Studies on the brittle-ductile transition in silicon

At the brittle-ductile transition (BDT), the fracture mode changes from brittle cleavage to ductile fracture depending upon both temperature and loading rate. In silicon, the fracture toughness increases sharply with a temperature increase of just 5°C, and the low mobility of dislocations is believed to control the transition behavior. If dislocation mobility controls the transition, then the same mechanism must be responsible for both the static transition behavior and the quasi-static crack-arrest behavior. Both behaviors are studied in an attempt to find a unified theory.; The behavior of discrete dislocations near a static crack is studied using numerical simulation to provide an understanding of the dislocation-mobility controlled BDT. Similar models with one or two slip planes have often failed to reproduce the sharpness of the transition curve. The current results show that the sharpness of the BDT in silicon may depend on the number of activated slip systems. This proposed mechanism is consistent with experimental results that show multiple slip planes around the crack tip at the BDT.; The role of dislocations emitted from the crack tip in causing crack arrest is investigated experimentally. Studies using dislocation-free, single-crystal silicon allow the effect of the crack tip to be isolated from all other microstructural features. In the ideal experiment, a quasi-static crack propagates through a temperature gradient. Since the dislocation mobility depends sensitively on the temperature, the crack is arrested at the point where the dislocations become just mobile enough to form a plastic zone and shield the crack from the applied load. It was found that such an experiment is extremely difficult to perform. Nevertheless, the results indicate that such an experiment could be performed with an improved experimental facility.; After a specimen geometry that ensures straight crack growth was found for the experiment described above, research on crack turning was continued. As a hot glass plate is dipped into cool water, a pre-crack may grow along many different paths depending upon the testing parameters. In silicon the crack favors the cleavage planes. A cohesive-element model modified for thermally-driven cracks reproduced many of the experimental crack morphologies.