Effects of strain and electropulse duration on elastic modulus of TiNi alloy

Effect of rolling strain on microstructural evolution

The microstructural features of the as-quenched and electroplastic rolled TiNi alloys are exhibited in Figure 1. In the as-quenched sample, large body-centered cubic austensite B2 grains with a size of ∼ 150 µm appear in metallographic images (not shown here). And a small amount of monoclinic martensite B19^′ coarse lamellas with a width of 150–750 nm precipitated in the B2 matrix (see Figure 1(a)). The phase structure of coarse B19^′ lamellas was confirmed by selected area electron diffraction (SAED) pattern made from circled region (see the inset in Figure 1(a)). After EPR deformation with a strain of 0.32, the coarse B19^′ lamellas are elongated along the rolling direction and its width decreased in the region of 70–250 nm (Figure 1(b)). However, increasing the strain to ε ∼0.92, the B19^′ lamellas in the matrix disappeared and high-density dislocations appeared as indicated by the red triangles in Figure 1(c). Further increasing the reduction to ε ∼1.70, there are a great many deformation-induced B19^′ lamellas with a width of 30–70 nm (see yellow arrows) and high-density dislocations appear in the B2 matrix. The phase structures ­­of martensitic B19^′ lamellas and austenitic B2 phase were confirmed by SAED pattern (see the inset) made from the circled region in Figure 1(d), in which the diffraction spots and arcs were indexed.

OPEN IN VIEWER

Figure 1

Bright-field TEM images of as-quenched sample (a) and the alloy after EPR deformation with strains of 0.32 (b), 0.92 (c) and 1.70 (d). The SAED patterns corresponding to circled regions in (a) and (d) are presented in the inset, respectively, and the diffraction spots are indexed. The dislocations are indicated by the red triangles.

To further characterise the microstructural evolution of TiNi alloy with rolling strain, XRD measurement was conducted (Figure 2(a–d)). For the as-quenched sample, strong diffraction peaks of B2 phase, e.g. (110)_B2, and weak peaks of B19^′ phase, e.g. (101) _B19′, observed in the XRD spectrum (Figure 2(a)), implying a high fraction of B2 phase and a tiny fraction (about ∼4.1% determined from the quantitative phase analysis) of B19^′ phase in the sample. For the EPR samples, the intensity of (101)_B19′ peak gradually enhanced with the rolling strain increasing from 0.32 to 1.70, indicating the deformation-induced martensitic transformation from B2 to B19^′ phase during EPR deformation (Figure 2(b-d)). The volume fraction of B19^′ martensitic phase increases with the rolling strain increasing, as shown in Figure 3. And the B19^′ martensite volume fraction of the EPR sample at ε = 1.70 has been calculated to be ∼16.2% by using the Rietveld refinement method [14,15,22]. Moreover, the B2 phase diffraction peaks of EPR samples broaden obviously with the increase of strain as compared with the as-quenched sample, demonstrating that high-density dislocations were introduced during the rolling deformation as can be seen by the TEM images in Figure 1(b–d). In addition, as the strain increases from 0.92 to 1.70 (see Figure 2(c,d)), the (110)_B2 diffraction peak was slightly narrowed, which might be resulted from the reduction of the introduced dislocation caused by EPR deformation [23,24].

OPEN IN VIEWER

Figure 2

XRD spectra of as-quenched (a) and EPR deformation samples with strains of ε = 0.32 (b), 0.92 (c), 1.70 (d) and electropulse durations of τ = 80 µs (e), 120 µs (f).

OPEN IN VIEWER

Figure 3

Volume fractions of the B19^′ martensitic phase of the TiNi alloy under different conditions.

