Zeolitization of a devitrified high-silica rhyolitic tuff producing dachiardite: A comparison of hydrothermal experiments with the corresponding reaction progress modeling

We have investigated the hydrothermal alteration of polished wafers of tuff reacted with dilute groundwater at 90 & DEG;C, 150 & DEG;C and 250 & DEG;C for time periods ranging from 2 months to nearly 1 year. The polished rock wafer provided a convenient surface upon which to grow secondary minerals. Reaction product minerals were iden-tified and analyzed at the end of each experiment and, along with the evolving fluid chemistry, were compared to computational results from corresponding reaction progress models. At 250 & DEG;C after a few months the run products in the experiment were dominantly the mordenite group zeolite minerals: dachiardite (a Ca-rich variety) and mordenite, itself. At 150 & DEG;C after a few months of reaction only minor amounts of clay were produced, but after 1 year of reaction at this temperature both mordenite group zeolites were again present. At this lower temperature the total amount of run products was much smaller. At 90 & DEG;C no run products could be seen at all, even after 1 year of reaction. The reaction progress modeling results for reaction products were in good relative agreement with the experimental results. The higher the temperature, and the greater the extent of reaction, the better the fluid phase modeling results agreed with the actual experimental results. At 250 & DEG;C the agreement was good for nearly all elements. At 150 & DEG;C agreement for pH, SiO2, Na and K were good, but less good for Al, Mg and Ca, especially after short reaction times. At 90 & DEG;C agreement for pH, SiO2 and Na was reasonable, but not as good for the other elements, and all modeling results for short reaction times did not match experimental results as well as the longer time results. This study demonstrates that reaction progress modeling provides a powerful tool for predicting hydrothermal rock-water interactions, with results expected to improve, as more and better quality thermodynamic and kinetic data become available and as process-oriented simulators incorporate better and more comprehensive sub-models for mineral dissolution and growth.