Enhanced tensile plasticity of a CuZr-based bulk metallic glass composite induced by ion irradiation

1. Introduction

The last decades have witnessed the discovery of metallic glasses (MGs) being quenched from the molten state at an ultra-high cooling rate of >10⁶ K/s. Owing to the disordered atomic configuration, MGs often demonstrate excellent properties like superior strength, large elastic limit, and high corrosion resistance, making them attractive for practical applications [[1], [2], [3], [4], [5], [6]]. More importantly, the inherent liquid-like structure of MGs has an enhanced resistance against displacive irradiation-induced damage [7]. Therefore, MGs are expected to be promising candidates for industrial applications in irradiation conditions, such as in nuclear reactors and aerospace, and have therefore attracted extensive attention of researchers. Up to now, numerous attempts have been devoted to studying the effect of irradiation on the structure, mechanical properties, physical and chemical properties of MGs [[8], [9], [10], [11]]. For instance, ion irradiation can induce the formation of oriented nanocrystals in a Ni-P glassy alloy [12], and cause the precipitation of nano-crystals in CuZrTi and ZrCuAlNi glasses [13,14]. Simultaneously, the surface roughness [8,15], the free volume content [11], and the mechanical properties such as plasticity and hardness [9,16] can be also finely tuned by irradiation.

The structural reliability of MGs is significantly limited by catastrophic failure [[17], [18], [19]]. Great efforts have been made recently to overcome their intrinsic brittleness [[20], [21], [22], [23]]. For instance, secondary ductile phases have been introduced into the glassy matrix to hinder the rapid propagation of shear bands [[24], [25], [26], [27], [28]]. Among the developed composites, bulk metallic glass composites (BMGCs) reinforced by B2 CuZr phase show an excellent combination of high strength and high plasticity [29]. Because of the martensitic transformation (MT) from B2 CuZr to B19’ and B33 CuZr, a “work-hardening” behavior occurs upon plastic deformation [30]. Therefore, CuZr-based BMGCs bear great potentials for irradiation protection due to their inherent irradiation resistance and improved mechanical performance. Herein a question arises: can the irradiation affect the microstructure of the BMGCs and, thus, affect their mechanical performance or not? In the present work, a Cu₄₈Zr_47.2Al₄Nb_0.8 (at.%) BMGC with excellent mechanical properties is chosen as a model material [31]. A series of tests and simulations are employed to study the BMGC samples irradiated by different ion fluences, aiming at unveiling the structural and mechanical modifications induced by ion irradiation. It is expected that the results obtained here are helpful to extend the industrial application of BMGCs under extreme conditions.

2. Materials and methods

CuZrAlNb alloy ingots were prepared in a Ti-gettered argon atmosphere by arc melting a mixture of Cu, Zr, Al and Nb with purities of higher than 99.9 wt%. To guarantee chemical homogeneity, the ingots were remelted at least four times, followed by suction casting the alloy melt into a copper mold to form cylindrical samples of 3 mm in diameter and $\widetilde{3}$5 mm in length. The structure of the as-cast alloy sample was characterized by X-ray diffraction (XRD, Panalytical Empyrean) with CuKα radiation and by optical microscopy (OM, Leica DM4000). Three types of samples were used in the subsequent experiments, which were cut from the middle part of the as-cast cylinders and their uniform structure was verified by OM. In order to reveal the structural changes in the irradiated surface, the type I samples were cut and polished to a mirror finish with a dimension of ϕ3 mm × $\widetilde{6}$0 μm. The polished type II samples with a dimension of ϕ3 mm × 3 mm were used to characterize the internal structure after irradiation. The room-temperature tensile tests were performed on the type III samples, which were cut and polished to a dog-bone-shaped with 1 mm × 1 mm cross-sectional area, and 4 mm gauge length. All the flat surfaces of the samples were irradiated in a high vacuum chamber of 6 × 10^-4 Pa by N⁺ ions with fluences of 0, 1.7 × 10¹⁵, 5 × 10¹⁵, 1 × 10¹⁶ or 2 × 10¹⁶ cm^-2, respectively. The samples were fixed and placed perpendicularly to the N⁺ ion beam with an energy of 120 keV and a flux of 1.2 × 10¹² cm^-2 s^-1. To clarify the ion irradiation effect on the structure, neutron diffraction experiments were carried out on type I samples using ³He-PSD diffractometer at STRESS-SPEC of Research Neutron Source Heinz Maier-Leibnitz (FRM II), Germany [32]. Using Ge (511) monochromators, the beam with a wavelength of 1.68 Å was acquired. Then, the incident neutron beam was set to 3 mm in height and 3 mm in width to ensure the most optimized sampling. The collected diffraction data was visualized in STeCa2 software. To evaluate the stopping depth of the ions in the samples, the software package of stopping and range of ions in matter (SRIM) was used [33]. The microstructure of the samples was examined in a transmission electron microscope (TEM, Talos F200X). The irradiated surfaces of the type I samples were protected by an insulator during low-temperature twin-jet thinning prior to the TEM experiments. The central part of the type II samples was kept and thinned to meet the requirements for TEM observation. Room-temperature tensile tests on the type III samples were carried out with an initial strain rate of 5 × 10^-4 s^-1 in an Instron 5569 system. The shear band morphology after fracture was examined by scanning electron microscope (SEM, Quanta 200FEG).

