Search

Kimberlite‐borne eclogite xenoliths having Precambrian oceanic crustal protoliths and entrained from ≥100 km depth can retain pristine geochemical features despite extended residence in the cratonic lithospheric mantle, making them valuable archives of deep chemical cycling including that of water. We determined, by Fourier Transform Infrared Spectroscopy, structural OH contents in clinopyroxene and garnet from 15 unmetasomatized eclogite xenoliths. Calculated total c(H2O) is 100–510 wt.ppm for clinopyroxene and below detection (∼2 wt.ppm) to 200 wt.ppm for garnet, while garnet δ18O, determined by Secondary Ion Mass Spectrometry, ranges from +5.0‰ to +7.3‰, (similar to high‐ and low‐temperature seawater‐altered oceanic crust). Estimated electrical conductivity in pristine eclogites increases with temperature (i.e., depth for conductive geotherms), while clinopyroxene‐garnet H2O partition coefficients decrease with increasing temperature and garnet grossular component (i.e., Ca#), similar to other incompatible components. Various considerations suggest the retention of primary H2O in the samples, likely occurring in km‐sized pods of coarse‐grained eclogite. High Al2O3 in clinopyroxene as omphacite component, stabilized during high‐pressure metamorphism, facilitates H2O uptake. Therefore, the high bulk c(H2O) estimated for samples with plagioclase‐rich, deep crustal protoliths (median 290 wt.ppm) may indicate an interaction with fluids expelled at depth from serpentinites. The c(H2O) of ancient and modern subducted bulk oceanic crust (∼220–240 wt.ppm) are similar, suggesting constant mantle ingassing since at least 3 Ga ago. This places constraints on factors, such as mantle temperatures, that determine the efficiency of deep water cycling. Plain Language Summary: Water in Earth's interior exists mostly as OH− anion in trace abundances in nominally anhydrous minerals. Despite these low concentrations, deep water exerts a strong influence on fundamental Earth processes, such as partial melting of the mantle and the operation of plate tectonics. However, the extent to which the loss of water via volcanism has been compensated over time by retention in downgoing oceanic plates, after their dehydration and metamorphism to eclogite, remains poorly known. Deeply buried Archean and Paleoproterozoic oceanic crust is sampled as remarkably pristine eclogite fragments quickly exhumed by volcanism from depths >100 km. High Al2O3 contents, characteristic of deep crustal plagioclase‐rich cumulates, facilitate H2O‐uptake in clinopyroxene, the main carrier of H2O in eclogite. Because clinopyroxene is rich in Al only at high pressure, where the crust is already dehydrated, but the underlying seawater‐altered mantle begins to liberate fluids, we suggest that interaction with this fluid explains the Al‐H2O association. Moreover, the eclogite‐based H2O estimate for ancient crust is similar to estimates for the modern crust, suggesting that deep water cycling in the crustal part of subducting slabs changed little in the last 3 billion years, with consequences for the factors determining the efficiency of ingassing. Key Points: Electrical conductivity in deep (>100 km) pristine eclogite xenoliths from Siberian and Slave cratons increases with depthH2O content increases with Al2O3, which is highest in gabbroic eclogites with deep oceanic crustal protolithsThe H2O content of subducted bulk oceanic crust sampled by eclogite 3 Ga ago may be similar to today [ABSTRACT FROM AUTHOR]

The origin and evolution of the subcontinental lithospheric mantle (SCLM) underlying orogenic belts bordering cratons, as either newly added domains or strongly reworked older cratonic lithosphere, remains controversial. This limits our understanding of deep lithospheric behavior and processes during plate convergence. Here, we use detailed petrology, whole-rock and mineral compositions, and in-situ Sr isotopic compositions for spinel-facies peridotite xenoliths from Jiaohe (northeastern China) in the southeastern Central Asian Orogenic Belt (CAOB), which is pinched between two NE-striking Mesozoic to Cenozoic giant trans-lithospheric fault systems. These data are combined with whole-rock Re-Os isotope and platinum-group element compositions and literature data, to explore the physicochemical evolution of the regional SCLM and examine crust-SCLM relationships for this vast swath of continental lithosphere. Lherzolites predominate at Jiaohe, and have higher whole-rock Al2O3 contents than harzburgites (2.27-3.46 wt% versus 0.84-1.02 wt%), but exhibit FeO enrichment similar to harzburgite (bulk FeO up to 9.54 wt%). The lherzolitic clinopyroxenes (Cpx) have higher heavy rare earth element (HREE) levels (2.79-5.11 ppm) and Ti/Eu (3882-6864), coupled with wider variation of 87Sr/86Sr (0.7021-0.7038) and lower average ƒO2 (oxygen fugacity relative to the fayalitemagnetite- quartz buffer; FMQ-1.92 ± 0.55) compared to the harzburgites (HREE: 0.94-2.11 ppm; Ti/Eu: 163-2044; 87Sr/86Sr: 0.7032-0.7036; ƒO2 = FMQ-1.25 ± 0.20). All these observations suggest that the lherzolites were produced by adding Cpx ± spinel to the protoliths from reducing silicate melts, whereas the harzburgites may have originated from a remnant, highly depleted mantle domain that variably interacted with silica-undersaturated oxidizing melts. Bulk-rock and mineral compositional relationships as well as ƒO2 for peridotite samples from the southeastern CAOB (including Jiaohe) are similar to those of the northeastern NCC (including Huinan). Their characteristics contrast with those from the northwestern CAOB, suggesting that they belong to distinct lithospheric provinces with respect to both composition and oxidation state. Jiaohe harzburgites yield Paleoproterozoic Re depletion Os model ages (TRD up to 1.76 Ga), similar to Huinan peridotites (up to 1.92 Ga), just 170 km away, and may also represent strongly reworked cratonic mantle. By analogy with evidence from recent continental collision zones, the Jiaohe SCLM may have originated as cratonic SCLM that migrated northeastward during plate convergence along giant translithospheric fault systems. [ABSTRACT FROM AUTHOR]