PREFERENTIAL WATER-LOSS FROM SYNTHETIC FLUID INCLUSIONS

A fundamental question in most fluid inclusion studies is whether incl usions behave as compositionally closed systems after trapping, and, t hus, represent samples of the fluid phase(s) present in the system at the time of their formation. This question was addressed in high-tempe rature laboratory experiments with synthetic fluid inclusions in quart z and it was found that at 825-degrees-C the inclusions exhibited open -system behavior with respect to water. Synthetic salt-water fluid inc lusions in quartz were reequilibrated for 12 hours to 35 days at 825-d egrees-C in a dry argon atmosphere under 1.5 kbar confining pressure. These conditions created initial internal overpressures (P(int) > P(co nf)) of 1.5-4 kbar in the inclusions and differential water fugacities in the same sense (i.e., f(H2O)int > f(H2O)conf). After 108 hours of reequilibration, preferential water loss had resulted in salinity incr eases as large as 22 wt% salt (e.g., from 57 to 79 wt% NaCl, as determ ined from measured temperatures of salt dissolution). Also, following reequilibration, a strong inverse correlation between salinity and inc lusion volume was observed, and this trend became more pronounced with increasing reequilibration time. These observations, together with a lack of evidence for selective H2O removal via hydration reactions, su ggest that water loss occurred by a diffusion-related mechanism. Fluxe s of almost-equal-to 4 x 10(-11) g/cm2-s and diffusion coefficients on the order of 10(-9) cm2/s are calculated for water loss from the incl usions. The calculated H2O diffusion coefficient is consistent with th e determination of Blacic (1981) derived from hydrolytic weakening exp eriments, but is much larger than the value obtained by Giletti and Yu nd (1984) for volume diffusion of oxygen in isotope exchange experimen ts. These observations suggest that the mechanism of water loss from o ur synthetic fluid inclusions may have been pipe diffusion along dislo cations, subgrain boundaries or other structural defects rather than b ulk volume diffusion. The results of this study are relevant to the in terpretation of fluid inclusions in quartz from several natural high-t emperature environments where water fugacities of included and ambient fluids are known to have evolved along separate paths over geologic t ime.