Understanding of multi-stage R-phase transformation in aged Ni-rich Ti–Ni shape memory alloys

Abstract

An abnormal three-stage martensitic transformation behavior (one-stage R and two-stage B19′) has been found in Ni-rich Ti–Ni alloys when aged at intermediate temperatures (350–500 °C). This strange behavior was successfully explained by considering the kinetics of Ti₃Ni₄ precipitation in polycrystalline samples. However, it was found that R-phase transformation occurs in two stages, when aging is done at low temperatures (250–300 °C). In order to find the origin of this abnormal phenomenon, we made a comparative study with Ni–49.4 at.% Ti single crystals and polycrystals, which were aged at 250 °C for different time. Differential scanning calorimetry showed that all single crystals undergo normal one-stage R-phase transformation, while polycrystals exhibit the abnormal two-stage R-phase transformation after short time aging. These different transformation behaviors can be understood by considering the kinetics of precipitation in supersaturated solid solution. The abnormal two-stage R-phase transformation is attributed to a large-scale heterogeneity of Ni-content between grain boundary region and grain interior, resulted from the preferential precipitation in grain boundary regions. Therefore, it is suggested that there is a unified scenario for seemingly different transformation behaviors of both low-temperature-aged and intermediate-temperature-aged Ti–Ni alloys.

Introduction

Ti–Ni alloys of near-equiatomic composition are technologically important materials because they combine excellent functional properties (shape memory effect, superelasticity) with good mechanical strength and ductility [1]. It is known that near-equiatomic Ti–Ni alloys may undergo three different martensitic transformations depending on thermomechanical treatment condition: B2-R, R-B19′ and B2-B19′, which are the key to the above functional properties and have been extensively discussed in literature [2], [3], [4], [5], [6]. These three kinds of transformations are characterized by different features of transformation strain and hysteresis. B2-R, called R-phase transformation, exhibits small transformation strain (∼1%) and hysteresis (∼2–5 K). In contrast, R-B19′ and B2-B19′ are characterized by much larger transformation strain (∼10%) and hysteresis (∼20–70 K).

Unlike, fully annealed and quenched near-equiatomic Ti–Ni alloys, which transform from B2 to B19′ directly, the aged Ni-rich Ti–Ni alloys normally transform in two stages (B2-R-B19′), and thus, show two distinct peaks on differential scanning calorimetry (DSC) curves, which are well documented in literature [7]. In addition to the generation of R-phase transformation, an abnormal three-stage (one-stage R and two-stage B19′) martensitic transformation in intermediate-temperature-aged Ni-rich Ti–Ni alloys was reported from time to time. The debate about the origin of this three-stage martensitic transformation has been going on for a long time and several explanations were given based on different experimental observations [1], [8], [9], [10], [11], [12]. Recently, by further studies, Fan et al. proved that those preceding mechanisms are not the ultimate origin of this abnormal phenomenon and they attributed it to the inhomogeneous precipitate distribution between grain boundary and grain interior [13]. In addition, three-stage martensitic transformation are identified to be B2-R-B19′ in grain boundary region, and B2-B19′ in grain interior by doing partial differential scanning calorimetry cycles. This scenario successfully explained this abnormal phenomenon and is consistent with all the existing data reported in literature.

As mentioned in the above, both the normal two-stage martensitic transformation and the abnormal three-stage transformation of intermediate-temperature-aged alloys involve only one stage of R-phase transformation. However, an abnormal two-stage R-phase transformation was found in Ti–50.9 at% Ni polycrystals aged at low temperatures (250–300 °C). By investigation with DSC, X-ray diffraction (XRD) and transmission electron microscopy (TEM) techniques, Kim et al. attributed it to the local stress and composition heterogeneity between Ti₃Ni₄ particles [14]. Apparently, this mechanism expects that this strange phenomenon is intrinsic for all Ti₃Ni₄-containing samples and is independent of whether or not the sample is single crystal or polycrystal. As reported in the following, this mechanism is not consistent with the fact that 250 °C-aged Ti–50.6 at% Ni (this will be abbreviated as 50.6Ni, hereafter) single crystals all undergo normal one-stage R-phase transformation in our experiment.

Since, grain boundary plays an important role in the occurrence of abnormal three-stage martensitic transformation, as Fan et al. pointed out, we consider that this abnormal two-stage R-phase transformation may be due to the existence of grain boundary. Based on this idea, we designed the following experiment. At first, low-temperature aging is done with 50.6Ni single crystals, which is the crucial test for the mechanism based on the local heterogeneity of composition and stress between Ti₃Ni₄ particles. Then, the same aging treatment is done with 50.6Ni polycrystals to clarify the effect of grain boundary on this abnormal phenomenon. As can be seen in the following, our experimental results direct to a simple explanation for the abnormal two-stage R-phase transformation in low-temperature-aged alloys. In addition, it is suggested that there is a unified scenario for both the abnormal three-stage martensitic transformation in intermediate-temperature-aged alloys and the above two-stage R-phase transformation in low-temperature-aged alloys.