Twinning and rotational deformation of nanocrystalline NiTi under shock loading

Understanding the formation of twins and new grains (NGs) in B2 austenite NiTi alloys under shock loading is of significance and importance for its science insights and engineering applications. However, the formation of $\left\{112\right\}$ twin in austenite phase under shock loading is still controversial, and the NGs' evolution under shock loading is unclear. The Electron Backscatter Diffraction characterizations and x-ray diffraction analyses of the NiTi samples recovered from the shock experiments reveal that the main deformation modes include dislocations, twins, and new grains, etc. Similar phenomena are also obtained in our nonequilibrium molecular dynamics simulations for nanocrystalline NiTi (nc-NiTi) with limited-duration-pulse shock loading. Simulations confirmed that $\left\{112\right\}$ twins in austenite phase can be formed by successively gliding a displacement of $\frac{a}{3}$[111] on ($\overline{2}11$) plane. It can occur in both shock compression and release stages, in addition, the grain boundaries triple junction and the interaction of slip bands with different slip systems can serve as the nucleation of twins. Moreover, that the nanoscale rotational deformation leads to the formation of NGs is found in our simulation of nc-NiTi, without experiencing the conventional disorder-recrystallization-grain refinement stages in the corresponding region. Related to shear stress $\tau$, the shock loading velocity $U_p$ plays the key role in the formations of twins and NGs. These early successes may hope to get some insights into the deformation mechanism of NiTi under shock loading.

Figure 1

(a) Particle velocity profile $\mathit{Up}$ for shock loading along $\langle 100\rangle ,\langle 110\rangle ,\langle 111\rangle$ and $\langle 112\rangle$, respectively, at $t=15.0$ ps. (b) The shock wave velocity-particle velocity (Us–Up) plots obtained from our NEMD simulations of single crystalline and polycrystalline NiTi experiments [21, 22, 23, 24, 25, 26]. The line represents the linear fitting to the experimental results: $\mathrm{Us}=4.351+1.588{\mathrm{U}}_{\mathrm{p}}$ [26]. Only the plastic wave velocities are plotted in Fig. 1 for clarity. (c)–(f) Shock-induced deformation in single crystalline NiTi along the $\langle 100\rangle ,\langle 110\rangle ,\langle 111\rangle$ and $\langle 112\rangle$ direction, respectively. Characterization by von Mises shear stain analysis.

Figure 2

(a) Schematic of magnetically driven shock compression experiment and sample recovery. 1: PMMA holders; 2: DLHV pins; 3: Anode panel; 4: Windows; 5: Sample; 6: Momentum trapping rings; 7: Cathode panel; 8: Copper momentum trapping. (b) Columnar nc-NiTi model characteristic by CNA methods: the blue region represents B2 austenite NiTi, the white region represents GBs or atoms in disorder.

Figure 3

(a)–(b) EBSD characterizations of as-received and experimentally recovered polycrystalline NiTi samples at shock loading velocity $U_p=0.927$ km ${\mathrm{s}}^{-1}$ at room temperature. (c)–(d) The corresponding amplified configurations for the marked rectangular areas in (b), which represent twins and new grains, respectively. (e)–(f) The histograms of frequency and distribution of the boundaries for (a) and (b), respectively, the red circle corresponding to $\left\{112\right\}$ austenite twins, $70.5^{\circ }$ misorientation angle, and $5^{\circ }$ for the maximum deviations. (g) The XRD analysis of as-received and experimentally recovered polycrystalline NiTi samples at shock loading velocity $U_p=0.927$ km ${\mathrm{s}}^{-1}$ at room temperature.

Figure 4

The (a) 1D pressure profiles (${\sigma }_{xx}$) in nc-NiTi under different shock-loading velocities $U_p$. Comparisons of simulated results corresponding to (a) at initial ambient temperature 300 K and different $U_p$: (b) 0.6 km ${\mathrm{s}}^{-1}$; (c) 0.8 km ${\mathrm{s}}^{-1}$; (d) 1.0 km ${\mathrm{s}}^{-1}$ (different states are distinguished by red long dashes). The microstructures are characterized by CNA methods: twin T1, T2, T3, and new grain (NG). The shock direction is labeled by black arrows. The simulated XRD patterns (e) of the nc-NiTi models for initial austenite [Fig. 2], Fig. 4, 4 and 4, respectively.

Figure 5

Nucleation, propagation and growth of $\left\{112\right\}$ twin T1 in austenite phase at shock loading velocities $U_p=0.6$ km ${\mathrm{s}}^{-1}$ in nc-NiTi, which is formed by successively gliding a displacement of $\frac{a}{3}$[111] on ($\overline{2}11$) plane. The atom color coding is based on the MRD magnitude of $\mathit{s}$, the cyan (corresponding to “4”) represent dislocation $\frac{a}{3}$[111]. The arrows denote slip vectors and it is also colored by MRD magnitude of $\mathit{s}$.

Figure 6

Growth of $\left\{112\right\}$ twin T1 in austenite at different shock loading velocities $U_p$ in nc-NiTi. The first four pictures corresponds to amplified configurations for the marked rectangular areas in the last two. The atom color coding is based on the MRD magnitude of $\mathit{s}$, the cyan and red (one each for 4 and 6) correspond to dislocations $\frac{a}{3}$[111] and $\frac{a}{2}$[111], respectively.

Figure 7

Formation of amorphous shear band at grain boundaries (GBs) triple junction at shock loading velocity $U_p=1.0$ km ${\mathrm{s}}^{-1}$, characterization by CNA analysis and distribution of von Mises shear stress ${\tau }^{\mathrm{Mises}}$ in the same region. We also calculate radical distribution function [$g(r$)] of the B2 structure sc-NiTi and amorphous shear band in nc-NiTi.

Figure 8

Details of twin T2 at shock loading velocity $U_p=0.8$ km ${\mathrm{s}}^{-1}$. The atom color coding is based on the MRD magnitude of $\mathit{s}$.

Figure 9

Details of twin T3 at shock loading velocity $U_p=0.8$ km ${\mathrm{s}}^{-1}$. The atom color coding is based on the MRD magnitude of $\mathit{s}$.

Figure 10

(a)–(c) The microstructural evolution of new grain NG for shock loading velocity $U_p=0.8$ km ${\mathrm{s}}^{-1}$ at 15.0 ps, 21.5 ps and 46.0 ps, respectively. The atom color coding is based on the MRD magnitude of $\mathit{s}$. (d)–(f) The corresponding region based on OM analysis along the $x$ axis. (g)–(h) Projection of B2 crystal position on ($0\overline{1}1$) plane before (g) and after (h) deformation.