Using the embedded-atom method (EAM) potential function for the interatomic interactions [34], LAMMPS, a molecular dynamics (MD) simulation software [35], was employed to model the structural changes of the CuZr-based BMGC in this work. The original sample containing 225,000 atoms with a 24.4 nm × 3.2 nm × 48.8 nm dimension was built applying a three-dimensional periodic boundary condition. The sample was firstly melted at 2000 K under the isothermal-isobaric ensemble (NPT ensemble). After equilibrating the system for 1 ns, the melt was quenched to 300 K with a cooling rate of 2 × 10¹² K/s. Spherical B2 CuZr crystals with an average radius of $\widetilde{3}$0 nm were subsequently introduced into the glassy matrix by replacing 50% of the atoms. After removing the atoms from the overlap regions, the as-cast sample was annealed at 500 K for 1 ns and then cooled to 300 K to stabilize the structure. Primary knock-on atoms (PKA) were randomly placed on the left surface of the sample with the dimensions of 3.2 × 48.8 nm². Owing to the uncertainty of collision, the kinetic energy transferred from the ions, i.e., the velocity of PKAs was chosen to be random from 0 to the maximum calculated using the laws of conservation of energy and momentum. Then, 300 PKAs entered the sample with an interval of 50 ps (time for equilibrating the structure) under the NPT ensemble. Finally, uniaxial tensile tests were carried out on both the as-cast sample and the irradiated sample along the Z-direction with a strain rate of 10^-4 ps^-1. The atomic strain was calculated by

$η_{i}^{Mises}$ = $\sqrt{\eta_{xy}^{2}+\eta_{yz}^{2}+\eta_{xz}^{2}+\frac{(\eta_{xx}-\eta_{yy})^{2}+(\eta_{yy}-\eta_{zz})^{2}+(\eta_{xx}-\eta_{zz})^{2}}{6}}$ (1)

where η_ab (a, b = x, y or z) represents the components of the strain tensor of atom i. The results were visualized by the Open Visualization Tool (OVITO) [36].

