PACROFI VI - Electronic Program

A link between melt inclusion studies and experimental petrology: exemplified by investigations on granitic rocks from the Erzgebirge, Germany

Rainer Thomas

GeoForschungsZentrum Potsdam, D-14473 Potsdam (Germany)

The exact determination of the solidus temperature (beginning-of-melting) of melts preserved in melt inclusions takes up a great deal of time. Quenching experiments on silicate melt inclusions in minerals of granitic rocks offer an alternative, rapid way to estimate solidus temperatures using experimental data from the haplogranite system Qz-Or-Ab-H2O. Together with the determined water content we can calculate a model solidus using a lndK-1/T relationship (Thomas, 1994). Here, dK is the critical diameter which is obtained as follows: after quenching from any temperature above the true homogenization temperature TH, the smallest melt inclusions will be completely homogenized, larger ones not, and some of intermediate size are just on the verge of homogenization. The latter are the "critical" inclusions and their size (critical diameter dK) is carefully measured with a calibrated micrometer ocular. In practice, values for dK at different run temperatures are collected for each sample and plotted on a log scale against run temperature (1/T). The linear extrapolation of the general equation

to a dK value of 1 micron (lndK = 0) gives us the temperature T0. Using the algorithm for the calculation of viscosity from inclusion and chemical data we can calculate the critical inclusion diameter at different temperatures from the bulk chemistry and different water concentration values CW. For water concentrations CW between 2 and 10 wt.% T0 varies from about 500 to 900oC. As an example of the method, we obtain the following relations of T0, TS and CW using inclusion data from the Kirchberg OG3 granite (Erzgebirge):

From the main inclusion group in the OG3-sample a water content of 6.7 +/- 0.1 wt.% was determined. Using equations (2) and (3) T0 is 632oC and TS is 663oC. According to Johannes (1992), we obtain a pressure at TS of 2.8 kbar.

This method connects the results of experimental petrology with melt inclusion data and therefore forges a link between geology and experimental petrology.