Uranium content of the oceanic upper mantle

Abstract

Fission track determinations of both the whole rock contents and the distribution of uranium in individual phases were made on twenty serpentinized ultramafic rocks from the Mid-Atlantic Ridge at 45°N (Hudson Geotraverse) and 52°N (Gibbs Fracture Zone). The rocks are thought to represent uppermost oceanic upper mantle material. Whole rock uranium concentrations, varying from 0.19 to 0.70 ppm, reflect more the subsequent histories of metasomatism (serpentinization, amphibolization and rodingitization) than concentrations of the original fresh rocks. Relic minerals reveal that, unlike continental mantle equivalents, most of the uranium is homogeneously distributed in primary orthopyroxenes (1 ppm), and to a lesser extent (0.2 ppm) in primary clinopyroxenes. Primary olivine is relatively depleted in uranium (0.03 ppm), as is primary chrome spinel (0.09 ppm). Extrapolation to pre-metasomatic conditions suggests that at the time of crystallization these ultramafic rocks had concentrations of at least 0.1 to 0.5 ppm uranium, up to an order of magnitude greater than expected. These concentrations suggest that the ultramafic rocks are unlikely to be directly genetically related to the overlying basalts and gabbros containing 0.25 ppm uranium, but are probably primary ultramafic material from which there has been no previous episode of basalt extraction. These uranium concentrations suggest that the oceanic upper mantle (plate) has quite high radioactive heat production in contrast to low heat production in the continental upper mantle. The equality of oceanic and continental heat flows is explained by the data, since the total heat produced in an oceanic plate is estimated to be about equal to that of the continental crust. One can construct a model that has an isothermal, low velocity (partial melt) layer at a shallower depth under the oceans than under the continents and that has the same heat flux from below the oceanic and shield plates. Lateral convective heat transfer in the low velocity layer is not required. High radioactive heat production of the oceanic plate can explain the high heat flows measured behind trenches with downgoing slabs.