Benchmarking the Sulfur Isotopic Composition of the Upper Mantle

Nobu Shimizu, Geology & Geophysics



This is a proposal for benchmarking the sulfur isotopic composition of the upper mantle by determining in-situ sulfur isotopic compositions of olivine-hosted melt inclusions from the global mid-ocean ridge system.  The sulfur isotopic composition of the mantle has historically been assumed to be the same as troilite (FeS) in the Canyon Diablo iron meteorite, but there have been indications that this might not be the case.  Since surface sulfur reservoirs possess enormous isotopic variations, deep recycling of rocks and sediments that existed on the surface could transfer the isotopic variations into the mantle and contribute to the shift of mantle isotopic composition, as well as to large isotopic variations in mantle-derived melts.  In order to make a step toward comprehensive understanding of the process, it is proposed here to benchmark sulfur isotopic compositions of the upper mantle. Primitive olivine-hosted melt inclusions have been shown to display large chemical and isotopic variations attributable to the mantle sources, and can be used here to assess statistical distribution of recycled components in the MORB source mantle.

Isotopic compositions of various elements measured in mid-ocean ridge basalts (MORB) have been used as benchmarks for the earth’s upper mantle, and models for deep recycling of elements and geochemical evolution of the mantle have been built upon them.  For the isotope geochemistry of sulfur, the mantle has historically been assumed to be uniform with the same isotopic composition as that of chondritic meteorites and of troilite (FeS) in Canyon Diablo (e.g., Thode et al., 1961).  This sulfide has been used as the standard for expressing sulfur isotopic compositions in ä34S notation [(34S/32S)sample/(34S/32S)std -1]x1000, where (34S/32S)std is the sulfur isotope ratio of the Canyon Diablo troilite (or the CDT scale; the new convention is based on the IAEA standard, VCDT scale). Studies made in the 1980’s on mantle-derived basalts and sulfide globules in them appear to reinforce the assumption. For instance, Sakai et al. (1984) found “a narrow but significant range of variation from -0.3 to +0.8‰” for MORB, and Torssander (1989) stated that the sulfur isotopic composition of the mantle could be estimated to be between -0.5 and +1.0‰. A close inspection of the data on MORB (Sakai et al., 1984) reveals, however, that there could be regional variations for the global mid-ocean ridge system as shown in Fig. 1. Furthermore, Chaussidon et al. (1989) concluded, on the basis of sulfur isotope mass balance for continental crust, oceanic crust, seawater, and the mantle, that the mantle contains 93.2 wt% of sulfur and ä34S of the mantle is +0.8‰, significantly off the historical assumption.