Introduction

Shape memory alloys (SMAs) such as nickel-titanium (NiTi) have a special capability to recover large strain induced by mechanical deformation. This capability is sufficiently exploited in many SMA applications (e.g., pipe couplings). As is well known, the TiNiFe alloy pipe coupling is widely used in aviation industry. Due to the low martensitic transformation temperature and narrow thermal hysteresis, the TiNiFe pipe coupling must be stored in liquid nitrogen. Many recent researches show that NiTiNb alloys have good mechanical properties, suitable martensitic transformation temperature, and very wide thermal hysteresis (Ref 1, 2). So, if pipe couplings can be made of NiTiNb alloy, it will be stored and transported conveniently at room temperature. And many researchers have started to exploit the potentiality of NiTiNb alloys (Ref 3-5).

Like other NiTi-based alloys, the recoverability of NiTiNb alloy is fulfilled when martensite variants return to the austenite. During the phase transformation, monoclinic martensite phase (B19′) with low symmetry converts to high-symmetry body-centered cubic (bcc) austenite phase (B2) (Ref 6). Therefore, different lattice orientations lead to distinct recoverability. Several experiments focusing on single crystal have shown that different martensitic transformation behaviors depended on crystallographic orientations (Ref 7-9). Following studies showed that textures played similar roles in processed polycrystalline NiTi alloys (Ref 10, 11). Miyazaki et al. examined the textures of cold-rolled NiTi thin plates, and discussed quantification of the relationship between transformation strain and crystal orientation (Ref 12). Texture and shape memory property of melt-spun TiNiCu ribbons (Ref 13) and hot-rolled TiNi bars (Ref 14) were also investigated.

In these foregoing studies, they usually focused on isolated relationships between crystallographic textures and transformation behaviors of single deformation processed NiTinol materials. In practice, deformation processes such as forging, drawing, and rolling are always used in combination, and the main object of deformation processes is to obtain appropriate microstructure and material size. Furthermore, most previous researches concentrated on thin sheets (Ref 12, 15, 16) or bars (Ref 17) rather than tubes which were widely used in engineering. In this paper, three different processes were investigated, and the study focused on understanding the fundamental physical links between processing, microstructure, and relevant properties. The notable difference of recoverability in two contrasting tube process flows was also observed and discussed.

Experimental Procedures

Materials and Specimens

Consumable electrode arc melting was used to produce Ni47Ti44Nb9 (at.%) ingot with a diameter of 200 mm, and then the ingot undergone cogging and multi-direction forging with 50% reduction per pass at 950 °C. Then we got 50-mm diameter rods (denoted as: original NiTiNb materials). After sampling, these rods were divided into two batches. With 20% reduction per pass at 850 °C, one batch of rods was hot-forged down to 12-mm diameter rods (denoted as: hot-forged NiTiNb materials). Those 12-mm diameter rods were cold-drawn, then annealed at 650 °C for 0.5 h, and finally cold-drawn down to 8-mm diameter bars (denoted as: cold-drawn NiTiNb materials). While another batch, with a pre-manufactured hole, was processed by cold three-roll rolling processes with intermediate annealing, and then obtained tubes with an external diameter of 8 mm and a wall thickness of 2 mm (denoted as: cold-rolled NiTiNb materials). The specimens were cut by spark cutting and heat treated at 850 °C for 1 h in an evacuated quartz tube. The normal direction (ND) of x-ray diffraction (XRD) and EBSD test zone paralleled the axial direction of rods or tubes.

Test Methods

Specimens for XRD were mechanically polished and then for tested by a Philips X’ Pert MRD XRD machine. Cu Kα X-ray was used at 40 kV, 40 mA, and with a scanning range of 20°-85°. The XRD pattern of NiTiNb alloy was shown in Fig. 1. Texture data were obtained from {110}, {200}, and {211} planes in the B2 parent phase and characterized in the form of orientation distribution functions (ODFs).

