EXPERIMENTAL STUDIES OF TRACE-ELEMENT PARTITIONING APPLICABLE TO IGNEOUS PETROGENESIS - SEDONA 16 YEARS LATER

In the 16 years since the Sedona Conference on the behaviour of trace elements in silicate systems, numerous studies providing new data have filled many of the gaps in knowledge of trace-element partitioning ev ident at that conference. The advent of new microbeam techniques for i n situ trace-element analysis has provided great impetus for this work . For example, values for large ion lithophile element (LILE) and high field strength element (HFSE) partitioning between olivine, pyroxene, garnet, amphibole and titanate minerals and silicate liquids have bee n determined. When plotted on mantle normalizing geochemical diagrams, partition coefficients for the main mantle silicate minerals show ste eply inclined patterns (over several orders of magnitude) from LILE to heavy rare-earth elements (HREE). Amphibole, however, has a relativel y flat pattern (though still favouring HREE over LILE by about an orde r of magnitude). Also, there is a notable flattening of the patterns f or HREE in pyroxenes and garnet relative to olivine. The effects of pr essure, temperature and melt composition on trace-element partition co efficients have been evaluated, as well as crystal-chemical controls o n the substitution of trace elements in minerals. This has led to form ulation of relationships between mineral compositions and trace-elemen t partition coefficients for olivine, low-Ca pyroxene and calcic pyrox ene. These studies have been motivated by the need of geochemists for partition coefficient data to apply to models of igneous petrogenesis and mantle evolution. Overall, the new data show systematic and consis tent behaviour, as determined in different laboratories. Also, partiti oning relationships for key elements between selected mantle minerals and H2O-rich fluids have been established experimentally, although fur ther work exploring the effect of variable fluid composition, for a wi der range of mantle minerals is needed. Future work will allow refinem ent and more precise ''fingerprinting'' of geochemical processes, incl uding the role of fluids and trace element-enriched accessory minerals in metasomatic events. Also, experiments are needed to extend the par tition coefficient determinations to much greater pressure, to resolve controversies over the geochemical evolution of the mantle at depths greater than similar to 130 km.