Fig. 9 Dislocation distribution in the model under monotonic loads. Loading directions are along (a) the y -axis and (b) thex -axis, respectively. The initial structure is provided for comparison in (c). Green and magenta lines in bcc-Fe matrix correspond to 1/2<111> and <100> dislocations, respectively. Gray planes in bcc-Fe matrix are defect meshes. Red dashed boxes contain defect meshes with the same orientation.
Cyclic softening and elastic-plastic transformation in Fig. 6 are closely related to the evolution of dislocations during cyclic shear deformation. Fig. 10 shows the dislocation distribution in the model underlying cyclic deformation. Fig. 10a-c correspond to the loading conditions of d1, d2 and d3 respectively, and the loading direction is along the y -axis. Fig. 10a shows that dislocations begin to appear at the end of the third cycle, and then increase and remain at a low level in the following cycles. Fig. 10b shows that dislocations begin to appear at the end of the first cycle. When the load is d3, a large amount of dislocations and defect meshes exist in the bcc-Fe phase at the end of the first cycle. The proportion of irregular structures is large, and the quantity of dislocations and defect meshes remains high throughout the whole cyclic loadings. The dislocation evolution under these three conditions corresponds to the variation of dislocation density shown in Fig. 7. The defect meshes shown in the red dashed boxes in Fig. 10c have the same orientation, which is similar with the results shown in Fig. 9b, indicating that there exists a lot of residual strain in the model. This also points out that the material has undergone severe plastic deformation and damage accumulation, and the cementite phase tends to be damaged.20 Note that there are more defect meshes in Fig. 10c than those in Fig. 9a, indicating that cyclic shear deformation is more destructive to the structure than monotonic shear deformation, and is more likely to result in fatigue failure of the material.
Fig. 10d-f correspond to the loading conditions of d1, d2 and d3 respectively, and the loading direction is along the x -axis. When the load is d1, the interfaces are free of dislocations at the end of the third cycle, after which the dislocations increase and remain at a low level. When the load is d2, dislocations begin to appear at the end of the first cycle. It is noted that under these two loading conditions, most of the dislocations in the model are annular, germinating from the interface and eventually remaining in the bcc-Fe phase. When the load reaches d3, dislocations and obvious defect meshes are generated at the end of the first cycle. In the subsequent cycles, although there are still many dislocations in the model, no obvious defect mesh could be found on the interface. The less residual shear strain in the model can be attributed to that the direction of the cyclic loadings is parallel to the interface. This indicates that the model produces less plastic accumulation, or the rate of plastic accumulation is slower, which may result in a relatively longer fatigue life of the material. It is worth noting that, under cyclic deformation, fewer defect meshes are generated in the model when loads are along the x -axis, which is contrary to the rule drawn from Fig. 9. This suggests that the plastic accumulation mechanism of the material is ascribed to the synergistic effects between the cyclic deformation and the loading direction.