Fig. 1
figure 1

XRD diffraction patterns of hot-forged NiTiNb alloy with β-Nb and austenite (B2) peaks identified

Full size image

The mechanical polished specimens were then etched using a solution consisting of 10 mL HF and 30 mL HNO3 and 100 mL H2O for 30 s. This etchant preferentially attracts on grain boundaries which consisted by eutectic structure, but does not result in their crystallographic grain boundary. So the metallography calculated from EBSD (Fig. 2) is different to optical micrography (Fig. 3c).

Fig. 2
figure 2

Three-color EBSD map of cold drawn materials

Full size image
Fig. 3
figure 3

Optical metallography of Ni47Ti44Nb9 alloy (a) original materials, (b) hot-forged materials, (c) cold-drawn materials, (d) cold-rolled materials

Full size image

Specimens for electron backscatter diffraction (EBSD) were mechanically polished followed by electrolytic polishing. The electrolyte solution was a mixture of 25% HNO3 and 75% methanol. The electrolytic polishing was conducted at 18 V in electrolyte solution with a temperature of −40 °C (Ref 18). A JEOL 6700 field emission SEM equipped with a backscattering diffraction detector was used for EBSD test. Crystallographic orientations were obtained by indexing of Kikuchi patterns with TSL-EDAX software. For convenient analysis, the ND of EBSD test zones paralleled the axial direction of pipe couplings, while the transverse direction (TD) paralleled along the circumferential direction.

The shape memory property of cold-drawn and cold-rolled materials was characterized by inner diameter recoverability of corresponding pipe couplings. The pipe couplings, with an inner diameter of 4 mm, a wall thickness of 1.8 mm, and a length of 16 mm, were machined by cold-drawn ribbons or cold-rolled tubes. Those pipe couplings were heat treated at 850 °C for 1 h in an evacuated quartz tube, and then were enlarged at −60 °C with an expanding ratios of 1.16. After the inner diameter (Φ1) was measured, the pipe coupling was heated to 100 °C for 5 min in water bath to obtain a completely inverse martensitic transformation. Then the inner diameter (Φ2) was measured again. The recoverability (η) of pipe coupling was calculated according to the following formula: η = (Φ1 − Φ2)/Φ1.

Results

Microstructure of Deformed Materials

The optical metallography of NiTiNb alloy in different deformation states (original materials, hot-forged materials, cold-drawn materials, and cold-rolled materials) were shown in Fig. 3(a-d). According to the XRD for phase identification as shown in Fig. 1, the microstructure was consisted of TiNi matrix phase that was bright in the micrography and eutectic structure comprising β-Nb particles that had been etched off (Ref 19, 20). The grain size of original and hot-forged NiTiNb material was about 10-20 μm, while both cold-rolled and cold-drawn NiTiNb material was about 5-10 μm. The eutectic structure displayed a similar change. Cold-working material (cold-drawn and cold-rolled material) has much finer and more dispersed eutectic structure than hot-working (original and hot-forged material) material. Otherwise, only the original material exhibited dendritic microstructure in the center of rods, which was circled in Fig. 3(a).

Texture After Deformation

Figure 4 showed that the original materials possessed mixed fiber texture of 〈113〉 (5-6 times random) and 〈331〉 (1-1.5 times random). The texture of hot-forged materials was clear fiber texture in 〈113〉 direction and it was 15 times random as shown in Fig. 4(b).

Fig. 4
figure 4

ODF sections and the ODF section at φ2=45° of (a) original NiTiNb materials and (b) hot-forged NiTiNb materials, revealing both NiTiNb materials contained a 〈113〉 fiber texture

Full size image

The inverse pole figures (IPFs) of cold-drawn and cold-rolled materials were shown in Fig. 5, which were calculated by EBSD. In Fig. 5(a), there was high density preferred orientation in 〈111〉 district of ND IPF and 〈110〉 district of TD IPF, which indicated that 〈111〉 crystal directions paralleled to axial direction and 〈110〉 crystal directions paralleled to circumferential direction of cold-rolled tubes. Compared with cold-rolled materials, the density of 〈111〉 texture was higher in the cold-drawn specimens (as high as 18 times random) in ND IPF, but the preferred orientation was weaker and undisciplined in TD IPF as shown in Fig. 5(b). It means that cold-drawn tubes contained stronger 〈111〉 texture paralleled to axial direction and weaker 〈110〉 texture paralleled to circumferential direction than cold-rolled tubes.

