Previous page     Contents     Next page

Explosive silicic volcanism is controlled by a complex interplay of physico-chemical properties, leading to a wide range of eruptive features. Dome explosions range from low-energy, coarse-grained pyroclastic flows to high-energy, fine-grained surges. We are determining the parameters that control fragment size through an experimental study using a fragmentation bomb apparatus. The fragmentation bomb simulates magma fragmentation under rapid decompression and allows recovery of the particles for analyses. We first hydrate a haplogranite powder (HPG8) containing Al₂O₃ crystals of ˜350 µm diameter (20, 35, 45 vol%) with 1 to 6 wt% of water in an internally heated pressure vessel. The hydrated, crystal-bearing glass cylinders, 20 mm long and 20 or 8 mm in diameter, are placed in the vertical autoclave of the fragmentation bomb. At temperatures of 600 or 800°C and pressures up to 30 MPa, water exsolves from the melt, so that the pre-fragmentation sample is a crystal- and vesicle-bearing melt. Temperature and pressure are held constant for 30 minutes and the subsequent rapid decompression of the sample is induced by the disruption of the three steel or copper diaphragms on the top of the autoclave. If the strength of the sample is lower than the applied stress, fragmentation occurs. The fragments are collected for determination of the grain size distribution (FSD, by laser particle sizer), porosity, vesicle size and texture (scanning electron microscopy), and glass water content (infrared spectroscopy and Karl-Fischer titration).

The very sharp outlines of the fragments and some glass failure textures suggest brittle fragmentation of the melt. The vesicle textures are different from those of the previous experiments performed on crystal-free samples (see Annual Report 1998). No tube pores are generated (vesicles are round) and a frame of radial and concentric fractures sometimes surrounds the largest bubbles (Fig. 3.6-5). The measured porosity of the fragments is in agreement with the calculated vesicularity, suggesting that even high crystal contents and temperatures as low as 600°C do not prevent bubble nucleation and complete expansion. However, the fracture frames around the pores in some experiments suggest that bubble expansion was temporarily and locally prevented, generating water overpressure in the bubble. This observation is quite interesting because water overpressure in bubbles is one of the hypotheses proposed for the generation of the high-energy fine-grained surges. A further attempt at generating strong water overpressures in pores is currently being performed on microlite-bearing samples (coexistence of the 350 µm and <180 µm Al₂O₃ crystals). FSD results highlight the complex role of crystal size and strength on the generated fragment sizes. For starting glasses containing no fractured crystals, FSDs are bimodal up to 14 MPa (main and secondary modes at crystal-size and 15 µm, respectively), but are progressively restricted to the crystal-size mode at higher pressures. For starting glasses containing fractured crystals, the main mode of the FSD is shifted to smaller sizes (˜100 µm). This implies that in addition to decompression magnitude, melt viscosity, and pore size (see Annual Report 1998), the fragment size of dome-generated pyroclastic flows is related to the strength and size of the dominant mineral phases in the magma.

Fig. 3.6-5: Fragment obtained at 800°C and 16 MPa. Crystal : vesicle: glass = 25:28:47 vol%. Note the fracture frame around the largest bubbles.