Search

We present an inventory of B, Cl and Li concentrations in (a) key minerals from a set of ultramafic samples featuring the main evolutionary stages encountered by the subducted oceanic mantle, and in (b) fluid inclusions produced during high-pressure breakdown of antigorite serpentinite. Samples correspond to (i) nonsubducted serpentinites (Northern Apennine and Alpine ophiolites), (ii) high-pressure olivine-bearing antigorite serpentinites (Western Alps and Betic Cordillera), (iii) high-pressure olivine–orthopyroxene rocks recording the subduction breakdown of antigorite serpentinites (Betic Cordillera). Two main dehydration episodes are recorded by the sample suite: partial serpentinite dewatering during formation of metamorphic olivine, followed by full breakdown of antigorite serpentine to olivine+orthopyroxene+fluid. Ion probe and laser ablation ICP-MS (LA ICP-MS) analyses of Cl, B and Li in the rock-forming minerals indicate that the hydrous mantle is an important carrier of light elements. The estimated bulk-rock B and Cl concentrations progressively decrease from oceanic serpentinites (46.7 ppm B and 729 ppm Cl) to antigorite serpentinites (20 ppm B and 221 ppm Cl) to olivine–orthopyroxene rocks (9.4 ppm B and 45 ppm Cl). This suggests release of oceanic Cl and B in subduction fluids, apparently without inputs from external sources. Lithium is less abundant in oceanic serpentinites (1.3 ppm) and the initial concentrations are still preserved in high-pressure antigorite serpentinites. Higher Li contents in olivine, Ti-clinohumite of the olivine–orthopyroxene rocks (4.9 ppm bulk rock Li), as well as in the coexisting fluid inclusions, suggest that their budget may not be uniquely related to recycling of oceanic Li, but may require input from external sources. Laser ablation ICP-MS analyses of fluid inclusions in the olivine–orthopyroxene rocks enabled an estimate of the Li and B concentrations in the antigorite breakdown fluid. The inclusion compositions were quantified using a range of salinity values (0.4–2 wt.% NaClequiv) as internal standards, yielding maximum average fluid/rockDB∼5 and fluid/rockDLi∼3.5. We also performed model calculations to estimate the B and Cl loss during the two dehydration episodes of serpentinite subduction. The first event is characterized by high fluid/rock partition coefficients for Cl (∼100) and B (∼60) and by formation of a fluid with salinity of 4–8 wt.% NaClequiv. The antigorite breakdown produces less saline fluids (0.4–2 wt.% NaClequiv) and is characterized by lower partition coefficients for Cl (25–60) and B (12–30). Our calculations indicate that the salinity of the subduction fluids decreases with increasing depths. fluid/rockDB/fluid/rockDCl