Fig. 5
figure 5

IPF of the (a) cold-rolled and (b) cold-drawn NiTiNb materials, representing the preferred crystalline orientation in the axial direction (ND of specimen) and circumference direction (TD of specimen)

Full size image

Recoverability of Materials

The recoverability of 50 pieces cold-drawn and 50 pieces cold-rolled pipe couplings were measured. And the average recoverability (η) of cold-drawn pipe coupling was 7.1%, while the cold-rolled pipe coupling was 8.4%, as showed in Fig. 6. With uniform size and the same expanding diameter process, this result reflected that cold-rolled materials had higher circumferential recoverability than cold-drawn materials.

Fig. 6
figure 6

Recoverability of cold-drawn and cold-rolled pipe couplings

Full size image

Discussions

The experimental results showed the effects of deformation processes on microstructure, texture, and recoverability. The relationship of processing-microstructure-property is discussed as following.

Effects of Deformation Processes on Microstructure

Some studies show the grain size and metallography of SMAs with different deformation processes (Ref 20-22). In this study, the 50-mm diameter original rods are produced from an ingot of 200 mm in diameter. Due to the preceding hot-working, grain size of original material is more fined than some of previous reports (Ref 17, 23). However, the grain refinement in following hot-forging processes (50-mm rods hot-forged down to 12-mm rods) is limited. Basing on our previous study (Ref 24), dynamic recrystallization might occurred in cogging and multi-direction forging with 50% reduction per pass at 950 °C, while dynamic recovery might be the major soften mechanism in following hot-forging process with 20% reduction per pass at 850 °C.

The micro-morphology of original and hot-forged materials also showed great difference. As for large material size causing uneven deformation, original materials contain some dendritic microstructure in the center of original rods (Fig. 3a). This dendritic microstructure causes the nonuniformity of elongation. With following hot-forgings, dendritic microstructure is eliminated, and eutectic structure turns into fined and dispersed which is advantageous to mechanical properties. Table 1 showed that hot-forged materials have higher strength and elongation than original materials. The homogeneous properties and higher elongation are the most important factors for cold workability. So, for NiTiNb alloy, hot-forging processes not only provide semi-manufacture in suitable material dimension, but also benefit for attaining dispersed and homogeneous microstructure for following cold processes, which is same with other NiTi alloys (Ref 25).

Table 1 Experimentally measured mechanical properties of hot-forged NiTiNb rods (without heat treatment)
Full size table

As shown in Fig. 3(c) and (d), the grain size of cold-workings (cold-drawn and cold-rolled materials) decreases to 5-10 μm, which is far more fined than hot-forged. The cold-working induces high-density dislocation walls and microbands, as shown in Fig. 7(a). With heating treatment, the dislocation walls and microbands develops to low angle grain boundaries or to trigger recrystallization (as Fig. 7(b) shows). And then, the grain boundary migrates and crystalline grain grows. After the heat treatment, cold worked materials get full static recrystallization (as Fig. 2 shows), and showed similar microstructure in grain size and micro-morphology both in cold-drawn and cold-rolled materials.

Fig. 7
figure 7

EBSD maps of NiTiNb alloy, (a) cold-rolled, the black district presenting the severe plastic deformation area and (b) cold-rolled and then annealed at 600 °C for 30 min, showing different rotation angle boundaries

Full size image

Effects of Deformation Processes on Texture

Deformation processes have an obvious influence on textures as well as the metallography. The texture of ingot induced by temperature-gradient-driven is always random and strong (Ref 17). Due to analogical deformation processes, both the original and hot-forged materials show a clear 〈113〉 fiber texture as Fig. 4(a) and (b). But the 〈113〉 texture density of hot-forged material was increased to 15 times random. It infers that 〈113〉 texture is induced and intensified by hot forging processes.

