| dc.description.abstracteng | Until today, the explosivity of volcanic eruptions cannot be forecasted by even the most modern monitoring techniques. The processes causing these events, endangering the lives of millions of people, happen deep inside the magma plumbing system. Some of these explosive eruptions are characterized by melts with high CO2 contents that originate from CO2 flushing directly from the Earth mantle or magma plumbing systems being rooted in carbonate substrata. In the latter case, CO2 is produced by interaction of the magma with the carbonate basement leading to carbonate break-down and assimilation. Making use of high pressure high temperature experiments, this PhD project focuses on reproducing of magma storage conditions and simulating magma ascent processes to improve the knowledge on the role of CO2 for volcanic degassing mechanisms and their effect on explosive volcanic eruptions.
The first study of this project focusses on the solubility of mixed CO2-H2O
volatile phases in K-rich leucititic and phonolitic melts as a function of pressure to shed light on the storage conditions of those melts in the Earth’s
crust. Synthetic melt analogues to the composition of the Pozzolane Rosse
eruption of the Colli Albani volcanic District 456 ka ago (SULm) and the
famous 79 AD Mt. Somma-Vesuvius eruption (VES79) were used. Solubility experiments were carried out in an internally heated pressure vessel
(IHPV) at 1250 °C a constant oxygen fugacity of NNO +3 ±1 and pressures
between 50 and 300 MPa for fluid compositions of XflH2O = 0.0, 0.2, 0.4, 0.6, 0.8, 1.0. It was found that the highly depolymerized SULm melt is capable of dissolving a very large amount of CO2 (up to 8500 ppm at 300 MPa) that is five times higher than in the polymerized VES79 composition at similar pressures. At the same time H2O solubility is fairly similar in both compositions throughout the studied pressure range. Furthermore, evidence for a
depolymerizing effect of H2O on the VES79 melt was found, leading to an
elevated CO2 solubility in CO2-H2O bearing samples.
In the second study we experimentally investigated the decompression induced volatile exsolution of mixed CO2-H2O bearing SULm melts that occurs during magma ascent through the crust. In an IHPV at 1250 °C, the
pressure was decreased from 200 MPa to about 150, 100, 50 and 30 MPa at
a fast, constant decompression rate of dP/dt= 1 MPa/s for sample series
with XflH2O = 0.0, 0.5, 1.0. After rapidly quenching the samples to room temperature, optical analysis was done to obtain parameters important for interpretation of the melt degassing behavior such as vesicle number density (VND) and porosity. We successfully implemented a new procedure to enable porosity measurements of the entire sample capsule matching calculated porosities very well. To preserve the entire capsule, we used a diamond wire saw to cut the closed capsule in half with minimal damage to the fragile sample. One half of the sample is embedded in epoxy and polished before taking an image in reflected light. To improve the statistics for porosity and VND measurements, we polished the sample down by approx. 500 μm while taking a total of three pictures at different depths. We demonstrate that the exsolution mechanisms of H2O strongly differ from those of CO2. In contrast to the single homogeneous nucleation event of H2O bubbles, CO2 bearing melts tend to continuously nucleate new bubbles after surpassing the critical supersaturation pressure. In CO2-H2O bearing melts we found that CO2 degasses exclusively at higher pressure before H2O exsolves from the melt, rapidly decreasing the volatile supersaturation at low final pressure. We propose that this causes increased magma ascent rates at shallow depths (where H2O exsolution starts) leading to enhanced volcanic explosivity.
The high CO2 concentrations in the SULm melts are at the limit of analytical capability of the conventionally used infrared-spectroscopic technique in transmission. Here the peakheights of the carbonate doublet easily exceed two absorbance units (depending on the sample thickness). Since the dependence of the peakheight on the concentration is not linear in such a
case, the carbonate content of these samples cannot be quantitatively evaluated. To account for this problem we implemented a novel application in the
third study, using micro Attenuated Total Reflectance (μ-ATR) FTIR Spectroscopy to quantify high CO32− contents (> 0.17 wt%) in silicate glasses. μ-ATR analyses only a very restricted sample volume (< 10 μm depth).
Here the carbonate signal shows absorbances that are roughly two orders
of magnitude lower compared to transmission measurements. The spectra
have shown to nicely separate the CO32− doublet from the SiO2 lattice vibration allowing for a simple straight baseline. A linear correlation coefficient of k= 2.28 ± 0.02 was determined for the studied SULm composition with CO2 contents between 0.17 and 4.28 wt%. | de |