Density measurements on melts at high pressure using the sink/float method: Limitations and possibilities

We investigated the use of the sink/float method to determine the density of silicate melts with a range of compositions intended to provide a general understanding of density changes as functions of composition and pressure. ch composition, estimates of both the melting temperature and the melt viscosity (which determines the required experimental run time) at the pressures of interest are required prior to the density experiments. We constrained the melting relationships for two sodium silicate melts and a melt of albite composition and were able to calculate the maximum possible viscosities of the sodium silicate melts. nk/float method is limited by the absence of high pressure-high temperature density data for minerals that may be used as buoyancy markers in the experiments. These data need to be sufficiently precise to justify their use in the determination of melt densities at high pressure. At present, high-quality data are available for only a few minerals, such as spinel, forsterite, quartz and some garnets. r limitation is that the markers should not react with the melt. Reaction between the buoyancy markers, melt and capsule material changes the density of both melt and markers. Errors on the order of ≥ 2% on the determined densities may be introduced by the changing density of the mineral markers through reaction with the melt. gh the sink/float method is very elegant, it has significant limitations that prohibit the precise determination of the density of an arbitrary silicate melt at high pressure; in most cases, only loose constraints on the density and compressibility of melts can be obtained as a function of pressure.