The recent and on-going eruption of the Soufriere Hills Volcano in Montserrat has been producing different non-predictable eruptive styles that are supposedly determined within the conduit as magma ascends. Indeed, magma degassing as an open system leads to quiescent extrusions, while volatile retention may generate explosive eruptions. We investigate the magma ascent-degassing relations at shallow depths through decompression (P) experiments under controlled conditions of pressure (P), temperature (T), oxygen fugacity and time.
The prerequisite step for such a study is the establishment of phase equilibria that provide the parageneses and phase compositions at equilibrium (constant P, T, fO2). The starting material is a synthetic glass powder representing the rhyolite composition of the matrix of Soufriere Hills andesite (Shill2: 75.2 SiO2, 13.6 Al2O3, 1.9 FeO, 4.2 Na2O, 1.7 K2O, 0.3 MgO, 0.2 MnO, 2.4 CaO, 0.3 TiO2 wt%), which is sealed together with excess water in gold capsules. Experiments are performed at 860°C, NNO+1, for 7 days at P between 1700 and 150 bars in rapid-quench vessels. Water contents for the melts are determined by infrared spectroscopy and electron microprobe (using the difference method and glass standards for which water contents were previously determined by Karl-Fischer titration). The compositions of the crystallized phases (Plag + Opx + Mt + Glass) are determined by electron microprobe, volume proportions and porosities are determined by image analysis. The results of the phase equilibria experiments are summarized in Fig. 3.6-1.
For the P experiments, the water added to the starting glass powder corresponds to the solubility determined in the phase equilibria experiments, so that we have direct quantitative information on degassing through measurement of porosity. The experiments consist of two steps: 7 days at constant P followed by a controlled P and a rapid quench. Run #D3 was decompressed from 1500 to 500 bars within ~4 days, simulating a magma ascent rate of ~1 cm/s. Both melt water content (2.6 +0.4 wt%) and porosity (39 +5 vol%) are in agreement with the equilibrium values expected at 500 bars. However, the matrix glass (77.5-79.5 wt% SiO2), plagioclase (An49.0-36.8), and orthopyroxene (En53.0-49.2) have compositions typical for the P-range between 1000 and 500 bars. These results suggest that this slow P rate does not inhibit the degassing process (water exsolution and bubble growth are completed), while it prevents compositional re-equilibration at 500 bars for some of the phases. Run #D1 was decompressed from 1500 to 500 bars within 12 min, simulating a magma ascent rate of ~4 m/s. The results on melt water content and porosity suggest that this fast P rate does not inhibit the degassing process. However, the compositions of the glasses (75.8-76.4 wt% SiO2), plagioclase (An49.3-43.7), and absence of orthopyroxene indicate that compositional re-equilibration of phases <1000 bars and the nucleation of orthopyroxene are inhibited.
Further P experiments are currently being performed, notably from 1500 to 500 bars within less than 12 min, aiming at inhibiting the degassing process and within up to 1 month, in order to reach complete compositional equilibration of the phases. A set of runs with decompression from 500 to 150 bars at comparable P rates is being run in parallel.