitution model Ti4+ + 2H+ = Mg2+ + Si4+. Hence, the Ti content is a useful proxy to estimate the maximum amount of water incorporated by this substitution-providing a tool to approximate the degree of water loss. Five of the investigated olivines provide evidence for no or very little water loss. Their water content of 16-21 ppm is presumably typical for the depleted uppermost mantle. Twelve samples with <1-15 ppm may have lost between 36 and >80% of their original water. Olivine from amphibole-bearing spinel peridotite has relatively low water contents as well as low amounts of Ca, Al, Cr, Ti, and V. Particularly low Ca, Al, and Cr contents suggest fluid infiltration, amphibole formation, and re-equilibration of the whole assemblage at comparably low pressure and/or temperature and may explain the low water content of olivine. Infrared spectra with dominant peaks in the 3200-3300 cm-1 range (spectrum type E) are confined to olivine from amphibole-bearing peridotite.

Water in natural olivine—determined by proton-proton scattering analysis

Abstract Here we present water concentration data for olivine from different host rocks, measured with a nuclear technique using proton-proton scattering. This method, which is used here for the first time on olivine, is very powerful for determining trace amounts of water. The studied olivine specimens differ in their H2O contents, ranging from 4 to 51 wt ppm (=10-117 atom ppm H). The lowest concentrations are found in olivine from spinel peridotite xenoliths, the highest concentrations in olivine from alpine-type peridotite; the contents of an ophiolitic and a hydrothermal olivine are intermediate. Infrared spectroscopy was applied to ensure that the measured water contents stem solely from hydroxyl defects in the mineral structure. The infrared spectra differ from sample to sample. Five of six olivine specimens show absorption bands typical of hydroxyl groups associated with Ti defects. These olivines differ in their Ti contents by two orders of magnitude. However, a correlation of water and Ti content was not observed.

Water in mantle orthopyroxene - no visible change in defect water during serpentinization

Earlier studies on xenolithic mantle rocks imply that the water contents originally present in the minerals at mantle depth may survive despite uplift and decompression. This study addresses the question if mantle minerals in slowly exhumed peridotites that are partially replaced by secondary hydrous phases still retain the original defect water contents that were present at mantle depths prior to alteration. Orthopyroxene from oceanic spinel peridotite was chosen for this investigation because (i) it has variable defect water contents and (ii) it is partially serpentinized. Infrared spectroscopy reveals that defect water is homogeneously distributed in coarse orthopyroxene grains. In contrast, the amount of secondary serpentine is highly variable. The intensity of defect water peaks stemming from hydroxyl defects in the orthopyroxene structure is constant over a crystal and shows no decrease or increase in areas where serpentine is present. This suggests that secondary hydration of mantle minerals ( i.e. partial replacement by hydrous phases) at low temperature either did not influence the defect water contents or, alternatively, completely re-homogenized water in the crystals. However, hydrogen diffusion in orthopyroxene at serpentinization temperatures of ≤300 °C, corresponding to lizardite serpentinization, is too slow to reset the water contents in mm-sized crystals. Therefore, the measured values of defect water in the tectonically exhumed oceanic peridotites may be used to infer the water contents of the upper (sub-oceanic) mantle. If serpentinization temperature exceeds 400 °C (= antigorite serpentinization), as in many Alpine-type peridotites, water in mantle minerals may be modified to a greater extent by secondary hydrothermal processes.

