LASER MICROPROBE ANALYSIS OF CARBON DIOXIDE AND WATER IN GLASS INCLUSION: AN ATTEMPT TO DETERMINE THEIR CONCENTRATIONS IN PRE-ERUPTIVE MAGMA

Volatile materials dissolved in magmlas play an important role in the evolution of magmas and volcanic eruptions. During magma ascent, the magma may be saturated with the volatiles because of decrease in solubilities of the volatiles in silicate melts with pressure. Formation of a gas phase would lead to decrease in the bulk density of magma and thereby accelerate the magma ascent. Expansion of the resulting gas phase may induce a volcanic eruption. The major volatile components in a nlagma are H(2)0, C0(2), S and Cl. Glass inclusions in phenocrysts are the most suitable samples for measuring volatile concentrations in pre-eruptive magmas. However, it has been difficult to measure the H(2)0 and CO(2) concentrations in glass inclusions, mainly because the size of the glass inclusions is very small (50-200 μm) and the sensitivity of analytical methods is not　sufficient. In this study, we have developed a laser microprobe　technique for determining a minute amount of H(2)0 and C0(2) dissolved in glass inclusions. The analytical system consisting of a Nd-YAG laser for selective heating of glass inclusions and a gas chromatograph-mass spectrometer for measuring a micro quantity of H(2)0 and C0(2) dissolved in glass inclusions. The H(2)0 and C0(2) were extracted from glass inclusions by laser heating under a vacuum. The extracted gas was collected in a cold trap and then introduced to GCMS by carrier gas. Gas concentration was calculated from the absolute amount of the extracted gas and the mass of glass melted. The optimum working conditions (lamp current, pulse frequency and shooting duration) of laser extraction were decided using a basaltic glass sample. A glass sample ground to less than 100 μm thick was pierced by a laser beam and measured for diameter of the hole produced to obtain the accurate determination of the volume of the glass melted. The mass of the melted glass was calculated from its volume and density. Efforts were made for reduction of background level of C0(2) and H(2)0. High purity helium carrier gas was farther purified with cold traps to remove C0(2) and H(2)0 contained in the carrier gas. Analytical line was thoroughly baked out at 300℃ before the analyses. With these procedures, the blank C0(2) was reduced to less than 0.07 ng and the blank H(2)0 to less than 6 ng. The detection limit of the present system was found to be 0.15 ng C0(2) and 15 ng H(2)0, considering the confidence limit of the calibration. The repeated analyses of C0(2) in basaltic glass indicate that the present technique allows us to analyze C0(2) concentration of glass inclusions as small as 70 μm in diameter within an accuracy of ±60 ppm, assuming that the glass inclusion contains 300 ppm C0(2). The H(2)0 analyses of the glass by the present technique gave significantly lower H(2)0 concentration than the bulk analysis, suggesting incomplete degassing of H(2)0 from molten glass during laser irradiation due probably to its slow diffusion rate. The glass inclusion samples from Killauea volcano, Hawaii, and Izu-Oshima volcano, Japan, were analyzed to determine CO(2) concentrations of pre-eruptive magma. The C0(2) concentration of pre-eruptive magma in South-West Rift Zone of Kilauea volcano was 230 ppm. This suggests that the magma has been significantly degassed during its storage in the summit magma chamber and migration to the rift system. The depth of the magma chamber was estimated to be 2.8-4.0 km from the C0(2) concentration of the pre-eruptive magma, in good agreement with the geophysical estimation. Comparison of the volatile budget of Kilauea volcano based on the C0(2) analysis of glass inclusion with the observed C0(2) flux suggests that parental magma contains 3000 ppm C0(2). The bulk density of magma containing exsolved gases is not snlall enough to ascend to the surface by its own buoyancy. This is consistent with the present settings of the Kilauean magma; the sumrnit magma chamber and rift magma system extending from the summit caldera. The C0(2) concentration of pre-eruptive magma of Izu-Oshima volcano was found to be 1 70 ppm. The measured C0(2) concentration indicates that the magma chamber exists at a depth of about 2 km beneath the volcano. At this depth, the magma should be saturated with C0(2). The estimated bulk densities of the magma (d=2.4-2.6 g/c㎥) containing C0(2)-rich gas phase are consistent with the density structure of the volcano, if the magma before degassing contains C0(2) more than 1700 ppm. The rate of magma supply was estimated from the C0(2) concentration of pre-eruptive magma and C0(2) fluxes based on measurements on surface volcanic gases. Assuming the primary magma contains 2500 ppm C0(2), the estimated rate of magma supply(6-19x 10(3) tons/day) is compatible with the observed rate of magma supply based on the total mass of ejecta of major eruptions during the last 1500 years.