Electrical conductivity of magma in the course of crystallization controlled by their residual liquid composition

摘要

The electrical conductivity of a magma in the course of crystallization was experimentally investigated in the temperature range of 1350–1018°C. Large samples of basaltic composition with a homogeneous crystal content were synthesized in a gas mixing furnace at 1 atm pressure. The samples were analyzed by electron microprobe. The relative proportions of the phases as a function of temperature were determined. Depending on temperature, the phase assemblies included quenched silicate liquid, ±plagioclase, ±pyroxene, ±Fe-Ti oxides. The crystal content varied from 0 to 80 wt %. In response to partial crystallization, the residual liquid changed composition from basalt, to andesite, to dacite liquid. The electrical conductivity of the partially crystallized basaltic samples was measured. In addition, above liquidus conductivity measurements were conducted on compositions matching the residual liquid at different temperature. These supplemental electrical measurements allowed us to discriminate the effect of crystal content from the effect of changing liquid composition associated with partial crystallization. Combining with the modified Archie's law a set of constraints describing the conductivity of the residual liquid versus chemical composition and temperature, we propose an equation to calculate changes in conductivity associated with partial magma crystallization. We showed how the composition of the residual liquid is critical on the electrical behavior of crystal-liquid system. The model overcomes the previous difficulties in finding a robust model for describing the electrical behavior of crystal-liquid systems. The effect of liquid composition on the electrical conductivity is related to diffusion mechanisms and transport properties in molten silicate. Combining known constraints on Na tracer diffusion and our conductivity results confirms the statements that sodium is the dominant charge carrier silicate liquids from basalt to rhyolite. These findings revealed that we need a comprehensive model that can predict the conductivity of molten silicate as a function of chemical composition.