Subduction zone metamorphic pathway for deep carbon cycling: I. Evidence from HP/UHP metasedimentary rocks, Italian Alps

• Effects of subduction zone metamorphism on carbon cycling were examined.
• Calculated decarbonation history matches history observed in an Italian Alps suite.
• Extent of CO2 retention dependent upon degrees of infiltration by H2O-rich fluids
• Exact flow fluid paths and infiltration histories likely to differ from margin to margin
• Subducting sediments retain most of their carbon to depths beneath volcanic fronts.

Metamorphosed Jurassic oceanic sediment in the Italian Alps experienced peak P–T conditions (1.5–3.0 GPa; 330–550 °C) similar to those experienced by sediments subducting through forearcs in most modern subduction zones. Integrated field, petrologic, and geochemical study of the devolatilization history of these rocks provides evidence regarding the extents of loss and mobility of oxidized and reduced C during subduction of sediments to depths beneath volcanic fronts, thus constraining models of the release of initially subducted C into the atmosphere via arc volcanism or contributing to the C budget of the deeper mantle.In this suite, occurrences of lawsonite and grossular-rich garnet in higher grade calcschists indicate some decarbonation reaction along a P–T gradient of ~ 8 °C/km. Across grade, carbonates largely retain δ13CVPDB values typical of marine carbonates (mostly ranging from − 1.5 to + 1.5‰), with some shifts to lower values in low-carbonate samples possibly in part reflecting decarbonation but likely largely due to isotopic exchange with abundant reduced C. Carbonaceous matter in some carbonate-poor samples shows increase in δ13C at higher grades, from values of − 25 to − 22‰ typical of marine organic matter, to values as high as − 10‰, consistent with the expected effects of devolatilization involving release of varying fractions of the original reduced C into fluids as CH4. In more carbonate-rich samples, shifts in the carbonaceous matter to higher values, by up to 15‰, reflect varying degrees of exchange with the carbonate in these samples. Carbonate across grade has δ18OVSMOW of + 17 to + 22‰ significantly lower than values typical of marine carbonates (the latter + 28 to + 30‰), likely in part reflecting exchange with silicate phases. Such shifts in the more carbonate-rich samples cannot be explained by closed-system exchange with silicates and require exchange with an external reservoir. The observed extents of decarbonation in these carbonate-bearing sediments subducted to depths of up to 90–100 km are consistent with limited open-system behavior, hence intermediate between models of Rayleigh-like CO2 loss and recent models supporting open-system flushing of carbonate-rich rocks by H2O-rich fluid from underlying subducting oceanic lithosphere. Based on a separate study of devolatilization history in the metapelites intercalated with the metacarbonates, and consideration of the O isotope compositions of the two lithologies, the metapelites appear to be sources for the H2O-rich fluid that drove decarbonation reactions and shifted the δ18O values of the carbonates to their present values. Calculated decarbonation histories of the calcschists, using the Perple_X software, predict significant release of CO2 related to the breakdown of CO2-bearing phases (dolomite, aragonite) and the growth of Ca-rich silicate phases (lawsonite, grossular-rich garnet). This decarbonation occurred at very low XCO2 (< 0.01), for the P–T range considered here, perhaps attained during pulses of infiltration by H2O-rich fluids from the nearby metapelites.These results demonstrate that a large fraction of the initially subducted oxidized and reduced C in pelitic–carbonate sediment can be retained during forearc metamorphism and thus be available for loss beneath arcs or addition to the deeper mantle. In forearcs, the C inventory in these sediments experience minimal isotopic shift due to devolatilization, and isotopic reequilibration of the oxidized and reduced reservoirs during heating likely results in little net change in the bulk δ13C of the subducting section. Further attention should be paid to understanding the extent of C release, and possible isotope fractionation, during decarbonation, partial melting, and carbonate dissolution at the 80–100 km depths over which subducting slabs experience greater heating by exposure to the convecting mantle wedge.