After cold-working, the texture is changed. Both cold-drawn materials possess a strong 〈111〉 preferred orientation paralleling to the axial direction. But in circumference direction (TD), only cold-rolled materials exhibit a clear preferred orientation in 〈110〉 district. This crystal orientation is dependent upon the tube deformation mode. With three-roll rolling processes, not only the axial direction but also the direction along the thickness of the tube undergoes deformation. Generally, this axisymmetrical tube deformation is between extreme cases of drawing and sheet rolling (Ref 26, 27). Accordingly, the cold-rolled tubes form 〈111〉 texture in the axial direction, which is similar to the cold-drawn materials, and while it also develops 〈110〉 crystalline orientation preferred along the circumferential direction, which exists in cold flat rolling sheets (Ref 28, 29).

Effects of Texture on Recoverability

The mathematic relationship between transformation strain and crystal orientation is calculated by crystallographic point of view called “phenomenological crystallographic theory” in single crystal. Based on the theory, the transformation strain produced by lattice distortion during martensitic transformation can be calculated by lattice constants (Ref 6, 8, 12, 28). As shown in Fig. 8, the above quantitative relationships could be well expressed in [001]-[011]-[111] standard stereographic triangle, which shows that the calculated recovery strain was dependent on orientation. The recoverable strain would increase if the crystal axis was preferentially located around the [011] or [111] district, while it is relatively small if the crystalline aligned along the [001] direction. The following studies showed that it also fit to polycrystalline NiTi thin films and sheets (Ref 6, 12, 28). As for pipe couplings, it should be analyzed in cylindrical coordinate. With inner diameter Φ increasing to Φ1, the mechanical strain ε is induced in circumference direction,

$$ \upvarepsilon = \frac{{2\uppi (\Upphi_{1} - \Upphi )}}{2\uppi \Upphi } = \frac{{\Upphi_{1} - \Upphi }}{\Upphi }. $$

During inverse martensitic transformation, circumferential strain ε recovered and inner diameter Φ1 decreases to Φ2. Thus, the recoverability of pipe couplings η depended on the recoverability of circumferential strain ε, which is affected by preferred orientation in the circumferential direction.

Fig. 8
figure 8

Orientation dependence of the calculated transformation strain associated with the martensitic transformation (Ref 7)

Full size image

As shown in Fig. 5, the cold-rolled materials possess 〈110〉 crystal preferred orientation in the circumferential direction. According to Fig. 8, this preferred orientation may causes higher recoverability for cold-rolled pipe couplings, though both cold-drawn and cold-rolled pipe couplings are made from one ingot and have similar micro-morphology. However, the pipe coupling is more complicated than thin film and sheet, the exact difference between calculated and measured still needs further study.

Conclusions

In summary, the objective of this study is to present microstructure and properties of deformation processed NiTiNb alloy. Some important conclusions are described as follows:

  1. 1.

    Under hot forging, dendritic microstructure can be eliminated and eutectic structure turns into fined and dispersed. The materials become homogeneous and elongation increases. 〈113〉 fiber texture is induced and intensified with hot forging.

  2. 2.

    After cold-workings (cold-rolled and cold-drawn), the heat treatment induces static recrystallization and effectively fines grain size. Cold-drawn materials form a strong 〈111〉 fiber texture paralleling to the axial direction. While cold-rolled materials contains a complicated orientation, which comprises 〈111〉 crystal direction paralleling to the axial direction and 〈110〉 crystal direction aligning to the circumferential direction of tubes.

  3. 3.

    The recoverability of pipe coupling is affected by crystalline orientations in the circumferential direction. With 〈110〉 crystalline orientation arranged along circumferential direction, the pipe couplings which made by cold-rolled tubes shows greater recoverability than those made by cold-drawn tubes.