In order to understand the processes leading to certain violent volcanic eruptions, the physical mechanisms of the fragmentation of highly viscous magma require closer investigation. The experimental approach given by Alidibirov and Dingwell (Annual Report 1996) is currently the only experimental method capable of permitting analysis of the influences of pressure and temperature on the products of magma fragmentation at conditions relevant to dome explosions.
Well-constrained rapid decompression experiments were performed on basaltic andesite samples of Merapi´s (central Java) November 1994 pyroclastic flow. The samples were obtained during a fieldtrip in the lower valley of the Kali Boyong to represent the range of textures and densities of rocks exposed in the deposits. In a second field campaign in 1998 we sampled 8 rock varieties of the June 1998 pyroclastic flow for further investigations. The vesicularity of the material used for the experiments covers ranges from 5 vol% to 65 vol% total porosity. The wide textural variety of the basaltic andesite makes Merapi an excellent subject for our techniques.
Experiments were performed in the fragmentation bomb at room temperature and 900°C - 950°C. The initial pressure difference (between lower and upper section) was used to define the fragmentation threshold of the different materials. A series of investigations was performed on the samples before the experiments, firstly to characterise the material and secondly to enable comparison with data derived from fieldwork. The investigations include the analysis of density, water content, chemical composition of the glass (by microprobe), determination of surface area by nitrogen sorption (BET) and a close examination by image analysis (crystal size and vesicle-form-parameter). We use sieving and laser particle sizing to characterise the fragmentation products prior further investigation. Selected grains were characterised in the SEM and polished sections were used for optical image analysis.
Here we wish to focus on a single significant aspect influencing the grain size distributions of experimentally and naturally fragmented samples. In nature a common observation of pyroclastic products of fragmentation is the presence of so-called crystal ashs, where whole phenocrysts are retained in a fine glassy ash matrix. We propose, as a result of the comparison of experimental and field data on grain size distributions, that crystal retention is due to the presence of vesicle-rich zones surrounding phenocrysts that serve to deflect fracture surfaces away from the phenocrysts.
The evidence is of two kinds. Firstly, the SEM observation that intact
phenocrysts are surrounded by the remains of highly vesicular matrix (Fig.
3.7-2). Secondly, the statistical
Fig. 3.7-2: Thin glass rim covering a plagioclase phenocryst in the product of a fragmentation experiment. The remains of a highly vesicular zone around the phenocryst are interpreted as having deflected fracture surfaces around single phenocrysts.
comparison of grain size distribution data between experimental and
field samples. We compared particle distribution curves derived by matrix
analysis (Fig. 3.7-3) of Merapi November 1994 pyroclastic flow (particles
>32 mm) to distribution curves obtained from decompression experiments
(Fig. 3.7-3). Both analyses show the characteristic second peak produced
by the preservation of crystal grains.
Fig. 3.7-3: Typical particle size distribution curves for experimental (A) versus field (B) samples of pyroclastic flow matrices. The characteristic second peak at Phi = 1 reflects phenocryst retention. b) Typical experimental data for the particle distribution curve from experiments with highly porous basaltic andesite (Phi = -log2 d, where d is the particle diameter in mm).
Further investigations are planned to test whether the natural fragmentation generating Merapi´s pyroclastic flows is a high or low energy event. Fragmented phenocrysts may demonstrate the existence of a high energy decompressive fragmentation wave.