Water transport by subduction: Clues from garnet of Erzgebirge UHP eclogite

Abstract A key question concerning the water budget of Earth’s mantle is how much water is actually recycled into the mantle by the subduction of eclogitized oceanic crust. Hydrous phases are stable only in quartz eclogite not coesite eclogite so that water transport to greater depths is mainly governed by structural water in omphacite and garnet. Here we explore if garnet can be used as a proxy to assess the amount of this water. Available data on the water contents of garnet in coesite eclogite vary over orders of magnitude, from a few up to ca. 2000 ppm. By implication, the maximum bulk-rock water contents are unrealistically high (wt% level). New data from the Erzgebirge indicate moderate amounts of structural H2O stored in garnet (43–84 ppm), omphacite (400–820 ppm), and in the bulk coesite eclogite (ca. 280–460 ppm). Higher garnet water contents occur, but these are not primary features. They are related to molecular water in fluid inclusions that can be attributed to eclogite-facies fluid influx postdating the metamorphic peak. Fluid influx also caused the uptake of additional structural water in garnet domains close to fluid inclusions. Such secondary H2O incorporation is only possible in the case of primary water-deficiency indicating that garnet hosted less water than it was able to store. This is insofar astonishing as comparably high H2O amounts are liberated by the breakdown of prograde eclogite-facies hydrous minerals as a result of ultrahigh-pressure (UHP) metamorphism. Judging from Erzgebirge quartz eclogite, dehydration of 5–10% hydrous minerals (±equal portions of zoisite+calcic amphibole) produces 1500–3000 ppm water. We infer that the largest part of the liberated water escaped, probably due to kinetic reasons, and hydrated exhuming UHP slices in the hanging-wall. Depending on the physical conditions, water influx in eclogite during exhumation (1) produces fluid inclusions and simultaneously enhances the structural water content of nominally anhydrous minerals—as in the Erzgebirge—and/or (2) it may give rise to retrograde hydrous minerals. We conclude that eclogite transports moderate quantities of water (several hundred parts per million) to mantle depths beyond 100 km, an amount equivalent to that in ca. 1% calcic amphibole.

Water Incorporation in Garnet: Coesite versus Quartz Eclogite from Erzgebirge and Fichtelgebirge

The water content of garnet was determined for eclogite from two Variscan complexes in Germany: the Erzgebirge (EG), Saxony, and the Fichtelgebirge (FG), Bavaria. Erzgebirge eclogites occur in three units, each of which experienced specific peak conditions (unit 1: 840–920 C/ 30 kbar, unit 2: 670–730 C/24–26 kbar, unit 3: 600–650 C/20–22 kbar). Peak conditions of the FG eclogite (690–750 C/25–28 kbar) are close to those of eclogite from EG unit 2. Coesite eclogite is restricted to the EG ultra-high pressure (UHP) unit 1. Garnet shows infrared absorption bands at ca. 3650, 3580–3630, and 3570 cm, ascribed to structural water. Many garnets also contain molecular water (in sub-microscopic fluid inclusions), which is irregularly distributed on the grain scale and of secondary origin. Grain volumes with molecular water invariably reveal a band at 3580–3630 cm attributed to a hydrogarnet substitution. Because structural water due to this substitution positively correlates with molecular water, the primary content of structural water can only be deduced from grain volumes that are free of molecular water as demonstrated by Schmädicke & Gose (2017; American Mineralogist 102, 975–986). This primary content is typically low in garnet from quartz eclogite (<2–50 ppm); averages for most samples fall in the range of 8–28 ppm. Garnet from coesite eclogite hosts more water (50–180 ppm) except for garnet from an unusual, phlogopite-bearing coesite eclogite that contains only 19–55 ppm. Structural water in garnet is unrelated to metamorphic peak pressure but governed by the presence (or absence) of eclogite-facies hydrous minerals such as calcic amphibole, zoisite, and, or, phlogopite. In the case that hydrous minerals were stable at peak metamorphism—as in quartz eclogite—garnet hosts little or no water. If hydrous minerals are not part of the peak assemblage— as in common coesite eclogite—garnet contains distinctly more water. The latter was apparently derived from eclogite-facies hydrous minerals, which decomposed and liberated their H2O due to overstepping their stability field during UHP metamorphism. Moreover, garnet in coesite eclogite is more Ca-rich than garnet in quartz eclogite. This is ascribed to the breakdown of prograde zoisite, liberating Ca and facilitating a higher grossular content, which, in turn, enhances the garnet’s capacity for water storage. This study further suggests: (1) post-peak metamorphic introduction of secondary fluid; (2) relatively dry conditions prior to fluid influx, because only water-deficient garnet is able to incorporate additional structural water; (3) The determined primary contents of structural water were probably not modified by decompressional water loss, because the latter should only occur if the water content at peak pressure is 75 % of the maximum storable amount; (4) Since garnet from both eclogite types was water-deficient at the metamorphic peak it is unlikely that the different water contents are related to pressure; (5) The mineral assemblage and the dehydration of hydrous minerals is definitely more important in this context; (6) Garnet and, by implication, omphacite from both eclogite types was able to incorporate only part of the water liberated by hydrous minerals, a great part must have been released to hanging-wall rocks; and (7) The study points to a VC The Author(s) 2018. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com 207 J O U R N A L O F P E T R O L O G Y Journal of Petrology, 2018, Vol. 59, No. 2, 207–232 doi: 10.1093/petrology/egy022 Advance Access Publication Date: 28 February 2018