The redox state of subduction zones: insights from arc-peridotites

Spinel peridotites from a variety of island arcs have been utilised to calculate the redox state of the mantle wedge above subduction zones. Oxygen fugacities (f(o2) values) calculated from the ferric iron content of spinels, measured by Electron Microprobe (EMP) using secondary standards [Wood, B.J., Virgo, D., 1989. Upper mantle oxidation state: ferric iron contents of Iherzolite spinels by Fe-57 Mossbauer spectroscopy and resultant oxygen fugacities. Geochim. Cosmochim. Acta, 53, 1277-1291.], yield values which range from 0.3 to 2.0 above the fayalite-magnetite-quartz (FMQ) buffer. These data provide further evidence that the mantle wedge is ubiquitously oxidised relative to oceanic and ancient cratonic mantle. There is no correlation between f(o2) values and the presence of hydrous phases and, in fact, the most oxidised samples contain no hydrous phases. Within individual suites there is no correlation between f(o2) and degree of depletion as indicated by spinel Cr#, except for a suite of reacted forearc-peridotites. However, when the data is viewed as a whole there is broad a positive correlation between f(o2) and spinel Cr# suggesting that partial melting processes may influence the redox state of the mantle wedge. We suggest that the ultimate source of the oxygen which oxidises the mantle wedge is from the subducted slab. It is not clear whether this oxidising agent is a solute-rich hydrous fluid or a water-bearing silicate melt. However, our data does indicate that silicate melts are effective oxidisers of the depleted shallow upper mantle. Simple mass balance calculations based on the ferric iron content of primitive subduction zone magmas indicates that the source region must contain 0.6-1.0 wt.% Fe2O3. This amount of Fe2O3 in a fertile spinel peridotite yields an oxygen fugacity of 0.5-1.7 log units above FMQ in the IAB source. If water is the sole oxidising agent in the mantle wedge then 0.030-0.075 wt.% H2O is required which is considerably less than the 0.25% H2O envisaged by Stolper and Newman [Stolper, E.M., Newman, S., 1994. The role of water in the petrogenesis of Mariana trough magmas. Earth Planet. Sci. Lett., 121, 293-325.], suggesting water is not necessarily an efficient oxidising agent. Alternatively, ferric iron may be added to the mantle wedge by addition of a ferric iron-rich sediment melt or more likely as a solute-rich hydrous fluid. This model would produce spinel, orthopyroxene or amphibole in the wedge with only a slight increase in f(o2) of the source region. Although it is unclear which model is correct the maximum f(o2) of the fertile mantle wedge is unlikely to be above FMQ + 2 and therefore some decompression melting in the mantle wedge is required to explain the higher f(o2) values of primitive are lavas than arc-peridotites. (C) 1999 Elsevier Science B.V. All rights reserved.