Temperature memory effect of stress annealing-induced anisotropy in metallic glasses

Stress annealing (SA)-induced magnetic anisotropy is known in iron, nickel, and cobalt-based ferromagnetic metallic glass ribbons and it has already been used in commercial processes. Uniaxial elastic strain is introduced by SA and is quenched into the ribbons even after cooling and removing the external stress. The release of the uniaxial quenched strains is clearly observed as an anomaly in the linear thermal-expansion coefficient (LTEC) when the ribbon is reheated without stress. The rate of strain release corresponding to the LTEC anomaly reaches a maximum at the temperature at which the original SA was performed. We have observed this temperature memory effect over the whole temperature range from 280 to 400 °C, which is below the crystallization temperature $T_x$. The observed results are explained well by the existence of a localized “flow unit” embedded in an elastic matrix, which is accepted as the origin of the shear band formation and rejuvenation of metallic glasses with $T_g$ (glass transition temperature) $<T_x$. Although $T_g$ is hardly visible in calorimetry measurements in most of the ferromagnetic metallic glass ribbons ($T_g>T_x$), the results here indicate that the same important structural feature is common to metallic glasses with both $T_g<T_x$ and $T_g>T_x$. Because magnetization behavior is very sensitive to the existence of residual elastic strain which is difficult to evaluate in most of the metallic glasses, detailed studies and a revival of interest in ferromagnetic ribbons will help us to understand more about the nature of the localized flow unit, as well as nonaffine deformations.

Figure 1

(a) Linear thermal expansion (LTE) and (b) LTE coefficient (LTEC) for the different SA temperatures. Arrows in (a) indicate the starting point of the anomaly corresponding to shrinking in the ribbon direction ($=$ tensile stress direction of SA).

Figure 2

Relation between SA temperatures and peak temperatures in LTEC. The difference in colors correspond to two experiments with different atmosphere (blue: in air; red: under ${\mathrm{N}}_2$ flow). The dotted line corresponds to the case that peak temperature perfectly agree with the SA temperature.

Figure 3

The transmission XRD profiles around the halo top position for the SA temperatures of (a) 280 °C, (b) 320 °C, (c) 350 °C, and (d) 400 °C. Solid and dotted curves correspond to the profiles measure with $q$ parallel and perpendicular to the tensile stress direction of SA, respectively.

Figure 4

The degree of structural anisotropy determined by transmission XRD and amount of release strain determined as the area of the peak in the LTEC. The integrals are calculated up to 50° above the peak temperatures.

Figure 5

Kissinger plots of the peak position in LTEC for the ribbons treated at different SA temperatures.

Figure 6

Models for the mechanism of the temperature memory effect. Panels (a)–(d) correspond to the lower SA temperatures at which the small size of flow unit is the source of the temperature memory effect. Panels (e)–(g) show that the larger flow unit corresponds to the source of temperature memory effects at higher SA temperatures. In (a)–(g), atoms inside the flow unit are shown in red and atoms in the elastic matrix are shown in blue. Dashed green lines are eye guides for the elastic strain, i.e., the distance between 3 atoms and the line 1-2 is perpendicular to the external stress. The atoms in blue can move only elastically through the SA, while the atoms in red have high mobility when ribbons are heated for SA. The dotted green curves show the original (before SA) shape of a cage formed by the innermost atoms in the elastic matrix. The solid curve shows the elastically deformed cage produced by the external stress (b), (c), (f) or the accommodation of atoms inside the cage (d), (g).

Figure 7

Overall view of the ribbon during SA. (a) The region starts to interconnect to adjacent flow units and it becomes a source of permanent deformation. Panel (b) represents the ribbon at a medium SA temperature at which the small flow unit starts to interconnect and the atoms in the bigger flow unit (shown in light red color) have not achieved high mobility. Only those units which have the proper size (isolated and shown in a deep red color) have enough mobility and accommodate to the elongated cage without changing the number density of atoms inside the cage. As a result, the elastic matrix surrounding the red flow unit has quenched strain. They are the source of temperature memory effect at medium SA temperatures.