Influence of dislocation structure on two-way shape memory effect in a CuZnAlMnNi alloy

Abstract

Repeated rounds of “training-aging” treatment is a new technology to improve the two-way shape memory effect (TWSME) and the stability of Cu-based shape memory alloy (SMA), during which the influence of dislocation structure on TWSME has been studied by transmission electron microscopy (TEM). It is found that different dislocation structures produce a different influence on TWSME. Disorderly distributing dislocations promote the pseudo-stabilization of the parent phase, which decreases the amount of parent phase participating in the transformation and finally results in TWSME degradation during aging, but the stabilization can be eliminated through re-training. The stable dislocation arrays distributing orderly contribute to improving TWSME and obtaining stable TWSME. Compared with NiTi-based SMAs, Cu-based SMAs have many advantages, such as low cost, simple processing, excellent electric and thermal conductivity, which are favorable to their wide use. The application of Cu-based SMAs is generally associated with the TWSME achieved by thermomechanical training. It is known that dislocations will be produced when an SMA sample is subjected to training under applied stress. Many researchers have studied the effect of dislocations on TWSME [1–12], but there is controversy. Some of the researchers believed that the formation of preferential thermoelastic martensite was the origin of the TWSME, and that the preferential dislocations and retained martensite generated during training were favorable to the formation of preferential martensite variants [1–6]. The others regarded the dislocations introduced during the training process as the origin of the TWSME, and the arrays of dislocation and residual stress would favor the nucleation and growth of some preferential martensite variants, and thus resulted in TWSME [7–12]. In spite of a large amount of literature on Cu-based SMAs, the application of the alloys is much less than that of NiTi-based alloys, because of the instability of SME and transformation temperatures during practical applications [13]. The degradation of TWSME in Cubased SMAs during aging at different temperatures was associated with two mechanisms, first, with the annihilation of dislocations and next, with the precipitation of an α phase [11, 14, 15]. The degradation of TWSME during thermal cycling was reported to be caused by the increment of dislocation and martensite stabilization [16–18]. So far, the problem of the instability of Cu-based SMAs has not been solved satisfactorily. However, on the basis of repeated experiments, we found [19] that repeated rounds of “training-aging” treatment evidently improved the TWSME and the stability of Cubased SMAs. The stable shape recovery rate reached 99.7% or so after four rounds of “training-aging” treatment. We also found that different dislocation structures will produce a different influence on TWSME of a CuZnAlMnNi SMA, as will be reported below. The alloy used in this paper was prepared from 99.99% Cu, 99.95% Zn, 99.7% Al, 99.9% Mn and 99.5% Ni by melting using a graphite crucible in an induction furnace. The ingot was homogenized at 850 ◦C for 12 h and then hot-rolled into sheets with thickness about 1 mm. The strip samples cut from the sheets which were of the size of 200× 4× 1 mm, were solution-treated at 840 ◦C for 20 min, quenched into boiling water for 30 min and then cooled to ambient temperature in normal atmosphere. The transformation temperatures of the alloy with composition Cu-23.6Zn4.47Al-0.23Mn-0.17Ni (in mass %) were as follows: Ms= 20 ◦C, Mf= 1 ◦C, As= 27 ◦C, Af= 42 ◦C. The sample was uniformly fixed by bending up into a “U” shape on a mound with a diameter of 78 mm, and conducted thermal cycling between boiling water and 18 ◦C water. When the shape memory amount of the sample no longer changed with the number of training cycles, the first round of training was ended. The sample was then placed into boiling water for 8 h, and thus completed the first round of “training-aging” treatment. Subsequently, the as-aged sample was subjected to the second round of training until the stable shape memory amount was achieved again, and then the sample was placed, without any constraint, into boiling water for 8 h, which finished the second round of “trainingaging” treatment. The process, as above stated, was repeated. This repeated treatment process, conducting alternate training and aging, was called the “trainingaging” treatment. This is a new technology for improving the SME and its stability of Cu-based SMAs as reported in our previous paper [19]. The variation of dislocation structures during the treatment was observed on an H800 transmission electron microscope, and specimens were jet-polished with 33.3% nitric acid and 66.7% methanol solution. As can be seen from Fig. 1a there appear a large number of dislocations after the first round of training, because the training process itself is a thermal cycling process under applied stress during which the dislocations multiply with increase in the number of training cycles [18]. The dislocations which are distributing