Previous page     Contents     Next page

Viscosity plays an important role in the rheological behavior of silicate melts. Thus, a better understanding of viscosity is required in order to predict features of magma ascent or eruption. Because of the pressure dependence of viscosity, high pressure viscosity data are needed for applications to geological models. Here, we have developed a new dilatometer (see Fig. 3.8-5) easily inserted into an internally heated pressure vessel (IHPV) in order to study melt rheology. In particular, we are interested in determining the viscosity of supercooled silicate melts at elevated pressures.
 

Fig. 3.8-5: Cross-section of the apparatus.

Our set-up uses the micropenetration method (for further details see Annual Report 1996) and works most effectively in the viscosity range 10⁸ < < 10¹¹ Pa s permitting measurements at conditions that are relevant to the Earth´s crust (P <= 10 kbar, T <= 1000° C). This new apparatus has now been calibrated at 1 bar and runs successfully up to 2 kbar. It will be further tested in order to extend the pressure range to 10 kbar. The advantages of the micropenetration method are the ability to use small samples and a simple sample geometry compared to other high viscosity methods such as parallel plate or fiber elongation. The use of small samples minimizes thermal gradients across the sample.

For calibration, we measured the viscosity of the DGG-1 standard glass from the "Deutsche Glastechnische Gesellschaft". All cylindrical samples were polished (3 mm in height, 4 mm in radius), heated at a rate of 10 K/min and equilibrated at the measuring conditions for 1 hour to ensure complete structural relaxation (Fig. 3.8-6).
 

Fig. 3.8-6: The measured viscosities determined using the micropenetration method for the DGG-1 standard glass (circles) and the standard viscosity-temperature relationship from the "Deutsche Glastechnische Gesellschaft" (line). Error bars are represented by the size of the circles.

This dilatometer has now been precisely calibrated for room-pressure measurements. In the range of 10⁸ - 10¹¹ Pa s the deviation is low and nearly constant. The viscosity data have been reproduced within an error of ± 0.06 log units. In several high-pressure experiments this set-up has run successfully up to 2 kbar. Our current aim is to carry out measurements up to 10 kbar. Further investigations are required to construct a complete calibration curve that will enable viscosity determinations of volatile bearing silicate melts under geologically relevant conditions to be made.