3. Results and discussion

Fig. 1(a) shows the typical OM image from type II sample of the as-cast CuZr-based BMGC with spherical precipitates embedded in a featureless glassy matrix. The volume fraction of the crystalline phase in all three types of samples was determined to be 50 ± 5 vol.% using ImageJ software, based on the average area ratio of the phases in the OM images. Fig. 1(b) presents the neutron diffraction patterns of the CuZr-based BMGC samples before and after ion irradiation. For the as-cast sample, two obvious Bragg peaks located at 2.298 Å and 1.621 Å corresponding to the lattice planes of (110) and (200) of the B2 CuZr phase, respectively, and one weak Bragg peak located at 2.069 Å corresponding to the (020) lattice plane of the B19’ phase superimposed on the amorphous hump are seen in the diffraction pattern. As the applied ion fluence increases, the intensities of the B2 CuZr peaks gradually decrease whereas the intensity of B19’ CuZr peak gradually increases. The intensity of each peak in the neutron diffraction patterns is gained using a Gauss fitting, and the corresponding intensity ratios of the amorphous hump and B19’ (020) peak to B2 (110) peak are summarized and plotted in the inset of Fig. 1(b). It should be noticed that the orientation of the B2 CuZr and transformed B19’ CuZr crystals are randomly distributed in the crystals, thus the intensities of the peaks should allow for a first estimation of the relative volume fraction. As can be seen, the increase in the applied ion fluence results in the increase in the intensity ratios of both the amorphous peak and the B19’ (020) peak to the B2 (110) peak, thereby revealing the enhanced content of the amorphous phase and the B19’ CuZr phase.

View original graphic| Download| PPT slide

Fig. 1.   (a) Typically optical microscopy image of the CuZr-based BMGC sample showing a composite structure consisting of spherical crystalline phases embedded in featureless matrix and (b) neutron diffraction patterns of the as-cast and the ion irradiated BMGC samples, with inset highlighting the intensity ratio of the amorphous phase peak and the B19’ (020) peak to B2 (110) peak.

The N⁺ ions are accelerated by a high voltage during the ion irradiation process. Once these ions knock on the near-surface atoms, the kinetic energy of high-velocity ions is transferred into the sample [11,13]. The succeeding cascade collisions cause a sudden increase in the atomic mobility, and thus a dramatic temperature rise of the near-surface volume. Subsequently, rapid quenching of this volume will be carried out by transferring energy to the atoms near the volume. The rapid heating-quenching may induce numerous defects at the surface [11], and will cause a temperature rise inside the sample. After the cascade collisions exhaust the kinetic energy, the ions will be stopped and implanted in the material. Then the depth-dependent ion distribution profile can display the range directly affected by cascade collisions, which is simulated using 10⁶ ions in TRIM, a component of SRIM, and shown in Fig. 2(a). The result reveals that the structural change caused by the cascade collisions will be concentrated near the surface in a range of $\widetilde{3}$50 nm. Fig. 2(b) is the high-resolution TEM (HRTEM) image, illustrating the effect of the cascade collisions on the surface of the crystals. The elliptic region marked by the dashed line shows a disordered feature, which is different from the surrounding uniform ordered structure. It means that the cascade collisions can cause partial amorphization of the crystals. But how does the structure change in the glassy region at the surface? The auto-correlation functions (ACF) are performed on all the captured HRTEM images of the surface amorphous phase to examine the periodicity and quantitatively determine the volume fraction of the ordered region. The details of the calculation were described in our previous paper [37]. Fig. 2(c) and (d) are typical ACF results of the as-cast sample and the sample with a fluence of 2 × 10¹⁶ cm^-2, respectively. The decrease of the ordered region marked by red squares in the glass at the surface can be clearly observed after ion irradiation. Less ordered regions mean higher Poisson’s ratio of the amorphous phase [37], which can be considered as the indicator reflecting the mechanical behavior of the amorphous phase [38]. It has been argued that more non-uniform stress fields can be achieved in glasses with a higher Poisson’s ratio, resulting in easier nucleation of shear bands [39]. The above results indicate that the irradiation leads to a more disordered structure in both the crystal and the glassy matrix, which causes partial amorphization and structural rejuvenation, respectively.

View original graphic| Download| PPT slide

Fig. 2.   (a) Depth of ions entering into the sample simulated by SRIM software, (b) FFT-filtered HRTEM image of irradiated sample surface revealing amorphization in the crystal induced by ion irradiation, and the typical results of the ACF calculation of (c) the as-cast samples and (d) the irradiated samples with a fluence of 2 × 10¹⁶ cm^-2.

