A half-shear-half-shuffle mechanism and the single-layer twinning dislocation for [formula omitted] mode in hexagonal close-packed titanium

Abstract

Among the twinning modes in hexagonal close-packed (HCP) metals, the mechanism for {112¯2}〈112¯3¯〉 mode is particularly confusing and controversial. In the literature reports, there are three possible second invariant planes, i.e. the K2 planes for {112¯2}〈112¯3¯〉 twinning mode: {112¯4¯} which has been widely accepted and corresponds to a three-layer zonal twinning dislocation; {112¯2¯} that is deemed unfavorable; and (0002) which has only been observed in atomistic simulations and corresponds to a single-layer twinning dislocation. {112¯4¯} was predicted by classical twinning theory and the experimentally measured magnitude of twinning shear s in titanium and zirconium seemed to agree well with the prediction. However, {112¯4¯} has never been verified in simulations which show that (0002) should be the K2 plane. This conflict has not been resolved due to the lack of experimental observation of the structure of twinning dislocations. In this work, scanning transmission electron microscopy (STEM) observations are conducted to resolve the twin boundary structure in deformed pure titanium on the atomic scale, combined with atomistic simulations. Atomic resolution STEM unambiguously shows that the twinning dislocation only involves a single twinning plane and the K2 plane is (0002), which is consistent with the atomistic simulations. The STEM results also reveal a half-shear-half-shuffle process which is manifested by a unique twin boundary structure generated by the glide of single-layer twinning dislocations. To explain these results, the lattice correspondences of all three K2 planes are examined in great detail. In particular, shear and shuffle required in the lattice transformations are analyzed inside the framework of classical theory. These analyses explain well why (0002) is the more favorable K2 plane than {112¯4¯} and {112¯2¯}, and properly resolve the conflict between the prediction of the classical twinning theory and the simulation results.