Effect of electropulse duration on microstructural evolution

Figure 4 exhibits the TEM images of EPR samples with a strain of ∼1.70 at electropulse durations of 80 and 120 µs. At electropulse duration of 80 µs (Figure 4(a)), the microstructure of the sample is characterised by some dislocation tangles (indicated by red arrows) and the SAED pattern exhibits slightly elongated spots and discontinuous arcs (Figure 4(b)), indicating high deformed defects stored in the sample. The phase composition of the EPR sample at the electropulse duration of 100 µs was displayed in Figure 2(d) and described in detail. With the electropulse duration further increased to 120 µs (Figure 4(c)), few nanoscale lamellas and high-density dislocations exist in the sample. The SAED patterns (Figure 4(b,d)) corresponding to circled region in Figure 4(a,c) were indexed as the B2 and B19^′ phases, and the elongated diffraction spots imply the existence of high-density dislocations in the EPR samples. The phase evolution of TiNi alloy with electropulse duration was also confirmed by XRD analyses (Figure 2(d–f)). For all the three samples, the austensite B2 phase is dominant along with a tiny amount of martensite B19^′ phase (spectra (d–f)). Specially, with increasing the electropulse duration from 80 to 120 µs, the (110)_B2 diffraction peaks narrowed gradually, whereas the intensity of (101)_B19′ diffraction peak first increases (Figure 2(e,d)) and then decreases (Figure 2(f)). The volume fraction of B19^′ martensitic phase increases first and then slightly decrease with the electropulse durations increasing from 80 to 120 µs, and the corresponding phase fraction was determined to be ∼6.5%, 16.2% and 10.1% for the EPR samples, respectively, by using Rietveld refinement procedure (see Figure 3). Above results demonstrate that the thermal (Joule heating) and athermal (electron wind) effects induced by high-energy pulse current might promote the dynamic recovery of dislocations, slightly grain growth and reverse martensite transformation from B19^′ to B2 phase with the electropulse duration increasing [23-27].

OPEN IN VIEWER

Figure 4

Bright-field TEM images of EPR samples with electropulse durations of 80 µs (a), and 120 µs (c). (b) and (d) are the SAED patterns corresponding to circled regions in (a) and (c), and the diffraction spots are indexed.

Mechanical properties

Figure 5 presents the tensile engineering stress–strain curves of the TiNi alloy at ambient temperature for various conditions. The as-quenched sample (curve A) shows a typical double yielding or stress plateau on account of the martensitic transformation during the tensile deformation [28,29], while the electroplastic rolled samples (curves B–F) exhibit continuous yielding and non-linear elastic behaviour (Figure 5(a)). The elastic modulus in this study was determined by calculating the slope of the tangent at zero strain on the tensile curves [30]. The as-quenched sample shows an incipient elastic modulus (E) of ∼61 GPa (curve A), which is much higher than that of EPR samples (see Figure 5(b,c)). After EPR with a strain of 0.32, the incipient E decreases to 48 GPa (Figure 5(b) curve B). And the incipient E decreases continuously from 42 to 30 GPa with the rolling strain further increasing from 0.92 to 1.70 (Figure 5(b) curves C and D). Researches have been reported that the monoclinic martensitic B19^′ phase presents a very low elastic modulus (∼20–50 GPa), which is only half that of cubic austenitic B2 phase [18,31,32]. With the increase of rolling strain, the volume fraction of B19^′ phase gradually increased (see Figure 2 spectra (b–d) and Figure 3), thus resulted in the decrease in elastic modulus of EPR TiNi alloy. Moreover, the deformation-induced martensitic transformation, on account of internal stress and high-density defects (Figure 1(b–d)) introduced by EPR deformation, is also favourable to the decrease in elastic modulus [29,30].

OPEN IN VIEWER

Figure 5

Uniaxial tensile engineering stress–strain curves of TiNi alloy for all the experimental conditions (a). (b) is an enlarged view of as-quenched sample (curve A) and the alloy after EPR deformation with strains of 0.32 (curve B), 0.92 (curve C) and 1.70 (curve D) marked with light blue shadow in (a); (c) is an enlarged view of the EPR samples with electropulse durations of 80 µs (curve E), 100 µs (curve D) and 120 µs (curve F) marked with light blue shadow in (a).

However, the changing trends are quite different with the electropulse duration. The incipient E first decreased from 53 to 30 GPa as the electropulse duration increasing from 80 to 100 µs, and then slightly increased to 40 GPa with the pulse duration further increased to 120 µs (see curves D–F in Figure 5(c)). The decrease in incipient E can mainly be attributed to the volume fraction increase of martensitic B19^′ phase induced by high-density pulse current at pulse duration of 100 µs as presented in Figure 2(d,e) and Figure 3. Moreover, the abundant phase boundaries may also contribute to the low incipient E due to the loose structure caused by the adjacent atoms deviating from its equilibrium state at phase boundaries [7,9,10]. Nevertheless, with the pulse duration further increased to 120 µs, the slight increase in incipient E is owing to the volume fraction decrease of martensitic B19^′ phase and dynamic recovery of dislocations induced by high-density pulse current.