Fig. 3 shows the TEM images and the corresponding select area electron diffraction (SAED) patterns of the samples before and after ion irradiation. It can be seen from Fig. 3(a) that two distinct phases exist in the as-cast sample. The SAED pattern shown in Fig. 3(a) inset (bottom right) corresponding to the featureless phase illustrates a typical feature of the amorphous phase, whereas the other phase shows a feature of the cubic B2 lattice in its SAED pattern shown in Fig. 3(a) inset (top right). After ion irradiation, numerous crystals in the form of laths can be found in the sample as shown in Fig. 3(b), and the corresponding inset SAED pattern reveals the appearance of the B19’ structure while other crystals remain their B2 structure. These results further verify that the MT from B2 to B19’ can be induced by ion irradiation, as shown in Fig. 1(b). Thermodynamically, the B19’ CuZr phase is more stable than the B2 CuZr phase at ambient temperature [40]. As mentioned above, the energy suddenly transferred from the ions results in a temperature rise in the sample. Such temperature rise will provide enough energy for the metastable B2 CuZr phase to overcome the energy barrier of the phase transformation. Therefore, with the same flux, the content of the martensite will increase with increasing the applied ion fluence, resulting in the gradually increasing content of B19’ CuZr phase.

View original graphic| Download| PPT slide

Fig. 3.   TEM images of (a) the interface between the glassy matrix and the crystal of the as-cast sample and (b) the crystals in the irradiated sample indicating the B2 and B19′ CuZr phase.

Next, tensile tests are conducted on all the as-cast and the ion-irradiated BMGC samples. The typical tensile stress-strain curves are displayed in Fig. 4(a). As the ion fluence increases from 0 to 2.0 × 10¹⁶ cm^-2, the tensile plastic strain of the composite dramatically increases from 8.0%±0.3% to 25.5%±0.4%, without sacrificing its fracture strength. The inset of Fig. 4(a) illustrates the work-hardening rate of the samples, which was calculated by:

Θ=dσ_true/dε_true (2)

where Θ is the work-hardening rate, σ_true and ε_true are the true stress and the true strain, respectively. The work-hardening rates of the irradiated BMGC samples vary with the applied ion fluences. At the initial stage of the plastic deformation, the work-hardening rate dramatically decreases from a larger value with increasing the strain. However, the initial work-hardening rate and its deceleration decrease as the applied ion fluence increases. It is generally accepted that the MT will cause an increase in the work-hardening rate whereas the shear banding decreases the work-hardening rate because of the work-softening behavior. The larger initial work-hardening rate implies that more B2 CuZr phase has transformed to B19’ CuZr phase at the beginning deformation stage. That means, the ion irradiation decreases the volume fraction of B2 CuZr phases occurring MT at the beginning deformation stage and then may homogenize the MT in the ongoing deformation. The shear localization facilitates the work-softening in the glassy matrix. However, the low deceleration of the work-hardening rate in the irradiated sample may indicate that the plastic deformation is homogeneously distributed in the amorphous phase. Subsequently, the work-hardening rate becomes stable and approaches a constant value, which should be a result of the balancing of work-hardening resulted from MT and work-softening resulted from shear banding. This suggests that ion irradiation can be considered as a powerful method to finely tune the plastic deformability of the materials. SEM observations were performed after tensile experiments. Fig. 4(b)-(d) show the side-view appearances of the fractured samples. Compared with the morphology of the as-cast sample shown in Fig. 4(b), the ion irradiated sample exhibits more branching and intersecting shear bands shown in Fig. 4(c) and (d). Furthermore, as the applied ion irradiation increases, shear bands distribute more homogeneously at the fractured surface. These results indicate that the enhanced plasticity should be closely related to the homogeneous distribution of plastic deformation and the interaction between shear bands.

View original graphic| Download| PPT slide

Fig. 4.   (a) Typical tensile stress-strain curves of the as-cast sample and the ion irradiated CuZr-based BMGC samples with different fluences, with inset showing the work-hardening rate of the samples. The surface morphologies of (b) the as-cast sample, and the ion irradiated sample with fluences of (c) 5 × 10¹⁵ cm^-2, and (d) 2 × 10¹⁶ cm^-2.

