LILE-DEPLETION IN GRANULITES: MYTH OR REALITY?

摘要

The notion of LILE-depletion in granulites, which in the early seventies has led to the concept of granulite formation by melt extraction, has been for a long time almost a dogma in metamorphic petrology. After a short review of the historical background, it is shown that the chemical changes at the granulite/amphibolite boundary are far more complex than commonly assumed. Besides the elements commonly invoked (essentially heat-producing elements like K, U, Rb, Cs ), many other elements are concerned: REE, Ti, Zr, etc. The behaviour of these elements is quite variable from occurrence to occurrence: depletion, increase, immobility. A number of evidences, from field observation to mass balance evaluations, show that melt extraction is only one, probably minor, mechanism among many others. Element mobility at the lower crustal level is dominated by fluid-rock interaetion. Two major types of fluids are concerned, CO_2 and brines, which are generated by a number of processes: inheritage from sedimentary formation waters, fluids resulting from 'vapor-absent melting' or introduced from external sources during and after peak-metamorphic conditions. These fluids, not granitic melts, are responsible for most of the typical granulite features: control of H_2O aetivity (with direct implications for the stability of anhydrous mineral assemblages), selective dissolution and transport of chemical elements, including some considered traditionally as immobile during regional metamorphism (Zr,Ti, REE). Hypotheses are presented for the respective roles of both fluids: CO_2 could be involved in the transport of Zr and REE, brines in the control of alkali mobility.