THE CARBON ISOTOPE GEOCHEMISTRY OF MANTLE XENOLITHS

Abstract

Carbon occurs in mantle samples in several chemical, mineralogical and morphological forms. It has been observed as CO2, CH4 and CO in fluid inclusions, as carbonate, graphite, diamond, moissanite, solid solution in silicates, and organic compounds. The total carbon concentration reported for mantle xenoliths varies by four orders of magnitude from below 1 ppm to close to 10 000 ppm. About 40% of these samples contain less than 50 ppm, 70% less than 100 ppm and 95% less than 500 ppm C. Carbon with δ13C of about −5‰ has been identified as a major isotopic composition signature for the mantle (carbonatite and kimberlite carbonates, diamonds and volcanic CO2 exhalations); it is also observed in mantle xenoliths. However, there may also be a minor signature of C depleted in 13C (δ13C=−22‰ to −26‰). Such light carbon has been observed in the dissolution residue of mantle minerals (olivine, pyroxene) and rocks, C fractions that have been interpreted as C dissolved in silicates, in diamonds, graphite, carbide, and hydrocarbons which are thought to be indigenous to the mantle. The data on xenoliths from basalts indicate that their δ13C distribution is essentially bimodal with peaks at −5‰ and −25‰ although the geologic occurrence of this light carbon has not yet been clearly delineated. Xenoliths from both hotspot and non-hotspot volcanoes cover the whole C isotopic composition range observed in mantle xenoliths; however, on average, xenoliths from non-hotspot volcanoes contain isotopically lighter carbon. Xenoliths from kimberlites cover the whole isotopic composition range as well, but, on average, probably show the lowest degree of 13C depletion. In addition, the second, low δ13C, mode may occur just above −20‰, coincident with the low 13C mode of southern African diamonds. Differences in the C concentration and isotopic composition have been observed between mantle minerals. The data are too few, however, to support firm conclusions on their size, or on how systematic these differences might be. Chemically more fractionated xenoliths tend to have higher δ13C values than less chemically fractionated ones. Processes that have been considered to be responsible for the considerable δ13C range in mantle C include the subduction of organic material and degassing. The observations on mantle xenoliths do not provide support for either, but indicate that as yet unexplored thermodynamic isotope effects, probably involving dissolved C in minerals and SiC bonds, may be responsible for the observed mantle carbon isotope distribution. The occurrence of such isotope effects would help to understand a number of observations on the carbon isotope geochemistry of diamonds. In so far as mantle-degassing models have been based, in part, on the carbon isotopic composition and C/3He ratios, an understanding of the mantle carbon isotope geochemistry is essential to support or refute their validity. The xenolith data do not support degassing models based on the assumption of limited indigenous carbon isotope variability within the mantle, nor the supposition that all 13C depleted carbon is of surface origin. The relative proportions of mantle C's of differing isotope signature are not known; they will have to be established for well-founded C cycle models to be developed.