Abstract:
We present an empirical technique for estimating crystallization pressure and H2O content of MORB and BABB glasses using glass major element compositions alone. This technique is based on models describing 1 atm mineral — basaltic melt equilibria for olivine (ol), plagioclase (pl), and clinopyroxene (cp) which allow calculation of liquidus temperatures for each mineral on the basis of melt composition alone. A necessary requirement for using the proposed technique is cosaturation of glasses with of andpl for estimation of H2O contents, and withol andcp for estimation of crystallization pressures. Mineralogical criteria can be used to assess cosaturation. The following logic is employed: Calculated temperatures ofol, pl andcp liquidus must be the same for a glass which represents a quenched melt which crystallized these minerals at 1 atm. The effect ofpl, forcingcp to appear earlier on the liquidus. For higher-pressure crystallization, calculated 1 atm temperature forcp must be lower than that forol due to stronger dependence ofcp crystallization temperature on pressure. Thus the difference between calculated 1 atm liquidus temperatures forol andcp must reflect crystallization pressure. It is well known that H2O can depress crystallization temperature ofpl compared to those ofol andcp. Thus forol-pl saturated H2O-bearing melts, calculated 1 atm anhydrous temperatures forpl must be higher than those forol, and the difference between calculated anhydrous 1 atm liquidus temperatures forpl andol must reflect melt H2O content. p ]As the results of this technique are entirely dependent on the quality of glass major element analyses, we emphasize that application of this method to a single glass analysis can lead to erroneous results. A preferable approach is to obtain a statistically meaningful number of glass analyses for each suite and use averages. We also emphasize the danger of extrapolation of our technique to compositions beyond the MORB — BABB compositional range.