Keywords:FTIR spectroscopy, hydration, submarine volcanology
Volatile degassing during magma ascent plays a fundamental role in determining eruption style and pyroclast formation, and the analysis of the dissolved volatile contents preserved in volcanic glasses (both melt inclusions and matrix glasses) is therefore an important tool for investigating eruption processes. Given its strong control on parameters such as melt viscosity, H2O is one of the most important magmatic volatiles, especially for felsic magmas where it is often the most volumetrically abundant volatile. In addition, the well-constrained pressure-dependence of H2O solubility means that the H2O contents of pyroclast glasses can provide a record of the pressure at which pyroclasts quenched. For submarine eruptions, where vent depth and pyroclast formation are rarely observed directly, this is particularly valuable information. Unfortunately, silicate glasses are vulnerable to secondary hydration (i.e. the addition of H2O from the surrounding environment at low temperature in the time following deposition), and thus volcanic glasses deposited in the submarine environment are often hydrated. To obtain meaningful H2O data from these hydrated glasses it is therefore necessary to distinguish between the original final dissolved H2O content and the H2O added subsequently during hydration. Since H2O added during hydration is added as molecular H2O (H2Om), and the species interconversion between H2Om and hydroxyl (OH) species is negligible at ambient temperature, the final OH content of the glass remains unaltered during hydration. By using H2O speciation models to find the original H2Om content that would correspond to the measured OH content of the glass, the original total H2O (H2Ot) content of the glass prior to hydration can be reconstructed. These H2O speciation data are obtained using Fourier Transform Infra-red (FTIR) spectroscopy. In many cases OH cannot be measured directly and instead is calculated indirectly as OH = H2Ot - H2Om. Here we demonstrate the importance of using a species-dependent H2Ot molar absorptivity coefficient to obtain accurate H2Ot and H2O speciation data and outline a methodology for calculating such a coefficient for both hydrated and unhydrated rhyolite and andesite glasses. Using this method we present reconstructed final H2Ot contents of hydrated felsic pyroclasts from submarine volcanoes in the Japanese Izu-Bonin Arc and use these data to investigate submarine felsic eruptions and the processes that produce submarine pyroclasts, in order to understand their associated hazards.