Quantifying the abnormal strain state in ferroelastic materials: A moment invariant approach

Abstract The strain glass transition has been found in many ferroelastic systems, but the microstructural nature of strain glass is still unclear. Here, two-dimensional second and fourth order moment invariants as well as image entropy are used to evaluate the presence of a strain-glass state in simulated microstructure images of a doped ferroelastic system. Four different microstructural states are identified, depending on the doping concentration of point defects and the temperature. The martensitic state is characterized by a broad moment invariant distribution peaking near the invariants for the circle, whereas systems that display the strain glass transition produce distinctly different distributions. The image entropy is found to increase with increasing defect concentration; above the critical defect concentration of 0.1, image entropy becomes nearly linearly dependent on temperature. The image analysis approach is capable of characterizing the range of strain domain shapes that occur in the different microstructural states of a doped ferroelastic system.