Ra-partitioning between phlogopite and silicate melt and ^(226)Ra/Ba–^(230)Th/Ba isochrons

In this study we experimentally determine phlogopite/melt partition coefficients of Ra and other trace elements in a lamproitic system. This work was achieved using an analytical technique (LA-ICP-MS) with low detection limits (~0.01 fg) permitting the measurement of the very low Ra concentrations feasible in experiments (~1 ppb). D_(Ra)^(phlogopite/melt) was determined to 2.28 ± 0.44 and 2.84 ± 0.47 in two experiments, the ratio D_(Ra)/D_(Ba) is around 1.6. The compatibility of Ra in phlogopite results from an ionic radius being close to the apex of the lattice strain parabola for earth alkalis in the large XII-coordinated interlayer site of phlogopite. A re-evaluation of D_(Ra) and D_(Ra)/D_(Ba) for magmatic minerals containing appreciable Ra, yields D_(Ra)^(mineral/melt) ranging from ~2.6 for phlogopite down to 2–3 • 10^(−5) for pyroxenes, and D_(Ra)/D_(Ba)^(mineral/melt) from ~4 for leucite to 2 • 10^(−2) for orthopyroxene. The influence of melt composition on D_(Ra)/D_(Ba) is less than 10%. All investigated minerals have different D_(Ra)/D_(Ba), strongly fractionating Ra from Ba. Thus, for magmatic systems, (^(226)Ra)/Ba in the various minerals is not constant, these minerals do not form a straight line in the (^(226)Ra)/Ba–(^(230)Th)/Ba system at the time of crystallization and thus, there is no (^(226)Ra)/Ba–(^(230)Th)/Ba isochron at t_0. 226Ra–230Th–Ba mineral dating is thus applicable only to model ages calculated from mineral–glass pairs with known D_(Ra).