Results

In accordance with the experimental observations (Kenkmann et al. 2000) we could reproduce a concentration of plastic work in an oblique zone log | qI slrain}

Fig. 7. Accumulated plastic strain field at the end of the model run. Logarithm of strain is displayed. The occurrence of a zone of concentrated shear strain with correspondingly elevated temperature near the material interface is reproduced by these simulations. Melt production is not observed because of the employed material model.

Fig. 7. Accumulated plastic strain field at the end of the model run. Logarithm of strain is displayed. The occurrence of a zone of concentrated shear strain with correspondingly elevated temperature near the material interface is reproduced by these simulations. Melt production is not observed because of the employed material model.

along the inclined interface, which leads to an additional temperature rise of 400 K above the average temperature in shock-compressed quartzite. This is due to the different compressibility of the materials which causes considerable localised shear at their interface. This temperature rise does not change any more for a finer mesh resolution, thus it reflects no numerical artifact. Shear heating of the material has occurred in its compressed state so that the differences between equilibrium release path (which is assumed in the employed equation of state) and experimentally observed release path (e. g. Ahrens and Rosenberg 1968, Swegle 1990) are not critical for the outcome of the simulations.

In recent additional simulations (Hertzsch 2003) with a slightly changed geometry (in particular with a material interface extending all the way across the sample width) the angle between material interface and shock wave plane has been varied in order to determine its effect on the magnitude of the shear heating. The dependence of the maximum temperature on the angle is shown in Fig. 9. The degree of shear heating is found to depend on the angle between incoming shock wave plane and

Fig. 8. Pressure and temperature recorded by two tracers at one cell distance from the interface (see tracer location in Fig. 6). The arrival of the shock wave reflected from the bottom boundary is shown as a vertical dashed line.
Fig. 9. Maximum temperature during passage of shock front depending on the angle of the material interface to the horizontal (initial shock wave) plane. The most pronounced temperature elevation is found for an angle of 45°.

material interface, and the temperature assumes its highest value for an angle of 45°. However, even with an improved material model, which would allow for melting, it is not possible to judge the amount of possible melt production from such information alone. A thorough comparison with the experimental results of Kenkmann et al. (2000) must be subject of more detailed investigations.

Although we observe temperatures at the material interface which are above those reached due to shock compression of the bulk of the material, the formation of melt could not be reproduced in the present simulations. This is due to the simple material model, which we have used, in which friction decreases with rising temperature and reaches zero at melt temperature. This results in a decrease of the heat production and a limited temperature rise. Consequently, the material cannot reach its melting point in this model. An additional mechanism of energy dissipation like viscosity of the material close to the melting point is needed to explain the experimentally observed melting. This is planned as a next step to improve our numerical model.

In addition to the model improvement the possibility of a thermo-chemical reaction at the quartz/forsterite interface may play a role in the onset of boundary melting observed in experiments by Kenkmann et al. (2000). The Mg-rich melt observed in these experiments may result from the reaction

Mg2SiO4 + SiO2 ^ 2MgSiO3

In this case low eutectic melting is taking place. Comparing melting temperatures for SiO2, forsterite, enstatite and forsterite/enstatite eutectic (Fig. 10, compiled according to Chen and Presnall (1975) and Presnall et al. (1998)) one can see that up to ~30 GPa the eutectic melting occurs ~1000 K below pure forsterite melting and ~2000 K below stishovite melting. Addition a of small amount of water from the sample preparation or from accessory minerals decomposed in the shock wave also may decrease the melting temperature at the mineral interface. Finally, a slightly imperfect fitting of the two rock half cylinders along their polished interface may have produced gaps of a few micrometer width in the experiments. The closure of such a fissure during shock compression causes an additional heat pulse (Heider and Kenkmann 2003) that is not considered here.

Was this article helpful?

0 0

Post a comment