Direct measurements of latent heat during crystallization and melting of a ugandite and an olivine basalt

Step-scanning calorimetric measurements using a Setaram HT1500 calorimeter were performed between 800 and 1400°C on two natural samples: a ugandite from the East African rift and an olivine basalt from the western Mexican arc. Our measurements provide the first in-situ quantitative assessment of enthalpy during melting of initially crystalline natural samples. The distribution of latent heat across the liquidus-solidus intervals of the two samples is distinctly different, reflecting significant variation in the sequence and abundance of mineral phases during melting (clinopyroxene and leucite in the ugandite; olivine, clinopyroxene, and plagioclase in the basalt). Our data further indicate that the common assumption of a uniform distribution of latent heat across the liquidus-solidus interval of a magma is a reasonable approximation for the olivine basalt, but is grossly in error for the ugandite. This is due to cotectic precipitation of leucite and clinopyroxene, leading to a large, disproportionate release of latent heat early in the crystallization sequence. The implication for the thermal history of a crystallizing ugandite magma is that the rate of heat loss during conductive cooling will unitially be more rapid than the average rate. The net result will be to produce lower magmatic temperatures after a given cooling interval relative to models assuming a uniform release of latent heat. An additional series of scanning calorimetric experiments were performed at variable rates (1,2 and 3°/min) to evaluate the role of kinetics on the distribution of enthalpy during both melting and crystallization of the ugandite and olivine basalt. The results indicate that clinopyroxene is the most important mineral phase in controlling the shapes of the enthalpy profiles during cooling; this is due to its large enthalpy of fusion and its tendency for sluggish nucleation, followed by rapid crystallization at temperatures that vary with cooling rate. The resolution of the calorimeter (in terms of heat detected per unit time) is also important in determining the shapes of the observed enthalpy profiles during these rapid scans. Estimates based on the observed calorimetric signal associated with melting of olivine, and the lack of a calorimetric signal during melting of leucite, combined with known enthalpies of fusion for the two phases, indicate detection limits of approximately 0.6-1.2 kJ per 5 min increments.