Sulfur content and isotope composition of the terrestrial mantle can provide major constraints on the internal dynamic of this element and on nature of mantle heterogeneities.
The mean S isotope composition of this reservoir can also characterize the potential mantle-core equilibrium and hence help to decipher the question of sulfur origin on Earth.
In this study, we report S content and relative abundances of 32S, 33S, 34S and 36S in more than 120 glassy basalts coming from 3 ocean ridges and a transform fault. We propose a new sulfur extraction
protocol for silicate glasses based on hydrofluoric acid. We evidenced an analytical bias aecting the previously published dataset, probably due to the use of the "Kiba reagent". The re-determined 34S
are approximately 1‰ lower than the previously published data. We also provide the rst 33S and 36S data for ocean ridge basalts. These data are, within uncertainty, indistinguishable from our
CDT estimate, conrming the idea that terrestrial mantle has a 33S and 36S close to 0.000‰ with respect to this international standard.
The 34S variability encompasses a 3‰range in ocean ridge basalts and correlations between 34S and radiogenic isotopes demonstrate that S isotopes are tracers of mantle heterogeneities. On the other
hand, the mean 34S of MORB mantle is -0,80 ± 0,38‰, signicantly lower than the chondritic value of 0,04 ± 0,31‰. The Pb-S isotope systematic in Pacic Antarctic basalts indicates that recycled
oceanic crust bears a positive 34S. As a corrollary, this result implies that the depleted part of the mantle is even more negative than the mean MORB value. We propose a 34S value of -1,40 ± 0,20‰
for this reservoir.
Sunrise on the Indian Ocean, southern summer 2010, during the mission ocean GEISEIR 2. The photo was taken by the author of these lines
In South Atlantic basalts, the 34S are correlated with Sr isotopes, trends being consistent with recycled sediment feeding the mantle source of these basalts. Taken together, the 34S variability
associated to the 33S and 36S constancy indicates that the recycled components are of postarcheanage, in agreement with independant estimation of their mean age using Pb isotopes. Finally,
the negative 34S of MORB mantle can be reconcilied only with a mantle-core equilibrium, implying an internal origin of sulfur as opposed to a late origin.