SEARCH

SEARCH BY CITATION

Keywords:

  • deformation;
  • deformation mechanisms;
  • granite;
  • microstructures;
  • partial melt

Abstract The effects of varying amounts of partial melt on the deformation of granitic aggregates have been tested experimentally at conditions (900°C, 1500 MPa, 10-4 to 10-6/s) where melt-free samples deform by dislocation creep, with microstructures approximately equivalent to those of upper greenschist facies. Experiments were performed on samples of various grain sizes, including an aplite (150 μm) and sintered aggregates of quartz-albitemicrocline (10–50 and 2–10 μm). Water was added to the samples to obtain various amounts of melt (1–15% in the aplite, 1–5% in the sintered aggregates). Optical and TEM observations of the melt distribution in hydrostatically annealed samples show that the melt in the sintered aggregates is homogeneously distributed along an interconnected network of triple junction channels, while the melt in the aplites is inhomogeneously distributed.

The effect of partial melt on deformation depends an melt amount and distribution, grain size and strain rate. For samples deformed with ˜ 1% melt, all grain sizes exhibit microstructures indicative of dislocation creep. For samples deformed with 3–5% melt, the 150 μm and 10–50 μm grain size samples also exhibit dislocation creep microstructures, but the 2–10 μm grain size samples exhibit abundant TEM-scale evidence of dissolution-precipitation and little evidence of dislocation activity, suggesting a switch in deformation mechanism to predominantly melt-enhanced diffusion creep. At natural strain rates melt-enhanced diffusion creep would predominate at larger grain sizes, although probably not for most coarse-grained granites.

The effects of melt percentage and strain rate have been studied for the 150 μm aplites. For samples with ˜ 5 and 10% melt, deformation at 10–6/s squeezes excess melt out of the central compressed region allowing predominantly dislocation creep. Conversely, deformation at 10-5/s produces considerable cataclasis presumably because the excess melt cannot flow laterally fast enough and a high pore fluid pressure results. For samples with 15% melt, deformation at both strain rates produces cataclasis, presumably because the inhomogeneous melt distribution resulted in regions of decoupled grains, which would produce high stress concentrations at point contacts. At natural strain rates there should be little or no cataclasis if an equilibrium melt texture exists and if the melt can flow as fast as the imposed strain rate. However, if the melt is confined and cannot migrate, a high pore fluid pressure should promote brittle deformation.