MD simulations were employed to illustrate the structural changes of the studied BMGC during ion irradiation and during tensile loading, and the results are displayed in Fig. 5(a)-(d). Compared with the as-cast sample (Fig. 5(a)), the crystals near the surface of the irradiated sample (on the left of the dashed line) are amorphized (Fig. 5(c)), further confirming the neutron diffraction and TEM results shown in Fig. 1(b) and Fig. 2(b). After applying 4% tensile strain, some regions demonstrating high strain (warm color regions marked by arrows) can be seen near the surface in the as-cast sample (Fig. 5(b)), which means the initiation of shear banding. For the irradiated sample (Fig. 5(d)), more shear banding events can be found near the surface. Considering the structural change at the surface revealed by the ACF results, it can be confirmed that ion irradiation enhances structural disorder at the surface, i.e., shear banding is easy to occur in the irradiated volume. However, as seen in Fig. 5(d), the crystals marked by the dashed circles hinder the rapid propagation of the shear bands previously initiated at the surface. On the other hand, the interface (the dashed line) between the as-cast glassy matrix and the rejuvenated surface could block shear band propagation because of the structural difference [41]. Therefore, the propagation of shear bands will be delayed near the surface of the irradiated sample [42]. If sufficient stress applied, these shear bands will propagate into the matrix and homogenize plastic deformation in the glassy matrix.

View original graphic| Download| PPT slide

Fig. 5.   Simulated structural configuration of (a) the as-cast CuZr-based BMGC sample, (b) the corresponding atomic strain field with 4% tensile strain, (c) the ion irradiated sample and (d) the corresponding atomic strain field with 4% tensile strain.

On the other hand, the MT from B2 to B19’ CuZr in the crystals during the ion irradiation exerts a crucial role on the mechanical behavior of the studied CuZr-based BMGC. As illustrated in Fig. 3(b), numerous interfaces between the B2 CuZr phase and the B19’ CuZr phase are formed because of the MT. These interfaces will act as the heterogeneous nucleation sites for the subsequent MT upon tensile loading. Thus, the stress-induced MT in B2 CuZr phase is susceptible to occur on those sites, i.e., the yield stress of the crystal and the initial work-hardening rate will decrease [43]. Because the first yielding of the CuZr-based BMGC is caused by the yielding of the B2 CuZr phase [25], the yield stress of the irradiated sample gradually decreases with increasing the fluence and, hence, the content of the B19’ CuZr phase. Furthermore, the MT in the crystals will result in volume dilation [44], and thus induce a stress concentration on the interface between the crystal and the amorphous matrix. Such stress concentration can facilitate heterogeneous nucleation of shear bands in the further deformation [19]. Compared with the as-cast sample, the irradiated sample will contain more nucleation sites for shear bands under identical stress. Considering the easier initiation of shear bands at the surface and inside the sample, more shear bands will be formed in the glassy matrix during the tensile tests when the applied ion fluence increases. In the ongoing deformation process, the relatively higher density of the shear bands in the irradiated sample directly manifests itself more pronounced plasticity through branching, multiplication and interaction of shear bands [45].

4. Conclusion

In summary, we have studied the effect of ion irradiation on the structure and plasticity of a CuZr-based BMGC. Ion irradiation induces numerous defects at the surface and favors the martensitic transformation from B2 CuZr phase to B19’ CuZr phase in the crystals inside the sample. The mechanical tests indicate that the tensile plastic strain of the studied BMGC increases dramatically from 8.0%±0.3% to 25.5%±0.4% as the applied ion fluence increases. The surface disordered structure and the martensitic transformation caused by ion irradiation and the resulting higher density of shear bands explain the enhanced plasticity. The obtained results suggest that ion irradiation, in general, can optimize the structure and thus the mechanical behaviors of BMGCs.

Acknowledgments

This work was financially supported by the National Natural Science Foundation of China (Nos. 51871076, 51671070, 51827801, 51671067, and 51671071), the Opening Funding of the State Key Laboratory of Advanced Welding and Joining, Harbin Institute of Technology, China (No. AWJ-Z16-02), the Chinese Scholarship Council (CSC) and the German Science Foundation (DFG) (Nos. PA 2275/2-1, PA 2275/4-1, and PA 2275/6-1).